COC :/7&/v. I
*X
^Lr
^e^
*fe
ri
«fl
^
^
Col lected
Reprints
1976
Atlantic Oceanographic and Meteorological Laboratories
Volume I
April 1977
U.S. DEPARTMENT OF COMMERCE
National Oceanic and Atmospheric
Administration
Q.
O
0
Collected
Reprints 1976
Atlantic Oceanographic and Meteorological Laboratories
Miami, Florida 33149
Volume I
April 1977
Boulder, Colorado
U.S. DEPARTMENT OF COMMERCE
Juanita M. Kreps, Secretary
g National Oceanic and Atmospheric Administration
v Robert M. White, Administrator
Environmental Research Laboratories
% Wilmot Hess, Director
FORWARD
This is the tenth consecutive year in which the
Collected Reprints of NOAA's Atlantic Oceanographic
and Meteorological Laboratories have been published
for distribution to scientists, institutions, and
libraries here and abroad. This series provides a
single reference source for articles by AOML personnel
which have appeared in numerous scientific journals
and various internal scientific and technical publi-
cations .
The Atlantic Oceanographic and Meteorological
Laboratories conduct research programs in the areas
of physical, chemical, and geological oceanography
and air-sea interaction. The 1976 edition presents
the papers published in that year. They are arranged
in alphabetical order by first author within each of
five groups:
Office of the Director
Physical Oceanography Laboratory
Marine Geology and Geophysics Laboratory
Sea Air Interaction Laboratory
Ocean Chemistry Laboratory
It is hoped that those recipients with whom we do not
already have an exchange arrangement would add the
AOML Library to the distribution list for any relevant
publications from their institution.
Harris B. Stewart, Jr. Atlantic Oceanographic and
Director, AOML Meteorological Laboratories
NOAA/Environmental Research
Laboratories
15 Rickenbacker Causeway
Virginia Key
Miami, Florida 33149
i n
Digitized by the Internet Archive
in 2012 with funding from
LYRASIS Members and Sloan Foundation
http://archive.org/details/collectedreprin1976v1atla
CONTENTS
VOLUME I
General
1. Apel, J. P.
Page
Ocean Science from Space. EOS, Vol. 57,
No. 9, 612-624. 1
2. Sawyer, C. B.
High-Speed Streams and Sector Boundaries. Journal of
Geophysical Research, Vol. 81, No. 13, 2437-2441. 14
3. Sawyer, C. B. and M. Haurwitz.
Geomagnetic Activity at the Passage of High-Speed
Stream in the Solar Wind. Journal of Geophysical
Research, Vol. 81, No. 13, °435-2436. 19
4. Stewart, H. B. , Jr.
Introduction. Proc. of CICAR-I I Symposium: Progress
in Marine Research in the Caribbean and Adjacent
Regions, Caracas, Venezuela, July 12-16, 1976, p. 126. 21
5. Stewart, H. B. , Jr.
Introduction to the CICAR-I I Symposium. Proc. of
CICAR-I I Symposium: Progress in Marine Research in
the Caribbean and Adjacent Regions, Caracas, Venezuela,
July 12-16, 1976, p. 241. 22
6. Stewart, H. B. , Jr.
Preliminary Bibliography of Published Results of Marine
Research by U.S. Scientists in the CICAR Area, 1968-
1975: Introduction. U.S. Department of Commerce,
NOAA/ERL/AOML-National Oceanographic Data Center
Publication, Washington, D.C., 50 p.* 23
7. Stewart, H. B. , Jr.
Where the Sea and Man Meet: The Coastal Zone.
Museum, Vol. 7, No. 11, 19-25, 44-48. 24
* Introduction only.
PHYSICAL OCEANOGRAPHY LABORATORY
Page
8. Beardsley, R. C, W. C. Boicourt, and D. V. Hansen.
Physical Oceanography of the Middle Atlantic Bight.
Middle Atlantic Continental Shelf and the New York
Bight. ASLO Special Symposia, Volume 2, 20-34. 34
9. Charnell, R. L., M. E. Darnell, G. A. Berberian,
B. L. Kolitz, and J. B. Hazel worth.
New York Bight Project, Water Column Characterization
Cruises 1 and 2 of the NOAA Ship Researcher, 4-15 March,
5-14 May 1974. NOAA Data Report ERL MESA-18 , 220 p.* 49
10. Festa, J. F. , and D. V. Hansen.
A Two-Dimensional Numerical Model of Estuarine Cir-
culation: The Effects of Altering Depth and River
Discharge. Estuarine and Coastal Marine Science ,
Vol. 4, 309-323. 50
11. Gordon, H. R.
Radiative Transfer: A Technique for Simulating the
Ocean in Satellite Remote Sensing Calculations.
Applied Optics , Vol. 15, No. 8, 1974-1979. 65
12. Hansen, D. V.
A Lagrangian Buoy Experiment in the Sargasso Sea.
Proc. AIAA Drift Symposium, Hampton, Va., May 22-23,
1974, NASA CP-2003, 175-192. 71
13. Herman, A.
Automated Contouring of Vertical Oceanographic
Sections Using an Objective Analysis. Proc. of
the Third Annual Conference on Computer Graphics,
Interactive Techniques, and Image Processing,
University of Pennsylvania, Computer Graphics 10,
No. 2, 218-223. 89
* Abstract only;
Complete text available on microfiche.
VI
Page
14. Herman, A. and A. C. Campbell.
An Automated Solution for Omega Navigation. Proc.
of the Fourteenth Annual Southeast Regional ACM
Conference, University of Alabama, Birmingham,
Alabama, 305-308. 95
15. Leetmaa, A.
Some Simple Mechanism for Steady Shelf Circulation.
Marine Sediment Transport and Environmental Management ,
D. J. Stanley and D. J. P. Swift, editors, John Wiley
and Son, Inc., Chapter 3, 23-28. 99
16. Leetmaa, A. and M. Cestari.
A Comparison of Satellite-Observed Sea Surface
Temperatures with Ground Truth in the Indian Ocean.
NOAA Technical Report ERL Z76-AOML 22, 10 p. 105
17. Maul, G. A.
The Study of Ocean Circulation from Space. Proc.
of the Thirteenth Space Congress: Technology for
the New Horizon, 3-27—3-36. 118
18. Maul, G. A., H. R. Gordon, S. R. Baig, M. McCaslin,
and R. DeVivo.
An Experiment to Evaluate SKYLAB Earth Resources
Sensors for Detection of the Gulf Stream. NOAA
Technical Report ERL 378-AOML 23, 69 p. 128
19. Mofjeld, H. 0.
Tidal Currents. Marine Sediment Transport and
Environmental Management , D. J. Stanley and
D. J. P. Swift, editors, John Wiley and Sons, Inc.,
Chapter 5, 53-64. 201
20. Molinari , R. L.
The Formation of the Yucatan Current Based on
Observations of Summer 1971. Journal of Physical
Oceanography, Vol. 6, No. 4, 596-602. 213
vn
21. Molinari, R. and A. D. Kirwan.
Page
Calculations of Differential Kinematic Properties
from Lagrangian Observations. Proc. AIAA Drift
Buoy Symposium, Hampton, Va., May 22-23, 1974,
NASA CP-2003, 193-209. 220
22. Starr, R. B., G. A. Berberian, and M. A. Weiselberg.
MESA New York Bight Project, Expanded Water Column
Characterization Cruise XWCC-1 of the R/V ADVANCE
II. NOAA Data Report ERL MESA- 22, 43 p.* 237
23. Voorhis, A. D., E. H. Schroeder, and A. Leetmaa.
The Influence of Deep Mesoscale Eddies on Sea
Surface Temperature in the North Atlantic Sub-
tropical Convergence. Journal of Physical
Oceanography, Vol. 6, No. 6, 953-961. 238
MARINE GEOLOGY AND GEOPHYSICS LABORATORY
24. Bennett, R. H., W. R. Bryant, W. A. Dunlap,
and G. H. Keller.
Initial Results and Progress of the Mississippi
Delta Sediment Pore Water Pressure Experiment.
Marine Geotechnology , Vol. 1, No. 4, 327-335. 247
25. Dash, B. P., M. M. Ball, G. A. King, L. W. Butler,
and P. A. Rona.
Geophysical Investigation of the Cape Verde
Archipelago. Journal of Geophysical Research,
Vol. 81, No. 29, 5249-5259. 256
26. Dietz, R. S.
Iceland: Where the Mid-Ocean Ridge Bares Its
Back. Sea Frontiers , Vol. 22, No. 1, 9-15. 267
27. Dietz, R. S. and K. 0. Emery.
Early Days of Marine Geology. Oceanus ,
Vol. 19, No. 4, 19-22. 274
* Abstract only;
complete text available on microfiche.
viii
Page
28. Dietz, R. S. and J. F. McHone.
El'gygtgyn: Probably World's Largest Meteorite
Crater. Geology, Vol. 4, No. 7, 391-392. 278
29. Freeland, G. L. and G. F. Merrill.
Deposition and Erosion in the Dredge Spoil and
Other New York Bight Dumping Areas. Proc.
American Society of Civil Engineers Specialty
Conference on Dredging and Its Environmental
Effects, Mobile, Al . , 26-28 January 1976, 936-946. 280
30. Freeland, G. L., D. J. P. Swift, W. L. Stubblefield,
and A. E. Cok.
Surficial Sediments of the NOAA-MESA Study Areas
in the New York Bight. Middle Atlantic Shelf and
the New York Bight, ASLO Special Symposia, Volume
2, 90-101. 291
31. Lavelle, J. W. , P. E. Gadd, G. C. Han, D. A. Mayer,
W. L. Stubblefield, D. J. P. Swift, R. L. Charnell,
H. R. Brashear, F. N. Case, K. W. Haff, and
C. W. Kunselman.
Preliminary Results of Coincident Current Meter and
Sediment Transport Observations for Wintertime
Conditions on the Long Island Inner Shelf. Geo-
physical Research Letters , Vol. 3, No. 2, 97-100. 303
32. Lowell, R. P. and P. A. Rona.
On the Interpretation of Near Bottom Water
Temperature Anomalies. Earth and Planetary
Science Letters, Vol. 32, No. 1, 18-24. 307
33. Nelsen, T. A.
An Automated Rapid Sediment Analyser (ARSA).
Sedimentology , Vol. 23, No. 6, 867-872. 314
34. Peter, G. and G. K. Westbrook.
Tectonics of Southwestern North Atlantic and
Barbados Ridge Complex. American Association
of Petroleum Geologists Bulletin, Vol. 60,
No. 7, 1078-1106. 320
IX
35. Richardson, E. and C. G. A. Harrison.
Page
Opening of the Red Sea With Two Poles of Rotation.
Earth and Planetary Science Letters, Vol. 30,
No. 1, 135-142. 349
36. Richardson, E. and C. G. A. Harrison.
Reply: Opening of the Red Sea With Two Poles of
Rotation. Earth and Planetary Science Letters ,
Vol. 30, No. 2, 173-175. 357
37. Rona, P. A.
Asymmetric Fracture Zones and Sea-Floor
Spreading. Earth and Planetary Science
Letters, Vol. 30, No. 1, 109-116. 360
38. Rona, P. A.
Book Review: Plate Tectonics and Oil.
Earth Science Reviews, Vol. 12, No. 1, 74-75. 368
39. Rona, P. A. , Editor.
Mid-Atlantic Ridge. Geological Society of
America, Microform Publication, Vol. 5, 490 p.* 369
40. Rona, P. A.
Pattern of Hydrothermal Mineral Deposition: Mid-
Atlantic Ridge Crest at Latitude 26° N. Marine
Geology, Vol. 21, No. 4, M59-M66. 371
41. Rona, P. A.
Resource Research and Assessment of Marine
Phosphorite and Hard Rock Minerals. Proc. of
N0AA Marine Minerals Workshop, March 1976, 111-119. 379
* Abstract only;
complete text on microform.
42. Rona, P. A.
Page
Salt Deposits of the Atlantic. Special Volume
of 'Annals of the Brazilian Academy of Sciences.
Anais Acad. Brasil Ciencies (Suplemento) , Vol. 48,
265-274. 388
43. Rona, P. A. and L. D. Neuman.
Energy and Mineral Resources of the Pacific
Region in Light of Plate Tectonics. Journal
of Ocean Management, Vol. 3, 57-78. 398
44. Rona, P. A. and L. D. Neuman.
Plate Tectonics and Mineral Resources of Circum-
Pacific Region. Papers from Circum-Pacif ic Energy
and Mineral Resources Conference, Honolulu, Hawaii,
August 26-30, 1974, publ . by Amer. Assoc, of
Petroleum Geologists, Memoir 25, 48-57. 420
45. Rona, P. A., R. N. Harbison, B. G. Bassinger, R. B. Scott,
and A. J. Nalwalk.
Tectonic Fabric and Hydrothermal Activity of Mid-
Atlantic Ridge Crest (lat 26° N). Geological Society
of America Bulletin, Vol. 87, 661-674. 430
46. Scott, R. B., J. Malpas, P. A. Rona and G. Udintsev.
Duration of Hydrothermal Activity at an Oceanic
Spreading Center, Mid-Atlantic Ridge (lat 26° N).
Geology, Vol. 4, No. 4, 233-236. 444
47. Stubblefield, W. L. and D. J. P. Swift.
Ridge Development as Revealed by Sub-Bottom
Profiles on the Central New Jersey Shelf.
Marine Geology, Vol. 20, No. 4, 315-334. 448
48. Swift, D. J. P.
Coastal Sedimentation. Marine Sediment Transport
and Environmental Management, D. J. Stanley and
D. J. P. Swift, editors, John Wiley and Sons, Inc.,
Chapter 14, 255-310. 468
XI
Page
49. Swift, D. J. P.
Continental Shelf Sedimentation. Marine Sediment
Transport and Environmental Management , D. J. Stanley
and D. J. P. Swift, editors, John Wiley and Sons,
Inc., Chapter 15, 311-350. 524
50. Swift, D. J. P. and J. C. Ludwick.
Substrate Response to Hydraulic Process: Grain-
Size Frequency Distributions and Bed Forms.
Marine Sediment Transport and Environmental
Management , D. J. Stanley and D. J. P. Swift,
editors, John Wiley and Sons, Inc., Chapter 10,
159-196. 564
51. Swift, D. J. P., G. L. Freeland, P. E. Gadd, G. Han,
J. W. Lavelle, and W. L. Stubblefield.
Morphologic Evolution and Coastal Sand Transport,
New York-New Jersey Shelf. Middle Atlantic Shelf
and the New York Bight, ASLO Special Symposia,
Volume 2, 69-89. 602
VOLUME II
SEA-AIR INTERACTION LABORATORY
52. Apel , J. R., H. M. Byrne, J. R. Prom", R. Sellers.
A Study of Oceanic Internal Waves Using Satellite
Imagery and Ship Data. Remote Sensing of Environ-
ment 5, No. 2, 125-135. Also appeared in Proc.
Thirteenth Space Congress, Technology for the
New Horizon, Cocoa Beach, Florida, April 7, 8, 9,
1976, 3-21--3-25. 623
53. Hanson, K. J.
A New Estimate of Solar Irradiance at the Earth's
Surface on Zonal and Global Scales. Journal of
Geophysical Research, Vol. 81, No. 24, 4435-4443. 634
XII
54. Hasselmann, K. , D. B. Ross, P. Muller, and W. Sell
Page
A Parametric Wave Prediction Model. Journal
of Physical Oceanography, Vol. 6, No. 2, 200-228. 643
55. McLeish, W. and S. M. Minton.
STD Observations From the R/V COLUMBUS ISELIN
During Phase III of GATE. NOAA Technical
Re-port ERL 379-AOML 24. 101 p. 672
56. Newman, F. C.
Temperature Steps in Lake Kivu: A Bottom Heated
Saline Lake. Journal of Physical Oceanography ,
Vol. 6, No. 2, 157-163. 776
57. Proni, J. R., F. C. Newman, D. C. Rona, D. E. Drake,
G. A. Berberian, C. A. Lauter, Jr., and R. L. Sellers.
On the Use of Accoustics for Studying Suspended
Oceanic Sediment and for Determining the Onset of
the Shallow Thermocline. Peep-Sea Research ,
Vol. 23, No. 9, 831-837. 783
58. Proni, J. R. , F. C. Newman, R. L. Sellers, and C. Parker.
Acoustic Tracking of Ocean-Dumped Sewage Sludge.
Science, Vol. 193, 1005-1007. 794
59. Thacker, W. C.
A Solvable Model of "Shear Dispersion." Journal
of Physical Oceanography, Vol. 6, No. 1, 66-75. 797
60. Thacker, W. C.
Spatial Growth of Gulf Stream Meanders. Geophysical
Fluid Dynamics , Vol. 7, 271-295. 807
61. Webster, W. J., Jr., T. T. Wilheit, D. B. Ross, and P. Gloersen.
Spectral Characteristics of the Microwave Emission
From A Wind-Driven Foam-Covered Sea. Journal of
Geophysical Research, Vol. 81, No. 18, 3095-3099. 832
Xlli
OCEAN CHEMISTRY LABORATORY
Page
62. Atwood, D. K.
Regional Oceanography as it Relates to Present
and Future Pollution Problems and Living Resources-
Caribbean. IOC/FAO/UNEP International Workshop on
Marine Pollution in the Caribbean and Adjacent
Regions, Port of Spain, Trinidad, IOC/FAO/UNEP/
IWMPCAR/8, 40 p. 837
63. Gilio, J. L. and D. A. Segar.
Biogeochemistry of Trace Elements in Card Sound,
Florida Inventory and Annual Turnover. Proc.
of the Sea Grant Symposium on Biscayne Bay,
April 2-3, 1976, 17 p. 879
64. Hatcher, P. G. and L. E. Keister.
Carbohydrates and Organic Carbon in New York
Bight Sediments as Possible Indicators of Sewage
Contamination. Middle Atlantic Continental Shelf
and the New York Bight, ASLO Special Symposia,
Volume 2, 240-248. 896
65. Hatcher, P. G. and D. A. Segar.
Chemistry and Continental Margin Sedimentation.
Marine Transport and Environmental Management ,
D. J. Stanley and D. J. P. Swift, editors,
Chapter 19, 461-477. 905
66. Segar, D. A. and G. A. Berberian.
Oxygen Depletion in the New York Bight Apex:
Causes and Consequences. Middle Atlantic
Continental Shelf and the New York Bight,
ASLO Special Symposia, Volume 2, 220-239. 922
67. Segar, D. A. and A. Y. Cantillo.
Some Considerations on Monitoring of Trace Metals
in Estuaries and Oceans. Proc. of the International
Conference on Environmental Sensing and Assessment,
IEEE Annuals No. 75CH004-1, 6-5, 1-5. 942
xiv
Page
68. Segar, D. A. and A. Y. Cantillo.
Trace Metals in the New York Bight. Middle
Atlantic Continental Shelf and the New York
Bight, ASLO Special Symposia, Volume 2,
171-198. 947
69. Tosteson, T. R,.D. K. Atwood, and R. S. Tsai .
Surface Active Organics in the Caribbean Sea.
MTS-IEEE Oceans '76, 13C-1-13C-7. 975
xv
Reprinted from: EOS, Vol. 57, No. 9, 612-624.
Ocean Science From Space
John R. Apel
Introduction
The ocean plays as fundamental a
role in the natural scheme of things
as does the atmosphere, although its
functions, being considerably more
varied and diffuse, are probably
neither as well appreciated nor as
well understood. The sea profoundly
affects the weather and climate and
in turn is affected by the atmo-
sphere, acting as both a heat reser-
voir for storing, distributing, and
releasing solar energy and as the
source for most atmospheric
moisture. It interacts with the
bounding land and air over times
ranging from minutes to millennia.
Geological activity on all time and
space scales- takes place in and
under the seas, which serve as the
repository for the detritus of man
and nature and, just as important,
as practicable sources of petroleum
and a few useful minerals. Its cur-
rents and dilutant powers are called
upon to disperse sewage, poisonous
and nonpoisonous wastes, solid
trash, and excess heat, while it
maintains a role as the aqua viva
for an extremely complicated and
612
commercially important food chain
and a role as a means of recreation
and refreshment for people. In the
estuaries and the coastal zones
these conflicting demands are
especially severe.
This article attempts a rather
limited review of the types of
oceanic information that current
experimental results and planning
indicate should be available from
spacecraft in the near future. To the
author's knowledge, plans exist to
orbit sensors that will yield
measurements or observations of all
of the parameters discussed here,
albeit often only on an experimental
basis.
Questions on the usefulness of
satellites for ocean science have
been raised by oceanographers since
the first space-derived imagery was
returned to earth. It was not at all
obvious what relationships such
data might have to the physical
oceanographer's usual repertory of
salinity, temperature, and depth
measurements, the biologist's con-
cerns with flora or fauna, or the
geologist's interests in rocks or sedi-
ments.
After nearly two decades of ac-
tivity in space it is becoming ob-
vious that for several limited but
nevertheless important classes of
phenomena it is possible to make
observations and measurements
from spacecraft of considerable
usefulness to oceanographers. In a
few isolated instances it even ap-
pears one may do so with a breadth
and accuracy exceeding anything
attainable from ships or buoys. For
these types of observations, the
satellite represents a new tool of
great power, and the information on
physical and biological processes ob-
tained from it will be worthy of in-
clusion in the data banks and in the
minds of researchers.
However, for a sizable percentage
of physical and biological ocean sci-
entists, much of these data may fall
far afield or might be too indirect or
perhaps even too esoteric for their
tastes. The value of the data to
these workers will chiefly be in the
concomitant enlargement of the
general fund of oceanic knowledge.
By and large, satellite oceanogra-
phy is confined to surface and near-
surface phenomena. This constraint
is not as severe as it appears at first
glance, hecause data taken from
spacecraft will be appended to
other, conventionally derived sur-
face and subsurface measurements
of parameters such as vertical cur-
rent or temperature profiles in
order to construct a more nearly
three-dimensional view of the
ocean. In addition, near-surface
data are useful in their own right,
since the coupled nonlinear interac-
tions between ocean and atmo-
sphere largely take place in the few
tens of meters above and below the
sea-air interface, at least for shorter
time scales. Man's marine activities
are mostly confined to near that
surface as well, so that the kind of
two-dimensional oceanography that
one can pursue from spacecraft is
often highly relevant.
Uses of Spacecraft Data
ble 1 is a listing [LaViolette, 1974;
Apel and Siry, 1974; NASA, 1975;
Koffler, 19751 of the spacecraft that
have been or will be sources of data
having oceanographic significance.
Of the several listed, the most
useful are probably NOAA 3 and 4,
ERTS 1/Landsat 2, Geos 3, the
SMS/GOES quintuplets, Tiros-N,
Seasat-A, and Nimbus-G. The data
types are diverse, as is discussed
below. The last three satellites,
which are to be launched in 1978,
are of much interest to oceanogra-
phy. Tiros-N is the first of a new
generation of operational
meteorological and environmental
polar-orbiting satellites. Seasat-A is
dedicated to oceanography, geodesy,
meteorology, and climatology \Apel
and Siry, 19741. Nimbus-G is
designed to serve experimental ends
for both pollution monitoring and
oceanography [AVISA, 19751.
Data Available From Satellites
Spacecraft data presently availa-
ble on any basis other than a pri-
marily experimental one are quite
limited and are effectively confined
to low- and medium-resolution visi-
ble and infrared imagery (NOAA,
GOES), from which sea surface tem-
peratures having accuracies of
order ±1.5°-2.0°C may be derived,
and small amounts of high-resolu-
tion Landsat images. However, the
near future promises a large in-
crease in the quantity, quality, and
coverage of oceanic data.
The estimates of data accuracy
and coverage cited below are
thought to be valid for the general
1978-1982 era, when Tiros-N,
Seasat-A, Nimbus-G, Landsat 3,
and the GOES system are all to be
active. In each case the dominant
instruments contributing to the
The answer as to who needs what
information from spacecraft ob-
viously depends on the type of infor-
mation that is obtainable. In
research areas the disciplines
served with some degree of useful-
ness are marine geodesy and gravi-
ty; physical, geological, and biologi-
cal oceanography; glaciology; boun-
dary layer meteorology; and
climatology. Various maritime
operations, shipping, offshore min-
ing, oil drilling, and fishing, all re-
quire an improved and expanded
data base and more accurate
marine forecasts. The ever-increas-
ing fraction of the population living
along the seacoasts needs improved
forecasting and warning services
for protection of life and property.
However, because of the great
length and breadth of the sea the
difficulties in obtaining timely
detailed information of sufficient
observational density across its ex-
panse have prevented an effective
monitoring and forecasting system
for the oceans.
Satellites of Utility
to Oceanography
The number of satellites carrying
sensors that yield data useful to
ocean science is large, and the value
of the data from them variable. Ta-
Fig. 1. Surface isotherms in degrees centigrade of Lake Huron derived from the
VHRR sensor on the NOAA 4 environmental satellite, August 7, 1975. Relative ac-
curacy is approximately ±1°C (NOAA, National Environmental Satellite Service).
613
TABLE 1. U.S. Satellites of Utility in Oceanography
Satellite
Launch
Date
Orbit
Utility
of Data
Character
Sensors
Oceanic
Parameters
Mercury
Gemini
Apollo
Apollo-Soyuz
Nimbus 4
Nimbus 5
Nimbus 6
Nimbus-G
ITOS 1-4
ESS A 1-9
NOAA 1-4
ATS 1-3
SMS/GOES 1-5
Geos 1-3
ERTS 1
Landsat 2
Landsat 3
Skylab
1962-
1975
1970
1973
1975
1978
1966-
1975
1966
1967
1974-
1978
1965
1975
1972
1974
1978
1973
Variable
Polar
Polar
Low to Exploratory Cameras
medium
Imagery
Polar
Medium Experimental IR and MW radiometers Temperature, ice cover,
progressing and bolometer; color radiation budget, wind,
to high scanner color
Medium and Operational
high
Synchronous Medium Prototype
Synchronous High Operational
Variable High Experimental
Medium Prototype
progressing
to high
Medium Experimental
progressing
to high
Visible vidicon; IR
scanner
Visible, IR scanners;
data channel
Visible, IR scanners;
data channel
Laser reflectors;
altimeter
Visible, near-IR scanner;
thermal IR scanner
Cameras, visible, IR
scanner: spectre radi-
ometer; MW radiom-
lU'is; altimeter; scatter-
ometer
Shuttle
1983
Varied
M
edium to
high
Varied
Varied
Tiros- N
1978
Polar
H.gh
Operational
Visible, IR scanners
Seasat-A
Seasat-B
1978-
1983
Near Po!
lar
High
Experimental
Altimeter; imaging
radar; scatterometer;
MW radiometer;
visible/IR scanner
Imagery, temperature
Imagery, temperature,
data relay
Imagery, temperature,
data relay
Geoid, ocean geoid
Imagery, temperature
Imagery, temperature,
wave height, wind
speed, geoid
Unknown
Imagery, temperature
Geoid, wave spectra,
wind speed, ice,
temperature
From LaVioh'ttv 119741, Apel and Siry 11974!, and S'ASA 119751.
TABLE 2. Sensors of Oceanographic Interest
Short Form
Sensor Name
Wavelength or Frequency
Spatial
Spacecraft
Resolution
NOAA 1-4
7 km
NOAA 1-4
1 km
GOES
1-7 km
Tiros-N
1 km
ERTS/Landsat 1 and 2
70 km
Landsat 3, Landsat 4
25 m, 100 m (IR
Nimbus-G
800 m
Nimbus 5
15 km
Nimbus-G, Seasat-A
15-140 km
Skvlab. Geos 3, Seasat-A
2 km
Skvlab, Seasat-A
25 km
Seasat-A
25 m
SR Scanning radiometer
VHRR Very high resolution
radiometer
VISSR Visible and infrared spin
scan radiometer
AVHRR Advanced very high
resolution radiometer
MSS Multispectral scanner
Thematic mapper
CZCS Coastal zone color scanner
ESMR Electronically scanned
microwave radiometer
SMMR Scanning multichannel
microwave radiometer
Alt Short pulse altimeter
Scatt Radar wind scatterometer
SAR Synthetic aperture radar
Visible and thermal IR
Visible and thermal IR
Visible and thermal IR
Visible and thermal IR
Four channels, visible and
reflected IR;
Thermal IR
Six channels, visible, reflected
and thermal I R
19 GHz
Five channels: 6.6. 10. 18. 21,
35 GHz
13.9 GHz
13.4 GHz
1.4 GHz
614
measurement are listed, although to
achieve the precision or accuracy
cited, ancillary data will usually be
required. There is every reason to
blend surface and satellite data
together, so that the space-derived
information can be calibrated and
verified by point surface measure-
ments and thus can often extend the
surface observations to near-plane-
tary scales. The sensors of prime in-
terest are also cited and with their
shortened forms are listed in Table
2.
One finds a diverse list of
features, or observables, that enter
into oceanic processes. In listing
these parameters it is convenient to
begin at the level of the action of the
atmosphere upon the sea; then
follow the ocean's response, waves
and currents, and its effects upon
the shore. Other land-sea interac-
tions are then listed. Identification
of water mass properties
established by natural and man-
made influences is discussed next.
Finally, some estimates of the role
of the ocean in establishing
climatology are given.
In many of the parameter values
and ranges given below the lack of
experimental verification requires
that the data be regarded as
preliminary estimates only, and the
reader is cautioned to remain skep-
tical. In most cases they represent
compromises between requirements
leveled by the ocean scientists and
the attempts of the instrument
designers to meet those require-
ments via remote sensing.
Air-Sea Interaction
The transport of matter, momen-
tum, and energy across the air-sea
interface is chiefly due to solar
radiation and atmospheric stress.
Such parameters as the air-sea tem-
perature difference, exchange of la-
tent and sensible heat, and the vec-
tor surface wind field are important
observables for climatological,
meteorological, and oceanic pur-
poses. For spacecraft the following
estimates appear reasonable.
Sea surface temperature. For the
estimated capability, in cloud-free
areas it should be possible to deter-
mine absolute temperature ac-
curacy to order 1°C and precision or
relative accuracy to approximately
±0.5°C. Over coastal waters and
lakes, space-time averaging of order
4 km and 1 day is needed \Koffler,
19751; for regional ocean areas, 10-
km and few-day averages are re-
quired; in the open ocean, 50-km
and several-day averages should
suffice [Bromer et al., 19761. The
sensors to be used are VHRR,
VISSR, and AVHRR (Table 2). In
cloudy areas or in light rain a tem-
perature precision of ±1.5°-2.0°C
should obtain with 100-km and few-
day averages away from coasts by
using SMMR. To the satellite-
derived temperatures should be ap-
pended ship surface and vertical
temperature profiles to the max-
imum extent possible.
Figure 1 shows isotherms for
Great Lakes surface temperatures
as an example of the current high-
resolution thermal mapping in a
limited region, derived from the
VHRR sensor on the NOAA 4
satellite [Koffler, 19751.
Surface rector wind field. As
referenced to a 20-m height, the
scatterometer may measure surface
wind speed from a very few to
perhaps 20 m/s, with a precision of
about ±2 m/s or 25% of the actual
value (whichever is larger) and
wind direction to ±20° through
clouds and light rainfall; 25-km
resolution over a several hundred
kilometer swath width will be the
case [Grantham et al., 19751. For
higher winds, attempts will be made
to determine speed from 5 to
perhaps 35 m/s within ±25'/!', of ac-
tual speed over a several hundred
kilometer swath through clouds and
light rain by using the SMMR [Apel
and Sirx, 1974, p. 14; NASA, 1975;
Baruth and Gloersen , 1975!.
Figure 2 shows radar backscatter
cross section of the ocean <r° as a
function of wind speed at 20 m, with
angle of illumination (measured
from nadir) as a parameter. This
effect forms the basis for the wind
speed measurement with the radar
wind scatterometer \Grantham et
al., 19751.
Radiation budget. Precision
Incidence angle
0°
JHH
Wind speed, m/s
Fig. 2. Radar cross section it mf of ocean surface at 13 GHz versus surface wind
speed measured at 20-m height. Angle off nadir is the parameter. Horizontal
polarization, cross-wind illumination (NASA Langley Research Center).
4
615
TRANSMITTED
PULSE
^Sni — »J
RECEIVED-
SMOOTH OCEAN
RECEIVED-
ROUGH OCEAN
Fig. 3. Short-pulse method for determining significant wave height with a 3-ns radar
altimeter.
radiometers are estimated to be able
to determine spectrally integrated
solar radiation absorbed in and
reflected by the global ocean, with a
precision of approximately ±5
Ly/day, with various spatial resolu-
tions [NASA, 19751.
Surface Wave Field
There is obviously a strong coup-
ling between the surface wind field
(°)
and ocean waves, with the wind in-
itially generating short-length
capillary waves which then cascade
toward longer wavelengths and
larger amplitudes, dependent upon
the strength, direction, duration,
and fetch of the wind. While signifi-
cant wave height //1/3 is a one-pa-
rameter specification of sea state,
the proper description of a
homogeneous surface wave field is
more detailed, e.g., the two-dimen-
(c)
WAVELENGTH
= 60 m
ANGLE
= 83'
WAVELENGTH
= 150 m
ANGLE
= -15";
60m
u±
Fig. 4. (a) band synthetic aperture radar image of 60-m and 150-m ocean waves off
Alaska; (b) two-dimensional Fourier transform of part o showing wave energy con-
centrations as bright spots; (c) interpretation of part b in terms of two dominant wave
trains, with densitometer traces of the figure taken at 83° and -15° (Jet Propulsion
Laboratory).
sional power spectral density as a
function of surface wave vector. A
reasonably complete determination
of this function near storms, when
used as input data to numerical
models, would allow wave forecasts
to be made at a distance of several
hundred kilometers from the high
wind regions. Where the field is
n o n h o m oge neou s , as near
shorelines, near intense low
pressure systems, or in shoaling
water, an image of the surface field
is more appropriate than a
spectrum.
Significant wave height. For the
estimated capability, it appears
possible to measure significant
wave height W1/3 with a precision of
±1 m or ±25% of the actual height
over a range of 1-20 m along the
subsatellite track on a near-all-
weather basis by using the short-
pulse altimeter I Walsh, 19741.
Figure 3 illustrates the effect of a
rough ocean in broadening a 3-ns
radar altimeter pulse, the measure-
ment of which forms the basis for
the determination of W1/3.
Surface wave spectrum. For the
surface wave power spectrum the
synthetic aperture radar (SAR) may
yield square amplitude measure-
ments consistent with the precision
for //1/3 (above) for all wavelengths
between 50 m and the largest obser-
vable length, measured at 10° inter-
vals for all angles of propagation;
the spatial and temporal resolution
is limited to small samples taken
near the United States or to more
intensive spectra in selected
regions. The instrument appears to
have an all-weather capability
[Brown era/., 19761.
Wave refraction pat-
terns. Surface waves reflect,
refract, and diffract under the in-
fluence of shoal water and may con-
verge or diverge, depending on bot-
tom topography. Heavy wave action
moves shoals and channels about
and damages ocean structures such
as jetties and offshore platforms.
Wave refraction studies for a given
region assist in shoreline protection,
channel maintenance, and under-
standing of wave-driven circulation.
Under these conditions, images
rather than spectra are required.
The SAR should image wave refrac-
616
tion patterns for wavelengths
greater than 50 m over swath
widths of up to 100 km on a selected
basis; it does so with a near-all-
weather capability \ Brown et <;/.,
19761.
Figure 4 shows a surface wave
field as obtained from the synthetic
aperture imaging radar and a
digital Fourier transform, which ap-
pears to yield a wave slope spectrum
I Brown vt «/., 19761.
Currents and Vertical Motions
Ocean currents are driven by
wind stress, by tidal forces, and by
uneven temperature and salinity
distributions in the body of the sea.
On the rotating earth a moving fluid
tilts its surface relative to the geoid
with a slope proportional to the fluid
velocity; this is called geostrophic
flow. In the case of western bound-
ary currents, e.g., the Gulf Stream,
the slopes are of order 10 ~ r' or less;
the resultant topographic elevations
across the stream, measured with
respect to the geoid, are about 1 m
or less.
Upwellings and downwellings are
slow vertical flows usually brought
about by wind stress and coastal
topography. Upwellings in particu-
lar are of interest because the cold
subsurface water often has a high
nutrient level that may lead to a
plankton bloom and ultimately an
enhanced fish population. From the
standpoint of spacecraft data the
speed of the current in an upwelling
is not observable, but rather the
timely identification and location of
the event are possible.
In order to determine the com-
plete dynamical current velocity
field, one must measure speed and
direction as a function of position
and time. In addition, the vertical
distribution of current velocity
throughout the water column is
needed for measuring total
transports of water, dissolved
chemicals, nutrients, etc. This is ob-
viously impossible from satellites,
and therefore to any surface current
measurements made from
spacecraft must be appended sub-
surface current profiles taken by
conventional means.
Present estimates \Kaitla, 1970;
Apel, 1972; Apel and Byrne, 19741
give roughly ±20 cm/s as the ulti-
mate achievable precision in the
determination of surface
geostrophic speeds from spacecraft
by way of surface slope measure-
ments using a radar altimeter and
perhaps several kilometers as the
time-averaged error in the position
of the current measurements along
the subsatellite track only.
Nevertheless, surface current
speeds considerably below 20 cm/s
are found in the ocean and are of in-
terest. No apparent means exist for
remotely determining such low
speeds from spacecraft. However,
drifting Lagrangian buoys may act
as near-surface water movement
tracers for these lower speeds
\Molman, 19741. When the drifting
buoys are equipped with satellite
positioning devices and data collec-
tion systems, they become ex-
tremely valuable adjuncts to the
remote sensors on board the
spacecraft.
However, it should be emphasized
again that spacecraft remote sen-
sors alone can by no means deliver
all of the required information.
Figure 5, taken from Defant
119611, shows the long-time mean
surface topography of the western
North Atlantic as calculated assum-
ing that geostrophy obtains, with
elevations above and below an
equipotential surface close to the
geoid given in centimeters. The
time-averaged Gulf Stream is
110° 100° 90° 80° 70° 60° 50° 40° 30°
Fig. 5. Long-term topographic setup of western North Atlantic as calculated from
oceanic density anomalies; elevations are given in centimeters relative to the geoid
\Defctnt, 1961!.
617
Fig. 6. Thermal infrared image made off U.S. East Coast on May 12, 1975, showing
Gulf Stream, meanders, and eddies in lighter shades; dark areas are cold clouds
(NOAA, National Environmental Satellite Service).
clearly visible; its instantaneous
position may depart from the
horizontal mean axis by amounts
approaching 200 km, moving slowly
(5- to 40-day periods) in comparison
with the time (a few days) required
to map the area with a satellite. The
hope is that satellite altimetry will
become sufficiently precise so that
this dynamic topography, and hence
surface current speed, can be deter-
mined by using it \Kaula, 1970;
Apel, 1972; Apel and Byrne, 19741.
This requires that both the back-
ground geoid and the topographic
departures from it be determined
with precisions approaching ±10
cm in the vertical. The requirement
inextricably links dynamical
oceanography and marine geodesy
if such schemes are to be pursued.
Figure 6 illustrates a NOAA 4
thermal infrared image off the
northeastern U.S. coast with the
warm water of the Gulf Stream in
lighter shades [Koffler, 19751. Such
imagery can be used to interpolate
between the altimetry traces in
order to obtain a more complete
mapping of the Gulf Stream or simi-
lar intense flows in regard to sur-
face position and current speed.
For upwellings it appears feasible
to determine position, temperature,
and areal extent of an upwelling
event to 5 km within 1 to 2 days of
its onset and to obtain estimates of
the near-surface chlorophyll con-
centration by using combined tem-
perature and color imaging devices
such as CZCS [NASA, 1975).
Tides: Open Ocean and Shelf
Deep-sea tides, being largely
astronomically driven by the moon
and sun, occur at precise frequen-
cies, some five of which contain
about 95% of the tidal energy. Their
amplitudes in the open ocean are
typically 0-1 m. Open ocean and
shelf tides are difficult and time-
consuming to measure, and their re-
lationships to coastal tides are hard
to establish. Worldwide deep-sea
tidal measurements would aid in
the theoretical understanding and
prediction of tides at arbitrary loca-
tions along the coastlines.
By using precision altimetry in
the way described earlier, it appears
that one may determine tidal range
to ±25 cm (relative to mean sea
level) and phase to ±20° for diurnal
and semidiurnal periods [Hen-
dershott et ai, 19741. The required
spacings are 25 km on continental
shelves and 100 km globally. Ap-
proximately 1 year of data is needed
for the solution.
Sea -Earth Interactions
In the category of interactions be-
tween the ocean and the solid earth
is found such a wide diversity of
features that no general discussion
will suffice. Instead, each observa-
ble will be taken up individually.
Storm surge and wind setup along
a coast. Storm systems pile up
water ahead of them as they ap-
proach a coastline from seaward. In
the case of hurricanes this surge is
often directly responsible for more
damage and loss of life than the
wind is. Hurricane surges are con-
fined to a few tens of kilometers and
a few hours of time during the land-
fall; amplitudes can exceed 9 m.
Wind setup is the accumulation of
water along a coast due to long-term
stresses such as trade winds; a typi-
cal elevation is about 1 m.
618
By altimetric means it should be
possible to measure storm surge
elevations to ±1 m in a storm
system on a target-of-opportunity
basis, along a single subsatellite
track [Apel and Siry, 19741. It
should be recognized that the space-
time coincidence of storm and
satellite is a low probability event,
however.
Tsunamis. Tsunamis are
seismically excited long-length
ocean waves capable of great
damage. Their peak-to-trough
amplitudes in midocean have never
been measured but theoretically
should be of order lh m;
wavelengths are a few hundred
kilometers, and the disturbance
ultimately fills an entire ocean
basin. As they approach shore, the
amplitude may increase to tens of
meters. Assessing the energy con-
tent of a tsunami is a difficult task,
and thus much overwarning results.
In principle, altimetric measure-
ments could yield a tsunami
amplitude to ±25 cm and a
wavelength to ±20% in the open
ocean on a target-of-opportunity
basis along a subsatellite track.
This is again a low-probability ob-
servation \Apel and Siry, 19741 .
Beach and shoal
dynamics. Waves and currents
erode and build shorelines and
shallow water features. Base line
data on shoreline and shoal con-
figurations allow assessment of
changes due to wave action. By
using an imaging radar it should be
possible under storm conditions to
image shorelines and shoal waters
with resolutions down to 25 m with
image centers located to ±500 m
over swath widths of up to 100 km
on a selected basis near the conti-
nental Unites States. High-resolu-
tion optical and near-infrared imag-
ery taken at several wavelengths
(such as will be available from the
Landsat 4 thematic mapper) can
yield data on subsurface conditions
as well under clear skies.
Shallow -water charting and
bathymetry. The positioning of
newly formed or poorly charted
shoals and some assessment of their
topography can be obtained by
using multispectral optical imagers
such as MSS or CZCS. It is possible
Fig. 7. Microwave image of the Antarctic continent with brightness scales affixed
made from the ESMR on Nimbus 5 during January 1973 (NASA Goddard Space Flight
Center).
to image shoals of depths less than
10-15 m where the water is clear
enough, with vertical resolutions of
2-5 m and horizontal resolutions of
order 70 m, with image centers lo-
cated to ±500 m, on a selected basis
[Polcyn and Lyzenga, 19741.
Near-surface sediment
transport. Wave action, river dis-
charges, tidal flushing, and advec-
tion by current systems result in
transport of sediments and sands
throughout the ocean. Surface sedi-
ment patterns and particulate con-
centrations are indicators of
transport of material, which can be
viewed at several optical1
wavelengths with 800-m resolution
over swath widths of up to 700 km
(MSS, CZCS). By designing
algorithms that use image
brightnesses at these wavelength
bands it may be possible to deter-
mine concentrations from approx-
imately 0.2 to 100 mg/ma on a
selected basis [Pirie and Steller,
19741.
Ice cover, dynamics, and
icebergs. Ice cover and ice move-
ments vary greatly with the time of
year and surface wind conditions.
The percentage of ice cover in polar
regions governs much of the
weather there, owing to the large
exchange of heat between air and
water occurring through open water
areas, especially in narrow leads
and openings. In coastal areas and
lakes, shipping depends upon an ac-
curate assessment of ice conditions
throughout the navigable waters.
Iceberg tracking and forecasting
are vital for protection and naviga-
tion of shipping. The observation of
ice from satellites is greatly com-
pounded by the persistent cloud
cover found in polar and subpolar
regions. Thus the synthetic aper-
ture radar will be very useful for
imaging ice cover and perhaps very
large icebergs, with a resolution of
25 m and with image centers lo-
cated to ± 500 m, over swath widths
of up to 100 km, on a near-all-
weather but very selected basis.
With the SMMR it is possible to im-
age ice cover with low resolution, 20
km, over the entire polar caps with
swaths of 1000 km on a near-all-
weather basis \Gloersen and
Salomonson, 19751.
Figure 7 is a brightness map of
the Antarctic continent as obtained
from the 19-GHz microwave
radiometer on Nimbus 5 and
gathered in the course of approx-
8
619
0.0
- 10.0
20.0
30.0
- 40.0
- 50.0
60.0
70.0
80.0
90.0
-100.0
17h15m20*
17h16mOO'
GMT-TIME-S
17h16m30'
1000
2000
3000
4000
5000
2
l
£ 6000
7000 _
8000
9000
10000
1 1 1
SKYLAB II
PASS 4 MODE 5
4 JUNE 1973
-
I
^ BOI
^ TOF
TOM
•OGRAP
«Y
I I
\l
>
\
A
\
17n15m20s
17n16m00'
GMT-TIME-S
17n16m30'
Let 198°
160°
Fig. 8. (Topi Altimeter geoid heights referenced to a spheroid, as measured across
western Puerto Rico by the Skylab S-193 altimeter; precision is about ± 1 m. (Bottom)
Bottom topography over a portion of the subsatellite track (NASA Wallops Flight
Center).
imately 5 days; the resolution is ap-
proximately 15 km \Gloersen and
Salomonson, 19751.
Marine geoid. In a quite separ-
ate category from the previous nb-
servables is the marine or ocean
geoid, defined as the surface
assumed by a motionless uniform
ocean under the influence of gravi-
tational and rotational forces only.
Geostrophic currents, tides, storm
surges, setup, and waves lead to an
ocean surface that departs from the
geoid; the latter must then be
known on a spatial grid with preci-
sion at least as fine as that with
which the observable is to be deter-
mined. Although only preliminary
data have been published, it appears
altogether possible to measure rela-
tive short-scale vertical variations
in the marine geoid to ±20 cm and
long-scale to perhaps ±100 cm
along the subsatellite track over a
grid spacing of order 25 km over all
open ocean areas by using the
altimeter and precise orbit deter-
mination \Apel and Stry, 1974;
Kaula, 1970; Apel, 1972; Apel and
Byrne, 1974; McGoogan et al., 1975!.
Data from Skylab I McGoogan et al.,
19751 and Geos 3 (H. R. Stanley, pri-
vate communication, 1975) support
this view. Some of the data from
Skylab are illustrated in Figure 8,
which shows the variation in rela-
tive geoid height and water depth
along the subsatellite track across
Puerto Rico. By using the altimeter,
whose noise figure was approx-
imately ± 1 m, the gravity anomaly
associated with the Puerto Rico
trench is clearly seen as a geoidal
depression of order 15-20 m
[McGoogan et al., 19751. Such short
wavelength data, taken globally,
can be combined with long
wavelength geoidal models obtained
via satellite tracking and orbit
analysis to obtain a precision geoid
over the ocean. While this has not
yet been done, attempts have been
made to combine marine gra-
vimetric measurements with
satellite geoids to produce a geoidal
map such as is shown in Figure 9 for
the western Atlantic I Vim cut et al.,
1972; Marsh et ui, 1973!. Here
heights are given in meters relative
to the reference ellipsoid. The
problem of measuring geostrophic
620
Fir. 9. Geoid calculated in the western Atlantic from satellite orbit perturbations and marine gravity measurements; elevations
are in meters relative to the reference ellipsoid. The track of Sky lab while taking the data of Figure 8 is shown as a stripe off Puer-
to Rico (NASA Goddard Space Flight Center).
currents is equivalent to discerning
the 100-cm setup due to current
shown in Figure 5 against the hack-
ground of 100-m geoidal undula-
tions illustrated in Figure 9 \ Kan la,
1970; A/W, 1972; Apcl an, I Byrne,
19741.
Climatology
The role of the ocean in climatic
change is not completely under-
stood, but it it clear that the
transformation of absorbed sunlight
into thermal energy in the upper
layers of the sea is an important
one, as is the poleward transport of
this heat by western boundary cur-
rents. Variations in the positions of
major ocean currents in part appear
to be induced by changing wind
stress, which apparently lead to the
El Nino phenomenon, for example.
The appearance of anomalous large
areas of warm water in the Pacific
has been hypothesized as the origin
of warm winters in the eastern
United States through poorly under-
stood processes involving motions of
the upper atmosphere [Gates and
Mintz, 19751.
The contributions which
spacecraft can make to ocean
climatology therefore appear to be
mainly related to the global deter-
mination of sea surface tern-
10
621
10
5
d.
5
4.13 mg m~3
9.2 xlCT'nn-'sr-1
GULF STREAM
Chl-aC13mg m-J
yS i 1.7xl0_sm*"sr-'
COASTAL
Chi a 0.28mg m-'
/3 :6.7xlO"V'sr-'
I I
400
500
600 700
WAVELENGTH
800
900
(nm)
Fig. 10.
the Gulf
surface i
Upwelling spectral irradiance as measured in three types ofwatei masses in
of Mexico; the shift toward the red end of the spectrum is due to increased
hlorophyll a iNOAA Environmental Research Laboratories).
perature and heat transport \Gates
and Mintz, 1975; Stommel, 1974;
NACOA, 19741. The five GOES-type
synchronous satellites appear capa-
ble of delivering the temperature
data over mid-latitude regions with
the required accuracy of ±0.5°C rel-
ative, if special processing is under-
taken. Over polar regions the Tiros-
N series is more suitable. Programs
for optimal extraction of the global
temperature fields, averaged over
approximately 100 x 100 km- areas
and several days, are in the
embryonic stages.
Water Mass Properties
Variations in the physical or
chemical composition of a water
mass lead to variations in its color
or reflectivity, for example. These
changes can be natural or man-
made; in either case they tend to be
more pronounced near continents.
The color is determined primarily
by molecular scattering and secon-
darily by nutrients, chlorophyll a in
plankton and algaes, suspended
sediment load, pollutants, and,
TABLE 3. Summary of Sensors and Observables
Imaging Radiometers
Short Pulse
Imaging
Observables
Visible
Thermal IR
Microwave
Altimeter
Radar
Chlorophyll and algaes
Current position
1
2
1
1
1
1
Current speed
Estuarine circulation
Fog
Ice cover
1
1
2
1
1
1
3
1
Icebergs
Internal waves
-
1
1
Marine geoid
Oil spills
-
-
1
1
Pollutant identification
Salinitv
-
3
•
-
Sea state and swell
Sediment transport
Setup
2
1
-
2
1
1
1
Shallow water bathymetrv
1
-
-
-
Storm surges
Surface winds
3
3
3
1
1
2
3
2
Temperature
Tides
1
1
1
-
Tsunamis
1
-
Upwellings
Water vapor
2
1
1
2
1
.
Wave refraction
1
-
-
-
1
Wave spectrum
2
1
Scatterometer
Numbers indicate order or importance in determining the observable with 1 for primary. 2 for secondary, and 3 for tertiary. Hy-
phens indicate no utility.
622
11
where water is sufficiently shallow,
water depth and bottom type. Other
environmental factors such as
atmospheric conditions, sun and
viewing angles, surface Winds, and
waves also influence the measure-
ment of ocean color.
Figure 10 shows surface measure-
ments of upwelling spectra from
three types of water masses and il-
lustrates the increase in energy in
the green and red regimes of the
spectrum as the transition "from
Gulf Stream to estuarine water is
made [Maul and Gordon, 1975].
Figure 11 is a computer-enhanced
Landsat image of a 140 x 140 km2
sector of the New York Bight, show-
ing suspended sediments from the
Hudson River, acid-dumping events,
water mass variations, and internal
waves, the last being visible because
of the sun glint [Apel et al, 1975].
Ocean color. The CZCS on Nim-
bus-G will image the ocean surface
and near-surface in multiple
wavelengths of visible light and
reflected and thermal infrared
radiation with 800-m spatial resolu-
tion over swath widths of 700 km
under controlled illumination condi-
tions; the observation interval will
be 1-6 days. The choice of
wavelength bands was dictated by
the requirement for making quan-
titative measurements relating to
chlorophyll and sediment concen-
trations (W. Hovis, private com-
munication, 1976).
Measurement of ocean color from
radiometric quality imagery of the
desired area in several spectral in-
tervals will perhaps allow measure-
ment, at least under certain limited
conditions, of the following features:
suspended near-surface sediment
distribution and concentration;
chlorophyll distribution and concen-
tration between perhaps 0.1 and 20
mg/m3 (W. Hovis, private com-
munication, 1976); fish stock loca-
tion via relationship to biosignifi-
cant observables [Stevenson et al.,
1973]; and pollutant distribution
and concentration [Wezernak and
Fig. 11. Image of the New York Bight
madeVith Landsat 1 on July 24, 1973.
The 'marbling' effect is due to light
winds; internal waves are visible in the
southeast section (NOAA Environmen-
tal Research Laboratories).
Thomson, 1972]. The CZCS sensor
may be used to make most of the
measurements 17VAS/4, 1975].
Surface reflectivity. By viewing
toward rather than away from the
sun it is possible to observe surface
features in the sun glitter owing to
the changes in surface reflectivity.
A variable viewing angle is required
to measure either color or reflected
sunlight; viewing upsun allows
determination of oil spills, internal
waves via surface slicks, and varia-
tions in surface roughness (Figure
11).
Table 3 summarizes the various
parameters discussed above and
lists the sensors and instruments
contributing to their determination.
The estimates of their usefulness
are given by primary (1), secondary
(2), and tertiary (3) designations.
Surface Data Collection
From Spacecraft
The United States and France
have programs in data collection
from unmanned automatic buoys,
both anchored and drifting, with
methods for data transmission
through such satellites as
SMS/GOES and Tiros-N. In addi-
tion, the United States maintains
large archives for surface-derived
oceanographic and meteorological
data. It is felt that presently
planned systems are sufficient to
meet the buoy data collection and
positioning requirements in the
next 5 years.
Integrated Global Ocean
Station System (IGOSS)
A system called IGOSS is an
evolving cooperative services
system for international exchange
of ocean data proceeding under the
auspices of the Intergovernmental
Oceanographic Commission of
Unesco [Junghans and Zachariasont
1974]. The coordination activities
needed to amalgamate the quite dis-
parate oceanic data sources, includ-
ing some of the data coming from
the spacecraft systems discussed
here, will be undertaken by IGOSS
if present plans materialize.
However, much of the spacecraft
data are experimental, and their
reliability and accuracy not yet
established, and it is not clear how
the archiving will be accomplished.
The presently recommended
method of utilizing satellite-derived
data is to become involved with the
ongoing programs as a scientific
investigator or similar role.
Summary
It has almost invariably been the
case that the introduction of a sig-
nificant new instrument technology
has yielded for the science to which
it was applied a number of un-
suspected and often highly signifi-
cant results. Such serendipitous dis-
coveries can surely be expected
from instruments as advanced as
those being orbited on ocean-look-
ing satellites. Oceanographers have
been hard put to gain the overview
of their domain required to under-
stand synoptic or planetary scale
events in the sea; for a limited but
important group of phenomena,
satellites promise to provide the
vantage point for this vision.
References
Apel, J. R., Ed., Sea Surface Topography
From Space, vol. 1 and 2, Tech. Rep.
ERL 228, Nat. Oceanic and Atmos.
Admin., Boulder, Colo., May 1972.
Apel, J. R., and H. M. Byrne, Oceanogra-
phy and the marine geoid, in Applica-
tions of Marine Geodesy, p. 59, Marine
Technology Society, Washington,
D. C, 1974.
Apel, J. R., and J. W. Siry, A synopsis of
Seasat-A scientific contributions, in
Seasat-A Scientific Contributions,
NASA, Washington, D. C, July 1974.
Apel, J. R„ H. M. Byrne, J. R. Proni, and
R. L. Charnell, Observations of oceanic
internal and surface waves from the
Earth Resources Technology satellite,
J. Gcophys. Res.. 80, 865, 1975.
Barath, F. T., and P. Gloersen, The scan-
ning multichannel microwave
radiometer, paper presented at U.S.
Annual Meeting, Int. Union of Radio
Sci., Boulder, Colo., 1975.
Bromer, R. L., W. G. Pichel, T. L. Seg-
nore, C. C. Walton, and H. S. Gohr-
band, 'Satellite-derived sea-surface
temperatures from NOAA spacecraft,
NOAA/NESS Tech. Memo., in press,
1976.
Brown, W. E„ Jr., C. Elachi, and T. W.
Thompson, Radar imaging of ocean
surface patterns, J. Gcophys. Res., 81,
2657, 1976.
Defant, A., Physical Oceanography, vol. 1,
Pergamon, New York, 1961.
Gates, W. L., and Y. Mintz, Understand-
ing Climatic Change: A Program for
Action, National Academy of Sciences,
Washington, D. C, 1975.
Gloersen, P., and V. V. Salomonson,
Satellites: New global observing tech-
niques for ice and snow, J. Glaciol., 15,
373, 1975.
Grantham, W. L., E. M. Bracalente, W. L.
Jones, J. H. Schrader, L. C. Schroeder,
and J. L. Mitchell, An operational
satellite scatterometer for wind vector
measurements over the ocean, NASA
Tech. Memo. X-72672, 1975.
Hendershott, M. C, W. H. Munk, and B.
D. Zetler, Ocean tides from Seasat-A,
in Seasat-A Scientific Contributions, p.
54, NASA, Washington, D. C, July
1974.
Junghans, R., and R. Zachariason, The
integrated global ocean station system
(IGOSSl, in Environmental Data Ser-
vice, National Oceanic and Atmo-
spheric Administration, Government
Printing Office, Washington, D. C,
July 1974.
Kaula, W. M. (Ed.), The Terrestrial En-
vironment: Solid Earth and Ocean
Physics, MIT Press, Cambridge, Mass.,
April 1970.
Koffler, R., Uses of NOAA environmen-
tal satellites to remotely sense ocean
phenomena, in Ocean '75 Conference
Record, Institute of Electrical and
Electronics Engineers and Marine
Technology Society, Washington,
D. C, 1975.
LaViolette, P. E., Remote optical sensing
in oceanography utilizing satellite sen-
sors, in Optical Aspects of Oceanogra-
phy, edited by N. G. Jerlov and E. S.
Nielsen, Academic, New York, 1974.
Marsh, J. G„ F. J. Lerch, and S. F. Vin-
cent, The geoid and free air gravity
anomalies corresponding to the Gem-4
earth gravitational model,
NASA/GSEC X-592-73-58, Feb. 1973.
Maul, G. A., and H. R. Gordon, On the
use of the Earth Resources Technology
satellite (Landsat-1) in optical
oceanography, in Remote Sensing of the
Environment, p. 95, Elsevier, New
York, 1975.
McGoogan, J. T., C. D. Leitao, and W. T.
Wells, Summary of Skylab S-193
altimeter altitude results, NASA Tech.
Memo. X-69355, Feb. 1975.
Molinari, R. L., Buoy tracking of ocean
currents, Advan. Astronaut. Sci., 30,
431, 1974.
NACOA, Third annual report to the
President and Congress, Government
Printing Office, Washington, D. C,
1974.
NASA, Announcement of opportunity
Science support for the Nimbus-G sen-
sors, NASA A.O. OA-75-1, Wash-
ington, D. C, 1975.
Pirie, D. M., and D. D. Steller, California
coastal processes study. Third ERTS 1
Symposium I, NASA Spec. Publ. 351,
1413, 1974.
Polcyn, F. C, and D. R. Lyzenga, Updat-
ing coastal and navigational charts
using ERTS-1 data, Third ERTS-1
Symposium I, NASA Spec. Publ. 351,
1333, 1974.
NASA, Announcement of opportunity:
Science support for the Nimbus-G sen-
sors, NASA A.O OA-75-1,
Washington, D. C, 1975.
Stevenson, W. H., A. J. Kemmerer, B. H.
Atwell, and P. M. Maughan, A review
of initial investigations to utilize
ERTS-1 data in determining the
availability and distribution of living
marine resources, in Third ERTS-1
Symposium I, NASA Spec. Publ. 351,
1317, 1973.
Stommel, H., The Ocean 's Role in Climate
Prediction, National Academy of Sci-
ences, Washington, D. C, 1974.
Vincent, S., W. E. Strange, and J. G.
Marsh, A detailed gravimetric geoid of
North America, the North Atlantic,
Eurasia, and Australia, NASAIGSFC
X-553-72-331, September 1972.
Walsh, E. J., Analysis of experimental
NRL radar altimeter data, Radio Sci.,
9, 711, 1974.
Wezernak, C. T., and F. J. Thomson,
Barge dumping of wastes in the New
York Bight, ERTS-1 Symposium Pro-
ceedings, NASA Doc. X-650-73-10,
142, 1972.
John R. Apel is a supervisory oceanographer
and Director of Pacific Marine Environmental
Laboratory in Seattle, Washington, a component
of the NOAA Environmental Research Laborato-
ries. He holds B.S. and M.S. degrees in theoretical
physics from the University of Maryland and a
Ph.D. in applied physics from Johns Hopkins
University. His specialties are in the physics of
fluids and in remote sensing. Apel is a consultant
to numerous government organizations, and has
played a leading role in the development of
satellites for oceanography.
624
13
2
Reprinted from: Journal of Geophysical Research, Vol. 81, No. 13, 2437-2441
High-Speed Streams and Sector Boundaries
C. Sawyer
Ocean Remote Sensing Laboratory, Atlantic Oceanographic and Meteorological Laboratories
Environmental Research Laboratories, NOAA. Miami, Florida 33149
High-speed streams in the solar wind are located with respect lo interplanetary magnetic sectors, and
t lie i r location in the sector is analyzed. The relation lo the sector taken as a whole is clearer than the rela-
tion lo sector boundaries. The streams occur preferentially near the center of the sector. Although high-
speed streams, as do sector boundaries, show a clear pattern of recurrence with solar rotation and
although they are about equally frequent in the period studied (1965-1471). there is no one-to-one
relation between them: sectors w ith no stream or w ith more than one stream are common. Longer sectors
contain more streams, showing that streams occur at a certain rate per unit lime rather than at a constant
number per sector.
Introduction
This paper describes an investigation of the relation of high-
speed streams in the solar wind to interplanetary magnetic
sectors. Earlier studies showed that the average solar wind
velocity rises after passage of a sector boundary [ Wilcox and
Ae.vv. 1965; SesseiaL, 1971]. Here we find high-speed streams
to occur most frequently near the middle of a sector. This is
true for sectors of either polarity, although the frequency of
streams organized about sector boundaries may depend on the
sense of the boundary.
D^ta on High-Spfed Streams and on
Magnetic Sectors
Intriligator [1973. 1974] presents and discusses a com-
pilation of data on high-speed streams in the solar wind based
on measurements of velocity made at earth-orbiting Vela and
sun-orbiting Pioneer spacecraft. The times of peak speed at the
spacecraft are given, along with the corotation delay time
appropriate to the beginning of the stream. These delay times
were used to estimate the times of both beginning and max-
imum at earth passage of the stream. In the appendix the error
in applying to the maximum the delay time appropriate to the
beginning is shown to be relatively small. Times of maximum
were used to identify observations of the same stream at differ-
ent spacecraft. Grouping of these observations reduced the 349
entries in Intnligator's list to descriptions of 235 separate
streams, for 215 of which interplanetary magnetic data are
available.
A list of "observed and well-defined' sector boundaries given
by Wilcox [1973] includes 51 sector boundaries in the period
July 1965 to November 1970. These are supplemented by data
from charts presented bv Wilcox and Colburn [1969. 1970,
1972]. hxcept for one period of missing data, April to August
1967. a plausible map of the sector structure from July 1965
through 1970 can be completed. This defines 195 sectors, of
which 44 are 'well-defined" in the sense that both boundaries
appear on the Wilcox list, while the remainder fit into an
evolving recurrent pattern that includes all of the well-defined
boundaries. Sectors shorter than 4 days are included only
when they belong to a recurrent sequence.
Rk lrrence of Sectors and High-Speed Streams
Both sector boundaries and maxima of high-speed streams
are plotted in Bartel's 27-day recurrence scheme in Figure 1.
Copyright © I97d b> the \mencan Geophysical Union.
This plot is strongly compressed in the vertical direction, and
departures of recurrence period from 27 days are thus exagger-
ated. Shading connects velocity maxima considered to be
members of a sequence of rotational recurrences. Open circles
have been added to emphasize gaps in these sequences, i.e.,
when an expected stream does not appear on the list. In-
triligator points out that gaps exist in the data w hen there was
no ground tracking of the spacecraft. These data gaps are not
seriously detrimental to our purpose. Even if all the gaps in
sequences represent data gaps, no more than 10% of the
streams were missed. In any case, missing streams are expected
to have no systematic effect on the main conclusions of this
study.
The identification of sequences is of course not certain, and
the reader will have to judge to what extent a different identi-
fication of sequences is possible and how it might affect the
derived recurrence period, noting that this value is unaffected
by wiggles but is determined by the mean slope.
While one can pick out streams with recurrence patterns
that match a nearby sector boundary, there are also many
stream sequences that cross from one sector to another, seem-
ing to develop quite independently of the sector pattern. Fig-
ure 2 shows the distribution of values of recurrence period for
the sector boundaries and high-speed streams shown in Figure
1. Sector boundary recurrence periods are distributed more
broadly than those of stream maxima, no doubt because
boundaries run behind or ahead of the sector center when the
sector is waxing or waning. The mean recurrence period of
high-speed streams is shorter than that of sector boundaries in
the same epoch by 0.45 day, which is more than 4 times the
variance of the mean.
Gosling [1971] compared daily velocity measurements from
close and from distant spacecraft and found that solar wind
speeds at one location are almost uncorrelated with speeds
measured at another location when the separation corresponds
to a rotation delay time of more than 4 days. He concluded
that solar wind speed cannot be successfully predicted over a
span of more than 4 days. In contrast, the recurrence of high-
speed streams in Figure I is remarkably stable and would
permit accurate advance prediction of the occurrence and time
of passage of a high-speed stream, though perhaps not of the
peak velocity. The difference between the two conclusions lies
in the fact that here we consider the high velocities, forming
the top tenth of all the days. Gosling's results, which show a
high mean predicted speed for the highest category of observed
speeds, are not in conflict with the present conclusion.
14
2438
Saw. i r: Brih Rkport
Apr 24 65
May 21
Jun 17
Jul 14
Aug 10
• Sep 06
Oct 03
Oct 30
Nov 26
Dec 23
Jan 19 66
Feb 15
Mar 14
SApr 10
May 07
Jun 03
Jun 30
Jul 27
Aug 23
Sep 19
Oct 16
Nov 12
Dec 09
Jan 05 67
Feb 01
Feb 28
Mar 27
Apr 23
May 20
Jun 16
Jul 13
Aug 09
Sep 05
Oct 02
Oct 29
Nov 25
Dec 22
Jon 18 68
Feb 14
Mar 12
Apr 08
.May 05
WJun 01
<£jun 28
Jul 25
Aug 21
Sep 17
Oct 14
Nov 10
Dec 07
Jan 03 69
Jon 30
Feb 26
Mar 25
Apr 21
Moy 18
Jun 14
Jul II
Aug 07
Sep 03
Sep 30
Oct 27
Nov 23
Dec 20
Jan 16 70
Feb 12
Mor II
Apr 07
May 04
May 31
Jun 27
Jul 24
Aug 20
Sep 16
Oct 13
Nov 09
Dec 06
Jan 02 71
Jan 29
Feb 25
Mar 24
Apr 20
May 17
Jun 13
Jul 10
Hg. 1. Twenty-seven-daN recurrence diagram of inlerplanelar) magnetic sectors and high-speed streams in the solar
wind. Battel's da> zero is listed at the right and is located at the arrow, shown on the left. Dashed lines are in the toward
sector: continuous lines are in the aw a> sector. Time of earth passage of the peak of a high-speed stream is indicated b> solid
circles, and absence of a listed stream in a sequence is indicated b> open circles. Note that the strong vertical compression of
the chart exaggerates drills due 'o periods longer or shorter than 27 da\s.
Location of Strfam in Sfctor
In order to make a more quantitative investigation ol the
location of streams in magnetic sectors. I classified each stream
according to the day in its sector that peak speed occurred.
noting also the field direction in the sector and the duration of
the sector in days. Then 1 counted the number of streams in
each category; e.g.. there are two streams with maxima on day
zero (same day as boundary) of 6-day sectors of away (away
from sun) polarity.
15
Sawyer: Brief Report
2439
CC =>
LU O
CD CD
50 r-
40 -
30 -
20 -
10 -
0
180
sector
boundaries
mean
L
_i_
_i_
22 23 24 25 26 27 28 29 30 31 32
RECURRENCE PERIOD, DAYS
Fig. 2. Frequency distribution of recurrence period values for (a)
high-speed streams and (h) sector boundaries, in sequences indicated
in Figure I .
Number oj streams in sectors of different duration. First, let
us examine the density of high-speed streams as a function of
sector length. The data for away and toward sectors both show
the same trend and are combined in Figure 3, where the
quantity (number of high-speed streams in sectors of length
/(/(number of sectors of length /) is plotted against /. If high-
speed streams were closely associated with sector boundaries,
we should expect to see one stream in each sector, since the
total number of streams and sectors is approximately equal. In
fact, the number of streams per sector increases with sector
length, in agreement with the hypothesis that streams occur at
a constant rate of 3. 1 per rotation regardless of sector length.
Occurrence frequency of streams at each sector day. From
the results of Wilcox and Ness [1965] and Ness el al. [1971] we
expect the maximum of a high-speed stream to tend to fall 2-4
days after passage of a sector boundary. Figure 4 shows the
number of maxima occurring in each day of a sector, summed
over all away sectors and separately over all toward sectors.
Because a day 0 occurs in every sector but later sector days are
less frequent (day 20 occurs only in sectors of a duration of at
least 21 days), the number of maxima decreases away from the
beginning of the sector. In order to interpret the distribution of
stream maxima among sector days we need to know the ex-
pected frequency, given the distribution of sector lengths. The
smooth continuous curves in Figures 4a and 4b show the
Fig.
"0 " 2 4 6 8 10 12 14 16 18 20 22 24 26 28
I LENGTH OF SECTOR, DAYS
3. Number of high-speed streams per sector as a function of
sector length. Streams occur at the rate of about three per rotation
regardless of sector length
1 1 1 1 1 1 1 1 1 r
+ SECTORS
ALL SECTORS
8 10 12 14 16 18 20 22 24 26 28
DAY IN SECTOR
Fig. 4. (o) The number of high-speed streams occurring on each
day of a sector for away (plus) sectors, (b) The number of high-speed
streams occurring on each day of a sector for toward (minus) sectors.
The continuous curve shows the expected number (see text) with
dashed curves at one standard deviation (square root of counted
number), (c ) All data are combined, and the quantity plotted is the
reduced number, the difference between the observed and expected
numbers, divided by the variance. Error bars show the square root of
the observed number for the peak values. Eighty percent of the points
are expected to lie between the dashed horizontal lines.
expected number as a function of sector length /: (total number
of high-speed streams/total number of sector days) X number
of sectors with duration greater than /. The general tendency
for the frequency of high-velocity streams to fall as distance
-12-10-8 -6 -4 -2
DAYS BEFORE
2 4 6 8 10
DAYS AFTER
16
Fig. 5. The reduced number of high-velocity streams is plotted (a)
for away/toward (plus/minus) sector boundaries, {b) for to-
ward/away (minus/plus) boundaries, and (c ) for all boundaries. One
third of the points are expected to lie beyond the horizontal dashed
lines.
2440
Sawv ir: Brih Ri fori
from sector beginning increases follows the expected trend.
The dashed curves differ from the expected number by plus or
minus the square root of the expected number, which we take
as an estimate of the variance. In a normal error distribution
we expect a proportion of 0.32, or about one third of the
values, lo diller by this amount or more from the mean or
expected value. Of the 21 points of figure 4a or the 22 points
of figure Ah we expect 7 to fall beyond the dashed curves and
observe 7 and 5. In figure 4c. data from away and toward
sectors have been combined. The plotted quantity is the re-
duced number:
observed number of high-speed streams - expected number
(expected number)' 2
Again, deviant points are scattered across the sector, the total
number deviating beyond the expected variance about as ex-
pected, from this analysis the high-speed streams seem to be
randomly located with respect to the leading boundary of the
sector, though we shall lind a relation to the sector itself.
Location of stream with respect to sector boundary. Next.
let us designate the last day of a sector as day —I, the day
before as day -2, etc.. so that we can examine the stream
frequency on either side of a sector boundary. In Figures 5a
and 5A is plotted the reduced number, i.e.. (observed number
-expected number)/expected variance, for away/toward and
for toward/away sector boundaries, and in Figure 5c it is
plotted for all sector boundaries. The difference between this
organization and that of the preceding section and Figure 4
can be shown by an example. A stream that falls on dav 8 of a
10-day sector appears at dav 8 in Figure 4 and at dav -2 in
Figure 5. Again, difference from a random distribution is
difficult to demonstrate. Of interest, however, is the minimum
at dav - I, which corresponds to a minimum in geomagnetic
index Kp found by Shapiro [1974] to precede aw ay/ toward
sector boundaries and to be the outstanding feature of his
analv sis.
Stream location as Jraciion oj sector. Finally, each stream
maximum was characterized as being in the first tenth, second
tenth. ■ • ■ , nlh tenth of its sector. Figure 6 show s the plots for
SECTOR
BOUNDARY
SECTOR
BOUNDARY
I -I
-
i i — l 1 1 1 1 1 1
1 r— i t— \ t
;„,
-
\ *
A
-
\ /
\ rv>
-J v\
■'/
V
,6
\>
^%>s.
/
N4
\ y
y N
i < i i > i i
1 1 1 1 1 1
0 12 3 4 5 6 7
90 12 3 4 5 6 7
STREAM POSITION IN SECTOR
(FRACTION OF SECTOR LENGTH)
Fig. 6. The reduced number of high-velocity streams is plotted for
each tenth ol a sector (a) I'lot lor away sectors, [b) Plot lor toward
sectors, (r) Plot lor all sectors, One third of the points arc expected to
kill hcvond the hon/ontal lines
away sectors, for toward sectors, and for all sectors. Although
deviations from expected values are not large, these plots show
a consistent trend through the sector, with below-average oc-
currence frequency near the beginning and end of the sector
and above-average occurrence frequency in the center of the
sector for both away and toward sectors. In the combined data
we lind 6 of 10 points falling beyond the expected variance
where 3 are expected. Although high-speed streams may fall
anywhere in a sector, they fall more frequently near the center
of the sector and least frequently near the sector boundary.
Thus we see that a clearer relation to high-speed streams
emerges when we consider the sector as a whole rather than
sector boundaries. The majority of sectors do not fit the simple-
picture of a stream in midsector, however. Of 195 sectors, only
91 (47^ (contain a single stream, 56(29^ ) have no stream, and
48 (24%) have more than one stream.
Conclusion. Although earlier studies showed the average
solar wind speed to be organized around sector boundaries,
specific high-speed streams are nearly randomly distributed
with respect to sector boundaries, with a more obvious pattern
of occurrence with respect to the whole sector. The total
number of streams is similar to the total number of sectors, but
fewer than half the sectors have just one stream. Longer sec-
tors have more streams, so that the occurrence rate of streams
is nearly constant instead of the number of streams per sector.
Individual high-speed streams show a stable and predictable
pattern of recurrence, in contrast to Gosling's conclusion
about the unpredictability of solar wind speeds in general.
Appendix: Effect of Applying Delay Time Appropriate
to Beginning Velocity to Locate
Maximum of Stream
The delay from observation of the beginning of the stream
at (he spacecraft to observation at the earth, given by In-
triligator [1973], is
u
17
where 0 is the longitudinal displacement of the spacecraft from
the earth and Ir is the radial displacement. The angular veloc-
ity of rotation of the stream is taken as 2.6934 10"" rad s '.
corresponding to a synodic rotation period of 27 days. U is the
velocity at the beginning of the stream. The mean peak veloc-
ity for all measurements is 577 km s"1. The mean number ol 50
km s ' steps of velocity increase is 3.77. giving a mean velocity
increase of 189 km s ' and mean beginning velocity of 388 km
s" '. Taking a typical value of 0. 1 Au for Sr. we find the error
from using beginning rather than peak velocity to be between
3 and 4 hours. In Figure 2. showing the distribution of
recurrence period values, the full width at half maximum is
3.0 days, so 1.5 days is#an estimate of the uncertainly in the
value of the period. This leads to an uncertainty in r of 9
hours when 0 = 90°. The median difference in time of earth
passage of streams observed at dilferenl spacecraft but deemed
to be the same stream is 0.9 day. The ditference from the mean
is half this value, or about I I hours. We conclude that the
error due to using beginning velocity rather than peak velocity
is small relative lo the errors due to uncertainty in the time
of maximum and in the appropriate rotation period.
■icknowledgments. The Editor thanks D. S. C olburn and R. Sha-
piro lor their assistance in evaluating this report
S VVVV. IK BKII I Rl I'OKI
2441
Rl I I Rl N( is
Gosling. .1 P . Variations in the solar wind speed along the earth's
orbit". Solar Phys . 17, 4W, 1971
Intnligaior. I) . High speed streams in the solar wind. Rep L AG--7.
World Data Center \ For Solar Terr Phys.. Boulder. Colo., 1973
Iruriiigator. I) . I videnee ol solar-cycle variations in the solar wind,
Astrophy* J Leu., lli/i. I 2.1-1 2b. 1474
Ness, V. V Hundhausen. and S. Bame. Observations ol the inter-
planetary medium Vela 3 .md Imp 3. 1965-1967.7 Geophvs Re\
76, 6tv4.1. |47|
Shapiro. K . Cieomagnetie activity in the \icinit\ ol" sector boundaries,
J. Geophvs Res . 7V. 289, 1474
Wilcox. J.. Solar activity and weather. Rep. 544. In si lor Plasma Res..
Stanford Uni\ . Stanford, C alii".. 1973.
Wilcox, J., and I). Colburn, Interplanetary sector structure in the
rising portion ol ihe sunspot cycle. J Geophvs. Res . 74, 23XX. 1969.
Wilcox, J., and I). Colburn. Interplanetary sector structure near the
maximum of the solar cycle, J Geophvs. Res . 75. 6366. 1470.
Wilcox. J., and 1). Colburn. Interplanetary sector structure at solar
maximum, J. Geophvs. Res., 77, 751. 1472.
Wilcox. J., and N. Ness. Quasi-stationary corolaiing structure in the
interplanetary medium. J Geophvs Res., 70. 5793. 1965.
(Received October 6. 1975:
accepted January 13, 1476.)
18
Reprinted from:
\OL SI. NO. 13
Journal of Geophysical Research, Vol. 81, No. 13,
JOURNAL OF GEOPHYSICAL RESEARCH
3
2435-2436.
MAY I. 1976
Geomagnetic Activity at the Passage of High-Speed Streams
in the Solar Wind
C. Sawyer
Ocean Remote Sensing Laboratory. Atlantic Oceanographic and Meteorological Laboratories
Environmental Research Laboratories, SO A A, Miami, Florida 33149
M. Haurwitz
Fori Collins, Colorado 8052!
The times of maximum velocity of high-speed streams in the solar w ind are used lo organize the analysis
of planetary geomagnetic activity index Ap. and this organization of the data is shown to give a clearer
pattern than the organization of the data around sector boundaries. Geomagnetic activity is highest on
the day preceding peak velocitj in the high-speed stream. The sector boundary analysis confirms the
minimum in geomagnetic activity preceding sector boundary crossing found by Shapiro ( 1974) but shows
little dependence on the sense of the boundary.
Hirshberg and Colburn [1973] discussed the mechanism of
geomagnetic disturbance that involves merging of southward-
directed interplanetary field with earth's field. They showed
that the presence of southward-directed held tends to be short-
li\ ed. lasting only about 6 hours before the vertical component
ol the interplanetary Held returns to normal. On the other
hand, geomagnetic indices Kp (planetary index I and AE (au-
roral zone substorm index) remain elevated for a day or
longer. They suggested that the disturbance-prolonging factor
may be a high-speed stream in the solar wind and showed for
the period 1965-1967 that the disturbance index AE increased
as clearly following passage of a high-speed stream as it did
following passage of an interplanetary magnetic sector bound-
ary
Patterson [1973] found geomagnetic activity to be markedly
higher in away sectors, where the magnetic field is directed
predominantly outward from the sun. than in toward sectors,
when the sectors are defined by geomagnetic diurnal variation
at high latitude, although the difference disappears in space-
observed sectors. Shapiro [1974] analyzed Kp about sector
boundaries (space-observed) and concluded that the salient
feature is a Kp minimum preceding passage of a toward/away
boundary. He emphasized the difference between away /to-
ward and toward away boundaries, suggesting that it tends
to confirm the Hirshberg-Colburn model of disturbance
initiated by held line merging and prolonged by a high-speed
stream.
A list of such high-speed streams measured from earth-
orbiting Vela and sun-orbiting Pioneer satellites has been pub-
lished and discussed by Intriligaior [1973. 1974]. The data
presented there allow determination of the time of earth pas-
sage of each observed stream and matching of observations at
various satellites to obtain a single description of each of 235
high-speed streams in the period July 1965 through June 1971.
The distribution of the observed streams in a recurrent pattern
indicates that no more than I0°t of the streams were missed
through lack of ground tracking of the satellites. The times of
earth passage of peak speed in these streams are used as zero
days in a superposed epoch analysis of the geomagnetic index
Ap. and this analysis is compared to similar analyses in w hich
Ap is organized around magnetic sector boundaries. These
include the 'well-defined' sector boundaries determined by
Wilcox [1973] as well as less certain boundaries found from
charts published by Wilcox and Colburn [1969, 1970, 1972].
The 'well-defined' boundaries are preceded and followed by at
least 4 days of consistent polarity. Additional sector bound-
aries that do not meet this criterion are included only if they
are members of a 27-day recurrent series. In Figure 1 the
average Ap at away/toward sector boundaries is compared to
that at toward/away boundaries. The bold lines show mean
Ap values for all boundaries, and the light lines show those for
well-defined boundaries. In order to estimate the significance
of departures from the mean value we need an estimate of the
variance of the distribution of values of Ap. For each of 20
months in 1972 and 1973 the variance a was computed accord-
ing to the definition
E Ap2
n Ap'
n - 1
where n is the number of days in the month and zip is the mean
for the month. A plot of a versus Ap showed that <r increases as
\p increases and allowed determination of values of o(Ap)
given below:
Ap
5
3.8
id
7.5
15
12.6
20
18.6
25
24.5
Copyright © 19"ti by the American Geophysical Union.
19
It also allowed determination of the 1% and 5% confidence
levels indicated at the right side of the figure. Shaded portions
of the curves fall beyond the 1% confidence level. These values,
computed on the assumption that the daily Ap values are
mutually independent, are not to be interpreted literally but
serve as a basis for comparing one curve to another.
None of the features in the sector boundary analysis are very
convincing in themselves, but the two most significant features
correspond to those found by Shapiro in his analysis of Kp: a
minimum in geomagnetic activity preceding entry into a to-
ward sector and a maximum at the beginning of an away
2436
Sawyer and Halrwit/: Briiv RhPORT
55 WELL-DEFINED
SECTOR BOUNDARIES
0
DAYS
Fig I. Superposed epoch analysis of geomagnetic planetary in-
dex 4p around magnetic sector boundaries with (lop) away toward
(plus minus) polarity and (bottom) toward away (minus plus)polar-
ity. The light curve is lor a smaller sample of well-defined boundaries.
Shaded portions deviate from the mean value by more than the devia-
tion corresponding to the \ni confidence level: we expect no more than
one point in 100 to deviate this much bv chance (see text)
17 -
16 -
15 -
14 -
— 13
Ap
12
II
10 -
9
8
241 SECTOR BOUNDARIES
1962 - 1970
_!_
-J_
-L.
-L.
_l_
-16
-12
8
12
_l_
16
-4 0
DAYS
Fig. 2. The pattern of Ap superposed about high-velocity streams
is stronger than the pattern about a similar number of sector bound-
aries in the same period of time. The sector boundary precedes the Ap
maximum, and the high-speed stream maximum follows it.
sector. These two features are echoed in the analysis of the
small and relatively noisy sample of well-defined sector bound-
aries. Because the similarity of the sector boundary analyses
seems more impressive than any difference, the two sector
boundary analyses are combined to give the comparison curve
in the bottom part of Figure 2. The curve in the top part of
Figure 2 shows mean Ap when organized about high-speed
streams, an organization that is obviously cleaner than that
around sector boundaries. The two analyses cover the same
period of time, with a similar number of zero days. The me-
dian value of beginning-to-maximum time for the high-speed
streams is 36 hours and that of the duration of the streams is 4
to 5 days. The Ap peak accompanies the beginning of the high-
speed stream, and Ap is decreasing by the time maximum
velocity is reached, the suggestion being that the beginning of
fhe high-speed stream would have been a wiser choice for zero
day of the analysis than the time of peak velocity. In any case,
the result is consistent with Hirshberg and Colburn's sugges-
tion that high-speed streams are important in the production
of geomagnetic variation. If we identify the Ap maximum with
that in the sector boundary analysis, supposing that geomag-
netic disturbance, high-speed solar wind stream, and sector
boundary each form a part of a single pattern, the high-speed
stream must fall 2 or 3 days after the boundary crossing.
To summarize, geomagnetic activity maximizes early in the
passage of a high-speed stream, on the day before peak speed
at the earth. The average geomagnetic response to earth cross-
ing of a sector boundary is weaker but is similar to that found
previously by Shapiro.
Acknowledgments. We are grateful to Catherine Candalena for
carrying out part of the analysis by computer.
The Editor thanks D. S. Colburn and R. Shapiro for their assistance
in evaluating this report.
References
Hirshberg. J., and D. Colburn. Geomagnetic activity at sector bound-
aries. J. Geophys. Res.. 78. 3952. 1973.
I nl r*i I iga tor. D . High-speed streams in the solar wind. Rep. UAG-27,
World Data Center A for Solar Terr. Phys.. Boulder. Colo.. 1973.
Intnligator. D., Evidence of solar-cycle variations in the solar wind.
Aslrophys. J Lett.. 188. L23-L26. 1974.
Patterson. V.. Forty years of implied interplanetary magnetic field
data related to the geomagnetic index Ap (abstract). Eos Trans
AGV. 54. 447. 1973.
Shapiro. R , Geomagnetic activity in the vicinity of sector boundaries,
J. Geophys. Res., 79. 289. 1974.
Wilcox. J., Solar activity and the weather. Rep 544. Inst, for Plasma
Res.. Stanford Univ.. Stanford, Calif, 1973.
Wilcox. J., and D. Colburn, Interplanetary sector structure in the
rising portion of the sunspot cycle, J. Geophys. Res.. 74. 2388, 1969.
•Wilcox. J., and D. Colburn, Interplanetary sector structure near the
maximum of the sunspot cycle. J. Geophys. Res.. 75, 6366. 1970.
Wilcox, J., and D. Colburn. Interplanetary sector structure at solar
maximum, J. Geophys. Res.. 77. 751, 1972.
(Received October 6, 1975;
accepted January 13, 1976.)
20
Reprinted from: Proc. of CICAR-II Symposium: °roqress in Marine Research in
the Caribbean and Adjacent Resions, Caracas, Venezuela, July 12-16, 1976,
p. 126.
-126-
INTRODUCTION
This volume contains more than one hundred abstracts
submitted to the CICAR-II Symposium Steering Committee for
consideration for oral presentation at the CICAR-II Symposium
held in Caracas, Venezuela, 12-16 July, 1976. In addition
to the abstracts of the invited papers and of those contrib-
uted papers that were accepted, we have also included the
abstracts of those contributed papers which could not be
accommodated in the limited time available for actual oral
presentations. It was the consensus of the Steering Com-
mittee that all submitted abstracts be included and that all
be presented in both Spanish and English. Although providing
translations in the second language created some problems
for both authors and editors, it was felt that the need to
make the information available to as many Caribbean scientists
and students as possible justified the extra effort.
The abstracts are arranged in eight separate groups
corresponding to the eight Sessions of the Symposium (Marine
Biology, Marine Geology and Geophysics, etc.), and within
each group they are arranged alphabetically by the last name
of the first author of each paper.
The volume was printed at the University of Miami's
Rosenstiel School of Marine and Atmospheric Science with fi-
nancial support provided by the Intergovernmental Oceanographic
Commission. Carol Wolverton, Rosemary Gutierrez, and Becky
Newell of NOAA ' s Atlantic Oceanographic and Meteorological
Laboratories in Miami, Florida, put in long hours typing
correspondence, preparing the abstracts in final form for
the printer, and in proof reading. Their contribution is
gratefully acknowledged.
Harris B. Stewart, Jr., Chairman
CICAR-II Symposium Steering
Committee, Editor
21
Reprinted from: Proc. of CICAR-I I Symposium: Progress in Marine
Research in the Caribbean and Adjacent Reqions, Caracas, Vene-
zuela, July 12-16, 1976, p. 241.
-241-
INTRODUCTION TO THE CICAR-II SYMPOSIUM
Harris B. Stewart, Jr.
National Oceanic and Atmospheric Administration
Atlantic Oceanographic and Meteorological Laboratories
Miami, Florida, U.S.A.
The first CICAR Symposium was held in Curacao in 1968.
That meeting provided a summary of the status of our knowledge
of the Caribbean and adjacent regions at that time, but more
importantly it pointed out the areas where additional knowledge
was required. Now, some eight years later, we are starting
five days of presentations of the results of our efforts to
provide some of that knowledge. But during these five days
we will also be considering what still needs to be done, and
ho v. we can best accomplish this through the continuation of
the fine regional scientific cooperation we have developed
during the CICAR Program.
The eight separate Sessions of the Symposium will be
described briefly with special attention paid to the Friday
afternoon Summary Session at which each of the Conveners will
summarize the results presented in his Session and what the
scientists involved feel are the directions future marine
scientific work in the area should take.
In addition to adding to our scientific knowledge of the
area, CICAR has established the basic mechanism for continued
cooperation in marine science among the many nations within
the region. The effectiveness of this mechanism has been
recognized by the Intergovernmental Oceanographic Commission
which is in the process of establishing the first IOC Regional
Association, and the Caribbean is the region selected for this
experiment. The degree to which such an Association is effec-
tive will depend heavily on you, the involved scientists, and
you are encouraged to put forth your ideas during the rest of
this week.
22
6
Reprinted from: U.S. Department of Commerce, NOAA/ERL/AOML-National
Oceanographic Data Center Publication, Washington, D.C. 50 p.
INTRODUCTION
The U.S. National Oceanographic Data Center as the CICAR Regional
Data Center published a series of CICAR bibliographies in 1972.
Ihese were of great utility to the scientists of many countries then
involved in the Cooperative Investigation of the Caribbean and
Adjacent Regions (CICAR) being sponsored by the Intergovernmental
Oceanographic Commission. However, these volumes could not include
references to capers published subsequently which presented the
results of work en the CICAR Program itself, the field phase of which
terminated in December of 1975.
This present preliminary bibliography is an attempt to provide in one
place a listing of the references to published papers documenting the
results of marine research by scientists, mainly from the United
States, working in the Caribbean and adjacent regions during the
1969-1975 CICAR period. Only those references provided by the
cooperating USA institutions or taken directly from the literature by
the staff of NOAA's Atlantic Oceanographic and tleteorological
Laboratories are included. Furthermore, only those references are
included which could be verified by the Environmental Science
Information Center of NOAA's Environmental Data Service. Therefore,
the listing is net corrplete, but it should prove useful to Caribbean
marine researchers, and it does document a portion. of the extensive
contributions of the United States in the CICAR Program.
Ihe final text of this bibliography has been prepared for
distribution at the CICAR-II Symposium, July 12-16, 1976, in Caracas,
Venezuela by the U.S. National Oceanographic Data Center using a
computerized text editing/f crmatt ing system. It is hoped that it
will stimulate the other cooperating CICAR nations to prepare similar
bibliographies and that United States marine scientists will pro/ide
references to their published papers that have not been included in
this preliminary listing.
Harris B. Stewart, Jr.
U.S. National Coordinator for CICAR
NCAA Atlantic Oceanographic and
Meteorological Laboratories
Virginia, Key, Miami, Florida 33149
U.S.A.
23
Reprinted from: Museum, Vol. 7, No. 11, 19-25, 44-48.
Where the Sea and Man Meet
THE COASTAL ZONE
Fishing, Recreation, Commerce, Energy,
Esthetics Must Be Considered in Plans
To Use Southeast Florida's Coastal Zone
(O
at
o.
<
By HARRIS B. STEWART JR., Ph.D.
The coastal zone has been described as the zone where the sea
and the land meet, but I submit that the problems come not be-
cause it is where the sea and the land meet, but because it is where
the sea and man meet. It is the injection of man himself into the
coastal system that creates very
real problems.
Let's look briefly at some of
the economic aspects of the wet
side of the Florida coastal zone.
Look, for example, at the val-
ue of the fish landings in the
South Florida area. The actual
value has gone up even though
the total volume of landings has
in fact gone down. This is re-
lated, as one might suspect, to
the increased price for fisheries
products.
Remember that a portion of
the life cycle of nearly all of
both the commercial and sport
fish which contribute so large-
ly to the South Florida economy
is spent in the shallow man-
grove fringes. These shallow
coastal mangrove areas are
areas of extremely high pro-
ductivity and areas where man
also has made very heavy in-
cursions. Once an area is
dredged and bulkheaded, there
Photo by Dr. Donald P. deSylva
Dr. Stewart is director of the Atlantic
Oceanographic and Meterological Labora-
tories, NOAA, on Virginia Key. The accom-
panying article is a condensation of an ad-
dress he made last fall to the Workshop on
Coastal Zone Economics.
19
24
is no chance for the fish to
spend that critical part of their
life cycle in the area anymore.
Boating More Popular
The recreational aspects of
the wet side of the coastal zone
certainly should not be neglect-
ed. Recreational boating is on
the rise. Between 1970 and 1973,
boating registrations in the
South Florida area rose 15.3 per
cent. A recent poll taken of tour-
ists coming into the south Flor-
ida area indicated that a sig-
nificantly large percentage of
tourists were attracted to the
area by the water sports avail-
able here. In 1974, the tourist
industry in southeast Florida
amounted to some two billion
dollars. Therefore, one must
keep in mind the fact that the
maintenance of good recreation-
al facilities within the south-
east Florida coastal zone is an
important aspect of the economic
development of this zone.
Commerce is a large portion
of the industrial economic ac-
tivity of southeast Florida.
Check the numbers of passen-
gers, the volume of freight and
the total monies coming into the
area as a result of our major
port areas - the Port of Miami
at Dodge Island, Port Ever-
glades at Ft. Lauderdale and
Port St. Lucie.
In addition to fishing, rec-
reation and commerce, I submit
that the esthetic value of the
coastal zone is one that should
25
CO
o.
<
not be overlooked. It may be
difficult to put cost figures on
it, but certainly an attractive
coastal area cannot be under-
rated for its economic value to
the south Florida coastal zone.
Industries and people come to
south Florida to work and to
live because it is a pleasant
place to work and live. In addi-
tion to our pleasant climate, the
major attractive force for these
people is our ocean.
Energy Problems
But we do other things with
our ocean that have economic
import. Let's take a look brief-
ly at energy. The minute you say
"energy" and "coastal zone" in
the same breath in south Flor-
ida, people immediately say
"Turkey Point," the major
power plant of Florida Power
& Light. It has had lots of
problems, but these problems
have been brought about pri-
marily because we are now liv-
ing in an era of eco-hysteria.
Because so much furor was
raised over the apparent de-
struction of approximately 60
acres of turtle grass on the bot-
tom of Biscayne Bay. the com-
pany was forced to go to some
other technique for cooling the
waters used to cool the power
plant. As a result, they con-
structed a "radiator" of 200-
foot wide cooling canals through
the mangroves south and west
of the plant. I would rather not
get embroiled in a controversy
on the relative merits of 60
"The major problem in
the coastal zone arises be-
cause it is a resource held
in common and a resource
for which many of the uses
are in mutual conflict."
acres of bay bottom versus
something over 600 acres of
mangroves.
So, the problems of the coastal
zone are not only economic ones.
They are environmental ones;
they are legal ones ; they are so-
cial ones; they are political
ones. This must be kept in mind.
There is another facet of the
energy aspect of the coastal
zone. This is the possibility that
the ocean side of the coastal
zone will in time be utilized for
the development of energy for
supplying, at least in part, the
needs of the growing southeast
Florida area. There are two sys-
tems that appear to be particu-
larly attractive at this point.
The first one is the one known
in the business as Delta-T. That
is, it is the system which uti-
lizes the temperature difference
between the warm surface wa-
ters and the cooler waters at
depth to run a closed-cycle
Rankin-type heat engine. The
other is the possible utilization
of the Florida current portion
of the Gulf Stream system as
an energy source. In addition
there is the possibility that off-
shore reactors — nuclear re-
actors or possibly fossil fueled
26
power plants — could be sited in
the waters off Florida where
the problems with cooling water
would be vastly reduced.
Offshore Wells
As long as we are talking
about energy, I feel I must say
that for the immediate future
the major energy source from
the ocean will be oil and gas
from wells drilled offshore. To
date, the northeast Gulf of Mex-
ico areas for which the oil com-
panies had such great hopes just
have not proved out. The oil
does not appear to be there. We
will worry about that one when
the possibilities look a little
brighter.
Another possible use of the
wet side of the coastal zone in
relation to the energy problem
is the potential development of
deep water ports. This involves
the establishment of what is
called a mono-buoy offshore —
pipelines running from the
mono-buoy up into the dry side
of the coastal zone to refineries
well inward from the coastal
area. To the mono-buoy would
come large tankers that could
not possibly negotiate the nar-
row channels and shallow areas
nearer to shore.
The major problem in the
coastal zone arises because it is
a resource held in common and a
resource for which many of the
uses are in mutual conflict. Per-
haps the best enunciation of this
problem is provided by Garrett
Hardin in an article written for
"If you want to use bays
like Biscayne Bay as dis-
posal sites for sewage, fine.
They are good ones; they
flush themselves twice a day
and are relatively effective
disposal mechanisms. But if
you want to use them for
that, don't plan to use them
for much else."
Science magazine several years
ago.
In the last century in Europe
it was a normal practice to have
a central common within a com-
munity. The idea here is that the
common could support a certain
number of cattle. Let us say that
the maximum sustainable yield
was 100 head of cattle on the
common. Any more than 100
cattle would mean a deteriora-
tion of the resource.
Let us now assume that there
were ten herdsmen in the town,
each of whom had ten cattle.
This meant that grass was grow-
ing just as fast as the cows could
eat it, and everything was just
fine; the common was able to
support the animals that were
feeding on it. Then one of the
herdsmen began to realize that
if he added one cow to his herd,
they would be then 1 per cent
over what the common could
maintain, but the deleterious ef-
fect would be shared by all ten
of them, and his share of this
would be only 0.1 per cent. How-
ever, the worth of his own herd
would be extended by 10 per
cent.
22
27
CO
o>
Q.
<
Photo by Dr. Donald P. deSylva
A Mangrove Thicket
Thus, even though he would
share in the depletion of the re-
source, his own net gain would
be some 9.9 per cent. So by
adding one cow to his herd, he
would gain nearly a 10 per cent
increase; so he added his one
cow. Other herdsmen saw what
he had done, and each of them
added one cow, and very soon
the grass was not able to keep
up with the munching, the whole
common disappeared, and none
ol them was able to benefit from
what originally had been a very
fine resource.
There is a lesson to be learned
from this analogy insofar as the
Florida coastal zone is con-
cerned. When there were 200
people living on Biscayne Bay,
they could dump all of their sew-
age directly into the bay and the
bay could accommodate it. But
now with several million people
living on the same bay, the pro-
blem has become acute: if you
want to have boats tied up at
your marinas with no system for
taking care of their sewage, then
you can not expect to swim in
the boat slips with impunity — at
least esthetic impunity.
If you build causeways, you
cannot sail through them.
If you bulkhead your man-
grove areas to put up condomi-
niums, you can not expect to
have your sport and commercial
fisheries as productive as they
have been in the past.
If you want to use bays like
Biscayne Bay as disposal sites
for sewage, fine. They are
good ones ; they flush themselves
twice a day and are relatively
effective disposal mechanisms.
But if you want to use them for
23
28
Photo by Dr Donald P. deSylva
A Baij Squatter
that, don't plan to use them for
much else.
Coastal Zone Program
The question then is : how do
you avoid totally destroying your
"common" and still maintain
viable uses for many of those
who would use it even though
the uses are in conflict? Its solu-
tion will, I suspect, be in large
measure a political solution. This
bothers me a little bit as a scien-
tist who is in love with the ocean
on the wet side of the coastal
zone, but again we have come to
learn to live with life as it really
is.
Through the Coastal Zone
Management Act and the inter-
action between the federal gov-
ernment and the states there
has developed a really fine pro-
gram whereby the federal gov-
ernment will provide funds for
the development of individual
state plans for management of
the coastal zone.
There is one other aspect: the
need for new ideas, particularly
exciting new concepts. Somehow,
though, "the system" just does
not seem ready for new ideas.
There seems to be an allegiance
to the status quo which I find
quite disconcerting. But let me
give you an example of what I
mean.
I can think of five specific
south Florida coastal zone prob-
lems which could be solved with
one solution : the problem of our
incredible beach erosion ; the
problem related to storm surge
associated with hurricanes; the
problem of inadequate beach
frontage ; the problem of pro-
viding adequate offshore sport
fishing areas ; and the esthetic
problem or the environmental
enhancement problem.
My contention is that all five
of these could very neatly be
solved by the establishment of
offshore islands. My proposal is
24
29
(O
Q.
<
that the southeast Florida area
look into the engineering and
economic feasibility of develop-
ing offshore islands. These is-
lands would be man-made is-
lands situated in 50-60 feet of
water.
Off Miami Beach, for example,
this would probably be in the
order of a mile or a mile and a
half offshore. These would be
linear islands, maybe a mile to a
mile and a half long, maybe a
hundred yards wide. They would
be built up from the sea floor,
and once they broke the surface
they would be covered with top
soil, and one would plant sea
grapes and palms that can sur-
vive in that fairly rugged coast-
al zone environment.
But how would these solve the
problems? Let me take the
points I made in reverse order.
First, the esthetics. As you look
out of a Miami Beach hotel in-
stead of seeing an empty ex-
panse of ocean, beautiful though
I consider that happens to be,
you would see a series of palm-
covered islands parallel to the
shore, which would help to erase
our image as spoilers of the en-
vironment and switch it to one
of improvers of the environ-
ment.
Sport Fishing
What about the sport fishing?
It is well known that the devel-
opment of offshore reefs, par-
ticularly artificial reefs wheth-
er they be made up of rocks or
old automobile tires or sunken
"My proposal is that the
southeast Florida area look
into the engineering and
economic feasibility of de-
veloping offshore islands . . .
These would be lineal is-
lands, maybe a mile to a
mile and a half long, maybe
a hundred yards wide."
vessels or serpulid worms, pro-
vides an ecological niche where
fish gradually congregate in in-
creasing numbers until you have
developed a very good sport
fishery. So, my offshore islands
would also do this.
Within the south Florida area
we have relatively few beaches
which are open to the public.
With the increasing population,
there is increasing pressure for
recreational use of beaches, and
these islands would in fact pro-
vide additional frontage for
swimming, sunning, surfing and
all the other things that people
do on beaches.
These offshore islands would
also reduce the waves and it is
on that aspect that my last two
benefits from these offshore is-
lands rest. With an approach-
ing hurricane, there is a build-
up of sea level because of the
waters being pushed shoreward
by the strong winds. The rising
water is bad enough, but the
real problem comes from the
strong storm waves on the sur-
face of these rising waters.
(Continued on Page 44)
25
30
Coastal Zone
Continued from Page 25)
Waves in a storm surge begin to
attack areas where normally
waves do not cause problems. I
am thinking in terms of the up-
per berm on beaches, hotel lob-
bies, the living rooms of beach
homes. The development of these
offshore islands would very def-
initely reduce the amount of
wave action and thus would re-
duce the damage resulting from
these hazardous waves riding on
top of a hurricane storm surge.
Politically Controversial
Perhaps the most politically
controversial aspect of these
islands relates to beach erosion.
The erosion of south Florida's
beaches results from two phe-
nomena. The first of these is the
longshore current which moves
sediment in suspension general-
ly on this coast from north to
south. The second is the waves
themselves. As waves break on
the beach, they throw sand
grains into suspension. The sand
grains then start to fall back
down to the bottom, but the
place where they fall is some-
what to the south of the place
where they were picked up be-
cause of the longshore current.
Therefore, if you can reduce the
wave action at the beach, you
will then reduce the southerly
movement of sand; that is, the
sand grains can not be thrown
into suspension by the smaller
(Continued on Page 46)
Coastal Zone
(Continued From Page 44)
waves in the lee of these off-
shore islands. What this means
is that sand migrating down the
coast with the longshore current,
once coming in the "shadow" of
these islands, will then be de-
posited. Gradually you will have
a buildup of sand on the beach
in the lee of the islands.
For something over five years
I have been trying to get the
federal government, the state
government, the Miami Tourist
Development Authority, anyone,
to sponsor a relatively inexpen-
sive engineering study of the
feasibility of offshore islands
for the south Florida area.
Maybe the idea is no good ; if
so, it should be discarded. On
the other hand it may be a good
idea, and I would hope that
someday some group could say
this is in fact worth investigat-
Limit Acreage
One possibility that should be
considered is the limiting of ac-
tual ocean front acreage to those
industries or other uses which
require their being right on the
water. If an activity can be lo-
cated equally well in West Palm
Beach or in west Dade County
or in the western part of
Broward County rather than on
the waterfront, it should be de-
nied access to the waterfront.
Virginia Key is an island be-
tween Key Biscayne and the
mainland connected to both of
31
them by causeways and bridges.
Several years ago the Miami
City Commission and the Dade
County Commission were con-
vinced that with a heavily tour-
ist-dependent economy, it was
important to develop other ac-
tivities within the area which
could lure new industry and new
dollars into the county. Thus we
were able to convince both Metro
Dade County and the City of
Miami to zone 162 acres spe-
cifically for marine research
work.
Presently the complex has the
Miami Seaquarium, Planet
Ocean of the International
Oceanographic Foundation, the
world-renowned Rosenstiel
School of Marine and Atmo-
spheric Science of the Univer-
sity of Miami, and two NOAA
Laboratories : the Southeast
Fisheries Center of NOAA's Na-
tional Marine Fisheries Service,
and my own Atlantic Oceano-
graphic and Meteorological Lab-
oratories.
Over the past year, we have
worked with Dade County, and
three acres of county land have
just been transferred to Miami-
Dade Community College for its
marine technician training pro-
gram, and five acres have gone
to a group from industry known
as Palisades Geophysical Insti-
tute which does primarily un-
derwater acoustic research work
for the navy.
In ten years I can see Vir-
32
ginia Key being the major ma- stitution itself has had to build
I'ine research area in the United a second campus several miles
States, if not the entire world, away and inland from their
Today one thinks of the Scripps main coastal lab and ship fa-
Institution of Oceanography in cility.
California and the Woods Hole The sort of thing that is hap-
Oceanographic Institution in pening with the growing marine
Massachusetts as the major science complex on Virginia Key
oceanographic research places in is not something that happens
the United States. But neither by itself. It takes concerned and
group had enough foresight to dedicated citizens willing to ap-
provide space for long term de- proach their local governments
velopment. If you wanted to lo- and to lobby, if you will, to see
cate near the Scripps Institu- that the things that have to be
tion of Oceanography today, you done are in fact accomplished,
could get probably no more In conclusion. I would like to
closer than eight or ten miles be sure that I leave with you
and be way back on the mesa. If only really one major point.
you wanted to be at Woods Hole. That is in consideration of the
there would be no chance. Even economics of the Southeast Flor-
Woods Hole Oceanographic In- ida Coastal Zone, you can not
afford to neglect the ocean. We
must con-icier it : we must take
care of it : we must utilize it ef-
fectively. But in order to do this
and to assure the continuing eco-
nomic growth of the Southeast
Florida Coastal Zone, we must
continue to consider the ocean
as the major aspect that makes
our coastal zone and its eco-
nomic and industrial develop-
ment problems considerably dif-
ferent from those of Saint Louis,
Kansas City, or Cedar Rapids.
Iowa. t->
33
8
Reprinted from: Middle Atlantic Continental Shelf and the New York Bight,
ASLO Special Symposia, Volume 2, 20-34.
Section 2
Physical
ysical processes
Physical oceanography of the Middle Atlantic Bight1,2
R. C. Beardsley
Woods Hole Oceanographic Institution, Woods Hole, Massachusetts 02543
W. C. Boicourt
Chesapeake Bay Institute, The Johns Hopkins University, Baltimore, Maryland 21218
D. V. Hansen
Atlantic Oceanographic and Meteorological Laboratories, NOAA, Miami, Florida 33419
Abstract
Kinetic energy spectra from moored current meters in the mid-Atlantic Bight reveal
marked differences in current variability between the inner shelf and the outer shelf and
slope regions. The nearshore subtidal current variability appears to be dominated by mete-
orological forcing. The amplitude of the semidiurnal and diurnal tidal peaks decreases in
the offshore direction. Shallow water records show little or no inertial energy, while at the
shelf break and over the slope, inertial motion contributes significantly to the current vari-
ance. A simple conceptual model is presented to explain how intense winter low pressure
systems ( "northeasters" ) drive strong alongshore currents which are coherent over much of
the bight. A map of "mean" currents measured in recent moored array experiments demon-
strates subsurface water flow along the shore toward the southwest. The average currents
generally increase in magnitude offshore and decrease with closeness to bottom. At most
sites, the mean current veers toward shore with increasing depth. The alongshore volume
transport measured at three transects across the bight shows surprising uniformity, consider-
ing die possible sources for discrepancy. This transport (order 2.0 X 105m3s— *) of water
within the 100-m isobath implies a mean residence time of the order % year. Much of the
shelf water observed flowing westward south of New England must originate in the Gulf of
Maine-Georges Bank area.
Before 1970, information on the circula-
tion of the mid-Atlantic Bight came mostly
from temperature and salinity measure-
ments and from drift bottles and seabed
drifters. Bigelow (1933) and Bigelow and
Sears (1935) first described seasonal tem-
1 Contribution 3701 of the Woods Hole Oceano-
graphic Institution and 228 of the Chesapeake Bay
Institute.
2 The collection of these data has been supported
by the NOAA-MESA New York Bight project, the
National Science Foundation, and the Office of
perature and salinity changes on the con-
tinental shelf, where vernal warming and
freshwater runoff build a strong stratifica-
tion which is subsequently destroyed in the
fall by storms and cooling. Iselin ( 1939,
1955) postulated an offshore motion in the
Electric and Gas Company and B. Magnell (EG&G)
contributed ideas and data. Preparation of this re-
port has been supported by the National Science
Foundation under grants DES-74-03001 (B.C.B.)
and DES-74-03913-A02 (W.C.B.) and the MESA
Naval Research. The New Jersey Public Service project (D.V.H.
AM. SOC. LIMNOL. OCEANOGR.
20
34
SPEC. SYMP. 2
Physical oceanography 21
upper layers of the shelf water and corre- momentum across it are not well under-
sponding shoreward flow in the lower layers stood.
because salinitv generally increases with
depth. He also' noted that the circulation Current variability, circulation, and
obeys the "rule of coastal circulation." water structure
whereby the average flow is parallel to the Self-contained current meters, tempera-
coast with land on the right-hand side of an hire and pressure gauges, and other instru-
observer facing downstream. Bumpus ments deployed in moored arrays in the
(1973) in his summary ol a L 0-year pro- mid- Atlantic Bight are now beginning to
gram of drift-bottle and seabed drifter re- provide records of sufficient' length to char-
leases and occasional drogue and drift-pole aetcri/e the variability of the subsurface
measurements, concluded that a mean current field in this region. We will present
alongshore flow of order 5 cm s ' occurs here some preliminary results of these field
from Cape Cod to Cape llatteras. except programs with an emphasis on describing
during periods of strong souther!) winds the "mean" circulation and subtidal current
and low runoff (Bumpus 1969). Nantucket variability.
Shoals and Diamond Shoals appeal" to be We begin by examining in Fig. 1 several
oceanographie "barriers" which limit the kinetic energy spectra computed from 1-
alongshore flow. At the Cape llatteras end. month or longer current records obtained
the alongshore How turns seaward and be- at several different sites on the middle At-
comes entrained in the Cult Stream. Oc- lantic continental shelf and rise. The four
casionally strong northeast winds drive a sites, labeled "A" through "D." and the loca-
small amount ol mid-Atlantic Bight water tion. water depth, depth at which the cur-
southward around Cape llatteras ( Bumpus rent record was taken, and other pertinent
and Pierce 1955). ml on nation for each site are given in Table
The transition /one between shell water 1. (Sites A through D correspond respee-
and warmer, saltier slope water often oc- tivelv to stations IS. 1. 4. and 11 shown in
curs during winter as a sharp inclined front lrig. 2. )
located near the shelf break. In summer, the The site A data has been taken and re-
front is less distinct but large temperature ported by KG&C (1975) under contract to
and salinity gradients still occur in the oft- the Public Service Electric and Cas Corn-
shore direction below the seasonal thermo- panv ol New Jersey. Flagg et al. (1976)
dine. These gradients are due to a band ol obtained the site B and C data. The low-
cold, low-salinity water located near the frequency cutoff for each estimated spec-
bottom on the outer shell. Described by trum is inversely proportional to the length
Bigelow (1933) as a remnant from the of the particular current record analyzed.
previous winter's cooling, these waters can The Woods Hole Oceanographic Institution
have temperatures of 6 -S C in August. has maintained moored arrays at site D for
Temperatures are around 16 C only 20 km almost a decade and the very long current
offshore of the "cold pool." The mechanisms records obtained there allowed Webster
governing the movement of this front. d (1969) and Thompson (1971) to make a
/one and exchanges of heat. salt, water, and reliable estimate of the kinetic energy spec-
Table 1. Location and other pertinent information for the current and wind spectra shown in Fig. 1.
Water
[nstr.
Sta. Nn
depth
depth
Data
Site
( Kin. 2 )
Location
Time
On )
(m )
source
A
IS
39°2S\\ 75°15\Y
Dec73-Feh74
12
5
FG&G (1975)
B
1
40°54X. 71 ()4\V
Mar 74
58
28
Flagg et al. (1976)
C
4
40°1S\. 75=51 W
Mar 74
112
30
Flags et al. (1976)
1)
1 1
39°20X, 70°0()\V
Se\ cral years
2.640
100
Webster (1969);
Thompson ( 1971 )
35
22
Physical processes
trum in the slope water over a seven-decade
range frequency. The power density of the
wind stress observed at site A is also shown
in Fig. 1. Wind stress has been computed
using the quadratic drag law t =
CD|Wio|W10 where W10 is the observed
wind vector at 10-m height and the as-
sumed constant drag coefficient is CD =
3.2 x 10- T in c.g.s. units.
The spectra have been visually smoothed
within the estimated uncertainties to sim-
plify graphical presentation and our spec-
CL
LU
>-
h;
CO
-z.
LU
O
>-
o
LT
Id
O
H
LU
10
_ 10"
10'
10"
10*
PERIOD T (hours)
I03 I02 I01 I0C
T_,— i — n — rrm — I — r1 *-
60 30 15 10 54 3 2 I 0.5 PERIOD (days)
wind stress power density at site A
inst. water
SD peak depth depth
■*- site A 5 m 12 m
♦-site B 28m 58m
♦-site C 30m 112m
♦ site D 100 m 2640m
site A
wind stress at site A •
site B
site C —
site D
Tidal frequencies
SD = semidiurnal = l/l2.4h =.081 cph
D = diurnal = l/24h = .042 cph
Inertiol frequency at 40.5°N
1=1/ 18. 5h = .054 cph
■+-
.-3
H —
10"
D I SD
10'
LU
Q
10° I
o
a.
V)
CO
UJ
rr
\-
co
o
z
10'
\&
FREQUENCY f (cph)
Fig. 1. Spectra of currents and wind in the Middle Atlantic Bight. Locations of current meter
ings are listed in Table 1. (See text for explanation of different formats.)
36
Physical oceanography
23
PERIOD T (hours'
4
10"
12
m
T3
>-
c
h-
i
(/)
13
h_
7*
o
o
I
LU
^_
**"
n
O
rr
—
F
tr
CO
UJ
O
0J
CO
6
o
CO
o>
c
LL
b
X
(->
■o
>-
CO
5
o
o
—
z
ll)
, — ,
- — .
D
O
jO
o
UJ
tr
c/>
4
Ll
c
3
rO
CM
Code
site A —
wind stress at site A ,
site B
site C
site D —
i ' i i i rrn — r
60 30 15 10 5 4 3 2
PERIOD T ( doys)
10'
I0L
i r
I 0 5
I
— 26.90(site A)
10"
■—10.75 (site B )
•3.70 (site C)
site D-
1.15 (site D)
10 10
FREQUENCY f (cph)
1 1 Tk
D I SDI0 D,B,C
\o"
Fig. 1. Continued
37
24
Physical processes
tral characterization of wind and current
variability over the continental shelf. The
reader should remember that spectra ob-
tained from much longer records will pre-
sumably show more structure than the
smoothed estimates shown in Fig. 1. The
spectra have been plotted in both the log
£(/) versus log / format (Fig. la) and the
area-preserving linear 2.3 X / x £(/) versus
log / format (Fig. lb). The first format best
displays the functional form of the energy
density E as a function of frequency, e.g.
£(/) oc l/f'" corresponds to a straight line
plotted in Fig. la with a slope of —m. The
second format in Fig. lb is used to illustrate
how much different frequency bands con-
tribute to the total variance of the current
record. The total area under the 2.3 X / x
£(/) curve is equal to the variance, and the
area under the curve between two specific
frequencies is the contribution from that
frequency range to the variance.
The spectra shown in Fig. 1 illustrate sev-
eral fundamental features of wind and cur-
rent variability on and near the mid-At-
lantic continental shelf. Wind stress and
current spectra are inherently "red," with
the power or kinetic energy density gener-
ally decreasing with increasing frequency.
Wind stress power density at site A is ap-
proximately constant at lower frequencies
with a transition occurring at periods be-
tween 2 and 4 davs and a higher frequency
falloff of about f1:>/2.
Most of the fluctuation in the wind stress
at site A is caused by rather wideband
meteorological transients which have char-
acteristic periods between 1 and 8 days.
The intense low pressure disturbances or
cyclones which generally form over the
southeast United States and intensify while
propagating up along the eastern seaboard
have characteristic periods of 2 to 4 days
and cause the peak in the site A wind stress
power density curve shown in Fig. lb.
In addition to being red at lower fre-
quencies, the four current spectra exhibit
relatively sharp peaks at the semidiurnal
(SD) and diurnal (D) frequencies. Ampli-
tude of the semidiurnal peaks generally
increases across the shelf toward shallower
water; at sites B and C, the kinetic energy
density at the semidiurnal frequency
crudely follows the relationship E oc h~3/2
as predicted by shallow-water wave theory.
The large semidiurnal peak observed at
site A is probably caused by the proximity
of Little Egg Inlet which can channel and
intensify local tidal currents. Semidiurnal
and diurnal tidal currents are weakest at
site D on the continental rise. While the
kinetic energy density at the diurnal fre-
quency shows a general increase with de-
creasing depth across the shelf, the spatial
structure of the diurnal tidal currents is not
yet understood.
It is important here to note, however, that
semidiurnal and diurnal tidal currents on
the continental shelf are in part predictable
since the astronomical forcing is determin-
istic and periodic. The accuracy of this pre-
diction depends on the basic accuracy of
the initial calibration of local tidal currents
with the astronomical forcing, the degree of
local nonlinearity (e.g. the phase shifting
of the surface tide by strong storms), and
the relative importance of baroclinic or "in-
ternal" tides, i.e. internal waves of tidal fre-
quency. We expect baroclinic effects to be
important perhaps all the time in the deeper
water near the shelf break and over most of
the shelf during the warmer months when a
strong seasonal pycnocline has formed.
Wunsch and Hendry (1972) observed bot-
tom-intensified semidiurnal tidal currents
in about 850 m of water on the New En-
gland continental slope. They described
these observations as a train of internal
waves of semidiurnal frequency generated
at greater depth on the slope and propagat-
ing up the slope toward the shelf. How far
these internal tides penetrate onto the
shelf and how much mixing is caused by
their dissipation is as yet unknown.
Tidal flow over topographic features can
also generate higher frequency internal
waves via nonlinear mechanisms. For ex-
ample, trains of large- amplitude internal
waves have been observed by remote sens-
ing to propagate almost across the shelf
during summer stratified conditions. Apel et
al. (1975) believed such wave trains are
formed near the shelf break by diurnal and
semidiurnal tidal currents. The question of
38
Physical oceanography 25
how much energy is really drained from the frequency peaks; hence the large question
barotropic tides via topographic generation mark shown in Fig. 1. It is not known at this
of internal waxes remains unanswered. time how much of the lower frequency end
Current spectra at sites C and D show an of the spectra is caused by local or regional
additional kinetic energy density peak near meteorological forcing or by the transmis-
the local inertia! frequency. The contribu- sion (or leakage) of lower frequency en-
tion of the near-inertial frequency band to ergv onto the continental shelf from the
the current variance is considerable at these deeper ocean. We have used the model of
two sites and especially so at site C near Xiiler and Kroll (in prep.) to estimate the
the shelf break (Fig. 1). The local genera- possible transmission of topographic Rossby
Hon of near-inertial currents lw meteoro- wave energy from the rise onto the shelf
logical transients has been well documented and find that this flux of energy across the
at site D by Pollard and Millard (1970); shelf break is comparable with the direct
fast moving fronts or strong veering winds kinetic energy input due to a surface wind
which rotate clockwise with near-inertial stress of 1 dyne /cm- acting over the width
frequency clearly excite nearly vertically of the continental shelf. Based on this and
propagating internal waves. The absence of other preliminary observations, we suggest
near-inertial peaks in the kinetic energy that the open ocean causes energetic low-
spectra at sites A and B nearer shore is prob- frequency motion on the outer continental
ably due to the existence of other "natural" shelf of the mid-Atlantic Bight. Longer cur-
modes like edge and shelf waxes (see Reid rent records (8 months or longer) are
1958) which are preferentially excited dm- needed to quantify accurately the impor-
ing any transient adjustment period. The tancc of low frequency energy transmission
observed lack of strong near-inertial energy onto the shell.
in shallow nearshore water should simplify Having shown that much of the current
the local current prediction problem. variability in the shallower section of the
We now turn to the lower frequency end bight is directly wind driven, we now de-
of the current spectra. Long records at site scribe a simple conceptual model for the
D show that much of the current variance dynamics of the response of this region to
at 100 m in slope water is caused bv low strong wind events. This model, suggested
frequency motion with characteristic pe- by Beardsley and Butman (1974), has been
riods centered at about 30 days. Propaga- supported by other observations (Boicourt
Hon of topographic Rossby waxes up the and Hacker 1976; Beardsley et al. in prep.),
continental rise ( perhaps generated by the Intense winter lows, the "northeasters"
Gulf Stream), meandering of the Gulf which pass to the east of the mid-Atlantic
Stream itself, and formation of antic\ clonic Bight, produce strong wind stress fields
(warm core) eddies can all generate strong toward the south and west over the shelf,
low frequency currents at site D which generally paralleling the coast from Cape
cause the spectral shape shown. In con- Cod to Cape Hatteras. The transient mass
trast with site D. kinetic density spectrum flux in the surface Ekman layer has a com-
at site A ( Fig. lb) shows that subtidal ire- ponent to the right of the wind stress vector
fluency currents in nearshore shallow water and a component parallel to the wind stress,
are strongly wind driven and cause most During northeasters, the Ekman component
of the total current variance (sec EG&G directed to the right of the wind stress is
1975). We thus suggest that the current onshore, causing sea level to rise along the
prediction problem in shallow nearshore coast. Wunsch ( 1972) and Brown et al.
water further simplifies to the development ( 1975) have shown that sea level over the
of a model which relates the subtidal cur- deep ocean (and presumably the outer
rent to measurable meteorological forcings. slope) is nearly constant over time scales of
Current records obtained at sites B and several days, so that the coastal rise in sea
C are too short for the computed spectra to level creates a large onshore pressure gradi-
indicate locations and magnitudes of lower ent that is roughly in geostrophic balance
39
26
Physical processes
with the strong alongshore flow. Since the
wind stress field tends to parallel the coast-
line, the intense northeaster generates
strong alongshore currents and cross-shelf
pressure gradients which appear to be co-
herent over the entire shelf from Cape Cod
to Cape Hatteras. Boicourt and Hacker
( 1976 ) observed that the more energetic
subtidal current fluctuations (especially
those associated with northeasters) ori-
ented along the 35-m isobath off Maryland
and Delaware are coherent and approxi-
mately in phase over distances of 230 km.
They report typical maximum daily mean
speeds of 40 cm/s at depths of 10 and 20 m,
which can produce alongshore fluid particle
excursions of 40-80 km during the several
days of the storm. Beardsley et al. (in
prep. ) found that subsurface pressure gra-
dients caused by sea level changes are co-
herent over the mid-Atlantic shelf from
Cape May to Cape Cod. These observa-
tions suggest that the wind-driven com-
ponent of the alongshore flow may be pre-
dicted from the more easily measured
wind-stress and pressure fields and coastal
sea level fluctuations.
We will now focus on the "mean" or very
low frequency current field on the mid-
Atlantic Bight. We have plotted in Fig. 2
the average currents which have been mea-
sured in recent moored array experiments.
Only records of 1 month or longer duration
have been used and information on the in-
dividual measurements (e.g. local water
depth, instrument depth, time of measure-
ment, current values, source of data, etc. )
is given in Table 2. The mean currents are
plotted as vectors with the magnitude equal
to the average speed. The same current
meter stations are numbered in Fig. 2 se-
quentially starting from the north and the
same key is used in Table 2. The depth (in
meters) of an individual measurement is
indicated in Fig. 2 by a small number lo-
cated near the head of the current vector.
We have separated the measurements into
winter ( unstratif ied ) measurements (de-
noted by solid vectors ) and summer ( strati-
fied) measurements ( dashed vectors ) . Mea-
surements from several sites (5-11) on the
continental rise and outer slope are in-
cluded to show the mean westward flow of
slope water. The mean position of the north-
ern edge of the Gulf Stream is also shown,
with the reminder that the actual position
of the Gulf Stream in this region is highly
variable ( Hansen 1970) .
These direct measurements of the mean
current field on the shelf demonstrate sub-
surface water flow along the shore toward
the southwest. The mean currents generally
increase in magnitude offshore and de-
crease with closeness to the bottom. At
most sites, the mean current veers toward
shore with increasing depth. With the ex-
ception of station 21, a net southwestward
transport is observed at all sites.
Measurements made along the three tran-
sects labeled I (New England), II (New
York), and III (Norfolk) in Fig. 2 have
been used to estimate mean alongshore
volume transport. The transects cover the
bulk of the continental shelf out to the
100-m isobath. Calculated transport values,
cross-sectional area, and mean speeds for
each transect are listed in Table 3. Although
Wright and Parker (1976) estimated that
roughly half of the volume of the shelf
water from Cape Cod to Cape Hatteras lies
in a thin surface wedge outside the 100-m
isobath, there are essentially no direct mea-
surements of mean current in the shelf
water wedge beyond the 100-m isobath.
The estimated volume transports for the
three transects are surprisingly consistent,
considering that the northern transects (I
and II ) are early spring measurements in
two different years, while the southern
transect (III) value represents summer
measurements. In addition, the transects
were made at different depths, with differ-
ent instruments, and with varying spatial
resolutions. For these reasons, we hesitate
to speculate about exchange of shelf water
and slope water based on continuity argu-
ments and assumed stationary flow through
the transects. We are uncertain, for ex-
ample, whether the higher mean alongshore
speed shown in transect III is due to a con-
tinuity of transport within the 100-m iso-
bath which forces the mean speed to in-
crease through the smaller cross-sectional
40
Physical oceanography
27
area, or whether it is clue to a more con- sistency of the transports lead us to specu-
sistent southward flow in summertime (for late that there may be little significant sea-
which there is some evidence). The con- sonal change in alongshore transport. Only
5? .J /
30 '
^ ^PE 'J
/ HATTERAS/
~$ ^" CAPE
/
76= / 75°
74o
I
73°
72°
70°
Fig. 2. Mean velocities as measured by moored current meters in the Middle Atlantic Bight region.
Winter measurements are indicated by solid arrows, summer velocities by dashed arrows. Individual sta-
tions are numbered according to Table 2; station numbers are circled. Measurement depths (in meters)
are shown near the head of the arrows.
41
28
Physical processes
Table 2. Tabulation of the recent direct measurements of sub- and near-surface mean currents shown
in Fig. 2. ( NA = not applicable. )
Sta.
No.
Location
Start
time
Record
length
( days )
Water
depth
(m)
Instr.
depth
(m)
E
(cm/s )
N
(cm/s)
Data
source*
1
40°54N/71°04W
28 Feb 74
35
58
28
57
-2.1
-0.2
-0.5
0.8
A
3
40°33N,70°56W
28 Feb 74
35
72
24
44
62
71
-5.7
-2.2
-2.2
-0.1
0.5
1.0
1.8
-0.5
A
4
40°18N,70°51W
28 Feb 74
35
110
30
50
70
109
-7.8
-7.4
-5.9
-0.8
0.7
3.0
3.3
0.0
A
7
39°23N,70°59W
20 Aug 70
46
2,527
1,504
-2.8
0.3
B
9
39°35N,70°58W
20 Aug 70
111
2,263
2,163
-6.4
0.2
B
5
39°50N,70°40W
20 Aug 70
104
876
776
-6.4
1.6
B
6a
b
39°50N,70°56W
39°50N,70°56W
20 Aug 70
20 Aug 70
45
45
45
451
86
943
993
846
933
941
880
990
-2.0
-2.0
-1.5
-4.6
-3.6
1.0
0.4
0.4
1.8
-0.7
B
8
39°37N,71°15W
20 Aug 70
111
2,150
2,052
-5.1
-0.4
B
10
39°23N,71°01W
20 Aug 70
56
2,509
2,394
-2.4
-0.5
B
2
40°45N,71°03W
8 Mar 73
33
60
42
-6.4
0.0
C
11
39°20N,70°00W
NA
NA
2,640
10
100
500
1,000
2,000
-13.0
-5.7
-3.7
-3.5
-1.6
-0.6
1.1
-0.6
0.3
-0.1
D
12
40o25N,73°28W
22 Mar 74
59
89
23
2
20
-0.7
-0.6
-3.5
-0.7
E
13
40°16N,73°13W
25 Feb 75
25 Feb 75
29 Apr 75
25 Feb 75
111
37
48
111
38
2
25
23
37
-4.7
-3.6
-5.5
-1.5
-4.8
-2.2
-3.9
-0.3
E
14
40°06N,72°54W
25 Feb 75
25 Feb 75
25 Feb 75
112
62
112
48
10
26
40
-2.4
-2.8
-1.6
-2.6
-1.1
-1.3
E
16
40°03N,72°42W
1 Mar 75
1 Mar 75
1 Mar 75
59
59
108
59
2
27
42
-2.8
-3.1
-2.9
-2.8
-2.8
-1.9
E
17
39°39N,72°38W
24 Feb 75
23 May 75
24 Feb 75
24 Feb 75
30 Apr 75
24 Feb 75
64
25
64
64
42
64
76
2
2
26
41
42
75
-2.6
-5.4
-4.9
-4.0
-6.2
-2.6
-1.7
-9.1
-0.7
-0.6
-7.4
1.0
E
18
39°28N,74°15W
1 Jul 72,73,74
1 Dec 73,74
1 Jul 72,73,74
1 Dec 73,74
60
60
60
60
12
5
5
10
10
-2.1
-2.0
-1.7
-1.3
-2.9
-2.8
-2.3
-1.8
F
15
40°07N,72°51W
18 Jun 74
18Jun74
18 Jun 74
18 Jun 74
35
50
2
13
26
46
-3.8
-6.8
-4.1
-1.7
-6.9
-5.5
-3.0
-1.8
E
42
Physical oceanography 29
Table 2. Continued
Record Water Instr.
Sta. Start length depth depth E N Data
No. Location tone (da\si ( m ) (m) (em's) (cm/s) source*
19 38049N,74o12\V 29 Oct 74 36 43 9
29 Oct 74 23
29 Oct 74 35
20 37°55\,74°39\Y 26 . Tun 74 22 35 24
21 36°50X.75°42\V 21 Jul 74 29 16 4
15
22a 36o50\",75°02\Y 21 Jul 74 37 36 9
21 Tul 71 20
21 Tul 7 1 30
b 36°50X,75002\V 15 Tan 74 29 36 7
15 Tan 74 20
15 .Tan 74 32
23 36°o0X,74°48\V 21 Jul 74 26 70 11
21 Tul 74 30
21 Jul 74 58
24 36°50\.74o40YV 21 Jul 74 18 70 76
21 Jul 7 t 104
6.2
-3.2
G
4.9
-1.0
3.0
2.8
5.7
-7.5
II
1.7
1.5
11
0.2
3.1
2.4
-8.4
II
2.6
-5.5
2.4
-1.3
2.6
-8.6
1.6
-6.9
0.7
-4.7
3.7
-10.7
H
2.3
-16.6
0
-13.8
3.1
-12.6
H
1.4
-6.6
* A-Beardsle> and KlaRg (1976): B-Schmitz ill)74>. C-Beardsle> and Butnvan (1974); D-\Vebster (1969); E-
N'OAA-MESA I in prep.); F-EGfcG i in prep.); (• Boicmirt i personal i imitation); H-Boicourt and Hacker (1976).
.simultaneous measurements will provide 100-m isobath, while 60 times the river run-
conclusive evidence, off, is only about 0.3' < of the northward
If the fluxes through the three transects transport of the Gulf Stream. The volume
are approximately the same, we postulate of the shelf water within the 100-m isobath
that there is little net flow between the is estimated to be VSi, — 6,000 km:! ( Ketchum
shelf and slope regions; Wright (1976) esti- and Keen 1955; Wright and Parker 1976)
mated that as much as 2,000 knv'Vyr might and the estimated alongshore mean flux of
leave the shelf region off New England via shell water within the 100-m isobath is
the "calving" process. This number, how- TS|, 8,000 knv'/yr. The mean residence
ever, was determined on the basis of a much time is then r = \*si,/TSi, - !4 yr. This im-
larger alongshore gradient in transport than plies that the shelf water between Cape
we observed. Hatteras and Cape Cod is removed from
The various volume fluxes for the mid- the shelf and entrained into the Gulf Stream
Atlantic Bight are shown schematically in in less than a year. This estimate is slightly
Fig. 3. For comparison, note that the along- less than the 1.3 years estimated by Ket-
shore transport of shelf water within the chum and Keen ( 1955 ) who knew the fresh-
water inflow and the salinity distribution
on the shelf and who assumed no flow
Table 3. Alongshore transport to the 100-m iso- t,nterin„ ()r leaving the bight via the Nan-
bath estimated through three transects across mid- , , P1 , , S TT ^
Atlantic Bight. Position of individual transects tucket Shoals and Cape Hatteras.
shown in Fig. 2. The Gulf of Maine and Georges Bank
region must supply the low salinity water
observed flowing westward through tran-
sect I, i.e. Tsii T,;M + T,;n following the
notation in Fig. 3. This conclusion is im-
plicit in the hydrographie structure of the
shelf water in this region, namely that the
shelf-slope water salinity front is a persis-
tent and continuous feature from New York
to the southern flank of Georges Bank and
43
Cross-
sectional
area tn
Transport
100-m
to 100-m
Mean
isobatli
isobath
speed
A
T
ii = TA
Transect
( km- )
( knv1 'yr )
( cm ;s )
Period
i
6.4
5.300
2.7
Mar 74
ii
7.6
8.800
3.7
\1
ir— Apr 75
in
3.6
8,200
7.2
J"
1-Aug 74
30
Physical processes
**->- *~1
,pCAPE TGM
C ~~~^^ y
If J
1 'GEORGES /'
BANK rJ
Tr ■—
1 , ' •
(
"%.(I>\ \ .
Job ^y
/
*
\ %
jiW
' ■
-V- --*"'
J
•J-
. *
r vN
•/
S
/
^
y
^"
r
S"
s
VF/
i !
"
JSh(ni)
/ -
Tgs
^
CAPE
HATTERA
H/
s"
/
//
y
••'■'£' '
Tr =
125 km'
/,r
/
/
TSh ■
8000 km3
/yr
VSh '
6000 km'
'y
r =
Tsh " 4
y
|TSL
S 2000 kr
nVy,
Note
60 TR;TSh;
0 3 % TGS
Fig. 3. Schematic diagram of the important vol-
ume transports for the Middle Atlantic Bight. Tr is
the total annual freshwater runoff ( of which over
50% occurs via the Chesapeake Bay), Tsh is the
alongshore transport over the shelf out to the
100-m isobath, Tsl is the net flux of slope water
into the shelf water, Tgs is the transport of the Gulf
Stream, and Tgm and Tob are the unknown fluxes
of shelf water from the Gulf of Maine and southern
flank of Georges Bank. The volume of the shelf
water mass out to the 100-m isobath is Vsh, and
the average residence time r is simply Vsh/Tsh.
the "cold pool" is also continuous during
spring and summer along this same section
of the shelf (see Bumpus 1976). Any sub-
stantial flux of more saline slope water oc-
curring across the 100-m isobath must be
balanced by an increased flux of low sa-
linity water (above TSh) from the Gulf of
Maine and Georges Bank to maintain a
steady salt balance. This conclusion is also
dictated by simple continuity arguments
which require a northern source region to
maintain the observed westward flux of
shelf water shown in Fig. 2.
The summer current measurements in
transects II and III show that the along-
shore currents in the cold pool water equals
or exceeds the mean southward current of
the surrounding warmer water. These mea-
surements counteract the traditional im-
pression that the cold pool, formed by win-
ter cooling, remains stationary throughout
the spring and summer seasons (Ketchum
and Corwin 1964). There is good evidence
(Ford et al. 1952; Boicourt 1973) that the
cold pool moves southward and is entrained
by the Gulf Stream. High alongshore veloci-
ties of the cold water, as measured in tran-
sects II and III, imply that the cold water
found near Cape Hatteras in August must
have formed by winter cooling near Cape
Cod or perhaps in the Gulf of Maine.
Two large unknowns in the calculation of
water and salt budgets in the mid-Atlantic
Bight are fluxes of water and salt into the
region from the north and amounts of water
and salt exchanged across the shelf-slope
boundary. Although we cannot yet quantify
shelf-slope exchanges, we can describe
some processes involved. In summer and
winter, much exchange appears to be wind
controlled, with onshore-offshore flows in
the upper Ekman layer compensated by
opposite flows in the lower layer ( Boicourt
and Hacker 1976). In winter the cross-shelf
flows driven by northeast winds enhance
the thermal front at the shelf break and
vertically mix the midshelf region. Winds
from the south and southwest, on the other
band, cause offshore flows in the upper
Ekman layer and intrusions of warm salty
slope water along the bottom, thereby tend-
ing to stratify the outer shelf region.
Summertime cross-shelf circulation is
larger and has a more complex vertical
structure. Boicourt (1973) and Boicourt
and Hacker (1976) found that southerly
winds can drive an intrusion of high salinity
slope water onto the shelf at middepths in
the southern mid-Atlantic Bight. Because
these intrusions have been commonly ob-
served on the outer shelf, they may be an
important process in shelf water-slope
water exchange. Gordon et al. (1976) ob-
served a high salinity layer at middepth in
the New York Bight, indicating that such
intrusions may occur widely in the bight.
The cold pool and strong thermocline are
evident in the water temperatures in the
southern mid-Atlantic Bight (Fig. 4). The
salinity distribution shows an intrusion of
high salinity slope waters in the upper ther-
44
Physical oceanography
31
2 3 4 5 6 7 70 6 84 9 94 10
Fig. 4. Distributions of temperature, salinity, and (Tt in a cross-shelf vortical section off Ocean City,
Maryland, July 1975.
mocline. This intrusion extends about 30 kin
inshore of the slielr break, apparently driven
by southerly winds. A small parcel of cold
(<8°C), low salinity water may have been
detached from the cold pool and moved off-
shore. Because such parcels are commonly
found in this position, however, the amount
of water actually detaching is uncertain.
South of New England, Bigelow ( 1933 )
and Cresswell (1967) described calving of
the cold pool with parcels or bubbles of
shelf water moving into slope water. Wright
(1976) suggested that significant inter-
change of shelf and slope water may occur
via this mechanism. This process may be
related to the formation of anticyclonic Gulf
Stream eddies and their subsequent south-
west drift along the edge of the slope. Satel-
lite infrared photographs (e.g. Hughes
1975) suggest some exchange of shallow
surface water, and Saunders' (1971) aerial
temperature survey of one warm-core eddy
suggests that some deep shelf water is
pulled off the shelf and entrained into the
trailing side of the eddy. How much shelf
water is exchanged via these processes and
45
32
Physical processes
with what frequency (i.e. the intermittency
of these processes) is not known.
We conclude this section with a brief dis-
cussion of the physical processes that gov-
ern the mean circulation in the mid-Atlantic
Bight. Stommel and Leetmaa (1972) have
constructed a theoretical model ( with linear
dynamics) for the winter shelf circulation
driven by a mean wind stress and a dis-
tributed freshwater source at the coast.
They then applied this model to the bight
and concluded that an alongshore sea level
slope of about 10 cm drop from Cape Cod
to Cape Hatteras must exist to drive the
mean flow toward the southwest (as ob-
served! ) against the mean eastward wind
stress. This same basic conclusion was also
reached by Csanady ( in prep. ) who ex-
amined the influence of wind stress vari-
ability on the Stommel and Leetmaa model.
This inferred alongshore pressure gradient
can be either created by a succession of
long, shore-trapped waves as suggested by
Csanady ( in prep. ) , who showed evidence
for this process in Lake Ontario, or main-
tained by an upstream source of fresh shelf
water, presumably here the St. Lawrence
system and inshore Labrador Current. Sut-
cliffe et al. (1976) reported evidence that
fluctuations in the transport of the St. Law-
rence system can be traced down the
Scotian Shelf and into the Gulf of Maine.
This, together with our early point that
most of the fresh shelf water observed
flowing westward through transect I (Fig.
2) must be supplied by the Gulf of Maine
and outer Georges Bank regions, suggests a
continuous freshwater pathway from the
St. Lawrence to Cape Hatteras. The along-
shore pressure gradient inferred to occur
over the mid-Atlantic Bight may be par-
tially supported by a northward rise in sea
level found by oceanic leveling in the
slope water by Sturges ( 1974) .
Special features of the New York Bight and
adjacent nearshore zone
The New York Bight contains several fea-
tures of general interest that have been in-
tensely studied. Special topographic fea-
tures of this region include a relatively
deeply incised inner shelf region into which
enters one of the major river systems of the
region and the Hudson Shelf Valley and
Hudson Canyon.
The Hudson-Baritan estuary has a char-
acteristic circulation consisting of a sea-
ward flow of relatively brackish estuarine
water in the near-surface water, and a
shoreward flow of more saline water near
the bottom. The relatively great width and
complicated channel system in the Sandy
Hook-Bockaway transect allows inertial
and Coriolis effects to further modify cur-
rents such that seaward flow tends toward
the southern side of the entrance, and the
inflow occurs mainly in the navigation
channels and along the northern side of the
entrance (see Parker et al. 1976). This
mean flow of a few centimeters per second
is a weak residual superimposed on stronger
tidal flow but causes most of the material
exchange between the estuary and shelf re-
gions.
The Hudson Shelf Valley is the offshore
expression of the Hudson estuary. Current
measurements in this valley (30 km off the
New Jersey shore ) indicate that the average
flow in the valley over intervals as long as a
month can be shoreward with an average
speed of a few kilometers per day. Such
flows are more than ample, if coherent in
space, to return suspended materials to the
harbor entrance from far out on the con-
tinental shelf.
The combination of the Hudson estuary,
the complex bottom topography, and the
nearly right-angle bend in the shoreline
produces quite complicated flow patterns
over the inner shelf. There is evidence in
the water properties that the near-surface
flow from the estuary tends to move south-
ward along the New Jersey shoreline. Be-
covery of seabed drifters suggests the sta-
tistical occurrence of a mean clockwise
circulation within the inner bight, counter
to the flow over the shelf farther offshore.
This circulation is sometimes reflected in
current measurements (Charnell and
Mayer 1975 ) , but the current regime is best
described as more dispersive than advective,
especially during spring and summer, the
seasons of maximum stratification.
46
Physical oceanography
33
An interesting and significant aspect of
tiie flow in the inner bight is a shoreward
velocity component in the bottom boundary
layer. Numerous current measurements
have been made for the NOAA-MESA pro-
ject at distances of 1-5 m above the bottom.
Averages of such measurements over any
significant time frequently show a distinctly
shoreward component. In 19 out of 21 cases
examined in which a clear distinction could
be made, there was a shoreward component
in the bottom boundary layer. Furthermore,
subdividing the data into sets in which the
flow is east or west along Long Island, for
instance, yields the same result: flow in the
bottom boundary layer is shoreward in both
cases. It is not yet ascertained whether this
shoreward veering is a result of surface
winds or whether it may be a manifestation
of estuarine circulation generally over the
shelf, but in any case it suggests a tendency
for near-bottom materials to be carried in-
shore. Such a process is a plausible explana-
tion for the relatively high and constant
rate of return of seabed drifters from bight
waters ( Charnell and Hansen 1974) and
supports previous reports ( Bumpus 1973).
Some remaining problems
Although progress has been made in de-
termining the current variability and cir-
culation pattern over the mid-Atlantic shelf,
we are still unable to provide unambiguous
answers to many questions of a basic en-
gineering sort posed by environmental man-
agers. Only a general estimate of the flush-
ing rate of the shelf is available, and critical
evaluation of the importance of the shelf-
break exchange is not yet possible. Al-
though a first-order description of flow to
be expected can now be given for main
parts of the bight, our ability to predict de-
tails and events remains poor. The domi-
nant forces controlling the circulation are
believed known but their relative impor-
tance and region of influence are not.
Neither conceptual nor observational tools
are adequate to the task for modeling of
other than tides and tidal currents. Local
models have useful applications but must
be posed very carefully (especially bound-
ary conditions) in the context of what is
and what is not known about the physics of
water movement over the shelf. It cannot be
safely assumed that the way to solve a
given management problem will be pointed
by a mathematical model in any straight-
forward sense. Finally, there remain funda-
mental questions related to smaller scale
phenomena, especially mixing and other
dissipative processes. Smaller scale topo-
graphic features like the inner New York
Bight embayment, the ridge and swale
areas, and the shelf valleys and submarine
canyons must exert some steering influ-
ences on the local flow. Some of these
smaller scale problems will be immediately
addressable when the physics of shelf cir-
culation are better known; others must
await improvement of observational instru-
ments and techniques.
References
\i>el, J. H.. II. M. Byrne, J. R. Prom, and R. L.
Charnell. 1975. Observations of oceanic
internal and surface waves from the Earth Re-
sources Technology Satellite. J. Geophys.
Res. 80: 865-881. '
Hi ariislev. R. C, and B. Butman. 1974. Cir-
culation on the New England continental
shelf: Response to strong winter storm. Geo-
phys. Res. Lett. 1: 181-184.
. and C. Flago. 1976. The water struc-
ture, mean currents, and shelf water/slope
water front on the New England continental
shelf. Proc. 1975 Liege Colloq. Ocean Hy-
drodynam., in press.
Bk.elow. 11. B. 1933. Studies of the waters on
the continental shelf. Cape Cod to Chesapeake
Ray. 1. The cycle or temperature. Pap. Phys.
Oc'eanogr. Meteorol. 2(4): 135 p.
, and M. Sears. 1935. Studies of the
waters on the continental shelf, Cape Cod to
Chesapeake Bay. 2. Salinity. Pap. Phys.
Oceanogr. Meteorol. 4(1): 94 p.
Boicouht, \V, 1973. The circulation of water on
the continental shelf from Chesapeake Bay to
Cape Hatteras. Ph.D. thesis, The Johns Hop-
kins Univ., Baltimore. 183 p.
, and P. Hacker. 1976. Circulation on
the Atlantic continental shelf of the United
States, Cape Mav to Cape Hatteras. Mem.
Soc. R. Sci. Liege Ser. 6 10: 187-200.
Brown, W„ W. Munk, F. Sxoix.rass, H. Mofjeld,
and B. Zetler. 1975. MODE bottom ex-
periment. J. Phys. Oceanogr. 5: 75-85.
Bumpus, D. F. 1969. Reversals in the surface
drift in the Middle Atlantic Bight area.
Deep-Sea Res. 16( suppl. ) : 17-23.
47
34
Physical processes
— . 1973. A description of the circulation
on the continental shelf of the east coast of
the United States. Prog. Oceanogr. 6: 1 11—
157.
— . 1976. Review of the physical oceanog-
raphy of George's Bank. Int. Comm. N. Atl.
Fish. Res. Bull. In press.
and E. L. Pierce. 1955. The hydrogra-
tal shelf south of Long Island, New York.
Limnol. Oceanogr. 9: 467-475.
and D. J. Keen. 1955. The accumula-
phy and the distribution of chaetognaths over
the continental shelf off North Carolina.
Deep-Sea Res. 3(suppl.): 92-109.
Charnell, R. L., and D. V. Hansen. 1974.
Summary and analyses of physical oceano-
graphic data collected in the New York Bight
apex during 1969 and 1970. MESA Rep. 74-3.
NOAA-ERL. 44 p.
, and D. A. Mayer. 1975. Water move-
ment within the apex of the New York Bight
during summer and fall of 1973. NOAA
Tech. Memo. ERL MESA-3. 29 p.
Cresswell, G. M. 1967. Quasi-synoptic monthly
hydrography of the transition region between
coastal and slope water south of Cape Cod,
Massachusetts. Woods Hole Oceanogr. Inst.
Tech. Rep. Ref. 67-35. 114 p.
EG&G. 1975. Summary of oceanographic obser-
vations in New Jersey coastal waters near
39°28'N latitude and 74°15'W longitude dur-
ing the period May 1973 through April 1974.
Rep. B4424, EG&G Environ. Consultants,
Waltham, Mass.
Flagg, C, J. Vermersch, and R. Beardsley.
1976. Report on the 1974 MIT New England
shelf dynamics experiment, Part 2; the moored
array. Dep. Meteorol. Mass. Inst. Techno!.,
Lab. Rep. (Unpublished.)
Ford, W. L., J. R. Longard, and R. E. Banks.
1952. On the nature, occurrence and origin
of cold low salinity water along the edge of
the Gulf Stream. J. Mar. Res. 11: 281-293.
Gordon, A. L., A. F. Amos, and R. D. Gerard.
1976. New York Bight water stratification-
October 1974. Am. Soc. Limnol. Oceanogr.
Spec. Symp. 2: 45-57.
Hansen, D. V. 1970. Gulf Stream meanders be-
tween Cape Hatteras and the Grand Banks.
Deep-Sea Res. 17: 495-511.
Hughes, P. 1975. The view from space. En-
vironmental Data Service, January Bull.,
NOAA.
Iselin, C. O'D. 1939. Some physical factors
which may influence the productivity of New
England's coastal waters. J. Mar. Res. 2:
74-85.
. 1955. Coastal currents and the fisheries.
Deep-Sea Res. 3(suppl.): 474-478.
Ketchum, B. H., and N. Corwin. 1964. The
persistence of "winter" water on the continen-
tion of river water over the continental shelf
between Cape Cod and Chesapeake Bay.
Deep-Sea Res. 3(suppl.): 346-357.
Parker, J. H., I. W. Duedall, H. B. O'Connors,
Jr., and R. E. Wilson. 1976. Raritan Bay
as a source of ammonium and chlorophyll a
for the New York Bight apex. Am. Soc. Lim-
nol. Oceanogr. Spec. Symp. 2: 212-219.
Pollard, R. T., and R. C. Millard, Jr. 1970.
Comparison between observed and simulated
wind-generated inertial oscillations. Deep-Sea
Res. 17: 813-821.
Reid, R. O. 1958. Effect of Coriolis force on
edge waves (I) investigation of the normal
modes. J. Mar. Res. 16: 109-141.
Saunders, P. M. 1971. Anticyclonic eddies
formed from shoreward meanders of Gulf
Stream. Deep-Sea Res. 18: 1207-1219.
Schmitz, W. J. 1974. Observations of low-fre-
quency current fluctuations on the continental
slope and rise near site D. J. Mar. Res. 32:
233-251.
Stommel, H., and A. Leetmaa. 1972. Circula-
tion on the continental shelf. Proc. Natl.
Acad. Sci. 69: 3380-3384.
Sturges, W. 1974. Sea level slope along conti-
nental boundaries. J. Geophys. Res. 79: 825-
830.
SUTCLIFFE, W. H., R. LOUCKS, AND K. DRINK-
water. 1976. Considerations of ocean cir-
culation and fish production on the Scotian
Shelf and in the Gulf of Maine, Pt. 1. Ocean
circulation and physical oceanography. J. Fish.
Res. Bd. Can., in press.
Thompson, R. 1971. Topographic Rossby waves
at a site north of the Gulf Stream. Deep-Sea
Res. 18: 1-19.
Webster, F. 1969. Vertical profiles of horizon-
tal ocean currents. Deep-Sea Res. 16: 85-98.
Wricht, W. R. 1976. The limits of shelf water
south of Cape Cod, 1941 to 1972. J. Mar.
Res. 34: 1-14.
, and R. Hendry. 1972. Array measure-
ments of the bottom boundary layer and the
internal wave field on the continental slope.
Geophys. Fluid Dynam. 4: 101-145.
-, and C. E. Parker. 1976. A volumetric
temperature/salinity census for the Middle At-
lantic Bight. Limnol. Oceanogr. 21: 563-
571.
Wunsch, C. 1972. Bermuda sea level in relation
to tides, weather, and baroclinic fluctuations.
Rev. Geophys. Space Phys. 10(1): 1-49.
48
9
Reprinted from: NOAA Data Report ERL MESA-18, 220 p.
ABSTRACT
During April 1974, two oceanoqraphic cruises were made by the NOAA
Ship Researcher in the New York Bight. The cruises were used for
deployment and recovery of three bottom-mounted pressure gauges and
to collect physical and chemical oceanographic data from the water
column. Thirty-one oceanographic stations were occupied on a seg-
ment of the continental shelf bounded on the east by Block Island,
on the south by Cape May, and extending outward to the edge of the
continental shelf. This report presents the corrected water column
data from these two cruises and describes the measurement methods
and corrections applied to the data.
49
10
Reprinted from: Estuarine and Coastal and Marine Science^ Vol. 4, 309-323,
A Two-dimensional Numerical Model of
Estuarine Circulation: The Effects of
Altering Depth and River Discharge
John F. Festa and Donald V. Hansen
Atlantic Oceanographic and Meteorological Laboratories,
15 Rickenbacker Causeway, Virginia Key, Miami, Fla 33149, U.S.A.
Received 22 December 1974 and in revised form 2 s June 1975
Steady-state numerical solutions are obtained for a two-dimensional,
vertically stratified model of a partially mixed estuary. The boundary at the
seaward end of the estuary is considered to be open, with the profiles of
salinity, vorticity and streamfunction obtained by extrapolating interior
dynamics out to the boundary. A salinity source is maintained at the bottom
at the mouth. Zero salt flux is required at a free-slip top and no-slip bottom
boundary. Zero salinity and a parabolic velocity profile are maintained at the
head of the estuary.
A number of cases are run for various estuarine parameters; the river
transport and Rayleigh number being the two parameters that have the most
pronounced effect. The river transport is varied by adjusting the mean
freshwater velocity, Uf. Decreasing Us allows salt as well as the stagnation
or null point to penetrate upstream. The estuarine circulation weakens, but
expands over a larger portion of the estuary. The position of the stagnation
point, with respect to the seaward boundary, varies as £/f_6/8 for Uf>i cm/s
and as U( ~5/6 for Ut<i cm/s. Increasing the Rayleigh number, by deepening
the estuarine channel, H, results in an increased circulation as well as strong
intrusion of salinity and inward migration of the stagnation point. The
horizontal location of the stagnation point is found to be proportional to Ra
and therefore, varies as H3.
Introduction
Estuaries are the part of the ocean most subject to modification, sometimes to enhance their
commercial utility, sometimes as an unanticipated side effect of engineering works elsewhere
in their watershed. One of the earliest, and still most common, modifications stems from the
attractiveness of estuaries as sites for coastal cities and seaports. Most of the major seaports
of the world consist of estuaries improved for large vessel navigation by dredging of entrance
bars and channel deepening; however, undesired side effects have sometimes accompanied
these improvements. Freshwater supplies on the Delaware River estuary are periodically
threatened. Marine borers have invaded previously inhospitable regions in San Francisco
Bay. Both of these have resulted from alteration of the salinity distribution as a consequence
of channel deepening. The growth of uiban areas on estuaries sometimes has more subtle
effects. Filling bay marshes for development results in a reduction of the tidal prism, thus
inhibiting natural water exchange and flushing processes. River flow modifications for
agriculture, flood control or hydroelectric power generation have also had unpredicted
309
50
3io J. F. Festa & D. V. Hansen
adverse downstream consequences due to alteration of the salinity distribution and circu-
lation in the river estuary. The classic example of this type of misadventure was the diver-
sion of Santee River water into the Cooper River in South Carolina. Readjustment of the
estuarine circulation to the new dynamic regime resulted almost immediately in severe
shoaling problems in Charleston Harbor. More subtle effects, such as the reduction of water
exchange within estuarine embayments because of the regulation of seasonal peaks of river
discharge, probably occur, but we know of no clear documentation of this effect.
Mathematical models have come into frequent use in estuaries, especially in application to
problems of distribution of heat, salinity and solutes. Nearly all of these models are of the
barotropic or vertically integrated numerical type. These are very useful for wave phenomena
such as tides and tidal currents, or transports of solutes for which effective dispersion co-
efficients can be determined empirically. They are not well suited for predicting responses
of circulation and salinity distributions to engineering modifications, because estuarine
dynamics depend strongly upon the vertical structure of the circulation and stratification and
their interaction. Baroclinic models in which vertical variations are retained have been
presented by Rattray & Hansen (1962) and Hansen & Rattray (1965) by the method of
similarity solutions. Although instructive as to the interpretation and generalization of
physical processes observed in estuaries, these models are of limited utility in predicting
responses to alterations. A characteristic feature of estuarine flows, the transition from
estuarine to riverine dynamics is precluded by the mathematical constraints of the similarity
approach.
In this paper we present some results from a two-dimensional, numerical model of the
gravitational circulation within estuaries. The model is used to investigate effects of channel
deepening and variation of stationary river discharge volumes on the circulation and salinity
distribution in estuaries. Turbulent transports of salt and momentum are expressed by
Fickian type diffusion coefficients. The vertical structure of circulation and stratification, and
their interaction, are retained, in contrast to the model of Harleman et al. (1974) which has
inconsistent modeling of density advection. A time-dependent version of the problem has
been modeled by Hamilton (1975), but he seems not to have run the model long enough to
ascertain that the circulation had come to equilibrium with the density field.
Governing equations
The problem considered is that of a steady state, two-dimensional, laterally homogeneous
estuary (Pritchard, 1956). The co-ordinate system is Cartesian in * and z, where z is positive
upward and x increases toward the river. A linear equation of state, p = />o(I+^)' 's
assumed and the Boussinesq approximation (Spiegel & Veronis, i960) is employed.
The horizontal and vertical momentum balances, continuity of flow and conservation of
salinity are:
ut+uux+wuz = -/?0~1Px+(^h^)x+(^vwz)z, (la)
wt+iavx+zewz = -p0"1PzMAh^x)x+(Avwz)z-/]gS, (ib)
ux+wz = o, (ic)
St +uSx+toS, = (KhSx)x+(KyS:)z. (id)
where u and w are the horizontal and vertical components of velocity respectively, P is the
hydrostatically reduced pressure, S is the salinity field, /? is the coefficient of 'salt contraction',
p0 is the density of fresh water, Ah, Ay, and Kh, KY are the horizontal and vertical exchange
coefficients of momentum and salt, respectively, and g is the gravitational acceleration.
51
Two-dimensional circulation model 311
Tidal fluctuations have been averaged out; however, the tides are considered to be the
primary source of energy for turbulent mixing. The exchange coefficients therefore repre-
sent a measure of the strength of tidal mixing. For simplicity, these coefficients are chosen to
be constant.
The vorticity and salt equations corresponding to (1) are:
rjt = -J(V, l)+AriXx+Ayr!zz—PgSx, (2a)
St = -j(ys,S)+KhSxx-{-KvSzz. (2b)
where y/ is a streamfunction with 11 = pVxj'O being a unit vector in the -\-y direction), rj =
A22y/ is the vorticity, J is the Jacobian and A22 is the two-dimensional Laplacian operator.
Non-dimensional equations, corresponding to equation (2), are obtained by scaling t and
77 by rd = H2/Kv and ra~\ respectively, x and z by H, y/ by Ky and 5 by ASh. rd is the
vertical diffusive time scale, H is the depth of the estuary, and ASh is the horizontal salinity
difference between the river and mouth of the estuary. In non-dimensional form the vorticity
and salt equations are:
It = -%¥, ri)+o{Atixx+rjzl-RaSx), (3a)
St = -J(W,S)+KSXX+SZZ, (3b)
where // = A2y/ = Vxx^Wzz, Ra ~ PgdSh H3/(AVKW) is the estuarine Rayleigh number,
a = Av/Kv is the Prandtl number, A = AhjAy and K = KJKV. Non-dimensionalizing
both horizontal and vertical distances by the estuarine depth, H, while arbitrary, forces an
aspect ratio, s = H/L, to enter only through the boundary conditions. Here, L is defined as
the computational length of the model estuary, that is, the location of the upstream bound-
ary. This length should not be confused with the dynamical length of the estuary, Ld,
roughly equivalent to the extent of salinity intrusion. The determination of the dynamical
length is a major object of analysis.
Boundary conditions
The boundary conditions to be satisfied at the river end are zero salinity and a parabolic
velocity profile (consistent with constant density and viscosity) having a transport per unit
width TT = UtH, where Ut is the vertically averaged river flow per unit width. At the
bottom boundary, a no-slip condition and zero vertical flux of salt are specified. At the top
boundary, a free-slip condition and zero vertical flux of salt are specified. These are expressed
by:
5=o, y/{z) = i-^R(z2—z3/2), and t]{z) = t,R{i—z) at x = e_1
Sz = o, if/ — o and y/z = o at z = o, (4)
Sz = o, y/ = R and 77 = 0 at z = 1,
where R = TT/Ky is the non-dimensional river transport. Inclusion of non-zero wind stress
at the surface is an easy modification to the model, but is not pursued herein. The remaining
boundary conditions to be considered are those at the mouth of the estuary, x = o. These
are perhaps the most difficult part of the model and will be discussed at some length.
Estuaries empty either into a larger bay or directly onto a continental shelf (see Figure 1).
They usually widen abruptly, allowing geometrical and rotational effects to become import-
ant. A two-dimensional model is no longer appropriate. To investigate estuarine dynamics in
its simplest form, attention must be focussed landward of this outer legion. The inshore
limit of this region is herein considered to be the mouth of the estuary.
Salinity and velocity distributions at the mouth are functionally dependent upon river
flow, depth, horizontal density difference and other parameters. The surface layers become
52
312
J. F. Festa & D. V. Hansen
Figure i. An idealized estuarine system.
fresher as the river transport increases. The estuarine circulation becomes stronger for
increasing Rayleigh numbeis. Consequently internal dynamics determine seaward boundary
profiles as well as those within the estuary. The boundary conditions given at the seaward
end of the model must be consistent with these internal dynamics. Thus, salinity and
velocity profiles cannot be specified as boundary conditions. Preliminary numerical experi-
mentations support this result, since unrealistic seaward boundary layers occur where
salinity and velocity profiles are specified as seaward boundary conditions. Experimentation
also showed that unless a source of salt in the form of a definite salinity value is specified
somewhere in the region, the solution S = o is obtained. Observations suggest that, although
the salinity distribution everywhere within estuarine regions is strongly influenced by varia-
tion of river discharge and other parameters, the salinity of the deep water near the seaward
boundary is least influenced. We have therefore made salinity at the bottom of the seaward
boundary invariant, >S(0,0) = i, to assure estuarine behavior. In order to obtain the seaward
boundary conditions, attention is focused on the dynamics near the estuarine mouth.
In the vicinity of the seaward boundary for the model it is expected that the estuarine
circulation is relatively well developed. Pritchard (1954, 1956) has shown that in this
situation the salt balance is maintained primarily by a dynamic balance between horizontal
advection and vertical diffusion of salt and a vorticity balance is maintained primarily by a
balance between buoyancy forces due to horizontal density gradients and vertical diffusion
of vorticity. Horizontal diffusion of salt, especially, while shown by Hansen & Rattray (1965)
to be essential to the overall estuarine regime, does not appear to be locally important where
the gravitational circulation is well developed. In addition, horizontal diffusion of vorticity
and horizontal shear in the vertical velocity field are also assumed to be locally unimportant.
These conditions are:
and
t]xx = o, at x = o
¥xx - o.
53
(5)
Tzvo-dimensional circulation model 313
Thus, horizontal diffusive fluxes of salt and vorticity are required to be constant, but
unspecified, at the open boundary. Although we are unable to provide a completely rigorous
justification of these conditions, they do provide a means of completing the mathematical
specification of the problem, without inducing boundary layer behavior near the seaward
boundary.
Numerical formulation and procedures
A finite-difference grid is chosen to be uniform in x and z, such that
xt = iAx, i = o, 1, / ]
Zj = jAz, j = 0, i, J (6)
t" = nAt, n — o, 1, j
where Ax = (sl)~1, Az =J~X and At is the time step whose magnitude depends upon the
stability of the differencing scheme that is chosen. The Laplacian operator is approximated
by the usual five-point difference scheme. The advection of salt and vorticity, expressed in
terms of the Jacobian, J, is approximated by using the % and J3 forms of Arakawa (1966),
respectively. These conserve salinity and salinity squared and vorticity and kinetic energy.
Diffusion is approximated by the time-centered scheme of DuFort-Frankel (1953). The
resulting finite difference analogs to (3) are:
faf ' - tffl)l(zAt) = -J\(y,,n)l(4A xAz) + aAAx~%rj1+ „ + tft_ „ - tfi+ ' - rftf «)
-aRa(S1+li-S^j)l(2Ax), (7)
(Sf ' - Slr')l{2At)= -j\{¥,S)!{±AxAz) + KAx-\S1+Xj + SUj-Sl^- S«fl)
+ Az-\S"ini+S1]_x~ S$+i- STf1). (8)
The streamfunction at the latest time step, y/"j+i, is calculated by means of a direct solution
(Buzbee et ah, 1970) to the finite difference Poisson equation,
nnu+i = ^x-^i+xi + w7-ii-2Vu+l) + **-%Vu+\ +rt±\-Wij+1)- (9)
In finite difference schemes, the size of the time increment, At, is often limited by two
stability requirements. The first of these is the advective stability condition, the Courant-
Freidricks-Lewy criterion, which requires that for a given grid spacing, S, AtVm/3<i,
where Vm represents the maximum velocity of the fluid. The other condition, the diffusive
stability criterion, requires At<S2/8k, where k is a diffusion coefficient. This condition
associated with most explicit difference schemes (Richtmyer & Morton, 1967), which would
greatly limit the size of the time step in this particular model, is eliminated by using the
DuFort-Frankel scheme. This scheme may produce spurious transient results if the time
increment is much larger than the diffusive requirements; however, transient fluctuations
are unimportant when steady-state solutions are desired. Comparisons between the DuFort-
Frankel and the usual leapfrog (time and spacially centered) difference approximations
indicate that the DuFort-Frankel scheme does produce similar steady-state solutions.
The salinity, streamfunction and vorticity fields are averaged periodically over adjacent
time steps to hasten convergence and maintain computational stability. The fields are
considered to represent steady-state solutions when the difference between the kinetic
54
314
J. F. Festa & D. V. Hansen
energy, salinity and streamfunction fields at two adjacent time steps are less than i% of the
aveiage kinetic energy, salinity and streamfunctions, respectively.
In all of the computations, the choice of a value for the model estuary length, L, or
equivalently the aspect ratio, e, depends upon the estuarine dynamics. The dynamical
length of the estuary, Ld is parameter dependent, with Ld<L. For large river flows, as in the
Columbia River, the dynamical length is small compared to thegeomorphic length classically
associated with this estuary, while for low river flows, such as that of the Delaware Estuary,
they are likely to be of the same order. The location of the seaward boundary, x = o, is
fixed by assigning a bottom salinity value and thus determining ASh. The location of the
river boundary, x = £_1 and S = o, is then adjusted for each experiment to effectively
resolve the gravitational circulation, salinity intrusion and stagnation point. Thus, L is
optimized to provide efficient use of the horizontal grid points. If L is too small, a smooth
transition to the freshwater of the river will be prevented, since the upstream boundary will
be too close to the mouth. If L is too large, there will be too few grid points to resolve the
dynamics close to the estuary mouth. Initially, L is estimated at the beginning of a calculation
and then is adjusted after a small number of iterations, if necessary.
Initial conditions
Optimum initial fields are desired to shorten the time needed to reach a steady-state solution.
In most cases the initial conditions were adapted from the similarity solution of Hansen &
Rattray (1965)
where
fJij
We
mxi
(10)
Sg(i-
-sin-
i-52?(*/-*//3),
v2te(-2#/+5s/-3*/)/48,
3*(i-*,-)>
vRa(—4zf+$Zj- 1)/8,
vRa(— 2Zj5-\-6zJi— 5Zj3)j 240
+R(-zj*+iz/-zf)18
(")
and where v is a function of Ra, R and K.
Finite difference boundary conditions
The computational grid is extended one grid point beyond the bottom, top and mouth of the
estuary. This allows the finite difference representation of derivatives at the boundary to be
consistent with interior computations. Values of S, if/ and tj outside the boundaries are
defined as
(1) Bottom (j = o)
5?,.
y/"j+l; {u = o).
Wi)-\
(2) Top(j = J)
Wu+i = 2W"j - Wv-u ("z = °)-
(3) Mouth (z = o)
S'i-\j — 2S(j — S"+\j'i(Sxx = °)>
Wi-u = 2¥ij ~ Vt+u\ (Vxx = °).
fJi-lj = 2*1 u - >/< + l/> (f!xX = o).
(12)
55
Two-dimensional circulation model 315
Solutions at the boundaries are obtained by substituting these values into the difference
equations, (7), (8) and (9). The bottom vorticity is evaluated using the first order Taylor
series approximation developed by Bryan (1963):
7?o+1 = ayii+IM*a- (13)
At the mouth of the estuary, a three point forward difference scheme for the salinity
gradient,
S"x\ ,_„ = (2AV)-1 (-355, + 4S7,. - S"2j), (14)
is used in equation (7) to calculate vorticity boundary values. We double integrate the
boundary vorticity field by means of a Gaussian-elimination procedure to calculate y/ at the
seaward end of the estuary. Given y/ on all boundaries, we invert Poisson's equation to
obtain the interior streamfunctions. Salinity and velocity profiles obtained by this method
are slightly closer to the similarity solution than are those obtained by using the central
difference representation of the salinity gradient. A third method, in which the boundary
vorticity is simply extrapolated out from the interior,
«nof1 = 2T,l+1-TI°2tl, (15)
also produces acceptable solutions.
Simple extrapolation, however, does not work well when calculating the salinity dis-
tribution at the mouth. We have found that simple smoothing produces significantly lower
values for the salinity throughout the estuary. Smaller horizontal salinity gradients occur
and as a consequence lower values of the streamfunction and velocity fields result through-
out the interior. The full salinity equation (8) must therefore be used for the calculation at
the open boundary.
Discussion of results
Steady-state solutions to the model equation were obtained using a 33X33-point finite
difference grid. Initially, a 17-point vertical resolution seemed adequate; however, tests with
similarity solutions indicated that truncation errors may produce significant differences
between analytical and numerical values of y/ and 77.
Results of computations not included here showed that variations of A between 1 and io6,
and ct from 1 to io2 have a negligible effect upon the results. Prandtl number independence
is common in thermal convection problems (Beardsley & Festa, 1972). Variation of K from 1
to io7 does produce significant change in the solutions, but this effect will not be fully
explored here.
The model was run for a range of parameters characteristic of, or centered on, nominal
values typical of coastal plain estuaries for which data have been published. These nominal
values are ASh = 30%0, K = io6 (Ky = 1 cm2/s), Ra = 3 X io9 (H = 10 m), and R =
2 X io3 (Us = 2 cm/s).
Contours of the streamfunction and salinity distribution obtained are presented in Figures
3 and 8. The salinity distribution shows a stratified intrusion into the estuary, and the
streamfunction shows the typically estuarine pattern of seaward flow of near surface water
and a landward flow of deeper water.
The vertical component of flow is of considerable interest in connection with estuarine
problems, but is not directly, or in many cases indirectly, measurable. It is of interest there-
fore, to explore the magnitude and structure of the vertical flow associated with the
56
316
J. F. Festa & D. V. Hansen
conditions and parameter ranges of the model. The longitudinal variation of vertical velocity
at three levels for H = 10 m and Uf = 2 cm/s is shown in Figure 2. The boundary conditions
require the vertical flow to be identically zero on the top and bottom boundaries. Both
the order of magnitude and the vertical structure of the vertical flow are consistent with the
determination of vertical velocity in the James River estuary made by Pritchard (1954).
The principal feature of this longitudinal variation is that at all levels a maximum occurs
somewhat seaward of the stagnation point (intersection of the internal zero of the stream-
function with the bottom) and a rapid falloff across the position of the stagnation point.
70
1 1 1 1 ! 1 1
1 1
60
-
-
50
(/I
'*' S^ \ \
£
40
: -.-"" ^^^ \ \
O
30
— \ \\
20
-
\ V
10
1 1 1 1 1 1 1 ^^-^
3a_„ 1
0
10
20
30
70
HO
90
100
40 50 60
*(km)
Figure 2. Longitudinal variations of the vertical velocity field at z = 0-25, 0-5 and
0-75 for Ut = 2 cm/s. Ra = 3X io9 (H = 10 m), <r= 10, K = io6. Stagnation
point is at x = 60 km. , z = 0-50 ; , z = 0-75 ; , z = 0-25.
The model provides a mutually consistent system of salinity, advection and diffusion fields
which can be used for investigation of a variety of kinematic problems of biological or
geological origin in estuaries. Such applications will require (empirical) determination of the
model parameters required to represent a particular estuary, and provide a vehicle for
modelling biological or geochemical processes. Our focus here, however, is on the inter-
actions of purely physical processes in estuaries.
Influence of river discharge
Effects of river discharge on the estuarine circulation and the salinity distribution are
presented for Ra = 3 x io9 (H = 10 m). The river transport parameter, R, is varied
between 5X102 and 6xio3, corresponding to a range in Ut of 0-5 to 6-o cm/s. Stream-
function and salinity distributions from this series of numerical experiments are shown in
Figure 3. Both the stratification and the strength of the estuarine circulation increase with
increased freshwater discharge. Vertical profiles of salinity and horizontal velocity, at the
seaward end of the modelled region (Figure 4), where the estuarine circulation is well
developed, are very similar to the solutions obtained by Hansen & Rattray (1965). The
principal response to changes in volume of freshwater discharge occurs in the upper half of
the water column. Velocity profiles contain a more or less invariant region near z = 0-4,
below which velocity profiles differ little. Current measurements at this level, or elsewhere in
the bottom half of the estuary, would be unable to discriminate between these profiles;
57
Two-dimensional circulation model
317
120 km
: 1 — 1 — 1 — 1 — 1 — 1— 1 — 1 — 1 — 1 — 1 — 1 — 1 — 1 — 1 — 1 — 1 — ■ — 1 — 1 — 1 — 1 — 1 — \ 1 i__i ■ 1 1 — 1 — :
'V
■ ' 1
1
\
1 1 t 1 1 1 1 1 1
Ut = 2 cm/s
1
s
\\
\
\
\
\
1
\
: >
\0-8>
I0'6 1
0-4 \
0-2 \
0-01
:
:
1 1
1 ,1 ,
, 1, J
,1,1
1
• . . li 1 1 A i._ i
, !
120 km
ry-r.
: \ \ \
:
K. N. \ \
Ui = 4 cm/s
:
^s\\ \
\
:
S\\\ ^k \
\
■
V\\\V\\\
\
\
j
\os\o-^0-4\ 0-2 '
1 0-01
!
ill .L 1. 1.1 .1.1..
60 km
Figure 3. Salinity and streamfunction fields as a function of river flow, Uc. Salinity
fields are contoured from o to 1 in intervals of o-i. The streamfunction fields are
scaled by io3 and contoured from o in intervals of 0-4. Ra = 3 X io9 (H = 10 m),
a = 10, K = io6 (Kv = 1 cm2/s).
58
3i8
J. F. Festa & D. V. Hansen
5(%„)
12 15 18 21 24 27 30
I ' ■ V
T
1
\
h \ \
\
- \
\
\
\
\
\
N^ \
\
\ \
-
\\\ "
-
X^V
-
^V
-
1
- (b)
i i i 1
-20 -15 -10 -5 0 5 10 0-3 0-4 0-5 0-6 0-7 0-8 0-9 1-0
i/(cm/s) S
Figure 4. Variations of (a) velocity and (b) salinity profiles at the seaward boundary,
x = o, with river flow, Ut. Ut (cm/s) : , 1 ; , 2 ; , 4.
nevertheless, they are quite different overall. The total landward transport into the lower
layer is almost independent of the amount of freshwater discharged. This is a somewhat
surprising result, but may be due in part to making the bottom salinity at the seaward
entrance independent of river flow. Upstream attenuation of the landward flow is consider-
ably more rapid for larger river flows.
This model varies qualitatively from the similarity solutions in that the vertical profiles are
not constrained to be similar throughout the estuary. Two features of particular interest are
the length of the salinity intrusion into the estuary and the position of the stagnation point.
The horizontal variations of salinity at the surface and at the bottom are shown in Figure 5.
The longitudinal patterns are similar to the exponential forms given by Hansen & Rattray
(1965), except that here we obtain what is not available as an analytic similarity solution: a
complete transition to zero salinity. The length of the salinity intrusion, defined by the
_! L
0 10 20 30 40 50 60 70 80 90 100
Xlkm)
Figure 5. Longitudinal variations of surface and bottom salinity as a function of
river flow, U,. Uf (cm/s): , 1 ; — , 2; , 4.
59
Two-dimensional circulation model
319
distance over which 5^o-i5%0 at the bottom of the estuary, is a strong function of fresh-
water discharge. It increases from 44 km for Uf = 4 cm/s to 170 km for U( = 0-5 cm/s as
shown in Figures 5 and 6. This behavior has been empirically, if qualitatively, known to
hydraulic engineers for many years. The functional form of the dependence of salinity
intrusion upon freshwater discharge is of special interest because a characteristic value of Uf
in coastal plain estuaries is approximately 1 cm/s. It is apparent from Figure 6 that for the
values of other parameters used, the length of salinity intrusion changes behavior in the
100 120 140
X(km)
200 220 240
Figure 6. Influence of river runoff on salinity intrusion and stagnation point
location. , S = 0-05 ; — , stagnation point; , S = 0005.
vicinity of Uf = 1 cm/s. Increases of U{ from 1 cm/s lead to modest retreat of salinity
intrusion, but reductions of Uf result in greatly increased salinity intrusion. The implica-
tion of this non-linear relationship for reduction of freshwater discharge into already critical
estuaries is fairly obvious. Within the range of values explored, the salinity intrusion varies
approximately as £/f~4/7 for C/f>i cm/s and as t/f_5/6 for Uf<i cm/s.
The location of the stagnation point is closely related to the extent of salinity intrusion as
may be seen from Figure 3. For Uf>i cm/s, the position of the stagnation point varies
approximately as C/f_5/8, and as C/f_5/6 for Uf<i cm/s. Its position is shown in Figure 5 to
be bracketed by the (dimensional) salinity values i-5%0and 0-15%,, at the bottom. This result
could be of great value to engineers, but it must be understood that the particular values
obtained here are a function of the parameters used for these model runs. This fact is
demonstrated also in Figure 6 by means of results from model runs using the same set of
parameters except for K, which was increased by a factor of 5. Establishment of a criterion
for the location of the stagnation point based on salinity intrusion, for particular estuaries in
which good estimates of the exchange coefficients are obtainable, seems possible in principle.
The exchange coefficients of course cannot be those inferred from application of a one-
dimensional model to data.
Figure 7 shows contours of the vertical velocity field at mid-depth, z = 0-5, as a function
of x, and Uf, for the range of parameters explored. Increasing river discharge results in an
increase in the magnitude of the vertical velocity as well as a seaward displacement of its
spatial maximum.
60
320
J. F. Festa & D. V. Hansen
X(km)
Figure 7. Vertical velocity contours at z = 0-5 as a function of river discharge, Us.
Vertical velocity contours are scaled by io-5 cm/s with a contour interval of 20.
Influence of depth
Effects of the depth of an estuary on the estuarine circulation and salinity distribution are
presented for a non-dimensional river discharge, R, of 2 X io3. The depth is varied from 7-5
to 12-5 m by varying the Rayleigh number, Ra, from 1-3 X io9 to 5-9 X io9 (this of course
results in a variation of Uf inversely as the depth). Streamfunctions and salinity distributions
are shown in Figure 8. Increasing the depth increases the strength of the gravitational
circulation and results in greater salt penetration. The landward and seaward flow both
increase with increasing depth, but their integral transport is constant and equal to the river
discharge. Vertical profiles of salinity and horizontal velocity at the seaward boundary are
shown in Figure 9. At the mouth, the speed of the seaward flow near the upper surface is a
weak function of the depth of the estuary, but the landward flow is increased except very near
the bottom. A current measurement within a meter of the top or the bottom would not
discriminate between these profiles, although, to be sure, boundary layer phenomena are not
well resolved here. In general, there is a striking difference between the responses of the
vertical profiles, at the mouth, to depth and to freshwater discharge. Whereas changes in
river discharge affect the horizontal velocity primarily in the top half of the water column,
depth changes influence primarily the landward transport in the bottom half of the estuary.
The salinity stratification decreases with increasing depth. This result is attributable to
the tendency for the greater horizontal current shear found in the shallower estuary to
increase stratification as described by Hansen (1964).
The general effect of changing depth on the extent of salinity intrusion is inverse to that of
river discharge. Salinity intrusion increases from 34 km for a depth of 7-5 m to 118 km for a
depth of 12-5 m, as shown by Figures 8, 10 and 11. The position of the stagnation point in
the circulation also has a similar behavior, as shown by Figures 8 and n. As might be
expected, the position of the stagnation point shows strong Rayleigh number dependence,
varying very closely as H3. Diffusive processes weaken the dependence of the salinity
intrusion on Rayleigh number however, causing the length of salinity intrusion to vary
approximately as H5'2.
Figure 12 shows contours of the vertical velocity field at z = 0-5 as a function of x and H
for the range of parameters explored. Decreasing the depth of an estuary results in an
61
Two-dimensional circulation model
321
48 km
120 km
-1 — 1 — 1—1 — r— 1 — 1 — 1 — 1 — 1 — t— ;
#=l2«5m
I
J 1 1 1 1 — 1 — 1 — *-Xi — 1 — ■ — 1 — ■ — 1 — 1 — 1 — ■ — 1 — 1 1 i_
192 km
' ■ « ■ 1 1 — • 1 — ■ — 1 — ■ — ■ — 1 — 1 — 1 — 1 — ■ — 1 — ■ — u — 1 — 1 — 1 — ■ — 1 — 1 — 1 — 1 — ■ — 1__ 1 — 1 :
Figure 8. Salinity and streamfunction fields as a function of estuarine depth, H.
Salinity fields are contoured from o to 1 in intervals of o-i. The streamfunction
fields are scaled by io3 and contoured from o in intervals of 0-4. i? = 2Xio3,
a = 10, K = io6 (Kv = 1 cm2/s).
62
322
Jf. F. Festa & D. V. Hansen
-5 0
i/(cm/s)
Figure 9. Variations of (a) velocity and (b) salinity profiles at the seaward boundary,
x = 0, with depth, H. : , 12-5 m; , 10 m; , 75 m.
1-0
t 1 1 1 1 1 1 1 1 1 > '
ft\
0-9
i\\
\\ v^
0-8
A \
, \ V
0-7
• \ \ \
, \ \ \
0-6
V^ \ \
<o 0-5
\\\\
0-4
0-3
A \ \
\ \\ \ SN
0-2
' V \\\ \ \
0-1
■ v\ W ^- \
\\^ V^ ^~v>->-_
0 10 20 30 40 50 60 70 80 90 100 110 120
* (km)
Figure 10. Longitudinal variations of surface and bottom salinity as a function of
depth, H. H: , 75 m; , 10 m; , 12-5.
15-0
12-5
3 10-0 ~
7-5 -
5-0
-
S
**
1 1
-
-
4*
-
_J 1 1
1 i_
20 40 60 80 100 120 140
*(km)
Figure 11. Influence of estuarine depth on salinity intrusion and stagnation point
location. , S1 = 0-05 ; , stagnation point; , S = 0-005.
63
Two-dimensional circulation model
323
t 10
X(Wm\
Figure 12. Vertical velocity contours ats =
Vertical velocity contours are scaled by 10"
0-5 as a function of estuary depth, H.
' cm/s with a contour interval of 20.
increase in the magnitude of the vertical velocity as well as a seaward displacement of its
spatial maximum.
The caveat regarding dependence of particular values upon the choice of values for the
exchange coefficients used in the model must also be accepted here, but the general behavior
will be unchanged.
Acknowledgements
The progress of this research has benefitted from numerous discussions and criticism of the
manuscript by Drs H. Mofjeld, A. Leetmaa and C. Thacker. The manuscript was prepared
by Ms K. Phlips.
References
Arakawa, A. 1966 Computational design for long-term numerical integration of the equations of fluid
motion. Two-dimensional incompressible flow. Part 1. Journal of Computational Physics I, 1 19-143.
Beardsley, R. C. & Festa, J. F. 1972 A numerical model of convection driven by a surface stress and
non-uniform horizontal heating. Journal of Physical Oceanography 2, 444-455.
Bryan, K. 1963 A numerical investigation of a non-linear model of a wind-driven ocean. Journal of
Atmospheric Science 20, 594-606.
Buzbee, B. L., Golub, G. H. & Nielson, C. W. 1970 On direct methods for solving Poisson's equations.
SIAM. Journal of Numerical Analysis 7, 627-656.
DuFort, E. C. & Frankel, S. P. 1953 Stability conditions in the numerical treatment of parabolic
differential equations. Mathematical Tables and other Aids to Computation 7, 135.
Hamilton, P. 1975 A numerical model of the vertical circulation of tidal estuaries and its application to
the Rotterdam Waterway. Geophysical Journal of the Royal Astronomical Society 40, 1-22.
Hansen, D. V. 1964 Salt balance and circulation in partially mixed estuaries. Proceedings from the
Conference on Estuaries, Jekyll Island.
Hansen, D. V. & Rattray, M. Jr 1965 Gravitational circulation in straits and estuaries. Journal of
Marine Research 23, 104-122.
Harleman, D. R. F., Fisher, J. S. & Thatcher, M. L. 1974 Unsteady salinity intrusion in estuaries.
Technical Bulletin No. 20, U.S. Army Corps of Engineers.
Pritchard, D. W. 1954 A study of the salt balance in a coastal plain estuary. Journal of Marine Research
I3» I33-I44-
Pritchard, D. W. 1956 The dynamic structure of a coastal plain estuary . Journal of Marine Research 15,
33-42.
Rattray, M. Jr & Hansen, D. V. 1962 A similarity solution for circulation in an estuary. Journal of
Marine Research 20, 1 21-123.
Richtmyer, R. D. & Morton, K. W. 1967 Difference Methods for Initial-value Problems. 405 pp. Inter-
science, New York.
Spiegel, E. A. & Veronis, G. i960 On the Boussinesq approximation for a compressible fluid. Astro-
physics Journal 131, 442-447.
Printed in Great Britain by Henry Ling Ltd., at the Dorset Press, Dorchester, Dorset
64
Reprinted from: Applied Optios, Vol. 15, No. 8, 1974-1979.
Reprinted from Applied Optics, Vol. 15, page 1974, August 1976
Copyright 1976 by the Optical Society of America and reprinted hy permission of the copyright owner
11
Radiative transfer: a technique for simulating the ocean in
satellite remote sensing calculations
Howard R. Gordon
A method is presented for computing the radiative transfer in the ocean-atmosphere system which does not
require detailed knowledge of the optical properties of the ocean. The calculation scheme is based on the
observation that the upwelling radiance just beneath the sea surface is approximately uniform, which
implies that the effect of the ocean can be simulated by a lambertian reflector just beneath the sea surface.
It is further shown that for aerosol concentrations up to ten times the normal concentration, the radiative
transfer in homogeneous and vertically stratified atmospheres (of the same optical thickness) is nearly iden-
tical. Examples indicating the applicability of these results to the remote sensing of ocean color from space
are discussed in detail.
Introduction
It is now well established that the upwelling light
field above the oceans can contain significant infor-
mation concerning the oceanic concentration of sedi-
ments and organic material. However, when the oceans
are viewed from satellite altitudes, the increased ra-
diance due to the intervening atmosphere has made
quantitative interpretation (in terms of oceanic prop-
erties) of the radiance observed by the satellite difficult.
We can realize the full potential of oceanic remote
sensing from space in the visible portions of the spec-
trum only if we can learn to relate the radiance which
reaches the top of the atmosphere to the optical prop-
erties of the ocean. To effect this, the radiative transfer
equation must be solved for the ocean-atmosphere
system with collimated flux incident at the top of the
atmosphere. In such calculations, the optical properties
of the ocean which must be varied are the scattering
phase function Pn(#) and the single scattering albedo
o)(j (defined as the ratio of the scattering coefficient to
the total attenuation coefficient). Furthermore, unless
the ocean is assumed to be homogeneous, the influence
of vertical structure in these properties must be con-
sidered. To describe the cloud-free atmosphere, we
must know the optical properties of the aerosols and
their variation with wavelength and altitude as well as
the ozone concentration. Considering the ocean for the
present to be homogeneous, we can relate the radiance
The author is with University of Miami, Physics Department, and
Rosenstiel School of Marine & Atmospheric Sciences. Coral Cables.
Florida 33124.
Received 20 December 1975.
at the satellite to the ocean's properties by choosing an
atmospheric model and solving the transfer equation
for several oceanic phase functions and u>o's at each
wavelength of interest. The number of separate com-
putational cases required is then the product of the
number of phase functions, the number of values of u>o,
and the number of wavelengths. Even if the multiphase
Monte Carlo method (MPMC)1 is used, the wo resolu-
tion of Gordon and Brown2 would require a number of
simulations equal to ten times the number of wave-
lengths for each atmospheric model considered. In this
paper an alternate method of computation that does not
require detailed knowledge of the ocean's optical
properties is presented.
Calculations
From the work of Plass and Kattawar3-4 on radiative
transfer in the ocean atmosphere system, it is seen that
when the solar zenith angle is small, the upwelling ra-
diance just beneath the sea surface is approximately
uniform (i.e., not strongly dependent on viewing angle)
and hence determined by the upwelling irradiance. It
is possible for remote sensing purposes to utilize this
observation in simulations of the transfer of radiation
in the ocean-atmosphere by assuming that a fraction
R of the downwelling photons just beneath the sea
surface are reflected back toward the surface with a
uniform radiance distribution, while the rest of the
downwelling photons are absorbed. The ocean is then
treated as if there were a lambertian reflecting surface
of albedo R just beneath the sea surface. In this case,
Gordon and Brown ' have shown that any radiometric
quantity Q can be written
Q = Qt + \Q,R/<\ - rR\\. (1)
where Q] is the contribution to Q from photons that
1974 APPLIED OPTICS / Vol. 15. No. 8 / August 1976
65
Table I. Three Ocean Scattering Phase Functions
0
KA
KB
AT
Cde.n)
( • 10")
(\ 10)
(V H)J)
0
1092 1
10171
9521
1
4916
4 577
4285
o
57 3.5
53 1.0
499.9
10
169.3
157.7
147.6
20
29.5
29.39
29.31
30
12.56
1 1.95
11.42
4 5
3.059
3.661
4. 1 89
60
1.092
1.57 7
1.999
i 0
0.546
0.915
1.190
90
0.34 !
0.661
0.952
1 05
0. 3 1 1
0.611
0.928
120
0. 3 1 7
0.7 32
1.094
135
0.410
0.8 29
1.309
150
0.4 92
1.017
1.618
165
0.5 79
1.261
1.856
ISO
0.617
1.357
1.999
never penetrate the sea surface (hut may he specularly
reflected from the surface). Q> is the contribution to Q
from photons that interact with the hypothetical 1am-
hertian surface once for the case R = 1, and r is the ratio
of the number of photons interacting with the lamber-
tian surface twice to the number of interacting once,
again fori? = 1. I 'sing Eq. ( 1 ) any radiometric quantity
can then be computed as a function of R. Physically the
quantity R is the ratio of upwelling to downwelling ir-
radiance just beneath the sea surface and is known as
the reflectance function'5 [/?(().— )| in the ocean optics
literature. Spectral measurements of the reflectance
function i?(,\) have been presented for various oceanic
areas by Tyler and Smith.' Henceforth in this paper
R(\) will be referred to as the ocean color spectrum.
A series of Monte Carlo computations has been car-
ried out to see if an approximate simulation ( ASl ) using
this assumption of uniform upwelling radiance beneath
the sea surface yields results that agree with computa-
tions carried out using an exact simulation (ES) in
which the photons are accurately followed in the ocean
as well as the atmosphere. The Monte Carlo codes used
in Rets. 2 and 5 were modified by the addition of an
atmosphere. The atmosphere consisted of fifty layers
and included the effects of aerosols, ozone, and Rayleigh
scattering, using data taken from the work of Klterman.8
The aerosol scattering phase functions were computed
by Fraser9 from Mie theory assuming an index of re-
traction of 1.5 and DeirmendjianV" haze C size distri-
bution. Also, to determine the extent to which the
vertical structure of the atmosphere influences the ap-
proximate simulation, a second approximate simulation
(AS2) was carried out in which the atmosphere was
considered to be homogeneous, i.e., the aerosol scat-
tering, Rayleigh scattering, and ozone absorption were
independent of altitude. The oceanic phase functions
in the ES are based on Kullenberg's11 observations in
the Sargasso Sea and are given in Table I. KA is
roughly an average of Kullenberg's phase function at
632.8 nm and 655 nm, and KC is his phase function at
460 nm. KB is an average of KA and KC. These phase
functions show considerably less scattering at verv small
angles 0 < 1° than observed by Petzold12 in other clear
water areas; however, the exact form of the oceanic
phase function is not very important, since it has been
shown '•'' to influence the diffuse reflectance and /?(0,— )
only through the backscattering probability (B)
R= 2;
£
P„(tf) sinHdf).
In all the computations reported here, the solar beam
is incident on the top of the atmosphere from the zenith
with unit flux. At visible wavelengths, the variable
atmospheric constituent that will most strongly influ-
ence the radiance at the top of the atmosphere is the
aerosol concentration. (Plass and Kattawar14 have
shown that for the range of expected variation of the
ozone concentration, the radiance at 700 nm is essen-
tially independent of the concentration.) Thus, we
have carried out computations for aerosol concentra-
tions in each layer of one, three, and ten times the nor-
mal concentration given by Elterman. These aerosol
models are henceforth labeled N, 'A X N and 10 X N.
All computations presented here are carried out at 400
nm, the wavelength in the visible portion of the spec-
trum where the atmospheric effects are expected to be
most severe.
Results
A sample of the results is given in Table II where the
upward flux at the top of the atmosphere for the AS
cases is compared with the ES case for oceanic phase
function KC and ujo = 0.8. The values of i? used to ef-
fect the AC computations were taken from the EC
computation of this quantity; however, if R is taken
from
H = 0.0(H)] + 0.3244.x + 0.1425*'- + 0.1308* ■'.
(2)
where x = u.-nft/[l - u.-o(] - B)\ which, according to
Cordon rt o/.,1, reproduces the in-water reflection
function for the corresponding case, but with no at mo-
sphere present, the results of the AS computations agree
with those listed to within 0.2%. The numbers in the
parenthesis next to each flux value represent the sta-
tistical error in the flux based on the actual number of
photons collected in each case. It is seen that ES and
AS simulations generally agree to within the accuracy
of the computations. Notice also the excellent agree-
ment between the ASl and AS2 fluxes.
In Fig. 1 the comparison between the ES, ASl, and
AS2 upward radiances at the top of the atmosphere is
presented for the three aerosol models. The steplike
curve in the figure is for ES, the solid circles for ASl,
and the open circles for AS2, and n is the cosine of the
Table II. Comparison of the Flux at the Top of the
Atmosphere for the ES, ASl, and AS2 Simulations
Aerosol
concen-
tration
Model
ES
ASl
AS2
.V
0.
22
2(
•0.002)
0.
224 (
'0.001)
0.
226 0
0.001)
x N
0.
27
1 (
•0.003)
0.
273 (
■0.001)
0.
275 (
0.001)
■ .V
0.
42
3(
•0.004)
0.
126 (
• 0.002)
0.
4 25 (
0.002)
August 1976 / Vol. 15. No. 8 / APPLIED OPTICS 1975
66
015-
010
<t 005
000
00
Fig. 1. Comparison between ES (steplike curve), ASl (solid circles),
and AS2 (open circles) upward radiances at the top of the atmosphere
for an ocean with wo = 0.8 and phase function KC and an atmosphere
with a normal (1 X N), three times normal (3 X A/), and ten times
normal (10 X N) aerosol concentration.
angle between the nadir and the direction toward which
the sensor is viewing. The radiances in Fig. 1 for the ES
cases are accurate to about 3% in the ^ = 1 to about 0.4
range, while for the AS cases the accuracy is about 1%.
It is seen that again, to within the accuracy of the com-
putations, the three simulations agree for all the aerosol
concentrations except within the ft = 0 to about 0.3
range, i.e., viewing near the horizon. It is felt that these
computations demonstrate that the transfer of the
ocean color spectrum through the atmosphere can be
studied with either the ASl or AS2 model as long as
radiances close to the horizon are not of interest.
Furthermore, from the reciprocity principle, 16 the nadir
radiance, when the solar beam makes an angle An with
the zenith, can be found by multiplying the radiance
/(m) in Fig. 1 by n where n is taken to be cosflo- This
implies that as long as the sun is not too near the hori-
zon, the ASl and AS2 methods of computation can be
used to determine the nadir radiance at the top of the
atmosphere as a function of the ocean's properties
through Eq. (2). The fact that the AS2 model (homo-
geneous atmosphere) yields accurate radiances is very
important in remote sensing since it implies that only
the total concentration (or equivalently the total optical
thickness) of the aerosol need be determined to recover
the ocean color spectrum from satellite spectral radio-
metric data. In principle, as suggested by Curran,1 ' this
can be accomplished by observing the ocean (assumed
free of white caps) in the near ir where R(\) s 0.
It should be pointed out that these results also
strongly suggest that R{\) is the quantity relating the
subsurface conditions that can be quantitatively de-
termined from space and hence is the most natural
definition of the ocean color spectrum. Moreover, it has
been shownls that R(\) is not a strong function of the
solar zenith angle [the maximum variation in R(0,— )
with 0O is of the order of 15% for 0 < 90 ^ 60°] in contrast
with other definitions.1718 Of course any attempt to
relate satellite measurements to subsurface conditions
must be through an understanding of how the concen-
trations of suspended material, dissolved organic ma-
terial, chlorophyll, etc. influence R(X) or more funda-
mentally influence u>o(M and R(X).
Application: Minimum Detectable Change in R
As an example of the application of ASl to oceanic
remote sensing, we compute the minimum change AR
in R at 400 nm, which can be detected with a sensor of
given sensitivity, or conversely specify the sensor sen-
sitivity required to detect a given change in R.
Applying Eq. ( 1 ) to the radiance I(^) at the top of the
atmosphere with the sun at the zenith, we have
l(n) = h(n) + \[RI2(n)]/d -rR)\. (3)
/i(m) and I-z(n) are presented in Figs. 2 and 3, respec-
tively, for the three aerosol models discussed above as
well as an aerosol-free model (0 X N) and a model with
seven times the normal aerosol concentration (7 X N).
Now I\in) and 1 2(11) depend only on the direction of the
incident solar beam, the properties of the atmosphere,
and ocean surface, but not on R, so if we assume these
latter properties remain essentially constant over hor-
izontal distances large compared to those over which R
changes significantly, we can directly relate changes in
I(n) to changes in R. Noting that in general R < 0.1, we
have
[a/(/i)]/(afl) * hin). (4)
Figure 3 shows that nI/aR is not an extremely strong
function of the aerosol concentration for concentrations
10 0 8 0 6 0 4 0 2
H
Fig. '2. /](<;) as a function of h for various aerosol concentrations.
1976 APPLIED OPTICS / Vol. 15, No. 8 / August 1976
67
12
i.
i i i i i — r — i r
OxN
UN 1
10
3xN ' 1
-
8
_
2
_J«N i
-
(xlOO) 6
_J0xN 1 1 1 "
4
1
1 1 -
2
-
\=400nm
0
_l 1 1 1 L . 1 i
1.0 0 8 0 6 0 4 0 2-
H
Fiji, :f. /;j(m' as a function of m for various aero
ncent r.il ions
up to three times normal and viewing angles up to 35°
from nadir. This suggests that horizontal gradients in
R can he estimated without knowing the aerosol optical
thickness with great accuracy.
We can now use Eq. (4) to relate changes in radiance
A/(/i) to changes in R(AR). i.e..
Equation (f>a) enables one to determine the minimum
radiance change the sensor must be able to detect for
a given AR. For example, suppose that observing at ^
= O.K.") it is desired to detect a 5"<> change in R lor clear
ocean water at 400 nm (ft * 0.1 ) through an atmosphere
with three times the normal aerosol concentration.
Figure 3 shows that /;>((). 85) is about 0.092. and noting
that the extra terrestrial flux at 400 nm is about 140
juW/cm- nm, we find from Eq. (4) that A/(0.85) is 0.064
(iW/cm- nm sr. In a similar way we can relate radiance
changes to AR for a nadir viewing sensor and any solar
zenith angle. As mentioned previously from the reci-
procity principle. /nadir = /(^o)//o, where ju.i = cosfln, "n
is the solar zenith angle, and /(/uu) is the radiance at the
top of the atmosphere seen by a sensor viewing at j*n
when the sun is at the zenith. Following through with
the same arguments that led to Fq. (5a). we find
A/ni,H,r = Htl\:iHlln)/ilR\±R * *Jn/j(K,,>A/(\ (5b)
Clearly, for a given AR, A/nHriir decreases substantially
with increasing solar zenith angle due to the presence
of the mo factor in Eq. (5b). For example, with a three
times normal aerosol concentration, a nadir viewing
sensor would have to have about 2.5 times more sensi-
tivity at Ik) = 60° as compared to 0t) * 0 to detect the
same A/?.
The above examples indicate how ASl can be used
in the design of a satellite sensor system for estimating
some ocean property such as the concentration of sus-
pended sediments or organic material. Specifically, one
must first determine the effect of the property to be
investigated on R, then, based on the sensitivity desired,
find A/?, and, finally, use Eqs. (5a) or (5b) to find the
minimum radiance change the sensor must be capable
of detecting. If the sensor has a limited dynamic range,
Eq. (3) can be used with Eqs. (5a) or (5b) to aid in the
sensor performance design tradeoffs. Unfortunately
at this time relationships between R(X) and sea water
constituents are not well established .
Conclusions
It has been shown that the upward radiance at the top
of the atmosphere can be accurately computed by as-
suming the radiation entering the ocean is diffusely
reflected from a hypothetical lambertian surface (be-
neath the ocean surface) of albedo r?(0\— ). This leads
to the natural definition of R(\) [/?(0,— ) as a function
of wavelength) as the ocean color spectrum. The de-
termination of subsurface oceanic properties from space
can thus be divided into two problems: (1) the deter-
mination of R[\) from satellite radiance measurements
and (2) the establishment of relationships between R(\)
and the desired ocean properties. Since the method of
computation conveniently separates the radiance into
a component that interacts with the ocean (I2) and a
component due to reflection from the atmosphere and
sea surface (/]), it is easy to relate changes in radiance
to changes in R(X). It was found that for viewing angles
up to 35° from nadir, I> is a relatively weak function of
the aerosol concentration for concentrations up to three
times normal. This suggests that spatial gradients of
R(\) can be determined with only a rough estimate of
the aerosol concentration.
Presently, computations are being extended to sev-
eral wavelengths in the visible and near ir portions of
the spectrum. When these are complete, it will be
possible to determine the radiance for any R(\) and
perhaps point the way toward recovering R(\) from
satellite radiances.
The author is also affiliated with N0AA Atlantic
Oceanographic and Meteorological Laboratories,
Physical Oceanography Laboratory, Miami, Florida
33149.
Appendix: Influence of Aerosol Phase Function on /,
and l2
It is natural to inquire how strongly the computations
of I\(n) and /_>(/j) presented in Figs. 2 and 3 depend on
the shape of the aerosol phase function. To effect a
qualitative understanding of the influence of the aerosol
phase function, computations of I\ and 1% have been
carried out using the well known Henyey-Cireenstein
(HCi) phase function,
II -A'-)/4tt
' h<;(«> = ; 7^.
(1 + H- - 2(! rns/M''-
where the asymmetry parameter g is defined according
'-"J>"
) cos«sin«rf«.
August 1976 / Vol. 15, No 8 / APPLIED OPTICS 1977
68
I I I I I I I I I I I I I I I ' I
• •• "HAZE C"
_ HENYEY-
GREENSTEIN
2 co"1
i ' I ' I I L_l I l__l I — I— I — I — 1—1-
120 140 160 180
0 (degrees)
Fig. 4. Comparison between the haze (' and various Henvey-
('■reenslein phase functions characterized by asymmetry parameters
0.6. 0.7. and 0.8.
and II is the scattering angle. Since i,' for the haze C
phase function used in the computations described in
the text is 0.690, computations have been made with
Phi\W for i,' values of 0.6, 0.7. and 0.8. Figure 4 com-
pares these PhCiWs with the haze C phase function.
The HG phase function for# = 0.7 clearly fits the haze
C phase function quite well in the range of 5° < 0 <
140°: however, as is well known, the HG formula is in-
capable of reproducing phase functions computed from
Mie theory in the extreme forward and backward di-
rections. The HG phase functions with asymmetry
parameter 0.6 and 0.8 are seen to be substantially dif-
ferent from the haze C distribution at nearly all scat-
tering angles. On the basis of Fig. 4. it should be ex-
pected that / 1 and /_> computed with Phc.C) will be in
close agreement with the haze C computations only for
ii close to 0.7. Figures 5 and 6, which compare the re-
sults of computations of I\ and /_., respectively, for
Phc.UH with a' = 0.6, 0.7. 0.8 (steplike lines) and the haze
C phase function (solid circles) for the normal aerosol
concentration, show that this is indeed the case. It is
seen that except for apparent statistical fluctuations,
the HG phase function for# = 0.7 yields values of /j and
1 2 in good agreement with the haze C computations.
This suggests that the detailed structure of the phase
function is not of primary importance in determining
1 \ and 1 2, and it may be sufficient for remote sensing
purposes to parameterize the phase function by i,'.
To get a feeling for the importance of variations in the
phase function in the remote sensing of ocean color,
consider the effect of changing the aerosol phase func-
tion from a HG with # = 0.6 to g = 0.8 over an ocean
with R =0.1. From Figs. 5 and 6, it is found that the
normalized radiance at n = 0.85 (the assumed obser-
vation angle) decreases by 4.9 X 10-:'; this decrease in
radiance would be interpreted under the assumption of
no atmospheric change as a decrease in R from 0.10 to
0.056. This clearly indicates then that variations in the
aerosol phase function in the horizontal direction could
be erroneously interpreted as horizontal variations in
the optical properties of the ocean. It is, however,
probably unlikely that, except in extreme cases, the
clear atmosphere oceanic aerosol phase function will
exhibit variations as large as considered in this example.
o
O 6
X
1 1 r
"HAZE C"
HENYEY-GREENSTEI
X = 400nm
9=0.7
J I I I I I l_
1.0 09 08 07 06 05 04 03 0.2 01
• H-
Fitj. :"). Comparison between /i(^i) computed for the haze C and
Henvev-Creenstein phase functions for an atmosphere with a normal
aerosol concentration.
1
1 T
i i
"HAZE C"
1
1?
HENYEY-GREENSTEIN
X =400nm
-
II
=w-Ti~l .
1 9=08
-
-^ • |9=C
7
10
IVa
361
9
8
o
Q 7
X
•
h-7 6
5
4
3
2
l
i i
1 1
1
1.0 09 08 07 06 05 04 03 02 01
H-
Fig. 6. Comparison between /^(m! computed for the haze C and
Henyey-Oreenstein phase functions for an atmosphere with a normal
aerosol concentration.
1978 APPLIED OPTICS / Vol 15, No. 8 / August 1976
69
Assuming that the aerosol concentration of the atmo-
sphere can be determined, the uncertainty in the aerosol
phase function will still of course provide a limit to the
accuracy with which the ocean color spectrum can be
retrieved from satellite radiance measurements.
Quant. Spectrosc. Radiat.
References
1. H R Gordon and 0. B Brown,
Transfer 15,419 (1975).
2. H. R. Gordon and 0. R. Brown. Appl. Opt. 12, 1544 (1973).
3. G. N. Plass and G. W. Kattawar. Appl. Opt. 8, 455J1969).
4. G. W. Kattawar and G. N. Plass. J. Phys. Ocean 2, 146 (1972).
5. H. R. Gordon and 0. B. Brown, Appl. Opt. 13, 2153 (1974).
6. R. VV. Preisendorter, C.G.G.I. Monogr. 10, 1 1 (1961).
7. .1. E. Tyler and R. C. Smith. Measurement.-, of Spectral Irra-
diance I'nderwater (Gordon and Breach. New York, 1970).
8. L. Klterman, VV, Visible, and IR Attenuation for Altitudes to
50km. 1968. Air Force Cambridge Research Laboratories, Report
AFCRL-68-0153U968).
9. R. S. Fraser. Goddard Space Flight Center, Greenbelt, Md.,
Personal Communication.
10. D. Diermendjian, Appl. Opt. 3, 187 (1964).
11. G. Kullonberg, Deep Sea Res. 15, 423(1968). Note that all the
phase functions in the present paper are normalized according
to
«/0
P(8)sin6d6 = 1.
12. T. J. Petzold, Volume Scattering Functions for Selected Waters
(Scripps Institution of Oceanography, University of California
at San Diego. 1972), SIO Ref. 72-78.
13. H. R. Gordon, Appl. Opt. 12, 2803 (1973).
14. G. N. Plass and G. W. Kattawar, Appl. Opt. 11, 1598(1972).
15. H. R. Gordon, O. B. Brown, and M. M. Jacobs, Appl. Opt. 14,417
(1975).
16. S. Chandrasekhar. Radiative Transfer (Clarendon, Oxford, 1950).
17. R. Curran, Appl. Opt. 11, 1857 (1972).
18. J. L. Mueller, "The Influence of Phytoplankton on Ocean Color
Spectra," Ph.D. Thesis, Oregon State University ( 1973).
70
12
Reprinted from: Proc. AIAA Drift Symposium, Hampton, Va., May 22-23, 1974,
NASA CP-2003, 175-192.
A LAGRANGIAN BUOY EXPERIMENT IN THE
SARGASSO SEA
by
Dr. Donald V. Hansen
Atlantic Oceanographic and Meteorological Laboratories
Environmental Research Laboratories
National Oceanic and Atmospheric Administration
Miami , Florida
As indicated, we'll hear from a group of distinguished drifters this
morning. In order to be sure we don't run out of time for me, I'll say
my piece first. I can make mine a little bit shorter -than I'd planned
because a number of comments that have already been given set the stage
for it. The genesis of my story begins back about 1970 when a number of
people in the physical oceanographic community in this country and abroad
began thinking and talking about a project to be called the Mid-Ocean Dyna-
mics Experiment (MODE). It was referred to yesterday by Doug Webb and others
as the M0DE-1 Project. About that time, I began talking to Sam Stevens about
the possibility of hitching a free ride, or at least an inexpensive ride,
on the French EOLE satellite system, and through the yery good offices of
Sam and his crack team, we were indeed able to do that. The engineering
for the project was done by the Miami Branch of the Engineering Development
The Author: Dr. Hansen received his Ph.D. in Oceanography from the
University of Washington in 1964. He worked as a
Research Assistant Professor at the University for 1
year before becoming a Research Oceanographer with the
Department of Commerce in 1965. He is presently
Director, Physical Oceanography Laboratory, Atlantic
Oceanographic and Meteorological Laboratories, NOAA,
Miami, Florida.
71
Laboratory of NOAA's National Ocean Survey in Miami. Charlie Kearse
described yesterday some of the shipboard procedures and arrangements
that were developed by them for us to get these buoys in the water,
but what he did not mention was that they also were entirely in charge
of the engineering and fitting out of these buoys, and in getting them
into the water on what turned out to be extremely short notice. As the
project developed, it really didn't go quite as we had planned to have
it no, because, due to changes in the scheduling of the MODE Project and
of the EOLE Satellite Project, it appeared at a critical time that the
two after all were not going to be coincident in time. The EOLE Project
was to terminate before the MODE Project went to sea. However, it seemed
an interesting and important enough experiment to do in its own right,
so we pressed on and did it anyway, almost totally independent of MODE.
There was about a 1 month overlap between the termination of this
project and the initiation of MODE and, in fact, the buoy that we
initially had deployed farthest from the MODE area passed within 30 miles
of the central mooring of MODE during the second month of that project.
I want to show you a few slides first to indicate some of the motivation
fcy~ having done the experiment in the way we did it, and to set the stage
to address the question of interpretation which Dean Bumpus raised yesterday
w^th some vigor. If I can see the first slide now, please.
This is an example of a publication that is put out by the Navy. They're
called Pilot Charts and show currents and wind to be expected in this
reqion of the Sargasso Sea, what mariners and, in fact, what the rest
of us know about surface currents in the Sargasso Sea. I might mention in
passing, that all of the data that you can find anywhere on such atlases or
cherts are, in fact, derived by Lagrangian means. These currents summarized
'•<■■ atlases are about 99 44/100% pure ship drift calculation. They're
currents inferred from the deviation of ships from their navigational
calculations. The major feature I want to point out here is the fact that
all of these current vectors show a very smooth steady flow to the west at
72
speeds ranging from about a knot to speeds on the order of 1/2 a knot.
The MODE Project which you saw illustrated in one of Doug Webb's slides;
I believe, was conducted in a circle of about 200 kilometer radius.
"Figure 2 is a copy of a slide taken from some Soviet work in this region.
The Soviets have an active interest in the oceanography of the low latitude
Atlantic because they conduct vigorous fisheries activities out there and
they have conducted intensive research cruises in this region in 1969
and again in 1971. Figure 2 shows their interpretation of those obser-
vations. They're a rather intensive set of observations. Soviet
literature is a bit hard to interpret as many of you know, in that they
don't document their conclusions by Western standards, but as best one
can determine, the observations themselves are good. The interpretation
is that the solid dark vectors represent the conventional wisdom about
the Antilles Current - the northward and westward flow. Imbedded within
them are open vectors which are directed to the southeast, which they
interpret as a major countercurrent within the Antilles Current and
flowing from someplace just off Florida, all the way down, as a con-
tinuous feature, joining the complicated equatorial current system and
then flowing off to the east. The light lines you see are where they
have intensive sets of observations. The observations consist of moored
current meter measurements and shipboard measurements of temperature
and salinity, from which are computed the velocity field by classical
methods. This is the interpretation of what looks like a rather good set
of conventional measurements in the region. When I first saw it, I was
a little skeptical to say the least - if it's true, it certainly is rather
exciting news to the oceanographic community in general and, in fact,
rather embarrassing news to the American oceanographic community: that
the Soviets should discover right on our doorstep a very major oceano-
graphic feature about which we have no knowledge. This is a very major
current. It is a surface current which, however, extends to about a
kilometer deep in the ocean and it has a volume transport approximately
equivalent to that of the Gulf Stream or Florida Current as it issues
from the Florida Strait and heads up the east coast, which all of you are
aware, I am sure, is the major oceanographic feature off the U.S. east coast.
73
So to try to serve two purposes here--one, we recognized before we
went to sea that we would not be able to conduct an experiment in close
coordination with the rest of the MODE operations; nontheless, it seemed
worthwhile to try to obtain a direct measure of the near surface current
structure and its variability in the MODE region. Hence we deployed
our buoys along 67°W, immediately to the east of the MODE area, presuming
that with the northward and westward drift they would sweep through the
MODE area and probably be gone, along the lines of the rather imaginative
sketch that Vukovich showed us yesterday, before MODE-I operations began.
That was my preliminary guess as to what we might expect in the way of a
trajectory development of these buoys when they were deployed, but as you
will see, it didn't go quite that way. The idea then was to deploy the
buoys so that they would sweep through the surface water in the MODE area
before MODE ships came out for that project, except for the southernmost
buoy. We learned of the Russian work fairly late in the game and modified
the plan to some extent. The buoys were deployed 1° of latitude, 60 miles
apart, between 28 north and 25 north. We placed the last one an additional
30 miles south, to place it in the middle of the region where the Soviets
claimed to have discovered the countercurrent , to test that particular
hypothesis.
Figure 3 shows one of our buoys in the water, using the EOLE satellite
tracking system which is exhibited in the side room.
The next slide is of some interest because I think there probably will be
additional discussion of this EOLE system today. Figure 4 shows the dis-
tribution of position fixes in time for the No. 5 buoy. It shows the hour
of the day from midnight to midnight versus day of drift, so the points
show the hour and day from time 0 that positions were obtained through
the satellite system. They have a quasi-random pattern providing generally
2-5 fixes per day which round the clock slowly. The satellite "day"
turns out to be something on the order of 23 1/2 hours. This is not a
p<i*-Moul3rly good data distribution for most kinds of analysis we anticipated
doing. Once we saw how the data were evolving, we did polynomial '"itting
74
to the X and Y coordinates of the position to provide some smoothing,
then we interpolated positions on these polynomial fits at 1-day
intervals so we could deal in terms of a fixed time interval. I have
a film animation, which we will see, that is the same sort of thing
that Doug Webb showed yesterday. It runs rather rapidly so I want tc
take just a moment to tell you what it contains.
EDITOR'S NOTE: At this point, an animated
film sequence showing the drift
history of all five buoys was shown.
Figure 5 shows the complete trajectory for the buoy, No. 4, that survived
longest. It was retrieved and returned to the laboratory in April !973.
To speak very breifly about interpretations now, I think that even from
this fairly simple experiment, one must begin to make some interpretations
and begin to think seriously about how to interpret such data. Soir.e things
come fairly immediately to mind--in particular the region wnere we were
exploring the possibility of a major counter-current. Three of the buoys
moved into the region of the supposed countercurrent and pretty much
negate the possibility of there being any such countercurrent, and in fact,
identified the source of confusion about a countercurrent, It's a sampling
problem. The Soviet observations I think are good observations. Their
current observations are usually good ones and their shipboard observations
are good also. However, they have sampled at fairly widely spaced sections,
as one must by traditional methods, and in each of these sections they have
found some sort of eddy motion. The error is not in the observation, but n
interpretation, in assuming continuity between these various sections. A
Lagrangian technique appears to offer much potential for exploring spatial
structure in the flow, and offers a fairly economic means for exploring or
answering questions about spatial distributions or the existence of particular
phenomena.
Another thing that we are working on now--I just have a bare beginning of
some things to say about it, another kind of application that has been of
interest in oceanography for many decades now is an interest in trying to
75
predict in a semiquantitative sense the transport or the distribution
or dispersion of things dissolved or fine objects scattered in the sea,
etc. As an example, the interest is in being able to predict the
concentration or the probability of a particular object being in a particular
place by an equation in the form:
8P = K 92
at ,xij 3 3y
X. Xj
In this formulation, K. . is a dispersion coefficient relating the
concentration change to the spatial gradient in two dimensions,
essentially saying the time change is a diffusion type process related
to the gradient, but with a tensor diffusion coefficient. In some classic
work by G. I. Taylor dating back to 1921, it's shown that the kind of
information that is needed to approach this kind of a problem is, in fact,
the Lagrangian information—not Eulerian information, and is laboratory
and wind tunnel dynamics, a lot of work over several decades has gone into
the problem of trying to establish a relation between Lagrangian and Eulerian
statistics. The Eulerian statistics are easier to measure, but for certain
problems the Lagrangian statistics are the ones that you really want.
Given a particular particle or particular buoy that has a particular path,
one can consider the mean path and the deviations from it, and compute the
time lagged autocorrelation. That's what we have done but only for the
diagonal components to date.
'.■.'hat we did was to take position data, differentiate it to obtain velocity
data, and then compute a statistical function.
The autocorrelation, call it R, is the ensemble average v (t)V-(t+x)
V-(t) denotes east or north component of velocity at some time t, and t a
time lag interval .
76
This is averaged over the ensemble or averaged over all time for the
buoy motion. It's a measure of how rapidly the motion loses similarity
with itself. At no lag at all the velocity looks exactly like itself so
the autocorrelation is equal to the variance. The correlation decays
with some structure as time runs on. Figure 6 shows the nature of the
autocorrelation function for buoy 4.
The major features of the curve show that the correlation drops to zero on
a time scale of about a week or 10 days. That is the Lagrangian time scale
for motion in the Sargasso Sea. It also has an oscillatory structure that
damps out with increasing lag. It really has validity only out to about
120 days. After that there are too few data points, to draw even tentative
conclusions. There are roughly 200 observations going to make up each data
point in the beginning of the curve. I don't believe anything out in the
tail end of the curve, so basically what is revealed at least during this
set of observations is a periodic variability in current having a time
scale of about a month—peak to peak here is a lag time of about a month.
The next step to apply this to the dispersion problem is to relate the K. .
to the autocorrelation function using logic of G. I. Taylor and others.
The result is that the K. . is obtained from the integral of the autocorrelation
over time. From a quick calculator integration of the function for buoy 4,
7 2
I obtained a value 10 cm /s, which turns out to be a number popular among
oceanographers. If one had to guess without knowing anything else it
probably would be slightly higher than this, perhaps by a factor of 5 or so.
The other thing you can do—this is a thing that I think Lagrangian techniques
are in fact more appropriate for, relative to other kinds of observation and
fixed moorings, etc., is to explore not the time correlation behavior but
the thing that's really hard to get from moored current meters, the space
correlation behavior, because it's really an expensive undertaking to put
down a lot of moorings with fixed current meters to explore how currents
vary on a time scale of 1 mile - 10 miles - a hundred miles, etc.
77
By deploying an array of drifters such as we've been discussing here,
one can set the initial scale, but as the pattern evolves, it covers
quite efficiently a considerable band of space scale. We tried doing
that with the buoy data we have here using the buoys in pairs, and
in order to get the bulk of data up to some1 usable level it turns out
we don't really have very much data at all yet. In order to try to
get the statistics as well behaved as possible, we borrowed a ploy f rom
the field of homogeneous turbulence and worked with buoys in pairs which
are separated by a vector having some direction and some length L and
decomposed them, presuming that the flow field is isotropic. I really
don't have any very good argument to defend that except that the r.m.s.
speed in the east directions are approximately the same at about 15 cm/sec,
so with a little bit of hand waving we must pass over that question.
Then we decomposed the velocity components at buoy pairs into components
parallel to and orthogonal to the separation vector between them and
computed spatial correlations at fixed times for the parallel and per-
pendicular components so defined. It turns out, however, that for the
ipa^e scales ^nyered by this data set, 100-400 km, the correlations are
evidently so iow that tney cannot be distinguished from zero in the
quantity of dar.- available. Indications are '.hat probably the spatial cor-
relation is lowest someplace here in the first 100 kilometers or so
which is essentially the same sorr of thing that was found before and
during the M0DL experiment for the deep water circulation—deep currents
in this same area. I think I've taken about as much time as I ought to.
Thank you for /our attention. If anyone has any comments, I : 1 1 try to
respond.
78
QUESTIONS
JIM RUSSELL -- U. S. Naval Avionics Facility:
When you assume your isotrophy in your turbulence, what kind of scales
are you really looking at in your measurements? Are they fairly large?
Answer -- DR. HANSEN:
Right, the scales we're looking at here are roughly in the 100-500 kilometer
range for enough data to be of any significance at all.
JIN RUSSELL -- U. S. Naval Avionics Facility:
And it's also in the surface water rather in the deeper water that we're
talking about?
Answer -- DR. HANSEN:
This is strictly the surface water. This was using the buoy that Charlie
Kearse showed some slides of yesterday. We had a parachute drogue on them
which was at 30 meters depth, so it's really very much in the upper layers
of the ocean. The thermocline there is 800 meters deep or so.
JIM RUSSELL -- U. S. Naval Avionics Facility:
Something does bother me about assuming isotrophy there. Did the results
you got indicate that assumption may have been o.k.?
Answer -- DR. HANSEN:
I really don't think I can address that. I haven't looked at it carefully.
The only think I can say in justification is that the variance in the north-
south and in the east-west direction is approximately equal, about 13 and 15
centimeters per second for the r.m.s. speed. There is some indication that
in deeper water there probably is some anistrophy, higher energy levels in
the north-south direction as compared to the east-west, but it does not
show up in this surface data set.
79
BOB HEINMILLER -- Woods Hole:
There is a little event on your film that caught my eye. There were
two buoys--looked like they were very close together--just estimating
from the scale 5 and 10 miles--the tail of one about 10 times the tail
of the other, both going in the same direction which implies that the
speeds for one were considerably, an order of magnitude, higher than
the other. I didn't notice that that occured any other time during
the film. Have you seen any sort of that? That seems like an awfully
high differential .
Answer -- DR. HANSEN:
It does happen other times. You have to see the film several times to
detect more of these events, but when we first deployed the buoys I thought
we had discovered the center of the ocean circulation because for a period of
about 10 days the No. 4 didn't move within the resolution of the satellite,
which is about a kilometer there, while buoys north and south of it,
particularly one of them north of it, turned and moved toward it and came
by at a good rate of speed within about 30 miles, yet the one that was
initially deployed there hardly moved for about 10 days. After 10 days
it suddenly took off and moved to the south as rapidly as any of them.
I interpret that as being indicative of large lateral shears in the flow.
In the movies that Doug showed yesterday, you see very much the same sort
of thing in the S0FAR float measurements. It looks there as if there are
jets imbedded in the flow. They seem to be north-south oriented there but
didn't show up quite so much here perhaps because a lot of the statistics
may be biased by the fact that the buoys spent a fraction of their time
fairly near the Bahama Banks where presumably north-south motion is
strongly inhibited and east-west motion parallel to the banks is favored.
CHRIS WELSH -- Virginia Institute of Marine Science(VIMS) :
It occurs to me that if you were to put a current meter section out where
the Russians did for a long length of time and average over the time to get
a cl imatological circulation, you would still see the countercurrent structure
80
that they apparently saw simply because when the currents going south, the
western boundary, if you want to call it that, structure is apparently more
intense from the little worms you have than when they go off to the north.
Answer -- PR HANSEN:
I think that's probably true - if you'd observe just those sections, you
likely would see what you interpret as a countercurrent migrating onshore and
offshore and north and south or something. I suspect different eddies or
different waves or whatever they are occur there at various times. You're
probably right. You'd really have to have a very dense set of current meter
moorings to be able to resolve the spatial structure in the flow to disabuse
yourself of that idea.
PETER HACKER -- JOHN HOPKINS:
I'm worried a little bit about the slippage of the drogues in regions where
you do have high lateral shear from the currents you observed and from the
winds thai are typical in that area. Do you have any kind of a percentage
estimate of slippage of the drogue with respect to a water mass?
Answer -- DR. HANSEN:
I haven't put a number on it. We're investigating. We just got all the
tropical weather information. We will correlate the local winds with the
buoy movements; however, I haven't put a number on it. Maybe Charlie has, I
don't know. I think the wind drift for this particular buoy is probably
negligable in terms of the currents and the things we see for two reasons:
one, the dominant periodicities in the major flow features have a time scale
of about a month and you just don't see things like that down there in the
weather pattern. You don't expect major wind events in a time scale of a
month. Strictly from the engineering point of view, this buoy was about
40-41' long with the major portion of the cross section submerged and, in
addition, it has a parachute drogue on it. All indications are that the
parachute droges did indeed survive for a time scale of 6 months or better.
Bob Heinmiller was one of the last people, I think, to see buoy No. 5 and the
reports I have from the appearance of the buoy in the water, the way accessory
31
floats were arrayed and so on, indicate that the parachute drogue hardware
apparently was still on at that time. We recovered one in December after
3 months at sea, and the whole subsurface hardware was essentially in
perfect condition then. The one we recovered after 8 months outside the
Bahamas had lost its parachute. I don't think it's a serious problem. Did
you ever put a number on the windage Charlie?
CHARLIE KEARSE:
I guess I'm just worried. You know, even if it's just 5 or 10 percent--if a floi*
drifts 100 kilometers downstream or something, at the same time it can be
going cross stream 5 of 10 kilometers in a region where you do have intense
sheer, it may in fact drift from a countercurrent into the other part of the
countercurrent if you do have closely spaced currents.
82
» \ \ \
i
T
\
^ — — — •
\,~ >, y i/
#-
o
to
-t->
c
>- CD
<c o
O r—
•-H Q.
D. <D
>- -o
»—
■a
co c
<: <o
< c
UJ o
CO t- c
—J -a
o cd
LU +->
0) u
.c •!-
+-> -o
c
cr>T-
o a>
CO $-
CO
< r-
cn i
q; lu
< Q
CO o
LiJ
CD
•i- o
to to
Z3 i—
to
>> CD
O X-
.O
E
C7> 0)
C +J
•r- tO
4J >>
M- to
CO CO T3 C
2 CO
UJ <£
en _j
a: i—
c_>
UJ i— (
o
< 2:
U. 3
cc o
id 3:
CO CO
S-
Z3
CTl
ir> +j
o o
to
c c
o o
+-> +->
to t/1
O O
a. a.
83
Figure 2 THE ANTILLES COUNTERCURRENT AS HYPOTHESIZED BY
V.G. KORT
84
Figure 3 DRIFTING BUOY USING EOLE TRANSPONDER DEPLOYED AT SEA
85
>-
I >
±>
CD
■~1 1 5
J -I
Is
—i o
<\l
O
O
O- -I CM
0 „ i °
o — { «r
o
? o
10
0)
o
o
— o
05
o
o
I o u.
) 3
.}<> u.
! S a
IS
o
o
0
o
'
o
o
*
o
-I ,'J
1
0 I
.'J O
o
o
O
O
o
v I
0J
u
to
avq iO
co
HDC H
86
>-
o
ZD
co
CJ
O
1/0
X
O0
o
o
o
I — I
I—
co
C£
I—
oo
t — I
Q
0J
S-
en
15c -
j I i ■ ' » \ ■ ■ i »
I i i
15c
85° 80° 75<
70° 65<
April 22, 1973
25(
V
1>^
20c
CO
Sept. 22, 1972
J i
— 25'
20c
75<
70°
65<
Figure 5 DRIFT TRAJECTORY FOR BUOY No. 4
87
240.1
200.
160.
oo
cc
o
1—
o
120.
LU
■>
>-
1-
t— 1
o
o
80.
_1
LxJ
>
I/O
I
u_
40.
o
s:
o
►— 1
\-
<c
0.
_J
UJ
oc
C£
o
o
o
-40.
-80,
■120.
■160.
40,
; r5-1. 160. 20(j. ?_/\o
!llil!W :> ■:>! ;;/\y.S
280. 320
rioutp 6 ! Af ;PA';r, I, AN MMf CORRELATION riJNCTIOM FOR BUOY Mo. 4
88
13
Reprinted from: Proc. of the Third Annual Conference on Computer Graphics,
Interactive Techniques, and Image Processing, University of Pennsylvania,
Computer Graphics 10, No. 2, 218-223.
AUTOMATED CONTOURING OF VERTICAL OCEANOGRAPHIC
SECTIONS USING AN OBJECTIVE ANALYSIS '
A. HERMAN
National Oceanic and Atmospheric Administration
Atlantic Oceanographic and Meteorological Laboratories
15 Rickenbacker Causeway
Miami , Florida
This paper describes a group of computer programs developed for contouring vertical
sections of oceanographic parameters. The vertical profiles can be constructed from data
collected in a variety of ways. The input data for the driver subroutine need not be
equally spaced horizontally or vertically. The routines are written in Fortran for a
UNIVAC 1 10 S with an offline Gould Plotter, but can easily be adapted to any computer
with a Fortran compiler and a plotter which accepts Calcomp-like commands. The routines
are of modular construction.
1
INTRODUCTION
This paper describes an automated tech-
nique for producing vertical profiles of
oceanographic data using quasi-objective
analysis. The computer program based on
the method is also described. Though obj-
ective analysis is well established in
meteorology, it has seldom been used in
oceanography, and when used, it has been
restricted to a specific geographical area
[Bretherton (2)]. This program is not
restricted to a specific geographical area
and is currently being used by oceano-
graphers at the Atlantic Oceanographic and
Meteorological Laboratories (AOML) of the
National Oceanic and Atmospheric Adminis-
tration (NOAA) for studying profiles and
creating contour naps in the Gulf of
Mexico and the New York Bight areas. The
advantages of this technique over others
generally available to oceanographers are:
1. The sampling points for the input
data need not be uniformly spaced in either
the vertical or the horizontal direction.
2. Contours are not extrapolated beyond
the input data.
3. Interpolation of data is done using
statistically based correlation coeffi-
cients .
The program is based on the assumptions
that all stations in a single profile lie
close enough to a straight line connecting
the two farthest stations that no signi-
ficant errors will be introduced bv
assuming all stations to be on that line,
and that time differences in the collec-
tion of dna points may be ignored. These
assumptions arc needed when data do not
exist for accurate spatial and time
corrections. Khen studies are made with
adequate resolution in time and space to
make such corrections the program can be
modified to include them.
2. PROGRAM FLOW PLAN
The driver subroutine Versex, calls many
subroutines which together accomplish the
fol lowing :
1. The conversion of latitude and longi-
tude to rectangular coordinates on a
Mercator projection.
2. The projection perpendicularly of all
stations onto a line connecting the end
stations, and the computation of the
distance of each station from an end of
the line. The distances are in inches at
a scale of four inches equal to one degree
of longitude (Figures 1 and 3) .
3. The determination of the depth scaling
is such that the deepest depth is equiva-
lent to the distance between the farthest
stations. This will produce a square chart
with a fixed horizontal scale and the ver-
tical scale being a function of depth and
horizontal size.
4. The fitting of data to a matrix with
user controlled smoothing with matrix
values below deepest depths being logged.
5. The construction of a contour matrix
such that a chart is left blank below
deepest data, all data points are marked,
and the vertical scale is labeled (Figures
2 and 4) .
218
89
o
o
m
o
>-
o
2
ro
<u
N
*£
c
LU
71
cr
c
o
Z>
'Z
CD
u
o
L_
r-i
<+,
o
C
0
^r
1— t
n
0)
219
90
o
o
o
O
o
o
o
o
CJ
u
G>
•H
X
V
<4-l
o
3
b
ro
c
■H
o
o
UJ
■s,
c
tr
o
ZD
ID
u
(3
o
U-
>H
4-1
O
o
a.
o
o
m
Cf>
0>
>~
o
o
o
CO
en
3
220
91
09-
O
o
o
ro
O
O
if)
I O
O
O
h cvj
\- O
o
in
L o
o
o
I-
LjJ
UJ
X
»-
a.
UJ
Q
UJ
cr
Z)
e>
Ll.
3
o. •
*-> n!
3 ^
O ctt
Q
3 +->
O JZ
C 'H
O PQ
<_> O
O >-
2
0) 0>
U 2
3
■m C
aj o
(-.
0) T3
e <n
O 03
HI
3
M
a)
a
o
u
■H
X
V
?-
3
T3
4)
O
p.
e
4)
l-
3
221
92
3. CONVERSION OF LATITUDE AND LONGITUDE
TO CARTESIAN COORDINATES
Subroutine Merc converts latitude and longi-
tude to cartesian coordinates on a Mercator
projection chart at a scale of four inches
equal to one degree of longitude. The chart
has its origin at 80° S and the Greenwich
meridian. Chart values increase to the
east and the north. Formulae for the Mer-
cator projection can be found in Honford(l)
4. CREATING THE DATA MATRIX
The interpolation of data onto a grid is
normally subjective. Thus the quality of
the result depends on the skill of the
analyst. In order to obtain a more uni-
form quality, subroutine V.'TMAT performs
the interpolation in a quasi-objective
manner. First an array of correlation
functions is establ i shed , and then the
interpolation is done based on the correla-
tion function. The method of objective
analysis is described in Gandin (31. The
equations are of the following form:
The autocorrelation function for the field
f(r1 is:
Mf Cr2
r,)
fTfprifp
(i)
where r and r., are two points in the
vertical plane'of interest. Instead of
using f, it is convenient to use f'=f-7,
the statistical deviation from the norm.
(r
r2j = f'Crj) f'(r2;
(2)
Because of assumptions of homogeneity and
isotropy [Gandin (31], the correlation
function is expressed in terms of separa-
tion distance.
1
(3)
Both noisy data and wave data will have a
correlation function which crosses :ero
more than once. To reduce the effects
of noise, the correlation function, Mf,
is taken to be ;ero past the first zero
crossing. The effect of the presence
of waves on the data will not be smoothed
out if the interpolation grid is fine
enough (grid si:e is a user option in
VERSEX1 .
The correlation coefficient is a normal-
ized correlation function:
(p)
(o)
ToT
(4)
U '(d represents a correlation coefficient
between the values of f at two points
apart and is a'function of distance.
ft ■*
A bar denotes an average and an arrow a
vector .
The interpolation is done using the follow-
ing equation:
n
£
f = i = l UCrQfj
£OCr,)
i = l
where fQ= deviation of value from known
dos it ion .
£:= values at known distances r. .
n = number of observation points.
The actual contouring is done on a Gould
electrostatic plotter using a Fortran pro-
gram with Calcomn type nlot instructions.
The contouring algorithm involves only
linear interpolation between grid points.
This method was chosen for simolicity.
6. INSTRUCTIONS FOR USE OF SUBROUTINE
VERSEX
VERSEX
called
Appendi
f ol lows
Call VE
DTMIN,
where N
M
XLAT
XLON
DAT
is the only routine that must be
by the user as illustrated in
x A. The calling statement is as
DEP
CI
DTMIN
RSEX (N, M, XLAT, XLON, DAT, DEP, CI
DTMAX, DPMIN, DPMAX, NRO'.VS)
= number profiles to be inDUt
= maximum number of data noints
in a nrofile.
= array of latitudes in degrees
= array of longitudes in degrees
= a tv;o-dimensional array (N,M)
of data values (emnty locations
should be set to a negative
number . 1
= a two-dimensional array (N,M)
of denths corresDonding to data
points .
= contour interval (if zero or
negative, CI is set = maximum
data value -minimum data values
divided by 101 •
= minimum data value to be con-
sidered. If it is negative,
then it is reset to the lowest
value in DAT.
lue to be consid-
negative, then
the largest data
If it is nega -
s reset to the
ue at which there
If it is nega-
hen it is reset
denth at which
value .
rol is the num-
e matrix to be
ntrol the amount
he number of rows
Minimum
DTMAX
maximum data va
ered. If it is
it is reset to
value in DAT.
DPMIN
minimum depth,
tive, then it i
least denth val
is a data value
DPMAX
maximum denth.
tive or zero, t
to the greatest
there is a data
NROWS
(i
f
greater than ze
ber of
row
s you want in th
compute
d
i
f you wish to co
of smoo
th
ing . The fewer t
the gre
ater the smoothing.
222
93
smoothing occurs when N'ROKS = M-l.
REFLRLNCLS
(1) Bomford, G., "GEODESY", Oxford,
Clarendon Press, 1971.
(2) Brethertcn, P.P., "A Technique for obj-
ective Analysis and Design of Pceano-
graphic Experiments", to be published
in Deep Sea Research.
(5) Gandin, L.S., "Objective Analysis of
Meteorological Fields", Gimiz, Lenin-
grad, 19o3.
223
9*
14
Reprinted from: Proc. of the Fourteenth Annual Southeast Regional ACM
Conference, University of Alabama, Birmingham, Alabama, JUb-JU«.
AN AUTOMATED SOLUTION FOR OMEGA NAVIGATION
A. Herman, NOAA/AOML
A. C. Campbell, U.S. Naval Oceanographic Office
ABSTRACT
Omega is a commercially available, electronic, world wide navigation system.
Navigators and other Omega users generally determine geographic positions from
Omega lane counts either by scaling charts and tabulated corrections or by a
programmed iterative technique. This paper describes an algorithm which determines
position directly.
INTRODUCTION
Omega is a worldwide navigation system usable anywhere and at any time.
Other navigation systems have the limitation of being available only locally or
at certain times of day. Electronic navigation systems can be classified by the
kind of line of position they generate. These include hyperbolic, concentric
circles, radials or any combination.. Omega is a hyperbolic system. A hyperbola
is defined as the locus of a point moving such that the difference in distances
from two fixed points to the moving point remains a constant. Thus, a hyperbolic
navigation system is one in which a mobile user measures the difference in trans-
mission time between signals from two shore stations. This difference, as can be
seen from fiqure 1, determines one line of position. Two lines of position are
required for a fix. Omega is a pulsed hyperbolic system. A master station transmits
an encoded series of pulses at short intervals; these pulses are retransmitted
each time they are received by a slave station. The system has fixed time delays
such that the mobile user always receives first the master pulse, then the slave.
The receiver measures the time difference between master and slave signals, removes
the time delays, and displays the resulting time differences in lane counts. Two
time differences (lane counts) define two hyperbolic lines of position. The signals
are affected by: (1) The conductivity of the surface over which the signal propagates
and (2) refractivity of the atmosphere through which the signal propagates.
This paper introduces a new method for calculating the point of intersection
of two hyperbolic lines of position. The computation differs from the usual method
in that it is a noniterative solution. The method is an application of the solution
for hyperbolic systems described in Campbell (1955). The Defense Mapping Agency
of the Department of Defense publishes the data necessary to make calculations
including station locations, datum, spheroid specifications, and transmission
frequencies. The program described in this report is for 10.2 KHz with a phase
velocity of 300574 Km/sec. (Omega lane widths are equal to 1/2 wave length along
the base line, which is equal to velocity/2x frequency or 14734.02 meters.) The
time delay mentioned above is such that the base line bisector is 900 lines.
Therefore, the length of the baseline in wavelength plus coding delay = 900.
MATHEMATICAL METHOD
The method requires an approximate initial position from which x and y
distances from the approximate postion, Pi , to the true position, P, are computed.
Trie geometry of the problem is displayed in figure 2. Four normal equations of
the form:
95
306
x sin a + y cos a, + a . ' - ai = 0 (1)
i = 1,2,3,4,
are utilized to obtain x and y. As is shown in Figure 2, x, y, and the four a.
are the unknowns. Lane counts must be translated into distances along the baseline.
Ko = cj " bj where j ° lf 2* ^
K. = H. x 147342.02 meters. (3)
The differences a', - a. can be written as:
a^ - a] = a^ - (a2 + k^ = (a^ - kj) - a2
a'2 a2 = a2' - a2
a3* " a3 = V ' W + y = ^3 " k2J ' a4
V - a4 = = a4' - a4
Then the 4 observational equations (1) become
x sin a-j + y cos ou - a2 + (a-,' - k,) = 0
x sin a2 + y cos a2 - a2 + a2' = 0
x sin cu + y cos <*3 - a. + (aJ - k2) = 0
(4)
(5)
(6)
x sin a^ + y cos a4 - a^ + a^' = 0
A simplified notation for (5) is:
Ax + By + a2 + C = 0
Dx + Ey + a2 + F = 0
Gx + Hy + a4 + J = 0
Kx + Ly + a4 + M = 0
This according to Cramer's Rule reduces to:
y = (A-D) (J-M) + (C-F) (K-G)
'(A=D) (L-H) + (B-E) (g£K)
x and y are then used in a forward geodetic position computation as described in
Bomford (1971) to determine the latitude and longitude of the actual position.
Implementation Procedure: A fortran program has been written which accomplishes
the following.
1. Transfers DMA Omega correction tables from DMA supplied tape to a high speed
drum.
96
2. Interpolates a correction to the hyperbolic rate based on month, day, time of
day, and approximate position.
3. Applies the outlined mathematical methods to determine latitude and longitude.
4. Uses the position in 3 as an approximate position to compute a new approximate
position for the next point, then goes to 2, and continues until all positions have
been computed.
RECOMMENDATIONS FOR BEST RESULTS
Assuming all Omega stations that can be received are broadcasting with the
same reliability*, the accuracy of a fix is a function of the network geometry.
According to Bigelow (1953) the accuracy of a fix is determined by the angle of
intersection of the lines of position. Bigelow says "those lines intersecting
with the smaller angle between 60 and 90 give strong fixes," and "a net should
not- be used where the smaller angle of intersection is less than 15 ". Thus
if one utilizes station monitoring information, DMA corrections, and network
geometry considerations, he can decide on the best networks to use.
GENERAL COMPUTER PROGRAM DESCRIPTION
The* Atlantic Oceanographic and Meteorological Laboratories, Physical Oceano-
graphy Laboratory program uses correction tables for time and location that are
made available by the Defense Mapping Agency on magnetic tape. The program works
as follows: A position is computed based on the hyperbolic- rates. This position
is then used as input to an interpolation routine along the time information which
utilizes the DMA tape to derive a correction to the hyperbolic rates.- These
rates are used to recompute the position. When there are many fixes to be computed,
the prior point to the one being computed is used as an approximate position for
the next calculation.
*Reli ability here means transmitters are functioning properly.
REFERENCES
1. Bigelow, Henry W., "Electronic Surveying: Accuracy of Electronic Positioning
Systems", Journal of the Surveying and Mapping Division, Proceedings of the
American Society of Civil Engineers, October 1963.
2. Bomford, G., "Geodesy", Oxford, Clarendon Press, 1971,
3. Campbell, Andrew C, "Geodesy at Sea", an unpublished masters thesis, the Ohio
State University Press, 1965.
4. Sodano, Emmanuel M.,' "General Non-iterative Solution of the Inverse and Direct
Geodetic Problems." Proceedings of the XIII I.U.G.G. General Assembly at
Berkley, Calif., 1963.
5. Defense Mapping Agency "Specifications for Omega", revised 9 January 1973.
97
_- . i_
HYPERBOLIC TRIAD
FIGURE
SLAVE
MASTER^
SLAVE
^a, (GEODETIC DISTANCE P'S.)
(COMPUTED USING A
SODONO INVERSE)
y ^ACTUAL POSITION
APPROXIMATE POSITION
GEODETIC DISTANCE
MATH MODEL GEOMETRY
FIGURE 2
98
15
Reprinted from: Marine Sediment Trans-port and Environmental Management ,
D. J. Stanley and D. J. P. Swift, editors, John Wiley and Son, Inc.,
Chapter 3, 23-28.
CHAPTER
3
Some Simple Mechanisms for Steady Shelf Circulation
ANTS LEETMAA
Atlantic Oceanographic and Meteorological Laboratories, Miami, Florida
An understanding of the mechanisms of sediment
transport on the continental margins depends on a
knowledge of the oceanic shelf circulations. This, in
itself, is a complex phenomenon, and its study has just
begun. Ideally, the geologist is generally not interested
in the totality of the circulation pattern. Although
sediment transport can occur at all levels, the primary
interest of the geologist is in knowing the magnitude
and direction of the flow close to the bottom; this can be
obtained either from observations or theory. However,
there is little of either. Bumpus (1973) describes what
is known about the circulation on the continental shelf
off the east coast of the United States. There is little
theory to describe these observations. The motions
appear to be complex and highly variable. Factors
that determine the circulation on one shelf likely are
not as important on another. Seasonal effects are
dominant.
Some of the simplest models for shelf circulation are
examined in this chapter. With these models, various
concepts about forcing mechanisms and the nature of
the dynamics can be explored. The relevance of the
model for the real world depends on how well its results
compare with observations. More complex models can
be developed by using the simpler models whose dy-
namics are well understood and hopefully verified by
observations, as building blocks. Realistic models
ultimately permit calculating the nature of the flow
close to the bottom, and determining which types of
forcing are of importance for sediment transport.
The simplest models to treat theoretically are steady
state ones. In this chapter attention is confined to these.
Following chapters discuss wave effects, tidal flows, and
other time-dependent phenomena. We start by exploring
several models that differ only in their forcing mecha-
nisms. The forces that are most important for steady
motions on the shelf are thermohaline effects and wind
stress. The former are caused by spatial differences in
the temperature and or salinity distributions. These
produce pressure differences that drive water motions.
Results will depend strongly on the value of the
frictional coefficients chosen. For very "viscous" models,
the flows are little influenced by the rotation of the
earth. However, for small values of viscosity, the solu-
tions are strongly influenced by the earth's rotation.
The magnitude of the viscosity is observationally
extremely difficult to measure, and perhaps is best
estimated from theoretical models.
THERMOHALINE FORCING
One example of thermohaline forcing is freshwater
runoff from land. This produces fresher and conse-
quently lighter water close to the coast than further
offshore. This also is the dominant driving force in
estuaries. To understand the general type of motion
this creates, consider the following problem. We assume
that the dynamics are governed by the following set of
equations:
dv
ds d*u
-J--z- = -gP-r- + A.—
dr dx Or3
(1)
99
23
24
MECHANISMS FOR STEADY SHELF CIRCULATION
bu_
bz~
= A,
b3v
bz3
0 =
bit
bs
u — h w
Ox
bz
bw
bz
= A',
bz2
(2)
(3)
(4)
For this right-handed coordinate system, x is perpen-
dicular to the coast, v is parallel to it, and z is vertically
upward; ;/, r, w are the v. r, Z components of velocity,
respectively; .1, and A, are the vertical eddy mixing
coeilicients for momentum and salt; s is the salinity; and
0 is the coefficient of contraction for salt. The effects of
the earth's rotation are contained in the terms fv2 and
///-, where / is known as the C'oriolis parameter and
/ = 20 sin 8 where it is the angular velocity of the
earth and 6 is the latitude.
To arrive at these equations it is assumed that the
effects of lateral mixing are small, and that the motions
are slow and steady so that nonlinear effects and time
dependence can be neglected. To simplify the problem
further it is assumed that the motion does not vary in the
direction parallel to the coast, i.e., d( ) by = 0. The
geometry is shown in Fig. 1.
At the coast there is a laterally distributed transport
of fresh water (river runoff) denoted by Tr. This sets
up a salinity gradient normal to the coast. What arc the
implications of this gradient? Equation 1 is
where
( h
dc
~f1'' = —gP** + Asu„,
( ), ( )„ = d- ( ),
In an estuary where there are lateral side walls, o is
either very small or absent. Similarly, on the shelf it can
be shown that in shallow' water, or for large values of
A, and A',., this is also true. In such situations the term
fv2 can be neglected and
vuzzz - g(3sx
U
:m/sec
-2.0
-1.0 0 1.0 2 0
1
1 1
1 1 V 1
.8 /
-2/h
.A
..2
FIGURE 2. Vertical profile of the offshore velocity for the fol-
lowing parameters: Av = 102 sec'1; g/3sx = 6.25 X 10~s; H =
5 X 10 cm; TK = 50 cm2 /sec.
Situations exist on the shelf or in estuaries in which
sx is essentially independent of z. In these cases the
solution for u is given by
gfah3
Av
J©'
16 \h
+
1
+ 3
As boundary conditions we have assumed that the stress
is zero at the sea surface (uz = 0 at z = h), the velocity
is zero at the bottom (a = 0 at z = 0), and the net
transport of water offshore is given by
u dz = TR.
This solution is illustrated in Fig. 2. Although there is a
net offshore transport of water, Tr, the magnitude of the
flow toward the source of the river transport is much
larger than Tr.
This perhaps is of importance for sediment or sewage
transport onshore. Similar velocity profiles are obtained
for more general and dillicult problems. The circulations
in estuaries are of this type. It should also be noted that
the inflow velocity is at least an order of magnitude
larger than the river outflow velocity Tr 'h, 0.01 cm/sec.
This type of problem becomes considerably more
dillicult when the distribution of s in x and z is solved
for simultaneously.
The earth's rotation can play an important role in
determining the nature of the flow. In the preceding
example by the choice of geometry (in the case of
estuaries) or by the assumption that the flow was very
viscous, this effect was not present. If the value of the
eddy coellicicnt is decreased or the water depth is
increased, rotational effects become important. Then
the solution, over most of the water column, becomes
FIGURE 1. Geometry for shelf models.
u = 0,
-&'£)
100
WIND FORCING
25
Close to the surface and the bottom this solution is
modified by friction. Note the differences between this
solution and the preceding one. The primary flow is now
parallel to the coast and northward. Only in thin
regions close to the top and bottom is there on- or
offshore transport (it is in these "boundary layers" that
the offshore transport of fresh water occurs). The most
likely direction for sediment transport is now parallel to
the coast. The primary factor that determines whether
this solution or the previous one applies is the value of .-1 ,..
Unfortunately this parameter is extremely dilhcult to
measure directly, and within the range of physically
possible .1 , either solution can occur.
WIND FORCING
Consider a situation where a wind stress is applied at
the surface, and l'n = 0 (sx = 0). The governing
equations are
1
—jv = px + AVUZ1
Po
fu = AvVzz
0 = -pz - Pog
0 = ux + ws
where p is the pressure. At the sea surface we apply a
stress p0Av(Pv dc) = r„. The boundary conditions are
dv du
at z = h, poAv — = Ty\ — = 0; w = 0
oz oz
at z '— 0, u = v = w = 0
For small values of A,, the solution consists of two parts.
In the interior of the fluid there is a "geostrophic" part
where
u = 0,
1
)
Po/
P.
p.W/
In addition, there are contributions that die away
exponentially from the surface and bottom that match
the interior solution to the boundary conditions. These
are called Ekman layers. They have the property that
the net transport in each layer is given by Tel = ryf,
which is independent of the eddy coeiheicnt. The
velocities are of the order of a few centimeters per
second for reasonable values of the wind stress.
The transport in the surface layer is to the right of the
wind stress, or offshore in this example. The transport
in the lower layer is equal to that in the upper layer,
but onshore. The detailed solutions for these layers are
given in textbooks on dynamical oceanography.-
We can understand this system of currents in the
following way. When the wind begins to blow, the upper
Ekman layer starts to transport water away from the
coast. This lowers the sea surface next to the coast and
creates a pressure gradient perpendicular to the coast.
This is balanced by a flow parallel to the coast [fv =
( 1 po)px]- Close to the bottom, friction acting on this flow
causes another Ekman layer to form with onshore
transport. The pressure gradient continues to build
until the transports in the upper and lower Ekman
layers balance. A steady state is then obtained. This is
the solution that was presented earlier.
In this problem the transports close to the bottom are
again directed onshore and would support sediment
transport in that direction.
The thickness (Del) of each Ekman layer is propor-
tional to (Ar /)1/2, and this depends on the square root
of the eddy coefficient. As the flow becomes more viscous
(i.e., A „ increasing) Del increases, and the upper and
lower Ekman layers merge.
When this happens the solution becomes (in the limit
of Del > h)
u = 0,
-(f)
This is known as Couette flow. Rotational effects are no
longer important and all the flow is parallel to the coast
in the direction of the stress. The velocities close to the
bottom are small.
In the examples presented so far, the magnitude of the
eddy coellicient plays an important role in determining
the nature of the solution. This is another reason for
dilhculty in formulating satisfactory models of shelf
circulations. The criterion which determines the nature
of the solution is the ratio of Del, the Ekman depth, to k,
the depth of water. When Del is less than h the flows
tend to be rotationally dominated. When Del is com-
parable to or larger than h, the effects of rotation diminish
or disappear. Consequently, the nature of the solutions
for a given value of Av can depend also on the depth of
the water. In deep water offshore, the solutions may be
rotationally dominated, whereas inshore, where the
water is shallower, they might become more Couette or
estuarine in nature.
The simplest possible effects of two types of forcing in a
simple model have now been briefly examined. As
should be obvious the problems can become extremely
complex even for this simple model, when the depth
varies, when sx is to be determined, and when rotational
and viscous effects are equally important.
101
26
MECHANISMS FOR STEADY SHELF CIRCULATION
A MODEL OF CONTINENTAL SHELF CIRCULATION
The ideas exploit d in the preceding sections can be used
to form a model of shelf circulations driven by freshwater
runoff from land and by wind stress (Stommel and
Leetmaa, 1972). As before a shelf of infinite length
() direction) and a semiinfinite width (extending from
the deep ocean at x = 0 to negative x— infinity) is
considered. The depth of the shelf is /;. A mean flux
of freshwater, Tr per unit length of coastline, flows
toward the sea, due to the cumulative effects of river
discharge along the coast. The steady wind stress com-
ponents at the surface z = h are tx and ry. The salinity
and density are related by p = p„( 1 + 0s). Assume
linear dynamics:
A ,fc™ - (3gsx - fr: = 0
A,V„ +f/f, = 0
Krs,. + \p:sx — \pxs: = 0
where the motion is independent of y; x derivatives of
diffusion terms are neglected because of the large ratio
of horizontal to vertical scales; and the stream function
\p defines the velocity components u = — i/-.-, ic = if/x. The
boundary conditions in * are that
at c = 0, \p = \p: = sz = v = 0
at z — h, \p = — TR\ —puA,\p:: = rx;
p{,A ,.!'- = Ty\ st = 0
This problem models wintertime conditions on the east
coast continental shelf of North America. Winter is
attractive because ( 1 ) density is primarily controlled by
the salinity distribution and (2) the weak vertical density
gradient in winter permits a simplification in the treat-
ment of the third equation above, which is nonlinear.
Even for this simple model, a complete solution is
dillicult. Instead the model is used to estimate the
natural horizontal scale length L = SqVx Vj where Vx is
the observed width of the shelf and Vs is the decrease in
salinity over that distance from the ocean value, sn.
The details of the solution are given by Stommel and
Leetmaa ( 1972). The solution for L for various values of
the wind stress and the eddy coefficient A ,, is shown in
Fig. 3, where E (the Ekman number) is equal to A , fir.
For these solutions, values of the parameters which are
appropriate for the eastern L'.S. continental shelf from
Nantucket Shoals to Cape Hatteras have been chosen:
/ = 0.7 X 10"4 sec"1, /( = 5 X 10:t cm, (3gs0 = 30 cm'
sec2, 'I'u = 50 cm2 sec, A, A',, = 1.
The upper curves of the diagram correspond to the
purely wind-driven regime. The lowest curve, which is
convex upward, is the pure density-driven model. For
IU'
\\T, * Ty = ' \ Tx - 0. Ty ' 2
\> \
V \
\> \
>> \
V \
L
cm
T. °\ \
T„ ■ ' \SN \
V ¥ \ * \
X \ X \
\ \ x \
\ \ * \
\ \ * \
v\ \\\
108
T. ■ 0.5 \. \ \ \
\. \* \
\. \* \
T> ■ 1. T„ ■ 0,_»\ \\ \
"~ ~" ^'vNaX
^^^ ^\ xX>\
^^ ^^+*S*<Si\
^^^^>\
x v ^^^.
107
0.01
0.1
E
FIGURE 3. Solutions fori* for different values of the wind stress
and the Ekman number (E = Av fH2).
large values of the mixing coefficient all curves coalesce
and the motion is basically density driven and of the
estuarine nature that was described earlier.
For small values of the vertical mixing coefficient,
with v stress predominant, the Ekman transports which
convect salt onshore and offshore are independent of Av
(as was pointed out earlier). However, vertical mixing,
A',., "short-circuits" these transports. This is propor-
tional to A „ since we assumed that the Prandtl number,
A , A',, was unity. Thus small values of Av correspond to
small mixing between the upper and lower Ekman
layers and large penetration of salt occurs (i.e., large L).
When there is no applied wind stress, the Ekman trans-
ports are driven by stress associated with the shear
produced by the horizontal salinity gradient, i.e., Avv2 =
(Avg(S j)sx. This diminishes as A,, becomes smaller, and
despite a partial compensation because k is also smaller,
the salt penetration diminishes. This accounts for the
different behavior for L (A ,.) for density as compared
to wind forcing.
For large values of mixing, as pointed out earlier, the
dynamics of the flow become nonrotational and also
vertical mixing is enhanced. Thus all the curves coalesce
in the purely salinity-driven case.
Attempts to compare this theory with the observations
are dillicult because a priori the appropriate values of Av
and k are unknown. It is assumed that Av/Kv = 1. The
observations in this area indicate that L is about 3.2 X 108
cm. Wintertime mean wind stress in this area can be
estimated from Hellermann's (1967) world charts. These
indicate that the magnitude of the x and^> components of
stress is about 1 dyne cm2. Thus with tx = tv = 1 and
L = 3.2 X 10s, Fig. 3 indicates that Av is about 37
cm2, sec. This is consistent with the observations. It
102
SUMMARY
27
should also be noted that values of /. as large as those
observed imply that the shelf circulation, at least for
this model, is basically wind-driven. As another test of
the model the difference between the salinity at the top
and bottom can be computed. This turns out to be
0.14r/f. Again this' is of the right order of magnitude
according to the observations.
Despite these limited successes of the model, there is a
serious discrepancy between the predicted v component
of velocity and the observed value. All the observations
indicate a negative v velocity of an order of 5 cm sec.
The theoretical v component is positive, about 20 cm sec.
If the observations arc correct, this indicates that this
simple model is not adequate to describe the observed
shelf circulation.
There Is some observational evidence to indicate that
there is a northward rise in sea level along the coast
(Sturges, 1974). If this feature is introduced into the
model, the discrepancy in the direction of flow parallel
to the coast can be resolved (Stommel and Leetmaa.
1972). However, the observations are not conclusive on
this point. The theoretical flow close to the bottom then
is in the right direction and is on the order of a centi-
meter or two per second.
CONCLUSIONS
As contradictions occur between model results and
observations, more details can lie added to the models.
At some point, however, the question has to be asked as
to how applicable steady state models arc to shelf
circulations and in particular to sediment transport.
Examination of daily wind records at Nantucket Shoals
light vessel shows that the wintertime root-mean-squarc
wind stress is 5 to 10 times larger than the mean. Thus
the transient fluxes are possibly an order of magnitude
larger than the mean ones. For sediment transport this
could be the dominant factor since the steady models
give rather low near-bottom velocities.
Better observations are needed to indicate the direc-
tion that modeling should go. Long-term series of current
and density measurements are needed to obtain an
observational verification of the mean fields and their
vertical structure. Time series of currents and density as
functions of depth are needed; without these the more
complex transient theories of shelf circulations cannot
be adequately attained. Finally, for the results of the
physical oceanographer to be of relevance to those
interested in sediment transport we need to know
whether the means or the transients are important in
sediment transport.
In this chapter the reader is introduced to some of the
problems facing a shelf modeler. Other more compli-
cated models exist that were not examined. Two of these
are the models by Csanady (1974) and Pietrafesa (1973).
Csanady discussed the barotropic (depth independent)
response of a shelf to an imposed wind stress or external
pressure gradient. Pietrafesa considers a steady state,
nonlinear, wind-driven model of an eastern meridional
coastal circulation. Both are considerably more complex
analyses than the one presented here. A more extensive
list of references can also be found in them.
SUMMARY
This chapter provides an introduction to steady state
models of the oceanic circulation on the continental
margin. Horizontal salinity gradients comprise a major
forcing mechanism for shelf circulation. A gradient of
seaward-increasing salinity will result in a seaward net
transport of surface water and a larger landward net
transport of bottom water, if the flow is relatively
viscous (low values for the eddy coefficient and depth).
With decreasing viscosity, the earth's rotation plays an
increasingly important role in determining the nature of
the flow . The primary flow tends to parallel the coast,
while onshore and offshore transport is confined to the
surface and bottom layers.
Wind forcing is effected by the application of wind
stress to the sea surface. For small values of the eddy
coefficient, the solution for the horizontal components of
motion occurs in two parts. There is a coast-parallel
"geostrophic" component of flow in the interior of the
fluid. Additional flow components are experienced at
the upper and lower boundaries, which die away
exponentially toward the interior of the flow. Net
transport in these Ekman layers is given by 7"el = tvJ,
where ry is the component of shear stress parallel to the
coast and / is the Coriolis parameter. Net transport is
independent of the eddy coefficient. The thickness of each
Ekman layer is proportional to the square root of the
eddy coefficient.
As wind-driven flow becomes more viscous (because
of increasing eddy coefficient or decreasing depth), the
upper and lower Ekman layers merge and the vertical
velocity gradient becomes linear in nature (Couette
flow). Rotational effects are no longer important, and all
flow is parallel to the coast in the direction of stress.
These relationships may be combined into a single
steady state model for shelf circulation. When applied
to the Middle Atlantic Bight of North America, the
model predicts an eddy coefficient of 37 cm2 sec, and for
the observed horizontal length scale, a primarily wind-
driven circulation. However, it is necessary to postulate
a northward rise in sea level along the coast, in order for
the model to predict net flow to the south, as observed.
103
28 MECHANISMS FOR STEADY SHELF CIRCULATION
SYMBOLS
.4,, vertical edd>- mixing coefHcient for momentum
Del thickness of the Ekman layer
E Ekman number
f Coriolis parameter:/ = 2Q sin d
h depth of water
A',, vertical eddy mixing coelhcient for salt
L natural horizontal length scale
p pressure
s salinity
T h river transport
u x component of velocity
v y component of velocity
w z component of velocity
x horizontal distance from origin perpendicular to
y
coast
horizontal distance from origin parallel to coast
vertical distance upward from origin
coeilicient of contraction for salt
0 latitude
r shear stress
p density
\p stream function
U angular velocity of the earth
REFERENCES
Bumpus, D. F. (1973). A description of the circulation on the
continental shelf of the East Coast of the United States.
Prog. Oceanogr.,(t: 111-158.
Csanady, G. T. (1974). Barotropic currents over the continental
shelf. J. Phys. Oceanogr. 4(3): 357-371.
Hellermann, S. (1967). An update estimate of the wind stress on
the world ocean, hlon. Weather Rev., 95: 607 -626.
Pietrafesa, L. J. (1973). Steady baroclinic circulation on a con-
tinental shelf. Ph.D. Dissertation, Dept. of Oceanography and
Geophysics Group, University of Washington, Seattle.
Stommcl, H, and A. Leetmaa (1972). Circulation on the conti-
nental shelf. Proc. Sail. Acad. Sci. U.S.A., 69(11): 3380-3384.
Sturges, \V. (1974). Sea level slope along continental boundaries.
J. Geophys. Res., 79(6): 825-830.
104
16
Reprinted from: NOAA Technical Report ERL 376-AOML 22, 10 p,
NOAA Technical Report ERL 376-AOML 22
^wsr^
A Comparison of
Satellite-Observed
Sea-Surface Temperatures With
Ground Truth in the Indian Ocean
Ants Leetmaa
Matthew Cestari
Atlantic Oceanographic and Meteorological Laboratories
Miami, Florida
August 1976
U.S. DEPARTMENT OF COMMERCE
Elliot Richardson, Secretary
National Oceanic and Atmospheric Administration
Robert M. White, Administrator
Environmental Research Laboratories
Wilmot Hess, Director
.CA-UT/Ov ,
Boulder, Colorado
105
NOTICE
The Environmental Research Laboratories do not approve,
recommend, or endorse any proprietary product or proprietary
material mentioned in this publication. No reference shall
be made to the Environmental Research Laboratories or to this
publication furnished by the Environmental Research Labora-
tories in any advertising or sales promotion which would in-
dicate or imply that the Environmental Research Laboratories
approve, recommend, or endorse any proprietary product or
proprietary material mentioned herein, or which has as its
purpose an intent to cause directly or indirectly the adver-
tised product to be used or purchased because of this Envi-
ronmental Research Laboratories publication.
106
CONTENTS
Page
1. INTRODUCTION 1
2. THE SATELLITE-OBSERVED SEA-SURFACE TEMPERATURE MAPS 2
3. SEA-SURFACE TEMPERATURE VARIATIONS ACCORDING TO SATELLITE DATA 3
4. COMPARISON OF SATELLITE DATA WITH SHIP REPORTS 5
5. COMPARISON OF SATELLITE DATA WITH 1963 SURFACE OBSERVATIONS 8
6. SUMMARY 10
7. REFERENCES 10
107
A COMPARISON OF SATELLITE -OBSERVED SEA-SURFACE
TEMPERATURES WITH GROUND TRUTH IN THE INDIAN OCEAN
Ants Leetmaa
Matthew Cestari
Daily worldwide sea-surface temperature maps are produced by
the National Environmental Satellite Service. For the first half
of 1975, sea-surface temperatures recorded on these maps were com-
pared with concurrent ship observations in the Indian Ocean. Addi-
tional comparisons were made with historical data. These show sys-
tematic differences between the satellite and sea-surface observa-
tions. The satellite-derived temperatures appear to be too low
along the equator and along the East African coast in the vicinity
of the equator. Furthermore, in April, May, and June the areas off
the equator (and not along the coast) appear to have temperatures
that are too high. Although the mean differences are not large
(1°-2°C), the fact that the errors vary in time and space made it
difficult to apply the satellite data for oceanographic interpre-
tations.
1. INTRODUCTION
Numerical experimentation has shown that the tropics are an important
area for interactions and feedbacks between the ocean and the atmosphere.
From present planning, it is clear that during the First GARP Global Experi-
ment (FGGE) equatorial regions will receive special attention in the ocean
as well as in the atmosphere. The Indian Ocean, because of the monsoons,
will also have a special observing period during FGGE, the Monsoon Experi-
ment (MONEX).
Because of the importance of equatorial regions to climatic studies,
and because FGGE will provide relatively complete meteorological coverage,
a group of oceanographers has started planning an Indian Ocean Experiment
(INDEX). The primary goal of INDEX will be to study the transient reponse
of a low latitude ocean to a strong regular forcing by the atmosphere. Pilot
experiments, whose results will aid in the design of the final experiment,
are now taking place. Sea-surface temperature maps from satellite data could
be a valuable tool to study the onset of the Somali Current, upwelling along
the Arabian coast, and heat budgets in the Arabian Sea. At the present time
such maps are available from the National Environmental Satellite Service.
However, as with e\/ery new product or technique, they have to be examined
carefully to ascertain their limits of accuracy and applicability. This
study reports on a number of intercomparisons between the satellite-observed
sea-surface temperatures and "ground truth" in the Indian Ocean during the
first half of 1975. The results suggest that more work has to be done before
reliable sea-surface temperatures can be obtained from satellites.
108
2. THE SATELLITE-OBSERVED SEA-SURFACE TEMPERATURE MAPS
The National Environmental Satellite Service provides daily worldwide
satellite sea-surface temperature (SSST) maps. This product is known as
the Global Sea-Surface Temperature Computation (GOSSTCOMP). One form of this
is an uncontoured computer printout with sea-surface temperature values for
each one-half degree of latitude and longitude. With each numerical value
for temperature is a code that indicates the estimated reliability of the
data. If the code is "+4" , then the last reading had been taken four days
before the date of the map, etc. If the number of days exceeds nine, the
code space is blank, and the temperature value given is from historical data.
If data are available for the day of the map, a letter appears in the code
space. An "+A" indicates that the temperature listed is an average of five
readings. A "+B" indicates an average of five to eight values and so on up
to "+H" which indicates that over 25 values were averaged. The better maps
in our analysis had mostly D's through H's associated with the temperature
readings.
For this study, the daily map with the highest code letter was selected
to represent an entire week. One day was chosen to be representative of a
whole week because changes from day to day were observed to be small, and
weekly representations were more readily compared than daily maps. They
start with the week of January 3-9 and end with June 1-7, 1975. Each map
selected was contoured in the area of the Indian Ocean off the coast of
Africa from 6°S to 15°N and 35°W to 65°E in latitude and longitude. From
the collection of maps, one was selected from the early portion of each
month to illustrate any monthly differences (Figs. 1-3).
50° E
IO°N
IO°N
FlguA.2. 1. SatdJUUitz Ana-AuAfiace. Ftau/i<£ 2. SatdUUXz. A&a-Au/i&ace.
tmpeAotuAe. data ^on. January 1975. tempeAotuAn data fati V nbtiuaAy 1975.
109
3. SEA-SURFACE TEMPERATURE VARIATIONS ACCORDING TO SATELLITE DATA
The seasonal variations of sea-surface temperatures in the Indian Ocean
is strongly related to the NE and SW monsoons, the transition periods be-
tween them, and the ocean current systems established by the winds. The
features observed on the maps must be interpreted in the context of these
phenomena. Figure 1 shows that there was not a wide range of temperatures
in January. Most of the readings were either slightly greater or less than
26°C. On either side of the equator the temperatures are somewhat warmer
than at the equator. During February (fig. 2), the sea surface immediately
north and south of the equator warms, while temperatures at the equator re-
main cool, as in January. North of approximately 8°N the temperatures begin
to decline with areas containing temperatures lower than 24°C. This is colder
than in January. In March (fig 3) the same pattern persists, but a warming
trend is evident. The area of the equator continues to remain cool, and the
areas immediately north and south of the equator (5°S-8°N) are warmer.
Larger areas of 28°C and higher temperatures are visible, with 30°C tempera-
tures reported for some locations. Temperatures for April (fig. 4) show an
increased warming trend, with many areas containing temperatures of 30°C and
higher.
50° E
10° N
10° N
VlaixAd 3. ScuteUUte. 4ea--6uAtfa.ee tern- F-tguAe 4. Sattttitz *ea-4uAtface tern-
pojia&viz. data Ion MaAch 1975. poAatuAn data tfoA kpnJX 7975.
110
Temperatures in the area of the equator, as noted in all previous
months, remain between 26°C and 28°C. Areas immediately north and south of
the equator have become considerably warmer. In the north there are isolated
areas with temperatures higher than 31°C. The warmest month from January
to June, 1975, is May (fig. 5). Again the area at the equator remains cool.
Large areas of 30°C and higher temperatures are visible north and south of
the equator. The north exhibits a slight cooling trend in June (fig. 6).
The equatorial band remains cool, and areas to the north and south become
cooler. Fewer and smaller areas of 30°C and higher temperatures are still
present, and the major portion of the entire region contains temperatures
of 28°C or slightly higher. The areas of 30°C and higher temperatures seem
to have moved toward the north and south, away from the area immediately
north and south of the equator. In all months, temperatures along the East
African coast were cooler than those offshore.
On first glance the seasonal variation in the sea-surface temperature
pattern as seen from these maps appears to be reasonable. The transition
period between the northeast and the southwest monsoon occurs during March.
April, and through the middle of May. A major factor in the heat budget of
the surface layers is evaporation, which is proportional to wind speed. Dur-
ing the transition, the evaporation decreases and the sea-surface temperature
increases. The cooler coastal areas could be related to upwelling or to
north-south transport of cooler water by the Somali Current along the coast.
A feature that is anomalous, however, is the cool band of water along the
equator. In the Pacific and Atlantic Oceans, such a cool band is indicative
of equatorial upwelling. However, in the Indian Ocean the winds are not
favorable for upwelling, and this feature is rarely present. To examine the
validity of this indication and others in more detail, a comparison was made
of these satellite data with data from a number of other sources.
50° E
--. IO°N
IO°N
VIqvjul 5. SaJtoJULLtd texi-Au/iiaae. torn- TIquJiz 6. SatoJUUte. beja-buxiaao, tem-
peAatusiz data ion Hay J 975. peAatu/ie. data, ion June J 975.
Ill
4. COMPARISON OF SATELLITE DATA WITH SHIP REPORTS
We can compare the satellite data with actual ship observations obtained
at the same time in the same area. For February through May 1975 data are
available from a chartered research vessel, La CutUzuaz, in the vicinity of
the equator. Bucket thermometer readings (estimated accuracy ±0.2°C) were
taken periodically along 55°40'E from 3°S to 2°N. The National Weather Ser-
vice also provides information on air temperature, dew point, and sea surface
temperature at ship positions through its twice-daily surface-weather maps.
There were 106 cases in which bucket thermometer readings from La
CuAi&uAe. could be compared with data from the satellite. The mean differ-
ence for the whole data set was +0.4°C. Temperatures recorded from the
ships were, on the average, higher. The standard deviation was 0.9°C.
This indicates that the scatter was quite large. The mean difference actu-
ally is rather small. However, if the data are studied in more detail, it
becomes clear that there are obvious trends. The satellite temperature data
for the equator are always lower than the surface observations and the differ-
ence becomes greater as time goes on. For example, for all intercomparisons
(32) in the region from 0.5°S to 0.5°N the mean difference is +0.9°C. The
standard deviation is 0.7°C. Clearly the equator is systematically colder in
the satellite data. This conclusion supports our previous speculations.
The satellite data were also compared with merchant ship reports. Un-
fortunately, there is a yery limited amount of ship data available in real
time. Also, frequently only the air temperatures are available rather than
the sea-surface temperatures. In the tropics this is not a serious problem
because the differences between these are usually small. Thus, to maximize
the data set, the ship reports were compared three ways. First the Satellite
Sea Surface Temperatures (SSSTs) were compared with the reported air tempera-
tures. The SSSTs were then compared with the sea-surface temperatures, and
finally with air and sea temperatures that were within one degree of each
other. Approximately 270 ship reports were available in the period from
January to Jiine 1975.
The difference between air temperature from ship reports and satellite
sea-surface temperature was determined from three 2-month groups. For Janu-
ary-February, the mean difference was +0.85. For March-April it was -0.59,
and for May-June it was -0.97. This indicates that the satellite tempera-
tures are lower than actual temperature measurements for January and February,
but higher than actual for March-April and May-June. These differences also
appear to have a geographic dependence. Figure 7 shows the geographical
distribution of these differences. For May-June the satellite reads low
in the vicinity of the equator and along the Somali coast, and high else-
where. March-April shows the same trend. In January-February the satellite
reads systematically low almost everywhere.
The difference between sea-surface temperatures from merchant vessels
and SSST was investigated for only May-June because not enough data were
available for other months. The geographic distribution of the differences
is shown in figure 8. Again the satellite appears to read low in the vicin-
ity of the coast and the equator and high elsewhere.
112
40° E
50'
60*
\-LU
1
+ 1.4
3
+ 1.23
1
-0.1
1
+ 1.0
2
+ 1.08
4
+0.7
1
+1.1
2
+27
+s.»
1
+1.3
2
+0.1
1
+ 1.0
i
+ !•
t
+2.05
1
-o.e
l
+1.4
I
+ 14
1
-o.2
1
+ 10
i
+3.0
1
+ 15
1
+ 1.0
2
+ 1 «
1
-1.5
t
♦s.os
1
-t.O
13
+0.13
II
fO.«4
IO°N
1
)
-S.i
1
-3.3
-l
4
S+0.&
1 *
-i
0
1
+ 1.3
i
+ 1.2
2
+ 0.2
1
+0.2
1
-0
i
45 -0
24
i -0.73
«
-0.0«
IO°N
January-February
March-April
\L_UW
— I.I
J ,
1
-2.6
2
-1.6
Si
^1.78
4
-1 OS
1
-1.2
^1.7
1
+ 1.1
2
+0»5
1
+0.3
1
+0.6
t
+2.7
1
+ 1.3
1
+0*
1
-l.«
1
-1.3
1
+ 3.6
1
+2.9
t
-IJB
1
-o.e
a
-i.i*
2
-2.1
1
-0.3
1
-1.0
i
-P. 5
4
-0.3
1
+ 0*
+ 1.2
4»
-1.7
2
-103
1
-3.4
IO°N
May-June
Figune. 7. GioQKa.pkid dLUtAibutlon oi di{i{i2Ae.nczA between
oJji tmp&icutivite and batdUUXd. i>ojx-i>uJi{cui<i twpeAatuAeA,
Thz uppeA numbeA In nach 6quaA2. <LndLc.cut<ii> the. numbeA o&
comp<viL£>oni> .
113
40° E
50<
60*
\|_UJ
|
1
-2 2
J
;
1
+ 0«
2
-2 1
&0.7S
*
-i 2a
1
-2 2
f+0.7
2
+045
1
+ 1.3
1
+ 1 «
1
+ 3.1
1
-0.7
1
+ 0.7
1
+oe
2
-1.2
3
-223
4
-I.I
1
-0.3
1
+ 10
-05
3
-0.17
tot
1
-0 2
i
-2.4
IO°N
F-iguAe. 8. GnognapKLc diA&UbLutLon o& di^&iznceA be-
tuxizn Ada-iuA^ace tempeAcutuAU {nam mzAchant i>hipi>
and tatoJUUXi 6<ia-6uA^ac<i tojapoJiatuJidi^ ^oK. May- June.
1975.
For a final intercomparison, we computed the monthly mean (SSST) for
April, May and June, 1975 for the areas: REGION I - 12°-13°N, 54°-56°E;
REGION II - 11°-12°N, 54°-55°E; REGION III - 10°-11°N, 53°-55°E (fig. 9).
We compared these with the 21-year mean temperatures in these regions as de-
rived from all ship reports on file at the National Climatic Center. These
were computed by Fieux and Stommel (1975). The results are presented in
Table 1.
It should be noted that the satellite data, except for one comparison
always give higher temperatures than the historical data do in this region.
In April and June the satellite temperatures are nearly always two standard
deviations away from the historical temperatures.
114
3°N
VIqvjkl 9. Reg-cow-6 faofi which compositions
weAe made between sateltlte bea-buA-
l0°N ^ace tempenatvJie data and ku> to ileal
6 kip data.
Table. 1. CompaA-Uon o& Average SSST' 6 with HlAtontcal Skip Vata.
SSST SHIPS
Apri 1
SSST SHIPS
May
SSST SHIPS
June
REGION
28.49 28.52 29-64 29-51 29-32 27-90 Mean Temp,
0.48
0.66
0.86 Standard Deviation
REGION
29.25 28.38 30.34 29.50 29.74 27.17 Mean Temp,
0.60
0.78
0.90 Standard Deviation
REGION III 29.93 28.84 30.82 29.43 29.49 26.65 Mean Temp
0.56
1.06
1.20 Standard Deviation
5. COMPARISON OF SATELLITE DATA WITH 1963 SURFACE OBSERVATIONS
From the International
sea-surface temperature for
surface observations from s
fig. 10) may *be compared wi
obvious difference is that
temperature minimum. Also
along the coast sea-surface
are in the satellite data,
that the International Indi
Indian Ocean Expedition, data are available on
1963 (Wyrtki , 1973). These were accumulated from
hips. Maps for January through June 1963 (see
th satellite temperature maps from 1975. One
in January-June, 1963, there was no equatorial
in the 1963 data, there is no indication that
temperatures are lower for January-April as they
Another difference between the two data sets is
an Ocean Expedition maps for April and May contain
115
50° E 60°
50° E 60°
^~yr>25
^r
C
**\
I0°N
January
February
I0°N
March
April
*v O A
^
^
May
I0°N
June
VZguAz 7 0. S£a.-.6uAf)<xce tzmpeAatuAz data ^on. 1963 obtcuimd hiom the. Iwtojt-
natlonal Indian Ocean Expedition.
116
temperatures of 30°C and slightly above as maximum temperatures for the area
being studied, while the satellite maps contain areas with temperatures above
31°C for April and above 32°C for May. The maximum temperatures are there-
fore higher in the satellite readings than in the 1963 International Indian
Ocean Expedition data. This, of course, might be related to year-to-year
variations.
The final major difference between 1963 and 1975 data is the north-south
temperature gradient for May. The satellite temperatures range from 27°C at
the equator to greater than 32°C at about 10°N along 55°E, while the Inter-
national Indian Ocean Expedition data range from 29°C to between 30°C and
31°C in the same area. Clearly the satellite temperature data have a wider
range. Satellite-measured temperatures are higher in warm areas and lower
in cool ones than temperatures measured at sea-surface.
6. SUMMARY
It is difficult to obtain reliable sea-surface temperature data for the
Indian Ocean. Potentially, satellite IR data could satisfy this need. How-
ever, the quality of these data has to be assessed before extensive use can be
made of them. Since ground truth is difficult to obtain in this region, we
tried to evaluate the SSST's by comparing them with data from a number of
independent sources. In each case, the intercomparisons showed serious dis-
crepancies between the satellite data and the ground truth. Although any one
result might be suspect, a clear trend emerged from the analyses. The SSST's
appear to be too low along the equator and along the East African coast in
the vicinity of the equator for the time period examined. Furthermore, in
April, May, and June in the areas off the equator and not along the coast,
they appear to be too high.
On the basis of this study, one would have to conclude that the SSST's,
at least for the Indian Ocean, are suspect. The errors appear to vary geo-
graphically in time and more work must be done if SSST's are to be of great
usefulness to oceanography. Meteorologically, improved SSST's are very impor-
tant because during FGGE, SSST's will be used in the numerical models. An
enhanced north-south temperature gradient in the models will probably create
unrealistic atmospheric circulation patterns in the tropics.
7. REFERENCES
Fieux, M. L. and H. Stommel , 1975: Personal communication. M.I.T., Cambridge,
Mass. 02139.
Wyrtki, K. 1973: Oceanographic Atlas of the International Indian Ocean
Expedition. NSF-I0E-1, Washington, D.C. '531 pp.
117
17
Reprinted from: Proc. of the Thirteenth Space Congress: Technology for the
New Horizon, 3-27 — 3-36.
THE STUDY OF OCEAN CIRCULATION FROM SPACE
George A. Maul
National Oceanic and Atmospheric Administration
Atlantic Oceanographic and Meteorological Laboratories
15 Rickcnbackcr Causeway
Miami, Florida 33149
ABSTRACT
Major ocean currents have surfarp manifpsta i i r _■ _ i
t:' . _,,„_ _,_, «... r ace manuesta- t0 level surfaces (.surfaces of equal geo-
tions that make them observable hv cmn>. _ » »■ ^^ ti • _ _• r __
r-rif* <:„„ = - i. i lu • sPace potential). The intersection of these
lit Lf «! : "nder.Cfcrtam conditions, densitv surfaces with the sea surface
!»v k. , /,"" lnaT ^f th° fol]cwjl>e marks the so-called cyclonic boundary of
„ ^ed }° ldentify the current's thc current, which is the left-hand side
;°„a{;, changes m sea surface tempera- facing downstream in the northern hemisphere,
ture, salinity, color (diffuse), sea state
ifSffJJ^'J* Se* sur£ace topography, wave Figure 1 schematically represents a
Jhl lnwiS -? EnS* a"d.r,odifi"tions to cross-section of a geostrophically adjusted
the lower atmosphere. Infrared sensors , T, . ? , ' '. .r
l_,,_ u-*.,, j R" sensois current. The view is downstream in the
have been used most extensively to study northern hemisphere. Several important
ocean circulation; however, new instruments features should be noted: The mean density
s^n.o^c PasSlvc and actlvc microwave (p) in the current is slightly less than in
"" 5°[ tCan SV'se temperature, salinity, the juxtaposed water to left. Typically
, ?t ' e; a"d surface topography, and the surface temperature (TJ in the current
J"!?"*?""1 vlsIble scanners and spectro- ranges fr0I- 2oC to 10°C warmer and salinity
raaioneters are providing new information rr n/ \ ■ ■, i mi ->n i _ ■ \
__ _____ „_-, ' j AUX"i> llL* i nunndiion (s o/00) 1S usually lo/oo to 2°/oo higher;
on ocean color and sea state Man ' s ml? *u ■ ■ i i ■ _ • t .. ■ j
_c ._ _!,___, . . ;>Lc»LC' 'wn s roie this is due to high insolation and evapora-
IL m Z ana pnotographer provides tion in the tropi2al source region of the
f f .P.J"1 resolution to dale for Stream. In terms of density, thermal ex-
describing visible changes across boundaries pansion is largeT than saline effects, and
as rfexl as sea and swell patterns. ~he average density of Gulf Stream waters
is less than the slope waters along the
INTRODUCTION
left-hand side looking north. From the
hydrostatic equation,
Ocean currents have several sea surface 0W
manifestations that can be used singularly u-. t °P
or in concert, to locate their, boundaries • «V> PQ C1)
Coastal currents typically have significant
energy at the tidal(l-2 cycles per day; cpd) where g is gravitv and p is pressure, it is
or local inertia] frequencies (0.5 - 1.5 cpd seen that the height of the sea surface
for mid-latitudes) . These frequencies are (H) is larger in the current when the in-
too high for the ocean's density field to tegration is to some deep base pressure,
adjust to the motion. Adjustment of the say p=2000 db (i.e. approximately 2000 m) .
density iicld provides the conditions by That is, there is a physical rise in the
winch many satellite 'sensing techniques sea surface of thc order of 1 m when cress-
can be employed in studying the ocean. ing into the Gulf Stream. The last feature
on this figure to be noticed is the hori-
lhc Gulf Stream off the east coast of North r.ontal velocity profile drawn at the top.
America is an example of a quasi-stationary In a geostrophically balanced system, the
current system that is well described by surface velocity (Vc) is given by
its density field alone. The density _
distribution defines that portion of the V„ - -- -°".
pressure field which is used to measure S f dX <>'•■'
the flow called a geostiophic current. where f is thc Coriolis parameter, and x
Frequencies associated with the boundary of is the cross-stream horizontal dimension,
this current arc approximately 0.25-0.1 cpd The horizontal velocity shear, 3V./3.X,
in the Straits of Florida (1) and 0.03 - becomes a valuable feature in the study
0.01 cpd in thc meander region off New of ocean circulation from space, because
tnglandU). Geostrophic adjustment associ- it has surface manifestations. More
ated with low frequencies requires that thc importantly, there is some prospect of
density surfaces are inclined with respect determining V's directly from remote sensing;
118
hrl-
G
o
in
r\
SURFACE VELOCITY
Ts= 21°c
Ss= 357^
^$=10245
o
U
Ps--1.02'l3
I-*— 10&m — ►-
I GULF STREAM
SEA SURFACE
GEOID
ISOPYCNALS
Figure I. Schematic cros a- section of a
western boundary current in the northern
hemisphere . Hove the exaggeration of
the sea surface as compared to the sub-
surface feazurcs, where a 102 scale
change is used for clarity.
one example of that will be discussed
later.
Asso
ther
whic
Deta
full
cont
euph
Thes
thet
bear
blue
tion
the
asso
prop
ter i
sedi
dele
wh i c
the
riated
e appe
h is a
i 1 s of
y unde
inuous
otic z
e nutr
i c org
ing pi
light
, shif
green,
ciated
c r t i e s
iig oi
ments
ct ed a
h also
water .
with t
a r s to
1 s o s k e
this i!
rstood ;
1 y b r i n
one a ] o
i e n t s a
ani s;;is
ants .
, and w
t the c
The p
o r g a n i
of the
light.
from th
long th.
modify
l-i)
he h o r i
be a v e
t c h e d o
pwel 1 in
howeve
g n u t r i
n g the
re u t ] 1
w h i c h a
Pigment
hen in
olor o f
hytopla
sms cha
water
Freque
c coast
e curre
the op
zon
rt i
n F
g m
r ,
erit
eye
izc
re
ed
suf
th
n k t
n^c
bv
n'tl
al
nt
1 1C
the
to
motion ,
c i rcu ] at ion ,
e 1. (■>)
n are not
effect is
to the
c edge.
photos yn -
r o p h >■ 1 1 -
culcs absorb
ent concentra-
a towards
nd other
opt ica 1
eased scat-
nt rained
ons can be
a
he
c r
e
undaries ,
properties of
When wind and waves run in opposition to
the current, the local sea builds higher
than when they run in the same direction.
Thus the sea "ay be higher or lower in the
current depending on the relative wind/cur-
rent directions. The latter feature often
translates into changes in white cap and
foam distributions, changes in glitter
patterns, and changes in surface wave
refraction patterns.
To reiterate, when crossing into the Gulf
Stream from the west, one typically encount-
ers an increase in temperature, salinity,
and perhaps sea state; the color shifts
from the green to deep blue, and the parti-
culate scattering decreases; there is a
rise in sea level due to steric conditions,
and a sudden increase in horizontal velocity
The detection of these features of the edge
of the currents from spacecraft is dis-
cussed in the following sections.
INFRARED SENSING
The ocean surface acts very nearly as a
blackbody. Its radiative behavior closely
3-28
119
follows Plank's Law with e?s;ii ssi v i t.i cs (c)
greater than 0.99. As j- consequence ol'
Kirchoff's law, this requires that the
source of the radiation he fro::, the upper
millir.eter. Evaporation and condensation
can thus play ;. rok in the radiative
temperature of water v.hich is frequently
0.5"C or so less than its thermodynamic'
temperature. I ° J . This is of fundamental
importance in det err.:ini:.p, '!' fro;:; space,
but for ocean current beu:- Jary determina-
tion, the observe! is looking for thermal
gradients rather than absolute temperatures
Thermal radiation leaving the earth is modi'
ficd by atmuspheri c absorption and emis-
sion. Clouds arc opaque to the earth's
radiation, which peak.- at about H)..:i. The
radiative transfer equation for spectral
thermal radiation is
Nn=N,
1-or
tran
subs
pher
top
the
pher
iii I c
but i
ncgl
is v
thi
smi
cr i
c ,
of
los
ic
gra
on
cc t
cry
tt;
pt:
re:
th<
s
ri-
ot
ed
s
formula
, p
and
t i v
o i
a r. s
L e i"
t;\
i n
mal
tmc
5 I! 1
m i 1
t i on , N is radiance,
i s p res sure , and t h
a are s u r f a c e a n d a
el)'. The radiance a
phere iNc, ) is the su
ace radiance due to
ance (\'s . -, ) plus
i cii i o i iiiu 1 a t es I he c
n o s o h c r c . S c a 1 1 c r i n
(3)
T 1 P
e
trios -
t the
!il of
atmos-
the
ontri •
o ] 5
se it
this formulation, he can
1 at there wave] cnglhs ( > ) .
One cons
is that
lessened
moist at
vapor ah
in T ' c a
Thi i" is
red re no
a r i e s .
diffcrcn
and the
infrared
equence oi
ocean surf
a t s a t e 1 1
mo sphere i
s o r p t i o n w
a be rcduc
a fundamen
tc sensing
In 1 o \\ 1 a t
cos across
a tmospheri
tcchni que
at
ace
itc
n t
and
ed
tal
o f
1 tu
00
c r.i
When
thro
view
to t
liqu
red
dept
clos
hi nd
sate
coas
ance
wave
ocea
that
and
iy,
c leu
rad
ugh
s c 1
he f
id w
phot
h.
e to
cran
Hit
tal
s .
long
n i s
oi
i n f r
then
d-fr
lation
a c 1 o u
0 U d ~ t O
act t h
a t cr ,
ons ar
I.ow cl
the o
cc to
c d a t a
lands
1 1 o w e v e
ths (-
an or
clouds
a red d
the p
e e can
t rom
a j
he
to
re
oc
dc s
c a;i
oi s
re
c r i
ma 1
tud
. 5 -
:; S
C l
r ic
n c
her
fr
re
m i t
tran sun s? ion
radients are
. lor a very
. 5 _n wa tor
difference
on in infra-
rent bound -
the the r in a 1
t s a re sma 1 1
high, the
p te
at c
and
e em
ouds
ccan
prop
S
can
r , i
0 . 0 u
der
and
at a
i c tu
be
a '
mper
loud.
as n
ittc
oft
1C V
ei l
ii,;i 1
have
n t ii
m) l
o f in
i a :i
are
re e
id en
surf a c c
nd the
u r e s . T
a r e m a d
e d earl
from a
have t
u e s a n d
e r d r c t a
ly', lots
e a r - o c e
longer
r e f 1 e c
n i t udc
I !.' bo
qui red
::.ents IV
ficd.
p a s s
space
his i
e up
icr,
v e r >'
emper
bene
t i o n
1 >■ in
anic
v l s i b
t a n c e
less
th vi
simul
li l c h
Dual
es
craft
s due
of
infra-
s ha 1 low
a t u l e s
e are a
cf
r a d i -
le
of the
than
s i b 1 e
t a n e o u s ■
a r e
channel
visible and
craft are pr
int crprct a t i
Figure 2 .
The left -ban
infrared I 1 u
Atlantic P. i g
Cape Matters
hand panel i
visible (0.6
1975, at 145
positive pri
are lighter;
print with 1
tops) 1 i giit e
Scotia is sc
cloud, but i
clear where
comparing th
seen to be c
Stream no a rid
surface t h e r
hints of the
image as v.el
the section
d i s c r i m i n a t i
visible clian
(approximate
highly absor
cloud deterin
synchronous
earth in the
30 minutes i
day. C9. 10)
3 0 - minute i
notion p i c t u
rates sevcrj
than ocean f
b e r e a d i ly i
t hernia] back
interval is
all f requeue
cpd can be 5
spatial r e s o
and the mid-
geometry are
sampling sys
infrared scanners aboard space
oviding the data for such
on; an example is given in
(8)
d panel of Figure 2 is an
.S - I2.5nmj image of the Mid-
ht showing the coastline from
s to Nova Scotia. The right -
'he simultaneously scanned
'c.7:im) image taken on 11 May
0 GMT. The visible image is a
nt , that is, higher radiances
the infrared is a negative
over radiances (i. e. cloud
r. The large feature off \ova
en in the visible panel to be
n the infrared alone it is not
the cloud-sea boundary is. By
e images, radiance patterns are
loud-free expressions of Gulf
ers, rings, eddies, and other
mal features. There arc some
se patterns in the visible
1. Tli is will be discussed in
on visible imagery. The cloud
on would be improved if the
ael were in the near infrared
ly lam) where the water is more
bing. An alternate method of
i nation involves use of the
satellites that observe the
same infrared band, but every
nstead of several times a
Using only infrared data, the
mages arc made into time-lapse
res. Clouds have advection
1 orders of m a g n i t u d e faster
catures, and therefore they can
dcntificd against the ocean
ground. Since the sample
3 0 minutes, ocean features of
ies down to the Nyquist of 24
tudied if detectable. The
lution of approximately 10 km,
latitude limits of the viewing
the limiting features of this
t em .
Inf
pri
Man
to
fil
Ian
res
pro
occ
ran
don
app
2
bla
syn
i s
for
abl
rare
m a r i
y oc
the
m an
d, a
olu t
cess
an ic
ge-
c , m
car
the
ck.
chro
be in
the
c .
d sat
ly fo
can f
match
d the
nd cl
ion c
ing t
r a n g
Mi en
os t c
whi te
land
A sp
nou s
g con
ocea
lite d
meteor
tures
n the
a d i a n c
ds. I
be o b
d a t a
is mat
his t y
uds sa
and, a
d vcrv
ially
tel 1 i t
r u c t c d
commun
a ar
ogic
e no
ay s
rang
rove
ined
ch t
ed t
of
rate
in t
a r m
sign
to
so f
y i.'i
e pro
a 1 pu
t emp
cale
e of
d oce
by s
ha t o
o the
proce
the
h e c a
water
ed de
accom
ilra 1
11 be
cess
rpos
ha s i
rang
the'
an r
peci
nly
gra
ss in
film
se o
app
vice
pi 1 s
oo p
come
ed
es .
;ed due
e of the
sea ,
ad i ance
ally
the
y scale
g is
and
f I" i g u r e
ca r
for the
h this
product s
avai 1 -
3-29
120
Figure 2. Infrared Cleft-panel) and
visible (right panel) image pair of
Gulf Stream off [Jew England. These
aata are from the very high resolu-
tion •scanning radiometer of the NOAA
polar orbiting satellite (8).
VISIBLE SEN'S l.\G
N0=NQ + r
Ns + °Nd
(4)
Detection of ocean
wavelengths (0.4-0.
currents in the visible
';ini) depends on the
change of ocean color and the change of sea
state associated with the boundary"; The
-spectra! radiance at the top of the atmos-
phere
by
in the visible region (Nn ) is given
O'
where N is the contribution of the atmos-
phere alone (most ly . Rayleigh scattering),
Ns is the contribution at the surface due
to reflection from the surface, N, is the
diffuse radiance at the surface due to
photons that have penetrated the surface;Y
and a arc atmospheric t ransmi ttance factors
3-30
121
for Ns and N'j respectively. Changes in sea complex patterns in the slope water off New
state, by which is meant changes in white York, are observable in both the visible
caps, foam, glitter, etc., enter the equation and infrared images. In the visible they
only through the N*? term. Similarly arc regions of higher radiance embedded in
changes in the optical properties of the a zone of low and probably uniform specular
water itself, winch is information on the return. This suggests that what is seen is
absorption and scattering of light by due to patterns in the diffuse radiance.
particles in the water, is represented by Thus the interpretation is that these
d' circulation features are being detected due
to variations in the optical properties of
The spectrum of diffuse radiation is a the water and are variations in biochromes
complex function oi scattering and absorption. and particulate scatterers . ( 13)
In a simple single- scatter ing model, the
independent variables were shown to be the The interpretation is difficult here
total attenuation coefficient, the total because only two channels of data are
scattering coefficient and the fraction of available. The 0.6-0.7 Mm channel is a
backscattercd light. ^ UJ Each variable is good -choice for visible radiance and was
also wavelength dependent, so that an the most useful in applying LAN'DSAT (Earth
infinite variety of optical conditions can Resources Technology Satellite) data to the
combine to produce the same Nd . In the marine environment . >-4J However, since it
case of current boundary determinations, is a valuable channel in the study of
the highly productive water along-side the optical oceanography, it is not the best
Gulf Stream cyclonic front is high in spectral interval for cloud and sea state
pigmented molecules as well as in particir descr imination . Experience with LANDSAT
late natter. The net result is to shift and the experimental scanner on SK'YLAB have
the peak of the upwelling radiance spectrum shown that 0 . 95- 1 . 05 vm is the optimum band
toward longer wavelengths, that is towards f0r this purpose. Incorporation of this
green colors. At the sane time, the opti- wavelength interval in future multispectral
cal intensity increases due to increased imagers is strongly recommended in order to
scattering. When these conditions hold, overcome the ambiguities of oceanic inter-
the water along the cyclonic side of the pretation in the present system.
stream will have a higher radiance in a
multispectral image such as on the NOAA MICROWAVE SENSING
polar orbiting satellite. In Figure 2 the
edge of the current can be seen in the Microwave sensing may be considered in two
visible imagery as well. as the infrared. ways: Active systems, by which is meant
The Gulf Stream appears to have lower radar type devices such as altimeters,
radiance than the slope waters in agreement scat terometers , and imaging radars; and
with the above explanation. passive or radiometer- type devices that
sense the emitted microwave energy in the
However, the S's term also contributs to same sense as infrared or visible radio-
NQ, and its behavior requires discussion. meters do. Many features of the edge of a
Reflection from the surface at these wave- current (see Figure 1) can be identified in
lengths changes the radiance spectrum in a microwave data, including sea state changes,
wavelength dependent fashion. In the 0.6- temperature and sal inity" changes , the
0.7um interval of Figure 2, a higher sea physical shape of the sea surface, and
state in the Gulf Stream could raise the actual current speeds. t 14 )
radiance to the sane level as the N'j
influence on the slope water (and thus Passive microwave energy is sensitive to
there would be no visible signature), or it changes in surface temperature, salinity,
could exceed the radiance and once again a roughness, and foam coverage, and to the
signature would exist. Kinds for this day presence of sea ice.1- J The transfer of
were from the southwest at less than 2 m s"1 radiation follows Equation 3, except that
due to a high-pressure ridge lying parallel the N«. term is a function of polarization
to the east ccast. Under conditions of as well as nadir angle and wavelength,
such weak and variable winds no foam or Since foam transmits energy from below,
white caps arc ant icipated ll 2 ' , and the transmission (t ), emissivity (e ) and
contribution to N0 is dominated by N'j, with reflectivity (p ) all contribute to N as
Ns contribution at a uniform value over the follows: s
whole image.
Some effects of what is a probable lack of Ns= T\ Nj "•" £\ NBB+ P\Nfl ("1
specular return can be seen in Long Island
Sound and south of Nova Scotia. These dark
areas arc interpreted to be zones of calm
seas where glitter does not reflect the
morning sun (P"?>0 F.ST). The major features The subscripts in this expression denote
of the Gulf Stream front, the large eddy- the blackbody radiance (BB), the incident
like structure south of Cape Cod, and the
3-31
122
radiation from tl
inc i dent rad iat ic
the foam. '!
microwave en
due to water
100 cm (50-0
r.i i s s i v i t y is
ic
c
V(
a tine splie re (a ) , ant!
f rom t he ocean ( i ) cm
atmosphere is opaque tc
y at wave I ei:;; t lis he low ] cm
por, -but between 1 c;:i an J
Gil.; the atmospheric trans-
r v " c 1 o s c to 1.0.
Measurements oft
lennodynam i c temperature
relate to N.,„ by
' lank ' s 1 aw. N..., can be
approxir.at e J by t
io Raylcigh-JcanS Law at
microwave wave 1 en
',ths, and this is proper -
tional to '!_. -V-l .
This s impl i f ies the
c a ?. c u I a t i o n s in t
iat the radiance is dir-
ectly proper Lion a
t o the first po w e r o f
the temperature.
Salinity affects ' , and
absolute measure!'!
:nts by aircraft of the
'oo arc be ins reported. (■■'■"J
order of l°/oo-2fJ
Salinitv deter."', in
. ny techniques measure
enissivity at 21 cm and require the ancil-
lary measurement of temperature. Salinity
effects on the dielectric constant of sea
water are small at 8 cm. Thus a two-
channel microwave sensor can provide salin-
ity and temperature measurements in the
absence of foam coverage (p=0 ) .
Sea state effects on an imaging radar are
shown in Figure 5. These7data from a side-
looking airborne radar1- ' show two
narrow lineatiens which were shown by
simultaneous infrared thermometers to be
associated with a large temperature grad-
ient in the Gulf Stream front region off
Cape llatteras. These lineations are changes
in the radar backscattering cross- section
which suggest changes in sea state asso-
ciated with the thermal change. This is
^
St rear: •'.••
oj tit j u:rJ/
<ao (17).
an be seen
in t nc ov i .i" ; ':
one application of active radar in locating
boundaries .
3-32
123
Another active
sion a] t imeter .
(on])r) measure
sea surface. I
area of i n 1: e r e s
eliminated, the
used to estimat
gradients at th
Figure 1 ) . Abs
will provide th
measure of surf
can be inferred
of caution must
ing the + 10 cm
meter (i.e. SEA
assuming that
(i.e. the width
following error
several latitud
radar s
It pro
of the
f the g
t , and
surf ac
e t he a
e sea
olute p
e ocean
ace cur
from i;
be men
accura
SAT) .
H = + 1
of the
s in V
es (♦)?
ystem is t
vides a su
topography
eoid is kn
tidal ef f c
c topograp
b s o 1 u t e p r
urface (se
res sure gr
ographer w
rents (on!
qua tion 2.
tioned her
c y of a r a
Using Equa
0 cm , x =
Gulf St re
are calcu
he preci-
b o r b i t a 1
of the
own in the
c t s are
hy can be
'. 30
e s s u r e
£
c again
a d i c n t s
ith a
y) which
A note
<
K
<
Z
ec
c concern-
T
O
d a r a 1 1 i -
u
tion 2 and
] 0 "' cm
am) , the
lated for
u
o
n
w
o
K
50"
4 0°
3&
20
1CP
V5
+ 8.8 cm sec
^10.5 cm sec
+13.5 cm sec
_+19. 7 cm sec
+38.8 cm sec
CURRENT SPEEO.V.
As can be seen, at the Gulf Stream Lati-
tudes ( $ = -10°) the error is 10° of the
flow; this in not an improvement over
conventional measurements. At lower
1 a t i t u d e s the error approaches values of
the order of S0» or higher. Other ocean
currents have typical speeds of 15-20 cm
sec l, and the error can be greater than a
factor of 2. One may argue that if a large
number of observations were taken, + 10 cm
would lead to a good absolute determina-
tion. 'This is only valid for a static
system, which the ocean is not. Baro-
tropic motions in the Gulf Stream at
inertial and tidal frequencies (1) would
require many radar altimeters in orbit at
the same time to begin getting the coverage
needed for such averaging. This somewhat
diminishes the enthusiasm for a + 10 cm
altimeter for strictly ocean current
determinat i on .
Studies of
show promis
problems in
particular!
that of cur
Figure 4 is
as a funct i
cross- sect i
wind speeds
dctermi ncd
ent tcchniq
inferred. A
determined
drifter t19
These and o
new and exc
the ba
e for
radio
>• into
rent
a plo
on of
on per
(K) .
throyg
lterna
by an
' , a r
t h e r m
iting
cksoattcr ing
appl i cat ion
oce allograph
resting appl
peed determi
t of current
radar backsc
unit area f
If the wind
h ^ne or sev
10) .then Vs
tivcly, if V
altimeter or
efinement on
i c r o w a v e tec
horir. on for
from radars
to many
y(18) .A
i cat ion is
nat i on (14)
speed (Vs)
a 1 1 e r i n g
or several
speed can be
eral independ-
c a n be
can be
s .
a L a g r a n g l a n
W is possible,
hninues are a
remote sensing,
Figure 4. Plot of current speed versus
radar backscattering as a function of
different wind speeds (12). Gulf
Stream current speeds are typically
1-2 -meters per second.
Photographic Sensing
Man ' s
appre
none
for t
servi
from
the r
hopef
futur
such
other
obser
Thesc
the d
resol
syste
An ex
Figur
turbu
of th
South
south
Curre
At Ian
the a
horde
sea .
m i x i n
role
ciated
so ful
h e fir
ng and
space .
eccnt
u 1 1 y w
e. Ma
as con
pstte
v e d an
data
etails
ution
ms .
as an o
from t
1 y as S
st t ime
p h o t o g
This
Apollo-
ill be
n >■ o c e a
f luence
rns tha
d photo
are pro
of the
unavail
bserv
lie ea
KYLAB
give
raplii
progr
Soyuz
s t r c n
n ci r
s , up
t ref
graph
v i d i n
ocea
able
er in sp
rlier mi
Astro
n traini
ng ocean
am was c
test p r
gthened
culation
welling ,
1 c c t cur
ed from
g new in
n at a s
by other
to
ace was
ssions , but
nauts were
ng in ob-
f eatures
arried on
oject and
in the
features
eddies, and
rents were
space L20?
sight into
patial
imaging
ample
c 5 w
lent
e Fal
Amer
crn h
nt/Gu
t ic
stron
r and
The
g acr
of
hich
mixi
klan
ica
emis
If S
The
auts
was
colo
oss
such d
is a
ng in
d and
phere
trcam
c on f 1 u
near
folio
r boun
the fr
etai
hand
the
Braz
The
anal
syst
ence
the
w e d
dary
ont ,
1 is
-hel
conf
il c
sec
og o
em i
was
Braz
for
sho
unl
give
d pho
luenc
urren
urren
f the
n the
firs
il/Ur
over
wed v
ike t
n in
tograph of
e region
ts off
ts are a
Labrador
North
t seen by
uguay
1000 km to
ery little
he eddies
3-33
124
Figure 5.
snrf-
,ij
of c
si otis cf the bozo:.:
Current taken fron
have beet: : ho to.j ra
by laboratory tc.ah
es tima tea
O DC
shown in
isure
The edd
be seen because plankton o
materials cause color v.iri
sea. The eddy appears to
and is poss'.ibly caused by
shear alon" the Falkland C
Photographic derails like this arc usual
chance happenings. In the case of SKYLAH
irbi: I ■
, >: t
r :■:: r c .
'if i ',
:,..
1 .<' (• < I a. :•
LA B ' '
~'i)
. Pciai
cr.
'": ; >:? cd
: e s .
y'r
c eddy
i n a
c i c r .
y i n
i"i;
ure 5
r s u s
p e :
ulcd
a t i o r
s
n the
be a n t i c y c 1 o n i
the v
eh
) C 1 t >•
u rrt n
t .
Ic
ly
'i
ii o w c v c r , 1 1 1 e s
several times on sever
and features such as t
not on I \ v i oil ed sevi i a
ecraft transited the area
were s t iu! i i
hart h
t
vo sou ice
con s ecu t i \e d a y
se in f i <: u re 5 w
times, but they
t he m.my device s i n t h
pe r i men t i'ac ka go .
ere
An important reason for the successful
observations was the real-time communica-
tions bet w eon the crew and occanog rap hers
at mission control. Future manned earth
observing missions must exploit this com-
munications feed-back even further, so that
ships and aircraft can conduct detailed
studies into these interesting new- found
features of the ocean's circulation.
Conclusions
Several techniques have been described that
offer useful application to the study of
ocean circulation from space. Each has
3-3-1
125
advant ag
es and disadvantages. The very
high spatial resolution of visible and
infrared
sensors is not practical at
microwave frequencies because of limita-
tions in
antenna design. On the other
hand, the microwave devices offer all-
weather
sensing capability but at reduced
radiant
resolution. The sens hip, of ocean
currents
should be approached with multi-
spectral
techniques using visible, infra-
red, and
microwave observations. This
will not
only provide the best opportunity
to o b s o r
ve the feature, but will also
increase
the degrees - of - freedom in an
automated identification scheme which is
an important goal for studying ocean
physics
from space
Acknowledgement: The author wishes to
express his appreciation to his colleagues
at AOMI, for assistance in preparing this
manuscript, and particularly to D.V.
Hansen with whom there have been many
fruitful discussions.
References
(1) Duing, W. (1975). Synoptic Studies of
Transients in the Florida Current, J. Mar.
Res. , 33(1) , pp. 53-73.
(2) Hansen, D.V. (1970). Gulf Stream
Meanders between Cape Hatteras and the
Grand Banks. Deep Sea Res . ,17, pp. 495-
511.
(3) Neumann, G. and W.
(1966)
Pierson, Jr.
iles of Phvsical Oceano-
graphy , Prentice-Hall, Fnglewood Lli'rts,
N.J., pp. 2 24-228.
(4) Maul, G.A. and H.R. Gordon (1975). On
the Use of the Earth. Resources Technology
Satellite (LANDSAT-1) in Optical Oceano-
graphy, Remote Sensing of Environ . , 4_, pp
95-128.
(5) Teague, W.J. (1974). Refraction of
Surface Gravity Waves in an Eddy. Scienti-
fic Report, U. of Miami, Coral Gables,
Fla'. , UM-RSMAS-No. 74034, 94 pgs.
(6) Ewing, G. and E.D. McAnster (1960).
On the thermal Boundary Layer of the
Ocean, Sc i once , 1_31_, pp. 1574-1576.
(7) Maul, G.A. and M. Sidran (1973).
Atmospheric Effects on Ocean Surface
Temperature Sensing from the NOAA Satellite
Scanning Radiometer, J. Geophys . Res . ,
28(12), pp. 1909-1916.
(8) Data from the NOAA-4 meteorological
satellite provided by H.M. Byrne (NOAA-
AOML) .
(9) Maul, G.A. and S.R. Baig (1975). A
new Technique for Observing Mid-latitude
Ocean Currents from Space, Proceedings,
Amer. Soc. Photogram. , Wash. D.C., pp.
713-716.
(10) Legeckis, R. (1975). Application of
Synchronous Meteorological Satellite Data
to the Study of Time Dependent Sea Surface
Temperature Changes Along the Boundary of
the Gulf Stream, Geophys. Res . Ltrs. ,
2(10), pp. 455 - 438.
(11) Gordon, H.R. (1973). A Simple Cal-
culation of the Diffuse Reflectance of
the Ocean., AgpJ^ Opt. , 12, pp. 2804 - 2805.
(12) Ross, D.B. and V. Cordone (1974).
Observations of Oceanic Whitecaps and
their Relation to Remote Measurements of
Surface Wind Speed., J. Geophys . Res . ,
79(5) , pp. 444-452.
(13) Compare with the interpretation under
different wind conditions given by Strong,
A.E. and R.J. DeRycke, Ocean Current Moni-
toring Employing a New Satellite Sensing
Technique, Science , 181 , pp 482-484.
(14) Parsons, C. and G.S. Brown (1976).
Remote Sensing of Currents Using Back-
scattering Cross-Section Measurements by a
Satellite Altimeter. (Submitted to J .
Geophys . Res . )
(15) Hanson, K.J. (1972). Remote Sensing
of the Ocean. In: Remote Sensing of the
Troposphere , V.E. Derr, ed., U.S. Gov't.
Printing Office, Washington, D.C., pp. 22-1
to 22-56.
(16) Thomann, C.G. (1975). Remote Sensing
of Salinity. Proceedings of the NASA Earth
Resources Survey Symposium, NASA TM X-
5816S, pp. 2099-2126.
(17) Data provided by D.B. Ross, NOAA-
AOML .
(18) Eisenbcrg, R. P. (1974). Practical
Considerations to the Use of Microwave
Sensing from Space Platforms. In: Remote
Sensing Applied to Energy-Related Problems ,
T.N. Veziroglu, FJ. , U. ot Miami, Coral
Gables, Fla., pp. S3-29 to S3-41.
(19) Molinari, R.L. (1973). Buoy Tracking
of Ocean Currents. In: Advances in _the
Astronaut ical Sciences , F.S. Johnson, ed . ,
Amer. Astron. Soc., Tarzana Calif., pp.
431-444.
(20) Kaltenbach, J.L., W.B. Lenoir, M.C.
McEwen, R.A. Weitcnhagen, and V.R. Wilmarth,
eds. (1974). SKYLAB-4 Visual Observations
Project Report, NASA, JSC- 09055, TM X-
58142, Houston, Texas, 250 pgs.
3-35
126
(21) Johnson, W.K.
A Mul t i spcct ral Ai:
bet Keen the Bra : j 1
from SKYLAB. Subm
of Knvi ron.
and U.K. Nor is (1976) .
.'.lysis of the Interface
and i'a 3 k] and Curv< nts
itted to Remote Sens inn
3-3G
127
18
Reprinted from: NOAA Technical Report ERL 378-AOML 23, 69 p,
NOAA Technical Report ERL 378-AOML 23
.uOMM»S^
NOflfl
An Experiment to Evaluate
WM SKYLAB Earth Resources Sensors
for Detection of the Gulf Stream
*>%£!?* &
George A. Maul
Howard R. Gordon
Stephen R. Baig
Michael McCaslin
Roger DeVivo
Atlantic Oceanographic and Meteorological Laboratories
Miami, Florida
August 1976
U.S. DEPARTMENT OF COMMERCE
Elliot Richardson, Secretary <?°^
National Oceanic and Atmospheric Administration
Robert M. White, Administrator %
Environmental Research Laboratories
Wilmot Hess, Director Boulder, Colorado
128
NOTICE
The Environmental Research Laboratories do not approve,
recommend, or endorse any proprietary product or proprietary
material mentioned in this publication. No reference shall
be made to the Environmental Research Laboratories or to this
publication furnished by the Environmental Research Labora-
tories in any advertising or sales promotion which would in-
dicate or imply that the Environmental Research Laboratories
approve, recommend, or endorse any proprietary product or
proprietary material mentioned herein, or which has as its
purpose an intent to cause directly or indirectly the adver-
tised product to be used or purchased because of this Envi-
ronmental Research Laboratories publication.
129
Page-
CONTENTS
ABSTRACT
1. INTRODUCTION 1
1.1 Background £ Purpose 2
1.2 Test Site 2
1.3 Prior Investigations 3
2. SURFACE TRUTH DATA 3
2.1 Cruise Report 4
2.2 Trackline Profile Data 4
2.3 Spectrometer Data 8
3. PHOTOGRAPHIC EXPERIMENT 11
3.1 Measurements 11
3.2 Data Analysis 12
3.3 Discussion 16
4. SPECTROMETER EXPERIMENT 17
4.1 Tracking Data 17
4.2 Infrared Radiance 18
4.2.1 Theoretical calculations 19
4.2.2 Comparisons of S-191 and models 21
4.3 Visible Radiance 23
4.3.1 Theoretical calculations 23
4.3.2 Technique for atmospheric correction 3 5
4.3.3 Recovery of R(A) from the S-191 data 36
130
Page
5. MULTISPECTRAL SCANNER EXPERIMENT 3 7
5.1 S-192 Data 38
5.2 Computer Enhancement 39
5.3 Discussion 44
6. SUMMARY 46
7 . ACKNOWLEDGMENTS 4 8
8. REFERENCES 4 8
131
AN EXPERIMENT TO EVALUATE SKYLAB EARTH RESOURCES
SENSORS FOR DETECTION OF THE GULF STREAM
George A. Maul
Howard R. Gordon
Stephen R. Baig
Michael McCaslin
Roger DeVivo
An experiment to evaluate the SKYLAB Earth Resources
Package for observing ocean currents was performed in the
Straits of Florida in January 19 74. Data from the S-190
photographic facility, S-191 spectroradiometer, and the
S-19 2 multispectral scanner were compared with surface
observations made simultaneously by the R/V VIRGINIA
KEY and the NASA C-130 aircraft. The anticyclonic edge
of the Gulf Stream could be identified in the SKYLAB
S-190 A and B photographs, but the cyclonic edge was
obscured by clouds. The aircraft photographs were
judged not useful for spectral analysis because vig-
netting caused the blue/green ratios of selected areas
to be dependent on their position in the photograph.
The spectral measurement technique could not identify
the anticyclonic front, but a mass of Florida Bay water,
which was in the process of flowing into the Straits
could be identified and classified. No calibration
was available for the S-191 infrared detector, so the
goal of comparing the measurements with theoretical
calculations was not accomplished. Monte Carlo simu- •
lations of the visible spectrum showed that the aerosol
concentration could be estimated and a correction
technique was devised. The S-19 2 scanner was not useful
for detecting the anticyclonic front because the
radiance resolution was inadequate. An objective cloud
discrimination technique was developed; the results
were applied to the several useful oceanographic channels
to specify the radiance ranges required for an ocean
tuned visible multispectral scanner.
1. INTRODUCTION
An important problem in physical oceanography is deter-
mining the boundaries of surface currents. Many techniques
have been proposed to study such boundaries from space, but the
actual process of extracting the correct information from
satellite data is in the early stages of development. SKYLAB,
with its several types of sensors (NASA, 1975), afforded the
means of testing three techniques simultaneously: photography,
spectroscopy, and multispectral imagery.
132
1.1 Background and Purpose
Major ocean currents are known to have several observable
surface features that make them distinguishable from the surround-
ing waters. The Gulf Stream system is used as an example to
typify these changes because it is one of the most important
ocean currents, and because understanding of its features can be
applied to the study of other current systems.
Because of its subtropical origin, the Gulf Stream is
typically warmer than surrounding waters and thus has a surface
thermal signature that often can be detected in infrared (IR)
imagery. The waters of the current are also much lower in
biological productivity and hence there are fewer particles and
biological pigment molecules in the Stream; this translates to
a deep blue color of water. Conversely, the juxtaposed water
masses are frequently higher in biological productivity, and
this can cause that water to be greener. Another feature of
the current that makes it visibly distinguishable is caused by
the large horizontal velocity shear. Frequently the faster
moving water in the current has a different sea state than
surrounding waters. Just as common are the many slick lines
associated with the shear. Finally, modifications of the at-
mosphere above the Gulf Stream, under certain conditions , can
also give an indication of the current's location.
Several other features of the Gulf Stream, potentially
detectable by satellite altimetry and other microwave techniques,
will not be discussed in this report. The approach here is
confined to visible and infrared wavelengths. The goal of this
experiment was to contribute to the determination of the loca-
tion of the Gulf Stream by visible and infrared measurements of
radiance .
1.2 Test Site
The site chosen for the experiment was in the Straits of
Florida along a suborbital track in the vicinity of Key West,
Florida. In this channel, the Gulf Stream runs approximately
perpendicular to the satellite ground track. This track would
maximize the changes in oceanic variables while minimizing the
impact on the data acquisition facility onboard SKYLAB. Further-
more, the logistics of obtaining the surface-truth data from a
6 5-foot vessel in January weather made the choice of a semi-
protected body of water mandatory.
Hydrographic conditions in the test site are controlled by
the location and intensity of the Gulf Stream (also called
the Florida Current in this vicinity). The cyclonic edge,
defined as the left hand edge facing downstream, has horizontal
excursions of approximately 50 km; that is, at some times of the
year the current's edge may be found 20 km south of Key West
and at other times 70 km to the south (Maul, 19 75). The location
133
of the cyclonic front determines the location of the major
hydrographic features of the Straits. Materials from Florida Bay
are also known to flow into the Straits and at times become
entrained in the current . Occurrences involving mixing of Gulf
Stream and Florida Bay waters are of fundamental importance to the
understanding of the dispersal of natural and man- introduced
materials .
1.3 Prior Investigations
A general review of remote sensing of ocean color was given
by Hanson (1972); Maul (1975) discussed the application of visible
spectroscopy to locating ocean current boundaries. Gordon, in a
series of papers (e.g. Gordon, 19 73; Gordon and Brown, 19 73;
Maul and Gordon, 19 75) discussed the spectra of upwelling irradi-
ance as a function of the optical properties of the water as
calculated by Monte Carlo simulations; those studies are directly
related to the current boundary location problem because the
spectrum of light changes from Gulf Stream water to coastal water.
Techniques for determining ocean chlorophyll (e.g. Baig and
Yentsch, 1969; Mueller, 1973; Duntley et al. , 1974) are also
related to current boundary determination because pigment-forming
molecules, along with suspended materials, affect the light
spectrum.
Remote sensing of ocean currents in the1 infrared region of
the electromagnetic spectrum has been attempted for many years
(e.g.: Warnecke, et al. , 1971; Hanson, 1972; Richardson, et al. ,
19 7 3). However, there have been questions concerning the radia-
tive transfer model dependency of the atmospheric correction
(Maul and Sidran, 19 72; Anding and Kauth, 19 72) that have awaited
SKYLAB to be addressed. Adequate atmospheric correction tech-
niques are required for ocean current boundary determination
using once-or twice-daily observations because compositing of
images is required in order to fill in the areas covered by
clouds; composites must be based on a common measurement, that
of the sea surface temperature itself.
SKYLAB provided the first opportunity to evaluate photon-
graphic, spectrometric, and multispectral imagery in a specific
experiment designed for current boundary location. It will be
seen that each instrument has unique advantages, disadvantages,
and limitations. It is the intent of this report to objectively
evaluate each technique and to provide recommendations for
future equipment and measurements.
2. SURFACE-TRUTH DATA
This section gives the details of how the ocean surface
data were obtained, calibrated, and analyzed. In many cases
surface optical measurements are useful indicators of the pro-
perties of the water that need to be measured. This is because
the theory is well ahead of the measurements, and adequate
134
instruments are not yet designed or built. In the case of
spectrometer measurements, the ideal observations are in fact
physically impossible.
2.1 Cruise Report
The at-sea observations were designed to provide simultan-
eous measurements of the ocean while the aircraft and satellite
transited the area. Since the speeds of the three vehicles are
mismatched, the assumption must be made that the oceanic con-
ditions are a steady-state for 12 hours or so. While it is
recognized that this is not strictly true, it is a necessary
assumption in view of resources available.
Underway operations on the Virginia Key included gathering
data on ocean salinity, chlorophyll-a concentration, surface
nutrients, seawater scattering properties, sea surface tempera-
ture by bucket and by a continuous radiometric profile, and
ocean temperature down to 450 m with expendable bathythermographs;
ship was hove-to for these spectrometry observations. Collection
of data started at 24° 39'. 1 N, 81° 08.1 W at 1253 GMT, 8 January
19 74. This point is about 7 km SSE of Marathon in the Florida
Keys. The track was directed SW and ended at 23° 33'. 2 N,
81° 55'. 5 W at 0150 GMT 9 January. This was 41 km off the north
coast of Cuba on the evening of the same day.
Weather conditions for the experiment were not ideal, with
partly cloudy skies and moderate seas. Wind was from 045° at
4 ms_l and remained steady most of the day. Air temperature
ranged from 23.8° to 26.1°C. Wet bulb values were 23.0°C most-
of the day. Visibility was 20 km except in a rain shower at
1700 GMT when it dropped to less than 5 km. Barometer was steady
at 10 2 5 mb until 19 0 0 GMT when it abruptly dropped to 10 2 3 mb
and remained so thereafter. Wave height was one-fourth meter
until 2100 GMT when it abruptly increased to one-half meter;
wave period remained 4 seconds throughout the day.
The Virginia Key traveled at 8 knots while on track. The
boat stopped only for spectrometry stations; all the trackline
profile data were collected while making speed.
2.2 Trackline Profile Data
Figure 2.1 is a plot of the trackline data, after reduction.
The stippled profile below the 22°C isotherm shows the bottom
profile along the track. The arrow shows where the 2 2°C isotherm
crosses a depth of 100m, a point interpreted to be approximately
15 km south of the boundary of the Gulf Stream (Maul, 1975).
The B(45) curves show the volume scattering function at 45°, and
are in units of meter"-'- steradian ""-*■ (m~l sr~l) .
A detailed discussion of the data collection, reduction, and
interpretation follows .
135
.34-
400
24° 39.1 'N
81°08.fW
23°33.2'N
81°55.5'W
Figure 2.1 Surface truth profiles across the Straits of Florida on
8 January 1974. The profiles 3 from top to bottom are: Continuous
ohlorophyll-a (mg m~3; discrete salinity (°/oo) : discrete volume
scattering function at 45° (m-lsr-1): discrete thermometric tempera-
ture (°C); continuous radiometric temperature in 10.5 - 12.5 \xm
band (°C) discrete depth of 22°C isotherm.
136
a) Chlorophyll-a CCL-a)
Cl-a concentrations were obtained continuously by measuring,
the fluorescence when Cl-a was exposed to blue light. The
data are reported as if all the pigment-forming molecules (in-
cluding pheophytins) were chlorophylls. The continuous record
was obtained by using a Turner fluorometer, Model 111, which
measured the fluorescence of surface water drawn through a
continuous-flow intake system. This method is as described
in Strickland and Parsons (1968), with the addition of a bubble
trap. In order to calibrate the continuous record, three dis-
crete samples were obtained by filtration and measured against a
known standard after the cruise. This also is as described by
Strickland and Parsons (1968), and uses the SCOR/UNESCO equation.
Table A. 1 in Appendix A shows the times and positions of the
three samples.
The large gap in the Cl-a curve on Fig. 2.1 is due to a
combination of drift on a particularly long spectrometry ob-
servation and a delay in turning on the fluorometer after leaving
the station. The three other breaks in the record represent a
change in fluorescence during stops for short spectrometry
observations. The degree of variability in Cl-a concentration
over the short distances indicated in the record shows the
desirability of a continuous record instead of discrete samples
as a source of the profile. The high values at the northern
(left hand) end of the line occur over the reefs of the Florida
Keys .
In addition to the three discrete surface samples, discrete
samples at various depths were obtained during stops at two of
the spectrometer stations. This was achieved by acquiring water
at the various depths for filtration and measurement later with
the surface samples. Times, depths, and positions are given in
Table A.l.
b) Salinity (S °/oo)
Ten salinity samples were obtained on the trackline. These
were surface water samples which were bottled for measurement
after the cruise. The times and positions of the salinity
samples are given in Table A. 2. The salinity profile in Fig.
2.1, which starts at 120 0 GMT, is a straight-line plot of the
ten values obtained.
c) Volume scattering (3)
The volume scattering function is a measure of the amount
of scattering at various angles by a sample of seawater irradiated
by a beam of light. In this case a single angle of 45° was
measured, for a beam with a blue filter (4 36 run) and a beam with
137
a green filter (546 ran) , with, a Br ice- Phoenix light-scattering
photometer. BC45) was calculated by using:
3(450) = a TD D(45°) x sin 45°
tt h DCOO)
where a is the ratio of the working standard diffuser to the
reference standard diffuser, TD is the transmittance of the
reference standard diffuser, h is the dimension of the irradiat-
ed element, D is the deflection of the galvonometer , and t is
the transmittance of the neutral density filters.
Measurements were obtained by collecting water samples with
PVC sampler bottles. Thirty surface samples were measured, and
at five stations samples were collected at various depths.
Values, times, and positions appear in Table A. 3. The curve in
Fig. 2.1, which starts at 120 0 GMT, is a straight-line plot of
the surface values.
d) Bucket Temperatures (T )
As is customary, bucket temperatures were acquired at each
XBT cast. Additional bucket temperatures were acquired at
spectrometry stations and at samplings for scattering measure-
ments. The curve in Fig. 2.1, which starts at 1215 GMT, is a
straight-line plot of the 2 8 total temperatures obtained. See
Table A. 4 for values, times, and positions.
e) Radiometric Temperature
A continuous sea-surface temperature profile was obtained
by using a Barnes, Model PRT-5, precision radiometric thermo-
meter. This radiometer has a special 10.5-12.5 ym filter that
approximates those in the SKYLAB multispectral scanner. The
instrument's voltage output was converted to temperature based
on a calibration performed in March 19 74. The profile in Fig.
2.1 begins at 1215 GMT. The large gap in the temperature profile
matches the gap in the Cl-a profile and exists for the same
reasons .
f) Expendable Bathythermograph (XBT)
The 23 XBT casts used the Sippican XBT system with 450-m
probes. These casts provided the information to plot the depth
of the 22°C isotherm. The point where the 22°C isotherm crosses
the 100-m depth (arrow in Fig. 2.1) is taken to mark the zone
of maximum horizontal velocity shear of the Gulf Stream. This
crossing happened at 23° 40.9' N, 81° 51.0' W, about 57 km
north of the coast of Cuba. See Table A. 4 for times and
positions of casts.
138
Position
Solar Zenith
angle
24° 38.9
N
81°
08.0
W
64°
24° 30.7
N
81°
16.3
w
51°
24° 19.1
N
81°
27.3
w
47°
24° 08.1
N
81°
34.6
w
56°
24° 04.0
N
81°
37.0
w
65°
23° 58.8
N
81°
40.4
w
76°
id
<D
>,
o
*
0
CO
rtf
id
Mh U
<+H k
0 -P
O -H
o
O
• Q)
• <U
O ft
O ft
S W
S to
3
1
3
2
3
2
4
1
2
0
3
1
2.3 SPECTROMETER DATA
The spectral signature of the ocean and the sky at various
points along the trackline was measured in the visible range of
light. The vessel was stopped for the spectrometry observations;
six observations were made to obtain a total of 2 5 spectra.
Table 2.1 shows the times and positions of these stations. A
detailed discussion of the instrument, data reduction and wave-
length calibration follows .
TABLE 2.1
Spectrometry Observations
Time (GMT)
1406
1603
1815
1956
2100
2210
a) Instrument
The instrument was a Gamma Scientific, Model 2400 SR,
spectroradiometer in a special water tight case. The Model
2400 SR scans in a wavelength range of 350 to 750 nm by
rotating a high efficiency diffraction grating that faces a
narrow aperture slit. It has a wavelength accuracy of +2 . 5 nm.
For this experiment the instrument was set to scan from 3 70 to
725 nm with a Wratten 2b filter installed over the entry slit.
This filter effectively cuts transmission below 400 nm, insuring
that the results will not contain secondary diffraction return.
An opal glass diffuser plate was used as a cosine (Lambertian)
collector; all measurements are irradiances. The data were
recorded on a dual-channel strip chart recorder.
b) Data Reduction
Fig. 2.2 is a typical spectral scan. It is a scan of ocean
upwelling light from the station at 2 210 GMT (see Table 2.1).
The dashed curve with the many peaks is the original unsmoothed
data. The peaks are due to changes in the angle of the water
surface relative to the instrument as waves pass underneath.
These changes impose a fairly regular periodic variation over
the general trend of the spectral return; all of the spectra
acquired on the cruise showed this variation to some extent. It
is necessary to remove the peaks if the data are to be useful.
139
Digital low-pass time series filtering techniques were used
to produce the smooth curve in Fig. 2.2. These techniques, to be
explained in some detail below, permit an objective filtering
of the unwanted periodicities with a minimum loss of significant
data trends. The filtering was performed on a UNIVAC 110 8 using
a FESTSA (Herman and Jacobson, 19 75) software system at the
Atlantic Oceanographic and Meterological Laboratories.
T
UJ
<
o
CO
>-
CC
<
or
CD
a:
<
UJ
o
<
<
300-
200-
100 -
400
450
500
550
600
650
700
WAVELENGTH (nm)
Figure 2.2 Example of upwelling spectral irradiance before (broken line)
and after (solid line) filtering to eliminate the effect of ocean,
surface glitter variations due to surface waves.
The spectra were digitized off the strip chart at intervals
of 20 points per inch; the points were sufficiently close to
retain the shape of the original traces. The trace in Fig. 2.2,
provided 248 data points. As different wavelength drive speeds
were used on different stations, this number varied from scan to
scan .
In order to choos
frequency energy with
is important to identi
signals. Fig. 2.3 con
relative strengths of
scan shown in Fig. 2.2
me try scan using Tukey
One can see strong per
30 data points per eye
signals that give the
appearance .
e a filter that successfully removes high
a minimum loss of significant trends, it
fy the periods of the high frequency
tains a plot (light, broken line) of the
various periodicities of the spectrometry
This is a power spectrum of the spectro-
's method (see Herman and Jacobson, 1975).
iodicities at approximately 6,7,9,20, and
le. These are the dominant high frequency
original scan in Fig. 2.2 its sawtooth
140
cr
<
CD
cc
<
o
o.
■ 100
0.80
o
<
or
'■'
060
UJ
7
o
Q.
(f>
0.40
UJ
cc
or
UJ
i-
020
u.
90| 30 20 1513|ll| 9 8 7 6 5 4°°°
45 12 10
SAMPLE INTERVAL
Figure 2. Z Tower spectrum of upwelling spectral irradiance shown
in Figure 2.2 The narrow line is the high frequency motion
caused by surface waves reflecting specularly; the heavy line
is the response of the Fourier filter to low-pass the data.
The response of the filter chosen to remove these high
frequency signals is shown by the heavy, solid curve in Fig. 2.3.
This response is in terms of a ratio of the contribution various
frequencies make to the form of the original trace, to the con-
tribution the same frequencies are allowed to make to the form
of the filtered result. Thus in the example chosen, no contri-
bution is allowed for periods smaller than about 9 data points
per cycle (10 nm) ; full contribution is allowed for periods
greater than about 9 0 points per cycle (10 0 nm) , and half
contributions are allowed at about 25 points per cycle (30 nm) ,
which was the longest period of the major peak in Fig. 2.3.
In general, all filters were chosen to remove the short
period (high frequency) signals in the same way. Power spectra
of different spectrometry scans did not always closely resemble
each other however, and each filter had to be chosen on the
basis of an individual inspection of each scan.
c) Wavelength Calibration
Irradiance and wavelength are indicated by separate voltage
outputs. The wavelength voltage is produced by a potentiometer
directly connected to the diffraction grating, voltage varying
with angle. In addition, an inscribed wavelength scale is
connected directly to the potentiometer. Thus it is possible to
compare the voltage of the wavelength output as recorded by
whatever strip chart recorder is used with the wavelength
indicated by the scale. Such a comparison was performed on the
strip chart recorder used during the cruise, and is the basis of
141
the wavelength calibration employed for these data. The drift-
free nature of the grating-scale design permits this type of
calibration.
Recorder outputs were compared with scale readings at 5 nm
intervals over the entire wavelength range ased in this experi-
ment. A third-order polynomial was fitted to the resulting
numbers to obtain an equation giving wavelength in terms of the
position within the spectra on the time axis of the strip chart.
This form of equation was chosen because it does not require a
fixed number of data points for all spectra; the only require-
ment is that the digitizing interval remain constant.
Comparison of this calibration with known spectral lines
indicates that the calibration is within 3 nm of true wavelength
over the whole range of visible light.
3. PHOTOGRAPHY
The use of color photographs to measure phytoplankton con-
centrations in natural waters is based on the argument that the
photographic material responds in a quantitative, reproducible
manner to the variance in the light field of the water. The vari-
ance is associated in a fixed manner with the concentration of the
phytoplankton, its distribution with depth, and its species and
nutritive history. For a number of years there have been appli-
cations of these variance techniques to remote sensing of the
ocean Ce.g. Baig and Yentsch, 1969; Mueller, 19 7 3). The SKYLAB
experiments provided the opportunity to extend some of the tech-
niques developed from laboratory tanks and low-level aircraft
flights to synoptic mesoscale coverage. The field program
(section 2.1) was to provide surface truth for the variance
analysis; the SKYLAB photography was to provide the photographic
products .
3.1 Measurements
A photograph is merely the record of the integral of the
intensity of the illuminant falling on a subject and the re-
flectivity of the subject. In a color photograph a third
variable is introduced in the spectral properties of both the
illuminant and the subject. A color photograph is satisfactory
if it produces, in a viewer's eye, a response similar to that
produced by the actual subject. The satisfactory spectral
response brought about by the color photograph is a result of
the eye's inability to distinguish between a pure spectral
source of light and a mixture of such sources. A color photo-
graph does does not produce in each pixel (picture element) the
exact spectral reflectivity of the corresponding spot of the
subject (with the exception of a subject that is itself a color
photograph) . Instead the photograph produces in each pixel a
mixture of three colors which the eye perceives as a single
color, and it produces only these three c&lors; this is called
a "metameric" match.
142
A color can be thought of as a vector in n-space, with pure
light as the origin. An infinity of coordinate systems may be
created around this vector, but once one of the coordinate axes
is fixed the others also become fixed. Common varieties of
color films need only three colors to reproduce the variety of
colors seen in the real world. To the eye these colors look like
yellow, magenta, and cyan (blue-green). These colors are the
coordinate axes of the color space . Every other color is an
unique combination of these three primary colors. If the
subject is composed of two different colors then the eye will
perceive it as if it were a single color. The eye in this
case performs the metameric match. A color photograph will do
the same. If the color of a subject is changing then the change
may be noted as differing quantities of the three dyes in a
color photograph of the subject.
The utility of monitoring the changes in dye concentration
of a color photograph can be carried a step further. It has
been shown that provided the continuous spectrum of the subject
has previously been measured, the dye concentrations can be
used to generate a new spectrum without re-measuring the
spectrum (Baig and Yentsch, 1969). The new spectrum must have
been part of a "training set" of spectra for which the color
photographs exist, or must be any combination of the original-
training set spectra. Then, through a multivariate analysis
and regression technique, a synthetic spectrum can be generated
using only the dye concentrations. The technique is especially
useful when the concentration of one of the components of a
mixture is changing. Tank (Baig and Atwell 19 75) and low- level
aircraft flights (Baig, 19 73) have amply demonstrated that
phytoplankton concentration in natural waters can be easily and
accurately measured with the technique. If the spectra of the
phytoplankton are not of immediate interest, then the multi-
variate reduction is not necessary. The problem then reduces to
a correlation between the concentration of phytoplankton in a
training sample and the variation in the dye concentrations in
the color photograph.
3.2 Data Analysis
The first photographs to be analysed were the 9-inch color
transparancies from the aircraft aerial cameras . On a number
of frames there were subjective color differences. However,
similar differences were noted on frames in which the color of
the subject area would have been expected to be uniform. A
densitometric analysis of one of these frames revealed a
variable blue/green ratio as a function of radial distance from
the principal point. These data are presented graphically in
Fig. 3.1. Attempts to use these data to "correct" data from
other frames of aircraft transparencies were not successful.
This vignetting problem was so severe that the images were
displayed on the non-linear portion of the film's D-log e curve.
This introduced an unknown non-linear error which could not be
143
corrected without an in-flight calibration with targets
known spectral response.
of
432101 2345
CENTIMETERS FROM CENTER OF PHOTOGRAPH
Figure 5,1 Example of the. effect of vignetting in aircraft film
data on the -percent transmittance of blue C450nm) and green
(550 nm) light > and on the blue /green ratio. Aerial color-
positive film was used in the RC-8 camera.
Because the scale of the satellite photos is so much smaller
than that of the aircraft photos , discontinuities such as fronts
and eddies are recorded in only a small area of the photo. By
comparison, similar elements have to be recorded over large areas
of single aircraft photos, or even in a sequence of such aircraft
photos. The practical effect of this scale difference is that
variations in film density related to position on the photo can
be ignored where the area of interest covers only a small area
of the photo. Of course comparisons between areas widely scat-
tered over the photo are still subject to the problem of spatial
density in the photo.
All of the pertinent SL-M- duplicate films were analysed on
a hybrid transmissometer. The light table and associated aper-
tures, filters, and diffuse acceptor are from a Welch Densichron
densitometer. The light sensor and associated electronics are
a Gamma Digital Photometer. The transmission of each of the
three color filters and of the white light setting was calibrated
with a non-silver standard step wedge which is traceable to N.B.S.
standards. Such a step-wedge is a better approximation to the
actual attenuation characteristics of color film than the usual
silver grain step wedges. This is because color films do not
have any silver in the final image, depending instead on dye
144
*£ 83d *TS 08P V8VN
ZIC-B8
Figure Z. 2 S-190B panchromatic (SO-022) photograph of the Straits of
Florida and the northern coast of Cuba. The box in the lower left
brackets the anticy clonic front of the Gulf Stream, and defines the
area where densitometric measurements were made.
145
rm oat vsvn
♦IE-68
Figure 3.3 S-190B panchromatic (S)-022) photographs of the Straits
of Florida and the western Florida Keys. The box in the lower left
brackets a plume of water from Florida Bay3 and defines the area
where densitometrio measurements were made.
146
densities for attenuation of the illuminant . Data in the follow-
ing paragraph are reported as the percent fraction of transmitted
light rather than as density, since transmission ratios will
have more meaningful interpretation than density ratios. Stand-
ard deviations (a) follow each ratio.
The first area analysed was a plume of water off the coast
of Cuba, in frame 97, roll 64 SL-4, near reseau #8 (Fig. 3.2).
The average Blue/Green transmission (B/G) ratio is 5.9, cr* 0.1.
Just to the left of the interface of this plume with the Gulf
Stream, the B/G ratio changes to 7.0, a+_ 0.1 in Gulf Stream
water. In frame 98 a similar plume on the Florida Keys side
of the Gulf Stream has a B/G ratio of 6.0, o +_ 0.1, while the
Gulf Stream water immediately to the right of the plume has a
B/G ratio of 6.8, a + 0.1 (Fig. 3.3). This particular plume
shows in frames 97, 98, and 99. The B/G ratios for the plume
are 5.5, 6.0, and 6 . 2 Respectively ; the B/G ratios for the
Gulf Stream water adjacent to the plume are 6.3, 6.8, and 7.1,
respectively. Thus, while the absolute values of the ratio are
changing, the difference between the plume ratio and the Gulf
Stream water ratio is nearly constant from frame to frame.
Both filtered panchromatic films SO-0 2 2 showed some apparent
density changes in the same areas as those in roll 64. Roll 65,
filtered to pass 0.6 to 0.7 micron light showed a transmission
change from 40.0x10"! to 34.0x10"-'- on going from the plume off
Cuba to adjacent Gulf Stream water. Roll 66, filtered for 0.5
to 0.6 micron light showed no transmission change between these
two areas, both noted as 34xl0~l.
Neither of the two b/w IR films showed any transmission
differences among the areas analyzed. The color IR film did
show some differences that are considered to be statistically
significant. The plume off of Cuba showed a B/G ratio of 7.0,
while adjacent Gulf Stream water showed a ratio of 7.5. The
plume off the Florida Keys showed a B/G ratio of 8.5, while
across the interface of the Gulf Stream the water showed a
ratio of 7.2. It should be noted that the B/G ratio of the
color infrared film is really a ratio of the green to the
visible red radiation.
The conclusions that can be drawn from this limited data
set are that a significant variation in ocean color can be
observed by changes in dye concentrations in color photographs
of the scene. When the surface truth is considered the evidence
tends to favor the variation in suspended chlorophyll as the
most probable cause of the color variation.
3 . 3 Discussion
It is immediately apparent from the data that the transmission
ratio technique is a useful means of analyzing variations in color
of satellite-derived photography. At the same time the data
147
might have been more useful had certain precautions been taken.
Reference is specifically made to the aircraft-derived photography.
To achieve a flat spectral response across the film the associated
optics should have been fitted with anti-vignetting filters. The
space craft cameras suffered to a lesser extent with the same
problem. In the latter case each of the associated optics had
been calibrated so that the error in transmission was known.
There is however, no indication that such care was taken in prep-
aration of the subsequent duplicate images. While care was taken
to ensure that duplicate grey scales were reproduced at the same
levels as those on the on-board films, apparently no account was
taken of the variation in illumination across the print head of
the printer. All of these problems taken together substantially
reduce the possible intercomparisons that might have been attempt-
ed.
The photographs in Figs. 3.2 and 3.3 are both S-190B products
that have been enhanced by printing on high-contrast film. Ex-
posure levels were set to saturate the details in the non-oceanic
features. This is a trial and error technique that extracts
markedly more low radiance level information. The change in
texture marking differing sea states in Fig. 3.2 is not measurable
by the densitometer technique, but it is clearly noticeable to
the eye. The boundary between the two levels of radiance is
probably the anticyclonic edge of the Florida Current. The de-
tection of the features in Figs. 3.2 and 3 . 3 by the S-192 scanner
is discussed in section 5.
4. SPECTROMETER EXPERIMENT
The SKYLAB S-191 steerable spectroradiometer was to be used
in this experiment to study changes in the visible (0.4-0. 7ym)
and infrared (7.0-14.0 ym) spectra of the ocean across the
current's cyclonic boundary. The plan was for the crew to acquire
a cloud- free oceanic area with the S-191 looking 45° forward of
nadir, and to track that site until 0°. Thereafter, the spectro-
radiometer was to be locked into nadir viewing across the cyclonic
front and up into the waters of Florida Bay.
The experiment proved to be not very successful for several
reasons: although the crew did as the plan said, the data acqui-
sition camera (DAC) was turned off and the exact tracking data
(angles, times, locations) were never recorded; the calibration
of the S-191 infrared detectors is not known; the visible region
data radiance values do not agree with theoretical or observed
values reported by other investigators .
4.1 Tracking Data
Location of the data was made difficult by the lack of DAC
output. The voice log was the best clue to what actually was done
by the crew. According to the transcript of the voice tape, at
148
16:29:33 GMT the pilot had the S-191 set at 35° looking forward
along the track, although, all :rew instructions were to set the
S-191 target acquisition at 45°. The word "thirty-five" was not
clearly audible however, and the pilot may have followed the in-
structions sent up to SKY LAB just prior to the pass. At 16:29:35
GMT, the pilot reported tracking a clear area of water; the
assigned start time was 16:29:33. The exact time of reaching
nadir is difficult to tell from the voice log however, 16:30:45
is the approximate time.
The location and time of the nadir point were calculated from
geometrical considerations assuming a spherical Earth with a ra-
dius of 6 378 km and a satellite altitude of 443 km. If the nadir
angle was 45° at 16:29:33 GMT, the position of the point tracked
was 2 3°53'.4 N, 81°5 8'.0 W; the time of arrival of the spacecraft
over this point from the best available positioning data was
16:30:41.1 GMT, which is in good agreement with the voice log
estimate. The message sent up to the crew had the finish of the
tracking at 16:31:05.
The S-19 2 line-straightened data show that the position given
above was in the middle of a clear ocean area and it appears
reasonable to have tracked this as the site. The vehicle was over
Florida Bay at 16:30:55 according to the S-192, but the pilot
commented at 16:31:10 that they were going across the Keys. This
discrepancy cannot be accounted for unless the Florida Keys were
observed well after the spacecraft transit.
If the above analysis is correct, then according to the sur-
face truth data in Fig. 2.1, the S-191 probably never acquired
data from the Gulf Stream. The position of the 22°C isotherm at
10 0 meters depth indicator was 2 5 km SW of the point where the
pilot tracked a clear area. Although the exact location of the
front cannot be identified in the ship track data it appears that
it was also SW of the nadir tracking point. Maul (19 75) reported
the mean separation between the indicator isotherm and the front
to be 11 km in this area, and that further supports the contention
that the S-191 did not obtain spectra in the Gulf Stream. The
objective of analyzing the change in spectra across the front can-
not be accomplished with these data.
4.2 Infrared Radiance
The infrared experiment was designed to study the accuracy of
atmospheric transmission models. This objective could not be
accomplished because the calibration of the infrared detector is
an unknown function of wavelength (Barnett ,NASA-JSC personal
communication; Anding, and Walker, 1976). Several relative tests
were made however, which provide some information on the atmos-
pheric transmission model dependency, and these will be discussed
below.
149
4.2.1 Theoretical calculations
Emitted infrared radiation (7 ym<_A<_14y) leaving the Earth
passes through the atmosphere before detection at the S-191
sensor. The atmosphere modifies the infrared radiation by ab-
sorption and, to a very minor degree, by scattering. Details
of the theory are given by Chandrasekar (1960) and recent reviews
on its application to oceanography are given by Hanson (19 72)
and Maul (1973). The radiative transfer equation through an
absorbing but non-scattering atmosphere is:
r
*/
N(6,A) = e(9r , A) L(T,A) t(6,A)
P
s L(Ta, P,A) 9-r(p,6,A) p
9P
o
p(6' ,A) N (G",A) t(6,X) (4.1)
as
where 9,0', A" are the nadir angle, angle of reflectance, and
angle of incidence, respectively. Radiance (N) at the satellite
is wavelength-dependent, and is a function of the surface black-
body radiance (L), the emissivity of the surface (e) and the
transmittance of the atmosphere (t); these three _ parameters
describe the absorption of emitted blackbody radiation by the
atmosphere. The second term in the equation, the integral term,
describes the atmospheric (a) modification of the radiance as a
function of pressure (P). The third term describes the contribu-
tion of the reflected (p) atmospheric radiance at the surface
(N ) , again as modified by transmittance.
as
The theoretical calculations discussed herein are an ex-
tension of the model used by Maul and Sidran (19 73) which uses
the transmissivity data of Davis and Viezee (1964). The area of
interest is the 10.5 - 12.5 ym band that is used on many space--
craft including the SKYLAB S-19 2 multispectral imager. In this
spectral interval e>0.99 at low nadir aggies; hence p(=l-e) is
very small and equation (4.1) may be written
N(6) = 4>(A) L (Ts,A) x(9,A)dX
00 ps
U
(A) L (Ta, p,X) 3r_(p,6,A) dpdA (4.2)
'o /0 ap
The filter function (<j>) is zero outside the interval discussed
above. The radiance may be converted to equivalent ^blackbody
temperature by inverting the Planck equation (L) which has been
integrated over the same 10 . 5<_<j><_12 . 5 interval.
The calculations were carried out on the AOML computer. A
special radiosonde was released by the Key West office of the
150
National Weather Service at tha time of SKYLAB transit Csee Fig.
4.1). Before the radiative transfer from a radiosonde is computed
the data must be inspected to insure that no clouds are in the
path of ascent, in order to compute a cloud-free radiance. Clouds
are ^ readily identified by their characteristically high relative
humidity and _ isothermal temperature. There is evidence of clouds
in the data in Fig. 4.1, so calculations were made to test the
effect of clouds.
RELATIVE HUMIDITY (%)
0 20 40 60 80
AIR TEMPERATURE
1000
100
-I —
-80 -60 -40 -20 0 *20
AIR TEMPERATURE (°C)
►40
Figure 4.1 Vertical ■profiles of atmospheric pressure and rela-
tive humidity taken at the times of SKYLAB transit. The dotted
lines on the relative humidity profile are the oloud-free
estimate of atmospheric moisture.
Two cloud layers are in evidence, one centered at 744 mb
and one centered at 6 71 mb . Clouds are characterized by a sudden
increase in relative humidity and a small (near zero) lapse rate.
An equivalent clear sky estimate is made by assuming the clouds
are absent; the estimated relative humidity profile in the clouds
region is given by the dotted curve. The calculated equivalent
blackbody temperatures for TQ 2 9 8.15°1C are:
Wavelength Observed (Appendix B)
Cloud-free Equivalent
llym
12. 5ym
293.22°
290.46°
K
K
293.85° K
291.28° K
The differences in this case are small, 0.6 3°K at llym, and 0.8 2°K
integrated over the 10.5-12.5ym region where the S-19 2, NOAA-4/5,
and SMS-1/2 observe. Other experience with this type of cloud-
free equivalence has been as high as 5°K over the Gulf of Mexico.
151
4.2.2 Comparison of S-191 and models
As stated in section 1.3, the wavelengths chosen for the
two-channel technique, CAnding and Kauth., 1970) of atmospheric
correction depend on the radiative transfer model. SKYLAB was
to be used to study that question but since the calibration of
the S-191 infrared detector is unknown, the problem cannot be
investigated.
The mean sea surface temperature along the trackline was
2 5.0°C. This value has been used in the calculations shown in
Fig. 4.2. The Davis and Viezee (1964) model does not include
absorption due to the ozone molecules which show up as a maximum
at 9.6ym in the S-191 observation. The comparison shows that
ozone does not affect the 10.5-12.5um window and hence is not a
factor in the S-19 2 infrared scanner data. At 11.0ym, the ap-
parent difference between the observed and calculated equivalent
blackbody temperature CTgg) is 3.5°C. This seems to be the
approximate error estimate of other SKYLAB investigators (personal
communications), but no conclusions can be drawn.
CALCULATED
7.0
8.0 9.0 100 1 1.0 12.0 13.0 14.0 15.0
WAVELENGTH (/im)
Figure 4.2 Spectral infrared radiance observed by the S-191
spectroradiometer (dashed line) 3 and calculated by the ozone
excluding model of Davis and Viezee (fine solid line). Heavy
solid lines are blackbody curves.
Since the calibration uncertainity is wavelength-dependent
the data at llym were studied to determine the shape of the nadir
angle dependence curve. In the lowe^ half of Fig. 4.3 is a least
squares fourth-order polynomial fit to all the observed radiances
as a function of time (dashed line). Since the same ocean spot
was to be tracked, radiance should be a function of nadir angle
up to 16:30:45 GMT. The maximum on this curve (arrow) is at
152
16:30:19. The upper curve is the theoretical calculation using
equation 4.2 for the same atmosphere in Fig. 4.1 and for T=25°C.
Nadir angles were computed using a start time of 16:29:35 GMT
(45°) and a stop time of 16; 30: 45 GMT (0° ) following the discus-
sion in section 4.0. The match in the curves maxima would be
approximately coincident if the tracking started with a 3 7° nadir
angle, which is in agreement with the voice tape transcript.
NADIR ANGLE
45*
40*
35* 30° 25* 20° 15° 10*
5» 0°
E
— I-
1 —
i i i i i >
1 l
(§j 860
-CALCULATED _
E 850
_
-
T* S
-»
■i_
i .,.
.
<t> 830
7E
o 820
-
■ .
• ."*"•• •
~
$
-*'u> •
m\
4.
I1* .
>'^^' * •
*~*^^- • "'
— 810
LlI
~
0|0 •
-re...- •♦( ...
MAXI IMAX
• . \^7
O
1 •
Z 800
-
A '
-
<
/
t \
O «,„..
• r
< 790
t *
~
or
i ' i
1 1 1 1
i i
I6'29'30 40 50 l&30'00 10 20 30 40 50 I&3I-00
GMT
Figure 4.3 Radianoe at 11 \im as a function of nadir angle of the
same ocean spot as that tracked on the S-191 (dots). The dashed
vertical line separates channel Al and channel A6 data. Dashed
curve is fourth-order polynomial fit to all data; fine solid
line is fourth-order polynomial fit to A6 data only. Heavy
solid line is the calculated nadir angle dependence.
The S-191 spectroradiometer uses a series of detectors that
cover a segment of the spectrum. In some regions these overlap
and ambiguity often exists as to which detector to use. In the
sections to follow, those detectors (channels) chosen were as
recommended in the NASA reports on instrument performance. It
is suggested that only those data that are well calibrated be
reported to non-instrument engineering investigators in the
future .
Barnett (personal communication) cautioned against the use
of radiometer channel Al in the S-191 (see again Fig. 4.3).
Accordingly a second fourth-order polynomial Csolid line) was
fitted to the channel A6 data only. The rms spread of the ra-
diance about this polynomial is 4.48yW cm~2 sr~i. This corresponds
to a noise-equivalent temperature difference (NEAT) of +0.3°K at
2 89. 6 5° K. Since the atmospheric attenuation tends to diminish
surface gradients, this results in a calculated NEA.T of
153
approximately +_0.6°C in T for this model at 11pm on this day.
The equivalent blackbody temperature at the maximum in the. poly-
nomial is 16. 5° C at the top of the atmosphere. This implies a
temperature correction of 8.5°C which is not unreasonable for a
tropical winter atmosphere whose precipitable water vapor is
3.6 cm (cf. Maul and Sidran, 1973).
4. 3 VISIBLE RADIANCE
The visible radiance experiment was designed to study the
accuracy with which spectral changes across oceanic fronts can be
observed and interpreted from satellite altitudes. Unfortunately,
the strongest front expected in the experiment area was missed by
the S-191 so this objective, as discussed before, could not be
accomplished. However, during the course of this work, a theore-
tical technique for recovering the "ocean color spectrum" through
the atmosphere was developed. This is discussed in detail below
and an attempt is made to compare the predictions of the theory
with the S-191 data and the associated ground truth.
4.3.1 Theoretical calculations
It is clear that the full potential of oceanic remote sens-
ing from space in the visible portions of the spectrum can be
realized only if the radiance that reaches the top of the atmos-
phere can be related to the optical properties of the ocean. To
effect this , the radiative transfer equation must be solved for
the ocean-atmosphere system with collimated flux incident at the
top of the atmosphere. In such calculations the optical proper-
ties of the ocean that must be varied are the scattering phase
function (PQ(6)) and the single scattering albedo (wQ; defined
as the ratio of the scattering coefficient to the total attenua-
tion coefficient). Furthermore, unless the ocean is assumed to
be homogeneous , the influence of vertical structure in these
properties must be considered. To describe the cloud-free at-
mosphere, the optical properties of the aerosols and their
variation with wavelength and altitude as well as the ozone
concentration must be known. Considering the ocean for the pres-
ent to be homogeneous, the radiance at the satellite can be
related to the ocean's properties by choosing an atmospheric
model and solving, the transfer equation for several oceanic
phase functions and oo0's at each wavelength of interest. The
number of separate computational cases required is then the
product of the number of phase functions, the number of values
of w , and the number of wavelengths. Even if the multi-phase
Monte Carlo method (MPMC) (Gordon and Brown, 1975) is used, the
co resolution of Gordon and Brown C19 7 3) would require a number
or simulations equal to ten times the number of wavelengths for
each atmospheric model considered. It is possible, however, to
obtain the necessary information without modeling the ocean's
optical properties in such detail.
154
Th_e model is based on an observation evident in results of
computations given by Plass and Kattawar C19 69) and by Kattawar
and Plass (197 2) on radiative transfer in the ocean-atmosphere
system, namely, that when the solar zenith angle is small, the
upwelling radiance just beneath the sea surface is approximately
uniform, (i.e., not strongly dependent on viewing angle) and
hence determined by the upwelling irradiance . This observation
is utilized in simulations of oceanic remote sensing situations
by assuming that a fraction R of the downwelling photons are
absorbed. The ocean is then treated as if there is a Lambertian
reflecting surface of albedo R just beneath the sea surface. In
this case Gordon arid Brown (19 74) have shown that any radiometric
quantity Q-, can be writte
n
Qo R
Q = Q, + J_ (4.3)
1-rR
Qj_ is the contribution to Q from photons that never penetrate the
sea surface (but may be specularly reflected from the surface) .
Q2 is the contribution to Q from photons that interact with the
hypothetical "Lambertian surface" once for the case R=l. r is
the ratio of the number of photons interacting with the
"Lambertian surface" twice, to the number of photons interacting
once, again for R=l.
By use of equation 4.3, any radiometric quantity can then be
computed as a function of R. Physically the quantity R is the
ratio of upwelling to downwelling irradiance just beneath the
sea surface and is known as the reflectance function [R(0,-)] in
the ocean optics literature (Preisendorfer , 1961). Spectral
measurements of the reflectance function R(A) have been pre-
sented for various oceanic areas by Tyler and Smith (19 70).
Henceforth, R(A) will be referred to as the "ocean color spectrum"
A series of Monte Carlo computations have been carried out
to see if an approximate simulation (AS1) , using this assumption
of uniform upwelling radiance beneath the sea surface, yields
results that agree with computations carried out using an exact
simulation (ES) , in which the photons are accurately followed in
the ocean as well as the atmosphere. The Monte Carlo codes used
in Gordon and Brown (19 73, 19 74) were modified by the addition of
an atmosphere. The atmosphere consisted of 50 layers and includes
the effects of aerosols, ozone, and Rayleigh scattering, using
data taken from the work of Elterman (1968). The aerosol scat-
tering phase functions were computed by Fraser (NASA-GSFC,
personal communication) from Mie theory assuming an index of re-
fraction of 1.5 and Deirmendjian ' s (1964) "haze-C" size
distribution. Also, to determine the extent to which the vertical
structure of the atmosphere influences the approximate simulation,
a second approximate simulation (AS2) was carried out in which
the atmosphere was considered to be homogeneous; i.e., the aerosol
scattering, Rayleigh scattering, and ozone absorption were inde-
pendent of altitude. The oceanic phase functions in the ES
155
are based on Kullenherg's C19 6 8) observations in the Sargasso Sea,
and are given in Table 4.1 CNote that all the phase functions in
the present paper are normalized according to 27T/71" PCB) sin d9 = l).
o
Table 4.1 The Three Ocean Scattering Phase Functions
e
KA
KB
KC
(deg)
(xlO2)
(xlO2)
(xlO2)
0
10924
10171
9521
1
4916
4577
4285
5
573.5
534.0
499.9
10
169 .3
157.7
147.6
20
29.5
29. 39
29.31
30
12.56
11.9 5
11.42
45
3.059
3.661
4.189
60
1.092
1.577
1.999
75
0.546
0.915
1.190
90
0.344
0.661
0.952
105
0.311
0.641
0.928
120
0. 317
0.732
1.094
135
0.410
0. 829
1.309
150
0.492
1.017
1.618
16 5
0.579
1.261
1.856
180
0.617
1.357
1.999
KA is roughly an average of Kullenberg's phase function
at 632.8 nm and 655 nm, and KC is his phase function at 460 nm.
KB is an average of KA and KC . These phase functions show con-
siderably less scattering at very small angles (8<1°) than was
observed by Petzold (19 72) in other clear-water areas; however,
the exact form of the oceanic phase function is not very
important, since it has been shown (Gordon, 1973) to influence
the diffuse reflectance and R(0,-) only through the back-scattering
probability (B)
J TT.
= 2tt f P (6) sinede.
72
In all of the computations reported here the solar beam incident
on the top of the atmosphere is from the zenith, and _ with unit
flux. At visible wavelengths the variable atmospheric constit-
uent that will most strongly influence the radiance at the top
of the atmosphere is the aerosol concentration, so the computations
have all been carried out as a function of the aerosol computa-
tion .
Table 4.2 gives a sample comparison of upward fluxes at the
top of the atmosphere at 40 0 nm in the three simulation models
(ES, AS1, and AS2) as a function of the aerosol concentration.
N, 3xN, and lOxN refer to aerosol concentrations in each layer
156
of 1, 3, and 10 times the normal concentration given by Elterman.
400 nm is chosen because in the visible portion of the spectrum
it is the wavelength at which the atmospheric effects are ex-
pected to be most severe. The ES case uses w = 0.8 and phase
function KC. The values of R used to effect the AC computations
were taken from the EC computation of this quantity. However,
if R is taken from
3
R = 0.0001 + 0.3244x + 0.1425x + 0.1308x (4.4)
where x = u> B/Cl-u (1-B) which, according to Gordon, Brown and
Jacobs (1975), reproduces the in-water reflection function for
the corresponding case but with no atmosphere present, the results
of the AS model computations agree with those listed to within
0.2%. The numbers in the parenthesis next to each flux value
represent the statistical error in the flux based on the actual
number of photons collected in each case. It is seen that ES
and AS simulations generally agree to within the accuracy of the
computations. Notice also the excellent agreement between the
AS1 and AS2 fluxes.
Table 4.2: Comparison of the flux at the
top of the atmosphere for the
ES, AS1, and AS2 simulations .
Aerosol
Concentration ES AS1 AS2
N 0.222 (+.002) 0.224 (+.001) 0.226 (+.001)
3xN 0.274 (+.003) 0.273 (+.001) 0.275 (+.001)
lOxN -.423 (+.004) -.426 (+.002) 0.425 (+.002)
Fig. 4A presents a comparison between the ES, AS1, and AS 2
upward radiances at the top of the atmosphere. The step-like
curve in the figure is for ES , the solid circles for AS1, and
the open circles for AS2 , and u is the cosine of the angle be-
tween the nadir and the direction toward which the sensor is
viewing. The radiances in Fig. 4.4 for the ES cases are accurate
to about 3% in the range y=l to about 0.4, while for the AS cases
the accuracy is about 1%. To within the accuracy of the computa-
tions , the three simulations again agree for all the aerosol
concentrations except within the range y=0 to about 0.3; i.e.
viewing near the horizon. These computations appear to demons
strate that the transfer of the ocean color spectrum through the
atmosphere can be studied with either the AS1 or AS2 model as
long as radiances close to the horizon are not of in+^rest.
Furthermore, from the reciprocity principle (Chandrasekhar , 1960)
157
the nadir radiance, when the aolar heam makes an angle 6Q with
the zenith, can be found by multiplying the radiance I Cy) in
Fig. 4.4 by p where p is taken to be cos 9 . This implies that
as long as the Sun is not too near the horizon, the AS1 and AS 2
methods of computation can be used to determine the nadir radiance
at the tip of the atmosphere as a function of the ocean's prop-
erties through equation 4.4. The fact that the AS2 model
(homogeneous atmosphere) yields accurate radiances is very im-
portant in remote sensing since it implies that only the total
concentration (or equivalently the total optical thickness) of
the aerosol need be determined to recover the ocean color spec-
trum from satellite spectral radiometric data.
015-
»
Figure 4.4 Comparison between ES (step-like curve) AS1
(solid circles) and AS2 (open circles) upward radiances
at the top of the atmosphere for an ocean with w0 * 0.8
and phase function KC and an atmosphere with a normal
(lxN)3 three times normal (3xN) and ten times normal
(lOxN) aerosol concentration.
158
It should be noted that these results also strongly suggest
that R(A) is the quantity relating to the subsurface conditions
that can be determined from space, and hence, is. the most natural
definition of the "ocean color spectrum". Moreover, it has been
shown (Gordon, Brown, and Jacobs, 1975) that R(A) is not a strong
function of the solar zenith angle Cthe maximum variation in
R(0,-) with 60 is of the order of 15% for 0<eo<60°),
with other definitions (Curran, 19 7 2; Mueller, 1973)
in contrast
I.C/0
(xlOO)
14
1
i i
1 1
1 1 I
-n
IOxNr
. r
"" 1
12
7xNr
1
r-1
1_r—
1 — -
10
-
1
8
3xN[
r~
-d
_
6
-"OxN, ...J
r -
4
-
" X
-400
nm'
-
2
A
•
1 1
' '
i i i
1.0
0.8
0.6
0.4
02
X
Figure 4.5 IjCv) as a function of y for carious aerosol
concentrations; wavelength of calculations is 400 nanometers,
The only way spacecraft data can be used to obtain informa-
tion concerning subsurface conditions (such as concentrations of
chlorophyll, suspended sediments, etc.) is through determination
of R(A).Clt is assumed here that the relationship between R(X)
and the ocean constituents is well known , whereas in fact much
work still remains to be carried out before such a relationship
can be established!! . This can be effected by applying equation
(4.3) to radiance I(u) at the top of the atmosphere with the Sun
at the zenith, which yields,
159
R iCy)
iCy) = I Cy) +
1 - rR
InCvO and I2 Cy) are presented in Figs. 4.5 and 4.6 for the
three aerosol models discussed above as well as for an aerosol
free model (OxN) and a model with seven times the normal aerosol
concentration ( 7xN ) .
12
1 "I —
OxN
1 -1 1 r 1 1
IxN
^ 1
10
3xN
"^-1
—
~l
-
8
-
1(H)
2
_,7xN
1
-
(xl00)6
1
lOxN
I
1
1
1
4
^S_
L-, -
2
1
\=400
nm
0
' •
1 1 1 1 1 1
1.0 0.8
0.6
0.4 0.2
Figure 4.6 .TgfiJ as a function of y for various aerosol
concentrations; wavelength of calculations is 400 nanometers.
For the cases considered, RSO . 5 , and since R is usually about
0.0 3 to 0.10 at this wavelength we can rewrite equation (.4.5)
approximately as
so
I (y ) = I, (y ) + Rig Gj )
R^IO ) - I^Cy )
C4.5
Cy )
160
Applying the reciprocity theorem to equation 4.6 it is found for
nadir viewing that
R^ I
nadir-
■1C^
I2 Cp0)
(4.7)
where y0 is the cosine of the solar zenith angle . Noting again
that R is 0 . 10 it is seen that the difference between Inadir
and y0 I-^(y0) must be small, which implies that the accuracy in
R will be limited by knowledge of I-,(y0). Since It(Vo) depends
strongly on the aerosol concentration, it is absolutely necessary
to be able to determine the aerosol concentration if an accurate
value of R is desired. Curran (19 72) has suggested that this
can be accomplished by observing the ocean (assumed free of white
caps) in the near infrared where R(A) 0. In section 4.3.2 of
this report, Curran' s suggestion is utilized to find the aerosol
concentration from the S-191 data.
Before trying to apply the relationships developed here to
the S-191 spectra, there are several important implications of
the theory to be discussed. Noting that I^Cu) and IoCu) depend
only on the direction of the incident solar beam, the properties
of the atmosphere and ocean surface, but not R (if it is assumed
these latter properties remain essentially constant over horizon-
tal distances large compared with those over which R changes
significantly), one can directly relate changes in I(y) to
changes in R. From equation 4.5
3I(y) a
3R
i2(y)
Figure 4.6 shows that 3I/3R is not an extremely strong function
of the aerosol concentration for concentrations up to three times
normal and viewing angles up to 35° from nadir. This suggests
that horizontal gradients in R can be estimated without an
accurate aerosol optical thickness.
When equation 4.5 is used to relate changes in radiance
(AI(y)) to changes in R(AR),
I(y) = [3I(y)/3R ]AR2rI2(y)AR-
(4.8a)
Equation (4
ance change
For example
detect a 5%
through an
centration.
noting that
140 W cm" 2
. 8a) makes possible a determination of minimum radi-
the sensor must be able to detect for a given AR.
, suppose that observing at ^=0.85 it is desired to
change in R for clear ocean water at 400 nm CR~.l)
atmosphere with three times the normal aerosol con-
Figure 4.6 shows that I 2 C. 0 85} is about 0.11 and
the extraterrestrial flux 40 0nm is about
nm -1, we find from equation (4.6) that AIC0.85) is
161
0.077 Wcm~2 nm~1sr~1. In a similar way radiance changes can be
related to AR for a nadir-viewing sensor and any solar zenith .
angle. As mentioned previously from the reciprocity principle,
nadir - u °
where y0 = cos 60 , 60 is the solar zenith angle and I(^0) is the
radiance at the top of the atmosphere seen be a sensor viewing
at 60 when the Sun is at the zenith. Following through with the
same arguments that led to equation (4.8a) it is found that
Al ,. =yo[3I(y0)/3R)> R^ y0I0(yn)AR. (4.8b)
nadir ° ° -020
Clearly, for a given AR, Ina^ip decreases substantially with
increasing solar zenith angle because of the presence of the y0
factor in equation (4.8b). For example, with a three-times
normal aerosol concentration, a nadir-viewing sensor would need
about 2.5 times more sensitivity at 8o = 60° as compared with 6o~0
to detect the same R.
The above examples indicate how the theory (AS1) can be used
in the design of a satellite sensor system for estimating some
ocean property such as the concentration of suspended sediments
or organic material. Specifically, one must first determine the
effect of the property on R. Then, on the basis of the sensi-
tivity desired, find AR, and finally, use equation (4.8a) or
(4.8b) to find the minimum radiance change the sensor must be
capable of detecting. If the sensor has a limited dynamic
range, then equation (4.5) can be used with equation (4.8a) or
(4.8b) to aid in the sensor performance design trade-offs.
EUnfortunately at this time, relationships between R(X) and sea-
water constituents are not well established.]
Considering the fact that we have used only the "haze-CM
aerosol phase function (which is clearly only approximately
characteristic of the actual aerosol scattering) it is natural
to inquire how strongly the computations of Ii(y) and I;?(y)
presented in Figs. 4.5 and 4.6 depend on the shape of the aerosol
phase function. To effect a qualitative understanding of the
influence of the aerosol phase function, computations of I]_ and
1 2 have been carried out using the well known Henyey-Greenstein
(HG) phase function
P ( 6) = (l-g2)/4TT
HG (l+g2-2g cos 6)3/2 ,
where the asymmetry parameter g is defined according to
g = 2-rr / PC6) cos 6 sin 6 d 6,
and 6 is the scattering angle. Since g for the haze-C phase
162
function is 0.69Q, computations haye been made with P^p C 6) for
figure 4 . 7 compares these P„p ( 8) ' s
g values of 0.6, 0.7, and 0 .
with the haze-C phase function. The HG phase function for g=0.-7
clearly fits the haze-C phase function quite well in the range
5°<0^140°; however, as is well known, the HG formula is incapable
10
i -
S 10"*
10'-
1 i i i — r
i i i i i i i i i i i i i
••• "HAZE C"
• \
HENYEY-
GREENSTEIN
\
\ •v •
^^^9 = 0.7
g=0.8
iiii
• i • i i 'I'
I0"3
0 20 40 60 80 100 120 140 160 180
9 (degrees)
Figure 4. 7 Comparison between the "haze-C" and various Eenyey-
Greenstein phase functions characterized by asymmetry parameters
0.63 0.7 j and 0.83 as a function of scattering angle ( 6 ).
of reproducing phase functions computed from Mie theory in the
extreme forward and backward directions. The HG phase functions
with asymmetry parameter 0.6 and 0.8 are seen to be substantially
different from the haze-C distribution at nearly all scattering
angles. On the basis of Fig. 4.7 it should be expected that I]_
and I2 computed with PhG^6^ will be in close agreement with the
haze-C computations only for g close to 0.7. Figures 4.8 and
4.9, which compare the results of computations of Iq_ and I2 for
P C e) for the normal aerosol concentration, show that this is
indeed the case. It is seen that except for apparent statistical
fluctuations, the HG phase function for g=0.7 yields values of
163
II
10
i
1 1
1 1 1 1
••• "HAZE C
HENYEY -
i
.■I
GREEI
1 1
OSTEIN
r— — -
r
_•
X = 400nm
9
— r-=
8
•
— r — r i i i i i i
■
7
g=o.6
•
o
g=o.7
O 6
•
•
X
•
g»o.8
h-r 5
J
4
3
2
1
-
"
i
i
i
'
1.0 0.9 0.8 0.7
0.6 0.5
0.4 0.3 0.2
Figure 4.8 Comparison between Ij(\x) computed for the "haze-C"
and Henyey-Greenstein phase functions for an atmosphere with
a normal aerosol concentration, as a function of cosine Q(\i) .
wavelength of calculations is 400 nanometers.
I]_ and 1 2 in good agreement with the haze-C computations. This
suggests that the detailed structure of the phase function is
not of primary importance in determining I]_ and l2» and it may
be sufficient for remote sensing purposes to parameterize the
phase function by g.
To get a feeling for the importance of variations in the
phase function in the remote sensing of ocean color, consider
the effect of changing the aerosol phase function from an HG
with g = 0.6 to one with g = 0.8 over an ocean with R = 0.1.
From Figs. 4.8 and 4.9 it is found that the normalized radiance
at \x - 0.85 (the assumed observation angle) decreases by 4. 9x10 3.
This decrease in radiance would be interpreted under the assump-
tion of no atmospheric change as a decrease in R from 0.10 to
0.056. This clearly indicates then that variations in the aerosol
phase function in the horizontal direction could be erroneously
interpreted as horizontal variations in the optical properties
of the ocean. However, it is probably unlikely that the clear
164
12
II
10
9
8
O
Q 7
X
i— i 6
5
4
3
2
■fc-TTL
T 1 1 1 1
••• "HAZE C"
HENYEY-GREENSTEIN
X = 400nm
9=0.8
am.
glofi
1.0 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1
Figure 4.9 Comparison between IgfyJ computed for the "haze-C"
Eenyey-Greenstein phase functions for an atmosphere with a
normal aerosol concentration 3 as a function of cosine 6 (v-) ;
wavelength of calculations is 400 nanometers.
atmospheric oceanic aerosol phase function will exhibit varia-
tions as large as that considered in this example, except in
extreme cases. Assuming that the aerosol concentration of the
atmosphere can be determined, the uncertainity in the aerosol
phase function will still of course provide a limit to the ac-
curacy with which the ocean color-spectrum can be retrieved from
satellite radiance measurements.
In summary then, the theory CAS1) leads to the natural
definition of RC*) [RC0,-)J as a function of wavelength]] as the
"ocean color- spectrum" . The determination of subsurface oceanic
properties from space can thus be divided into two problems:
165
1) the determination of RCA) from satellite radiance measure
ments , and 2) th_e establishment of relationships between RO)
and the desired ocean properties. Since the method of computa-
tion conveniently separates the radiance into a component that
interacts with the ocean CI?) and a component due to reflection
from the atmosphere and sea surface CI-, ) , it is easy to relate
changes in radiance to changes in R(A). It is found that for
viewing angles up to 35° from nadir, In is a relatively weak
function of the aerosol concentration for concentrations up to
three times normal. This suggests that spatial gradients of
R(A) can be determined with only a rough estimate of the aerosol
concentration. It is further found that variations in the aerosol
phase function can strongly influence the interpretation of the
radiance at the satellite. Clearly then, it is vital to under-
stand the magnitude of aerosol phase function variations.
4.3.2 Technique for Atmospheric Correction
As discussed in section 4.3.1 it is necessary to know
the aerosol concentration in order to recover the ocean color-
spectrum R(A) from the nadir radiance spectrum observed at the
satellite. In this section a method based on Curran's suggestion
of using the near-infrared radiance to determine the concentra-
tion is developed and applied to the S-191 data.
The method involves finding a band of wavelengths in the
near infrared for which the absorption by ozone and water vapor
is negligible. Since the Rayleigh scattering by air is very
small in the near-infrared, the greatest contributor to the
optical thickness at the wavelength in question is the aerosol.
It is found that at 7 80 nm ozone and water vapor do not absorb
significantly, and the Rayleigh scattering contributes only about
0.023 to the total optical thickness of the atmosphere. This
implies that aerosols play the dominant role in the radiative
transfer here with the normal aerosol concentration yielding an
optical thickness of about 0.2. Also since R(0,-) for wave-
lengths greater than about 70 0 nm is essentially zero, the
upward radiance at the top of the atmosphere at 7 80 nm simply
becomes
I(y) = I-,(y)
F° I
where FQ is the solar irradiance (mW cm"2ym" ) at the top of the
atmosphere CKondrat'ev 19 73). By use of the reciprocity principle
for nadir viewing and any solar zenith angle ( 6Q ) , Ina(^-Lr
can be written
I . = u _F I-i Cp « ) •
nada_r o lv-po/
IjCjj) and I2GO have been computed for aerosol concentrations
OxN, lxN, 2xN, and 3xN at 400 nm, 500 nm, 600 nm, and 780 nm.
166
The results for the OxN and lxN computations are presented in
Appendix C.
Using I-j_Cy) for 780 nm and noting that the S-191 nadir
radiance was recorded with 8 40° the upward radiance at the top
of the atmosphere for nadir viewing is found to be 0.32 3 and
0.9 38 mWcm-2 sr~lym~l for aerosol concentrations of OxN and lxN
respectively. Using the S-191 radiances at 780 nm for the nadir
viewing spectra taken on Jan. 8, 19 75 at 16:30:45.75 GMT (spec-
trum A) and 16:30:52.2 GMT (spectrum B) which respectively were
0.64 and 0.72 mW cm-2 sr-1, it is found that the theory suggests
the aerosol concentration was 0.51xN for spectrum A and 0.64xN
for spectrum B. In order to compute R(A) from the S-191 data
the assumption is made that the variation of the aerosol extinc-
tion coefficient with wavelength is exactly as given by Elterman.
4.3.3 Recovery of R(A) from the S-191 data
As mentioned above, in order to recover R(A) from the
S-191 data it is necessary to assume that the variation of the
aerosol extinction coefficient with wavelength is identical to
that given by Elterman. Also, since I-j_(y) and T^Cy) at 7 80 nm
were derived using the haze-C phase function for the aerosols,
the assumption is implicit that this phase function is correct.
With these assumptions the nadir radiance at the top of the
atmosphere has been computed at 400, 500, 600, and 7 80 nm for
aerosol concentrations OxN and lxN , assuming that R(A) is zero.
These radiances are presented in Table 4.3 along with those from
spectra A and B.
Since the actual aerosol concentration is known to be be-
tween OxN and lxN, it appears that the S-191 data at 400 nm are
in error. It is virtually impossible for the nadir radiance to
be less than that for a OxN atmosphere. (It should be noted
that the discrepancy here is great, i.e., the S-191 radiances at
400 nm appear to be too small by more than a factor of 2.) The
radiances at the other wavelengths listed in Table 4.3 seem to
be reasonable and were used in equation (4.7) to estimate R(\) .
The results are shown in Table 4.4.
It is seen that R(A) is negative except in the spectral
region 500-550 nm where the values shown compare well with the
Tyler and Smith Gulf Stream data for R(0,-). As discussed above,
the 400 nm data are apparently in error. However, the data at
other wavelengths appear realistic, so the negative R(0,-) values
are probably due to the assumptions that the haze-C phase func-
tion characterizes the aerosol, and that the spectral variation
of the aerosol scattering coefficient is correctly described by
Elterman' s data. It is clear that considerably more experimental
work is needed to test the ability of the theory discussed in
4.3.1 4o obtain an accurate RCA) from the satellite radiance.
167
Table 4.3
Wavelength
F
nadir
(R=0)
(nm)
lmW
cm"
• 2
-1
mW cm" 2
ym sr~l
157
OxN
lxN
Spect A
Spect B
400
5.61
6.58
2.30
2.50
500
201
2.93
4.10
4.13
4.39
600
184
1.25
2.13
1.55
1.67
780
125
0. 323
0.938
0.64
0. 72
Table 4.4
Wavelength R(0,-)
(nm) Spectrum A Spectrum B
400 -0.274 -0.268
450 -0.0242 -0.0235
500 0.0318 0.0370
550 0.0295 0.0303
600 -0.00715 -0.00547
780 0 0
5. MULTISPECTRAL SCANNER EXPERIMENT
SKYLAB's multispectral scanner was a unique design that had
13 spectral channels of data spread over the visible and infra-
red bands. The system used a conical scan which had the ^ advantage
of keeping the atmospheric path length the same at all times.
The visible region of the spectrum (0.4 - 0.75 ym) was divided
into 6 channels, each about 0.05 ymwide. Two reflected infra-
red (0.75 - 1.0 ym) channels and one in the emitted infrared
(10.2 - 12.5 ym) were also provided. The channels useful to
Table 5.1.
LANDSAT-1 has been shown to have several useful applications
of visible region imagery to marine science (Maul, 1974). The
much finer spectral resolution of the S-19 2 provided an opportun-
ity to expand those results to ocean current boundary
determination and to test if the lower wavelength (0.0 5ym) inter-
vals were useful through the intervening atmosphere.
168
Tahle 5 . 1
Spectral Channels Useful for Oceanography
BAND DESCRIPTION RANGE (ym)
1
Violet
0.41 -
0.46
2
Violet-Blue
0.46 -
0.51
3
Blue-Green
0.52 -
0.56
4
Green-Yellow
0.56 -
0.61
5
Orange-Red
0.62 -
0.67
6
Red
0.68 -
0.76
7
Reflected Infrare
d
0. 78 -
0.88
8
Reflected Infrare
d
0.98 -
1.03
13
Thermal Infrared
10.2 -
12.5
5.1 S-19 2 Data
S-192 data were collected from 16:29:22 GMT (over the open
sea just north of the Cuban coastline) to 16:31:04 GMT (over the
mainland Florida coast north of Florida Bay) . All channels listed
in Table 5.2 were carefully examined in the analog format pro-
vided by NASA to the principal investigator. The data in the
images were compared with the S-19 0A and S-190B photographs to
see if what is interpreted in section 3.2 as the anticyclonic
edge of the current could be detected. This feature was not
observable in the standard data product.
The cyclonic edge of the stream appears to be obscured try
clouds. This is often a useful means of locating the edge of the
current but unfortunately made the objective of directly sensing
the edge an impossibility.
However, an unexpected opportunity to evaluate the S-19 2
developed by the photographic detection ( section 3) of a mass of
water from Florida Bay flowing south into the Straits of Florida
just west of Key West. This water is milky in appearance and
somewhat greener in color. No ocean surface spectra were ob-
served inside or outside of the plume of Florida Bay water,
although it could have been easily accomplished if the SKYLAB
crew had observed the feature and notified the ship of its
presence. Upwelling spectral irradiance reported by Maul and
Gordon (1975) probably describes the essential features of the
plume and water in the straits.
An intensive effort was made by Norris (NASA-JSC) , Johnson
(Lockheed-JSC) , and Maul (NOAA-AOML) to identify from S-192 data
the plume and the anticyclonic edge, using the computer enhance-
ment facilities at NASA-JSC. After approximately 10 hours of
169
machine time on both, conical and line straightened data, the
feature described as the anticyclonic edge could not be identi-
fied, although it is clearly brought out in the photographic
enhancements Csee Fig. 3.2). Further effort to bring out the
anticyclonic edge was judged to be unwarranted and attention was
turned to the plume feature which is visible in Fig. 3.3, and
which preliminary computer enhancement showed to be a useful area
in which to work.
20
15 12 10 9 8 7 6
DATA SAMPLE INTERVAL
Figure 5.1 Power spectitwi of the radiance in the unfiltered
S-192 ooniaal format. Significant noise is noted every 15,
8-9 } and 6 data points.
Before a general computer enhancement technique was develop-
ed, the data were examined for periodic features in a spectrum.
Figure 5.1 is a spectrum of data specially provided for this
experiment that was to be high-pass filtered only; the calibrat-
ion of the S-19 2 data is considered a high-pass filter.
Significant periods at about 15 data sample interval? are noted
in these conical data as has been reported (Schell, Pnilco-JSC,
personal communication, 1975). The line-straightened data (see
Figure 5.2) have been band-pass filtered to remove this 15-data-
sample periodicity. The wavy patterns near the edge of clouds
are the result of filter ringing.
5 . 2 Computer Enhancement
Computer enhancement of S-19 2 data was an objective of the
experiment. The technique described below is a step toward
automatic detection of clouds in multispectral data. The goal
is to use a near infrared channel Cchannel 8 in this case) to
specify where cloud-free areas are, for analysis of sea surface
temperature or ocean color.
Channel 8 CO. 9 8 -1.0 3 jim) is selected as the cloud discri-
mination channel because there is a maximum in the atmospheric
transmissivity at this wavelength, and a maximum in the
170
Figure 5.2 S-192 Line -straightened, filtered, scanner data over the Straits of
Florida near the western Florida Keys. The appropriate S-192 channel number
is at the top of each panel.
171
absorption coefficient of water. The high absorption coefficient
of water at 1 urn causes the ocean surface to have a very low
radiance when compared with, land or clouds. Thus there should
be two modes in the frequency distribution of radiance: one mode
for the clear ocean and another mode for land and/or clouds. An
example of such a bimodal distribution is given by the histogram
in Fig. 5.3.
N - 2.15 /*W cm-«tr-»
a* ±4.00fiW cm-'sr"'
CLASS INTERVAL -0.5
01 23456789 10 II
RADIANCE (^.W cm^sr"1)
Figure 5. 3 Histogram (normalized to unity) of the radiance over
the area shown for channel 8 in figure 5.2. The primary peak
at the left is clear ocean; the broad peak centered at 7 vm' cm~2sr~l
is due to clouds and land.
In this figure, the low ocean radiances are clustered at
the mode centered at N = 0.2 yW cm"2 sr'1. The other mode,
centered at N = 5 . 7 pW cm" 2 sr"l is a contribution of the clouds.
(There is no land in this example.) If these modes can be
identified and separated, a statistical identification of cloud-
free ocean pixels can be made.
Cox and Munk (.19 5 4) observed that the radiance reflected from
the ocean is essentially Gaussian in character. The problem then
is to fit a curve of the form
y =
ni
exp E-CN - N)2/sa2]
172
(5.1)
to the data at the. lower valued mode. In this equation, the
normal frequency curve Cy) is a function of the total number of
observations Cn) , the class interval CD , and the standard de-
viation (a); the overbar on the dependent variable CN) denotes
ensemble average. Fitting equation C5.1) to the data is done
in an iteration scheme that uses the lower valued mode as a
first estimate of N. (.Only the values M +_ 2a from the original
ensemble are used In this first Iteration; this eliminates many
of the cloud contaminated data. ) After the first fit using a
predescribed N, the scheme is_to iterate, the data using only +_2a
of each new fit. When a (or N) changes less than 0.1% between
iterations, the fit is considered acceptable and the cloud- free
pixels are defined as those between 0<y<y +3. This guarantees
that 99% of the values around the lower mode are accepted and
included in the ocean data. Experience with this algorithm
suggests that only five or six iterations are usually necessary
for the scheme to converge.
OBSERVED N = 0.37 ± 0.22 /aW cm^sr-'
CALCULATED Y = 0.22 ± 0.09 /iW cm"2 si-'
CLASS INTERNAL = 0.05
Ql 0.2 0.3 0.4 Q5 0£ 0.7 0.8 0.9 10
RADIANCE (/iW cm"2 sr"1)
Figure 5.4 Expansion of low radianee portion of the histogram in Figure
5.3. The smooth curve is the fitted Gaussian approximation to the
observed radiance distribution.
In Fig. 5.4, the data from the left hand portioned Fig. 5.3
are plotted along with the fitted frequency distribution given
by equation C5.1). This fit required five iterations. Cloud- free
data are conservatively Identified as all those whose
173
Figure 5.5 Coniaal S-192 imagery of the area shown in figure 5.2. Channel 8
in this figure is a binary mask with clear ooean blaok, and clouds s land,
and other unwanted -pixels are white.
174
N<0.5 W cm~2 sr"-'- CN + 3a). Data from any other channel can
new be statistically examined on a pixel -by-pixel comparison
with channel 8; only "cloud-free" values go into the statistics.
Cloud-free data from any other channel are analyzed for
their mean and standard deviations. These calculations automa-
tically provide the limits over which the ocean data are to be
stretched. Following the technique suggested by Maul, Charnell,
and Qualset (19 74) , the formulation used by Maul (1975) is used
here. The stretch variable (c)for a negative image is defined
by:
X, - 0 for N>_N + Ka
X, - M RN+ko) - N] for (N-<a )<N<(N+ko )
2kcj
C = M for N<N - ko
In this formualtion M is the maximum value allowed by the digital-
to-analog (D/A) output device; N is the mean radiance of the
cloud- free data, and is a constant. Considering N a continuous
variable, setting k=2 would stretch 95% of the cloud-free data
over the full range of the photographic enhancement device.
The technique described above allows an objective specifica-
tion of both the range of settings for an optimum enhancement
of an ocean scene, and a statement of the radiance range
required of an ocean color sensor under these conditions. In
Fig. 5.5, the data from the sediment plume flowing through the
Florida Keys are presented for channels 1-6, 8, and 13. These
data are not line-straightened and they are only high-pass
filtered. Although the data are distorted geographically, they
represent the best radiometric information from the S-19 2.
5 . 3 Discussion
Without actual spectra as surface truth it is not possible
to intrepret the scanner data in the vicinity of the plume.
Other spectral data (see Section 4.2 and Maul, 19 75) suggest
that there should be a detectable difference in the data from
channels 1 through 5. Little information, let alone information
difference, is contained in channel 1 (see Figs. 5.2 £ 5.5).
These findings are in agreement with Hovis (NASA-GSFC, personal
communication, 19 75) who found that little or no information is
detectable through the atmosphere for wavelengths much shorter
than 0.45 um. The Rayleigh scattering at shorter wavelengths is
so intense that multispectral scanners such as the S-19 2 or the
coastal zone color scanner destined for NLMBUS-G should not have
a violet (0.41 - 0.46 um) channel. This means that the wavelengths
that contain much of the oceanographlc information are of little
175
value at orbital altitudes.
Analysis of the conical data used In Pig. 5.5 allows a
specification of the radiance range encountered. Since an ocean
color sensor should be allowed to saturate over land or clouds ,
the range appears narrow. The range is based on +_3a, which
describes 99% of the data encountered herein. Saturation at
higher signal levels is recommended in order to expand the
quantization commensurate with an acceptable data flow rate.
Table 5.2 lists the radiance ranges.
Table 5.2
Radiance Ranges of Infrared Channels
Oceanic
Radiance Range
Channel Mean +_ a mW cm~2sr~lym~l
1
2
3
4
5
6
7
8
13
This range of values applies only to this data set, which
represents low latitude, winter conditions. A similar analysis
of S-19 2 data from other areas of the oceans and at other times
could lead to an objective statement of the radiance range speci-
fication for an oceanic color sensor. Note that the lower
value on the shorter wavelength channels is non-zero. This
reflects the radiance of the atmosphere only.
Neither the line-straightened (Fig. 5.2) nor the conical
(Fig. 5.5) imagery was capable of detecting any ocean thermal
variations. The thermal data (channel 13) were processed in
both negative and positive format in an attempt to show some of
the 2°C range recorded in the surface observations (Fig. 2.1).
In comparing the data, note that the area covered is not identi-
cal because of geographical distortion in the line-straightened
processing.
The actual mask generated from using equation 5 . 1 is illus-
trated in the channel 8 data in Fig. 5.2. Here the values 0 or
3 are assigned for contaminated or ocean (respectively) radiances
176
3.56 +
0.45
2.21-4.91
4.71 +
0.9 7
1.80-7.62
4.07 +
1.33
0.08-8.06
2.83 +
1.46
0.00-7.21
1.73 +
0.92
0.00-4.49
0.93 +
0.43
0.00-2.22
0.57 +
0.29
0.00-1.46
0.37 +
0.22
0.00-1.03
0.80 +
0.03
0.71-0.89
The channel 8 data in Fig. 5.5 were simply stretched over the
+_2cr range by equation C 5 . 21. A significant error in the auto-
matic stretch units of other channels would occur if the extra
step of 0 or 1 assignment were not made, because of the exclu-
sion of some ocean radiances. Note also that the effect of
filter ringing is eliminated in the Fig. 5.2 masking; this can
also lead to wrong stretch limits if care is not excercised in
application of the technique. Filtering should always be done
after the data are masked.
The conical scan technique is an unusual approach to multi-
spectral scanning. Quality of the S-192 data is judged to be
poorer than the quality of data from LANDSAT MSS which uses a
linear scan. If poorer data quality is inherent in the design
of all conical scanners then the conical scan technique cannot
be recommended for the NIMBUS-G coastal zone color scanner. If
this is not the case there is no reason based on this investi-
gation to not consider such a future design.
6 . SUMMARY
The objectives of this experiment were to obtain simultan-
eous ship, aircraft, and spacecraft data across the Gulf Stream
in the Straits of Florida in order to evaluate several techniques
for remote sensing of this ocean current. Calibration of the
S-191 infrared detectors was not known; this precluded comparing
the atmospheric transmission models, which was an objective.
Detection of the current with the photographs was possible but
the S-19 2 scanner was not useful in this respect because the
gain settings were not adequate for ocean radiances. A tech-
nique to determine atmospheric corrections for visible radio-
meters based on theoretical considerations was shown to be
promising. Details of those results are enumerated below:
1. Observation and data reduction techniques for obtaining
surface truth data included measurement of ocean color spectra
from 3 meters above the surface. Objective filtering techniques
were developed to remove periodic specular return caused by wave
facets. (section 2.3)
2. The limited photographic (S-190) data set in this
experiment provides a baseline against which satellite photos of
more biologically productive waters may be compared. Inter-
comparisons with aircraft derived photography may be more
difficult. This problem was not resolved in this experiment
because of unforseen variation in the exposure level of these
photos. Csection 3.)
Non-uniformity in the satellite space-derived photography is
probably most due to the effects of variable lens transmission.
It is recommended that tbe production of multiple generation
photos C'dupes") be documented in a manner similar to that of the
177
original films. This should include some documentation on the
printer light variation across the platen, suspected to be a
major source of the variability measured in the dupes used in
this experiment. Csection 3.2)
3. The data acquisition camera on SKYLAB was not turned on
during the experiment; this made any quantitative comparisons
with spectra impossible. There should be a positive interlock on
all future missions to prevent a recurrence. Spectra were not
obtained across the Gulf Stream front because a) no data just
prior to the mission were obtained, and b) no direct communica-
tion link from the ship to the spacecraft could be set up. Future
experiments must include a mechanism to communicate to surface-
truth investigators the position of variable features such as
ocean currents, (section 4.1)
4. Unfortunately the S-191 visible near-IR data could not
be used to study the observability of oceanic fronts from satel-
lite altitudes because the Gulf Stream front was missed. However,
a theoretical method for recovering the ocean color spectrum
through the atmosphere was developed. This was used to try and
retrieve the ocean color spectrum from the nadir-viewing S-191
data, with limited success. The results agreed well with measure-
ments (Tyler and Smith, 19 70) for wavelengths greater than 5 00 nm,
but in the blue the S-191 radiance was substantially smaller than
previous aircraft observations, and even a factor of 2 less than
theoretical predictions for an aerosol-free atmosphere. This is
either due to a malfunction of the sensor in the blue or to the
presence of a very strongly absorbing aerosol. Our present
knowledge of the optical properties of marine aerosols does not
appear to be sufficiently complete to effect a quantitative
retrieval of ocean color spectrum from spacecraft data. Clearly
further research on this problem is indicated, (section 4.3)
5. Ocean features that were visible in photographs (sec-
tion 3.2) were not visible in the S-192 multispectral scanner
data because of the lack of radiance resolution. The range of
radiances needed to observe ocean features adequately is present-
ed in table 5.2. Tnese data support the view that an ocean
color multispectral sensor can do without a channel equivalent to
channel 1 (0.41 - 0.46 ym) of the S-19 2, as no information on a
very strong color boundary was contained in those data (section
5.3).
An objective S-19 2 cloud detection technique was developed
that uses Gaussian statistics to identify cloud-free areas in
channel 8 (0.98 -1.03 ym) . Cloud-free pixels are then analyzed
in other channels so that contrast stretching based on the same
statistics is automatically accomplished Csection 5.2).
178
7. ACKNOWLEDGMENTS
This research was supported in part by the National Aero-
nautics and Space Administration under the SKYLAB Earth Resources
Experiment Program. The authors wish to express their apprecia-
tion to the crews of the SKYLAB, the R/V VIRGINIA KEY, and the
NASA C-130 without whose enthusiastic support the experiment
could not have been accomplished. Specifically we would like to
acknowledge the assistance of A. Yanaway for computer program-
ming, D. Norris and W. Johnson for Image enhancement experiments,
N. Larsen, S. O'Brien, and A. Ramsey for secretarial, drafting,
and photographic contributions. The special radiosonde was
taken by the Key West office of the National Weather Service;
helpful arrangements by P. Connors of AOML are also gratefully
acknowle dge d .
8. REFERENCES
1. Anding, D. and J. Walker (1976): Use of SKYLAB EREP data
in a Sea-Surface Temperature Experiment, NASA CR-144479,
NITS Report E 76-10217, 54 pgs.
2. Anding, D. and R. Kauth (1970): Estimation of sea surface
temperature from space, Remote Sensing of Environ. , J_(4),
pp. 217-220.
3. Anding, D. and Kauth, R. (1972): Reply to comment by G. Maul
and M. Sidran, Remote Sensing of Environ, , 2_(3), pp. 171-
173.
4. Baig, S.R. (1973): Spectral properties of some marine
phytoplankters , J. Soc. of Mot. Pict . S T. V. Engrs. ,
82, pp. 1007-1012.
5. Baig, S.R. and Atwell, B.H., (1975): Spectra of Phytoplankters
in Aqueous Suspension, (in preparation).
6. Baig, S.R. and Yentsch, C.S., (19 69): A photographic means
of obtaining monochromatic spectra of marine algae , Appl.
Opt. 8(12), pp. 2566-2568.
7. Chandrasekhar, S. (1960): Radiative Transfer, Dover Press,
N.Y. , 393 p.
8. Cox and W. Munk (.1954). Measurements of the roughness of the
sea surface from photographs of the Sun's glitter. J. Opt .
Soc. Am., 44(11), pp. 838-850.
9. Curran, R. C1972) : Ocean color determination through a
scattering atmosphere, Appl. Opt. ,11, pp. 1857-1866.
179
10. Davis, P. A. , and Viezee, W. C19 64): A Model for computing
infrared transmission through, atmospheric water vapor and
carbon dioxide, J. Geophys. Res. , 69, pp. 3785-3794.
11. Deirmendj ian , D. C1964): Scattering and polarization prop-
erties of water clouds and hazes in the visible and
infrared, Appl. Opt . , 3, pp. 187-96.
12. Duntley, S.Q., Austin, R.W. Wilson, W.H. , Edgerton, S.F.
and S.E. Moran, (1974): Ocean Color Analysis , Scripps
Institution of Oceanography, San Diego, Calif., Ref.
74-10, 72 p.
13. Elterman, L. , (1968): UV , Visible, and IR Attenuation for
Altitudes to 50 km. , Air Force Cambridge Research Labora-
tories Report AFCRL-6 8-0153 .
14. Gordon, H.R. (1973): Sample calculation of diffuse reflec-
tance of the ocean, Appl. Opt. 12, pp. 2803-2804.
15. Gordon, H.R. and Brown, O.B., (19 73): Irradiance reflectivity
of a flat ocean as a function of its optical properties:
Appl. Opt. 12, pp. 1549-1551.
16. Gordon, H.R. and Brown, O.B., (1974): Influence of bottom
depth and albedo on the diffuse reflectance of a flat
homogenous ocean, Appl. Opt. 13, pp. 2153-2159.
17. Gordon, H.R., and Brown, O.B., (1975): A multiphase Monte
Carlo technique for simulation of radiative transfer,
J. Quant. Spectrosc. Radiat. Transfer, 15, pp. 419-422.
18. Gordon, H.R., Brown, O.B., and M.M. Jacobs, (1975): Computed
relationships between inherent and apparent optical
properties of a flat homogenous ocean, Appl. Opt . 14,
pp. 417-42 7.
19. Hanson, K.J. (1972): Remote sensing of the troposphere
IV. Derr, Ed.), U.S. Government Printing Office,
Washington, D.C. 22-1 to 22-56.
20. Herman, A. and Jacobson, S.R., (1975). FESTSA: A System for
Time Series Analysis, NOAA/AOML, Miami, Fla. , Unpublished
manu script .
21. Kattawar, C.W., and Plass, G.N., (.19.72): Radiatiye transfer
in the Earth.' s atmosphere ocean system: II Radiance in
atmosphere and ocean, J. Phys. Ocean. 2, pp. 146-15 6.
22. Kondrat'ev, K.Y* , CEd). (1973: Radiation Characteristics of
the Atmosphere on the Earth.' s Surface,, Amerind. Pub. Co.
Fvt. Ltd. , New Eelhi.
180
23. Kullenberg, G. , (.1968): Scattering of light by Sargasso
Sea water, Deep Sea Res . , 15, pp. 423-32.
24. Maul, G.A. , C1975): An Evaluation of the use of the Earth
Resources Technology Satellite for observing ocean
current boundaries in the Gulf Stream system, NOAA
Technical Report ERL 335-AOML 18, Boulder, Colorado,
125 pp.
25. Maul, G.A. (1974): Applications of ERTS data to oceanography
and the marine environment, COSPAR: Approaches to
techniques , Akademie-Verlag, Berlin, pp. 3 3 5-347.
26. Maul, G.A. (1973): Infrared Sensing of Ocean Surface Temper-
ature, In: The Second Fifteen Years in Space , A.A.S. , 31,
pp. 451-464.
27. Maul, G.A., and Gordon, H.R., (1975), On the Use of the
Earth Resources Technology Satellite (LANDSAT-1) in
Optical Oceanography, Remote Sensing of Environment , 4( 2) ,
pp. 95-128.
28. Maul, G.A. and Sidran, M. , (1973): Atmospheric effects of
ocean surface temperature sensing from the NOAA satellite
scanning radiometer, J. Geophys . Res., 78(12), pp. 1908-
1916.
29. Maul, G.A. and Sidran, M. (1972): Comment on "Estimation of
Sea Surface Temperature from Space" by D. Anding and
R. Kauth , Remote Sensing of Environment , 2(3), pp. 165-169.
30. Maul, G.A. , Charnell, R.L., and Qualset R.H. (1974).
Computer Enhancement of ERTS-1 Images for ocean radiances
Remote Sensing of Environ. , 3, pp. 237-253.
31. Mueller, J.L., (1973): The influence of Photoplankton on
Ocean Color Spectra, PhD. Dissertation, Oregon State
University, Corvallis , Oregon, 237 p.
32. NASA (1975): Report on Active and Planned Spacecraft and
Experiments, NS SDC1 WDC-A-RSS, 7501, Goddard Spaceflight
Center, Greenbelt, Maryland, 30 3 p.
33. Petzold, T.J. (1972): Volume Scattering Functions for
Selected Waters, Scripps Institution of Oceanography,
Univ. of Caligornia at San Diego, SIO Ref. 72-78.
34. Plass, G.N. and Kattawar, G.W. , C1969): Radiative Transfer
in an atmosphere ocean system, Appl. Opt . , J3, pp. 455-466.
181
35. Preisendorfer, R,W« , (JL9.6.1) : Application of Radiative
Transfer Theory to Lrght Measurements in Sea, U.G.G. I . ,
Monogr. 10, pp. 11-30.
36. Richardson, P.L., Strong, A.E., and J. A. Knaus , (1973):
Gulf Stream Eddies: recent observations in the western
Sargasso Sea, J. Phys . Ocean. , 3(3), pp. 297-301.
37. Strickland, J.D.H. and Parsons, R.R. (1968): A Practical
Handbook of Seawater Analysis , Fisheries Research Board
of Canada, Ottawa, 311 p.
38. Tyler, J.E., and Smith, R.C., (1970): Measurements of
Spectral Irradiance Underwater , Gordon and Breach, N.Y.
103 p.
182
Appendix A
Surface Data Obtained During
Cruises of R/V Virginia Key
This appendix lists the surface data obtained during the
8-9 January 1974 (GMT) cruises of R/V VIRGINIA KEY. The geo-
graphic positions were determined after the control data were
carefully replotted and a best fit of the vessel's location made
Values reported were determined by using standard oceanographic
techniques unless modified as explained in section 2.2 of the
text.
183
TABLE A.l
Cl-a
SURFACE SAMPLES
3
Cl-a (mg/m ) TIME (GMT) Position
0.15 1957 24° 08.1 N, 81° 34.6 W
0.09 2200 23° 58.8 N, 81° HO. 4 W
0.11 2400 23° 47.5 N, 81° 47.2 W
SAMPLES at depth at 1957 GMT at 24° 08.1 N, 81°34.6 W
3
Depth Cl-a (mg/m )
10 m 0.79
20 m 0.21
30 m 0.25
40 m 0.42
50 m 0 . 30
SAMPLES at depth at 2200
GMT
at
23'
D 58.8 N, 81° 40.4 W
Depth
3
Cl-a (mg/m )
10 m
0.10
20 m
0.20
30 m
0.3 5
40 m
0.31
50 m
0 .30
184
TABLE A. 2
SALINITY SURFACE SAMPLES
Sal(°/oo) Time (GMT) Position
35.996 1314 24° 38.9 N, 81° 08.0 W
35.924 1449 24° 34.5 N, 81° 12.2 W
35.929 1515 24° 30.7 N, 81° 16.3 W
35.921 1719 24° 23.3 N, 81° 21.8 W
36.009 1806 24° 19.1 N, 81° 27.3 W
36.023 1940 24° 10.2 N, 81° 33.7 W
36.017 1958 24° 08.1 N, 81° 34.6 W
36.119 2200 23° 58.8 N, 81° 40.4 W
35.874 2400 23° 47.5 N, 81° 47.2 W
35.918 0110 23° 38.2 N, 81° 52.6 W
185
Table A. 3
Volume Scattering 3 (45)
Surface Samples
Blue Green GMT
(m-1sr-1 x 10-3) (m_1sr-1 x 10~3) Time Position
17.3 13.5 1300 24° 40.2 N 81° 07.3 W
13.2 9.9 1314 24° 38.9 N 81° 08.0 W
3.03 3.02 1434 24° 36.5 N 81° 10.6 W
2.61 2.08 1449 24° 34.5 N 81° 12.2 W
2.40 1.60 1504 24° 32.2 N 81° 14.7 W
2.82 2.37 1515 24° 30.7 N 81° 16,3 W
2.69 1.95 1630 24° 27.7 N 81° 16.3 W
3.07 2.15 1645 24° 26.1 N 81° 18.1 W
3.06 2.16 1700 24° 24.6 N 81° 20.2 W
2.16 2.31 1715 24° 23.3 N 81° 21.8 W
3.42 3.09 1750 24° 20.5 N 81° 25.4 W
3.10 2.24 1802 24° 19.1 N 81° 27.3 W
4.02 3.49 1855 24° 16.3 N 81° 31.0 W
2.74 2.49 1910 24° 14.6 N 81° 32.5 W
3.08 1.75 1925 24° 12.4 N 81° 33.5 W
4.54 3.29 1940 24° 10.2 N 81° 33.7 W
5.82 8.44 1957 24° 08.1 N 81° 34.6 W
2.76 2.20 2040 24° 04.5 N 81° 35.6 W
2.73 1.45 2115 24° 03.2 N 81° 37.7 W
3.08 1.98 2130 24° 01.5 N 81° 39.0 W
2.99 1.90 2145 24° 00.5 N 81° 39.8 W
1.65 1.77 2200 23° 58.8 N 81° 40.4 W
186
Table A. 3 (cont'd)
Volume Scattering 3 (45)
Blue Green GMT
(m~1sr""1x 10-3) (m-1sr-1x 10-3) Time Position
3.35 2.04 2255 23° 55.8 N 81° 42.0 W
2.63 1.85 2310 23° 53.8 N 81° 43.5 W
2.85 1.83 2325 23° 51.9 N 81° 44.6 W
2.58 1.67 2400 23° 47.5 N 81° 47.2 W
1.92 1.30 0050 23° 40.9 N 81° 51.1 W
1.93 1.28 0110 23° 38.2 N 81° 52.6 W
1.96 1.35 0130 23° 35.7 N 81° 54.1 W
2.19 1.50 0150 23° 33.2 N 81° 55.6 W
Volume Scattering 6 (45)
Samples at depth at 1314 GMT at 24° 38.9 N, 81° 08.0 W
depth Blue (ni'^r"1 x 10~3) Green (m'-^sr"1 x 10~3)
10 m 2.11 1.83
Samples at depth at 1600 GMT at 24° 29.3 N, 81° 15.5 W
depth
Blue
(m 1sr~ l
X
io-3)
Green
(m !sr
xx 10"3)
10 m
13.4
11.1
20 m
2.39
1.77
30 m
2.88
2.80
40 m
2.27
1.86
50 m
1.46
1.42
187
Table A.3 (cont'd)
Volume Scattering 3(45)
Samples at depth at 1806 GMT at 24° 19.1 N, 81° 27.3 W
Blue (m^sr"1 x 10~3) Green (m^sr"1 x 10"3)
8.07 8.43
2.57 2.19
4.37 3.37
3.34 2.46
6.87 5.05
Samples at depth at 1957 GMT at 24° 08.1 N, 81° 34.6 W
depth
10
m
20
m
30
m
40
m
50
m
depth
Blue
Cm"
-isr-i
x lO"3)
Green (m~1sr~1
x 10"3)
10 m
5.68
3.78
20 m
4.26
3.66
30 m
4.61
3.73
40 m
3.87
3.26
50 m
4.41
3.55
Samples at depth at 2200 GMT at 23° 58.8 N, 81° 40.4 W
Blue (m^sr"1 x 10~3) Green (m"1sr~1 x 10~3)
4.12 2.99
2.60 1.94
5.70 4.06
3.98 3.42
3.28 2.60
depth
10
m
20
m
30
m
40
m
50
m
188
T
emp
.(°C)
24
.5
*
24
.8
*
25
.3
*
25
.4
s'«
25
.0
25
.0
JU
24
.5
ft
24
.3
ft
24
4
*
24
,5
A
24
5
ft
24.
7
ti
24.
4
24.
4
*
24.
6
24.
9
•*•
24.
9
ft
24.
8
25.
0
•♦*
25.
0
ft
25.
1
A
25.
5
4\
25.
5
25.
5
Table A. 4
Bucket Temperatures
TIME GMT Position
1315 24° 38.9 N, 81° 08.0 W
1434 24° 36.5 N, 81° 10.6 W
1449 24° 34.5 N, 81° 12.2 W
1504 24° 32.2 N, 81° 14.7 W
1630 24° 27.7 N, 81° 16.3 W
1645 24° 26.1 N, 81° 18.1 W
1700 24° 24.6 N, 81° 20.2 W
1750 24° 20 .5 N, 81° 25.4 W
1805 24° 19 .1 N, 81° 27.3 W
1855 24° 16.3 N, 81° 31.0 W
1910 24° 14.6 N, 81° 32.5 W
1925 24° 12.4 N, 81° 33.5 W
1940 24° 10.2 N, 81° 33.7 W
2010 24° 07.0 N, 81° 34.6 w
2040 24° 04.5 N, 81° 35.6 W
2115 24° 03.2 N, 81° 37.7 W
2130 24° 01.5 N, 81° 39.0 W
2145 24° 00.5 N, 81° 39.8 W
2200 23° 58.8 N, 81° 40.4 W
2255 23° 55.8 N, 81° 42.0 W
2310 23° 53.8 N, 81° 43.5 W
2325 23° 51.9 N, 81° 44.6 W
2340 23° 49.9 N, 81° 45.8 W
2400 23° 47.5 N, 81° 47.2 W
189
Table A. 4 (cont. )
Bucket Temperatures
Temp. (°C) TIME GMT Position
26.1 0050 23° 40.9 N, 81° 51.1 W
TIME GMT
0050
0110
0130
0150
26.1 Oil : 23° 38.2 N, 81° 52.6 W
U i 3D 23° 35.7 N, 81° 51.1 W
26.3 0150 23° 33.2 N, 81° 55.6 W
* Denotes XBT casts.
190
APPENDIX B
Radiosonde data
Key West
1600 GMT,
8 January 1974
Pressure (mb)
Temp. (°C)
R . H . ( % )
Surface
1021
024.8
79
1011
023.5
82
Dry Bulb: 24.°9C
1000
023.0
85
Wet Bulb: 22.°2C
902
017.1
86
RH : 7 9 %
869
015.9
63
Wind Dir: 050°T
850
015.1
64
Wind Spd : 5 mps
808
012.8
66
Clouds 1/10 cu
799
012.2
34
<J> 24035'N
788
011.2
72
A 81°42 'W
744
009.9
66
740
008.3
72
722
008.5
30
700
007.8
31
671
005.2
55
640
003.3
42
621
003.1
22
530
-06.3
21
500
-09.4
13
470
-12.4
10
384
-24.9
14
300
-39.1
14
250
-48.9
224
-50.3
212
-48.9
200
-50. 3
150
-62.6
100
-77.1
070
-75.7
066
-75.2
061
-71.5
058
-72.7
054
-67.6
050
-67.8
045
-64.7
043
-58.1
030
-51.2
023
-46. 3
020
-47.4
017
-48.0
016
-48.2
13.5
-43.9
191
APPENDIX C - Monte Carlo Simulations
This appendix lists the Monte Carlo simulations of
radiances 1-^ and Io as described in Section 4.3.2. Wave'
lengths at which the OxN and lxN atmospheric aerosol
concentrations were computed are 400, 500, 600, and 780
nm. The cosines of ten zenith angles (m) were the in-
dependent variables. 1^ and I2 are normalized to unit
solar flux on a surface normal to the solar beam.
192
Wavelength = 400 nm Aerosol = OxN
y !]_( y) I2 ( y)
0.00
0.10
0.20
0.30
0.40
0.50
0.60
0.70
0.80
0 .90
0 .95
1.00
.10089+00 .53255-01
.88850-01 .72697-01
.78011-01 .87278-01
.67433-01 .98883-01
.57839-01 .10661+00
.52975-01 .11159+00
.48109-01 .11406+00
.46607-01 .11634+00
.44244-01 .11615+00
.42872-01 .11701+00
.75049-01 .11782+00
193
Wavelength = 4-00 nm
v
0.00
0.10
0.20
0.30
0.40
0.50
0.60
0.70
0.80
0 .90
0 .95
1.00
IxCy)
.93753-01
.90653-01
.83232-01
.76449-01
.67122-01
.61046-01
.57118-01
.54662-01
.52434-01
.52278-01
.71513-01
Aerosol = lxN
I2U)
.51310-01
.63140-01
.77826-01
.87640-01
.96479-01
.10118+00
.10457+00
.10757+00
.10919+00
.10916+00
.10906+00
194
Wavelength = 500 nm Aerosol = OxN
0.00
0.10
0.20
0.30
0.40
0.50
0-60
0.70
0. 80
0.90
0.95
1.00
.63253-01 .59700-01
.52971-01 .91621-01
.37429-01 .10902+00
.29710-01 .12474+00
.24246-01 .13138+00
.22112-01 .13641+00
.19925-01 .13898+00
.19015-01 .14171+00
.18002-01 .14225+00
.17044-01 .14326+00
.67308-01 .14203+00
195
Wavelength = 500 nm
v
0.00
0.10
0.20
0.30
0.40
0.5«0
0.60
0.70
0 .80
0 .90
0 .95
1000
i-l(p)
.62318-01
.59056-01
.49220-01
.41716-01
.35218-01
.31928-01
.27678-01
.26638-01
.25573-01
.26419-01
.57385-01
Aerosol = lxN
I2(u)
56038-01
75227-01
96058-01
11022+00
11781+00
12576+00
12875+00
13063+00
13228+00
13315+00
13155+00
196
Wavelength = 600 nm
V
0.00
0.10
0.20
0.30
0.40
0 .50
0 .60
0 -70
0 .80
0 .90
0 .95
1.00
I^v)
.27632-01
.36531-01
.29628-01
.25526-01
.21423-01
.17841-01
.16190-01
.15107-01
.14509-01
.16199-01
.52310-01
Aerosol = lxN
I2(.v)
.36991-01
.67189-01
.92357-01
.10823+00
.11883+00
.12614+00
.13059+00
.13303+00
.13448+00
.13524+00
.13560+00
197
Wavelength = 600 nm
y
0.00
O.ffiO
0.20
0.30
0.1*0
0.50
0.60
0.70
0.80
0.90
0.95
1.00
Ix( u)
.30110-01
.24508-01
.17292-01
.13224-01
.10725-01
.98865-02
.90777-02
.88360-02
.80934-02
.77006-02
.61865-01
Aerosol = OxN
I2(y)
.42002-01
.80947-01
.10497+00
.12254+00
.13037+00
.13672+00
.14052+00
.14304+00
.14358+00
.14664+00
.14694+00
198
Wavelength = 780 nm
V
0.00
0.10
0.20
0.30
0 .40
0 .50
0 .60
0 .70
0 ,80
0 .90
0 .95
1.00
i-lCh)
.27838-01
.13026-01
.76052-02
.51505-02
.42167-02
.37095-02
.34865-02
.33889-02
.30901-02
.30811-02
.65909-01
Aerosol = OxN
I2(v)
.69923-01
.10854+00
.13086+00
.14849+00
.15502+00
.16096+00
.16342+00
.16269+00
.16464+00
.16649+00
.16607+00
199
Wavelength = 780 rim
V
0.00
0.10
0.20
0.30
0.40
0.50
0.60
0.70
0.80
0.90
0.95
1.00
i-lCv)
.40506-01
.33748-01
.24905-01
.18224-01
.14512-01
.11931-01
.10491-01
.98166-02
.98476-02
.11295-01
.57149-01
Aerosol = lxN
i2C u>
.67910-01
.95346-01
.11832+00
.13459+00
.14180+00
.14856+00
.15236+00
.15399+00
.15395+00
.15613+00
.15511+00
200
19
Reprinted from: Marine Sediment Transport and Environmental
Management, D. J. Stanley and D. J. P. Swift, editors, John Wiley
and Sons, Inc., Chapter 5, 53-64.
Tidal Currents
HAROLD O. MOFJELD
Atlantic Oceanographic and Meteorological Laboratories, Miami, Florida
CHAPTER
In regions where they are sufficiently strong, tidal
currents constantly rework bottom sediment. Weaker
currents combine with storm-generated wave motion
and currents to move sediment both at the water-
bottom interface and in suspension. Tidal currents
are especially effective agents of sediment transport
because they persist throughout the year, whereas other
types of water motion, particularly storm events, tend
to be seasonal. They are the background upon which
are superimposed other kinds of currents causing
sediment transport.
The tides typically rise and fall twice a day (semi-
daily tides), once a day (daily tides), or occur as a com-
bination of daily and semidaily components. Figure 1
illustrates tides at different locations along the east
coast of the United States and in the Gulf of Mexico.
The tide is semidaily along the eastern seaboard and
dominantly daily in the Gulf of Mexico (Pensacola and
Galveston).
As the earth rotates about its axis, the forces producing
the tides move across the earth's surface from east to
west. The motion of the moon around the earth and the
earth-moon system around the sun produce variations
of the tides and tidal currents with periods of about two
weeks, a month, six months, a year, and longer.
Approximately twice monthly the range of the tide is
a maximum; that is, the difference in sea level between
successive high tide and low tide is largest. These spring
tides occur for both the daily and semidaily tides,
although not necessarily on the same day. The term
neap tides refers to the tides with minimum range.
If the ocean covered the entire earth to a constant
depth, the pattern of sea-level changes caused by the
tides would be simple. However, since the oceans have
complicated shapes, these patterns in the real oceans
are also complicated. The global distributions of the
daily and semidaily tides are given in Figs. 2 and 3.
The numbers in small type along the coast and at
islands in Figs. 2 and 3 are the spring ranges averaged
over a year.
The lines traversing the oceans in these figures give a
general idea of the stage of the tide, i.e., when high,
mean, and low water occur. For example, assume that
in a particular region of the North Atlantic the semi-
daily high water is occurring along the cotidal line
marked 0 hour. Then along the 3 and 9 hour lines, the
semidaily tide is passing through mean sea level; and
along the 6 hour line, the tide has reached low water.
About 3 hours later, high water for the semidaily tide
will occur along the 3 hour line, low water along the
9 hour line, and so forth. The pattern rotates counter-
clockwise around a point in the North Atlantic where
the range of the semidaily tide is zero. Such a rotating
pattern is called an amphidromic system, and the point
is called an amphidromic point. Amphidromic systems
are a general feature of both the daily and semidaily
tides in the deep ocean. Amphidromic systems also
occur in shallow seas, such as the North Sea, as shown
in Fig. 4.
The tides and tidal currents on the continental shelves
and seas bordering the open seas are propagated as
waves from the open oceans. These waves are partially
201 53
54
TIDAL. CURRENTS
TYPICAL TIDE CURVES FOR UNITED STATES PORTS
12 13 14 15 16 17 19 19 20
Luni' dltj mil S dBChnit.on 9th ipogw 10!h lilt qulrtti I3ttv on fouilo' !6ln rww moon 20th panftt
22d mil N dtcnnttion 23d
FIG! RE 1 . Predicted tides at selected ports along the I S. east
coast and in the Gulf of Mexico. The range and character of tides
differ significantly along coasts and between distinct regions. From
the I .S. Department of Commerce Tide Tables (1974).
reflected back out to sea by the shoaling bottom which
rises toward the beach. The combination of the in-
coming, or incident, wave and the reflected wave is
called a standing tide; the semidaily tide on the conti-
nental shelf between Cape Hatteras and Long Island
is an example of a standing tide. The wave may also
propagate along the coast, in which case it is called a
progressive tide; that is, the stages of the tide progress
down the coastline. An example of a progressive tide is
the daily tide along the U.S. cast coast. Both of these
examples can be seen in Figs. 2 and 3. The technical
term for a tide that is generated elsewhere and propa-
gates into a given region is cooscillation. The water on
the continental shelf off the U.S. east coast cooscillates
with the North Atlantic.
Coastal lagoons, bays, and estuaries cooscillate with
the water on the continental shelves. The tides often
enter these coastal bodies of water through inlets and
over bars, both of which can significantly attenuate the
tides so that the tides inside the embayment are smaller
than the tides along the open coast.
EQUATIONS OF MOTION
To compute the tides and tidal currents within a region
having a complicated shape and realistic bathymetry,
computer programs have been developed which use
information at a number of points to compute the
motion at those or other points. How well the results of
the calculations describe the motion depends on the
spacing between points (how well the grid of points
resolves depth variations), approximations to the
fundamental equations, knowledge of the motions at
the boundaries of the region, and estimates of the
bottom stress coefficient determining the drag of the
sediment on the water.
When idealized depth variations are assumed and
when less significant forces are neglected, the tides and
tidal currents can sometimes be described by simple
formulas from which considerable insight can be gained
into tidal phenomena. Historically, intensive research
was done on the behavior ol tides in channels having
constant depth and vertical side boundaries. The
channel theory of tides is used in the present discussion
and then extended to open regions such as the conti-
nental shelves.
In a channel where bottom stress and the Coriolis
eflect due to the earth's rotation can be neglected, a tide
causes the water to accelerate through the downchanncl
slope of the sea surface. Horizontal differences in the
resulting tidal currents in turn cause sea level to change.
The interplay between these two effects produces a
wave that propagates down the channel, away from the
source of the tide. A general theory of waves has been
presented by Mooers in Chapter 2; tides are waves
whose wavelengths are long compared with the water
depth but whose amplitudes are small compared with
the depth. With these assumptions, the tidal motion in a
channel is described by the pair of equations
dt
-g
dx
dit
dx
(1)
(2)
202
80
•5 8
fig .»
<S »
« -2
60 *
* 9
■s
|
*
s
=i
^*
0
§
E
•8 $
-§
s
8
5t
s
4)
0
•**>
<o
>
p
.g
-3
St
!*
"S
**
HI
1
>->
■*
3
x
^
(N
s
^1
LU
<s
OS
Ij
3
te
0
5
203
o
2 +
Q
I
6*
c
•3
o
si
0
56
204
EQUATIONS OF MOTION
57
60° -
58°
56° -
54° -
52° -
50°
FIGURE 4. Amphidromic systems of the Mj tidal constituent
(semidaily lunar tide) in the North Sea. The cotidal lines show
the progress of the tide each constituent hour (30 ' phase change),
the dotted corange lines show the decrease in feet of the Mj tidal
range away from the shore. From Doodson and Warburg ( 1941)-
Equation 1 states that the time rate of change of the
downchannel horizontal velocity u is equal to the
acceleration of gravity g multiplied by the downchannel
slope of the sea surface, whose displacement above mean
water is r\\ t is the time elapsed after high water at the
source; and x is the distance away from the source of the
tide as measured along the axis of the channel. Equation
2 states that the time rate of change of the sea surface
displacement rj is equal to the mean depth h multiplied
by the horizontal rate of change of the downchannel
velocity u.
Assuming that the mean depth h is uniform through-
out the channel, a tide with a period T and an amplitude
a (one-half the range) would be described by
t? = a cos
H-t)]
■GT-GK'-f)]
(3)
(4)
where c = (gh)112 is the speed of propagation at which
the shape of the sea surface moves down the channel.
For oceanic and shelf depths c is 200 and 31 m/sec,
respectively.
If the depth h were representative of the open ocean
(h = 4000 m) and the amplitude of the tide were
a = 0.5 m, the maximum tidal current according to (4)
would be 2.5 cm/sec, a relatively small speed. On the
other hand, if the depth were representative of the
continental shelves (h = 100 m), the corresponding
current would be 15.8 cm/sec. For a given tidal ampli-
tude, the maximum tidal currents are inversely pro-
portional to the square root of depth. In very shallow
water, where (4) would predict unrealistically large
currents, the formula is not applicable since turbulent
dissipation and bottom stress which would limit the
currents have not been included.
In Fig. 5, the tide and tidal current are shown along a
vertical section parallel to the channel axis; both are
uniform across the channel. At any given time, the
pattern repeats itself downchannel with a horizontal
distance equal to the wavelength X = cT of the wave.
The water velocity u is the same at every depth because
no bottom stress is allowed in this idealized model.
DIRECTION OF PROPAGATION
SEA SURFACE
MEAN SEA LEVEL
TIDAL CURRENTS
////////////////////////////////// BOTTOM
FIGURE 5. Idealized progressive tide propagating in a narrow channel in which bottom
stress effects on the tidal currents are neglected. The vertical scale is greatly exaggerated.
205
58
TIDAL CURRENTS
WAVE DIRECTION
-* — * •* »o« <- < — « «— «- <o> -» — * —
COMPONENT OF CURRENT IN WAVE DIRECTION
4 5 6 7 6
HOURS AFTER HICH TIDE
TIDAL HOURS AfTCR HW
6 5 4 3 2 10
VELOCITIES AT
TIDAL.HOURS
HW
777777777777/7777777
16 8 0 3 16 Cl^/SEC
DISTRIBUTION OF
VELOCITY COMPONENT
ABOVE BOTTOM
TRAJECTORY OF
A WATER PARTICLE
FIGURE 6. Typical tidal currents on a continental shelf where
bottom stress forces the currents to zero speed at the bottom. From
Fleming and Revelle (1939, p. 130); after Sverdrup (1927).
With turbulent stresses, a more realistic profile of u is
shown in Fig. 6, in which u decreases within the bottom
boundary layer to essentially zero at the water-sediment
interface.
Since the tide propagates down the channel, it is a
progressive tide. In this case, the maximum horizontal
currents occur at higrrwater and low water where the
current is in the direction and opposite to the direction
of propagation, respectively. As the sea level passes
through mean sea level, t; = 0, the current is momen-
tarily zero.
As a water parcel moves in the channel, its position X
is the integral in time of the horizontal velocity:
X
u dt + A'0
(5)
where A'0 is the position of the parcel at the initial time
/ = 0. Using (4), the horizontal displacement of the
parcel is given by
^W'^M'-t)]
+ A'0 (6)
Water subject to the tidal motion oscillates about an
average position A'n with an amplitude equal to half the
total excursion. For a semidaily tide (7" = 12.5 hours)
and a tidal range 2a = 1 m, the excursion for open sea
depths (h = 4000 m) is 0.35 km, whereas for shelf
depths {h = 100 m) it is 2.25 km.
In (6), tidal currents would produce no net displace-
ment of water or suspended particles. That is, if a water
parcel were tagged using dye and observed throughout a
tidal cycle, the parcel would return to the same location
at the end of each tidal cycle.
STANDING TIDES
In bays and many estuaries, the incident progressive
tide is reflected. The tide in this region is the combina-
tion of the incident and reflected tides. An analogous
tide can be created in a channel by inserting a reflecting
barrier. The sea surface displacement tj above mean sea
level and horizontal water parcel velocity for the
standing tide are
17 = a cos
&Hf)
uma\v sin(^jsinlvj
(7)
(8)
where a is the tidal amplitude at the reflecting barrier,
x is the distance seaward from the barrier, and X =
(gh)lt2T is the wavelength of the tide.
The tide and tidal currents are out of phase. At mean
water, the strongest flood tide current occurs. It brings
water into the embayment whose sea level must then
rise. The incoming current finally ceases when the tide
outside the embayment has reached high tide. As sea
level begins to drop outside, an ebb current develops
which continues until low tide outside the embayment.
On the next rising tide, the flood tide again refills the
embayment.
An important parameter determining the characters
of tides in bays, sounds (large bays), and estuaries is the
ratio R = L/X, of the distance L between the reflecting
barrier and the mouth of the embayment to the wave-
length X of the tide. Where the ratio is small, such as a
deep bay or a small indentation in the coastline, the
tide as expressed in (7) has the same range as the tide
outside the embayment. This is because the factor
cos(27rx/X) determining the variation of the sea surface
displacement t\ over the embayment is equal to unity if
27rx/X is always much less than unity. The tidal currents
are small since the factor sin(27rx/X) is equal to 2irx/X, a
small number.
For example, a semidaily tide (T ~ 12.5 hours) in a
fjord with a depth h of 500 m and a length L of 100 km
has a wavelength X = (ghyi2T of 3180 km and hence a
maximum tidal current of 2.8 cm/sec at the mouth of
206
EFFECTS OF THE EARTHS ROTATION
59
the fjord (x = L) if the amplitude a of the tide is 1 in.
The current decreases linearly from the mouth to the
head where the reflecting barrier forces the horizontal
tidal current to be zero. Usually, fjords arc separated
from the outside by a shallow sill which can strongly
inhibit the tidal penetration into the fjord. The tidal
currents over sills can reach several meters per second.
The example above treats currents inside the sill.
For large, shallow embayments, areal variations in
tides can occur, if the length L is comparable to one-
fourth of the semidaily or daily tidal wavelengths.
One measure of this variation is the ratio between ampli-
tudes of the tide at the head of the embayment and at
the mouth:
■q{x = 0) 1
v(x = L) cos(2ttL X)
(9)
If L is close to X 4, the ratio is large so that the tidal
range inside the embayment is much larger than the
tide outside. The embayment is near resonance with
the tide.
The Gulf of Mexico is near resonance with the daily
tides. In shallow bays, such as the Bay of Fundy, the
Straits of Georgia, and Long Island Sound which are
near resonance, bottom stress and other effects limit the
motion. By assuming that the tides consist of progressive
waves that diminish exponentially with the distance
of propagation, Redfield (1950) has been able to repro-
duce most of the tidal characteristics in these bays. As
the incident wave propagates up a bay, its amplitude
diminishes; after reflection and return to the mouth of
the bay, the wave is significantly smaller in amplitude
than when it entered the bay. Near the mouth the tide
is progressive, whereas near the closed end of the bay
it is standing. The maximum amplification of the tide in
naturally occurring bays is about four times the incident
tide.
undistorted tides is a shallow water tide. The flood tidal
current is also greater than the ebbing current.
Ebb and flood channels occur in shallow water in
which the tidal currents in one direction are largely
confined to one set of channels and the currents in the
opposite direction are confined to other channels.
Ebb-flood channel systems are described in detail in
Chapter 10.
A shallow bar separating the shelf from an embayment
offers considerable resistance to tidal flow. Large
differences in sea level develop during the tidal cycles,
which generate strong tidal currents. While the velocity
of the water is large across the bar, the total amount of
water that can flow in and out of the embayment is
severely limited by the constriction. As a result, the
tide in the embayment is less than it would be without
the bar.
In an embayment having a complicated bathymetry,
a water parcel wanders into a variety of tidal regimes,
such as tidal flats and channels, bars, and shoals; the
simple theory predicting a return of the parcel to its
original position at the end of each tidal cycle does not
apply in this case. The Stokes drift induced by the
distinct tidal regimes is a net drift of the parcel which
may be thought of as a steady current.
Under some circumstances, tides should cause a net
transport of sediment through the Stokes drift. However,
this phenomenon has not been adequately documented
by field study.
In the frictional boundary layer near the bottom where
the horizontal tidal currents increase with height above
the bottom, the increase in wave tidal momentum with
height produces a steady current. This steady current,
which can advect suspended sediment, is driven by
variations in tidal momentum and is limited by turbulent
friction.
TIDES IN SHALLOW WATER
Where the tidal range is a significant fraction of the
depth, processes that cause the waveform to propagate
act in the deeper water under the wave crest to move
that part of the waveform more rapidly than the wave-
form near the trough. The tide becomes distorted, with
the slope of the sea surface greater on the leading side
of the crest. Where this distortion is large, there is
significantly more landward water discharge associated
with the crest of the tidal wave than there is seaward
discharge associated with the trough of the tidal wave.
This landward net transport of water is known as
Stokes drift. The difference between the distorted and
EFFECTS OF THE EARTH'S ROTATION
In larger bodies of water, the tides and tidal currents
are subject to the Coriolis effect, caused by the earth's
rotation. A moving water parcel experiences a force
proportional to its speed which, looking down on the sea
surface, is to the right in the northern hemisphere and
to the left in the southern hemisphere. This Coriolis
effect, when not counteracted by another force, drives
the water in an elliptical path: the direction of the
tidal current rotates clockwise in the northern hemi-
sphere and counterclockwise in the southern hemisphere;
the speed of the current is never zero. The semidaily
tidal currents in the Middle Atlantic Bight are an
example of this type of motion, shown schematically in
207
60
TIDAL CURRENTS
COT10E UNES
40°
33°
30°
70°
FIGURE 7. Theoretical corange chart jor the M: semidaily tide off the L'.S. east
coast. Ranges are in feet. From Redfield (1958).
Fig. 6; the corange and cotidal charts are given in
Figs. 7 and 8.
In regions where bottom stress can be neglected, the
motion is determined approximately by the following
equations:
du
-fo = ~g
dt
bv
dt
dr,
dt
^1
dx
+ /« = ~g
dr,
dy
= -h
/du dv\
\dx+ dy)
(10)
(ID
(12)
The second terms, — fv and +/«, in (10) and (11)
represent effects of the earth's rotation.
There are two ways in which the Coriolis effect can
alter tides and tidal current. In the case of a Poincare
wave, the water parcel trajectories are ellipses whose
major (larger) axis is in the direction of propagation;
the ratio of the major to minor axis is the inertial
period Te divided by the period T of the tidal con-
stituent. This type of tide can occur only where Te > T
and is generally found in exposed regions such as conti-
nental shelves. The semidaily tide in the Middle Atlantic
Bight (Fig. 7) is a standing Poincare wave (Redfield,
1958).
In restricted cmbayments such as the North Sea
(Fig. 4) or on continental shelves where the direction of
propagation of the tide is parallel to the coastline, a
slope in the sea surface set up against the shore can
balance the Coriolis effect. The result is a Kelvin wave.
For a coastline parallel to the x direction and located
at y = 0, a Kelvin wave has the form
(13)
(14)
(15)
H> ~ 7 )]
v = 0
208
IOTTOM STRESS 61
40°
35<
30c
CORANGE UNES
70°
FIGURE 8. Theoretical cotidal chart for the Mj semidaily tide of] the U.S. east
coast. The cotidal lines are in hours after the Greenwich transit of the M2 moon.
From Redfield (1958).
The tidal currents are parallel to shore (v = 0). At a
latitude of 45° and with a depth of 50 m, a Kelvin wave
decays to e~x (36.8%) of its magnitude at the coast in a
distance y = c/f of 286 km. Conversely, a Kelvin wave
propagating at 45°N along a continental shelf 150 km
wide with a depth of 50 m has a tidal amplitude 59%
of the amplitude at the coast.
A Kelvin wave propagating around a sea or ocean
produces an amphidromic system. When a Kelvin wave
enters an embayment in the northern hemisphere, such
as the North Sea, it propagates counterclockwise around
the embayment with the maximum tides and currents
nearshore. Because the Kelvin waves do not decay
rapidly away from their respective coasts, the motion at
any given location is a combination of Kelvin waves. As
a result, the tidal currents may not be colinear with the
bathymetry. The sense of rotation of the tidal current
direction is counterclockwise in this case, which is
opposite to the direction for a Poincare wave on a
continental shelf.
209
BOTTOM STRESS
Bottom stress modifies tides and tidal currents;^ its
effect is greatest where strong tidal currents occur in
shallow water. To model quantitatively the stress
applied by the sediment on the water above, the flow is
assumed to consist of a slowly varying tidal current
superimposed on turbulence. The distribution of
turbulent stress within the water determines the varia-
tion of tidal currents with distance above the bottom
(velocity profile) and the dissipation of tidal energy.
The details of flow near the bottom and estimates of
bottom stress are central to the study of sediment trans-
port.
The turbulent stress r is often modeled as proportional
to the rate of change of the current with increasing
distance z above the bottom:
- = A,
P
du
(16)
62
TIDAL CURRENTS
where the stress vector t is that part of the horizontal
stress caused by vertical changes in the horizontal
current u and Av is the vertical eddy or turbulent
viscosity. There are other terms caused by horizontal
variations in u which could be added to the stress, but
the term in (16) dominates turbulent processes in
shallow water. A layer of water will produce a force
opposite to the relative motion of the water just above
the layer. The slower moving water near the bottom
therefore acts as a drag on the water above. In general,
Av is determined by the spatial variations of currents,
distance from boundaries, stratification of the water
density, and the past history of the motion.
In turbulent boundary layers, a sublayer near the
boundary layer exists where the stress is constant and
the current speed increases logarithmically with distance
from the boundary:
1 ( t,, \ - 30c
« = -■•(- In-— (17)
k\ p ) Co
where c<> is the roughness length of the boundary which
is determined by bottom irregularities, t>, is the magni-
tude of the bottom stress, and /, is von Kai man's constant
(Ci;0.4). The effects of turbulence generally diminish
with height above the water; the inertia of the water and
the Coriolis effect become more important in balancing
the pressure force due to the sea surface slope. In an
oscillating tidal flow, the water farther from the bottom
is moving faster and therefore has more inertia than tin-
water near the bottom. In Fig. (> the water farther from
the bottom takes longer to respond to the pressure force
and lags in time the motion near the bottom.
To model the attenuation of progressive tides due to
bottom stress, an empirical formula is often used which
relates the bottom stress t<, p to the vertically averaged
tidal current I ':
*b = -C/pL'2 (18)
The bottom stress is proportional to the square of the
tidal current and opposite in direction to the current.
The stress depends on the depth // through the current
I ', which is inversely proportional to y/'h. The stress is
inversely proportional to // and hence is greater in
shallower regions. The constant of proportionality ('.,
lias been found from field studies to be about 0.0025.
A number of such studies are described in Proudman
(1952) for shallow regions around England.
INTERNAL TIDES
An internal tide is a wave with tidal period, associated
with displacements within the water column and with
very little displacement of the sea surface. Where there
are two layers, the currents are in opposite directions in
the two layers. The speed of propagation in this case is
/ Ap h,h2 V"
c = \g 7 ' aTT^] (l9)
where Ap p is the fractional change in water density
between the lower and upper layers, g is the acceleration
of gravity, and Ai and ht are the thicknesses of the upper
and lower layers. On a typical shelf with Ap/p ~
0.002, g = 980 cm sec2, h = 10 m ,and h2 = 50 m, an
internal wave would propagate with a speed c of 40.4
cm sec, which is about 60 times slower than the surface
tide's speed of propagation.
On the continental slope and at the shelf break, tidal
currents interact with bathymetry to produce vertical
displacements of density layers within the water column.
The resulting undulations propagate both shoreward
and seaward as internal tides.
As internal tides propagate inshore, the shoaling
bottom thins the lower layer and hence slows the wave.
Since the wave energy then becomes more concentrated,
the amplitude of the currents increases as does the dissi-
pation into turbulence. Sufficiently strong currents
produce an internal bore in analogy to tidal bores in
rivers. The internal tide becomes a series of pulses of
waves with periods of several minutes, the pulses
separated in time by the tidal period. The formation of
internal bores occurs when the internal tidal currents
equal the speed of propagation of internal waves.
Internal waves in a two-layered fluid cannot propa-
gate shoreward of the intersection of the density interface
with the bottom. Any internal waves that have not
dissipated will lose the remainder of their energy to
turbulence at the location where the water becomes
unstratified. On some narrow shelves with strong
stratification intercepting sharply rising bottom topog-
raphy, internal tides are reflected back to sea, producing
an internal standing tide.
A more realistic description of internal tides requires a
continuously stratified water column and the Coriolis
effect. The tides then propagate in the vertical as well as
the horizontal direction. Whether an internal wave as it
reflects off the bottom continues to propagate shoreward,
or whether it is reflected seaward, is determined by the
slope of the bottom and the direction of wave energy
propagation (slope of the wave characteristic). A bottom
slope steeper than the wave characteristic produces
reflection seaward. Smaller slopes allow the wave to
continue in the incident direction. A discussion of the
reflection process may be found in Cacchione and
Wunsch (1974).
Since the water density structure depends on the time
of year, the existence and behavior of internal waves are
210
SUMMARY
63
also seasonal. In summer when the shelf water is strati-
fied, internal waves ean exist over most of the shelf
regions; in winter, lbs lack of stratification precludes
occurrence of internal waves.
ADDITIONAL READING
This chapter was written to provide a qualitative intro-
duction to the study of tidal currents. There is a large
literature on tidal phenomena; as in any scientific field,
the recent research is presented in succinct journal
articles which presuppose a knowledge of the field.
There are a number of texts which treat tides and tidal
currents in much more detail and more quantitatively
than was possible in this chapter. The general texts
by Sverdrup et al. (1942), Proudman (1952), and
Dietrich (1963) provide such treatments. The text by
Neumann and Picrson (1966) is more recent and more
advanced.
SUMMARY
Equations may be written to describe the propagation
of an idealized tidal wave down a straight-walled
channel. If bottom stress and the Coriolis effect are
neglected, the wave is seen to propagate as a result of
the interaction between water level displacement and
the flow of water induced by this displacement. The
speed of the tidal wave form (c) is equal to (gf/)] '"', where
g is the acceleration of gravity and h is water depth,
while the speed of the associated current («) is propor-
tional to this value, In very shallow water, u is reduced
by turbulent dissipation of energy and frictional loss of
energy to the bottom.
In nature, tides arc propagated onto the continental
margin as waves from the open ocean. Such marginal
tides are said to cooscillate with the oceanic tide. Since
the incoming wave is rarely parallel to the coast, it
appears to propagate along the coast. Tidal waves
behaving in this fashion are referred to as progressive
tidal waves. The tidal wave may be partially reflected
back out to sea by the shoaling bottom and interact
with the next incoming wave so as to produce a standing
tidal wave. In a progressive tidal wave, maximum flood
velocity occurs at high water, while maximum ebb
velocity occurs at low water; in a standing tidal wave the
tide and tidal currents are out of phase, so that maximum
flood velocity occurs during the rising tide, and maxi-
mum ebb velocity occurs during the falling tide.
An important parameter determining the character
of tides in bays and estuaries is the ratio R = Z./X, where
L is the distance between the reflecting barrier and the
mouth of the embayment, and X is the wavelength of
the tide. When the ratio is small, the tide within the bay
has the same range as outside, and tidal currents are
small. However, if L is comparable to one-fourth of the
scmidaily or daily tidal wavelength, the embayment
resonates with the outside tide. Ranges are up to four
times higher, and currents are more intense.
When the tide range is a significant fraction of the
depth, the wave form becomes distorted, with the slope
of the sea surface becoming greater on the leading side
of the crest. The difference between the time-water
height curves of the undistorted and distorted tides is
called a shallow water tide. Where this distortion is
large, the velocity and discharge associated with the
crest are greater than those associated with the trough.
The resulting net transport of water is known as Stokes
drift.
In larger bodies of water, the tides and tidal currents
are subject to the Coriolis effect, caused by the earth's
rotation. A moving parcel of water experiences a force
proportional to its speed, which looking down at the sea
.surface, is to the right in the northern hemisphere, and
to the left in the southern hemisphere. On open conti-
nental margins, the pressure force associated with the
passage of the tidal wave, together with the apparent
Coriolis force, results in a water parcel following an
elliptical trajectory with right-hand sense of rotation.
A tidal wave behaving in this fashion is a Poincare wave.
It occurs where the inertial period 7 ',. is greater than the
period T of the tidal constituent. In restricted embay-
ments such as the North Sea, or on continental shelves
where the tidal wave propagates parallel to the coastline,
coastward water flow induced by the Coriolis effect
is blocked by the coast, and there results a slope of the
sea surface up toward the coast. A tidal wave thus
modified is a Kelvin wave. A Kelvin wave propagating
around a sea or ocean is known as an amphidromic
system. The sense of rotation is counterclockwise.
The turbulent stress r is often modeled as proportional
to the rate of change of the current with increasing
distance above bottom. The proportionality constant Av
is the vertical eddy viscosity. It is determined by the
spatial variation of the currents, distance from bound-
aries, stratification of the water density, and the past
history of the motion. In turbulent boundary layers, a
sublayer near the boundary exists where stress is con-
stant and the current speed increases logarithmically
with distance from the boundary. The slope of velocity
profile is determined in part by the degree of roughness
of the bottom, as measured by a bottom roughness
length Z0.
An internal tide is a wave with a tidal period, asso-
211
64
TIDAL CURRENTS
ciated with displacements within the water column,
and with very little displacement of the sea surface.
The wave may occur at the interface between fluids of
two densities, or may occur in a continuously stratified
fluid. On a typical shelf, an internal wave would
propagate with a speed about 60 times slower than the
surface tide's speed of propagation.
As the internal tide propagates inshore, the shoaling
bottom thins the lower layer and hence slows the wave.
Amplitude increases as does dissipation into turbulence;
eventually the wave becomes a bore. Internal waves in a
two-layered fluid cannot propagate shoreward of the
intersection of the density interface with the bottom.
At this point the waves lose their energy to turbulence,
or if the bottom slope is steep enough, are reflected.
SYMBOLS
A, vertical eddy coelhcient
a amplitude
Cf drag coefficient
c phase velocity of tidal wave
g acceleration of gravity
h water depth
A' a constant; von Karman's constant (~0.4)
L horizontal length scale
T period of tidal wave
/ time
IS vertically averaged tidal current
u current velocity
x horizontal distance
Z0 roughness length
Z vertical distance
X wavelength
r\ vertical displacement of sea surface with respect to
mean water level
p density
REFERENCES
Cacchione, D. and C. I. Wunsch (1974). Experimental study of
internal waves over a slope. J . Fluid Mech., 66: 233-239.
Dietrich, G. (1963). General Oceanography. New York: Wiley-Inter-
science, 588 pp.
Doodson, A. T. and H. D. Warburg (1941). Admiralty Manual of
Tides, London: HM Stationery Office, 270 pp.
Fleming, R. H. and R. Revelle (1939). Physical processes in the
ocean. In P. D. Trask, ed., Recent Marine Sediments. New York:
Dover, pp. 48-141.
Neumann, G. and \V. J. Pierson (1966). Principles of Physical
Oceanography. Englewood Cliffs, N.J.: Prentice-Hall, 545 pp.
Proudman, J. (1952). Dynamical Oceanography. New York: Dover,
409 pp.
Rcdfield, A. C. (1950). The analysis of tidal phenomena in nar-
row cmbayments. Pap. Phys. Oceanogr. Meteorol., 11(4): 1-36.
Rcdfield, A. C. (1958). The influence of the continental shelf on
the tides of the Atlantic coast of the United States. J. Mar.
Res., IT: 432-448.
Sverdrup, H. V. (1927). Dynamics of tides on the North Siberian
shelf, results from the Maud Expedition. Ceofys. Publ., 4: 5.
Sverdrup, H. V., M. VV. Johnson, and R. H. Fleming (1942).
The Oceans. Englewood Cliffs, N.J.: Prentice-Hall, 1087 pp.
U.S. Department of Commerce, National Oceanic and Atmos-
pheric Administration, National Ocean Survey (1974). Tide
Tables, East Coast of North and South Americas, 7973. National
Ocean Survey, Rockville, Maryland, 288 pp.
212
20
Reprinted from: Journal of Physical Oceanography, Vol. 6, No. 4, 596-602.
596 JOURNAL OF PHYSICAL OCEANOGRAPHY Volume 6
The Formation of the Yucatan Current Based on Observations
of Summer 1971
Robert L. Molinari
NO A A Atlantic Oceanographic and Meteorological Laboratories, Miami, Fla. 33149
2 January 1975 and 15 January 1976
ABSTRACT
Temperature, salinity and Lagrangian current data collected during the summer of 1971 in the western
Caribbean Sea are employed to evaluate the ageostrophic components of the flow in the formation region of
the Yucatan Current. The ratio of tangential and centripetal accelerations to Coriolis acceleration for data
averaged over 24 h periods remain less than 10% except in two areas. An anticyclonic turn, centered at
19°30'N, 86°W, has the largest centripetal accelerations, and in the region of Cozumel Island significant tan-
gential accelerations occur. The large-scale accelerations and additional evidence support the hy pothesis that
inertia! effects dominate in the formation of the Yucatan Current.
1. Introduction
The Yucatan Current is considered the first segment
of the Gulf Stream system, in the sense that current
speeds similar to those measured further downstream
are first observed in this region. Recent studies by
Molinari (1975) and Molinari and Kirwan (1975)
suggest that inertial effects dominate in the formation
of the Yucatan Current.
Fig. 1. The depth (m) of the 10°C isotherm, during the interval A, 14 to 23 July, 1971. The station locations are indicated by
solid circles. Trajectories for the center of mass of buoy triads are also shown (see text).
213
July 1976
NOTES AND CORRESPONDENCE
597
90°
88°
86c
84°
82°
i
80°
DEPTH OF THE 10°C
ISOTHERM IN METERS
24 JULY TO 1 AUGUST
CICAR SURVEY MONTH I
20 m CONTOUR INTERVAL
TRAJECTORY
%£ca
24°
22°
20°
90c
88° 86° 84° 82°
Fig. 2. As in Fig. 1, except for the interval B, 24 July to 1 August, 1971.
18°
16°
80fl
A two-ship operation was conducted in the summer
of 1971 to investigate the formation of the Yucatan
Current, and in particular to determine the nature and
extent of the ageostrophic component in the western
Caribbean Sea (Fig. 1) where the Current forms. In
this field study, the ageostrophic component of the
flow was determined by measuring the acceleration of
water parcels tagged by drift buoys.
The surface buoy was described by Molinari (1973);
nominally it senses the motion of the water at 40 m by
a parachute drogue. Data reduction techniques, and
the reduced surface drifter, temperature and salinity
data collected by the NOAA ship Researcher were also
discussed by Molinari (1973). These data, supplemented
by additional temperature data collected by Discoverer
(Hazelworth and Starr, 1975) are used in the following
sections to describe the formation of the Yucatan
Current in the summer of 1971.
2. Temperature-salinity data
Certain isotherm topographies are useful surrogates
for the density distribution in this area. In particular,
Molinari (1975) found that the 10°C topography
closely maps the geostrophic current regime in the
western portion of the Caribbean Sea. The study period,
14 July to 22 August, 1971, is divided into three time
intervals. The intervals are selected to provide spatial
resolution of the temperature field over the shortest
time span. Figs. 1-3 show the 10°C topographies during
these intervals (identified as A, B, C).
The 10°C topography at 83°\V (Fig. 1) has a slope at
18°N consistent with a narrow and fast geostrophic
current. The maximum slope occurs along a band
bounded bv the 460 m and 520 m contours. The band
extends continuously from 83°\V, 18°N to the Yucatan
Strait. The acceleration along this band is not mono-
tonic. Rather, the direction of and gradients along the
band varv, suggesting accelerations, decelerations and
meanderings of the current.
The temperature structure indicates the presence of
eddies on either side of the maximum slope. The flows
around these eddies are both cyclonic and anticyclonic.
Figs. 1 and 2 suggest that there is a south to north
214
598
JOURNAL OF PHYSICAL OCEANOGRAPHY
88° 86° 84° 82°
Volume 6
88° 86° 84° 82°
Fig. 3. As in Fig. 1, except for the interval C, 8 to 22 August, 1971.
80e
decrease in eddy size, with no resolvable eddies found
at the Yucatan Strait.
The migrations of the 460 to 520 m band show the
temporal changes which occur in the main current.
The only significant temporal variability occurs in the
southern basin where the large eddy initially centered
at 17°30'N, 86°W (Fig. 1) apparently drifts to the
west. There is no significant movement of the band at
the Yucatan Strait during the six-week time period.
3. Drifter data
The technique for reducing the drifter data was
described by Molinari (1973) and Molinari and Kirwan
(1975). This procedure provides geographic positions
every 2h. The individual drogue trajectories are shown
in Fig. 4.
Four sets of trajectories were obtained, and with at
least three drifters being deployed in each set. Legs 1
and 2 employed the same buoys which drifted un-
attended from points 4 to 5 during an emergency port
call. The buoys were retrieved at the end of leg 2, and
redeployed in an unsuccessful attempt to sample the
cyclonic flank of the current (leg 3). After a schedulep
port call leg 4 was initiated, and then prematurely
terminated when a storm threatened the operations
area.
The results of the drifter analysis are presented below
in increasing order of the derivatives of the buoy
coordinate-vs-time functions, i.e., trajectories, speeds
and accelerations. Finally, the ageostrophic components
of the flow, as evaluated from the drifter data, are
discussed.
a. Trajectories
Trajectories are computed for the center of mass of
the buoy triads, and are given in Figs. 1-3. The current
fields inferred from these trajectories are very similar
to the circulation fields inferred from the 10°C topog-
raphies. For instance, in the Yucatan Strait region the
trajectories closely follow the depth contours of the
10°C isothermal surface during all three time intervals.
However, in the southern basin the trajectories parallel
only those contours obtained concurrently, i.e., the
leg 1 trajectory parallels the phase 1 temperature field
215
July 1976
NOTES AND CORRESPONDENCE
599
(Fig. 1), and the leg 4 trajectory parallels the phase 3
temperature field (Fig. 3). This result is consistent
with the temporal variability observed in the tempera
ture field and discussed above.
Visual inspection of the trajectories in Fig. 4 indicates
that large-amplitude meanders did not occur in the
area. The largest curvature in the trajectories occurs in
the anticyclonic turn indicated in the temperature
fields at 19°30'N, 86°W (Figs. 1-3). The average
anticyclonic radius of curvature from points 7 to 10
shown in Fig. 4 is 75 km.
b. Speeds and accelerations
Buoy speeds have been computed from the position
data by using a centered difference approximation to
the differential. The accelerations have been computed
in natural coordinates, i.e., downstream s, crossstream
n (positive to the left of the downstream axis), and
vertical z (positive up). The accelerations are tangential
(dV/dt), centripetal (KV2) and Coriolis (fV), where
d( )/dt = d( )/dt+Vd( ) ds, V is the measured speed,
K the horizontal curvature (positive for downstream
cyclonic turning), and / the Coriolis parameter.
The trends in the individual buoy speeds have been
determined by fitting in a least-squares sense a cubic
polynomial in time to the speed values. The fitted
polynomial trend has been subtracted from the observed
values to arrive at residual speed curves. The speed and
residual speed curves for each drifter, and the poly-
nomial fit curve for one drifter, are given in Fig. 5.
The large-scale accelerations inferred from the
temperature data are apparent in the fitted polynomial
curves. For instance, the temperature gradients of the
anticyclonic turn centered at 19°30'N, 86°\V suggest
a deceleration in the flow. This deceleration occurs at
206/1930 (Fig. 5), as the drifters enter this turn. The
temperature gradients increase as do the buoy speeds
(leg 3, 210/0315 to 213 1315), as the drifters approach
the Yucatan Strait. The largest average downstream
accelerations occur in the vicinity of and north of
Cozumel Island, where the speeds approach those
observed in the Gulf Stream. These large-scale accelera-
tions occur on time scales of days and/or space scales
of hundreds of kilometers.
Smaller time and or space scale disturbances are
superimposed on these lar^e-scale features. In Fig. 5,
the speed curves with the trends removed show that the
amplitude of these oscillations are relatively constant
Fig. 4. Drogue trajectories as determined from 2 h positions. The trajectory of buoy 2 is continuous
during legs 1 and 2 (identified by circled numbers), although the buoy was not continuously tracked
from interval 4 to 5. The first and last position times (Julian day/hour) are: leg 1, 200/1830 to 204/
1630; leg 2, 205/1930 to 209/0730; leg 3, 210/0315 to 213/1315; and leg 4, 228/1600 to 232/1000.
216
600
JOURNAL OF PHYSICAL OCEANOGRAPHY
Volume 6
I.Oi-
Ll) .
a.
in.
o
Ul
Ul
Q.
W
_l
<.
3
Q.
C/5
UJ
q:
a 8
2.0
16
1.2 -
J L
i LkJ I L
200/1830 201/1830 202/1830 203/1830 204/1830
2P 1 1 1 1 1 1 1 1
^M
I I I L
O
UJ
UJ
CL
in
BUOY I
BUOY 2
BUOY 3
BUOY 4
POLYNOMIAL FIT CURVE
I.Or
o 8 -
4
210/0315 211/0315 2I2/03I5-
2,1 ri ' ' '
4 I
213/0315
1 1
205/1930 206/:930 207/1930 208/1930
p — r^
5
hlA-
3
TMy*
1
_i_. V i i
8 P ^^Vy*
228/1600 229/1600 230/1600 231/1600
2P 1 1 1 1 1 1 1 1
FlG. 5. The observed and residual speed curves for the trajectories of Fig. 4 as a
function of time (Julian day/hour). The residual speeds are determined by sub-
tracting the third-degree polynomial fit curve (a representative curve is shown in
the upper panels of each time interval) from the observed curve.
throughout the basin, although their primary period
appears to vary. A visual inspection of the records
indicates that the principal period in the southern
speed data is 24 h, and in the northern data, 12 h.
As indicated, no large-amplitude meanders, similar
to Gulf Stream meanders, appear in the trajectory
data (Fig. 4). Thus, the velocity oscillations discussed
above occur along the axis of the flow, rather than
normal to it. The downstream spatial extent of the
oscillations vary from approximately 25 km in the
low-speed regions to 75 km in the high-speed region
of the basin.
c. Ageostrophic components
The horizontal equations of motion can be written in
natural coordinates as
dV dD
—+—=Ru
dt ds
dD
KV2+JV+ — =i?2)
dn
(1)
(2)
where D is the dynamic height relative to a level of no
motion, and Ri and /?•> include all the forcing and
retarding functions. In a frictionless system the
ageostrophic components are the centripetal accelera-
tion KV'2 and the tangential acceleration dV /dt.
The terms on the left-hand side of (1) and (2) are
evaluated for those portions of the trajectories where
density data are available. Dynamic heights are
computed relative to 600 m, since the majority of the
hydrographic casts were to this depth. Accelerations for
the center of mass trajectories and the density gradients
are averaged over 24 h periods to eliminate the small-
scale oscillations shown in Fig. 5. Table 1 lists these
average properties at 12 h intervals.
If i?> = 0, Eq. (2) becomes the gradient equation. The
gradient balance expressed in percentages as [_Ri/
(A'V2+/V)]X100 was maintained on the average to
within 10% durings legs 1, 2 and 3. The gradient balance
computed relative to 600 m was not maintained during
leg 4 for undetermined reasons (although internal wave
aliasing of the density field may be a cause).
The terms in (1) are an order of magnitude less than
the terms in (2) and therefore more difficult to evaluate
realistically. However, the large-scale tangential accel-
erations are consistent ■• with downstream pressure
gradients observed durirtg legs 2 and 3 (Fig. 4).
217
July 1976
NOTES AND CORRESPONDENCE
601
Table 1. Ageostrophic and geostrophic components averaged over 24 h periods.
\KV*\/fV
\dV/dt\//V
Time Julian
V
fv
dD/dn
xioo
X100
Fig. 4 point
day/hour
(cm s"1)
(cm s~2X 104)
(cm s-2X104)
(percent)
(percent)
1
202/1330
0.69
0.31
0.27
8
2
2
203/0130
0.68
0.30
0.25
11
2
3
203/1330
0.71
0.32
0.31
6
4
4
204/0130
0.71
0.33
0.36
7
3
5
206/0830
0.61
0.28
10
4
6
206/2030
0.58
0.27
9
4
7
207/0830
0.58
0.28
0.20
12
3
8
207/2030
0.61
0.30
0.22
17
10
9
208/0830
0.69
0.34
0.28
12
8
10
208/2030
0.76
0.38
7
3
11
211/0215
0.66
0.32
4
3
12
211/1415
0.75
0.37
0.29
13
4
13
212/0215
0.94
0.47
0.34
8
13
14
212/1415
1.15
0.59
0.50
1
8
15
213/0215
1.49
0.70
1
7
16
229/0500
0.32
0.14
2
5
17
229/1700
0.33
0.15
9
4
18
230/0500
0.30
0.14
3
1
19
230/1700
0.29
0.13
7
3
20
231/0500
0.28
0.13
9
8
21
231/1700
0.26
0.12
0
6
The data listed in Table 1 verify the results described
qualitatively in previous sections ; that is, the large-
scale flow undergoes little acceleration in the mid-basin
(points 1-4, Fig. 4), the largest centripetal accelerations
occur in the anticyclonic turn centered at 19°30'N,
86°W (point 8, Fig. 4), and the largest accelerations
occur in the vicinity of Cozumel Island (point 13,
Fig. 4). If Ri = R2 = 0, the large-scale flow is geostrophic
to within 20%. In particular, although the average
velocity more than doubles from points 1 1 to 15 (Fig. 4),
the maximum value of the ratio [(f/t/<//)//F]Xl00 is
only 13%.
If the ageostrophic components are evaluated for
intervals of monotonically accelerating or decelerating
flow (Fig. 5), the ratios of ageostrophic to geostrophic
terms are considerably higher than those listed in
Table 1. For instance, over these shorter averaging
intervals the centripetal acceleration is as large as
one-third the Coriolis acceleration, and the tangential
acceleration is as large as one-fourth the Coriolis
acceleration.
4. Discussion
For the average accelerations listed in Table 1, the
balances expressed in (1) and (2) are confirmed qualita-
tively for (1), and quantitatively for (2). Molinari and
Kirwan (1975) demonstrate qualitatively that potential
vorticity is conserved for portions of legs 2 and 3
(points 5-9 and 11-15, Fig. 4). In addition, the absolute
value of their relative vorticity is approximately 40%
of the value of the planetary vorticity on the northern
portion of these trajectories.
The 1971 measurements also suggest that if lateral
friction plays a role in the formation process of the
Yucatan Current it is limited to a narrow band along
the Yucatan Peninsula. A frictional boundary layer
can be characterized by a cyclonic shear zone, but the
westernmost buoy of leg 3 (Fig. 4) exhibits the highest
speeds measured (Fig. 5), indicating the buoys are in
the anticyclonic zone of the current. The only indication
that the buoys may be on the cyclonic flank of the
current occur as this buoy drifts over Arrowsmith Bank.
However, it is difficult to ascertain if the speed reduction
is a local effect of the bank topography, or if the buoy
has indeed crossed the speed axis. The preceding results
provide additional support for the contention of
Molinari (1975) that inertial effects dominate in the
formation of the Yucatan Current.
The 24 h average ageostrophic components listed in
Table 1 are seldom greater than 10% of the Coriolis
accelerations. As discussed, if shorter averaging periods
are used the ratio of ageostrophic to geostrophic
components increases. The effect of these smaller
scale features on the formation process are still matters
for speculation. Plausible explanations are wind and/or
tidal forcing of the upper layers. For instance, the
dominant semi-diurnal and diurnal periods of the
speed oscillations (Fig. 5), their constant amplitude
throughout the basin, and the predominantly down-
stream orientation of these disturbances suggested a
tidal modulation of the flow. However, the results of a
tidal analysis performed on the data were inconclusive.
218
602
J O U R N A L OF PHYSICAL OCEANOGRAPHY
Volume 6
Acknowledgments. The invaluable assistance in the
collection of data of the officers and crew of the Re-
searcher is gratefully acknowledged.
The data analysis was facilitated by the programming
of Mr. A. Herman, and the efforts of Mr. D. Tidwell.
The figures were drafted by the publication and
presentation group of Mr. R. L. Carrodus.
This work was supported in part by the National
Science Foundation, under International Decade of
Ocean Exploration Grant AG-253.
REFERENCES
Hazelworth, J., and R. B. Starr, 1975: Oceanographic conditions
in the Caribbean Sea during the summer of 1971. NOAA
Tech. Rep. ERL 344-AOML 20, 144 pp.
Molinari, R. L., 1973: Data from the Lagrangian current measure-
ment project conducted aboard the NOAA ship Researcher
during CICAR Survev Month I. NOAA Tech. Memo. ERL-
AOML-19, 81 pp.
, 1975: A comparison of observed and numerically simulated
circulation in the Cayman Sea. /. Pliys. Oceanngr., 5, 51-62.
, and A. D. Kirwan, 1975 : Calculation of differential kinematic
properties from Lagrangian observations in the western
Caribbean Sea. /. Phys. Oceanogr., 5, 483-491.
219
21 Reprinted from: Proc. AIAA Drift Buoy Symposium, Hampton, Va., May 22-23,
1974, NASA CP-2003, 193-209.
CALCULATIONS OF DIFFERENTIAL KINEMATIC PROPERTIES
FROM LAGRANGIAN OBSERVATIONS
by
Dr. R. Molinari
Atlantic Oceanographic and Meteorogical Laboratories
Dr. A. D. Kirwan, Jr.
Texas A&M University, Department of Oceanography
INTRODUCTION
In the past oceanographers have used Lagrangian data, primarily to obtain
elementary fluid properties such as trajectories, velocities, and accelarations
However, meteorologists have recognized the utility of Lagrangian data in
determining estimates of the differential kinematic properties, divergence,
vorticity, shearing deformation, and stretching deformation. These properties
are important ingredients in any description and/or explanation of fluid
motions. For instance, divergence is an important factor in determining
vertical motion in the ocean, vorticity can be related to the field of force
that drives ocean flows, and the two deformations are important in the
formation and dissipation of fronts.
The Authors: Robert Molinari received his Ph.D. in Physical Oceanography
from Texas A&M University in 1970. Since 1971 he has held
the position of Research Oceanographer with N0AA/A0ML. His
work has centered on observational and theoretical study of
the Cayman Sea and the Gulf of Mexico.
Dennis Kirwan, Jr. also received his Ph.D. from Texas A&M.
He has been an associate Professor at New York University
and worked as a Program Director with the Office of Naval
Research. More recently, he has become Research Scientist
at Texas A&M and has been involved in drift buoy studies.
220
iwo methods are presented for the calculation of these properties. One
method is more readily applicable to a large number of buoys. The other
approach is given to provide estimates to verify the results of the more
general technique. A short description of the experiment and data analysis
is given.
DATA COLLECTION AND ANALYSIS
Figure 1 is a schematic diagram of the ship-tracked buoy used in the
experiment. The buoy nominally was tied to a water parcel at 40 m by a
35-ft diameter parachute.
The experiment was conducted in the western Caribbean Sea in the summer of
1971, aboard the NOAA ship RESEARCHER. The prime navigational control was
supplied by a satellite positioning unit. In that region, satellite fixes
can be obtained on the average ewery 1.5 hours. The satellite positions were
supplemented by Omega fixes collected every 15 minutes. The buoys were
positioned relative to the ship at each fix.
Errors are introduced into the buoy positions by the imprecision of the
satellite, Omega, and radar systems. Assuming the satellite system to be
the more precise of the two positioning techniques, an estimate of the Omega
errors was made. An individual Omega position is accurate approximately to
±2 km. Thus, there is a yery small signal-to-noise ratio when considering
the 15-minute fixes.
The following smoothing procedure was applied to the Omega fixes to eliminate
some of the noise in the trajectory data. Hourly fixes were obtained by
taking 5-point running averages of the 15-minute component coordinates. A
second degree polynomial curve was then fitted to 13 consecutive hourly fixes
to arrive at the data used in the analysis.
Kirwan, in a previous talk, indicated other possible sources of error when
attempting to tag a particular water parcel. Using his analysis for the
drifter configuration used in this experiment, it was found that a 10 m/sec
221
wind could cause a 5 percent error in the estimate of the true current.
This effect was not considered in reducing the data from this experiment.
A. Least Square Method
Consider a small, but finite, parcel of water, and assume that within
this parcel the velocity at any point is adequately represented by the
linear terms in a Taylor's expansion about the center of mass of the
parcel. For a cluster of N drifters located within the parcel, the
expansion yields for the velocity components of the i^h drifter.
U. = U + gi + {(D + N) X.} II + {(S - c)Yi 1/2
i = 1....N (1)
V. = V + h1 + {(S + c) X.} II + {(D - N)Yi }/2
The U and V are the components of the velocity of the center of mass of
the parcel. The coordinates with respect to the cluster center of mass
of drifter i are X. and Y.. The g. and h. represent the sum of the higher
order non-linear terms in the expansion.
The differential kinematic properties are:
D = 3U/3X + 3V/3Y (Divergence)
3 = 3V/3X - 3U/3Y (Vorticity)
S = 3V/3X + 3U/3Y (Shearing deformation rate) (2)
N = 3U/3X - 3V/3Y (Stretching deformation rate)
The divergence, D, is a measure of the parcel volume change without change
of orientation or shape. 5, the vorticity, is a measure of the orientation
change without volume or shape change of the parcel. Shape changes without
change of volume or orientation are given by S and N respectively.
In equation (1), U ■ , V., and U and V are computed from the buoy coordi-
nates. The g and h functions, and D, N, S, and i can be computed by noting
222
that at each time the total kinetic energy density of the cluster
due to small-scale turbulence is:
N
KE = I gf + hf II (3)
i = 1 n ]
Substituting (1) into (3) shows that the kinetic energy density depends
on the kinematic properties. These four parameters can be estimated by
selecting values which give a minimum for the kinetic energy density.
The g and h functions can then be determined from (1).
The minimum number of drifters that can be used to determine D, S,
N, and r, is three. However, this approach is readily extended to con-
sider larger numbers of drifters. In addition, the approach generates
time series of the turbulent velocities, g. and h. , from which direct
estimates of turbulent stresses can be made.
Area Method
Horizontal divergence can be expressed as the fractional time rate of
change of the horizontal area, A, of a parcel:
dU 9V 1 DA (4)
9X 3Y A Dt
For a triad of drifters, A is readily evaluated from the buoy positions.
From the time series of A's an appropriate numerical technique is used
to estimate the time rate of change.
Vorticity, shearing and stretching deformation can be evaluated by
selected rotations of the velocity vectors of the three drifters. Saucier
(1955) describes this technique.
223
RESULTS
Figure 2 is a schematic diagram of representative drogue trajectories and
speeds. Four trajectories are shown. Beginning with the trajectory over
the Cayman Ridge and preceeding counterclockwise around the basin, the
trajectories will be numbered 1, 2, 3, and 4 for purposes of identification.
The area of figure 2 is the formation region of the Yucatan Current. The
accelarations which occur during legs 2 and 3 are indicative of the forma-
tion processes occuring in the vicinity of the Yucatan Channel.
Figure 3 is a more detailed plot of the drogue trajectories of leg 1.
A Universal Transverse Mercator projection is used, and the x and y coor-
dinates are marked in kilometers. The apexes of the triangles represent
drogue positions.
Also given on the figure are the velocity, accelaration, and radius of
curvature of the triad center of mass. The last two curves indicate the
difficulty of obtaining from these data smooth estimates for higher order
derivative terms.
Figure 4 gives the divergence and vorticity as determined by the two methods.
The solid lines connect the values computed by the least square approach,
and the crosses represent the values computed by the area method. The triad
areas (figure 4) are small and the estimates of the kinematic properties are
very irregular with respect to time.
Figures 5 and 6 present the buoy trajectories and all the kinematic properties
for leg 2. Again, the agreement between the two methods is good. The estimates
of these parameters are smoother functions of time for this leg.
224
Figures 7 and 8, and 9 and 10 display the results for legs 3 and 4 respectively
The buoy speeds were lowest during leg 4, on the average 0.3 m/sec, and the
triangle areas small. The kinematic property estimates given on figure 10 are
yery ragged, with frequent crossings of the axis. It is doubtful that these
values are reliable estimates of the differential kinematic properties.
The value of the measurements is increased if the resulting data can be
used to explain the dynamics of the circulation. An attempt to incorporate
the data of leg 2 (figure 6) into a dynamic expression is made.
Figure 11 gives the conservation of potential vorticity relation, and the
evaluation of the terms in this relation using the data of leg 2. This
equation is derived by assuming no external forces (tides, winds) are acting
on the flow. The terms in this expression are Z, the relative vorticity, f,
the Coriolis parameter, and V-V, the divergence. The qualitative balance of
the terms for the first two days of the trajectories suggests a balance exists.
To summarize, it appears feasible to compute differential kinematic properties
from drifting buoy data. In addition, if estimates are sufficiently well-
behaved, some dynamical statements about the flow can be offered.
225
Aluminum alio1
Buoy ^
Polyurethanc foon
Filling "^
r
l
6
1
-K
i
-13 J--.
^.Radar
r — \^"~ Tronsponder
Battery
-1 Compartments
f-L «, 90'.
100 (b Load
Weigh* -
Tether Balls
200'
v-^_35 ft. Diameter
Parachute
figure l .
226
CD©
Zl o
CD
=3
a
0
o_
rt
CM
227
in
CM
oo
O CO U5 t CM O
-- o o o o o
(03S/W) A±IOO"13A SSVW JO M31N30
O
o
00
in
UJ
(-)
o
o
m
z
CD
N
<
<.-
h-
'."3
cn
1
Q
I
X
in
CM
o
o
o
m
o
CM
m
CM
o
CM
O
O
O
CM
m
MX) 30NV1SIQ-A
228
u
-2-"
-4
I \
I \
A Av
1\
Ml V
M
-i »-
-i 1 i 1-
i >
202
0530
time:
a'0 3
0 5 30
TIME
204
0530
Figure 4
229
2275
2250-
2225-
2200
o
-z.
(/)
o
I
2175-
2150 -
2125
2100
550
575 600
X-DISTANCE (KM)
625
650
Figure 5
230
0 J30
TIMH
Figure 6
231
2
Li
O
2
<
H
co
o
i
>-
2375
2350
2325
2300 -
2275 -
2250 -
2225 -
2200 -
2175 -
2150 -
500
525
550
575
600
X- DISTANCE (KM)
Figure 7
232
u
<^
2
U 2
ID *
vc
hi
>
s o^
-2--
-4
H 1 1 I
TIME
t-
q:
a
>
-2-
210
2015
201 5
TIME!
212
2015
213
201S
Figure 8
233
to
m
o
lO
lO
^^
S
*:
<y>
uj
o
<u
z
j-
<
3
i-
O)
C/)
•r-
m Q
Li_
cvi ■
in x
o
o
m
m
o
(VJ
O
in
o
in
cvj
O
(IAI>I) 33NV1SIQ-A
234
219
no
J3I
I 100
1100
2 100
TIME
Figure 10
235
o
CO
O
^°
O
C\i
cc
o
:r>
CO
o
o
o
X
v.
oo
\
o
CM
>-
f— i
Ck
^
<s
C7)
— m^
•r-
u
U_
:d
~^
**— ►
C
CO
O
UJ
o
C\J
K
o
CO
o
^°
o
z03S/ni_OlX
01
236
22
Reprinted from: NOAA Data Report ERL MESA- 22, 43 p.
ABSTRACT
During January 1975, an oceanoqraphic cruise, denoted XWCC-1 was
made by R/V Advance II in the New York Bight. The objective of the
cruise was to supply data for analysis of the water characteristics
in the New York Bight. This report presents the physical and
chemical data from this cruise, and describes the parameters meas-
ured, the measurement methods, and the procedures for reducing the
data.
237
23
Reprinted from: Journal of Physical Oceanography, Vol. 6, No. 6, 953-961.
Reprinted from Journal oi" Physical Oceanography, Vol. 6, Xo. 6, November 1976
American Meteorological Society
Printed in U. S. A.
The Influence of Deep Mesoscale Eddies on Sea Surface Temperature
in the North Atlantic Subtropical Convergence1
Arthur D. Voorhis and Elizabeth H. Schroeder
Woods Hole Oceanographic Institution, Woods Hole, Mass. 02543
Ants Leetmaa
Atlantic Oceanographic and Meteorological Laboratory, NOAA, Miami, Fla. 33149
(Manuscript received 24 February 1976, in revised form 8 July 1976)
ABSTRACT
Maps of sea surface temperature in the North Atlantic subtropical convergence during the 1973 MODE
field experiment (and recent satellite imager)') show large meridional and zonal features on a scale of 40-400
km which are superimposed on the seasonal meridional temperature gradient. After comparing these maps
with dynamic topography relative to 1500 db it is argued that these features are mainly due to advective
distortion by surface currents associated with the deep baroclinic mesoscale eddy field. Wind-induced
surface currents appear to have a lesser effect in generating such structure. Surface frontogenesis observed
during MODE and by earlier workers in the area suggests that jet-like shallow surface density currents
may be also significant in advecting and distorting the surface temperature field on scales of 10 km and less.
Finally, rough calculations indicate that these advective processes of the sea surface may supply annually an
amount of heat to the surface water mass of the northern Sargasso Sea which is significant compared with
that lost to the atmosphere.
1. Introduction
An important goal of contemporary oceanography
is to understand the horizontal distributions of proper-
ties at the sea surface and the mechanisms that
produce them. Major programs such as the North
Pacific Experiment (NORPAX) in the United States
and the Joint Air-Sea Interaction Experiment (JASIX)
in the L'nited Kingdom are actively working in this
area. Although progress has been made in modeling
the vertical structure of the upper ocean, much remains
to be done in modeling its horizontal structure, particu-
larly over intermediate oceanic scales of 100 to 500 km.
Presented here are observations in the subtropical
convergence of the western North Atlantic which
provide a reasonably detailed look at one aspect of
this problem.
The subtropical convergence is one of the classical
transition zones (Wiist, 1928) separating two meteoro-
logical regimes. In the western North Atlantic it lies
roughly between 22°N and 32°N latitude and sepa-
rates the prevailing westerlies to the north from the
easterly trades to the south. Maps of monthly mean
sea surface temperature in this zone are relatively
simple eastward of the Gulf Stream's influence. This
can be seen in Fig. 1. In general, the temperature
1 Contribution No. 3709 from the Woods Hole Oconographic
Institution.
decreases northward at all times of the year by an
amount which varies seasonally. The maximum me-
ridional gradient occurs in late winter (approximately
0.5°C per degree of latitude) and the minimum in
late summer (approximately 0.1°C per degree of
latitude). The large-scale zonal temperature variation
is small.
How does the synoptic temperature distribution,
that is, the actual temperatures at any one time,
differ from the above average picture? Surveys of
thermal fronts in the area (Voorhis, 1969) suggested
that major differences occurred on surprisingly large
scales of hundreds of kilometers. More recent evidence
comes from satellite infrared imagery of the sea sur-
face, such as shown in Fig. 2. In this image the syn-
optic temperature is dominated by large meridional
and zonal variations on the same scale. What is the
reason for this large-scale structure?
Except for the Aries data (Crease, 1962) little was
known about the sub-surface currents in this region
until the Mid-Ocean Dynamics Experiment (MODE),
which was conducted over a period of four months
during the spring of 1973 in a 400 km square area
centered at 28°99'N, 69°40'W. The principle finding
from MODE was that subsurface currents were
dominated by eddy-like motions, having spatial scales
of several hundred kilometers and time scales (resi-
dence time) of two to three months, which were
238
954
JOURNAL OF PHYSICAL OCEANOGRAPHY
Volume 6
APRIL 17,589 OBS
AVERAGE SURFACE TEMPERATURES ,"C
Fig. 1. Climatic mean April sea surface temperature in western
North Atlantic after Schroeder (1966). MODE was conducted
during March-April, 1963, in cross-hatched area.
nearly in geostrophic balance with vertical deforma-
tions of the main thermocline. We have collected all
useful sea surface temperature measurements made
during MODE and constructed maps which, in the
following, are presented and compared with the sur-
face geostrophic circulation. From this we argue that
most of the large-scale synoptic surface temperature
structure in the subtropical convergence is due directly
to surface advection by the geostrophic eddy field.
2. The surface temperature field
MODE was designed to study deep currents and
water structures and very little effort was expended
in surface measurements. In addition, no adequate
satellite thermal images of the surface were obtained
during the experiment. Nevertheless, surface tempera-
tures were recorded continuously from three ships2
as they maneuvered about the area during the four
months of thi experiment. We have used these data
plus that from CTD and STD casts to describe the
large-scale evolution of the surface temperature field.
In order to retain adequate spatial coverage, we chose
1 R. V. Chain from the Woods Hole Oceanographic Institution,
Woods Hole, Mass. ; R. R. S. Discovery from the United Kingdom ;
and NOAA ship Resecrcher.
to group the data in successive time periods of about
15 days. The spatial coverage during the period 31
March to 14 April, which was typical, is shown in
the upper half of Fig. 5.
For each period the records of surface temperature
were sub-sampled every 10 min to the nearest 0.2°C
and plotted along the ship's tracks. These were com-
bined with temperatures from the STD and CTD
casts, which were also used to calibrate the surface
temperature records. The total data set was then
subjectively contoured by one of us (Schroeder) in
half-degree intervals without any prior knowledge of
the field of surface currents discussed in the next
section. For the most part spurious spatial effects
introduced by diurnal heating and cooling along the
tracks of the moving ships could be recognized and
eliminated in the contouring. Data were rejected,
however, on about ten days when afternoon diurnal
heating during calm weather exceeded 0.5°C. The
resulting maps from 9 March through 13 July arc-
shown in the upper part of each picture in Figs. 3
and 4.
The average surface temperature and its average
meridional gradient over the MODE area were com-
puted from each map and these are shown as a func-
tion of time in the lower half of Fig. 5. The spatially
averaged temperature was close to its late winter
minimum (approximately 22°C) at the start of the
experiment. Thereafter it increased, due to surface
heating, to almost 27°C at the end of the experiment.
The average meridional gradient is, perhaps, not too
meaningful because of the fluctuations introduced by
eddy distortion in the MODE area. Nevertheless, it
shows an overall decrease during the experiment from
approximately 0.007 to 0.002°C km-1, with cooler water
always" to the north.
All of the temperature maps in Figs. 3 and 4 show
a large, changing zonal and meridional structure
superimposed on the mean meridional gradient.
Usually, this structure is dominated by long intrusive
features or tongues of alternate warm and cool water,
40-50 km wide, which can extend for distances of
several hundred kilometers. The resemblance between
the temperature pattern in most of these maps and
that shown in the satellite image in Fig. 2 is remarkable.
3. Mesoscale eddy surface currents
The mesoscale eddy field found during MODE has
been discussed by Robinson (1975), McWilliams
(1976), and by other participants in the report of
the MODE-I Dynamics Group (1975) cited in the
references. The purpose here is to describe the motion
of the sea surface in a way which allows one to see
its effect on the distribution of surface temperature.
Over 800 CTD, STD, and hydrographic lowerings
were made during the experiment from the sea sur-
face to depths greater than 2000 m. In addition, cur-
239
November 1976 A. D. VOORHIS, E. H. SCHROEDER AND A. LEETMAA
955
Fig. 2. Enhanced infrared satellite image of sea surface (courtesy of A. Strong and R. Legechis,
NOAA/NESS, Suitland, Md.) showing large -scale advective patterns on 2 April 1974 along latitude
30°N in western North Atlantic. Dark areas are warm water, grey are cool water, and white are clouds.
The warm water of the Gulf Stream appears all along the left on the image. MODE was conducted the
previous year in the outlined quadrangle.
rents were measured extensively at depths below 400 m,
primarily by moored current meters and by drifting
neutral buoyant floats. These currents, when averaged
over a period of several days, have been shown by
Bryden (1974) to be in geostrophic equilibrium with
the density field within the limits imposed by mea-
surement noise. Furthermore, the vertical structure
of these time-averaged currents is highly baroclinic,
with most of the eddy energy confined to the main
thermocline and above. One concludes, in fact, from
the analysis of Schmitz el al. (1976) that about 80%
of the geostrophic surface current is, on average, due
to density structure above 1500 m. To describe the
eddy surface motion, therefore, we computed the
d)-namic height of the sea surface relative to 1500 db
from the CTD and STD data and constructed the
maps of dynamic topograph}- (in meters) shown by
the solid contours in the lower portion of each picture
in Figs. 3 and 4. The mapping time interval was
chosen to be the same as that for the accompanying
surface temperature maps. For each map the direction
of geostrophic current is indicated by the arrows and
the magnitude can be estimated from the geostrophic
speed scale shown. The positions of the CTD and
STD stations are given by small dots. The maps
were contoured by computer in intervals of 4 dyn cm
using an objective analysis program which smoothed
the dynamic heights over a space scale of 60 km and
a time scale of 30 days. The ability of the contours
to resolve the spatial structure of the field depends
on the density of the station data. Near the map
centers about 80% of the total dynamic height vari-
ance has been resolved. On the map periphery the
resolution is much poorer, only 20 to 30% of the
variance being resolved.
The sequence in Figs. 3 and 4 shows a slowly
moving, close-packed array of cyclonic and anticyclonic
surface pressure disturbances having a spatial perio-
dicity of about 400 km (eddy diameter of 200 km)
and an amplitude of about 0.1 dyn m. The field is
clearly irregular and unsteady. Features on scales of
several hundred kilometers tend to persist throughout
the mapping sequence (4 months) while on the smallest
scale (60 km) they cannot be traced from one map
to another (15 days). In March the center of the
MODE area appears to be in a saddle between two
high pressure cells (anticyclonic eddies) to the east
and west and two low pressure cells (cyclonic eddies)
to the north and south. The eastern anticyclonic eddy
moves into the center during the first half of April
and then enlarges and dominates the MODE area
until the end of June. During this period it slowly
drifts westward, moving out of the area in the first
half of July at the end of MODE.
240
956
JOURNAL OF PHYSICAL OCEANOGRAPHY
Volume 6
9 MARCH -30 MARCH
31 MARCH - 14 APRIL
re to* 69* e»'
15 APRIL-29 APRIL
rO" 69* 68'
30 APRIL-14 MAY
Fig. 3. Sea surface temperature maps (upper) and surface dynamic topography (lower), relative to
1500 dl), of the MODE area for four successive periods in the first half of the experiment. The cross-
hatched area on the maps of dynamic height show all surface water cooler than the mean for that period.
241
Novemher l')76 A . D . VOORHIS, E . H . SCIIROEDER AND A . LEETMAA
957
r
J j
-
1
27"C
(
,~~~,
2(5.5*
/
-
1 (',"
-'/'"""
. — 27»
\y0 \ '■
275'
Fig. 4. As in Fig. 3 except for the second half of the experiment.
242
958
JOURNAL OF PHYSICAL OCEANOGRAPHY
Volume 6
Fig. 5. (Upper) Ships tracks showing typical spatial coverage
used to construct temperature maps in Figs. 3 and 4. (Lower)
Mean temperature (T) and mean meridional temperature gradient
(Tv) averaged over MODE area as a function of time during the
experiment.
Geostrophic surface currents in the maps vary in
speed from zero at the eddy centers to as high as
30 km day-1 in the high-gradient regions between
the eddies. The latter is undoubtedly limited by
resolution. The average speed around the periphery
of an eddy is about 20 km day-1. If an eddy were
stationary it would take about one month for surface
water to go around it once at this speed.
4. Surface temperature advection
By comparing maps in Figs. 3 and 4 it becomes
very clear that large-scale features in the surface
temperature field are due primarily to eddy surface
currents which advectively distort the mean meridional
gradient.3 To facilitate this comparison we have cross-
hatched on the maps of dynamic topography all those
areas of the sea surface which are cooler than the
mean temperature of the corresponding temperature
map. Note, in particular, the situation during the
period 9-30 March in Fig. 3. Here, the interplay be-
3 Ideally one would like to quantify this statement by numerical
correlation between temperature and eddy fields. The density of
our data set is not high enough, unfortunately, to make such a
correlation statistically significant.
tween four apparent eddies advects southward a 200-
300 km tongue of cool northern water along longitude
70°30'\V, and advects northward a similar tongue of
warm southern water along 69°00'\Y. Similar patterns
occur in the other maps. At times, isolated pools of
warm water (30 April- 14 May) or cool water (14 June-
28 June) are formed as a result of the eddy currents.
Important changes can occur within the mapping
periods of Figs. 3 and 4. It is possible to examine
some of this in more detail on a one-week time scale
by examining selected STD data. These fields are
shown for the weeks of 14 May and 21 May in Fig. 6.
The temperature pattern at a depth of 50 m looks
very similar to the surface pattern for the same time
in Fig. 4. Two tongues of water are evident, a warm
tongue extending to the north, and a developing cold
tongue extending to the south. Even on a weekly
time scale considerable changes occur (in both the
depicted fields). It is interesting to note that the
temperature field is as\"mmetric relative to the dy-
namic topography. The warm tongue sits over the
eddy defined by closed contours of dynamic height,
suggesting a recirculation of the warm waters within
the tongue. The cold tongue, however, is situated
over d\"namic height contours that do not close.
In general, surface isotherms do not coincide on the
large scale with contours of dynamic height. The
pattern of the first is intrusive or finger-like while
the second is circular or eddy-like. This is in direct
contrast with the deep horizontal temperature struc-
ture in the main thermocline where the patterns were
remarkably similar, with cyclonic eddies having cool
centers and anticyclonic eddies warm centers.
The anomalous surface temperature structure in our
maps extends at least over a decade of decreasing
scales, from 400 km down to the order of 40 km.
The former corresponds to the average wavelength
between eddies and the latter is typical of the width
of the long intrusions of warm and cool water. (Fea-
tures on a smaller scale, although occasionally resolved,
were likely to be advected and distorted beyond
recognition in the mapping intervals.) Over this range
it is reasonable to suppose that temperature variance
is extracted from the mean meridional temperature
gradient at the large scale and cascades to the small
scales. The Lagrangian time for this cascade would
depend on how quickly the surface currents can
distort the temperature field. From the spatial struc-
ture on the maps of dynamic height one estimates
that large temperature features are stretched and
thinned by a factor of 2 in a period of 5 days to a
week. Hence, one estimates that the long intrusions
are formed in 15 to 30 days, that is, in the time re-
quired for surface water to move one-half to once
around the periphery of a typical eddy.
The above cascade extends to much smaller scales
than those resolved in our maps. The most obvious
243
November 1976 A. D. VOORHIS, E. H. SCHROEDER AND A. LEETMAA
959
67" 72'W
MAY 21-27
Fig. 6. Temperature maps at 50 m (left) and dynamic topographs' (right), relative to
1500 db, of the MODE area for two successive weekly periods.
of such features observed during MODE were those
associated with surface frontogenesis. Voorhis and
Hersey (1964), Voorhis (1969) and Katz (1969) had
shown prior to MODE that these surface fronts are
very narrow transition zones separating adjacent sur-
face water masses of different temperatures, salinities
and densities which can meander along the sea surface
for distances of several hundred kilometers. Surface
temperature changes of 1-2°C are frequently observed
across a front in a distance of only 100 to 200 m.
Beneath the surface a frontal pycnocline slopes down-
ward beneath the lighter water and becomes level in
a horizontal distance of the order of 10 km and at
depths usually of the order of 100 to 200 m, although
occasionally it is much deeper. Associated with the
sloping pycnocline is a geostrophic current jet flowing
along the front with surface speeds as large as 50
to 100 cm s_1.
The continuous shipboard records of surface tem-
perature from MODE showed numerous frontal
crossings along the boundaries of the long intrusive
warm and cool tongues in Figs. 3 and 4, and fronts
were frequently seen in these areas by the high con-
centrations of surface debris. The nature of the pro-
gram and the haphazard sampling, however, made it
impossible to map particular frontal features. It is
significant, however, that a s\Tioptic image of the
surface temperature field from a satellite (Fig. 2)
usually shows the boundaries between the intrusive
tongues to be much sharper (greater thermal gra-
dients) than in our 7-15 day maps. We suggest
that frontogenesis is common along these boundaries
and that frontal currents may contribute an important
near surface circulation around the boundaries of the
long intrusive features which is superimposed on
broader scale eddy surface currents along their axis.
5. Surface wind drift
We have so far neglected the advection and dis-
tortion of surface temperature structure by surface
currents other than those due to mesoscale eddies or
possible near surface geostrophic currents associated
with frontogenesis. The most important of the former
are the shallow surface currents driven by wind stress.
During MODE all ships routinely reported wind
speed and direction once daily. The weather from
March through mid-May was dominated by a suc-
cession of moderate high and low pressure disturbance
every 5 to 10 days with mainly veering winds which
varied in speed from less than 1 m s-1 to no more
than 15 m s~
Conditions were somewhat steadier
from mid-May onward with disturbance every 10 to
15 days. The wind backed and veered with maximum
speeds less than 10 m s-1.
Surface drift current was computed using the model
of Gonella (1971), which assumes an Ekman current
244
960
JOURNAL OF PHYSICAL OCEANOGRAPHY
Volume 6
Table 1. Mean surface stress magnitude (t) and direction (<j>)
computed from observed wind speed and direction, assuming a
drag coefficient of !.2X10~:'; mean mixed layer depth (//) from
CTI) observations; and menn surface wind drift speed (I) and
direction (0) computed from Gonella (1971), assuming an eddy
viscositv of 102 cm2 s""1.
T
4>
//
r
e
Period
(d;
in cm"2)*
, (oT)
(m) (ki
ii day-1
) (°T)
9 Mar-30 Mar
0.16
68
38
1.7
115
31 Mar-14 Apr
0.30
37
34
3.1
81
15 Apr-29 Apr
0.82
260
48
8.5
305
30 Apr-14 Mav
0.40
260
28
4.1
305
15 Mav-29 Mav
0.26
353
11
i.2
68
30 Mav-13 June
0.50
261
13
6.1
335
14 Jun-28 Jun
0.21
329
12
3.1
36
29 Jun-13 Jul
0-20
i2i
11
3.0
ii
Mode mean
0.20
293
25
2.2
356
* 1 dvn cm-2 =
0.1 N m"2.
completely confined to the surface mixed layer (zero
stress at the bottom of the layer). The surface stress
was determined from the usual relation GpA | Vir j W,
where Vw is the reported wind velocity, pA is the
standard air density (1.2X10~3 g cm-3), and C is a
drag coefficient taken to be 1.2X10-3. From Gonella
(1971) we assumed a constant and conservative value
of 102 cm'2 s_1 for the eddy coefficient of viscosity in
the layer. A larger viscosity will reduce the currents,
but not less than about 25% of the values computed.
A smaller value will increase the current (by a factor
roughly proportional to the inverse square root of the
viscosity).
In Table 1 we have listed the spatially averaged
net vector wind stress and surface drift current for
each of the mapping intervals in Figs. 3 and 4. Also
shown is the average depth of the mixed layer, which
clearly shows the effect of decreasing wind stress and
increasing surface heating over the duration of MODE.
The computed currents are all less than 10 km day-1
and except for the two periods 15 April-29 April
and 30 May-13 June, when there were periods of
persistent wind direction, less than 5 km day-1. It is
apparent, therefore, from the maps of dynamic height
(Figs. 3 and 4) that advection by the eddy surface
currents (of order 20 km day-') around the periphery
of the eddies dominates the surface wind drift. The
latter, however, is significant over large areas where
there is little relief in the dynamic topography. The
average surface displacement due to the wind drift
was approximately 60 km per mapping interval.
Also shown in Table 1 is the net vector wind stress
and surface current over the entire 127 days of MODE.
The stress is quite comparable in magnitude and direc-
tion to the typical long-term stress in the MODE
area computed by Saunders (1976) for the years
1959-1971. The overall surface drift is predominantly
northward and the total northward volume transport,
computed from the surface stress, is 2.6X102 m3 s_1
across each kilometer in an east-west direction.
The spatial variation of surface wind drift con-
tributes, of course, to the distortion of surface tem-
perature structure. Superficially it appeared small
because the reported daily wind speed and direction
were remarkably similar from ship to ship. Never-
theless, there were differences. By comparing wind
data from one ship to another we found over an
average horizontal scale of 100 km that there were
variations of net surface drift speed of about l the
spatially averaged value in Table 1, and variations
of drift direction of about ±20°. The distortion of
surface temperature features by such variations is an
order of magnitude less important than that due to
the spatial variations of the eddy surface currents.
We tentatively conclude, therefore, that the dominant
effect of the surface wind drift is to shift but not
distort the temperature pattern shown in Figs. 3 and 4
according to the drift currents in Table 1.
6. Conclusions and discussion
In the MODI-: area we conclude that the surface
temperature field, on time scales less than about one
month and over space scales from 400 to about 40 km,
tags primarily the surface currents associated with
the baroclinic mesoscale eddy field of the main ther-
mocline. Surface currents induced by wind stress
appear to be of secondary importance in generating
spatial structure in this scale range. Relatively little
can be said about scales less than 40 km. However,
there is some evidence from MODE but mostly from
previous measurements in the same area that much
of the spatial structure on scales less than about
10 km 4 tags not only the mesoscale current field
but also a relatively shallow field of currents which
are in geostrophic equilibrium with horizontal density
gradients in the near surface layers. This new field
of currents is often jet-like and is associated with
surface frontogenesis.
Mesoscale eddies appear to be an effective mecha-
nism for stirring the large-scale thermal (and haline)
field imposed on the near surface layers by the atmo-
sphere. One can speculate that this surface process
on an eddy time scale may generate a net meridional
heat transport in the surface layer on a longer time
scale. For example, if a single anticyclonic eddy
develops in the convergence zone one would expect
warm water to move initially northward on its western
side and cool water southward on its eastern side.
(The flows will change sides if the eddy is cyclonic.)
In time both flows will simply circulate in a complex
manner around the eddy with a great deal of stirring
but no net heat transport if there is no heat exchange
between the warm and cool water. However, if there
4 It is significant that this scale is of the order of the internal
radius of deformation of the near-surface pycnocline.
245
November 1976 A. D. VOORH1S
H. SCHROEDER AND A. L E E T M A A
961
are many eddies, which are evolving, moving and
decaying, it is highly likely that surface water is ex-
changed5 from eddy to eddy and one might expect
to observe at times long tongues of warm and cool
surface water running north and south. This is very
similar to what one sees in Fig. 2. The result would
be a mean meridional heat transport northward in
the MODE area of the order of Nil per eddy, where
V is the geostrophic advecting surface velocity, and
// is the anomalous heat carried by each tongue.
The latter can be approximated by pwCT,DL AT, where
AT is the temperature difference between north and
south flowing tongues, L is the zonal width of the
tongue, and D is the depth of the heat anomaly.
Representative values for these parameters are
F=20 cm s-1, pw=l g cur3, C'p=4.18 J g"1 K~\
D=50 m, L=100 km, Ar=2°C. Using these values
one computes a transport of 8.2 X1012 W per eddy.
Taking 200 km as a mean zonal spacing between
eddies one finds a northward eddy heat transport of
4.2X1010 W across each kilometer in an east-west
direction. Assuming the northern Sargasso Sea to be
bounded on the north and west by the Gulf Stream,
on the east by 50°W longitude, and on the south by
30°N latitude, one computes an annual heat input
of 32X1020 J across its southern boundary (length
2400 km) by the eddy mechanism. This is of the
same order as the annual heat loss to the atmosphere
across its surface area (2.2X106 km2) computed from
Bunker and Worthington (1976), using an average
net heat surface flux of 66 W m-2 (50 kcal cm-2 year-1).
Speculating on a still larger scale and assuming that
the observed mesoscale eddy activity extends across
both the northern Atlantic and Pacific Oceans at
mid-latitudes, a total distance of the order of 1.6
X 104 km. one finds an annual poleward heat transport
by the eddies at these latitudes of the order of 6.8
X 1014 W. This can be compared with the annual
oceanic poleward energy transport of about 22.6
X1014 W (1.7X1022 cal year-1) estimated by Vonder
Haar and Oort (1973). Considering the uncertainties
in all of these estimates one concludes that the meso-
scale eddy heat transport may not be inconsequential.
Finally, our results can have important implications
for oceanographers and meteorologists interested in
annual or longer term changes in sea surface tem-
perature and their effect on world climate. The fluc-
tuating mesoscale temperature field is unwanted noise
from their point of view and introduces an uncertainty
to estimates of mean temperatures. For data collected
from a fixed point (or within an eddy radius of this
point) this uncertainty is of the order of (ATe)/yjn,
where (ATe) is the rms temperature change due to
a typical eddy, and n is the number of eddy events
5 This may be greatly enhanced by the unusually strong surface
currents associated with surface frontogenesis.
in the averaging time. Assuming (ATe)«0.5°C and
no other sources of noise, one would have to average
over 25 eddy events in the MODE area in order to
resolve a climatic 0.1 °C change in mean surface tem-
perature. If the eddy residence time is of the order
of 2 months this would take 4 to 5 years.
Acknowledgments. This work was supported by the
Office of Naval Research under Contract N00014-74-
C-0262, XR 083-004 and by the Office of the Inter-
national Decade for Ocean Exploration of the National
Science Foundation under Funding Agreement AO-385.
The data used in this paper were collected and
processed by many people in the MODE program
and the authors wish to acknowledge all of this work
and to express their gratitude. We would also like to
thank N. Fofonoff of the Woods Hole Oceanographic
Institution, Woods Hole, Mass., who programmed
and computed the objective maps of dynamic height
in Figs. 3 and 4.
REFERENCES
Bryden, H. L., 1974: Geostrophic comparisons using moored
measurements of current and temperature. Nature, 251,
409-410.
Bunker, A. I'., and I,. V. Worthington, 1976: Energy exchange
charts of the North Atlantic Ocean. Bull. Amer. Meteor. Soc,
57, 670-678.
Crease, J., 1962: Velocity measurements in the deep water of the
western North Atlantic. J. Geopliys. Res., 67, 3173-3176.
Dynamics and the- Analysis of MODE-1, March 1975: Report of
the MODE-1 dynamics group (unpublished manuscript).
[The MODE Executive Office, 54-1417, M.I.T., Cambridge,
Mass. 02139.]
Gonella, }., 1971: The drift current from observations made on
the Bouee-Laboratoire. Call. Oceanogr., 23, 1-15.
Katz, E. J., 1969: further study of a front in the Sargasso Sea.
Tell us, 21, 259-269.
McWilliams, J. C, 1976: Maps from the Mid-Ocean Dynamics
Experiment. I. Geostrophic streamfunction. /. Pliys.
Oceanogr. (accepted for publication).
Robinson, A. R, 1975: The variability of ocean currents. Rev.
Geopliys. Space Pliys., 13, 598-601.
Saunders, P. M., 1976: On the uncertainty of wind stress curl
calculations. J. Mar. Res. (submitted for publication).
Schmitz, W. J., J. R. Luyten, R. E. Payne, R. H. Heinmiller,
G. H. Volkmann, G. 11. Tupper, J. P. Dean and R. G. Walden,
1976: A description of recent exploration of the eddy field in
the western North Atlantic with a discussion of Knorr Cruise
49. W1IOI Tech. Re]), (to be published).
Schroeder, E. H., 1966: Average surface temperatures of the
western North Atlantic. Bull. Mar. Set., 16, 302-323.
Vonder Haar, T. H., and A. H. Oort, 1973: New estimate of
annual poleward energy transport by Northern Hemisphere
oceans. J. Pliys. Oceanogr., 3, 169-172.
Voorhis, A. D., 1969: The horizontal extent and persistence of
thermal fronts in the Sargasso Sea. Deep-Sea Res., 16,
331-337.
and J. B. Hersey, 1964: Oceanic thermal fronts in the
Sargasso Sea. Deep-Sea Res., 69, 3809-3814.
Wiist, G., 1928: Der Ursprung der atlantischen Tiefenwasser.
Z. Ges. Erdk. Berl., Sonderband zur Hundertjahrfeier,
506-534.
246
24
Reprinted from: Marine Geoteoknology , Vol. 1, No. 4, 327-335.
Initial Results and Progress
of the Mississippi Delta
Sediment Pore Water
Pressure Experiment
RICHARD H. BENNETT,* WILLIAM R. BRYANT,t
WAYNE A. DUNLAP,t AND GEORGE H. KELLERtt
Abstract This report describes the instrumentation, initial results,
and progress of an experiment designed to measure and monitor
submarine sediment pore water and hydrostatic pressures in a selected
area of the Mississippi Delta. The experiment also is intended to
monitor significant pressure perturbations during active storm
periods. Initial analysis of the data revealed excess pore water
pressures in the silty clay sediment at selected depths below the
mudline. Continuous monitoring of the pore water and hydrostatic
pressures was expected to reveal important information regarding
sediment pore water pressure variations as a function of the geo-
logical processes active in the Mississippi Delta.
Introduction
The NOAA-Atlantic Oceanographic and Meteorological Laboratories is
presently engaged in a NOAA program directed toward the delineation and
understanding of important processes and mechanisms related to submarine
sediment stability. A unique situation arose to test some of the equipment and
concepts being developed in this program on Project SEASWAB (Shallow
Experiment to Assess Storm Waves Effecting Ztottom), which is part of a larger
study of the Mississippi Delta being conducted by the U.S. Geological Survey.
NOAA-Atlantic Oceanographic and Meteorological Laboratories, Miami, Florida.
'Texas A&M University, College Station, Texas.
ft School of Oceanography, Oregon State University, Corvallis, Oregon.
(Received December 4, 1975; Revised February 13, 1976.)
Marine Geotechnology, Volume 1 , Number 4
Copyright © 1976 Crane, Russak & Company, Inc.
327
247
328
RICHARD H. BENNETT ETAL.
29°30' -
Mississippi Birdfoot Delta
29°20' -
29°10' -
29°00-
28°50' -
89°30' 89°20' 89°10' 89°00'
Figure 1. General study area.
88°50'
The purpose here is to describe the instrumentation, progress, and initial
results of an experiment designed to measure and continuously monitor, over a
period of several months, submarine sediment pore water pressures in a selected
area of the Mississippi Delta (Figure 1). Objectives of the experiment are to
measure not only pore water and hydrostatic pressures at various depths below
the mudline, but also to monitor significant pressure perturbations during active
storm wave periods. It is interesting to note that while engineers have known for
decades that pore water pressures are an important geotechnical consideration,
the first reported attempt to measure pore water pressures in submarine sedi-
ments was made by Lai et al. (1968) and Richards et al. (1975).
The Mississippi Delta is well known for being a very dynamic region charac-
terized by the interaction of riverine and marine processes and the large dis-
charge of bedload and suspended sediment. Large plumes of sediment extend
considerable distances beyond the subaerially exposed delta and deposit vast
248
MISSISSIPPI DELTA SEDIMENT PORE WATER PRESSURE EXPERIMENT 329
quantities of silt and clay in the prodelta environment. This environment is
characterized not only by the rapid deposition of fine-grained sediment having
very high water contents, but also by the accumulation of organic material
(Coleman et al., 1974). Methane and carbon dioxide gases, intimately related to
decomposition of the organic material, influence substantial portions of the
Mississippi Delta submarine sediments (Whelan et al., 1975). Knowledge of the
sediment geotechnical properties in this complex and dynamic environment is of
great importance to engineers faced with the design and construction of offshore
structures, and to geologists investigating sedimentological processes relating to
submarine diagenesis, environments of deposition, mass movement, and sedi-
ment stability (Morelock and Bryant, 1966; Keller and Bennett, 1968; Bennett
and Bryant, 1973). Not only will the measurement of pore water pressures in the
Mississippi Delta sediment aid in understanding and interpreting the sediment
geotechnical properties, but it also should provide an insight into the behavior of
these sediments in response to dynamic and static loads.
Instrumentation
The NOAA sediment pore water pressure probe (piezometer) system consists
of the following components:
1 . Probe and Sensing Units (Transducers)
2. Signal Conducting Cables
3. Signal Conditioner Units
4. Voltage and Frequency Regulator Units
5. Recording Unit.
The probe enclosing the pressure sensing devices is a 0.10-m O.D. steel pipe
having a total length of 17.12 m. A weight stand mounts to the top of the probe
on four steel gusset plates. The weight stand is fastened to the probe with two
steel pins and the probe assembly is lowered into the seafloor by a steel cable
fastened to the top of the weight stand. Four sensing units (variable reluctance
pressure transducers) were placed in the probe at selected intervals; two sensors
measured pore water pressure and two measured hydrostatic pressure (Figure 2).
The pore water was transmitted through 0.05-m diameter porous corundum
stones, which were inset in the pipe and ground to the pipe radius. Hydrostatic
pressure was transmitted from the mudline through the seawater-filled steel pipe
to the sensors installed inside the probe. Sensors were mounted inside oil-filled
capsules in the probe and connected to the appropriate pressure ports with short
tubing. Separate measurement of the pore water and hydrostatic pressures was
necessary since one objective of the experiment was to determine the effect of
249
330 RICHARD H. BENNETT ETAL.
bottom pressures from storm waves on the pore water pressures. This approach,
however, is feasible only in shallow water. A sensing unit which is robust enough
for use in deeper water will generally not have sufficient resolution for the
purposes desired. For measurement of static pressures, this limitation is largely
overcome by a differential piezometer of the type described by Hirst and
Richards (1976).
The signals from each transducer are transmitted through a conducting cable
to signal conditioners and filtering systems before being recorded on a strip chart
recorder. Electronic units and pressure transducers were tested and calibrated
prior to the assembly of the probe. Calibration was' carried out using a fused
quartz Bourdon tube pressure gage having a sensitivity of 1 part in 200,000. The
transudcers have a maximum working range of 689.5 kPa (1 psi =6.89 kPa) and a
reproducibility of ± 3.5 kPa. The electronic units were checked frequently
during various phases of the probe assembly.
Installation of the Probe
The probe and electronics system we.e assembled in the field aboard the
Texas A&M University Ship R/V 1yr°. during 18-19 September, 1975. The total
weight of the probe and weight stanrl loaded with four train wheels was 1.115
Mg in sea water; this weight having been calculated as adequate to implant the
probe to the desired depth of penetration. On the afternoon of 19 September,
1975, the probe was lowered from the Gyre in the Mississippi Delta sediment
(Block 28, Soutn Pass Area, slightly south of 29°00'N, 89°15'W) at a preselected
site 145 m from an offsnore production platform where the recorder and signal
conditioner units were installed later. The water depth was approximately 19 m
at the site. After installation, divers removed the steel pins and made a general
inspection of the exposed portion of the instrument. The weight stand and
weights were returned to the ship.
Pore water and hydrostatic pressures were monitored from the ship during
installation and for 40 min afterward. In the ensuing 4 h period, no readings
were made while the electronic units were transferred to the platform. During
this time divers installed the signal conducting cables along the sea floor to the
platform. After reconnection, all systems appeared to be functioning properly.
The probe was implanted only a few days before the passage of Hurricane Eloise
near the site.
Discussion
Sediment pore water pressures, uw, were measured at depths of approxi-
mately 8 and 15 m below the mudline. Hydrostatic pressures, us, were measured
simultaneously at depths of approximately (actual mudline difficult to de-
250
MISSISSIPPI DELTA SEDIMENT PORE WATER PRESSURE EXPERIMENT
331
-A
Steel Cable
1.98m
m
^
'T7
Pressure
Transducer #1
(Hydrostatic
Pressure)
1.41m
6.95m
Pressure
Transducer #2.
(Pore Water
Pressure )
Po r o u s
St o ne
6.88m
Pressure Transducer #4
(Pore Water Pressure)
Steel Pins
Weights (Train Wheels)
0.91m Dia.
Welded Steel Plates
Cable Clamps
Multiconductor
Armored Cables
0.10m O.D. Steel Pipe
6.40m Sections
Fastened With
Inner Couplings
Pressure Transducer #3
1 Or — ■"" (Hydrostatic Pressure)
7
Figure 2. Sediment pore water pressure probe and weight stand. Drawing not to
scale.
termine) 1 and 15 m below the mudline (Figure 2). A value of 68.9 kPa,
equivalent to 7 m of sea water, has been added to Pressure Transducer 1 data for
direct comparison with the sediment pore water pressures recorded by Trans-
ducer 2. Comparison of the hydrostatic pressure and sediment pore water
pressure for a given depth below the mudline may reveal one of three possible
conditions:
Condition 1. Sediment pore water pressure equals the hydrostatic pressure
(uw ~ us).
Condition 2. Sediment pore water pressure exceeds the hydrostatic pressure
("w > "*)•
Condition 3. Hydrostatic pressure exceeds the sediment pore water pressure
(uw<us).
251
332 RICHARD H. BENNETT ETAL.
Condition 1 is common for normally consolidated sediments assuming there
has been no movement or shearing of the sediment. The pore water pressure is in
equilibrium with the hydrostatic pressure. Condition 3 may occur with some
overconsolidated sediments and possibly with other sediments while responding
to complex dynamic conditions. This condition is not directly related to the
results presented here and therefore will not be treated. When the sediment pore
water pressure exceeds the hydrostatic pressure (Condition 2) the difference is
termed excess pore water pressure, ue. Thus, ue = uw - us. This condition is
associated with underconsolidated sediments wherein very low sediment
permeabilities hinder dissipation of the pore water pressure which builds up
under rapid rates of deposition and loading (Bryant et al., 1975). Sediment
movement and shearing may also contribute to the presence of excess pore water
pressures, as may the undissolved gases present in some muds. Since the possible
existence of excess pore pressures was the major reason for conducting the
Mississippi Delta pore pressure experiment, the results reported here are ex-
pressed in terms of excess pore pressures, ue.
The data reveal relatively high excess pore water pressures of 99.3 kPa at 15
m below the mudline and 49.6 kPa at a depth of 8 m below the mudline
immediately following probe insertion (Figure 3). High excess pore water pres-
sures were expected to occur due to implanting the probe and this condition also
was observed by Richards et al. (1975). It was expected that these pore pressures
would dissipate to the static condition following a typical log time consolidation
relationship, and, in fact, this appeared to be the trend at the 8 m depth.
However, little value can be given to these early readings owing to stabilization
of the electronic system including temperature equilibration of the pressure
transducers. Six hours after inserting the probe, excess pore water pressures were
still relatively high at 81.4 kPa (15 m depth) and 37.2 kPa (8 m depth). They
appeared to become relatively constant after approximately 7 h at the 8 m
depth and 10-12 h at the 15 m depth, although at the latter depth, the excess
pore pressures began to decline again just prior to the initial effects of Hurricane
Eloise. Excess pore water pressures averaged approximately 72 kPa (15 m depth)
and 32 kPa (8 m depth) after 7 h of initial stabilization of the system and prior
to the storm. Clearly, significant sediment excess pore pressures were observed
for a considerable period of time prior to the initial effects of the storm activity
that began 21 September, 1975.
The records gathered during the passage of Hurricane Eloise indicate that the
system was functioning satisfactorily. Pore water pressures appear to have varied
significantly in response to the storm wave activity; however, these data require
considerably more analysis for a complete assessment of the pore water condi-
252
MISSISSIPPI DELTA SEDIMENT PORE WATER PRESSURE EXPERIMENT
333
100.04-
80
60 -
40
20 -
0
-20
Probe Entered Mud
Al 1326, Sept. 19,1975
\
Approximate
Electronic -
Stabilization
Period
No Readings
During Cable
■"— Installation —
15m Depth
8 m Depth
Initial Time
of
Approaching Storm
1.0
10.0 100.0
Elapsed Time In Minutes
1000.0
10,000
Figure 3. Sediment excess pore water pressures at approximate depths of 8 and
15 m below the mudline off the Mississippi Delta.
tions relating to the influence of the hurricane. Further data analysis for the
period of time from 22—23 September, during which Hurricane Eloise passed in
close proximity to the probe, is expected to reveal important information
regarding pore water pressure fluctuations, particularly observed increases, and
the degree of pore water pressure dissipation as recorded by the sensors im-
planted at various depths below the mudline. Poststorm data also are being
collected in order to assess any possible ldfig term changes in the pore water
pressures of these prodelta muds.
A soil boring was completed in the immediate area approximately six weeks
after the probe was installed. The sediment contained large amounts of gas as
evidenced by the appearance of the recovered cores. It is not known whether
undissolved gas was present in the area of the sensors. If this were the case, then
the sensors may have responded to possible higher pore gas pressure rather than
pore water pressures. Based on the success of this experiment, it is anticipated
that future pore pressure probes will include high air entry ceramic stones to
separate the water from the gas pressure.
253
334 RICHARD H. BENNETT ETAL.
Summary
Significant excess pore water pressures have been observed in the submarine
sediment of South Pass, Block 28, Mississippi Delta. Excess pressures averaging
72 and 32 kPa at depths of 15 and 8 m below the mudline respectively, appear
to be characteristic of the general conditions prior to the passage of Hurricane
Eloise. Further data analysis and study may result in a slight refinement of these
initial observations and may also reveal important information regarding sedi-
ment pore water pressure variations during and after the passage of the storm.
Initial evaluation of the data collected during Hurricane Eloise indicates that
sediment pore water pressures varied significantly in response to the storm
activity.
Acknowledgments
The writers wish to express their appreciation for support given by NOAA
Atlantic Oceanographic and Meteorological Laboratories (AOML). Considerable
support for the Delta Project was given by the U.S.G.S. Marine Geology Branch,
Corpus Christi, Texas, and the Conservation Division, New Orleans, Louisiana.
From NOAA/AOML we thank John Burns for his assistance in assembling the
electronics system, Frances Nastav for the drafting of the figures, and Thomas
Clarke for the computer processing of the data. Calibration of the pressure
transducers was made at the Naval Coastal Systems Laboratory, Panama City,
Florida, through the consideration and help of G. W. Noble. The writers also
appreciate the cooperation and assistance given by various members of the Shell
Oil Company staff. Critical review of this paper was made by Drs. T. Hirst, A.
Richards, B. McGregor, H. B. Stewart, Jr., L. Garrison, and Mr. R. Bea.
References
Bennett, R. H., and W. R. Bryant, 1973. Submarine sediment microstructure (abstracted).
Clay and Clay Minerals Society 22nd Annual Clay Minerals Conference. Banff, Canada,
Program Abstracts, p. 22.
Bryant, W. R., W. Hottman, and P. Trabant, 1975. Permeability of unconsolidated and
consolidated marine sediments, Gulf of Mexico. Marine Geotechnology, vol. 1, pp. 1—14.
Coleman, J. M., J. N. Suhayda, T. Whelan, and L. D. Wright, 1974. Mass movement of
Mississippi Delta sediments. Transactions of the Gulf Coast Association of Geological
Societies, vol. 24, pp. 49-68.
Hirst, T. J., and A. F. Richards, 1976. Excess pore pressure in Mississippi Delta front
sediments: initial report. Marine Geotechnology, vol. 1, pp. 337—344.
Keller, G. H., and R. H. Bennett, 1968. Mass physical properties of submarine sediments in
the Atlantic and Pacific Basins. Proceedings of the 23rd International Geological Con-
gress, Prague, vol. 8, pp. 33—50.
Lai, J. Y., A. F. Richards, and G. H. Keller, 1968. In place measurement of excess pore
water pressure of Gulf of Maine clays (abstract). Transactions of the American Geo-
physical Union, vol. 49, p. 221.
Morelock, J., and W. R. Bryant, 1966. Physical properties and stability of continental slope
254
MISSISSIPPI DELTA SEDIMENT PORE WATER PRESSURE EXPERIMENT 335
deposits, northwest Gulf of Mexico. Transactions of the Gulf Coast Association of
Geological Societies, vol. 16, pp. 279-295.
Richards, A. F., K. 0ien, G. H. Keller, and J. Y. Lai, 1975. Differential piezometer probe
for an in situ measurement of sea-floor pore pressure. Ge'otechnique, vol. 25, pp.
229-238.
Shepard, F. P., 1956. Marginal sediments of the Mississippi Delta. American Association
Petroleum Geologists Bulletin, vol. 40, pp. 2537-2623.
Whelan, T., J. M. Coleman, and J. N. Suhayda, 1975. The geochemistry of recent Mississippi
River Delta sediment: gas concentration and sediment stability. Proceedings of the
Offshore Technology Conference, vol. 3, pp. 71-84.
255
25
Reprinted from:
VOL. 81,
NO.
Journal of Geophysical Research, Vol. 81, No. 29, 5249-5259,
JOURNAL OF GEOPHYSICAL RESEARCH OCTOBER in, 1976
GEOPHYSICAL INVESTIGATION OF THE CAPE VERDE ARCHIPELAGO
B. P. Dash,1 M. M. Ball,* G. A. King ,'
L. W. Butler,5 and P. A. Rona-
Abstract ■ The Cape Verde Islands are emerged
portions of a Mesozoic-Cenozoic volcanic accretion
in the form of a westward-opening horseshoe along
fracture zones converging from the mid-Atlantic
ridge toward Africa. An interior abyssal plain
slopes westward, increasing in depth from 2.7 to
4.5 km. The plain is underlain by low relief on
acoustic basement that is associated with a
300-gamma negative magnetic anomaly. The flanks
of the Sal-Maio ridge appear bounded by large-
displacement normal faults; superficial slumping
is common. The trends of magnetic anomalies are
linear N-S north of the islands and less linear
within the islands and may change coincident with
E-W bathymetric trends south of the islands. A
triangular pattern of reversed refraction lines
200-250 km long along the north and east ridges
and NW-SE across the interior abyssal plain
indicated 2-3 km of semiconsolidated sediments
underlain by 3-6 km of basalt and 6-8 km of
plutonic rocks. The depth of the Meho is between
16 and 17 km. A deep NW-SE trending fault inter-
sects the Sal-Maio ridge near Boa Vista. The
consistent depth to Moho and the regional Bouguer
anomaly indicate lack of local relief at the base
of the crust. The crustal load of the entire
archipelago is regionally adjusted.
Introduction
The Cape Verde archipelago is situated on the
continental rise about 500 km west of the north-
west African continental shelf and 2000 km east
of the mid-Atlantic ridge. The archipelago is
encompassed by the 37C0-m isobath which outlines
the Cape Verde plateau extending seaward from
northwest Africa [Ror.a, 1971; Eglof f , 19 72 J .
The islands are aligned along three bathymetric
ridges which form a horseshoe opening westward
(Figure 1) . The segment of the mid-Atlantic
ridge facing the Cape Verde archipelago ex-
hibits a progressive change in trend of the
ridge axis from NE-SW to N-S accompanied by
an abrupt change in the displacement of the
Department of Geophysics, Imperial College of
Science and Technology, London SW 7 2BP, England
Rosenstiel School of Marine and Atmospheric
Science, University of Miami, Miami, Florida
33149
-'National Oceanic and Atmospheric Administra-
tion, Atlantic Oceanographic and Meterological
Laboratories, Miami, Florida 33149
Copyright 1976 by the American Geophysical Union.
fracture zones from right to left lateral at
the Kane fracture zone [Heezen and Tharp , 1968].
It is problematic how the structure of the
archipelago relates on the one hand to that
of the continent and on the other to that
of the ocean basin. Rock compositions on
the islands exhibit both continental and
oceanic affinities.
A geophysical investigation of the islands,
including the adjacent sea floor, was under-
taken to complement prior geological investi-
gations in order to determine the structure
of the Cape Verde archipelago. The Geophysics
Department of Imperial College performed
geophysical studies around the islands. The
Rosenstiel School of Marine and Atmospheric
Science of the University of Miami and the
National Oceanic and Atmospheric Administration's
(NOAA) Atlantic Oceanographic and Meteorological
Laboratories, Miami, collaborated in the pro-
ject, with the university research vessel
John Elliott Pillsbury and the NOAA Ship
Discoverer as part of the NOAA Trans-Atlantic
Geotraverse (TAG) project.
Lithology and Structure of the Islands
The oldest known rocks in the Cape Verde
Islands are Lower Cretaceous (possibly Upper
Jurassic) limestones exposed on Maio. An
aptychus of Lamellaptychus angulocos tat us
atlanticus occurring in the Lower Cretaceous
limestones on Maio suggests an open marine
origin for the limestone. Outcrops of
questionable Mesozoic age are present on
Sao Nicolau, Sal, and Boa Vista. Paleogene
sediments and lavas are known on Maio, and
Neogene rocks are present on most of the
islands. Eruptions have occurred on Fogo
as recently as 1951 [Machado, 1965].
Part [1950] cites early observations of
structural trends in the island made by
J.B. Bebiano of the Portuguese Geological
Survey who noted (1) the marked linear
distribution of Santo Antao, Sao Vicente,
Sao Nicolau, and Boa Vista, marking a
possible WNW-ESE fault trend; (2) NE-SW
linear trends on the west side of Santo
Antao within the island itself and between
Santo Antao and neighboring Sao Vincente;
(3) generally NW-SE structures running
through Sao Nicolau, Sao Tiago, and Fogo;
and (4) the N-S Sal-Maio ridge. Ballard
and Hemler [1969] have reported an eastward-
facing fault scarp on the east side of the
Sal-Maio ridge. Of key importance in a
structural analysis of the island group are
5249
256
5250
Dash et al. : Geophysical Investigation of the Cape Verde Archipelago
I5°N
14° N
26°W
25°W
24°W
23° W
22°W
Fig. 1. The Cape Verde archipelago and its location with respect to the African
continent. A, B, and C are the refraction profiles showing the recording
stations on the islands of Sao Vicente, Sao Tiago, and Sal. The Sal-Maio
fracture zone is indicated. Contours are at 100 fm (182.88m) intervals.
the recently recognized trends on the
island of Maio [Serralheiro, 1970].
It appears that this Cretaceous sequence
on Maio represents a section of marine limestones
and shales invaded by highly undersaturated
alkaline intrusive and extrusive rocks along
a NNW-SSE zone of weakness in the North Atlantic
sea floor. Carbonatites are present on
Maio, Fogo , Brava, Sao Vicente, and possibly
Sal [Assuncao et al. , 1968]. On Brava the
syeni te-carbonati te series forms a ring
complex intruded into a series of palagonitic
basaltic pillow lavas of submarine origin and
probable Cretaceous age [Machado et al ■ , 1967].
Essexite-syeni te-carbonati te associations
commonly occur as a compound central plug
enclosed in a ring complex, a structural
pattern suggesting emplacement of the carbonate
as well as the alkaline basic rocks by
intrusion. This is the prevailing pattern in
some of the east African rift valleys [Turner and
Verhoogen , I960]. Such intrusions are thought
to be characteristic of a prerifting up-doming
phase in continental areas [LeBas , 1971]
Area of Investigation and Data Collection
The first phase of the geophysical investi-
gation carried out in 1969 consisted of refraction
seismic work supplemented by magnetic profiling.
Three seismic stations were established on the
islands of Sal, Sao Tiago, and Sao Vicente,
forming a triangle with sides of over 200 km
(Figure 1) . Shots ranging from 50 to 350 lb
(23-159 kg) were fired on each line while all
three stations were recording. This system
meant that for every shot fired, two stations
were in line, whereas the third was 'broadside1.
All the data collected were recorded on magnetic
tape for subsequent digital processing.
Refraction Seismic Data Processing and Integration
Digital processing. The seismic data were
digitized at a sampling interval of 10 is, giving
a maximum recoverable frequency for the digital
data of 50Hz. The first stage of processing
was to increase the signal to noise ratio. A
number of methods were tried, including
predictive deconvolution and band-pass frequency
filtering. The latter was found to be the most
effective as well as the most economical in
terms of computer storage and operating
time. Having obtained the filtered records,
we applied stacking and correlation techniques.
These methods were used in conjunction with
routine operations such as correction of
the data to a datum plane which allows
for differences in water depths at shot
points. These methods were applied to
synthetic data as well as to the field data
for purposes of comparison.
The best method for use on the field data
was the stacking method in which several traces
are added, or stacked, after being given time
shifts defined as the ratio of the offset
distance to the apparent velocity being
examined (T = X/V) . The correlation technique
was less successful, since it is heavily
dependent on wave shape and is extremely suscep-
257
Dash et al.: Geophysical Investigation of the Cape Verde Archipelago
5251
25 n
r25
(.0 60 80 100
TRAVEL TIME OF DIRECT SOUND
Fig. 2.
Time-distance plot of profile A.
the time-distance graph.
tible to noise.
Synthetic seismograms. Synthetic seismo-
grams were produced for various crustal models
inferred from the seismic and gravity data
by an iterative program based on ray theory
[Dash et al. , 1970]. This approach assumes
that the geological units constituting the model
are isotropic, elastic, and homogeneous media
with no attenuation of seismic energy and are
two-dimensional with interfaces between the
rock units normal to the plane of section.
Results
All the seismograms obtained were subjected
to the processing techniques described above, and
the results obtained for each profile are indicated
below.
Line A. This line was shot between Sao Tiago
on the south and Sao Vicente on the north with Sal
as the broad side listening point (Figure 1). The
lateral separation between Sao Tiago and Sao
Vicente is about 200 km, and as can be seen from
the time-distance plot (Figure 2) , the two stations
provide a reversed profile. The mean water depth
along this line was about 3000 m. The observed
times were corrected for shot and detector
locations. The thickness and velocity of the
various layers are shown in Table 1.
Line B. This line was shot between the islands
of Sao Tiago and Sal. The station at Sao Vicente
acted as the broadside listening point. The
time-distance plot is shown in Figure 3. As
Figure 3 shows, this line presented an interesting
problem. The station at Sao Tiago registered the
seismic arrivals from above the Moho up to a
distance of about 100 km. However, the station
at Sal failed to register any arrivals at all
for shots fired beyond 30 km. Within 30 km the
arrivals were strong and readily recognizable.
The most plausible explanation for this loss of
energy is that the shots were being fired beyond
a heavily faulted or fractured zone where seismic
energy was dissipated. Once past this zone there
was no interference , and normal arrivals were
registered at Sal.
bathymetry along the profile is shown below
Table 2 shows the velocities and thickness
of various layers calculated from the data
obtained from Sao Tiago. The low apparent
velocity of 7.9 km/s would suggest a slight
rise of the Moho toward Sal.
Line C. Line C was shot betwwen the islands
of Sal and Sao Vincente with Sao Tiago acting
as the broadside recording station. The line
was about 200 km long. The time-distance graph
is shown in Figure 4. A situation similar to
that along line B was encountered here. At
Sal, arrivals were registered up to about
30 km west of the island. From then on for
a further 30 km, no arrivals were registered
at either Sal or Sao Vicente. From a point
approximately 65 km away from Sal, Sao Vicente
began to receive distinct arrivals from the Moho,
while Sal recorded nothing. This anomalous
behavior indicates that seismic energy was
being dissipated by a near-surface fractured or
faulted zone 39 km west of Sal. Once past
this region the energy propagated in the usual
TABLE 1. P Wave Velocities and Thicknesses of
Various Refractors Along Refraction
Profile A
Depth to
Velocity
Thickness
Top of Layer
Horizon
km/s
km
km
Sao Vicente
Water
1.5
3.4
0.0
1
3.2
1.9
3.4
2
4.8
3.6
5.3
3
6.4
7.2
8.9
4
8.1
16.1
S
ao Tiago
Water
1.5
3.4
0.0
1
3.1
2.1
3.4
2
4.8
2.9
5.5
3
6.3
8.0
8.4
4
8.1
16.4
258
Dash et al . : Geophysical Investigation of the Cape Verde Archipelago
PROFILE B
VELOCITIES IN KM/S
20 40 60 80
TRAVEL TIME OF DIRECT SOUND
Fig.
Time-distance plot of profile
the time-distance graph.
bathymetry along the profile is shown below
manner, and arrivals were registered at Sao
Vicente. However, Sal remained in a shadow
zone, and no deep seismic arrivals were
registered on the island. Table 3 shows the
velocities and thicknesses of several layers.
For the purposes of calculation, profile C was
considered a reversed line as far as the first
two layers were concerned. The depths to the
layers of velocities of 6.6 and 8.0 km/s were
calculated from the data obtained on Sao
Vicente .
Broadside at Sal and Sao Vicente. Analysis
of the later arrivals of the broadside data
suggested the existence of a fault zone with a
throw of about 900-1000 m. The large shot point
to detector distance, inadequate charge size, and
associated noise in the seismograms prevented
us from drawing a final conclusion. Nevertheless,
from the results it was possible to postulate the
general trend of the fault system, which is in the
NW-SE direction.
Refraction Summary. Seismic velocities ob-
tained in this survey, can be successfully correl-
ated between profiles and are inferred to repre-
sent the following materials :
1. Surface velocity of 3.1-3.2 km/s consists of
semiconsolidated sediments.
2. Layer with velocities ranging from 4.4 to
4.8 km/s may be basaltic pillow lava.
3. Velocities of 6.3 km/s are typically
gabbroic rocks.
4. The velocity of 8.1 km/s signifies the
Moho .
Extrapolation of the seismic results obtained
at Sal and Sao Vicente suggests that the Moho rises
toward the Sal-Maio ridge.
A wide faulted and fractured zone is associat-
ed with the Sal-Maio ridge (Figure 1) . The
attitude of the Moho in this area suggests that
the western islands of the Cape Verde archipelago
lie on a crust with a thickness of between
16 and 17 km. the implication thus being that
the islands are structurally not part of the
African continent.
Seismic Reflection Profiling
In 1968 the U.S. research vessel Kane carried
out a reconnaissance reflection seismic, magne-
tic, and gravity survey from Dakar to the eastern
Cape Verde Islands [Lowrie and Escowitz, 1969].
In 1970, NOAA participated in the present research
project with their ship Discoverer. Their work
was confined to reflection seismic, gravity,
and magnetic profiling within the Cape Verde
archipelago. The tracks of Discoverer and Kane
are shown in Figure 5. For the sake of better
understanding the reflection data obtained by
Discoverer the results of the Kane survey are
also presented here (Figures 6, 7, and 8). The
following is a description of these records.
West of Dakar to the Sal-Maio ridge, line AA
e Verde block
fied deposit
annels . The
al and Mauritania
sediment
ratified deposit
ion of sediments
lowing seaward
ick sequence
the Cape Verde
or blankets the
(Figure 6) . In the east the Cap
is bounded by a 1-s-thick strati
incised by canyons and leveed ch
present drainage system of Seneg
is insufficient for considerable
accumulation, and the present st
can only be explained by deposit
from the continent or by water f
across the shelf break. This th
ends abruptly about 185 km from
block. A shallow opaque reflect
TABLE 2. P Wave Velocities and Thicknesses of
Various Refractors Along Refraction
Profile B as Recorded at Sao Tiago
Depth to
Veloci
ty
Th
ickness
Top of Layer
Horizon
km/s
kms
km
Water
1.5
2.9
0.0
1
3.2
1.9
2.9
2
4.5
5.6
4.8
3
6.4
6.3
10.4
4
7.9
16.7
259
Dash et al . : Geophysical Investigation of the Cape Verde Archipelago
25 -i
5253
Fig. 4.
Time-distance plot of profile C. Bathymetry along the profile is shown below
the time-distance graph.
island block. The opaque reflector may be
cherty, volcanic, or calcareous, or perhaps a
combination of these lithologies. A seamount east
of the Cape Verde block is easily recognizable.
At the western end of line AA the Sal-Maio
ridge is steep sided and flat topped. The
basement beneath the ridge steps down to the east.
A fault occurs at the first of these steps of
the eastern slope. Ballard and Hemler [1969]
have noted a N-S striking fault with a dis-
placement of about 450 m at this location.
Some indications of sediment infill, possibly
due to sediments derived from the ridge and
the seamount, are seen on the western side
of the seamount.
Sal-Maio ridge to the abyssal plain, line BB
(Figures 6 and 7) ■ This section of the profile
runs northwest through the Cape Verde 'horseshoe'
At the eastern end of this line (Figure 7) ,
transparent poorly stratified presumably
pelagic deposits cap a protruding and elon-
gated hill. The opaque reflector, as was
indicated earlier, can be found almost in-
variably on this profile. Two diapiric
structures protruding through the seabed
can be seen. At the extreme end of thj
profile a partly buried abyssal hill covered by
1-s-thick stratified sediments is apparent.
South of the Fogo-Sao Tiago ride to north of
Sao Nicolau, lines DP and CC (Figure 8). The
N-S profile across the Fogo-Sao Tiago ridge
suggests some small displacements indicated by
irregular topography and hyperbolic echoes.
Although they are not very clearly defined,
there are suggestions of the presence of
several slumped blocks on the ridge. Numerous
sediment masses have slid down the flanks of
the ridge and are best seen on its south side.
In the area north of the ridge and south of
Sao Nicolau, stratified sediment partly covered
by the dark reflector can be seen. the presence
of hyperbolic echoes south of Sao Nicolau
suggests a rough shallow basement. Seismic
refraction data (line C, Figure 1) in the
vicinity of this profile confirm the presence
of this layer at a depth of 1.9 in 2.1 km.
This basement appears to rise stepwise toward
Sao Nicolau from about 15 n. mi. (28 km) south
of the island. Slumped blocks are evident at
the base of the island rise. Farther north,
beyond Sao Nicolau, thick deposits of sediments
are noticeable. The basement in this area is
rugged, with indication of faulting, consistent
with the N-S fault lineation suggested by
Ballard and Hemler [1969].
The seismic reflection records within the
Cape Verde archipelago show several successive
areas having well-defined to poorly defined
stratified zones. The basement configuration
is also ill defined except in certain areas.
Wherever there are clear reflections, correlation
with refraction seismic data is excellent. From
the refraction data it is apparent that the aver-
age thickness of sediments varies from 1.9 to
2.4 km. This is in close agreement with the
indications from the reflection results.
According to these observations the area inside
the Cape Verde 'horseshoe' seems to be built
up by sediments supplied from either the con-
tinent or the islands. Erosion seems to be of
minor importance, while slumping was observed
more frequently.
TABLE 3. P Wave Velocities and Thicknesses of
Various Profile C as Recorded at
Sao Vicente
Depth to
Velocity
Th
ickness
Top of Layer
Horizon
kr?/s
km
km
Water
1.5
2.4
0.0
1
3.2
2.4
2.4
2
4.4
5.8
4.8
3
6.6
5.6
10.6
4
8.0
16.2
260
5254
Dash et al.: Geophysical Investigation of the Cape Verde Archipelago
Fig. 5. Track chart of air gun reflection lines AA, BB , CC, and DD. Data on lines AA
and CC were collected by Kane, the rest by Discoverer. The contours are in
fathoms (1 fm = 1.8288m).
Gravity Investigation
The seismic data were supplemented by the
measurements of the earth's gravitational field
on and around the islands (Figure 9). The
marine data were collected during two research
cruises in the area. In 1970 the NOAA ship
Discoverer carried out a limited survey (3500 km)
around the islands. The ship's position was
accurately known at all times by means of
satellite and Omega navigation systems. The course
speed, and position of the ship and the
bathymetry of the area were recorded along
the profiles. All these data together with
the gravity readings taken, on the average,
every 5 to 6 km were fed into the on-board
computer (Univac 1208) which applied latitude
and Eotvos corrections to the data and thus
yielded the free air anomaly value at each
station. In 1972 the same ship recorded a
further 1350-km line of gravity profiles.
This survey was conducted mainly within the
island 'horseshoe' and provided a tight grid
of data to the west of the Sal-Maio ridge.
In addition to these marine data, gravity
values at 168 land gravity stations (collected
by Servico Meteorologico Nacional de Portugal
and kindly made available to us) were used.
The land data were corrected to a sea level
datum, and all three surveys were combined.
Since the gravity data were still influenced
by variations in the ocean bottom topography
and the subaerial topography of the islands,
it was necessary to remove these effects in
order to observe abnormalities in the structure
of the crust. The method used was that pro-
posed by Talwani and Ewing [1960] for a
three-dimensional body. To this end a new
bathymetric chart was contoured from the
existing data together with those collected
261
Dash et al . : Geophysical Investigation of the Cape Verde Archipelago
52S5
O CN f O
LU
I
CO
if
9
4-1
o
0
£,
V)
C
o
■D
■H
4J
9
O
QJ
<U
01
60
•o
O
■H
1
M
co
O
•H
0)
rcl
CO
a
oo
•H
'-flt
CO
>
o
u
Q
CO
o
o
o
o
o
o
m
o
m
•"■
CO
^-
T3 H
0) -H
4J
n
CO
ca
01
S
c
<T)
cu
J3
-o
4-1
c
c
u
o
•H
•H
p.
4-)
CD
u
■H
01
T3
01
o
c
0
3
•H
4-J
o
cu
a
rH
■H
00
•H
262
5256
Dash et al.: Geophysical Investigation of the Cape Verde Archipelago
z
o
r
ii >
I;
r
A
z "■
<
z
I
C/3
o
is
.c
y)
n
■iH
o
C
c
Hi
CD
.C
4J
(J
«
o
T3
C
c
HI
3
O
o
l_>
«
V)
(1)
(U
X.
c
4-1
by the Pillsbury and Discoverer (Figure 1) .
In order to estimate the most likely
replacement density the terrain-corrected
Bouguer anomalies were computed for a range of
reasonable density contrasts. If the gravity
field varies smoothly, it can be represented
by a low-order polynomial. Various low-order
polynomials were calculated for each density
contrast. The root mean square deviations
between the predicated values of the field
and the actual values were computed for each
surface fitted. A fifth-order polynomial
was accepted for which a least squares best
fit was obtained. Comparison of the results
of this polynomial fitting method with those
of the density profiling method of Nettleton
[19 39] revealed that a value of 2.3 g/cm
for the subaerial parts and 2.4 g/cm for the
submarine portions should be accepted for the
replacement bulk densities o be used for
detailed investigations in tne Cape Verde
region. These values are all in accord with
those predicted from the Nafe and Drake
[1963] curve. The detailed terrain-corrected
Bouguer anomaly map is shown in Figure 9.
There are numerous positive anomaly closures
which are assumed to be caused by high-density
plutonic bodies. To the southeast of the
island of Maio there occurs a sharply angled
bend in the isogals coupled with extremely
high gradients perpendicular to both arms
of the bend. The northward continuation of
the bend passes through major deviations in
the isogals to the southeast of Boa Vista
and emerges to the east of Sal, where another
steep gradient is observed. A major roughly
N-S fault is postulated to the east of the
islands of Maio and Sal.
Bebiano [1932] postulated a major fault
through the four northern islands extending
to Boa Vista. There is no extreme anomaly
gradient perpendicular to this proposed
fault trace. It is possible that the
observed minor and major deviations of the
isogals could indicate a large tectonic
feature whose field has been partially
obscured by the interfering effects of
extruded magma occupying a large area.
Analysis of the data corrected with the
regional replacement density of 2.58 g/cm
indicates that the Moho ranges between
16 and 18 km beneath the island block. These
values are in excellent agreement with those
inferred from the seismic refraction data.
Magnetic Anomalies
The magnetic quiet zone boundary in the
eastern North Atlantic lies between the Cape Verde
archipelago and the African continent and trends
basically N-S [Heirtzler and Hayes, 1967;
Rona et al. , 1970]. A sequence of oceanic
magnetic anomalies, the Keathley sequence or
J anomalies, forms a 350-km-wide band seaward
of the quiet zone boundary and extends up to
the eastern margin of the Cape Verde archipelago
(Figure 10). Vogt et al. [1970] along with
Rona et al. [1970] have shown that this J anomaly
band occurs with almost mirror image similarity on
the east and west sides of the central North
Atlantic.
263
Dash et al . : Geophysical Investigation of the Cape Verde Archipelago
5257
18 w
TERRAIN-CORRECTED BOUCUER ANOMALY
CAPE VERDE ISLANDS
CONTOUR INTERVAL 10 MGAL
26
25
24
23
22
Fig. 9.
Terrain-corrected Bouguer anomaly map around the Cape Verde archipelago with a
contour interval of 10 mGal.
Tentative correlation of oceanic magnetic
anomalies in the Cape Verde area based on our
new data and added on to those published by
Rona et al. [1970] substantially confirms the
general N-S trends established by previous
work. There is some indication of minor offsets
of anomalies, although no fracture zones can
be mapped except the possible eastern extension
of the Kane fracture zone at about 21°N latitude.
Where deviations from the general N-S anomaly
trend occur, the preferred orientation is NNW-SSE,
parallel with structural features and morpholo-
gical lineations within the island group as
well as with the postulated deep fault trends
derived from the seismic and gravity data.
Conclusions
The present crustal structure beneath the
Cape Verde archipelago determined by our seismic
refraction and gravity measurements is transi-
tional in that the Moho lies at a crustal depth
between 16 and 17 km, midway between dimensions
typical for continental and oceanic crust.
It is interesting to compare the crustal
structure of the Canary Islands, lying
off the coast of Spanish Sahara with
that of the Cape Verde archipelago. Dash and
Bosshard [1969] postulated that the five western
islands of the Canaries group are not related
structurally to the African continent. The
crust is of oceanic thickness on the west. In
the central part of the island group the
thickness of the crust is transitional. The
depth to Moho varies from 12 to 14 km. Roeser et
al. [1971] , from their refraction seismic and
gravity studies of the area between Africa and
Gran Canaria, suggest a Moho depth of 21 km
with a crust originally of oceanic character
having presumably been depressed to the depth
of 21 km with 10 km of differentially meta-
morphosed sediments deposited on it. The
transitional zone between the oceanic and
continental crust is characterized by major
faults of NE-SW strike.
The original composition of the crust
underlying the Cape Verde archipelago was
probably oceanic as is evidenced by (1) the
264
S2S8
Dash et al.: Geophysical Investigation of the Cape Verde Archipelago
16'|
■J\ v
j s.iA
■
^»
■ ,v. >
CAP BLANC
- ^
*y\
v v
v^vu
>",/.
V. ^ . /
, t
1/
'v;- •*,;_,
"'•:,
^"v^.
CAPE VERDE ISLANDS
DAKAR
Fig. 10. Magnetic anomaly map around Cape Verde Islands and continental shelf of Africa
to 24°N. The Keathley sequences, a possible fracture zone, and the magnetic
quiet zone boundary are indicated. No major bottom topographic features occur
along tne magnetic track lines.
occurence of a typically oceanic layer 3 with
a velocity of 6.3 km/s^ (2) the presence
of the magnetic quiet zone boundary landward of
the islands [Rona et al. , 1970]; and (3) the
presence of open water aptychus limestone (Upper
Jurassic-Lower Cretaceous) exposed on Maio
similar to limestones recovered from analogous
locations of the western North Atlantic deep
ocean basin [Hollister et al. , 1972]. The
origin of the Canaries is closely linked with the
faults through the islands [Dash and Bosshard ,
1969]. The Cape Verde archipelago also
originated as igneous intrusions and extru-
sions along a fault system. The similarities
between the Canary and Cape Verde islands end
at the comparison of their crustal structures.
Owing to a lack of any detailed seismic refrac-
tion data in the area between the t^o island
groups it is not possible to draw any
extension of similarities between them.
The magnetic signature of each of the two
island groups is singularly distinctive. The
Cape Verde islands lie west of the magnetic
quiet zone, whereas in the Canaries the
magnetic quiet zone passes through the island
of Tenerife. The magnetic quiet zone almost
follows the transitional crustal zone on the
west coast of Africa.
To reconcile the present transitional
thickness and the inferred former oceanic
composition of the crust beneath the Cape Verde
archipelago, we postulate that these islands
originated as igneous intrusions and extrusions
along the N-NW, NW, and W-SW trending fault
systems. The Moho was depressed from oceanic
to transitional depth as a result of the loading
of accumulated materials of the islands. It is
suggested that the building of the Cape Verde
archipelago began at an early stage in the
Mesozoic opening of the central North Atlantic
and at no time did the islands belong to the
African continent. The position of the
archipelago was apparently controlled by the
convergence of fracture zones as a consequence
of the marked change in the trend of the adjacent
mid-Atlantic ridge.
Acknowledgements . The authors wish to
thank A. Richardson, F. Machado , P. Hubral ,
B. Buttkus, and C. J. M. Hewlett for their
265
Dash et al.: Geophysical Investigation of the Cape Verde Archipelago
5259
help during the collection of the data.
Thanks are also due to the officers and crew
of the R/V John Elliot Pillsbury and the
NOAA ship Discoverer. Financial support for
the project was kindly made available by the
Natural Environment Research Council of
Great Britian, U.S. ONR grant N00014-67-A-
0201-0013, NSF grants GA-39744, GA-19471,
GB-27252, and GA-27465, and NOAA. The
authors gratefully acknowledge this support.
References
Assuncao, C.F.T. de , F. Machado and A. Serralheiro,
New investigations on the geology and
volcanism of the Cape Verde Islands,
Int. Geol. Congr. Rep. Sess. 2 3rd, 2, 9-16,
1968.
Ballard, J. A., and L.G. Hemler, Structure of the
Cape Verde rise (abstract), Eos Trans. AGU,
50, 210, 1969.
Bebiano, J.B., A geologia do arquipelago de
Cabo Verde, Comun. Serv. Geol. Port. , 18,
1-275, 1932.
Dash, B.P., and E. Bosshard, Seismic and gravity
investigations around the western Canary
Islands, Earth Plan et. Sci. Lett. , 72,
169-177, 1969.
Dash, B.P. , K.O. Ahmed, and P. Hubral, Seismic
investigation in the region of Poulo Panjang,
offshore from southwestern Vietnam, ECAFF,
Tech. Bull. , 3, pp. 37-54, 1970.
Egloff, J., Morphology of the ocean basin sea-
ward of northwest Africa: Canary Islands
to Monrovia, Liberia, Artier. Ass. Petrol.
Geol. Bull. , 56, 694-706, 1972.
Heezen, B.C., and M. Tharp, Physiographic
diagram of the North Atlantic Ocean, Geol.
Soc. Amer. Spec. Pap. , 65, 1968.
Heirtzler, J.R. , and D.E. Hayes, Magnetic
boundaries in the North Atlantic Ocean,
.Science, 157, 185-187, 1967.
Hollister, CD., et al . , Initial Reports of
the Deep Sea Drilling Project, vol. 11,
p. 1077, U.S. Government Printing Office,
Washington, D.C. , 1972.
LeBas, M.J., Per-alkaline volcanism, crustal
swells and rifing, Nature Phys . Sci. ,
230, 1971.
Lowrie, A., and E. Escowitz, Eds. Kane 9, in
Global Ocean Floor Analysis and Research
Data Series, 971 pp. , U.S. Naval Oceano-
graphic Office, Washington, D.C, 1969.
Machado, F. , Vulcanismo das Ilhas de Cabo
Verde e das outras ilhas Atlantidas ,
Junta Invest. Ultramar Port. , Estud.
Ensaios Doc. , 117, 1-83, 1965.
Machado, F. , J. Azeredo Leme , and J. Monjardino,
0 complexo sienitocarbonatitico dal Ilha
Bravo, Cabo Verde, Garcia de Orta, 15, (1)
9 3-98, 1967.
Nafe, J.E., and C.L. Drake, Physical properties
of marine sediments, in The Sea, vol. 3,
The Earth Beneath the Sea, edited by
M.N. Hill, pp. 794-813, Interscience ,
New York, 1963.
Nettleton, L.L. , Gravity and magnetic calcula-
tion, Geophysics, 4, 176-185, 1939.
Part, CM., Volcanic rocks from the Cape Verde
Islands, Bull. Brit. Mus . Natur. Hist.
Geol. , 1(2) , 27, 1950.
Roeser, H.A., K. Einz, and S. Plaumann,
Continental margin structure in the Canaries,
The Geology of the East Atlantic Continental
Margin, 2, Africa, edited by F.M. Delany ,
rep. 70/16, pp. 27-36, Inst. Geol. Sci.,
Rons, P. A., Bathymetry off central northwest
Africa, Deep Sea Res. , 18, 321-327, 1971.
Rona, P. A., J. Brakl, and J.R. Heirtzler,
Magnetic anomalies in the northeast Atlantic
between the Canary and Cape Verde islands,
J. Geophys. Res. , 75(35), 7412-7420, 1970.
Serralheiro, A. , Geologia da Ilha de Maio (Cabo
Verde) , Junta Invest. Ultramar Port. ,
Estud. Ensaios Doc. , 103, 1970.
Talwani , M. , and M. Ewing , Rapid computation
of gravitaional attraction of three-
dimensional bodies of arbitrary shape,
Geophysics, 25, 203-205, 1960.
Turner, F.J., and J. Verhoogen, Igneous and
Metamorphic Petrology, p. 694, McGraw-Hill,
New York, 1960.
Vogt, P.R., C.N. Anderson, D.R. Bracey , and
E.D. Schneider, North Atlantic magnetic
smooth zones, J. Geophys . Res . , 75(20),
3955-3968, 1970.
(RSceived July 31, 1975;
revised February 13, 19 76;
accepted February 29, 1976.)
266
26
Reprinted from: Sea Frontiers, Vol. 22, No. 1, 9-15.
Iceland
Where the Mid-Ocean Ridge Bares Its Back
By Robert S. Dietz
NOAA. Atlantic Oceanographic and Meteorological Laboratories
Miami, Florida.
Only in ICELANO can man walk on the Mid-Atlantic Ridge. This is die one place where
pan of the 45.0<)(l-kilonieter-/ong ocean rift is exposed above sea level.
Robert S Dietz
ICELAND, a bleak, windswept island
in the far North Atlantic, touching
on the Arctic Circle, lies on rock
hotter than lands at the equator. It is
not entirely a foolish joke to say that
an inhabitant of this island who runs
short of hot water in his bathroom
has only to drive a pipe down through
the floor toget plenty for his hot bath.
But the interest of geologists runs
deeper and concerns more funda-
mental aspects of the earth's history
than hot water.
A Grand Scheme
Only at Iceland does the 45.000-
kilometer-Iong mid-ocean ridge, a rift
marking the pulling apart of the
earth's crustal plates, breach the sur-
face of the ocean. This island is. there-
45.000 kilometers - ?7'I00 miles
LOOKING north along Iceland's central
nft. one can see where the earth's crust
is slowly being pulled apart. To the left
of the rift, the western Atlantic Ocean
and North America to as far as the San
Andreas Fault in California are drifting
west at a rate of I centimeter each year.
To the right of the rift, the eastern Allan-
tic Ocean and all of Tunisia are drifting
eastward to as far as the Pacific trenches
off Kamchatka and Japan. In Iceland, the
rifting is strongh overprinted h\ com-
panion effects — the formation of vol-
canoes and the effusion of lava above a
vast ascending plume of magma, rising
from deep within the earth's mantle.
10
fore, crucial to the revolutionary new
concept of plate tectonics, or struc-
tural geology of the earth's crust.
{Also see "A Magnificent Revolu-
tion." Sea Frontiers. Vol. 18. No. 6,
November-December. 1972.)
According to plate tectonics, the
earth's crust is a mosaic of about
eight 100-kilometer-thick rigid plates,
or shells, which slowly drifts over a
100 kilometers = 62 miles
268
plastic upper mantle. The plates do
not collide with one another: instead,
one edge subducts, or descends, into
the earth's mantle while the opposite
edge accretes new ocean floor to its
margin. The latter process occurs at
the mid-ocean ridge and is called sea-
floor spreading. Along still other
boundries of a crustal plate are giant
zones of shear, or transform faults,
where a plate slides past its neighbor.
Geologists only recently have come
to understand this grand scheme of
earth tectonics because the evidence
is largely not on land but out of sight
beneath the sea. Tectonism. or per-
manent displacement of the earth's
crust, is confined to the plate bound-
aries which, although over 100.000
kilometers in total length, are nearly
all in oceanic crust. Major exceptions
100 000 Hometers = 62 000 miles
Roberts Dietz
269
are California's San Andreas Fault (a
transform fault); Africa's Afar tri-
angle (a triple junction where three
plates join) at the nexus of the Red
Sea and Gulf of Aden: and Iceland,
the only place on earth where the mid-
ocean ridge is above sea level.
Three Types of Volcanism
Geologists recognize three distinct-
ly different types of volcanism. or
lava production, on the earth. The
first is suhduction (calc-alkalic) vol-
canism associated with the oceanic
trenches and island arcs. It is caused
by the return of molten rock to the
surface from crustal plates that are
being subducted, or carried down,
into the earth's mantle. The lavas are
charged with steam and thus are high-
ly explosive. They create the classic
volcanoes around the Pacific "ring of
fire" such as Mount Rainier in the
United States and Mount Fuji in
Japan.
The second type of volcanism
(tholeiitic) is that which injects the
dikes and pillow lavas that fill in the
mid-ocean ridge as it spreads apart.
This process, which generates new
ocean floor by symmetrical accretion
to the plates that are moving apart,
is called sea-floor spreading. This vol-
canism is effusive and quiet, produc-
ing dikes and flows, but not a single
volcanic cone. Although never di-
rectly observed, this type of volcan-
ism adds more to the earth's crust
than either of the others. It repaves
the ocean floor along the mid-ocean
ridges around the world at the rate
of 2 square kilometers per year-
enough to renew the entire ocean
floor in only 1 50 million years.
A third type of volcanism is plume
(alkalic) volcanism, caused by lavas
that rise as ascending columns from
deep within the earth's mantle. Upon
breaking through the earth's crust,
they create volcanoes that may be
compared to the thunderheads that
form over ascending columns of air.
Plume volcanoes usually form in a
row as the magma rises from the fixed
deep mantle over which the earth's
outer crust is drifting. The Hawaiian
Chain is a good example. The Pacific
plate is moving northwest at about 10
centimeters each year so that, as the
old volcanoes drift away, a new one
is created over the fixed plume site.
The only modern active volcanism is
on the big island of Hawaii, at the
southeast end of the chain.
Iceland has been built by the last
two named types of volcanism:
tholeiitic (rift injection) lavas and
plume lavas. Among the chief centers
of plume volcanism on earth, Iceland
probably ranks first, spewing out
about 20 percent of all surface lavas.
(Other major centers of plume vol-
canism are Hawaii, the Galapagos
Islands, and the Azores.)
In Iceland, the rift lavas are abun-
dantly augmented by plume lavas,
which have built more than 200 vol-
canoes, many of them active. This
volcanic pile that straddles the Mid-
Atlantic Ridge is thus of a composite
nature. The process of sea-floor
spreading (rifting) observable in Ice-
land is strongly "overprinted" by
plume lavas. Accordingly, the spread-
ing process within Iceland is more
complex than the beautiful simplicity
2 square kilometers = 0 77 square mile
10 centimeters = 39 inches
12
Sea Frontiers
270
ASTHENOSPHERE
This mcim simim n n i> sketch illustrates the process of marginal plate accretion, or
sea-floor spreading, which lakes place at ihc mid-Atlantic rift. As the America plate
1A1 and the Eurasia plan- 1D1 are pulled apart, a dike <>l hot lava is injected into the
earth s lithosphere. or crust. Partial melting in the soft asthenosphere (the region below
the crust I provides cm ever-present source of new magma. I he hot dike I speckled I cools
ihluci against the adjacent plates. With renewed extension tinsel I. the dike breaks
svtnmeiricullv along us warm and. hence, weak axis. L'pon congealing and passing
through the so-called Curie point at ^~5°C .. the upper portion of the dike takes on the
ambient sense of the earth's magnetic field. The while bands are intervals when the
earths magnetic field has been normal: the black ones indicate intervals of reversed
magnetic fields.
observed farther to the south along
the Reykjanes Ridge, and along most
other portions of the world-wide mid-
ocean ridge system. While it is true
that the mid-ocean ridge does hare its
back at Iceland, this exposure is
somewhat anomalous, complex, and
atypical.
The geologic structure of Iceland is
dominated by two giant rifts which
trend, generally, north-south. These
rifts are clearly discernible from the
air. The eastern rift is now compara-
tively inactive, anil it is believed that
active rifting, or spreading apart, is
lamely confined to the western rift.
January-Februdfy 1 ')/<»
13
271
The 0'
:f theOcea- : R c^es 3> Ej:-
1959 b> S: e-r ( c Arr,er can Inc. A;: '.ghts reserved
Flown bi na\ 'i aircraft over the cresi of the Reykjanes Ridge south of Iceland, this
magnetic survey shows, in red. the present period of normal magnetism (during which
the north magnetic pole has been near the north geographic pole). This period extends
back to "DO. 111)1) years ago. and it overlies and flanks the axis of the mid-ocean ridge.
Other rainbow colors mark earlier periods of normal magnetism back to anomaly five
which occurred about in million vears ago. Intervening periods of reversed magnetism
i w hen the north magnetic pole was near the south geographic pole) are shown in white.
The anomaly patterns are symmetric, as each injected dike eventually split into two
equal parts which accreted to opposite plates. This sun ey thus provides "back-to-back
tape recorders" ot ocean-floor growth. Each limb of anomaly patterns is lot) kilometers
wide which means, since anomaly five is 10 million years old. a growth, or spreading,
rate of ID kilometers per million vears or 1 centimeter per year. This is a separation
rate for the two plates of 2 centimeters per year. The Reykianes Ridge is. therefore,
splitting apart at the rate at which a fingernail grows.
14
Sea F-cn: ers
272
Down-dropped blocks of basalt, a
dark lava rock, reveal that there has
been extension within the earth's
crust. The earth's outercrust. or litho-
sphere, is computed to be pulling
apart at a rate of about 2 centimeters
per year. Attempts have been made
by scientists from Imperial College
in London to actually measure this
rifting, using a laser beam. The re-
sults, thus far. are not conclusive but
are said to be consistent with the
theoretically computed 2 centimeter-
per-year spreading rate. No clear pat-
tern of magnetic anomalies are ob-
servable along the Icelandic rift, but
this seems certainly related to two
factors: the confusion created by the
plume lavas and the fact that strong
magnetic imprinting occurs only in
the quickly quenched pillow lavas,
which must be erupted beneath water.
Magnetic Anomalies
Reference to the Reykjanes Ridge
laying athwart the Mid-Atlantic
Ridge, immediately to the south of
Iceland, is convincing evidence that
the Icelandic rift was created by sea-
floor spreading. In fact, this process,
although inferred earlier by geologic
considerations, was first demonstrat-
ed by an aerial magnetic survey flown
across this ridge. This survey revealed
a succession of stripes or bands of
strongly magnetized rock with their
magnetic signal being alternated, i.e..
normal and then reversed in sign. The
banding, or stripes, of this survey
quantitatively measured the growth
of the ocean floor in a manner some-
what analogous to the growth of a
tree by its annual rings. The anomalies
revealed that the earth's magnetic
field switches its polarity, so that the
north pole becomes the south pole
(and vice versa), about once every
one half million years. The ambient
direction of the earth's magnetic field
is frozen into the mid-ocean ridge
lavas as they pass through the so-
called Curie point at 575°C. when
solidifying.
A central band running along the
axis of the mid-ocean rift was found
to be normally magnetized. With re-
spect to this central band, the others
on either side of the ridge lay in mirror
image, so that, if the survey map was
folded into a V along the axis, the
anomalies on opposite sides of the
fold would be juxtaposed. Clearly
these were vertical growth lines dem-
onstrating that the ocean floor had
grown bv some process whereby new
sea floor was being slowly accreted
to crustal plates moving apart from
the mid-ocean rift locus. Although
these remarkable magnetic anomalies
could not be traced through Iceland,
lava ages showed that a similar pat-
tern existed. The strips of lava are
progressively older on both wings of
Iceland as one moves away from the
central rifts.
Iceland is thus not only a remote
island of vivid contrasts in the far
North Atlantic touching on the Arctic
Circle. Its mountains, volcanoes.
geysers, and thermal springs have a
deeper significance. Its rugged youth,
with no portion being older than 15
million years, can now be understood.
It is the only place on earth where
one can actually observe the earth's
crust being pulled apart.
2 centimeters = 0 78 inches
575°C = 1.035°F
January-February 1976
15
273
27
Reprinted from: Oaeanus, Vol. 19, No. 4, 19-22.
EARLY DAYS
OF
MARINE GEOLOGY
BY R.S. DIETZ AND K.O. EMERY
We hold no brief for the "good old days" but
perhaps it adds to the perspective of marine sciences,
and certainly to humor, to recall something of the
beginnings of marine geology in the United States
by citing some of our early experiences. This
stibdiscipline of geology commenced almost
simultaneously in the mid- 1930s on the East Coast
at the Woods Hole Oceanographic Institution with
the research of Henry C. Stetson and on the West
Coast with the studies of Francis P. Shepard.
Stetson died at sea aboard Atlantis off Chile in 1955,
while Shepard is still actively working at the Scripps
Institution of Oceanography in La Jolla, California.
We were the first of Shepard's sixty or
so marine geology students, shuttling with him
between the University of Illinois and Scripps. We
met at the University of Illinois, where we arrived
via modes of transportation that were the norm for
those Depression days. Dietz arrived by hitchhiking
from the East Coast; Emery came by train, riding
boxcars from San Diego.
In 1936 Shepard received a grant from the
Penrose Fund of the Geological Society of America
for studying submarine canyons and the sea floor
generally off the coast of California. The amount
was SI 0.000, which was a handsome grant for those
days— in fact, the largest ever given by the GSA in
prewar years. With the money he was able to
charter the 96-foot schooner E. IV. Scripps of the
Scripps Institution of Oceanography for six one-
month cruises, build the necessary scientific
equipment, employ us as his assistants at a salary of
S30 per month, and support the abortive
development (to the tune of S 1 ,000) of the Varney-
Redwine hydrostatic corer. It was hoped that this
latter device would outperform the famous
C. S. Piggot gun corer, which shot the barrel into
the ocean bottom. We should add "in principle,"
because the Piggot device, when used from Atlantis,
seemed to obtain cores of equivalent length whether
or not the gun actually fired. A subsequent grant
provided for three more months in the Gulf of
California during the fall of 1940. Since bed and
board was provided aboard ship, we both signed
on for $1 for the three months to make us official
expedition members (but Scripps never paid the
Si -perhaps fearing that we'd spend it unwisely).
An interesting guideline also was that students
should not receive any pay for research that
pertained to their own thesis projects.
The low funding at least required us to
develop some ingenuity in devising simple,
inexpensive instrumentation. For example, we
used the 2-meter-long Roger Revelle, later director
of Scripps Institution of Oceanography, as a wave
staff by having him stand at various distances from
shore in the buffeting surf. This rather absent-
minded wave staff also was noted for having stepped
into a bucket while measuring cores aboard ship and
wearing it for a couple of hours. As another
example, we organized a rock preparation and
sedimentation laboratory for which a budget of
S50 per year was arranged. This was considered a
reasonable proportion of the Scripps' overall budget
of SI 25,000 per year.
Notably also. Woods Hole Oceanographic
Institution was founded in 1930 with a gift of
S3, 500,000 from the Rockefeller Foundation
received over a period of several years; this was
sufficient to construct the large brick Bigelow
Building and the ketch Atlantis, and to cover all
operations for ten years. The annual budgets for
274
19
The 96-foot schooner E. W. Scnpps. principle research
vessel of the Scripps Institution of Oceanography from
1937 to 1 950. (Courtesy of SIOj
the two institutions have remained about equal,
nowadays almost S22 million for Scripps and
S20 million for Woods Hole.
Life aboard E. W. Scripps was somewhat
different from shipboard duty today. The ship's
crew consisted of only four persons— captain,
engineer, deck hand, and cook; the scientific party
was seven-the number of bunks available. We
generally worked around the clock, six hours on
and six off. The scientific party was expected to be
sailors to run the ship and technicians to operate
oceanographic winches, assemble and use the water
and bottom samplers, and do various shipboard
analyses for water chemistry. Among our duties
while steering the ship was to tabulate by hand the
water depth every two minutes. We did this with
great enthusiasm since we had installed aboard the
latest Submarine Signal Co. fathometer, which
indicated the depth on a revolving red-flashing
neon light. Graphic recorders had not yet been
invented, so this instrument represented to us a
remarkable advance over the sounding lead. And.
in fact, we continued to use the hand-powered
wire-and-lead sounding winch installed on a rowboat
for making hydrographic surveys of the inner heads
of several submarine canyons. Rather remarkably.
it was possible to demonstrate that canyon heads
were repeatedly filling with sediment and then
emptying out.
Prior to the cruises we built dredges, grab
samplers, sediment traps, and corers. The best
corer that we constructed was a 600-pound open-
barrel gravity model that increased the weight of
such devices over earlier models by a factor of
ten. We purchased junk load at 34 per pound, used
scrap 2^-inch pipe, and built two corers for about
S50 each. It was not until after the war that we
heard about Kullenberg's invention of the piston
corer. Nevertheless, we commonly obtained cores
12 feet long, and in one instance, a diatomaceous
ooze core in the Gulf of California 1 7 feet long for
a new record. Of course, things were considerably
cheaper in those times. By way of example, an
apparently wealthy American tourist at the local
swinging bistro named El Tecolote (The Owl) in
Guaymas. Mexico, generously offered to buy beer
for our ship's staff. When he discovered that the
bartender could not change his U.S. S10 bill, he
gallantly said, "Set up the whole amount in beer."
One hundred and twenty bottles of Carta Blanca
were lined up along the bar, and as was customary
then in Guaymas, bowls of unshelled shrimp were
thrown in like the free peanuts of today. A side
advantage was that the long row of beers
immediately stopped the girls' pestering us for
drinks.
After the cruises, when the ship was
unloaded so that the biologists or physical
oceanographers could take their turns, we were able
to study the samples and other results. Since all
was new, both to us and to others, we had no
difficulty in finding problems. Our masters' theses
were on mechanics of coring and on the extensive
phosphorite deposits we discovered covering many
of the offshore banks. Our doctoral dissertations
were on clay minerals of the deep areas ;md on rocks
of the shallow banks. We recognized that the
offshore basement geology of the Southern
California borderland belonged to the Franciscan
province. Articles on terraces, currents, barite
concretions, and transport of rocks by kelp and
sea lions were by-products. The overall results were
incorporated into Special Paper 31 of the Geological
Society of America by Shepard and Emery, a
monograph treating submarine canyons and the
general bathymetry of the sea floor off California.
These must have been our most productive years in
terms of variety and number of investigations,
because of newness of the field and, probably, the
aid of funds too small to permit much diversion of
time and energy.
275
Local transportation was provided by a
succession of old cars, starting with a 1928 Chevy
that Shepard bought for us for $50. By the time
We drove it 1 5 miles, cork in the transmission wore
out and serious noises developed. Replacement by
junk gears extended the life of the Chevy for a year
or so. After tiring of having to tie a rope around
the car to keep the doors closed, we swapped it for
a 1928 Reo that had a good engine but bad tires.
Eventually, this was swapped for Walter Munk's
1028 Buick (The Queen Mary). The state of the
Reo's tires is illustrated by a blowout of the spare
tire in the hot California sunshine when he drove
northward too long. In time the differential of the
Buick disentegrated, and a 1928 Ford was next.
The total cost of these four cars was $200-nothing
•?
Emery with hollow giant worm or animal tube of enigmatic
origin dredged from the wall of Dume Canyon off
California, May 1 938.
Dietz with gravity coring device at rail of E. W. Scripps,
about June J 938.
compared to their present value as antiques if they
had been stored until now.
The four years of cross-country commuting,
cruising (at least I 2 months aboard E. W. Scripps),
and study came to an end in 1941. Just before
receiving his doctorate, Emery wrote 135 individual
letters, blanketing the entire country, seeking
employment. Dietz, being congenitally lazier (or
possibly more efficient), trusted that this blizzard
of inquiries would produce several plums of which he
might select one after Emery made his acceptance.
But the market for marine geologists, like the job
market for poets then and now, was bleak. Not a
single position was tendered. As with many
products, there is commonly no demand for the
first ones off the line. Even the U.S. Navy saw no
particular need to know anything about
oceanography; in fact, its interest, when it did
develop, probably stemmed from the initiative
shown by the Army Air Corps in setting up a group
of officers and civilians to predict the paths of
downed airmen in their rubber life rafts carried by
surface currents of the western Pacific.
In retrospect, the "good old days" were
both the best of times and the worst of times.
Happily, one tends to recall the ups rather than the
downs— and there is no substitute for the bouyancy
276
21
of youth. Oceanography of today is, of course,
much more sophisticated and the results ever more
quantitative. But there was a certain enjoyable
simplicity, and even beauty, in working with
instruments that had less than one vacuum tube, let
alone one transistor. The need to do all kinds of
work gave us a broad view of the ocean such that
we were oceanographers and not just marine
geologists. We even thought we understood physical
chemical, and biological oceanography. Working on
the low-freeboard E. W. Scripps with decks awash
gave an intimate feel for the oceans such as
experienced today only by scuba divers.
As we write this note Charley Hollister is
putting out to sea with his "Super Straw," the giant
4'/i-inch coring device, and a new generation of
marine sedimentologists. They will study complex
seabed forms, subbottom acoustically reflecting
layers, and mass physical properties of muds. In
this work they will be guided by the multisensor
MPL Deep-Tow, a real-life dream machine. All in
all, a million-dollar effort. Yes, times have
changed-and for the better.
R. S. Dietz is a research oceanographer at the National
Oceanographic and Atmospheric Administration, Miami,
Florida. K. O. Emery is Henry Bryant Bigelow
Oceanographer at Woods Hole Oceanographic Institution.
Dietz with sediment trap and Emery and Shepard with wire sounding machine
setting one for the survey of La Jolla Submarine Canyon, November 1 938.
Ill
28
Reprinted from: Geoloqy , Vol. 4, No. 7, 391-392.
El'gygtgyn: Probably world's largest
meteorite crater
+<*'
»*^^*
o
KM e
10
15
20
IE
EE
Figure 1. LANDSAT images of the Siberian meteorite crater
El'gygytgyn. Remarkable circularity of feature is revealed in upper snow-
covered winter view (dark spot near center is cloud shadow, not island).
Lower scene shows central lake in ice-free summer conditions.
GEOLOGY, v. 4, p. 391-392
Robert S. Dietz
National Oceanic and Atmospheric Administration
Miami, Florida 33149
John F. McHone
Department of Geology, University of Illinois
Urbana, Illinois 61801
ABSTRACT
LANDSAT imagery indicates that El'gygytgyn in northern
Siberia is probably a giant meteorite crater, the largest Quater-
nary impact structure on Earth, and not a tectonic depression.
This probability is supported by the remarkable circularity of
the crater, as outlined by (he ring-mountain rim, remoteness
from modern volcanic sites, and lack of collapse scalloping
of the margin.
Siberia appears to have an unusual attraction for cosmic
bodies. The Tunguska comet head exploded above central Siberia
in 1908, and the Sikhote-Alin nickel-iron meteorite struck eastern
Siberia in 1947. A million or so years earlier, Siberia probably
was the site of the largest crater-forming meteorite impact to
strike the continents in modern times. We refer to the El'gygytgyn
crater (sometimes transliterated El'gytkhyn; lat 67°30'N,
long 172°00'E), in the remote Anadyr Mountains of eastern
Siberia; this crater, by morphologic criteria, appears to be a
meteorite impact site. Its 18-km diameter would make it by far
the largest meteorite crater on Earth, far exceeding both the Lake
Bosumtwe crater in Ghana, 10.5 km across, and the New Quebec
crater of Canada, 3.2 km across. There are, of course, larger
ancient impact sites or astroblemes, but these are now so deeply
eroded that they are the roots of craters that are no longer
craterform.
El'gygytgyn has been previously listed as a possible impact
site by Dence (1972), citing Zotkin and Tsvetkov (1970), who
listed the diameter as only 12 km, which is approximately that
of the rudely circular lake occupying the center of this depression.
LANDSAT imagery (Fig. 1) reveals the remarkable circularity,
symmetry, and elevated rim of the overall crater, 18 km across.
This circularity is easily overlooked on maps because of carto-
graphic emphasis of the lake shoreline, as large sediment aprons
have filled parts of the crater (Fig. 1, lower) and produced an
irregular form.
391
278
KM 0 5 10
3_
a:
Figure 2. LANDSAT image of Crater Lake, Oregon, showing
scalloped walls, central volcanic island, and asymmetric perimeter typical
of calderas.
El'gygytgyn is a unique feature of the maturely dissected and
nonglaciated Anadyr mountainland. The nearest active volcanoes
lie 600 km away on the Kamchatka Peninsula. This craterform
depression is asymmetrically filled with a 170-m-deep lake about
12 km across and of squarish outline. The crater is outlined by a
ring mountain that attains its greatest relief to the west. The rim
is breached by a river in the southeast quadrant, the outflow of
which eventually reaches the Belaya River of the Pacific watershed.
The highest elevations along the rim are about 1,060 m, according
to the U.S. Air Force operational navigational chart. This rim
thus stands about 450 m above lake level, or 620 m above the
lake bottom. Extensive talus aprons along the western half of the
crater are now being entrenched, suggesting that the lake level
was once higher than at present. The almost perfect circularity
of the depression is enhanced in the winter LANDSAT image
because snow masks the outline of the lake shore. The mature
degree of erosion suggests that the crater was created one to a
few million years ago.
El'gygytgyn was discovered in 1933 by S. V. Obruchev from
an aircraft, according to Nekrasov and Raudonis (1973). The lake-
filled crater immediately attracted attention because of its unusual
shape. Obruchev expressed the opinion that it was a volcanic
crater or caldera of vast dimensions, yet there are no young vol-
canic rocks associated with the feature. Further negative evidence
is provided by LANDSAT images of calderas that are quite unlike
El'gygytgyn. Figure 2, for example, is a LANDSAT image of
Crater Lake, Oregon, one of the world's best examples of a
caldera. Crater Lake is situated on top of a large volcanic dome,
and although it is rudely round, its circularity is spoiled by its
scalloped margin. This was created by land slippage, subsidence
associated with magma withdrawal, and subsequent internal
evisceration by ash eruptions. A caldera has been aptly described
as "a volcanic crater whose head has fallen in when its insides
were blown out." El'gygytgyn does not have the geomorphic
aspect of a caldera, which is invariably situated atop a large vol-
canic come and which occurs in chains or groups and not as a
solitary feature. El'gygytgyn is also larger than most calderas
and more symmetrical.
Nekrasov and Raudonis (1973) have described El'gygytgyn
as a collapse feature of unspecified origin. One must assume
that a subsidence of this magnitude would necessarily be tectonic.
Such an origin, however, would not account for the ring of moun-
tains unless the structure underwent domal uplift by injection
of magma prior to collapse. In this event, the remarkable circu-
larity would remain unexplained. Nekrasov and Raudonis (1973)
studied eight rock samples collected from the north and northeast
parts of the ring mountain and found them to be an assortment
of silicic, intermediate, and mafic igneous rocks, including both
intrusive and extrusive types, of probable Mesozoic age. These
specimens appear to be country rocks rather than products of
the crater-forming event. Nekrasov and Raudonis concluded that
the crater could not be an impact site because they detected no
coesite in thin sections. This conclusion is unjustified, as coesite
is virtually unrecognizable in thin section, and, in any event,
shock overpressures at an impact-crater rim are already far below
those needed to create this high-pressure silica polymorph.
In general, shock metamorphism and shatter coning are never
found in situ beyond one-half of the radius of an impact crater
from ground zero.
We conclude that El'gygytgyn is probably the world's
largest modern impact crater.
REFERENCES CITED
Dence, M., 1 972, The nature and significances of terrestrial impact
structures: Internat. Geol. Cong., 24th, Montreal 1972, Hroc,
sec. 15, 1'lanetology, p. 77-89.
Nekrasov, I., and Raudonis, P., 1973, Meteorite craters (translation from
Russian ms.): Ottawa, Canada. Canadian Translation Bureau.
Zotkin, I. T., and Tsvetkov, V. I.. 1970. Searches for meteorite craters
on earth: Astron. Vestnik, v. 4, p. 55-65.
ACKNOWLEDGMENTS
Reviewed by Peter Rona.
The work of John Mcllone was supported by a grant-in-aid for
meteoritic research from the Barringer Crater Company.
MANUSCRIPT RECEIVED MARCH 22, 1976
MANUSCRIPT ACCEPTED APRIL 27, 1976
392
JULY 1976
279
29
Reprinted from: Proa. American Society of Civil Engineers Specialty
Conference on Dredging and Its Environmental Effects, Mobile, Al . ,
26-28 January 1976, 936-946.
DEPOSITION AND EROSION IN THE DREDGE SPGI- AND
OTHER NEW YORK BIGHT DUMPING AREAS
By: Ge&rge L. Freeland' and George F. Merrill'
INTRODUCTION
The disposal of solid wastes from the New York C'ty metropolitan
area is the cause of considerable environmental conce'"", as most of these
wastes are dumped in marine waters outside of the harrcr mouth (Table 1).
Dredge spoil and sewage sludge constitute over 94" of the volume of
material dumped containing solids.
In 1973 the National Oceanic and Atmospheric Administration (NOAA)-,
under the Marine Eco Systems Analysis (MFISA) Project, initiated research to
determine the effect of dumping in the New York Bight. A new hydrographic
survey of the Bight apex was immediately started to determine what changes
had occurred in bottom topography since the last previous survey in 193&.
Some results from this survey are presented here.
HYDROGRAPHIC SURVEYS
Hydrographic surveys have been made in the New York Bight since
1845 by the U.S. Coast and Geodetic survey (now the National Ocean
Survey, NOS, part of NOAA). Trie last U.S. C.&G.S. survey to cover the
Bight apex, the area immediately adjacent to the harbor mouth where dumping
is most intense, was in 1936. Comparison with the 1S45 revealed the
development of several knolls due to early dumping (4).
Our 1973 survey had depth sounding lines spaced 1000 ft (305 m) apart
over an area approximately 15 nautical miles (28 km) square (Fig. I).
Data from this survey were then compared with data from the 1935 survey to
produce a net-change map.
NET BATHYMETRIC CHANGE
Examination of the boat sheets (detailed maps showing final data plots)
from the 1S36 survey (NOS No. H-6193) revealed trackline spacing of approx-
imately 0.5 nautical miles (900 m) versus 1000 ft (305 m) spacing for the
1973 survey, and divergence of trackline directions. In order to compare
the two surveys, boat sheets from both surveys were cqntoured on a 3 ft.
(0.92 m) contour interval (see Fig. 1 for the 1973 map). A 1000 ft. (305 m)
grid was then prepared for the entire area and overlaid on both naps. From
TT National Oceanic and Atmospheric Administration, Atlantic Oceanographic
and Meteorological Laboratories, 15 Rickenbacker Causeway, Miami, Florida
33149
l>36
280
Nl W YORK I'.iCHl
93
Fig. 1. 1973 Bathymetric map of the Mew York' Bight apex. From a NOAA survey.
Contour interval one meter.
281
938 DREDGING EFFECTS
plotted data on the boat sheets and interpolations between contour lines,
a value was picked for the center of each 1000 ft (305 m) square for each
survey. These numbers were added algebraically to produce a positive or
negative number for each square representing erosion or deposition "'n that
square for the 37 years between surveys. The values were corrected for a
sea level rise of 0.62 ft (0.189 m) from N0S mca monthly sea levels at
Sandy Hook, N.J. They 'were then contoured to produce the net-change map
(see Fig. 3).
Volumes of erosion and deposition were calculated by pi animetering
all contours and multiplying these areas by-the appropriate contour inter-
val. Appropriate voluir.es were added for slope sediment between contours.
Areas, volumes of erosion and deposition, and net changes for Bight
apex features are listed in Table 2.
DISCUSSION: ANTHROPOGENIC SEDIMENTS
Sediments are introduced into the Bight apex almost entirely in the
form of fine-g-ained matter. Dredge spoil constitutes the most important
source of sol'ds brought in by man (anthropogenic sediment) (Table 1 ) .■
Estimates of the total amounts of dredgings ba-ged from 1936 to 1973
(records are unreliable prior to 1954) indicate that about 136 x 10° cu.
yd. (142 x 10D m3) were dumped, compared to 162 x. 106 cu yd (124 x 106 r:3)
calculated on the basis of net bathyinc-tric change for the dredne sooil
duinpsite and the dumping areas near Ambrose and Sandy Hook Channels
(Table 2). This indicates that approximately 87% of the material barged
is still in place on the bottom.
Detailed mapping of the dredge spoil dumpsite shown that shoaling of
up to 34 ft. (10.36 m) has occured over an area of 11 square nautical miles
(36. Km?) south of a knoll which itself was formed by earlier dumping
(Figs. 4-6).
udge is appa,
easily resuspended and dispersed, as only traces of sludge can be found
at the designated site. An unknown fraction settles in the Cnristiaensen
Basin and the upper Hudson Shelf Valley where it mixes with natural muds,
the remainder beifj dispersed by the water column. For this reason, the
sludge dumpsite is not listed in Table 2. Differentiation of sludge from
natural muds, mostly by chemical means, is currently undergoing intensive
study by NOAA. Although the volume of sludge barged is considerable, and
will increase in the future as more plants come on line in the New York
area and percentage treatment improves, solids by weight do not constitute
an important addition to sediment volume on the bottom (Table 1). Contami-
nation of bottom sediment and sediment suspended in the water column by
282
NLW YORK BIGHT
43V
Source of Solids Transported into Marine «ate
of the New York Bight
SOURCE
, VOLUME
10 ,Cu,yd/yr1 .. of
(10 m) J barqod
c WEIGHT
10 .-short tons/yr
(10 metric tons)
of
barged
. ' of
total input
Dredge spoil
8.35
(6.38)
62.4
5.21
(4.73)
85.8
1
Sewage sludge
(3.27)
32
0.2C3
(0.184)
3.3
2.1
Cellar dirt
O./fc
(0.5?)
b. ,'
0.55;'
(0.r,S,
l,.8
5.S
Total Earned
13. 39
(10.24)
100. i
(hh
Atirosoheric
0.403
1
e.i
1.0.447:
Wastewater*
Municipal
0.39
i (0.35)
4.0
Industrial
| U.2
(0.2)
0.2
Runoff*
Gaged
1.5
(1.4)
1 16
Urban
1.2 12
(i.D ! ;
Total input
9.6C
(8.76)
100. 2
From Ref. 3
* 98' of these coastal zone inputs come through the Pockaway - Sandy Hook
Figur1:. do not include shelf-derived sediment from outside the Bight.
TABLE 2
Volumes of Erosion and Deposition in the New York Bight Apex
between 1936 and 1973
Area
nm^
(Km?)
Volume
106 cu yd. (106 m3)
Erosion
Deposition 1 Net Change
1. Entire Apex
209
(718)
182
(140)
212 29 D
(162) | (?7)
i.
TJredge Spoil Oumpsite
11
(36)
122 122 0
(93) , (<)<)
T
Cellar Dirt Dumps i te
(8)
6 6 0
77
Ambrose 4 Sandy Hook
Channels
25
(86)
63
(48)
40 ,_ E
Ml) ;i-'
5 .
I Anthropogenic
38
.(130)
63
(48)
ies io6 ■.'
(1?1) (8!)
57
Christiaensen Basin
(83)
13
(im
8 5 E
(61 ; (4i
/.
Hudson Shelf Valley'
7
(23)
12
(10)
3( IC E
ii.
v. non-anthropoqenic
171
(507)
120
(9!)
43 76 C-
1. Area between 14 and 20 fathoms (26-37 m) north of 4024'N.
2. Area deeper than 20 fathoms (37 m) north of 40'19.22'N.
3. Equal to a layer 0.106 in. (2.7 mm) thick per yen'.
Some figures may not agree due tc rounJing off.
283
940
DREDGING EFFECTS
73°45'
40°35'
40°30'
40°25'
40°20' -
73°40'
_L
_L
l
I
_l_
74°00' 73855" 73°50' 73°45' 73°40'
40°35'
40° 30'
40°25
40820'
Fig. 2. Tracklines of the 1973 bathymetric survey. Light lines show track-
lines for bathymetry only. On heavy lines both bathymetric and geo-
physical data were collected.
284
M W YORK BIGHT
94 i
BATHYMETRIC NET CHANGE
1936-1973
74°00' 73°55' 73°50' 73°45'
73°40'
74°00' 73°55'
73°50
73°45
73°40'
DEPOSITION 0-6 FT
DEPOSITION >6FT
| EROSION 0-2 FT
I EROSION >2FT
Fig. 3. Net bsthymetric change, N.Y. Bight Apex, from 1936 to 1973.
235
9a:
DREDGING EFFECTS
Fig. 4. N.Y. Bight dredge spoil dumpsite. 1936 Bathymetry. Line marked 198°T
and hachured area show the designated dumpsite (Figures 4-6) based on 1936
soundings to lie within the 90 ft. isobath.
286
NEW YORK BIGHT
94?
JJJ^tJ »»"«««]|; i j j (aa 8 otitis J» nS5» J*^** gO
„„„» ,.„,r-K..,-. ..S8'- ...•■•v^:t[;.;; '*$$« : * c e I i
nmiimiiiliiii'M"'""6, 1
50' IMftitiggfS
NEW YORK BIGHT APEX
DREDGE SPOIL DUMPSITE
1973 BATHYMETRY
CONTOUR INTERVAL 5 FEET
0.L. FREELAND NOAA AOML 2-74
Fig. 5. N.Y. Bight dredge spoil dumpsite. 1973 Bathymetry.
287
944
DREDGING EFFECTS
I
73*52'
73°51'
73" 'so' 40«26"
DREDGE SPOIL DUMPSITE, N.Y BIGHT
NET CHANGE MAP
1936-1973
CONTOUR INTERVAL 5 FEET
G.L. FREELAND 2/74 NOAA-AOML
73°50'
I
40"25-
40«24-
3
40«23'—
40°22 •
Fig. 6. N.Y. Bight dredge spoil dumpsite. Net change in depth from 1936 to
1973. Note that the 47 ft. knoll in the 1936 map is essentially un-
changed, and that a large volume of material has been dumped north
of the start (northeast end) of the arrow designating the minimum
distance (4 nm) from Ambrose Light that dumping should be initiated.
288
N1W YORK MICH ! 94!
organic pesticides and heavy metals in sludge is, however, a serio r, concern.
Cellar dirt, the third anthropogenic sediment, consists of c ■ r.tion
rubble from demolition, foundation rock and dirt, and slag. Brie- .
norphic rocks, and red sandstone are commonly recovered in grab S3 ;• - ; .
Cellar dirt, while making a recognizable spoil mound, is not considered
an important pollutant because of low volumes and the absence of toxic
chemicals.
YATURAl SEDIMENTS
Natural sediment input from land sources comes mainly from stream
runoff from the Hudson River drainage basin and urban runoff from the New
York mecropol itan area (Table 1). These sources are relatively easy to
-easure compared to sediment transported from other areas of the shelf.
Various estimates of sediment transport indicate that, for the eastern
U.S. continental nwgin, a) 90% of the sediment from land sources is
deposited in estuaries and wetlands; b) net suspended fine sediment transport on
the shelf is probably landward, with possibly much of the material finally settling
in estuaries; and c) recycling (resuspension and settling) of sediinent on
the shelf may transport orders of magnitude more sediment than either
enters or leaves the shelf (1, 2).
From our net change map, we have calculated volumes of natural sediment eroded
and deposited in the Bight apex (Table 2). After subtracting the anthropogenic
naterial in the Ambrose - Sandy Hook Channels area and in the dredge spoil and
cellar dirt dumpsites, the volume of material eroded exceeds deposition by
76 x 106 cu yd (58 x 1 0& m^), equivalent to a layer 0.106 in. (2.7 mm) thick
over the non-anthropogenic areas. From other ongoing studies, it appears
that this erosion occurs primarily during storms which pass most frequently
in winter months.
ERROR SOURCES
Tidal corrections from the Sandy Hook tidegage were made on all records
and are not considered to be a significant error source.
Bar checks for fathometer error were made before and after daily operations
at sea. Surveying was not done or was rerun if fathometer readings were
incorrect.
After contouring W3S completed, data from some small areas in the northern
Christjaensen Basin became suspect because of sinusoidal wiggles in contour-
lines which varied systematically with tracklines, amounting to a maximum
of about 1.5 ft. (0.457 m) difference between adjacent tracklines. After
reruns of bathymetry lines in an east and west direction (perpendicular to
the original tracklines), it was determined that the "waves" originally mapped
in the bottom were real, although of lower amplitude, in one area, and non-
existant in other areas where the original waves were lower than maximum
amplitude. Further analysis indicated the source of error to be "squat" of the
289
9-56 DR! DGING 1ITK TS
survey boat (level cf ride of the boat in the water) on geophysical tracklir.es
in certain areas due to towed instruments. These errors were corrected in
the final contour nap and at least partially compensated for in the net chancie
map. Maximum error is estimated to be + 0.5 ft. (0.1524 m).
ACKNOWLEDGEMENTS
Grateful appreciation is given to the Corps of Engineers Operations
Section of the new York District, contractor for the 1973 survey: Mr. Lewis
Pinata, Acting Chief, Mr. Dennis Suszkowski , Oceanographer, Mr. Herbert w--':-
and Mr. Bill Musak, Chief and Assistant Chief of the Survey Branch, and :■ -:•
crew of the survey vessel HATTO'I , which did tne northern half of the survey.
Thanks are also given to Mr. Robert Spies, and Mr. Robert Wagner, Chief and
Assistant Chief of the Survey Branch of the Corps Philadelphia District, and
to the crew of the survey vessel SHUMAN, which did the southern part of the
survey. Mr. George Lapiene, electronic technician at AOML in 1973, was aboard
both survey vessels during the three months of work on geophysical track! ines
Dr. Anthony E. Cok, Professor of Geology at Ade 1 phi University, Garden City,
New York, was aboard the HATTON during geophysical trac'kline work. Dr. John
J. Dowling, Associate Professor, Marine Sciences Institute, University of
Connecticut, Groton, Connecticut, made the net ciiange calculations. The
Tidal Datum Planes Section, Oceanographic Division, of the national Ocean
Survey (NOAA), Rockville, Maryland, supplied monthly mean sea level data for
the Sand Hook Station for 1936 and 1973. Cdr. R. L. Swanson, Project
Manager, MESA New York Bight Project, Stony Brook, New York, provided
additional tidal correction data. Finally, Drs. H. B. Stewart, Jr., and
D. J. P. Swift of AOML reviewed the manuscript.
REFERENCES CITED
1. Meade, R.H., Sachs, P.L., Manheim, F.T., Hathaway, J.C., and Spencer, D.W.,
"Sources of Suspended Matter in Waters of the Middle Atlantic Bight",
Journal of Sedimentary Petrology, Vol. 45, 1975, pp. 171-188.
2. Milliman, J.D., Pilkey, O.H., and Ross, D.A., "Sediments of the Continental
Margin of the Eastern U.S.", Bulletin of the Geological Society of
America, Vol. 83, 1972, pp. 1315-1334.
3. Mueller, J. A., Anderson, A.R., and Jen's, J.S.,( "Contaminants Entering the
New York Bight - Sources, Mass Loads, Significance", Report sent to
NOAA, MESA N.Y. Bight Project Office, Stony Brook, N.Y., 11794, 1975.
4. Williams, S.J., and Duane, D.B., "Geomorphology and Sediments of the
Inner New York Bight Continental Shelf", Technical Memorandum 45,
U.S. Army, Corps, of Engineers, Coastal Engineerina Research Center,
July, 1974.
290
30
Reprinted from: Middle Atlantic Shelf and the New York Bight, ASLO Special
Symposia, Volume 2, 90-101.
Surficial sediments of the NOAA-MESA study areas in the New York Bight
George L. Freeland, Donald J. P. Swift, and William L. Stubblefield
Atlantic Oceanographic and Meteorological Laboratories, NOAA, 15 Rickenbacker Causeway, Miami,
Florida 33149
Anthony E. Cok
Department of Earth Sciences, Adelphi University, Garden City, New York 11530
Abstract
In the New York Bight apex, extensive sedimentological studies and a 1973 bathymetric
survey reveal that the only significant change in bottom topography since 1936 occurred
at the dredge spoil dumpsite where the dumping of 98 X 10fi m3 of dredged material has
caused up to 10 m of shoaling. The center of the Christiaensen Basin, a natural collecting
area for fine-grained sediment, is no doubt contaminated with sludge but shows no apparent
sediment buildup during the intervening 37 years. The apex outside of the Christiaensen
Basin is floored primarily by sand ranging from silty fine to coarse, with small areas of
sandy gravel, artifact (anthropogenic) gravel, and mud. Nearshore mud patches appear to
be covered at times with sand and occasionally scoured out. Sidescan sonar records show
linear bedforms, indicative of sand movement, over most of the apex area.
Two midshelf areas have been proposed as interim alternative dumping areas. The
northern area is in a tributary valley of the ancestral Long Island river system. Fine sands
cover the northeast part and medium sands predominate to the west and south. Bottom
photographs show a smooth, slightly undulatory, mounded or rippled sea floor.
In the southern alternative dumping area coarse sand and gravel deposits lie on the crest
and east flank of the Hudson divide, while medium and fine sand occurs in the ridge and
swale topography to the west. These distributions suggest fine sediment is winnowed from
the crest and east flank of the divide and deposited to the west. Veatch and Smith Trough
contains a veneer of shelly, pebble sand with large, angular clay pebbles and occasional
oyster shells derived from exposed early Holocene lagoonal clay. These studies suggest that
if sewage sludge were dumped, widespread dispersion, mostly to the southwest, could be
expected, with winter resuspension and transport of fine material on the bottom. Possible
permanent buildup on the bottom could be expected if dredged material were dumped.
The nature of bottom sediments and sed- maps at 1 -fathom ( Stearns and Garrison
iment particles suspended in the water col- 1967 ) and 4-m intervals on the shelf and
umn becomes of interest to environmental 200-m intervals on the continental slope
managers when man's activities in the ocean (Fig. 1; LIchupi 1970) were made from
disturb the sea floor or the near-bottom 1936 survey data. A new survey of the bight
water column. In addition to the immediate was made in 1975; results should be avail-
results, one must also consider the effect on able in 1977.
long term natural phenomena. How are Surficial morphology of the New York
these processes affected by what man has Bight, and sediment distribution across this
done, or perhaps more importantly, how do surface, may be explained by sea level flue-
natural processes modify what man has tuations caused by continental glaciation
done to disturb the natural environment? during the past several million years. The
Here we report work done at the Atlantic last glacial stage ended 15,000 years ago
Oceanographic and Meteorological Labora- (Milliman and Emery 1968) when the
tory as part of the NOAA-MESA New York eastern North American ice sheet extended
Bight Project. as far as Long Island and northern New
Hydrographic surveys of the New York Jersey. During maximum glacial advance
Bight were initiated in 1936 by the Coast sea level was lowered about 160 m ( Veatch
and Geodetic Survey (now the National and Smith 1939) so that the shoreline of the
Ocean Survey) in nearshore areas and have bight was in the vicinity of Hudson Canyon
been repeated periodically. Bathymetric ( see Fig. 1 ) . Since the ice melted, the shore-
AM. SOC. LIMNOL. OCEANOGR. 90 SPEC. SYMP. 2
291
Surficial sediments
91
Fig. 1, Index to detailed study areas and topo-
graphic features in the New York Bight. ( Bathym-
etry from Uchupi 1970. ) Contour intervals 4 and
200 m. 1A — New Jersey nearshore ridge and swale
study area and the Atlantic generating station site;
IB — -New Jersey central shelf ridge and swale study
area; LINS — Long Island nearshore study area;
SCOA— Suffolk County outfall area; 2D1, 2D2—
proposed interim alternative dumpsites.
line advanced to its present position; many-
features of the shelf are the result of sev-
eral sea level fluctuations. Morphologic fea-
tures are discussed in our companion paper
in this volume (Swift et al. 1976) and else-
where (e.g. McKinney and Friedman 1970;
McKinnev et al. 1974; Stubblefield et al.
1974, 1975; Knott and Hoskins 1968; Duane
etal. 1972; Williams 1976).
Surficial sediments
A comprehensive sampling program for
the outer shelf was conducted by the Woods
Hole Oceanographic Institution and the
U.S. Geological Survey, who sampled on
an 18-km spacing. The Corps of Engineers
Coastal Engineering Research Center has
collected about 4,200 km of geophysical
data and over 300 cores as a part of its stud-
ies on the inner shelf of the bight ( Duane
1969; Williams and Duane 1974; Williams
1976). MESA work had been conducted
primarily in New Jersey nearshore and cen-
tral shelf areas, the bight apex, the near-
shore of Long Island eastward to Fire Is-
land, two central shelf alternative dumping
areas, and the Hudson Shelf Valley ( Fig.
1 ) . Emphasis here is on the bight apex and
the central shelf alternative dumping areas.
Source and age of sediments — Sediments
covering the floor of the bight were mostly
deposited during lowered sea level and were
reworked during the landward-seaward
migrations of the shoreline. As transgression
progressed, fluvial and older sediments
were covered by estuarine and lagoonal
sediments behind barrier islands or directly
reworked by littoral processes associated
with the advancing shoreline. During a
transgression, bottom currents of the inner
shelf interact with the shelf floor to form a
concave surface whose profile resembles an
exponential curve, with the steep limb com-
prising the shoreface (Swift et al. 1972).
With a loose, sandy substrate, the inner
shelf shoreface tends to extend itself later-
ally across the mouths of bays, closing them,
except for inlets, by the deposition of sand
in the form of spits and barrier islands.
Estuaries and lagoons behind these spits
and islands then trap suspended fine sedi-
ment (mud), while the barrier islands are
nourished by littoral drift from eroding
headlands and by sand moving landward
from the shelf.
As sea level rose during the Holocene
transgression, the inner shelf profile moved
shoreward by means of shoreface erosion.
Some eroded sand was swept onto the bar-
rier islands by storm overwash and buried,
only to be re-exposed again at the eroding
shoreface. Most material from shoreface
erosion, has, however, been washed down-
coast and seaward to form a discontinuous
sand blanket 0 to 10 m thick (Stahl et al.
1974 ) . Thus, the New York Bight shelf floor
is dominantly sand-sized sediment (Schlee
1973). Fine-grained sediments are gener-
ally absent, having been transported either
into the Hudson-Raritan estuary, behind
barrier islands, or off the shelf edge. Lo-
cally, underlying strata of transgressed la-
goonal and estuarine semiconsolidated mud
deposits or resistant coastal plain strata are
exposed on the sea floor (Swift et al. 1972;
Stahl et al. 1974; Sheridan et al. 1974).
Sediment types — Sediment types have
been mapped in the New York Bight pri-
marily by dominant grain size (Fig. 2).
292
92
Geological processes
Generally, the shelf is covered by sand-
sized sediment with isolated gravel patches
(Schlee 1973, 1975; Williams and Duane
1974; Williams 1976). In deeper water, gen-
erally seaward of the 60-m isobath, in the
Hudson Shelf Valley, and in lagoons and
estuaries where wave action is less pro-
nounced, silt is the dominant sediment
( Freeland and Swift in press ) . In the Long
Island nearshore zone west of Fire Island,
small mud patches, some of which are sea-
sonal, are of considerable environmental
concern owing to contamination of the fines
by pollutants.
Suspended sediments — Meade ( 1972a,
/;) noted the following: Pleistocene glacia-
tions and sea level fluctuations drastically
altered the composition and distribution of
sediments on continental margins; it is not
always immediately evident whether pres-
ent shelf deposits reflect modern or Pleisto-
cene conditions. Fine sediment transport
studies are hindered by the fact that de-
posited sediments may reflect processes act-
ing over thousands of years, whereas our
Fig. 2. Sediment types in the bight area ( depth
in meters). Hatching — gravel, sandy gravel, and
gravelly sand; speckling — sand; stippling — silty
sand, sandy silt, and clayey silt; dappling — glau-
conitic sand, silty sand, and sandy silt. ■■■ — Pyrite-
filled foraminiferal tests. 1 — Zone of rounded
quartz grains; 2 — zone of limonitic pellets. ( From
Uchupi 1963.)
Table 1. Source of suspended solids in the New
York Bight.*
XlO0 tonnes /yr
Direct bight ( 68% )
Dredged (54%) 4.73
Sludge (2.1%) 0.18
Cellar dirt (6.8%) 0.60
Total barged (62.9%) 5.51
Atmospheric (5% ) 0.45
Coastal zone (32%)
(98% of coastal zone input is through the Rock-
away-Sandy Hook transect)
Municipal wastewater (4% ) 0.35
Industrial wastewater ( 0.2 % ) 0.02
Gauged runoff (16%) 1.4
Urban runoff (12%) 1.1
Total coastal zone
Total input
2.87
8.83
* From Mueller et al. 1976.
studies of suspended sediment are com-
monly limited to a few days or months of
observations. Natural processes may be im-
possible to separate from the changes pro-
duced by human activities, particularly in
estuaries ( and at the present dumpsites ) .
Fine sediment sources to estuaries and
the shelf — Fine sediment discharged into
the bight is shown in Table 1 ( Mueller et al.
1976). Fluvial sediment is comprised of
roughly 85% inorganic and 15% combust-
ible organic material (Table 2). The fine
inorganic fraction is mostly illite, chlorite,
feldspar, and hornblende from the Hudson
River ( Hathaway 1972 ) .
Shelf erosion and coast-parallel transport
appear to be significant but unmeasured
sources of suspended material and were
probably major sources during the Holo-
cene transgression. Hathaway (1972)
showed that fine sediments near the mouths
Table 2. Composition of suspended matter.
Rivers 80-90% minerals
Estuaries 60-80% minerals
+ biogenic shells
Shelf 10-70% minerals
+ biogenic shells
10-20%
combustible
organics
20-40%
combustible
organics
30-90%
combustible
organics
293
Surficial sediments
93
of coastal plain estuaries differ significantly
from the composition of overborne sedi-
ments. It is probable that much estuary-
mouth sediment is being eroded from shelf
deposits and returned to and trapped in
estuaries (Meade 1969). The fact that the
sediments from modern rivers have not
obscured this conclusion implies that either
the modern sediment is bypassing the lower
portions of the estuary, or it is trapped al-
most completely near the river mouths.
Along the east coast, the heads of the Ches-
apeake and Delaware estuaries are far up-
stream from the estuary mouth, therefore,
most river sediment is deposited far inland
from the sea. Although saline tidal water is
present in the Hudson River up to Albany,
fine fluvial sediment is carried by low-
salinity surface water to Upper and Lower
New York Bays where some fines settle out
( Folger 1972/; ) and the remainder is car-
ried with estuarine sediment into the bight
apex and mixed with recirculated shelf sed-
iment. In the northeast United States, most
of the fluvial suspended sediment is effec-
tively trapped in estuaries and coastal wet-
lands (Millimanl972).
At the present, the annual suspended
sediment discharge of Atlantic coastal rivers
is about equal to the annual deposition on
marsh surfaces (Meade 1972a). However,
much of the deposited material re-enters
the shelf water column after the shoreline
has passed over the marsh, through the
process of shoreface erosion ( Fischer 1961 ) .
Particles derived from biologic processes
are also a significant component of sus-
pended matter in estuaries and on the shelf
(Table 2), ranging from 20-90% in surface
waters ( Manheim et al. 1970). However,
concentrations of combustible biogenic mat-
ter decrease rapidly with depth, and little
of this material is preserved in sediment de-
posits (Folger 1972a; Gross 1972).
Atmospheric fallout over the New York
Bight is small relative to other sediment
sources (Table 1), but it may be a signifi-
cant transport path for specific pollutants
(e.g. lead from vehicular exhaust emis-
sions ) .
Highest concentrations of organic and in-
organic suspended materials in the water
occur within 10 km of the coastline and de-
crease nearly exponentially seaward ( Man-
heim et al. 1970). Mineral grains larger than
4 fim (silt-size) comprise 10-25% of near-
shore suspended sediment and only 2-5%
of offshore samples; the remainder is or-
ganic matter. The zone of strong terrige-
nous influence is restricted to nearshore
waters and, specifically, to the inner shelf
zone of turbid water drifting away from the
estuary mouth. The coarser grains in this
zone are effectively trapped in the "estua-
rine" circulation (which serves to reinforce
the surface concentrations) and are trans-
ferred from one estuary to the next along
the path of the longshore current.
Studies of other areas (Postma 1967) sug-
gest that volumes of suspended sediment
transported on the many feedback loops in
the bight are probably orders of magnitude
greater than both the net volume from the
Hudson River that is transported across the
shelf and the much larger amounts intro-
duced by dumping.
Although the factors which influence sus-
pended sediment dispersal can be readily
defined, many large gaps in our knowledge
must be closed before quantitative sediment
transport budgets can be constructed on a
regional scale. The most important of these
are: shelf circulation patterns and mecha-
nisms, particularly during storms; hydraulic
properties of suspended sediments, particu-
larly resuspension and settling properties;
and the influence of flocculation and bio-
logic aggregation on settling.
Detailed studies in the
New York Bi^ht apex
A 1973 bathymetric map (Fig. 3) of the
bight apex was made as the result of a
NOAA-Corps of Engineers survey. The
principal topographic features are the
northern end of the Hudson Shelf Valley,
Cholera Bank, and the Christiaensen Basin,
an amphitheaterlike feature terminating the
Hudson Shelf Valley (Veatch and Smith
1939). Dumpsites for dredge spoils (the
mud dump ) , cellar dirt, sewage sludge, and
acid wastes are shown. Knolls immediately
northwest of Ambrose Light and north and
northwest of the dredge spoil dumpsite
294
94
Geological processes
74*00'
Fig. 3.
73* 55' 73* 50' 73* 45' 73* 40'
Bathymetric map of the New York Bight apex. Contour interval, 1 m. Data ( in meters ) from
1973 NOAA-Corps of Engineers survey.
were formed from early 20th century dump-
ing of assorted building excavation material
and sand and gravel from the dredging of
Ambrose and Sandy Hook Channels (Wil-
liams 1975).
Comparison of the 1973 bathymetric sur-
vey results with data from the 1936 survey
reveals that only the anthropogenic areas
have changed significantly. Figure 4 shows
the 1973 and 1936 bathymetry of the dredge
spoil site, as well as the net change between
the two surveys. The 50-ft knoll on the 1936
map ( relatively unchanged in 1973 ) is itself
the result of earlier dumping (Williams
1975). The amount of anthropogenic ma-
terial accumulated during these years
(1936-1973) has been calculated to be
about 124 xlO6 m3. This compares with
about 142 xlO6 m3 dumped. The difference
easily can be accounted for by settling
alone.
Surficial sediments have been mapped by
analyzing over 700 bottom grab samples
collected at 1-km spacing (Fig. 5). The
topographically low Hudson Shelf Valley
and the Christiaensen Basin are floored
295
Surficial sediment.'}
95
SOlflS 90 95
—22' « i « '>^->80 "
DREDGE SPOIL DUMPSITE
1936 BATHYMETRY
j 95 isaa
DREDGE SPOIL DUMPSITE
1973 BATHYMETRY
-:,;;:■ 9* 100 115
vV»o ■;..;;.;;;:;:
li.lll!i:B'i!;;-l:..-1v..i;ii-l|-
NET CHANGE MAP
1936-1973
with fine-grained sediment, whereas the
rest or the area contains assorted sizes of
sand and both anthropogenic (artifact) and
natural gravel deposits. Artifact gravels
consist of recognizable construction rubble
■ — brick, schist, concrete, etc.
Geophysical data taken during the 1973
survey consisted of 3.5-kHz shallow-pene-
tration seismic reflection records and side-
Fig. 4. Bathymetric maps ( 5-ft contour inter-
vals ) of the dredge spoil dumpsite, New York Bight
apex. The 198"T azimuth (minimum distance 4
nmi from Ambrose Light ) and the 90-ft isobath
define the designated site (hatched). Upper left —
1936; upper right — 1973; left — net change from
1936-1973.
scan sonar records with 150-m range on
each side of 610-m-spaced tracklines. Al-
though data interpretation is incomplete,
bottom roughness patterns and trends of
linear bedforms (sand ribbons and de-
graded sand waves) have been mapped
from sidescan data (Fig. 6). These bed-
forms appear as alternating light and dark
bands corresponding to fine- and coarse-
296
96
Geological processes
40°30'N -
40°20'N
74°00'W
CONTOUR INTERVAL: 5fm
= MUD
Mill SILTY-FINE SANDS
73°50'W
73°40'W
FINE-MED. SANDS Xvl SANDY GRAVEL
COARSE SANDS jgggg ARTIFACT GRAVEL
Fig. 5. Distribution of surficial sediment based on visual sample examination. Bathymetry from
1936 data.
grained sediment or as isolated dark bands.
Streaky, patchy, and rough textures are as-
sociated with the dredge spoil and cellar
dirt dumpsites and may be related to indi-
vidual dumps.
Preliminary analysis of seismic data shows
filling of the Hudson Shelf Valley from
Cholera Bank.
Suspended sediment studies are particu-
larly important in the bight apex because of
the large amounts of fine particles dispersed
in the water by waste disposal operations.
These particles are in addition to the fine
sediments discharged from the Hudson
River, other river mouths, and tidal inlets
connected to coastal wetlands. Fine-grained
sediment is also eroded from the sea floor
during storms. Of immediate concern is
sewage sludge which contains bacterial,
viral, and heavy metal contaminants that
adhere to fine sediment particles in the
water column. The suspended fraction of
dredge spoils is also probably similarly con-
taminated. All of these fines are largely re-
297
Surficial sediments
97
PATCHES OF DEGRADED
SAND WAVES
LARGE, IRREGULAR
SAND RIBBONS
- STREAKY TEXTURE
PATCHY TEXTURE
£& ROUGH TEXTURE
Fig. 6. Distribution of bottom roughness pat-
terns from sidescan sonograplis. Blank area NW
and SE of Ambrose Light (A) shows no bedforms.
M — Dredge spoil dumpsite; CD — cellar dirt site;
SS — sewage sludge site.
tained in the nearshore water column as a
consequence of the bight circulation pat-
tern.
Suspended sediment studies were initi-
ated in the bight apex during 1973 when
sample stations were occupied to collect
chemical and physical oceanographic data.
Water samples were collected, filtered, and
examined from the surface, 10-m depth,
and the bottom at 25 stations. Preliminary
10 METEKSTOIAl SUSPENDED LOAD-mg/L
results for data taken in fall 1973 (Drake
1974; Figs. 7-10) indicate the existence of
a fair-weather, clockwise current-circula-
tion gyre, driven in part by the southwest
drift of offshore shelf water. This has been
verified by current meter studies in the
T^W
sm^
Fig. 7. Total suspended sediment load in waters
at 10-m depth, late November 1973. ( From Drake
1974.)
Fig. 8. Distribution of ferric hydroxide particles
in the water column in late November 1973 ( grains
X lOVliter). A — Surface; B — midwater; C — bot-
tom water. ( From Drake 1974. )
298
98
Geological processes
FERREL 2
SEP 16-20, 1973
Fig. 9. Vertical distribution of total suspended
load (in mg/liter) seaward of Long Beach, Long
Island. ( From D. E. Drake unpublished. )
apex (Charnell and Hansen 1974). Part of
the total suspended load in the bight apex is
easily identifiable, red-orange ferric hydrox-
ide particles. These particles are formed by
precipitation of iron in seawater as the re-
sult of acid waste dumping. They consti-
tute an excellent tracer of suspended sedi-
ment circulation. The vertical distribution
of suspended sediment shows high values
( 1.0 mg/liter) near the surface, and 2.0 mg/
liter in the near-bottom "nepheloid" layer,
typical of shelf areas (Fig. 9). It is expected
that this layer will transport much of the
suspended particulate matter and its asso-
ciated contaminants.
CtNEDALIZEO FINE SEDIMtNl T8ANSPQ8T FALL 1973
— 40*20' N
Fig. 10. Fine sediment transport system as in-
ferred from distribution of suspended sediments
during fall 1973. Dashed line is mean position of
boundary between more turbid coastal water and
less turbid offshore water. Clockwise gyre is ap-
parently driven by southwesterly drift of offshore
shelf water, and, on the bottom, by influx of saline
water into New York Harbor. Regional currents
which appear to be persistent are indicated by
solid arrows. ( From Drake 1974. )
Preliminary results show there is a con-
centration of fine-grained sediment in en-
closed lows in the Hudson Valley axis,
sandy mud in the remainder of the valley
axis, and coarser sediment up the flanks of
the valley and onto the shelf.
Alternative dumping area studies
Two midshelf areas have been designated
as possible interim alternative dumping
areas for sewage sludge and dredge spoils
from the New York metropolitan area (see
Fig. 1). The northern area is to be a mini-
mum of 46 km from the Long Island shore-
line, 18 km from the axis of the Hudson
Shelf Valley, and 120 km from the entrance
to New York Harbor. The southern area is
seaward of the 36-m isobath and the same
distances from the Hudson Valley axis and
the New York Harbor entrance as the
northern area (areas 2D1 and 2D2 on Fig.
1 ) . Each area is 18.5x18.5 km.
Northern area — In the northern area
( Fig. 1, 2D1 ), the sampling grid was placed
seaward of the center of the location-criteria
triangle to investigate, in part, a shallow
tributary valley of the ancestral Long Island
drainage system. The surficial sediments
consist of sand with some areas of over 5%
gravel ( Fig. 11 ). Fine sands lie in the north-
eastern part of the area, medium sands
cover the western and southern parts, with
a gravel deposit ( —39% gravel ) at one sta-
tion associated with an area of coarser med-
ium sand in the southern part of the area.
Only two stations contained >5% mud.
Bottom photographs indicate that the area
is characterized by a smooth, slightly un-
dulatory, mounded or rippled bottom. Side-
scan sonar records reveal elongate dark
areas which may be erosional windows in
the Holocene sand sheet that expose the
basal Holocene pebbly sand or may be
areas of abundant large shell fragments.
Grab samples were spaced too far apart to
be definitive. Bottom photo and submers-
ible-observation data support the existence
of windrows of shell fragments.
Southern area — The southern study area
in Fig. 1 (2D2) is centered over the broad,
flat high of the Hudson divide ( Fig. 12). To
299
Surficial sediments
99
72°50
72"45
MEDIUM SAND
.".•. 100-124*1
:':■'.'■ 1.23-1.49*/
I I 130-1.74 ♦ (
•£S: 175-199 * I
Mil > 200+ FINE SAND
Fig. 11. Northern proposed interim alternative
dumping area (2D1 on Fig. 1). Grain-size distri-
bution of sand-sized fraction. Large dots — sample
stations. ( Bathymetry from Stearns and Garrison
1967; 1-fm contour intervals.)
ing ridge and swale topography. Geophysi-
cal data, sediment samples, and two dives
in submersibles showed that grain-size pat-
terns appear to be related to bottom topog-
raphy; coarser sand and gravel deposits lie
on the crest and east flank of the Hudson
divide, while medium- and fine-grained
sand occur in the ridge and swale topog-
raphy (Fig. 13). These distributions sug-
gest that fine sediment is winnowed from
the crest and east flank of the divide and
deposited to the west. Observations from a
submersible in Veatch and Smith Trough
reveal a veneer of shelly, pebbly sand with
large, angular clay pebbles and occasional
oyster shells derived from the underlying
early Holocene lagoonal clay. Seismic data
also reveal that the reflector associated with
this surface, outcrops on the ridge flank. It
appears that storm-generated currents from
the northeast have winnowed the east flank
of the Hudson divide and formed or main-
tained the ridge and swale topography on
the west side of the divide.
the northeast the bottom grades gently into
the Hudson Shelf Valley, while the western
section is characterized by northeast-trend-
Fig. 12. Southern proposed interim alternative
dumping area (2D2 on Fig. 1). (Bathymetry from
Stearns and Garrison 1967, 1-fm contour intervals. )
Solid lines — geophysical tracklines; bars — sites of
dives by submersibles.
C
73°30' 73*25' 73*20' 73°15'
;.?S <100 + COARSE SAND
;.;•;■ 1.00-1.24+
••':;■'.• 125-1.49 +
I I 150-174 +
iSS 175-199 +
Mill >200+ FINE SAND
MEDIUM SAND
Fig. 13. Southern proposed interim alternative
dumping area, (2D2 on Fig. 1). Grain-size distri-
bution of sand-sized fraction. Large dots — sample
stations. (Only the 20-fm isobath is shown.)
300
100
Geological processes
Suspended sediment — As previously men-
tioned, most fluvial suspended sediment is
effectively trapped in estuaries and coastal
wetlands. Consequently, the terrigenous
fraction of the suspended matter decreases
rapidly seaward. Suspended solids through-
out the water column in the alternative
dumping areas were predominately plank-
ton and their noneomhustible remains. Total
suspended matter concentration in surface
water is from 100-500 fig/ liter, comprised
of 5% or less terrigenous matter, 80% com-
bustible matter, and 15% siliceous and cal-
careous noncombustible planktonic remains
(D. E. Drake personal communication).
Subsurface water-suspended matter con-
centration is similar or somewhat less, ex-
cept in the nepheloid layer 5-10 m above
bottom. There, suspended matter concen-
trations are 500-2,000 fig/ liter, consisting of
30-60% combustible matter and 50-80%
noncombustible matter which includes 10-
20% terrigenous matter. Textural proper-
ties of sediment deposits in the alternative
dumping areas show that very little sedi-
ment finer than 62 microns is present.
References
Charnell, R. L., and D. V. Hansen. 1974.
Summary and analysis of physical oceanogra-
phy data collected in the New York Bight apex
during 1969-1970. NOAA-MESA Rep. 74-3.
74 p.
Dhake, D. E. 1974. Suspended particulate mat-
ter in the New York Bight apex: September-
November 1973. NOAA Tech. Rep. ERL
318-MESA 1.
Duane, D. B. 1969. Sand inventory program. A
study of New Jersey and northern New En-
gland coastal waters. Shore Beach October.
, M. E. Field, E. P. Meisburber, 1). J.
Swift, and S. J. Williams. 1972. Linear
shoals on the Atlantic inner continental shelf,
Florida to Long Island, p. 447-498. In D. J.
Swift et al. [eds.], Shelf sediment transport:
Process and pattern. Dowden, Hutchinson &
Ross.
Fischer, A. G. 1961. Stratigraphic record of
transgressing seas in light of sedimentation on
Atlantic coast of New Jersey. Bull. Am.
Assoc. Pet. Geol. 45: 1656-1666.
Folger, D. W. 1972a. Texture and organic
carbon content of bottom sediment in some
estuaries of the United States, p. 391-408. In
B. W. Nelson [ed.], Environmental framework
of coastal plain estuaries. Geol. Soc. Am.
Mem. 133.
. 1972i>. Characteristics of estuarine sedi-
ments of the United States. U.S. Geol. Surv.
Prof. Pap. 742. 94 p.
Freeland, G. L., and D. J. Swift. In press. Sur-
ficial sediments. NOAA-MESA New York
Bight Atlas Monogr. 10.
Gross, M. G. 1972. Geologic aspects of waste
solids and marine waste deposits, New York
metropolitan region. Geol. Soc. Am. Bull. 83:
3163-3176.
Hathaway, J. C. 1972. Regional clay mineral
facies in estuaries and continental margin of
the U.S. East Coast, p. 293-316. In B. W.
Nelson [ed.], Environmental framework of
coastal plain estuaries. Geol. Soc. Am. Mem.
133.
Knott, S. T., and H. Hoskins. 1968. Evidence
of Pleistocene events in the structure of the
continental shelf of the northeastern U.S.
Mar. Geol. 6: 5-43.
McKinney, T. F., and G. M. Friedman. 1970.
Continental shelf sediments of Long Island,
N.Y. J. Sediment. Petrol. 40: 213-218.
, W. L. Stubblefield, and D. J. Swift.
1974. Large-scale current lineations on the
central New Jersey shelf: Investigations by
side-scan sonar. Mar. Geol. 17: 79-102.
Manheim, F. T., R. II. Meade, and G. C. Bond.
1970. Suspended matter in surface waters of
the Atlantic continental margin from Cape
Cod to the Florida Keys. Science 167: 371-
376.
Meade, R. H. 1969. Landward transport of
bottom sediments in estuaries of the Atlantic
Coastal plain. J. Sediment. Petrol. 39: 222-
234.
. 1972a. Transport and deposition of sedi-
ments in estuaries, p. 91-120. In B. W. Nel-
son [ed.], Environmental framework of coastal
plain estuaries. Geol Soc. Am. Mem. 133.
1972i>. Sources and sinks of suspended
matter on continental shelves, p. 249-262. In
D. J. Swift et al. [eds.], Shelf sediment trans-
port: Process and pattern. Dowden, Hutchin-
son & Ross.
Milliman, J. D. 1972. Marine geology, p. 10-1
to 10-91. In Coastal and offshore environ-
mental inventory, Cape Hatteras to Nantucket
Shoals. Mar. Publ. Ser. 6. Univ. R.I.
, and K. O. Emery. 1968. Sea levels
during the past 35,000 years. Science 162:
1121-1123.
Mueller, J A., A. R. Anderson, and J. S. Jeris.
1976. Contaminants entering the New York
Bight: Sources, mass loads, significance. Am.
Soc. Limnol. Oceanogr. Spec. Symp. 2: 162-
170.-
Postma, H. 1967. Sediment transport and sedi-
mentation in the estuarine environment, p.
158-179. In G. H. Lauff [ed.], Estuaries.
Publ. Am. Assoc. Adv. Sci. 83.
Schlee, J. 1973. Atlantic continental shelf and
slope of the U.S. Sediment texture of the
301
Surficial sediments
101
northeast part. U.S. Geol. Surv. Prof. Pap.
529-L. 64 p.
1975. Sand and gravel. MESA New
York Bight Atlas Monogr. 21. 26 p.
Sheridan, R. E., C. E. Dill, Jr., and J. C. Kraft.
1974. Holocene sedimentary environment of
the Atlantic inner shelf off Delaware. Geol.
Soc. Am. Bull. 85: 1319-1328.
Stahl, L., J. Koczan, AND D. J. Swift. 1974.
Anatomy of a shoreface-connected sand ridge
on the New Jersey shelf: Implications for the
genesis of the surficial sand sheet. Geologv
2: 117-120.
Stearns, F., and L. E. Garrison. 1967. Bathy-
metric maps, middle Atlantic U.S. continental
shelf, 1:125,000. NOAA, Natl. Ocean Surv.
Stubblefield, W. L., M. Dicken, and D. J. Swift.
1974. Reconnaissance of bottom sediment on
the inner and central New Jersey shelf.
NOAA-MESA Rep. 1.
, J. W. Lavelle, T. F. McKinney, and D. J.
Swift. 1975. Sediment response to the
present hydraulic regime on the central New
Jersey shelf. J. Sediment. Petrol. 45: 337-
358.
Swift, D. J., G. L. Freeland, P. E. Gadd, G. I Ian,
J. W. Lavelle, and W. L. Stubblefield.
1976. Morphologic evolution and coastal
sand transport, New York-New Jersey shelf.
Am. Soc. Limnol. Oceanogr. Spec. Symp. 2:
69-89.
, J. W. Kofoed, F. P. Saulsbury, and P.
Sears. 1972. Holocene evolution of the
shelf surface, central and southern Atlantic
shelf of North America, p. 499-574. In D. J.
Swift et al. [eds.], Shelf sediment transport:
Process and pattern. Dowden, Hutchinson &
Ross.
Uchupi, E. 1963. Sediments on the continental
margin off eastern U.S. U.S. Geol. Surv. Prof.
Pap. 475-C, p. C132-C137.
. 1970. Atlantic continental shelf and
slope of the U.S. — shallow structure. U.S.
Geol. Surv. Prof. Pap. 529-1.
Veatch, A. C., and P. A. Smith. 1939. Atlantic
submarine valleys of the United States and the
Congo Submarine Valley. Geol. Soc. Am.
Spec. Pap. 7.
Williams, S. J. 1975. Anthropogenic filling of
the Hudson River shelf channel. Geology 10:
597-600.
. 1976. Geomorphology, shallow subbot-
tom structure, and sediments of the Atlantic
intercontinental shelf off Long Island, New
York. U.S. Army Corps Eng. Coastal Eng.
Res. Center Tech. Pap. 76-2. 123 p.
, and D. B. Duane. 1974. Geomorphology
and sediments of the inner New York Bight
continental shelf. Tech. Memo. 45. U.S.
Army Corps Eng. Coastal Eng. Res. Center.
81 p.
302
Reprinted from: Geophysical Research Letters, Vol. 3, No. 2, 97-100,
31
Vol. 3, No. 2
Geophysical Research Letters
February 1976
PRELIMINARY RESULTS OF COINCIDENT CURRENT METER AND SEDIMENT TRANSPORT OBSERVATIONS
FOR WINTERTIME CONDITIONS ON THE LONC ISLAND INNER SHELF
J.W. Lavelle, P.E. Gadd , C.C. Han, D.A. Mayer, W.L. Stubblef ield , and D.J. P. Swift
NOAA/Atlantic Oceanographic and Meteorological Laboratories,
15 Rickenbacker Causeway, Miami, Florida 33149
R.L. Charnell
NOAA/Pacific Marine Environmental Laboratory, 3711 15th Avenue N.E.,
Seattle, Washington 98105
H.R. Brashear, F.N. Case, K.W. Haff, and C.W. Kunselman
Oak Ridge National Laboratory, P.O. Box X, Oak Ridge, Tennessee 37830
Abstract . We have observed late fall and winter
bedload sediment transport and the overlying cur-
rent field in ridge and swale topography on the
inner continental shelf south of Long Island, and
can report movement of bed material at a water
depth of 20 m to a distance of approximately 1500
m after several storm events. Movement over an
11-week observation period was longshore and
oblique to the ridge crest at the experimental
site. Currents were also predominately longshore,
but long term averages demonstrate that a vertical
shear existed in the fluid motion. Although the
number of sediment transport "events" suggested by
the current meter data is nearly balanced in east-
ward and westward directions, both estimates of
transport from current speeds and sand tracer dis-
persion patterns show that several westward flow-
ing events dominated the transport during a two
and one-half month period. A quantitative upper
bound of 31 cm/sec on the threshold velocity for
sediment movement in this size range is also set
by the data.
Introduction
Increasingly widespread interest in the charac-
terization and quantification of shelf sediment
transport stems from the requirements of the
growing number of shelf and nearshore users to
understand the dynamics of an area on which they
may have potential impact. The uses and interests
are myriad, but more common expressions of concern
are phrased in terms of recovery rates of contami-
nated sediments by replacement, the stability of
the substrate for offshore structures, the in-
fluence of offshore work on beach and nearshore
features, and the temporal variability of sediment
transport as an influence on faunal habitats.
While considerable efforts have been made in ob-
serving and describing water-sediment coupling in
the laboratory, under riverine flow, and in the
nearshore area, few direct measurements of off-
shore sediment movement and the associated near-
bottom water velocity field have been made. For
these reasons, we are reporting preliminary re-
sults of an experiment recently completed in the
New York Bight to directlv measure offshore co-
hesionless sediment movement and its immediate
forcing mechanism, the overlying water velocity
field.
The Long Island Near-Shore (LINS) Study
(Figure 1) was centered at 40°33'N and 73°25'W,
halfway between Jones and Fire Island Inlets, Long
Island, New York, some 9 km offshore. The study
area, an 8 x 10 km rectangle, was located in an
area of undulating morphological features described
in Duane et ai. (1972) as ridge and swale topo-
graphy. Bedforms at the study site have wave-
lengths of approximately 1 km with wave heights of
4-7 m, intersect the shoreline obliquely, and are
composed of relatively clean, medium to fine
sands; the ridges are asymmetrical with steeper
southwest facing flanks. The experimental design
was twofold: to gather sediment dispersion and
current meter data which could be used to aid in
quantifying sediment transport; and to gather
qualitative data on the construction and/or the
maintenance mechanism of ridge and swale features
which are widespread on the Atlantic continental
VERTICAL CUK«[NT Will STATIONS
HIST DBOP
Copyright 1976 by the American Geophysical Union.
Fig. 1. Bathymetry and current meter station
locations for the Long Island Nearshore Study
(LINS). Stations CI, C6, C7, and C8 returned no
usable data.
Q7
303
98
Lavelle et al.: Results of Meter and Sediment Observations
shelf (Swift et al. , 1973). Field work was divi-
ded into two concurrent operations: a sediment
tracer experiment and a current meter array of
high spatial resolution. We present here a quali-
tative, preliminary view of the data collected in
those efforts.
Current Meter Observations
During the first six weeks of the current meter
operation (October 16 to December 4, 1974), nine-
teen stations (Figure 1) were occupied. A single
current meter string was retained in the area
during the remainder of the experiment. Aandaraa
RCM-4 Savonius rotor current meters which record
instantaneous direction and integrated average
speed at 10-minute intervals were used throughout.
Measurement emphasis was placed on a well-defined
ridge and trough; meters were located on a crest,
flank, and trough on each of three transects (B,
C, and D of Figure 1) as well as the adjacent
flank of transect C. Additional meters were set
outside the central study area to measure far-
field velocities.
Flow during the observation period trended both
east and west, parallel to the coast. Figure 2 is
a vector time series of velocities at station 2C
(1.5 m above the bottom) and is representative of
near-bottom water movement during one of the most
active periods of flow. The data presented here
have been subjected to a 40-hr low pass filter and
then resampled at hourly intervals. Although east
is the dominant flow direction during this sampling
interval, the most intense flow was westward during
a three day period near the end of this period.
Predominance of eastward flow is consistent with
15 -
16-
,*os*-
17
T
ia
(
19
m
20
f
21
L
22
Wr
23
i
24
V
25
r
26
1
27
k
28
f
29
j
30
!.!
DEC 74
"
2
<^k\
3
™d»
LONG TERM BOTTOM, MIODEPTH, AND
NEAR -SURFACE CURRENT MEANS -
EASTWARD FLOW
(c) LONG TERM BOTTOM, MIODEPTH, AND
NEAR-SURFACE CURRENT MEANS-
WESTWARD FLOW
CURRENT SPEED AND DIRECTION
Fig. 2. Near bottom current vector time
series and velocity averages. Shoreline
direction represents the trend of the 5 fathom
isobath between Jones and Fire Island Inlets.
the observation of Charnell and Mayer (1975) who
reported the existence, in the statistical sense,
of a clockwise gyre in the long term mean flow
within the New York Bight apex during the fall and
winter of 1973. The strong westward flow (Figure
2A) occurred during the storm of December 1 -
December 4, 1974, an event which was reported to
have been the most damaging northeaster since the
Ash Wednesday storm of 1962 (C. Galvin, Coastal
Engineering Research Center, personal communica-
tion). Winds from the east-northeast up to
16 m/s were recorded at John F. Kennedy Inter-
national Airport during the initial 36 hrs of
this period; winds from the northwest at an aver-
age speed of 10 m/s followed on December 3 and 4.
The second most important sustained flow during
the observation period, that which began during
December 16, also followed east winds. Periods
of high speed winds from the west and northwest
cause less intense near-bottom water movement.
The asymmetry of the fluid response to easterly
and westerly winds in this area has been noted by
Beardsley and Butman (1974).
Vertical shear in current velocities was unmasked
in the data (Figures 23 and 2C) when long term
velocity averages were made on data from meters
grouped by position in the water column. Flow
recorded by meters 1.5 to 4 m from the bottom (B) ,
5 and 6 m from the bottom (M) , and 6 to 11 m from
the surface (S), were averaged separately in time
over periods when flow had eastward and westward
components. Water depths at the stations varied
from 15 to 22 m. These data show that near sur-
; face water flow had an offshore component for both
eastward and westward flow, while bottom flow
tended to be more inshore, parallel to the iso-
baths during westward flows and more strongly in-
shore during eastward flows. Speeds decreased in
a relatively uniform fashion from the upper to
the bottom meters. Wind records document a
northerly wind component throughout much of the
observation period; the observed shore-normal
components may be an indication of upwelling
contributions to fluid motion.
Sand Tracer Measurements
In order to directly assess the flow response of
the sediment to the observed water movement, we
employed the Radioisotope Sand Tracer (RIST)
system developed at Oak Ridge National Laboratory
(Duane, 1970; Case et al. , 1971). Indigenous sand
was sorted to produce a fraction whose size dis-
tribution was roughly Gaussian, with a mean dia-
meter of .15 mm (fine to very fine sand), a
standard deviation of .03 mm, and no material
larger than .25 mm or smaller than .06 mm. Approx-
imately 500 cm3 of this material was surface coat-
ed with 10 Curies of the isotope 103Ru (T*j = 39.6
days). On November 12, equal portions of the
tagged sand in water soluble bags were released
at three points at the east end of the main trough
(Figures 1 and 3). The injection points formed an
equilateral triangle with sides roughly 100 m in
length. The ensuing dispersal pattern of labeled
sand was surveyed at intervals by scintillation
detectors mounted in a cylindrical vehicle which
was towed across the bottom. Raydist precision
navigation with 10 m resolution was used. Four
post-injection surveys were made during the 11-
week tracer experiment.
304
Lavelle et al.
Results of Meter and Sediment Observations
99
400
A £
UJ200
z
25 NOV 74
$0 -L
1000
800
600 400 200
200 400 METERS
B
400-
10 JAN 75
: j M
V)
■\\\\\\vM'-
"• '". '■J&K
K
• ' ; *,' . l •
bi
f-
LJ200-
2
n.
1000
800
600
400
200
200
400 METERS
S 100
£ 80
2 60
z 40
£ 20
, JM I Threshold Velocity =
OV'74 20 28 6DEC'74 14 22 30 7JAN'75
r80
o
■60 w
to
■40 v.
20 o
* I2N0V'74 20
6DEC'74
i — i — i — p*
7JAN'75
Fig. 3. a and b) Dispersion patterns measured 13 and 59 days after injection of
tagged sand. Point sources are represented by dots. Broken line is the survey
trackline; stippling represents radiation intensity. c) Near bottom current speed
record over the length of the experiment, and calculated sediment transport infor-
mation (see text) .
Dispersion patterns mapped two and eight weeks
after injection are shown in Figure 3; each has
been corrected for background radiation and decay.
After two weeks (November 25) roughly ellipsoidal
smears trended east from each of the three in-
jection points (Figure 3a). Each smear could be
traced for about 200 m before the signal was lost
in the background radiation. After eight weeks
(January 10) , the three eastward smears had been
replaced by a single, more extensive pattern
extending 700 m to the west (Figure 3b) . The
reversal of the patterns from east to west was
more markedly demonstrated by preliminary data
from a survey made during mid-December (December
17-19) . Although those data have not yet satis-
factorily been processed, the data at the time of
that survey were sufficient to indicate that the
reversal of the dispersion pattern of Figure 3a
had already occurred, and in fact extended approx-
imately 1,500 m to the west. We should point out
that the patterns of Figure 3 must be regarded as
minimum transport patterns, in the sense that
tagged material which has been buried or has dif-
fused downward into the bed is attenuated in sig-
nal strength by the overburden. For this reason,
the signal measured is an underestimate of the
true signal and the observations must be regarded
as a lower bound to the true transport.
The temporal pattern of sediment transport over
a 60-day period may be inferred from Figure 3c.
The basic record is current speed, measured 1.5 m
from the bed, versus time. The horizontal line at
18 cm/sec is an estimated threshold, based on the
work of Shields and subsequent workers for shear
velocity (Graf , 1971) and a choice of 3.0 x 10-3
for the drag coefficient (Sternberg, 1972). We
believe that this choice of threshold velocity is
in part verified by empirical evidence obtained
during the course of the experiment (see below).
We have made estimates of the relative role each
transport event played in the overall transport
record, based on the concept of the proportionality
of frictional energy expenditure to the transport
volume (Bagnold , 1963) . For each event where
velocities exceeding threshold were recorded, we
have calculated a transport volume:
^H
(|v| -
|vTH|)
dt
(I)
v
TH1
where |V| is measured current speed,
threshold speed, a is a constant of proportional-
ity, and T. is the duration of the transport event.
Expression of sediment transport as a power of the
difference of measured and threshold velocity is
supported by analysis of stream transport data
(Kennedy , 1969) . Without assigning a value to a,
one may calculate relative rates of transport, one
event to the next, or one event to the total trans-
port evidenced by the current meter record. We
have taken the second of these options, and have
represented relative sand transport by solid bars
superimposed on the current record (Figure 3c).
Despite the exceedence of the sediment transport
threshold at many points in the record, only the
solid bars centered on December 2 and December 16-
305
100
Lavelle et al.
Results of Meter and Sediment Observations
17 arc visible in the figure, bearing witness to
the dominance of the calculated transport by these
two events. Furthermore, the figure also shows
that most of the calculated transport occurred
during the early December storm. While this cal-
culated transport index may be biased by the choice
of threshold speed as well as the functional de-
pendence on velocity, we believe any other reason-
able parameterization is likely to lead to the
same general conclusion: the storm event of
December 1 - December U moved more sand at 20 m
water depth than the combination of all other
transport events.
The reversing nature of sediment flow during the
observation period provides a constraint on the
entrainment velocity. A threshold speed greater
than approximately 31 cm/sec at 150 cm off the
bottom would eliminate transport during the first
14 days of the record, in contradiction to the
observation of eastward transport (Figure 3a).
Setting the threshold much below 12 cm/sec would
result in more eastward transport during the
entire tracer experiment than was the case. Based
on the relative extent of the dispersion patterns
in Figures 3a and 3b, we believe that the calcu-
lated threshold velocity of 18 cm/sec is realistic.
Summary
Water movement on the Long Island Inner Shelf at
depths of 10 to 20 m and at frequencies below
1/40 hr was predominately alongshore with a net
flow over the observation period to the east. The
non-tidal flow reversals at these depths suggest
domination by winds associated with frontal pas-
sages; the net eastward flow likely reflects the
average winds from the north and west through the
fall and winter months. Vertical shear of the
flow is observable in long term averages of the
current records; small offshore mid-depth flows
and some onshore bottom flow may reflect as an
upwelling circulation the net offshore component
of the wind. The most intense water movements
recorded during the experimental period followed
high northeasterly and easterly winds.
Sediment is transported both eastward and west-
ward parallel to the shoreline, and oblique to the
ridge and trough system. Current speeds recorded
150 cm from the bed show that the sediment en-
trainment threshold is exceeded only intermit-
tently; sand transport occurs only during storm
events, separated by periods of quiescence. Mean
water movement was to the east over the observa-
tion period in sharp contrast to the observed
mean westward sediment transport. Some eastward
sediment transport was observed, but the most
intense water movement and resultant sand move-
ment were associated with several "northeaster"
storm events. Asymmetry of the ridges (steeper
southwest facing flanks) suggests that westward
flows associated with such storms constitute the
primary sediment flow mechanism in this ridge and
swale topography.
Acknowledgements .
Support for this work has come from NOAA's New
York Bight Marine Ecosystems Analysis (MESA)
Project, NOAA's Environmental Research Labora-
tories, and ERDA's Division of Biomedical and
Environmental Research. Oak Ridge National
Laboratory is operated by Union Carbide Corpora-
tion for the U.S. Energy Research and Development
Administration.
References
Bagnold, R.A., Beach and near-shore processes,
part I, mechanics of marine sedimentation, In:
The Sea, vol. 3, pp. 507-528, Interscience
Pub. , New York, 1963.
Beardsley, R. , and B. Butman, Conditions on the
New England continental shelf: response to
strong winter storms, Geophys. Res. Letters,
1, 181-184, 1974.
Case, F.N., E.H. Acree, and H.R. Brashear,
Detection system for tracing radionuclide-
labeled sediment in the marine environment,
Isotopes and Radiation Technology, 8, 412-414,
1971.
Charnell, R.L., and D.A. Mayer, Water movement
within the apex of the New York Bight during
summer and fall of 1973, Tech ■ Memo. , National
Oceanic and Atmospheric Administration,
Boulder, Co. (in press).
Duane, D.B., Tracing sand movement in the littoral
zone: progress in the Radio Isotopic Sand
Tracers (RIST) study, July 1968-February 1969,
Coastal Eng. Res. Center Misc. Paper, Washing-
ton, D.C. , 1970.
Duane, D.B., M.E. Field, E.P. Meisburger,
D.J. P. Swift, and S.J. Williams, Linear
shoal on the Atlantic inner continental
shelf, Florida to Long Island, I_n: Shelf
Sediment Transport: Process and Pattern,
pp. 447-498, Dowden, Hutchinson and Ross,
Stroudsburg, Pa., 1972.
Graf, W.H., Hydraulics of Sediment Transport,
p. 96, McGraw Hill, New York, 1971.
Kennedy, J.F., The formation of sediment ripples,
dunes, and antidunes, _In: Annual Review of
Fluid Mechanics, vol. 1, pp. 147-168, Annual
Reviews, Inc., Palo Alto, Calif., 1969.
Sternberg, R.W., Predicting initial motion and
bedload transport of sediment particles in
the shallow marine environment, In: Shelf
Sediment Transport: Process and Pattern,
pp. 61-82, Dowden, Hutchinson and Ross,
Stroudsburg, Pa. , 1972.
Swift, D.J. P., D.B. Duane, and T.F. McKinney,
Ridge and swale topography of the Middle
Atlantic Bight, North America: secular re-
sponse to the Holocene hydraulic regime,
Mar. Geol., 15, 227-247, 1973.
(Received October 14, 1975;
accepted November 21, 1975.)
306
32
Reprinted from: Earth and Planetary Science Letters , Vol. 32, No. 1, 18-24.
18
Earth and Planetary Science Letters, 32 ( 1976) 1 8-24
© Elsevier Scientific Publishing Company, Amsterdam - Printed in The Netherlands
[5|
ON THE INTERPRETATION OF NEAR-BOTTOM
WATER TEMPERATURE ANOMALIES
R.P. LOWELL
School of Geophysical Sciences, Georgia Institute of Technology, A tlanta, Ga. 30332 (USA)
and
P.A. RONA
National Oceanic and A tmospheric Administration, A tlantic Oceanographic and
Meteorological Laboratories, Miami, Fla. 33149 (USA)
Received November 16, 1975
Final revised version received June 21, 1976
A positive water temperature anomaly of 0.1 LC and an inverse gradient of potential ternperature ot 1 .5 X
jq-2 aQjm nus bCL,n mcasu^d llt the TAG hydrothermal field in the rift valley of the Mid-Atlantic Ridge at latitude
26°N by means of a thermistor array towed between 2 and 20 m above the seafloor. This anomaly appears to be as-
sociated with hydrothermal discharge from the oceanic crust. Thectemperature data are interpreted in terms of (1) a
steady, turbulent thermal plume rising in a homogeneous, neutrally buoyant medium, and (2) turbulent diffusion in
the ocean-bottom boundary layer. The calculations indicate that the thermal output of the TAG anomaly area is of
the order of several megawatts, which is of the same order of magnitude as some continental geothermal systems.
The thermal output from the TAG anomaly area represents a significant fraction of the total heat loss resulting from
the generation of new lithosphere at the Mid-Atlantic Ridge at 26°N.
1. Introduction
Conductive heat flow measurements in the erestal
zone of various sections of the ocean ridge system
exhibit a high degree of scatter [1 ,2], in a manner
opposite to that expected from heat flow refraction
effects [3]. Moreover, the mean conductive heat flow
in the erestal zone is frequently less than the conduc-
tive heat flow from a uniformly spreading lithospheric
plate generated at the ridge axis [4,5]. This discrepancy
between the observed heat flow and conductive heat
flow models based on a uniformly spreading litho-
sphere is usually attributed to convective heat losses
due to hydrothermal circulation in the newly created
oceanic crustal rocks. Additional evidence for hydro-
thermal circulation comes from the occurrence of
hydrothermally altered rocks [6] and deposits of
hydrothermal origin [7,8]. Lastly, thermistor probes
towed over segments of the axial zone within 10-20
m of the seafloor have measured temperature anom-
alies which appear to be associated with hydrothermal
discharge [2,9,10].
Theoretical models for hydrothermal circulation in
the oceanic crust have been based both on models
of convection in porous rock [3,1 1 ], as well as on
models of convection in fractured rock [12,13].
Ocean ridge hydrothermal systems are exceptionally
complicated and the theoretical modelling is still in
its initial stages of development. Near-bottom water
temperature data, however, may provide useful, quan-
titative information with regard to the thermal regime
in the oceanic crust. Such information may, for ex-
ample, give estimates of the heat flux through the
ocean floor and place some constraints on acceptable
hydrothermal convection models. This is of particular
importance in regions of young crust, where the sedi-
307
ment layer is too thin for standard conductive heat
flow measurements to be made.
The purpose of this paper is to examine the im-
plications of near-bottom temperature anomalies mea-
sured with towed thermistors. Since the crestal zone
of the Mid-Atlantic Ridge at 26°N has been studied
in some detail [8-10,14-16], the data from this
area will be used in discussing the theoretical results.
2. The TAG hydrothermal field
During the 1972 and succeeding cruises of the
NOAA Trans-Atlantic Geotraverse (TAG) project,
anomalously thick manganese oxide crusts were re-
peatedly dredged from the southeast wall of the rift
valley at 26°N (Fig. 1). Radiogenic dating of the
manganese crusts, which attain thickness of 42 mm
only 5 km from the axis of the rift valley, show them
to be accumulating at about 200 mm per 106 years,
about two orders of magnitude faster than hydrog-
enous ferromanganese [8]. The crusts are almost pure
manganese (40%), with only trace quantities of Fe,
Cu, Ni, and Co [8]. The rapid accumulation rate and
pure composition evidence a hydrothermal origin for
crusts.
Fig. 1. Bathymetric map contoured in hundreds of meters [161
showing locations of profiles A and B, along which water tem-
perature measurements and bottom photographs were concur-
rently made at the southeast wall of the rift valley of the Mid-
Atlantic Ridge at 26° N. A water temperature anomaly (AT)
was measured between 2950 and 3000 m along profile A
[10]. No water temperature anomaly was present along pro-
file B. The floor of the rift valley is shaded. The TAG hydro-
thermal field is outlined (dashed lines).
TABLE 1
Temperature profiles at the TAG hydrothermal field (Fig. 2)
Tempera-
Cumulative
Depth (m)
Potential tempe
rature (°C).
Thermistor position
Vertical
ture
profile
distance along
in vertical array
(m above lowermost thermistor)
gradient
(n Mfn^Ar^
no Pin hottn m
rc/m)
u v. t. a 1 1 uuuuni
(m)
4
3
0
i
0- 115
3080-3068
2.461
2.449
2.434
+0.007
2
115- 230
3068-3055
2.460
2.454
2.433
+0.007
3
230- 345
3055-3043
2.464
2.453
2.434
+0.008
4
345- 460
3043-3030
2.464
2.455
2.440
+0.006
5
460- 575
3030-3015
2.491
2.482
2.462
+0.007
6
575- 690
3015-2997
2.484
2.470
2.464
+0.005
7
690- 805
2997-2990
2.529
2.518
2.561
-0.014
8
805- 920
2990-2975
2.542
2.510
2.559
-0.016
9
920-1035
2975-2965
2.603
2.599
2.574
+0.007
10
1035-1150
2965-2950
2.617
2.610
2.598
+0.005
11
1150-1265
2950-2935
2.615
2.599
2.469
+0.037
12
1265-1380
2935-2915
2.482
2.481
2.470
+0.003
13
1380-1495
2915-2910
2.508
2.499
2.476
+0.008
14
1495-1610
2910-2905
2.514
2.501
2.482
+0.008
Based on Rona et al. [ 10].
308
20
CO
>
-*—
o
<f>
-Q
o
<
E
fl)
1-
o
c
o
o
^E4
2 3
E
03
Q
0
3 4
PROFILE NUMBER
6 7 8 9 10
246
246 246 246 249 248 2 56
2 56 2 60 262
" , " r-J r
243
243 243 244 246
11 12 13 14
246 252 2 51 2 57 2 60 2 47
POTENTIAL TEMPERATURE (°C)
2 61 248 2 51
247 247 248
251
Pig. 2. A plot of the potential temperature vs. height above the lowest thermistor based on the data in Table 1 from Rona et al. [10]
The hydrothermal manganese oxide occurs both
as crusts on basalt talus and as veins filling fractures
in the talus along the inner margins of steps on the
southeast wall of the rift valley. Bottom photographs
[15] and narrow-beam bathymetry [16] reveal that
the steps range from tens to hundreds of meters in
width, are tens of meters in height, and kilometers to
tens of kilometers in length. The steps are interpreted
as fault scarps.
The manganese oxide is hypothesized to have been
deposited by a sub-seafloor hydrothermal convection
system involving the circulation of seawater through
basalt [16], driven by intrusive heat sources beneath
the rift valley. The discharge is thought to be focused
by fractures in the rift valley wall which are overlain
by a porous and permeable body of talus that may
act to diffuse the fracture focused flow [16]. An
abrupt temperature anomaly was measured in the
water column over one of the steps on the southeast
wall of the rift valley within the area of hydrothermal
deposits, suggesting persistence of hydrothermal ac-
tivity [9,10]. The temperature anomaly of +0.1 1°C
associated with an inverse gradient .of 1.5 X 10"2 °C/m
was measured within 20 m of the bottom, along a hor-
izontal distance of 250 m, between water depths of
3000 and 2950 m using three thermistors mounted in
a 4 m long towed vertical array (see Table 1 and Fig. 2).
A second temperature profile made 5 km away on the
southeast wall of the rift valley, showed no tempera-
ture anomaly [10] (see Fig. 1). The evidence for past
and present activity led to designation of this area,
the TAG hydrothermal field [14].
3. Thermal models
In order to simplify the interpretation of the water
temperature data, we will assume that the tempera-
ture anomaly and superadiabatic gradient results from
steady-state heat transfer to the seafioor. This is a
reasonable assumption because the thickness of the
hydrothermal deposits suggest a time scale for the dis-
charge of 2.5 X 105 years [8], which should be long
enough for a steady state to be achieved [12]. More-
over, the existing data is insufficient to develop a
meaningful time dependent model.
The water temperature data will be interpreted in
two ways. First we will assume that the temperature
anomaly is due to a steady, turbulent thermal plume
discharging at the seafioor. The plume model will
give an estimate of the heat transfer due to the as-
sumed hydrothermal discharge. Secondly, we will
estimate the heat transfer to the seafioor on the basis
of a simple turbulent diffusion model. The heat trans-
fer estimates based on these models provide useful in-
formation with regard to the hydrothermal circula-
tion in the oceanic crust.
3. 1. Thermal plumes
We will first assume that the positive water temper-
ature anomaly of 0.1 1°C measured over the south-
east wall of the rift valley at the TAG area is due to
hydrothermal discharge. Following Williams et al. [2],
we will assume that this anomaly is due to a steady,
turbulent, thermal plume rising by free convection in
309
21
a neutrally buoyant medium. We will assume that near-
bottom currents are negligible. Expressions for the
heat transfer in such a plume have been derived by
Batchelor [18] based on the experimental data of
Rouse et al. [19]. They may be written as Q = (ps/ag)F
where:
„ i=sl"ag(AT) ,_, 2/2 )3/
F = (— — exp(71r \z ),
11
for a three-dimensional plume and:
F = fe^exp(41*W3/2
(1)
(2)
for a two-dimensional plume. In the above equations
a, p, s andg represent the thermal expansion coeffi-
cient, the fluid density, specific heat and accelaration
of gravity respectively. AT represents the magnitude
of the temperature anomaly at height z above the
bottom, and at a distance from the axis of the plume
given by r and x for the three-dimensional and two-
dimensional plume respectively.
Since the temperature anomaly was measured with
the lowest thermistor ranging from 2 to 20 m above
the bottom, we will for convenience choose z- 10 m.
We will also let p = 1 03 kg/m3 , s = 4.2 X 1 03 J/kg °C,
a= 1.6 X 10-4/°C,£= 10 m/s2. Lastly, we will as-
sume that the temperature measurement was made at
the axis of the plume. The results are:
Q=53X 104 watts
for a three-dimensional plume, and:
<2 = 4.6X 104 watts/m
for a iwo-dimensional plume. These results are approx-
imately an order of magnitude greater than those
given by Williams et al. [2] for a similar temperature
anomaly at a similar height measured on the Galapa-
gos spreading center. The reason for the difference is
that Williams et al. [2] used eqs. 1 and 2 directly to
obtain their estimate for the heat transfer. These
equations must be multiplied by ps/ag in order to be
made dimensionally correct (see Batchelor [18]).
We must admit, here, that there is a significant dif-
ference between the rather narrow temperature anom-
alies observed over the Galapagos spreading center
[2, fig. 8] and the anomaly observed over the TAG
area. The anomaly feature over the TAG area is quite
broad (250 m at a height of approximately 10 m from
the bottom). At a height of 10 rn, however, the plume
given by Batchelor [18] would be less than 10 m wide.
On the plume model, it may be possible to partially
explain the width of the temperature anomaly be
means of (1) advection by currents, (2) plume dis-
charge into a stably stratified surrounding medium,
or (3) discharge from several vents along the profile
and mixing of the plumes rising from each. Any or all
of these mechanisms may be operating in the TAG
area.
Bottom photographs were made concurrently with
the thermistor profiles [15]. When the camera com-
pass suspended 5 m below the camera hit pockets of
sediment, sediment plumes rose vertically, indicating
that currents were negligible at the time of the tem-
perature measurements. The presence of ripple marks
in the sediment indicated the existence of intermittent
currents. Sizeable near-bottom currents (<25 cm/s)
have been measured elsewhere in the median valley of
the Mid- Atlantic Ridge [20] Plumes discharging into
a stable environment and forced plumes have not been
as well investigated as the free plume model which
has been used here. However, Morton et al. [21 ] have
shown that plumes discharging in a stable environment
reach a finite height and tend to spread laterally near
the maximum height. An STD profile near the TAG
field shows that the water column there is stable [10]
and this may also be the case in the region of the
thermistor profile. Since numerous faults were en-
countered along the thermistor profile, the tempera-
ture anomaly may in part be due to discharge from
several vents with mixing of the individual plumes.
These are all ad hoc hypotheses, however, and
there is no reliable data available to either substan-
tiate or disprove them. Moieover, the width of the
anomaly is of the same order of magnitude as the
distance traversed by typical semi-diurnal tidal cur-
rents. Therefore we will examine an alternative mod-
el by which to interpret the water temperature anom-
aly. This model is based on the theory of turbulent
diffusion.
3.2. Turbulent diffusion
Wimbush and Munk [22] have recently reviewed
the structure of the ocean-bottom boundary layer.
Only the essential points will be stated here.
310
22
(1) In nearly all cases the boundary layer is turbu-
lent.
(2) There is a "constant stress layer" near the
boundary such that a "friction" velocity may be de-
fined by:
U* = (T0/p)112
where t0 is the stress on the boundary.
(3) Within the constant stress layer there often
exists a viscous sublayer in which heat is transferred
by conduction. This layer has a thickness of the order
of a few centimeters.
(4) Above the viscous sublayer, if it exists, there is
a "logarithmic layer" in which the mean velocity and
temperature increase as the natural logarithm of the
height. This layer usually extends above the constant
stress layer. Within the logarithmic layer we can de-
fine an "eddy diffusion" coefficient:
K=kU*z
(3)
where k — 0.4 is von Karmen's constant.
Thus we may write for the heat flux in this part of
the boundary layer:
H=~psK
bz
(4)
where p and s are the density and specific heat of sea-
water, respectively,// is the heat flux from the earth's
interior, and dT/dz is the gradient of the potential
temperature. For the TAG area we will assume ps =
4.2 X 106 J/m3 °C and z = 10 m. From Table 1 , the
gradient of the potential temperature is -1.5 X 10-2
C/m. The value U* is uncertain, but observations in
the ocean bottom boundary layer have yielded values
from 2 X 10"4 to 2 X 10"3 m/s [22]. Table 2 gives the
heat flux per unit area, the heat flux per unit length of
the ridge axis assuming a width of 250 m fur the TAG
temperature anomaly, and the total heat output from
the anomaly region, assuming lengths of 1 km and 10
km. The heat flux values in Table 2 appear to be rather
large. This may partially be due to the fact that the
logarithmic layer is expected to extend to a height of
the order of 1 m above the seafloor, whereas the tem-
perature gradient was measured at a height of the
order of 10 m. Wimbush and Sclater [23] suggest that
application of eqs. 3 and 4 at heights above the log-
arithmic layer may lead to overestimates of the heat
flux. Nevertheless, for a "typical" value of U* = 0.1
cm/s [22], and a length of 10 km, the total heat out-
put from the TAG anomaly area is of the same order
of magnitude as for the Wairakei high-temperature
area in New Zealand [24], Furthermore, in high-tem-
perature continental geothermal areas, the heat flow
per unit area is often found to be of the order of sev-
eral tens of watts per square meter [25,26]. It is pos-
sible that such heat transport could be achieved by
hydrothermal circulation in the upper few kilometers
of the oceanic crust. The absence of an observed
temperature anomaly a few kilometers away (Fig. 1)
suggests that the 1 km length may be more appropri-
ate for the TAG area. Thus the results in Table 2 do
not appear to be too unreasonable. In any case, it
would appear that the heat flux in the region of the
TAG anomaly is a significant fraction of the heat loss
due to the creation of new lithosphere at the ridge
axis.
4. Conclusions
In order to estimate the heat flux through the
ocean floor from temperature anomaly data, much
TABLE 2
Heat transfer estimates based on turbulent diffusion
u*
H
Heat output per meter of
Total heat out-
Total heat out-
X10-2
(W/m2)
ridge axis assuming 250 m
put assuming 1
put assuming 10
(m/s)
anomaly width (kW/m)
km length (MW)
km length (MW)
0.02
5.0
1.26
1.26
12.6
0.05
12.6
3.15
3.15
31.5
0.1
25.2
6.30
6.30
63.
0.2
50.4
12.60
12.60
126.
311
23
better measurements are needed. Towed thermistor
data can give only semi-quantitative results. Thermis-
tors should be separated by no more than 1 m, and
the array should be towed as close to the bottom as
possible. Since it is generally not feasible to tow the
thermistor array within the logarithmic layer (z < 1
m) because of the irregular topography on ridge
crests, we recommend that ocean floor heat flux mea-
surements in the crestal zone be made by the tech-
niques described by Wimbush and Sclater [23]. This
would involve determination of the velocity and tem-
perature spectra within the logarithmic layer by
means of a bottom mounted device. Such measure-
ments have not been made in regions of young oceanic
crest, and they would be especially useful in regions
where temperature anomalies have been measured
with towed thermistors. Measurements of this type
would show (1) whether the large superadiabatic
temperature gradients measured by towed thermistors
are real, (2) whether strongly unstable layers persist
in the ocean-bottom boundary layer, at least on a
time scale of a tidal cycle, and (3) whether the turbu-
lence in the boundary layer is shear generated or
buoyancy generated.
It may also be useful to measure temperatures in
the upper 10-20 cm of sediment in regions of suspect-
ed hydrothermal activity. Dawson [26] has used soil
temperatures to infer heat flow in convection domi-
nated regions of the Wairakei area. The technological
problems are, of course, somewhat more difficult for
seafloor measurements. It may be difficult to correct
for variations in sediment temperature due to periodic
variations in bottom water temperature.
The two models which we have presented here for
interpreting ocean-bottom water temperature anom-
alies have rather apparent limitations. This is espe-
cially true in view of the qualtiy of the existing data.
The results presented here, however, do sugge that
small, localized near-bottom water temperature
anomalies may be associated with a convective heat
transfer through the seafloor of a significant magni-
tude. This suggests that water temperature anomalies
may not be steady-state phenomena, but rather are
indicative of transient cooling of very young oceanic
crust by episodic hydrothermal circulation. Measur-
able water temperature anomalies may therefore be
somewhat rare.
Acknowledgements
We thank the reviewers for their valuable com-
ments with regard to the original manuscript. In par-
ticular, we thank Dr. G. Bodvarsson for suggesting
that the water temperature anomaly be interpreted
on the basis of turbulent diffusion theory.
This work is part of the NOAA Trans-Atlantic
Geotraverse (TAG) project. This work was supported
by NOAA and the Oceanography Section of the
National Science Foundation under NSF Grant DES
74-00513 A01.
References
1 M. Talwani, C.C. Windisch and M.G. Langseth, Jr.,
Rcykjanes Ridge Crest: a detailed geophysical study, J.
Geophys. Res. 76(1971)473.
2 D.L. Williams, R.P. von Herzen, J.G. Sclater and R.N.
Anderson, Galapagos spreading center: lithospheric cool-
ing and hydrothermal circulation, Geophys. J. R. Astron.
Soc. 38(1974)587.
3 C.R.B. Lister, On the thermal balance of a mid-ocean ridge,
Geophys. J. R. Astron. Soc. 26 (1972) 515.
4 J.G. Sclater and J. Francheteau, The implications of ter-
restrial heat flow observations on current tectonic and
geochemical models of the crust and upper mantle of the
earth, Geophys. J. R. Astron. Soc. 20 (1970) 509.
5 R.L. Parker and D.W. Oldenburg, Thermal model of ocean
ridges, Nature 242(1973) 137.
6 F. Aumento, B.D. Loncarevic and D.I Ross, Hudson
geotraverse: geology of the Mid-Atlantic Ridge at 45°N,
Philos. Trans. R. Soc. Lond., Ser. A, 268 (1971) 623.
7 J.B. Corliss, The origin of metal-bearing submarine hydro-
thermal solutions, J. Geophys. Res. 76 (1971) 8128.
8 M.R. Scott, R.B. Scott, P.A. Rona, L.W. Butler and A.J.
Nalwalk, Rapidly accumulating manganese deposit from
the median valley of the Mid-Atlantic Ridge, Geophys.
Res. Lett. 1 (1974) 355.
9 P.A. Rona, B.A. McGregor, P.R. Betzer and D.C. Krause,
Anomalous water temperatures over the Mid-Atlantic
Ridge crest at 26°N, EOS 55 (1974) 293.
10 P.A. Rona, B.A. McGregor, P.R. Betzer, G.W. Bolger and
D.C. Krause, Anomalous water temperatures over Mid-
Atlantic Ridge crest at 26°N latitude, Deep-Sea Res. 22
(1975)611.
11 E.R. Lapwood, Convection of a fluid in a porous medium,
Proa Cambridge Philos. Soc. 44 (1948) 508.
12 G. Bodvarsson and R.P. Lowell, Ocean-floor heat flow
and the circulation of interstitial waters, J. Geophys. Res.
77 (1972)4472.
13 R.P. Lowell, Circulation in fractures, hot springs and
convective heat transport on mid-ocean ridge crests,
312
24
14
15
Geophys. J. Astron. Soc. 40 (1975) 351.
R.B. Scott, P.A. Rona, B.A. McGregor and M.R. Scott,
The TAG hydrothermal field, Nature 251 (1974) 301.
B.A. McGregor and P.A. Rona, Crest of the Mid-Atlantic
Ridge at 26°N, J. Geophys. Res. 80 (1975) 3307.
16 P.A. Rona, R.H. Harbison, B.G. Bassinger, R.B. Scott
and A.J. Nalwalk, Tectonic fabric and hydrothermal
activity of Mid-Atlantic Ridge crest (26° N) Geol. Soc.
Am. Bull. 87(1976)661.
17 E.T.C. Spooner and W.S. 1'yfe, Sub-sea-floor metamor-
phism heat and mass transfer, Contrib. Mineral. Petrol.
42(1973) 287.
18 G.K. Batchelor, Heat convection and buoyancy effects in
fluids, Q. J. R. Meteorol. Soc. 80 (1954) 339.
H. Rouse, C.-S. Yih and H.W. Humphreys, Gravitational
convection from a boundary source, Tellus 4 (1952) 201.
G.H. Keller, S.H. Anderson and J.W. Lavelle, Near-bottom
currents in the Mid-Atlantic Ridge rift valley, Can. J.
Earth Sci. 12(1975) 703.
B.R. Morton, Sir G. Taylor and J.S. Turner, Turbulent
19
20
gravitational convection from maintained and instanta-
neous sources, Proc. R. Soc. Lond., Ser. A, 234 (1956) 1.
22 M. Wimbush and W. Munk, The benthic boundary layer,
in: The Sea, Vol. 4, A.E. Maxwell, ed. (Interscience, New
York, N.Y., 1970) 731.
23 M. Wimbush and J.G. Sclater, Geothermal heat flux eval-
uated from turbulent fluctuations above the sea floor, J.
Geophys., Res. 76 (1971)529.
24 G.E.K. Thompson, C.J. Banwell, G.B. Dawson and D.J.
Dickenson, Prospecting of hydrothermal areas by surface
thermal survey, in: Proceedings of United Nations 1961
Conference of New Sources of Energy, Geothermal
Energy 2, No. 1 (1964) 386.
25 D.E. White, Rapid heat flow surveying of geothermal
areas, utilizing individual snowfalls as calorimeters, J.
Geophys. Res. 74 (1969) 5191.
26 G.B. Dawson, The nature and assessment of heat flow
from hydrothermal areas, N.Z. J. Geol. Geophys. 7 (1964)
155.
21
313
33
Reprinted from: Sedimentology , Vol. 23, No. 6, 867-872.
Sedimentology (1976) 23, 867-872
An automated rapid sediment analyser (ARSA)
TERRY A. NELSEN
NO A A, Atlantic Oceanographic and Meteorological Laboratories,
15 Rickenbacker Causeway, Miami, Florida 33149, U.S.A.
ABSTRACT
The automated rapid sediment analyser (ARSA) is a pressure-transducer grain-
size analysis system. This basic Woods Hole-type fall tube was automated by the
addition of a digital voltmeter, Hewlett-Packard 9810A calculator, and an x-y
plotter. Eight min after sample introduction, the system automatically produces
size distribution data in 025-9 intervals, distribution statistics, and a plotted
frequency histogram.
INTRODUCTION
As early as 1938, Emery (1938) turned to settling tubes as an alternative to tradi-
tional sieves for a more rapid method of sediment textural analysis. Since then others
have modified the original sand accumulation vs. time technique (Emery, 1938;
Poole, 1957) by measuring pressure changes in the water column with the transit of
falling grains (Zeigler, Whitney & Hays, 1960; Schlee, 1966; Bascomb, 1968) or by
weight accumulation on a balance pan (Felix, 1969) similar to earlier Dutch work.
Although the settling tubes achieve a significant time savings over sieving, the
reduction of the analogue data produced still requires operator time for interpretation,
statistical analysis, and graphic display. Only one early attempt at automated data
acquisition from a settling tube is in the literature (Zeigler, Hayes & Webb, 1964), but
it does not provide for real time data reduction, statistical analysis, and graphic dis-
play. For laboratories analysing hundreds of samples, it is desirable to have a rapid
sediment analyser (RSA) which is as fast as those previously built and also yields
highly accurate and precise data while eliminating the human element from the time
of sample introduction to final statistical treatment of the data. Although the concept
of the settling tube does not limit the analysis range to sand size particles, the long fall
times required for silt and clay sized particles would negate the benefits of rapid
867
314
868 Terry A. Nelsen
analysis gained by the settling tube. Hence the analysis of fines (< 62 urn) is best
undertaken by alternative methods (pipette or electronic particle counters), and the
rapid sediment analyser is most efficiently employed for the textural analysis of sand.
The instrument described here was therefore developed for only the size analysis of
sand. It is a computer based data acquisition system coupled to a Woods Hole type
(Schlee, 1966) rapid sediment analyser and is hereafter referred to as an automated
rapid sediment analyser (ARSA).
ARSA COMPONENT HARDWARE
The fall tube used in this system is clear plastic and the inside diameter measures
10 cm by 200 cm in total length. Previous work (Gibbs, 1972) on the accuracy of
particle-size analysis by settling tube indicated that tubes 7-5 cm and 12-7 cm in dia-
meter were burdened with fall time inaccuracies of up to 34-8% respectively. It should
be noted that these inaccuracies cited by Gibbs (1972) were the result of comparing
the differences in fall velocities of a given size particle for a single sphere against
samples of up to 4 g. Although the ARSA system described here is 10 cm in diameter,
the calibration of the system was conducted relative to sieve analysis (Sanford &
Swift, 1971), and the processes accounting for fall velocity errors were compensated
for in the calibration technique.
Pressure ports are located at 0-5 and 133 cm below the water level in the tube. This
separation is necessary to insure that all introduced particles are in the sensing zone
after the time required to damp surface oscillations resulting from sample introduction.
Pressure and pressure changes within the water column are detected by a Hewlett-
Packard Model # 270 differential gas pressure transducer and are interpreted as voltage
changes resulting from the displacement of the transducer diaphragm. The rate of
change of pressure represents the size distribution of the sample being analysed. This
analogue voltage signal is conditioned by a Sanborn (Hewlett-Packard) Model
350-1 100C carrier preamplifier before it is sent to a Hewlett-Packard Model 3480B
digital voltmeter (with Model 3482A DC range unit) where it is transformed into a
digital voltage signal. A Hewlett-Packard 2570A coupler/controller with crystal clock
provides a reference time base for the calculator's predetermined 0-25- <J> fall times.
Initial fall times were derived from Schlee's (1966) work and adjusted for the longer
tube length of this system. The coupler/controller also provides electronic compati-
bility between the digital voltmeter and the calculator memory. The memory-calcula-
tion function of this system is provided by a Hewlett-Packard Model 9810A calculator.
Final histogram display is generated on a Hewlett-Packard Model 9862A plotter.
Total system compatibility dictated the exclusive use of a single electronics system. It
should also be noted that line voltage fluctuations can introduce spurious transient
signals into the system which cause erroneous voltmeter readings. Therefore it is
necessary to supply power through a voltage regulator.
The ARSA system is pictured in Fig. 1. The fall tube is suspended from a wooden
frame by metal turnbuckles with foam rubber separation pads. The entire system is
shock mounted from the floor by additional foam pads. This minimizes vibration
transmission to the tube mounted transducer. Spirit levels secured to the fall tube at
right angles insure a perfectly vertical tube orientation through turnbuckle adjustments.
315
An automated rapid sediment analyser ( ARSA)
869
Fig. 1. View of the total ARSA system showing fall tube and associated electronics.
Approximately 150-200 samples can be run before accumulated sediment must be
removed through the bottom drain valve and the tube refilled with deionized or
distilled water.
Figure 2 shows the sample introduction device. A controlled electric motor
mounted above the tube depresses a sediment coated screen onto the surface of the
water column. The inverted sub-62 micron screen holds the sample in place (as seen
in Fig. 2b) by water surface tension. Parallel contact of the sediment-laden screen and
the water surface releases the particles and permits a gentle and simultaneous discharge
of the grains. Sample sizes between 5-7 g are used for all ARSA analyses.
316
870
Terry A. Nelsen
Fig. 2. (a) Showing variable speed sample introduction device, top of fall tube, and upper transducer
pressure port, (b) Samole introduction device with sediment on screen ready for sample run.
DATA ACQUISITION AND COMPUTATION PROGRAM
A Hewlett-Packard Model #9810A calculator provides the heart of this ARSA
data acquisition system. The calculator program includes subroutines for data
acquisition, storage, computation, and hard copy output.
Before sample introduction the transducer's output voltage is adjusted to an
arbitrary small positive value which is simultaneously displayed on the digital volt-
meter. Data acquisition starts when sample introduction causes a predetermined
threshold millivoltage to be exceeded. The program then pauses for 5 s while surface
oscillations caused by sample introduction damp. Following this, three voltage values
are read in the next second and placed in memory. Later these values will be averaged
and this average used in the computational subroutine as the 100% reference value.
Based on fall-time values in the memory bank, the calculator then runs time com-
parison do-loops against the system's crystal clock. When the time value for each
0-25-<> interval of the sand range ( — 1 -00-4-00 §) is satisfied, the calculator commands
the digital voltmeter to read the transducer voltage and place this value in memory
for future use in the data reduction subroutine. Successive voltage values decline in
magnitude as grain fallout past the lower pressure port causes the transducer's dia-
phragm to return to the null (baseline) position. After gathering digital voltage values
317
An automated rapid sediment analyser (ARSA)
871
for all the 0-25-<> intervals in the memory, the computational subroutine takes over.
Since all samples analysed in the ARSA have been prescreened (wet and dry) at 4-0 <?
to remove sub-62u material, the program assumes that no material remains in the
water column and considers the final voltage reading as the zero baseline value. In
reality, this is a valid assumption since the occasional trace amounts left in the water
column are below the transducer's detection threshold.
Statistics computed are frequency distribution, cumulative distribution, phi mean,
standard deviation, skewness, and kurtosis. All computations are based on the moment
statistical methods described by Krumbein & Pettijohn (1938). These values are pre-
sented in hard copy by the calculator printer. The graphic display is a 0-25-'> frequency
histogram on the x-y plotter. An example of this graphic display is shown in Fig. 3b.
The entire process from sample introduction to printed statistics and plotted histo-
gram takes 8 min.
-2-0 -1-0 0-0
0 2-0 3-0 4-0 5-0
Fig. 3. Examples of the system's x-y plotter output (sample 10-0-A) for (a) manually imputted sieve
data, cp mean = 1-39, s.d. 091, and (b) an ARSA run data, 9 mean 1-40, s.d. 0-93.
SYSTEM PERFORMANCE
Although the ARSA was developed as an instrument which gave phi means
similar to sieve phi means (final correlation coefficient of 0-99), the final system also
showed a remarkable similarity between ARSA frequency distributions and sieve
frequency distribution data (Fig. 3) with an overall correlation coefficient for 0-25-4
intervals of 0-86.
ACKNOWLEDGMENTS
Throughout the development of this system, Donald J. P. Swift, Patrick G.
Hatcher, and Charles Lauter offered constructive criticism and sound advice which
was greatly appreciated.
318
872 Terry A. Nelsen
REFERENCES
Bascomb, C.L. (1968) A new apparatus for recording particle s;ze distribution. J. sedim. Petrol. 38,
878-884.
Emery, K.O. (1938) Rapid method of mechanical analysis of sands. J. sedim. Petrol. 8, 105-111.
Felix, D.W. ( 1969) An inexpensive recording settling tube for analysis of sands. J. sedim. Petrol. 39,
777-780.
Gibbs, R.J. (1972) The accuracy of particle-size analysis utilizing settling tubes. J. sedim. Petrol. 42,
141-145.
Krumbein, W.C. & Pettijohn, F.J. (1938) Manual of Sedimentary Petrography. Appleton-Century-
Crofts, New York, U.S.A.
Poole, D.M. (1957) Size analysis of sand by a sedimentation technique. J. sedim. Petrol. 27, 460-468.
Sanford, R.B. & Swift, D.J. P. (1971) Comparison of sieving and settling techniques for size analysis,
using a Benthos rapid sediment analyser. Sedimentology, 17, 257-264.
Schlee, J. (1966) A modified Woods Hole rapid sediment analyser. J. sedim Petrol. 36, 403-413.
Zeigler, J.M., Whitney, G.G. & Hayes, C.R. (1960) Woods Hole rapid sediment analyser. J. sedim.
Petrol. 30, 490-495.
Zeigler, J.M., Hayes, C.R. & Webb, D.C. (1964) Direct readout of sediment analysis by settling tube
for computer processing. Science. 145, 51.
(Manuscript received 21 January 1976; revision received 23 March 1976)
319
34
Reprinted from:
Bulletin, Vol
American Association of Petroleum Geologists
60, No. 7, 1078-1106.
Tectonics of Southwestern North Atlantic and
Barbados Ridge Complex1
Abstract More than 40,000 km of bathymetric, mag-
netic, and gravity data and 2,000 km of seismic-reflec-
tion data were obtained in 1971 and 1972 aboard the
NOAA ships Researcher and Discoverer over the Bar-
bados Ridge complex and the ad|acent southwestern
North Atlantic. Most of the tracklines were oriented
east-west and spaced closely (20 km) to attempt corre-
lation between adjacent lines. About a dozen long,
north-south-trending tracklines provided c >ntrol on the
structural variations in that direction
From bathymetric and magnetic dala t .vat. • Jtab-
lished that from the Late Cre'-jceous '■:■■ v.;/ '.ie the
development of the Mid-Atla;.i.c Ric^w i ENia aiea is
essentially the same as in the rest of the North Atlantic.
Indications for relatively recent tectonic activity were
found on some seismic records along several east-
west faults, some of which were in alignment with off-
set zones of the magnetic-anomaly lineations. The im-
plicit suggestion is that intraplate tectonic activity is
common, and that the western extension or "dead
traces" of transform faults may provide avenues where
the accumulated tectonic energy within the oceanic
plate is released.
The influence of many of the major east-west faults
extends westward from the Atlantic basin across the
Barbados Ridge complex to the platform of the Lesser
Antilles volcanic arc. Major topographic changes, as
well as changes in the character of the geophysical
anomalies and in the chemistry of the volcanic rocks
across the fault lines suggest that the faults have
played a significant role in the evolution of this area. As
these faults apparently have affected the structure
west of the shallow earthquake belt and the axis of the
gravity minima, this area appears ideal to study possi-
ble anomalies in the subduction process or, perhaps,
the applicability of the concept itself.
Introduction
The area from the Romanche fracture zone,
near the equator, to the Barracuda Ridge, about
16°N, is one of the more complex geologic areas
of the Atlantic Ocean floor. Whereas the general
evolution of the North and South Atlantic
Oceans, on the basis of magnetic-anomaly linea-
tions, had been understood by 1970, this region
remained a problem area because of the many
fracture zones, the close proximity of the magnet-
ic equator, and the lack of adequate survey cover-
age. Yet this area is in a key position for critical
tests or refinements of the plate-tectonics hypoth-
esis (Isacks et al, 1968; Le Pichon, 1968; Morgan,
1968). In addition to the still unresolved problem
of the overlap of Central and South American
Paleozoic rocks in the Bullard reconstruction of
Pangea (Bullard et al, 1965), and the controversy
GEORGE PETER2 and GRAHAM K. WESTBROOK^
Miami, Florida 33149, and Keele, Staffordshire, England
about the age and origin of the Barbados Ridge
(Meyerhoff and Meyerhoff, 1972), the various
tectonic concepts contained in the papers that
discuss the evolution of this part of the Atlantic
include north-south extension (Funnell and
Smith, 1968), north-south extension and left-lat-
eral shear (Ball and Harrison, 1969, 1970), and
sea-floor spreading along east-west-trending mid-
oceanic-ridge segments (Dietz and Holden, 1970;
Freeland and Dietz, 1972).
To test some of these hypotheses, in 1971 and
1972 a systematic geophysical study of the sea
floor was undertaken between the Lesser Antilles
island arc and the Mid- Atlantic Ridge (Fig. 1).
The northern and southern boundaries of the
study area were approximately 18°N and 10°N,
respectively. The specific scientific objectives
were to: (1) investigate the possible presence of
magnetic-anomaly lineations east of the Lesser
Antilles island arc, establish their trend, and iden-
tify them; (2) define topographic and structural
trends, and establish the development of the Mid-
Atlantic Ridge and the associated fault zones in
the area; (3) determine the east-west extent of the
Barracuda fault zone and its role as a major
©Copyright 1976. The American Association of Petroleum
Geologists. All rights reserved.
'Manuscript received, July 23, 1975; accepted. January 20,
1976.
2NOAA, AOML. MG&GL.
3The University.
The success of this work was due largely to the cooperation
and dedication of the captains, officers, and crews of the
NOAA ships Researcher and Discoverer.
We are grateful to Omar E. DeWald, George Merrill, and
Sam A. Bush of the Atlantic Oceanographic and Meteorological
Laboratories (AOML) of the NOAA for their significant
contributions to the data-collection and processing phases of
this work, and to the preparation of some of the bathymetric
and magnetic maps.
We acknowledge George H. Keller and Bonnie C. McGregor
for their critical reviews of the manuscript, and Claire Ulanoff
for her cheerful editorial and typing work.
Data presented in this paper over the Barbados Ridge
complex south of 14°N were provided before their publication
by Graham K. Westbrook. University of Keele. England.
This work was supported by the Marine Geology and
Geophysics Laboratory of AOML, NOAA, with contributions
from NSF-IDOE Grant NO. AG-253 and AG-489.
Although the writers generally agree on the interpretation
presented, the senior writer takes full responsibility for
challenging some of the concepts of plate tectonics.
1078
320
Southwestern North Atlantic and Barbados Ridge
1079
Fig. 1 — Trackline coverage east of Lesser Antilles island arc (NOAA 1971, 1972).
transform fault or plate boundary; (4) describe in
detail the southeast extension of the Puerto Rico
Trench and the area of transition between it and
the Barbados Ridge (of special interest was the
determination of the role of the Barracuda and
other fault zones as barriers to sediment deposi-
tion); and (5) determine the structure of the Bar-
bados Ridge, and investigate subduction and un-
derthrusting as possible mechanisms for its
formation.
Data-collection techniques, instrumentation,
and comments on data processing and accuracy
were given by Peter et al (1973a, b) and Dorman
et al (1973). In this paper the bathymetric, mag-
netic, gravimetric, and seismic-reflection results
will be discussed in light of the basic objectives.
Mid-Atlantic Ridge
Bathymetry
Previous investigations have established that
the overall trend of the Mid-Atlantic Ridge east
of the Lesser Antilles island arc is north-south
(Heezen and Tharp, 1961; Collette et al, 1969;
van Andel et al, 1971; Collette and Rutten, 1972).
Most of the NOAA tracklines (Fig. 2) were ori-
ented perpendicular to this trend, and were ex-
pected to reveal the development of the Mid-At-
lantic Ridge with little interference from
*Z
35
36-
-r-
55*
-15*
-T-
50*
T-
45*
20-
-T
40-
Fig. 2 — Identification of selected east-west tracklines.
321
1080
George Peter and Graham K. Westbrook
60°W
55°W
50°W
45°W
n — t — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — r
CO
<r
UJ
h-
UJ
2
Q
3000
6000
0
3000
6000
0
3000
6000
0
3000
6000
0
3000
6000
0
3000
6000
0
3000
6000' L
36
35
Royal Deep
30
Researcher Ridge
500
I000
I500
2000
DISTANCE (KILOMETERS)
Fig. 3 — Bathymetric profiles along northern half of east-west tracklines.
322
Southwestern North Atlantic and Barbados Ridge
1081
3000
6000
0
3000
6000
0
3000
6000
5J °
tr
UJ
ui
— 3000
I
I-
Q.
UJ
a
6000
0
3000
6000
3000
6000
0
3000
6000
60°W 55*W 50°W 45°W 40° W
0rT 1 1 1 1 1 1 1 1 1 1 1 1 1 I I I I I I ' ' '
30
J ^L
29
72
41
42
44
70
^>v^\
-i — i — i — i — i — i — i i i i_i i i i i i i i i ■ ' '
500
2000
1000 1500
DISTANCE (KILOMETERS)
Fig. 4 — Bathymetric profiles along southern half of east-west tracklines.
323
2500
1082
George Peter and Graham K. Westbrook
in
cc
£ 6
s
"o 7
1 —
I I I I
I I I I
I I
i i
i i
1000
500
0
Km
2 —
s
\
3-
^i J
AWy
111 lUi
i '\i/j
'\ rwf
lit - ' fh Akjl
4 —
i ii iuU fiilfyw
iwv
\
5-
$
. ,.li.*Uiil
\MMrf
• K
6-
PiWf"^"«l
<H '
I
I I I
I I I I
i i
i i
1 1 1
Fig. 5 — "Standard" Mid-Atlantic Ridge profile S and comparison of its average elevation
curve to bathymetric profile 36.
transform faulting. However, it was found that
changes in the development of the ridge are sig-
nificant even between relatively closely spaced
lines (Figs. 3, 4). On most of the northern track-
lines (Fig. 3) the crestal province is so rugged that
it is hard to pick even the central rift valley or the
axis of the ridge. On the southern lines the re-
placement of the central rift valley by a major
axial mountain is illustrated on the adjacent pro-
files of 42 and 41, and 70 and 44 (Fig. 4). Other
dramatic changes among adjacent profiles are
due to fracture zones that cross the Mid-Atlantic
Ridge obliquely. Profiles 35 and 34 cross the Roy-
al deep (Collette et al, 1973), which reaches a
depth of nearly 6,000 m only at a distance of 200
km from the axis of the ridge (after the southeast
extension of the Puerto Rico Trench, this is the
greatest depth east of the Lesser Antilles). Profiles
30 to 33 show the largest ridge and fracture zone
of the area, the Researcher Ridge (Peter et al,
1973c), which appears to be a branch of (or en
echelon with) the Fifteen-Twenty fracture zone
(Collette and Rutten, 1972).
From the bathymetric sections in Figure 4, and
from the sediment-thickness determinations of
the area (Ewing et al, 1973) it is obvious that the
flank of the Mid-Atlantic Ridge is buried by pro-
gressively more sediments as one approaches the
South American continent. When a "standard"
Mid-Atlantic Ridge section from the North At-
lantic is compared with ridge sections in this area
(Figs. 5, 6), the average ridge-elevation curve
matches all ridge segments well (Peter et al,
1973c). The identical height relations imply iden-
tical age (Sclater et al, 1971) and identical devel-
opment; i.e., a continuity of the central half (from
the axis toward the flank to about 1,000 km) of
the Mid-Atlantic Ridge from the area of the
"standard" section to the southern North Atlan-
tic, adjacent to the Lesser Antilles island arc.
Two north-south profiles on Figures 7 and 8
illustrate the segmentation of the Mid-Atlantic
Ridge due to transverse faulting. On profile 1 1-48
the deepest point is the Royal deep, the tallest
peak is on the Researcher Ridge. The step-like
deepening of the sea floor south of the Research-
er Ridge is, of course, due to the eastward offset
of the Mid-Atlantic Ridge at 15°N. Profile 13-46,
north of 15°20'N, is over the eastern flank of the
Mid- Atlantic Ridge; on the south the ridge offset
at 15°20'N has shifted the ridge axis to the east of
this profile, so the topography shown is the west
flank of the ridge. Farther south along this profile
are smaller offsets at approximately 13°30'N and
12°30'N, and the deepest valley at 11 °N is part of
the Vema fracture zone.
324
Southwestern North Atlantic and Barbados Ridge
1083
Fig. 6 — Comparison of average elevation curve of "standard" Mid-Atlantic Ridge section 5 to
bathymetric profiles 29 and 41.
-i i i i i i ' ' i i i ' i i i_
-20°
4
* w
17
67
58
I
50
48
/
13
46
-15c
-10"
-r
1 1 1 1 r
"i 1 1 1 1 1 r
60° 55° 50° 45°
Fig. 7 — Identification of selected north-south tracklines.
40°
325
1084
George Peter and Graham K. Westbrook
6000
10°N 11° 12° 13° 14° 15° 16° 17°N
0 500
DISTANCE (KM)
Fig. 8 — Magnetic anomaly and bathymetric profiles
along north-south tracklines.
Our interpretation of the fracture pattern of the
Mid-Atlantic Ridge is somewhat different from
that of Collette et al (1974). This difference essen-
tially resulted from the orientation of the survey
lines; theirs ran mostly north-south, ours mostly
east-west. When the transform faults are close to-
gether, it is difficult to distinguish whether a ma-
jor peak and trough sequence is part of a fracture
zone or part of the north-south-trending ridge to-
pography. To solve this minor discrepancy be-
tween the two interpretations, more east-west
lines are needed between the Researcher and the
Vema fracture zones, and more north-south lines
west of 50°W.
Within the area of the more detailed survey
coverage (between 14°N and 17°30'N) a change
in the strike of the fracture zones was detected
between 50°W and 53°W (Figs. 9, 10). The three
major features on the map, the Royal trough, 15°
20'N fracture zone, and the Researcher Ridge
and fracture zone all strike west-northwest near
the crest of the Mid-Atlantic Ridge. Trackline
spacing was too wide to show clearly the ridge
flank south of the 15°20'N fracture zone, but it is
expected that the topography is just as rough
there as north of the Royal trough. Due to the
same problem, the transition from northwest to
north-south trend south of the Researcher Ridge
and northwest of the Royal trough is highly inter-
pretive. What can be detected without doubt is
that the Researcher Ridge terminates at 51°40'W,
and that an east-west trending fault emerges from
it at about 15°15'N (Fig. 9). The 15°20'N fracture
zone and the Royal trough also seem to have their
western limit around 51°20'W, and two east-west
faults, one along 16°20'N and the other along
17°N, take their place.
Magnetics
Magnetic-anomaly lineations were identified,
and a complete evolutionary history of the North
Atlantic was given by Pitman and Talwani
(1972). They noted that, compared to the Pacific
Ocean, the identification and correlation of the
magnetic anomalies are more difficult because of
the much slower Atlantic-sea-floor-spreading rate
and the more common transform fault zones. In
the area east of the Lesser Antilles the reduced
anomaly amplitudes due to the nearness of the
magnetic equator, and the more common fracture
zones, make the anomaly identifications and cor-
relations even more of a problem. The technique
and the set of criteria used to overcome these dif-
ficulties were described by Peter et al (1973c).
Briefly, it involved the use of several "standard"
North Atlantic profiles taken from Lattimore et
al (1974) for comparison purposes (one is shown
on Fig. 10), and the use of theoretical magnetic-
326
Southwestern North Atlantic and Barbados Ridge
1085
spreading models computed at different spread-
ing rates. In addition, the relation between the
ridge elevation and age of the oceanic crust, de-
rived by Sclater et al (1971), was utilized by com-
paring both the magnetic and topographic pro-
files to the "standard" North Atlantic sections.
The profiles obtained by Lattimore et al (1974)
were used as "standard" because those were run
parallel with and between the Atlantis and Kane
fracture zones and, therefore, are not expected to
be influenced by any other significant transform
faults.
An amended version of the correlations of Pe-
ter et al (1973c) is shown in Figure 10. It is em-
phasized that several other identifications could
have been made, depending on one's criteria.
However, the differences among those would
have changed only the identification of certain
individual peaks without negating the overall pat-
tern, or the existence of the late Mesozoic and
Cenozoic magnetic lineations east of the Lesser
Antilles island arc (Fig. 1 1). The lineation pattern
shown in Figure 1 1 matches well the pattern de-
rived by Pitman and Talwani (1972) north of
18°N, and several of the offsets correlate with
fracture zones that also are present in the bathy-
metry.
Although the correlation of individual anom-
alies cannot be demonstrated convincingly by a
single illustration alone. Figure 12 is offered to
show that the overall character of the magnetic-
anomaly profiles is the same on the north and
south sides of the Vema fracture zone, between
the 100- and 600-km marks.
The magnetic-anomaly profiles shown in Fig-
ure 8 dramatize the point made by Schouten
(1974) about the large amplitude of the magnetic
anomalies over the east-west-trending fracture
zones in this area. Part of the correlation prob-
lems discussed before is the result of these large-
amplitude anomalies, which because of the fre-
quency of the east-west fracture zones mask the
effect of those anomalies that are caused by the
north-south-oriented, normally and reversely
magnetized bands of crustal rocks associated with
the Mid-Atlantic Ridge.
Atlantic Basin
Bathymetry
The area between the western flank of the Mid-
Atlantic Ridge and the eastern margin of the Bar-
bados Ridge complex can be divided into several
physiographic units. South of 14°30'N and west
of 54°30'W, adjacent to the Barbados Ridge, the
sea floor is cut into several regionally uplifted
crustal blocks, characterized by gentle southern
dips of the individual blocks, downdropped to the
north along east-west-bounding faults (Figs. 8, 11,
13, profiles 7-17-21 and 58-67). Due to the north-
easterly dip of the regional bathymetry, the east-
ern margin of these uplifted blocks is not obvious
north of 13°N, but south of 13°N a clear boun-
dary fault is present (Fig. 4, profile 72). However,
south of 12°N the large accumulation of sedi-
ments of the South American continental rise ob-
scures this boundary. The uplifted blocks are
arched gently along a north-south axis. Their
highest elevation is between 56°W and 57°W,
and they gently dip away from this region toward
the Barbados Ridge and the abyssal plain on the
east (see Fig. 14).
With the exception of these regionally uplifted
blocks, the area between the western flank of the
Mid-Atlantic Ridge and the Barbados Ridge
complex is occupied by the northwest extension
of the Guiana basin, or the Demerara abyssal
plain. As Embley et al (1970) have noted, the ele-
vation of this abyssal plain is 350 m higher south
of 13°N than between 14°30'N and the Barracu-
da Ridge. The change of elevation between 13°N
and 14°30'N is controlled by the westward exten-
sion of the east-west faults described previously.
The northwest margin of the Demerara abyssal
plain is a gentle topographic arch which connects
the Barracuda Ridge and the foothills province of
the Barbados Ridge complex (Fig. 13). East of
this topographic high several small east-west-
trending steps on the sea floor (see Fig. 15) indi-
cate additional faulting; on the west, the sea floor
is smooth and gently slopes westward to 58°30'W,
where the northward continuation of the Barba-
dos foothills province forms its limit.
The other two significant bathymetric features
of this part of the Atlantic basin are the Barracu-
da Ridge and the Barracuda abyssal plain. The
Barracuda Ridge extends approximately between
54°30'W and 59°W, and rises more than 1,600 m
above the surrounding ocean floor (Paitson et al,
1964; Birch, 1970). Between 54°30'W and 57°W
it forms part of the east-west structural fabric of
the area; west of 57°W its overall strike is north-
westward, but there are possible suggestions in
the bathymetric (Fig. 13) and seismic data that
the bulk of the ridge may be composed of smaller,
en-echelon, east-west-trending ridge segments.
The Barracuda abyssal plain lies between the
northern fault scarp of the Barracuda Ridge and
another fault (possible transform fault shown on
Fig. 11) along 17°N.
Magnetics
There is no clear pattern to the magnetic-
anomaly distribution between the Barbados
Frontal Hills zone and the western flank of the
Mid-Atlantic Ridge (Fig. 16). Over the abyssal
plains the magnetic anomalies reflect the com-
327
1086
George Peter and Graham K. Westbrook
17
16
Barracuda Abyssal Plain
1 5"
1 4'
Demerara Abyssal Plain
^?
56
55"
54'
Fig. 9 — Bathymetric map of western flank of Mid-Atlantic Ridge and adjacent abyssal plain between 14°N and
17°N. Contour interval 200 fm (fm= 1.83 m).
323
Southwestern North Atlantic and Barbados Ridge
1087
329
1088
George Peter and Graham K. Westbrook
Fig. 10 — Identification and correlation of magnetic
anomalies along east-west tracklines. Anomaly numbers
are after Pitman et al (1968).
bined effect of (1) possible vestiges of Late Creta-
ceous magnetic lineations; (2) segments of frac-
ture zones; and (3) topographic features. Over the
uplifted crustal blocks the anomalies are related
to the east-west-trending faults.
In amplitude, the largest magnetic anomalies
are associated with segments of the east-west
fault zones (Fig. 8; Peter et al, 1973b; Collette et
al, 1974). Within the area of closer control (be-
tween 14°N and 17°30'N) many anomalies are
aligned along 15°N, indicating the westward con-
tinuity of the fracture zone from the Researcher
Ridge (Figs. 11, 16). The Barracuda Ridge does
not have a large continuous sequence of magnetic
anomalies associated with it. The only significant
anomalies are at 55°43'W and at 57°W, but these
anomalies may be parts of the north-south mag-
netic lineations, which appear to cross the Barra-
cuda Ridge east of 58°W.
Seismic-Reflection Data
The bathymetric map (Fig. 13) and the magnet-
ic lineations map (Fig. 11) show a general east-
west fault pattern in this area. The north-south
seismic-reflection lines confirm this general pat-
tern and reveal several smaller faults which in
some cases permit estimations of timing of the
tectonic events along these faults.
One of the seismic-reflection lines along 54°
30' W (Fig. 17) is over the Demerara abyssal plain
(north of 14°40'N) and over the faulted and
uplifted crustal blocks.
The sea floor of the northern area is flat except
near 16°N, where the Barracuda Ridge, only a
few tens of meters high, crosses the profile. South
of the ridge, a sediment-filled trough, approxi-
mately 30 km wide and 1.8 km deep (assuming an
average of 2 km/sec sediment velocity), lies along
the entire southern margin of the ridge (Merrill et
al, 1973) and is well developed even this far east.
The 15°20'N fault zone is represented by a nar-
row raised basement block, centered on 15°19'N,
followed on the south by an 800-m drop of the
basement (centered on 15°16'N). Increasing dis-
placement of the deeper reflectors along this
fault, and along the fault at 15°04'N, suggests
prolonged tectonic activity. The step on the sea
floor above these faults indicates that this activity
has been continuous to the present (seen better on
original records).
Most of the prominent reflectors are traceable
across the northern part of the Demerara abyssal
plain from the Barracuda Ridge to about 14°
38'N. The northward dip of the reflectors be-
tween the faults at 15°18'N and at 14°38'N is
caused either by the concurrent displacement
along the faults at 15°18'N, 15°04'N, and 14°
38'N with the deposition of the sediments, or it
may be primary, and the result of the depositional
environment north of the raised crustal block
(south of 14°38'N).
Several strong reflectors can be traced across
most of the first uplifted block (south of 14°
48'N). These reflectors, and occasionally the sea
floor as well, are offset by the larger faults, which
again suggest tectonic activity that continued un-
til the present. Many warps and small faults with-
in the sedimentary strata, and small undulations
of the sea floor over basement highs probably are
related to differential compaction.
The sequence of sedimentary deposits overly-
ing the basement starts with a transparent zone
that mostly fills the basement depressions
throughout the entire profile (Fig. 17). It has been
suggested that the massively stratified sequence
overlying the transparent zone represents late
330
Southwestern North Atlantic and Barbados Ridge
1089
i i i_
.. .. i_30__i25 i20_Jli_>J.i _f_ }_
30 25 20 13 6 5 ! 5
t_ t_ J I t / ( \
l3^_l2L _J!° J_ ; I i
[30 }25 " |20 fa |6 I5 f'
i l i i i i i i i i i i_
VEMA F Z
16°
14°
12°
10°
60°
58"
56°
54°
52°
50°
48°
46°
44°
42°
40°
Fig. 1 1 — Identification and offset pattern of magnetic-anomaly lineations east of Lesser Antilles arc. Anomaly
numbers are after Pitman et al (1968). Anomaly no. 1 is over axis of Mid-Atlantic Ridge.
Tertiary, Quaternary turbidites (Damuth and
Fairbridge, 1970; Embley et al, 1970). The block
south of 12°54'N appears to be an exception, be-
cause here the transparent zone is overlain by a
sedimentary unit approximately 1,000 m thick
(intermediate turbidite unit) characterized by
strong incoherent reflections and by three or four
weak reflecting horizons, which locally fade into
the incoherent zone. From 12°20'N southward
this unit is overlain by strongly stratified se-
quence of late Tertiary -Quaternary turbidites.
Whereas normal, reverse, and thrust faults
commonly can be recognized on seismic records
run normal to the trends of these faults, strike-slip
motion cannot be established on the basis of
these records alone. The studies of Currey and
Nason (1967) of the seaward extension of the San
Andreas fault revealed that a zone of chaotic re-
flections, abruptly terminating coherent reflecting
horizons, can be expected in a strike-slip fault
zone. This zone also may involve complementary
normal faulting. From these observations and the
fact that some of the faults on Figure 17 lie at the
~T~V
t
-i500
400
300
200
100
0
300
Km
Fig. 12 — Magnetic-anomaly profiles 41 and 70 showing
correlation across Vema fracture zone.
westward extension of magnetic-offset zones, the
apparent normal faults centered on 15°15'N and
14°38'N, and the fault zones centered on 12°53'N
and 11°05'N also may have had strike-slip dis-
placements.
The importance of near-bottom currents in car-
rying and eroding the sediments is illustrated at
15°15'N (Fig. 17) where the trough left by the
fault already is filled by transparent sediments,
and at 11°40'N where there is a small erosion
channel at the point where the dip of the sea floor
changes from a southerly to a northerly direction.
Profile A (Figs. 13, 15) illustrates currently ac-
tive faults south of the Barracuda Ridge. Their
relative youth is indicated by the step-like dis-
placements of the sea floor, and the increased dis-
placement of the deeper reflectors also indicates
activity along these faults in the geologic past.
Most of the faults are associated with scarps or
steep slopes of the basement, suggesting that the
origin of the tectonic activity is within the oceanic
crust. The northernmost fault at 15°39'N appears
to be related to the relative uplift of the Barracu-
da Ridge, the troughs at 15°N and at 15°15'N are
related to the east-west trending fault systems of
this area.
Barbados Ridge Complex
Bathymetry
The Barbados Ridge complex is an outer sedi-
mentary-island arc that consists of a system of
the north-south and east-west-trending ridges,
troughs, scarps, and topographic lineaments,
whose respective development varies along the
331
1090
George Peter and Graham K. Westbrook
c
o
u
o
o
C
o
E
.a
332
Southwestern North Atlantic and Barbados Ridge
1091
O - M
>
S-i
^t/fm^y^%^^-.»,<-}-.y-V^'Mii If/a,
~-^;MM
0 ,.^
8-
-
s
i \ :.jft 4 . ;
o-
% t
: . .,■
|| ;
| ;
.,...,. ■• :,.. ,0
B~ »
m «
1 ■£ *
>
00
Q.
B
o
4
o
e •» « «
S0NO33S Nl 3WU 13AVU AVM'OMl
S0NO33S Nl 3WI1 13AVH1 AVM-OM1
333
1092
George Peter and Graham K. Westbrook
I6°00' N
Fig. 15 — Retouched photograph of north-south seismic-reflection line along 56°30'W, directly south of
Barracuda Ridge.
60° W
I8°N
16°-
14°-
18° N
-16*
_l4<
Fig. 16 — Magnetic-anomaly map over northern half of Barbados Ridge and adjacent Atlantic sea floor.
334
Southwestern North Atlantic and Barbados Ridge
1093
k?- >
S0NO33S Nl 3WI1 13AV81 AVMOM1
mi.
• >iM
o
Z
g
u.
SQNOD3S Nl 3WI1 13AV81 AVMOAAi
335
1094
George Peter and Graham K. Westbrook
Fig. 18 — Key physiographic elements of Lesser Antilles arc-Barbados Ridge complex.
Contour interval 1 km.
arc. The north-south-trending topographic ele-
ments are part of the overall island-arc trends, the
east-west elements are related to the fault systems
of the adjacent Atlantic basin. These extend west-
ward underneath the Barbados Ridge complex
and have cut or modified the major north-south
elements. The interaction of the north-south and
east-west tectonic trends often produced locally
complex structural and topographic patterns:
compared to the trackline spacing, the wave-
length of these features is too short and, there-
fore, these were not incorporated into the bathy-
metric presentation (Fig. 13).
Four major north-south-trending topographic
elements may be distinguished within the Barba-
dos Ridge complex. These are: (1) the Tobago
trough and Lesser Antilles Trench; (2) the Barba-
dos Central Ridge province; (3) the Barbados
Trough province; and (4) the Barbados Frontal
Hills zone. The major east-west-trending topo-
graphic elements are: (1) the Dominica Trans-
verse Valley system; (2) the Martinique Trans-
verse Valley system; and (3) the Sta. Lucia-
Barbados Transverse Ridge system (Fig. 18). The
north-south topographic elements are prominent
from the continental shelf of South America
northward to about 13°N; there is a transition
zone between 13°N and 14°N; and the east-west
elements are more abundant north of 14°N. The
14°N parallel also cuts the Barbados Central
336
Southwestern North Atlantic and Barbados Ridge
1095
w
KM
Fig. 19 — Photograph of seismic section L (Fig. 13).
Ridge, and the average depth of the Barbados
Ridge complex is more than 1,000 m greater
north of here.
South of 14°N, the backbone of the Barbados
Ridge complex is the Barbados Central Ridge, on
which the island of Barbados is located. In addi-
tion to the change of strike at 13°N, the single
Central Ridge also may have broken into a
broader "ridge province," if one assumes that the
isolated peaks north of Barbados are part of the
Central Ridge. As an alternate interpretation,
these peaks also could be related to the Sta. Lu-
cia-Barbados Transverse Ridge system. The Bar-
bados Central Ridge and the other elements of
the Barbados Ridge complex are well developed
south of 13°N.
North of 14°N, the topographic high between
the Dominica and the Martinique Transverse
Valley systems may represent the northernmost
element of the Barbados Central Ridge province.
The Lesser Antilles Trench separates it from the
volcanic arc on the west, and a prominent, nar-
row depression at 58°45'W forms its eastern limit.
The Lesser Antilles Trench has the same over-
all width as the Tobago trough, but its northern
half is "V" shaped with a graben 10 to 15 km
wide in the center (Fig. 19).
It is difficult to trace the Barbados Trough
province between 13°N and 14°N because, as
stated before, there are several minor valleys and
ridges there, occupying an area more than 100 km
wide. North of 14°N, the valley at 58°45'W is a
337
feature that may be a structural equivalent or a
northern extension of the Barbados Trough prov-
ince. Its northern terminus is at 15°N, where it
merges with the Dominica Transverse Valley sys-
tem.
Among the major east-west-trending topo-
graphic elements the Dominica Transverse Valley
system is the northernmost (15°N). Although
there is a limited amount of data available, it ap-
pears that whereas the overall strike of the system
is almost due east-west, it consists of four en-ech-
elon, northwest-southeast-trending valleys, sepa-
rated by narrow, sharp peaks. Its eastern terminus
seems to be at 58°45W. On the west it merges
with the Lesser Antilles Trench, although the
bench on the slope of the volcanic-arc platform
also may be genetically related to it.
The Martinique Transverse Valley system is lo-
cated between 13°55'N and 14°25'N. It is more
than 50 km wide and has a definite east-west
strike. In detail this valley system also is highly
fragmented, but here by a combination of narrow
troughs and ridges, striking both east-west and
north-south.
The broad saddle between the Tobago trough
and the Lesser Antilles Trench is the western
third of the Sta. Lucia-Barbados Transverse
Ridge system (Weeks et al, 1971; Bassinger and
Keller, 1972). The east-west-trending axis of the
system shifts southward to 13°35'N east of the
Central Ridge, and the system becomes much
narrower. It may extend from the Central Ridge
1096
George Peter and Graham K. Westbrook
o
Z
o
u
LU
in
60° 00' W
>
<
i—
>-
<
i
o
KM
Fig. 20 — Retouched photograph of western edge of the Barbados
Central Ridge (sec. B, Fig. 13).
to the Frontal Hills zone in the form of irregular
topographic highs. Data are inconclusive to de-
cide whether the topographic highs between
59°W and 59°30'W are part of this Transverse
Ridge system or part of a broader Central Ridge
province.
North of the Barbados Frontal Hills zone, be-
tween about 59°W and the island platform of the
volcanic arc, lies a generally hummocky topog-
raphy that resembles in geologic and geophysical
character the Frontal Hills zone. The area has a
roughly triangular shape with the volcanic-arc
platform, the Frontal Hills zone, and the topo-
graphic extension of the Puerto Rico Trench
forming the sides of the triangle. The effect of
both east-west and north-south faulting is indi-
cated in the general bathymetric trends and in the
basement configuration below the sediments
(Schubert and Peter, 1973; Schubert, 1974).
Seismic-Reflection Data
Marine seismic-reflection data over several ele-
ments of the Barbados Ridge complex already
have been discussed by Chase and Bunce (1969);
Collette et al (1969); Bunce et al (1970); West-
brook (1973); and Peter et al (1974). Here only a
few comments will be made about the NOAA
1971-1972 seismic-reflection data, as these illus-
trate the overall structural division of the Barba-
dos Ridge complex and the relative age of some
of the tectonic events.
Line J (Fig. 14) is a continuous line from the
Atlantic basin, across the Barbados Ridge com-
plex, the volcanic arc, into the eastern margin of
the Grenada trough. Because the east-west topo-
graphic elements are more significant north of
14°N, one can see from the topography (Fig. 13)
that a 10-km north-south shift of the trackline
would have highlighted different topographic ele-
ments. The line extends across the southern mar-
gin of the Lesser Antilles Trench and across the
northern tip of the Barbados Central Ridge. On
the west flank of the Central Ridge, subbottom
reflectors of the Lesser Antilles Trench arch up
and pinch out, or terminate at the sea floor over
the Central Ridge. Only a thin layer of sediments
covers the acoustic basement on the Central
Ridge. This is shown well on line B (Fig. 20). Line
J runs along the southern part of the Martinique
Transverse Valley and, although good reflecting
horizons are lacking, approximately Vi to 1 sec of
penetration is indicated on the record. East of
59°W, the prominent valley on the record lies in
the same structural province as the Barbados
trough farther south; on the east lies the Frontal
Hills zone. This zone is characterized by lack of
reflectors and penetration along this line. The At-
lantic basin on the east of this area (Fig. 13) is
338
Southwestern North Atlantic and Barbados Ridge
1097
O — Nov»n,0'^eo0-
8? iU
:>
*
'"«
h
t»' ,
o~
"V
m
O •* np>^>n>ON.coo»
S0NCO3S Nl 3WI1 13AVS1 AVM"OMl
SGNOD3S Nl 3Wli 13AVS1 AVMOAAi
o
E
o
U
0
u.
339
1098
George Peter and Graham K. Westbrook
Fig. 22 — Photograph of seismic section M (Fig. 13).
gently arched, and the sediments are dipping
away from 57°W, both toward the Atlantic and
toward the island arc. A good onlap sequence of
sediments is present east of the arch at 57°W and
there is an abrupt termination of a strong, shallow
reflector under the toe of the Frontal Hills. It of-
ten is said that the lack of continuous reflectors
and poor penetration under the toe of island arcs
in general, and in this area in particular, is the
result of very complex faulting (Chase and Bunce,
1969). Records will be presented here that show
that when good reflectors are present, they can be
detected even in cases of very intense faulting.
The strong reflector that dips toward the Frontal
Hills from the Atlantic basin does not show up on
the Frontal Hills because it has not been deposit-
ed there. The termination point of that reflector
W
marks a former edge of the Atlantic basin and the
toe of the Frontal Hills zone, which subsequently
became covered by transparent sediments, ex-
tending the toe eastward.
Most of the reflectors of the Lesser Antilles
Trench on line K (Fig. 21) are uplifted over that
part of the Central Ridge that lies between the
two transverse valleys. Despite intense faulting
the reflectors are recognizable. Along line M (Fig.
22) this entire upper sequence is shown clearly,
suggesting that the valleys and ridges are the
product of tectonics rather than erosion. The At-
lantic basin dips gently toward the west along line
K, and it appears that the reflectors become less
coherent toward the west. Some hints of weak re-
flectors suggest that this entire incoherent se-
quence of sediments is uplifted in the toe of the
KM
Fig. 23 — Photograph of seismic section D (Fig. 13).
340
Southwestern North Atlantic and Barbados Ridge
1099
59°00'W
> 10
i
10
20
KM
Fig. 24 — Photograph of seismic section P (Fig. 13).
Frontal Hills. Farther south, line D (Fig. 23)
shows the uplift of the Atlantic sediments quite
clearly; the well-stratified beds are progressively
thinner as they are uplifted successively along at
least three faults to form the toe of the Frontal
Hills. These faults are either high-angle reverse,
or normal faults. Farther up on the slope the re-
cord shows an approximately 1-km thick zone of
incoherent reflectors, which are similar to those
shown in line K and on some other NOAA pro-
files (unpub.) that were run directly east of the
Frontal Hills zone.
The uplifted sediments of the Atlantic basin on
the Frontal Hills zone are in direct contrast with
the gently westward-dipping (dip approx. 2°) sed-
iments of the Atlantic, north of the Barbados
Ridge complex. On line P (Fig. 24) these sedi-
ments are overlain by a thick accumulation of an
acoustically incoherent sediment pile and, as far
as the instrumentation allowed them to be seen (3
sec penetration, 45 km from the edge of the over-
lying sediment pile), reflecting horizons within
these sediments are undisturbed.
Line L (Fig. 19) illustrates the central graben
and youthful tectonism of the Lesser Antilles
Trench.
Gravity
Several aspects of the gravity anomalies of the
Lesser Antilles island-arc system have been dis-
cussed by Talwani (1966), Bush and Bush (1969),
and Bunce et al (1970). The NOAA and Universi-
ty of Durham investigations (Kearey, 1973; West-
brook, 1973, 1975; Peter, 1974; Peter and West-
brook, 1974a, b; Westbrook, 1974a, b; Kearey et
al, 1975; Westbrook, 1975) allowed the mapping
of the gravity field of this area in much greater
detail than previously, and established the exten-
sion of certain structural trends from the Atlantic
basin into the Barbados Ridge complex.
The most noticeable feature of the gravity field
is the continuation of the negative free-air anom-
aly band of the Puerto Rico Trench, which turns
away from the topographic axis as 18°N. As the
anomaly band extends farther south, it reflects
the east-west structural discontinuities similar to
the topography. These effects manifest them-
selves as: (1) reduced amplitude of the gravity
low (-192 mgal) between 19°N and 18°N (from
Schubert, 1974); (2) sharp increase of the ampli-
tude at 17°20'N (from -220 mgal to -276 mgal,
Schubert, 1974); (3) a 45-km sinistral offset of the
axis of the low at 16°30'N, and another similar
offset at 15°10'N; (4) the interruption of the neg-
ative, north-south anomaly band by a positive,
east-west-trending free-air anomaly band at 13°
50'N; and (5) the development of two approxi-
mately parallel negative free-air anomaly bands
south of Barbados (13°N) with amplitudes about
100 mgal less than the amplitude of the single
band on the north (Fig. 25).
The axis of the free-air anomaly minimum does
not follow the Lesser Antilles Trench south of the
Sta. Lucia-Barbados Transverse Ridge; the sinis-
tral offset at 15°10'N places it east of the Barba-
dos Central Ridge. South of Barbados, the axis of
341
1100
George Peter and Graham K. Westbrook
342
Southwestern North Atlantic and Barbados Ridge
1101
E
o
c
CL.
a
0
343
1102
George Peter and Graham K. Westbrook
one of the negative anomaly bands is over the
eastern edge of the Tobago trough, the axis of the
other is over the Barbados trough. These two
bands are parallel near Barbados, but separate at
the latitude of Tobago ( 1 1 ° 1 5'N ), where the west-
ern band veers west-southwest onto the Paria
shelf, and the eastern band continues south-
southwest until about 10°30'N, where it sharply
swings toward Trinidad.
The free-air anomaly contour lines trend essen-
tially east-west over the Atlantic basin, east of the
Barbados Ridge complex. There are two promi-
nent east-west-trending anomaly bands; one is a
positive band at about 13°50'N, the other is a
negative band north of it. These two seem to ex-
tend from the volcanic platform to about 56°
30'W in the Atlantic basin.
The Bouguer anomaly map of the area (Fig. 26)
also is dominated by essentially north-south and
east-west trends: the north-south trend follows
the Barbados Ridge complex, and the east-west
trend characterizes the Atlantic basin. When the
regional gradient is removed from these data,
some of the east-west-trending Bouguer anom-
alies also cross the complex, and extend to the
island platform of the Lesser Antilles arc.
The north-south-trending Bouguer anomaly
features include: (l) a gradient change under the
Frontal Hills zone due to the dipping mantle; (2)
a band of gravity lows over the greatest sediment
accumulation, which on the south is over the Bar-
bados Central Ridge (presumably this band
marks the location of a former trench); and (3) a
series of highs (170-240 mgal) that extend from
the island platform east-southeast of Guadeloupe
to the Sta. Lucia-Barbados Transverse Ridge
over the lower slope of the island platform. Many
of these highs are even larger than those reported
over the volcanic islands themselves, which only
reach 120-190 mgal (Kearey et al, 1975).
Bouguer anomalies also were used to study fur-
ther the east-west structures of the Atlantic basin.
Two-dimensional modeling of the crustal struc-
ture was performed using the Bouguer anomalies,
NOAA seismic-reflection data (Peter and West-
brook, 1974b), and University of Durham and
earlier seismic-refraction data (Ewing et al, 1957;
Westbrook et al, 1973). An interesting result that
emerged was the fact that when Bouguer anom-
alies were computed down to the acoustic base-
ment with a realistic sediment density (2.0 g
cm-3), then many of the large anomalies were
eliminated (Fig. 27). These results suggest that in
this area the buried topography of the basement
is responsible largely for the Bouguer anomalies,
with only very minor contribution from changes
of mantle elevation.
Discussion
One of the main objectives of this paper is to
present a large body of new data over a previous-
ly little studied region of the southwestern North
Atlantic, and the Barbados Ridge complex. Pa-
pers by Westbrook (1973, 1974a, 1975) discussed
in detail how some of these data may be fitted
into an overall plate-tectonics scheme. We intend
only to highlight here the relation of these new
data to (1) the scientific objectives outlined earli-
er; (2) some of the hypotheses advanced for the
evolution of this area; and (3) some of the corol-
lary assumptions of the plate-tectonics hypothe-
sis.
At the time this project was initiated in 1971,
the sea-floor-spreading history was not known
east of the Lesser Antilles arc. Although the mag-
netic-anomaly lineation pattern presented (Fig.
1 1) is admittedly debatable in detail, the magnetic
lineations and the east-west topographic profiles
clearly establish that there is a well-developed
Mid-Atlantic Ridge east of the Lesser Antilles,
and that this segment of the ridge has evolved
since the Late Cretaceous in the same way as in
the rest of the North Atlantic. Two tracklines in-
dicate similar development of the Mid-Atlantic
Ridge even south of the Vema fracture zone.
From geometric considerations of the original fit
of the continents, we propose that the possible
southern limit of this type of ridge development is
the Doldrums fracture zone (approx. 8°N). We
found neither topographic nor magnetic evidence
for major breaks in the continuity of the Mid-
Atlantic Ridge, which would have supported a
major north-south extension of this area during
the Cenozoic (Funnell and Smith, 1968; Ball and
Harrison, 1969, 1970). As the Late Cretaceous
magnetic lineations also are trending north-south,
there is no indication for the existence of a ridge-
ridge triple junction at that time, which might
have provided some indirect support for the inter-
pretation and identification of the east-west-
trending anomalies in the Colombia basin as
being Late Cretaceous (Christofferson, 1973), if
this part of the Caribbean were formed in an At-
lantic spreading regime.
As part of the topographic and structural stud-
ies of the Mid-Atlantic Ridge, several major
(transform) and minor faults were located. Data
from the northern half of the study area indicate
that the northwest-southeast trend of the faults
near the ridge axis changes to east-west between
magnetic anomalies 6 and 13. The eastern half of
the Barracuda Ridge follows an east-west fault
pattern, and there is no indication in the bathy-
metric data that it extends farther east and con-
344
Southwestern North Atlantic and Barbados Ridge
1103
425
250
400
225
mgal |g|
^ong
54 5C
w
TO ACOUSTIC
TO SEABED
8&SEMENT
-
425
250
400
225
375
200
350
I75
325
150
J50
175
325
150
0
5
km 10
0
"i
1 04
"
2 e
5 3
10
15
"
15
20
!
1 1 1
1 1 1 1
1
_i
20
Fig. 27 — Bouguer anomaly profiles and crustal structure along 54°30'W. Upper
half: Bouguer anomaly profile to seabed, lower numbers on scale; Bouguer anomaly
profile to acoustic basement, higher numbers on scale. Lower half: crustal section
from sea surface to mantle. Density of 1.04 g/cc = sea water, 2.0 g/cc = sediment,
2.8 g/cc = average crust, and 3.3 g/cc — mantle. Top of 2.8-g/cc layer was taken
from our seismic-reflection record, top of 3.3-g/cc layer was computed to fit "acous-
tic basement Bouguer profile," with nearby seismic-refraction data extrapolated as
guide. Horizontal and vertical scales are in kilometers; for map location, latitude
crossings also are shown.
nects with the Fifteen-Twenty or Researcher frac-
ture zones. Clearly additional bathymetric and
magnetic data are needed to define better the to-
pography and structure of the flank of the Mid-
Atlantic Ridge where changes in the trend of
these faults take place, and where major ridges,
like the Barracuda and Researcher, seemingly ter-
minate.
One of the interesting results of our study was
the discovery of relatively recent tectonic activity
along many major and minor faults west of the
exposed flank of the Mid-Atlantic Ridge. This ac-
tivity was especially obvious in the bathymetric
and seismic-reflection records from the Demarara
abyssal plain directly south of the Barracuda
Ridge (Fig. 15), and along the east-west faults on
the 300-km-wide, uplifted crustal block adjacent
to the Barbados Ridge complex. Our detailed sur-
veys and the seismic records suggest that faulting
and tectonic activity within an oceanic crustal
plate may be common. The lack of any clear pat-
tern of recorded earthquakes from these areas
may indicate that (1) there are no very recent
movements along these faults; (2) the displace-
ments occur through creep; or (3) the magnitude
of the earthquakes is so small that they cannot be
recorded at far removed measuring stations.
Because of the east-west-trending offset zones
of the magnetic lineations and the fact that the
faults on the seismic records often coincide, one
also may conclude that the offsets of magnetic-
anomaly bands may not be due entirely to trans-
form faulting, or that the so-called "dead-traces"
of transform faults remain zones of weakness
where subsequent tectonic adjustments occur.
Relatively recent motions along these faults
also may support one of the contentions of Ball
and Harrison (1969, 1970) that most of the faults
cutting the Mid-Atlantic Ridge in the equatorial
region are very slow-moving transcurrent faults.
They argued that if the transcurrent motion is
slow compared to the combined spreading mo-
tion between the offset parts of the ridge crest,
then the earthquake first-motion studies will show
the "transform" movement but the transcurrent
part may not be detected.
Bathymetric and seismic-reflection data show
that the Barracuda Ridge becomes a subbottom
feature west of 58°50'W, and extends westward
along the same trend to the island platform of the
Lesser Antilles arc (Schubert and Peter, 1973;
Schubert, 1974). However, it is only one of the
many east-west structural elements in this area,
and does not appear to have had a controlling
influence by itself on the transition, of the south-
eastward extension of the Puerto Rico Trench.
As suggested by Peter et al (1974), the Puerto
Rico Trench proper terminates at about 18°N,
60°W, where the zones of the gravity minima and
the shallow earthquakes separate from the topo-
graphic trough. If a definition of a trench as a
structural feature accompanied by the zones of
the gravity minima and shallow earthquakes is
accepted, then the topographic trough extending
345
1104
George Peter and Graham K. Westbrook
southeastward from the Puerto Rico Trench is
not part of this trench. We believe that in this
transition zone there is a buried trench, marked
by the axis of the gravity minima and the co-lo-
cated shallow earthquake zone, which connects
with the Puerto Rico Trench on the north, and
the Lesser Antilles Trench on the south.
The high acoustic reflectivity of the sea floor
and the manv incoherent subbottom reflectors
make the structural analysis of this area very dif-
ficult. However, the roughly east-west bathyme-
tric trends, and the abrupt changes in the ampli-
tude of the gravity minima most likely reflect the
westward extension of the fault systems that char-
acterize the Atlantic basin east of the Lesser An-
tilles arc and the Barbados Ridge complex (Schu-
bert, 1974). Some of these faults appear to have
caused the offset of the axis of the buried trench
in a left-lateral sense, and others probably are
partly blocking it. Basically, however, it is not
these faults, but the large amounts of sediments
that have collected in the trenches and fault de-
pressions east of the Lesser Antilles island arc
that are responsible for the termination of the to-
pographic expression of the Puerto Rico Trench.
The hummocky topography in the transition
zone (between the Puerto Rico Trench and the
Barbados Ridge complex) is most likely slump
and current-derived sediment that has been trap-
ped between the east-west faults. The sediments
overlie the gently westward-dipping Atlantic sea
floor (line P, Fig. 24), causing the topographic
low that extends southeast from the Puerto Rico
Trench. As an interpretation alternative to that
proposed by plate tectonics, it is possible that the
upper sedimentary horizon (strong upper reflec-
tor on line P) of the Atlantic sea floor is an un-
conformity, rather than a thrust surface. Accord-
ing to the subduction hypothesis the overlying
sediment pile represents scrapings of sediments
from the Atlantic sea floor. If such a process is
possible, then in case of underthrusting not only
the overlying sediment pile should be disturbed,
but at least the upper, unconsolidated parts of the
Atlantic sea-floor sediments as well. It is unlikely
that the approximately l-km-thick, mostly unli-
thified sediments covering the basement rocks of
the Atlantic would underthrust a 2 to 3-km thick
sediment pile to about 40 km, without undergoing
any noticeable internal deformation.
The increasing width and height of the Barba-
dos Ridge complex toward the south generally
are attributed to the availability of more sedi-
ments closer to the South American source, and
to the subsequent subduction of more sediments
on the south (Chase and Bunce, 1969; West-
brook, 1973). Our north-south seismic-reflection
profiles east of the Barbados Ridge complex show
that the thickness of sediments over the basement
is essentially uniform between 13°N and 14°30'
N; at about 15°30'N it increases rapidly toward
the Barracuda Ridge, and at 12°20'N it increases
substantially southward. West of the thickest sed-
iments on the north, the Barbados Ridge complex
terminates, and there are no changes in the Bar-
bados Ridge complex south of 12°20'N either.
These observations seem to indicate that there is
no simple relation between the present sediment
thickness of the Atlantic basin and the topog-
raphy of the Barbados Ridge complex, and that
other controlling factors should be considered.
For this alternative we have suggested that the
east-west faults of the Atlantic basin have extend-
ed across the trough now occupied by the Barba-
dos Ridge complex, and that these have formed
dams against the northward-advancing sediments
within the trough (Peter, 1974; Peter and West-
brook, 1974a, b; Westbrook, 1974a, b, 1975). The
influence of these fault zones also is manifested
by the abrupt changes in the elevation of the Bar-
bados Ridge complex, as they provided a suffi-
cient discontinuity within the crust that uplifts of
the sediment pile have occurred at different times
and in different degrees between the various seg-
ments of these faults. The seismic records are es-
pecially clear in indicating a relatively recent
(Pleistocene-Holocene) uplift of the Barbados
Ridge complex north of 14°N (Fig. 19), but the
uplift on the south has occurred earlier (Pliocene-
Pleistocene?; Peter et al, 1974; Westbrook, 1975).
In this context of decreasing age of the elements
of the Barbados Ridge complex toward the north,
the triangular-shaped hummocky region on the
north — from its geophysical characteristics — is
really the youngest member of this complex, that
has not been subjected to significant uplift. If the
major faults do define the boundaries of uplift,
then it is reasonable to assume that the area be-
tween the Barracuda Ridge and the present edge
of the complex (approximately 16°N) will be
uplifted next.
The data presented here for the area of the Bar-
bados Ridge complex show vertical uplift at the
toe of the Frontal Hills zone (Fig. 2). Further,
east-west fault zones, revealed by bathymetric,
seismic, and gravity data, extend across the Bar-
bados Ridge complex, the axis of the gravity mi-
nima, and the shallow-earthquake zone to the vol-
canic arc platform. At the largest of these fault
zones, near 14°N, major changes occur both
within the structure of the Barbados Ridge com-
plex and the volcanic arc. These changes include:
(1) the Barbados Ridge complex becomes 1,000 m
deeper on the north; (2) the free-air gravity
anomalies change character; (3) the positive
Bouguer anomalies extending southward from
Desirade over the lower slope of the island plat-
form terminate; (4) the chemistry of the volcanic
346
Southwestern North Atlantic and Barbados Ridge
1105
rocks of the Lesser Antilles arc shows marked dif-
ferences on the two sides of this fault (Stipp and
Nagle, 1974; Wills, 1975); and (5) the trend of the
volcanic arc changes.
These observations and changes may be ex-
plained by some anomaly in the subduction pro-
cess (Westbrook, 1973, 1974a, b, 1975), and some
of them may be only coincidences. In all certain-
ty, however, these data are sufficient to question
some of the current concepts of the subduction
process as they are applied to this area, and sug-
gest the need for serious, further investigations.
Conclusions
1. The Mid-Atlantic Ridge in the area east of
the Lesser Antilles arc developed from about the
Late Cretaceous to the Holocene much as in the
rest of the North Atlantic. Thus, the Barracuda
Ridge and fracture zone is not a major disconti-
nuity between oceanic crusts of different spread-
ing history.
2. Relatively recent tectonic activity along the
western extension of some transform faults sug-
gests that these "dead traces" actually may pro-
vide avenues for the release of tectonic energy in
the oceanic plate.
3. Several east-west faults extend from the At-
lantic basin to the island platform of the volcanic
arc. These have cut the former trench east of the
arc, have dammed the northward advance of sed-
iments in the trough, and probably have caused
the segmented differential uplift of the Barbados
Ridge complex. Bathymetric and seismic-reflec-
tion records indicate that the area south of 14°N
was uplifted before the area on the north; the
area north of 16°N has geophysical characteris-
tics similar to the Barbados Ridge complex and
may be thought of as the youngest member of the
complex that has not received substantial uplift.
4. The Puerto Rico Trench terminates at ap-
proximately 18°N, 60° W, where, because of sedi-
ment fill and the influence of east-west faults, the
topographic trough veers sharply eastward of the
shallow earthquake zone and the axis of the gravi-
ty minima. These, however, most likely mark the
now buried part of the trench that connects
southward with the Lesser Antilles Trench.
5. Because the east-west faults of the Atlantic
basin cross the axis of the gravity minima and the
shallow earthquake zone, and seemingly even in-
fluence the structure of the volcanic arc and the
chemical composition of its rocks, a simple sub-
duction model probably is not applicable for this
area.
References Cited
Ball, M. M„ and C. G. A. Harrison, 1969, Origin of the
Gulf and Caribbean and implications regarding
ocean ridge extension, migration, and shear: Gulf
Coast Assoc. Geol. Socs. Trans., v. 19, p. 287-294.
1970, Crustal plates in the central Atlan-
tic: Science, v. 167, p. 1128-1129.
Bassinger, B. G., and G. H. Keller, 1972, Marine geo-
physical observations across the Barbados Ridge-St.
Lucia cross warp (abs.): 6th Caribbean Geol. Conf.,
Margarita, Venezuela, Trans., p. 379-380.
Birch, F. S., 1970, The Barracuda fault zone in the west-
ern North Atlantic: geological and geophysical stud-
ies: Deep-Sea Research, v. 17, p. 847-859.
Bullard, E. C, J. E. Everett, and A. G. Smith, 1965, The
fit of the continents around the Atlantic: Royal Soc.
London Philos. Trans., ser. A, v. 258, p. 41-51.
Bunce, E. T., J. D. Phillips, R. L. Chase, and C. O.
Bowin, 1970, The Lesser Antilles arc and the eastern
margin of the Caribbean Sea, in The sea, v. 4: New
York, Wiley-Intersci., p. 359-385.
Bush, S. A., and P. A. Bush, 1969, Isostatic gravity map
of the eastern Caribbean region: Gulf Coast Assoc.
Geol. Socs. Trans., v. 19, p. 281-285.
Chase, R. L., and E. T. Bunce, 1969, Underthrusting of
the eastern margin of the Antilles by the floor of the
western North Atlantic Ocean, and origin of the Bar-
bados Ridge: Jour. Geophvs. Research, v. 74, p.
1413-1420.
Chnstofferson, E., 1973, Linear magnetic anomalies in
the Colombia basin, central Caribbean Sea: Geol.
Soc. America Bull., v. 84, p. 3217-3230.
Collette, B. J., and K. W. Rutten, 1972, Crest and frac-
ture zone geometry of the Mid-Atlantic Ridge be-
tween 10° and 16°N: Nature, Phys. Sci„ v. 237, p.
131-134.
H. Schouten, and K. Rutten, 1973, Fracture
zones and flow trajectories in the southern North At-
lantic (abs.): Eos (Am. Geophys. Union Trans.), v.
54, p. 325.
J. I. Ewing, R. A. Lagaay, and M. Truchan,
1969, Sediment distribution in the oceans: the Atlan-
tic between 10° and 19°N: Marine Geology, v. 7, p.
279-345.
H. Schouten, K. Rutten, and A. P. Slootweg,
1974, Structure of the Mid-Atlantic Ridge province
between 12° and 18°N: Marine Geophys. Research-
es, v. 2, p. 143-179.
Curray, J. R., and R. D. Nason, 1967, San Andreas fault
north of Point Arena, California: Geol. Soc. America
Bull., v. 78, p. 413-418.
Damuth, J. E., and R. W. Fairbridge, 1970, Equatorial
Atlantic deep-sea sands and ice-age aridity in tropical
South America: Geol. Soc. America Bull., v. 81, p.
189-206.
Dietz, R. S., and J. C. Holden, 1970, Reconstruction of
Pangea: Breakup and dispersion of continents. Perm-
ian to present: Jour. Geophys. Research, v. 75, p.
4939-4956.
Dorman, L. M., B. G. Bassinger, E. Bernard, et al, 1973,
Caribbean Atlantic geotraverse, NOAA-IDOE 1971,
Rept. 3, Gravity: NOAA Tech. Rept. ERL 277-
AOML 11.
Embley, R. W., J. I. Ewing, and M. Ewing, 1970, The
Vidal deep-sea channel and its relationship to the
Demerara and Barracuda abyssal plains: Deep-Sea
Research, v. 17, p. 539-552.
347
1106
George Peter and Graham K. Westbrook
Ewing, J. I., C. B. Officer, H. R. Johnson, and R. D.
Edwards, 1957, Geophysical investigations in the
eastern Caribbean: Trinidad shelf, Tobago trough,
Barbados Ridge, Atlantic Ocean: Geol. Soc. America
Bull., v. 68, p. 897-912.
Ewing, M., G. Carpenter, C. Windisch, and J. Ewing,
1973, Sediment distribution in the oceans: the Atlan-
tic: Geol. Soc. America Bull., v. 84, p. 71-88.
Freeland, G. L., and R. S. Dietz, 1972, Plate tectonic
evolution of the Caribbean-Gulf of Mexico region:
6th Caribbean Geol. Conf., Margarita, Venezuela,
Trans., p. 259-264.
Funnell, B. M., and A. G Smith, 1968, Opening of the
Atlantic Ocean: Nature, v. 219, p. 1328-1333.
Heezen, B. C, and M. Tharp, 1961, The physiographic
diagram of the South Atlantic, the Caribbean, and
the Scotia Sea, and the eastern margin of the South
Pacific Ocean: Geol. Soc. America.
Isacks, B., J. Oliver, and L. R. Sykes, 1968, Seismology
and the new global tectonics: Jour. Geophys. Re-
search, v. 73, p. 5855-5900.
Kearey, P., 1973, Crustal structure of the eastern Carib-
bean in the region of the Lesser Antilles and Aves
Ridge: PhD thesis, Univ. Durham, 182 p.
G. Peter, and G. K. Westbrook, 1975, Geophysi-
cal maps of the eastern Caribbean: Geol. Soc. Lon-
don Jour., v. 131, p. 311-321.
Lattimore, R. K., P. A. Rona, and O. E. DeWald, 1974,
Magnetic anomaly sequence in the central North At-
lantic: Jour. Geophys. Research, v. 79, p. 1207-1209.
Le Pichon, X., 1968, Sea floor spreading and continen-
tal drift: Jour. Geophys. Research, v. 73, p. 3661-
3697.
Merrill, G, G. Peter, and W. Sproll, 1973, Seismic re-
connaissance east of the Lesser Antilles arc (abs.):
Eos (Am. Geophys. Union Trans.), v. 54, p. 326.
Meyerhoff, A. A., and H. A. Meyerhoff, 1972, Conti-
nental drift, IV: the Caribbean "plate": Jour. 'Geol-
ogy, v. 80, p. 34-60.
Morgan, J. W., 1968, Rises, trenches, great faults, and
crustal blocks: Jour. Geophys. Research, v. 73, p.
1959-1982.
Paitson, L., C. H. Savit, D. M. Blue, and W. A. Knox,
1964, Reflection survey at Barracuda fault: Geophy-
sics, v. 29, p. 941-950.
Peter, G, 1974, Tectonic evolution of the eastern mar-
gin of the Caribbean region (abs.): Geol. Soc. Ameri-
ca Abs. with Programs, v. 6, p. 910.
and G. K. Westbrook, 1974a, Tectonics of the
Barbados Ridge and adjacent Atlantic basin (abs.):
Eos (Am. Geophys. Union Trans.), v. 55, p. 284.
1974b, Interconnection between the tec-
ERL 293-AOML 13, 29 p.
C. Schubert, and G. K. Westbrook, 1974,
tonic framework of Barbados Ridge and the adjacent
Guiana basin (abs.): 7th Caribbean Geol. Conf. Abs.,
p. 50-51.
O. E. DeWald, and B. G Bassinger, 1973b, Car-
ibbean Atlantic geotraverse, NOAA-IDOE 1971,
Rept. 2, Magnetic data: NOAA Tech. Rept. ERL
288-AOML 12.
G. Merrill, and S. Bush, 1973a, Caribbean At-
NOAA-IDOE Caribbean Atlantic geotraverse: Geo-
times, v. 19, p. 12-15.
R. K. Lattimore, O. E. DeWald, and G. Merrill,
1973c, Development of the Mid- Atlantic Ridge east
of the Lesser Antilles island arc: Nature, Phys. Sci., v.
245, p. 129-131.
Pitman, W. C, III. A. M. Herron, and J. R. Heirtzler,
1968, Magnetic anomalies in the Pacific and sea-floor
spreading: Jour. Geophys. Research, v. 67, p. 2069-
2985.
and M. Talwani, 1972, Sea floor spreading in
the North Atlantic: Geol. Soc. America Bull., v. 83, p.
619-646.
Schouten, H., 1974, Magnetic anomalies over fracture
zones in the lower magnetic latitudes of the central
North Atlantic (abs.): Eos (Am. Geophys. Union
Trans.), v. 55, p. 232.
Schubert, C, 1974, Seafloor tectonics east of the north-
ern Lesser Antilles (abs.): 7th Caribbean Geol. Conf.
Abs., p. 62-63.
and G. Peter, 1973, Sea floor tectonics west of
the Barracuda Ridge (abs.): Eos (Am. Geophys.
Union Trans.), v. 54, p. 326.
Sclater, J. G., R. N. Anderson, and M. L. Bell, 1971,
Evolution of ridges and evolution of the central East-
ern Pacific: Jour. Geophys. Research, v. 76, p. 7898-
7915.
Stipp, J. J., and F. Nagle, 1974, A geochemical study of
the Lesser Antilles island arc: regional distribution of
Sr 87/86 initial ratios (abs.): 7th Caribbean Geol.
Conf. Abs., p. 65.
Talwani, M., 1966, Gravity anomaly belts in the Carib-
bean in continental margins and island arcs, in Conti-
nental margins and island arcs: Canada Geol. Survey
Paper 66-15, p. 177.
van Andel, Tj. H., R. P. Von Herzen, and J. D. Phillips,
1971, The Vema fracture zone and the tectonics of
transverse shear zones in oceanic crustal plates: Ma-
rine Geophys. Researches, v. 1, p. 261-283.
Weeks, L. A., R. K. Lattimore, R. N. Harbison, et al,
1971, Structural relations among Lesser Antilles,
Venezuela, and Trinidad-Tobago: AAPG Bull., v. 55,
p. 1741-1752.
Westbrook, G. K„ 1973, Crust and upper mantle struc-
ture in the region of Barbados and the Lesser An-
tilles: PhD thesis, Univ. Durham, 228 p.
1974a, The structure of the Barbados Ridge and
buried trench of the Lesser Antilles subduction zone
(abs.): 7th Caribbean Geol. Conf. Abs., p. 71.
1974b, The structure of the Barbados Ridge: 7th
Caribbean Geol. Conf. Trans, (in press).
1975, The structure of the crust and upper man-
tle in the region of Barbados and the Lesser Antilles:
Royal Astron. Soc. Geophys. Jour., v. 43, p. 1-42.
M. H. P. Bott, and J. H. Peacock, 1973, The
lantic geotraverse, NOAA-IDOE 1971, Rept. 1, Pro-
ject introduction — bathymetry: NOAA Tech. Rept.
Lesser Antilles subduction zone in the region of Bar-
bados: Nature, Phys. Sci., v. 244, p. 18-20.
Wills, K. J. A., 1975, The geologic history of southern
Dominica and plutonic nodules from the Lesser An-
tilles: PhD thesis, Univ. Durham, 244 p.
348
35
Reprinted from: Earth and Planetary Science Letters, Vol. 30, No. 1, 135-142
Earth and Planetary Science Letters, 30(1976) 135-142
© Elsevier Scientific Publishing Company, Amsterdam - Printed in The Netherlands
135
OPENING OF THE RED SEA WITH TWO POLES OF ROTATION *
EVAN S. RICHARDSON '<2 and C.G.A. HARRISON ]
University of Miami. Rosenstie! School of Marine and Atmospheric Science, Miami, Fla. (USA)
NO A A - Atlantic Oceanographic and Meteorological Laboratories, Miami, Fla. (USA)
Received August 12, 1975
Revised version received January 5, 1976
Previous studies have shown that the Red Sea was formed by two stages of sea-floor spreading, with a quiescent
period in between. We suggest that these two phases have occurred in different directions. The shape of the central
trough indicates that the present-day motion is almost E-W, whereas the total opening, deduced from the shape of
the coastlines, is NE-SW. If the axial trough has opened in an E-W direction, the earlier stage of opening was in a
direction which made the Dead Sea Rift fall along a small circle to the pole of early opening, and hence suggests
that the Dead Sea Rift was a transform fault during this early stage. The later movement gives almost pure extension
along the Dead Sea Rift, and this should be seen by normal faulting. Available first-motion studies are not precise
enough to confirm or deny this hypothesis.
1. Introduction
2. Coast to coast opening
It has been suggested that there have been two
stages of sea-floor spreading in the Red Sea [1,2]. The
latter stage was responsible for the narrow axial trough
in the center of the Red Sea, which is approximately
50 miles wide and is believed to have been caused by
spreading over the last 3.5 m.y. The earlier stage was
responsible for the formation of the bulk of the Red
Sea. Between the two stages of spreading there was a
period of quiescence, which allowed thick salt deposits
to be accumulated. These thick salt deposits indicate
that the oceanic crust formed during the earlier stage
of spreading is at a greater depth than the crust form-
ed during the later stage. The axial trough is caused by
the absence of salt accumulations over the younger
crust. Bathymetric and magnetic evidence has been
published to support this two-stage concept [2].
* Contribution from the University of Miami's Rosenstiel School
of Marine and Atmospheric Science, 4600 Rickenbacker
Causeway, Miami, Florida 33149.
Several poles of rotation for the opening of the Red
Sea have been published in the past [1 ,3— 6]. These
TABLE 1
Poles of rotation for the Red Sea
Pole of rotation Method of calculation
lat.
long.
Refer-
ence
32.5° N 22.5° E shear along Dead Sea Rift and [3]
Gulf of Aden data
29.0°N 32.0°E fit of coastlines and rates of [4]
opening in Red Sea and Dead
Sea Rift
36.5°N 18.0°E fit of coastlines and Suez Rift [5]
32.0°N 22.0°E fit of lines 20-30 km seaward [6]
of coasts and fit of Danakil
Horst
31.5°N 23.0°E fit of lines 52 km seaward of [1]
coasts, Gulf of Suez
349
136
poles and the methods by which they were calculated
are summarized in Table 1 .
An important factor in obtaining a pole of rotation
for the Red Sea is whether or not the sea opened from
coast to coast. Davies and Tramontini [7| believe that
much o\ the Red Sea is underlain by oceanic crust.
The\ felt that the most important result to emerge
from then work was the clear indication that the Red
Sea is underlain by rocks whose seismic velocity is
higher than the majority of velocities reported from
the continents. They reported a layer with an approxi-
mate seismic velocity of 6.63 km/sec which is in good
agreement with the oceanic layer 3 average of 6.69 km/
sec. From where their refraction lines end to the shore-
line, they were not prepared to generalize except to
note that there is. apart from a superficial change in
sediment reflection characteristics, no reason to suppose
that the structure changes.
Magnetic anomalies in the axial trough are lineated
parallel to the strike of the topography. The pattern
tits the anomaly sequence expected from sea-floor
spreading, and it seems absolutely certain that the
axial trough has been formed by sea-floor spreading
over the past few million years. Recently, Girdler and
Styles [2] have shown that at one place outside the
axial trough there are also lineated magnetic anomalies.
They believe that these have also been caused by sea-
floor spreading during an earlier episode, thus confirm-
ing the seismic evidence that almost all of the Red Sea
is oceanic in nature.
Girdler and Styles proposed that there was a period
of quiescence between the formation of the main por-
tion of the Red Sea, which they thought was formed
between 41 and 34 m.y. ago, based on the magnetic
time scale of Heirtzler et al. [8], and the axial trough,
which was formed during the past 3.5 m.y. Although
the timing of the earlier stage of opening seems to be
in doubt, the difference in structure and the change in
character of the older magnetic anomalies, compared
with the younger ones, suggesting a deeper source for
the older ones, certainly is strongly indicative of two
stages of spreading with quiescence in between.
3. Establishment of the later pole of rotation
The establishment of the pole of rotation for the
total opening of the Red Sea by fitting together the
Fig. 1. Oblique Mercator projection of the Red Sea with projec-
tion pole at 36.5° N, 18.0°E. This was the position of the pole
of rotation used by McKenzie et al. [5] to describe the coast-
line fit on either side of the Red Sea. Lines connecting correspond-
ing points on opposite coastlines are horizontal and represent
small circles about the pole. The shorter lines connecting corre-
sponding points across the axial valley are not horizontal.
coastlines gives a pole at 36.5°N, 18.0°E [5]. Fig. 1 is
an oblique Mercator projection of the Red Sea region
using this point as the projection pole. It can be seen
that lines connecting congruent points on each coast-
line are horizontal lines, representing small circles
about the pole. But we noticed on a detailed chart of
the Red Sea prepared by Laughton [9] that the axial
trough appeared to have been formed by a more or
less E-W separation. On this chart, the western bound-
350
137
ary of the axial trough, which is marked by the 500-
fathom contour, can be superimposed on the same
contour on the eastern boundary, by an eastward
translation. We digitized the 500-fathom margins of
the axial valley at 15' intervals of latitude. These mar-
gins are plotted on the transverse Mercator projection
of Fig. 1 . Congruent points on either side of the axial
trough are connected by the short lines, and it can be
seen that these lines are not horizontal. This suggests
that the axial valley was not formed by relative motion
about the pole used to plot Fig. 1 , but about a differ-
ent pole.
In order to establish a possible pole of rotation for
the formation of the axial trough, we traced the western
500-fathom boundary of the axial trough from Laugh-
ton's map and slid it eastwards so that it matched the
eastern boundary. Since the movement was almost
exactly latitudinal, it is possible to do this because the
Fig. 2. Oblique Mercator projection of the Red Sea with projec-
tion pole at 15.2°S, 32.8° E. This is the pole of rotation calcu-
lated for the later stage of opening. The lines connecting con-
gruent points on either side of the axial trough (500-fathom
isobath) are now horizontal, whereas the lines connecting the
coastlines are not.
map of Laughton is plotted on a Mercator projection.
We were able to measure a vector of opening for the
northern end of the axial trough and one for the south-
em end. These two vectors enabled us to establish a
pole of opening which was at 1 5.2°S, 32.8°E, with a
total opening angle of 1 .08° about this pole. This is
the pole for the clockwise rotation of Arabia from
Nubia (or that part of Africa to the northwest of the
East African rift valley). Using this rotation as the pro-
jection pole of an oblique Mercator projection (Fig. 2),
we can see that now congruent points on either side of
the axial valley are horizontal, and therefore lie on
small circles about this pole.
One problem in establishing a pole for the later
spreading episode is that there may have been flow-
age of salt deposits into the axial trough, which may
therefore be narrower than the amount of sea floor
created during the more recent episode of sea-floor
spreading. Girdler and Darracott [1 | have also suggest-
ed that the later opening may be about a pole different
from that describing the early opening, but without
going into details concerning what the difference might
be. However, in a more recent paper. Girdler and Styles
[2] make no mention of different poles of rotation.
Girdler and Whitmarsh [10] have found evidence at
two DSDP sites in the southern and central Red Sea
(sites 227 and 228 in Fig. 3) that there has been lateral
flow of salt deposits into the axial trough. At both
sites Pliocene sediments and Miocene evaporites were
found overlying oceanic crust which was predicted
by magnetic anomaly evidence to be younger than
2.5 m.y. Coleman [11] has also mentioned the possi-
bility of lateral flow of Red Sea evaporites. He suggest-
ed that the irregular bathymetry of the axial trough
at 16.67°N may have resulted from salt flowage.
At present there is no evidence that there has been
similar flowage in the northern Red Sea. Clearly the
bathymetric shape of the axial valley is less smooth
in the north. If there has been flowage of evaporites
in the south but not in the north, this will cause an
error in our calculation of the pole of opening of the
axial trough. The error will be mainly in the determi-
nation of the latitude of the pole and not in the longi-
tude. The true pole would lie further to the south than
the position given above. If the axial valley in the south
has been narrowed by 12 km due to flowage of salt,
then the pole of opening should be shifted from a
latitude of about 1 5°S to a latitude of about 70°S.
351
138
THE RED SEA
EGYPT
SUDAN
ETHIOPIA
Fig. 3. Map of the Red Sea showing the location of three earth-
quakes (a, b, and c) [5,13,14) for which fault plane solutions
have been obtained, and the locations of two DSDP sites (227
and 228) where it has been suggested that lateral flow of evap-
orites has occurred tending to smooth the boundaries of the
axial trough [10].
However, since the amount of flowage is difficult to
assess quantitatively, we shall assume that the pole at
15.2°S, 32.8°E is correct, but will mention what the
effect of moving it further south would be.
A better way of establishing the distance to a pole
of rotation is to rotate magnetic anomalies on either
side of the spreading center into coincidence, as done,
for instance, by Pitman and Talwani [12]. Magnetic
anomalies represent oceanic crust which is more precise-
ly dated than the locus of the thick salt deposits out-
lined by the central valley described above. However,
the development of magnetic anomalies is generally
poor within the Red Sea. Estimates of spreading rate
can only be obtained from a rather narrow latitudinal
band from about 17°N to 21°N [13-15]. The uncer-
tainty in observed spreading rates is so large that no
trend from south to north can be seen. All that these
data tell us is that the pole of rotation lies sufficiently
far away from the Red Sea such that a variation of
spreading rate of 20% or more is not produced within
this 4° band. This puts the pole of rotation further
away than about 20°. Beyond that the magnetic data
cannot go.
An alternative method of determining the direction
of movement is to measure the strike of transform
faults offsetting the axis of spreading. However, in the
Red Sea we do not believe that there are any features
which have been definitely identified as transform
faults. Therefore this method of analysis is not available
for this portion of the earth.
4. Establishment of early pole
We accept the pole of McKenzie et al.(36.5°N,
18.0°E) [5] as a good approximation of the mean pole
of opening for the complete history of the Red Sea.
However, if a two-stage model is accepted, this pole
becomes a resultant of the first and second episodes
of spreading. Therefore it is possible to obtain a pole
of rotation for the early stage of spreading by vectori-
ally subtracting the axial trough pole from the resul-
tant or total pole of McKenzie et al. [5]. Employing
this method, we have calculated the early pole at
29.6°N, 20.6°E, some 6° south of the total pole open-
ing. If there has been flowage of evaporites in the south-
ern Red Sea, the later pole will be further south and
this will tend to move the early pole to the north.
5. Dead Sea Rift
One supposition of plate tectonics is that a trans-
form fault always falls on a small circle to the relative
pole of rotation of the two plates involved. The Dead
Sea Rift has generally been accepted as such a bound-
ary between the Arabian and Nubian plates. Fig. 4
shows the pole of McKenzie et al. [5] plotted on a
map of the Mediterranean and Red Sea region. It is
evident here that the Dead Sea Rift does not fall on a
small circle to this pole. However, the Dead Sea Rift
does fall quite close to a small circle about the early
pole presented in this paper (Fig. 5). We feel that this
is more than a fortuitous occurrence. In fact, depending
352
139
/L,
sT
L--
—
l^__-L— -T \ \ \ \^^ \
i \ \ \ \^\ \
1 Ni -—\-~~\ \ \
1 \ r \ \ ^-V""^\
' X
h
N
^
\ / \L_V-"T^ \ \
■^
|
Fig. 4. Azimuthal equidistant projection of the Mediterranean
and Red Sea region with projection pole at 36.5° N, 18.0°E
marked with a cross. This is the position of the McKenzie et al.
[5] pole. The Dead Sea Rift is marked with plus signs which
are located from south to north as follows: south end of the
Gulf of Aqaba, north end of the Gulf of Aqaba, south end of
the Dead Sea, north end of the Dead Sea, Sea of Galilee. These
points do not lie on a small circle with respecj to the rotation
pole.
^1 ixTVl
---L \ \ \J\^\ \
—\—L-\- ^
\ vl \ \ ^ ^ \ \ \^-^\
\ jK N^L» \ A-- \ \
\t y —
JL
1 r — i
\
/ \jLr— -\ — \
jN
v
^
'-■--
Fig. 5. Same as Fig. 4, except that the projection pole is at
29.6° N, 20.6° E, which is the rotation pole calculated for the
early stage of opening of the Red Sea. The points along the
Dead Sea Rift now lie close to a small circle about the rotation
pole.
on the amount of flowage of the evaporites from the
main trough, the precise location of the early pole may
be slightly north of its calculated position, therefore
causing the Dead Sea Rift to fall even closer to a small
circle about the pole.
By assuming a two-stage model for the development
of the Red Sea we may therefore find a key to under-
standing the history of the Dead Sea Rift as a plate
boundary. The strike of the fault was determined
during the first stage of spreading in the Red Sea when
most of the motion along the fault consisted of a left-
lateral strike-slip component. Then when the second
stage of spreading began, the fault took on a compo-
nent of rifting. So, today the major source of earth-
quakes along the Dead Sea Rift should be caused by
normal rather than strike-slip faulting. We suggest that
this hypothesis be tested by first motion studies.
It should be noted that several authors have suggest-
ed that a significant amount of left -lateral shear has
occurred along the Dead-Sea Rift during the Quaternary.
Quennell [16] inferred a Pleistocene movement of
45 km from geomorphic features, the most prominent
of which is the shape of the deep depression of the
Dead Sea.
Zak and Freund [17] have recorded horizontal dis-
placements (which are younger than the Lisan Marl -
23,000 years) of 1 50 m along the fault in the Dead
Sea area. However, Freund et al. [18] do not hesitate
to admit that a general agreement has not yet been
reached concerning the Dead Sea Rift's lateral displace-
ment. They refer to Neev and Emery [19] in discussing
the geology of the Dead Sea as accepting the shear
hypothesis, and Picard [20] as not accepting it.
We also refer to Bender [22] who lists several reasons
why he does not accept the shear hypothesis. Along the
entire east side of the rift, he notes that there is over-
whelming evidence of dip-slip movement along hundreds
of faults and fault zones with vertical throws of up to
1000 m. He reports that evidence of lateral displace-
ment (horizontal slickensides, etc.) is very rare (observ-
ed at three places) and in the order of centimeters up
to a few meters. Bender suggests that these minor later-
al movements and some minor folding due to tangental
compression can be explained as secondary structural
phenomena.
Perhaps the confusion throughout the literature con-
cerning movement on the Dead Sea Rift is because both
strike-slip and normal faulting have occurred, but at
353
140
different times. We feel that our model of strike-slip
and then normal faulting is not inconsistent with the
observed data.
The Dead Sea Rift follows approximately a small
circle about the early pole from its intersection with
the Red Sea to as far north as the Huleh Depression
in Lebanon. North of this, the Yamuneh Fault trends
more or less northeasterly, clearly departing from the
small circle. For this reason we postulate that the struc-
tural continuation of the Dead Sea transform fault
follows the Roum Fault which trends north toward
the Mediterranean near Beirut. Dubertret [23] also
suggested a similar continuation, but for different
reasons. He postulated that motion (strike-slip) be
taken up along the Roum Fault rather than along the
Yamuneh because the structures in Lebanon are far
too gentle to accommodate the amount of shortening
necessary to explain the displacement along the Dead
Sea Rift to the south. Of course today the Roum Fault
is not active, but was probably a continuation of the
Dead Sea Rift as a transform fault during the first
stage of spreading in the Red Sea.
One difficulty with our later pole is that, being
south of the Red Sea, a much greater amount of exten-
sion would be expected across the Dead Sea Rift than
is actually observed. The largest amount of extension
appears to be less than 20 km in the Gulf of Aqaba.
One possible solution to this problem is that again,
depending on the amount of flowage of evaporites in
the southern Red Sea, the later pole may be pushed far
enough to the south so that its anti-pole may be north
of, and not far from the Dead Sea Rift. In this case, a
smaller amount of extension along the Dead Sea Rift
than in the Red Sea's axial trough would be expected.
However, it is also possible that some extension has
been taken up in crustal thinning rather than faulting
and could not be ascertained by field investigations
alone.
6. Red Sea spreading rate
Girdler and Styles [2] published a spreading rate for
the axial trough (recent episode) of 0.9 cm/yr. This
rate was computed from a magnetic profile across the
southern Red Sea trending N58°E. However, because
we postulate a second-stage direction that is almost
E-W, we calculate a new spreading rate. We use the
same profile used by Girdler and Styles but have ob-
tained a spreading rate of 1 .0 cm/yr.
7. Fault plane solutions
Fault plane solutions have been previously publish-
ed for three earthquakes that have occurred in the Red
Sea [5,24,25]. The locations of these earthquakes are
shown in Fig. 3 and given in Table 2, and the solutions
are shown in Fig. 6. The earthquake occurring in the
northern Red Sea (a in Fig. 3, Fig. 6A and Table 2)
occurred at the extreme northern boundary of the
axial trough and was generated by a normal fault. Be-
cause of the relative paucity of data for this earth-
quake, it would be possible to draw the fault plane
and the accessory plane so that they were striking
approximately N-S, in agreement with what we would
expect for a normal fault generated within an axial
valley where the motion was E-W. We feel that first
motion data from this earthquake are not inconsistent
with the motion which we propose is happening in the
Red Sea today.
Two of the earthquakes, one in the central and one
in the southern Red Sea (b and c in Fig. 3, Fig. 6B and
C, and Table 2), appear to have occurred on transform
faults, although offsets of the axial trough are not evi-
dent. However, some disagreement exists between
various solutions of these two earthquakes which have
appeared in the literature. For the earthquake in the
central Red Sea, Fairhead and Girdler [24] have obtain-
ed a focal plane with a strike of N53°E and dip 82°SE,
whereas McKenzie et al. [5] published a solution for
the same earthquake and obtained a focal plane with
TABLE 2
First motion studies of earthquakes in the Red Sea
Origin time Lat. Long. Ref.
Northern
Red Sea
Central
Red Sea
Southern
Red Sea
31 Mar. 1969 27.5°N 33.8°E [5,24]
13 Mar. 1967 19.7°N 38.9°E [5,24]
11 Nov. 1962 17.1°N 40.6°E [5,24,25]
354
141
Fig. 6. Focal mechanism for three earthquakes in the Red Sea
(see Table 2 and Fig. 3). The dashed focal planes are those
that have been previously published. A. Solution for earth-
quake on March 31, 1969. B. March 13, 1967. C. November 11,
1962.
magnitude for the East Africa Rift. We use the same
pole of rotation for the Gulf of Aden as presented
by McKenzie et al. [5]. We find that this new pole for
the East Africa Rift lies at 51 .7°S, 3.5°E and that the
magnitude of rotation is 2.65 X 10~7 degree/year. This
pole lies farther to the southwest than that presented
by Girdler and Darracott [1] but lies in the same gener-
al direction with respect to the rift. They refer to seis-
mological studies and gravity anomalies indicating a
tensional stress field of approximately S30°E across
the rift. These geophysical data are as consistent with
our pole as they are with theirs.
strike N68°E and dip 80° SE, a difference of 15° in
the strike. We have no means of knowing which solu-
tion is the more accurate, but point out that a slight
modification to the solution of McKenzie et al. would
give a motion close to the one which we predict.
The earthquake in the southern Red Sea has even
less data (see Fig. 6), and a slight modification to the
fault planes would give good agreement to our predict-
ed motion. In this case, the fault would be a right-
lateral fault along the approximately E— W plane, rather
than the left-lateral motion along the approximately
N— S plane, as proposed by McKenzie et al. [5].
We conclude that the first motion studies of these
three earthquakes are sufficiently inaccurate that it is
impossible to use them to decide whether our proposed
E— W motion is more correct than the NE— SW motion
derived from the total opening of the Red Sea. The
first motion studies fit either hypothesis equally well.
9. Summary
We have shown that the shape of the axial valley
of the Red Sea suggests that recent spreading has been
in an E-W direction, rather than the NE-SW direc-
tion of the earlier phase of spreading. If this is the
case, then the Dead Sea Rift was a plate boundary
with almost pure transform fault motion along it during
the earlier phase of spreading. According to our model,
present-day motion of the Dead Sea Rift is extensional,
and should be marked by predominantly normal fault-
ing. The available earthquake evidence is not capable
of distinguishing between the E— W motion suggested in
this paper and the motions suggested previously. The
rotation pole position calculated for the later phase of
spreading is probably not accurate, because of salt
flowage. However, it can be taken as the northernmost
limit of the true rotation pole.
8. Implications for the East Africa Rift
McKenzie et al. [5] and Girdler and Daracott [1]
have both made use of three-dimensional vector addi-
tion to arrive at a pole and magnitude of rotation for
the East Africa Rift system. This is possible because
the rift is part of a three-plate spreading system; and
by vectorially adding the poles and magnitudes of rota-
tion for the Red Sea (Arabia— Nubia) and the Gulf of
Aden (Arabia— Somalia), a pole and magnitude of
rotation for the East Africa Rift (Nubia-Somalia)
may be obtained.
Because we have obtained a new pole of rotation
for the Red Sea, we can also calculate a new pole and
Acknowledgments
We thank M.M. Ball, R.S. Dietz and F. Nagle for con-
structive criticisms. The maps in Figs. 1,2,4 and 5
were drawn using FORTRAN program HYPERMAP,
written by R.L. Parker, whom we thank. Research
supported by NSF grant GA-42979 from the oceanog-
raphy section and from NOAA.
References
1 R. Girdler and B. Darracott, African poles of rotation,
Comments Earth Sci.: Geophys. 2 (1972) 131-138.
355
142
2 R. Girdler and P. Styles, Two-stage Red Sea floor spread-
ing, Nature 247 (1974) 7-11.
3 A. Laughton, The birth of an ocean, New Sci. 27 (1966)
218-220.
4 D. Roberts, Structural evolution of the rift zones in the
Middle East, Nature 223 (1969) 55-57.
5 D. McKenzie, D. Davies and P. Molnar, Plate tectonics
of the Red Sea and East Africa, Nature 226 (1970) 1-6.
6 R. Freund, Plate tectonics of the Red Sea and East
Africa, Nature 228 (1970) 453.
7 D. Davies and C. Tramontini, The deep structure of the
Red Sea, Philos. Trans. R. Soc. Lond.,Ser., A 267 (1970)
181-189.
8 J.R. Heirtzler, CO. Dickson, E.M. Herron, W.C. Pitman
III and X. Le Pichon, Marine magnetic anomalies, geo-
magnetic field reversals, and motions of the ocean floor
and continents, J. Geophys. Res. 73 (1968) 2119-2136.
9 A.S. Laughton, A new bathymetric chart of the Red Sea,
Philos. Trans. R. Soc. Lond., Ser. A 267 (1970) 21-22.
10 R. Girdler and R. Whitmarsh, Miocene evaporites in Red
Sea cores and their relevance to the problem of the width
and age of oceanic crust beneath the Red Sea, in: Initial
Reports of the Deep Sea Drilling Project 23 (1974)913-
921.
11 R. Coleman, Geologic background of the Red Sea, in:
Initial Reports of the Deep Sea Drilling Project 23 (1974)
813-819.
12 W.C. Pitman III and M. Talwani, Sea-floor spreading in the
North Atlantic, Geol. Soc. Am. Bull. 83 ( 1972) 619-646.
13 T.D. Allan, Magnetic and gravity fields over the Red Sea,
Philos. Trans. R. Soc. Lond., Ser. A 267 (1970) 153-180.
14 J.D. Phillips, Magnetic anomalies in The Red Sea, Philos.
Trans. R. Soc. Lond., Ser. A 267 (1970) 205-217.
15 F.J. Vine, Spreading of the ocean floor: new evidence,
Science 154 (1966) 1405-1415.
16 A.M. Quennell, The structural and geomorphic evolution
of the Dead Sea Rift, Q.J. Geol. Soc. Lond. 114 (1958).
17 1. Zak and R. Freund, Recent strike slip movements along
the Dead Sea Rift, Israel J. Earth Sci. 15 (1966) 33-37.
18 R. Freund, Z. Garfunkel, I. Zak, M. Goldberg, T. Weissbrod
and B. Denn, The shear along the Dead Sea Rift, Philos,
Trans. R. Soc. Lond., Ser. A 267 (1970) 107-130.
19 D. Neev and K.O. Emery, The Dead Sea, depositional
processes and environments of evaporites, Bull. Geol. Surv.
Israel 41 (1967) 147.
20 L. Picard, Thoughts on the graben system in the Levant,
Geol. Surv. Can. Paper 66-14 (1966) 22-32.
21 L. Picard, On the structure of the Rhinegraben with com-
parative notes on Levantgraben features, Israel Acad.
Sci. Hum. 9 (1968) 34.
22 F. Bender, The shear along the Dead Sea Rift: discussion,
Philos. Trans. R. Soc. Lond., Ser. A 167 (1970) 127-129.
23 L. Dubertret, Remarques sur le fosse de la Mer Morte et
ses prolongements au nord jusqu'au Taurus, Rev. Geogr.
Phys. Geol. Dyn. 9 (1967) 3-16.
24 D. Fairhead and R. Girdler, The seismicity of the Red Sea,
Gulf of Aden and Afar Triangle, Philos. Trans. R. Soc.
Lond., Ser. A 267 (1970) 49-74.
25 L. Sykes, local mechanism solutions for earthquakes along
the world rift system, Bull. Seis. Soc. Am. 60 (1970) 1749-
1752.
356
36
Reprinted from: Earth and Planetary Science Letters, Vol. 30, No. 2, 173-175,
Earth and Planetary Science Letters, 33(1976)173 175
© Klsevicr Scientific Publishing Company, Amsterdam Printed in The Netherlands
173
[4]
OPENING OF THE RED SEA WITH TWO POLES OF ROTATION REPLY
E.S. RICHARDSON K2 and C.G.A. HARRISON '
1 Rosensticl School oj Marine and Atmospheric Science, University of Miami, Miami. Fla. (USA)
2NUAA - Atlantic Oceanographic and Meteorological Laboratories, Miami, Fla. (USA)
Received August 2. 1976
We welcome the comments by Girdler and Styles
1 1 | concerning our recent paper [2]. However, the
reason for their first argument evades us.
Girdler and Styles state that the bathymetry of the
Red Sea's axial trough is wider in the south than in
the north, and that the extension in the Dead Sea Rift
is even less than that in the Red Sea.
However, it is obvious that the axial trough from
Laughton's bathymetric chart of the Red Sea [3] (de-
fined by the 500-fm contour) is narrower in the south-
ern Red Sea than in the north. (We have examined some
of the seismic reflection profiles published in the lit-
erature [4,5] and have found that in the majority of
profiles, the slope break on the walls of the axial
trough occurs quite close to the 500-fm depth). With
this in mind, we were not surprised to calculate a pole
of rotation for the later stage of opening to the south
of the Red Sea.
Two criteria, however, caused us to doubt the lati-
tudinal accuracy of this pole:
( 1 ) The Dead Sea rift has undergone a maximum
amount of extension of less than 20 km (Gulf of
Aqaba).
(2) Girdler and Whitmarsh [6] and Coleman [7]
have reason to believe that lateral flow of evaporites
into the axial trough has occurred in the southern Red
Sea, causing the axial trough to be narrower than the
amount of sea floor created during the later stage of
spreading.
These two lines of evidence point to a pole of ro-
tation to the north of the Dead Sea Rift. And we made
note of this in our orginal paper, on page 1 40 (which
Girdler and Styles obviously did not see).
We stated that depending on the amount of evapo-
rite flowage into the axial trough, the later pole may
be pushed far enough to the south so that its anti-pole
may be north of, and not far from, the Dead Sea Rift.
Our example was that, if the axial valley in the
south has been narrowed by 12 km due to salt flow-
age, the rotation pole should be shifted from a lati-
tude of about 1 5°S to a latitude of about 70°S. This
is only an example, as Girdler and Whitmarsh [6]
show evidence that the narrowing of the axial trough.
in the south is much more than 12 km. They note the
presence of Miocene evaporites overlying oceanic
crust with a magnetic age of 2.5 m.y. Assuming an
average sea-floor spreading rate of 1.0 cm/yr. we esti-
mate that the flowage of evaporites is close to 25 km
(12 km on either side). This would push the pole of
rotation so far south that its antipole would approach
55°N.
Girdler and Styles [1] object to our description of
the motion of Arabia away from Nubia as clockwise.
We wish simply to point out that an anticlockwise
rotation about a pole requires a clockwise rotation
about its anti-pole. Since the pole we describe is lo-
cated south of the Red Sea, it would naturally require
a clockwise rotation to moce Arabia away from Nubia.
The situation is illustrated in Fig. 1.
The other major point to which Girdler and Styles
address themselves is that of northeast-southwest fea-
tures within the axial trough. They go as far as to mea-
sure 67 azimuths from magnetic, gravity, bathymetry,
and interpretation maps to prove their point of recent
northeast-southwest motion. However, they note that
due to the small width of the axial trough their mea-
surements "are not very accurate" and "give rise to
large errors". We agree with this conclusion. In addi-
357
174
Or RED SEfl
Fig. 1 . Movement of pole of rotation to account for 25 km of
salt flowage. Note that anti-pole moves to a position north of
the Red Sea. Rotation of Arabia from Nubia is clockwise
about the pole, when the motion is viewed from outside the
earth, but anti-clockwise about the anti-pole.
tion, we note that none of the data referred to gives
conclusive evidence of transform motion within the
axial trough except the fault-plane solutions from
Fairhead and Girdler [8]. Here, strike-slip motion is
indicated. But because of scanty data, a precise azi-
muth of the nodal planes cannot be determined, as
already discussed in our paper. We have checked care-
fully all the references used by Girdler and Styles [1]
to make their measurements of azimuths, except that
by Backer et al. [9]. To our surprise we found that in
most cases the original authors made no suggestion
that these features were transform faults. Searle and
Ross [5] did suggest that the magnetic anomalies stud-
ied by them could be best explained by northeast-
southwest motion, but other interpretations are also
possible. Phillips [10] suggested three possible models
for the magnetic anomalies he studied in the Red Sea.
He was unable to choose between these models ex-
cept on the basis of other evidence for directions of
motion. This other evidence was the direction estab-
lished from earthquake first-motion studies, which we
have already discussed in our paper, and which we
have suggested do not make great constraints on the
actual motion because of the rather poor recording of
earthquakes in this region. One of the models suggest-
ed by Phillips [10] was one in which there was east-
west motion between Arabia and Nubia, in agreement
with our model.
Several authors have made note of magnetic linea-
tions striking N60°E and N70°E [5,10-12]. In fact,
Allan [13] makes special mention of five earthquake
epicenters which are aligned in an east-west direction
and show remarkable coincidence with his postulated
offsets in the axial trough. He states that this is con-
vincing proof of a tranform fault in this region.
Girdler and Styles [14] suggest that there was a ces-
sation in spreading in the Red Sea of about 30 m.y.
Even though the later stage of spreading took place at
the same geographic location as the earlier stage (in
the center of the Red Sea) it would be surprising in-
deed if the new direction of spreading were along ex-
actly the same azimuth as the old direction, as the
direction of spreading is controlled by processes un-
derlying the lithosphere. If the processes ceased for
30 m.y., there is no reason to believe that the move-
ments woul regenerate in the same direction as be-
fore.
Our conclusions are that the shape of the axial
trough suggests an east-west movement of Arabia
away from Africa (as shown by super-position of the
500-fm contours). Problems of salt flowage, however,
preclude the calculation of an accurate latitude for
the pole of rotation, whereas the meridian of the rota-
tion pole is much more accurately known.
Research supported by the National Science Foun-
dation and by NOAA.
References
1 R. Girdler and P. Styles, Opening of the Red Sea with
two poles of rotation - some comments, Earth Planet.
Sci. Lett. 33 (1976) 169-172.
2 E.S. Richardson and C.G.A. Harrison, Opening of the Red
Sea with two poles of rotation, Earth Planet. Sci. Lett. 30
(1976) 135-142.
3 A.S. Laughton, A new bathymetric chart of the Red Sea,
Philos. Trans. R. Soc. Lond., Ser. A, 267 (1970) 21-22.
4 J.D. Phillips and D.A. Ross, Continuous seismic reflexion
profiles in the Red Sea, Philos. Trans. R. Soc. Lond., Ser.
A, 267 (1970) 143-152.
5 R.C. Searle and D.A. Ross, A geophysical study of the
Red Sea axial trough between 20° S and 22° N, Geophys.
J.R. Astron. Soc. 43 (1975) 555-572.
6 R. Girdler and R. Whitmarsh, Miocene evaporites in Red
Sea cores and their relevance to the problem of the width
and age of oceanic crust beneath the Red Sea, in: Initial
Reports of the Deep Sea Drilling Project 23 (U.S. Govern-
ment Printing Office, Washington, D.C., 1974) 913-921.
358
175
R. Coleman, Geologic background of the Red Sea, in: Ini-
tial Reports of the Deep Sea Drilling Project 23 (U.S.
Government Printing Office, Washington, D.C., 1974)
813-819.
D. Fairhead and R. Girdler, The seismicity of the Red Sea,
Gulf of Aden, and Afar Triangle, Philos. Trans. R. Soc.
Lond., Ser. A, 267 (1970) 49-74.
H. Backer, K. Lange and H. Richter, Morphology of the
Red Sea Central Graben (Valdivia Enzschlamme A & B,
Preussag).
J.D. Phillips, Magnetic anomalies in the Red Sea, Philos.
Trans. R. Soc. Lond., Ser. A, 267 (1970) 205-217.
1 1 J.D. Phillips, J. Woodside and CO. Bowin, Magnetic and
gravity anomalies in the Central Red Sea, in: Hot Brines
and Recent Heavy Metal Deposits in the Red Sea, E.T.
Degens and D.A. Ross, eds. (Springer-Verlag, New York,
N.Y., 1969)98-113.
F.K. Kabbani, Geophysical and structural aspects of the
central Red Sea valley, Philos. Trans. R. Soc. Lond., Ser.
A, 267 (1970) 89-97.
T.D. Allan, Magnetic and gravity fields over the Red Sea,
Philos. Trans. R. Soc. Lond., Ser. A, 267 (1970) 153-181
R.W. Girdler and P. Styles, Two-stage Red Sea floor
spreading, Nature 24 7 ( 1 9 74 ) 7 1 1 .
12
13
14
359
37
Reprinted from: Earth and Planetary Saienae Letters, Vol. 30, No. 1,
74-75.
Earth and Planetary Science Letters, 30 ( 1 976) 109-116
© Elsevier Scientific Publishing Company, Amsterdam - Printed in The Netherlands
109
[6]
ASYMMETRIC FRACTURE ZONES AND SEA-FLOOR SPREADING
PETER A. RON A
NOAA, Atlantic Oceanographic and Meteorological Laboratories, Miami, Fla. (USA)
Received August 12. 1975
Revised version received January 29, 1976
An asymmetric pattern is observed in the orientation of minor fracture zones about the axis of the Mid-Atlantic
Ridge at five sites where relatively detailed studies have been made between latitudes 22°N and 51°N. The minor
fracture zones intersect the axis of the Mid-Atlantic Ridge in an asymmetric V-shaped configuration. The V's point
south north of the Azores triple junction (38°N latitude) and point north south of that junction.
The rates and directions of sea-floor spreading are related to the asymmetric pattern of minor fracture zones at
the sites studied. Half-rates of sea-floor spreading averaged between about 0 and 10 m.y. are unequal measured per-
pendicular to the ridge axis. The unequal half-rates of spreading are faster to the west north of the Azores triple
junction and faster to the east south of that junction. The half-rates of sea-floor spreading calculated in the directions
of the asymmetric minor fracture zones are equal about the ridge axis within the uncertainty of the direction deter-
minations.
A discrepancy exists between minor fracture zones that form an asymmetric V about the axis of the Mid-Atlantic
Ridge, and major fracture zones that follow small circles symmetric about the ridge axis. To reconcile this discrepancy
it is proposed that minor fracture zones are preferentially reoriented under the influence of a stress field related to
interplate and intraplate motions. Major fracture zones remain symmetric about the Mid-Atlantic Ridge under the
same stress field due to differential stability between minor and major structures in oceanic lithosphere. This inter-
pretation is supported by the systematic variation in the orientation of minor fracture zones and the equality of sea-
floor spreading half-rates observed about lithospheric plate boundaries.
1. Introduction
A discrepancy is becoming apparent between the
overall symmetry of the Atlantic ocean basin and asym-
metry of both topography and sea-floor spreading about
the Mid-Atlantic Ridge. The overall symmetry of the
Atlantic (Fig. 1) was first recognized from bathymetric
profiles along widely spaced tracklines that revealed
the nearly median position of the Mid-Atlantic Ridge
[1], the nearly mirror image distribution of physiograph-
ic provinces about the ridge axis [1], the trajectories
of major fracture zones which follow small circles
symmetric about the axis of the Mid-Atlantic Ridge
[2,3], and the sequences of remanent magnetic anom-
alies attributed to polarity reversals that indicate a
grossly similar history of sea-floor spreading in the
eastern and western basins [4,5].
Recent work summarized in Table 1 reveals asym-
metry of both topographic features (Fig. 1) and half-
rates of sea-floor spreading about the axis of the Mid-
Atlantic Ridge at sites where relatively detailed in-
vestigations have been made between major fracture
zones. A problem exists in reconciling the overall
symmetry of the Atlantic ocean basin with the asym-
metry revealed by the recent work.
2. Asymmetric topography of the oceanic ridge crest
Valleys with intervening ridges oriented transverse
to the rift valley have been delineated where relatively
detailed bathymetric surveys have been made at sites
along the crest of the Mid-Atlantic Ridge (Fig. 1). The
spacing between the transverse valleys ranges between
360
110
TABLE 1
Relation between topography and sea-floor spreading on the Mid-Atlantic Ridge
Reference
Topography
Azimuth of transverse
valleys
Average
Sense of off-
and ridges
spacing be-
set of axis of
bounding
latitude on
azimuth of
tween trans-
MAR at trans-
lithospheric
plates
MAR
axis of MAR
Side of MAR
verse valleys
(km)
verse valleys
W
E
[33]
America and
61-62°N
033°
280°
_
30
left lateral
Eurasia
61-62°N
033°
280°
-
30
left lateral
61-62°N
033°
280°
-
30
left lateral
[30]
America and
Eurasia
47-51°N
335 and
360°
280-295°
080-
090°
30
left lateral
47-51°N
335 and
360°
280-295°
080-
090°
30
left lateral
[10,23,30]
America and
45_46°N
019°
285 ± 10°
087 ±
10°
30
left lateral
Eurasia
45-46°N
019°
285 ± 10°
087 ±
10°
30
left lateral
left lateral
Azores triple
function
[8,9,11,
America and
36-37°N
018°
270 ± 10°
108°
50
right lateral
24,25]
Africa
36-37°N
018°
270 ± 10°
108°
50
right lateral
36-37°N
018°
270 ± 10°
108°
50
right lateral
[6,7,26,27]
America and
25-27°N
025°
265°
115°
55
right lateral
Africa
25-27°N
025°
265°
115°
55
right lateral
25-27°N
025°
265°
115°
55
right lateral
[12]
America and
22-23°N
020°
260 ± 10°
110 ±
10°
50
left lateral
Africa
22-23°N
020°
260 ± 10°
110 ±
10°
50
left lateral
22-23°N
020°
260 ± 10°
110 ±
10°
50
left lateral
[28,34]
America and
6-8°S
350°
231 ± 10°
080c
—
left lateral
Africa
6-8°S
350°
231 ± 10°
080°
-
left lateral
6-8°S
350°
215 ± 10°
080°
-
left lateral
6-8°S
350°
215 ± 10°
080°
-
left lateral
MAR = Mid-Atlantic Ridge; - indicates no data. Italicized values are computed values (this paper).
30 and 55 km at the various sites (Table 1). The trans-
verse valleys and intervening ridges are distinctly de-
lineated by those surveys that include tracklines at
spacings closer than 20 km oriented parallel to the
axis of the Mid- Atlantic Ridge, such as the surveys at
26°N (Fig. 2) [6,7] , and at 36°N [8] . The transverse
features are less distinctly delineated by surveys based
only on tracklines oriented perpendicular to the axis
of the Mid-Atlantic Ridge, as at the other sites (Table 1).
Where distinctly delineated at 26°N (Fig. 2) [6,7]
and at 36°N [8,9], the transverse valleys exhibit the
characteristics of minor fracture zones associated with
small ridge-ridge transform faults. These characteristics
include ridge-ridge offset of the rift valley up to about
20 km, the association of earthquake epicenters with
the zone of offset, the presence of a basin several
hundred meters deep at the intersection of the zone of
offset with the rift valley, and relief of hundreds of
meters between the floors of the transverse valleys and
the crests of the intervening ridges. It is inferred by
analogy with their characteristics at 26°N and 36°N
that the transverse valleys and intervening ridges at the
361
Ill
Amount of offset
of ,i\is o\ MAR .it
transverse valleys
(km)
Ace o\ crust
(m.y. B.P.)
Sea-floor spreading
average half- averaging side of
rate of spread- interval MAR
ing (cm/yr) (m.y.)
Azimuth of spreading
direction (relation to
axis of MAR)
<10
<10
<10
<10
<10
3-7
3-7
0-10
0-10
1.0
/./
1.10
0-10
0-10
0-10
0-10
0-10
vv
w
E
E
w
303° (normal)
280° (oblique)
123° (normal)
070° (normal)
250° (normal)
<10
<10
<10
0-10
0-10
1.10
1.28
1.19 ± 0.10
0-10
0-10
0-10
E
w
E
109° (normal)
289° (normal)
087 ± 10° (oblique)
20
20
20
<10
<10
<10
<10
<10
<10
0-10
0-10
0-10
0-10
0-10
0-10
0-10
0-10
0-10
0-4.5
0-4.5
4.5-10
4.5-10
1.3
1.0 ± 0.1
/./ ± 0.2
1.3
1.1
1.3
7.5
1.4
1.5
0.1
0.1
2.16 ± 0.24
1.89 ± 0.04
1.59 ± 0.24
1.12 ± 0.08
0-10
0-10
0-10
0-10
0-10
0-10
0-10
0-10
0-10
0-4.5
0-4.5
4.5-10
4.5-10
E
W
w
E
W
W
E
W
w
E
W
E
W
108° (normal)
288° (normal)
270 ± 10° (oblique)
115° (normal)
295° (normal)
265° (oblique)
110° (normal)
280° (nearly normal)
260 ± 10° (oblique)
080° (normal)
260° (normal)
080° (normal)
260° (normal)
other sites described (Table 1) are also minor fracture
zones associated with small ridge-ridge transform faults.
The transverse valleys and intervening ridges at each
site intersect the east and west sides of the rift valley
in an asymmetric V-shaped configuration (Fig. 1;
Table 1). The two sides of the rift valley are parallel.
North of the Azores triple junction between 45°N and
46°N the angle of intersection of the transverse valleys
and intervening ridges with the rift valley is oblique
on the east side and nearly normal on the west side
[10]. Between 47°N and 51°N the transverse valleys
and intervening ridges appear to retain the same orienta-
tion as between 45°N and 46°N, but the azimuth of
the axis of the Mid-Atlantic Ridge changes from north-
east to northwest resulting in oblique intersections on
both sides of the rift valley. At both sites north of the
Azores triple junction the V formed by the intersection
of the transverse valleys and intervening ridges with the
two sides of the rift valley points southward (Fig. 1).
South of the Azores triple junction at 36°N (Table 1)
[8,9,1 1 ], at 26°N (Fig. 1 ; Table 1 ) [6,7] , and at 22°N
[12], the angle of intersection of the transverse valleys
362
112
80° W 70° 60° 50° 40° 30° 20° WW 0°
60" N -
- 60° N
- 10° N
10°S
80° W 70
30° 20°
10° W 0°
Fig. 1. Map of the Atlantic ocean basin showing principal litho-
spheric plates, axis of the Mid-Atlantic Ridge, major fracture
zones that form small circles symmetric about the ridge axis,
minor fracture zones that form V-shaped configurations asym-
metric about the ridge axis delineated in areas of relatively de-
tailed investigations (boxes). The configuration of inferred
minor fracture zones in the area at 6°S is predicted rather than
observed, as discussed in the text.
and intervening ridges with the rift valley reverses be-
coming nearly normal on the east side and oblique on
the west side. The V formed by the intersection of the
transverse ridges and intervening valleys at the sites
south of the Azores triple junction points northward
(Fig. 1).
The characteristics of the minor fracture zones de-
scribed are distinct in several respects from major frac-
ture zones. Major fracture zones of the Atlantic like the
Gibbs (latitude 52°N) [13], the Oceanographer (latitude
35°N) [14], the Atlantis (latitude 30°i\ J [15], the
Kane (latitude 24°N) [16], and the Vema (latitude
10°N) [17], follow families of small circles symmetric
about the axis of the Mid-Atlantic Ridge, generally ex-
hibit ridge-ridge offsets of at least 100 km, and are
spaced hundreds of kilometers apart along the ridge
axis.
3. Apparent and true relative rates of sea-floor spread-
ing
Determination of rate and direction ol sea-floor
spreading are related in that the apparent and true
relative rates of spreading are a function ol direction.
The principle method to determine spreading rate is
based on the Vine and Matthews hypothesis [18]. Strips
of crustal material that are alternately magnetized dur-
ing spreading about an oceanic ridge are identified m the
magnetic polarity reversal time scale. Relative hall-rates
of sea-floor spreading may be derived from the distance
between the axis of the oceanic ridge and the identified
magnetic anomaly. The distance measured perpendicular
to the axis of the oceanic ridge yields an apparent rela-
tive half-rate of spreading. The distance measured par-
allel to transform faults and their continuation as frac-
ture zones that may be oblique to the axis of the ocean-
ic ridge, yields a true relative half-rate ol spreading,
because these features indicate the true direction of
relative motion between diverging lithospheric plates
[3,19-21 ]. In the case that the direction of a fracture
zone is perpendicular to the axis of an oceanic ridge
the apparent and true relative half-rates of spreading
are equal.
The apparent relative half-rates of sea-floor spread-
ing determined perpendicular to the axis of the Mid-
Atlantic Ridge at the sites studied are unequal (Table 1 ).
To facilitate comparison between sites and to suppress
shorter period variations [22] the spreading rates are
averaged over the period 0-10 m.y. B.P. The average
apparent relative half-rates of spreading are faster to
the west of the Mid-Atlantic Ridge axis at latitude
45°N north of the Azores triple junction [23] , and are
faster to the east at latitudes 36°N, 26°N, and 6°S
south of that junction [24—28],
The average true relative half-rates of sea-floor
spreading in the directions of the minor fracture zones
were calculated from the average apparent relative
half-rates perpendicular to the ridge axis using simple
trigonometric relations (Fig. 3). At latitude 26°N where
the azimuths of the minor fracture zones are accurately
known (Fig. 2; Table 1 ), the average true relative half-
spreading rates are equal about the ridge axis. At the
other sites at latitudes 45°N, 36°N, and 22°N, where
the azimuths of the minor fracture zones are less ac-
curately known (Table 1 ). the average true relative
half-spreading rates are equal about the ridge axis
363
113
27° al46°00 W
N
45°00 W
44°00 W 27c
46°00 W
45°00 W
44°00 W
Fig. 2. Bathymetric map [7] contoured in hundreds of meters of a 180-km square on the Mid-Atlantic Ridge crest at 26°N latitude
(Fig. 1; Table 1). Sounding tracks are dashed. Depths exceeding 3400 m are shaded to delineate the rift valley and transverse valleys
that trend normal to the east side and oblique to the west side of the rift valley, as sketched in the inset.
within the uncertainty of the azimuth determinations.
The five sites described are the only sites known in
sufficient detail to reveal the systematic variation in
orientation of minor fracture zones and the equality
of average true relative half-rates of sea-floor spreading
about the Mid-Atlantic Ridge. If the relations observed
between the orientation of minor fractures zones and
half-rates of spreading are consistent, then the orienta-
tions of minor fracture zones may be computed from
half-rates of spreading. For example, at the Mid-Atlantic
crest between latitudes 6° and 8°S the apparent relative
half-rates of spreading are known [28] , and orienta-
tions of minor fracture zones are unknown. The pre-
dicted orientations of the minor fracture zones are
computed from the apparent relative half-rates of spread-
ing (Table 1). Detailed studies (line spacing closer than
364
114
Fig. 3. Geometry of spreading normal and oblique to the axis
of an oceanic ridge, to = zero isochronal = isochron at unit
time; /] and l^ = lengths of crust generated in t\ , normal to the
axis of an oceanic ridge; rx and r^ = half-rates of spreading
corresponding to l\ and /j ; '3 = length of crust generated in t\
along a direction defined by angle a oblique to the axis of the
ridge; r$ = half-rate of spreading corresponding to 1$. An angle
a exists such that: I3 = /2/cos a = /) and r$ = rj/cos a = rx .
20 km both perpendicular and parallel to the ridge axis)
are needed at more sites along the Mid-Atlantic Ridge
to test these relations.
4. Discussion
Hypotheses to explain the observations presented
of fracture zones and sea-floor spreading must consider
the various characteristics described. In particular, the
asymmetric V-shaped intersection of the inferred minor
fracture zones with the axis of the Mid-Atlantic Ridge,
the inversion of the V north and south of the Azores
triple junction, the inequality of apparent relative
spreading half-rates determined perpendicular to the
ridge axis, the equality of true relative spreading half-
rates determined parallel to minor fracture zones nor-
mal and oblique to the ridge axis, and the existence
of asymmetric features within the symmetric frame-
work of the Atlantic ocean basin.
Two alternative hypotheses are considered to ac-
count for the observed relations between topography
and sea-floor spreading, as follows:
(1) Original orientation. The asymmetric orienta-
tion of minor fracture zones about the axis of an oce-
anic ridge is produced by asymmetric processes of
development of the oceanic lithosphere. According
to this hypothesis the relative motions of the litho-
spheric plates follow the directions of the asymmetric
minor fracture zones. This hypothesis poses problems
in reconciling asymmetric with symmetric features
of the ocean basin because asymmetric plate motions
at minor fracture zones would be incompatible with
symmetric plate motions at major fracture zones.
(2) Reorientation. The processes of development of
oceanic lithosphere are essentially symmetric and pro-
duce both symmetric minor and major fracture zones
associated with symmetric sea-floor spreading. The
minor fracture zones are continuously reoriented while
the major fracture zones maintain their original orienta-
tions. As a consequence of this reorientation apparent
relative half-rates of spreading determined perpendicu-
lar to the axis of an oceanic ridge are unequal. True
relative half-rates of spreading determined in the direc-
tions of the reoriented minor fracture zones normal
and oblique to the axis of an oceanic ridge are equal.
This hypothesis is supported by the relations between
spreading directions and rates determined at sites
along the Mid-Atlantic Ridge (Table 1), and offers
promise of reconciling the discrepancy between asym-
metric and symmetric features of the Atlantic ocean
basin.
The continuous reorientation of minor fracture zones
according to hypothesis 2 may be caused by the applica-
tion of an external stress field deriving from different
sources, as follows:
( 1 ) Forces related to magma tic processes. These
forces are related to vertical and horizontal magmatic
movements associated with the axial region of an oce-
anic ridge. A type of regional magmatic movement
proposed by Vogt [29] and applied by Johnson and
Vogt [30] to account for V-shaped topography about
the axis of an oceanic ridge depends on the principle
of a geopotential gradient to drive asthenospheric flow
from topographic highs over inferred mantle plumes
such as at Iceland and the Azores. According to their
hypothesis, the V should point in the direction of
flow away from the high as the result of astheno-
spheric flow along and sea-floor spreading about an
oceanic ridge. The Vogt-Johnson hypothesis does not
account for the orientation of the V-shaped topography
365
15
described to the north and south of the Azores because
the V points toward rather than away from the Azores
(Fig. 1 ). Forces related to magmatic processes un-
doubtedly contribute to the stress field, but are con-
sidered secondary rather than primary components.
(2) Forces rehired re tectonic processes. These forces
are related to interplate and intraplate motions and
may be primary components of the stress field in-
ferred to be reorienting the direction of minor fracture
zones and sea-floor spreading along the Mid-Atlantic
Ridge. The role oi' interplate and intraplate forces as
primary components o[' the stress field is supported by
the observation that the orientation of the minor frac-
ture zones and of sea-floor spreading systematically
changes about lithospheric plate boundaries. The
orientation of minor fracture zones and of sea-floor
spreading is different on the two sides of the rift valley
of the Mid-Atlantic Ridge, a divergent plate boundary,
and differs between the America and Eurasia plates
north of the Azores triple junction and the America
and Africa plates south of that junction (Fig. 1 ; Table
1). The Azores triple junction has been a major in-
fluence in the development of the Atlantic at least
since the early Mesozoic opening of the central North
Atlantic [31],
5. Differential stability of symmetric and asymmetric
structures in oceanic lithosphere
The reorientation hypothesis allows the simultaneous
development of small asymmetric structures and large
symmetric structures in oceanic lithosphere. Minor
fracture zones associated with small transform faults
(offset <30 km) behave in an unstable manner at the
relatively slow average half-rates of spreading (<2
cm/yr) prevalent at the Mid-Atlantic Ridge. The minor
fracture zones are continuously reoriented under the
influence of an external stress field as they are gen-
erated by sea-floor spreading about the small transform
faults. Major fracture zones associated with large trans-
form faults (offset >50 km) behave in a stable manner
at relatively slow average half-rates of spreading (<2
cm/yr). The major fracture zones maintain their orien-
tation under the influence of the same external stress
field as they are generated by sea-floor spreading about
the large transform faults. Thickness of lithosphere re-
lated to distribution of isotherms at a transform fault
may be a determinant of the stability of fracture zones
[32]. Asymmetric small structures may then develop
within the overall symmetry of the Atlantic ocean
basin as a consequence of the differential stability
between minor and major fracture zones of the oceanic
lithosphere.
Acknowledgement
1 thank Walter C. Pitman, 111, for a helpful review.
References
1 B.C. Heezen, M. Tliarp and M. Ewing, The floors of the
oceans, I. The North Atlantic, Geol. Soc. Am. Spec. Paper
65 (1959) 122 pp.
2 B.C. Heezen and M. Tharp, Physiographic diagram of the
North Atlantic Ocean, Geol. Soc. Am. Spec. Paper 65
(1968) revised.
3 W.J. Morgan, Rises, trenches, great faults and crustal
blocks, J. Geophys. Res. 73 (1968) 1959.
4 W.C. Pitman, III and M. Talwani, Sea-floor spreading in
the North Atlantic, Geol. Soc. Am. Bull. 83 (1972) 619.
5 G.O. Dickson, W.C. Pitman, III and JR. Heirtzler, Mag-
netic anomalies in the South Atlantic and ocean-floor
spreading, J. Geophys. Res. 73 (1968) 2087.
6 P. A. Rona, R.N. Harbison, B.G. Bassinger and R.B. Scott,
Asymmetrical bathymetry of the Mid-Atlantic Ridge at
26°N latitude (abstract), EOS Trans. Am. Geophys. Union
54 (1973) 243.
7 P. A. Rona, R.H. Harbison, B.G. Bassinger, R.B. Scott and
A.J. Nalwalk, Tectonic fabric and hydrothermal activity
of Mid-Atlantic Ridge crest (26°N), Geol. Soc. Am. Bull.
(1976) in press.
8 R.S. Detrick, J.D. Mudie, B.P. Luyendyk and K.C. Mac-
donald, Near-bottom observations of an active transform
fault (Mid-Atlantic Ridge at 37°N), Nature 246 (1973) 59.
9 I. Reid and K. Macdonald, Microearthquake study of the
Mid-Atlantic Ridge near 37°N, using sonobuoys, Nature
246(1973) 88.
10 P.J. Bhattacharyya and D.I. Ross, Mid-Atlantic Ridge near
45°N, computer interpolation and contouring of bathy-
metry and magnetics, Marine Sciences Directorate, Dep.
of Environment, Ottawa, Mar. Sci. Paper 11 (1972) 9 pp.
1 1 H. Fleming, Naval Research Laboratory, personal com-
munication.
12 T.H. van Andel and CO. Bowin, Mid-Atlantic Ridge be-
tween 22° and 23° north latitude and the tectonics of mid-
ocean rises, J. Geophys. Res. 73 (1968) 1279.
13 H.S. Fleming, N.Z. Cherkis and J.R. Heirtzler, The Gibbs
fracture zone: a double fracture zone at 52°30'N in the
Atlantic Ocean, Mar. Geophys. Res. 1 (1970) 37.
366
116
14 P.J. Fox, A. Lowrie, Jr. and B.C. Heezen, Oceanographer
Fracture Zone, Deep-Sea Res. 16 (1969) 59.
15 J.D. Phillips, B.P. Luyendyk and D.W. Forsyth, Central
North Atlantic plate motions. Science 1 74 (1971) 846.
16 P.J. Fox, W.C. Pitman, III and I . Shephard, CniMal plates
in the Central Atlantic: evidence for at least two poles of
rotation. Science 165 (1969)487.
17 T.H. van Andel, J.B. Corliss and V.T. Bowen, The inter-
section between the Vema Fracture Zone and the Mid-
Atlantic Ridge in the North Atlantic, J. Mar. Res. 25
(1967) 343.
18 F.J. Vine and D.H. Matthews, Magnetic anomalies over
oceanic ridges, Nature 199 (1963) 947.
19 B.C. Heezen and M. Tharp, Tectonic fabric of the Atlantic
and Indian oceans and continental drift. Philos. Trans. R.
Soc. London, Ser. A 258 (1965) 90.
20 D.P. Mckenzie and R.L. Parker, The North Pacific: an
example of tectonics on a sphere. Nature 216 (1 967) 1 276.
21 X. LePichon, Sea-floor spreading and continental drift,
J. Geophys. Res. 73 (1968) 3661.
22 D.L. Anderson, Accelerated plate tectonics, Science 187
(1975) 1077.
23 B.D. Loncarevic and R.L. Parker, The Mid-Atlantic Ridge
near 45°N, XVII. Magnetic anomalies and ocean-floor
spreading. Can. J. Earth Sci. 8(1971) 883.
24 H.D. Needham and J. Francheteau, Some characteristics
of the Rift Valley in the Atlantic Ocean near 36 48' North,
Earth Planetary Sci. Lett. 22 (1974) 29.
25 D. Greenewalt and P. I . Taylor. Deep-tow magnetic mea-
surements across the axial valley ot the Mid-Atlantic
Ridge, .1. Gcoph>s. Res. 79 (1974) 4401.
26 R.K. Lattimore, P. A. Rona and O.I . DeWakL Magnetic
anomaly sequence in the Central Niutli Atlantic. .1. Geo-
phys. Res. 79 (1974) 1207.
27 B.A. McGregor, C.G.A. Harrison, J.W. Lavellc and PA.
Rona, Magnetic anomaly pattern on Mid-Atlantic Ridge
crest at 26''N. J. Geophys. Res. (1976) in press.
28 T.H. van Andel and T.C. Moore, Magnetic anomalies and
sea-floor spreading rates in the northern South Atlantic.
Nature 226 (1970) 328.
29 P.R. Vogt, Asthenosphere motion recorded by the ocean
floor south of Iceland, Larth Planetary Sci. Lett. 13 (1971)
155.
30 G.L. Johnson and P.R. Vogt, Mid-Atlantic Ridge from 4 7"
to 51° north, Geo!. Soc. Am. Bull. 84 (1973) 3443.
31 P. A. Rona and U.S. Fleming, Mesozoic plate motions in
the eastern central North Atlantic. Mar. Geol. 14 (1973)
239.
32 P.R. Vogt, O.F. Avery, E.D. Schneider, C.N. Anderson and
D.R. Bracey, Discontinuities in sea-floor spreading. Tectono-
physics 8 (1969) 285.
33 P.R. Vogt and G.L. Johnson. Seismic reflection survey of
an oblique aseismic basement trend on the Rcykjanes
Ridge, Earth Planet. Sci. Lett. 15 (1972) 248.
34 T.H. van Andel and G.R! Heath, Tectonics of the Mid-
Atlantic Ridge, 6-8° south latitude, Mar. Geophys. Res.
1 (1970) 5.
367
38
Reprinted from: Earth Science Reviews, Vol. 12, No. 1, 74-75.
PLATE TECTONICS AND OIL
Alfred G. Fischer and Sheldon Judson (Edi-
tors), 1975. Petroleum and Global Tecton-
ics. Princeton University Press, Princeton,
N.J., 322 pp., US. $16.50.
Petroleum and Global Tectonics is a col-
lection of nine papers discussing geological
processes relevant to the occurrence of oil
from the point of view of plate tectonics.
The papers, by scientists from universities
and the petroleum industry, were presented
at a symposium held at Princeton University
in 1972 to honor Hollis D. Hedberg.
The papers are meaningfully arranged
and introduced by the editors In an over-
view of plate tectonics, Sir Edward Bullard
points out that while petroleum exploration
is largely concerned with vertical crustal
movements which allow the accumulation
of sediments, plate tectonics is primarily
concerned with horizontal movements. Five
successive papers demonstrate how both ver-
tical and horizontal movements determine
the evolution of sedimentary basins as the
sites of petroleum generation, accumulation
and storage.
W. Jason Morgan contributes theoretical
background on the relation between heat
flow and vertical movements of the oceanic
lithosphere, one of the most thoroughly un-
derstood of the phenomena producing verti-
cal crustal displacements. AG. Fischer dem-
onstrates that the vertical movements of
oceanic lithosphere, combined with horizon-
tal movements of plates, provide a plausible
mechanism for basins that develop on conti-
nental margins; basins developed on conti-
nental interiors remain problematic. Those
basins that originated by rifting contain the
largest volume of prospective sediment.
D.J.J. Kinsman explains their development
by initial uplift and post-rifting subsidence
related to subcrustal temperature and densi-
ty distributions. J.D. Lowell, G.J. Genik,
T.H. Nelson, and P.M. Tucker consider the
evolution of the southern Red Sea as an
example of how structural arching, rifting,
subsidence, and breaching of continental li-
thosphere act to control the occurrence of
petroleum. J.R. Curray synthesizes different
assemblages of marine sedimentary facies
that constitute basins and shows that oroge-
nic histories of these sedimentary assem-
blages follow almost infinite variations in
plate tectonic settings, rather than an invari-
ant geotectonic cycle.
The generation of petroleum is treated
by J.G. Erdman who reviews the processes
by which an organic fraction of sediment
may be transformed into hydrocarbons by
inorganic processes related to temperature,
degree of oxidation, and the mineral matrix.
H.D. Klemme assembles substantial data to
examine relations between hydrocarbon
occurrence and both the tectonics and ther-
mal regime of productive basins. His evi-
dence indicates that basins associated with
significantly higher heat flow located along
continental margins and rift zones provide
optional conditions for the generation, mi-
gration, and accumulation of petroleum.
The papers of this symposium demon-
strate that plate tectonics provides insight to
problems relevant to the occurrence of pe-
troleum including the origin of basins,
sources of sediment, open basins, restricted
marine circulation, basin geometry, basin re-
lations on opposite continental margins, and
thermal regimes within basins. Most of the
198 known giant oil fields discussed by J.D.
Moody in the concluding paper were found
prior to the advent of plate tectonics, but
plate tectonics will play a significant role in
finding the estimated 200 to 300 remaining
giant fields. This book exemplifies the kind
of creative interplay between academe and
industry that results in intellectual and ma-
terial advances. It is worthwhile reading
both for scientists and informed laymen.
Peter A. Rona, Miami, Fla.
368
39
Reprinted from: Geological Society of America, Microform Publication, Vol. 5,
490 p.
Mid-Atlantic Ridge: Selected
Reprints and Bibliography
Edited by
Peter A. Rona
INTRODUCTION
The following articles from publications of the Geological Society of
America are assembled in chronological order and provide perspective of the
development of geological knowledge of the Mid-Atlantic Ridge spanning a
quarter century of research from early studies to the present frontier.
Early bathymetric reconnaissance gradually revealed the regional
morphology of the Mid-Atlantic Ridge (Tolstoy and Ewing, 1949; Tolstoy, 1951).
Cross-sections of the deep crustal structure underlying the Mid-Atlantic Ridge
determined by the two-ship seismic refraction method (Ewing and Ewing, 1959)
are only now being refined by new methods. Groundwork on the regional
distribution of sediment type by coring (Ericson and others, 1961) and of
sediment thickness by seismic reflection profiling (Ewing and others, 1964)
preceeded studies of sedimentary processes at representative sites on the
Mid-Atlantic Ridge (van Andel and Komar, 1969; Ruddiman, 1972). Sampling
of rocks from emerged (Le Maitre, 1962) and from submerged (Quon and Ehlers,
1963; Engel and others, 1965; Switzer and others, 1970; Melson and Thompson,
1973) portions of the Mid-Atlantic Ridge has contributed to recognition
of the distinctive petrology of oceanic rocks, and has stimulated their
comparison with ophiolites (Thayer, 1969; Green, 1970).
369
The designation of the Mid-Atlantic Ridge as a divergent plate boundary
in the theory of plate tectonics has focused research on processes at the
axial region of the ridge. Advances in magnetic interpretation made it possible
to determine the history of generation about the ridge crest of Atlantic
oceanic lithosphere (Pitman and Talwani, 1972). Studies of the thermal regime
of the Mid-Atlantic Ridge are related both to its characteristic profile
(Sclater and Detrick, 1973) and to the petrologic effects of hydrothermal
activity (Anderson, 1972).
Increasing realization of the complexity of axial processes has led to
the concentration of studies at representative areas of the Mid-Atlantic
Ridge crest (Ward, 1971; Johnson and Vogt, 1973; van Andel and others, 1973;
Phillips and others, 1975). Interdisciplinary, cooperative investigations
have been adopted as an effective research approach. These investigations
of the crestal region of the Mid-Atlantic Ridge include work near lat 36°N
by project FAMOUS (French-American Mid-Ocean Undersea Study; Heirtzler and
Le Pichon, 1974), near lat 45°N by Canadian scientist (Loncarevic, this
publication), and near lat 26°N by the Trans-Atlantic Geotraverse (TAG) project
of the National Oceanic and Atmospheric Administration (Rona and others, 1976).
No longer an enigmatic geographic feature, the Mid-Atlantic Ridge is being
studied as the locus of processes that affect the entire Earth.
370
Reprinted from: Marine Geology, Vol. 21, No. 4, M59-M66,
Marine Geology, 21 (1976) M59— M66 M59
© Elsevier Scientific Publishing Company, Amsterdam — Printed in The Netherlands
Letter Section
PATTERN OF HYDROTHERMAL MINERAL DEPOSITION: MID-
ATLANTIC RIDGE CREST AT LATITUDE 26° N
PETER A. RONA
National Oceanic and Atmospheric Administration, Atlantic Oceanographic and Meteor-
ological Laboratories, Miami, Fla. 33149 (U.S.A.)
(Received February 25, 1976; revised and accepted May 20, 1976)
ABSTRACT
Rona, P.A., 1976. Pattern of hydrothermal mineral deposition: Mid-Atlantic Ridge crest
at latitude 26° N. Mar. Geol., 21:M59-M66.
Interdisciplinary studies of the TAG Hydrothermal Field on the Mid- Atlantic Ridge
crest at latitude 26° N reveal two principal depositional patterns of hydrothermal minerals:
(1) A pattern of deposition controlled by physical and chemical processes within the
hydrothermal field. A major process in determining depositional pattern within the
hydrothermal field is inferred to be sealing of talus by deposition of hydrothermal minerals
from solutions discharged through underlying faults at and adjacent to the wall of the
rift valley. The sealing of a given volume of talus is inferred to occur during a period of the
order of 1 • 10 yr, causing successive migrations of the zone of discharge. The resulting
pattern of hydrothermal mineral deposition within the hydrothermal field would be expec-
ted to be a mosaic of deposits overlapping in time and space with a predominantly fault-
controlled trend parallel to the axis of the rift valley.
(2) A pattern of deposition controlled by sea-floor spreading encompassing the entire
hydrothermal field. A linear zone of hydrothermal deposits will extend from an active
depositional locality at a rift valley along the direction of sea-floor spreading depending
both on the continuity of sea-floor spreading and the persistence in time of the special
structural and thermal conditions that concentrate the hydrothermal activity. The special
structural and thermal conditions that have concentrated hydrothermal activity at the
TAG Hydrothermal Field have persisted during sea-floor spreading for at least 1.4 • 10 yr.
INTRODUCTION
Concentrated hydrothermal mineral deposits are known from several local-
ities along divergent plate boundaries, including the Red Sea (Degens and
Ross, 1969), the Galapagos spreading axis (Moore and Vogt, 1976), and the
Mid- Atlantic Ridge at latitudes 36° N (ARCYANA, 1975), 26° N (M.R. Scott
et al., 1974), and 23°N (Thompson et al., 1975). The minerals are deposited
by sub-sea floor hydrothermal convection systems inferred to involve the
circulation of seawater through oceanic crust driven by intrusive heat sources
at sea-floor spreading centers (Spooner and Fyfe, 1973). Knowledge of the
371
M60
depositional pattern of hydrothermal minerals in time and space at localities
along divergent plate boundaries would help to elucidate the nature of sub-
sea floor hydrothermal convection systems and metallogenesis in oceanic
crust (Rona, 1973; Bonatti, 1975). Interdisciplinary studies by the NOAA
Trans- Atlantic Geotraverse (TAG) project of concentrated hydrothermal
mineral deposits on the Mid- Atlantic Ridge at latitude 26° N provide the
basis for a preliminary interpretation of their pattern of deposition (Rona
et al., 1976). Evidence for past and present concentration of hydrothermal
activity including at least a 10-km2 area at the southeast side of the rift valley
has led to designation of this locality as the TAG Hydrothermal Field (Fig.l;
R.B. Scott et al., 1974).
OBSERVATIONS
The bathymetric setting of the TAG Hydrothermal Field is a ridge that
44°55'W
44°30*W
00*N
Fig.l. Bathymetric map (McGregor and Rona, 1975) contoured in hundreds of meters
showing axis (solid line) of rift valley (shaded) of the Mid-Atlantic Ridge at latitude 26 N,
two profiles (X, Y) along which temperature measurements and bottom photographs were
made concurrently, and dredge stations at the southeast wall of the rift vallev (TAG 1972-
13 [72-13], TAG 1973-2A [73-2A] TAG 1973-3A [73-3A] and along a ridge (unshaded)
trending orthogonal to the axis of the rift valley (TO-75AK61-1A [ 75-1 A ], TO-75AK59-1B
[75-1B], TAG 1973-6A [73-6A]). The approximate known area of the TAG Hydrothermal
Field is indicated (dashed line).
372
M61
trends orthogonal to the rift valley (Fig.l). The west end of the ridge forms
the southeast wall of the rift valley. Abundant manganese oxide crusts were
recovered from three dredge stations where the orthogonal ridge forms the
southeast wall of the rift valley (Fig.l; dredge stations TAG 1972-13, TAG
1973-2A, TAG 1973-3A). The hydrothermal origin of the manganese oxide
crusts is evidenced by their rapid rates of accumulation (200 mm per 106 yr)
determined radiogenically, and by their extreme purity of composition (40%
Mn), with only trace quantities of metals other than manganese (M.R. Scott
et al., 1974), occupying the Mn-rich end member of Bonatti's hydrothermal
classification (Bonatti, 1975). Manganese oxide crusts were also present at
one station situated along the crest of the orthogonal ridge (Fig.l; dredge
station TO-75AK61-1A). Manganese oxide crusts were absent at two other
dredge stations along the crest of the orthogonal ridge (Fig.l; dredge stations
TAG 1973-6A, TO-75AK59-1B), where ropy-textured, apparently fresh basalt
with high K2 O content (0.3%) characteristic of off-axial extrusion was recov-
ered (R.B. Scott et al., 1976).
The hydrothermal manganese oxide occurs as a crust on talus of basalt frag-
ments (Fig.2), as veins in the basalt fragments, and as a crust on and matrix in
breccia of altered basalt fragments (Fig.3). The talus and breccia occur along
the inner margins of steps on the southeast wall of the rift valley revealed by
narrow-beam bathymetry and bottom photographs (McGregor and Rona,
1975). The steps are interpreted as the topographic expression of faults that
may act as conduits for hydrothermal solutions.
Two profiles combining water temperature measurements and bottom
photographs were made over steps on the southeast wall of the rift valley
between depths of about 2500 and 3500 m (Rona et al., 1975). A temperature
anomaly about 300 m wide consisting of an increase in ambient temperature
(+0.1°C) and an inversion of potential temperature gradient (0.015°C per m
warming downwards), was measured within 20 m of the sea floor over a talus-
covered step at a depth of about 3000 m on one of the two profiles (Fig.l;
profile X; Rona et al., 1975). The character and geologic setting of this tem-
perature anomaly favor its interpretation as due to convective transfer of
heat by discharge of hydrothermal solutions focussed by faults in the wall of
the rift valley and diffused by a porous and permeable layer of talus overlying
the faults. A temperature anomaly was absent at the second profile (Fig.l;
profile Y) situated 5 km along the step on the wall of the rift valley where the
temperature anomaly was measured on profile A. Bottom photographs reveal-
ed that breccia and pillow lavas are the predominant rock types present along
profile B (McGregor and Rona, 1975, their fig.6).
DISCUSSION
The duration and position of deposition of the manganese oxide crusts
observed at the TAG Hydrothermal Field may be deduced from their rates
of accumulation and the local half-rate of sea- floor spreading (1.3 cm per yr;
373
M62
Fig.2. Bottom photographs (field of view approximately 4 X 6 m) showing talus of basalt
fragments at 3000-m depth along profile X (Fig. 1) on the southeast wall of the rift valley
where a water temperature anomaly was measured (Rona et al., 1975). The camera water
current compass (length 34 cm) is visible suspended 5 m below the camera.
Lattimore et al., 1974). The manganese oxide crust sampled attains a thick-
ness of 42 mm at the southeast wall of the rift valley, 5 km from the axis of
the rift valley (Fig.l; dredge station TAG 1972-13). U-Th dating of the man-
ganese crust shows it to have accumulated at a rate of about 200 mm per
106 yr, with cessation of accumulation about 15 • 103 yr ago (M.R. Scott et
al., 1974, their table lb, fig.3). Assuming a constant rate, the manganese
oxide crust at dredge station TAG 1972-13 began to accumulate about 2 • 10s
yr ago at a position 2 km from the axis of the rift valley and continued to ac-
cumulate during sea-floor spreading nearly up to its present position at the
wall of the rift valley.
The manganese oxide crust at dredge station TO-75AK61-1A, situated
17 km from the axis of the rift valley (Fig.l), consists of two layers. An un-
derlying layer of hydrothermal manganese up to 10 mm thick accumulated
at a rate of about 35 mm per 106 yr (R.B. Scott et al., 1976), during a period
of 3 • 10s yr, at distances between 8 and 12 km from the axis of the rift val-
ley. The deposition of hydrothermal manganese oxide ceased 12 km from
the axis of the rift valley when the underlying basalt was about 1 • 106 yr old.
Then a layer of hydrogeneous ferromanganese oxide up to 3 mm thick accu-
mulated at a rate of about 8 mm per 106 yr on the hydrothermal manganese
(R.B. Scott et al., 1976) during a period of about 4 • 105 yr, at distances
between 12 and 17 km from the axis of the rift valley. The reconstructed
374
M63
Fig. 3. Bottom photograph (field of view approximately 4 X 6 m) showing hydrothermal
manganese oxide crust (upper left and central portion of photograph) on breccia of basalt
fragments at 2600-m depth along profile Y (Fig. 1 ). The camera water current compass
(length 34 cm) is visible suspended 5 m below the camera.
sequence of events indicates that hydrothermal deposits began to accumulate
on the floor of the rift valley and continued to accumulate through uplift
of the floor to form the walls and adjacent orthogonal ridge during a period
of about 8 • 105 yr.
The duration of deposition of the hydrothermal manganese crusts is of the
same order of magnitude at dredge stations TAG 1972-13 (2 • 10s yr) and
TO-75AK61-1A (3 • 105 yr), in spite of the different distances from the axis
of the rift valley at which the accumulation occurred. The similar duration of
accumulation implies the operation of a process that may limit hydrothermal
discharge at a given site adjacent to the rift valley to a period of the order of
1 '10s yr. It has been suggested that development of an impermeable sedi-
ment cover may suppress hydrothermal discharge on an oceanic ridge (Lister,
1972). However, our studies indicate that sediment cover is negligible in the
area of the TAG Hydrothermal Field (Rona et al., 1976).
The observations presented of near-bottom water temperature, bottom
photographs, and petrology, support the hypothesis that hydrothermal dis-
375
M64
charge becomes suppressed at sites within the hydrothermal field when
deposition of hydrothermal minerals seals off a portion of the discharge zone.
Self-sealing by mineral precipitation is a mechanism that has been recognized
to operate in certain geothermal systems on continents (Elder, 1966; Facca
and Tonani, 1967; Helgeson, 1968; Elders and Bird, 1974; Batzle and
Simmons, 1976). The temperature anomaly attributed to convective transfer
of heat occurs where the wall of the rift valley is covered by talus, a porous
and permeable material through which hydrothermal discharge from under-
lying faults can flow (Figs. 1,2; profile X). Hydrothermal manganese oxide
recovered near profile X occurs as a crust on basalt talus (Fig.l; dredge stations
TAG 1972-13 and TAG 1973-3A; Rona et al., 1976, their table 3). A tem-
perature anomaly was absent along profile Y (Fig.l), where a high proportion
of breccia and pillow lava was photographed (Fig.2), materials which are
impermeable to hydrothermal flow. Hydrothermal manganese oxide occurs
as a crust on and matrix in breccia recovered near profile Y (Fig.l; dredge
station TAG 1973-2A; Rona et al., 1976, their table 3). According to this
interpretation, the breccia may fcrm by alteration and cementation of the
talus by concentrated hydrothermal activity.
CONCLUSIONS
A preliminary pattern of deposition of hydrothermal minerals at a locality
along a divergent plate boundary is emerging from interdisciplinary studies of
the TAG Hydrothermal Field. Two principal patterns may be discerned, that
will require testing by more detailed studies at this and other localities:
(1) A pattern of deposition controlled by physical and chemical processes
within a hydrothermal field. A major process is inferred to be sealing of inter-
stices in talus by deposition of hydrothermal minerals from solutions dis-
charged through underlying faults at and adjacent to the wall of the rift valley.
The duration of accumulation of hydrothermal manganese oxide crusts
determined at two sites (Fig.l; dredge stations TAG 1972-13, TO-75AK61-1A),
indicates that the sealing process, inferred to involve the conversion of talus to
breccia, occurs during a period of the order of 1 • 10s yr. As the talus overly-
ing fracture-focussed hydrothermal discharge becomes sealed, the zone of dis-
charge gradually migrates to areas of unsealed talus. The migration of the
hydrothermal discharge zone is probably controlled by the characteristics of
the fracture system that focusses the flow. Consequently, the migration will
follow the direction of faults at and near to the wall of the rift valley, which
are primarily aligned parallel to the axis of the rift valley. Once sealed, the
breccia may be fractured by tectonic forces opening the possibility of another
generation of hydrothermal deposition; however, the fractured breccia
probably would not regain the original porosity and permeability of the talus.
An additional process that may suppress hydrothermal activity within the
area of a hydrothermal field is off-axis intrusive and extrusive volcanism (Rona
et al., 1976). The resulting pattern of hydrothermal mineral deposition within
376
M65
the hydrothermal field would be expected to be a mosaic of hydrothermal
deposits overlapping in time and space, partially covered by extrusive volcanic
rocks, with a predominant fault-controlled trend parallel to the axis of the
rift valley.
(2) A pattern of deposition of hydrothermal minerals controlled by sea- floor
spreading encompassing an entire hydrothermal field. It was previously pro-
posed (Rona, 1973; Rona et al., 1976) that a linear zone of relict hydrothermal
deposits will extend along the direction of sea-floor spreading from an active
depositional locality at a rift valley. The length of the linear zone would de-
pend both on the continuity of sea- floor spreading and the persistence in time
of the special structural and thermal conditions that concentrate the hydro-
thermal activity. The width of the linear zone of relict hydrothermal deposits
would equal the width of the associated hydrothermal field, which is 10 km
in the case of the TAG Hydrothermal Field. The relict hydrothermal manganese
crust recovered 17 km in the direction of sea- floor spreading from the axis of
the rift valley (Fig.l; dredge station TO-75AK61-1A), indicates that the
special structural and thermal conditions that have concentrated hydrothermal
activity at the TAG Hydrothermal Field have persisted during sea- floor spread-
ing for at least 1.4 • 106 yr. Off-axis extrusive volcanism, such as that eviden-
ced at dredge stations TAG 1973-6A and TO-75AK59-1B (Fig.l), may cover
linear zones of hydrothermal deposits that may extend along flow lines of sea-
floor spreading.
REFERENCES
ARCYANA, 1975. Transform fault and rift valley from bathyscaph and diving saucer.
Science, 190: 108—116.
Batzle, M.L. and Simmons, G., 1976. Microfractures in rocks from two geothermal areas.
Earth Planet. Sci. Lett., 30: 71—93.
Bonatti, E., 1975. Metallogenesis at oceanic spreading centers. Annu. Rev. Earth Planet.
Sci., 3: 401—431.
Degens, E.T. and Ross, D.A. (Editors), 1969. Hot Brines and Recent Heavy Metal Deposits
in the Red Sea— A Geochemical and Geophysical Account. Springer, New York, N.Y.,
571 pp.
Elder, J.W., 1966. Heat and mass transfer in the Earth: hydrothermal systems. N.Z.D.S.I.R.
Bull., 169: 115 pp.
Elders, W.A. and Bird, D.K., 1974. Investigations of the Dunes geothermal anomaly,
Imperial Valley, California, II. Petrological studies, presented at the International
Symposium on Water— Rock Interaction of the International Union of Geochemistry
and Cosmochemistry, Prague, 14 pp.
Facca, G. and Tonani, F., 1967. The self-sealing geothermal field. Bull. Volcanol., 30: 271.
Helgeson, H.C., 1968. Geologic and thermodynamic characteristics of the Salton Sea geo-
thermal system. Am. J. Sci., 266: 129.
Lattimore, R.K., Rona, P.A. and DeWald, O.E., 1974. Magnetic anomaly sequence in the
central North Atlantic. J. Geophys. Res., 79: 1207—1209.
Lister, C.R.B., 1972. On the thermal balance of a mid-ocean ridge. Geophys. J. R. Astron.
Soc, 26: 515—535.
McGregor, B.A. and Rona, P.A., 1975. Crest of Mid- Atlantic Ridge at 26°N. J. Geophys.
Res., 80: 3307-3314.
377
M66
Moore, W.S. and Vogt, P.G., 1976. Hydrothermal manganese crusts from two sites near
the Galapagos spreading axis. Earth Planet. Sci. Lett., 29: 349—356.
Rona, P.A., 1973. Plate tectonics and mineral resources. Sci. Am., 229 (1): 86—95.
Rona, P.A., McGregor, B.A., Betzer, P.R. and Krause, D.C., 1975. Anomalous water tem-
peratures over Mid-Atlantic Ridge crest at 26 north latitude. Deep-Sea Res., 22: 611—
618.
Rona, P.A., Harbison, R.N., Bassinger, B.G., Scott, R.B. and Nalwalk, A.J., 1976. Tectonic
fabric and hydrothermal activity of Mid- Atlantic Ridge crest (lat. 26 N). Geol. Soc.
Am. Bull., 87: 661-674.
Scott, M.R., Scott, R.B., Rona, P.A., Butler, L.W. and Nalwalk, A.J., 1974. Rapidly accu-
mulating manganese deposit from the median valley of the Mid- Atlantic Ridge. Geophys.
Res. Lett., 1: 355— 358.
Scott, R.B., Rona, P. A., McGregor, B.A. and Scott, M.R., 1974. The TAG hydrothermal
field. Nature, 251: 301-302.
Scott, R.B., Malpas, J., Rona, P.A. and Udintsev, G., 1976. Duration of hydrothermal ac-
tivity at an oceanic spreading center, Mid-Atlantic Ridge (lat. 26 N). Geology, 4: 233—
236.
Spooner, E.T.C. and Fyfe, W.S., 1973. Sub-sea floor metamorphism, heat and mass trans-
fer. Contrib. Mineral. Petrol., 42: 287—304.
Thompson, G., Woo, C.C. and Sung, W., 1975. Metalliferous deposits on the Mid- Atlantic
Ridge. Geol. Soc. Am. Abstr. Progr., 7: 1297—1298.
378
41
Reprinted from: Proc. of NOAA Marine Minerals Workshop, March 1976, 111-119,
Resource Research and Assessment of Marine Phosphorite
and Hard Rock Minerals
Peter A. Rona
National Oceanic and Atmospheric Administration
Atlantic Oceanographic and Meteorological Laboratories
INTRODUCTION
The National Oceanic and Atmospheric Administration (NOAA) is
involved in six projects related to assessment of marine phosphorite
and hard rock minerals (Table 1) . NOAA involvement constitutes support
through the Sea Grant Program of four of the projects (S-3, S-9, S-28,
S-32) , and actual implementation of two of the projects (M-4 and S-37,
NOAA Metallogenesis) . Brief summaries and a list of publications are
presented for each of the six projects.
PROJECT SUMMARIES
Evaluation and Economic Analysis of Southern California Phosphorites
and Sand-Gravel Deposits (S-3).
The Principal Investigators of this project are Peter J. Fischer
of California State University, Northridge, and Walter Mead of the
University of California, Santa Barbara (Table 1). The project objective
is to make a geological evaluation, integrated with economic and socio-
economic assessment, of offshore and onshore sand and gravel and
phosphorite deposits.
The assessment of the sand and gravel resource potential of the
southern California shelf is nearing completion. The study extends
from the Mexican border north to Point Conception, a distance of 460 km.
Based upon preliminary estimates, the volume of unconsolidated shelf
sediments is 26.5 km^. Economic studies are in progress to determine
which, if any, of these deposits are viable resources.
With regard to phosphorite, a set of maps of the southern
California continental borderland has been completed showing all
available phosphorite resource data.
Undersea Mineral Survey of the Georgia Continental Shelf (S-9) .
The Principal Investigator of this project is John Noakes of the
University of Georgia. The project was completed in 1975, accomplishing
the following :
1. The technique of neutron activation analysis using a
Californium 2 52 neutron source has been applied to both
379
shipboard and In situ identification of elements in seafloor
minerals.
2. Field tests have demonstrated the potential of using a
mobile sled equipped with radiation detection equipment to
locate and differentiate between thorium associated with
heavy mineral deposits and uranium associated with phosphorites,
3. Over 300 miles of Georgia coastal area have been covered
by reconnaissance surveys.
Lake Superior Copper Survey (S-20) .
R. P. Meyer of the University of Wisconsin is the Principal
Investigator of this project which was completed in 197 5. Accomplish-
ments of the project include the following:
1. Five areas adjacent to the copper producing area of the Keweenaw
Peninsula were investigated and were identified as possible target
areas for future development.
2. Bottom-towed and surface-towed resistivity arrays were success-
fully applied to the location of known copper-bearing veins and
sand deposits with high heavy mineral content.
3. An active-source audiomagnetotelluric system with towed
receivers successfully detected conductivity anomalies associated
with known copper-bearing veins.
4. A first-order analytical method was developed to distinguish
resistivity anomalies related to bottom topography from those due
to changes in conductivity.
Marine Lode Minerals Exploration (S-32)
The Principal Investigator of this project is J. R. Moore of the
Marine Research Laboratory of the University of Wisconsin. The project
objective is to provide basic chemical, mineral, and textural explora-
tion clues that will indicate the presence of sub-seafloor lode bodies,
particularly ores of copper, lead, zinc, nickel, and barite. The
project has already received cooperative assistance from Chromalloy
Corp. and ASV Corp. for surveys at industry mining sites at Castle
Island (barite) and Kllamar (copper), Alaska.
Coronado Bank Phosphorite Deposit (M-4)
The Principal Investigator of this project is B. B. Barnes of
the former Marine Minerals Technology Center. The project was completed
380
in 1971, accomplishing the following:
1. A typical marine phosphorite deposit on Coronado Bank offshore
southern California, was investigated to test equipment and techniques for
phosphorite deposit delineation. The investigation included bathymetry,
seismic reflection profiling, bottom photography, and dredging.
2. Areas of Coronado Bank that yielded the nighest percentage of
P20- (nodules) were related to zones of deep weathering, fractures
in the sea floor, and organic activity.
Metallogenesis at Dynamic Plate Boundaries (A-])
In 1972 the NOAA Trans-Atlantic Geotraverse (TAG) project
(P. 'A. Rona, Chief Scientist-) of the Atlantic Oceanographic and
Meteorological Laboratories (AOML) , dredged hydrothermal manganese
oxide crusts frcm the wall of the rift valley of the Mid-Atlantic Ridge
at latitude 26° N. Subsequent multidisciplinary investigations in-
cluding narrow-beam bathymetry, gravity, magnetics, bottom photography,
near-bottom water temperature and chemistry measurements , dredging and
coring revealed both active and relict hydrothermal manganese oxide de-
posits covering at least a 15 km square area, in and adjacent to the
rift valley, that has been designated the TAG Hydrothermal Field.
The TAG Hydrothermal Field is hypothesized to be the discharge
zone of a voluminous sub-seafloor hydrothermal convection system
involving the circulation of seawater through oceanic crust driven
by intrusive heat sources beneath the rift valley. From geochemical
considerations and analogy with ophiolites, such as the Troodos Massif
of Cyprus, massive copper - iron stratabound sulfide bodies, are inferred
to underlie the hydrothermal manganese oxide crusts, although only dissem-
inated sulfides have been sampled to date.
A new NOAA project, Metallogenesis at Dynamic Plate Boundaries
(see A-l) is being proposed to increase understanding of the hydrothermal
process of metal concentration in oceanic crust, to develop exploration
criteria for both active and relict hydrothermal deposits in oceanic
.rust in situ and in ophiolites, and to determine the distribution of
hydrothermal deposits in oceanic crust. The Principal Investigator of
this project is P. A. Rona (AOML, Miami). Ophiolices, slices of oceanic
i rust formed about an oceanic ridge and incorporated into certain islands
and continents are presently accessible to exploitation, and are being
mined for base and precious metals at certain localities such as Cyprus.
1 1
381
TABLE 1.
NOAA Activities in Assessment of Marine
Phosphorite and Hard Rock Minerals
Project Principal
Identification* Investigator (s) Title Term
S-3 P- J- Fischer and Evaluation and economic 1975-76
W. Mead analysis of southern
California's phosphorite
and sand-gravel deposits
S-9 J. Noakes Undersea mineral survey 1970-75
of the Georgia continental
shelf
S-28 R. P. Meyer Lake Superior copper 1971-75
survey
S-32 J. R. Moore Marine lode minerals 1975-78
exploration
M-4 B. B. Barnes Coronado Bank phosphorite 1968-71
deposit
A-l P- A. Rona Metallogenesis at Dynamic 1976 (pursuant
Plate Boundaries to work initio-
in 1972) - 196
* S - Sea Grant Program
* M - Marine Minerals Technology Center
* A - Atlantic Oceanographic and Meteorological Labs, NOAA
114
382
REFERENCES BY PROJECT (TABLE 1)
Evaluation and Economic ^malysis of Southern California's
Phosphorite and Sand-Gravel Deposits (S-3)
Ashley, R. , Berry, R. , and Fischer, P.J., 1975, Geology of
the northern continental shelf of the Santa Barbara
Channel from Gaviota to El Capitan : in, Studies on the
Geology of Camp Pendleton and Western San Diego County,
California, p. 77-79.
Ashley, R. , Berry, P.., and Fischer, P.J., 1976, Geology of
the northern continental shelf of the Santa Barbara Channel
from Gaviota to El Capitan: Journ. cf Sedimentary Petrology,
in press.
Byrd, R. , Berry, R. , and Fischer, P.J., 1975, Quarternary
geology of the San Diego - La Jolla Underwater Park: in,
Studies on the geology of Camp Pendleton and Western San
Diego County, California, p. 77-79 and p. 300.
Drake, D. , Kolpack, R. , and Fischer, P.J., 1972, Sediment
transport on the Santa Barbara - Oxnard shelf, Santa
Barbara Channel, California: in, Swift, D.J. P., and
others, editors/ Shelf sediment transport: Dowden ,
Hutchinson and Ross, Inc., p. 307-331.
Mead, W. J., 1969, and Sorensen, P.E., 1969, A new economic
appraisal of marine phosphorite deposits: Marine Technology
Society, The Decade Ahead.
Mead, W. J., and Sorensen, P. E., 1970, The principal external
costs and benefits of marine mineral recovery: Offshore
Technology Conference, Proceedings, V. 1.
Wilcox, S. , Mead, W., and Sorensen, P.E., 1972, A preliminary
estimate of the economic potential of marine placer mining:
Marine Technology Society, Proceedings.
383
Undersea Mineral Survey of the Georgia Continental Shelf (S-9)
Noakes, J. E. and Harding, J. L.,1971, New techniques on seafloor
mineral exploration: Marine Technology Society, V. 5, No. 6,
p. 41.
Noakes, J. E. , Harding, J.L., and Spaulding, J.D. , 1974, Locating
offshore mineral deposits by natural radioactive measurements:
Marine Technology Society, V. 8, No. 5, p. 36-39.
Noakes, J. E. , Harding, J. L. , Spaulding, J. D. and Fridge, D. S. ,
Surveillance system for sub-sea survey and mineral exploration
Offshore Technology Conference, Paper 2239, p. 909-914.
Noakes, J. E., Harding, J. L. , Spaulding, J. D. , and Hill, J.,
Radioactive monitoring of offshore nuclear power stations:
Offshore Technology Conference, Paper OTC 1988, p. 501-506.
Noakes, J. E., Smithwick, G. , Harding, J. and Kirst, A., 1971,
Undersea mineral analysis with Californium-252 : Proceedings
Am Nuclear Society Meeting, April.
Lake Superior Copper Survey (S-28)
Brzozowy, C. P., 1973, Magnetic and seismic reflections surveys
of Lake Superior: University of Wisconsin, Sea Grant College
Technical Report WIS-SG-74-220, 40 pp.
Goodden, J. J. P., 1973, Surveying the lake floor in search of
underwater copper reserves to revive an ancient mining district,
Keweenaw Peninsula, Northern Michigan: University of Wisconsin -
Madison Marine Research Laboratory, Sea Grant Underwater Minerals
Program, 7 pp.
Goodden, J. J. P., 1974, Sedimentological aspects of underwater
copper exploration in Lake Superior : University of Wisconsin -
Madison, Master's Thesis.
Meyer, R. P., Moore, J. R. and Nebrya, E., 1975, Underwater copper
explorations in Lake Superior II: Specific targets charted
in 1974: Offshore Technology Conference, Paper OTC 2291, 16 pp.
116
384
Moore, J. R. , Meyer, R. P., and Wold, R. J., 1972, Underwater
copper exploration in Lake Superior - prospects mapped in
1971: Offshore Technology Conference, Paper OTC 1648, p. II - .
307-322.
Nebrija, E., Young, C. , Meyer, R. , and Moore, J. R. , 1976,
Electrical prospecting for copper veins in shallow water:
Offshore Technology Conference, in press.
Smith, P. A., and Moore, J. R. , 1972, The distribution of trace
metals in the surficial sediments surrounding Keweenaw Point,
Upper Michigan: International Assoc. Great Lakes Res.,
Proc. 15th Conf. Great Lakes Res., p. 383-393; The University
of Wisconsin Sea Grant College Reprint WIS-SG-73-341.
Thornton, S. E., A shipboard geochemical prospecting technique
for determining copper in Lake Superior sediments: University
of Wisconsin, Sea Grant Underwater Minerals Program, 7 pp.
Tuerkheimer, F. M. , 1974, Copper mining from under Lake Superior:
The legal aspects: Natural Resources Lawyer, Winter issue,
p. 137-155, University of Wisconsin, Sea Grant College Reprint
WIS-SG-74-354.
Marine Lode Minerals Exploration (S-32)
Moore, J. R. , and Welkie, C. W. , 1975, Metal-bearing sediments of
economic interest, coastal Bering Sea: Anchorage, Proc.
Conference of the Alaska Geological Society, April.
Moore, J. R. , and Van Tassel, J., 1976, Exploration research for
marine gold placers: Grantley Harbor - Tuksuk Channel region,
Seward Peninsula, Alaska: Sea Grant Technical Report, in
preparation.
Panel on Operational Safety in Marine Mining, Moore, T. R. ,
Chairman, 1975, Mining in the outer continental shelf and in
the deep ocean: Washington, D.C., National Academy of
Sciences, 119 pp.
117
385
Otjen, R. P., 1975, Texture and composition of surficial sediments
between Cape Home and Rocky Point, Alaska: University of
Wisconsin - Madison,, M.S. Report, 89 pp.
Owen, R. M. , 1975, Sources and depositions of sediments in Chagvan
Bay, Alaska: University of Wisconsin - Madison, Ph. D. Thesis,
201 pp.
Welkie, C J., 1976, Noble metals placer formation: An offshore
processing conduit: University of Wisconsin - Madison, M.S.
Thesis, in preparation.
Coronado Bank Phosphorite Deposit (M-4)
Barnes, B. B. , 1970, Marine phosphorite deposit delineation
techniques tested on the Coronado Bank, Southern California:
Offshore Technology Conference, Paper OTC 1259, p. II - 315-347.
Metallogenesis at Dynamic Plate Boundaries (A-l)
Betzer, P. R. , Bolger, G. W. , McGregor, B. A., and Rona, P. A.,
1974, The Mid-Atlantic Ridge and its effect on the composition
of particulate matter in the deep ocean: EOS (Am. Geophys.
Union Trans.), V. 55, No. 4, p. 293.
McGregor, B. A. and Rona, P. A., 1975, Crest of Mid-Atlantic Ridge
at 26° N: Jour. Geophys. Research, V. 80, p. 3307-3314.
Rona, P. A. ,1973, Plate tectonics and mineral resources: Scientific
American, V. 229, #1, pp. 86-95.
Rona, P. A., Harbison, R. H., Bassinger, B. G. , Scott, R. B. , and
Nalwalk, A. J., 1976, Tectonic fabric and hydrothermal activity
of Mid-Atlantic Ridge Crest (lat. 26° N) : Geol. Soc. Am. Bull.,
V. 87, 661-674.
118
386
Rona, P. A., McGregor, B. A., Betzer, P. R. , and Krause , D. C. , 1975,
Anomalous water temperatures over Mid-Atlantic Ridge Crest at
26° North latitude: Deep-Sea Research, V. 22, p. 611-618.
Scott, M. R. , Scott, R. B. , Rona, P. A., Butler, L.W. , and
Nalwalk, A. J., 1974, Rapidly accumulating manganese deposit
from the median valley of the Mid-Atlantic Ridge: Geophysical
Research Letters, V. 1, p. 355-358.
Scott, R. B., Rona, P. A., McGregor, B. A., Scott, M. R. , 1974,
the TAG hydrothermal field: Nature, V. 251, p. 301-302.
J 19
337
4 2
Reprinted from: Special Volume of 'Annals of the Brazilian Academy of
Sciences. ' Anais Acad. Brasil Ciencies (Suplemento) , Vol. 48, 256-274.
SALT DEPOSITS OF THE ATLANTIC
PETER A. BONA
National Oceanic and Atmospheric Administration
Atlantic Oceanographic and Meteorological Laboratories
15 Rickenbacker Causeway, Miami, Florida 33149 U.S.A.
ABSTRACT
The distribution in space and time of salt
deposits beneath Atlantic continental margins
and the adjacent ocean basin is presented in a
map and synthesized in a table. Criteria for
detecting the salt deposits are defined. The
major features of the distribution of the salt
deposits are summarized. The distribution of
the salt deposits corresponds to the independently
determined history of opening of the North
Atlantic and South Atlantic.
INTRODUCTION
The present paper reviews the occurrence
of salt deposits beneath continental margins and
the adjacent ocean basin around the Atlantic
(Fig. 1). Major features of the distribution in
space and time of Atlantic salt deposits are
deduced from this review.
DISTRIBUTION OF ATLANTIC SALT
DEPOSITS
The distribution of salt deposits of the
Atlantic is illustrated in Figure 1 and synthesized
in Table 1. The location of each salt deposit
and the association of salts present are listed
in Table 1. Only those salt deposits that include
halite (rock salt) are listed.
The mode of occurrence of the salt is speci-
fied in Table 1 as either diapirs or strata. Strata
refers to beds of salt that may be undeformed
or partially deformed. The thickness of salt
present is given where known from physical
evidence (drilling, outcrop, seismic reflection
and /or refraction). The thickness given is that
of stratified deposits and not of diapirs. Theore-
tical computations are not used as evidence for
thickness of a salt deposit. However, it is useful
to recall that theoretical studies indicate that
salt thicknesses of the . order of hundreds of
meters beneath a sedimentary overburden of at
least 600 m are generally necessary to produce
diapirism (Nettleton, 1934; Parker and McDo-
well, 1955).
Evidence for the occurrence of the salt
deposits described in Table 1 derives from the
distinctive physical and chemical properties of
salt given in Table 2. Drilling and outcrops
furnish direct evidence of the presence of salt
deposits. Indirect evidence of the presence of
salt deposits is furnished by the following
methods:
1 . Seismic reflection and refraction measur-
ements based on density-velocity contrasts
between salt and surrounding sediment.
2. Magnetic measurements based on the amag-
netic properties of salt relative to surround-
ing sediment.
3 . Gravity measurements based on the density
differential between the salt and surround-
ing sediment. In practice, the density of
salt deposits varies widely depending on
the mass of associated caprock and other
factors.
An. Acad. bras. Cienc. (1976). 48. (Suplemento)
388
266
PETER A. RONA
60"N
RONA
Fig-. 1 — Distribution of Atlantic salt deposits. T: Tertiary period; K: Cretaceous period; J: Jurassic period;
TR: Triassic period: M: Mississipean (Carboniferous) period; S: Silurian period; DSDP 139, 140, 144; Deep
Sea Drilling Project sites
4. Thermal gradient measurements based on
the high conductivity of salt relative to
surrounding sediment.
5. Salinity and chlorinity gradient measure-
ments of interstitial water in unconsolidated
sediment over salt deposits based on the high
solubility of certain salts, especially halite,
in water. The measured salinity gradients
are given in Table 1 in parts per thousand
(ppt).
Seismic, magnetic, and gravity measurements
alone may be inadequate to unambiguously dis-
tinguish diapirs of salt from diapirs of mud or
igneous rock. Thermal gradient and salinity
measurements used in conjunction with the other
geophysical methods can distinguish salt diapirs
from those of other materials.
Salinity gradients in interstitial water of
unconsolidated sediment have been effectively
used by the Deep Sea Drilling Project to indicate
the presence of both salt diapirs and strata
beneath the seabed (Manheim et al., 1973). As
a consequence of the solubility of halite and
rapid rates of ionic diffusion, vertical salinity
389
SALT DEPOSITS Of THE ATLANTIC
267
n
03
w
"O
&q
73
c
O
in
§
^
S
E-i §
•* <
^ ft
6q aj
2 W"
^
bo
.. dj pq
10 .3, •-
2 B c_
£ 2
s» o -
O w "1
§ -3 ? O
O a H 2
2 t» <U
.— 02 ■*-■
rt s
o) 5
t3 ft
© £
2 J3
oi
b !X
o O
^2S
3
S co
W 2
cii
e
o
>.
CO
CO
c
bo
So
o
0
ttf)
o
CO
CO
co
pcj
il)
33
co
&3
O
ft;
81
S ft o
8.8
Q
5 § £
^ ii 02
2 £
t5
«
+j
*;
t-
0>
•a
>>
o
"3
c
5
o
CO
>>
3
£
a
K
fi
CO
^ S3
£ CO
CO
c
a "2 .2
PL, O
~ 1i W
Ul
ui
H
P
W
W
H
3
w
T3
1-5
O
M
*d
1-5
Ul
a)
CUD
"w
a)
ui
s
J
S
CO
<
01
S
CO
•a
B
CO
£ .5
"O
B
cO
CO
;-
CO
s
sd
CO
- u
CO T-,
3 x
3 -B
a
CO 02
fi
a
o
p
v-<
CO
CO
B
8
0
CO
CO
"co
CO
r?
^
-
Of
13
>>
.-
J=
X
8
CO
■B £
2 "« 2 ?
3 2 "C «
riJo
b" rco ..
B ft
W ft w
3 £
« Eg
cS
a
£
cO
02
03
3
0
CO
ft
-5
S "d p
B «<
° w
° p
s 5 .S
< s g
co oT ,_,"
— t>
¥ '■* 2
S ra •
pq -g
.. -a 3
CX) c ^
CO CO 5"
05 CO
t-i a
>; •- tj a!
co S S «n
— CO ^-t
E c
CO
3 C3 i
co ^ •"
2 g W g
h , g 5
3 ~ B; s
hoS<
T2
B
A
ft
>>
be
1
CO
CO
03*
o
CO
03
o
o
a
5
CO
B
O
CO
o
fi
'~
"co
t^-i
CO
CO
B
O
CJ
w
f-l
'*'■<
J i Hi
bo
B
V
bs
o
>i
B
CO
bo
c
>.
B
CO
CJ
5
CO
s
CO
|
-5
CO
3
fi
■5
cO
Q
CO
C
bo
g
•g
co"
3
03
>0
"3
K
a
X
bo
B
CO
m --^ 2 «
r"C E
CO ^ p
■a * »
S bo
Dh o
H
w
1-5
>
H
3
CO
o
1-5
CO
CO
CO
^z:
CO
CO
C3
rt
o
O
CO
B
iJ
)J
hi
CO
•a
B
CO
22
ft ^2 "S
co p 5°
S 'O
CO
•sis -5
■3 2fi^ o
CO bo — w il
co CO ^
O ft m CO
O g
CO JC
J3 B
S u
CO co
CO .-
M CQ
Q «
390
268
PETER A. RONA
M >
« st
e.
K)
t>
fc
he
Bq
a H
Q
S S
•~ 03
Bq
Q S
Q as
Kl g
s s
^ £
15 0
s
^ fa
OS tt)
0
K) El
0
1O
Bq
O
£
T3
K)
e
CO
«
1$
01
£
.i: cs
O
CO "
O
.2 t-
C
Q "5
Cow
c o
o .2 ■=
1 .2
2-3
CH K ^ bC 03
SI
— 03
a 3 >e
■- 03 >>
■3 ft a
• S S J5
a 01
o" 03 « g
. .0
9 01"
o 01
cu — v ai Ti a
w < W H u w
3
(Li S
w Ch
3 *d
0! >.
a .a
'■? s
O «3
.co a
S *
CD
05
>
£
rH
•c
TS
*
b
CO
a
3
c
0
a
«
a
'£
CJO ,Q
0)
0
a
s
n
£
A
:p
M
a
M
CO 03
- E
Ml O
CO co £
s~
s o
.2 t.
W 0) H->
0
"3
E
a
>
■a
3
0}
a
h- 1
3
>»
c
>,
H
0
03
a
to
0
■a
S
J
c
CO
'53
CO
„
£>
0
CO
0
C
N
0
E
0
O
3
tJj
0
0}
;-
3
0
a
a
O
0
0
O
O
0.
rH 3 C--.
t- X ■>
en 1; en
0>
-J -
tU r-J
13 .. B
< £ **
^ OS 41
is n s
U <- rH O
to H rs
S -S a
M > t> W
T3
C
OJ
—
.v
.-0
03
O
.2
c
CO
B
O
>>
0
CO
8
3
CO
bo
0
0
bo
O
O
*5
ri
03
«
Uj
7,
1>
J-
ft
- -S 'K w &
CO — . co >*
Eco CU co
t« M CM
F< - Q cc 10
a o
CO —
xs B
co .2
co cc b*? 03
8 S
0 fl ~
J 5 S
. " £
a>
CO » _
-
NOB
0
c bo 0
co B £
03
3 O CO
Vh
OOO
0
as uj
Ph o
o « 2
— cm
w CO 03
03 w CO
CO ^
-S Z 3
« « CO J
(-1 ^ (D w
■3 r= &
PS
3
O1
J=
a
CO
03
be
0
d
cd
GQ
c «
■a r
Cu **
r «
« I
•5 £
c
<src .-
^ o
I-
- CJ3
CT3 rH
CO
Oi .
"-1 E
o
co" m
E S c-1
o o fc
Kh2
K So
r< « m
rJ CO
to
8
£
0
s
f/.'
CO
>>
a
«
c
ro
U
a,
Q
co
0
0
•r
en
or
-a
B
H3
CO
c
91
CO 03
5 g s
.i: co
11
0 %
"cO « ft
i § «
10 C3
0) 01
si*
03 tT -p
■^. o d
O r? <!
391
SALT DEPOSITS O F THE ATLANTIC
269
Srt
a ro
3 CO
43 t.
o >,
0)
KB W
■CP .-
o g
3 CO
2 «• f
■«Ka
3 -'3
<j 8 ft
Wl
:0
n3
*« a
a £
T3 j-
C 3
cO >"}
cO
w
b
e
bo
o
cS
cS
u
TA "S ^
| S £ciiSa
§ 3 *
Q a o
S" 2 *
< oj 42
"5 "8 cO
S » in
i-S ,-
CO
a
3 -
ce to
•m u .
a) s ;>>
*3 » •£>
■C -3
a .- 'm
CO 03 to
H s o
2 S I "o
4i q d C
M h c 3
53 §
ho 13
CO J2
o £ ££ f
g m 5 «££
g .5 o « * S
43
,£]
a
0.
cO oi
CO 03
s
3
M §
co
a
2 «s
« cO
>,
♦J q;
+J cy
u
in -~
OT -
® ?
*J CD
J 5
M
o
CO
"S
OJ
cfl
-
CJ
T3
°
c
'm
to
CO
to
3
03
CO
«
U
3
13
EH
ft
1-1
t.
V
3
>>
tO
9
M
js
63 °i
on "5
6J Eh
c
o
o
a
S
In
w
~
0
ft
o
g
3
£•§
3 o •
•S 2
ft o
ho -^
C Ch
O
a ft
a w
1
«9
0
03
'3
-
to
i
3
>
s
0>
£
s|
Q 3
o
CD
s
'0
S
3
CO
>.
03
25
4=
ft
CO
>.
C
>.
i-t ^^
A
42
CO
W)
C
o
—
a
n
s-
ft
03
03
a £
._ 03
co to
73
T3
C
C
CO
CO
03
03
C pi
.& d
.2 t-
b w
.is
ft m
.2 a
to a
■a
a
CO
03
ra
a co
& t;
'5.
.2 2
CO
ft to
5
a co
a «
V T3
a !>.
o
• £ "C
S 3 "O
a w >.
&l' CO
M
W CO
m c °
to 3 £
3
.2 co
OJ CO
to 43
"" CO
<t 73
s a
to £
IS
= B
U *oj
(^ J2
CO co
03
TO to c
V 'u W •—
w *0 3 JS
H C —
T3 ?
a c
O" CO
< 43
392
270
PETER
RONA
Cl
CO
Kl
CO
h
*l
"!
'<
*
'<
i~i
\J
fci
£*
on
uq
h
O)
w
—
_
u
&
w
H
-
ML|
H
cu
Kl
o
,C
CD
3
a,
0)
^
J
N
a
a
hj
N
J
«
Q 1
g
o —
cu A
•1 W
Q £
c
-.;
t3
3
•a
3
aj
£
£
C8
■§
ft
^
C3
CX
HH
c;
to;
^ U) i"
Fh
k)
^-v
cq
CD
fcj
Oi
o
w
M
«
<
^
-J
K|
CQ
P.
i<
fc<
..,
o
a,
£
i!
w
CO
^
o
co
^
•£)
►J
S3 p
O
ik
R!
CO
O
>-
R,
W
BP|
rt
&
5
§,
ci
H
J
N
<
H
<5
o o o o o
t> OJ C4 t>
<N W CC
O m LO ic
lO 00 CO o
h o ^ ir
i? CJ V TT
N CJ OJ H r(
w s
o m m f
5 cs b H
<s o o W
<b -d S js
ffi < O to O t>
393
SALT DEPOSITS OF THE ATLANTIC
271
gradients may develop over halite deposits
through several kilometers of water-saturated
unconsolidated sediment overburden (Manheim,
1970). The horizontal salinity gradients that
develop are approximately equal to the vertical
salinity gradients, so that the distribution of
vertical salinity gradients delineates the hori-
zontal extent of an underlying salt deposit (Ma-
nheim and Bischoff, 1969).
The ages of the salt deposits listed in Ta-
ble 1 are based on various criteria, in order of
increasing reliability:
1. Stratigraphic relations: The stratigraphic
position of salt strata or the base of salt
diapirs in a known stratigraphic sequence.
2. Associated strata: Paleontologic or radio-
genic dating of strata associated with stra-
tified salt or incorporated in salt diapirs.
3. Caprock: Palynologic dating of the caprock
of a salt diapir.
4. Salt: Palynologic dating of the salt of
diapirs or strata.
MAJOR FEATURES OF THE DISTRIBUTION
OF ATLANTIC SALT DEPOSITS
The information presented in Figure 1 and
Table 1 on the distribution of Atlantic salt
deposits is probably incomplete. Salt deposits
are generally detected in their most spectacular
manifestation as diapirs. Extensive areas of
relatively thick salt strata may remain unde-
tected beneath Atlantic continental margins and
the adjacent ocean basin. For example, layers
of competent materials such as carbonates and
basalt flows and sills may suppress diapirism
and mask underlying salt beds. However, major
features of the distribution of Atlantic salt de-
posits may be deduced from Figure 1 and Table
1, as follows:
I. Salt deposits are present along those rifted
portions of the continental margins of
North America, South America, Africa,
and Eurasia that trend nearly perpen-
dicular to fracture zones of the Atlantic
ocean basin.
II. Salt deposits are absent along those
sheared portions of the equatorial conti-
nental margins of South America and
Africa that trend nearly parallel to frac-
ture zones of the Atlantic ocean basin.
III. Salt deposits are absent in the South
Atlantic south of the Rio Grande Rise
and the Walvis Ridge.
IV. The salt deposits of the rifted continental
margins appear to extend continuously in
basins opening seaward from the conti-
nents to the deep ocean basin.
V. The farthest seaward known extent of salt
deposits in the Atlantic is beneath the
lower continental rise off northwest Africa
at least 450 km from the coast, as
predicted from geophysical measurements
(Rona, 1969, 1970) and confirmed by
measurement of salinity gradients ( DSDP
sites 139, 140; Waterman et al., 1972).
Salt deposits beneath the Sao Paulo Pla-
teau extend 700 km seaward from the
coast. According to continental drift re-
constructions of the Mesozoic opening of
the Atlantic (Dietz and Holden, 1970), the
extent of salt deposits off northwest
Africa represents a half- width of opening.
The extent of salt deposits of the Sao
Paulo Plateau represents a full-width of
opening.
VI. Atlantic salt deposits exhibit a systematic
distribution in time and space, as follows:
1. Late Silurian period: Eastern North
America including the Michigan basin.
2. Mississippian period: Northwestern
Atlantic including the Maritimes basin,
Scotian shelf, and Grand Banks.
3. Late Permian period: Northeastern
Atlantic including the North European
basin and the North Sea.
4. Late Triassic and Jurassic periods:
a. North Atlantic including the Grand
Banks, Scotian shelf, Atlantic con-
tinental margin of North America,
Cuba, Bahama Banks, Senegal basin.
Aaiun basin, offshore northwestern
Africa. Essaouira basin, Portugal
basin, Aquitaine basin, North Euro-
pean basin, North Sea, and British
Isles.
b. Gulf of Mexico.
c. Mediterranean region including the
Atlas Mountains and the Algerian
Sahara.
5. Aptian stage of the Cretaceous period:
South Atlantic including the south-
eastern continental margin of South
America (Sergipe Alagoas, Reconcavo,
Espirito Santo, Campos, and Santos
basins), and the southwestern conti-
nental margin of Africa (Mocamedes,
Cuanza, Lower Congo, and Gabon
basins).
394
272
PETER A EONA
6. Miocene epoch of the Tertiary period:
Western Mediterranean Sea.
7. The age and geographic relations of
inferred salt deposits beneath the De-
merara Rise off the northeastern con-
tinental margin of South America
remain problematic.
VII. Salt deposits of two different ages sepa-
rated by intervals of nonsaliferous sedi-
ments are superposed in at least two
regions of the North Atlantic:
1 . Mississippean and Late Triassic through
Jurassic salt deposits are superposed in
the Scotian shelf-Grand Banks region.
2. Late Permian and Late Triassic salts
are superposed in the North European
basin — North Sea region.
VIII. The distribution of Atlantic salt deposits
in space and time (Fig. 1; Table 1) is
genetically related to the independently
determined history of opening of the North
Atlantic in the Late Triassic and Jurassic
(Rona, 1969, 1970; Schneider and Johnson,
1970; Pautot et al., 1970; Olson and
Leyden, 1973), and the opening of the
South Atlantic in the Early Cretaceous
(Belmonte et al., 1965; Campos et al.,
1974).
ACKNOWLEDGEMENTS
I thank Professor F. F. M. de Almeida,
other members of the Organizing Committee,
Professor H. Martin, and the Brazilian Academy
of Sciences for the opportunity to participate in
the International Symposium on Continental
Margins of Atlantic Type. The research was
supported by the National Oceanic and Atmos-
pheric Administration (NOAA) as part of the
Trans-Atlantic Geotraverse (TAG) project.
REFERENCES
ANTOINE, J. W., MARTIN, R., PYLE. T.. and
BRYANT. W. R. — 1974 — Continental margins
of the Gulf of Mexico, in Eurk. C. A., and Drake,
C. L.. Geology of continental margins. New York.
Springer Vet-lag, 1009 pp.. pp. 6S3-G94.
AUDLEY-CHARLES. M. G. — 1970 — Triassic palaeo-
geography of the British Isles. Geol. Soc. London
Quart. Jour., v. 126, pp. 49-90.
AUZENDE. J. M.. EONNIN, J., OLIVET, J. O.,
PAUTOT. G., and MAUFFRET, A. — 1971 —
Upper Miocene salt layer in the Western Medi-
terranean basin. Nature Phys. Sci., v. 230, pp.
&2-S4.
AYMe. J. M. — 1965 — The Senegal salt basin, in
Salt Basins Around Africa. London, Institute of
Petroleum, 112. pp., pp. S3-90.
BALL, M. Jtf., LOCK, W.. HARRISON, C. G. A.,
NAGLE, F., JR., and WILLIAMS, G. J. — 1974 —
Diapirs of the Old Bahama Channel. EOS, Amer.
Geophys. Union Trans., v. 55, pp. 284.
BALL, M, M., DASH. E. P., HARRISON, C. G. A.,
and AHMED, K. O. — 1971 — Refraction seismic
measurements in the northeastern Bahamas (abs.).
EOS, Trans. Amer. Geophys. Union, v. 52, n. 4,
pp. 252.
BARTLETT. G. A., and SMITH. L. — 1971 — Me-
sozoie and Cenozoic history of the Grand Banks
of Newfoundland. Canadian Jour. Earth Sci., v. 8,
n. 1. pp. 65-S4.
BAUMGARTNER, T. R., and VAN ANDEL, T. II, —
1971 — Diapirs of the continental margin of Angola,
Africa. Geol. Soc. America Bull., v. 82, pp. 793-802.
BELL, W. A. — 1958 — Possibilities for occurrences
of petroleum reservoirs in Nova Scotia Dept. Mines,
277 pp.
BELMONTE, Y., HIRTZ, P., and WENGER, R. —
1965 — The salt basins of the Gabon and the
Congo (Brazzaville). A tentative paleogeographic
interpretation, in Salt Basins Around Africa. Lon-
don, Institute of Peti-oleum, 112 pp, pp. 55-74.
BELT, E. S. — 1965 — Stratigraphy and paleogeo-
graphy of Mabou Group' and related Middle Car-
boniferous facies, Nova Scotia, Canada. Geol. Soc.
America Bull., v. 76, pp. 777-802.
EIDGOOD, D. E. T. — 1969 — The distribution and
and diapiric nature of some Nova Scotia evaporites
— a geophysical evaluation. 3rd Symp. on Salt,
Cleveland, Ohio.
BENZ.^A. — 1958 — Relations between oil fields and
sedimentary troughs in Northwest German Basin.
in Weeks, L. D., editor, Habitat of Oil. Tulsa,
Amer. Assoc. Petroleum Geologists, 13S4 pp., pp.
1054-1066.
BIJU-DUVAL, B., LETOUZEY, J., MONTADERT, L.,
COURRIER, P., MUGNIOT, J. F., and SANCHO, J.
— 1974 — Geology of the Mediterranean Sea basins,
in Burk, C. A., and Drake, C. L. editors, The
geology of continental margins. New York, Springer-
Verlag, 1009 pp., pp. 695-721.
BONNARD, E., DEBOURLE, A., HLAUSCHEK, H.,
MICHEL, P., PEREBASKINE, V.. SCHOEFFLER,
J., SERONIE-VIVIEN, R., and VIGNEAUX, M. —
1958 — The Aquitanian basin, southwest France, in
Weeks, L. G., editor, Habitat of Oil. Tulsa, Amer.
Assoc. Petroleum Geologists, 1384 p., pp. 1091-1122.
BRINKMANN, R., and LoGTERS, H. — 1968 — Diapirs
in Western Pyrenees and foreland Spain, in Brauns-
teiw, J., and O'Brien, G. D., Diapirism and diapirs.,
a symposium. Tulsa, Amer. Assoc. Petroleum
Geologists, Memoir 8, 444p., pp. 275-292.
BROGNON, G. P., and VERRIER, G. R. — 1966 —
Oil and geology in Cuanza basin of Angola, Amer.
Assoc. Petroleum Geologists Bull., v. 50, pp. 108-158.
BURK, C. A., EWING, M., WORZEL, J. L., BEALL,
A. O., JR., BERGGREN, W. A.. BUKRY. D..
FISCHER, A. G., and PESSAGNO, E. A., JR. —
395
SALT DEPOSITS OF THE ATLANTIC
273
1969 — Deep-sea drilling' into the Challenger Knoll,
central Gulf of Mexico. Amer. Assoc. Petroleum
Geologists Bull., v. 53, pp. 1338-1347.
BUROLLET, P., and DUMESTRE, A. — 1952 — Le
diapir du Djebel Rhoeuis. Algeria, 19th Interna-
tional Geological Congress, S.2, p. 6.
BUTLER, L. W. — 1970 — Shallow structure of the
continental margin, southern Brazil and Uruguay.
Geol. Soc. America Bull., v. 81, pp. 1079-1096.
CAMPOS, C. W. M., PONTE, F. O, MIURA, K. —
1974 — Geology of the Brazilian continental margin,
in Burk, C. A., and Drake, C. L., editors, The
geology of continental margins. New York, Springer-
Verlag, 1009 pp., pp. 447-461.
CLARK, S. O., JR., editor — 1966 — Handbook of
physical constants: Geol. Soc. America, Memoir 97,
587 pp.
DEMAISON, G. J. — 1965 — The Triassic Salt in
the Algerian Sahara, in Salt basins around Africa.
London, Institute of Petroleum, 122 pp., pp. 91-100.
DIETZ, R. S. and HOLDEN, J. C. — 1970 — Recons-
truction of Pangaea. Breakup and dispersion of
continents, Permian to present. Jour. Geophys. Res.,
v. 75, pp. 4939-4956.
DUPOUY-CAMET, J. — 1953 — Triassic diapiric salt
structures, southwestern Aquitaine basin, France.
Amer. Assoc. Petroleum Geologists Bull., v. 37,
pp. 2348-2388.
EMERY, K. O., UCHUPI, S., PHILLIPS, J. D.,
BOWEN, C. O., BUNCE, E. T., and KNOTT, S. T.
— 1970 — Continental rise off eastern North
America. Amer. Assoc. Petroleum Geologists, v.
54, pp. 44-108.
EVANS, R. — 1970 — Sedimentation of the Mississipian
evaporites of the Maritimes: an alternative model.
Canadian Jour. Earth Sci., v. 7, pp. 1349-1352.
FONSECA, J. I. — 1966 — Geological outline of the
Lower Cretaceous Bahia supergroup, in Van Hinte,
J. E., ed., Proc. 2nd West African Micropaleon-
tological Colloquium, Ibadan, pp. 49-71.
GUSSOW, W. C. — Carboniferous stratigraphy and
structural geology of New Brunswick, Canada.
Amer. Assoc. Petroleum Geologists Bull., v. 37,
pp. 1713-1816.
HODGMAN, C. D., WEAST, R. C, SHANKLAND, R.
S., and SELBY, S. M., editors — 1963 — Handbook
of chemistry and physics. Cleveland. Ohio, The
Chemical Rubber Pub. Co., 3604 pp.
HOWIE, R. O., and CUMMING, L. M. — 1963 —
Basement features of the Canadian Appalachians.
Geol. Surv. Canada, Bull. n. 89, 18 pp.
JUX, V., 1961 — The palynologic age of diapiric and
bedded salt in the Gulf coastal province. Louisiana
Geol. Survey, Geol. Bull. 38, 46 pp.
KEEN, M. J. — 1970 — A possible diapir in the Lau-
rentian Channel. Canadian Jour. Earth Sci., v. 7,
pp. 1561-1564.
KENT, P. E. — 1969 — The geological framework
of petroleum exploration in Europe and North
Africa and the implications of continental drift
hypotheses, in Hepple, P., editor. The exploration
for petroleum in Europe and North Africa. London,
Institute of Petroleum, pp. 3-17.
KING, L. H., and MAC LEAN, B. — 1970 — A diapiric
structure near Sable Island-Scotian shelf. Maritime
Sediments, v. 6, pp. 1-4.
KIRKLAND, D. C, and GERHARD, J. E. — 1971 —
Jurassic salt, central Gulf of Mexico, and its
temporal relation to circum-Gulf evaporites. Amer.
Assoc. Petroleum Geologists Bull., v. 55, pp. 680-686.
KR6MMELBEIN, K, and WENGER, R. — 1966 — Sur
quelques analogies remarquables dans les miero-
faunes cretacees du Gabon et du Bresil Oriental
(Bahia et Sergipe), in Reyre, D., editor. Sedi-
mentary basins of the African coasts, Part 1,
Atlantic Coast. Paris, Association of African Geo-
logical Surveys, 304 pp., pp. 193-196.
LACOSTE, J. — 1934 — Etudes gcologiques dans le
Rif Meridional. Protectorat Repub. Francaise au
Maroc, Serv. Mines, Notes et Mem. 31 and 32, v
vols., 660 pp.
LANDES, K. K., EHLERS, G. M., and STANLEY,
G. M. — 1945 — Geology of the MacKinac Straits
Region. Michigan Geol. Survey, Pub. 44, Geol. Ser.
37, pp. 192.
LANDES, K. K. — 1945 — Preliminary map no 40, The
Salina and Bass Island rocks in the Michigan basin.
U. S. Geol. Survey, Oil and Gas Inv.
LEYDON, R., BRYAN, G., and EWING, M. — 1972 —
Geophysical reconnaissance on African shelf: 2.
Margin sediments from Gulf of Guinea to Walvis
Ridge. Amer. Assoc. Petroleum Geologists Bull.,
v. 56, pp. 682-693.
LEYDEN, R., DAMUTH, J. E., ONGLEY, L. K,
KOSTECKI, J., and VAN STEVENICK, W. —
1975 — Salt diapirs on the Sao Paulo Plateau,
Brazilian continental margin. Geol. Soc. America,
Abstracts with programs, v. 7. n. 7, 1350 pp., pp.
1168-1169.
MAHER, J. C, and APPLIN, E. R. — 1971 — Geologic
framework and petroleum potential of the Atlantic
coastal plain and continental shelf. U. S. Geol.
Survey Prof. Paper 659, 98 pp.
MANHEIM, F. T.. and BISHOFF, J. L. — 1969 —
Geochemistry of pore waters from Shell Oil Com-
pany drill holes in the continental slope of the
northern Gulf of Mexico. Chemical Geol., v. 4, pp. 63.
MANHEIM, F. T., SAYLES, F. L., and WATERMAN.
L. S. — 197.'! — Interstitial water studies on small
core samples, Deep-Sea Drilling Project: Leg 10, in
Worzel, J. L., Bryant, W., ct al., Initial Reports
of the Deep Sea Drilling Project, v. 10. Washington,
D. C, U. S. Government Printing Office, 748, pp.
615-623.
MANHEIM, F. T., and HORN, M. K. — 1968 — Com-
position of deeper subsurface waters along the
Atlantic continental margin. Southeastern Geology,
v. 9, pp. 215-236.
MATTICK, R. E., FOOTE, R. Q., WEAVER, N. L..
and GRIM, M. S. — 1974 — Structural framework
of United States Atlantic outer continental shelf
north of Cape Hatteras. Amer. Assoc. Petroleum
Geologists Bull., v. 58, Part II of II, pp. 1179-1190.
MAUFFRET, A., FAIL, J. P., MONTADERT, L.,
SANCHO. J., and WINNOCK, E. — 1973 — North-
western Mediterranean sedimentary basin from
seismic reflection profile. Amer. Assoc. Petroleum
Geologists Bull., v. 57, pp. 2245-2262.
MC IVER, N. L. — 1972 — Cenozoic and Mesozoic
stratigraphy of the Nova Scotia shelf. Canadian
Jour. Earth Sci., v. 9, pp. 54-70.
396
274
PETER A. RONA
MEYERHOFF, A. A., and HATTEN, C. W. — 1968 —
Diapiric structures in central Cuba, in Braunstein,
J., and O'Brien, G. D., editors, Diapirism and
diapirs, a symposium. Tulsa, Amer. Assoc. Pe-
troleum Geologists, Memoir 8, 444 pp., pp. 315-357.
MONTADERT, L., WINNOCK, E., DELTEIL, J. R.,
and GRAU, G. — 1974 — Continental margins of
Galicia-Portugal and Bay of Biscay, in Burk, C. A.,
and Drake, C. L., editors, The geology of conti-
nental margins. New York, Springer- Verlag, 1009
pp., pp. 323-342.
MURRAY, G. E. — 1968 — Salt structures of Gulf
of Mexico basin — a review, in Braunstein, J.,
and O'Brien, G. D., editors, Diapirism and diapirs,
a symposium. Tulsa, Amer. Assoc. Petroleum
Geologists, Memoir 8, 444 pp., pp. 99-121.
NETTLETON, L. L. — 1934 — Fluid mechanics of
salt domes. Amer. Assoc. Petroleum Geologists Bull.,
v. 18, pp. 1175-1204.
OFFSHORE EXPLORATION STAFFS, AMOCO CA-
NADA PETROLEUM COMPANY LTD., AND
IMPERIAL OIL LTD. — 1974 — Regional geology
of Grand Banks. Amer. Assoc. Petroleum Geolo-
gists, v. 58, Pt. II of II, pp. 1109-1123.
OLSON, W. S., and LEYDEN, R. J. — 1973 — North
Atlantic rifting in relation to Permian-Triassic salt
deposition; in Logan, A., and Mills, D. V. editors,
The Permian and Triassic systems and their mutual
boundary. Canadian Soc. Petroleum Geologists,
Memoir 2, 766, pp., 720-732.
OLSON, W. S. — 1974 — Structural history and oil
potential of offshore area from Cape Hatteras to
Bahamas. Amer. Assoc. Petroleum Geologists Bull.,
v. 58, Part II of II, pp. 1191-1200.
PARKER, T. J., and MC DOWELL, A. N. — 1955 —
Model studies of salt-dome tectonics. Amer Assoc.
Petroleum Geologists Bull., v. 39, pp. 2384-2470.
PAMENTER, B. — 1971 — Sands aplenty in Onondaga
well. Oilweek, 22 November, pp. 22-24.
PAUTOT, G., AUZENDE, J. M., and LE PICHON, X.
— 1970 — Continuous deep salt layer along North
Atlantic margins related to early phase of rifting,
Nature, v. 227, pp. 351-354.
QUEROL, R. — 1966 — Regional geology of the Spanish
Sahara, in Reyre. D., editor, Sedimentary basins of
the African coasts. Part 1, Atlantic coast. Paris,
Association of African Geological Surveys, 304 pp.,
pp. 27-38.
REYRE, D. — 1966 a — Particularites geologiques des
Bassines de l'Ouest African (essai de recapitulation),
in Reyre, D., editor, Sedimentary basins of the
African coasts, Part I, Atlantic coast. Paris,
Association of African Geological Surveys, 304 pp.,
pp. 253-304.
REYRE, D. — 1966 b — Recapitulation generale : Les
formations permiennes, Triassiques, Jurassiques et
Cretacees dans les bassins cotiers Ouest-Africains,
in Reyre, D., editor, Sedimentary basins of the
African coasts, Part I, Atlantic coast. Paris. Asso-
ciation of African Geological Surveys, 304 pp.. 3
tables.
RONA, P. A. — 1969 — Possible salt domes in the
deep Atlantic off North-west Africa. Nature, v.
224, pp. 141-143.
RONA, P. A. — 1970 — Comparison of continental
margins of eastern North America at Cape Hatteras
and northwestern Africa at Cap Blanc. Amer. Assoc.
Petroleum Geologists Bull., v. 54, pp. 129-157.
RYAN, W. B. F., HSU, K. J., ct al. — 1973 — Initial
reports of the Deep Sea Drilling Project, v. XIII,
Parts 1 and 2. Washington, D. C, U. S. Gov't:
Printing Office, 1447 pp.
SANNEMANN, D. — 1968 — Salt-stock families in
northwestern Germany, in Braunstein, J., and
O'Brien, G. D., editors. Diapirism and diapirs,
a symposium. Tulsa, Amer. Assoc. Petroleum
Geologists, Memoir 8, 444 pp., pp. 261-270.
SCHNEIDER, E. D., and JOHNSON, G. L. — 1970 —
Deep-ocean diapir occurrences. Amer. Assoc. Pe-
troleum Geologists, v. 54, pp. 2151-2169.
SCIENTIFIC PARTY — 1975 — Basins and margins of
the eastern South Atlantic, Leg 40 of the Deep
Sea Drilling Project. Geotimes, v. 20, n. 6, pp. 22-24.
SHERIDAN, R. E. — 1974 — Atlantic continental
margin of North America in Burk, C. A., and
Drake, C. L., editors, The geology of continental
margins. New York, Springer-Verlag, 1009 pp..
pp. 391-407.
SOCIeTe CHERIFIENNE DES PeTROLES — 1966 —
Le bassin du sud-ouest Marocain, in Reyre, D.,
editor, Sedimentary basins of the African coasts,
Part I, Atlantic coast. Paris, Association of
African Geological Surveys, 304 pp., pp. 5-12.
SWAIN, F. M. — 1952 — Ostracoda from wells in
North Carolina. 2. Mesozoic ostracoda. U. S. Geol.
Survey, Prof. Paper 234-B, pp. 59-192.
TEMPLETON. R. S. M. — 1971 — The geology of the
continental margin between Dakar and Cape
Palmas, in Delany, F. M., editor, ICSU/SCOR
Working Party 31 Symposium. Cambridge 1970.
The geology of the East Atlantic continental
margin. 2. Africa. Rep. N» 70/16, Inst. Geol. Sci.,
209 pp., pp. 44-60.
VIANA, C. F. — 1971 — Revisao estratigrafica da
Bacia do Reconcavo-Tucano. Bol. Tec. Petrobras,
Rio de Janeiro, v. 14, n. 3/4, pp. 157-192.
VIANA, C. F. — 1966 — Stratigraphic distribution of
ostracoda in the Bahia supergroup (Brazil), in
Van Hinte, J. E., ed., Proc. 2nd West African
Micropaleontogical Colloquium, Ibadan. pp. 240-256.
VON HERZEN, R. P.. HOSKINS, H, and VAN
ANDEL, T. H. — 1972 — Geophysical studies in
the Angola diapirs field. Geol. Soc. America Bull,
v. 83, pp. 1901-1910.
WATERMAN, L. S., SAYLES, F. L., and MANHEIM.
F. T. — 1972 — Interstitial water studies on small
core samples, Leg 13, in Hayes, D. E., Pimm,
A. C, et al.t Initial reports of the Deep Sea
Drilling Project, v. XIV. Washington, U. S. Go-
vernment Printing Office, 975 pp., pp. 753-762.
WATSON, J. A., and JOHSON, G. L. — 1970 —
Seismic studies in the region adjacent to the Grand
Banks of Newfoundland. Canadian Jour. Earth Sci.,
v. 7, pp. 306-316.
WEBB, G. W. — 1973 — Salt structures east of
Nova Scotia. Earth Science Symposium on Offshore
Eastern Canada, Geol. Survey Canada, Paper 71-23,
pp. 197-218.
WINNOCK, E. — 1973 — Expose sussinct de revo-
lution paleogeologique de l'Aquitaine\ Bull. Soc.
G6ol. France, 7eSerie, v. XV, pp. 5-12.
ZIEGLER, P. A. — 1975 — Geologic evolution of the
North Sea and its tectonic framework. Amer. Petro-
leum Geologists, v. 59, pp. 1073-1097.
397
43
Reprinted from: Journal of Oaean Management, Vol. 3, 57-78,
Ocean Management, 3 (1976) 57—78
© Elsevier Scientific Publishing Company, Amsterdam — Printed in The Netherlands
Energy and Mineral Resources of the Pacific Region
in Light of Plate Tectonics
Peter A. Rona * and Lawrence D. Neuman **
ABSTRACT
The Pacific is a closing ocean basin that is diminishing in size as it is consumed at
convergent plate boundaries around three-fourths of its perimeter. Geothermal energy
sites, areas of offshore petroleum potential, deposits of precious, base, iron and ferro-
alloy metals are distributed along the convergent plate boundaries of the Pacific including
the surrounding continents. The energy and mineral resources of the Pacific region are
concentrated by geologic processes at the convergent plate boundaries.
INTRODUCTION
The Pacific is a region of geologic diversity. The Pacific region encom-
passes the largest ocean basin on earth, extensive chains of volcanic islands
that follow arcuate trends around the western margin of the Pacific Ocean,
and seas occupying marginal basins between the island arcs and eastern Asia
(Fig. 1). The theory of plate tectonics has gained wide scientific acceptance
during the past five years, and offers a conceptual framework to unify the
diverse geological phenomena of the Pacific region. The conceptual frame-
work of plate tectonics is leading to a new understanding of the relation be-
tween the geology and the distribution of energy and mineral resources of the
Pacific region.
An earlier paper treated principles of the relation between plate tectonics
and mineral resources (Rona, 1973). This paper aims to apply these princi-
ples to develop a basic understanding of the distribution of energy and min-
* National Oceanic and Atmospheric Administration (NOAA), Atlantic Oceanographic
and Meteorological Laboratories, Miami, Fla. 33149, U.S.A.
** Office for Ocean Economics and Technology, United Nations, New York, N.Y. 10017,
U.S.A.
57
398
eral resources of the Pacific region. The approach taken is first to present
salient features of the geology of the Pacific region from the point of view of
plate tectonics (Figs. 1—3). Then to view the distribution of selected energy
and mineral resources of the Pacific region with respect to the geology in a
series of maps (Figs. 4—9) compiled from various sources (Van Roy an and
Bowles, 1952; ECAFE, 1962, 1963, 1970; McKelvey and Wang, 1969;
Dr. Peter A. Rona is presently Senior Research
Geophysicist with the NO A A, Atlantic Oceanograph-
ic and Meteorological Laboratories in Miami, Florida
and Adjunct Professor of Marine Geology and Geo-
physics at the University of Miami. He is Chief Scien-
tist of the Trans- Atlantic Geotraverse (TAG), an
international cooperative project to investigate the
Earth's crust along a corridor across the Atlantic be-
tween southeastern North America and northwestern
Africa to gain an understanding of continental drift,
sea-floor spreading, and the occurrence of seabed
minerals. He generally spends several months of the
year at sea leading explorations of the seabed.
Prior to joining NOAA in 1969 he was Exploration
Geologist with Standard Oil Company (N.J.) between
1957 and 1959. From 1960 to 1969 he was with
Columbia University, Hudson Laboratories where he
developed and became Head of Marine Geophysics.
He received the degree of Ph.D. in 1967 from the
Department of Geology and Geophysics at Yale Uni-
versity. He has published about 50 scientific papers
and is a member of 12 professional societies.
Lawrence D. Neuman joined the Ocean Economics
and Technology Office as Scientific Affairs Officer in
1973, specializing in the economic potential of the
sea and the economic development of coastal areas. A
native New Yorker, he received his bachelor's degree
in physics at Columbia College and his doctorate in
geology and marine geophysics at Columbia Univer-
sity's Lamont-Doherty Geological Observatory.
58
399
Anon., 1972; DEMR, 1972; Jones, 1972; Eimon, 1974). Finally, to attempt
to understand the distribution of the energy and mineral resources in terms
of geologic processes (Fig. 10). Attention is focused on those energy and
mineral resources associated with present plate boundaries of the Pacific
region. Other deposits may then be interpreted in terms of past plate bound-
aries following the uniformitarian principle of geology that the present is the
key to the past.
GEOLOGY OF THE PACIFIC
Lithospheric plates
The conceptual framework of plate tectonics, developed by many work-
ers, views the earth as comprised of a rigid outer shell about 100 km (60
miles) thick, the lithosphere, that behaves as if it were floating on an under-
lying plastic layer, the asthenosphere (Fig. 2). The upper, more brittle part
of the lithosphere is termed crust, of which there are two major types, the
granitic continental crust (about 30 km thick) and the basaltic oceanic crust
LITHOSPHERIC PLATES
120*e 150* 180* 150* 120" 90* 60*
DIVERGENT PLATE
BOUNDARY
... CONVERGENT PLATE
BOUNDARY
_ TRANSFORM PLATE
BOUNDARY
__ UNCERTAIN PLATE
BOUNDARY
HALF RATE OF
* * SEA FLOOR SPREADING
(CM/YR)
„ RELATIVE PLATE MOTION
(CM/YR)
„. DIP OF BENIOFF ZONE
(UPPER 100 KM)
Fig. 1. Lithospheric plates of the Pacific region showing directions (arrows) and half-rates
of sea-floor spreading about divergent plate boundaries in the eastern Pacific, directions
(arrows) and rates of convergence at convergent plate boundaries bordering the Pacific
ocean basin, and angles of inclination (in degrees from horizontal) of Benioff zones
beneath the convergent plate boundaries (Fig. 2). (Le Pichon et al., 1973)
59
400
(about 10 km thick). The lithosphere is segmented into a number of major
plates, each of which may encompass a continent and part of an ocean basin,
and numerous minor plates. The Pacific region includes portions of the Pa-
cific, China, America, and Antarctic major plates, and several minor plates
(Fig. 2).
PLATE BOUNDARIES
The boundaries of lithospheric plates are delineated by narrow earthquake
zones where the plates are moving with respect to each other. Two types of
boundaries are considered (Fig. 2). At the first type, a divergent plate bound-
ary, two adjacent plates move apart as new lithosphere is added to each plate
by the process of sea-floor spreading. Divergent plate boundaries extend
around the globe through all the major ocean basins as part of a 65,000 km-
(40,000 mile-) long undersea mountain chain. Divergent plate boundaries of
the Pacific region including the East Pacific Rise are located in the eastern
Pacific ocean basin off South America, Central America, and North America
(Fig. 1). The sea floor is spreading about different segments of the East Pa-
cific Rise at rates ranging between about 1 and 10 cm per year (Fig. 1).
Fig. 2. Diagram showing plate motions at divergent and convergent plate boundaries.
Lithospheric plates move like conveyor belts from a divergent plate boundary (oceanic
ridge) to a convergent plate boundary where they either descend along a Benioff zone
at an oceanic trench (subduction) or they override the adjacent plate (obduction).
60
401
At the second type of boundary, a convergent plate boundary, two adja-
cent plates come together. In the general case, one plate descends under
another plate along an inclined plane (Benioff zone) and is resorbed into the
asthenosphere (subduction; Fig. 2). In the special case, one plate may
temporarily override the other plate (obduction) until the situation reverts
to subduction. The Pacific is bounded on three sides by convergent plate
boundaries marked by oceanic trenches where lithosphere descends along
Benioff zones at rates comparable to the rates of sea floor spreading. As a
consequence of the crustal consumption at the convergent plate boundaries
bounding the Pacific, the Pacific is a closing ocean basin that is diminishing
in size, in contrast to the Atlantic which is an opening ocean basin that is
growing larger. The coexistence of divergent plate boundaries where litho-
sphere is created, and convergent plate boundaries where lithosphere is de-
stroyed, implies that the diameter of the earth is not radically changing.
The inclination of Benioff zones at convergent plate boundaries plays an
important role in the development of basins marginal to continents and the
generation of volcanism. The inclination of a Benioff zone is inversely pro-
portional to the rate of convergence of adjacent plates at a convergent plate
boundary (Luyendyk, 1970). Marginal basins are present in the western
Pacific where the rates of plate convergence are relatively slow and the incli-
nation of the Benioff zone exceeds about 35° (Fig. 1; Karig, 1971; Oxburgh
and Turcott, 1971; Sleep and Toksoz, 1971; Bracey and Ogden, 1972). Mar-
ginal basins are absent in the eastern Pacific where the rates of plate conver-
gence are relatively fast and the inclination of the Benioff zone is less than
about 35°. As will be shown, the presence or absence of marginal basins af-
fects offshore petroleum potential. The inclination of Benioff zones also
affects the composition of igneous rocks (rocks solidified from molten mate-
rial) and associated metal deposits.
AGE OF SEA FLOOR AND OF CONTINENTS
The range and distribution of ages differs markedly between the Pacific
Ocean basin and the surrounding continents (Fig. 3). The age of the Pacific
sea floor, determined by dating of rock samples recovered by the Deep Sea
Drilling Project (Fischer et al., 1971) and by the magnetic polarity reversal
time scale (Pitman et al., 1974), ranges between about 150,000,000 years
(Late Jurassic period) and the present. The distribution of the ages about the
divergent plate boundaries in the eastern Pacific is regular, the sea floor being
youngest adjacent to the boundaries and becoming progressively older away
from the boundaries as a consequence of sea-floor spreading. The distribu-
tion of ages at the convergent plate boundaries around the Pacific is irregu-
61
402
120'E 150* 1B0* 150* 120* 90* 60* 30*^
AGE OF SEA FLOOR
% CENOZOIC
::: MESOZOIC
CONTINENTAL
STRUCTURAL
PROVINCES
% CENOZOIC
• ::: MESOZOIC
v. HERCYNIAN
lllll CALEDONIAN
PRECAMBRIAN
'CONVERGENT PLATE BOUNDARY = DIVERGENT PLATE BOUNDARY
Fig. 3. Age of the Pacific sea floor and of continental structural provinces. Lithospheric
plate boundaries are shown.
lar, the age of the sea floor varying along the boundaries as a consequence of
subduction.
The circum-Pacific continents are divided into structural provinces accord-
ing to the time of their most recent deformation during mountain building
episodes (Fig. 3). Structural provinces are predominantly Cenozoic (0—
70,000,000 years) in western South America, and Mesozoic (70,000,000—
200,000,000 years) in western North America. The island arcs of the western
Pacific are Cenozoic (0—70,000,000 years). Structural provinces of eastern
Asia and Australia exhibit a more complex distribution spanning nearly the
entire age range of the earth. The ages of most circum-Pacific metal deposits
range from Mesozoic through Cenozoic (0—200,000,000 years ago), corre-
sponding to the known history of lithospheric plate motions in the Pacific
Ocean basin.
DISTRIBUTION OF SELECTED ENERGY RESOURCES
Geothermal Energy
Geothermal phenomena including active volcanos, thermal springs, fuma-
roles, geysers, and high values of heat flow are distributed along convergent
62
403
GEOTHERMAL ENERGY
120'E ISO" 180" 150" 120" 90" 60" 30* V
'.'■>//////! 1H\\\
■ GEOTHERMAL ENERGY SITES
* ACTIVE VOLCANOES
• THERMAL SPRINGS
AND FUMAROLES
:SS NUMEROUS THERMAL SPRINGS
a GEYSERS
HEAT FLOW
•.••V MEAN HEAT FLOW > 3.0
mi MEAN HEAT FLOW > 2.0
m MEAN HEAT FLOW < 2.0
III! MEAN HEAT FLOW < 1.0
CONVERGENT PLATE BOUNDARY
DIVERGENT PLATE BOUNDARY
Fig. 4. Map of geothermal energy of the Pacific region. Lithospheric plate boundaries are
shown.
plate boundaries around the Pacific (Fig. 4; Kennedy and Richey, 1947;
Waring, 1965; Karig, 1971; Snead, 1972). These geothermal phenomena re-
sult from heating due to mechanical factors (friction), chemical reactions
(dehydration), and to the internal heat of the earth, as the lithospheric plates
descend into the asthenosphere at convergent plate boundaries. Similar geo-
thermal phenomena are inferred to occur along the East Pacific Rise and the
divergent plate boundaries of the Pacific region where heat flow is high, as a
consequence of the upwelling of magma during sea-floor spreading (Lang-
seth, 1969;Sclater, 1972).
Geothermal energy is being tapped at sites in the western United States,
Japan, and New Zealand. In addition to its direct utilization, geothermal
energy drives hydrothermal processes involving the circulation of hot solu-
tions through the lithosphere which act to concentrate metals both at diver-
gent and convergent plate boundaries, as will be discussed. Hydrothermal
mineral deposits, that is, mineral deposits precipitated from hot aqeous solu-
tions, constitute a major part of useful metallic ores on continents and may
be important in ocean basins.
63
404
Organic energy: petroleum
Areas of offshore petroleum potential conform with convergent plate
boundaries around the Pacific (Fig. 5; McKelvey and Wang, 1969). Both the
circum-Pacific trenches and the island arcs of the western Pacific create a
habitat that is favorable for the accumulation of petroleum in several re-
spects. The trenches and island arcs act as barriers that catch sediment and
organic matter from the continent and ocean basin. Deep-sea sediment with
variable content of organic matter is continuously transported into the
trenches on a conveyor belt of spreading sea floor (Sorokhtin et al., 1974).
The island arcs divide the ocean basin into marginal basins such as the South
China Sea, the East China Sea, the Yellow Sea, the Sea of Japan, the Sea of
Okhotsk, and the Bering Sea. The shape of the trenches and marginal basins
acts to restrict the circulation of the ocean, so that oxygen is not replenished
in the seawater and the organic matter is preserved. Geothermal heat in the
trenches and marginal basins may facilitate the conversion of organic matter
to petroleum (Fig. 4; Tarling, 1973; La Plante, 1974). Finally, geological
structures that develop as a result of deformation of the sediment in the
trenches and marginal basins by tectonic forces form traps that favor the
accumulation of petroleum. In contrast to the areas of offshore petroleum
ORGANIC ENERGY
150° 180* 150" 120* 90° 60" 30* \
PETROLEUM PRODUCING
AREAS
ONSHORE PETROLEUM
POTENTIAL
OFFSHORE PETROLEUM
POTENTIAL
SEDIMENTARY ROCKS
%& CRYSTALLINE ROCKS
CONVERGENT PLATE BOUNDARY = DIVERGENT PLATE BOUNDARY
Fig. 5. Areas of petroleum potential and production of the Pacific region (adapted from
Rona and Neuman, 1974, 1975; after McKelvey and Wang, 1969).
64
405
potential, the sedimentary basins from which petroleum is produced on con-
tinents around the Pacific exhibit no apparent spatial relation to plate bound-
aries (Fig. 5; Irving et al., 1974; Rona and Neuman, 1974).
DISTRIBUTION OF SELECTED MINERAL RESOURCES
Metal deposits at divergent plate boundaries
Knowledge of the distribution of metal deposits with respect to divergent
plate boundaries is limited because, as submerged oceanic ridges, these
boundaries are less accessible to observation than convergent plate bounda-
ries. This knowledge is necessary to evaluate the metallic mineral potential of
oceanic lithosphere. All oceanic lithosphere is generated by sea-floor spread-
ing about oceanic ridges and underlies all ocean basins which cover two-
thirds of the earth. The Red Sea and the Atlantic Ocean provide evidence of
the nature of processes that may be concentrating metals in oceanic litho-
sphere of the Pacific.
Evidence from the Red Sea and the Mid-Atlantic Ridge indicates that
hydrothermal processes are concentrating metals in oceanic crust at diver-
gent plate boundaries. The Red Sea represents the earliest stage in the
growth of an ocean basin, the stage when a divergent plate boundary rifts a
continent in two. About five years ago the richest known submarine metallic
sulfide deposits were found in basins along the center of the Red Sea at a
depth of about 2,000 m (6,600 ft.) below sea level (Degens and Ross, 1969).
The sulfide minerals, in which various metals are combined with elemental
sulfur, are disseminated in sediments that fill the basins to a thickness esti-
mated between 20 m (66 ft.) and 100 m (330 ft.). The top 10 m (33 ft.) of
sediment, which has been explored by coring the largest of the basins, has a
total dry weight of about 80 million tons, with average metal contents of
29% iron, 3.4% zinc, 1.3% copper, 0.1% lead, 0.005% silver, and 0.00005%
gold (Bischoff and Manheim, 1969). The deposits are saturated with (and
overlain by) salty brines carrying the same metals in solution as those present
in the sulfide deposits. The salty brines are considered to be the hydrother-
mal solutions from which the sulfide minerals are precipitated.
The most advanced growth stage of a divergent plate boundary is the
oceanic ridge system including the Mid -Atlantic Ridge and the East Pacific
Rise. An active area of submarine hydrothermal mineral deposits, the TAG
Hydrothermal Field, was recently discovered at the crest of the Mid- Atlantic
Ridge (26°N) by the Trans-Atlantic Geotraverse (TAG) project of the Na-
tional Oceanic and Atmospheric Administration (NOAA). As the first of its
kind discovered, the distribution of such hydrothermal fields along oceanic
65
406
ridges is unknown, but it is suspected that the TAG Hydrothermal Field may
represent an important class of features.
The TAG Hydrothermal Field includes both active and relict areas (Rona
et al., 1976). The active area (15X15 km), including the east wall of the rift
valley between depths of 2000 and 3500 m, is covered by a discontinuous
layer of manganese oxide at least 5 cm (2 in.) thick (R. Scott et al., 1974;
McGregor and Rona, 1975), that is being deposited by hydrothermal solu-
tions enriched in various metals (Betzer et al., 1974). The hydrothermal so-
lutions emanate as hot springs from fractures in the ocean bottom (Rona et
al., 1975). The relict area comprises hydrothermal material that was deposit-
ed in the active area adjacent to the rift valley and transported at least tens
of kilometers away from the ridge crest on a conveyor belt of spreading sea
floor (Rona, 1973). A hydrothermal origin for the metallic oxide present is
indicated by its chemical purity (40% manganese with only trace quantities
of iron and copper compared with manganese nodules which generally con-
tain about 10% manganese and appreciable quantities of iron and copper),
and rapid rate of accumulation (about 200 mm per 1,000,000 yr. which is
about one hundred times faster than manganese nodules) (M. Scott et al.,
1974). The TAG Hydrothermal Field not only confirms that metals are con-
centrated in normal oceanic crust by hydrothermal processes, but indicates
that such processes may occur at a divergent plate boundary more-or-less
continuously from early (Red Sea) to advanced (Mid-Atlantic Ridge) stages
of growth.
Sediment samples directly overlaying the basalt that forms the foundation
of the Pacific and other ocean basins recovered by the Deep Sea Drilling Pro-
ject both at and away from oceanic ridges, reveal widespread enrichment by
certain precious, base, iron and ferro-alloy metals (Bostrom and Peterson,
1969; Dymond et al., 1970; Von der Borch and Rea, 1970; Von der Borch
et al., 1971; Cook, 1972; Piper, 1973; Sayles and Bischoff, 1973; Dasch,
1974). The observation that the enrichment is limited to sediment in the
basal layer directly overlying basalt indicates that it occurred soon after the
generation of the underlying basalt by sea-floor spreading about an oceanic
ridge. The metal enrichment is ascribed to hydrothermal processes similar to
those that produced the metalliferous sediments in the Red Sea and the
metallic oxides at the TAG Hydrothermal Field. The concentration of metals
in the widespread enriched sediments of the Pacific is only a fraction of that
observed in the Red Sea, but higher concentrations may exist locally.
Processes of metal concentration at divergent plate boundaries
A model of metallogenesis at divergent plate boundaries, based on various
lines of evidence (Spooner and Fyfe, 1973), considers that certain precious,
66
407
base, iron and ferro-alloy metals may be concentrated as deposits by sub-sea
floor hydrothermal convection systems involving the circulation of seawater
as a hydrothermal solution through rocks to a depth of about 5 km (Hart,
1973) beneath the ocean bottom. The development of such hydrothermal
convection systems is favored by the supply of seawater, heat, and the in-
tensely fractured basaltic rocks at divergent plate boundaries. According to
the model, cold, dense seawater descends through fractures in the basalt of
an oceanic ridge and is heated by contact with hot, intrusive bodies of mag-
ma (molten rock material) and rock that upwell to form new lithosphere at
the ridge crest. The warm, less dense seawater rises through the features and
leaches metals disseminated in the basalt that are then transported in solu-
tion as complexes with chlorides in the seawater. A fraction of the metals in
solution combines with sulfur in the seawater and precipitates to form mas-
sive statiform bodies of metallic sulfide including copper and iron, possibly
associated with gold. It is suspected that such copper-iron sulfide bodies may
underlie the TAG Hydrothermal Field, but it is technically infeasible at
present to drill into the ocean bottom to test this idea (Rona, 1973; R. Scott
et al., 1974; Rona et al., 1976). Metallic oxides, like the manganese oxide at
the TAG Hydrothermal Field, precipitate under oxidizing conditions as the
hydrothermal solutions discharge from the ocean bottom in hot springs.
Amorphous particles of ferric hydroxide precipitate from the hydrothermal
solutions in the overlying seawater. The ferric hydroxide scavenges the re-
maining metals from solution and settles to deposit a layer of metalliferous
sediment on basalt of the ocean bottom, like the metalliferous sediments ob-
served in the Red Sea and the Pacific Ocean.
Metal deposits at convergent plate boundaries
Precious-metal deposits including gold, silver, and platinum are distributed
along convergent plate boundaries around the Pacific Ocean (Fig. 6). In the
eastern Pacific precious-metal deposits occur landward of convergent plate
boundaries along the western margins of North America and South America.
In the western Pacific, precious-metal deposits occur on island arcs situated
along convergent plate boundaries including Japan, the Philippines, and In-
donesia. Deposits are also present in eastern Asia and Australia, where they
are separated by a gap from the active convergent plate boundaries.
The distribution of light-metal deposits including aluminum, beryllium,
lithium, and titanium appears unrelated to plate boundaries of the Pacific
(Fig. 7). These metals are associated with granitic rocks of the continents
that are compositionally different from the basaltic rocks of the ocean basin.
The distribution of light-metal deposits on the circum-Pacific continents is
67
408
PRECIOUS METAL DEPOSITS
A SILVER
• COLD
i PLATINUM
CONVERGENT PLATE BOUNDARY =• DIVERGENT PLATE BOUNDARY
Fig. 6. Map of precious-metal deposits of the Pacific region (adapted from Rona and
Neuman, 1974, 1975). Lithospheric plate boundaries are shown.
LIGHT METAL DEPOSITS
120°E 150" 180° 150* 120* 90* 60* 30**
■ ALUMINUM
* BERYLIUM
a LITHIUM
• TITANIUM
CONVERGENT PLATE BOUNDARY = DIVERGENT PLATE BOUNDARY
Fig. 7. Map of light-metal deposits of the Pacific region. Lithospheric plate boundaries
are shown.
68
409
BASE METAL DEPOSITS
120°E 150° 180" 150° 120" 90° 60° 30"W
* ANTIMONY
• COPPER
' LEAD
d MERCURY
" TIN
o ZINC
* * * 0 CONVERGENT PLATE BOUNDARY
.DIVERGENT PLATE BOUNDARY
Fig. 8. Map of base-metal deposits of the Pacific region (adapted from Rona and Neuman,
1974, 1975). Lithospheric plate boundaries are shown.
related to the occurrence of particular minerals in granitic rocks and to the
concentration of these minerals by processes of subaerial weathering.
Base-metal deposits including antimony, copper, lead, mercury, tin, and
zinc are distributed along convergent plate boundaries around the Pacific
Ocean (Fig. 8), similar to the distribution of precious metal deposits (Fig. 6).
The base-metal deposits occur landward of convergent plate boundaries
along the western margins of the Americas in the eastern Pacific, and on is-
land arcs along convergent plate boundaries of the western Pacific. In south-
east Asia deposits of tin associated with tungsten, fluorite, bismuth, and
molybdenum occur in belts of granites of predominantly Mesozoic age
(70,000,000-200,000,000 years ago). Base-metal deposits also occur in east-
ern Asia and Australia where they are separated by a gap from active conver-
gent plate boundaries.
The copper occurs associated with other metals along convergent plate
boundaries of the Pacific region in two economically important classes of
ore deposits — massive statiform sulfide bodies and porphyry ore bodies.
69
410
Massive statiform sulfide bodies, deposits confined to layers within bedded
sequences of volcanic or sedimentary rocks, are present in western North
America, Japan, and the Philippines (Eimon, 1974) where they range in age
from Paleozoic through Cenozoic (0—500,000,000 years old). Porphyry ore
bodies, disseminated deposits of copper-sulfide minerals associated with vol-
canic rocks, constitute over one-half of the world's copper production. The
majority of porphyry copper deposits lie in two belts of the Pacific region
(Eimon, 1974): 1) the western Americas belt extending from Chile to Alaska
where the deposits are Mesozoic and Cenozoic in age (0—200,000,000 years
old), 2) the southwest Pacific belt including Taiwan, the Philippines, Borneo,
West Siam, New Guinea (Papua), and the Solomon Islands, where the depos-
its are Cenozoic in age (0—70,000,000 years old).
Iron and ferro-alloy metal deposits including chromium, cobalt, manga-
nese, molybdenum, nickel, tungsten, and vanadium are distributed along
convergent plate boundaries around the Pacific (Fig. 9), similar to the distri-
bution of precious (Fig. 6) and base (Fig. 8) metals. Iron and ferro-alloy
metal deposits occur landward of convergent plate boundaries along the
western margins of the Americas in the eastern Pacific, and on island arcs
IRON AND FERROALLOY METAL DEPOSITS
• IRON
D CHROMIUM'
g COBALT
O MANGANESE
* MOLYBDENUM
a NICKEL
■ TUNGSTEN
a VANADIUM
MANGANESE NODULES
UP TO 20% OF BOTTOM
= 20-50% OF BOTTOM
BOUNDARY BETWEEN
RED CLAY & BIOGENIC OOZES
CD COPPER CONTENT IN WEIGHT
1.5-2.0%
© NICKEL CONTENT
1.5-2.0%
'CONVERGENT PLATE BOUNDARY = DIVERGENT PLATE BOUNDARY
Fig. 9. Map of iron and ferro-alloy metal deposits including manganese nodules of the
Pacific region (adapted from Rona and Neuman, 1974, 1975). Lithospheric plate bound-
aries are shown.
70
411
along convergent plate boundaries of the western Pacific. These deposits
also occur in eastern Asia and Australia where the deposits are separated by
a gap from active convergent plate boundaries.
Nodules containing variable percentages of manganese, copper, nickel, and
cobalt are present on about two-thirds of the Pacific sea floor (Fig. 9; Strak-
hov et al., 1968; Horn et al., 1972; Skornyakova and Andrushchenko, 1974).
A zone of nodules of anomalously high copper and nickel content (1.5—
2.0%) and areal density (20—50% of sea floor) extends east- west across the
central Pacific between latitutes 5° and 20° N coinciding with a region of
high biological productivity and may be related to additional concentration
of these metals from seawater by organisms (R.M. Garrels, pers. comm.). No
apparent relation exists between the positions of the plate boundaries and
either the overall distribution of the enriched zone of nodules in the Pacific
Ocean basin. The nodules are formed by hydrogenous processes in which the
metals are precipitated both from seawater and from interstitial water of un-
derlying sediments. The metals are at least partially derived from hydro-
thermal sources.
In the special case of obduction, the upper layer of oceanic lithosphere is
thrust up and overrides an adjacent plate at a convergent plate boundary
(Fig. 2). The slice of oceanic lithosphere, up to several tens of kilometers
thick, may contain the various types of precious, base, iron and ferro-alloy
deposits that were described to form at oceanic ridges (divergent plate
boundaries). Areas of obduction in the southwest Pacific are Papua, New
Guinea (Davies and Smith, 1971), where gold and copper prospects exist
(Eimon, 1974; Grainger and Grainger, 1974), and the island of New Caledonia
(Arias, 1967), where nickel and chromium deposits are mined. The island
arcs of the western Pacific, the Kamchatka Peninsula, and western North
America from Alaska to Baja California are areas of former obduction which
incorporate slices of oceanic lithosphere (Coleman, 1971, fig. 4).
Processes of metal concentration at convergent plate boundaries
The observation that precious, base, iron and ferro-alloy metal deposits
are associated with convergent plate boundaries of the Pacific region (Figs.
6—9) has led to the interpretation of these deposits as genetically related to
plate convergence. Models are being developed to gain an understanding of
the sources of the various metals and the processes that concentrate the
metal deposits.
Prior to the theory of plate tectonics, the source for metals was generally
considered to be anomalous metal concentrations in continental crust and
mantle underlying the deposits (Krauskopf, 1967; Noble, 1970). Plate tec-
71
412
tonics has turned attention to the oceanic lithosphere as a source for a signif-
icant fraction of the metals in deposits at convergent plate boundaries of the
Pacific. Early models stressed metals concentrated by hydro thermal pro-
cesses in particulate phases (metalliferous sediments) and in solid phases
(oxides and sulfides) in oceanic crust as primary sources (Sawkins, 1972; Sil-
litoe, 1972a). However, the amounts of those precious, base, iron and ferro-
alloy metals disseminated in oceanic crust by magmatic processes more than
suffice to quantitatively account for the majority of deposits of these metals
observed along convergent plate boundaries (see NOTES, p. 75).
Adequate sources and supplies of various metals exist to account for the
metal of ore deposits (Krauskopf, 1967). The principal problems in metallo-
genesis are extraction of metals from the sources, transport of the metals,
their concentration and deposition. In simplest form, models of metallo-
genesis at convergent plate boundaries envisage the extraction of metals from
sea water-saturated oceanic crust as it undergoes partial melting under condi-
tions of increasing temperature and pressure during descent of the oceanic
lithosphere along a Benioff zone (Fig. 10; Sawkins, 1972; Sillitoe, 1972a).
The metals ascend as components of magmas, are concentrated in fluids
released from the magmas, and are deposited.
The models are becoming increasingly complex to account for the actual
characteristics of metal deposits at convergent plate boundaries of the Pacific
region (Mitchell and Bell, 1973; Ridge, 1972). The distribution of metal de-
posits parallel to convergent plate boundaries in metal provinces of the west-
ern Americas (Fig. 10) may be related to progressive increase in temperature
and pressure and change in chemical environment down the inclined plane
of the Benioff zone which together act to separate different components of
the oceanic lithosphere during partial melting (Sillitoe, 1972b). Different
associations of metals and igneous rocks may be related to variation in com-
position of magmas controlled by changes in the inclination of Benioff zones
resulting from changes in rates of lithospheric plate convergence and sea
floor spreading through time (Mitchell, 1973). The actual inclinations of
Benioff zones are not constant as shown in models (Fig. 10), but vary with
depth.
Metals other than those present in oceanic crust such as tin, as well as ad-
ditional quantities of metals and sulfur present in oceanic crust, may be
derived from the asthenosphere and continental lithosphere overlying
Benioff zones .(Fig. 10). The proportion of metals and sulfur derived from
the various potential sources is unknown and is the subject of studies using
sulfur, lead, and strontium isotopes as tracers (Corliss, 1974; Dasch, 1974).
Volatile components such as hydrogen fluoride and carbon dioxide liberated
from dry oceanic lithosphere at depths exceeding 200 km along a Benioff
zone may lower melting points, assist in transporting metals, and liberate tin
72
413
and associated metals (tungsten, bismuth, fluoride, and molybdenum) from
granite in overlying continental crust (Mitchell and Garson, 1972; Stern and
Wylie, 1973; Oyarzun and Frutos, 1974). The tin and associated metals in
eastern Asia and the various metal deposits in eastern Australia may have
been deposited above former Benioff zones of shallow inclination adjacent
to the continental margins related to relatively fast plate convergence. Sub-
sequent increase in inclination of the Benioff zones related to relatively slow
plate convergence has resulted in the seaward migration of the Benioff zones
as a consequence of the growth of marginal basins (Fig. 10), leaving the ob-
served gap between the deposits of the continental margins and active con-
vergent plate boundaries of the western Pacific (Mitchell, 1973).
Convergent plate boundaries are the loci of a multiplicity of interacting
geologic processes that are difficult to differentiate. The models incorporate
different processes to explain the factors that control the locations of ore
deposits along the convergent plate boundaries of the Pacific region: (1)
deep processes: variations in sources of metals, physico-chemical mecha-
nisms, magmatic processes, seismic activity, rate and inclination of litho-
spheric descent, and geologic structure associated with subduction along
Benioff zones (Krauskopf, 1967; James, 1971; Sawkins, 1972; Sillitoe,
1972a, 1974; Mitchell, 1973); (2) shallow processes: regional and local vol-
canism, magmatic processes, hydrothermal activity, geologic deformation
and structure of circum-Pacific mountain belts and island arcs (Minato et al.,
1965; Hollister, 1973; Solomon, 1974). The models of metallogenesis at con-
vergent plate boundaries are becoming more complex as factors are added to
successively approximate the actual deposits. The models are still inter-
pretive in that they explain the distribution of known deposits. With further
development these models may predict the locations of new deposits.
0 KM
OKM
15,200 18,700
T
19,000
T
T T 1
SOUTH AMERICA
ANDES
PETROLEUM j Co ^Cu.Au -i^-Eb'Zn'Cu'A'!»
Fe
ili^f
. >-x
:litho-
SPHERE;
:;::%v:
tlASTHENpSPHERE
::%•>.
Fig. 10. A diagrammatic east-west section across the central Pacific region shows the rela-
tion of petroleum and metal deposits to divergent (East Pacific Rise) and convergent
(Pacific margins) plate boundaries, as discussed in the text. Metals are disseminated in the
rocks and concentrated as oxides and sulfides in oceanic crust represented by the black
layer at the top of the oceanic lithosphere.
73
414
SUMMARY
Despite the geologic diversity of the Pacific region, the distribution of
selected energy and mineral resources follows a pattern with respect to
lithospheric plate boundaries (Fig. 1), as follows:
(1) Hydro thermal processes acting at divergent plate boundaries (oceanic
ridges) concentrate metals in the upper layers of oceanic lithosphere as
metalliferous sediments, metallic oxide deposits, and possibly as massive
stratabound copper-iron sulfide deposits. Because all of the oceanic litho-
sphere is generated by sea-floor spreading about oceanic ridges (Figs. 1, 2),
the processes of metal concentration at divergent plate boundaries affect the
oceanic lithosphere underlying two-thirds of the earth including the entire
Pacific ocean basin (Fig. 10).
(2) Oceanic trenches along convergent plate boundaries of the eastern,
northern, and western Pacific, and marginal basins formed between island
arcs and eastern Asia are areas of offshore petroleum potential (Figs. 5, 10).
(3) Deposits of precious, base, iron and ferro-alloy metals occur along the
landward side of convergent plate boundaries on continents and island arcs
of the Pacific region (Figs. 6—9).
(4) Models suggest that the observed distribution of petroleum and metal
deposits of the Pacific region are genetically related to geologic processes
acting at the circum-Pacific convergent plate boundaries (Fig. 10). The
development of oceanic trenches and marginal basins create conditions that
favor the accumulation of sediment and organic matter, the conversion of
the organic matter to petroleum, and the trapping of the petroleum. Metals
undergo multiple stages of concentration from various sources in two prin-
cipal regions (Fig. 10):
(a) Divergent plate boundaries: certain precious, base, iron and ferro-alloy
metals are disseminated in oceanic lithosphere by magmatic processes and
concentrated by hydrothermal processes.
(b) Convergent plate boundaries: metals are concentrated from oceanic
lithosphere descending along Benioff zones and from the overlying astheno-
sphere and continental lithosphere by various physical and chemical pro-
cesses. The deposits at convergent plate boundaries are products of complex
histories that are only beginning to be understood.
In conclusion, the conceptual framework of plate tectonics may be ap-
plied to predict areas hundreds to thousands of kilometers in extent of the
Pacific region where certain types of energy and mineral resources are likely
to occur. Conventional geological, geochemical, and geophysical methods
must then be employed to locate the deposits that may be only tens to thou-
sands of meters in extent within the areas of potential occurrence identified
from plate-tectonic considerations. The resolution of plate tectonics in pre-
74
415
dieting the locations of deposits will improve as models of geologic processes
at plate boundaries are refined, but conventional exploration methods will
continue to complement plate tectonics.
ACKNOWLEDGMENTS
We thank V. Baum, Chief of the Resources and Transport Division, and
J.P. Levy, Chief of the Office for Ocean Economics and Technology of the
United Nations for their important encouragement. The United Nations and
the National Oceanic and Atmospheric Administration supported this work.
NOTES
The amount of many precious, base, iron and ferro-alloy metals dis-
seminated in oceanic crust (basalt) is considerably greater than in con-
tinental crust (granite). For example, the copper content of basalt (about
100 ppm) is approximately five times that of granite (Turekian and Wede-
pohl, 1961; Vinogradov, 1962). An orebody containing 1,000,000 tons
of copper is equivalent to only about 4% of the copper disseminated in a
100-km3 volume of oceanic basalt (1 ppm = 104 tons per mile3 or 25 • 102
tons per km3). Sulfur is also present both in oceanic crust (100—400 ppm;
Vinogradov, 1962) and in seawater (1 g/1) in sufficient quantities to form the
various sulfide ores. A continuous supply of metals and sulfur is provided by
the conveyor belt of seawater-saturated oceanic crust that is generated at
divergent plate boundaries, moves across the ocean basin, and is consumed at
the convergent plate boundaries. An estimated volume of 100,000—250,000
km3 of oceanic crust has been overridden for every kilometer of leading edge
along the western Americas (Gilluly, 1973). Four percent of the copper dis-
seminated in this volume of oceanic crust is equivalent to between
1,000,000,000 and 2,500,000,000 tons, only a fraction of which is known
to be concentrated in massive stratabound and porphyry copper deposits
along the western margins of North and South America.
REFERENCES
Anon., 1972. The Times Atlas of the World. Bartholomew, Times Newspapers Ltd.,
Edinburgh, 4th rev. ed., 272 pp.
Avias, J., 1967. Overthrust structure of the main ultrabasic New Caledonian massives.
Tectonophysics, 4: 531—541.
Betzer, P.R., Bolger, G.W., McGregor, B.A. and Rona, P.A., 1974, The Mid-Atlantic Ridge
and its effect on the composition of particulate matter in the deep ocean. Am. Geo-
phys. Union Trans., 55: 293.
75
416
Bischoff, J.L. and Manheim, F.T., 1969. Economic potential of the Red Sea heavy metal
deposits. In: E.T. Degens and D.A. Ross (Editors), Hot Brines and Recent Heavy Metal
Deposits in the Red Sea. Springer, New York, N.Y., 600 pp.
Bostrom, K. and Peterson, M.N.A., 1969. The origin of aluminum-poor ferromanganoan
sediments in areas of high heat flow on the East Pacific Rise. Mar. Geol., 7: 427—447.
Bracey, D.R. and Ogden, T.A., 1972. Southern Mariana Arc: geophysical observations
and hypothesis of evolution. Geol. Soc. Am. Bull., 83: 1509—1522.
Coleman, R.G., 1971. Plate tectonic emplacement of upper mantle peridotites along
continental edges. J. Geophys. Res., 76: 1212—1222.
Cook, H.E., 1972. Stratigraphy and sedimentation. In: J.D. Hayes et al. (Editors), Initial
Reports of the Deep Sea Drilling Project, IX. U.S. Govt. Printing Office, Washington,
D.C., 1205 pp., pp. 933—943.
Corliss, J.B., Dasch, E.J., Dymond, J.D., Field, C.W., Heath, G.R., Senechal, R.G. and
Veeh, H.H., 1974. Metallogenesis in southeast Pacific Ocean: Nazca Plate Project.
Am. Assoc. Pet. Geol. Bull., 58: 1434 (abstract).
Dasch, E.J., 1974. Metallogenesis in the southeastern Pacific: a progress report on the
IDOE Nazca Plate Project. Phys. Earth Planet. Inter., 9: 249—258.
Davies, H.L. and Smith, I.E., 1971. Geology of Eastern Papua. Geol. Soc. Am. Bull., 82:
3299-3312.
DEMR, 1972. Principal mineral areas of Canada, scale 1 : 7,600,200. Dept. Energy, Mines
and Resour., Min. Resour. Branch and Geol. Surv. of Canada, Map 900A, 22nd ed.
Dymond, J., Corliss, J.B., Heath, G.R., Field, C.W., Dasch, E.J. and Veeh, H.H., 1973.
Origin of metalliferous sediments from the Pacific Ocean. Geol. Soc. Am. Bull., 84:
3355-3372.
ECAFE, 1962. Oil and natural gas map of Asia and the Far East, scale 1 : 5,000,000.
UN Econ. Comm. for Asia and the Far East (ECAFE), No. 16351, UN Publ. Sales
No. 62-1-16, 4 sheets.
ECAFE, 1963. Mineral distribution map of Asia and the Far East, scale 1 : 5,000,000.
UN Econ. Comm. for Asia and the Far East (ECAFE). No. 63-15714, UN Publ. Sales
No. 63-1-18, 4 sheets.
ECAFE, 1970. Energy atlas of Asia and the Far East. UN Econ. Comm. for Asia and the
Far East (ECAFE). UN Publ. Sales No. E.70.1.11, E/CN. 11/900, 25 pp.
Eimon, P. I., 1974. World copper resources. Mineral Res. Company, Salt Lake City, Utah,
map in 3 sheets.
Fischer, A.G., Heezen, B.C., Boyce, R.E., Bukry, D., Douglas, R.G., Garrison, R.E., Kling,
S.A., Krasheninnikov, V., Lisitsin, A.P. and Pimm, A.C., 1971. Initial Reports of the
Deep Sea Drilling Project, VI. U.S. Govt. Printing Office, Washington, D.C., 1329 pp.
Gilluly, J., 1973. Steady plate motion and episodic orogeny and magmatism. Geol. Soc.
Am. Bull., 84:499-514.
Grainger, D.J. and Grainger, R.L., 1974. The minerals deposits map of Papua New Guinea,
1: 2,500,00. Aust. Bur. Miner. Resour., Geol. Geophys., Bull., 148, PNG 8, 171 pp.
Hart, R.A., 1973. A model for chemical exchange in the basalt-seawater system of ocean-
ic layer II. Canadian J. of Earth Sci., 10: 799—816.
Hollister, V.F., 1973. Regional characteristics of porphyry copper deposits of South
America. Soc. of Mining Engineers Preprint No. 73-1-2.
Horn, D.R., Horn, B.M. and Delach, M.N., 1972. Metal content of ferromanganese depos-
its of the oceans. Lamont-Doherty IDOE/NSF GX33616 Tech. Rep. 3. Office for Int.
Decade of Ocean Exploration, Natl. Sci. Found., Washington, D.C., 51 pp.
Irving, E., North, F.K. and Couillard, R., 1974. Oil, climate, and tectonics. Can. J. Earth
Sci., 11: 1-17.
Isacks, B., Oliver, J. and Sykes, L.R., 1968. Seismology and the new global tectonics. J.
Geophys. Res., 73: 5855—5899.
James, D.E., 1971. Plate tectonic model for the evolution of the Central Andes. Geol.
Soc. Am. Bull., 82: 3325—3346.
76
417
Jones, D.B. (Advisory Editor), 1972. Oxford Economic Atlas of the World. Oxford Uni-
versity Press, 4th ed., 239 pp.
Karig, D.E., 1971. Origin and development of marginal basins in the western Pacific. J.
Geophys. Res., 76: 2542-2561.
Kennedy, W.Q. and Richey, J.E., 1947. Catalog of the active volcanoes of the world. Bull.
Volcanol. Suppl. Ser. II, VII: 1-11.
Krauskopf, K.B., 1967. Source rocks for metal-bearing fluids. In: H.L. Barnes (Editor),
Geochemistry of Hydrothermal Ore Deposits. Holt, Rinehart and Winston, New York,
N.Y.,pp. 1-33.
Langseth, M.G., Jr., 1969. The flow of heat from the Earth and the global distribution at
the surface. Am. Inst. Aeronaut. Astronaut., Pap, 69-589: 1—10.
La Plante, R.E., 1974. Hydrocarbon generation in Gulf Coast Tertiary sediments. Am.
Assoc. Pet. Geol. Bull., 58: 1281—1289.
Le Pichon, X., 1968. Sea floor spreading and continental drift. J. Geophys. Res., 73:
3661-3697.
Le Pichon, X., Francheteau, J. and Bonnin, J., 1973. Plate Tectonics. Elsevier, Amster-
dam, 300 pp.
Luyendyk, B.P., 1970. Dips of downgoing lithospheric plates beneath island arcs. Geol.
Soc. Am. Bull., 81 : 3411-3416.
McGregor, B.A. and Rona, P.A., 1975. Crest of the Mid-Atlantic Ridge at 26°N. J. Geo-
phys. Res., 80: 3307-3314.
McKelvey, V.E. and Wang, F.F.H., 1969. World subsea mineral resources. U.S. Geol.
Surv., Misc. Geol. Invest., Map 1-632.
Minato, M., Gorai, M. and Hunabashi, M., 1965. The Geological Development of Japanese
Islands. Tokyo, Tsukiji Shoka.
Mitchell, A.H.G., 1973. Metallogenic belts and angle of dip of Benioff zones. Nature
Phys. Sci., 245: 49-52.
Mitchell, A.H. and Bell, J.D., 1973. Island-arc evolution and related mineral deposits. J.
Geol., 81: 381-405.
Mitchell, A.H.G. and Garson, M.S., 1972. Relationship of porphyry copper and circum-
Pacific tin deposits to palaeo-Benioff zones. Inst. Min. Metall., Trans., Sec. B, 81:
B10-B25.
Oxburgh, E.R. and Turcotte, D.L., 1971. Origin of paired metamorphic belts and crustal
dilation in island arc regions. J. Geophys. Res., 76: 1315—1327.
Oyarzun, J.M. and Frutos, J.J., 1974. Porphyry copper and tin-bearing porphyries: a dis-
cussion of genetic models. Phys. Earth Planet. Inter., 9: 259—263.
Piper, D.Z., 1973. Origin of metalliferous sediments from the East Pacific Rise. Earth
Planet. Sci. Lett., 19: 75-82.
Pitman, W.C. Ill, Larson, R.L. and Herron, E.M., 1974. Age of the ocean basins deter-
mined from magnetic anomaly lineations. Geol. Soc. Am., map.
Ridge, J.D., 1972. Annotated bibliographies of mineral deposits in the western hemi-
sphere. Geol. Soc. Am. Mem., 131: 681 pp.
Rona, P. A., 1973. New evidence for seabed resources from global tectonics. Ocean Man-
agement, 1 : 145—159.
Rona, P. A. and Neuman, L.D., 1974. Plate tectonics and mineral resources of the Pacific
(abstract). Am. Assoc. Pet. Geol. Bull., 58: 1456.
Rona, P. A. and Neuman, L.D., 1975. Plate tectonics and mineral resources of the circum-
Pacific. Circum-Pacific Energy and Mineral Resources — Am. Assoc. Pet. Geol., Mem.
25, in press.
Rona, P. A., McGregor, B.A., Betzer, P.R. and Krause, D.C., 1975. Anomalous water tem-
peratures over Mid-Atlantic Ridge crest at 26°N. Deep-Sea Res., 22: 611—618.
Rona, P.A., Harbison, R.N., Bassinger, B.G., Scott, R.B. and Nalwalk, A.J., 1976. Tec-
tonic fabric and hydrothermal activity of Mid-Atlantic Ridge crest (26°N). Geol. Soc.
Am. Bull., in press.
77
418
Sawkins, F.J., 1972. Sulfide ore deposits in relation to plate tectonics. J. Geol., 80:
377-397.
Sayles, F.L. and Bischoff, J.L., 1973. Ferromanganoan sediments in the equatorial East
Pacific. Earth Planet. Sci. Lett., 19: 330—336.
Sclater, J.G., 1972. Heat flow and elevation of the marginal basins of the western Pacific.
J. Geophys. Res., 77 : 5705-5709.
Scott, M.R., Scott, R.B., Rona, P. A., Butler, L.W. and Nalwalk, A. J., 1974. Rapidly ac-
cumulating manganese deposit from the median valley of the Mid-Atlantic Ridge.
Geophys. Res. Lett., 1: 355—358.
Scott, R.B., Rona, P.A., McGregor, B.A. and Scott, M.R., 1974. The TAG hydrothermal
field. Nature, 251 : 301-302.
Sillitoe, R.H., 1972a. A plate tectonic model for the origin of porphyry copper deposits.
Econ. Geol., 67: 184—197.
Sillitoe, R.H., 1972b. Relation of metal provinces in western America to subduction of
oceanic lithosphere. Geol. Soc. Am. Bull., 83: 813—818.
Sillitoe, R.H., 1974. Tectonic segmentation of the Andes: implications for magmatism
and metallogeny. Nature, 250: 542—545.
Skornyakova, N.S. and Andrushchenko, P.F., 1974. Iron-manganese concretions in the
Pacific Ocean. Int. Geol. Rev., 16: 863-919.
Sleep, N.H. and Toksoz, M.N., 1971. Evolution of marginal basins. Nature, 233: 548—
550.
Snead, R.E., 1972. Atlas of World Physical Features. Wiley, New York, N.Y., 158 pp.
Solomon, M., 1974. Massive sulfides and plate tectonics. Nature, 249: 821—822.
Sorokhtin, O.G., Ushakov, S.A. and Fedynskiy, V.V., 1974. Dynamics of crustal plates
and origin of petroleum deposits. Akad. Nauk SSSR, Doklady, 214: 1407—1410.
Spooner, E.T.C. and Fyfe, W.S., 1973. Sub-sea floor metamorphism, heat and mass trans-
fer. Contrib. Miner. Pet., 42: 287—304.
Stern, C.R. and Wylie, P. J., 1973. Water-saturated and undersaturated melting relations of
a granite to 35 kilobars. Earth Planet. Sci. Lett., 18: 163—167.
Strakhov, N.M., Shterenberg, L.E., Kalinenko, V.V. and Tikhomirova, E.S., 1968. Geo-
chemistry of a sedimentary manganese ore-forming process. Akad. Nauk. SSSR Geol.
Inst., Trans., 185: 495 pp.
Tarling, D.H., 1973. Continental drift and reserves of oil and natural gas. Nature, 243:
277-279.
Turekian, K.K. and Wedepohl, K.H., 1961. Distribution of the elements in some major
units of the Earth's crust. Geol. Soc. Am. Bull., 72: 175—192.
Van Royan, W. and Bowles, O., 1952. The mineral resources of the world. V. II, Atlas of
the World's Resources. Prentice-Hall, New York, N.Y., 181 pp.
Vinogradov, A.P., 1962. Average contents of chemical elements in the principal types of
igneous rocks of the earth's crust. Geokhimiya, 7: 555—571.
Von der Borch, C.C., Nesteroff, W.D. and Galehouse, J.S., 1971. Iron-rich sediments
cored during Leg 8 of the Deep Sea Drilling Project. In: J.I. Tracey, Jr. et al. (Editors),
Initial Reports of the Deep Sea Drilling Project, VIII. U.S. Govt. Printing Office,
Washington, D.C., 1037 pp., pp. 725—819.
Von der Borch, C.C. and Rex, R.W., 1970. Amorphous iron oxide precipitates in sedi-
ments cored during Leg 5, Deep Sea Drilling Project. In: D.A. McManus et al. (Edi-
tors), Initial Reports of the Deep Sea Drilling Project, V. U.S. Govt. Printing Office,
Washington, D.C., pp. 541—544.
Waring, G.A., 1965. Thermal springs of the United States and other countries of the
world: a summary. U.S. Geol. Surv., Prof. Pap., 492.
78
419
44
Reprinted from: Papers from Circum-Pacific Energy and Mineral Resources
Conference, Honolulu, Hawaii, August 26-30, 1974, publ. by Amer. Assoc,
of Petroleum Geologists, Memoir 25, 48-57.
Plate Tectonics and Mineral Resources of Circum-Pacific Region1
PETER A. RONA2 and LAWRENCE D. NEUMANP
Abstract Distribution of selected energy (petroleum and
geothermal) and mineral (precious, base, iron, and ferro-
alloy metals) resources of the Pacific Ocean basin and Cir-
cum-Pacific continents appears to be related to lithospheric
plate boundaries. Divergent plate boundaries (oceanic
ridges) are related in time to the development of strati-
graphic traps in sedimentary basins of the continents, and in
space to metalliferous deep-sea sediments and the possible
occurrence of massive stratabound metallic sulfide deposits
in oceanic crust. Convergent plate boundaries are related in
space to areas of offshore petroleum potential and to on-
shore deposits of precious, base, iron, and ferro-alloy met-
als. Models suggest genetic relations between the observed
distribution of deposits and geologic processes at plate
boundaries, and may lead to the discovery of new re-
sources.
Introduction
The Pacific Ocean basin and surrounding con-
tinents provide a natural laboratory to develop
and test ideas on the relation between plate tec-
tonics and mineral resources. Our approach is as
follows:
1. Outline the geologic framework of the Pacif-
ic with particular attention to boundaries of the
lithospheric plates.
2. Determine the spatial distribution of various
mineral resources with respect to the plate bound-
aries (Figs. 1—4).
3. Consider models of mineral-concentrating
processes related to plate boundaries that may ex-
plain the observed distribution of mineral re-
sources (Fig. 5).
The boundaries of lithospheric plates are delin-
eated by narrow earthquake zones where the
plates are moving with respect to each other. The
theory of plate tectonics recognizes three types of
plate boundaries (Isacks et al, 1968). One type, a
convergent plate boundary, is where two adjacent
plates move together and collide or where one
plate descends under the other plate along a Be-
nioff seismic zone and is subducted into the up-
per mantle. The second type, a divergent plate
boundary, is where two adjacent plates move
apart because new lithosphere is added to each
plate by the process of seafloor spreading. The
third type is the transform plate boundary, where
two adjacent plates slide past one another.
A series of resource maps (Figs. 1-4), using the
same base map (Van der Grinten projection) as
that used by McKelvey and Wang (1969), was
compiled from numerous sources (Van Royan
and Bowles, 1952; Roberts and Irving, 1957;
Anon., 1962, 1963, 1970, 1972a, b; Lafitte and
Rouveyrol, 1965; McKelvey and Wang, 1969;
Bonine et al, 1970; Dengo and Levy, 1970; Jones,
1972).
Pacific Geologic Framework
Lithospheric Plate Boundaries
Divergent plate boundaries expressed as ocean-
ic ridges divide the Pacific into several lithospher-
ic plates (Fig. 1). The oceanic ridge system of the
Pacific is not midoceanic but is located in the
eastern Pacific off Central America and South
America and northwestern North America. The
average half-rates of seafloor spreading about the
oceanic ridges of the Pacific, determined from the
geomagnetic polarity-reversal time scale, range
from 1.2 cm/yr at the Gorda Rise off the north-
western United States to about 10 cm/yr off
Chile and Peru (Table 1).
The Pacific is a closing ocean basin surrounded
on three sides by convergent plate boundaries ex-
pressed as oceanic trenches (Fig. 1). In the global
crustal budget, the Circum-Pacific convergent
plate boundaries account for the major portion of
crustal consumption. Linear rates of crustal con-
sumption around the Pacific, assumed to be equal
to rates of plate convergence, range between
about 1 and 15 cm/yr (Table 2).
The dip of the Benioff zone at convergent plate
boundaries (Table 3) is inversely proportional to
the relative rate of convergence of the adjacent
plates (Luyendyk, 1970). For example, the dip of
the Benioff zone in the upper 100 km is about 12°
beneath Peru, where the relative rate of plate con-
vergence is 8.8 cm/yr (Table 2), and the half-rate
of seafloor spreading about the corresponding
section of the East Pacific Rise is 9.5 cm/yr (Ta-
ble 1). The dip of the Benioff zore increases to
about 55° beneath the Kermadec Islands (Table
3), where the relative rate of convergence decreas-
es to 5.8 cm/yr, and the half-rate of seafloor
1 Manuscript received. December 26, 1974.
^National Oceanic and Atmospheric Administration. Atlantic
Oceanographic and Meteorological Laboratories, 15
Rickenbacker Causeway, Miami, Florida 33149.
3Office for Ocean Economics and Technology, United
Nations. New York. New York 10017.
48
420
Plate Tectonics and Mineral Resources
49
180' 150* 120* 90* 60* 30*W
• PETROLEUM PRODUCING
AREAS
ONSHORE PETROLEUM
POTENTIAL
m OFFSHORE PETROLEUM
POTENTIAL
CRYSTALLINE
ROCKS
SEDIMENTARY
ROCKS
DIVERGENT PLATE
BOUNDARY
CONVERGENT PLATE
BOUNDARY
TRANSFORM PLATE
BOUNDARY
UNCERTAIN PLATE
BOUNDARY
FIG. 1 -Areas of petroleum production and potential of Pacific region. Lithospheric plates and plate boundaries
are shown.
spreading about the East Pacific Rise decreases to
about 5.5 cm/yr.
Oceanic and continental crust are juxtaposed
at convergent plate boundaries in the eastern Pa-
cific (Fig. 1). Marginal basins generally underlain
by oceanic crust intervene between convergent
Table 1 . Seafloor-Spreading Half-Rates about
Pacific Oceanic Ridges
Latitude
Longitude
Half-spreading Location
rate
(+North,
(+East,
(cm/yr)
-South)
- West)
46.5
-129
1.5
Juan de Fuca Ridge
41.3
-127.5
1-2
Gorda Rise
12.5
-103.5
4.5
East Pacific Rise
3.2
-102
6.4
East Pacific Rise
3.0
- 83.5
3.2
Galapagos rift zone
2.0
- 97.5
3.2
Galapagos riff zoneJ
0.75
- 87.6
3.2
Galapagos rift zone
- 5.6
-106.8
6.5
East Pacific Rise
-9.2
-110.3
7-5
East Pacific Rise
-19
-113
9-9.5
East Pacific RiseJ
-28
-112
8-10
East Pacific Rise
-35.6
-110.9
5.0
East Pacific Rise
-43.2
- 82.5
3.0
Chile Ridge3
'Vine and Wilson, 1965.
JAtwater and Mudie, 1973.
*Herron, 1972.
plate boundaries and continental crust in the
western Pacific. The development of marginal ba-
sins may be related to the dip of Benioff zones
(Oxburgh and Turcotte, 1971; Karig, 1971; Sleep
and Toksoz, 1971; Bracey and Ogden, 1972).
Where the marginal basins are present in the
western Pacific, the dip of the Benioff zone ex-
ceeds about 35°; where marginal basins are ab-
sent in the eastern Pacific, the dip is less than
about 35° (Table 3).
Ages of Seafloor and Continents
The age of the Pacific seafloor, as determined
by dating of remanent magnetic anomalies in the
geomagnetic polarity-reversal time scale (Pitman
et al, 1974) and dating of rock samples recovered
by the Deep Sea Drilling Project (Fischer et al,
1971), ranges between Late Jurassic (about 150
m.y. ago) and recent. The distribution of ages
about divergent plate boundaries is regular, and
the seafloor becomes progressively older with dis-
tance from these boundaries. The distribution of
ages at convergent plate boundaries is irregular,
and the seafloor is of various ages at these bound-
aries. The distribution of ages on continents is
delineated by structural provinces that reflect the
youngest deformational event, as distinguished
from the ages of the seafloor, which reflect the
origin of the constituent rocks. Structural prov-
inces of the eastern Pacific are predominantly Ce-
nozoic along western South America and Meso-
zoic along western North America. In the western
Pacific, structural provinces of eastern Asia and
Australia exhibit a complex distribution and
421
50 Peter A. Rona and Lawrence D. Neuman
Table 2. Rate of Plate Convergence at Convergent Plate Boundaries of the Pacific1
Latitude
Longitude
Rate
Azimuth
Location
(+ North,
(+ East,
(cm/yr)
- South)
- West)
51
160
7.2
114
Kurile Trench
43
148
7.5
107
Kurile Trench
35
142
8.6
101
Japan Trench
27
143
7.5
265
Japan Trench
19
148
4.9
282
Mariana Trench
11
142
2.3
243
Mariana Trench
- 3
142
14.5
78
New Guinea
-13
-172
9.9
97
N. Tonga Trench
-34
- 178
5.8
95
S. Kermadec Trench
-45
169
3.5
72
S. New Zealand
-55
159
2.6
29
Macquarie Island
-50
- 75
3.1
240
Cape Horn
-35
- 74
8.7
74
S. Chile Trench
- 4
- 82
8.8
77
N. Peru Trench
7
- 79
8.3
75
Panama Gulf
20
- 106
6.4
39
N. Middle America Trench
57
- 150
5.6
144
E. Aleutian Trench
50
- 178
6.9
126
W. Aleutian Trench
54
162
7.0
115
W. Aleutian Trench
'From Le Pichon et al, 1973, Table V.
range in age from Precambrian through Ceno-
zoic.
Distribution of Geothermal Phenomena
The distribution of geothermal phenomena, in-
cluding heat flow, active volcanoes, thermal
springs, fumaroles, and geysers, is spatially relat-
ed to lithospheric plate boundaries. Values of
heat flow in the Pacific Ocean basin (Langseth,
1969; Sclater, 1972) are relatively high (>2 HFU)
at divergent plate boundaries and decrease basin-
ward away from these boundaries. Values of heat
flow at convergent plate boundaries and in mar-
ginal basins (Karig, 1971) around the Pacific are
variable. The distribution of heat-flow values on
the Circum-Pacific continents is complex. Rela-
tively high values (>2 HFU) are present in limit-
ed areas of eastern Australia, eastern Asia, and
western North America. Active volcanoes, ther-
mal springs, and fumaroles are aligned along the
landward side of convergent plate boundaries on
continents and island arcs around the Pacific
(Kennedy and Richey, 1947; Waring, 1965;
Snead, 1972). These features are not evenly
spaced but are grouped (Kelleher, 1972).
It is apparent from the distribution of heat-flow
values and thermal springs on islands and conti-
nents around the Pacific that numerous potential
sites exist for the development of geothermal en-
ergy. Geothermal energy is being utilized at sites
in New Zealand, Japan, and the western United
States.
Distribution of Petroleum Resources
Areas of offshore petroleum potential conform
with convergent plate boundaries around the Pa-
cific (Fig 1 ; McKelvey and Wang, 1969). Areas of
petroleum potential in the eastern Pacific com-
prise thick sedimentary accumulations underlying
the continental margins of western North Ameri-
ca and South America and partially filling ocean-
ic trenches seaward of the margins of Central
America and South America along convergent
plate boundaries. In the western Pacific, island
arcs directly landward of convergent plate
boundaries divide the ocean basin into marginal
basins partially enclosed between the islands and
eastern Asia — for example, the South China Sea,
the East China Sea, the Yellow Sea, the Sea of
Japan, the Sea of Okhotsk, and the Bering Sea.
422
Plate Tectonics and Mineral Resources
51
Table 3. Dip of Benioff Zone (upper 100 km)1
Dip (degrees)
Location (counter-clockwise
around Pacific)
45
New Zealand
55
Kermadec
50
Tonga (south)
30
Tonga (north)
65
New Hebrides (south)
55
New Hebrides (north)
65
Solomons
45
New Britain
40
Sunda: Flores Sea
60
Sunda: Java
50
Sumatra
50
Burma
55
Mindanao
45
Marianas
25
Izu-Bonin
40
Ryukus
35
Kurile
30
Honshu
40
Aleutians
35
Middle America
12
Peru
15
Chile (north)
'isacks and Molnar, 1971; Oliver et al, 1973.
Areas of petroleum potential in the western Pacif-
ic comprise sedimentary accumulations in the
marginal basins as well as in oceanic trenches
along convergent plate boundaries.
The oceanic trenches along the eastern and
western sides of the Pacific Ocean basin receive
deep-sea sediments that presumably are trans-
ported into the trenches on a "conveyor belt" of
spreading seafloor during subduction of the
oceanic lithosphere. The content of organic mat-
ter in the deep-sea sediments that are transported
into the trenches varies in space and in time; for
example, a zone of high organic productivity ex-
tending across the equatorial Pacific affects the
composition of the sediments deposited in that
region. Both the amount of organic matter deliv-
ered to the oceanic trenches and the petroleum
potential are expected to vary accordingly. Tem-
perature and pressure conditions beneath the
trenches and the marginal basins may facilitate
the conversion of organic matter in the sediments
to petroleum (Tarling, 1973; Sorokhtin et al,
1974).
Areas of onshore petroleum production are
sedimentary basins on continents with no appar-
ent spatial relation either to divergent or conver-
gent plate boundaries of the Pacific (Fig. 1).
However, the development of the sedimentary se-
quences in the basins may be related to divergent
plate boundaries in time, if not in space, through
the influence on global sea level of reversible vol-
ume changes of oceanic ridges (Rona, 1973b;
Rona and Wise, 1974). The volume of the world
oceanic ridge system is not constant but appears
to vary directly with rates of seafloor spreading.
A volume increase in the world oceanic ridge sys-
tem reduces the cubic capacity of ocean basins,
resulting in transgression of the sea onto all the
continents and deposition of a sedimentary se-
quence that may contain organic source material
and reservoir rocks. Conversely, a volume de-
crease in the oceanic ridge system increases the
cubic capacity of ocean basins, resulting in re-
gression of the sea from all the continents and the
development of widespread unconformities that
may be associated with traps for the accumula-
tion of petroleum. Reversible volume changes of
the worldwide oceanic ridge" system have operat-
ed on a time scale of tens of millions of years, as
evidenced by the presence of six sedimentary se-
quences separated by surfaces of unconformity in
the Phanerozoic stratigraphy of North America
(Sloss, 1963). The inferred relations of the volume
of oceanic ridges to sedimentary sequences and
unconformities may prove useful in exploration
for stratigraphic traps.
Distribution of Selected Mineral
Resources
Light Metal Deposits
Deposits of light metals, including aluminum,
beryllium, lithium, and titanium, exhibit no ap-
parent relation to plate boundaries of the Pacific.
These metals are associated with granitic rocks of
the continents as opposed to basaltic rocks of the
ocean basins. The distribution of aluminum, be-
ryllium, lithium, and titanium on continents is re-
lated to the occurrence of particular minerals in
granitic rocks and to conditions of weathering.
Metal Deposits at or Near Convergent Plate
Boundaries
Precious metal deposits — Deposits of precious
metals, including gold, silver, and platinum, ex-
hibit a spatial relation to convergent plate bound-
aries (Fig. 2). In the eastern Pacific, precious met-
al deposits are found along the western margins
of North America and South America. Addition-
al deposits are present in eastern North America
423
52
Peter A. Rona and Lawrence D. Neuman
I2Q*E 150* 180* IW 120* 90* 60* 30*<
* SILVER
• GOLD
O PLATINUM
120* E 190* ltO* 190* 120*
FIG. 2-Map of precious metal deposits of Pacific region. Lithospheric plate boundaries are shown.
in the area of the Canadian shield and in eastern
South America in the area of the Guianian and
Brazilian shields. In the western Pacific, precious
metal deposits occur on island arcs (including Ja-
pan, the Philippines, and Indonesia) situated
along convergent plate boundaries. Deposits also
occur in eastern Asia and Australia, where they
are separated by a gap from active convergent
plate boundaries. Precious metal deposits have
not been found along those sections of conver-
gent plate boundaries where oceanic lithosphere
is juxtaposed and island arcs are absent (e.g., the
eastern side of the Philippine Sea and the section
between New Zealand and Samoa).
Base metal deposits — Deposits of base metals
exhibit a spatial relation to convergent plate
boundaries (Fig. 3). Their distribution is grossly
similar to that of precious metal deposits (Fig. 2).
In the eastern Pacific, base metal deposits are
present along the western margins of North
America and South America. In the Peruvian
Andes, porphyry copper and vein-type minerali-
zation are associated with Neogene silicic volcan-
ic rocks (Mitchell, 1973). Additional base metals
occur in central and eastern North America in-
cluding the area of the Canadian shield.
In the western Pacific, base metal deposits are
found on island arcs along convergent plate
boundaries. Deposits of tin, tungsten, and fluorite
with minor bismuth and molybdenum occur in
belts of predominantly Mesozoic granites and
acidic eruptive rocks along the southern margin
of Alaska and the eastern margin of Asia (Mit-
chell, 1973). Base metal deposits also occur in
eastern Asia and Australia, which are separated
by a gap from active convergent plate boundaries.
As in the case of precious metals, base metal de-
posits have not been found along those sections
of convergent plate boundaries where oceanic
lithosphere is juxtaposed and island arcs are ab-
sent.
Sediment samples recovered by coring near the
crest of oceanic ridges and by deep-sea drilling
away from the crest reveal widespread enrich-
ment by base metals of those strata directly over-
lying basalt of the Pacific Ocean basin (Bostrom
and Petersen, 1969; von der Borch et al, 1971;
Cook, 1972; Cronan et al, 1972; Dymond et al,
1973; Sayles and Bischoff, 1973; Piper, 1973).
The base metals include antimony, copper, lead,
mercury, and zinc, but no tin. The observation
that the base metal enrichment is limited to the
basal sedimentary layer overlying basalt implies
that the enrichment occurred soon after the gen-
eration of the underlying basalt by seaflooi
spreading about an oceanic ridge at a divergent
plate boundary.
Iron and ferro-alloy metal deposits — Deposits of
iron and ferro-alloy metals exhibit a spatial rela-
tion to convergent plate boundaries around the
Pacific (Fig. 4). Their distribution is grossly simi-
lar to that of precious and base metals. In the
eastern Pacific, iron and ferro-alloy metal depos-
its occur along the western margins of North
America and South America. Additional deposits
occur in central and eastern North America, in-
cluding the Canadian shield, and in the Guianian
and Brazilian shields of South America. In the
western Pacific, iron and ferro-alloy deposits oc-
cur on island arcs along convergent plate bound-
424
Plate Tectonics and Mineral Resources
53
150* 120* W
FIG. 3 -Map of base metal deposits of Pacific region. Lithospheric plate boundaries are shown. Sediments
i enriched in base metals (shaded) overlie basalts of ocean basin.
aries and in eastern Asia and Australia, where
they are separated by a gap from active conver-
gent plate boundaries. Iron and ferro-alloy metal
deposits have not been found along those sections
of convergent plate boundaries where oceanic
lithosphere is juxtaposed and island arcs are ab-
sent. Sedimentary strata directly overlying basalt
of the Pacific Ocean basin are enriched in iron
and ferro-alloy metals, in addition to base metals.
Nodules containing variable percentages of
manganese, copper, nickel, and cobalt are present
on about two thirds of the Pacific seafloor (Fig. 4;
Strakhov et al, 1968). A zone of nodules of anom-
alously high copper and nickel content (1.5-2.0%)
and areal density (20-50% of seafloor) extends
east-west across the Pacific between latitudes 5°
and 20° N. No apparent spatial relation exists be-
tween plate boundaries and either the overall dis-
tribution or the enriched zone of nodules in the
Pacific Ocean basin.
Metal Deposits at Divergent Plate Boundaries
Knowledge of the distribution of metal depos-
its with respect to divergent plate boundaries is
* IRON
Q CHROMIUM
© COBALT
O MANGANESE
* MOLYBDENUM
A NICKEL
■ TUNGSTEN
* VANADIUM
MANGANESE NODULES
' UP TO 20% OF BOTTOM
= 20-50% OF BOTTOM
BOUNDARY BETWEEN
RED CLAY & BIOGENIC OOZES
BE COPPER CONTENT IN WEIGHT
1.5-2.0%
© NICKEL CONTENT
1.5-2.0%
FIG. 4 -Map of iron and ferro-alloy metal deposits, including manganese nodules, of Pacific. Lithospheric plate
boundaries are shown.
425
54
Peter A. Rona and Lawrence D. Neuman
limited because, as submerged oceanic ridges,
these boundaries are less accessible to observa-
tion than are convergent plate boundaries. Sedi-
ments deposited about divergent plate boundaries
in the Pacific Ocean basin are enriched in certain
base, iron, and ferro-alloy metals. The recently
discovered TAG hydrothermal field (R. Scott et
al, 1974; Rona et al, in press), named after the
Trans-Atlantic Geotraverse (TAG) of the Nation-
al Oceanic and Atmospheric Administration, rep-
resents a type of metal deposit present in oceanic
crust and may provide insights to processes of
metal concentration at divergent plate bounda-
ries, including those of the Pacific.
The TAG hydrothermal field occupies an area
at least 15 km by 15 km including the east wall of
the rift valley of the Mid-Atlantic Ridge at 26°N.
Manganese oxide was recovered consistently by
dredging from deposits concentrated along steps
in the wall that are interpreted as faults which
have acted as conduits for hydrothermal fluids.
Concentration of the manganese oxide by hy-
drothermal processes capable of extreme chemi-
cal fractionation is evidenced by an exceptionally
rapid rate of accumulation, about 200 mm per
million years, and extreme purity, about 40%
manganese with only trace quantities of iron and
copper (M. Scott et al, 1974). Present activity is
indicated by a positive temperature anomaly with
an inverted temperature gradient (Rona et al,
1975) and by metal enrichment of suspended par-
ticulate matter (Betzer et al, 1974) in the seawater
overlying the TAG hydrothermal field.
Model of Metal Deposits at Divergent
Plate Boundaries
A model based on various lines of evidence
(Spooner and Fyfe, 1973) considers that certain
precious, base, iron, and ferro-alloy metals may
be concentrated as deposits by sub-seafloor hy-
drothermal convection systems that may develop
at crests of oceanic ridges. According to the mod-
el, cold dense seawater enters fractures in the ba-
salt of an oceanic ridge. The seawater is heated as
it encounters hot material intruded at the diver-
gent plate boundary. The warm, less dense seawa-
ter rises through fractures in the basalt and leach-
es metals that are transported in solution as
complexes with chloride in the seawater. A frac-
tion of the metals then combines with sulfur in
the seawater and precipitates as a massive strata-
bound sulfide body under reducing conditions
beneath the basalt-seawater interface. Manganese
oxide precipitates under oxidizing conditions at
the basalt-seawater interface as the hydrothermal
solutions discharge from the seafloor. Colloidal
ferric hydroxide precipitates from the hydrother-
mal solutions in the overlying seawater. The ferric
hydroxide scavenges the remaining metals from
solution by colloidal absorption and settles to de-
posit a layer of metalliferous sediments on basalt
of the adjacent seafloor.
An example of an economic mineral deposit
interpreted as the product of a sub-seafloor hy-
drothermal convection system occurs in the Troo-
dos massif of Cyprus. The Troodos massif is in-
terpreted as an obducted slice of oceanic crust
generated at a divergent plate boundary (Gass
and Masson-Smith, 1963; Moores and Vine,
1971). An "umber" of metallic oxides overlies
massive stratabound copper sulfide ore bodies in
basaltic rocks of the Troodos massif. The manga-
nese oxide of the TAG hydrothermal field chemi-
cally resembles the umber and may be underlain
by massive copper sulfide bodies (Rona, 1973a;
R. Scott et al, 1974; Rona et al, in press). Relict
metallic oxide and sulfide deposits may be ex-
pected to extend in belts along flow lines of sea-
floor spreading transverse to the axis of an ocean-
ic ridge; the extent of the belts would depend on
the continuity of seafloor spreading and the per-
sistence of the sub-seafloor hydrothermal convec-
tion system which concentrates the deposits
(Rona, 1973a). The metal deposits may be buried
by off-axis volcanism.
Model of Metal Deposits at Convergent
Plate Boundaries
A model to interpret the observed association
of precious, base, iron, and ferro-alloy metal de-
posits with the convergent plate boundaries of the
Pacific is based on the following concepts:
1. The dip of Benioff zones is inversely propor-
tional to the rate of plate convergence (Tables 1-
3; Luyendyk, 1970).
2. Marginal basins develop where the dip of
Benioff zones exceeds about 35°.
3. Calc-alkalic andesitic volcanic rocks and to-
nalitic plutons occur above steeply dipping Be-
nioff zones (Mitchell, 1973).
4. Silicic volcanic rocks and granitic plutons
occur above shallow-dipping Benioff zones (Mit-
chell, 1973).
5. Metals in deposits along convergent plate
boundaries are primarily derived from oceanic
crust descending along the associated Benioff
zone (Sawkins, 1972; Sillitoe, 1972) and from
continental crust. The role of the mantle as a
source of material remains to be evaluated.
6. The nature and volume of volatile matter ex-
pelled from oceanic crust descending along Be-
nioff zones influence the extraction, transport,
426
Plate Tectonics and Mineral Resources
55
concentration, and deposition of metals (Rub,
1972).
The model presents two regimes to account for
the past and present distribution of metals along
convergent plate boundaries of the western and
eastern Pacific, as follows:
1. Relatively fast seafloor spreading and plate
convergence are associated with a shallow Be-
nioff zone and the generation of silicic volcanic
rocks and granitic plutons (Fig. 5a, b). High-level
porphyry copper deposits occur in the silicic vol-
canic rocks, and deep-level tin, tungsten, bismuth,
fluorite, and molybdenum occur in the granites.
The copper is primarily derived from metallifer-
ous sediments and massive stratabound metallic
sulfide deposits of the oceanic crust that descends
along the underlying Benioff zone. The tin and
associated metals, along with a portion of the
granite (Stern and Wyllie, 1973), are derived from
continental crust and their segregation is aided by
the rise of volatile matter expelled from the de-
scending oceanic crust. This regime is exemplified
by western South America at present and by east-
ern Asia in late Mesozoic time.
2. Relatively slow seafloor spreading and plate
convergence are associated with a steep Benioff
zone, the development of marginal basins, and
the generation of calc-alkalic andesitic volcanic
rocks and tonalitic plutons with associated por-
phyry copper deposits (Fig. 5c, d). The calc-alkal-
ic volcanic rocks and the copper are primarily de-
rived from the oceanic crust (Jakes and White,
1972) that descends along the underlying Benioff
zone; this crust includes metalliferous strata and
massive strata-bound metallic sulfide deposits.
The tonalitic plutons and some copper may be
derived from the base of continental or island-arc
crust (Jakes and White, 1971; Brown, 1973).
Granitic plutons emplaced during the first regime
(Fig. 5b) are unroofed by erosion to expose de-
posits of tin and associated metals (Fig. 5d). The
development of marginal basins (Fig. 5c, d) re-
sults in the separation of island arcs from the con-
tinent, producing a gap such as that observed be-
tween the metal deposits of eastern Asia and
Australia and the active convergent plate bound-
aries of the western Pacific (Figs. 2-4). This re-
gime is exemplified by eastern Asia at the end of
the Mesozoic and at present.
Summary
Consideration of the distribution of selected
energy and material resources with respect to li-
thospheric plate boundaries of the Pacific (Figs.
1-4) leads to the following conclusions:
1. Divergent plate boundaries (oceanic ridges)
are inferred to be related in time to the devel-
opment of stratigraphic traps for petroleum in
sedimentary basins on continents through the
control of eustatic sea level by reversible volume
changes of oceanic ridges.
2. Divergent plate boundaries are related in
space to metalliferous deep-sea sediments and to
the possible occurrence of massive strata-bound
metallic sulfide deposits in oceanic crust.
3. Convergent plate boundaries are related in
space to areas of offshore petroleum potential
FIG. 5 -Model of metallogenesis at convergent plate boundaries of Pacific (modified from Mitchell, 1973).
Diagrammatic cross sections through convergent plate boundaries show Benioff zone, oceanic crust (black)
incorporating metalliferous sediments and massive strata-bound metallic sulfide bodies (white semicircles in
oceanic crust), rising magma (solid vertical arrows), rising volatile matter (dashed vertical arrows), silicic volcanic
rocks and granitic plutons (+), andesitic volcanic rocks and tonalitic plutons (X ), and known deposits of porphyry
copper (PCu), tin (Sn), tungsten (W), fluorite (F), and antimony (Sb). For explanation, see text.
427
56
Peter A. Rona and Lawrence D. Neuman
and to onshore deposits of precious, base, iron
and ferro-alloy metals.
4. Certain precious, base, iron, and ferro-alloy
metal deposits that are present in Precambrian
shields and old orogenic belts distant from pre-
sent plate boundaries may be related to former
plate boundaries.
5. Models suggest genetic relations between the
observed distribution of metal deposits and geo-
logic processes at plate boundaries. The models
are interpretative in that they attempt to explain
the observed distribution of deposits. With fur-
ther development these models may predict the
locations of undiscovered metal deposits.
References Cited
Anonymous, 1962, Oil and natural gas map of Asia and
the Far East, scale 1:5,000,000: U.N. ECAFE, no.
16351, U.N. Pub. Sales No. 62-1-16, 4 sheets.
1963, Mineral distribution map of Asia and the
Far East, scale 1:5,000,000: U.N. ECAFE, no. 63-157
14, U.N. Pub. Sales No. 63-1-18, 4 sheets.
1970, Energy atlas of Asia and the Far East:
U.N. ECAFE, U.N. Pub. Sales No. E-70-1-11,
E/CN-1 1/900, 25 p.
1972a, Principal mineral areas of Canada, scale
1:7,600,200, 22nd ed.: Canada Dept. Energy, Mines
and Resources, Mineral Resources Br., and Canada
Geol. Survey, Map 900A.
1972b, The Times atlas of the world, 4th ed.
(revised): Edinburgh, John Bartholomew and Son
Ltd., Times Newspapers Ltd., 272 p.
Atwater, T., and J. D. Mudie, 1973, Detailed near-bot-
tom geophysical study of the Gorda Rise: Jour. Geo-
phys. Research, v. 78, p. 8665-8686.
Betzer, P. R., et al, 1974, The Mid- Atlantic Ridge and
its effect on the composition of particulate matter in
the deep ocean: Eos (Am. Geophys. Union Trans.), v.
55, no. 4, p. 293.
Bonine, M. E., et al, 1970, Atlas of Mexico: Univ. Texas
at Austin, Bur. Business Research, 138 p.
Bostrom, K., and M. N. A. Petersen, 1969, The origin of
aluminum-poor ferromanganoan sediments in areas
of high heat flow on the East Pacific Rise: Marine
Geology, v. 7, p. 427-447.
Bracey, D. R., and T. A. Ogden, 1972, Southern Maria-
na arc; geophysical observations and hypothesis of
evolution: Geol. Soc. America Bull., v. 83, p. 1509-
1522.
Brown, G. C, 1973, Evolution of granite magmas at
destructive plate margins: Nature; Phys. Sci., v. 241,
p. 26-28.
Cook, H. E., 1972, Stratigraphy and sedimentation, p.
933-943 in J. D. Hays et al, eds., Initial reporjts of the
Deep Sea Drilling Project, v. IX: Washington, U.S.
Govt. Printing Office, 1205 p.
Cronan, D. S., et al, 1972, Iron-rich basal sediments
from the eastern equatorial Pacific; Leg 16, Deep Sea
Drilling Project: Science, v. 175, no. 4017, p. 61-63.
Dengo, G, and E. Levy, 1970, Estudios et mapa metalo-
geneticos de America Central: Inst. Centroamericano
Inv. y Tecnologia Indus., Pub. Geol. No. 3, 57 p.,
scale 1:200,000.
Dymond, J., et al, 1973, Origin of metalliferous sedi-
ments from the Pacific Ocean: Geol. Soc. America
Bull., v. 84, p. 3355-3372.
Fischer, A. G., et al, 1971, Initial Reports of the Deep
Sea Drilling Project, v. VI: Washington, U.S. Govt.
Printing Office, 1329 p.
Gass, I. G., and D. Masson-Smith, 1963, The geology
and gravity anomalies of the Troodos Massif, Cy-
prus: Royal Soc. London Philos. Trans., Ser. A, v.
255, p. 417-467.
Herron, E. M., 1972, Sea-floor spreading and the Ceno-
zoic history of the east-central Pacific: Geol. Soc.
America Bull., v. 83, p. 1671-1692.
Isacks, B., J. Oliver, and L. R. Sykes, 1968, Seismology
and the new global tectonics: Jour. Geophys. Re-
search, v. 73, p. 5855-5899.
and P. Molnar, 1971, Distribution of stresses in
the descending lithosphere from a global survey of
focal mechanism solutions of mantle earthquakes:
Rev. Geophys. and Space Physics, no. 9, 103 p.
Jakes, P., and A. J. R. White, 1971, Composition of
island arcs and continental growth: Earth and Plane-
tary Sci. Letters, v. 12, p. 224-230.
and — 1972, Major and trace element
abundances in volcanic rocks of orogenic areas:
Geol. Soc. America Bull., v. 83, p. 29-39.
Jones, D. B., advisory ed., 1972, Oxford economic atlas
of the world, 4th ed.: London, Oxford Univ. Press,
239 p.
Karig, D. E., 1971, Origin and development of marginal
basins in the western Pacific: Jour. Geophys. Re-
search, v. 76, p. 2542-2561.
Kelleher, J. A., 1972, Rupture zones of large South
American earthquakes and some predictions: Jour.
Geophys. Research, v. 77, p. 2087-2103.
Kennedy, W. Q., and J. E. Richey, 1947, Catalogue of
the active volcanoes of the world: Bull. Volcanol.,
Supp. Ser. 2, v. 7, 11 p.
Lafitte, P., and P. Rouveyrol, 1965, Carte miniere du
globe sur fond tectonique au 1:20,000,000 (projection
equiareale): Paris, Bur. Recherches Geol. et Minieres,
74, ed. provisoire, pt. 2.
Langseth, M. G., Jr., 1969, The flow of heat from the
Earth and the global distribution at the surface: Am.
Inst. Aeronautics and Astronautics, Paper No. 69-
589, 10 p.
Le Pichon, X., J. Francheteau, and J. Bonnin, 1973,
Plate tectonics: New York, Elsevier, 300 p.
Luyendyk, B. P., 1970, Dips of downgoing lithospheric
plates beneath island arcs: Geol. Soc. America Bull.,
v. 81, p. 3411-3416.
McKelvey, V. E., and F. F. H. Wang, 1969, World sub-
sea mineral resources: U.S. Geol. Survey Misc. Geol.
Inv. Map 1-632.
Mitchell, A. H. G., 1973, Metallogenic belts and angle
of dip of Benioff zones: Nature; Phys. Sci., v. 245, p.
49-52.
Moores, E. M., and F. J. Vine, 1971, The Troodos Mas-
sif, Cyprus, and other ophiolites as oceanic crust;
evaluation and implications: Royal Soc. London Phi-
los. Trans., Ser. A., v. 268, p. 443-466.
428
Plate Tectonics and Mineral Resources
57
Oliver, J., et al, 1973, Dynamics of the downgoing litho-
sphere: Tectonophysics, v. 19, p. 133-147.
Oxburgh, E. R., and D. L. Turcotte, 1971, Origin of
paired metamorphic belts and crustal dilation in is-
land arc regions: Jour. Geophys. Research, v. 76, p.
1315-1327.
Piper, D. Z., 1973, Origin of metalliferous sediments
from the East Pacific Rise: Earth and Planetary Sci.
Letters, v. 19, p. 75-82.
Pitman, W. C. Ill, R. L. Larson, and E. M. Herron,
1974, Age of the ocean basins determined from mag-
netic anomaly lineations: Geol. Soc. America, map.
Roberts, R. J., and E. M. Irving, 1957, Mineral deposits
of Central America: U.S. Geol. Survey Bull. 1034,
205 p.
Rona, P. A., 1973a, Plate tectonics and mineral re-
sources: Sci. American, v. 229, no. 1, p. 86-95.
1973b, Relations between rates of sediment ac-
cumulation on continental shelves, sea-floor spread-
ing, and eustacy inferred from the central North At-
lantic: Geol. Soc. America Bull., v. 84, p. 2851-2872.
and D. U. Wise, 1974, Symposium: global sea
level and plate tectonics through time: Geology, v. 2,
no. 3, p. 133-134.
et al, 1975, Anomalous water temperatures over
Mid-Atlantic Ridge crest at 26° North latitude: Deep
Sea Research, v. 22, p. 611-618.
et al, in press, Tectonic fabric and hydrothermal
activity of Mid-Atlantic Ridge crest (lat- 26°N):
Geol. Soc. America Bull.
Rub, M. G., 1972, The role of the gaseous phase during
the formation of ore bearing magma tic complexes:
Chem. Geology, v. 10, p. 89-98.
Sawkins, F. J., 1972, Sulfide ore deposits in relation to
plate tectonics: Jour. Geology, v. 80, p. 377-397.
Sayles, F. L., and J. L. Bischoff, 1973, Ferromanganoan
sediments in the equatorial East Pacific: Earth and
Planetary Sci. Letters, v. 19, p. 330-336.
Sclater, J. G., 1972, Heat flow and elevation of the mar-
ginal basins of the western Pacific: Jour. Geophys.
Research, v. 77, p. 5705-5709.
Scott, M. R., et al, 1974, Rapidly accumulating manga-
nese deposit from the median valley of the Mid-At-
lantic Ridge: Geophys. Research Letters, v. 1, p.
355-358.
Scott, R. B., et al, 1974, The TAG Hydrothermal Field:
Nature, v. 251, p. 301-302.
Sillitoe, R. H., 1972, A plate tectonic model for the ori-
gin of porphyry copper deposits: Econ. Geology, v.
67, p. 184-197.
Sleep, N. H., and M. N. Toksoz, 1971, Evolution of
marginal basins: Nature, v. 233, p. 548-550.
Sloss, L. L., 1963, Sequences in the cratonic interior of
North America: Geol. Soc. America Bull., v. 74, p.
93-114.
Snead, R. E., 1972, Atlas of world physical features:
New York, Wiley, 158 p.
Sorokhtin, O. G., S. A. Ushakov, and V. V. Fedynskiy,
1974, Dynamics of crustal plates and origin of petro-
leum deposits: Akad. Nauk SSSR Doklady, v. 214, p.
1407-1410.
Spooner, E. T. C., and W. S. Fyfe, 1973, Sub-sea floor
metamorphism, heat and mass transfer: Contr. Min-
eralogy and Petrology, Springer- Verlag, p. 287-304.
Stern, C. R., and P. J. Wylie, 1973, Water-saturated and
undersaturated melting relations of a granite to 35
kilobars: Earth and Planetary Sci. Letters, v. 18, p.
163-167.
Strakhov, N. M., et al, 1968, Geochemistry of a sedi-
mentary manganese ore-forming process: Akad.
Nauk SSSR Geol. Inst. Trans., v. 185, 495 p.
Tarling, D. H., 1973, Continental drift and reserves of
oil and natural gas: Nature, v. 243, p. 277-279.
Van Royan, W., and O. Bowles, 1952, The mineral re-
sources of the world, in Atlas of the world's resources,
v. II: New York, Prentice-Hall, 181 p.
Vine, F. J., and J. T. Wilson, 1965, Magnetic anomalies
over a young oceanic ridge off Vancouver Island: Sci-
ence, v. 150, no. 3695, p. 485^189.
von der Borch, C. C, W. D. Nesteroff, and J. S. Gale-
house, 1971, Iron-rich sediments cored during Leg 8
of the Deep Sea Drilling Project, in J. I. Tracey, Jr., et
al, eds., Initial reports of the Deep Sea Drilling Pro-
ject, v. VIII: Washington, U.S. Govt. Printing Office,
p. 725-819.
Waring, G. A., 1965, Thermal springs of the United
States and other countries of the world: a summary:
U.S. Geol. Survey Prof. Paper 492, 383 p.
429
45
Reprinted from: Geological Society of America Bulletin. Vol. 87, 661-674.
Tectonic fabric and hydrothermal activity of
Mid-Atlantic Ridge crest (lat 26°N)
PETER A. RONA
REGINALD N. HARBISON*
BOBBY G. BASSINGER*
National Oceanic and Atmosphertc Administration, Atlantic Oceanographtc and Meteorological Laboratories, IS
Rickenbacker Causeway, Miami, Florida 13149
ROBERT B. SCOTT Department of Geology; Texas A&M University, College Station, Texas 77843
ANDREW J. NALWALKf Marine Sciences Institute, University of Connecticut, Croton, Connecticut 06 ?40
ABSTRACT
An asymmetric tectonic fabric was de-
lineated by narrow-beam bathymetric
profiles in a 180-km- area of the Mid-
Atlantic Ridge crest at lat 26°N. Features of
the tectonic fabric are a continuous rift val-
ley offset by small (<10-km) transform
faults and minor fracture zones expressed
as valleys with intervening ridges that trend
normal and oblique to the two sides of the
rift valley. The discharge zone of a pos-
tulated sub-sea-floor hydrothermal con-
vection system is focused by faults on the
southeast wall of the rift valley and driven
by intrusive heat sources beneath the rift
valley.
The rift valley has a double structure
consisting of linear segments, bounded by
ridges, and basins at the intersections of the
minor fracture zones. The double structure
of the rift valley acts like a template that
programs the reproduction of the tectonic
fabric. The minor fracture zones form an
asymmetric V about the rift valley at var-
iance with the symmetric small circles
formed by major fracture zones. To recon-
cile the asymmetry of minor fracture zones
with the symmetry of major fracture zones,
it is proposed that the minor fracture zones
have been preferentially reoriented by an
external stress field attributed to interplate
and intraplate motions. Major fracture
zones remain symmetric under the same
stress field owing to differential stability be-
tween minor and major structures of
oceanic lithosphere. Key words: oceanic
ridges, Mid-Atlantic Ridge, tectonic fabric,
fracture zones, transform faults, rift valley,
hydrothermal activity, hydrothermal min-
" Present address: U.S. Geological Survey, P.O. Box
7944, Metaine, Louisiana 70011.
t Deceased.
eral deposits, sub-sea-floor hydrothermal
convection system.
INTRODUCTION
The crest of the Mid- Atlantic Ridge at lat
26°N in the corridor of the Trans-Atlantic
Geotraverse (TAG) of the National Oceanic
and Atmospheric Administration (NOAA)
was selected for study because the oceanic
crust in this region is believed to have un-
dergone a long history of relatively normal
development isolated from mantle plumes
and multiple plate boundaries (Fig. 1;
Rona, 1973a). The tectonic fabric of the
ridge crest at lat 26°N should be useful as a
standard of crust produced from a slow-
spreading ridge (Menard, 1967).
An interdisciplinary study of the Mid-
Atlantic Ridge crest at lat 26°N performed
as part of the TAG project in 1972 and
1973 included narrow-beam bathymetric,
gravimetric, and magnetic profiling, dredg-
ing, coring, measurements of the thermal
structure and chemistry of the water col-
umn, and ocean-bottom photography. The
study resulted in the discovery of the TAG
Hydrothermal Field, an area of at least 15
km2, including the southeast wall of the rift
valley, where manganese oxide of hy-
drothermal origin is present (Fig. 2; Scott,
R. B., and others, 1974) and hydrothermal
solutions may be discharging from the sea
floor (Rona and others, 1975). Previous
evidence of the concentration of metals by
hydrothermal processes in ocean basins has
come from widespread metal enrichment in
sediment immediately overlying basalt gen-
erated by sea-floor spreading about di-
vergent plate boundaries (Degens and Ross,
1969; Bostrom and Peterson, 1966;
Bostrom and others, 1972; von der Borch
and Rex, 1971; von der Borch and others,
1971; Hollister and others, 1972; Dvmond
and others, 1973; Sayles and Bischoff,
1973; Piper, 1973).
This report describes the regional tec-
tonic fabric of a 180-km square of the
Mid-Atlantic Ridge crest at lat 26°N cen-
tered on the TAG Hydrothermal Field and
considers the relation of the tectonic fabric
to hydrothermal activity. The tectonic fab-
ric was delineated by a rectilinear grid of
narrow-beam bathymetric, gravimetric,
and magnetic profiles spaced about 10 km
apart parallel to and 20 km apart perpen-
dicular to the axis of the rift valley (Figs. 3
through 8). Directions of features are given
in the azimuth system (0° to 360° clockwise
0
7
/
/ ^l^>
- — -C_££cro,5 [ „.,
' ' ,' 40'N"_
' LWfT' — ^-—-LJ
Figure 1. Inset shows location of study area
on the crest of the Mid-Atlantic Ridge in the cor-
ridor (striped) of the NOAA Trans-Atlantic
Geotraverse (TAG) in the central North Atlantic
Ocean between southeastern North America and
northwestern Africa. Locations and trends of
transverse valleys (minor fracture zones) de-
lineated within the study area are shown by solid
lines. The axis of the Mid-Atlantic Ridge is
dashed. The locations of major fracture zones
(Atlantis and Kane) are shown.
Geological Society of America Bulletin, v. 87, p. 661-674, 1 1 figs.. May 1976, Doc. no. 60504.
661
430
662
RONA AND OTHERS
Figure 2. Photograph of first specimen of hydrothermal manganese oxide, 15.0 cm in diameter and
4.2 cm thick, recovered from the site of the TAG Hydrothermal Field in 1972 (Table 2, Station TAG
1972-13, see footnote 1; photograph by Wayne Stevens).
from north). Primary position control was
by a satellite navigational system with esti-
mated accuracy of ±1.0 km.
BATHYMETRY
Rift Valley
Features of the tectonic fabric in the
study area are the rift valley of the Mid-
Atlantic Ridge, three valleys with interven-
ing ridges transverse to the rift valley, and
valleys that transect the transverse ridges
(Figs. 3, 4). The rift valley trends northeast
and consists of two structural elements that
alternate along its axis. Linear segments of
the rift valley, 20 to 40 km in length, alter-
nate with irregularly shaped areas 15 to 25
km in length. The azimuth of the linear
segments is 25°. The width of the linear
segments at the 3,400-m isobath, which
continuously delineates the floor of the rift
valley, ranges between 5 and 15 km; the
width of the irregularly shaped areas ranges
between 15 and 25 km. The irregularly
shaped areas between the linear segments
are occupied by topographic depressions
that form basins between 200 and 500 m
below the 3,800-m isobath. The linear seg-
ments of the rift valley are successively off-
set in a right-lateral sense by as much as 10
km at the irregularly shaped areas.
Transverse Valleys and
Intervening Ridges
The transverse valleys and intervening
ridges on the southeast side of the rift valley
trend about 115°, nearly perpendicular to
the rift valley axis (Figs. 3, 4). Those on the
northwest side trend about 265°, oblique
(about 60°) to the axis of the rift valley
(Rona and others, 1973). The trend of the
transverse valleys and intervening ridges
appears to change near the northwestern
margin of the study area.
Intersections of the transverse valleys and
intervening ridges with the rift valley are
topographically complex (Fig. 3). Certain
of the transverse valleys are continuous
with the rift valley (the southernmost trans-
verse valley and the western half of the cen-
tral transverse valley). Other transverse val-
leys are blocked by topographic highs
within 25 km of the rift valley (the north-
ernmost transverse valley and the eastern
half of the central transverse valley). The
actual or projected juncture of each trans-
verse valley with the rift valley occurs at
one of the irregularly shaped areas where
the linear segments of the rift valley are off-
set. All transverse ridges continue up to the
rift valley, where they form the walls along
the linear segments.
Narrow-beam bathymetric profiles (ef-
fective total beam width about 20°) across
the three transverse valleys and intervening
ridges are shown in Figure 5. The normal
and oblique orientations of the transverse
valleys and ridges with respect to the axis of
the rift valley are apparent from the align-
ment of the transverse valleys. Width of the
transverse valleys at their floors is about 10
km, and the average spacing between adja-
cent valley floors is 55 km. Relief of the
intervening ridges measured from the valley
floors ranges between 1,000 and 2,000 m.
The transverse valleys appear linear and
continuous along their axis on the basis of
the bathymetric profiles (Fig. 5) and
isobaths (Fig. 3), with the exception of the
topographic complexities noted at certain
junctures with the rift valley. The floor of
each transverse valley progressively deepens
away from the rift valley, and the mean
depth successively increases from northeast
to southwest (Fig. 5).
Narrow-beam bathymetric profiles that
cross the study area transverse to the axis of
the rift valley (Fig. 6; also see McGregor
and Rona, 1975, Fig. 5) reveal four things:
(1) Echo returns are predominantly diffrac-
tions, as distinguished from specular
reflections (Hoffman, 1957), indicating that
the rock surface is irregular relative to the
12.5-cm wavelength of the projected acous-
tic pulse. (2) The floor of the rift valley is
irregular and has relief up to several
hundred metres that is related to the pres-
ence of discontinuous linear topographic
prominences that project from either wall
(Fig. 6, profiles C, D, E, H, J) or stand near
the center of the rift valley (Fig. 6, profiles B
through G), subparallel with the axis. (3)
Steplike topographic levels with relief of
hundreds of metres and widths of
kilometres are present on both walls of the
rift valley; the steplike levels may be corre-
lated along either wall for distances corre-
sponding to at least the width of each
transverse ridge (about 30 km; Fig. 6,
profiles B through G). (4) The mean inclina-
tion of the two walls of the rift valley ranges
between about 5° and 35°, with no systema-
tic difference apparent between the walls.
Branches df the Rift Valley
The rift valley has branches that extend
from either side (Figs. 3, 4). Branches ex-
tending from the southeast side trend nearly
parallel to the rift valley (25°). Some
branches that extend from the northwest
side trend nearly parallel (25°), and others
trend oblique (355°) to the rift valley.
Valleys parallel to the branches of the rift
valley transect the transverse ridges at ir-
regular intervals, with a minimum spacing
of 5 km. These valleys generally extend
only partly across the transverse ridges, al-
though some valleys extend across the
ridges. The valleys that transect the trans-
verse ridges on the southeast side of the rift
valley trend nearly parallel to the rift valley
(25°) and perpendicular to the transverse
ridges. The valleys that transect the trans-
verse ridges on the northwest side of the rift
valley have two trends corresponding to the
two trends of the branches on the north-
west side of the rift valley: (1) a trend (355°)
oblique to the axis of the rift valley and
nearly perpendicular to the transverse
ridges, and (2) a trend (25°) parallel to the
rift valley and oblique to the transverse
ridges. The two trends appear to intersect.
TAG Hydrothermal Field
The TAG Hydrothermal Field occupies a
salient of the southeast wall that projects
into the rift valley between depths of about
3,500 and 2,000 m (Fig. 3). The salient is
431
45°30'W
TECTONIC FABRIC AND HYDROTHERMAL ACTIVITY OF MID-ATLANTIC RIDGE CREST
45°00' 44°30' 44°00'
663
43°30'W
46°00'W
45°30
45°00'
44°30'
Figure 3. Isobath map of the study area (Fig. 1) contoured in hundreds of metres corrected for ship's draft and vertical sounding velocity (Matthews,
1939). Depths exceeding 3,400 m are shaded. Sounding tracks are dashed. The known portion of the TAG Hydrothermal Field on the southeast wall of
the rift valley is outlined (trapezoid). Dredge and core stations are marked with dots (Table 2). Dots are omitted for three dredge stations within the
TAG Hydrothermal Field (Table 2, stations TAG 1972-13, TAG 1973-2A, TAG 1973-3A). Values of heat flow in heat-flow units are shown at five
stations marked by crosses (Langseth and others, 1972). The northeast-trending sounding tracks of bathymetric profiles 1 through 4 shown in Figure 5
are labeled (solid lines). The northwest-trending sounding tracks of bathymetric profiles A through K shown in Figure 6 are labeled. Note the north
arrow in the upper left corner. Contour interval 200 m.
the end of a transverse ridge and forms the
wall along a linear segment of the rift valley
just north of the intersection with the cen-
tral transverse valley. This transverse ridge
is transected by more valleys perpendicular
to its axis than any other ridge in the study
area.
The southeast wall of the rift valley at the
TAG Hydrothermal Field has steplike
topographic levels several kilometres wide
at depths of about 3,200 and 2,500 m (Fig.
6, profile F). Two stereophotograph tran-
sects of this area resolve steps on the wall
that are 100 to 300 m wide with 75-m av-
erage relief between depths of 3,400 and
3,100 m, 30 to 400 m wide with 40-m av-
erage relief between depths of 2,800 and
2,400 m, and smaller steps between depths
of 2,500 and 2,400 m, indicating that steps
in the wall range in scale from metres to
kilometres (McGregor and Rona, 1975).
GRAVITY
Measurements of gravity were made with
a shipborne Graf-Askania gravimeter to an
estimated accuracy of ±5 mgal. The free-air
432
664
RONA AND OTHERS
gravity anomalies of the study area range
between -40 and +80 mgal (Fig. 7). Nega-
tive free-air gravity anomalies coincide with
the rift valley and the transverse valleys.
Positive free-air anomalies coincide with
the transverse ridges, and largest values are
on the highest parts of the transverse ridges
adjacent to the rift valley. The TAG Hy-
drothermal Field occurs adjacent to such
an anomaly. The values of gravity agree in
magnitude with a rough calculation of the
terrain effect. These observations suggest
that the free-air anomalies are primarily
due to topography.
MAGNETIC MEASUREMENTS
Linear magnetic anomalies parallel the
rift valley (Fig. 8). The linear anomalies are
offset at transverse valleys in direction and
amount corresponding to the offset of the
linear segments of the rift valley. Low to
negative values of residual magnetic inten-
sity are associated with the transverse val-
leys. The linear magnetic anomalies are in-
terpreted to have been generated during the
Brunhes (axial anomaly), Matuyama,
Gauss, and Gilbert magnetic polarity
epochs between 0 and 5.8 m.y. B.P. The
positive axial anomaly does not coincide
with the axis of the rift valley but is cen-
tered over the southeast wall 5 km from the
axis. About 2 km of the 5-km offset may be
attributed to shift in the magnetic field
owing to superposition of the sea floor and
regional magnetic fields. The polarity of the
residual magnetic field is indistinct between
the end of the Gilbert epoch and anomaly 5
(10 m.y. B.P.) of the magnetic polarity re-
versal time scale (Heirtzler and others,
1968) identified about 120 km to either side
of the rift valley. Half rates of sea-floor
spreading measured perpendicular to the
axis of the rift valley and averaged over var-
ious intervals to 10 m.y. B.P. are asymmet-
ric, being faster to the southeast than to the
northwest (Table 1).
The TAG Hydrothermal Field is situated
within the positive axial anomaly and is as-
sociated with irregularities in the shape of
that anomaly (Fig. 8). A more detailed
magnetic survey reveals a pronounced low
in the axial anomaly that coincides with
the TAG Hydrothermal Field (Fig. 9;
McGregor and Rona, 1975). Magnetic
modeling indicates that the magnetic low is
due to reduction in intensity of remanent
magnetization (McGregor and others,
1976) that may be attributed to alteration
of the magnetic mineral component in
basalt by hydrothermal solutions (Luyen-
dyk and Melson, 1967; Watkins and Pas-
ter, 1971; Ade-Hall and others, 1971).
Hydrothermal alteration of basalt is evi-
denced by the presence of greenstone at
4S°30
— r-
1
\ ... '
vj*
IRREGULARIY.
■SHAPta I
\AREA
TRANS
RIDGE
TRANSVERSE
VAUEY
AV VA>IlLl|l°>„I i
• I N z. <•* .\0°v \lRREGUlARtY .™*NSV"SE VAllEY
<-. Vf. \ - N SHAPED
TRANSVERSE RIDGE
S$&\ \ . IRREGUIARIY
\* \ 0cA \ SHAPED
1* IV" A,EA I»ANSVE>SE
.^ , r
^\
Figure 4. Diagram-
matic representation
of the tectonic fabric
of the study area out-
lined from Figure 3.
Azimuths of the major
trends are noted. The
TAG Hydrothermal
Field is outlined (trap-
ezoid). Note the north
arrow in the upper left
corner.
several sites along transverse ridges (Table
2"; stations TAG 1972-8, TAG 1972-15).
HEAT FLOW
Five measurements of conductive transfer
of heat through the sea floor were made by
Langseth and others (1972). The heat-tlow
measurements show higher heat flow
through the transverse ridges (3.26 to 8.60
HFU) than through the intervening trans-
verse valleys (2.04 to 2.69 HFU; Fig. 3).
From the limited number of measurements,
it is ambiguous whether this distribution of
values is related to topography or to dis-
tance from the rift valley. A large variation
in heat flow occurs over a horizontal dis-
tance of 5 km on one of the transverse
ridges (6.75 and 3.26 HFU).
A water-temperature profile parallel to
the ocean bottom over the southeastern
wall of the rift valley at the TAG Hydro-
thermal Field was made with a 4-m-long
vertical array of three thermistors mounted
on a towed deep-sea camera (Rona and
others, 1975). The profile revealed an ab-
rupt anomaly of +0. 1 1°C associated with a
gradient of 0.008°C/m, warming down-
ward within 20 m of the bottom along a
horizontal distance of about 350 m be-
tween depths of 3,030 and 2,950 at a step-
like level on the southeast wall (Fig. 6,
profile F), where hydrothermal material
'Table 3, "Rocks recovered from the Study Area of
the NOAA Trans-Atlantic Geotraverse (TAG) on the
Mid-Atlantic Ridge Crest at lat. 26°N," GSA sup-
plementary material 76-4, may be ordered from Docu-
ments Secretary, Geological Society of America, 3300
Penrose Place, Boulder, Colorado 80301.
DISTANCE (KILOMETERS)
0 50 100 150
SW
Figure 5. Digitzed reproductions of narrow-
beam bathymetric profiles 1 through 4 recorded
along sounding tracks parallel to the rift valley
located in Figure 3. The three transverse valleys
correlate (dashed lines) as continuous features on
either side of the rift valley, consistent with the
isobath map (Fig. 3). Vertical exaggeration is
about x60.
433
TEC IONIC FABRIC AND HYDROTHKRMA1 ACTIVITY OF MID-ATLANTIC RIDGE CREST
665
was dredged .Tabic 2. dredge station I AC.
1973-3A). The source of the water-
temperature anomah may be either dis-
charge of hvdrothermal solutions or con-
ductive transfer ot heat from the ocean bot-
tom. 1 he abrupt, narrow character of the
anomalv, inverted gradient, and associated
geological and geochemical evidence favor
a hvdrothermal source.
PETROLOGY
Rocks recovered by dredging and coring
from the different features of the tectonic
fabric in the study area are described in
Table 2 i see footnote 1); sampling sites are
marked in Figure 3. Fresh pillow basalt was
recovered from the topographic promi-
nences t)ii the floor of the ritt valley (Table
2, stations TAC, 1972-17, TAG 1973-4C,
TAG 19~3-4G). Fresh basalt was also re-
covered from the walls of the rift valley. A
varied suite of altered and metamorphosed
rocks and limestone in addition to basalt
was recovered from the transverse ridges. A
coarse-grained cumulate gabbro exposed in
the wall of a valley that transects one of the
transverse ridges may be a magma chamber
formed in laver 3 of oceanic crust (Table 2,
stations TAG 1973-7A, TAG 1973-7B, TO
75AK60-2A).
Manganese oxide was consistently
dredged from the southeast wall of the rift
valley within the area of the TAG Hydro-
thermal Field (Fig. 3; Table 2, stations TAG
1972-13, TAG 1973-2A, TAG 1973-3 A).
The manganese oxide occurs as a crust up
to 42 mm thick on basalt talus, as veins in
the talus fragments, and as a crust on and
matrix in a breccia of basalt fragments.
Stereophotograph transects of the rift valley
wall show that the manganese oxide is as-
sociated with sediment-free talus and
breccia-covered inner portions of the steps
observed on the southeast wall between
depths of 3,100 and 2,500 m (McGregor
and Rona, 1975). The accumulation rate of
the manganese oxide measured from the
growth rate of Th-:i"and Pa-" toward secu-
lar equilibrium with their uranium parents
is about 200 mm/10" yr, two orders of
magnitude faster than deep-sea hydrogen-
ous ferromanganese nodules and crusts
(Scott, M. R., and others, 1974). Atomic
absorption spectrophotometry of the man-
ganese indicates that the Mn content is 40
percent, Fe less than 0.1 percent, Al 0.1
percent, Zn 100 ppm, Cr 15 ppm, Ni 300
ppm, Co 20 ppm, and Cu 30 ppm; deep-sea
hydrogenous ferromanganese nodules and
crusts contain less Mn (10 to 20 percent)
and considerably more of the other ele-
ments (Scott, M. R., and others, 1974). The
relatively rapid rate of accumulation and
the pure composition of the manganese
oxide indicate concentration by a hy-
drothermal mechanism capable of extreme
chemical fractionation. A crust consisting
of an upper layer of hydrogenous man-
ganese oxide up to 2 mm thick underlain by
a lower layer of hydrothermal manganese
oxide up to 10 mm thick was recovered
from a site 12 km southeast of the sites at
the southeast wall of the rift valley on the
same transverse ridge (Fig. 3; Table 2, sta-
TABLE
HALF RATES OF SEA-FLOOR SPREADING ABOUT THE MID-ATLANTIC RIDGE
Reference
Latitude on Mid-
Atlantic Ridge
Orientation
Averaging interval
van Andel and Moore
(1970)
Lattimore and others
(1974)
McGregor and others
(1976)
McGregor and others
(1976)
Present paper
Needham and
Francheteau (1974)
Needham and
Francheteau (1974)
Greenewalt and Taylor
(1974)
Macdonald and others
(1975b)
Macdonald and others
(1975 b)
Loncarevic and Parker
(1971)
6° to 8°S Perpendicular to axis
of rift vallev
26CN Perpendicular to axis
of riff valley
26°N Perpendicular to axis
of rift vallev
26°N Perpendicular to axis
of rift valley
26°N Normal (10"°) and
oblique (265°) to axis
of rift valley (25°)
36°N Perpendicular to axis
of rift valley
36°N Perpendicular to axis
of rift valley
36°N Perpendicular to axis
of rift valley
36°N Perpendicular to axis
of rift valley
36°N Perpendicular to axis
of rift valley
45°N Perpendicular to axis
of rift valley
Axis of rift valley
(0 m. v. B.P.) to anomaly
3 (4.5 m.y. B.P.)
Anomaly 3 (4.5 m.y. B.P.)
to anomaly 5 (10 m.y. B.P.)
Axial anomaly to anomaly
5 (10 m.y. B.P.)
Axis of rift valley
(0 m.y. B.P.) to Brunhes-
Matuyama boundary
(0.69 m.y. B.P.)
Matuyama epoch
(0.69 to 2.43 m.y. B.P.)
Axis of rift valley
(0 m.y. B.P.) to anomaly
5 (10 m.y. B.P.)
Axis of rift valley
(0 m.y. B.P.) to Brunhes-
Matuvama boundary
(0.69 m.y. B.P.)
Matuyama epoch
(0.69 to 2.43 m.y. B.P.)
Axis of rift valley
(0 m.y. B.P.) to Brunhes-
Matuyama boundary
(0.69 m.y. B.P.)
Axis of rift valley
(0 m.y. B.P.) to anomaly
2 (1.8 m.y. B.P.)
Anomaly 2 (1.8 m.y. B.P.)
approximately to anomaly
4 (8 m.y. B.P.)
Center of Brunhes
(0 m.y. B.P.) to anomaly
5 (10 m.y. B.P.)
Average h;
alf rate
Direction
(cm/yr)
2.16 ±
0.24
E
1.89 ±
0.04
W
1.59 ±
0.24
E
1.12 ±
0.08
W
1.3
SE
1.1
NW
1.7
SE
0.7
NW
1.3
SE
1.1
NW
1.3
SE
1.3
NW
1.5
E
0.7
W
1.3
E
0.9
W
1.3
E
1.0
W
1.33
E
0.70
W
0.95
E
1.35
W
1.10
E
1.28
W
434
666
RONA AND OTHERS
nun TO 75AK61-1A; Scott and others,
1975). Seven attempts to dredgt the north-
west wall of the n ft valley opposite the
TAG Hydrothermal Field to determine
whether hydrothermal deposits were sym-
metrically disposed about the rift valley re-
covered only a few fragments of basalt
(Table 2, station 1 AC, I97.3-5C). The hy-
drothermal material is sufficiently triable
that samples would probably have been re-
covered if present on the northwest wall.
DISCUSSION
Comparison of Tectonic Fabric
The continuous rift valley at lat 26°N
consisting of linear segments and basins
(Figs. 3, 4) is similar to that observed along
the Mid-Atlantic Ridge between lat 22CN
and 23°N (van Andel and Bowm, 1968), at
lat 36CN (Needham and Francheteau,
1974), at lat 45°N (Aumento and others,
1971), and between lat 47°N and 51CN
(Johnson and Vogt, 1973). The topo-
graphic prominences on the floor of the rift
valley at lat 26°N, from which fresh basalt
was recovered, appear similar to the discon-
tinuous medial ridge as much as 250 m high
by 1 km wide and as much as 4 km long
described from the rift valle\ at lat 36CN
(Bellaiche and others, 1974; Needham and
Francheteau, 1 9^4; Moore and others,
1974; Macdonald and others, 1975a), in-
terpreted as a locus of crustal emplacement
and basalt eruption.
Steplike levels and steps in the walls of
the rift valley on scales ranging from metres
to kilometres similar to those at lat 26°N
have been observed in walls of the rift val-
ley at lat 36GN (Needham and Francheteau,
1974) and at the Gorda Rise, where the
steps are interpreted as fault blocks (Atwa-
ter and Mudie, 1973). Block faulting involv-
ing uplift is evidenced in the study area by
the exposure of the cumulate gabbro from a
deeper crustal level in an inferred fault
scarp 1.2 km high (Table 2, stations TAG
1973-7A, TAG 19~3-7B, TO T"5AK6()-2A)
and by the exposure of greenstone in the
walls of transverse ridges. Transverse ridges
with intervening valleys that intersect and
offset the rift valley at lat 26°N at a spacing
of 55 km occur at an average spacing of 65
km along the Mid-Atlantic Ridge between
lat 10°N and 40°N (Perry and Feden, 1974)
and are present at lat 36°N (Derrick and
others, 1973). The portions of the trans-
verse valley that transect the rift valley are
the loci of earthquake epicenters at lat 26°N
(McGregor and Rona, 1975) and at lat
37°N (Reid and Macdonald, 1973).
The asymmetry in tectonic fabric, rates of
sea-floor spreading, and position of the
axial magnetic anomaly at lat 26CN is ob-
served at other intensively studied sites
along the Mid-Atlantic Ridge. Transverse
valleys and intervening ridges that trend
normal and oblique to the southeast and
northwest sides of the rift valley, respec-
tively, at lat 26°N are also present at lat
36°N (Detnck and others, 1973; H. Flem-
ing, 1975, personal commun.), and at lat
36°N, where the normal and oblique sides
reverse (Bhattacharyya and Ross, 1972).
Between lat 47°N and 51°N, the azimuths
of linear segments of the rift valley alternate
north and northwest; transverse ridges
occur adjacent to the former and interven-
ing transverse valleys occur adjacent to the
latter, forming a V-shaped pattern that is
asymmetric about the rift valley (Johnson
and Vogt, 1973, Fig. 7).
Half rates of sea-floor spreading mea-
sured perpendicular to the axis of the rift
valley of the Mid-Atlantic Ridge and aver-
aged over corresponding time intervals are
faster to the east than to the west at lat
36°N, similar to the half rates at lat 26°N
AXIAL VALLEY
D
H
K 3000 h
4000_
OKM 100 200
Figure 6. Digitized reproductions of narrow-beam bathymetric profiles recorded along sounding
tracks A through K transverse to the rift valley located in Figure 3. The rift valley and the TAG
Hydrothermal Field (arrow on profile F) are noted. Steplike topographic levels on the rift valley walls
are tentatively correlated by dashed lines between the profiles. Faults inferred between fault blocks
are indicated by dashed lines on profiles. Vertical exaggeration is about X 15.
435
TECTONIC FABRIC AND HYDROTHERMAL ACTIVITY OF MID-ATLANTIC RIDGE CREST
667
(Table 1). Conversely, average half rates of
spreading at lat 45°N are faster to the west
than to the east (Table 1). The axial anom-
aly is centered-over the southeast wall offset
about 5 km from the axis of the rift valley
at lat 36°N (Needham and Francheteau,
1974) and at lat 22°N (van Andel and
Bowin. 1968, Fig. 13), similar to the posi-
tion of the axial anomaly at lat 26CN (Fig.
8). Earthquake epicenters along the linear
segments of the rift valley may favor the
southeastern side at lat 26°N (McGregor
and Rona, 1975) and lat 36°N (Spindel and
others, 1974) and the west side between lat
47°N and 51°N (Johnson and Vogt, 1973).
Tectonic Fabric and Hydrothermal Activity
The concentration of a hydrothermal
mineral deposit in the Earth's crust requires
special physical and chemical conditions.
The emplacement of igneous rocks and the
propagation of fractures and faults at di-
vergent plate boundaries are conducive to
the development of sub— sea-floor hydro-
thermal convection systems (Spooner and
Fyfe, 1973; Hutchinson, 1973; Sillitoe,
1973; Rona, 1973b; Lister, 1974b; Lowell,
1975; Bonatti, 1975). Dense, cold sea water
may penetrate down fractures and acquire
heat at depth, and less dense hydrothermal
water may ascend and discharge as sub-
marine springs through fracture systems.
The circulating sea water may remove ele-
ments including heavy metals from the
oceanic crust through which it circulates by
high-temperature leaching and mass trans-
fer (Krauskopf, 1956; Holland, 1967;
Helgeson, 1964; Corliss, 1971; Hart,
1973). Metals including Fe, Mn, Cu, and
Ni have been experimentally leached by sea
water from basalt at 200° to 500°C and 0.5
to 1 kb (Mottl and others, 1974; Bischoff
and Dickson, 1975). The convection of sea
water as a hydrothermal solution through
rocks emplaced at divergent plate bound-
aries is indicated by several lines of evi-
dence. (1) Low values of heat flow from
oceanic ridge crests imply that heat must be
removed by water circulation (Palmason,
1967; Deffeyes, 1970; Talwani and others,
1971; Lister, 1972; Anderson, 1972; Wil-
liams and others, 1974; Sclater and others,
1974). (2) Hydrous metamorphosed
oceanic crust and oceanic serpentine re-
quire a voluminous source of water
(Miyashiro and others, 1971; Christensen,
1972). (3) Isotopic compositions of the
hydrated rocks require a low 501S isotopic
source such as sea water (Muelenbachs and
Clayton, 1972; Spooner and others, 1974).
v The TAG Hydrothermal Field is hypoth-
Figure 7. Map of free-air gravity contoured in milligals. Gravity highs (H) and lows (L) are
indicated. The ship's tracklines are dashed. The TAG Hydrothermal Field is outlined (trapezoid). The
axis of the rift valley along the linear segments is shown. Note the north arrow in the upper left corner.
esized to be the zone where a voluminous
sub — sea-floor hydrothermal convection
system discharges through faults between
steps of the southeast wall of the rift valley
that act as conduits for the hydrothermal
solutions (Figs. 9, 10). The system may be
charged with sea water through fractures
that underlie both the adjacent transverse
valleys and the many valleys that transect
the adjacent transverse ridge. Experimental
(Elder, 1965) and theoretical (Donaldson,
1962; Elder, 1967) modeling shows that
hydrothermal discharge is localized, and re-
charge is delocalized. Discharge is confined
to vertically rising narrow jets (Wooding,
1963) and fracture-focused streams (Elder,
1965). Recharge occurs over large areas
and involves downward water flow at rates
that are fast enough to reduce the upward
conductive heat flux (Wooding, 1963).
Heat-flow measurements from the study
area and from other areas on oceanic ridges
are consistent with downwelling of sea
water at valleys and upwelling at ridges
(Lister, 1972), a circulation pattern favored
by a geometric forcing effect of the topog-
raphy (Lister, 1974a, 1974b). Such hy-
drothermal circulation can account for the
low-intensity hydration and metamorphism
under the influence of geothermal gradients
that are higher than background value in
certain rocks recovered from the transverse
ridges (Table 2).
The southeast wall of the rift valley at the
TAG Hydrothermal Field projects over the
locus of crustal emplacement beneath the
rift valley, the source of heat that vigor-
ously drives the ascending limb of the
hypothetical hydrothermal convection sys-
tem. As hydrothermal solutions enriched in
heavy metals complexed with sea-water
chlorides ascend through the rocks, metals
may combine with sulfur in the sea water
and precipitate sulfides under reducing
conditions within the basalt (Meyer and
Hemley, 1967). Chalcopyrite, pyrrhotite,
and marcasite have been experimentally
grown from sea water during reaction with
oceanic tholeiite at 500°C and 0.8 kb (Ha-
jash, 1974). Metallic oxides may form
under oxidizing conditions at the basalt-
sea-water interface. Amorphous ferric hy-
droxide may precipitate as a colloid in the
overlying sea water (Zelenov, 1964) and
scavenge the metals remaining in solution
or in admixed sea water by absorption
(Krauskopf, 1956). Manganese oxide of
hydrothermal origin has been sampled from
the basalt -sea-water interface (Scott, M.
R., and others, 1974), a positive tempera-
ture anomaly (0.1 1°C) and inversion of the
gradient have been measured in near-
bottom water (Rona and others, 1975), and
amorphous particulate matter enriched in
iron and manganese has been sampled from
the water column overlying the TAG Hy-
drothermal Field (Betzer and others, 1974).
436
668
RONA AND OTHERS
Metallic sulfides have not been sampled
from the TAG Hydrothermal Field, but
their presence is suspected from geochemi-
cal considerations and analogy with ophio-
lites. Disseminated pyrite occurs in green-
stone dredged from a transverse ridge in the
study area (Table 2, station TAG 1972-15),
and metallic sulfides are common in altered
oceanic rocks (Dmitriev and others, 1970;
Bonatti, 1975). Certain ophiolites, includ-
ing those of Newfoundland (Upadhyay and
Strong, 1973) and the Troodos Massif of
Cyprus (Constantinou and Govett, 1973;
Hutchinson, 1973; Robertson and Hudson,
1973), exhibit an association of cupreous
pyrite bodies, altered pillow lava, and over-
lying manganiferous oxides attributed to
hydrothermal processes similar to those in-
ferred to be operating on the Mid-Atlantic
Ridge at lat 26°N. Troodos-type massive
stratiform cupreous sulfide bodies may be
forming in pillow lava under the manganese
oxide deposits at the TAG Hydrothermal
Field (Fig. 10; Rona, 1973b; Scott, R. B.,
and others, 1974). It is infeasible to test this
hypothesis by deep-sea drilling at this time
because techniques constrain drilling to
areas of sediment accumulation that, as
topographic lows, are the inferred inflow
areas of sub-sea-floor hydrothermal con-
vection systems, and hydrothermal deposits
would be absent. The massive stratiform
sulfides would be expected to underlie hy-
drothermal discharge areas at topographic
highs expressed as transverse ridges.
Development of Tectonic Fabric
Why is a crustal layer of uniform thick-
ness not generated about the rift valley in-
stead of a layer of varying thickness that
forms topographic highs (transverse ridges)
and lows (intervening transverse valleys)?
Why is the tectonic fabric not symmetric
about the rift valley? How can the asym-
metry of the tectonic fabric be explained?
How is hydrothermal activity related to tec-
tonic fabric? These questions pinpoint
problems addressed in a model that de-
scribes the development of the tectonic fab-
ric of the study area (Fig. 11).
In the initial configuration of the model,
the sea floor spreads symmetrically at equal
half rates (r ,) perpendicular to either side of
the axis of a rift valley (Fig. 11a). The rift
valley consists of alternating linear seg-
ments and irregularly shaped areas. The ir-
regularly shaped areas are transected by
minor fracture zones that successively offset
the linear segments in a right-lateral sense.
These minor fracture zones are considered
to originate as small ridge-ridge transform
faults on the basis of their characteristics in
the study area, including offset of the linear
segments of the rift valley and linear rema-
nent magnetic anomalies, presence of
topographic depressions, and seismicity.
46'00'W 45-30'w J500W <U°30'W
Figure 8. Map of residual magnetic intensity contoured in hundreds of gammas. The regional field
was removed (IAGA, 1969), and a constant 300 y was added to the resultant field to balance the
distribution of positive and negative values. The TAG Hydrothermal Field is outlined (trapezoid). The
axis of the rift valley along the linear segments is shown. Note the north arrow in the upper left corner.
The sequence of normal (shaded) and reversed (unshaded) magnetic polarity epochs is labeled as
follows (Talwani and others, 1971, Fig. 11): 1. Brunhes, 0 to 0.69 m.y.; 2. Matuyama, 0.69 to 2.43
m.y.; 3. Gauss, 2.43 to 3.32 m.y.; 4. Gilbert, 3.32 to 5.18 m.y.
Spacing between the small transform faults others, 1970). The propagation of the frac-
and sense of offset may be related to the
geometry of the rift between the continental
margins of Africa and North America (Wil-
son, 1965), as demonstrated for the mar-
gins of the Red Sea rift (McKenzie and
ture zones may be related to thermal con-
traction of the lithosphere (Turcotte, 1974;
Collette, 1974).
Crustal material is emplaced by a dike-
injection mechanism beneath the rift valley,
TABLE 2. PARTIAL LIST OF POSITIONS OF ROCKS
RECOVERED FROM THE STUDY AREA
Station
Latitude
Longitude
Depth
(corrected m)
TAG 1972-2A
TAG 1972-3A
TAG 1972-8
TAG 1972-13
TAG 1972-15
TAG 1972-17
TAG 1973-2 A
TAG 1973-3A
TAG 1973-4C
TAG 1973-4G
TAG 1973-5C
TAG 1973-7A
TAG 1973-7B
TO 75AK60-2A
T0 75AK61-1A
26°09.7'N
44°47.4"W
3,240
26°07.3'N
44°48.8'W
3,170
25°22.5'N
44°54.8'W
2,820
26°08.0'N
44°45.0'W
3,080
26°33.9'N
44°30.0'W
3,400
26°44.5'N
44°37.2'W
3,590
26°09.7'N
44°47.4'W
3,240
26°07.3'N
44°48.8'W
3,170
26°18.2'N
44°42.1"W
4,262
26°19.2'N
44°44.1'W
4,060
26°13.8'N
44°57.0'W
3,010
26°15.3'N
44°27.0'W
3,390 to 2,410
26°15.3'N
44°27.0'W
3,390 to 2,410
26°17.8'N
44°24.0'W
2,520 to 1,920
26°07.6'N
44°40.5'W
2,600 to 2,000
437
IU |n\li FAKRU \ND HYDROTHFRMAI AC I IVlTi OF MID-ATLANTIC RIDGE CREST
669
20 0 20
DISTANCE (KILOMETERS
20 0 20
DISTANCE (KILOMETERS)
BATHYMETRY IN METERS MAGNETIC ANOMALY IN GAMMAS
Figure 9. Bathvmctric and residual magnetic profiles across rift valley of Mid-Atlantic Ridge near
lat 26°N McGregor and Rona, r>_s). a. Profiles 12 km north of TAG Hydrothermal Field; this
magnetic profile has large axial lBrunhes) positive anomaly over rift valley typical of profiles outside
TAG H\ drotherm.il Field, b. Profiles across ~[ AG Hydrothermal Field. Arrow on magnetic profile
points to magnetic low in positive axial Rrunhes) anomaly. Reduction in magnetic intensity may
provide a useful criterion in exploration for active or relict submarine hydrothermal mineral deposits.
which is a zone' of extension [Hebzen and
others. 1959; Matthews and Bath, \^b~;
Harrison, |96S; C ami. |9_0). The dikes
and the sides of the rift valle) are parallel.
The crustal material remains near its level
of emplacement in the form of relative
topographic lows at the irregularly shaped
areas because the cold walls at the intersec-
tion of the rift vallev. with the transverse
tractnres ma\ cause upwelling material to
solidify below its isostatic equilibrium level
(Sleep and Biehler, 1970).
The crust is at least 1 . s km thinner at the
transverse tractnres than at the intervening
transverse ridges interred from the differ-
ence of topographic relief. Crustal material
also solidifies below its eventual isostatic
equilibrium level beneath linear segments of
the rift valley (Sleep, 1969), where the ma-
terial is isostaticallv uplifted in fault blocks
as a result of processes of crustal thickening
and extension (Deffeyes, 1970; Osmaston,
1971; Lachenhruch, J973; Parker and Old-
enburg, 19-s; S'eedham and Francheteau,
1974).
Sea-floor spreading at equal rates per-
pendicular to either side of the rift valley
results in the generation of transverse val-
leys about the topographic lows at the ir-
regularly shaped areas of the rift valley and
intervening transverse ridges about the
topographic highs at the linear segments.
The transverse ridges are constructed of
fault blocks that are uplifted from the floor
and accrete at the walls of the rift valley;
each fault block is several kilometres wide,
and its long axis lies parallel to the ntt
valley.
The transverse ridges are transected per-
pendicular to their axes by valleys. I he \ al-
ley s originate as branches nearly parallel to
the rift \ allev that remain near the level of
the rift valley floor during differential uplift
of the adjacent fault blocks. The branches
are inactive compared with the rift valley,
which is active because it is coincident with
the locus of crustal emplacement. The rift
zones of Iceland (Kjartansson, 1960, 1962,
1965, 1968, 1969) and the Afar region
(Lowell and Genik, 1972) also exhibit
branches subparallel to the active rift. As
the sea floor spreads, the branches split off
from the rift valley and form the valleys
that transect the transverse ridges. Lengths
of crust / 1 and /_. are generated per unit time
perpendicular to the axis of the rift valley,
such that /, = /j.
The second configuration of the model
(Fig. lib) introduces apparent asymmetric
directions and half rates of sea-floor spread-
ing. The half rate of spreading r
perpendicular to the rift valley decreases 15
percent to the left (r.,) relative to the right
side (r,), corresponding to half rates of 1.1
and 1.3 cm/yr averaged over an interval of
10 m.y. in the study area (Table 1); the cor-
responding lengths of crust generated per
unit time are / ., and / ,, such that I , < I ,. The
direction of spreading remains perpendicu-
lar to the axis of the rift valley to the right
and reorients 30° to the left, corresponding
to the trends of the transverse valleys on
either side of the rift valley in the study area
(Figs. 3, 4). Solving for the half rate of
spreading in the reoriented direction to the
left (r:1) using values from the study area, r:l
= r.,/eos 30° = 1.3 cm/yr. The crustal length
generated at spreading half rate r , per unit
time is/-,. LIsing values from the study area,
r | = r ., and /, = /.,.
The asymmetric directions and equal half
rates (r,, r:!) of spreading introduced in the
second configuration of the model (Fig.
I lb) continue for 6.5 m.y. to produce the
third configuration (Fig. I 1c). Transverse-
valleys and intervening ridges are generated
about topographic highs and lows, as previ-
ously described (Fig. I la). However, these
features are oriented normal (right side)
and oblique (left side) to the rift valley, that
is, aligned with the true relative directions
of spreading. In spite of the reorientation of
spreading direction, the long axes of the
fault blocks uplifted from the rift valley
floor remain parallel to the axis of the rift
valley. This parallelism accounts for the
parallelism of the walls observed along
linear segments of the rift valley. Do the
long axes of the fault blocks remain parallel
to the axis of the rift valley during oblique
spreading, or do the blocks rotate until the
long axes become perpendicular to the re-
oriented spreading direction? If the fault
blocks rotated during spreading, then a
delta-shaped gap would be expected to
form between the unrotated blocks ad-
jacent to the rift valley and the rotated
blocks away from the rift valley on the side
of oblique spreading. The apparent absence
of such gaps in the study area (Fig. 3) makes
such rotation unlikely. Rather, the long
axes of the fault blocks probably remain
nearly parallel to the axis of the rift valley
during both symmetric and asymmetric
spreading, as shown in the model (Fig. 1 1).
Two sets of branches of the rift valley
form, corresponding to the two sets ob-
served in the study area (Figs. 3, 4). One set
is parallel to the rift valley and the long axes
of the fault blocks. This set controls the
structural development of the valleys that
transect the transverse ridges perpendicular
to their axes on the normal spreading
(right) side of the model and controls de-
velopment of a secondary trend that tran-
sects the transverse ridges on the oblique
spreading (left) side of the model. The sec-
ond set of branches is oblique to the rift
valley and to the long axes of the fault
blocks. This second set appears only on the
oblique spreading (left) side, where it con-
trols the structural development of the val-
leys that transect the transverse ridges per-
pendicular to their axes. The actual
bathymetric features transecting the trans-
verse ridges on the northwest side of the rift
valley in the study area appear to be a com-
posite of at least these two trends (Fig. 3).
The central ridge on the right side of the
rift valley is transected by several closely
spaced valleys that originated as branches
of the rift valley (Fig. lie). These valleys
facilitate the inflow of sea water to charge a
sub — sea-floor hydrothermal convection
system. Hydrothermal deposits shown in
black in Figure 1 lc form at and adjacent to
the wall of the rift valley. Relict deposits
extend along the transverse ridge as a con-
sequence of sea-floor spreading (Rona,
1973b). The relict hydrothermal man-
ganese recovered 12 km along flow lines of
438
670
RONA AND OTHERS
0 (KM)
Figure 10. Diagrammatic sketch of a sub-sea-floor hydrothermal con-
vection system like that hypothesized to exist at the TAG Hydrothermal
Field (drawn from profile F, Fig. 6; vertical exaggeration is about x2).
Arrows indicate directions of hydrothermal flow. Slant lines indicate direc-
tions of maximum permeability controlled by structural grain, including
fractures, faults, and dikes. Hydrothermal deposits are forming adjacent to
the rift valley, and relict deposits are present away from the rift valley as a
consequence of sea-floor spreading. Actually, relict deposits may be covered
by off-axis volcanism. Symbols: +, zone of recharge; -, zone of discharge;
x, zone of igneous intrusion.
sea-floor spreading southeast of the active
site at the wall of the rift valley (Table 2;
station TO 75AK61-1A) indicates the per-
sistence of the special structural and ther-
mal conditions that maintain sub— sea-floor
hydrothermal convection for at least 1 m.y.
Off-axis extrusive volcanism may act both
to suppress hydrothermal activity and to
cover hydrothermal deposits. Consequent-
ly, the actual extent of hydrothermal de-
posits along flow lines of sea-floor spread-
ing may be difficult to determine.
Relative to magnetic measurements, the
model (Fig. 11) is consistent with the ob-
served general parallelism between rema-
nent magnetic lineations and the rift valley
(Fig. 8). Because remanent magnetization
resides in the rocks of the fault blocks, the
inferred parallelism between the long axes
of these blocks and the rift valley ensures
that the gross pattern of linear residual
anomalies remains parallel to the axis of
the rift valley in spite of the asymmetric tec-
tonic fabric. The 3-km southeastward offset
of the center of the axial (Brunhes) mag-
netic anomaly from the rift valley that re-
mains after removal of the magnetic field ef-
fect may be alternatively interpreted as fol-
lows: (1) if it is possible for the locus of
crustal emplacement to shift from the rift
valley to one of its branches, then a north-
westward shift to the present rift valley
could account for the off-center position of
the axial anomaly, or (2) asymmetric half
rates of sea-floor spreading perpendicular
to the rift valley would produce a wider
magnetic anomaly on the faster spreading
side, resulting in a displacement of the
geometric center of the anomaly in the di-
rection of faster spreading while the locus
of spreading remained at the rift valley. The
second interpretation is supported by the
: ' ^
■ m
• 1 ; '
\ \ * *■ ■
: W: ■■>.
i 1 1 1
i m
»•
.-—.x'
> r ; i
:4>:
. .♦:•:•
Ul ;J-*
:: ::; ::::
i mir,;
j crj
i B ;•;•
\ lp.r?.;
< ml
^ ■:•:«•:•:♦
m ;i!i
- — .'
-■■■
mm i
l:;i m
miii
\ m
t- ij
<
:•:•*•:•:
¥ Si:?"
*
*
..*:§
•:■:•• ■> :■ :♦::».
*:»W» •: "t'J'::
I:*?:
» t X
:■:♦:•:•
» « t |
?:+:•:•
>:*¥:
'5 * I
**::::
m ■; ,„U„ :i
____<»
r
'* i *
n
'.».'.» If.'*'*.'*
{ ♦ * * ♦
:*. ¥,:♦. .*:•:♦.
:*::*::±::*:;:*
■ t -y- •*• -T- ■ t
urn 1 1
I:::*::*:*::::
J=>:*:::*:::*W
f.-.*-'*:*:.-::
» * »
I Hi
i t « ■ ■■* J » * *
t t < i ■{ » ± *
( ■.•:{ 1 I* f I *
lit I tt
t t
t 1
' > t ' ; !
ii i t i l 1
rf
» t
i in
it
..! i
0
50
■ i i
100
-I— I
KM
Figure 1 1 . Diagrammatic model for development of tectonic fabric and hydrothermal activity observed on Mid-Atlantic Ridge crest at lat 26°N (Figs. 3,
4). Axis and two sides of rift valley (heavy solid vertical lines) are parallel. Arrows indicate apparent directions of sea-floor spreading at apparent half rates
r„ r2, and r , about rift valley. Transverse ridges (shaded) and intervening transverse valleys (open) are shown. Transverse ridges are transected by valleys
perpendicular to their axes (open). Most of valleys only partly transect transverse ridges, and valleys to left of rift valley exhibit composite trends (Fig. 3).
Transverse ridges are constructed of fault blocks with long axes (light vertical dashed lines) parallel to rift valley. Lengths of crust generated per unit time
are bracketed between sides of the rift valley (heavy solid vertical lines) and solid vertical lines on transverse ridges to either side of the rift valley (b, c).
Sites of active and relict hydrothermal mineralization (black) are shown along one of transverse ridges (c). Successive configurations of model (a, b, c)
are explained in text. Distinct tectonic fabric of model is obscured in ocean basin by off-axis volcanism.
439
TECTONIC FABRIC AND HYDROTHERMAL ACTIVITY OF MID-ATLANTIC RIDGE CREST
671
asymmetric spreading rates in the study
area (Table 1) and poses fewer mechanical
problems than the first interpretation.
Symmetric and Asymmetric Processes
of Oceanic Ridges
The asymmetric tectonic fabric of the
study area occurs within the overall sym-
metric framework of the central North At-
lantic, as indicated by the nearly median
position of the Mid-Atlantic Ridge; the
nearly mirror-image distribution of
physiographic provinces about the ridge
axis (Heezen and others, 1959); the trajec-
tories of major fracture zones such as the
Atlantis and Kane (Fig. 1), which follow
small circles symmetric about the axis of
the Mid-Atlantic Ridge (Morgan, 1968);
and the sequences of remanent magnetic
anomalies that indicate a grossly similar
history of sea-floor spreading in the eastern
and western Atlantic (Pitman and Talwani,
1972) since the time of the magnetic quiet
zone boundary (Rona and others, 1970).
Two alternative hypotheses are considered
to reconcile the development of the ob-
served asymmetric tectonic fabric within a
symmetric framework:
1. Original orientation. The asymmetric
orientation of minor fracture zones about
the axis of an oceanic ridge, such as the
normal and oblique orientation of the
minor fracture zones of the study area, is
produced by asymmetric processes of de-
velopment of oceanic lithosphere. This
hypothesis poses problems in reconciling
asymmetric with symmetric fractures of the
ocean basin, because asymmetric plate mo-
tions at minor fracture zones would be in-
compatible with symmetric plate motions
at major fracture zones.
2. Reorientation. The processes of de-
velopment of oceanic lithosphere are essen-
tially symmetric and produce both symmet-
ric minor and major fracture zones as-
sociated with symmetric sea-floor spread-
ing. The minor fracture zones are continu-
ously reoriented, whereas the major frac-
ture zones maintain their original orienta-
tions. As a consequence of reorientation,
apparent- half rates of spreading deter-
mined perpendicular to the axis of an ocean-
ic ridge are asymmetric. True- half rates of
spreading determined in the directions of
the reoriented minor fracture zones normal
and oblique to the axis of an oceanic ridge
are equal. This hypothesis is supported by
the relations between spreading directions
and rates determined in the study area and
may reconcile the discrepancy between
asymmetric and symmetric features of the
ocean basin.
The continuous reorientation of minor
fracture zones according to hypothesis 2
' rcl.Hivt*.
may be caused by the application of an ex-
ternal stress field deriving from different
sources, as follows:
1. Forces related to magmatic processes.
These forces are related to magmatic
movements associated with the axial region
of an oceanic ridge. A type of regional
magmatic movement proposed by Vogt
(1971) and applied by Johnson and Vogt
(1973) and Vogt and Johnson (1975) to ac-
count for V-shaped topography about the
axis of an oceanic ridge depends on the
principle of a geopotential gradient to drive
asthenospheric flow along the axis of the
ridge from topographic highs over inferred
mantle plumes, for example, the Azores
about 1,000 km north of the sudy area. Ac-
cording to their hypothesis, the V should
point in the direction of flow away from the
high as the vector resulting from astheno-
spheric flow along and sea-floor spreading
about an oceanic ridge. The Vogt-Johnson
hypothesis does not account for the
V-shaped configuration of the minor frac-
ture zones about the Mid-Atlantic Ridge in
the study area because the V points toward
rather than away from the Azores (Fig. 1).
Forces related to magmatic processes un-
doubtedly contribute to the stress field, but
they are considered secondary rather than
primary components.
2. Forces related to tectonic processes.
These forces are related to interplate and
intraplate motions and may be primary
components of the stress field that we infer
to be reorienting the direction of minor
fracture zones and sea-floor spreading
along the Mid-Atlantic Ridge. The role of
interplate and intraplate forces as primary
components of the stress field is supported
by the observation that the orientation of
minor fracture zones and of sea-floor
spreading differs about lithospheric plate
boundaries. The orientation of minor frac-
ture zones and of sea-floor spreading is dif-
ferent on the two sides of the rift valley of
the Mid-Atlantic Ridge and differs between
the American and Eurasian plates north of
the Azores triple junction and the American
and African plates south of that junction
(Table 1).
The reorientation hypothesis allows the
simultaneous development of small-scale
asymmetric structures and large-scale
symmetric structures in oceanic lithosphere.
Minor fracture zones associated with small
transform faults (offset ^ 30 km) like those
in the study area (Fig. 1) behave in an un-
stable manner at the relatively slow average
half rates of spreading (=£2 cm/yr) prevalent
at the Mid-Atlantic Ridge. The minor frac-
ture zones are continuously reoriented
under the influence of an external stress
field as they are generated by sea-floor
spreading about the small transform faults.
Major fracture zones like the Atlantis and
Kane associated with large transform faults
behave in a stable manner at relatively
slow average half rates of spreading. The
major fracture zones maintain their orienta-
tion under the influence of the same exter-
nal stress field as they are generated by sea-
floor spreading about the large transform
faults. Thickness of lithosphere related to
distribution of isotherms at a transform
fault may be a determinant of the stability
of fracture zones (Vogt and others, 1969).
Asymmetric small-scale structures may then
develop within the large-scale symmetry of
the Atlantic Ocean basin as a consequence
of the differential stability between minor
and major fracture zones (Rona, 1976).
CONCLUSIONS
The tectonic fabric of oceanic crust that
is slowly spreading about an oceanic ridge
develops according to a definite geometry,
which has been deduced from analysis of
the asymmetric tectonic fabric of the Mid-
Atlantic Ridge crest at lat 26°N within the
overall symmetric framework of the central
North Atlantic Ocean basin (Figs. 3, 4, 11),
as follows: (1) The double structre of the
rift valley consisting of linear segments be-
tween transform faults, alternating with ba-
sins at transform faults, acts as a template
that programs the reproduction of tectonic
fabric through control of the formation of
topographic highs and lows. (2) The trans-
verse ridges are constructed of fault blocks
that are uplifted from the floor and accrete
at the walls along the linear segments of the
rift valley. (3) The transverse valleys are
minor fracture zones aligned with the direc-
tion of sea-floor spreading about the topo-
graphic lows. (4) Branches of the rift valley
extend to either side oriented parallel to the
rift valley and perpendicular to the axis of
the transverse ridges; the branches split off
from the rift valley as a consequence of
sea-floor spreading and form valleys that
transect the. transverse ridges. (5) Minor
fracture zones expressed as transverse val-
leys between ridges may be asymmetric
about the axis of a rift valley, tending to
remain normal to one side and to reorient
oblique to the other side of the rift valley.
(6) Where tectonic fabric is asymmetric
about the rift va|ley, apparent half rates of
spreading measured perpendicular to the
rift valley are also asymmetric, with faster
half rates on the normal side and slower
half rates on the oblique side. (7) The half
rates of spreading measured in the direc-
tions of the minor fracture zones normal
and oblique to the rift valley tend toward
equality over averaging intervals of millions
of years. (8) The observed tectonic fabric
may be explained by preferential asymmet-
ric reorientation of minor fracture zones
relative to symmetric major fracture zones,
resulting from differential structural stabil-
440
6 2
RONA AND OTHERS
irv under the influence of an external stress
Held.
Structural .uul thermal conditions at di-
vergent plate boundaries .ire conducive to
hydnithenn.il activity. The concentration
of stllv- se.l-tloor h\ drotherm.il systems is
favored by special conditions in the tectonic
fabric ol an oceanic ridge crest, such as
close spacing oi valleys and proximity to
intrusive heat sources that promote vigor-
ous circulation (Fig. 10). The distribution
of hydrotherm.il convection systems along
divergent plate boundaries, like the inferred
system at the TAG Hydrothermal Field, can
only be conjectured from the known dis-
tribution of 17 active hydrothermal systems
over a distance of 250 km in the neovol-
canic /one of Iceland on the Mid-Atlantic
Ridge (Bodvansson, 1961) and at least 14
systems over a distance of 900 km in the
Red Sea (Degens and Ross, 1969; Backer
and Schoell, 19~2). Because all of the
oceanic crust that covers two-thirds of the
Farth has been generated about divergent
plate boundaries, the relation between tec-
tonic fabric and hydrothermal activity on
the Mid-Atlantic Ridge crest at lat 26°N is
relevant to the metallic mineral potential of
ocean basins and regions where oceanic
crust has been incorporated into islands
and continents.
ACKNOWLEDGMENTS
We lament the early death of our col-
league and friend, Andrew J. Nalwalk, who
generously contributed his prowess at sea
to this work.
We thank Louis W. Butler of the Na-
tional Oceanic and Atmospheric Adminis-
tration (NOAA) for his help in all phases of
the work and Bonnie A. McGregor of
NOAA for recontouring the bathymetnc
map. We are grateful to Bruce C. Heezen
for encouraging us to determine the charac-
teristics of normal oceanic crust.
We thank Gleb B. Udintsev and others
onboard the R'V Akademik Kurchatov for
obtaining the AK series of dredge samples
(Table 2, see footnote 1) on a cooperative
TAG cruise in 1975 as part of the
U.S. -USSR Agreement on Cooperation in
Studies of the World Ocean.
We acknowledge the excellent coopera-
tion of Captain Floyd J. Tucker, Jr., Cap-
tain Lavon L. Posey, Cdr. Richard H.
Allbntton, Cdr. Walter S. Simmons, Lt.
Paul M. Duernberger, and the other officers
and crews of NOAA Ship Discoverer and
NOAA Ship Researcher during the 1972
and 1973 TAG cruises.
REFERENCES CITED
Ade-Hall, J. M„ Palmer, H. C, and Hubbard, T.
P., 1971, The magnetic and opaque pet-
rological response of basalts to regional
hydrothermal alteration: Royal Astron.
Soe. Geophys. Jour., v. 24, p. 137-174.
Anderson, R. N., 1972, Petrologic significance of
low heat flow on the flanks of slow-
spreading mid-ocean ridges: Geol. Soe.
America Bull., v. 83, p. 2947-2956.
Atwater, T., and Mudie, J. D., 1973, Detailed
near-bottom geophysical study of the
Corda Rise: Jour. Geophys. Research, v.
78, p. 8665-8686.
Aumento, F., Loncarevic, B. D., and Ross, D. I.,
1971, IV. Regional studies; Hudson geo-
traverse: Geology of the Mid-Atlantic
Ridge at 45°N: Royal Soe. London Philos.
Trans., ser. A, v. 268, p. 623-650.
Backer, H, and Schoell, M., 1972, New deeps
with brines and metalliferous sediments in
the Red Sea: Nature Phys. Sci., v. 240,
p. 153-158.
Bellaiche, G., Cheminee, J. L., Francheteau, J.,
Hekinian, R., Le Pichon, X., Needham, H.
D., and Ballard, R. D., 1974, Inner floor of
the rift valley: First submersible study: Na-
ture, v. 250, p. 558-560.
Betzer, P. R., Bolger, G. W., McGregor, B. A.,
and Rona, P. A., 1974, The Mid-Atlantic
Ridge and its effect on the composition of
particulate matter in the deep ocean: EOS
(Am. Geophvs. Union Trans.), v. 55,
p. 293.
Bhattacharyya, P. J., and Ross, D. 1., 1972,
Mid-Atlantic Ridge near 45°N, computer
interpolation and contouring of bathymetry
and magnetics: Canada Dept. Environment,
Marine Sci. Directorate, Marine Sci. Paper
11, 9 p.
Bischoff, J. L., and Dickson, F. W., 1975,
Seawater-basalt interaction at 200°C and
500 bars: Implications for origin of
sea-floor heavy mineral deposits and regula-
tion of seawater chemistry: Earth and
Planetary Sci. Letters, v. 25, p. 385-397.
Bodvarsson, G., 1961, Physical characteristics of
natural heat resources in Iceland: Jokull, v.
11, p. 29-38.
Bonatti, E., 1975, Metallogenesis at oceanic
spreading centers, in Annual Reviews of
Earth and Planetary Sciences, Vol. 3: Palo
Alto, Calif., Annua] Reviews, Inc.
p. 401-431.
Bostrom, K., and Peterson, M.N. A., 1966, Pre-
cipitates from hydrothermal exhalations on
the East Pacific Rise: Econ. Geology, v. 61,
p. 1258-1265.
Bostrom, K., Farquarson, B., and Eyl, W., 1972,
Submarine hot springs as a source of active
ridge sediments: Chem. Geology, v. 10,
p. 189-203.
Cann, J. R., 1970, New model for the structure
of the ocean crust: Nature, v. 226,
p. 928-930.
Christensen, N. I., 1972, The abundance of ser-
pentimtes in the oceanic crust: Jour. Geol-
ogy, v. 80, p. 709-719.
Collette, B. J., 1974, Thermal contraction |oints
in a spreading seafloor as origin of fracture
zones: Nature, v. 251, p. 299-300.
Constantinou, G., and Govett, G.J.S., 1973,
Geology, geochemistry, and genesis of
Cyprus sulfide deposits: Econ. Geologv, v.
68, p. 843-858.
Corliss, J. B., 1971, The origin of metal-bearing
submarine hydrothermal solutions: Jour.
Geophys. Research, v. 76, p. 8128-8138.
Deffeyes, K. S., 1970, The axial valley: A
steady-state feature of the terrain, in
|ohnson, FL, and Smith, B. L., ed>, The
megatectonics of continents and oceans:
Camden, N.J., Rutgers Univ. Press,
p. 194-222.
Degens, E. T., and Ross, D. A., eds., 1969, Hot
brines and recent heavy metal deposits in
the Red Sea: New York, Springer, 600 p.
Detnck, R. S„ Mudie, J. D., Luyendyk, B. P., and
Macdonald, K. C, 1973, Near-bottom ob-
servations of an active transform fault
(Mid-Atlantic Ridge at 37°N): Nature Phvs.
Sci., v. 246, p. 59-61.
Dmitnev, L. V., Barsukov, V. L., and Udintsev,
G. B., 1970, Rift zone of the oceans and the
problem of ore formation: Geokhimiva, v.
4, p. 937.
Donaldson, I. D., 1962, Temperature gradients
in the upper layers of the Earth's crust due
to convective water flows: Jour. Geophvs.
Research, v. 67, p. 3449-3459.
Dymond, J., Corliss, J. B., Heath, J. R., Field, C.
W., Dash, E. J., and Veeh, H. H, 1973,
Origin of metalliferous sediments from the
Pacific Ocean: Geol. Soe. America Bull., v.
84, p. 3355-3372.
Elder, J. W., 1965, Physical processes in geo-
thermal areas, in Terrestrial heat flow: Am.
Geophvs. Union Geophvs. Mon. 8,
p. 211-239.
1967, Steady free convection in a porous
medium heated from below: Jour. Fluid
Mechanics, v. 27, p. 29-48.
Greenewalt, D., and Taylor, P. T., 1974, Deep-
tow magnetic measurements across the
axial valley of the Mid-Atlantic Ridge:
Jour. Geophvs. Research, v. 79,
p. 4401-4405.
H.i|.ish, A., 1974, An experimental investigation
of high temperature seawater-basalt in-
teractions: Geol. Soe. America Abs. with
Programs, v. 6, p. 771.
Harrison, C.G.A., 1968, Formation of magnetic
anomaly pattern by dyke injection: Jour.
Geophys. Research,' v. 73, p. 2137-2142.
Han, R. A., 1973, A model for chemical ex-
change in the basalt-seawater system of
oceanic layer II: Canadian Jour. Earth Sci.,
v. 10, p. 799-816.
Heezen, B. C, Tharp, M., and Ewing, M., 1959,
The floors of the oceans. I. The North At-
lantic: Geol. Soe. America Spec. Paper 65,
122 p.
Heirtzler, J. R., Dickson, G. O., Herron, E. M.,
Pitman, W. C. Ill, and Le Pichon. X., 1968,
Marine magnetic anomalies, geomagnetic
field reversals, and motions of the sea floor
and continents: Jour. Geophvs. Research, v.
73, p. 2119-2136.
Helgeson, H. C, 1964, Complexing and hy-
drothermal ore deposition: New York, Per-
gamon Press.
Hoffman, )., 1957, Hyperbolic curves applied to
echo sounding: Internat. H\drog. Rev., v.
?4, p. 45-55.
Holland, H. D., 19f>-, Gangue minerals in hy-
drothermal systems, in Barnes, H. 1... ed.,
Geochemistry of hydrothermal ore de-
posits: New York, Holt, Rinehart &: Win-
ston, p. 382-436.
441
IFCTONK FABRIC AND HYDROTHFRMAI ACTIVITY OF MID- ATLANTIC RIDGE CREST
673
Hollister, C . D.. huiit, |, I., Habih, IV, H.nh.i-
wav, | (. .. Lancelot, V., l.uterhacker. H.,
Paulus, 1 |.. Poag, C . W.. Wilcoxon. |. C ..
and \X oi mcII. P.. I9~2, Initial reports of the
Deep Sea Drilling Project, Vol. II: Washing-
ton, D.C., U.S. Clovt. Printing Office,
10- p.
Hutchinson, R. W.. |9 5, \'olcanogenic sulfide
deposits and then metallogenic significance:
Econ. Geology, \. 68, p. 1223-1246.
IAGA Commission 2. Working Croup 4. |9<,9,
International geomagnetic reference field
1965.0: |our. Geophys. Research, \. ~4,
p. 440~-440X.
Johnson. Ci. 1.. and Vogt, P. R. 19~3, Mid-
Atlantic Ridge from 4~ to si North: C.eol.
Soc. America Bull., v. S4. p. 5443-3462.
Kjartansson, C. I960, 1^2. I^hs, 19hK, 1969,
Geological map of Iceland, sheets 1,2, \ 5,
6: Reykjavik, Mus. Nat. History.
Krauskopf, K. B., 1956, Factors controlling the
concentrations of thirteen rare metals in
seawater: Geochim. et C osmochim. Acta, \.
9, p. 1-12B.
Lachenbruch, A. H., 19~3, A simple mechanical
model for oceanic spreading centers: Jour.
Geophys. Research, \. -S. p. 3395-3417.
Langseth, M. G-, |r.. Malone, I. E., and Book-
man, C. A., I9~2, Sea floor geothermal
measurements tram Win.i Cruise 25:
Palisades, N.Y., Lamont-Dohertv Geol.
Observatory Tech. Rept. 4, CL'-4-_2; Teeh.
Rept. 2, CU-2-72, 159 P.
Latrimore, R. K.. Rona, P. A., DeWald, O. E.,
1974, Magnetic anomaly sequence in the
central North Atlantic: |our. Geophys. Re-
search, \. _9, p. 120"- 1209.
Lister, C.R.B., 19~2, On the thermal balance of a
mid-ocean ridge: Royal Astron. Soc.
Geophys. Jour., v. 26, p. 515-535.
1974a, Water percolation in the oceanic
crust: EOS (Am. Geophvs. Union Trans.),
v. 55, p. -40-"42.
19-4h, On the penetration of water into hot
rock: Roval Astron. Soc. Geophvs. |our.,
v. 39, p. 465-509.
Loncarevic, B. D., and Parker, R. L., 1971, The
Mid-Atlantic Ridge near 45°N. XVII.
Magnetic anomalies and sea floor spread-
ing: Canadian Jour. Earth Sci., v. X,
p. 883-898.
Lowell, J. D., and Cenik, J. D„ 1972, Sea-tloor
spreading and structural evolution of
southern Red Sea: Am. Assoc. Petroleum
Geologists Bull., v. 56, p. 247-259.
Lowell, R. P., 1975, Circulation in fractures, hot
springs, and convective heat transport on
mid-ocean ridge crests: Royal Astron. Soc.
Geophys. |our., v. 40, p. 351-365.
Luyendyk, B. P., and Melson, W. G., 196",
Magnetic properties and petrology of rocks
near the crest of the Mid-Atlantic Ridge:
Nature, v. 215, p. 147-149.
Macdonald, K., Luyendyk, B. P., Mudie, ). D.,
and Spiess, F. N., 1975a, Near-bottom
geophysical study of the Mid-Atlantic
Ridge median valley near lat 37°N: Pre-
liminary observations: Geology, v. ?,
p. 211-215.
1975h, Near-bottom magnetic measure-
ments, asymmetric spreading, oblique
spreading, and the zone of intrusion in the
FAMOUS area of the Mid-Atlantic Ridge
I i~°N): EOS (Am. Geophys. Union Trans.),
v. 56, p. s-V
Matthews, D. H., and Bath, J., 196"7, Formation
ot magnetic anomaly pattern of Mid-
Atlantic Ridge: Roval Astron. Soc.
Geophys. |our., v. 13, p. 349-357.
Matthews, D. |., 19 19, Tables of the velocity of
sound in pure water and sea water for use in
echo-sounding and sound ranging (2nd
ed.): London. Admiralty, Hvdrographic
Dept. Pub. 2S2, S2 p.
McGregor, B. A., and Rona, P. A., 1975, Crest of
Mid-Atlantic Ridge at 26°N: Jour.
Geophvs. Research, v. 80, p. 3307-3314.
McGregor, B. A., Harrison, C.G.A.. Lavelle, J.
W., and Rona, P. A., 1976, Magnetic
anomaly pattern on Mid-Atlantic Ridge
crest at 26°N: Jour. Geophys. Research, v.
SI (in press).
McKen/ie, 1). P., Davis, D., and Molnar, P.,
1970, Plate tectonics of the Red Sea and
Fast Africa: Nature, v. 226, p. 243-248.
Menard, H. W., 1967, Sea floor spreading, to-
pography, and the second laver: Science, v.
157, p. 92 i- 924.
Meyer, C, and Hemley, J. J., 1967, Wall rock
alteration, in Barnes, H. L., ed., Geochemis-
try of hydrothermal ore deposits: New
York, Holt, Rmehart & Winston, p.
166-235.
Miyashiro, A., Shido, F., and Fwing, M-, 1971,
Metamorphism in the Mid-Atlantic Ridge
near 24°N and 3()°N: Royal Soc. London
Philos. Trans., ser. A, v. 268, p. 589-603.
Moore, J. G., Fleming, H. S., and Phillips, J. D.,
1974, Preliminary model for extrusion and
rifting at the axis of the Mid-Atlantic Ridge,
36°48'N: Geology, v. 2, p. 437-440.
Morgan, W. J., 1968, Rises, trenches, great faults
and crustal blocks: Jour. Geophvs. Re-
search, v. 73, p. 1959-1982.
Mottl, M. J., Corr, R. F., and Holland, H. D.,
1974, Chemical exchange between sea
water and mid-ocean ridge basalt during
hydrothermal alteration: An experimental
study: Geol. Soc. America Abs. with Pro-
grams, v. 6, p. 879-880.
Muehlenbachs, K„ and Clayton, R. N., 1972,
Oxygen isotope geochemistry of submarine
greenstones: Canadian Jour. Earth Sci.,
y. 9, p. 471-478.
Needham, H. D., and Francheteau, J., 1974,
Some characteristics of the rift valley in the
Atlantic Ocean near 36°48' North: Earth
and Planetary Sci. Letters, v. 22, p. 29-43.
Osmaston, M. F., 1971, Genesis of ocean ridge
median valleys and continental rift valleys:
Tectonophysics, v. 11, p. 387-405.
Palmason, C, 1967, On heat flow in Iceland in
relation to the Mid-Atlantic Ridge, in
Bjornsson, S., ed., Iceland and mid-oceanic
ridges: Soc. Sci. Islandica, v. 38,
p. 111-127.
Parker, R. L., and Oldenburg, D. W., 1973,
Thermal model of ocean ridges: Nature
Phvs. Sci., v. 242, p. 137-139.
Perry, R. K., and Feden, R. H., 1974, Bathymetry
of the Mid-Atlantic Ridge from 10°N to
40°N latitudes: Geol. Soc. America Abs.
with Programs, v. 6, p. 909.
Piper, D. Z., 1973, Origin of metalliferous sedi-
ments from the East Pacific Rise: Earth and
Planetary Sci. Letters, v. 19, p. 75-82.
Pitman, W. C. Ill, and Talwani, M., 1972,
Sea-floor spreading in the North Atlantic:
Geol. Soc. America Bull., v. 83,
p. 619-649.
Reid, 1., and Macdonald'K., 1973, Microearth-
quake study of the Mid-Atlantic Ridge near
37°N, using sonobuoys: Nature, v. 246, p.
88-90.
Robertson, A.H.F., and Hudson, J. D., 1973,
Chemical precipitates on a Tethyan ocean
ridge: Earth and Planetary Sci. Letters,
v. 18, p. 93-106.
Rona, P. A., 1973a, Marine geology, in
McGraw-Hill yearbook of science and
technology: New York, McGraw-Hill Book
Co., p. 252-256.
1973b, Plate tectonics and mineral re-
sources: Sci. American, v. 229, no. 1,
p. 86-95.
1976, Asymmetric fracture zones and
sea-floor spreading: Earth and Planetary
Sci. Letters (in press).
Rona, P. A., Brakl, J., and Heirtzler, J. R., 1970,
Magnetic anomalies in the northeast Atlan-
tic between the Canary and Cape Verde Is-
lands: Jour. Geophys. Research, v. 75,
p. 7412-7420.
Rona, P. A., Harbison, R. N., Bassinger, B. G.,
Butler, L. W., and Scott, R. B., 1973,
Asymmetrical bathymetry of the Mid-
Atlantic Ridge at 26°N latitude: EOS (Am.
Geophys. Union Trans.), v. 54, p. 243.
Rona, P. A., McGregor, B. A., Betzer, P. R.,
Bolger, G. W., and Krause, D. C, 1975,
Anomalous water temperatures over Mid-
Atlantic Ridge crest at 26° North latitude:
Deep-Sea Research, v. 22, p. 611-618.
Sayles, F. L., and Bischoff, J. L., 1973, Fer-
romanganoan sediments in the equatorial
East Pacific: Earth and Planetary Sci. Let-
ters, v. 19, p. 330-336.
Sclater, J. G., Von Herzen, R. P., Williams, D. L,
Anderson, R. N., and Klitgord, K., 1974,
The Galapagos spreading center: Heat flow
low on the north flank: Royal Astron. Soc.
Geophys. Jour., v. 38, p. 609-626.
Scott, M. R., Scott, R. B., Rona, P. A., Butler,
L. W., and Nalwalk, A. J., 1974, Rapidly
accumulating manganese deposit from the
median valley of the Mid-Atlantic Ridge:
Geophys. Research Letters, v. 1,
p. 355-358.
Scott, R. B., Rona, P. A., McGregor, B. A., and
Scott, M. R„ 1974, The TAG Hydrother-
mal Field: Nature, v. 251, p. 301-302.
Scott, R. B., Malpas, J., Udintsev, C, and Rona,
P. A., 1975, Submarine hydrothermal activ-
ity and sea-floor spreading at 26°N, MAR:
Geol. Soc. America Abs. with Programs,
v. 7, p. 1263.
Silhtoe, R. H., 1973, Environments of formation
of volcanogenic massive sulfide deposits:
Econ. Geology, v. 68, p. 1321-1336.
Sleep, N. H., 1969, Sensitivity of heat flow and
gravity to the mechanism of sea-floor
spreading: Jour. Geophys. Research, v. 74,
p. 542-549.
Sleep, N. H., and Biehler, S., 1970, Topography
and tectonics at the intersections of fracture
zones with central rifts: Jour. Geophys. Re-
search, v. 75, p. 2748-2752.
Spindel, R. C, Davis, S. B., Macdonald, K. C,
Porter, R. P., and Phillips, J. D., 1974, Mi-
442
h~4 RON A WD OTHKRS
croearthquake survev of median vallev of ocean rises: |our. Ceophvs. Research, v. "3, Drilling Project, Vol. 8: Washington, D.C.,
the Mid-Atlantic Ridge .it 56°30'N: Na- p. 12~V-129X. U.S. Govt. Printing Office, p. 829-833.
turc, v. 248, p. s~"-5~9. van Andel, T. H., and Moore, I. C. I9~(), Watkins, N. [)., and Paster, 1. P., 1971, The
Spooncr, F..T.C .. and F\ fe, XV. S., 1973, Sub-sea Magnetic anomalies and sea floor spreading magnetic properties of igneous rocks from
floor metamorphisrn, heat and mass trans- rates in the northern South Atlantic: Na- the ocean floor: Royal Sue. London Philos.
ter: Contr. Mineralog\ and Petrolog\, ture, v, 226, p. 528-330. Trans., ser. A, v. 268, p. 507-550.
v. 42. p. 28~-304. Vogt, P. R., 19~1, Asthenosphere motion re- Williams, D. L, Von Herzen, R. P., Sdater, J. G.,
Spooner. E.T.C .. Beckinsale, R. D., Fyfe, W. S., corded hy the ocean floor south of Iceland: and Anderson, R. N.. 1974, The Galapagos
and Sniewing, |. D., 19~4, O'" enriched Earth and Planctar\ Sci. letters. \. |3, spreading center: Lithospheric cooling and
ophiolitic metahasic rocks from E. Liguria p. 155 — 160. hydrothcrmal circulation: Renal Astron.
ItaK . Pindos Greece), and Troodos Vogt, P. R.. and Johnson. G. 1 . I9~5, 1 tans- Soc. Geophys. Jour., v. IN, p. 587-608.
i. \prus : C onrr. Mineralog) and Petrology, form faults and longitudinal flow beneath Wilson. |. 1., 1V6.5, A new class of faults and
v. 4". p. 41— "4, the midoceanic ridge: |our. Geophys. Re- their bearing on continental drift: Nature,
Talwani. M.. VX'indish, C. ( .. and Langseth, search, v. 80, p. 1399-1428. v. 20" p. 543-347.
M.C.. lr. |9~1, Reykjanes Ridge crest: A Vogt. P. R. Avery, O. F... Schneider. E. I).. An- Wooding, R. A.. 1 963, Convection in a saturated
detailed geophysical study : Jour. Geophvs. derson, C. N.. and Braces. 1). R.. 1969, porous medium at large Ra\ leigh numbers
Research, v. ~6, p. 473-51". Discontinuities in sea-floor spreading: Tec- and Peclet number: Jour. Fluid Mechanics.
Turcotte, D. 1., 19~4, Are transform faults tonophysics, v. 8, p. 285-3 1". \. 15, p. 527-544,
thermal contraction cracks?: Jour. von der Borch. C . C .. and Rex, R. NX.. |9~1, Zelenov, K. K., 1964. Iron and manganese in
Geophys. Research. \. -9, p. 25~5-25--. Amorphous iron oxide precipitates in sedi- exhalations from the submarine volcano
L'padhyay, H. D., and Strong, D. F., 1973, ments cored during Leg 5, Deep Sea Drilling Bano XX'uhu (Indonesia): Akad. Nauk SSSR
Geological setting of the Betts Cove copper Pro|ect. in Initial reports of the Deep Sea Dokladv. v. 155, p. 1515-1520.
deposits. Newfoundland: An example of Drilling Project, Vol. v Washington. D.C.,
ophiolite sulfide mineralization: Econ. U.S. Govt. Printing Office, p. s41-s44
Geology, v. 68, p. 161-167. von der Borch. C. C .. Nesteroft, \X . D., and Manuscript Received by the SOCIETY August
van Andel, T, H.. and Bowm. C. O., 1968, Galehouse, |. S„ 1 9" 1. Iron-rich sediments 21, 1V~4
Mid-Atlantic Ridge between 22c and 25c cored during Leg 8 of the Deep Sea Drilling Revised Manuscript Received July 24, 1975
North latitude and the tectonics of mid- Project, in Initial reports of the Deep Sea Manuscript Accepted August 20, 1975
443
46
Reprinted from: Geology, Vol. 4, No. 4, 233-236.
Duration of hydrothermal activity
at an oceanic spreading center,
Mid-Atlantic Ridge (lat 26°N)
Robert B. Scott
Department of Geology, Texas A&M University, College Station, Texas 77843
John Malpas
Department of Geology, Memorial University, St. Johns, Newfoundland,
Canada A1C5S7
Peter A. Rona
National Oceanic and Atmospheric Administration-Atlantic Oceanographic and
Meteorological Laboratories, 15 Rickenbacker Causeway, Miami, Florida 33149
Gleb Udintsev
Institute of Oceanology, USSR Academy of Sciences, Moscow, USSR
ABSTRACT
Hydrothermal manganese oxide coats
talus on the Mid- Atlantic Ridge at lat 26° N
until spreading moves the rock past the
thermally and structurally active rift-
valley wall. Hydrothermal activity is re-
placed by hydrogenous ferromanganese
oxide precipitation on ocean crust older
than 0.7 m.y. on the ridge-crest highlands.
INTRODUCTION
Abundant hydrothermal manganese
oxide crusts were found coating basalt
talus on the rift-valley wall at the Mid-
Atlantic Ridge at lat 26°N during the
Trans- Atlantic Geotraverse project of the
National Oceanic and Atmospheric Ad-
ministration in 1972 and 1973 (Scott, R. B.,
and others, 1974; Scott, M. R., and others,
1974). These crusts are almost pure man-
ganese oxide with only a trace of Fe,
GEOLOGY
Co, Cu, and Ni and compositionally fall
into the Mn-rich end member of Bonatti's
hydrothermal classification (Bonatti, 1975).
Very rapid growth of the manganese
deposits is required because they are as
great as 50 mm thick only 5 km from the
rift axis. U-Th dating of the outermost
layers of these manganese crusts shows
them to be accumulating at 200 mm/m.y.,
nearly two orders of magnitude faster than
typical hydrogenous ferromanganese crusts
or nodules (Scott, M. R., and others, 1974).
Bottom photographs of the hydrothermally
active portion of the rift wall (McGregor
and Rona, 1975) show the presence of
similar coatings over talus. Betzer and
others (1974) found abnormally high con-
centrations of weak-acid-soluble Fe- and
Mn-bearing particulate matter suspended
over the Mid-Atlantic Ridge at lat 26° N.
A region of low magnetic intensity within
the Brunhes normal in the hydrothermal
area may be related to the destruction of
magnetic domains during hydrothermal
alteration of basalt (McGregor and Rona,
1975). From these data, R. B. Scott and
others (1974) concluded that cold, dense
sea water flows down fracture systems
(Lister, 1974), reacts with hot rocks or
magma under the ridge crest, produces
less dense hydrothermal fluids enriched in
Ca, Si, Fe, Mn, and H2S and depleted in
Mg (Hajash, 1975; Mottl and others, 1974;
Bischoff and Dickson, 1975), and then is
emitted as submarine springs, where oxy-
genated sea water causes precipitation of
manganese oxides on talus overlying frac-
tured fault scarps.
However, the distribution of dredge sites
did not define the limits of this hydro-
thermal activity in time or space. Defini-
tion of the limits of activity at lat 26°N
became one of the objectives of participa-
tion of the NOAA Trans-Atlantic Geo-
traverse project in the U.S.-U.S.S.R. Agree-
ment for Cooperative Studies of the World
Ocean. The hydrothermal region at lat
26° N was dredged during the spring of
1975 aboard the R/V Akademik Kurchatov
by Soviet and American scientists.
233
444
DREDGED ROCKS
Manganese oxide crusts and associated
veins in basalt talus were recovered from
site 75-1 A, 18 km from the rift axis (Fig. 1).
Unlike the hydrothermal crusts found at
sites 72-13, 73-2A, and 73-3A, the crusts
at site 75-1 A have two distinct layers. The
basal layers on the altered basalt have the
same physical appearance as the hydro-
thermal crusts dredged in 1972 and 1973.
The basal layers have a smooth, slightly
botryoidal surface over an undulating
laminated interior, a uniform brownish-
black to submetallic gray color, and a
thickness as great as 10 mm. The slightly
undulating botryoids are about 1 mm in
diameter. In contrast, the upper layers are
most similar to hydrogenous crusts; these
upper layers have highly irregular micro-
botryoidal surfaces. Small columnlike bot-
ryoids are less than 0.1 mm in diameter
and vary from a shiny grayish-black to an
earthy, dark yellowish orange color.
Organism tests are trapped between
botryoid columns. The upper crusts are
less than 2 mm thick. The two crust types
were carefully separated for chemical,
mineralogical, and scanning electron
microscope (SEM) studies.
The basalts have been altered from
nearly fresh basalt to extremely friable
grayish yellow green material. Numerous
veins of manganese oxides similar to the
lower crusts fill fractures that cut the talus
of basaltic breccia. One vein of free-
growing zeolite crystals (75-1A28) as much
as 5 mm in diameter was found coating
altered basaltic glass; a thin (<0.1 mm
thick) irregularly botryoidal manganiferous
crust coated these zeolites.
RESULTS OF CHEMICAL, MINERAL-
OGICAL, AND SEM INVESTIGATIONS
The lower crusts (sample 75-1A24) only
have x-ray diffraction patterns of birnessite
(7.1, 3.5, and 2.46 A d-spacings). The
upper crust (75-1A24) is apparently amor-
phous to x-rays. The zeolite (75-1A28) is
obviously an analcite from powder camera
patterns, and chemically it is a potassic
Na analcite (10 percent Na20, 1 percent
K2Ot and trace of CaO). Atomic absorp-
tion spectrophotometry shows the lower
crusts to contain 40 percent Mn, less than
0.1 percent Fe, and 0.15 percent
Cu+Co+Ni; in contrast, the upper crusts
contain more than 16 percent Fe and 0.4
percent Cu+Ni+Co (Fig. 2). The low
Cu+Ni+Co contents and low Fe contents
of the basal crust are similar to hydro-
thermal crusts found in 1972 and 1973.
Figure 1. Hydrothermal field at lat 26°N. The 4-km-deep rift valley is on left; year and dredge
number are used to identify dredge sites; contour intervals are 0.1 km. Hydrothermal sites 72-13,
73- 2A, 73-3A, and 75-1 A lie along an irregular ridge that trends southeast perpendicular to rift val-
ley. This ridge seems to be cut into blocks by northeast-trending depressions parallel to rift that
may represent normal faults. Bathymetry from McGregor and Rona (1975).
(Cu + Co+Ni)xlO
k40
15-2
.2B
19-3
• »A
•I0G
»4?£e/?
--'A?«
Fe
50
25
Mn
Figure 2. Composition of hydrothermal and hydrogenous Mn deposits plotted on Mn portion of
the Mn-Fe-(Cu+Co+Ni)10 ternary diagram. Average composition of three analyses of upper and
lower crust of sample 75-1A24 are attached by dashed line. At Mn apex, 1972 and 1973 analyses
of hydrothermal crust are shown. Hydrogenous materials are represented by points on left: P = aver-
age Pacific and A - average Atlantic nodule; 10G and 2B are Mn crusts from the Atlantis Fracture
Zone; 19-3 and 15-2 are Mn crusts from pillow lavas close to hydrothermal field at lat 26 N (Scott,
M. R., and others, 1974). By atomic absorption spectrophotometry; precision expressed as percent
of value determined: Fe±2 percent, Mn±l percent, Cu±l percent, Ni±5 percent, Co±5 percent.
234
APRIL 1976
445
The coating on the zeolites (Fig. 3d) with
a 0.3 percent Co+Cu+Ni content is tran-
sitional between hydrothermal and hydro-
genous crust compositions; the high Fe
content in this coating (18 percent) may be
indicative of a hydrogenous origin. Even
though no Fe-rich hydrothermal man-
ganese deposits have been identified at
lat 26°N, Bonatti (1975) showed their
existence elsewhere. SEM photographs
show the lower crust to have the texture
of well-crystallized birnessite (Fig. 3a);
this boxwork of plates is very similar to
the texture found in birnessite from widely
differing environments (Swanson, 1975;
Brown and others, 1971; Fewkes, 1973).
In contrast, the upper crust at the same
scale (Fig. 3b) shows a smooth featureless
surface with no textural indication of
crystallinity. The smaller scale view
in Figure 3c shows the overall micro-
botryoidal form of the lower crust that is
similar to columnar zones described by
Sorem and Foster (1972) in hydrogenous
ferromanganese.
DISCUSSION
Clearly, from the chemical and physical
description and SEM observations, the
manganese oxide lower crust 75-1A24
appears to be most similar to other hydro-
thermal manganese crusts; the ferro-
manganese oxide upper crust 75-1 A24 has
strong affinities to hydrogenous ferro-
manganese crusts and nodules. A compari-
son of the compositions shown in Figure 2
with those of Bonatti (1975, Fig. 4) support
these conclusions. The dramatic differences
in iron and Cu-t-Co+Ni contents between
the upper and lower crust samples taken
only a few millimetres from one another
suggest drastically different mechanisms
of formation or different sources of fluids.
Several authors (Bonatti and others,
1972; M. R. Scott and others, 1974)
have noted an inverse relation between
the rate of manganese crust accumulation
and the content of trace metals. If data
given by M. R. Scott and others (1974,
Table 2, Figs. 2, 3) are representative
of this relationship, then the approximate
growth rate in millimetres/106yr
= e14.8S- 1.54 In Cu+Co+Ni ppm
The Cu+Co+Ni for the lower crust equals
1,520 ppm and for the upper crust equals
3,910 ppm. Growth rates for the lower
and upper crusts are calculated to be 35
and 8 mm/106yr, respectively. The maxi-
mum thickness of the upper crust is about
3 mm, and the underlying hydrothermal
layer is as great as 10 mm thick; this im-
plies that the hydrothermal activity may
have continued to affect the talus for 0.3
m.y. before hydrogenous activity began.
The magnetic anomaly age of the ocean
crust under site 75-1 A is approximately 1.4
m.y. (McGregor and Rona. 1975). With this
spacial relationship (Fig. 1), the half-
spreading rate of about 1.3 cm/yr, and
estimated growth rates of crusts, • se-
quence of events can be postulated. When
site 75-1 A rocks spread from the rift axis
to the rift wall position of modern hydro-
thermal activity, they were 0.4 m.y. old.
Hydrothermal activity continued until this
site was 0.7 m.y. old and had moved be-
yond the influence of hydrothermal activity.
Hydrogenous growths of manganese then
ensued for 0.4 m.y. The total age of the
site to account for this sequence of events
would be 1.1 m.y., close to the magnetic
anomaly age of 1.4 m.y. Thus, it seems
that the most active hydrothermal region
is on the rift wall near both high geo-
thermal gradients in the rift and active
faults scarps on the rift wall. A ^trip of
hydrothermally altered oceanic crust
results.
A ropy-textured, seemingly fresh, thin
basalt flow was dredged at site 73-6A and
at 75-1 B; chemical analysis of basalt
73-6A2 shows an abnormally high K20
content of 0.3 percent, whereas fresh typi-
cal rift-valley tholeiites in this region have
only 0.05 to 0.10 percent K20 (Scott and
others, 1973). Samples 6A and IB may be
Figure 3. SEM photographs of manganese crust, (a) Lower crust 75-1A24 showing crystal plates
of birnessite. (b) and (c) Botryoidsof upper crust 75-1A24 showing the absence of crystallinity.
(d) Smooth interior of a ferromanganese crust that caps analcite crystals of a vein 75-1 A28.
GEOLOGY
235
446
slightly alkalic younger off-axis basalts
(Strong, 1974) that may have covered older
hydrothermal areas. The Lag.f./Sme.f.
ratio (0.6) and the rare-earth element
abundances fit intraplate or ridge-crest
criteria (Schilling and Bonatti, 1975).
However, the possibility of simple low-
temperature addition of K during weather-
ing cannot be discounted (Scott and
Hajash, 1975). Thus, no definite limit
to the size of the hydrothermal field can
be established.
Three other localities of hydrothermal
manganese have recently been located;
one is close to the Galapagos spreading
axis, and it has nearly identical physical,
chemical, isotopic, and growth rate
characteristics to the lat 26°N deposits
(Moore and Vogt, 1975). Another occurs
at lat 23° N on the Mid- Atlantic Ridge
(Thompson and others, 1975). French sci-
entists have also found a hydrothermal
manganese deposit in a transform fault in
the FAMOUS area (ARCYANA, 1975).
Recognition of such occurrences imme-
diately following publication of findings
at lat 26°N suggests that these deposits
may be common and were overlooked in
the past.
The most common hydrothermal deposit
in oceanic spreading centers and related
structures besides manganese crusts are
sulfides precipitated as veins within the
crust (Dmitriev and others, 1970; Bonatti,
1975). It is probable that these two phe-
nomena are both part of the same com-
plex hydrothermal process operating in
the ocean crust. The association is strength-
ened by Hajash's (1975) observation of
experimental chalcopyrite and pyrrhotite
precipitation in Fe- and Mn-rich sea water
resulting from reaction with basalt at 400°
to 500°C. Some mechanisms may even exist
to precipitate sulfides at the water-rock
interface from chloride-rich brines (Sillitoe,
1972; Constantinou and Govett, 1972;
Searle, 1972; Upadhyay and Strong, 1973;
Sato, 1973). Thus far, no massive sulfides
have been observed in rocks dredged from
open-ocean centers either within or on the
rock-water interface. Obviously, both
hydrothermal manganese and sulfides will
have to be studied to understand the total
chemical effects of cycling sea water
through cooling oceanic crust.
REFERENCES CITED
ARCYANA, 1975, Transform fault and rift
valley from bathyscaph and diving saucer
Science, v. 190, p. 108-1 16.
Better, P. R., Bolger. G. W., McGregor, B. A.,
and Rona. P. A., 1974, The Mid-Atlantic
Ridge and its effect on the composition of
particulate matter in the deep ocean: EOS
(Am. Geophys. Union Trans.), v. 55, p. 193.
Bischoff, J. L., and Dickson, F. W., 1975,
Seawater-basalt interaction at 200 C and
500 bars: Implications for origin of sea-
floor heavy-metal deposits and regulation
of seawater chemistry: Earth and Planetary
Sci. Letters, v. 25, p. 385-397.
Bonatti, E., 1975, Metallogenesis at oceanic
spreading centers, in Annual review of Earth
and planetary sciences: Palo Alto, Calif.,
Annual Reviews, Inc., p. 401-431.
Bonatti, E., Kramer. T., and Rydell, H. S.,
1972, Classification and genesis of sub-
marine iron-manganese deposits, in Ferro-
manganese deposits on the ocean floor:
Palisades, N.Y., Lamont-Doherty Geol. Obs.,
Columbia Univ., p. 146-166.
Brown, F, H., Pabst, A., and Sawyer, D. L.,
1971, Birnessite on colemanite at Boron,
California: Am. Mineralogist, v. 56,
p. 1057-1064.
Constantinou, G., and Govett, G.T.S., 1 972,
Genesis of sulphide deposits, ochre and
umber of Cyprus: Inst. Mining and Metal-
lurgy Trans., v. 81, p. B33-B46.
Dmitriev, L. V., Barsukov, V. L., and Udintsev,
G. B., 1970, Rift zones of the ocean and the
problem of ore formation: Geokhimiya,
v. 4, p. 937-944.
Fewkes, R. H., 1973, External and internal
features of marine manganese nodules as
seen with the SEM and their implications in
nodule origin, in Morganstein, M., ed., The
origin and distribution of manganese nodules
in the Pacific and prospects for exploration:
Honolulu, Hawaii Inst. Geophysics, Univ.
Hawaii, p. 21-29.
Hajash, A., 1975, Hydrothermal processes along
mid-ocean ridges: An experimental investi-
gation: Contr. Mineralogy and Petrology
(in press).
Lister, C.R.B., 1974, Water percolation in the
ocean crust: EOS (Am. Geophys. Union
Trans.), v. 55, p. 740-742.
McGregor, B. A., and Rona, P. A., 1975, Crest
of Mid-Atlantic Ridge at 26 N: Jour. Geo-
phys. Research, v. 80, p. 3307-3314.
Moore, W. S., and Vogt, P. R., 197S, Hydro-
thermal manganese crusts from two sites
near the Galapagos spreading axis: Earth and
Planetary Si. Letters (in press).
Mottl, M. J., Corr, R. E., and Holland, H. D.,
1974, Chemical exchange between sea water
and mid-ocean ridge basalt during hydro-
thermal alteration: An experimental study:
Geol. Soc. America Abs. with Programs,
v. 6, p. 879-880.
Sato, T., 1973, A chloride complex model for
Kuroko mineralization: Geochem. Jour.,
v. 7, p. 245-270.
Schilling, J. G., andoBonatti, E., 1975, East
Pacific Ridge (2 S-10 S) versus Nazca
intraplate volcanism: Rare-earth evidence:
Earth and Planetary Sci. Letters, v. 25,
p. 93-102.
Scott, M. R., Scott, R. B., Rona, P. A., Butler,
L W., and Nalwalk, A. J., 1974, Rapidly
accumulating manganese deposit from the
median valley of the Mid-Atlantic Ridge:
Geophys. Research Letters, v. 1, p. 355-358.
Scott, R. B., and Hajash, A., 1975, Initial sub-
marine alteration of basaltic pillow lavas:
A microprobe study: Am. Jour. Sci.
(in press).
Scott, R. B., Hajash, A., Kuykendall, W. E.,
Rona, P. A., Butler, L. W., and Nalwalk,
A. J., 1973. Petrological and structural
significance of the Mid-Atlantic Ridge be-
tween 25°N and 30°N: EOS (Am. Geophys.
Union Trans.), v. 54, p. 249.
Scott, R. B., Rona, P. A., McGregor, B. A.,
and Scott, M. R., 1974, The TAG hydro-
thermal field: Nature, v. 251, p. 301-302.
Searle, D. L., 1972, Mode of occurrence of the
cupriferous pyrite deposits of Cyprus: Inst.
Mining and Metallurgy Trans., v. 81,
p. B189-B197.
Sillitoe, R. H., 1 972, Formation of certain
massive sulphide deposits at sites of sea
floor spreading: Inst. Mining and Metal-
lurgy Trans., v. 81, p. B14I-B148.
Sorem, R. K . and Foster, A. R., 1972, Marine
manganese nodules: Importance of struc-
tural analysis: Internal. Geol. Cong., 24th,
Montreal 1972. sec. 8, p. 192-200.
Strong, D. F., 1 974, An "off-axis" alkali vol-
canic suite associated with the Bay of Islands
ophiolites, Newfoundland: Earth and
Planetary Sci. Letters, v. 21, p. 301-309.
Swanson, S. B., 1975, Two examples of second-
ary alteration associated with mid-ocean
ridges (Master's thesis | : College Station,
Texas A&M Univ.
Thompson, G., Woo, C. C, and Sung, W., 1975,
Metalliferous deposits on the Mid-Atlantic
Ridge: Geol. Soc. America Abs. with Pro-
grams, v. 7, p. 1297-1298.
Upadhyay, H. D.. and Strong, D. F., 1973,
Geological setting of the Betts Cove copper
deposits, Newfoundland: An example of
ophiolite sulfide mineralization: Econ.
Geology, v. 68, p. 161-167.
ACKNOWLEDGMENTS
Reviewed by Enrico Bonatti and Ronald
Sorem.
Research supported by the Institute of
Oceanology of the USSR Academy of Sciences,
the National Oceanic and Atmospheric Admin-
istration, and National Science Foundation
Grant DES 74-18567.
We are indebted to the scientific colleagues
and crew aboard the R/V Akademik Kurchatov
Mark DiStefano and Steve Swanson of Texas
A&M University significantly aided our SEM
and x-ray research.
MANUSCRIPT RECEIVED DEC. 4, 1975
MANUSCRIPT ACCEPTED JAN. 13, 1976
236
APRIL 1976
447
47
Reprinted from: Marine Geology, Vol. 20, No 4, 315-334.
Marine Geology, 20(1976) 315—334
© Elsevier Scientific Publishing Company, Amsterdam — Printed in The Netherlands
RIDGE DEVELOPMENT AS REVEALED BY SUB-BOTTOM PROFILES
ON THE CENTRAL NEW JERSEY SHELF
W. L. STUBBLEFIELD and D. J. P. SWIFT
National Oceanic and Atmospheric Administration, Atlantic Oceanographic and
Meteorological Laboratories, Miami, Fla., (U.S.A.)
(Received December 12, 1974; accepted August 4, 1975)
ABSTRACT
Stubblefield, W. L. and Swift, D. J. P., 1976. Ridge development as revealed by sub-
bottom profiles on the central New Jersey shelf. Mar. Geol., 20: 315—334.
Closely-spaced 3.5 kHz seismic profiles were collected over the north-easterly trending
ridge and swale system 50 km east-southeast of Atlantic City, New Jersey. They yield infor-
mation on the Late Quaternary depositional history of the area, and on the origin of the ridge
system. Four of the sub-bottom reflectors identified were sufficiently persistent to
warrant investigation and interpretation. These reflectors, which have been cored, litho-
logically identified, and radiocarbon dated, are stratigraphically higher than the reflectors
dealt with by the majority of previous studies. The upper three reflectors are definitely
mid- and post-Wisconsin in age and present a record of the most recent glacial cycle. The
upper three units associated with the observed reflectors appear to exert a pronounced
influence on the bathymetry. The gently corrugated ridge system of Holocene sand is
formed over the regionally flat-lying upper unit, an Early Holocene lagoonal silty clay.
The characteristically flat, broad depressions of the area are floored by this lagoonal
material. Locally, however, marine scour has cut through the silty clay into an underlying
unit of unconsolidated fine Pleistocene sand. Several stages of trough development appear
to be represented. After penetrating the lagoonal clay, troughs are initially narrow, but
when incised through, the sand into a lower, Pleistocene, silty -clay unit, the troughs
become notably wider. As downcutting is inhibited by the lower clay, the upper clay is
undercut as the trough widens in a fashion similar to a desert blowout.
The sub-bottom reflectors indicate that ridge development on the central shelf has
involved aggradation as well as erosion. Some ridges seem to have grown by vertical and
lateral accretion from small cores. The internal structure of other ridges suggests that they
formed by the coalescence of several small ridges. Others appear to have undergone
appreciable lateral migration.
The ridges appear to be in a state of continuing adjustment to the hydraulic regime of
the deepening post-Pleistocene water column.
INTRODUCTION
A prominent system of northeasterly trending ridges and depressions
exists on the shelf floor 50 km east-southeast of Atlantic City, New Jersey.
Examination of a 1:125,000 scale ESSA bathymetric map contoured by
Stearns (1967) suggests that the ridges comprise two basic populations in
448
316
2 J^^'AP"
15'
Fig.l. Index map of the study area with the New Jersey coastline inset for regional setting.
The bathymetric contour lines are in fathoms.
terms of spacing and height (Fig.l). A small-scale ridge system is super-
imposed on a large system. The latter appears to be impressed onto a broad
shoal-retreat massif, a constructional feature resulting from the retreat of a
littoral drift depositional center associated with a retreating estuary mouth
(Swift, 1973). The ridge spacing of the larger ridge system averages 3.1 km
with a mean flank dip of 0.4°. In addition to the two basic populations a
third, smaller system of contrasting sediment bands of negligible relief, has
been observed from side-scan records, bottom photographs, and submersible
dives (McKinney et al., 1974).
Genesis of the ridge topography of the surficial sand sheet on the inner
and central continental shelf has remained an enigma to workers since the
pioneer work of Veatch and Smith (1939). A historical school of thought
suggests that the pronounced undulations of the sandsheet are fluvial or
littoral features formed during the lower sea-level stands of the Pleistocene
(Veatch and Smith, 1939; Shepard, 1963; Kraft, 1971; McKinney and
449
317
Friedman, 1970). Others questioned the feasibility of these structures
surviving a marine transgression and suggest instead a post-transgressive
response to the Holocene hydraulic regime (Uehupi, 1968; Swift et al., 1972;
Stubblefield et al., 1975). Uehupi (1970) subsequently abandoned the
hypothesis of recent reworking of the Holocene sands and proposed a
mechanism of terraces and barrier beaches overstepped by a transgressive
Holocene sea.
In order to resolve the controversy surrounding the origin of the ridges, a
dense network of high-resolution, shallow-penetration seismic-reflection
profiles was collected in a 400 km2 area (Fig.l). The investigation was
directed toward the internal structure of the sand sheet and the role which
the sub-bottom reflectors contribute to the existing bathymetry. In addition,
some of the reflectors were correlated with data of previous workers in an
effort to establish geological continuity with other sections of the New
Jersey continental shelf.
METHODS
Field methods
The seismic reflection data were collected from the NOAA ship "Peirce"
during August 1973 using a 3.5 kHz transducer. The transducer, with a
0.2 m/sec pulse length was towed 5—6 m beneath the surface at ship's speeds
varying from 3.5 to 4.0 knots. The seismic record was recorded at a 250
m/sec scan rate.
The seismic lines were run normal to the ridged features in an area
previously vibracored (Fig.l). This approach ensured maximum delineation
of the sand sheet's structure and permitted correlation between the core
record and seismic reflectors.
Raydist navigation provided an accuracy of ±10 m.
Laboratory methods
The 115 km of seismic records were scanned for bottom and sub-bottom
reflectors and "hand-smoothed" to compensate for sea surface waves. Each
reflector was converted to X— Y values, placed on computer data cards by
means of a X— Y digitizer unit, and subsequently plotted by a Univac 1108
computer. With this method, the horizontal scale was reduced by a 1:10 ratio
and the vertical by 1:2 resulting in a vertical exaggeration of 5. This
exaggeration, together with that resulting from the speed of the vessel, yields
a composite vertical exaggeration of 12:1. Such a degree of vertical
exaggeration enables delineation of subtle features in the original record.
A travel time of 1.65 km/sec was used through both the water and uncon-
solidated sediment. The negligible error induced by a slightly fast travel time
through the water (1.65 km/sec as opposed to 1.50 km/sec) is not in conflict
with the purpose of the study.
450
318
LATE-QUATERNARY STRATIGRAPHY
The seismic profiles reveal changes in acoustic impedence (reflectors)
which, in turn, can be correlated with the lithology sampled by vibracores.
As many as eleven reflectors were observed in the seismic records but only
four were of sufficient consistency throughout the area to warrant discussion.
The four reflectors of interest have been lithologically identified, strati-
graphically dated from vibracores (Stubblefield et al., 1975) and described
(Fig.2). Radiocarbon dates were obtained from analysis of shell material in
I V-l
ss (CREST)
-W
SEA LEVEL
100-
200
CM
I I I i I I i_
10 14 18 2 2
500 BP»
100---
3,760 BP*'
.**
4> -|-^HS
200-
» v-4
(UPPER FLANK)
V-3
(LOWER FLANK)
CM
| | SAND
fHf SILTY CLAY
\t*\ SHELL MATERIAL
<^ CROSS -BEDDING
-?- DISCONFORMITY
Hs HOLOCENE SHELF SAND
H,£.£l.£ SEISMIC REFLECTORS
I I I 1 I J L_
10 14 18 2 2
10,050 B.P.'
10,950 BP1
I00i
22,035 BP:
200
25,300 B.P.;
300-
36,600 BP.
400-
CM '
*i*
V-2
(TROUGH)
29,700 B P«:
IOO-f>:*j
i i i i i i i i ,
10 14 I 8 2 2 4> fc^
32,150 BP
iP1
Si
200-?-^
"--- G
CM
S
as<
I I I I I I I I
10 1.4 18 2 2<
Fig.2. Lithic log of four vibracores and mean grain size in quarter phi (0) units within the
sand horizons. (Modified from Stubblefield et al., 1975).
451
a±.
a
3
£
o
0
JQ
X!
3
C
0
£
4J
03
s
"a!
'w
^
t/3
m
0
0
Oh
o
03 c
-a c 2
* 2 S
- ■- §
Ma?
3 Oi c
WD J)
taJO
03
15
O
■c £
<*H
tm
0
<
e
0
t-H
-3
W
a
J
•e
<
u
(A
'2
H
Q
D
0
S2
03
01
e
3
Oh
U-i
o
2
c
■3
c
41
03
03
£
44>
Oi
4J
'£
o
0
c
0)
C
H- H
5?
£
£
44>
C
5
03
0
'3
J
u
Oh
o
o
CO
O
0) CO +*
1« si
o
■s: --1 a) a s
§ « -2 g "C
s. a « ° *
in <2 ii c rb
Oh < Ph 3 w
0
03
-u
CD
CIS
0
0>
=3 c
.S «
-a c
o
03 _S
03 s
c
03
& 2
So 2
>
0 0
0 o
-a
b -C
a ■&
03
a. hS
a «
I*
a w
I I
C 03
si I
£ & £
3 rt a
PH wfc
o
o
o
o
CO
CO
to
CD
CM
CO
V
A
3 0
00
Irt
o
CO
o
<o
CO
N
lO
(M
(M
V
V
01
XI
T3
Oi
4^
2
n
01
'3
0
S
&
4^
U3
<*4
3
c
T3
C
03
0
41
4>
a
o
03
%
0
61)
OJ
J2
I/J
u
441
"5
04
a
0
1/3
T3
IH
£
O
O
I
o
-a
>>
03
>>
03
i %
Li
41 £
tac _
C w
ZT >>
C 3
£ JS
« -a
£ 2,
>S -g
- a;
3 O
a v
3 "3
g c
u tab
-5 >>
03 £
-o
^ >>
4) *_;
£ '33
-2 o
C
03
!/)
oi *j
£ '35
o
LO
Irt
CO
05
©_
o"
(N
H
C<1
V
452
320
the vibracores. To avoid the often confusing situation of reflector labeling
versus stratigraphic-horizon labeling, each unit in Fig. 2 is denoted by the
same label as its upper boundary reflector; e.g., reflector H marks the upper
boundary of unit H. H denotes Holocene deposits, P the Pleistocene material,
and sediment of questionable age is marked as G. Various observed features
of the four seismic reflectors are summarized in Table I.
QUATERNARY STRATIGRAPHY
Unit G
Unit G is the lowest unit in the section. It has been penetrated by vibra-
coring only in its uppermost 10 cm (core V2, Fig. 2), where it appears as a
clean, fine-grained, shell-free sand. Its textural character and its minimum
constraining age of >36,000 B.P., suggests that this unit was deposited in a
near-shore environment, perhaps during a period of an advancing sea marking
the commencement of the Plum Point Interstadial (mid-Wisconsin). This
inference is supported by the apparent absence of an unconformity between
this supposed basal sand and the overlying offshore silty-clay deposit. Strati-
graphically, however, this unit appears to correlate with the unit underlying
Garrison's (1970) Pleistocene unconformity, which he suggests as Late
Tertiary in age. Garrison's work was over a broad area of the continental
shelf south of New England and to the northeast of our work area. The small
size of the study area relative to Garrison's work may have resulted in the
lack of detection of an unconformity. Until a more detailed coring program
is completed, the age of unit G and thus an interpretation of its depositional
environment remains uncertain.
Unit PI
This medium gray, silty clay is perhaps the most widespread of the Late-
Quaternary sequence, as indicated by the persistency of its reflective surface,
reflector PI (pp.324— 325). This unit is of Pleistocene age, with dates ranging
from 25,300 ± 1040 B.P. to >36,000 B.P. The younger section of this unit
was probably an offshore deposit formed in advance of the prograding
shoreline represented by unit P. However, the older part of unit PI, approxi-
mately 36,000 B.P. in age, may reflect the maximum glacial retreat during
the Plum Point Interstadial as described by Goldthwait et al. (1965) and
Milliman and Emery (1968).
The age of unit PI, mid-Wisconsin including the Farmdalian substage, is
comparable to that proposed by McMaster and Ashraf (1973) for their
reflector, P2. They made a tentative correlation of their reflector with
Garrison's (1970) Pleistocene unconformity. McMaster and Ashraf's work
was to the east of this study on the eastern fringe of Long Island extending
south to the shelf break. They traced their P2 reflector across most of the
shelf at sub-bottom depths of 17—34 m, but fail to mention the amount of
453
321
regional dip of their P2 other than that it parallels the present shelf
surface. In the present study area of this study the regional dip of PI was
calculated to be 0.04° to the S61°E. If 0.04° dip is assumed, an approxi-
mation of 17m/97 km (17 m/l° latitude) depth compensation may be
applied. By projecting McMaster and Ashraf's reflector for an additional
80—90 km in a plane normal to the strike of the eastern Long Island coast-
line, a depth comparable to that of our reflector PI results.
Unit P
The uppermost Pleistocene sand, dated at 22,035 ± 665 B.P., possesses a
slightly irregular reflective surface and ranges in thickness from 1 to 8 m.
Throughout most of the sample area, however, the thickness varies from 2 to
4 m. The maximum thickness of unit P is in the southeast sector.
The upper reflective boundary of this unit, reflector P, has a dip of 0.02°
and a strike of S38°E. The dip angle and strike direction were calculated
using the reflector's depth throughout the study area. The strike direction is
within 5° of the present beach orientation in the vicinity of Little Egg Inlet,
New Jersey (Fig.l).
Unit P is a clean, medium-grained upward-coarsening sand (Fig. 2). This
characteristic and its date of 22,035 B.P. suggest a deposition environment
of a prograding shoreline. If this inference is valid, the advance of the
Holocene seas was controlled by the regional gradient established during
periods of lower sea level, since the coast-concordant strike indicates only a
slight reorientation of the beach during the last 20 millenia.
After the Plum Point Interstadial, the ice sheets readvanced, the marginal
seas withdrew, and the Pleistocene sand of this unit was exposed to subaerial
processes. Fig. 3 suggests that the Pleistocene sand, which is 40—50 m below
THOUSANDS OF YEARS BEFORE PRESENT
4 8 12 16
ii i
20 24
i i i i i
28 32
i i i i
36
i i
^~""^N.
\ \
\ \
/ /
\
\
\ \
1 1
\
\
>•
\ \
/ 1
\
\ \
/ /
\
\ \
/ /
\
/ /
\ /
\
\J
MIUIMAN & EMERY
(1968)
CURRAY (1965)
20
40
60 S 5
uu
80 X at
100
-120
140
Fig. 3. Comparison of data from the vibracores with sea level curves of Milliman and
Emery (1968) and Curray (1965). The sample ages are represented by dots and the range
in age by error bars. (From Stubblefield et al., 1975)
454
322
present sea level, was a positive area from 12,000 to 20,000 B.P. Radio-
carbon dates from the four vibracores are in general agreement with the sea-
level curves of Milliman and Emery (1968) and Curray (1965). Visual
evidence in Core V-3 (Fig.2) indicates a possible disconformity at the top of
the clean sand (unit P). Sheridan et al. (1974) report an extensive erosion
surface on their upper Pleistocene unit, which probably correlates with P.
Core V-2 (Fig.2) demonstrates further evidence for erosion, in that the
Holocene lagoonal deposit and the bulk of the Pleistocene sand are both
absent. However, much of the missing section in Core V-2 is thought to have
been removed by the modern marine erosion as explained in a later
discussion, rather than by subaerial erosion during Early Holocene times.
Unit H
The upper unit in Core V-3 (Fig.2) is a medium gray silty clay with a
locally irregular and discontinuous upper boundary (reflector H). The
reflector appears as an undulating surface with the deepest part found under
a topographic high in the eastern sector of the study area. The depth of
reflector H, below present sea level, ranges from 36 to 52 m and the thick-
ness of unit H varies from 0 to 6 m. Unit H thickens to the southeast in the
direction of its regional slope.
The absence of this upper silty clay at various places throughout the study
area is probably the result of both erosion and non-deposition. Examination
of the bathymetry in Fig.l fails to suggest recent downcutting in those places
where the upper silty clay is missing within the substrate, suggesting that its
absence may be due to a positive area during deposition, rather than subse-
quent erosion. However, where unit H intersects the surface, particularly as
an outcrop in the deep topographical troughs (McKinney et al., 1974), and
in those places where an underlying unit is surficially exposed (profile A,
Fig. 4a; profile K, Fig. 4b) erosion of unit H is obviously occurring.
The depositional environment of the upper silty clay is inferred from
radiometric ages, lithology, depth of unit, and fragmented shell material. The
silty clay is underlain by medium to fine sand (unit P) which is dated at
22,035 ± 665 B.P., and is overlain by coarse sand with shell material dated at
10,950 ± 360 B.P. These limiting ages indicate that the unit was deposited
during a period in which landward passage of the shoreline occurred. The
lithology of cored sections of this unit is similar to that described by
Sheridan et al. (1974) as a Holocene lagoonal deposit, near the Delaware
coast. In addition, the average depth of reflector H is approximately 42 m
below present sea level which places the unit in that portion of the Emery
et al. (1967, fig. 4) diagram described as lagoonal. The geographic limits of
the Emery et al. (1967) study is sufficiently close to this work to allow
application of its interpretations to our data. Shell material, too small to
radiocarbon date, has been identified by Don Moore, University of Miami, as
organisms capable of living in shallow, brackish environments (Crassostrea
virginica and Mercenaria mercenaria), a conclusion which supports our
inference of lagoonal deposition.
455
323
Holocene lagoonal deposits tend to occur during glacial retreat and marine
transgression. The bracketing dates for unit H (> 10,950 <22,035 B.P.)
include the time of maximum glacial advance which occurred 18,000 to
22,000 B.P. during the Woodfordian glacial cycle (Goldthwait et al., 1965;
Schafer and Hartshorn, 1965). If this unit does in fact reflect deposition
during glacial retreat subsequent to maximum Late Wisconsin ice advance,
and if a date of approximately 16,000 years B.P. is accepted for the
Pleistocene— Holocene boundary (Emery and Uchupi, 1972) a date of post-
Pleistocene may confidently be applied to this silty clay.
These four units, their related seismic reflectors, and their time-stratigraphic
framework provide a record of a complete glacial cycle on the central New
Jersey shelf. The retreat of the ice sheet, accompanied by the advance of the
ocean is indicated by the lower unit G. PI possibly represents maximum
glacial retreat and marine transgression during late mid-Wisconsin time. Unit
P is then representative of the subsequent ice advance and retreat of the
ocean. Unit H was deposited by the advancing Holocene lagoonal belt.
SURFICIAL SAND SHEET
Topography and internal structure
The surficial sand sheet above reflector H is complexly structured. Large-
scale ridges (shaded pattern of inset, Fig.l) appear as half-cylinders of sand
resting on a relatively level reflector H (Fig. 4a, b). In some cases, internal
structure may be observed. This may take the shape of apparent ridge
"cores", formed by internal strata which parallel the ridge flanks (station 86,
profile C, Fig.4a; record A, Fig. 5). Elsewhere, multiple "cores" within a ridge
suggest coalescence of several nuclei during growth (station 77 to 80, record
C, Fig. 4a). Internal reflectors with consistent direction of dip occur in some
ridges (station 251 to 255, record G, Fig.4b) suggesting lateral ridge migration.
Internal patterns are locally very complex; in record B, Fig.5, truncated
reflectors suggest that a former ridge on the northwest side of the record has
been leveled and the adjacent trough filled in; a new ridge has appeared to
the southeast.
Large-scale troughs (stations 120—140, profile D, Fig. 4a) appear to
bottom in reflector H which is thinly mantled with a few centimeters of
coarse, shelly or pebbly sand locally grading upward into centimeters of finer
sand. This fine sand thickens towards the ridge flanks (Stubblefield et al.,
1975). Locally, reflector H is without this surficial coating. Small-scale ridges
(linear pattern of inset, Fig.l) likewise appear to be half cylinders of sand
resting on reflector H. In the few cases where internal structure have been
resolved, it appears to be similar to that of the large-scale ridges.
Small-scale troughs, unlike large-scale troughs, locally penetrate through
reflectors H and P, into unit PI (Fig.6). Two variants of such apparently
erosional troughs appear. Small-scale troughs which penetrate into the sand
of unit P tend to be "V" shaped in cross-section (station 60 to 62, profile B,
Fig.4a; record A, Fig.6). Other small-scale troughs extend completely through
456
324
b 2,
o
in
.
\Ti
b
J
o
vk
b
b
c
3WI1 13AV«i AVM0M1
457
325
C*-H
>>
0
n
(U
CU
£
£
^3
>,
Si
0>
>
03
0
-^>
<
T3
j
C
0)
03
>
*
m
y
s-<
CS
c«
Gfl
>i
+->
-a
c
3
01
09
01
V
a
J3
&
O
0
+j
0)
a>
ja
&M
s
0
u
a
£ 2.
3 Dh
« -5
■o 3
o
cfl
.fi y
• a
I 2
0 A
y «
_ y
GO JZ
a 53
!S o
Jh C|_|
°^
-3 .2?
y y
C
— T3
y c
3 2
o °
£
y
i/j
«
y
CO
£
^
-*
m
so
CO
458
326
100m
®
78
79
PROFILE C
*'*V* i*?**** -"' *"?*'
'"» ~-""*vV,,v •»■>,»
100m
Fig.5. Hand-enhanced, high-resolution seismic-reflection profiles. A. Ridge with internal
"core", suggesting upward ridge growth by the addition of conformable beds. B. Zone of
discontinuous ridge growth. Ridge at right was formed subsequent to filling of trough at
left by the progradation of incline strata. See Fig.l for location.
the unconsolidated Pleistocene sand and are floored by the silty clay of unit
PI and assume a more nearly parabolic cross-section, with a rounded bottom
and more gently inclined flanks.
Evolution of the ridge topography
Large-scale ridges are locally broken into segments by small-scale troughs
which cross them at a low angle, suggesting that small-scale troughs formed
after large-scale troughs (McKinney et al., 1974). The varieties of ridges and
their internal structure suggest the following model for ridge evolution ( Fig.7).
Large-scale ridges, hereafter called primary ridges, were initiated immediately
after the passage of the shoreline, at about 10,000 B.P. (Fig.7a). They
formed in the leading edge of the shelf sand sheet (Duane et al., 1972), which
459
327
~^&>Z!S&£^^F£!^!!Z!!!%!!c£
t
PROFILE B
100m
PROFILE K
, .,-, ,*»?'* «**■.'
r l^f! '
100m
Fig. 6. Hand-enhanced, high-resolution seismic-reflection profiles. A. Immature, small-scale
trough with a V-shaped profile. B. Mature trough with a parabolic profile with a broad
axis and gentle side slopes. Note that the flat-lying strata abut against sides of troughs.
The upper datum is approximately 5 m beneath the sea surface.
advanced as the shoreface underwent erosional retreat (Stahl et al., 1974).
Primary troughs formed concomitantly with primary ridges, by non-
deposition between ridges, or by the movement of sand off reflector H onto
the ridges.
As the Holocene transgression continued and the water column deepened,
the ridge topography underwent progressive modification. Ridge spacing, a
function of flow depth (Allen, 1968), increased. Internal ridge structure
suggests that this was accomplished by lateral ridge migration, or by coales-
cence of several smaller ridges. Ridge width appears to have increased as a
result of more intense sedimentation on ridge flanks than on ridge crests, so
that the ridges expanded laterally, rather than building upwards. As a result
internal reflectors are generally steeper than present ridge flanks.
Scour of the trough floors has locally breached reflector H. Where this has
460
328
PRIMARY RIDGE
DEVELOPMENT OF RIDGE TOPOGRAPHY
Fig. 7. Hypothetical model for ridge development. A. Large-scale ridges are initiated in the
nearshore environment. The ridges grow by vertical and lateral aggradation, resulting in
"concentric" stratification. Large-scale troughs are zones of non-deposition. B. Small-scale,
secondary trough is incised into older substrate. Profile is initially V-shaped. C. Secondary
trough widens; the Pleistocene sand is readily eroded and the Holocene clay is subject to
undercutting. Aggradation of secondary ridges is fed in part by sand released during
trough erosion.
occurred, rapid downcutting and removal of the noncohesive sand of under-
lying unit P appears to have resulted in a relatively steep- walled small-scale
trough, hereafter called a secondary trough (Fig.7B). At the same time,
secondary ridges began to appear in the primary troughs. Some ridges seem
to bear a levee-like relation to the secondary troughs (Fig.7B), as though
excavation of the former supplied material to the latter. Where secondary
troughs have penetrated as far as the silty-clay surface of reflector PI (Fig.
7C) the troughs are broader, as though the clay inhibited further downcutting,
and encouraged lateral erosion and trough widening, after the fashion of a
desert blow-out.
This model for ridge formation may be compared with the inshore portion
of the study area, traversed by profiles A— G (Fig.l). Here a broad primary
trough between two primary ridges develops a secondary topography as it is
traced southwest (profiles A— G, Fig.4). In Fig.8, bathymetry of transects A,
B and C are matched with a common datum, approximately 13 m below sea
level (computed with a travel speed of 1.5 km/sec through water), and
adjusted laterally so that successive crests of the landward primary ridge
(position 1) coincide. Since the secondary topography (positions 2—5)
becomes increasingly better developed through profiles C, B and A to the
south, this series may be approximately equivalent to a time series, and as
such may be compared with Fig.7.
461
329
Fig.8. Overlay of bathymetry from transects A, B, and C. See Fig.l for relative geographic
positions. The five labeled positions are explained in the text.
Possible relation of ridge topography to hydraulic regime
The progressive evolution of the ridge topography appears to reflect an
attempt on the part of the sea floor to maintain an equilibrium response to
the slowly changing hydraulic regime during the period of water column
deepening and shoreline retreat associated with the Holocene transgression.
While a morphologic progression may be inferred on the basis of seismic
reflection and related data, the character of the hydraulic forcing mechanism
must remain speculative until the nature of the flow field on the New Jersey
shelf is adequately documented. Our present, rather unsatisfactory state of
knowledge may be summarized as follows. Beardsley and Butman (1974)
have noted that sustained, high-velocity currents in the Middle Atlantic Bight
occur primarily during those winter storms whose trajectories, relative to the
shelf, permit a prolonged period of northerly winds. Such winds cause an
Ekman transport of surface water to the coast, and result in coastal setup on
the order of 40—60 cm (Beardsley and Butman, 1974). When this occurs, a
shelf-wide southward geostrophic flow ensues. Studies conducted by Csanady
and Scott (1974) under somewhat different circumstances suggest that the
coastal margin of such geostrophic flow may assume a jet-like character, with
velocities greater than those of the main flow. Limited data from the North
Carolina coast (Swift, 1975) suggests that during periods of peak flow such
accelerated coastal flows may experience downwelling and that inner shelf
ridges may in fact be initiated by such flow, with the downwelling jet local-
ized between the ridge and the shoreface.
Ridges left behind on the shelf floor by the retreating shoreface continue
to be maintained by the flow field. Pre-recent substrate continues to be
exposed in the troughs, and as noted above, secondary patterns of ridges
may appear. Continued ridge maintenance requires that the storm flow field
be structured; a homogeneous flow field would serve to degrade ridge crests
and fill in troughs. Such structure has not been observed, but it is predicted
by theoretical and experimental work of Ekman (1905), Faller (1963, 1971),
Faller and Kaylor (1966), Lilly (1966), Hanna (1969) and Brown (1971).
Ekman (1905) has shown that wind-driven shelf flows would tend to have a
three-layered structure. An upper boundary layer consists of wind-driven
water, whose speed is in excess of the geostrophic value induced by regional
set-up. The upper boundary layer is characterized by an Ekman spiral in
462
330
which wind-driven surface water moves at 45° to the right of the wind (in
the northern hemisphere). With increasing depth, velocity vectors shift
progressively to the right, and speed decreases to the geostrophic value.
Depth-averaged flow in the upper boundary layer is 90° to the right of the
wind direction. Beneath the upper boundary layer, the core flow is geo-
strophic in nature, moving approximately parallel to the coast in response to
the pressure field induced by wind set-up. Below the core flow, water is
sheared against the stationary sea floor, causing a lower boundary layer that
is characterized by a reverse Ekman spiral. As the bottom is approached,
flow is diverted progressively toward the left (in the northern hemisphere),
and the speed is progressively reduced.
The thickness of the boundary layers is a function of velocity at the top
of each layer, and the characteristic eddy viscosity of the layer. Very little is
known about the values that these parameters attain on the Atlantic shelf
during storms. However, it seems probable that during storm flows, the
boundary layers would expand at the expense of the core flow and would
partially or completely overlap (Leetmaa, 1975).
Above a critical Reynolds number the internal character of the boundary
layers, as well as their thickness, must change markedly. The boundary layers
become unstable. However, since the flows are still subject to the Coriolis
effect, the instability is not randomed but ordered (Faller and Kaylor, 1966;
Faller, 1971). The flow divides into relatively sharply defined zones of down-
welling, high-velocity surface water, alternating with more diffuse zones of
upwelling lower-velocity bottom water (Faller, 1971). The result is a series
of helical vortices, with alternate cells rotating with the opposite sense.
The extent to which this scheme applies to storm flows on the Atlantic
shelf is not known. However, if such a cellular flow structure should couple
with the cohesiveless substrate of the Atlantic shelf during periods of peak
flow, sand ridges might be expected to localize, and be localized at, zones of
bottom-current convergence; and troughs at zones of bottom-current
divergence. According to this model the ridges would be classic bedforms, in
the sense of morphologic responses to secondary flow patterns within a
sheared flow (Wilson, 1973).
The nature of such coupling is problematic. The flow cells, if they indeed
occur, are very flat. In the study area the ratio of ridge height to crest-to-crest
distances averages 1:12 for the secondary ridges and 1:120 for the larger
primary ridges. The ratio for the secondary ridges is similar to flow-cell
spacing described by Faller and Kaylor (1966) for rotating tank experiments.
They noted a spacing of 11 D for "small scale" cells where D is equal to the
thickness of the Ekman layer. The spacing is "much greater" for large-scale
cells. The sharply defined nature of the secondary troughs corresponds with
the sharply defined nature of downwelling zones. Large-scale, primary
troughs, however, are broad and flat. If they are responses to zones of down-
welling, then the focus of downwelling must shift across the trough during
the storm event in order to produce substrate mobility over large areas.
Faller and Kaylor note that small-scale cells are aligned to the left of the
463
331
mean flow, while the large-scale cells are aligned to the right of the smaller
cells. Large-scale, primary ridges, in fact, tend to be aligned to the right of
small-scale, secondary ridges. By this criterion, primary and secondary ridges
may be of synchronous origin. However, morphologic relationships suggest
that primary ridges formed prior to secondary ridges. The sequence of
primary ridges can be traced landward on the northern flank of the Great
Egg Shelf Valley (Swift, et al., 1972) where they appear to be presently form
forming on the shoreface. The offshore primary ridges tend to be J-shaped,
hooking landward into the Great Egg Shelf Valley, as though affected by the
tidal flow of the shelf valley when it was an active estuary mouth (McKinney
et al., 1974). Secondary ridges only occur in this offshore zone. Their
associated troughs pass through primary ridges, as though they were a later
overprinting. If this is the case, then the secondary ridges may be a response
to a change in the hydraulic regime initiated by a critical alteration in the
shoreline configuration and bathymetry during the course of the Holocene
transgression.
Source of the Holocene sand sheet
The seismic reflection observations presented above, including published
data, place some constraints on possible sources for the Holocene sand sheet
of the central New Jersey shelf. As noted by Meade (1969), Atlantic coastal
estuaries are not sources of fluvial sand, but instead serve as sinks for both
fluvial and littoral sands. Thus sands overlying the Holocene lagoonal carpet
must be of other than fluvial origin. Possible sources are: (1) erosional retreat
of the barrier face; (2) in-situ origin by breaching of the lagoonal carpet and
erosion of the underlying sand; and (3) southerly transport on the shelf
surface during storms.
The first possibility is difficult to evaluate. Since the Late Holocene
reduction in the rate of sea-level rise, New Jersey coastal barriers appear to
be at least locally growing upwards in place, being nourished by the inner
shelf sand sheet, rather than vice versa (McM aster, 1954). The reverse may
have been true during the earlier period of rapid sea-level rise, with the sand
sheet forming as a debris blanket resulting from erosional shoreface retreat
(Stahl et al., 1974). Since the barriers themselves rest on the lagoonal carpet
deposited landward of them, their source of sand during this period must
have been from updrift, from eroding headlands, or from zones where the
shoreface was incised through the lagoonal carpet (unit H) into the under-
lying Pleistocene sand (unit P). This hypothesis is in accord with the regional
ridge pattern (Fig.l) in which the sequence of ridges extending from the
study area back to the New Jersey coast appears to form a shoal-retreat
massif, marking the retreat path of the littoral drift depositional center on
the north side of the ancestral Great Egg Estuary (Swift et al., 1972).
The second potential source, from the excavation of secondary troughs
into the pre-recent substrate, is by itself inadequate to account for the
Holocene sand sheet. Examination of Fig.4 indicates that of the 115 km of
464
332
seismic record, less than 9 km reveal erosion through the silty clay of unit H
into the loose sand of unit P beneath. Furthermore, the surficial sand sheet
is 2—12 m thick, but the Pleistocene sand, where still capped by unit H, is
1— 8 m thick; its volume is inadequate to serve as a sole source.
We note, however, that northeast of the study area, the ridge topography
of the massif gives way to a nearly flat surface with broad shallow hollows
(Uchupi, 1970, pl.l). Deflation of this surface by shelf flows may also have
contributed to ridge growth in the study area.
SUMMARY
The ridges occur in a belt trending across the shelf normal to the shore
(Fig.l). The axis of individual ridges extend across the ridged zone, parallel
or sub-parallel to the shore. The ridge sequence is inferred to be a shoal-
retreat massif, the retreat path of the littoral-drift convergence localized on
the northeast side of the ancestral Great Egg Estuary. Nearshore members of
this sequence appear to be forming as shoreface-connected ridges in response
to coastal storm flows (Duane et al., 1972). A little further seaward, similar
ridges may have been detached from and abandoned by the shoreface during
Holocene sea-level rise. Yet further seaward, in the study area described by
this paper, larger ridges are spaced further apart. This may be an innate
characteristic, due to the more intense nature of tidal flows associated with
the Great Egg Estuary when it still received the drainage of the ancestral
Schuylkill River (Swift et al., 1972), or it may reflect an adjustment of the
ridge topography to the increasing depth of the flow. The character of
internal reflectors suggests that this response took the form of a dominance
of flank over crestal aggradation, so that narrow, steep-sided ridges became
broader with more gently inclined flanks, and that lateral migration of ridges
also occurred.
The large-scale, offshore ridges appear to have undergone a second stage of
evolution, in that a pattern of small-scale, more southerly trending ridges have
been imprinted over the first pattern. Secondary troughs have locally been
incised into the Early Holocene silty clay that underlies the surficial sand
sheet. These are relatively steep-walled features. However, where they have
penetrated to the underlying Pleistocene sand, undercutting and lateral
erosion have resulted in broader, more gently rounded features. Secondary
ridges may have been nourished in part by material released during the
formation of secondary troughs.
ACKNOWLEDGEMENTS
We are indebted to the officers and crew of NOAA ship "Peirce" for their
professional abilities and cooperative attitude. We thank Sue O'Brien and
Dave Senn for drafting, Thomas Clarke of the University of Virginia for
computer programming, and Drs. G. H. Keller and H. B. Stewart, Jr. for
critical review. Radiocarbon dates were provided by facilities at the
Department of Geology, University of Miami, Florida.
465
333
This study is part of the National Oceanic and Atmospheric Admini-
stration's Marine Ecosystem Analysis (MESA) program.
REFERENCES
Allen, J. R. L., 1968. The nature and origin of bed-form hierarchies. Sedimentology,
10: 161-182.
Beardsley, R. C. and Butman, B., 1974. Circulation on the New England Continental shelf:
Response to strong winter storms. Geophys. Res. Lett., 1: 181—184.
Brown, R. A., 1971. A secondary flow model for the planetary boundary layer. J. Atmos.
Sci., 27: 742-757.
Csanady, G. T. and Scott, J. T., 1974. Baroclinic coastal jets in Lake Ontario during
IFYGL. J. Phys. Oceanogr., 4: 524—541.
Curray, J. R., 1965. Late Quaternary history continental shelves of the United States.
In: H. E. Wright, Jr. and I. G. Frey, (Editors), The Quaternary of the United States.
Princeton Univ. Press, Princeton, N. J., pp.723— 735.
Duane, D. B., Field, M. E., Meisburger, E. P., Swift, D. J. P. and Williams, S. J., 1972.
Linear shoals on the Atlantic inner continental shelf, Florida to Long Island. In: D. J. P.
Swift, D. B. Duane and O. H. Pilkey, (Editors), Shelf Sediment Transport: Process and
Pattern. Dowden, Hutchinson and Ross, Stroudsburg, Pa., pp.499— 575.
Ekman, V. W., 1905. On the influence of the earth's rotation on ocean currents. Ark. Mat.
Astron. Fys., 2: 1—53.
Emery, K. O. and Uchupi, E., 1972. Western North Atlantic Ocean; Topography, rocks,
structure; water, life, and sediments. Am. Assoc. Pet. Geol. Bull., Memoir 17: p. 532.
Emery, K. O., Wigley, R. L., Bartlett, A. S., Rubin, M. and Barghoorn, E. S., 1967. Fresh
water peat on the continental shelf. Science, 158: 1301—1307.
Faller, A. J., 1963. An experimental study of the instability of the laminar Ekman
boundary layer. J. Fluid Mech., 15: 560—576.
Faller, A. J., 1971. Oceanic turbulence and the Langmuir Circulations. Ann. Review
Ecology and Systematics, 2: 201—233.
Faller, A. J. and Kaylor, R. E., 1966. A numerical study of the instability of the laminar
Ekman boundary layer. J. Atmos. Sci., 23: 466—480.
Garrison, L. E., 1970. Development of continental shelf south of New England. Am.
Assoc. Pet. Geol. Bull., 54: 109-124.
Goldthwait, R. P., Dreimanis, A., Forsyth, J., Karrow, P. F. and White, G. W., 1965.
Pleistocene deposits of the Erie Lake. In: H. E. Wright, Jr. and J. G. Frey (Editors),
The Quaternary of the United States. Princeton Univ. Press, Princeton, N. J., pp.85 — 97.
Hanna, S., 1969. The formation of longitudinal sand dunes by large helical eddies in the
atmosphere. J. Appl. Meteorol., 8: 874—883.
Kraft, J. C, 1971. Sedimentary facies patterns and geologic history of a Holocene trans-
gression. Geol. Soc. Am. Bull., 82: 2131—2158.
Leetma, A., 1975. Some simple mechanisms for steady shelf circulation. In: D. J. Stanley
and D. J. P. Swift, (Editors), Marine Sediment Transport and Environmental Manage-
ment. Wiley Interscience, New York, N. Y., in press.
Lilly, D. K., 1966. On the instability of Ekman boundary flow. J. Atmos. Sci., 23:
481-494.
McKinney, T. F. and Friedman, G. M., 1970. Continental shelf sediments off Long Island,
New York. J. Sediment Petrol., 40: 213-248.
McKinney, T. F., Stubblefield, W. L. and Swift, D. J. P., 1974. Large-scale current
lineations on the central New Jersey shelf: investigation by side-scan sonar. Mar. Geol.,
17: 79-102.
McMaster, R. L., 1954. Petrography and genesis of New Jersey beach sands. State of
New Jersey Dept, Conservation and Econ. Development, Geol. Surv. Bull., 63:
239 pp.
466
334
McMaster, R. L and Ashraf, A., 1973. Drowned and buried valleys on the southern New
England continental shelf. Mar. Geol., 17: 79—103.
Meade, R. H., 1969. Landward transport of bottom sediments in the estuaries of the
Atlantic Coastal Plain. J. Sediment. Petrol., 39: 222—234.
Milliman, J. I. and Emery, K. O., 1968. Sea levels during the past 35,000 years. Science,
162: 1121-1123.
Schafer, J. P. and Hartshorn, J. H, 1965. The Quaternary of New England. In: H. E.
Wright, Jr. and J. G. Frey (Editors), The Quaternary of the United States, Princeton
Univ. Press, Princeton, N. J., pp. 113—127.
Shepard, F. P., 1963. Submarine Geology. Harper and Row, New York, N. Y., 557 pp.
Sheridan, R. E., Dill, C. E. and Kraft, J. C, 1974. Holocene sedimentary environment of
the Atlantic inner shelf off Delaware. Geol. Soc. Am. Bull., 85: 1319—1328.
Stahl, L., Koczan, J. and Swift, D. J. P., 1974. Anatomy of a shoreface connected sand
ridge on a New Jersey shelf: Implications for the genesis of the shelf surficial sand
sheet. Geology, 2\ 117—120.
Stearns, F., 1967. Bathymetric Maps of the New York Bight, Atlantic Continental Shelf
of the United States, Scale 1:125,000. National Ocean Survey, National Oceanic and
Atmospheric Administration, Rockville, Md.
Stubblefield, W. L., Lavelle, J. W., McKinney, T. F. and Swift, D. J. P., 1975. Sediment
response to the present hydraulic regime on the Central New Jersey shelf. J. Sediment.
Petrol., 45: 337-358.
Swift, D. J. P., 1973. Delaware Shelf Valley: Estuary retreat path, not drowned river
valley. Geol. Soc. Am. Bull., 84: 2743-2748.
Swift, D. J. P., 1975. Coastal sedimentation. In: D. J. Stanley and D. J. P. Swift (Editors),
Marine Sediment Transport and Environmental Management. Wiley Interscience, New
York, N. Y., in press.
Swift, D. J. P., Kofoed, J. W., Saulsbury, F. P. and Sears, P., 1972. Holocene evolution of
the shelf surface, central and southern Atlantic shelf of North America. In: D. J. P.
Swift, D. B. Duane and O. N. Pilkey (Editors), Shelf Sediment Transport, Process and
Pattern. Dowden, Hutchinson and Ross, Stroudsburg, Pa., pp.499— 575.
Uchupi, E., 1968. Atlantic continental shelf and slope of the United States — Physiography.
U. S. Geol. Surv. Prof. Pap., 529-C: 30 pp.
Uchupi, E., 1970. Atlantic continental shelf and slope of the United States — shallow
structure. U. S. Geol. Surv. Prof. Pap., 529-1, 44 pp.
Veatch, A. C. and Smith, P. A., 1939. Atlantic submarine valleys of the United States:
The Congo submarine valley. Geol. Soc. Am. Spec. Pap., No. 7: 101 pp.
Wilson, I. G., 1973. Equilibrium cross-section of braided and meandering rivers. Nature,
241: 393-394.
467
48
Reprinted from: Marine Sediment Transport and Environmental Management, D. J.
Stanley and D. J. P. Swift, editors, John Wiley and Sons, Inc., Chapter 14,
255-310.
CHAPTER
14
Coastal Sedimentation
DONALD J. P. SWIFT
Atlantic Oceanographic and Meteorological Laboratories, Miami, Florida
The preceding chapters have discussed sedimentation in
the intracoastal zone of lagoons and estuaries which lie
seaward of the main shoreline, and on the open beach
and associated surf zone. This chapter looks at sedimen-
tation in the coastal zone as a whole, from the shoreline
out to an indeterminate distance on the order of 5 km,
where shelf flows are no longer affected by proximity to
shore. From this perspective, the system of longshore
sand transport beneath the zone of shoaling and break-
ing waves can be examined together with a deeper sys-
tem of longshore sediment transport driven by inter-
mittent wind or tidal flows. Time and space patterns of
sediment input into this double system, the character of
sediment transport, zones of temporary storage or per-
manent deposition, and the bypassing of sediment onto
the shelf surface are analyzed. More complex patterns
of sediment transport are also described, which result
when coastal flows associated with straight coastal com-
partments interact with circulation in the erosional re-
entrants of rocky coasts or constructional inlets of la-
goons and river mouths.
ONSHORE-OFFSHORE SEDIMENT TRANSPORT
In considering coastal sediment transport, it is convenient
to divide the movement of sediment into an onshore-
offshore component and a coast-parallel component, and
to consider these separately before examining the coastal
sediment budget as a whole. Coast-parallel transport is
many times more intensive than onshore-offshore trans-
port, but it is the latter that determines morphologic
changes at given coastal transects. Hence this chapter
begins by examining the coast in profile.
Hydraulic Zones and Morphologic Provinces
When examined in cross section, the inner shelf is seen
to consist of a regular succession of morphologic prov-
inces, each associated with a distinctive zone of hydraulic
activity (Fig. 1).
Subaerial environments of open coasts are most highly
developed on barrier islands, where a zone of storm
washover and eolian activity results in washover fiats and
dune belts, respectively. The intertidal swash zone builds
the beach foreshore. The foreshore progrades seaward dur-
ing fair weather by the addition of successive inclined
sand strata to form the beach prism, a body of stored sand.
The upper surface of the beach prism is the beach back-
shore. The zone of breaking waves may be divided into
the breaker line, which tends to maintain a breakpoint
bar, and a surf zone, in which a wave-driven littoral
current flowing parallel to the beach is overriden by the
bores of breaking waves. The littoral current tends to
scour a longshore trough.
On unconsolidated coasts capable of relatively short-
term response to the hydraulic regime, the inner shelf
seaward of the breakpoint bar tends to exhibit two mor-
phologic elements. A more steeply dipping shore/ace ex-
tends to depths of 12 to 20 m. Its upper slope may be
as steep as 1:10; its seaward extremity, at 2 to 20 km
from shore, may slope as gently as 1 : 200. Beyond it lies
468
255
256
COASTAL SEDIMENTATION
AEOLIAN ZONE
STORM WASHOVER
SWASH
SURF ZONE
BREAK
ZONE OF
ZONE
LITTORAL
POINT
SHOALING
CURRENT
WAVES
OLDER S
DEPOSITS
BERM SWASH BAR
DUNE FORESHORE BREAKPOINT
BACKSHORE LONGSHORE BAR SHOREFACE
TROUGH
SHOREFACE
FIGURE 1. Morphologic elements oj the open coast and corresponding hydraulic
provinces.
the flatter inner shelf floor proper; the transition may be
abrupt or very gentle. The upper shoreface, to a depth
of perhaps 10 m, corresponds to the hydraulic zone of
shoaling waves. The lower shoreface and inner shelf flow
also experience the surge of shoaling waves, but their
slopes, textures, and bed forms are equally a response
to unidirectional shelf flows.
The Beach Profile
Circulation in the surf zone and the morphologic response
of the substrate are described in Chapter 13. This sec-
tion deals with the net effect of such hydraulic process
and substrate response on the onshore-offshore sediment
budget.
As a consequence of the enormous and nearly continu-
ous expenditure of energy in the beach and surf zones,
the topographic features of cohesionless sand found there
may only exist as equilibrium or near-equilibrium re-
sponses to the circulation patterns described in the pre-
ceding chapter. The equilibrium is not a static one,
however, as the characteristics of the wave regime that
force the response are constantly changing, often more
rapidly than the morphologic response can accommo-
date. As a consequence, the nearshore beach and surf
zone topography is endlessly destroyed and rebuilt ac-
cording to a complex cycle, as the nearshore wave
regime and circulation pattern alternate between fair-
weather and storm configurations, and on a larger scale
between the summer season of infrequent storms and
the winter season of frequent storms (Davis and Fox,
1972); see Fig. 2.
FAIRWEATHER PHASE! BEACH AND BAR BUILDING. The
cycle is controlled by two mechanisms: the wave regime
and the net circulation pattern driven by it. During fair
weather, waves tend to be far-traveled swells, of low
amplitude and long period. The asymmetry of associated
bottom wave surge is marked, with the landward stroke
beneath the wave crest being significantly more pro-
longed and more intense than the seaward stroke be-
neath the trough (Chapter 8, Fig. 8). Peak orbital
velocities may be separated by periods of 8 seconds or
on windward coasts, markedly longer. These same fair-
weather swells tend to result in a relatively weak near-
shore circulation pattern. Momentum flux, which is a
function of wave height, is relatively low during fair
weather both seaward and landward of the breaker,
hence discharge through the littoral circulation cells is
relatively low.
During fair weather, these two mechanisms, bottom
wave surge and the littoral circulation pattern, cooperate
to store sand in the beach prism. The wave regime ap-
pears to serve as a fractionating mill, dividing the avail-
able sand into a fraction undergoing mainly bed load
transport, and a fraction undergoing mainly suspensive
469
(a)
WAVE DRIFT
RETURN FLOW NET FLOW
LITTORAL
CURRENT
RIP CURRENT
FIGURE 2. Comparison of (a) fair-weather and (b) storm hydraulic regimes. Based on
Longuet-Higgins (1953), Schiffman (1965), and Ingle (1966).
470
257
258
COASTAL SEDIMENTATION
transport. Sand coarser than a critical size threshold
will be driven landward as bed load, by the landward
asymmetry of bottom wave surge, toward the breakpoint.
Longuet-Higgins (1953) and Russell and Osorio (1958)
have undertaken calculations and experiments to
determine the shore-normal components of flow averaged
over a wave cycle, in the nearshore zone of shoaling
waves. Their results indicate an increase in net land-
ward flow with increasing height off the bottom, as a
result of the failure of wave orbitals to close. Super-
imposed on this is a mid-depth return flow resulting in a
three-layer flow system (Fig. 2). It is not entirely clear,
however, if the latter component of flow would exist
in nature as a response to wave setup, or if it was merely
a wave tank artifact, induced by the continuity require-
ments of a closed system.
At the breakpoint, much of the energy heretofore
available to drive sand landward over the bottom is lost
to turbulence, and sand tends to accumulate as a break-
point bar. Waves of oscillation turn into waves of trans-
lation (bores), in which water moves forward as a mass,
and there is some evidence to indicate that landward of
the breakpoint the velocity cross section averaged over
the wave cycle changes to a two-layer system (Schiffman,
1965); see Fig. 2. An upper layer moves landward as a
series of bores, and tends to be compensated by a basal
return flow. This two-layer system is of course super-
imposed on the generally much stronger coast-parallel
flow characteristic of the longshore trough. To the ex-
tent that the two-layer flow prevails, the bar crest is a
zone of flow convergence, and its sand storage capability
is readily understood. The bar builds upward until the
rate of deposition of sand at the conclusion of wave
breaking is equaled by resuspension during wave break-
ing. Depth of water over the bar at equilibrium is gen-
erally a third of the water depth prior to formation of
the bar (Shepard, 1950).
Breakpoint bars tend to orient themselves normal to
wave orthogonals. When deep-water orthogonals make
a high angle to the shore, wave refraction does not fully
eliminate this angle near the beach. Under these condi-
tions, the bar tends to consist of series of en echelon seg-
ments, each aligned obliquely with respect to the beach,
and alternating with rip current channels (Sonu et al.,
1967).
Bar position is very sensitive to wave height, as this
determines breakpoint position (Keulegan, 1948). If the
tide range is appreciable, bar position will shift detect-
ably through the tidal cycle. New bars tend to form
during the peak or waning phases of a storm and to be
slowly driven onshore as waves diminish during the en-
suing fair-weather period, although an abrupt decrease
in wave height may cause a second bar to form landward
of the first. During the period of landward migration of
the bar, coarser bed load sand may bypass the bar and
move onto the beach, if the waves are sufficiently long
in period to re-form after breaking (King and Williams,
1949). Such bypassed sand will tend to accumulate as a
swash bar (intertidal bar), or the plunge point bar itself
will tend to migrate landward to the point where it is
captured by intertidal processes, and becomes a swash
bar (Fig. 3). As noted by King (1972), a swash bar may
only form when the beach slope is lower than the maxi-
mum potential slope permitted by the grain size of the
available sand; swash bars thus comprise attempts by
the regime of wave swash and backwash to build to this
ideal beach profile. Unlike plunge-point bars which are
formed at a bottom current convergence, swash bars are
formed by an abrupt bottom current deceleration. Their
seaward slopes are swash current graded, but the land-
ward slopes are lower than the angle of repose, and have
the same net landward sense of sand transport.
Swash bars are the dominant bar on fine, flat beaches
such as those of the central Gulf of Mexico, where the
wave climate is mild and the supply of fine sand is
abundant. They also tend to form on beaches with a
high tidal range, where the bar is exposed to swash and
backwash throughout much of the tidal cycle (ridge and
runnel systems).
The landward movement of coarser fine sand during
fair weather on open beaches may thus proceed as a
sheet flow bypassing the bar, or migration of the bar up
the beach, or more commonly as both. The result of this
landward flux of sand is the formation of the beach
prism of gently inclined sand strata, differentiated into
the backshore beach (constructional upper surface sub-
ject to eolian action) and foreshore beach (swash-graded
forward surface) separated by the berm (Fig. 1). If swash
bar migration is the dominant mode of beach aggrada-
tion, then the berm will prograde seaward mainly by
the welding to it of successive swash bars, and the inter-
nal structure of the beach prism will consist of seaward-
dipping cross-strata sets, whose internal structures dip
more steeply landward (Davis et al., 1972).
The ease with which breakpoint and swash bars can
be constructed in wave tanks strongly suggests that these
are indeed basic genetic types of bars. These two rela-
tively simple types belong to a broad class of bed forms
that arise in response to the mutual interaction of flow
with the substrate. However, it has recently become ap-
parent that much more elaborate patterns of bars may
form more or less passively, in response to an innate
pattern within the velocity field. Crescentic bars that
form in response to standing edge wave patterns have
been described in Chapter 13. On gently inclined shore-
faces, shore-parallel bars may form in arrays of up to 30
471
ONSHORE-OFFSHORE SEDIMENT TRANSPORT
259
FIGURE 3. Sequence of maps showing bar migration and erosion at South
Beach, Oregon. Bars form below mean sea level and advance up beach at rate of
1 to 5 ml day. Under the influence of strong, southward-flowing currents they
migrate southward at 10 to I5m/day. When they reach midtide level, they become
stationary, or welded to the beach. From Fox and Davis (1974)-
units. Bowen (personal communication) has suggested
that such multiple bar systems may form in response to
standing waves generated by the partial reflection of low-
amplitude, long-period (1-2 minutes) incident waves.
Such complex bar patterns clearly amplify the fair-
weather storage capacity of the surf zone.
As noted above, the fair-weather littoral hydraulic
regime is a fractionating mill, which splits the available
sand into bed load and suspended fractions. The be-
havior of the bed load fraction has been traced above.
Sand thrown into suspension at the breakpoint and fine
enough to stay in suspension in the turbulent surf zone
will tend to be fluxed alongshore by the longshore flow
in the surf zone, and out through a rip channel to rain
out on the shoreface (Cook, 196*_,..
storm phase: beach and bar destruction. During
a storm, the wave regime and the littoral circulation
patterns cooperate to withdraw littoral sand stored dur-
ing the preceding fair-weather period. Wave steepness
(ratio of wave height to wavelength) increases beyond a
critical value (Johnson, 1949), at which point bottom
wave surge asymmetry is no longer efficient in driving
coarser sand landward as bed load. Waves during storms
are locally generated, and they tend to be shorter in
period and higher (more energetic) with higher maxi-
mum orbital velocities. More sand is thrown into sus-
pension and the critical grain-size threshold between
suspensive and tractive sand fractions is shifted to favor
suspension. Suspension is more nearly continuous. At the
same time, discharge through the littoral circulation cells
is increased manyfold.
During the advent of a severe storm the sudden sea-
ward shift in breaker position, plus the great intensifi-
cation of seaward sand transport, may be sufficient to
destroy the bar and beach prism altogether. Some sand
is driven across the back beach and over the dunes in
the form of a washover fan (if this area is low enough
to be so flooded), but most is transported seaward
through rip channels and in rip current plumes. Toward
the end of the storm, fallout from rip currents accumu-
lates as a series of coalescing aprons of sand on the
shoreface. Lagoons that are flooded during the period
of rising storm surge may cut new inlets and break out
472
260
COASTAL SEDIMENTATION
through their barrier islands. The associated sand-laden
jets may greatly add to this shoreface fallout (Hayes,
1967). As the storm wanes, the bar re-forms, and the
cycle begins anew.
The cyclic nature of sand storage on beaches has been
quantitatively assessed by Sonu and Van Beek (1971) in
a study of northern North Carolina beaches (Fig. 4).
They observed a sequence wherein a storm-degraded
concave beach profile, representing minimum storage,
passed by means of swash bar accretion to a. convex
profile of maximum storage, during a four-month pe-
riod. They noted that the sense of sedimentation (ero-
sion or accretion) was more strongly correlated with the
direction of wave approach (and hence with wind direc-
(a)
^ — -<■
>v >•
c \ c
I t
I
y
->■ Accretion
>- Erosion
@ 5/15-5/22
Sea level
0 10 20 30 40 50
Horizontal scale (meters)
FIGURE 4. (a) Characteristic sequences of beach profile (d) Time history of sand storage. From Sonu and Van Beek
change, (b) Observed sequences, (c) Observed sequences as a (1971).
function of sediment storage (Q) and beach width (S).
473
THE SHOREFACE PROFILE
261
(c)
50
40
30
n Ay,
-*- Accretion
-^. Erosion
30
40
Beach width, S (m)
50
A *
Wave height
K
A
A
\ */
A A
\w
WVi
^"\ rV
l\
V/v
L, . _L
i i
i i
i i
w
1 1
1 1
10
20 10
20
10 20
10 20
10 20
10 20
Dec,
1963 Jan.,
1964
Feb.
Mar.
Apr.
May
FIGURE 4.
(continued)
tion) than with the wave steepness. Waves arriving from
the northeast tended to cause erosion. These were asso-
ciated with strong onshore winds and probably a wind-
driven bottom return flow. It appears that during pe-
riods of strong alongshore or onshore winds, the system
of littoral sand transport is no longer a closed system
but discharges its sand into the wind-driven flow of the
adjacent shelf floor. This deeper, intermittent system of
transport is described in the following sections.
THE SHOREFACE PROFILE
Hydraulic Climate of the Shoreface
Far less is known about the circulation patterns of the
shoreface and inner shelf than is known about the cir-
culation patterns of the surf zone. Classical coastal
workers, long preoccupied by the surf, have been indif-
ferent to this topic, as have been physical oceanographers,
whose habit has long been to hurry in their ships across
the inner shelf, to the intellectual challenges of the large-
scale planetary flows of the deep ocean basins. This
situation is being rapidly reversed in view of rising public
concern over the coastal environment (see Chapter 2),
but old mental sets still linger.
The shoreface and inner shelf are a zone of transition,
where the wave climate is still a major factor in shaping
the seafloor, but where the shelf flow field is becoming
of increasing significance in a seaward direction. There
is some justice in the indifference of classical coastal
workers to this hydraulic province. During periods of
fair weather, the shelf flow on most coasts may be many
474
262
COASTAL SEDIMENTATION
times less intense than littoral drift (Fig. 2A). Its veloci-
ties, on the order of 1 to 10 cm/sec, are capable of mov-
ing whatever fines happen to be in suspension, but are
not significant transporters of sand, although sand is re-
peatedly suspended by bottom wave surge at the crests
of wave-generated ripples. Fair-weather flows, however,
may be relatively complex in pattern, with nearshore
reversals of the open shelf flow, induced by coastal prom-
ontories and by interaction with the tidal streams of
inlets and estuary mouths.
Two kinds of inner shelf flows are quite significant in
transporting sand and in molding coastal topography.
On coasts with high tidal ranges, midtide current veloci-
ties associated with the passage of the coastal tidal wave
may exceed 2 knots and locally attain 4 knots a few
hundred meters seaward of the surf. Enormous volumes
of sand are shifted on each tidal cycle, with significant
net transport in the direction of the residual tidal cur-
rent. Coastal tidal flows are poorly understood and tend
to be rather complex because of strong interactions be-
tween tide-built topography and the tidal flow. Some
examples are discussed in later sections (see pp. 294-295).
Intense coastal flows may also develop during storms
(see Chapter 4). Such flows are far more infrequent than
semidiurnal tidal currents, but unlike the latter, they
occur on every coast, whether or not strong tidal cur-
©
I
I
r^
HIGH
PRESSURE
>\. | •* GRADIENT
<^j FORCE
CORIOLIS \^
FORCE ^*%
\
\
PRESSURE GRADIENT
LOW
FIGURE 5. Geostrophic flow on the continental shelf.
(A) Parcel of water at a reference depth moves seaward in
response to pressure gradient force. As it accelerates, it
experiences a Coriolis force impelling it to the right of its
trajectory. Eventually trajectory parallels isobars, and
pressure force and Coriolis force balance. (B) Cross section
of hypothetical shelf experiencing geostrophic flow: illustrating
REFERENCE DEPTH
RESULTANT
FORCE .-',
PRESSURE
^ FORCE
>
HIGH
relationship of sea surface slope, isobaric surfaces, and
reference depth. (C) Relationship between geostrophic flow
and flow in bottom boundary layer. In latter case, a friction
term enters the equation of motion, and the balance of forces
occurs among a friction term, a pressure term, and a Coriolis
term. Modified from Strahler (1963).
475
THE SHOREFACE PROFILE
263
rents occur. They, too, are significant transporters of
sand. Without these storm-driven flows, the coasts of our
planet would have a markedly different appearance.
Storms, whether of tropical or extratropicalorigin, are
rapidly moving counterclockwise wind systems that may
be a thousand or more kilometers in lateral extent. Winds
intensify rapidly toward the storm center, and in hurri-
canes, by definition, exceed 74 mph. The extent to which
the shelf water column will couple with storm winds
depends on the trajectory of the storm with respect to
the geometry of the shelf. Sustained regional coupling
of water flow with wind flow appears to occur when the
winds blow equatorward along the length of eastward-
facing coasts (Beardsley and Butman, 1974) or blow
poleward along the length of westward-facing coasts
(Smith and Hopkins, 1972). Under such conditions,
water in the surface layer will be transported landward
as a consequence of the Coriolis effect. Coastal sea level
will rise until the coastal pressure head balances bottom
friction, and bottom water can flow seaward as rapidly
as surface water flows landward. Beardsley and Butman
report up to 100 cm of coastal setup under such condi-
tions. Since the sea surface is inclined against the coast,
there is a gradient of seaward-decreasing pressure at any
reference depth. A parcel of water, accelerated by the
pressure force, has its trajectory steadily deflected to the
right by the Coriolis "force," until finally, it is flowing
along the isobars and the pressure and Coriolis terms
balance (Fig. 5).
inner shelf velocity field. The complex velocity
structure of the coastal zone is best approached in terms
of the interaction of three major flow strata (Ekman,
1905; see Neumann and Pierson, 1966, p. 202). These
are an upper boundary layer, a core flow, and a lower
boundary layer (Fig. 6). The reader is advised to review
Chapters 3 and 4 for a better understanding of this
section.
The upper velocity boundary layer experiences strong
wave orbital motion and, much of the time, a vertical
velocity gradient imposed on it by wind stress. When
the surface boundary layer is fully developed, surface
water tends to move at 45° to the right of surface wind
as a consequence of the Coriolis effect. Each successive
lower layer moves at slower speed than the one above it,
and is deviated successively further to the right (Ekman
spiral). Net flow averaged over the depth of the layer
trends 90° to the right of the surface wind. Above a
critical Reynolds number this Ekman velocity structure
becomes unstable, and is overprinted by a more com-
plex structure, in which zones of upwelling and down-
welling alternate, forming a pattern of horizontal helical
vortices aligned parallel to or at a small angle to the
zor
WAVE DRIVEN ZQNE QF
OW FRICTION-DOMINATED
FLOW
ZONE OF
GEOSTROPHIC FLOW
UPPER BOUNDARY LAYER
LOWER BOUNDARY LAYER
FIGURE 6. Velocity structure of the shore/ace and inner shelf.
(A) General form of velocity profiles through the upper boundary
layer, core flow, and lower boundary layer, and relative values of
eddy viscosity. (B) V elocity structure during a period of relatively
mild flow. (C) Velocity structure during peak flow.
mean flow direction (Langmuir circulation: Langmuir,
1925). The transition tends to occur at surface wind
speeds of 10 km (Assaf et al., 1971). The coefficient of
eddy diffusion A, is relatively large in the surface layer
as a consequence of wave-generated turbulence (Fig. 6);
it must undergo an abrupt increase at the onset of
Langmuir circulation.
Below the base of the layer, core flow extends, un-
modified, down to the bottom boundary layer. In the
core, water flows in slablike fashion, with little vertical
shear. Core flows are generally geostrophic in the sense
that in the equation of motion, the pressure term is pri-
marily balanced by the Coriolis term (Fig. 5). However,
a steady state geostrophic balance is rarely maintained
for any length of time. The shelf pressure field is in a
state of continual change, in response to the passage of
the diurnal tidal wave and to the passage of weather
systems. As the pressure field builds up and then decays,
the flow must accelerate and decelerate in sympathy,
476
264
:OASTAL SEDIMENTATION
constantly changing direction so that the pressure and
Coriolis terms may balance. Such time-dependent flows
are referred to as rotary tidal currents if mainly tide-
forced, or inertial currents if mainly wind-forced.
The character of the bottom velocity boundary layer
differs fundamentally from its surface analog. The sur-
face boundary layer is externally forced, by the wind.
Its velocity gradient, wave surge, and secondary flow
patterns are overprinted on the core flow, and are car-
ried along with it. The bottom velocity boundary layer
is caused by frictional retardation of the core flow as it
shears over the motionless substrate. Its lowermost meter
exhibits a logarithmic velocity profile (Chapter 7), but
the lower boundary layer as a whole is a thicker stratum,
characterized by a velocity profile that is a reverse Ek-
man spiral. Frictional retardation of flow results in a
deviation of boundary flow direction to the left of core
flow, so that Coriolis and frictional terms may together
balance the pressure term (Fig. 5C). The lowest layers,
experiencing the greatest retardation, are deviated the
furthest. Theoretical studies (Faller, 1963; Faller and
Kaylor, 1 966) suggest that this layer is also subject to
helical flow structure above a critical Reynolds number.
However, no field studies of this phenomenon have been
undertaken. Such innate flow stabilities, and also turbu-
lence induced by bottom roughness elements, would
lead to an eddy coefficient larger than that of the core
flow (Fig. 6A).
Three hydraulic provinces may be defined on the
inner shelf on the basis of flow structure (Figs. 65, C).
Near the beach, the two boundary layers of the shelf
flow field must completely overlap. In this zone the
effects of the regional pressure gradient on water be-
havior are largely damped out as a consequence of fric-
tional retardation. Oscillatory wave surge is the domi-
nant water motion, giving rise to the complex nearshore
circulation pattern described in the preceding chapter.
A little farther seaward, the two boundary layers are
more or less separate, but still occupy most of the water
column. Flow is frictionally dominated; in the equation
of motion the wind stress is largely balanced by friction.
The effect of the Coriolis term is negligible in shallow
water and there is little or no deviation of boundary
flow with respect to core flow. The flow is Couette-like,
in that there is a more or less linear velocity gradient
from top to bottom. Still further seaward, the two
boundary layers diverge significantly. The geostrophic
core flow dominates the water column.
This pattern of coastal flow zonation must vary with
the intensity of the regional and local wind fields. An
intensified regional wind will accelerate core flow and
increase the thickness of the bottom boundary layer.
Intensification of the local wind field will cause the
upper boundary layer to thicken, though not necessarily
at the same rate. The intensification of local wind may
either lead or lag the intensification of wind on the
adjacent shelf, depending on the trajectory of the weather
system.
The net effect of a storm is to expand the width of the
coastal flow zones and to displace the outer two zones
seaward. There are few data available for such situa-
tions (see Chapter 4). From theoretical considerations,
it appears that the upper and lower boundary layers
may overlap far out on the shelf. Zonation becomes pri-
marily a function of depth (Fig. 5C). In the zone of
friction-dominated flow, the water accelerates in response
to direct wind stress until the stress is balanced entirely
by friction; the Coriolis term is not significant, and flow
in this zone may take on the dimensions of a coastal jet
(Csanady and Scott, 1974). The zone of friction-domi-
nated flow will be a downwelling zone if local winds
have an onshore component, or if regional coast-parallel
winds result in onshore surface transport. It will be an
upwelling zone if the reverse situation prevails (Cook
and Gorsline, 1972).
The deeper, offshore flow may retain a primarily geo-
strophic balance of forces during a storm, although the
friction term is necessarily more prominent. If overlap
of the boundary layers extends through this zone, it is
theoretically possible (Faller, 1971) that there be top to
bottom overturn as a consequence of Ekman instability,
with high-velocity, wind-driven surface water delivered
to the seafloor in zones of downwelling.
The velocity structure of the shelf water mass follows
a seasonal cycle that is coupled to the cycle of density
stratification. During the summer, this upper velocity
boundary layer is the same as the upper mixed layer.
Wave turbulence and Langmuir circulation maintain
the layer's mixed character, while the pycnocline tends
to decouple upper boundary flow from core flow. During
the fall, the thermal contrast is weakened by surface
cooling. The increasing frequency and severity of storms
cause steady erosion of the lower, stratified portion of
the water column by Langmuir circulation (Faller, 1971)
and the upper mixed layer thickens at the expense of
the stratified water below. Meanwhile, a lower mixed
layer may be induced by intensified turbulence in the
bottom boundary layer, and may thicken until the den-
sity structure has simplified to a two-layer system (Char-
nell and Hansen, 1974). Further vigorous storm action
will drive the weakening pycnocline down to the sea-
floor, so that there is no further impediment to top-to-
bottom overturn by secondary flow components.
Sedimentation on the Upper Shoreface
The shoreface slope, with its gradient of seaward-de-
creasing grain size, occurs primarily in the zone of wave-
477
THE SHOREFACE PROFILE
265
HSEDIMENT INPUT
WAVE CLIMATE
-Hgrain SIZE
^T
DEPTH AS A FUNCTION
OF DISTANCE FROM —
SHORE
FIGURE 7. Relationships of variables controlling slope of the shoreface.
driven flow (Fig. 6) although its lower portion tends to
extend into and be modified by the zone of friction-
dominated flow. The slope and grain-size gradient of
the shoreface have been generally considered to com-
prise a response to the regime of shoaling waves seaward
of the breakpoint, in which depth as a function of dis-
tance from shore is itself a function of littoral wave power,
sediment discharge, and grain size (Fenneman, 1902;
Johnson, 1919, p. 211; Johnson and Eagleson, 1966;
Price, 1954; Wright and Coleman, 1972; see Fig. 7.
Johnson (1919, p. 211) has described this equilibrium
relationship as follows:
The subaqueous profile is steepest near land where the debris
is coarsest and most abundant; and progressively more gentle
further seaward where the debris has been ground finer and
reduced in volume by the removal of the part in suspension.
At every point, the slope is precisely of the steepness required
to enable the amount of wave energy there developed to dis-
pose of the volume and size of debris there in transit.
The main line of inquiry into the forces maintaining
the shoreface profile has led to the null-line hypothesis,
evaluated in Chapter 8. The hypothesis has been ex-
pressed in its most complete form by Johnson and Eagle-
son (1966). It envisages shoreface dynamics in terms of
a Newtonian balance of forces experienced by a sand
particle on the shoreface, in which the downslope com-
ponent of gravitation is opposed by the net fluid force
averaged over a wave cycle. Since in shallow water,
bottom orbital velocities are asymmetrical, with stronger
landward surge (Chapter 8, Fig. 8), fluid forces are
directed upslope. The gravitational force becomes more
intense as the shoreline is approached and the slope
increases. However, the fluid force increases yet more
rapidly. As a consequence, for a given grain size there
should be a null isobath, seaward of which particles of the
critical size tend to move downslope, and landward of
which they tend to move upslope. The equilibrium grain
size should decrease with increasing depth. Hence, the
shoreface sand sheet should tend to become finer down-
slope, as indeed it does. The shoreface slope at each point
should be uniquely determined by the grain size of sub-
strate and the intensity of bottom wave surge.
However, attempts to utilize null theory in the field
have met with ambiguous or negative results (Miller
and Zeigler, 1958, 1964; Harrison and Alamo, 1964).
Objections include: (1) slopes are not sufficiently steep
over much of the shoreface (Zenkovitch, 1967, p. 120),
and (2) slope sorting by waves tends to be overwhelmed
by other processes, which as the authors of the theory
admit, are not accounted for in null theory. No account,
for instance, has been taken of the process of ripple
sorting as described in Chapter 8 (p. 117). Wells (1967)
has shown that divergence of onshore-offshore transport
of a given grain size from its null isobath should occur
as an innate response to higher order wave interactions,
without regard to the gravitational force acting on the
grains.
A perhaps more telling criticism of null-line theory is
that a significant portion of shoreface sand travels not
as bed load, but in suspension. Murray (1967) has per-
formed tracer studies that indicate that on the upper
shoreface, the dispersal of sand corresponds to the pre-
diction of diffusion theory. Field observations by Cook
and Gorsline (1972) have led them to conclude that the
seaward-fining grain-size gradients of the shoreface are
more likely to be caused by rip current fallout rather
than by the null-line mechanism.
It may be more fruitful to approach the problem of
shoreface maintenance from the point of view of ener-
getics, rather than from the point of view of a balance
of forces. Such an approach would view the depth at
each point of a shoreface profile as a function of wave
power at that point. The ideal wave-graded profile would
be one that experiences at each point a maximum bot-
tom orbital velocity equivalent to the threshold velocity
of the size class of available sand. It should be possible
to construct an algorithm for calculating water depth
as a function of wave characteristics and bottom sedi-
ment grain size, based on the equations for bottom or-
bital velocity, for friction energy loss to the bottom, and
for the shoaling transformations of waveform that have
been presented in Chapter 6.
Lower Shoreface Sedimentation : Onshore-Offshore Sand
Budget
It seems doubtful that such a model for maintenance of
the shoreface profile by the wave regime would be suf-
ficient to fully account for the distribution of slopes and
478
TRANSECTS
DEPTH
6? 20
20
20
Li
ULlLL
UULV
12
14
40
4 0
BERM
BREAKER
UPPER
SHORE
FACE
LOWER
SHORE
FACE
DIAMETER
75°30
75°00'
DEPTH
(M)
40
20
0
60
40
20
0
40
_ 20
55 o
i- 20
= 203'05
EC
l90
5 o
2ogns
0
112.5
20
0
1160
20
0
40320 ° |
2sl-A
moo
2S~
DEPTH
(M)
i3.0
BEACH
BREAKER
PPER
RE
FACE
q6.o I
■ UPPE
1 iMm )SHO
nS.O
J.
-,10 0
320.0
OWER
SHORE
FACE
DIAMETER
(H
FIGURE 8. Distribution of grain sizes on retrograding coasts.
(A) TAe storm-dominated coast of Virginia-northern North
Ijmuiden
^ Hoek van Holland
Carolina. Data from Swift et al. (1971). (B) Dutch coast. Data
from Van Straaten (1965).
266
479
THE SHOREFACE PROFILE
267
grain sizes associated with observed shoreface profiles.
For instance, modern coasts whose historical records in-
dicate that they are undergoing erosional retreat tend
to consist of two distinctive grain provinces. From the
breaker to a depth of about 10 m, the upper shoreface
consists of fine, seaward-fining sand (Fig. 8). Seaward
of 10 m, grain size on the lower shoreface and adjacent
shelf floor is far more variable and generally markedly
coarser.
We may account for the fine, upper shoreface sand
province as a mantle of rip current fallout, whose slope
is adjusted by the regime of shoaling waves (Cook,
1969). However, the lower shoreface province of coarse
variable sand does not fit the model for wave mainte-
nance of the shoreface. We may consider the hypothesis
that it is instead a response to the deeper, intermittent
high-intensity flows of the zone of friction-dominated
flow (Figs. 2B and 6B, C).
Observations by Moody (1964, pp. 142-154) on the
erosional retreat of the Delaware coast lend some sup-
port to this hypothesis (Fig. 9). In this area, the shore-
face steepens over a period of years toward the ideal
wave-graded profile, during which time the shoreline
remains relatively stable. The steepening is both a de-
positional and erosional process. Moody notes that steep-
ening was accelerated after 1934 because a groin system
initiated then "presumably trapped sand, causing the
upper part of the barrier between mean low water and
— 3 m to build seaward" (Moody, 1964, p. 142). How-
ever, erosion continued offshore at depths of 6 or 7 m
below mean low water. The steepening process is not
continuous, but varies with the frequency of storms and
duration of intervening fair-weather periods. The slope
of the barrier steepened from 1:40 to 1:25 between
1929 and 1954, but erosion on the upper barrier face
between 1954 and 1961 regraded the slope to 1 :40.
The steepening process is terminated by a major storm,
during which time the gradient is reduced and a signifi-
cant landward translation of the shoreline occurs. Moody
( 1 964, p. 1 99) describes the Great Ash Wednesday Storm
of 1962, bracketed within his time series, as having
stalled for 72 hours off the central Atlantic coast. Its
storm surge raised the surf into the dunes for six suc-
cessive high tides. The shoreline receded 18 to 75 m
during the storm. While much of the sand was trans-
ported over the barrier to build washover fans over 1 m
thick, much more was swept back onto the seafloor by
large rip currents and by the storm-driven seaward-
trending bottom flow of the shoreface (Moody, 1964,
p. 114); see Fig. 2B.
Moody's observations allow us to present a general
model of shoreface maintenance, in terms of the on-
shore-offshore sediment budget. There seems to be little
MEAN LOW WATER
V.N
« v.
1645
KILOMETERS
FIGURE 9. Retreat of the Delaware coast, based on U.S. Coast
and Geodetic Survey records and a survey by Moody. From Moody
(1964).
reason to doubt the applicability of the conventional
wave-grading model to the upper shoreface, even if we
cannot yet present this model in a quantitative manner.
Upper shoreface textures and slopes are time-averaged
responses to two opposing mechanisms, the seaward flux
of suspended sand in rip currents on one hand, and the
landward creep of bottom sand in response to the net
landward sense of bottom wave surge on the other hand.
The upper shoreface profile varies in cyclic fashion, with
storage of sand mainly on the beach during the fair-
weather summer season, and storage of sand mainly on
the upper shoreface during the winter. For long periods
of time, the upper shoreface profile may oscillate about
the ideal wave-graded configuration.
During major storms, however, the upper shoreface
system of wave-driven longshore sand flux interacts with
coastal boundary of the storm flow field. Sand eroded
from the beach and bar by storm waves passes seaward
480
268
COASTAL SEDIMENTATION
in intensified rip currents to the zone of friction-domi-
nated flow (Fig. 2B), which during storms may take the
form of a downwelling coastal jet. When this occurs,
bottom flow on the lower shoreface will have a seaward
component of flow, and the coastal sand transport system
is transformed from a closed system of net sand storage
to an open system of net sand loss. Sand raining out of
rip currents will not come to rest, but will be transported
obliquely seaward. If the storm is severe enough, the
mantle of rip current fallout that accumulated during
the preceding fair-weather period will be stripped off,
and the underlying strata will be exposed to erosion.
This hypothetical scheme has not yet been adequately
tested by field observations. However, as a hypothesis,
it has a number of advantages. It provides a rationale
for the Bruun model of erosional shoreface retreat (Fig.
10). Bruun (1962; see also Schwartz, 1965, 1967, 1968)
noted the characteristic exponential curve of the inner
shelf profile, and accepted the hypothesis that it consti-
tuted an equilibrium response to the hydraulic climate.
With this premise adopted, it follows that a rise in sea
level must result in a landward and upward translation
of the profile, as long as coastwise imports of sand into
the coastal sector under study are equaled by coastwise
exports. The translation necessitates shoreface erosion
and provides a sink for the debris thus generated be-
neath the rising seaward limb of the profile.
Moody's time series shows that over a 32 year period,
shoreface erosion on the Delmarva coast was in fact
nearly compensated by aggradation on the seafloor in
accordance with the Bruun principle (Table 1). The
small deficit is probably attributable to loss to washover
fans, and through littoral drift to nearby Cape Henlopen
spit.
Moody's studies provide us with insight into the proc-
esses governing the Bruun model. His observations indi-
cate that the process of erosional retreat of the shoreface
is not continuous. It is cyclic in a manner analogous to
the annual cycle of the upper shoreface profile, but the
period is related to the frequency of exceptional storms,
and is on the order of years.
The model also provides a more detailed and satis-
factory explanation for the origin of the surficial sands
of shelves undergoing transgression than does the relict-
Recent sediment model of Emery (1968). The surficial
sand sheet of the shelf is a lag deposit created during
the process of erosional shoreface retreat by the seaward
transfer of sand during storms and its deposition on the
adjacent shelf floor (Fig. \0A). The nearshore modern
sands of the upper shoreface are a transient veneer of
rip current fallout. Both textural provinces are "modern"
in the sense of being adjusted to the prevailing hydraulic
regime; both are "relict" in the sense of being derived
TABLE 1. Sediment Budget from the Delmarva Coast
Sediment Source
Period
Average Volumetric
Change* (m3/year)
Barrier
(mean low water to toe
of sand barrier)
Sand dunes
(mean low water to top
of sand dunes)
Offshore erosion
(principally on north-
west side of ridges)
Erosion from bay inside
Indian River Inlet
1929-1961
148,000
1954-1961 -100,000 (estimated)
1919-1961
100,000
-69,000
Total erosion -417,000
1939-1961 +120,000
1939-1961 +5,700
Site of Deposition
Tidal delta
Barrier
south of Indian River
Inlet
Offshore accretion 1919-1961 +256,000
Total accretion +381,700
Total erosion -417,000
Total accretion +318,700
Net erosion — 98, 300 m3/year
Source. From Moody (1964).
* "+" indicates accretion; "— " indicates erosion.
from the underlying substrate. The role of shoreface re-
treat in generating shelf sediments is explored further in
Chapter 15.
Deposits of the Coastal Profile: Textures and Bed Forms
textures of the shoreface. The patterns of onshore
and offshore sediment transport described in the pre-
ceding sections give rise to systematic distributions of
sediment types and bed forms over the beach and shore-
face. The beach and surf zones consist of alternating
belts of finer and coarser sand, the absolute grain-size
values depending on grain sizes available to the coast
and on the hydraulic climate of the coast (Bascom, 1 95 1 ) .
The coarsest grain sizes are found on the crest of the
berm, in the axis of the longshore trough during the
erosional phase of the beach cycle, and on the crest of
the plunge point bar.
The distribution of grain sizes on retrograding shore-
faces has already been described (Fig. 8). Upper shore-
face sands tend to be fine grained to very fine grained,
and become finer in a seaward direction. The grain-size
481
ONSHORE-OFFSHORE SEDIMENT TRANSPORT
269
®
®
VECTO« RESOLUTION OF
PROFILE TRANSLATION
AGGRADING OAR
Will CAPTURE
SHORELINE
ZONE OF AGGRAOATION
U^
WASHOVER CYCLE
OF BARRIER SANDS
HIGH HITO«Al 0«IFI
DISCHARGE
FIGURE 10. Dynamic and stratigraphic models for (A) a
retrograding and (B) a prograding coast during a rise in sea
level. IJ coastwise sand imports are balanced by or are less than
coastwise sand exports, the hydraulically maintained coastal
profile must translate upward and landward by a process of
gradient is perhaps originally the result of progressive
sorting (see p. 162) operating on the suspended sand
load of rip current plumes; the underlying deposit be-
comes finer down the transport direction in the manner
of loess or volcanic ash deposits. It is perhaps secondarily
the result of adjustment to the landward increase in bot-
tom orbital velocity, and to the mechanism of ripple
sorting (p. 117).
On retrograding coasts such as the North Carolina
and Dutch coasts (Fig. 8), the lower shoreface consists
of variable but generally coarser sand. It is texturally
adjusted to the coastal boundary currents associated
with peak flow events. This material is of negligible
thickness and constitutes a residuum mantling the erod-
ing surface of the underlying older deposits. Its final
resting place appears to be the adjacent seafloor, where
it forms a discontinuous layer up to 10 m thick (Stahl
et al., 1974).
Bimodal sands tend to occur at the contact between
the two provinces, where the rip current fallout blanket
thins to a feather edge. This contact advances down-
slope during fair-weather periods of upper shoreface ag-
gradation, and retreats upslope during periods of storm
erosion of the entire shoreface.
shoreface erosion and concomitant aggradation of the adjacent
seafloor (Brunn, 1962). If coastwise sand imports exceed exports,
as is the case for deltaic coasts, then the profile must translate
seaward and upward. Based on Curray et al. (1969).
On prograding coasts, such as the western Gulf of
Mexico (Bernard and Le Blanc, 1965) or the Costa de
Nayarit (Curray et al., 1969), more sand is delivered
by littoral drift during fair weather than can be removed
by storms. The fine, seaward-fining sand of the upper
shoreface extends down to the break in slope where it
may become as fine as mud, and continues across the
shelf floor (Fig. 105).
Visher (1969) has observed size frequency distribu-
tions in the surf zone and on the shoreface (see Fig. 11).
Moss' theory may be applied to his observations (see
p. 162), but caution must be used, as Moss' theory re-
lates to quasi-steady flows, while in the coastal marine
environment a high-frequency flow oscillation due to
wave surge tends to be superimposed on a steady flow
component. Peak wave surge regularly induces the up-
per flow regime (Moss' rheologic regime), while the in-
tervening flow may consist of one of the less intense
stages.
In Fig. 12, the supercritical flows of swash and back-
wash in the intertidal zone have resulted in complex
subpopulation assemblage (lower foreshore, 1.5 ft sam-
ples). The contact (C) population, consisting primarily
of shell debris, comprises up to 10% of the total distri-
432
270
COASTAL SEDIMENTATION
SHORE - PROFILE
TIDE GOING OUT
5 FEET
" 11 FEET
■ LOWER
. FORESHORE
0 12 3
PHI SCALE
4
- 15FEET
■
-
99
I PREDOM
: SHELL
i
90
50
10
1
0 1
LOW TIDE
FOREST BEACH S. C.=
FIGURE 1 1. Representative size frequency distributions of the shoreface. From Visher (i969).
bution. A large C population is characteristic of Moss'
rheologic regime; the B population, however, is less
than 1%. While the hydraulic microclimate of rheologic
flow is conducive to the incorporation of a large B popu-
lation into the bed, such a response is presumably in-
hibited by the gross hydraulic structure of the surf zone;
suspended fines are steadily flushed seaward through rip
channels. Two framework (A) populations are locally
present, reflecting perhaps discrete responses to swash
and the slightly higher backwash velocities. Subtidal
surf zone populations (5 and 6.5 ft) are similar, but the
framework population is better sorted (corresponding
segment of the cumulative curve is steeper). This sorting
is further improved in the upper foreshore sample ( 1 1 ft
sample). This sample has a reduced contact (C) popu-
lation and an enriched interstitial (B) population. B pop-
ulation enrichment reflects the heavy rip current fallout
of fine suspended sand experienced at this depth, and
perhaps also the presence of Moss's fine ripple regime.
BED FORMS OF THE SHOREFACE. CliftOn et al. (1971)
have noted that the high-energy shoreface of southern
Oregon is characterized by zones of primary structures
that reflect the hydrodynamic subenvironment (Fig. 12).
An "inner planar facies" occurs beneath the reversing
supercritical flows of the swash zone; the associated
structure within the deposit consists of thin beds and
laminae of gently inclined sand. The rhomboid ripple
marks and antidunes that form in each backwash are
rarely preserved.
Beneath the surf zone of gently sloping beaches lies an
"inner rough facies" of shore-parallel ridges and
troughs 1 to 2 m across and 10 to 50 cm deep. The
flat-topped ridges tend to be steepest on the seaward
side, and the ridges migrate seaward. During periods
of strong littoral currents, troughs are more nearly
perpendicular to land, and migrate downcurrent and
offshore. The internal structure of this facies consists of
medium-scale (units 4-100 cm thick) seaward-dipping
trough cross-bedding.
Beneath the breakpoint lies an "outer planar facies."
No bar existed here during the period of Clifton's study.
Small ripples may form during the initiation of trough
or crest surge, but the flow becomes supercritical during
maximum surge, and the ripples are destroyed. The in-
ternal structure is horizontal lamination.
Clifton et al. describe the upper shoreface of the Ore-
gon coast as the "outer rough facies." The characteristic
bed forms are lunate megaripples, 30 to 100 cm high,
and with spans (terminology of Allen, 1968a, pp. 60-62)
between 1 and 4 m. Concave slopes face landward, and
the ripples migrate landward at rates of 30 cm/hr.
483
LONGSHORE SEDIMENT TRANSPORT
271
WAVE ACTIVITY
WAVE TYPE
ORBITAL VELOCITY:
OFFSHORE
SWELL
NEARSHORE
c
sea surface
Fe" iCO
0
2C 4C
Meters
SEA ^LOOR-
STRUCTURAL FACIES
FIGURE 12. Relationship of depositional structures to wave type and activity. From Cl ij ton etal. (1971).
ASYMMETRIC
RIPPLE
INNER
PLANAR
Crestal sands are notably coarser than trough sands. The
resulting internal structure is a medium-scale cross-strati-
fication with foresets dipping steeply landward.
The lower portion of the upper shoreface of the Oregon
coast has an "asymmetrical ripple facies" of short-crested
wave ripples 3 to 5 cm high with chord lengths (Allen,
1968a. pp. 60-62) of 10 to 20 cm. They may reverse
asymmetry with each passing crest and trough, or reveal
a persistent landward asymmetry. Crests become lower,
longer, and straighter as the pattern is traced seaward.
Interfering sets are common, weaker sets tending to
occur as ladderlike rungs in the troughs of the stronger
set. The angle between the two sets is bisected by the
wave surge direction. The internal structure of this
facies tends to consist of small-scale (less than 4 cm)
shoreward-inclined ripple cross-lamination, inter-
fingering with gently dipping, medium-scale scour
and fill units.
A similar sequence of bed forms and textural prov-
inces has been reported from the Georgia coast by
Howard and Reineck (1972). The Georgia coast has a
milder wave climate than those described above, al-
though it is also characterized by strong tidal flows.
The equilibrium configuration of the shoreface is rather
different here (Fig. 13): the slope is much gentler, and
the break between the fine sand of the upper shoreface
and the coarse sand of the lower shoreface occurs as far
seaward as 14 km from the beach.
An inner planar zone of laminated sand is equivalent
to that of Clifton et al.'s i 1971 ) but is markedly wider,
extending from 0 to — 1 m. 200 m from the beach. An
inner rough facies (1-2 m depth) is equivalent to that
of Clifton et al.'s, but is expressed as rippled, laminated
sand, rather than megaripples. An outer planar facies
(5-10 m depth) consists of laminae and thin beds with
sharp, erosional lower contacts, grading upward into
bioturbate texture. Howard and Reineck (1972) suggest
deposition during storm intervals, alternating with pe-
riods of fair weather and bioturbation. An upper shore-
face facies of fully bioturbated, muddy fine sand has no
parallel in Clifton's study of the high-energy Oregon
coast, and is a consequence of the high input of fine sand
and reduced wave energy.
Seaward of 10 m, the muddy, gently sloping shoreface
becomes markedly coarser, then gives way to a flatter
seafloor of medium to coarse sand, characterized by
heart urchin bioturbation and trough cross-stratification.
Clifton and co-workers did not extend their study suffi-
ciently far seaward to detect such a coarse lower shore-
face and seafloor facies However, an equivalent facies
does appear on retreating coasts of both North Carolina
and Holland (Fig. 8).
Reineck and Singh (1971) have described shoreface
and inner shelf deposits from the low wave energy, high
mud input, prograding coast of the Gulf of Gaeta, Italy.
The inner facies are rather similar to those of the Georgia
coast. Ripple bedding is the main sedimentary structure
out to 2 m. Below 2 m, laminated bedding becomes the
main structure, and bioturbation becomes prominent,
increasing seaward. Laminae are inferred to be deposited
from graded suspensions after storms. At 6 m, sand gives
way to silty mud, heavily bioturbated by Echnocardium
cordatum. There is no equivalent of the coarse offshore
facies of retreating coasts.
LONGSHORE SEDIMENT TRANSPORT
The seasonal cycle of onshore and offshore sand migra-
tion in the surf is superimposed on a much more intensive
flux of sand parallel to the beach, under the impetus of
the wave-driven littoral current. The mechanisms driving
484
BB-N
BB-S
CEC-2
CAB
CEC-7
STRUCTURES
Laminated Sand
Small Scale Ripples
Meganpples
Bioturboied Sand
Sand|lt„„culo, B.dd.ng
and \ f'n« Interbedd.ng
Mud Icoone Interbeddii
Mud
Shells I-
27
28
29
30
£
FIGURE 13. Primary structures of the Georgia shorejace, as their traversing the north flank of an estuary mouth shoal,
revealed by box cores. Complexity oj lines N and S are due to From Howard and Reineck (1972).
272
485
LONGSHORE SEDIMENT TRANSPORT
273
this littoral drift and equations for determining its trans-
port rate are described in Chapter 13.
The propensity of this coastwise sand flux to aggrade
or erode the shoreline can be understood by reference
to a convenient graphical model presented by May and
Tanner ( 1973). As a consequence of refraction of waves
about a coastal headland, such a headland will tend to
concentrate the wave rays on it and hence wave energy
(see p. 7b). As a consequence, the wave energy density
(proportional to the spacing ol wave rays) decreases
steadily from point a on the headland to point e in the
bay. These relationships are shown in highly schematic
fashion in Fig. 14.
The longshore component of wave power Pi. is a
function of both the wave energy density and the breaker
angle (see Equation 2, Chapter 13). It must therefore
pass through a maximum between the point of greatest
wave energy density a, and the point of greatest breaker
®
©
FIGURE 14. Model for littoral sediment transport. (A) Wave
refraction pattern, with wave approach normal to coast. (B) Re-
sulting curves for energy density at the breaker E (dimensions
AIT-2)/ longshore component of littoral wave power PL {dimen-
sions MLT~3); and the littoral discharge gradient dq/dx (dimen-
sions L2T~l). (C) Advanced state of coastal evolution. After May
and Tanner (1973).
angle at c. The maximum is attained, however, closer to
point c, as the gradient of wave energy density on this
subdued model coast is relatively flat.
Both the sand transport rate //. (see p. 247) and the
discharge of sand (q) are proportional to Pi and vary
with it. Therefore, the longshore discharge gradient dq/
d.v varies as the derivative of Pi (Fig. 14). The sediment
continuity equation (p. 190) states that the time rate of
change of seafloor elevation along a streamline in the
littoral current is proportional to the littoral discharge
gradient under conditions of steady flow. In other words,
if more sand is moving into a given section of shoreface
than is moving out (negative dq/dx), then the seafloor of
that section must aggrade. If, on the other hand, more
sand is being exported than imported (positive dq/dx),
the seafloor of that sector must erode (see the discussion
on p. 190). In general, erosion occurs along a positive
discharge gradient, and deposition occurs along a nega-
tive discharge gradient.
In the model of Fig. 14, this relationship means that
the shoulder of the headland, from a to c, should erode,
with the material being transported into the bay, to fill
sections c through e. The same sort of process should
occur on the other side of the bay (not shown) and the
other side of the headland (not shown). If the direction
of wave approach were held constant and normal to the
regional trend of the coast, then a very peculiar coast-
line should eventually result. It would straighten out to
a nearly east-west line running through c, but with a
needlelike projection at a, the point of littoral drift
divergence, and a similarly narrow indentation at e, the
point of littoral drift convergence. On real coasts, how-
ever, the direction of wave approach is not constant, but
fluctuates about the mean value, with changes occurring
on a scale of hours to days. As the direction of wave ap-
proach fluctuates, so do the positions of points a and e,
and the development of coastal reentrants and projections
is suppressed.
If waves tend to approach at an angle instead of ap-
proaching normal to shore, and if coastal relief is more
deeply embayed (Fig. 15), then a rather different dis-
tribution of longshore wave power will result. The locus
of maximum deposition will be shifted from the bay head
toward the tip of the adjacent headland where the gradi-
ents of wave energy density and breaker angle are the
steepest. During a storm when the intensity of littoral
drift discharge is the greatest, deposition at this point
may be so intense that a discontinuity in the shoreface
may occur, in the form of a spit that builds out across
the bay as an extension of the headland shoreface. As
the shoreline matures, headland retreat, spit extension,
and bay head beach progradation occur simultaneously
and in this model also, the final coastline is again straight.
486
274
COASTAL SEDIMENTATION
FIGURE 15. Variant of the littoral transport model with a
more deeply embayed coast and an oblique direction of wave
approach. Conventions as in Fig. 14.
Thus, as a consequence of the submarine refraction o
waves about the shoals off headlands, the shoreface tends
toward an equilibrium plan view as well as an equilib-
rium profile. Headlands tend to be suppressed and bays
filled, because their existence leads to longshore wave
power gradients that transfer sand from headland to bay
head.
A similar smoothing process operates at deeper levels
on the shoreface, where sand transport occurs in response
to tide- and wind-driven currents. Such flows accelerate
past the projecting headlands which impede them, and
expand and decelerate off the bays. The result is a posi-
tive discharge gradient on the upcurrent sides of the
headlands, and a negative discharge gradient on the
downcurrent side. This pattern reverses when the cur-
rents themselves reverse, so that the lower shoreface off
headlands experiences net erosion, while the lower shore-
face off bays experiences aggradation. Thus the equilib-
rium plan view of a coast tends to be straight, to the
extent that variations in the homogeneity of the sub-
strate and the rate of sand supply to the surf zone will
permit.
A second characteristic of equilibrium coastal con-
figuration is the adjustment of the trend of the coast to
the angle of wave approach as mediated by the rates
and locations that sand is put into and taken out of a
littoral drift cell. A perfectly straight and infinitely long
coast could ideally maintain any angle to wave ap-
proach, if the only source of sand were its own shoreface
erosion. In fact, however, such an ideal straight coast is
rarely attained. Sea level is rising or has been until very
recently, and the straightening process must operate con-
tinuously as successive portions of the irregular subaerial
surface are inundated. Depending on the degree of in-
duration of the coast, an effective equilibrium is attained
with less than the "climax" degree of coastal straight-
ness; some irregularity usually persists, with at least sub-
dued headlands serving as sand sources, and embayments
serving as sand sinks. Locally, river mouths may serve
as point sources of sand with high sand input rates.
These result in deltas and an effective coastal equilib-
rium at less than climax straightness.
Komar (Chapter 13, Figs. 12, 13) has provided two ex-
amples of the adjustment of coastal trend to the angle of
wave approach, and to the location of sources and sinks.
His river mouth (Fig. 12) injects sand at a point in the
littoral drift system at a rate greater than the system can
initially accommodate, and the coastline progrades. As
it does so, orientation of the coast at each point adjusts
so that the river mouth protrudes as a delta. Eventually,
an equilibrium configuration is attained so that each
point along the shoreline maintains an incident wave
angle sufficient to bypass the same amount of sand at
every other point.
Komar's beach (Fig. 13) has no point source of sand.
He starts with a straight beach and a landward-convex
waveform, a form that might arise from offshore refrac-
tion over the rocky headlands enclosing a pocket beach.
The center of the beach becomes a source and the ends
become sinks; the shoreface adjusts to the wave refrac-
tion profile.
Komar's two examples correspond to two basic cate-
gories of coastlines. In the swash alignment (Davies, 1973,
p. 123), each point on the coast tends to be oriented
normal to the direction of wave approach, either as an
initial condition or because the configuration of the
shoreface and the wave refraction pattern have inter-
acted until this is the case. Such an adjustment is only
possible if there is a projecting headland in the down-
drift direction as in the case of Komar's beach, or
another reason to cause a sand sink and allow coastal
progradation. Drift alignments (Davies, 1973, p. 123) are
more nearly like Komar's delta model. These coasts are
487
TRANSLATION OF THE SHOREFACE
275
□ "
ecent sediments
| Olde
r bedrock
FIGURE 1 6. Zetajorm bays on the offset coast of eastern Malaya.
Major wave approach direction is from the northeast. From
Davies (2973).
stabilized by competition between two opposing trends
for control of the littoral transport system. As a coastal
compartment becomes more nearly normal to wave or-
thogonals, littoral energy density increases; however, the
longshore component of wave energy decreases. Maxi-
mum discharge tends to occur when the orthogonals of
prevailing wave trains make an angle of 40 to 50° with
the coast. A coast captured in this alignment will tend
to be stable, as it is the alignment of maximum transport.
Coasts with closely spaced barriers to littoral drift, in
the form of river mouths or projecting headlands, may
form offset coasts consisting of successive zetaform bays
(Halligan, 1906). These fishhook beaches have a curved
swash-aligned beach in the shadow zone immediately
downdrift from the barrier, and a straight drift-aligned
beach extending downdrift from the swash segment
to the next headland (Fig. 16). If the variance of
the direction of wave approach is high, then the shadow
zone behind the barrier will be exposed to the direct
approach of waves for a significant part of the time
and will intermittently operate as a drift-aligned beach
of reverse drift. The apex of the barrier will become a
zone of net drift convergence, and hence a self-sustaining
constructional feature.
The dynamics of zetaform bays have been discussed
in detail, with summaries of earlier observations, by
Silvester (1974, pp. 71-90).
TRANSLATION OF THE SHOREFACE
The preceding evaluation of the alongshore and on-
shore-offshore components of sediment transport per-
mits us to take a more general look at the coastal sedi-
ment budget and its effect on the shoreface stability.
It is clear from the preceding discussion that the
shoreface profile will translate either landward or sea-
ward (the coast will retreat or prograde) depending on
whether the effects of the fair-weather regime, which
tends to aggrade the shoreface, or the storm (or tidal)
regime, which tends to erode it, are dominant. The
sense of coastal profile translation further depends on
coastwise gradient of sand discharge (whether sand im-
ports and exports for the sector under consideration sum
to a surplus or deficit). As a consequence of the onshore-
offshore cycle of sand exchange, the nature of coastal
translation will finally depend on which shoreface prov-
ince, if any, is actually subjected to a sand surplus. The
coastal transport system may be visualized as two coast-
parallel pipes, corresponding to the wave-driven littoral
drift near the beach and the intermittent storm- or tide-
driven sand flux that occurs en the shoreface and inner
shelf seaward of the breaker. These two pipes are con-
nected by valves, corresponding to the onshore-offshore
cycle of sand exc iange. The factors listed above deter-
mine what vah s are open, for how long, and the net
sense of flow through the valves. We do not have the
measurements of onshore-offshore sand transport that
would allow us to document the manner in which this
system actually works. Until we do, we must be satisfied
with an exploration of possible limiting cases, by means
of deductive reasoning (Fig. 17).
Four basic cases may be distinguished. On modern
coasts, undergoing relatively rapid sea-level rise, the
gradient of littoral drift discharge (dq/dx) is either posi-
tive or is so slightly negative that the resulting sand
surplus is not sufficient to balance offshore transport
488
276
COASTAL SEDIMENTATION
A TRANSGRESSION
( RISING SEALEVEL )
A
UPPER vj ' ' 1
SHOREFACE ^^.V ^~"SJ5^7'->
LOWER I — I
SHOREFACE ' '
INNER SHEEF
^^i
DEPOSITIONAL REGRESSION
RISING SEALEVEL )
\\
EROSIONAL REGRESSION
( FALLING SEALEVEL
SOUR
TRANSPORT SYSTEM
MODERN DEPOSITS
FIGURE 17. Schematic models for the shore face sand budget. (A) Retrograding
coastal sector with rising sea level and balance or deficit in coastwise sand flux. (B) Near-
stillstand coastal sector with effect of rising sea level compensated by sand surplus associated
with coastwise sand flux. (C) Prograding coastal sector with effect of rising sea level and
reversed by sand surplus associated with coastwise sand flux. (D) Prograding coastal
sector with falling sea level and balance in or sand deficit resulting from coastwise sand
throughput.
during storms. Under these conditions Bruun coastal
retreat must prevail (Fig. 17.4). Storm erosion must pre-
dominate over fair-weather aggradation on the shore-
face as it translates landward and upward in response
to sea-level rise. During fair weather, sand may be tem-
porarily stored on the upper shoreface and beach (see
Fig. 4A). The barrier superstructure becomes a long-
term reservoir, receiving sand from the eroding shore-
face by storm washover, storing it, and finally releasing
it to the eroding shoreface. The inner shelf floor tends
to become a sand sink, retaining the coarser sand trans-
mitted to it by seaward bottom flow during storms, and
releasing the finer fraction to the coastwise shelf flows.
The coast of northern New Jersey appears to be under-
going such a retreat (Stahl et al., 1974).
Locally, however, the sand discharge gradient associ-
ated with the deeper storm-driven shelf flows may be
steeply negative (Fig. \7B), resulting in a sand surplus.
The surplus tends to be absorbed by inner shelf and
lower shoreface aggradation and also by storm washover
on the barrier. The upper shoreface, however, is rela-
tively unaffected. The trajectory of the shoreface profile
appears to be nearly parallel to the upper shoreface
slope, resulting in stillstand of the shoreline. The
barrier system of coastal North Carolina immediately
north of Cape Hatteras appears to be undergoing
such a depositional stillstand, resulting in the opening of
an anomalously wide lagoon behind it as sea level con-
tinues to rise (Swift, 1975).
If the littoral drift transport system has a strongly
negative discharge gradient, as is the case downdrift
from river mouths with a high sand discharge, the satu-
ration of the upper shoreface with sand causes the flooding
of all other inner shelf provinces as well, and the shore-
face progrades by the successive capture of upper shore-
face bars as beach ridges (Figs. 105 and 1 1C). The analy-
sis of the Holocene history of the Costa de Nayarit by
Curray et al. (1969) provides an excellent case history
of such a coast.
Finally, we must consider the case of falling sea level.
If, under such conditions, the net littoral drift input is
negligible, not all portions of the shoreface will prograde.
The shoreface must translate seaward down the gradient
of the shelf in a reversal of the Bruun process. Under
these conditions, successive beach-upper shoreface sand
prisms undergo subaerial capture, and the lower shore-
face and inner shelf undergo erosion as sea level drops
(Fig. \1D). However, should the sand surplus due to
489
TRANSLATION OF THE SHOREFACE
277
littoral throughput increase, the sand budget must ap-
proach that of the subsiding prograding coast and a more
rapid shoreline regression must result. Modern examples
of such falling sea-level budgets are confined to regions
of glacial rebound or tectonic uplift, but the surfaces of
the world's coastal plains were molded by it during the
withdrawal of the Sangamon (Riss-Wiirm) Sea (Oaks
and Coch, 1963; Colquhoun, 1969).
A word on the factors controlling the steepness and
curvature of the shoreface profile is in order at this
point. Grain size is the most obvious control (Bascom,
1951); the coarser the sediment supplied to the coast,
the steeper are the shoreface profiles. Shorefaces built of
shingle may attain 30° slopes near the beach; shorefaces
of sand are rarely more than 10° at their steepest, while
shorefaces on muddy coasts are so flat as to be virtually
indistinguishable from the inner shelf. Sediment input
and the wave climate also affect the shape of the profile.
In general, inner shelves experiencing a higher influx of
sediment and a lower wave energy flux per unit area of
the bottom are flatter, whereas inner shelves with a
lower influx of sediment and a higher wave energy flux
per unit area of the bottom are steeper (Wright and
Coleman, 1972). Because of the complex interdependence
of the process variables, cause and effect are difhcult to
ascertain; on a steeper shelf, for instance, grain size is
coarser because the steeper slope results in more energy
being released per unit area of the bottom; more energy
is released because the coarser grain size results in a
higher effective angle of repose. Or a reduced input of
sand will allow the profile to attain the maximum steep-
ness permissible under the prevailing wave climate, with
a resultant higher rate of energy expenditure on the
shoreface, and a consequent coarsening of its surface
(Langford-Smith and Thorn, 1969).
The relationship between the rate of sea-level dis-
placement and the shape of the profile requires some
thought. A number of workers have assumed that rap-
idly translating coasts are in a state of disequilibrium,
and that equilibrium can only be realized on very
slowly translating or stillstand coasts. This view results
from an inadequate appreciation of the equilibrium con-
cept and is tantamount to stating that only chemical
reactions that have gone to completion are equilibrium
reactions.
It is important to clearly distinguish between the con-
cept of coastal maturity on one hand, and the concept
of coastal equilibrium on the other. Davis (in Johnson,
1919) has assembled a spectrum of coastal types that
suggest that the coastal profile passes through stages of
'"youth, maturity, and old age" in which the profile be-
comes increasingly flatter, until a final profile of static
equilibrium is reached — ultimate wave base, in which
the continental platform has been shaved off to a level
below which further marine erosion occurs so slowly as
to be negligible. The scheme is unrealistic in that it fails
to recognize the continuous nature and mutual depend-
ence of the process variables of an equilibrium system.
Some of these stages will occur as transient states after
the sudden rejuvenation of a tectonic coast. But as the
profile becomes increasingly mature, its rate of change
decreases, until it attains the equilibrium configuration
required by existing rates of such other process variables
as sediment input and eustatic sea-level change. At this
point the profile must continue to translate according to
the Bruun (1962) model of parallel shoreface retreat,
until the rate of one or another variable changes again.
This equilibrium, of course, is only apparent if the coastal
profile is examined over a sufficiently long period of
time — on the order of decades. Shorter periods of obser-
vation will resolve the apparent "equilibrium" into a
series of partial adjustments to periods of fair weather
and periods of storms.
Only in cases of relatively rapid tectonism may hys-
teresis, or lagged response, occur, and strictly speaking,
the term "disequilibrium" should be applied only to
such cases. Slower changes in a process variable will
allow continuous and compensating adjustment of pro-
file, and while its shape changes, the profile is at all
times in equilibrium. Coastal disequilibrium tends to be
more apparent on rocky coasts, because of the greater
response time of the indurated substrate, and because
such coasts are more likely to be subject to tectonism.
Consequently, the effect of the rate of sea-level dis-
placement in the equilibrium profile must depend on the
initial slope of the substrate. On low coasts, where the
initial slope is flatter than the maximum potential slope
of the equilibrium profile, then the more rapid the sea-
level displacement, the flatter is the resulting equilibrium
profile (e.g., see Van Straaten, 1965). This relationship
may be viewed as a function of work done on a substrate
to build the optimum shoreface. As a coast advances
more rapidly, successive shorelines experience the ero-
sive effect of shoaling waves for shorter periods of time
and the resulting profile is flat (immature). If, however,
a coast undergoes stillstand, the climax or fully mature
configuration can develop, which is the steepest profile
possible for the available grain size of sand, rate of
sediment influx, and hydraulic climate.
On high, rocky coasts, however, the initial slope of
the substrate may be steeper than the mean, or even the
maximum slope of the steepest profile permitted by these
variables. Under such circumstances, the more rapidly
transiting shorelines, since these have the least work done
on them, have the least modified and hence steepest
(most immature) profiles, while the most slowly moving
490
278
COASTAL SEDIMENTATION
shorelines are the most modified and hence flattest
profiles.
As noted above, existing measurements of the coastal
hydraulic climate and resulting sand transport are gen-
erally inadequate to define the coastal sand budget. It
is possible, however, to extend the inferential models
presented in Fig. 1 7 so as to take into account the effect
of these variables on shoreface slope and curvature
(Fig. 18).
COASTAL ENVIRONMENTS
The preceding sections of this chapter have described
the onshore-offshore component of sediment transport,
and also the flow of sediment parallel to the coast. Modes
of shoreface displacement in response to rising and fall-
ing sea level have been considered. These insights are
prerequisites to an examination of specific patterns of
coastal sedimentation. But before we proceed to such an
EROSIONAL TRANSGRESSION
(PROFILE INVARIANT)
B
DEPOSITIONAL TRANSGRESSION
(PROFILE INVARIANT)
STILLSTAND
(PROFILE CURVATURE INCREASING)
DEPOSITIONAL REGRESSION RISING SEA LEVEL
(PROFILE INVARIANT)
3
EROSIONAL REGRESSION
(PROFILE INVARIANT)
DEPOSITIONAL REGRESSION FALLING SEA LEVEL
(PROFILE CURVATURE DECREASING)
;.;.;.:•; ZONE OF DEPOSITION
ZONE OF EROSION
FIGURE 18. Modes of shore/ace translation as a function of curvature. Envelopes of .erosion and aggradation are shown.
(1) direction of profile translation and (2) change in profile Terms from Curray (1964).
491
COASTAL ENVIRONMENTS
279
BEACH DOMINATED COASTS
DRIFT COASTS
SWASH COASTS
U_/uJ R5SP
INLET & RIVERINE COASTS ROCKY COASTS
INE
If
ESTUARINE COASTS
*«* **
INLET & RIVERINE
COASTSX
DELTAIC COASTS
LOBATE DELTAS
ft
A
DELTAIC
/COASTSX
& ^
RECESSED DELTAS ESTUARINE DELTAS
FIGURE 19- Descriptive taxonomy of coasts
analysis, we should consider a scheme for classifying the
coastal settings in which the transport patterns occur.
A study of maps of the world's coastlines suggests that
the apparently unlimited variety of coastal configurations
falls into a relatively small number of repeating patterns.
Considerable thought has gone into coastal classifica-
tion, and the reader is referred to the excellent summary
of existing classifications presented by C. A. M. King
(1972; also Chapter 1 5) .The operational classification that
is used in this text is presented in Table 2 and Fig. 19.
The most basic practicable division appears to be into
coasts with substrates of crystalline or lithified sedimen-
tary rock versus coasts bordering coastal plains, with
substrates of unlithified sediment. Both lithified and un-
lithified coasts may adjust their configurations in re-
sponse to the coastal wave climates but they do so at
different rates, and in response to somewhat different
mechanisms. Patterns of sedimentation may be relatively
simple on straight rocky coasts, but coasts of structured
rock may be so deeply embayed as to greatly complicate
the pattern (crenulate rocky coasts).
Unconsolidated coasts are floored by easily eroded
and flat-lying strata, and the surficial sediment tends to
be both abundant in quantity and continuous in extent.
On such coasts, wave-driven currents in the littoral zone
and wind- or tide-driven currents farther offshore tend
to build straight coastal segments of the sediment avail-
able to them. A basic second-order division in %he mor-
phology of unconsolidated coasts depends on the relative
importance of straight coastal segments versus inlets that
alternate with them.
TABLE 2. A Coastal Taxonomy
Criterion
Coastal Type
Substrate indurated
Coast-parallel anisotropy
Coast-transverse anisotropy
Rocky coasts
Straight rocky coasts
Crenulate rocky coasts
Substrate unconsolidated
Littoral drift dominant
Low drift angle
High drift angle
Transverse drainage dominant
Fluvial drainage dominant
Mild wave climate
Moderate wave climate
Moderate-strong wave climate
Strong wave climate
Tide modified
Tidal drainage dominant
Wave-modified tidal drainage
Unconsolidated coasts
Beach and barrier coasts
Drift coasts (straight)
Swash coasts (cuspate)
Inlet and riverine coasts
Deltaic coasts
Lobate deltas
Arcuate deltas
Cuspate deltas
Recessed deltas
Estuarine deltas
Estuarine coasts
Inlet coasts
492
280
COASTAL SEDIMENTATION
Inlets occur at river mouths, where they may be main-
tained by river and tidal flow or by purely tidal flow,
where there is a tidal exchange between lagoons and
the sea. A dense subaerial drainage net or a high tide
range may cause inlets with their coast-normal flow to
occupy over 509c OI tne shoreline, resulting in deltaic,
estuarine, or tidal inlet coasts. Open coasts with rigorous
wave climates and frequent strong wind-driven currents
tend to have fewer inlets than coasts not so affected,
resulting in mainland beach-barrier beach coasts.
This chapter has so far dealt mainly with the sedi-
mentary regime of such simple, two-dimensional coasts.
The succeeding sections examine in greater detail the
modes of sand storage on beach-barrier island coasts,
and also the modes of sediment storage on more complex
coasts. The following chapter on shelf sedimentation
stresses the role of varying coastal configurations in by-
passing sediment to the continental shelf, and thus modu-
lating the shelf sedimentary regime.
SAND STORAGE IN THE SHOREFACE
Storage in Low Retreating Shorefaces: Barrier Spits and
Islands
barrier formation. On most retreating coasts, the
most important form of sand storage is within the shore-
face itself, in the form of barrier spits and barrier islands.
It would seem that along many coastal sectors, the coastal
sedimentary regime rejects the primary shoreline formed
by the intersection of the subaerial continental surface
with the sea surface, and instead builds a secondary
"barrier" shoreline seaward of the primary one. A char-
acteristic of the equilibrium shoreface surface that as
much as any mechanism is the basic "cause" of barrier
islands and spits is its innate tendency toward two-
dimensionality, its tendency to be defined by a series of
nearly identical profiles in the downdrift direction. The
equilibrium shoreface does not "want" a lateral bound-
ary, since the wave and current field to which it responds
does not generally have one. The initial conditions dur-
ing a period of coastal sedimentation may, however, in-
clude such discontinuities, as in the case of a coast of
appreciable relief (bay-headland coast) beginning trans-
gression.
On such a coast shoreface surfaces will tend to be in-
cised into the seaward margins of headlands exposed to
oceanic waves, and will propagate by constructional
means in the downdrift direction as long as material is
available with which to build, and a foundation is avail-
able to build on. The basic mechanism is that described
by May and Tanner (1973); see Fig. 14. Where the
shoreline curves landward into a bay, the longshore
component of littoral wave power decreases, and the
alongshore gradient of sediment discharge (dq/dx) is
negative. The shoreface at that point must aggrade until
the gradient approaches zero at that point, and the
zone of negative gradient has moved downdrift. We give
the lateral propagation of the shoreface into coastal
voids the descriptive term "spit building by coastwise
progradation" (Gilbert, 1890; Fisher, 1968).
However, the tendency of the shoreface to maintain
lateral continuity also acts to prevent discontinuities as
well as to seal them off after they have formed. In order
to illustrate this, we may consider another set of initial
conditions — a low coastal plain with wide, shallow val-
leys after a prolonged stillstand during which processes
of coastal straightening by headland truncation and spit
building have gone to completion. As this coastline sub-
merges, the water, seeking its own level, will invade
valleys more rapidly than headlands can be cut back.
The oceanic shoreline, however, cannot follow, for if it
should start to bulge into the flooding stream valleys,
the bulge would become a zone of negative discharge
gradient; hence the rate of sedimentation would increase
to compensate for any incipient bulge. The shoreface
would translate more nearly vertically than landward at
this sector, until continuity along the coast was restored
(Fig. ME). Thus a straight or nearly straight oceanic
shoreline must detach from an irregular inner shoreline,
and be separated from it by a lagoon of varying width.
This process of mainland beach detachment was first
proposed by McGee (1890), and later described in de-
tail by Hoyt (1967); see Fig. 20.
COASTWISE SPIT PROGRADATION VERSUS MAINLAND
beach detachment. Much of the debate concerning
origin of barriers deals with the relative importance of
spit building versus mainland beach detachment (Fisher,
1968; Hoyt, 1967, 1970; Otvos, 1970a,b); see Chapter
12 (p. 223). The problem can be fully answered only by
careful study of the field evidence, and as noted by sev-
eral authors (Otvos, 1970a,b; Pierce and Colquhoun,
at
15 £
Ui
I3oS
12 3 4
KILOMETERS
FIGURE 20. Barrier island formation by mainland beach
detachment. Modified from Hoyt (1967).
493
SAND STORAGE IN THE SHOREFACE
281
1970) the evidence has frequently been destroyed by
landward migration of the barriers. However, it is pos-
sible, in the time-honored deductive fashion of coastal
morphologists, to consider the conditions most favorable
to these two modes of barrier formation. Spits are cer-
tainly characteristic of coasts ofhigh relief undergoing
rapid transgression as described above [see the papers in
Schwartz (1973)]. It seems probable that under such
conditions mainly beach detachment would be severely
inhibited. Even allowing for ideal initial conditions with
a classic coast of old age (Fig. 21), where alluvial fans
are flush with truncated headlands, detached mainland
beaches would have a limited capability for survival.
With significant relief, the submarine valley floors ad-
jacent to retreating headlands must lie in increasingly
deeper water after the onset of transgression. As the
barrier grows into the bay, its submarine surface area
must increase, and the capacity of littoral drift to
nourish it may eventually be exceeded. As this point is
approached, the combination of storm washover and
shoreface erosion will cause the barrier to retreat until
equilibrium is restored, a position which may be well
inland from the tips of headlands. Both littoral wave
power and sediment supply may be deficient in these
inland positions, further jeopardizing the survival of
A. STILLSTAND
B. BEACH DETACHMENT
C. CYCLIC SPIT
PROGRADATION
FIGURE 21. Barrier formation with spit-building dominant.
As a rugged coast passes from stillstand to transgression, a mature
configuration is replaced by a transient state of mainland beach
detachment, then by a quasi-steady state regime of cyclic spit
building. This diagram also illustrates the relationship between
the concepts of coastal equilibrium and coastal climax, since it
consists of Johnson's (1919) stages of coastal maturity — portrayed
in reverse sequence!
the barrier. As the loop of the barrier into the bay
becomes extreme, sediment supply from headlands is
liable to capture by secondary spits formed during
storms. These may prograde out toward the drowned
valley thalweg until capacity is again exceeded and
their tips are stabilized, further movement being
limited to retreat coupled with that of the headland to
which they are attached.
Finally the survival of primary barriers on such a
coast would be limited by the tendency of submerging
headlands to form islands. A spit tied to a promontory
that becomes an island can retreat no further if a drowned
tributary valley lies landward of it, but must instead be
overstepped. The few unequivocal examples of trans-
gressed barriers on the shelf floor appear to be over-
stepped, rock-tied spits (Nevesskii, 1969; McMaster and
Garrison, 1967).
On the other hand, transgression of a coast of very
subdued relief, such as is the case for most coastal plains,
would tend to promote mainland beach detachment at
the expense of spit formation, given initial conditions of
a straight coast (Fig. 22). The depth of water in which
detached bay mouth barriers would be built would be
less, because the relief would be less. The upper, ero-
sional zone of the shoreface (Fig. 10/1) would be more
likely to extend down into the pre-Recent substrate
(Fig. 23^4); hence erosion of the inner shelf floor would
become as important a source of sand for the barrier as
the erosion of adjacent headlands. With a rise in sea
level, valley-front dune lines would grow upward. River
mouths, initially deltaic, would flood as estuaries, while
lagoons would creep behind the beaches toward the
headlands on either side. Barriers would retreat in cyclic,
tank-trend fashion by means of storm washover, burial,
and reemergence of the buried sand at the shoreface
(Fig. 10.4). Coastal discontinuities sufficient to induce
coastwise spit progradation would occur only locally.
Thus, on a low, initially straight coast, barrier spits and
barrier islands would preferentially form by mainland
beach detachment rather than by coastwise progra-
dation.
Storage in Prograding Shorefaces
The preceding discussion has identified barrier islands
and spits as forms of sand storage on retreating coast-
lines. On prograding coastlines, sand storage occurs in
beach ridges and cheniers; the two forms differ in that
beach ridges are separated by sand flats, whereas che-
niers are separated by, and rest on, mud deposits.
Sequences of beach ridges 15 to 200 m apart may
form subaerial strand plains tens of kilometers wide.
These are smaller scale features than the barriers, which
494
282
COASTAL SEDIMENTATION
A STIUSTAND
B. BEACH DETACHMENT
C. BARRIER RETREAT
FIGURE 22. Barrier formation with mainland beach detachment as the
dominant process. A mature low coast passes via main land beach detach-
ment into a steady state regime oj barrier retreat.
characterize retreating and stillstand coasts; and bar-
riers may, in fact, be locally comprised of beach ridge
fields, as a consequence of minor frontal progradation
or more extensive distal, coast-parallel migration (Hoyt
and Henry, 1967). Curray et al. (1969) have presented
a detailed study of what has been recognized as a classic
strand plain coast, the Costa de Nayarit (Fig. 24). They
postulate that each ridge forms as a plunge point bar,
which in the presence of an oversupply of littoral sand,
builds up close to mean low water. During a period of
constant low swells, the bar may grow above this level
as tides rise to the spring tide value (0.98-1.25 m);
the bar becomes a subaerial feature during the subse-
quent neap phase, and continues to grow by eolian
activity (Fig. 105).
Chenier plains form on coasts with a high suspended
sediment input. In the classic chenier plain of the
Louisiana coast west of the Mississippi Delta, the sand
ridges support stands of live oaks (French, chene), hence
the name (Price, 1955). The formation of chenier plains
has been ascribed to rapid progradation of mud flats
during periods of high suspended sediment discharge
from nearby rivers or delta distributaries. When distribu-
taries crevasse and the subdeltas are abandoned, the
495
SAND STORAGE OFF CAPES
283
rrr^TTT777T7
LAGOONAL
DEPOSITS
FIGURE 2 3. Contrasting sand budgets of a barrier built (A) on
a gentle submarine gradient as in Fig. 22, and (B) on a steep
submarine gradient as in Fig. 21. In (A), zone of shoreface erosion
penetrates to Pre-Recent substrate, which becomes "income" for
barrier nourishment. In (B) the barrier may only "borrow" from
its own "capital" through shoreface erosion, and the heavy ex-
penditure involved in paving the shelf with sand during barrier
retreat may "bankrupt" the retreating barrier, which must
either accelerate its retreat or be overstepped. In either case
shoreface continuity is liable to be broken, resulting in cyclic spit
building.
downdrift coast becomes sediment starved. The mud flat
erodes back and a beach ridge formed of the coarse
sediment is thrown up by storm high tide. Todd (1968)
has stressed the role of estuary mouths and inlets in the
localization of chenier plains. He notes that littoral cur-
rents updrift of the inlet tend to decelerate during ebb
tide because of a reduction of the coastwise pressure
gradient by the ebb jet, resulting in sediment deposition.
Littoral currents in the same locality are accelerated by
proximity of the inlet during flood tide, but fine sediment
deposited requires a greater velocity for erosion than for
deposition, and in any case has already compacted.
Hence chenier plains tend to be localized on the muddy,
updrift sides of tidal inlets. Downdrift of the inlet, the
coast may instead be starved for fines as a result of sea-
ward transport or "dynamic diversion" of the littoral
current by the ebb jet, and the littoral sand deposits
have more nearly the character of a beach ridge sequence.
Otvos (1969) recognizes the role of inlets in localizing
deposition, but notes that chenier deposition goes on for
long distances beyond inlets. He cites new chronologic
evidence from the Mississippi chenier plains to indicate
that chenier ridge formation cannot be closely correlated
with the abandonment of a subdelta mouth, and sug-
gests that the intermittent shielding effect of nearby sub-
delta growth on the wave climate plays a greater role in
cyclic chenier plain growth.
Beall (1968) has examined in detail sediment distribu-
tion and stratigraphy in the present shoreface of the
western Louisiana shoreline (Fig. 25). He distinguishes
between three main stratigraphic patterns. Mudflats are
defined on their seaward margins by a break point bar
zone of very fine sand. Midtidal and upper tidal flats
are distinguished by progressively finer sand and in-
creasing percentages of silt and clay. A thin sand storm
beach may rest on eroded marsh sediments. The strati-
graphy is complex. Apparently a period of increasing
littoral sediment discharge results in progressive flatten-
ing of a shoreface, until the bar zone is triggered and
becomes the maximum locus of sedimentation, prograd-
ing both landward and shoreward. A (submarine) mud
flat zone is thereby initiated in the sheltered longshore
trough, and progrades toward the bar and landward.
Transitional beaches have largely erosional profiles, with
thin bar, beach, and washover sands overlying the ero-
sional surface near the high-water line. The sequence is
typical of that of erosional transgression, where the thin
sand cap is a transient fair-weather veneer. However,
Beall interprets these transitional beaches as prograda-
tional, with rates of progradation intermediate between
those of mud flats and those of "normal beaches."
Normal beaches consist of up to 1 .7 m of seaward-fining
fine sand, prograding seaward over an outer shoreface
facies. Washover fans of normal beaches are thicker than
those of transitional beaches. The three types of beaches
described by Beall would appear to illustrate a temporal
as well as a spatial sequence. Periods of rapid mud flat
progradation are presumably followed by erosion, then
the formation of transitional beaches, which prograde to
become cheniers, then prograde more rapidly as mud
flats.
SAND STORAGE OFF CAPES
South of the Middle Atlantic Bight of North America, the
generally southwest-trending coastline has been molded
into a series of large-scale cuspate forelands (Fig. 26).
They are the response of the shoreface regime to a mod-
erate to intense wave climate and a high variance in the
direction of wave approach (Swift and Sears, 1974).
Storm waves approach from the northeast, as is the case
in the Middle Atlantic Bight, but the coast is also exposed
to waves from more distant storms in the southeastern
Atlantic. As a result, the cuspate forelands have been
self-maintaining features throughout the postglacial pe-
riod of sea-level rise and erosional shoreface retreat.
Each foreland apex is a zone of littoral drift convergence.
496
284 COASTAL SEDIMENTATION
106°
I05°30'
23°N
FIGURE 24. The strand plain of the Costa de Nayarit, showing beach
ridge sequences. Rio de la Cahas meanders through interlocking spits,
indicating reversal of drift directions. From Curray et al. (2969).
The resulting surplus of sand at the apex creates a coastal
shoal. The shoal in turn maintains a pattern of wave re-
fraction that drives littoral drift convergence (Swift and
Sears, 1974); see Fig. 27.
The question arises as to how such a closely coupled
feedback system begins. The answer is that in a sense, it
does not matter. It is a truism that as process variables
approach the instability threshold, any singularity in a
water-substrate system will excite the feedback of the
process and response that lends to the formation of bed
forms. In the case of the cuspate Carolina coast, the
initial conditions were probably the sequence of shelf-
edge deltas during the Late Wisconsinan low stand, cor-
responding to the Peedee, Cape Fear, Neuse, and Pam-
497
SAND STORACE IN INNER SHELF RIDGE FIELDS
285
TIDAL -MUDFLAT
JTTTTTrUL.
LOCALITY A
HIGH TIDE-
PROTECTED MUDFLAT FACIES
OUTER SHOREFACE SEDIMENTS
OFFSHORE GULF BOTTOM MUD
^JHnNER- SHOREFACE
^ BREAKER BARS
WASHOyER FAN
EROSIONA
SURFACE
'TRANSITIONAL BEACH"
— HIGH TIDE'
irar
LOCALITY B
MUDFLAT SEDIMENTS
ERODED SHOREFACE/
OUTER SHOREFACE
OFFSHORE GULF BOTTOM MUD
"NORMAL BEACH" LOCALITY "C"
vWASHOVER FANS HIGH TIDE*
OFFSHORE GULF BOTTOM MUD
100
FIGURE 2 5. (A) The chenier plan of southeastern Louisiana. (B) Characteristic beach
configurations. See text for explanation. From Beall (1968).
lico rivers (note river system of Fig. 26 and compare
Fig. 27). The Appalachicola cuspate foreland of the
Florida Panhandle is particularly suspect as having been
formed by this mechanism (Swift, 1973). Other cape-
associated shoals may occur as a consequence of the re-
duction in intensity of littoral drift around a rock-de-
fended cape with consequent reduction in the compe-
tence of littoral drift (Tanner, 1961; Tanner et al.,
1963). On offset coasts (forelands separated by zetaform
bays; p. 275), the forelands may be triggered by cape
extension shoals, either river mouths or rocky promon-
tories (Davies, 1958). Forelands and cape-associated
shoals also occur on swash-aligned coasts. These coasts
tend to be inherently unstable, breaking into short "arcs
498
286
COASTAL SEDIMENTATION
H SHOAL RETREAT MASSIF
FIGURE 26. Cuspate coast of the Carolinas. Values for littoral drift are in yd/year X 10~3. From Langjelder et al. (1968).
of equilibrium," terminating in cuspate forelands with
neither rivers nor outcrops required for cusp formation.
Tidal inlets may evolve into cuspate forelands with asso-
ciated shoals. Chincoteague Shoals, on the Delmarva
coast, is an example of a barrier-overlap inlet that has
become a cuspate spit with associated shoal (Fig. 36).
Shoals developing over cuspate forelands may extend
seaward the width of the shelf. Such cape extension
shoals do not result from the seaward transport of sand,
but rather from the landward translation of the cuspate
foreland in response to rising postglacial sea level, to-
gether with the retreat of the associated littoral drift
convergence. The seaward-trending shoal marks the re-
treat path of this convergence. Its response to the shelf
regime is discussed in the next chapter.
SAND STORAGE IN INNER SHELF RIDGE FIELDS
Storm-Induced, Shoreface-Connected Ridges
A major category of inner shelf sand storage found on
low retreating coasts is storage in shoreface-connected
ridges (Fig. 28) and in associated inner shelf ridge fields.
These features are up to 10 m high, 2 to 5 km apart, and
their crestlines may extend for tens of kilometers. Side
slopes are rarely more than a degree. They typically
converge with the shoreface at angles of 25 to 35°
(Duane et al., 1972) and may merge with it at depths
as shoal as 3 m. The best known development is on the
coast of the Middle Atlantic Bight of North America,
but they may be found on coastal charts as far south on
the Atlantic coast as Florida, and around the Gulf coast
littoral as far as Alabama. They also appear locally on
the Texas coast. Allersma (1972) has reported them on
the muddy coast of Venezuela, where they are dom-
inantly composed of mud. They have been detected by
ERTS satellite imagery on the Mozambique coast (John
McHone, personal communication), and also appear on
the southern littoral of the North Sea. With the excep-
tion of the Venezuelan coast, most settings are that of a
low, unconsolidated coast undergoing Bruun erosional
retreat (Fig. \0A) in response to a moderate to strong
wave climate and periodic intense storm or tidal flows.
Where best studied, on the Virginia-northern North
Carolina coast (McHone, 1972), the ridges appear to
have some of the response characteristics of wave-built
bars at their inner ends where they merge with the
shoreface. Like wave-built bars, their landward ends are
asymmetrical, with steep landward flanks, although the
499
(a)
FIGURE 27. Model for the transformation of a stillstand delta into a retreating cuspate foreland. From
Swift and Sears (1974).
(b)
B
FIGURE 28. (a) Shoref ace-connected ridges of the Delmarva from (a). Dots represent hypothetical fixed points during a
inner shelf, contoured at 2 fathom intervals. From ESS A bathy- period of shoref ace retreat and downcoast ridge migration. See
metric map 0807 N-57. Ridges are in varying stages of detachment. text for explanation. From Swift et al. (1974) ■
(b) Schematic diagram of detachment sequence as inferred
500
287
288
COASTAL SEDIMENTATION
BASE
LINE
200
DISTANCE FROM BASE LINE IN METERS
400 600 800 1000
1200
1400
SMOOTHED
SEA FLOOR
BASIC PROFILE)
FIGURE 29. Superimposed profiles of the inner shorejace-connected ridge at False Cape,
Virginia. From McHone (1972).
reverse asymmetry tends to prevail further seaward. En-
velopes of profiles indicate that, as in the case of their
small-scale break-point counterparts, ridges built to a
height of approximately one-third water depth, at which
point wave agitation is sufficiently intense to preclude
further growth. The troughs between the ridges and the
shoreface are similarly excavated to one-third of water
depth below the smoothed profile (McHone, 1972); see
Fig. 29. At the False Cape Ridge Field, Virginia
(McHone, 1972; Swift et al., in press), analysis of the
wave climate suggests that waves are capable of breaking
on some part of the inner ridge crest about 10% of the
time. As a consequence of their oblique orientation and
varying crestal depth, such ridges may utilize energy
from a relatively broad spectrum of wavelengths.
As wave-built bars, however, the low-angle ridges are
anomalous. They are much larger than surf zone bars
and their oblique orientation is more nearly parallel to
the direction of wave approach than normal to it. The
ridges may be primarily a response to a downwelling
coastal jet that comprises the coastal margin of the
storm flow field (see p. 275), although storm wave action
is clearly a complementary mechanism. At False Cape,
Virginia, a 28 hour current-meter station revealed a
steady southward and offshore flow on the order of 15
cm/sec at a distance of 8 cm off the bottom, subsequent
to the passage of a cold front with winds in excess of
25 knots (Fig. 30). During this period, however, the
anchored observation vessel maintained a wake trending
southward and shoreward. The inferred structure of the
coastal flow field during the observation period is pre-
FIGURE 30. Progressive vector diagram of storm bottom flow
at the innermost ridge at False Cape, Virginia. Vectors represent
velocities taken for 3 minute intervals every 30 minutes by two
orthogonal Bendix Q-18 meters mounted in a plane parallel to
the seafloor, 16 cm off the bottom. After passage of a cold front,
bottom flow trended southeast obliquely seaward over ridge crest
at velocities up to 18 cm/sec, while wake of anchored observation
vessel streamed southeast, toward shore. Based on Holliday (1971).
501
SAND STORAGE IN INNER SHELF RIDGE FIELDS
289
SURFACE
WIND DRIVEN
CURRENT
BOTTOM
WIND DRIVEN
CURRENT
SURFACE
WAVE DRIVEN
FIGURE 31. Hypothetical structure of the coastal boundary of the storm
flow field, based on Figs. 2 and 30.
sented in Fig. 31. The observed pattern is interpreted
as downwelling coastal flow intensified by constriction
of the trough toward its southern end, and also by
the setup of waves breaking on the inshore end of the
trough. Mapping of the junction of this ridge with the
shoreface on four successive occasions has revealed the
presence of a shifting saddle, where storm flows presum-
ably break out over the ridge base (McHone, 1972).
A grain-size profile over the ridge is extremely asym-
metric (Fig. 32). Sands are coarsest in the landward
trough and become steadily finer up the landward flank,
are of relatively constant grain size across the crest, and
become finer again down the seaward flank. Sorting is
variable on the landward flank and crest but increases
steadily down the seaward flank.
The profile is characteristic of a flow-transverse sand
wave, and suggests that the ridges are responding as
would a sand wave to the cross-shoal component of flow.
As described in Chapter 10 (p. 166), bed shear stress
increases up the upcurrent flank of a sand wave, attain-
ing a maximum at the crest or just forward of it, then
decreases down the downcurrent flank. Grain size would
tend to decrease monotonically across such a shear stress
maximum as a consequence of the progressive sorting
mechanism; as sand is eroded out of the trough, the
coarser grains are more likely to be trapped out in the
initial portion of the transport path (Chapter 10, p.
162). On this particular ridge, however, size character-
istics do undergo a reversal on the landward side of the
crest, where maximum shear stress is to be expected.
502
290
COASTAL SEDIMENTATION
o -
i l3S
06 -
2 -
DEV
o
z
tat
O
= 05 -
4 -
DIA ^__ ^Ak \
z
o
6-
BOTTOM PROFILE^^^
STANDARD
o
Co
— 1
Uj
>
8 -
u 02 -
0
1
100
1
METERS
200 300
1 1
400
1
•
500
5
2
<t
oc
o
♦ 3 5
Q
FIGURE 32. Grain-size profile across the inner ridge at False Cape, Virginia (by Leonard Nero).
Shoreface ridges are not simply giant sand waves, how-
ever, because flow crosses them at an oblique angle. The
model presented in Fig. 31 suggests that they partake of
the characteristics of both flow-transverse sand waves
(Chapter 10, p. 166) and flow-parallel sand ridges
(Chapter 10, p. 172).
Thus the ridges appear to represent the attempt of an
intensified coastal flow to build an outer bank with
materials scoured from the high-velocity axis of down-
welling. The topography presented in Fig. 30 presents a
clue to the historical development. The initial perturba-
tion required to trigger these large-scale instabilities of
the shoreface might be something on the order of the
migrating "sand humps" of the surf zone described by
Bruun (1954) and Dolan (1971). As the ridge and trough
take form in response to the interaction of storm waves
with the coastal storm flow, they would tend to be self-
propagating. Enlargement of the trough by headward
erosion and downward scour of the trough, and aggra-
dation of the ridge crest by peak flow events would cre-
ate a morphology that would amplify successive trough
flows. Thus during a period of general shoreface retreat
due to storm erosion, ridges would be carved out of the
shoreface partly because the shoreface would retreat
away from them, and partly because they themselves
would tend to migrate obliquely offshore, extending their
crestlines to maintain contact with the shoreface as they
do so (Fig. 28). The sinuous pattern of crestlines on the
inner shelf floor of the Delmarva peninsula suggests that
ridge formation may be an episodic affair; troughs en-
large and trough flows amplify until the flow is intense
enough to cut through the ridge base, whereupon the
process is repeated farther down the shoreline. Shore-
face-connected ridges are seen in all stages of detachment
in Fig. 28. It is doubtful, however, if this is a full or
adequate explanation of ridge genesis even in a qualita-
tive sense. Any more comprehensive analysis must un-
dertake to explain the relative orientations of peak bot-
tom flow, the ridge crest, and the shoreline. As noted,
neither the breakpoint bar nor the sand wave models
meet this requirement.
The ridges are generally oriented along a trend that
is intermediate between the dominant direction of storm
wave approach and the coast-parallel trend of storm
currents (narrow-angle ridges), perhaps as a consequence
of the dual role of these elements in ridge genesis.
Locally, however, they may be aligned along the direc-
tion of storm wave approach (wide-angle ridges: Duane
et al., 1972); see Fig. 33. Such ridges would resemble
the "finger bars" of Niedoroda and Tanner (1970),
rather than break-point bars. While breaking waves
tend to drive sand across break-point bars, the bottom
surge of refracting waves tends to drive sand obliquely
crestward and landward, up both sides of a finger bar.
Wide-angle ridges exhibit the same textural and mor-
phologic asymmetry as do narrow-angle ridges, indicat-
ing that they too are shaped by storm flow as well as by
wave surge. But they presumably react to storm flow
more nearly as a flow-transverse bed form than as a
flow-parallel bed form, as in the case of narrow-angle
503
B
- 2
o
* 1
h-
ir
o
en
1 1 r
NORTH
SLOPE
o TROUGH
+ BARRIER
■ RIDGES
I I I ill.
, 'J S SLOPE
-2-101234
MEDIAN DIAMETER (0)
FIGURE 33. Substrate response and hydraulic process jor a wide-angle,
shorejace-connected ridge system, Bethany Beach, Delaware. (A) Bathym-
etry. From Moody {1964). (B) Median diameter versus inclusive graphic
standard deviation for sand samples. From Moody (1964). (C) Schematic
model jor the generation and maintenance of wide-angle ridges. Refracted
waves converge toward ridge crests. Breaker angle is most intense at heads
of trough. Storm flow is coast parallel and to the south.
504
291
292
COASTAL SEDIMENTATION
ridges. A full understanding of the genesis of these
coastal sand bodies must await more detailed field meas-
urements of the responsible flows.
Storm-Induced Inner Shelf Ridge Fields
Ridge fields on the inner shelf floor continue their mor-
phologic identity and their characteristic pattern of
grain-size distribution (Fig. 34). Relief continues at 10 m,
and slopes for isolated inner shelf ridges are very similar
to those of shoreface-connected ridges. Scour continues
in troughs; the erosional surface cut by shoreface trans-
lation extends beneath the ridges and is locally exposed
in trpugh axes (Swift et al., 1972a, b). Generally, how-
ever, it is veneered with a few decimeters of coarse,
pebbly sand, overlaid by finer sand. The coarser sand is
commonly exposed in elongate windows through the
finer sand veneer. Sidescan sonar records suggest that the
finer sand is moving as ribbonlike streamers over a
coarser lag substrate. Ridge crests consist of medium to
fine, well-sorted sand, with cross-stratified horizons (Swift
et al., 1972a; Stubblefield et al., 1975). Flanks consist
of fine to very fine sand and are distinctly asymmetrical
in their textural pattern; seaward flanks are notably
finer, and are locally steeper than landward flanks.
Crestal sands, however, may be distinctly coarser than
the flank sands of either side, probably a response to
winnowing by wave surge.
The inner shelf ridges themselves appear to be in a
state of slow transit, wherever there is a bathymetric
time series adequate to test this hypothesis (Figs. 35 and
36). The pattern of movement is a fairly consistent one,
in which both shoreface-connected and isolated inner
shelf ridges move along similar trajectories. Where the
angle of convergence of the ridge crest with the shoreline
is fairly large, the ridges are moving downcoast and
offshore, extending their crestlines so as to maintain
contact with the shoreface as they do so. Where the
ridges are nearly coast-parallel, they are extending these
crestlines downcoast, and may move either inshore or
offshore, but more commonly offshore.
The considerations just discussed strongly suggest that
inner shelf ridges continue to interact with the shelf
flow field after detachment, in such a way as to main-
tain their morphologic and textural characteristics. In
fact, ridged inner shelf topography occurs on sectors of
the North American inner shelf where it cannot have
74-20
74° 10'
FIGURE 34. Grain size distribution in the Brigantine inner
shelf ridge field, New Jersey. Data of M. Dicken.
75°00'
FIGURE 35. Bathymetric time series from the Bethany Beach
ridge field, Delaware, between 1919 and 1961. From Moody (1964).
505
SAND STORAGE IN INNER SHELF RIDGE FIELDS
293
FIGURE 36. Bathymetric time series of Chincoteague shoals,
Delmarva (Delaware-Maryland-Virginia) coastal compart-
ment. Ridges have migrated slightly offshore, and have extended
their crestlines markedly to the south between the Coast and
Geodetic surveys of 1881 (dashed line) and 1934 (solid line).
From Duane et al. (1972).
been formed by shoreface ridge detachment (Swift et al.,
1974); see Chapter 15. It appears that the shelf hydraulic
regime will adopt ridges from the retreating shoreline
or mold them afresh in the substrate if the hydraulic
regime is conducive to a ridged substrate. A possible
mechanism for ridge maintenance involving helical flow
structure in the storm flow field is presented in Chapter
10 (p. 173).
There is in addition a smaller scale bed form pattern
on the inner shelf whose patterns of distribution are
compatible with this hypothesis. These are the sand rib-
bons On the inner shelf of the Middle Atlantic Bight of
North America, revealed by sidescan sonar. They tend
to be 5 to 50 m wide, are of negligible relief, and tend
to make angles of 10 to 45° with the shore. They are
most commonly observed as dark streaks on sidescan
sonar, which means that they are not true sand ribbons
(streamers of finer sand over an immobile substrate of
coarser sand or gravel) but are instead erosional windows
in which a coarser substrate is locally exposed through a
discontinuous sand sheet (Chapter 10, p. 170). Locally,
however, the pattern anastomoses so as to create a true
sand ribbon pattern (Fig. 37). The dark streaks in many
areas are distinctly asymmetrical, with sharper landward
boundaries (Chapter 10, Fig. 17). The streaky patterns
occur on the smooth inner shelf or in ridge fields. In the
latter case they are largely confined to troughs, where
they tend either to parallel the trough axes or make a
somewhat larger angle with the shoreline.
FIGURE 37. Sidescan sonar record of sand ribbon
pattern in the trough of a central shelf ridge. From
McKinney et al. (1974)' Dark band is a window of coarse
trough sand traversed by streamers of fine (lighter toned)
sand.
506
294
COASTAL SEDIMENTATION
As noted in Chapter 10, the larger ridges are probably
responses to repeated flow events, whereas the smaller
sand ribbons and erosional windows may be formed
during a single flow event. The similar orientations and
asymmetries of the sand ribbons and ridges suggest that
both may be responses to a geostrophic flow regime in
which secondary flow cells occur at several spatial scales.
With fully developed secondary flow, the tendency for
regional landward transport of surface water and re-
gional seaward transport of bottom water (solid arrows,
Fig. 38) would be suppressed in order to maintain con-
tinuity. Instead, cells rotating with the sense of regional
shore-normal flow component would be enhanced, and
cells rotating with the opposite sense would be sup-
pressed. Detailed measurements of the velocity field in
the vicinity of such offshore ridges during peak flow
events would serve to test this model, and perhaps lead
to alternative models.
Tide-Induced Inner Shelf Ridge Fields
Tide-dominated coasts such as the Anglian coast of
England also tend to store sand in shoreface-connected
and inner shelf ridges. The forcing mechanism for ridge
formation must be in part the storm-augmented shelf
flow field as in the case of such storm-dominated coasts
as the Middle Atlantic Bight of North America, since
the Anglian coast is also subjected to severe storms
(Valentin, 1954). However, this coast experiences in
addition the progressive tidal edge wave associated with
the amphidromic tidal system of the North Sea (see
Chapter 5, p. 60) on a twice daily basis; midtide coastal
tidal velocities regularly exceed 2 knots.
As a consequence of the greater rate of energy ex-
penditure in tidal flow than in wave- and wind-current-
generated flow, storage of sand in the subaerial zone
as barrier superstructure, or in the surf zone as a beach
and surf prism, is greatly inhibited at the expense of
submarine storage in tide-maintained sand bodies. The
efficiency of this storage is greatly strengthened by the
tendency of the coastal tidal wave to interact with a
loose substrate by the formation of interdigitating ebb
and flood channels separated by sand shoals which form
effective sand traps (Robinson, 1966); see the discussion
on page 177. As in the case of the Middle Atlantic
WIND
^
COAST PARALLEL
FLOW COMPONENT
COAST NORMAL
FLOW COMPONENT
COAST NORMAL
HELICAL FLOW COMPONENT
OLDER SUBSTRATE
HOLOCENE SAND SHEET
FIGURE 38. Hypothetical scheme showing a possible
mode of coupling between Ekman flow cells and a mobile
inner shelf substrate during a period of strong downcoast
winds. See text for explanation.
507
SAND STORAGE IN INNER SHELF RIDGE FIELDS
295
FIGURE 39. Tide-maintained ridge typography on the inner
Anglian shelf. Shoreface-connected ridges separate ebb- and flood-
dominated channels. Ridges tend to migrate southward with time,
and to detach from retreating shoreface. Ridges are nourished at
the expense of shoreface, hence constitute cases of downdrift by-
Bight, the nearshore zone of sand storage is not a sta-
tionary one but is translating landward in response to
postglacial sea-level rise and erosional shoreface retreat.
The primary element of storage is again a shoreface-
connected sand ridge (Fig. 39). The angle between
these ridges and the coast opens northward, into the
direction of the advancing coastal tidal wave, and there-
fore the trough opens into the flood-dominated residual
tidal flow. Downwelling may also occur during the flood
tide, since the high velocity axis of trough flow will tend
to converge with the rising trough axis. The outside of
the ridge is shielded from the flood tide, and therefore
experiences a. greater ebb discharge. This diversion of
flow might be expected to result in a sand circulation
cell, with sand moving obliquely up the inner flank and
over the crest during the flood tide, to be returned along
the seaward flank during the ebb tide. As in the case of
storm-maintained shoreface-connected ridges, probably
any initial perturbation of the shoreface would result in
such a self-maintaining system.
As in the case of the Middle Atlantic Bight, the ridges
tend to migrate offshore and downcoast, in the direction
of the residual tidal flow (Robinson, 1966), and tend to
become detached and isolated on the inner shelf floor.
However, unlike storm-maintained ridges, tidal ridges
on the inner shelf floor tend to be unstable. Variations
in the rate of offshore migration along the length of the
ridges tend to result in self-propagating modifications of
ridge morphology (Caston, 1972; Chapter 10, Fig. 21),
whereby the ridge deforms into a sigmoidal pattern,
because of the growth of secondary ebb- and flood-
dominated channels, and may eventually split into three
ridges.
As in the case of the Middle Atlantic Bight, isolated
inner shelf ridges continue to be maintained, but the
character of interaction between the flow field and sub-
strate changes as the water column deepens. Ridge spac-
ing increases as a function of flow depth (Allen, 1968b)
as sand is partitioned between fewer, wider ridges. As
channels widen, they cease to become wholly ebb- or
flood-dominated, but are themselves partitioned into
ebb-dominant and flood-dominant sides (Caston and
Stride, 1970); see Chapter 10, Fig. 20). All channels
have the same sense of shear. Thus the offshore ridges
are sand circulation cells, but the sense of circulation is
the same from ridge to ridge, instead of alternating
between clockwise and counterclockwise as on the
shoreface.
passing. Offshore ridges are probably being nourished at expense
of nearshore ridges; if so, sand is moving seaward more rapidly
than are the ridge forms. From Robinson (1966).
508
296
COASTAL SEDIMENTATION
SAND STORAGE AT COASTAL INLETS
Categories of Coastal Inlets
In addition to the coastal flow discontinuities found at
capes and cuspate forelands and on ridged shorefaces,
discontinuities also occur at coastal inlets, which like-
wise result in sand storage systems. The term coastal
inlet is here used in its broadest sense, for a variety of
coastal reentrants, defined by the ratio of salt- to fresh-
water discharge in their two-layer circulation systems,
and in the extent to which their channel cross sections
have equilibrated with the discharge (Table 3).
TABLE 3. Categories of Coastal Inlet
Constructional form
(equilibrium channel)
Erosional form
(inherited river valley)
Delta distributary
Tidal delta distributary
or equilibrium estuary
■3 Tidal inlet
18
tZ3
Disequilibrium estuary
Tidal channel-mud flat
complex
Bay
The main sequence of coastal inlet morphologies trends
diagonally across Table 3, from delta distributaries en-
tering tideless seas, through tidal distributary mouths
and trumpet-shaped equilibrium estuaries to funnel-
shaped disequilibrium estuaries, to tidal channel-tidal
flat complexes and open bays. Tidal inlets are hybrid
cases, in which an equilibrium channel has been fitted
to a lagoon. Special effects such as mirror-image ebb,
flood "tidal deltas," and offset barrier coasts result.
Tidal inlet morphologies are continuous with equilibrium
estuary morphologies and grade into them through in-
termediate cases in which a central channel meanders
through a marsh-filled lagoon.
Equilibrium River Mouths
Delta distributaries (prograding river mouths) and equi-
librium estuaries (retrograding river mouths) belong to
a general class of river mouths. The cross-sectional area
of river channels is a power function of river discharge
(Leopold et al., 1964, p. 215). Where rivers enter the
sea, a salt wedge intrudes beneath the fluvial jet, whose
discharge is amplified by a two-layer (estuarine) circu-
lation (see p. 24). Most rivers enter tidal seas, and
river mouth discharge is further amplified by the dis-
charge associated with the semidiurnal tidal cycle, which
propagates for some distance upstream. Thus a river
mouth whose channel is in equilibrium with total dis-
charge must expand rapidly through the tide-influenced
zone toward the sea, resulting in a trumpet-shaped plan
configuration.
At the river mouth proper, a variety of processes con-
spire to construct an arcuate, seaward-convex sand shoal
(Fig. 40). The most fundamental factor is the hydro-
dynamic continuity relationship: Expansion of the fluvial
jet results in rapid deceleration and loss of competence,
and river sand is deposited in the form of a shoal. Estu-
arine circulation also plays a role; river mouth morphol-
ogy and the circulation interact, so that the crest of the
shoal becomes the leading edge of the salt wedge during
flood stage (Fig. 41), or, if the tidal component of river
mouth discharge is very large, the spring ebb tide ter-
minus of the salt wedge, or both (Wright and Coleman,
1974; Moore, 1970; Farmer, 1971). The crest of the
shoal thus becomes a bottom current convergence dur-
ing periods of maximum sediment transport, and hence
a reservoir for sand storage. A second major source of
sand maintaining the river mouth shoal is littoral drift,
which is diverted seaward along the shoal crest.
Sediment storage in tidal river mouths is mediated by
the behavior of the tidal wave as it passes over the shoal
crest. Here, as on open tidal coasts, tidal wave and sub-
strate tend to interact to form interdigitating ebb- and
flood-dominated channels separated by shoals that are
efficient sand traps. The tide within the estuary is re-
tarded by friction and is out of phase with the shelf tide;
it continues to ebb after the shelf tide has already begun
to flood. The two water masses tend to interpenetrate,
with the main ebb jet passing out over the center of the
shoal and the oceanic tide flooding on either side of it.
The response of the shoal surface to this periodic flow
pattern is an interdigitation of ebb- and flood-dominated
channels, separated by a discontinuous, zigzag system of
sandbanks (Ludwick, in press); see the discussion on
page 177.
A further process modifying the surface of the shoal
and enhancing its capacity for sand storage is the inter-
action between the tide-generated pattern of channels
and sand ridges and incident wave patterns. The arcu-
ate pattern of the shoal as a whole serves to focus wave
energy on it. Sand ridges between ebb and flood chan-
nels tend to build toward the level of mean high tide.
As their upper surfaces emerge into the intertidal zone
they become swash platforms, on which intersecting pat-
terns of wave trains tend to drive sand in the resultant,
landward direction (Oertel, 1972); see Fig. 40.
509
81°20'
81°10'
81°00'
80°50'
32°00'
- 32°00'
31°50'
31° 50'
31°40'
31°30'
31° 20'
81°20'
WAVE TRANSPORT CT> TIDAL TRANSPORT ► INTERTIDAL
SUSPENSIVE TIDAL TRANSPORT ►
V4 METERS
FIGURE 40. Sedimentation patterns at the mouths of Georgia estuaries as inferred from Oertel (1972) and Oertel
and Howard (1972).
510
297
Region I
Region III
Channel
Processes
Region IV
Buoyant expansion,
wave, wind, and tide induced mixing
Velocity scale
0 1 2
J I L^J
m/sec
6 8 10
Distance seaward. x/b„
Region I
Region IV
Weak buoyant expansion, wave
wind and tide induced mixing
6 8 10
Distance seaward, x/bn
FIGURE 41. (A) Cross section of density and flow during flood stage {Aprils, 1973). Both sections taken
structure of the South Pass of the Mississippi River during flood tide. From Wright and Coleman (1974).
during low river stage (October 25, 1969) and (B)
298
511
SAND STORAGE AT COASTAL INLETS
299
Despite the relatively rapid postglacial rise in sea level,
some river mouths have been able to maintain equilib-
rium channels in which cross-sectional area is adjusted
to discharge, as deltas (prograding river mouths) or as
equilibrium estuaries (slowly retrograding river mouths;
see Figs. 42A,B). Most, however, have not. Disequilib-
rium estuaries have resulted whenever aggradation of the
estuary floor in millimeters per year has been less than
the rate of sea-level rise, so that before any given segment
of channel could close down to the required cross-
sectional area, the main shoreline had passed it by.
Such "drowned" or disequilibrium estuaries are gen-
erally nearly funnel-shaped, rather than trumpet-shaped,
as are the equilibrium forms. As a consequence of their
higher ratio of saltwater to fluvial discharge, their river
mouth shoals are retracted into the throat of the estuary
and the interpenetration of ebb and flood channels be-
comes marked (Fig. 42C).
With a yet further increase of tidal over fluvial dis-
charge such a coastal indentation may no longer be ap-
propriately called an estuary, but simply a bay. Large
bays experiencing high tidal ranges may build a tide
mmm m
A. CONSTRUCTIONAL CHANNEL
RIVER DOMINATED FLOW
B. CONSTRUCTIONAL CHANNEL
TIDE DOMINATED FLOW
C. PARTLY EROSIONAL CHANNEL
TIDE DOMINATED FLOW
INTERTIDAL SHOAL
gig: SUBTIDAL SHOAL
1 FLOW DOMINANCE
OF CHANNEL
FIGURE 42. Varieties of river mouths. (A) Prograding delta distributary entering
tideless sea (based on Mississippi River Delta). (B) Equilibrium {discbarge adjusted) tidal
estuary mouth (based on Georgia coast estuaries). (C) Disequilibrium estuary mouth (based
on Thames estuary). From Swift (1975b).
512
300
COASTAL SEDIMENTATION
flat-tidal channel complex at their heads as a consequence
of net landward sediment transport by the shoaling
tidal wave. These deposits are the functional equivalent
of the tide-molded deposits of a disequilibrium estuary.
The patterns of sand storage in estuary mouths may
be extremely elaborate. These dynamic topographies are
of major concern to port authorities concerned with the
maintenance of deep-water approaches. As in the case
of the systems of open coasts, estuary mouth sand storage
systems are in a state of continuous reorganization in re-
sponse to the postglacial rise in sea level.
Kraft et al. (1974) have attempted to trace the trans-
gressive history of the mouth of Delaware Bay by equat-
ing a series of transects across the modern bay with the
time series of profiles that would be expected at a single
point during transgression (Fig. 43). Here ridges first
appear as subaqueous tidal levees on the edge of tidal
flats marginal to tidal channels. Unlike the tidal sand
ridges of open shelf seas, these ridges migrate away from
their steep sides (Weil et al., 1974). As transgression
proceeds, the channels service a larger and larger tidal
prism and tend to widen. The effect on the levees is
erosion on the steep, channel-facing side, and aggrada-
tion on the gently sloping side facing away from the
channel. Weil et al. (1974) have attributed the submarine
levees of Delaware Bay tidal channels to density-driven
secondary flow associated with the tidal cycle (Chapter
10, Fig. 26).
Inner estuary channels tend to be ebb-dominated
perhaps because the upper estuary water mass tends to
flood as a sheet, but tends to preferentially ebb through
the channel system under the impetus of gravity dis-
charge. Further down the estuary, as levees begin to
build, the interfluves tend to become flood-dominated
channels in their own right, although the dominance of
channel and interfluves may locally be reversed. As
previously noted, retardation of the tidal wave in the
estuary results in a phase lag across the estuary mouth
shoal, causing an interdigitation of ebb and flood
channels, separated by partition ridges, across the crest
of the shoal. Thus ridges initiated in the upper estuary
may undergo a complex evolution as successive estuary
environments and associated flow regimes pass over
them. Individual ridges may maintain their integrity
through this process or be replaced by related forms
maintained by somewhat different mechanisms.
Modification of ridge morphology intensifies as the
regional shoreline passes, and the lower estuarine regime
is replaced by an open shelf regime. If the wave climate
is intense, then the outer surface of the estuary mouth
shoal is pushed back by erosional shoreface retreat in a
fashion similar to that transpiring on the adjacent main-
1500 YR B P.
LIKE PRESENT
KITTS HUMMOCK
LIKE PRESENT
PORT MAHON
3000 YR. B,P.
LIKE PRESENT
BOMBAY HOOK
5000 YR B. P.
LIKE PRESENT
NEW CASTLE
7000 YR BP
FIGURE 43. Late Holocene evolution of the mouth of Delaware Bay, as inferred from cross sections across the
modern bay. Apron of sand extending into bay mouth is assumed to have prograded up the bay concurrently with
the landward movements of the shoreline on either side. From Kraft et al. (1974).
513
SAND STORAGE ON ROCKY COASTS
301
FIGURE 44. Representative examples of inlet morphology.
(a) Fire Island Inlet, Long Island, a barrier-overlap inlet on a
drift-aligned coast. Littoral drift dominates the ebb tidal jet.
(b) Ocracoke Inlet. North Carolina. Nearly symmetrical inlet
land coast. Frequently, however, the retreat path of the
estuary is visible in the form of a cross-shelf channel and
a ridge on the updrift side of the channel. On the Dela-
ware inner shelf, such a ridge can be seen to mark the
retreat path of the shoal on the north side of the estuary
mouth, while the associated channel has formed by the
retreat of the main flood channel of the estuary mouth
(see Fig. 12, Chapter 15).
Coastal Inlets and Littoral Drift
The morphology of narrow estuary mouths and their
analogs, coastal inlets, depends on the relative strengths
of the river mouth or inlet jet, the wave-driven littoral
current, and the tidal- and wind-driven components of
the shelf flow field. Distributary mouths, subject to pe-
riodic flooding and entering relatively tideless, wave-
sheltered seas, consist of subaerial levees capped by a
lunate distributary mouth shoal (Fig. 42^4). As a con-
sequence of the Coriolis effect, flow is more intense on
the right-hand side of the channel looking downstream,
and as a consequence, the right-hand levee tends to ex-
tend itself farther seaward as in the case of Mississippi
distributary mouths. If the inlet faces an open or tidal
sea, then the wave- and tide-driven coastal flow is
diverted seaward around the ebb tidal jet (Todd, 1968)
and the shoal assumes a half-teardrop shape (Fig. 425).
On barrier coasts, the pattern of sand storage at tidal
inlets tends toward one of three basic patterns: overlap,
symmetrical, or offset inlets (Fig. 44). While these pat-
terns have long been recognized, the responsible trans-
port systems and sand budgets are imperfectly under-
stood (Hayes et al., 1970; Byrne et al., in press; Gold-
smith et al., in press). As noted by Byrne (personal com-
munication), the patterns appear to reflect the relative
intensities of gross littoral drift (both up and down the
coast) and net drift (the difference between mean annual
upcoast discharge and mean annual downcoast dis-
charge). If both the gross rates of drift and the net rate
are high, a disproportionately high volume of sand stor-
age may occur in the updrift barrier segment, and an
overlap barrier may result (Fig. 44^4). Where moderate
gross rates of drift are associated with a strong net rate
of drift, the situation favors a barrier offset inlet, in
which the storage of sand on the downdrift side of the
exterior shoal is favored (Fig. 44C). In one of the best
studied barrier offset inlets, Wachapreague Inlet on the
flow on a swash-aligned coast has resulted in sand storage in the
wave-protected interior (flood delta) shoal. Ebb tidal jet domi-
nates over littoral drift, (c) Absecon Inlet, New Jersey. Ebb-
dominated flow has resulted in sand storage on the downdrift
side of the inlet and an offset of the flanking barrier islands.
514
302
COASTAL SEDIMENTATION
Delmarva coast, the role of the lagoonal reservoir in
modifying the hydraulic characteristics of the inlet is of
paramount importance (Byrne et al., in press a). In
lagoon-inlet systems where the ratio of the in tertidal water
prism to the subtidal volume is very large, a strong time-
velocity asymmetry develops (see Postma, 1967). The
strongest currents occur just before high tide, when the
tidal channels have filled and the vast marsh surface is
beginning to flood, and just after high tide, when the
marshes are draining. Flows around low tide are weaker,
as they are associated with the much slower discharge
and recharge of the tidal creek system.
In addition, flood and ebb durations are dissimilar,
with a greater ebb duration. This phenomenon is a con-
sequence of the lagoonal basin's morphology and fric-
tional characteristics (Byrne et al., in press a), and has
been predicted by shallow water tidal theory for storage
systems with sloping banks (Mota-Oliveira, 1970; King,
1974). In physical terms, the hydraulic head generated
across the inlet by the flood tide is imposed on the deepest
part of the lagoon relatively early in the tidal cycle.
Here frictional retardation of flow is least efficient, and
the resulting sea surface slope propagates rapidly across
the lagoon, resulting in rapid water influx. The greatest
potential drop across the lagoon surface during the ebb
half-cycle occurs when the marsh surface is still un-
covering. Frictional retardation of flow is more effective
in the thin landward portions of the lagoonal water
column, and the ebb is prolonged.
As a result of these modifications of the tidal cycle,
the inlet operates in a bypassing mode. Sand is swept
into the inlet from the updrift side, but does not pene-
trate very far before it is swept out again, and the pro-
longed ebb carries it into the storage area on the down-
drift side of the external shoal. Here sand storage is en-
hanced by the refraction pattern of shoaling waves
(Goldsmith et al., in press).
Symmetrical inlets are favored by swash-aligned coasts,
where the ratio of the littoral component of wave power
to tidal power is relatively low (Fig. 445). Symmetrical
inlets, particularly those backed by lagoons with rela-
tively small intertidal prisms and relatively large sub-
tidal volumes, tend to store sand primarily in the
interior shoal within the lagoon.
SAND STORAGE ON ROCKY COASTS
Rocky coasts display the greatest complexity in three
dimensions. Rocky hinterlands in a mature state of dis-
section result in embayed coasts with deep reentrants
between rocky salients. If the substrate consists of folded
metamorphic rocks, then it may have a well-defined
anisotropy of its own and truly baroque patterns may
result (Fig. 45). The fields of wave refraction developed
over the seaward extensions of headlands result in fre-
quent reversals of the sense of littoral drift cells and
closely spaced alteration of zones of littoral drift diver-
gence and convergence. Because of the relative steepness
of the regional seaward slope and the resistant nature of
the substrate, wave energy is concentrated along a very
narrow intertidal zone. Waves breaking against vertical
surfaces can generate enormous instantaneous forces of
tens of metric tons per square meter (Zenkovitch, 1967,
p. 139). Rocky shores yield along planes of weakness to
become mantled with boulders under this assault (Fig.
46) and the intertidal and subtidal talus slopes become
grinding mills where attrition produces finer debris and
continues to grind it finer until, at about the grade of
medium sand, the immersed weight of grains is no longer
adequate to result in significant chipping or cracking —
as long as the particles are able to escape the proximity
and nutcracker behavior of coarser particles. The inter-
action of intertidal and shallow subtidal wave forces
with the three-dimensional complexity of rocky coasts
results in such erosional forms as stacks, arches, and sea
caves, and the constructional forms of looped, fringing,
recurved, and cuspate spits, and tombolos that have long
been the delight of coastal morphologists (Fig. 47). The
constructional forms constitute localized depositional re-
gressions and are usually comprised of sets and subsets
of beach ridges reflecting stages in the feature's growth.
If the net rate of sedimentation is sufficiently high rela-
tive to the rate of sea-level rise, these forms tend to
grow and coalesce, and will ultimately form a continu-
ous shoreface.
Rocky coasts are more nearly likely to be tectonically
active than low, unindurated coasts, other things being
equal, and the resistant character of their substrate may
result in delays in the adjustment of the incised equilib-
rium profile to the crustal movement, if this adjustment
is indeed attained. Comparison of rocky coasts from dif-
ferent parts of the world has revealed a continuum of
adjustment from coasts as irregular as the margins of
newly dammed reservoirs, to coasts whose adjustment
has been complete, so that, by a combination of head-
land truncation and the filling in of bays, the coastline
has been straightened in plan view and the shoreface
has received the characteristic exponential curvature.
This continuum led Davis (1909) and Johnson (1919)
to the concept of a cycle of coastal evolution in which,
after an initial relative movement of sea level, the shore-
line is straightened and the equilibrium profile passes
through a cycle of youth, maturity, and old age.
Zenkovitch (1967) has objected to the simplified assump-
tions of the model and suggests that three types of em-
515
70°00'
43°30
69°50'W
69°45'
FIGURE 45. Lpper: A portion of the coast of southern Maine. Bedrock is iso-
clinally folded schist and gneiss. Lower: Beginning of formation of constructional
shoreface and estuary mouth shoal at mouth of Kennebec River; see the upper diagram
for location.
303
516
304
COASTAL SEDIMENTATION
la
MASS MOVfMENT
FIGURE 46. Diagrammatic representation of major processes
of cliff retreat and evolution. (7a) Undercutting and rapid
removal of collapsed material, (lb) Undercutting and slow
removal of collapsed material. (2) Mass movement and removal
at various rates. From Davies (1973).
bayed coasts may be distinguished on the basis of the
relationship between the submarine slope and the equi-
librium profile generated on it, as follows: (1) deep-
water coasts where the submarine bottom passes immedi-
ately below the equilibrium profiles; (2) coasts with
deep-water capes, where this is true only off capes, and
(3) shallow-water coasts, where the submarine slope is
everywhere above the equilibrium profile. The term
"effective wave base" is probably best substituted for
equilibrium profile here, for Zenkovitch concludes that
sectors of coasts "above the profile of equilibrium" are
those sectors that develop forms of accumulation (sandy
beaches, barriers, spits, and tombolos; Fig. 47), and
that shallow water coasts develop the most complex ar-
ray of these features. Zenkovitch further traces subcycles
of coastal evolution caused by feedback between evolving
accumulation forms and the rocky substrate, or between
two forms, whereby the growth of some spits into
wave shadows behind headlands may distort their sub-
sequent pattern, and the growth of other spits may
shield and starve younger spits, or induce yet others
where none existed.
These subcycles are probably more common than the
Davis-Johnson cycle, which requires an isostatic crustal
movement or eustatic sea-level jump for rejuvenation.
They may be observed on all stable rocky coasts under-
going transgression by postglacial sea-level rise. Such
coasts probably do not evolve at all in the Davis-Johnson
sense, but undergo steady state subsidence in a state of
perpetual youth, maturity, or old age, depending on the
degree of induration of the substrate and the amplitude
of the inherited relief.
The relationship between the rate of sea-level rise
and the relief and induration of the substrate also deter-
mines the geometry of sediment storage (Fig. 48). Cores
off transgressed crystalline coasts of high relief might be
expected to reveal a residual rubble overlain by fine-
grained bay deposits. Overlying sand deposits of com-
plex shape would reflect the passage of the outer shore-
line with its array of accumulation forms. The upper
surface of the sand horizon will have been beveled at
least locally by shoreward profile translation, and off-
shore sands or muds may locally have accumulated over
the surface of marine erosion. Off high crystalline coasts,
the full sequence will rarely develop and will be com-
pletely missing off capes, where surf-rounded boulders
may litter bare rock surfaces for kilometers offshore.
Pocket beaches and spits may locally survive the trans-
gressive process relatively intact; a rock-tied spit cannot
retrograde with the ease of a low coast barrier.
On steep coasts transgressive deposits may be minimal.
On steep coasts with very narrow shelves, submarine
canyons may penetrate almost to the shoreface, to tap
the littoral drift, through such gravity processes as sand
creep. On steep deep-water coasts, prisms of beach
shingle intermittently cascade to bathyal depths down
steep rock slopes that may be erosion-modified fault
scarps; sediment passes through the coastal zone by
gravity bypassing (Fig. 49). As the coast is lower and
softer, so will the sequence more nearly resemble the
uniform sequence typical of the low coast transgression.
Bay muds will more nearly resemble lagoonal muds,
capped perhaps by nearly uniform sheets of backbarrier
and shelf sands instead of lenticular remnants of spits
and tombolos.
Regressive deposits occur on some rocky coasts, as a
consequence of the Late Holocene reduction in the rate
of sea-level rise, where sediment input is sufficient to
reverse the sense of shoreline migration. In extreme cases,
alluvial gravel cones may build out across the transgres-
sive deposits. Bouldery topset beds may pass into foreset
sands and then into bottomset muds within a few hun-
517
SUMMARY
305
FIGURE 47. Types of coastal accumulation
forms, according to Zenkovitch (1967). Fringing:
a. beach nourished from offshore; h, beach nour-
ished from alongshore; c, beach filling an indenta-
tion; d, cuspate beach with bilateral nourishment;
e, asymmetrical, cuspate beach with bilateral
nourishment, attached at one end; (, spit with
unilateral nourishment; g, arrow (spit with
bilateral nourishment); h, spit on smooth coast;
i, bay mouth barrier; \, midbay barrier; k,
tombolo; \, interisland tombolo, doubly attached;
m, looped spit with bilateral nourishment;
n, looped spit with unilateral nourishment;
o, cuspate spit, detached; p, barrier island;
q, barrier island resulting from cutting of inlet;
r, estuary mouth swash bar; s, barrier sequence.
Symbols: (1) mainland and active cliff; (2) dead
cliff and coast with beach; (3, 4) major and
minor transport directions.
dred meters. On steep, unstable coasts, such masses may
periodically slump down the submarine slopes to the
basin floor.
SUMMARY
In considering coastal sediment transport, it is conven-
ient to divide the movement of sand into an onshore-
offshore component and a coast-parallel component.
Onshore-offshore transport occurs in two provinces. In
the nearshore province of beach, longshore trough,
plunge point bar, and upper shoreface, onshore-offshore
transport is controlled by the regime of shoaling and
breaking waves. Breakpoint bars are initiated during
the waning phases of storms. During the ensuing fair-
weather period they tend to migrate onshore, and weld
518
306
COASTAL SEDIMENTATION
LAND py-^j OUTCROP [p%3 GRAVEL
□ SAND I I MUD
1 2 3
KILOMETERS
FIGURE 48. Hypothetical stratigraphy of a rocky coast undergoing transgression.
to the berm. The high, steep waves of storms tend to
strip sand from the beach and transport it out to the
surf zone, and the cycle begins anew. The cycle tends
to be linked to the cycle of seasons in that offshore
transport dominates during the period of winter storms,
while onshore transport tends to dominate during the
summer season of fair weather. The lower shoreface is a
second province subject to onshore-offshore transport.
The corresponding hydraulic regime is the zone of fric-
tion-dominated unidirectional flow that constitutes the
coastal boundary of the shelf flow field. During storms
(or peak tidal flows) velocity in this zone may be more
intense than in the zone of quasi-geostrophic flow further
offshore. Downwelling and a seaward component of bot-
tom flow may occur in this zone during some storm
flows, at the same time that sand is moving seaward in
the surf zone, so that sand is transported off the shore-
face altogether.
The interrelated behavior patterns of the zone of
shoaling and breaking waves and zone of friction-domi-
nated flows give rise on many coasts to a long-term cyclic
pattern of advance or retreat of the coastal profile. The
upper shoreface undergoes net aggradation and pro-
gradation over a period of years tending toward the ideal
wave-graded profile. A major storm or period of severe
storms will result in large-scale seaward transport of
sand, causing flattening and significant landward trans-
lation of the profile. On coasts experiencing a net littoral
drift surplus, fair-weather progradation is more effective
than storm erosion, and the profile translates seaward
and (in compensation for postglacial sea-level rise) up-
ward. On coasts experiencing a net littoral drift deficit,
the storm regime controls the offshore-onshore sand
budget, and the coastal profile undergoes landward and
upward translation through a process of erosional shore-
face retreat. The debris resulting from this process nour-
ishes the leading edge of the surficial sand sheet that
mantles the shelf.
The cycle of onshore-offshore transport is superim-
posed on a much more intensive flux of sand parallel to
the beach, under the impetus of the wave-driven littoral
flows, and wind- and tide-driven coastal currents. As a
result, there is an innate tendency toward two-dimen-
sionality of the shoreface, in that successive downcoast
profiles tend to be very similar. Headlands experience a
greater littoral wave energy density, greater breaker
angles, and decreasing littoral sand discharge along the
beach toward the adjoining bay. A pattern of transport
away from headlands toward bays is superimposed on a
regional direction of littoral sand transport determined
by the prevailing direction of deep-water wave approach.
The resulting alternation of littoral drift divergences and
convergences may impose a three-dimensionality on an
unconsolidated coast in the form of alternate cuspate
forelands and zetaform bays. Three-dimensionality may
also be inherited from the relief of a rocky surface under-
going transgression, or may be induced on an unconsoli-
dated coast in the form of constructional river mouths
and tidal inlets.
The beach and shoreface comprise major reservoirs of
sand in the coastal sediment transport system. During a
transgression, the superimposition of a straight, wave-
maintained upper shoreface on an irregular surface re-
sults in the formation of two shorelines. An inner, la-
519
SUMMARY
307
VAR AND
PAILLON
CANYONS
Recen: Mud L__JSand
Q <o- D e stocene M jc IffigVI Grcvel
■?^j?*
EIj
':' ^Tj
EStSiBi
w^
r^- wi^H
^&^UteJH H
»**
B^,"'l^lk1i1^
Ik; ^folfcfe
->.*
^^J
cH
L ^>
■1 - ■-»- vsMH
SHU 'dJft*^ "*■'** tH
Hi. •- v v. *,- *!■
D
FIGURE 49. Gravity bypassing on a recently formed continental
margin, Provencal coast of France. (A) Axial fades of the Var
and Paillon canyons. (B) Paillon River mouth (left) and pebble
beach, Baie des Anges, Nice. (C, D) Boulder (up to 50 cm) mud
admixtures at diving locality shown in (B),- depth 25 m. (E) Large
blocks of Jurassic limestone overgrown with Poseidonia near
Cap Ferrat. From Stanley (1969).
goonal shoreline approximates the intersection of still
water level with the dissected subaerial surface under-
going erosion. An outer, oceanic shoreline of barrier
spits and islands results from ( 1 ) the detachment of
drift-nourished, wave-maintained beaches from the main-
land as the rising sea floods the swales behind them, and
(2) the lateral propagation of the shoreface from head-
lands across the mouths of adjoining bays. Sand is also
stored in shoals that form at littoral drift convergences,
and in oblique-tending, shoreface-connected sand ridges
that form at the foot of the shoreface in response to the
storm wave regime and storm coastal currents. Where the
littoral drift system intersects with the river- and tide-
driven jets of river mouths and inlets, sand is stored in
arcuate, seaward-convex shoals whose crests bear com-
plex patterns of interdigitating ebb and flood channels,
separated by sand ridges that build into the intertidal
zone as swash platforms.
520
308
COASTAL SEDIMENTATION
REFERENCES
Allen, J. R. L. (1968a). Current Ripples. Amsterdam: North-Hol-
land, 433 pp.
Allen, J. R. L. (1968b). The nature and origin of bedform hier-
archies. Sedimentology, 10: 161-182.
Allersma, E. (1972). Mud on the oceanic shelf off Guiana. In
Symposium on Investigations and Resources oj the Caribbean Sea and
Adjacent Regions. UNESCO, Paris (Unipub, N.Y.), pp. 193-
203.
Assaf, G., R. Gerard, and A. L. Gordon (1971). Some mecha-
nisms of oceanic mixing revealed in aerial photographs.
J. Geophys. Res., 76(27): 6550-6572.
Bascom, W. N. (1951). The relationship between sand size and
beach slope. Trans. Am. Geophys. Union, 32: 866-874.
Beall, A. O. (1968). Sedimentary processes operative along the
western Louisiana shoreline. J. Sediment. Petrol., 38: 869-877.
Beardsley, R. C. and B. Butman (1974). Circulation on the New
England continental shelf: Response to strong winter storms.
Geol. Res. Lett., 1: 181-184.
Bernard, H. A. and R.J. Leblanc (1965). Resume of the Quater-
nary geology of the Northwest Gulf of Mexico Province.
In H. E. Wright, Jr. and D. J. Frey, eds., The Quaternary oj
the United States. Princton: Princeton Univ. Press, pp. 137-185.
Bruun, P. (1954). Migrating sand waves and sand humps, with
special reference to investigations carried on in the Danish
North Sea coast. Con/. Coastal Eng., 5th Proc, 460-468.
Bruun, P. (1962). Sea level rise as a cause of shore erosion. Proc.
ASCE J. Waterw. Harbors Div., 88: 117-130.
Byrne, R. J., P. Bullock, and D. G. Tyler (in press a). Response
characteristics of a tidal inlet: A case study. In M. O. Hayes,
ed., Second International Estuarine Conference Proc, Myrtle Beach,
S.C., Oct. 15-18, 1973: New York, Academic Press.
Byrne, R. J., J. T. De Alteris, and P. A. Bullock (in press b). Chan-
nel stability in tidal inlets. 14th Conf. Coastal Eng. Proc,
Copenhagen, 1974.
Caston, V. N. D. (1972). Linear sand banks in the southern North
Sea. Sedimentology, 18: 63-78.
Caston, V. N. D. and A. H. Stride (1970). Tidal sand movement
between some linear sand banks in the North Sea off north-
east Norfolk. Mar. Geol., 9: M38-M42.
Charnell, R. L. and D. V. Hansen (1974). Summary and Analysis of
Physical Oceanography Data from the New York Bight Apex Col-
lected During 1969-70. MESA Rept. 74-3, National Oceanic
and Atmospheric Administration. Washington, D.C.: U.S.
Govt. Printing Office, 44 pp.
Clifton, H. E., R. E. Hunter, and R. L. Phillips (1971). Deposi-
tional structures and processes in the non-barred, high energy
nearshore. J. Sediment. Petrol., 41: 651-670.
Colquhoun, D. J. (1969). Geomorphology oj the Lower Coastal Plain
oj South Carolina. MS-15, South Carolina Development Board,
Div. Geol., 36 pp.
Cook, D. O. (1969). Occurrence and geologic work of rip-currents
in southern California. J. Sediment. Petrol., 39: 781-786.
Cook, D. O. and D. S. Gorsline (1972). Field observations of sand
transport by shoaling waves. Mar. Geol., 13: 31-55.
Csanady, G. T. and J. T. Scott (1974). Baroclinic coastal jets in
Lake Ontario during IFYGL. J. Phys. Oceanogr., 4: 524-541.
Curray, J. R. (1964). Transgressions and regressions. In R. L.
Miller, ed., Papers in Marine Geology — Shepard Commemorative
Volume. New York: Macmillan, pp. 175-203.
Curray, J. R., F. J. Emmel, and P. J. S. Crampton (1969).
Lagunas costeras, un simposio. In Mem. Simp. Int. Lagunas
Costeras. UNAM-UNESCO, Nov. 28-30, 1967, Mexico, D.F.,
pp. 63-100.
Davies, J. L. (1958). Wave refraction and the evolution of curved
shorelines. Geogr. Stud., 5: 1-14.
Davies, J. L. (1973). Geographical Variation in Coastal Development.
New York: Hafner, 204 pp.
Davis, R. A. and W. T. Fox (1972). Coastal process and near-
shore sand bars. J. Sediment. Petrol., 42: 401^412.
Davis, R. A., W. T. Fox, M. O. Hayes, and J. C. Boothroyd
(1972). Comparison of ridge and runnel systems in tidal and
non-tidal environments. J. Sediment. Petrol., 32: 413—421.
Davis, W. M. (1909). Geographical Essays, 1954 Dover edition.
New York: Dover Publications, 777 pp.
Dolan, R. (1971). Coastal landforms: Crescentic and rhythmic.
Geol. Soc. Am. Bull. ,82: 177-180.
Duane, D. B., M. E. Field, E. P. Meisburger, D. J. P. Swift, and
S. J. Williams (1972). Linear shoals on the Atlantic inner
continental shelf, Florida to Long Island. In D. J. P. Swift,
D. B. Duane, and O. H. Pilkey, eds., Shelf Sediment Transport:
Process and Pattern. Stroudsburg, Pa.: Dowden, Hutchinson &
Ross, pp. 447^-99.
Ekman, V. W. (1905). On the influence of the earth's rotation on
ocean currents. Ark. J. Mat., Astron. Phys., 2: 1-53.
Emery, K. O. (1968). Relict sediments on continental shelves of
world. Am. Assoc. Pet. Geol. Bull., 52: 445^64.
Faller, A. J. (1963). An experimental study of the instability of
the laminar Ekman boundary layer. J. Fluid Mech., 15: 560-
576.
Faller, A. J. (1971). Oceanic turbulence and the Langmuir cir-
culations. Ann. Rev. Ecol. System., 2: 201-233.
Faller, A. J. and R. B. Kaylor (1966). A numerical study of the
instability of the laminar Ekman boundary layer. J. Atmos.
Sci., 23: 466^80.
Farmer, D. G. (1971). A computer simulation model of sedimen-
tation in a salt wedge estuary. Mar. Geol., 10: 133-143.
Fenneman, N. M. (1902). Development of the profile of equi-
librium of the subaqueous shore terrace. J. Geol., 10: 1-32.
Fisher, J. J. (1968). Barrier island formation: Discussion. Geol.
Soc. Am. Bull., 79: 1421-1426.
Fox, W. T. and R. A. Davis (1974). Beach Processes on the Oregon
Coast, July 1973. Tech. Rep. 12, Office of Na"al Research,
Washington, D.C., 85 pp.
Gilbert, G. K. (1890). Lake Bonneville. U.S. Geol. Surv. Mono. 1,
438 pp.
Goldsmith, U., R. J. Byrne, A. H. Sallenger, and R. Driecker
(in press). The influence of waves on the origin and develop-
ment of the offset coastal inlets of the southern Delmarva
Peninsula, Virginia. In M. O. Hayes, ed., Second International
Estuarine Conference Proceedings, Myrtle Beach, South Carolina,
Oct. 15-18, 1973. New York: Academic Press.
Halligan, G. H. (1906). Sand movement on the New South Wales
coast. Proc Limnol. Soc. NSW, 31: 619-640.
Harrison, W. and R. M. Alamo (1964). Dynamic Properties oj Im-
mersed Sand at Virginia Beach, Virginia, Coastal Eng. Res.
Cent., Tech. Memo. 9, pp. 1-52.
Hayes, M. O. (1967). Hurricanes as Geological Agents: Case Studies oj
Hurricanes Carla, 1961, and Cindy, 1963. Texas Bureau of Eco-
nomic Geology, Report of Investigation 61, 54 pp.
521
REFERENCES
309
Hayes, M. O., U. Goldsmith, and C. H. Hobbs (1970). Offset
coastal inlets. In 12th Conference Coastal Engineering Proceedings,
pp. 1187-1200.
Holliday, B. VV. (1971). Observations on the hydraulic regime of
the ridge and swale topography of the inner Virginia shelf.
Masters Thesis, Old Dominion University, Inst, of Oceanog-
raphy, Norfolk, Va.
Howard, J. D. and H. E. Reineck (1972). Physical and biogenic
sedimentary structures of the nearshore shelf. Senckenbergiana
Marit.,4: 81-124.
Hoyt, J. H. (1967). Barrier island formation. Ceol. Soc. Am. Bull.,
78: 1125-1136.
Hoyt, J. H. (1970). Development and migration of barrier islands,
northern Gulf of Mexico, discussion. Geol. Soc. Am. Bull., 81:
3779-3782.
Hoyt, J. H. and V.J. Henry (1967). Influence of island migration
on barrier island sedimentation. Geol. Soc. Am. Bull., 78: 77-
86.
Ingle, J. C. (1966). The Movement oj Beach Sand. New York:
American Elsevier, 221 pp.
Johnson, D. VV. (1919). Shore Processes and Shoreline Development.
New York: Hafner (1965 facsimile), 584 pp.
Johnson, J. VV. (1949). Scale effects in hydraulic models involving
wave motion. Trans. Am. Geophys. Union, 30: 517-527.
Johnson, J. VV. and P. S. Eagleson (1966). In A. J. Ippen, ed.,
Estuary and Coastline Hydrodynamics. New York: McGraw-Hill,
pp. 404-492.
Keulegan, G. H. (1948). An Experimental Study oj Submarine Sand
Bars. Beach Erosion Board Tech. Rep. 3, 42 pp.
King, C. A. M. (1972). Beaches and Coasts. New York: St. Martins
Press, 570 pp.
King, C. A. M. and VV. VV. Williams (1949). The formation and
movement of sand bars by wave action. Geogr. J., 113: 70-85.
King, D. B., Jr. (1974). The Dynamics oj Inlets and Bays. Tech.
Rep. 22, College of Engineering, University of Florida,
Gainesville, 82 pp.
Kraft, J. C, R. E. Sheridan, R. D. Moose, R. N. Strom, and C. B.
Weil (1974). Middle-Late Holocene evolution of the mor-
phology of a drowned estuary system — the Delaware Bay. In
International Symposium on Interrelationships oj Estuarine and Con-
tinental Shelj Sedimentation. Inst. Geol. du Bassin d'Aquitaine,
July 9-14, 1973, pp. 297-306.
Langfelder, J., D. Stafford, and M. Amein (1968). A Reconnaissance
oj Coastal Erosion in North Carolina. Dept. Civil Eng., North
Carolina State University, Raleigh, N.C., 127 pp.
Langford-Smith, T. and B. G. Thorn (1969). New South Wales
coastal morphology. J. Geol. Soc. Aust., 16: 572-580.
Langmuir, I. (1925). Surface motion of water induced by the
wind. Science, 87: 119-123.
Leopold, L. B., M. G. Wolman, and J. P. Miller (1964). Fluvial
Processes in Geomorphology. San Francisco: Freeman, 522 pp.
Longuet-Higgins, M. S. (1953). Mass transport in water waves.
Phil. Trans. Roy. Soc. London, 245: 535-581.
Ludwick, J. C. (in press). Tidal currents, sediment transport, and
sandbanks in Chesapeake Bay entrance, Virginia. In M. O.
Hayes, ed., Second International Estuarine Conj. Proc. Myrtle
Beach, S.C., Oct. 15-18, 1973: New York, Academic Press.
McGee, W. D. (1890). Encroachments of the sea. The Forum, 9:
437-449.
McHone, J. F., Jr. (1972). Morphologic time series from a sub-
marine sand ridge on the south Virginia coast. Masters The-
sis, Old Dominion University, Norfolk, Va., 59 pp.
McKinney, T. F., W. L. Stubblefield, and D. J. P. Swift (1974).
Large scale current lineations on the Great Egg shoal retreat
massif: Investigations by sidescan sonar. Mar. Geol. , 17 : 79-102.
McMaster, R. L. and L. E. Garrison (1967). A submerged Holo-
cene shoreline near Block Island, Rhode Island. Mar. Geol.,
75: 335-340.
May, J. P. and W. F. Tanner (1973). The littoral power gradient
and shoreline changes. In D. R. Coates, ed., Publications in
Geomorphology. Binghamton: State University of New York,
404 pp.
Miller, R. L. and J. M. Zeigler (1958). A model relating dynamics
and sediment pattern in equilibrium in the region of shoaling
waves, breaker zone, and foreshore. J. Geol., 66: 417-441.
Miller, R. L. and J. M. Zeigler (1964). A study of sediment dis-
tribution in the zone of shoaling waves. In R. L. Miller, ed.,
Papers in Marine Geology — Shepard Commemorative Volume. New
York: MacMillan, pp. 133-153.
Moody, D. W. (1964). Coastal morphology and processes in re-
lation to the development of submarine sand ridges off Beth-
any Beach, Delaware. Ph.D. Thesis, Johns Hopkins Univer-
sity, Baltimore, 167 pp.
Moore, G. T. (1970). Role of salt wedge in bar finger sand and
delta development. Am. Assoc. Pet. Geol. Bull., 54: 326-333.
Mota-Oliveira, I. B. (1970). Natural flushing ability in tidal in-
lets. Proc. 12th Coastal Eng. Conj., Washington, D.C., pp. 1827-
1845.
Murray, S. P. (1967). Control of grain dispersion by particle size
and wave state. J. Geol., 75: 612-634.
Murray, S. P. (1970). Bottom currents near the coast during Hur-
ricane Camille. J. Geophys. Res., 75: 4579-4582.
Neumann, G. and W. J. Pierson, Jr. (1966). Principles oj Physical
Oceanography. Englewood Cliffs, N.J.: Prentice Hall, 545 pp..
Nevesskii, E. N. (1969). Some data on the postglacial evolution
of Karkinit Bay and the accumulation of bottom sediments
within it. In V. V. Longinov, ed., Dynamics and Morphology oj
Seacoasts. Akademiya Nauk SSSR Trudy Instituta Okeano-
logii, Vol. 48, 371 pp., translated from Russian by Israel
Program for Scientific Translations; reproduced by the Clear-
inghouse for Federal Scientific and Technical Information,
Springfield, Va., pp. 92-110.
Niedoroda, A. W. and W. F. Tanner (1970). Preliminary study
of transverse bars. Mar. Geol., 9: 41-62.
Oaks, R. Q., Jr. and N. K. Coch (1963). Pleistocene sealevels,
southeastern Virginia. Science, 14X): 979-983.
Oertel, G. F. (1972). Sediment transport of estuary entrance
shoals and the formation of swash platforms. J. Sediment.
Petrol., 42: 858-863.
Oertel, G. F. and J. D. Howard (1972). Water circulation and
sedimentation at estuary entrances on the Georgia coast. In
D. J. P. Swift, D. B. Duane, and O. H. Pilkey, eds., Shelj
Sediment Transport: Process and Pattern. Stroudsburg, Pa.: Dow-
den, Hutchinson & Ross, pp. 411-427.
Otvos, E. G. (1969). A sub-recent beach ridge complex in southern
Louisiana. Geol. Soc. Am. Bull., 80: 2353-2358.
Otvos, E. G. (1970a). Development and migration of barrier
islands, northern Gulf of Mexico. Geol. Soc. Am. Bull., 81:
241-246.
Otvos, E. G. (1970b). Development and migration of barrier
islands, northern Gulf of Mexico: Reply. Geol. Soc. Am. Bull.,
81: 3783-3788.
Pierce, J. W. and D. J. Colquhoun (1970). Holocene evolution of
522
310
COASTAL SEDIMENTATION
a portion of the North Carolina coast. Geol. Soc. Am. Bull.,
81: 3697-3714.
Postma, H. (1967). Sediment transport and sedimentation in the
marine environment. In G. H. Lauff, ed., Estuaries. Washing-
ton, D.C.: Am. Assoc. Adv. Sci., pp. 158-180.
Price, W. A. (1954). Dynamic environments: Reconnaissance
mapping, geologic and geomorphic, of continental shelf of
Gulf of Mexico. Trans. Gulf Coast Assoc. Geol. Soc, 4: 75-107.
Price, W. A. (1955). Correlation of Shoreline Type with Offshore Bot-
tom Conditions. Dept. of Oceanography, Texas A&M Univer-
sity, Project 65: 75-107.
Reineck, H. E. and I. B. Singh (1971). Der Golf von Gaeta
(Tyrrhenisches Meer). III. Die gefuge von vorstrand — und
schelfsedimenten. Senckenbergiana Marit., 3: 135-183.
Robinson, A. H. W. (1966). Residual currents in relation to
shoreline evolution of the East Anglian coast. Mar. Geol., 4:
57-84.
Russell, R. C. H. and J. D. C. Osorio (1958). An experimental
investigation of drift profiles in a closed channel. Proc. 6th
Conf. Coastal Eng., pp. 171-183.
Schiffman, A. (1965). Energy measurements in the swash-surf
zone. Limnol. Oceanogr., 10: 255-260.
Schwartz, M. L. (1965). Laboratory study of sea level rise as a
cause of shore erosion. J. Geol., 73: 528-534.
Schwartz, M. L. (1967). The Bruun theory of sea level rise as a
cause of shore erosion. J. Geol., 75: 76-92.
Schwartz, M. L. (1968). The scale of shore erosion. J. Geol., 76:
508-517.
Schwartz, M. L. (1973). Barrier Islands. Stroudsburg, Pa.: Dow-
den, Hutchinson & Ross, 451 pp.
Shepard, F. P. (1950). Longshore-Bars and Longshore-Troughs. Beach
Erosion Board Tech. Memo. 15, 32 pp.
Silvester, R. (1974). Coastal Engineering II. New York: American
Elsevier, 338 pp.
Smith, J. D. and T. S. Hopkins (1972). Sediment transport on
the continental shelf off of Washington and Oregon in the
light of recent current meter measurements. In D. J. P.
Swift, D. B. Duane, and O. H. Pilkey, eds., Shelf Sediment
Transport: Process and Pattern. Stroudsburg, Pa.: Dowden,
Hutchinson & Ross, pp. 143-180.
Sonu, C.J. and J. L. van Beek (1971). Systematic beach changes
in the outer banks, North Carolina. J. Geol., 74: 416-425.
Sonu, C. J., J. M. McCloy, and D. S. McArthur (1967). Long-
shore currents and nearshore topographies. Proc. 10th Conf.
Coastal Eng., pp. 55-549.
Stahl, L., J. Koczan, and D. Swift (1974). Anatomy of a shore-
face-connected ridge system on the New Jersey shelf: Impli-
cations for genesis of the shelf surficial sand sheet. Geology,
2: 117-120.
Stanley, D. J. (1969). Submarine channel deposits and their fossil
analogs (fluxoturbidites). In D. J. Stanley, ed., The New Con-
cepts of Continental Margin Sedimentation. Washington, D.C.:
Am. Geol. Inst., pp. DJS-9-1 to DJS-9-17.
Strahler, A. (1963). The Earth Sciences. New York: Harper and
Row, 681 pp.
Stubblefield, W. L., J. W. Lavelle, T. F. McKinney, and D. J. P.
Swift (1975). Sediment response to the hydraulic regime
on the central New Jersey shelf. J. Sediment Petrol., 45:
337-358.
Swift, D. J. P. (1973). Delaware Shelf Valley: Estuary retreat
path, not drowned river valley. Geol. Soc. Am. Bull., 84: 2743-
2748.
Swift, D. J. P. (1975a). Barrier island genesis: Evidence from the
Middle Atlantic Shelf of North America. Sediment. Geol., 14:
1-43.
Swift, D. J. P. (1975b). Tidal sand ridges and shoal retreat
massifs. Mar. Geol., 18: 105-134.
Swift, D. J. P., D. B. Duane, and T. McKinney (1974). Ridge
and swale topography of the Middle Atlantic Bight: Secular
response to Holocene hydraulic regime. Mar. Geol. ,15:227-247.
Swift, D. J. P., B. W. Holliday, N. F. Avignone, and G. Shideler
(1972a). Anatomy of a shoreface ridge system, False Cape,
Virginia. Mar. Geol., 12: 59-84.
Swift, D.J. P., J. W. Kofoed, F. P. Saulsbury, and P. Sears (1972b).
Holocene evolution of the shelf surface, central and southern
Atlantic coast of North America. In D. J. P. Swift, D. B.
Duane, and O. H. Pilkey, eds., Shelf Sediment Transport: Process
and Pattern. Stroudsburg, Pa.: Dowden, Hutchinson & Ross,
pp. 499-574.
Swift, D.J. P., R. B. Sanford, C. E. Dill, Jr., and N. F. Avignone
(1971). Textural differentiation on the shoreface during ero-
sional retreat of an unconsolidated coast, Cape Henry to
Cape Hatteras, North Carolina. Sedimentology, 16: 221-250.
Swift, D. J. P. and P. Sears (1974). Estuarine and littoral deposi-
tional patterns in the surficial sand sheet, central and southern
Atlantic shelf of North America. In International Symposium on
Interrelationships of Estuarine and Continental Shelf Sedimentation.
Inst. Geol. du Bassin d'Aquitaine, Bordeaux, Mem. 7, pp.
171-189.
Tanner, W. F. (1961). Offshore shoals in an area of energy deficit.
J. Sediment. Petrol., 31: 87-95.
Tanner, W. F., R. G. Evans, and C. W. Holmes (1963). Low
energy coast near Cape Romano, Florida. J. Petrol., 33: 713-
722.
Todd, T. W. (1968). Dynamic diversion: Influence of longshore
current-tidal flow interaction on chenier and barrier island
plains. J. Sediment. Petrol., 38: 734-746.
Valentin, H. (1954). Der landverlust in Holderness, Ostengland
von 1852 bis 1952. Die Erde, 6: 296-315. See also, Land loss
at Holderness, 1852-1952. In J. D. Steers, ed., Applied Coastal
Morphology. Cambridge, Mass.: MIT Press, 1971, pp. 1 16-137.
Van Straaten, L. M. J. U. (1965). Coastal barrier deposits in
south and north Holland — in particular in the area around
Scheveningen and Ijmuden. Meded. Geol. Sticht., NS 17: 41-75.
Visher, G. S. (1969). Grain size distribution and depositional
processes. J. Sediment. Petrol., 34: 1074-1 106.
Weil, C. B., R. D. Moose, and R. E. Sheridan (1974). A model
for the evolution of linear tidal built sand ridges in Delaware
Bay, U.S.A. In G. Allen, ed., Estuary and Shelf Sedimentation:
A Symposium. University of Bordeaux, July 7973.
Wells, D. R. (1967). Beach equilibrium and second order wave
theory. J. Geophys. Res., 72: 497-509.
Wright, L. D. and J. M. Coleman (1972). River delta morphology:
Wave climate and the role of the subaqueous profile. Science,
176: 282-284.
Wright, L. D. and J. M. Coleman (1974). Mississippi River
mouth processes: Effluent dynamics and morphologic devel-
opment. J. Geol., 82: 751-778.
Zenkovitch, V. P. (1967). Processes of Coastal Development. New
York: Wiley, 738 pp.
523
49
Reprinted from: Marine Sediment Transport and Environmental Management,
D. J. Stanley and D. J. P. Swift, editors, John Wiley and Sons, Inc.,
ley
Chapter 15, 311-350.
Offprints from:
Marine Sediment Transport and Environmental Management
Edited by D. J. Stanley and D. J. P. Swift
Copyright 1976 by John Wiley & Sons, Inc
CHAPTER
15
Continental Shelf Sedimentation
DONALD J. P. SWIFT
Atlantic Oceanographic and Meteorological Laboratories, Miami, Florida
The preceding chapter considered in detail the nature of
hydraulic process and substrate response along the
coast. This chapter examines patterns of sedimentation
on the shelf as a whole. It reexamines the coastal
boundary of the shelf as a source of sediment for the
rest of the shelf, and as a zone which thus regulates the
rate and character of sedimentation on the shelf surface.
Chapter 14 described a "littoral energy fence" imposed
upon coastal sedimentation by the landward-directed
asymmetry of wave surge in shoaling water, which
causes sediment to be retained on the shoreface. This
chapter concerns itself with the mechanisms by which
this dynamic barrier is penetrated, along the shoreface
or at river mouths, and by which sediment is injected
into the shelf dispersal system. The relative efheiencies
of shoreface and river mouth bypassing during periods
of transgression on one hand, and during periods of
regression on the other are described. These varying
efheiencies lead to two distinct shelf regimes: a passive
regime in which the shelf sand sheet is generated by
erosional shoreface retreat (autochthonous sedimenta-
tion) and a more active regime in which river mouth
bypassing causes deposition across the shelf surface
(allochthonous sedimentation). The chapter analyzes
the transport patterns associated with these two regimes,
and the resulting patterns of morphology, stratigraphy,
and grain-size distribution. Portions of the material in
this chapter have been presented elsewhere (Swift,
1974).
524
MODELS OF SHELF SEDIMENTATION
One of the first comprehensive models for the genesis of
clastic sediments on continental shelves was a by-
product of Douglas Johnson's (1919, p. 211) attempt to
apply Davis' geomorphic cycle of youth, maturity, and
old age to the continental shelf (see p. 277). Johnson
saw the shelf water column and the shelf floor as a sys-
tem in dynamic equilibrium, in which the slope and
grain size of the sedimentary substrate at each point
control, and are controlled by, the flux of wave energy
into the bottom. He described the resulting surface as an
exponential curve in profile, concave up, with the
steeper segment being the shoreface. Grain size was
considered to decrease as a function of increasing depth
with distance from shore, as ,a consequence of the
diminishing input of wave energy into the seafloor.
The model derived its sediment from coastal erosion
rather than from river input, a more broadly applicable
interpretation than many subsequent textbooks have
realized.
Despite its qualitative expression and limited appli-
cability, the model was in advance of its time in its
dynamical systems approach. However, this model
could not withstand in its initial form the subsequent
flood of data on the characteristic of shelf sediments.
Shepard (1932) was the first to challenge it, noting that
nautical charts of the world's shelves bore notations
indicating that most shelves were veneered with a
311
312
C O N T I N K N T A I SHE! F SEDIMENTATION
complex mosaic of sediment types, rather than a simple
seaward-fining sheet. He suggested that these patches
were deposited during Pleistocene low stands of the sea,
rather than during Recent time. Emery (1952, 1968)
raised this concept to the status of a new conceptual
model. He classified shelf sediments on a genetic basis, as
autlugenic (glauconite or phosphorite), organic (fcramini-
fera, shells), residual (weathered from underlying rock),
relict (remnant from a different earlier environment
such as a now submerged beach or dune), and detrual,
which includes material now being supplied by rivers,
coastal erosion, and eolian or glacial activity. On most
shelves, a thin nearshore band of modern detrital sedi-
ment' is supposed to give way seaward to a relict sand
sheet veneering the shelf surface.
A third, more generalized model for shelf sedimenta-
tion has been primarily concerned with the resulting
stratigraphy. It incorporates elements of both the
Johnson and Emery models. Like the Johnson model, it
views the shelf surface as a dynamic system in a state of
equilibrium with a set of process variables. The rate of
sea-level change, however, is one of these variables;
hence the effects of post-Pleistocene sea-level rise, as
noted by Shepard and Emery, may be accounted for.
The model may be referred to as the transgression-
regression model, since it is generally expressed in these
terms, or the coastal model, since it focuses on the
behavior of this dynamic zone. It was first explicitly
formulated by Grabau (1913), and more recently by
Curray (1964) and Swift et al. (1972). In this model,
the rate of sediment input to the continental shelf S, the
character of the sediment G (grain size and mineralogy),
the rate of energy input E, the sense and rate of relative
sea-level change R, and slope L are seen as variables
that govern the sense of shoreline movement (trans-
gression or regression) and ultimately the character of
shelf deposits.
The relationship may be expressed in quasi-quantita-
tive form as
SG
E
R
L
oc T
The processes controlling shelf sedimentation are
much too complex to be adequately described by this
equation and there is no way to evaluate its variables
adequately. The expression is useful, however, in
helping to sort out relationships. The first term, SG E,
might be called the effective rate of coastal deposition.
It increases with increasing 5\ the rate at which sedi-
ment is delivered to the shore. It increases with increas-
ing grain size G. since coarser sediments are less easily
bypassed across the shelf. It decreases with increasing E,
the rate of wind and tidal energy input, since a more
rigorous hydraulic climate causes more sediment to be
bypassed across the shelf.
The second term, R/L, might be called the effective
rate of sea-level movement. It increases with increasing
/t, the absolute rate of sea-level movement (eustatic or
tectonic), but decreases with increasing slope, L, of the
coast. The steeper the slope, the greater the fall of sea
level must be in order for the coast to advance a given
distance. Also, with a greater slope, a greater volume of
sediment must be delivered to the shoreline in order for
the shoreline to prograde a given distance shoreward.
The equation tells us that the rate and sense of shore-
line movement, T, whether landward (negative) or
seaward (positive), depends on the relationship between
these two terms. Basic elements of the relationship are
presented graphically in Fig. 1, according to a scheme
of Curray (1964). In Fig. 2, the history of the Nayarit
coast of Mexico has been plotted.
The Coastal Boundary as a Filter: Shelf Sedimentary
Regimes
The fundamental determinants of shelf sedimentation
are the areal extent of the adjacent continent undergoing
denudation, and its relief, climate, and drainage pattern.
These factors control the quantity of sediment delivered
to the shoreline, and its textural and mineralogical
composition. However, the rate and sense of shoreline
movement, as determined by the parameters described
above, have a modulating effect on the shelf sedi-
mentary regime.
It is helpful to think of the coastline as a '"littoral
energy fence" (Allen, 1970b, p. 169) in which the net
landward flow associated with bottom wave surge tends
RELATIVE SEA LEVEL
FALLING SEA LEVEL
OR EMERGENCE
RISING SEA LEVEL
OR SUBSIDENCE
RAPID
SLOW
STABLE
SLOW
RAPID
FIGURE 1. Diagram of the effects of sea-level movement and
the rate of coastal deposition on lateral migration of the shoreline.
See text for explanation. From Curray (1964).
525
0
25
50
£ 75
o
k 100
<
STRAND PLAIN > ^> J[[
I C=3
MODERN
ALLUVIUM
_L
SHELF hACIES
MUDS
X
MODELS OF SHELF SEDIMENTATION
313
BASAL SANDS
_L
20
15
10
0 5
NAUTICAL
10
MILES
15
20
25
30
RELATIVE SEA LEVEL
FALLING SEA LEVEL
OR EMERGENCE
RISING SEA LEVEL
OR SUBMERGENCE
RAPIO
SLOW
STABLE
SLOW
RAPID
o
S i
' 1
\
\ I
\ 1
\l
4
Z g
O -J
(
l\
1 \
irs
i r
h-
2 i
id
O I
(9
X
\
T
i
i
■
•
\*
FIGURE 2. Above: Diagrammatic section off the Costa de Naya-
rit, Mexico. See Fig. 10 A, Chapter 14 for details of coastal stra-
tigraphy. Below: Schematic representation of shoreline' migration.
From Curray (1964).
to push sediment toward the shore. There are two basic
categories of "valve" which regulate the passage of
sediment through this dynamic coastal barrier into the
transport system of the shelf surface. The shoreface
may serve as a zone of sediment bypassing. The erosional
retreat of the shoreface during a marine transgression
bevels the subaerial surface being transgressed (Fig. 10A,
Chapter 14), and spreads the resultant debris as a thin
sheet over the shelf floor. The process by which the sedi-
ment is so transferred is described in the accompany-
ing text. The process is a passive and indirect one; the
sediment that is released has undergone long-term
storage as flood plain, lagoonal, or estuarine deposits,
or has been derived from an earlier cycle of sedi-
mentation.
A second, more active route by which sediment may
pass through the littoral energy fence is via the ebb tidal
jet or flood stage jet of a river mouth. Patterns of river
mouth bypassing are illustrated in Chapter 14 (Fig. 42).
River mouth bypassing is more direct than shoreface by-
passing, but sediment must still undergo storage. Sand
is stored in the throat of the river mouth, and fines are
stored in marginal marshes and mud flats until the
period of maximum river discharge, when the salt
wedge moves to the shoal crest, and stored sediment is
bypassed to the shoreface of the shoal front (Wright and
Coleman, 1974); see Chapter 14, Fig. 41. It may undergo
a second period of storage on the shoreface and inner
shelf until the period of maximum storm energy (Wright
and Coleman, 1973).
The mode of operation of these valves is dependent on
basic parameters of coastal sedimentation. The spacing
of river mouths is the fundamental determinant of the
relative roles of shoreface versus river mouth bypassing.
The character of the hydraulic climate is also im-
portant; an intense tidal regime increases the efficiency
of river mouth bypassing, whereas an intense wave
climate increases the efficiency of shoreface bypassing.
The rate and sense of coastal translation as described
in the preceding section strongly affect the relative
526
314
CONTINENTAL SHELF SEDIMENTATION
©
k
'X
RIVER
ESTUARY
SHELF
'^L"'^L"^CJ\. \^l1""^1"'^L shoreface
///////
t..»»»,»t
RAPID TRANSGRESSION
/ / / / / /
©
SLOW TRANSGRESSION
^TTT^^ a-""^
/ / /
J
/ / /
REGRESSION
FIGURE 3. Sense of net sediment transport for (a) rapid trans-
gression, (b) slow transgression, and (c) regression. Offshore com-
ponent of transport is exaggerated for continuity. See text for ex-
planation.
roles of river mouth and shoreface bypassing (Fig. 3).
Rapid transgression results in disequilibrium estuaries
which become sediment sinks (see Chapter 14, Fig. 42
and associated text), and shoreface bypassing must
dominate (Fig. 3a). The resulting deposits consist of a
transient veneer of surf fallout on the upper shoreface,
and the residual sand sheet of the lower shoreface and
adjacent shelf (see Chapter 14, Fig. 8 and discussion,
p. 265. These two deposits correspond to the nearshorc
modern sand and shelf relict sand, respectively, of Emery
(1968). Both deposits are relict in the sense that they have
been eroded from a local, pre-Recent substrate, and both
are modern in the sense that they have been redeposited
under the present hydraulic regime. They are, in fact,
palimpsest sediments (Swift et al., 1971) since they have
petrographic attributes resulting from both the present
and the earlier depositional environment. The term
relict is best reserved for those specific textural attributes
reflecting the earlier regime. Perhaps the most effective
term for describing the relationship of these materials
to the present depositional cycle is autochthonous (of
local origin: Xaumann, 1858), and a shelf sedimentary
regime characterized by rapid transgression and by-
passing via shoreface erosion is described in this chapter
as a regime of autochthonous shelf sedimentation.
With a slower rate of translation (Fig. 3b), estuaries
can equilibrate to their tidal prisms (see Chapter 14, Fig.
42 and associated text). River mouth as well as shoreface
bypassing becomes a significant source of sediment.
More subtle, but equally important, is the effect of a slow
transgression on the grain size of bypassed sediment.
With a slower rate of shoreline translation the intra-
coastal zone of estuaries and lagoons can aggrade nearly
to mean sea level. The resulting surface of salt marshes
(or in low latitudes, mangrove swamps), threaded by
high-energy channels, tends to serve as a 'low-pass, or
bandpass filter, in the sense that the finer fraction of the
sediment load is preferentially bypassed, while the
coarsest fraction (and in the bandpass case, the finest
fraction as well) is preferentially trapped out. In this
process, migrating channels tend to select coarse ma-
terials for permanent burial in their axes. The surfaces
of the tidal interfluves receive the finest material for
prolonged storage or permanent burial. However, fine
sands and silts are deposited as overbank levees and
tend to be reentrained by the migrating channels;
hence they have the highest probability of being by-
passed to the shelf surface. This material is sufficiently
fine to travel in suspension for long distances.
The estuaries of the Georgia coast have built a gently
sloping shoreface of fine to very fine sand up to 20 km
wide (Pilkey and Frankenberg, 1964; Henry and Hoyt,
1968); see Chapter 14, Fig. 13. This unusually wide
and broad shoreface may be built by the combined
contributions of shoreface and river mouth bypassing.
Recent studies (Visher and Howard, 1974) suggest
that the reversing tidal flows within the estuary consti-
tute an efficient mechanism for the sorting of sands
into size fractions, the spatial segregation of these
fractions, and the bypassing of the finest sand out onto
the shoreface.
There is clearly a contribution of sand from shoreface
erosion; however, shoreface sands, like the adjacent
shelf sands, contain trace amounts of phosphorite
(Pilkey and Field, 1972), indicating erosion of the
Miocene strata which underlie the shoreface between
the closely spaced estuaries, and which floor the deep
scour channels of the estuary mouths (Barby and Hoyt,
1964).
As the sense of coastal translation passes through
stillstand to progradation (Fig. 3c), the shoreface
becomes a sink rather than a mechanism for bypassing.
Distributary mouths must further partition their pre-
filtered load between sand sulliciently coarse to be
captured by the littoral drift and buried on the shoreface,
and sand fine enough to escape in suspension in the ebb
527
MODELS OF SHELF SEDIMENTATION
315
Channtl
(C-F)
Cloyey silt and silly cloy
(C) - coortt gr
Sond | (M) ~ madium V
(F) - fin* gr
1VF)- vtry fin* gr.
FIGURE 4. Schematic illustration of the depositional environ-
ments and sedimentary Jacies of the Niger Delta and Niger shelf.
Progressive size sorting of sediment results in a decrease in grain
size through successive depositional environments in a seaward
direction. From Allen (1970a).
tidal jet, and be entrained into the shelf dispersal sys-
tem. The shoreface behaves more nearly as a sediment
trap, and bypassing occurs primarily through river
mouths.
The Niger-Benue delta system is one of the best
studied examples of differential sediment bypassing
through a prograding, deltaic environment (Allen,
1964, 1970a). The Niger-Benue river system delivers
about 0.9 X 106 m3 of bed load sediment and about
16 X 106 m3 of suspended sediment (Allen, 1964) to its
delta each year. During peak discharge from September
to May, average flow velocities range from 50 to 135
cm sec, and gravel as well as sand is in violent trans-
port. During low stages, flow velocities decrease to
37 to 82 cm sec, enough to transport sand and silt.
In the higher part of the flood plain, the Niger is
braided; in the rest the Niger shows large meanders
(Fig. 4). During high stages, levees are overtopped,
crevasse develops, and bottom lands are flooded. Gravel
and coarse sand are deposited as a substratum of braid
bars and meander point bars, respectively, and are
veneered with a top stratum of overbank clays. Silt
undergoes temporary deposition in levees in the lower
flood plain but these tend to be undermined, so that
their deposits reenter the transport system.
Thus the flood plain environment serves as a skewed
bandpass filter, with preferential bypassing of the
medium and finer grades, preferential entrapment of
the finest material over bank, and much coarse material
being deposited in channel axes. This process continues
through the tidal swamp environment, where the
entrapment of fines dominates. Reversing tidal flows
528
316
CONTINENTAL SHELF SEDIMENTATION
generate velocities of 40 to 180 cm, sec in tidal creeks,
enough to move sand and gravel. Entrapment of fines
overbank in the mangrove swamps is enhanced by the
phenomena of slack high water and the prolonged
period of reduced velocity associated with it. Fines then
deposited begin to compact, and require greater veloci-
ties to erode them than served to permit their deposition.
Major channels, which pass through the intertidal
environment to the sea, must store their coarser sedi-
ment during low water stages at the foot of the salt
wedge, where the landward-inclined surface of zero net
motion intersects the channel floor. During high water
stages, stored bottom sediment must be rhythmically
flushed out of the estuary mouth by the tidal cycle.
Sand coarser than the effective suspension threshold of
230 /u (Bagnold, 1966) will be deposited on the arcuate
estuary mouth shoals, where, after a prolonged period of
residence in the sand circulation cells of the shoal
(see Chapter 10, p. 177), it leaks into the downcoast lit-
toral drift system. Finer sand is entrained into suspension
by large-scale top-to-bottom turbulence in the high-
velocity estuary throat (Wright and Coleman, 1974) and
will be swept seaward with the ebb tidal jet, to rain out on
the inner shelf (Todd, 1968) where it is accessible to
distribution by the shelf hydraulic regime.
Shelves undergoing slow transgression or regression
(Figs. 3b,c) thus experience a contrasting regime of
allochthonous shelf sedimentation (Naumann, 1858)
characterized by significant river mouth bypassing. In
this regime there is a massive influx of river sediment
whose grain size has been modified by passage through
the coastal zone. Sheets of mobile fine sand and mud
stretch from the coast toward the shelf edge. Shorefaces
are broad and gentle and merge imperceptibly with a
shallow inner shelf.
Sedimentation on tectonic continental margins is a
special case of allochthonous shelf sedimentation so
distinctive as to warrant designation as a third and
equal category. Shelves subject to such a regime are
narrow and steep, if developed at all. River mouth
bypassing and fractionation of the sediment load occur
here also. Rubble subaerial fans may pass over short
distances into sandy marine deltas with bottomset
mud beds. Gravity dispersal becomes a significant
coastal bypassing mechanism. Submarine canyons may
cut completely across narrow shelves to tap the littoral
drift (Shepard, 1973, p. 140) and divert sand seaward
by slow or rapid mass movements. Where shelves are
altogether lacking, coarse littoral prisms cascade inter-
mittently down slopes that are nearly tectonic surfaces,
to bathyal depths (Stanley, 1969). Tectonic regimes on
incipient shelves are beyond the scope of this chapter,
partly because they are more appropriately discussed in
the chapter on slope sedimentation, and partly because
of our ignorance, as this category is one of the last to be
better known in the rock record (Stanley, 1969) than
in modern environments.
AUTOCHTHONOUS PATTERNS OF SEDIMENTATION
Morphologic-Stratigraphic Patterns
Shelves undergoing autochthonous sedimentation char-
acteristically have a varied and systematic pattern of
relief. The pattern tends to be correlated with both the
distribution of surficial sediment and the internal
structure of the surficial sediment mantle, and hence is a
morphologic-stratigraphic pattern. On shelves of high
relief, the pre-Holocene surface is exposed at the surface
over wide areas, and constitutes an additional control
of the pattern.
Survival of Subaerial Patterns
On high-latitude shelves, relief may exceed 200 m. Much
of this relief may be the consequence of pre-Holocene
fluvial and glacial erosion of a crystalline substrate
(Holtedahl, 1940, 1958), and of the dissection of flat-
lying or gently inclined Cenozoic strata into cuestas and
plateaulike remnants. On the North American Atlantic
shelf, the Fall Line, where turbulent piedmont streams
pass onto the coastal plain strata, intersects the shoreline
at New Jersey (Fig. 5). To the north, the Fall Line
cuesta, of gently inclined coastal plain strata, forms
first islands (Long Island, Nantucket), then offshore
banks (Georges Banks, the Nova Scotian Banks). Basins
landward of the drowned Fall Line (Long Island Sound,
the Gulf of Maine, the Nova Scotian basins) have inner
margins of crystalline rock thinly veneered with coarse
detritus. The basin centers have a lower stratum of
glacial lake deposits overlain by Holocene marine mud.
Shelves of lower relief tend to be divided into broad,
flat, plateaulike compartments by shelf valleys excavated
during Quaternary low stands of the sea (Figs. 6 and 7).
The outer margins of such shelves tend to consist of low-
stand deltas, whose fronts are seaward-bulging shelf-edge
scarps and whose landward margins may be marked by
V-shaped, seaward-facing scarps that rise to the level of
the inner shelf.
Subaerial morphologic elements smaller in scale
than cuestas, basins, and shelf valleys seem in general
to have been destroyed by erosional retreat of the shore-
face, and the larger scale elements have often been
subtly but pervasively modified by this process. This
point can usually be demonstrated by a comparison of
529
AUTOCHTHONOUS PATTERNS OF SEDIMENTATION
317
47° 67°
45°64c
38° 71°
70° 37°
FIGURE 5. Bathymetry of the Gulf of Maine and Nova Scotian shelf. Dashed line is submerged
extension of the Fall Line, separating gently dipping coastal strata from the crystalline substrate of
the Appalachian orogenic belt. From Uchupi (1968).
shelf morphology with the morphology of the associated
subaerial surface. The coastal plains of the world bear a
delicate fabric of high-stand scarps, separated by
terraces overprinted with beach ridge fields, commonly
dating from the last interglacial, or a high stand during
the Wurm-Wisconsin glacial epoch (see, e.g., Colquhoun,
1969; Oaks and Coch, 1963; Bernard and Le Blanc,
1965). However, most submarine shelves are relatively
featureless (the Aquitaine shelf: Caralp et al., 1972) or
bear complex patterns of sand ridges that are the conse-
quence of marine systems of sediment transport initiated
after the passage of the shoreline (ridge and swale
topography of Fig. 6).
Major exceptions to this rule are the littoral bed forms
of carbonate coasts; fringing reefs, beach rock, and
calcarenite dunes cement as they form, and are far
more resistant to the destruction during the passage of
the shoreline. Carbonate littoral and sublittoral features
have been reported from many shelves (Kaye, 1959;
Ginsburg and James, 1974; Van Andel and Veevers,
530
<3
S
T .
Z ~
«■£
•a st
** s
« *
a. 3
318
531
3
1
vS
4
14
■5 5
1 5
8 8
«• a.
»4
S <U
4PJ
r- -a
to
at
532
319
320
CONTINENTAL SHELF SEDIMENTATION
1967; Stanley et al., 1968; Sarnthein, 1972). Other ex-
ceptions are the "perilittoral" deltas of terrigenous sand
which seem to have survived transgression in the Gulf of
Mexico (Curray, 1964, p. 299). The latter are large
river-fed spits that grow in the direction of littoral drift,
causing the river to flow parallel to the coast before it
breaks out to the sea. End moraines have survived on the
New England-Canadian shelf (King et al., 1972), but
they were apparently emplaced seaward of the shoreline
by grounded ice; King notes the vulnerability of glacial
deposits on the present shoreline to glacial attack. The
rinnentaler of the North Sea (subicestream channels)
may likewise have been formed by an ice sheet seaward of
the shoreline (Brouwer, 1964).
Survival of Nearshore Marine Patterns
the surficial sand sheet. The most characteristic
aspect of shelves undergoing autochthonous sedimenta-
tion is the discontinuous surficial sand sheet 0 to 10 m
thick, deposited during the erosional retreat of the
shoreface during the Holocene transgression (Fig. 8).
On flat-lying constructional shelves such as the Middle
Atlantic Bight of North America, relief elements on the
surface of this sheet formed as the zones of nearshore
sand storage (estuary mouth and cape extension shoals;
shoreface-connected sand ridges: Swift et al., 1972), and
both the surface morphology and the internal structure
of the sand sheet bear little relation to the surface
morphology and the internal structure of the older strata
beneath (McClennen and McMaster, 1971). On shelves
of greater relief, the surficial sediment blanket occurs as a
thin drape over topographic highs, broken by substrate
outcrops. In the adjoining basins, marginal sands, shed
by highs, pass laterally into deposits of mud (Fig. 9).
The stratigraphy of the surficial deposits of the shelf is
twofold. On shelves bordering low coasts a lower unit
of fine sands and mud was deposited in the belt of
lagoons and estuaries in advance of the main shoreline
(Fig. 10). Its lower surface is ribbed with estuarine
channel fillings that fill the buried drainage pattern of
the pre-Recent substrate (Sheridan et al., 1974; Emery,
1968). Meandering of these channels in response to
DEPTH, FEET
— 0 MIW
— 20
500
I
1000
I
WEST
FEET
828
— 80
100
— 120
140
— 160
— 180
— 200
EAST
816 L29
•/> (*{{.'*, , ,
818
DRILL HOLE
DISCONFORMITY
MEDIUM-FINE SAND
PEBBLY SAND
FINE, SILTY SAND
SILTY CLAY
A A' PLEISTOCENE-HOLOCENE CONTACT
B B' TERTIARY-QUATERNARY CONTACT
MOTTLED, DESICCATED,
L^~ SILTY CLAY
© RADIOCARBON DATE ON PROFILE
FIGURE 8. Surficial sand of the inner New Jersey shelf. Stratum gressiou. The barrier superstructure is represented in this se-
Hi is a shelf sand. Stratum H2 is a backbarrier sand. Stratum Hi quence by an unconformity; its forward face underwent continuous
is a lagoonal mud. Thick zone in Hi is inferred to be a filled tidal erosional retreat (Chapter 14, Fig. 10A) and the resulting debris
scour channel whose axis is normal to the plane of the diagram. The accumulated seaward of the shoreface as the leading edge of Hi.
sequence was produced by coastal retreat during the Holocene trans- From Stahl et al. (1974).
533
NNW
GULLY TROUGH
SABLE ISLAND
BANK
AUTOCHTHONOUS PATTERNS OF SEDIMENTATION
SSE
321
Maximum late-Wisconsin
Low stand of seo level
outwash plain 66 fbthom terrace 20,000 to 18.000 years BP
^SUSPENDED SEDIMENT ^
JL— /Pv-r ?<RrK tw?3s* i^ifc»™ — *-
3g^j ^K0> m$ Cfer
FIGL'RE 9. Evolution oj the surficial sand shelf on a glaciated shelf of appreciable relief {Nova
Scotia shelf; compare with Fig. 5). From Stanley (1969)-
tidal flow has served to reduce the relief of the buried
subaerial surface. Tidal inlets scour trenches into the
lagoonal stratum, then backfill these trenches as they
migrate downdrift (Hoyt and Henry, 1967; Kumar and
Sanders, 1970). The lagoonal carpet is itself discon-
tinuous; Pleistocene beach ridges and other topographic
highs protrude through the lagoonal deposits as penin-
sulas or islands during their formation and are sheared off
by shoreface retreat during passage of the main shore-
line (Sheridan et al., 1974). On rocky coasts, lagoonal
and estuarine deposits are confined to shelf valleys.
Passage of the main shoreline results in destruction of
the barrier and the upper part of the lagoonal sequence,
and in the deposition of a second major stratum, a sheet
of residual sand. This sand sheet overlies a surface of
marine erosion whose areal geology is a patchwork of
remnant lagoonal deposits and older substrate. On
shelf sectors where the lagoonal carpet is well developed,
this sand must travel from eroding headlands along the
shoreface of retreating barriers, before being spread over
the lagoonal carpet; or it is released as the retreating
shoreface cuts into tidal inlet fills, or into estuarine
channel sands scoured out of the pre-Recent substrate
(Andrews et al., 1973). On shelves with poorly developed
lagoonal strata, the retreating shoreface may be incised
all the way through the lagoonal deposits and into
534
322 CONTINENTAL SHELF SEDIMENTATION
(HO)SUBAERIAL GRAVEL
Qjjj MAINLAND MARSH
(H2) TIDAL CREEK LAG GRAVEL
(H3)LAGOONAL SANDY MUD
(H4) MARSH, PEAT, WASHOVER SAND
(H5)lNLET SAND
(H6) DUNE, WASHOVER SAND
(H8) BASAL SHELF GRAVEL
(H9) SHELF SAND
(ni^) SHELF MUD
(Pj) PLEISTOCENE LAGOONAL SANDY MUD
(P2) PLEISTOCENE BARRIER SAND
(P3) PLEISTOCENE INNER SHELF MUD
(H7) BEACH, SHOREFACE SAND
FIGURE 10. Stratigrapbic model for a low coast undergoing erosional shoreface retreat,
Pleistocene sands, whose erosion provides material for
the surficial sand sheet.
The basal layer of the surficial sand sheet is a thin.
discontinuous gravel (Powers and Kinsman, 1953;
Belderson and Stride, 1966; Yeenstra, 1969; Xorris,
1972) or shell hash rich in backbarrier and beach
species (Fischer, 1961; Merrill et al., 1965; Milliman
and Emery, 1968; Field, 1974). More exotic clasts are
clay pebbles eroded from Early Holocene lagoonal
deposits, elephant teeth (Whitmore et al., 1967), and
concretions from Tertiary strata (Stanley et al., 1967).
The basal gravel is rarely more than a meter thick. It is
SUCCESSIVE POSITIONS OF SHORE EACE
DURING INTERMITTENT TRANSGRESSION
TRANSGRESSION (7-11)
AND BARRIER RETREAT
STIUSTAND 16,7)
WITH BARRIER NOURISHMENT
AND UPWARD GROWTH
TRANSGRESSION (1-6]
AND BARRIER RETREAT
LAGOON
TIDAL CHANNEL
BARRIER
RIDGED SAND SHEET
(DESTRUCTIONAll
TRUNCATED
BARRIER
SAND
LAGOONAL DEPOSITS
PRE-HOLOCENE SUBSTRATE
FIGURE 1 1. Above: Schematic illustration of intermittent shoreface retreat. As
shoreface profile translates primarily landward in response to rising sea level, material
eroded from shoreface accumulates on adjacent shelf as ridged sand sheet. Periods of
primarily vertical translation of profile followed by periods of resumed landuard trans-
lation result in truncated scarp. Below: Resulting stratigraphy. From Swift et al. (197 i).
535
AUTOCHTHONOUS PATTERNS OF SEDIMENTATION
323
commonly overlain by 1 to 10 m of sand, with a sub-
littoral molluscan fauna (Shideler et al., 1972).
terraces and scarps. Most continental shelves are
terraced, with terraces separated by scarps 10 m or more
in height (Fig. 6). Their counterparts on the adjacent
subaerial coastal plains mark Quaternary (or earlier)
high stands of the sea. Shelf scarps appear to have
resulted primarily from stillstands of the returning
Holocene sea, although they may in some cases be
reoccupied Pleistocene shorelines. On the Georgia coast,
for example, the modern barrier island system is perched
on the forward face of a Pleistocene shoreline, and the
modern tidal inlets are reoccupied Pleistocene inlets
(Hoyt and Hails, 1967).
Shelf scarps are drowned shorelines only in the
broadest sense; more specifically, they are relict lower
shorefaces. To form such a scarp would require a period
of near stillstand during a general transgression. The
shoreface profile will translate more nearly upward
than landward (Fig. 11) during this period, by means
of upper shoreface and barrier surface aggradation.
At the resumption of rapid transgression, the super-
structure of the stillstand barrier will resume its land-
ward migration through the process of shoreface
erosion and storm washover, leaving behind a truncated
lower shoreface. If both the lower and upper shoreface
undergo aggradation during the stillstand period, so
that the ideal profile is realized at all times, then there
will be no surface expression of the stillstand shoreline,
although seismic profiles may reveal a buried scarp
(Stanley et al., 1968).
SHELF VALLEY COMPLEXES AND SHOAL RETREAT MASSIFS.
A second nearshore marine pattern of relief and sedi-
ment distribution that may survive from the nearshore
environment during a marine transgression is a shelf
valley complex. This term refers to the groups of
morphologic elements that occur along the paths of
retreat of estuary mouths on autochthonous shelves.
Shelf valley complexes are composed of deltas, shelf
valleys, and shoal retreat massifs (Figs. 6 and 12). A
shoal retreat massif is a broad, shelf-transverse sand
ridge of subdued relief that marks the retreat path of a
zone of littoral drift convergence (Swift et al., 1972).
It may be dissected by subsequent storm or tidal flows
into a cross-shelf sequence of smaller coast-parallel
ridges, and the term massif is used in the sense of a
composite topographic high, itself consisting of smaller
highs.
Such complexes are locally well developed on low-
relief shelves such as the Middle Atlantic Bight of
North America. Here they are largely constructional
features molded into the Holocene sand sheet. The sand
MODERN ESTUARY
MOUTH SHOAL,
TIDAL CHANNELS
PAIRED FLOOD
CHANNEL RETREAT
TRACK, ESTUARINE
SHOAL-RETREAT MASSIF
40M SCARP
TRANSGRESSED
CUSPATE DELTA;
(CAPE SHOAL-
RETREAT MASSIF)
60M SCARP
FIGURE 12. The Delaware shelf valley complex, Delaware
shelf of North America. Southward littoral drift of New Jersey
coastal compartment is injected into reversing tidal stream of
mouth of Chesapeake Bay. The resulting shoal is stabilized as a
system of interdigitating ebb and flood channels, north of the main
couplet of a mutually evasive ebb and flood channel. The shelf
valley complex seaward of the bay mouth is the retreat path of the
bay mouth sedimentary regime through Holocene time. Retreat of
the main flood channel has excavated the Delaware shelf valley;
retreat of the bay mouth shoal has left a seaward-trending shoal
retreat massif on the shelf valley's north flank. From Swift (1973).
sheet tends to completely fill the former subaerial valley
cut by the river into the Pleistocene strata, and the
shelf valley complex and the buried river channel may
not everywhere coincide (Fig. 13).
Shelf valley complexes are built in serial fashion by
the retreating shoreline. It is important to remember
that the last high-energy depositional environment ex-
perienced by any given segment was the nearshore zone.
As a consequence of remolding of preexisting deposits in
this zone, elements of shelf valley complexes are not
always what they seem. For instance, in Fig. 12, the
topographic characteristics of the Delaware midshelf
536
324
CONTIN KNTAI. SIIKI.h SEDIMENTATION
REHOBOTH BEACH SECTION
HEN ♦ CHICKENS SHOAL
a.
ED LAGOONAL MUDS, CLAYS
(H) ESTUARINE - SHALLOW MARINE SILTS
E3 NEARSHORE MARINE SANDS
55 GRAVELS
Q SHELLS, ENSIS, MULINIA
60" 200"
DISTANCE (NAUTICAL MILES)
6 7 B 9
DISTANCE (KILOMETERS)
FIGURE 13. Section across the head of the Delaware shelf valley
complex based on vibracores and a 5.5 kHz seismic profile. Marine
sand sheet with constructional tidal topography rests on Holocene
lagoonal and older Pleistocene deposits. Delaware shelf valley
occurs entirely within Pleistocene sands. Note offset between shelf
valley and buried river channel. Prom Sheridan et al. (1974).
delta suggest that the surface of this stillstand feature was
successively remodeled at the resumption of transgression,
first as a retreating cuspate foreland, then as cape shoal
retreat massif, as illustrated in Chapter 14, Fig. 22. After
a further period of stillstand indicated by a 60 m scarp,
the coastal regime again changed, and the Delaware
River mouth resumed retreat, this time as an estuary. The
retreat path of this estuary mouth consists of a sharply de-
fined submarine channel (shelf valley) flanked by a shoal
retreat massif. The origins of these two features are
easily deduced from uniformitarian reasoning. The
shoal retreat massif may be traced into the modern
north side shoal of the Delaware estuary mouth. This
shoal is a sink for the littoral drift of the New Jersey
coastal compartment, and is stabilized by a system of
interdigitating ebb and flood channels. The shelf valley
may be traced into the flood channel of a large ebb
channel-flood channel couplet on the south side of the
estuary mouth that accommodates most of the tidal
discharge.
On the central and southern Atlantic shelf of North
America, four basic morphologic provinces may be
described on the basis of constructional morphologic
elements inherited from the retreating shoreline (Fig.
14). In the Middle Atlantic Bight (Fig. 6), widely
spaced master streams have resulted in widely spaced
shelf valley complexes. The plateaulike intcrfluves
between the shelf valley complexes bear ridge fields
that were also generated by shoreface retreat (see
Chapter 14, Fig. 28).
The more intense wave climate experienced by the
Carolina salient has elicited a different response from
the retreating river mouths. Capes Romain, Fear,
Lookout, and Hattcras may have originally been
cuspate deltas, associated with the Peedee, Cape Fear,
Neuse, and Pamlico rivers (Chapter 14, Fig. 26). Retreat
of these forelands has left large widely spaced shoal retreat
massifs. South of Cape Romain the retreat of small,
closely spaced cuspate forelands has generated a blanket
of coalescing shoal retreat massifs on the adjacent shelf
537
A tTOCHTHONOUS PATTERNS OF SEDIMENTATION
325
CAPE COD
SUSPENDED-
SEDIMENT
DISCHARGE
DEPOSITIONAL PROVINCES
SHELF VALLEY COMPLEX AND
SHOREFACE RETREAT BLANKET
CAPE RETREAT MASSIF AND
SHOREFACE RETREAT BLANKET
CAPE RETREAT BLANKET
ESTUARY RETREAT BLANKET
CAROLINA
SALIENT
WAVE CLIMATES
= >40% >5 FT.
— >30% >5 FT.
••• >20% >5 FT
SOUTHERN
ATLANTIC
BIGHT
\
100 KM
FIGURE 14. Coastal sediment discharge (Meade, 1969) and wave climate of the Middle Atlantic Bight
(Do/an et a/., Wl) and resulting depositional provinces. From Sui/t and Sears (W4)-
i Fie. 13). Vet further south, the Georgia Bight ex-
periences a high tide range, a milder wave climate, and
the closely spaced river mouths are estuarine in con-
figuration. Their retreat has generated a blanket of
coalescing shelf valley complexes (Fig. 16).
Initiation of Modern Patterns
TEXTL'RAL AND MORPHOLOGIC PATTERNS ON A STORM-
DOMiNATF.D shllf. On two of the best studied autoch-
thonous shelves, the Middle Atlantic Bight of North
America and the shelf around the British Isles, the
hydraulic climate is sufficiently intense to overprint
older subaerial and nearshore marine patterns of the
surficial sand sheet with a modern textural and morpho-
logic pattern.
In the Middle Atlantic Bight fair-weather flows are
driven by the geostrophic response of the stratified
shelf water column to freshwater runoff and to winds
(MeClenncn. 1973; Bumpus. 1973); see Fig. 17. How-
538
FIGURE 1 5. Cuspate forelanch ami cape shoal-retreat massifs (stippled) of the South Carolina shelf. Sote overprinting by ridge and
swale topograph). Contours in fathoms. From Swift et al. (19~2).
326
FIGURE 16. Morphologic pattern of estuarine shoal retreat blanket, overprinted by ridge
and swale topograph}. South Carolina coast. Highs are stippled. Contours in fathoms. Irom
Swift and Sears (IT 4).
539
AUTOCHTHONOUS PATTERNS OF SEDIMENTATION
327
39°30
- 39-00
38-30'
74-30'
74-00
73-30'
73-00'
72-30'
FIGURE 17. Fair-weather hydraulic regime of the Sew Jersey shelf as indicated by
Savoniits rotor current meters mounted 1.5 to 2.0 m above the seafloor at jour stations on
the New Jersey shelf, for periods of 9 to 11 days in late spring. Progressive vector diagrams
indicate a general southerly water drift, partly correlatable with wind directions (McC/en-
nen, 7973). Loops, spikes, and bulges on progressive vector diagrams are modulation by the
semidiurnal tide. In percent exceedence diagrams, current velocities are compared with
bottom wave surge calculated from wave climate data. All data from McC/ennen (1973).
ever, neither these unidirectional flow components nor
the superimposed wave oscillations (McClennen, 1973)
and tidal oscillations (Redfield, 1956) are strong enough
to result in significant bed load transport over broad
areas. During the winter period of frequent storms, the
water column is not stratified, and air-water coupling
is more efficient (see discussion, pp. 263-264). The
geometry of the Middle Atlantic Bight is especially
conducive to strong flows during this period. When
low-pressure systems pass over the bight, so that the
isobars of atmospheric pressure parallel the isobaths of
the shelf surface, the resulting winds blow southward
down the length of the Middle Atlantic Bio;ht, parallel-
ing the curve of the shoreline, and induce a uniform
setup of the shelf water mass against the coast of 40 to
60 cm. High-velocity "slablike" flows of remarkable
longshore coherence result (Beardslcy and Butman,
1974; Boicourt, personal communication).
The coastal boundary of these storm flows appears
to initiate ridge topography at the foot of the shorefacc
(I)uane et al., 1972); see Chapter 14, Figs. 28 and 31.
However, it is clearly an oversimplification to describe
the ridge topography of the Middle Atlantic Bight as a
purely inherited topographic pattern. The ridges
maintain their characteristic 10 m relief and textural
patterns across the central shelf (Swift et al., 1974);
see Fig. 18. Troughs retain erosional windows in which
Early Holocene lagoonal clays are thinly veneered
with a lag deposit of pebbly sand. Calculations based
on current-meter records suggest that the unidirectional
components of storm flows arc sufficient to mobilize the
sandy bottom (Fig. 19), and to slowly level the ridge
topographs', if the topography were not in fact a
continuing response of the seafloor to the modern
hydraulic climate.
In several areas on the Middle Atlantic shelf, there is
evidence to suggest that ridge topography may be
initiated on the central shelf, if not already present as a
survival from the nearshore environment. Off South
Carolina, shoal retreat massifs are overprinted by a
ridge topograpfiy even though the modern nearshore
zone is not apparently forming ridges (Fig. 15). Else-
where, the ride;e pattern appears to have changed as the
water column deepened and the shoreline receded
during the course of the Holocene transgression. The
estuary mouth shoal that is the landward end of the
Delaware Massif (Fig. 12) has impressed into it a
tide-maintained ridge pattern that trends normal to
the shoreline and parallel to the sides of the estuary
mouth. As the crest of the massif is traced seaward,
the trend of the ridges and troughs superimposed on it
shifts toward a shore-parallel orientation. The bay
540
328 CONTINENTAL SHELF SEDIMENTATION
39°I0'N
39°05'
74°00'
FIGURE 18. Distribution of grain sizes on the central New
Jersey shelf. Medium to fine sands occur on ridge crests. Fine
to very fine sands occur on ridge flanks and in troughs. Locally,
73°45' W
erosional contours in troughs expose a thin lag of coarse, shelly,
pebbly sand over lagoonal clay. From Stubblefield et al.
(in press).
mouth ridges are oriented parallel to the reversing tidal
flows of the bay mouth; the offshore ridges appear to par-
allel instead the geostrophic storm flows of the open shelf.
The Great Egg Massif, associated with the former
course of the Schuylkill River across the shelf, has
been heavily dissected into a transverse ridge pattern.
Seaward of a scarp whose toe lies at 90 m, a second,
small-scale ridge pattern with a somewhat different
trend has been superimposed on the first (Fig. 20).
Stubblefield and Swift (1975) have presented a model
for the evaluation of the compound ridge pattern based
on vibracores. and 3.5 kHz seismic profiles collected
in the area (Fig. 21). Radiocarbon dates indicate
that the large-scale ridges appear to have formed
immediately subsequent to the passage of the shoreline
at approximately 11,000 BP (Fig. 2\A). Internal
stratification indicates that large-scale ridges grow by
the accretion of conformable beds. Wide, large-scale
troughs appear as zones of bare Pleistocene substrate,
where the surficial sand sheet was never formed, or
where its material was swept away to nourish the
growth of adjacent ridges.
With continuing transgression and deepening of the
water column, the ridges appear to have increased
their spacing by means of lateral migration or the
coalescence of adjacent ridges. Internal strata tend to
dip more steeply than present ridge flanks, suggesting
that toward the latter part of their history, ridge growth
was mainly the consequence of lateral rather than
vertical accretion
Small-scale troughs transect large-scale ridges, and
tend to break large-scale ridges up into en echelon seg-
ments (Fig. 20). Where small-scale troughs cross large-
scale troughs they are incised into the flat-lying Early
Holocene and Pleistocene strata that floor the large-
scale troughs. Small-scale troughs are commonly
narrow features that do not penetrate through the
Early Holocene lagoonal clay (Fig. 2\B). Where this
clay is in fact breached, so that the small-scale troughs
penetrate the underlying sand, the troughs are notice-
ably wider, as though they had expanded by under-
cutting of the clay in a fashion analogous to the growth
of a blowout on a grass-covered eolian flat (Fig. 21C).
The ridge topography of the Middle Atlantic Bight
is accompanied by mesoscale bed form patterns, whose
relationship to the ridge pattern is not clearly under-
stood. The most ubiquitous mesoscale bed forms are the
current lineations, which occur as sand ribbons or more
541
AUTOCHTHONOUS PATTERNS OF SEDIMENTATION
329
75°48'
/S'M.V
• / I • • • •
36°32'
SIZE CLASS
COARSE SAND
MEDIUM SAND
CLASS % VOLUME
BOUNDARIES EXCEEDENCE TRANSPORT
jFINE SAND
jVERY FINE
J SAND
(PHI)
0-1
1-2
2-3
3-4
6.0
13.0
15.8
17.0
(m3/m/T)
3.4
8.0
75
5.2
30
u
u 20
S 10
CURRENT METER STATION • STATIONS
J NET TRANSPORT DIRECTION
12 16
DAYS
20
24
28
FIGURE 19- Sediment transport in response to the unidirectional component
of flow during the month oj November 1972, in an inner shelf ridge field, False
Cape, Virginia. Estimates based on Shield's threshold criterion, a drag coeffi-
cient of i X 10~3, and Laursen's (2958) total load equation. Values expressed
as cubic meters of quartz per meter transverse to transport direction for time
elapsed. Solid line is the 10 m isobath.
commonly as linear erosional furrows. They may trend
parallel to the trend of the ridge topography, or may
cut across it, so as to make a larger acute angle with the
shoreline (Fig. 22). Toward the southern end of the
Middle Atlantic Bight, the shelf surface shoals, narrows
and curves to the east. Sand wave fields appear, perhaps
indicative of the acceleration of storm flows in response
to the decreasing cross-sectional area of the shelf water
cokimn.
Ridges molded into the Albermarle shoal retreat
massif bear sand waves on their crests (Fig. 23). Sand
waves locally attain 2 m heights and angle of repose
slopes. Sand wave crestlines are not quite normal to
shore, suggesting that the ridge crests on which they are
found experience a seaward component of flow during
storms. At Diamond Shoals, the southern extremity of
the Middle Atlantic Bight, sand waves up to 7 m high
occur between sand ridges, forming a reticulate pattern
(see Fig. 27, Chapter 16).
Grain-size patterns in the Middle Atlantic Bight
suggest that the storm flows that interact with the ridge
topography and the mesoscale bed forms are capable of
transporting at least the finer grades of sand for appreci-
able distances. The inner shelf sectors before the seaward-
convex coastal compartments of the Middle Atlantic
Bight exhibit a repeating pattern of grain-size distribu-
tion (Fig. 24). The northern half of each of these inner
shelf sectors, where south-trending storm flows must
542
330
CONTINENTAL SHELF SEDIMENTATION
39°
00 NX
38°
45'NN
7^T
74°00 W
FIRST ORDER HIGHS
CRESTLINES. SECOND ORDER HIGHS
FIGURE 20. Great Egg shelf valley and shoal retreat massif. Large-scale
ridges in inset may date from a period when the ancestral Great Egg estuary
was active. Nearshore large-scale ridges were probably formed by shoreface de-
tachment during erosional retreat of the shoreface, after capture of the ancestral
Schuylkill River by the Delaware River, and consequent reduction in discharge
of the Great Egg estuary. See Fig. 6 for relationships of Schuylkill, Delaware,
and Great Egg rivers. From Stubblefield and Swift (in press).
presumably converge with the shoreline, tend to be
floored with primarily medium- and coarse-grained
sands, molded into a well-defined ridge topography.
On the southern halves of the coastal compartment,
where the shoreline tends to curve to the west, storm
flows might be expected to expand and decelerate.
Here the fine sand blanket of the shoreface extends
across the inner shelf floor, as though nourished by
material swept out of the ridge topography to the north.
The schematic flow pattern in the lowest panel of Fig. 24
is not basic on detailed observations. It is intended to
indicate that current flowing generally southwest parallel
to the long dimension of the shelf will tend to converge
with the northeastern portion of the shoreline and
diverge from the southwestern portion of the shoreline.
A somewhat closer relationship appears to exist
between flow geometry and sediment distribution in the
vicinity of the shoal retreat massifs (Fig. 25). The ridge
topography attains its maximum relief where it has
been molded onto the crests of the massifs. The massifs
do not exhibit bilateral symmetry; troughs are deepest
and widest on the northern sides. As a trough axis is
traced across the massif, erosional windows exposing
the basal pebbly sand or the underlying clayey sub-
strate become less frequent. The fine sands of the
trough flanks tend to bridge across the trough floor.
543
AUTOCHTHONOUS PATTERNS OF SEDIMENTATION
331
®
PRIMARY
RIDGE
PRIMARY TROUGH
PRIMARY RIDGE
■HOLOCENE SILTY CLAY
-^-PLEISTOCENE SAND
:':>:o:::o: ::>■ -'^ ' ■ ' T- ■"' ^m* -«- PLEISTOCENE SILTY CLAY
PRIMARY RIDGE
DEVELOPMENT OF RIDGE TOPOGRAPHY
FIGURE 21. Ridge evolution on the central New Jersey shelf. (A) Ridge nuclei are formed
during the process of ridge detachment and shoreface retreat, or by other means in the nearshore
zone. Sand continues to be swept out of troughs onto ridges as water column deepens during the
course of the Holocene transgression. (B) Seafloor scour during storms locally penetrates the Early
Holocene lagoonal clay carpet, and a secondary trough forms, initially by downcutting. (C)
Downcutting in the secondary trough decreases and lateral erosion increases as second silly clay
layer is exposed. Secondary trough widens by undercutting of upper clay in "blowout" fashion.
Sand from similar excavations upcurrent forms secondary ridges. From Stubblefield and Swift
(in press).
The trough axis tends to climb toward a low sill on the
southern side of the massif; beyond this the seafloor
drops off rapidly to the adjacent shelf valley. The valley
floor commonly consists of fine to very fine featureless
sand. The topography and grain-size pattern suggest
that south-trending flows converge with the rising sea-
floor and accelerate up the northern flanks of the massifs.
Fine sand swept out of the troughs is deposited in the
zone of flow expansion and deceleration over the shelf
valley south of the massif.
TEXTURAL AND MORPHOLOGIC PATTERNS ON A TIDE-
DOMINATED shelf. The tide-swept shelf around the
British Isles (Stride, 1963) provides an interesting con-
trast with the storm-induced sedimentation of the
Middle Atlantic Bight. Surges are at least as frequent
here as in the Middle Atlantic Bight (Steers, 1971).
However, much more work is done on the seafloor
by the semidiurnal tidal currents associated with the
amphidromic edge waves that sweep the margins of mar-
ginal shelf seas of western Europe (see Chapter 5, Fig. 4
and discussion, p. 60). The rotary tidal currents
associated with these tidal waves are in fact analogous
in some respects to the inertial wind-driven currents
generated by storms. Midtide surface velocities in
excess of 50 cm/sec (1 knot) are sustained over vast
areas, and locally exceed 200 cm/sec. Ebb-flood dis-
charge differentials result in currents residual to the
tidal cycle, whose velocities may be as great as a tenth
of the midtide value.
As a consequence of the higher rate of expenditure of
energy on the seafloor, morphologic and textural
patterns inherited from the retreating nearshore zone
have been largely erased. Erosional shoreface retreat
has resulted in a surficial sediment sheet that is com-
parable in many respects to that of the Middle Atlantic
Bight (see Belderson and Stride, 1966). However, the
poorly resolved sand transport patterns of the Middle
Atlantic Bight are replaced by well-defined transport
paths, with sand streams that diverge from beneath
544
332 CONTINENTAL SHELF SEDIMENTATION
39°05'N
73°58W
73°57'
73°56'
TREND OF BOTTOM UNEATION
TRACKLINE
FIGURE 22. Current I in eat ion patterns on the
central New Jersey shelf. Bars indicating lineations
are over 10 times as long as features that they rep-
resent. They locally represent sets of lineations. High
areas are stippled. Contours in fathoms. From
McKinney et al. (2974)-
tide-induced "bed load partings" and flow down the
gradient of maximum tidal current velocities until
either the shelf edge or a zone of "bed load conver-
gence" and sand accumulation is reached (Stride, 1963;
Kenyon and Stride, 1970; Belderson et al., 1970);
see Fig. 26.
Each stream tends to consist of a sequence of more or
less well-defined zones of characteristic bottom mor-
phology and sediment texture (Fig. 27). Streams may
begin in high-velocity zones [midtide surface velocities
in excess of 3 knots (150 cm/sec)]. Here rocky floors are
locally veneered with thin (centimeters thick) lag
deposits of gravel and shell. Where slightly thicker, the
gravel may display "longitudinal furrows" parallel to
the tidal current (Stride et al., 1972), a bed form
related to sand ribbons (see Chapter 10, p. 170).
Between approximately 2.5 and 3.0 knots (125-150
cm/sec) sand ribbons are the dominant bed form
(Kenyon, 1970). These features are up to 15 km long
and 200 m wide, and usually less than a meter deep.
Their materials are in transit over a lag deposit of shell
and gravel. Kenyon has distinguished four basic pat-
terns that seem to correlate with maximum tidal current
velocity and with the availability of sand (Chapter 10,
Fig. 15).
Further down the velocity gradient, where midtide
surface velocities range from 1 to 2 knots (50-100
cm/sec), sand waves are the dominant bed form. Where
the gradient of decreasing tidal velocity is steep or
transport convergence occurs, this may be the sector
of maximum deposition on the transport path. Over
20 m of sediment has accumulated at the shelf-edge
convergence of the Celtic Sea, although it is not certain
that this sediment pile is entirely a response to modern
conditions.
The Hook of Holland sand wave field off the Dutch
coast is one of the largest (15,000 km2) and the best
known (McCave, 1971). The sand body is anomalous
in that it sits astride a bed load parting; the sand patch
as a whole may be a Pleistocene delta or other relict
feature. Sand waves with megaripples on their backs
grow to equilibrium heights of 7 m with wavelengths
of 200 to 500 m in water deeper than 18 m; in shoaler
water, wave surge inhibits or suppresses them. Elongate
tidal ellipses favor transverse sand wave formation, and
the sand waves tend to be destroyed by midtide cross
545
AUTOCHTHONOUS PATTERNS OF SEDIMENTATION
333
FIGURE 2 3. Sand ridges with superimposed sand waves on the northern North Carolina inner shelf. Topographic highs are stippled.
Contours in feet. From Swift et ah (1973).
flow when the ellipse is less symmetrical. Under the
latter condition, linear sand ridges may be the pre-
ferred bed form, as midtide cross flow would tend to
nourish rather than degrade them (Smith, 1969). The
triangular sand wave field is limited by a lack of sand
on the northwest, by shoaling of the bottom and in-
creasing wave surge on the coast to the south, and by
fining of sand to the point that suspensive transport is
dominant to the north (McCave, 1971).
Further down the velocity gradient, beyond the zones
of obvious sand transport, there are sheets of fine sand
and muddy fine sand and in local basins, mud. They
lack bed forms other than ripples, and appear to be the
product of primarily suspensive transport (McCave,
1971) of material that has outrun the bed load stream
(see discussion, Chapter 10, p. 160). These deposits may
be as thick as 10 m (Belderson and Stride, 1966), but
where they do not continue into mud, they break up into
irregular, current-parallel or current-transverse patches
of fine sand less than 2 m thick, resting on the gravelly
substrate.
The complex pattern and mobile character of the
shelf floor around the British Isles have led British
workers to reject the relict model for the shelf sediments
(Belderson et al., 1970). They note that it correctly
draws attention to the autochthonous origin of the
sediment, but that it fails to allow for its subsequent
dynamic evolution. They propose instead a dynamic
classification:
1 . Lower sea-level and transgressive deposits, patchy
in exposure, but probably more or less continuous
beneath later material; largely the equivalent of a
blanket (basal) conglomerate.
546
76°
37'
76°
VERY COARSE TO
MEDIUM SAND
FINE SAND
.VERY FINE SAND,
I SILTY CLAY
WOODS HOLE DATA
VA. INST. MAR. SCI. DATA
LITTORAL DRIFT
STORM DRIVEN CURRENTS
♦ ■" TIDAL CURRENTS
FIGURE 24. Above: Bathymetry of the Delmarva inner direction of currents responsible for bed load sediment
shelf. From L'chupi (1970). Center: Distribution of sediment. transport. Reproduced from Suift (1975).
From Hathaway (1971) andSichols (1972). Belou: Inferred
334
547
ALLOCHTHONOUS PATTERNS O F SE DI M E N T A TI O N
335
76°
OO'W
% ; GRAVEL
:■:■:■:■ coarse sand
: : (INSET: 1.0-1.5 + )
□ MEDIUM SAND
(INSET: 2.0-2.5*1
FINE SAND
(INSET: 2.0-2.5*)
75°
29 W
xxvxxv VERY FINE
tt SAND
(INSET: 2.5-3.0*)
FIGURE 2 5. Grain-size distribution on a portion of the Virginia Beach massif, and
adjacent shelf valley.
2. Material moving as bed load (over the coarser
basal deposits) mainly well sorted sand and in places
first-cycle calcareous sand.
3. Present sea-level deposits (category 2 sediment
having come to permanent rest) consisting of large
sheets to small patches, which range from gravel and
shell gravel to sand and calcareous sands, muddy
sands, and mud.
The implication is that of a shelf surface moving
toward a state of equilibrium with its tidal regime.
The degree of adjustment appears to be greater than
in the case of the North American Atlantic Shelf, in that
there is less preservation of nearshore depositional
patterns. As a consequence of the intensity of the
hydraulic climate, there is less on shelf storage (category
3) and more material in transit.
Locally, sand ridges similar to those of the Middle
Atlantic Bight do occur. Like those of the Middle
Atlantic Bight, they tend to be grouped in discrete
fields. In some cases, it is possible to infer that these
ridge fields are in fact shoal retreat massifs, generated
by the retreat of a near shore depositional center during
the course of the Holocene transgression (Swift, 1975).
The clearest case may be made for the Norfolk Banks
(Houbolt, 1968; Caston and Stride, 1970; Caston,
1972); see Fig. 28. Here a series of offshore sand ridges
may be traced into a modern nearshore generating zone
(Robinson, 1966; see Chapter 14, Fig. 39) where sand
is packaged by the specialized tidal regime of the
shoreface into shapes hydrodynamically suited for
survival on the open shelf (see discussion, p. 180).
The Nantucket Shoals sector of the North American
Atlantic shelf appears to constitute a similar evolu-
tionary sequence of ridges (see Chapter 10, Fig. 30).
The Norfolk Banks are analogous to the cape shoal
retreat massifs of the Carolina coast of North America,
in that the generating zone is a coastal salient that
serves as a sink for the nearshore sand flux. Other, more
poorly defined ridge fields in the southern bight of the
North Sea (Fig. 28) may be analogous to the estuarine
shoal retreat massifs of the Middle Atlantic Bight in
that they may have been generated by the retreat of the
ancestral Rhine and Thames estuaries.
ALLOCHTHONOUS PATTERNS OF SEDIMENTATION
Shelves undergoing allochthonous sedimentation differ
from autochthonous shelves in a variety of character-
548
336
CONTINENTAL SHELF SEDIMENTATION
1
FIGURE 26. Generalized sand transport paths around the
British Isles and France, based on the velocity asymmetry of the
tidal ellipse and the orientation and asymmetry of bed forms. From
Kenyon and Stride (1970).
istics. The most obvious is that allochthonous shelves
tend to be floored by fine sands, fine muddy sands, or
muds that have escaped from adjacent river mouths:
autochthonous shelves in contrast are generally covered
by coarser grained sand of local origin. Although sur-
faces of allochthonous shelves are constructional in
nature, they tend to be smooth and featureless; their
fine materials have traveled primarily in suspension, and
the effective underwater angles of repose of the sediment
may be too low to result in such large-scale bed forms
as sand waves or sand ridges. However, such features
are not totally unknown. Allersma (1972) has reported
"mud waves" from the Venezuelan shelf that appear
to be very similar to the shoreface-connected ridges of
the Middle Atlantic Bight.
Transport on Allochthonous Shelves
Mechanisms of sediment transport on allochthonous
shelves have been generally described by Drake in
Chapter 9. Since this chapter stresses regional transport
patterns, it seems worthwhile to summarize Drake
et al.'s (1972) study of river-dominated sedimentation
on the southern California shelf. This carefully docu-
mented, real time study of the dispersal of flood sedi-
ment is probably the most detailed report on the nature
of allochthonous sediment dispersal available at the
time of writing.
In January and February of 1969, southern California
experienced two intense rainstorms which resulted in a
record flood discharge. The freshly eroded sediment was
a distinctive red-brown in contrast to the drab hue of
ZONE I
BEDROCK
&
GRAVEL
ZONE II
SAND
RIBBONS
ZONE III
SAND
WAVES
ZONE IV
SMOOTH
SAND
ZONE V
SAND
PATCHES
T^P^
DECREASING MID-TIDE SURFACE VELOCITY; BOTTOM GRAIN SIZE
FIGURE 27. Succession of morphologic provinces along a tidal transport path. Based on Belderson
et al. (1970).
549
AUTOCHTHONOUS PATTERNS OF SEDIMENTATION 337
FIGURE 2 8. Tidal ridge fields of the southern bight of the North Sea.
Northernmost ridge field appears to constitute a shoal retreat massif,
marking the retreat path of the nearshore tidal regime of the Norfolk
coast. Ridge fields in the approach to the English Channel may have been
initiated in an earlier, more nearly estuarine environment, from Houbo/t
(1968).
the reduced shelf sediments. The flood deposit on the
shelf could therefore be repeatedly cored and isopached,
and its shifting center of mass traced seaward through
time.
USGS stream records show that 33 to 45 X 106 metric
tons of suspended silt and clay and 12 to 20 X 106
metric tons of suspended sand were introduced by the
Santa Clara and Ventura rivers. By the end of April
1969, more than 70% of this material was still on the
shelf in the form of a submarine sand shoal extending
7 km seaward, and a westward-thinning and -fining
blanket of fine sand, silt, and clay existed seaward of
that (Fig. 29.4).
By the end of the summer of 1969, the layer extended
further seaward, had thinned by 20%, and had de-
veloped a secondary lobe beneath the Anacapa current
to the south (Fig. 29B). Eighteen months after the
floods, the surface layer was still readily detectable.
Considerable bioturbation, scour, and redistribution
had occurred south of Ventura, but the deposit was
more stable to the north (Fig. 29C).
A concurrent study of suspended sediment distribution
in the water column revealed the pattern of sediment
transport (Fig. 29D). Vertical transparency profiles,
after four days of flooding, showed that most of the
suspended matter was contained in the brackish surface
layer, 10 to 20 m thick. Profiles in April and May
revealed a layer 15 m thick, with concentrations in
excess of 2 mg 1, and a total load of 10 to 20 X 104
metric tons. Since this load was equal to river discharge
for the entire month of April, it must have represented
lateral transport of sediment resuspended in the near-
shore zone. Vertical profiles over the middle and outer
shelf for the rest of the year were characterized by
sharply bounded turbidity maxima, each marking a
thermal discontinuity. These also were nourished by
lateral transport from the nearshore sector where the
discontinuities impinged on the sloping bottom. The
near-bottom nepheloid layer was the most turbid zone
in the inner shelf. This nepheloid layer was invariably
the coolest, and was invariably isothermal, indicating
that its turbidity was the result of turbulence generated
550
34°20' -
34°10' -
34°20'-
34°10
34°20'
34°10'
_L_2 S ■ • '
119°50'
119°30'
119°10'
(d)
T3S<1 T3343 T3SS0
13588 13586 13372
sot- -"70
-1 - - .
70 - -.,
e° e-i
FIGURE 29. Thickness of flood sediment (centimeters) on the Santa Barbara-Oxnard shelf
in (a) March-April 1969; (b) ,M«>-v4»£».r/ i969/ and (c) February-June 1970, based on
cores, (d) East-west cross section showing vertical distribution of light-attenuating substances
over Santa Barbara-Oxnard shelf. For clarity, the bottom 20 m of the water column is not
contoured, but the percent transmission value at the bottom is noted. From Drake et al. (1972).
338
551
J^^y
6'
■ LLOCHTHONOUS PATTERNS OF SEDIMENTATION 339
YOUNGER SUITE
COARSEST GRADE
FINE SAND AND
COARSER
VERY FINE SAND
CLEAN VC. SILT
CLAYEY SILT
SILTY CLAY
5°
OLDER SANDS
V. COARSE SAND
COARSE SAND
MEDIUM SAND
FINE SAND
VERY FINE SAND
FIGURE 30. Distribution of sediments on the Niger shelf. Young suite is of allochthonous origin; older suite of autochthonous
sand is exposed in nondepositional windows. From Allen (1964).
by bottom wave surge. Bottom turbidities ranged from
50 mg 1 during the flood to 4 to 6 mg 1 during the next
winter, but were at no time dense enough to drive
density currents.
Drake et al.'s study suggests that the transport of sus-
pended sediment across shelves undergoing allochthon-
ous sediment action starts with introduction by a river
jet, and continues with deposition, resuspension, and
intervals of diffusion and advection by coastal currents
in a near-bottom nepheloid layer.
Depositional Patterns on Allochthonous Shelves
Fine sediments deposited on allochthonous shelves may
occur as a seaward-thinning sheet (Fig. 30), or as a
series of strips of fine sand or mud oriented generally
parallel to the shoreline, see Figs. 31 and 32 (see also
Venkatarathnam, 1968; McMaster and Lachance, 1969;
and Niino and Emery, 1966). On shelves of equant or
irregular dimensions, shelf sectors surfaced by far-
traveled, fine-grained sediment may be more irregular
in shape (Niino and Emery, 1966; McManus et al.,
1969; Knebel and Creager, 1973). Such allochthonous
deposits tend to be separated by, or to enclose,
nondepositional "windows" in which relatively coarse
autochthonous sands are exposed. The disposition of
these strips and sheets of allochthonous sediment is
generally meaningful in terms of what is known of
regional circulation patterns. Locally, the strips may
underlie turbid, brackish water plumes that extend
from river mouths under the impetus of buoyant ex-
pansion and inertial flow (Chapter 14, Fig. 41). Where
such flows of high-turbidity water extend for long dis-
tances parallel to the coast or seaward across the shelf at
promontories, they have been described by McCave
(1972) as "advective mud streams" (Fig. 33). He cites
Jerlov (1958) as describing such a mud stream running
south from the Po Delta, over the mud bed shown in
Fig. 34.
However, the presence of windows of older sand does
not necessarily mean that the sediment pattern is a
transient one, which must be eventually followed by a
total masking of the old surface of transgression by fine
sediment. Instead, the pattern may be a steady state one,
552
29*N .
*> neorshore drift
river outflow »^- residual
currents
FIGURE 31- Distribution of mud and generalized transport pattern on the western
Gulf shelf oj North America. I'rom McCave (l()72), after Van Andel and Curray (/960).
- zyoo'
- 22°30
22°00
21°30'
- 2I°00
05°00'
FIGURE 3 2. Distribution of sediment and generalized transport pattern off
the Nayarit coast. Pacific side of Mexico. Autochthonous sands occur in non-
depositional windows. Vrom Curray (1969).
340
553
AUTOCHTHONOUS PATTERNS OF SEDIMENTATION 341
Extinction (against distilled water)
I I (Ml
I I 0.4-0.6
0.2-0.4
0.6-0.8
0.8-1.5 (> 0.8 along Dutch coast) ^>1.5
^> Residual currents
FIGURE 33. Turbidity (light extinction) given by Joseph (1955) with the residual cur-
rent pattern in the southern North Sea. Two advective mud streams are illustrated, one
crossing from the English to the Dutch side of the area, the other running up the Dutch
coast from the mouths of the Plime River. Actual sediment ccncentrations are higher in the
latter. From McCave (1972).
determined by the local relationship between the hy-
draulic activity (primarily wave surge) on the bottom
and the near-bottom concentration of suspended sedi-
ment (Fig. 35), as well as by the regional transport
pattern.
On autochthonous shelves, sand transport is primarily
advective in nature, occurring during short, intense
episodes of wind-driven or tidal flow, and textural
gradients tend to reflect the direction of sand transport,
with transport becoming finer down the transport path
(see Figs. 25 and 27). The transport of fine sand and
mud on shelves undergoing allochthonous sedimentation
is also primarily advective in nature, in that the turbid
water tends to flow as a mass in response to the regional
circulation pattern. However, because of the greater
role of reversing tidal currents and wave surge in dis-
tributing fine sediment it is convenient to think of fine
sediment transport on allochthonous shelves as consisting
of a dominant advective component, driven by the
regional circulation pattern, and an important but
subordinate diffusive component, driven by reversing
tidal flows and wave surge.
The diffusive component of transport not only in-
fluences the regional pattern of fine sediment deposition
as noted in Fig. 35 and Chapter 9, Fig. 15, but may also
result in textural gradients within allochthonous sediment
sheets that trend at an angle across the advective trans-
port direction. On the Niger shelf, for instance, the domi-
nant, advective transport direction is from east to west,
under the impetus of the Guinea current (Allen, 1964).
However, bottom sediments tend to become finer in a sea-
ward (north to south) direction (Fig. 36). Sharma et al.
554
FIGURE 34. Distribution of sediment in the northern Adriatic Sea. From Van Straaten (7965).
342
muddy coast
nearshore
mud belt
■/////A
mid-shelf
mud-belt
\y/////////j%,
mz OAST
SHELF
EDGE
outer-shelf
mud-belt
* or under advective mud stream
FIGURE 3 5. Schematic representation of five cases of sites of shelf mud accumulation.
Compare with Fig. 7 5 in Chapter <J. From McCave (1972).
555
A LLOCHTHON OLS PATTERNS OF SEDIMENTATION
343
KILOMETRES
0 100 100
FROM AXIS OF SYMMETRY
-ACROSS DELTA
FIGURE 36. Grain size in relation to sedimentary environments in Siger Delta area. In
subaerial delta, all grades present are shown. In offshore part of delta, coarsest grade in near-
surface layers is projected onto vertical plane perpendicular to axis of delta symmetry. From
Allen (1964).
( 1972) have described grain-size gradients in Briston Bay
of the Bering Sea that are more nearly related to the iso-
baths than to the prevailing currents. They consider the
textural gradients to be the consequence of wave surge
diffusion (Fig. 37). The operative mechanism would be
progressive sorting during the seaward diffusion of sedi-
ment. In this process, sediment that drifts seaward into
deeper water during a transport event is likely to leave
its coarsest fraction behind when reentrained. because
of the weaker nature of deep-water wave surge (see dis-
cussion of progressive sorting. Chapter 10. p. 162.
Stratigraphy of Allochthonous Shelves
The tenfold reduction in the rate of eustatic sea-level
rise experienced between 4000 and 7000 years ago
(Milliman and Emery. 1968) has resulted in a shift from
autochthonous to allochthonous regimes in a number of
shelf sectors (Curray. 1964). River mouths servicing
such shelves have equilibrated with their tidal prisms.
and have begun to bypass fine sediment in quantities
sufficient to result in deposition on the shelf surface.
Two characteristic stratigraphies have resulted, which
may be correlated with the transport schemes illustrated
in Figs. 3a and 3c. Where the shift in the balance between
the rate of sedimentation and the rate of sea-level rise
has not been adequate to cause coastal progradation,
the coast has continued to undergo erosional shoreface
retreat, or has approached stillstand conditions (see
discussion of equation, p. 312). Patches and sheets of
fine-grained sediments have accumulated more or less
simultaneously over the sand sheet produced during
the earlier period of erosional shoreface retreat (Fig. 38).
Elsewhere, where the Late Holocene balance between
sedimentation and sea-level rise has resulted in coastal
progradation, the transgressive sand sheet passes land-
ward beneath a veneer of mud some few meters thick
into a thick littoral sand body deposited during stillstand,
and a second, subaerial, sand sheet extends seaward
over the inner portion of the mud veneer (see Chapter 14,
Fig. 10B).
556
FIGURE 37. The distribution of grain sizes in Bristol Bay. See text Jor analysis.
From Sharma et al. (7972).
RELATIVE SEA LEVEL
FALLING SEA LEVEL
OR EMERGENCE
RISING SEA LEVEL
OR SUBSIDENCE
RAPIO
SLOW
STABLE
SLOW
RAPID
POSITION
EROSION
EROSION
\
i 1
\ 1
\ 1
\ 1
K i
UJ ,
z § S
^
S
SI
H"
RATE OF
DEPOSIT
HIGH RATE
JAM
FIGURE 3K. Above: Generalized cross section oj Late Quater- Texas. Below: Schematic representation of shoreline migration,
nary sediments in a line perpendicular to the coast near Rockport, From Curray (Z964).
344
557
SUMMARY
345
The abrupt nature of the transition from Early
Holocene autochthonous regimes and the recent nature
of this transition have prevented us from observing on
modern continental margins a third characteristic
stratigraphy, which is widespread in the rock record,
namely "marine onlap" (Grabau, 1913, Fig. 144). In
this model, a prolonged period of relatively slow sea-level
rise is accomplished by the transport scheme shown in
Fig. 3b, where both shoreface and river mouth by-
passing occur in a regime of transgressive allochthonous
sedimentation. Subaerial and submarine depositional
environments are linked by a unified pattern of sedi-
ment transport, and their landward displacement results
in a threefold sequence of fluvial, marine marginal, and
open shelf lithosomes beneath the shelf surface.
During the slow eustatic transgression of the Creta-
ceous, such a sequence was deposited on the North
American shelf off North and South Carolina (Fig. 39).
The present erosional surface approximates a time plane,
and seaward decrease in grain size across this environ-
ment suggests progressive sorting through fluvial,
estuarine, and marine environments (Swift and Heron,
1969). Shoreface erosion was an important source of
sediment, as indicated by the internal unconformity
that largely replaces the littoral sand facies. River
mouth injection may also have been an important
mechanism of coastal bypassing, since the fine-grained,
open marine facies thickens seaward.
The Amazon shelf off Brazil, South America may
represent a modern analog of such a transgressive
autochthonous sequence (Milliman et al., in press).
Milliman et al. describe the mud deposits of the Amazon
shelf as a landward-thickening wedge, whose offshore
portions were deposited during lower stands of sea level,
by the predecessor of the coastal mud stream which
presently trends northwest, from the mouth of the
Amazon toward the Guyana coast (Fig. 40). Milliman
and his associates suggest that the offshore surface of
this mud deposit is at present experiencing an autoch-
thonous sedimentary regime. They suggest that the
net fine sediment budget of the offshore shelf is negative,
with more fine material being lost to erosion than is
replaced by diffusion from the coastal source, so that a
silty lag is accumulating over its surface. More detailed
investigations may indicate that this type of trans-
gressive allochthonous regime is more common than
now supposed.
SUMMARY
The rate and sense of shoreline movement have an
important modulating effect on the shelf sedimentary
regime. It is helpful to think of the coastline as a "littoral
energy fence" in which the landward-oriented net
surge of shoaling waves tends to push sediment back
toward the beach. There are two basic categories of
"valves" that serve to regulate the passage of sediment
through this barrier into the shelf dispersal system:
river mouths and the intervening expanses of shoreface.
During rapid transgressions, river mouths generally
cannot adjust to their combined river and tidal dis-
charges as fast as required by the rise of sea level, and
they become sediment sinks. Sediment is bypassed
through the coastal zone by the basically passive process
of erosional shoreface retreat, which leaves the shelf
surface veneered with a sandy residue.
During slow transgressions, estuarine channels are
more likely to equilibrate to their discharge. Such
channels are capable of bypassing sand as well as
finer sediment, and sediment is supplied by both
shoreface and river mouth bypassing.
During regressions, river mouth bypassing is domi-
nant. Shorefaces become sand sinks, which advance
seaward by means of the successive growth of beach
ridges.
Rapid transgressions result in autochthonous shelf
regimes, in which the surficial sediments are of in situ
origin. Slow transgressions result in allochthonous shelf
regimes. The sediment load is filtered during passage
through a broad intracoastal zone of estuaries and
lagoons, so that the fraction reaching the shelf is fine-
grained and mobile, and may be dispersed for long
distances across the shelf surface.
On autochthonous shelves, only such large-scale sub-
aerial features as cuestas and river valleys seem able to
survive transgression, and even these are strongly
modified by passage of the shoreline. On shelves of low
relief, most morphologic elements have formed at the
foot of the retreating shoreface. Shelf valley complexes
consist of shelf valleys, shoal retreat massifs, and deltas.
In many cases shelf valleys are the retreat paths of
estuary mouth scour channels, and do not always
overlie the buried subaerial river channels. They tend
to be paired with estuarine shoal-retreat massifs, the
retreat paths of estuary mouth shoals. Littoral drift
convergences at capes and headlands may also result in
shoal retreat massifs. Scarps on autochthonous shelves
do not seem to be drowned shorelines in the strict sense,
but instead are truncated lower shorefaces formed during
postglacial stillstands.
In the Middle Atlantic Bight of North America, both
morphology and grain-size distribution patterns can be
shown to be in part of post-transgressional origin, form-
ing in response to storm flows. The shelf surface is char-
acterized by a pervasive ridge and swale topograp'.y. It
is locally forming at the foot of the retreating shoreface,
558
BLACK CREEK
LITTORAL FACIES
MIDDENDORF FM
(FLUVIAL
' 100
METERS
0 30
KILOMETERS
(a)
SOURCE
(-1(J) THRESHOLD)
MEANDER
PLUG ClAyS
(00 THRESHOLD)
MARSH,
TlDEFLAT
CLAYS
(10 THRESHOLD)
(b)
FIGURE 39. (a) Schematic section through the Cre- lagoonal, and shelj environments, respectively, (b) Pat-
taceous Lumbee Group of North and South Carolina. tern of sediment transport as reconstructed from grain-
The Middendorf, Black Creek, and Peedee Formations size gradients and primary structures in outcrops. From
are deposits of landward-displacing fluvial, estuarine- Swift and Heron (1969).
346
559
SUMMARY
347
6CN
52° W
FIGURE 40. Sediment distribution of the Amazon shelf. Modern terrigenous muds
are deposited beneath a north-trending coastal mud stream. "Relict" muds are
believed to have been deposited by the same mud stream during lower stands of sea
level. From Milliman et al. (in press).
in response to coastal boundary flow, but elsewhere
appears to have developed more or less spontaneously
further out on the shelf surface. Current lineations (sand
ribbons and erosional furrows) are abundant. Coarse
sand lags occur on highs, finer sands occur on their
downcurrent slopes and in adjacent lows.
The shelf around the British Isles is an example of an
autochthonous shelf that has reacted in a more vigorous
fashion, in response to a high-intensity tidal regime.
A well-organized pattern of sand dispersal consists of
sand streams that extend from bed load partings to
bed load convergences, or to the shelf edge. Nearshore
morphologic elements have largely been obliterated.
However, well-defined fields of tide-maintained sand
ridges are probably analogous to the shoal retreat
massifs of the Middle Atlantic Bight.
Allochthonous shelves occur adjacent to large rivers
with sediment loads sufficiently large to locally slow or
reverse the sense of the postglacial transgression. Trans-
port is dominantly by water column advection (mud
streams) but diffusion in response to bottom wave surge
is important. Allochthonous deposits of fine sand and
mud are commonly not continuous, but tend to leave
"windows" in which autochthonous sands are exposed,
Such windows are not necessarily transient phenomena,
but may reflect areas in which the concentration of
suspended sediment in the bottom nepheloid layer is
counteracted by a relatively high level of "hydraulic
activity." Textural gradients in autochthonous deposits
may more nearly reflect the seaward, diffusive com-
ponent of transport, rather than the coast-parallel
advective component.
560
348
CONTINENTAL SNKI I SEDIMENTATION
ACKNOWLEDGMENTS
I ihank Paul E. Potloi and Orrin H Pilkcy
< riiioisms of this chapter.
their helpful
REFERENCES
Allen, J. R. L. (l'*t)4). Late Quaternary Niger Delta, and adjacent
air, is: Sedimentary environments and lithofacies. Am. Assoc,
Pet. Geol. Hull., 19: 547 600.
Allen, J. K. 1, (1970a). Sediments of the modern Niger delta:
A summary and review. In J. P, Morgan, ed., Deltaic Sedi-
mentation, Minimi and Ancient. Soc, F,con. Paleontologists and
Mineralogists Spec. Publ. 15, Tulsa. Okla., pp. 1:18-151.
Allen, J. R. L. (197l>b). Physical Processes of Sedimentation. New
York: American Elsevier, 248 pp.
Allersma. E. (1972). Mud on the oceanic shelf off Guiana. In
Symposium on Investigation', and Resources of the Caribbean Sea and
Adjacent Regions. UNESCO, Pans (Unipub, N.Y.). pp. 193-
203.
Andrews, E., D. G. Stephens, and D. J. Colquhoun (1973).
Scouring of buried Pleistocene barrier complexes as a source
of channel sand in tidal creeks. North Island Quadr, ingle.
South Carolina. Geol. Soc. Am. Hull., 81: 3659-3(5(52.
Bagnold, R. A. (1966), An approach to the sediment transport
problem from general physics, I'.S. Geol. Sun: Prof. Pap.
■122 I, 37 pp.
Barby, D. G. and J. II. Hoyt (19<)4). An upper Miocene dredged
fauna from tidal channels of coastal Georgia. ./. Paleontol.,
38: (57 73.
Beardsley, R. C. and B. Butman (1974). Circulation on the New-
England conlinent.il shelf: Response to strong winter storms.
Geol. Re. Lett., Is 181-184.
Belderson, R. 11. and A. II. Stride (1966). Tidal current fashion-
ing of a basal bed. Mar. Geol., I: 237-257.
Belderson, R. II., N. 11. Kenyon, and A. II. Stride (1970). Holo-
cene Sediments on the Continental Shelf II est of the Bulnh Isles.
Inst. Geol. Sci. Rep. No. 70, pp. 157 170.
Bernard, II. A. and R.J. Le Blanc (19(>5). Resume of the Quater-
nary geology of the northwest Gulf of Mexico province. In
II. E. Wright, Jr., and D.J. Irey, eels., The Quaternary of the
I'nited Slates. Princeton, N.J.: Princeton University Press, pp.
137 185.
Brouwer, J. E. (I9ti4). The North Sea. In L. D. Stamp, cd..
The Xorlh Sea and the Law of the Continental Shelf. Bude, Cornwall,
England, Geographical Publications Ltd., World Land
Use Survey Occ. P.ipers 5, pp. 3 14.
Rumpus, D. F. (1973). A description of circulation on the conti-
nental shelf of the east coast of the United States. Prog.
Oceanogr., 6:117 1 57.
Caralp, M., A. Klingebiel, C. Latouche, J. Moves, and R. Prud-
honinie (1972). liilan Cartografihique un le Plateau Continental
dWquitame. Rull. Inst Geol. Bassin d'Aquitaine, Numero
Special. 351 Cours cle la Liberation, \.'> Talence, France, 25 pp.
Carter, L. (1973). Surficial sediments of Rarklcy Sound and the
adjacent continental shelf, West Vancouver Island. Can. ,/.
Earth Sci., 10: 441 459.
Caston, V. N. 1). (1972). Linear sand banks in the southern
North Sea. Sedimentology, lit: l>3 78.
(aston, V. N. I), and A. II. Stride (1970). Tidal sand movement
between some linear sand banks in the North Sea off north-
( astern Norfolk. Mar. Geol., 9: M38- M42.
Colquhoun, I). J. (19(>9). Gecnwrphology of the Lower Coastal Plain
of South Carolina. Geological Notes, Division of Geologv, Co-
lumbia, >outh Carolina State Development Board, 3(5 pp.
(array, J. R. (1904). Transgressions and regressions. In R. L.
Mills, ed.. Papers in Marine Geology: Shepard Commemorative
I'olume. New York: MaeMillan, pp. 175-203.
(urray, J. R. (19(59). History of continental shelves. In The Xew
Concepts of Ci t.linental Margin Sedimentation, Washington, D.C.:
Am. Geol. Inst., pp. JC-6-1 toJC-6-18.
Dolan, R., B. Hayden, G. I loenenberg, J. Zeeman, and M.
Vincent (1973). Classification of the coastal environments of
the world. Part 1. 'The Americas. Tech. Rep. 1, Office of Naval
Research, Washington, D.C., 1(53 pp.
Drake, D. E., R. L. Kolpack, and P. J. Fischer (1972). Sediment
transport on the Santa Barbara Oxnard shelf, Santa Barbara
channel, California. In D.J. P. Swift, D. B. Duane, and O. H.
Pilkcy, eds., Shelf Sediment Transport: Process and Pattern.
Stroudsburg, Pa.: Dowden, Hutchinson & Ross, pp. 301-332.
Duane, D. B., M. E. Field, E. P. Meisburger, D. J. P. Swilt. and
S. J. Williams (1972). Linear shoals on the Atlantic inner
continental shelf, Florida to Long Island. In D. J. P. Swift,
D. B. Duane, and O. H. Pilkey, cds.. Shelf Sediment 'Transport:
Process and Pattern. Stroudsburg, Pa.: Dowden, Hutchinson &
Ross, pp. 447-499.
Emery, K. O. (1952). Continental shelf sediments of southern
California. Geol. Soc. Am. Bull., 6.3: 1105-1108.
Emery, K. O. (1968). Relict sediments on continental shelves of
the world. Am. Assoc. Pet. Geol. Bull., 52: 445-4t.4.
Field, M. E. (1974). Buried strandline deposits on the central
Florida inner shelf. Geol. Soc. Am. Bull., 85: 57-t>0.
Fischer, A. G. (19(>1). Stratigraphic record of transgressing seas
in the light of sedimentation Atlantic coast of New Jersey.
Am. Assoc. Pel. Geol. Bull.. 15: 1(55(5-1(5(5(5.
Ginsburg, R. N. and N. P. James (1974). Holocene carbonate
sediments of continental shelves. In C. A. Burk and C. C.
Drake, eds.. The Geology oj Continental Margins. Berlin and
New York: Springer-Verlag, pp. 137-15(5.
Grabau, A. \V. (1913). Principles of Stratigraphy, 19u0 facsimile
edition of 1924 revision. New York: Dover, 1185 pp.
Hathaway, J. C (1971). Data File, Continental Margin Program,
Atlantic Coast of the United States, Vol. 2, Samples Collection and
Analytical Data. Woods I lole Oceanographic Institution Ref.
71-15, U.S. Geol. Surv., Woods Hole, Mass., 44b pp.
Henry, V. J., Jr. and J. II. Hoyt (19(58). Quaternary paralic shelf
sediments of Georgia. Southeast. Geol., 9: 195-214.
Iloltedahl, II. (1958). Some remarks on the geomorphology of the
continental shelves off Norway, Labrador, and southeast
Alaska. ./. Geol., 66: 4(51 471.
Iloltedahl, O. (1940). The Submarine Relief off the Norwegian Coast.
Oslo: Norske Videnskaps-Akade, 43 pp.
Houbolt, J. J. II. C. (19(58). Recent sediments in the southern
bight of the North Sea. Geol. Mijnhouu; 17: 245-273.
Hoyt, J. II, and J. R. Hails (19(57). Pleistocene shoreline sediments
in coastal Georgia: Deposition and modification. Science, 155:
1541-1543.
Hoyt, J. II. and V.J, Henry, Jr. (19157). Influence of island mi-
gration on barrier island sedimentation. Geol. Soc. Am. Bull.,
78: 77 86.
561
REFERENCES
349
Jerlov, N. G. (1958). Distribution of suspended material in the
Adriatic Sea. Arch. Oceanogr. Limnol., 11: 227-250.
Johnson, D. (1919). Shore Processes and Shoreline Development. New
York: Wiley, 585 pp. (2nd ed., 1938).
Joseph, J. (1953). Die Trubangsverbaltnisse in der Sudwestlichon
Nordsee VVahrend der "Gauss Fahrt" im Februar, Marz,
1952. Ber. Dtsch. U'iss. Komm.f. Meeresforsch. 13: 93-103.
Kaye, C. A. (1959). Shoreline features and quaternary shoreline
changes, Puerto Rico. U.S. Geol. Surv. Pro/. Pap., 317-B, 140 pp.
Kenyon, N. H. (1970). Sand ribbons of European tidal seas.
Afar. Geol.. 9: 25-39.
Kenyon, N. H. and A. H. Stride (1970). The tide-swept conti-
nental shelf sediments between the Shetland Isles and France.
Sedimenlology, 14: 159-175.
King, L. H., B. Maclean, and G. Drapeau (1972). The Scotian
shelf submarine end moraine complex. In 24th Int. Geol.
Congr., Montreal, pp. 237-239.
Knebel, H.J. and J. S. Creager (1973). Sedimentary environments
of the east-central Bering Sea continental shelf. Mar. Geol.,
15: 25-47.
Kumar, N. and J. E. Sanders (1970). Are basal transgressive
sands chiefly inlet filling sands? Marit. Sediments, 6: 12-14.
Laursen, E. M. (1958). The total sediment load of streams. Proc.
Am. Soc. Civil Eng., 84(HV1): 1530.
McCave, I. N. (1971). Sand waves in the North Sea oflf the coast
of Holland. Mar. Geol., 10: 149-227.
McCave, I. N. (1972). Transport and escape of fine-grained sedi-
ment from shelf areas. In D. J. P. Swift, D. B. Duane, and
O. H. Pilkey, eds., Shelf Sediment Transport: Process and Pattern.
Stroudsburg, Pa.: Dowden, Hutchinson & Ross, pp. 225-248.
McClennen, C. F. (1973). New Jersey continental shelf near bot-
tom current meter records and recent sediment activity. ./.
Sediment. Petrol., 43: 371-380.
McClennen, C. E. and R. L. McMaster (1971). Probable Holo-
cene transgressive effects on geomorphic features of the conti-
nental shelf off New Jersey, United States. Marit. Sediments,
7: 69-72.
McKinney, T. F. (1974). Large-scale current lineations on the
Great Egg shoal retreat massif, New Jersey shelf. Investiga-
tion by side-scan sonar. J. Sediment. Petrol., It: 79-102.
McKinney, T. F., VV. L. Stubblefield, and D. J. P. Swift (1974).
Large-scale current lineations on the central New Jersey shelf:
Investigations by side scan sonar. Mar. Geol., 17: 79-102.
McManus, D. A., J. C. Kelley, and J. S. Creager (1969). Conti-
nental shelf sedimentation in an arctic environment. Geol. Soc.
Am. Bull., 80: 1961-1984.
McMaster, R. L. and T. P. Lachance (1969). Northwestern
African continental shelf sediments. Alar. Geol., 7: 57-67.
McMaster, R. L., T. P. Lachance, and A. Ashraf (1970). Conti-
nental shelf geomorphic features off Portuguese Guinea, Gui-
nea, and Sierra Leone (West Africa). Mar. Geol., 9: 203-
213.
Meade, R. H. (1969). Landward transport of bottom sediments
in the estuaries of the Atlantic coastal plain. J. Sediment.
Petrol., 39: 229-234.
Merrill, A. S., K. O. Emery, and M. Rubin (1965). Oyster shells
on the American continental shelf. Science, 147: 395-400.
Milliman, J. D. and K. O. Emery (1968). Sea levels during the
past 35,000 years. Science, 162: 1121-1123.
Milliman, J. D., C. P. Summerhayes, and H. T. Barretto (in
press). Quaternary sedimentation on the Amazon continental
margin: A model. Geol. Soc. Am. Bull.
Naumann, C. T. (1858). Lehrbuch der Geognosie, Bd. 1, 2nd ed.
Leipzig: Wilhelm Engelmann, 1000 pp.
Nichols, M. M. (1972). Inner Shelf Sediments off Chesapeake Bay. I.
General Lithology and Composition. Spec. Sci. Rep. 64, Virginia
Institute of Marine Science, Gloucester Point, Va., 20 pp.
Niino, H. and K. O. Emery (1966). Continental shelf off north-
eastern Asia. J. Sediment. Petrol., 36: 152-161.
Norris, R. M. (1972). Shell and gravel layers, western continental
shelf, New Zealand. N.Z. J. Geol. Geophys., 15: 572-589.
Oaks, R. Q., Jr. and N. K. Coch (1963). Pleistocene sea levels,
southeastern Virginia. Science, 140: 979-983.
Pilkey, O. H. and M. E. Field (1972). Onshore transportation of
continental shelf sediment: Atlantic southeastern United
States. In Shelf Sediment Transport: Process and Pattern. Strouds-
burg, Pa.: Dowden, Hutchinson & Ross, pp. 424-446.
Pilkey, O. H. and J. Frankenberg (1964). The relict-Recent sedi-
ment boundary on the Georgia continental shelf. Georgia
Acad. Set. Bull., 22: 1-4.
Powers, M. C. and B. Kinsman (1953). Shell accumulations in
underwater sediments and the thickness of the traction zone.
J. Sediment. Petrol., 23: 229-234.
Redfield, A. C. (1956). The influence of the continental shelf on
the tides of the Atlantic coast of the United States. J. Mar.
Res., 17: 432-448.
Robinson, A. H. W. (1966). Residual currents in relation to
shoreline evolution of the East Anglian coast. Mar. Geol., 4:
57-84.
Sarnthein, M. (1972). Sediments and history of the postglacial
transgression in the Persian Gulf and Northwest Gulf of
Oman. Mar. Geol., 12: 245-266.
Sharma, G. D., A. S. Naidu, and D. W. Hood (1972). Bristol
Bay: Model contemporary graded shelf. Am. Assoc. Pet. Geol.
Bull., 56: 2000-2012.
Shepard, F. P. (1932). Sediments on continental shelves. Geol.
Soc. Am. Bull., 43: 1017-1034.
Shepard, F. P. (1973). Submarine Geology. 3rd Edition. New York:
Harper & Row, 517 pp.
Sheridan, R. E., C. E. Dill, Jr., and J. C. Kraft (1974). Holocene
sedimentary environment of the Atlantic inner shelf off Dela-
ware. Geol. Soc. Am. Bull., 85: 1319-1328.
Shideler, G. L., D. J. P. Swift, G. H. Johnson, and B. W. Holliday
(1972). Late Quaternary stratigraphy of the inner Virginia
continental shelf: A proposed standard section. Geol. Soc. Am.
Bull., 83: 1787-1804.
Smith, J. D. (1969). Geomorphology of a sand ridge. J. Geol.,
77: 39-55.
Stahl, L., J. Koczan, and D. Swift (1974). Anatomy of a shore-
face-connected ridge system on the New Jersey shelf: Impli-
cations for genesis of the shelf surficial sand sheet. Geology, 2:
117-120.
Stanley, D. J. (1969). Submarine channel deposits and their fossil
analogs (fluxoturbidites). In D. J. Stanley, ed., The New Con-
cepts of Continental Margin Sedimentation. Washington, D.C.:
Am. Geol. Inst., pp. DJS-9-1 to DJS-9-17.
Stanley, D. J., G. Drapeau, and A. E. Cok (1968). Submerged
terraces on the Nova Scotian shelf. Z. Geomorphol., Suppl., 7:
85-94.
562
350
CONTINENTAL SHELF SEDIMENTATION
Stanley, D. J., D. J. P. Swift, and H. G. Richards (1967). Fossilif-
erous concretions on Georges Bank. J. Sediment. Petrol., 37:
1070-1083.
Stanley, D. J., D. J. P. Swift, N. Silverberg, N. P. James, and
R. G. Sutton (1972). Late Quaternary progradation and sand
spill-over on the outer continental margin cff Nova Scotia, Southeast
Canada. Smithsonian Contr. Earth Sciences, 8: 88 pp.
Steers, J. A. (1971). The East Coast floods, 31 January-1 February,
1953. In J. A. Steers, ed., Applied Coastal Geomorphology. Cam-
bridge, Mass.: MIT Press, pp. 198-224, 227 pp. (reprinted).
Stride, A. H. (1963). Current-swept sea floors near the southern
half of Great Britain. Q. J. Geol. Soc. London, 119: 175-199.
Stride, A. H., R. H. Belderson, and N. H. Kenyon (1972). Longi-
tudinal Furrows and Depositional Sand Bodies of the English Chan-
nel. Mem. Bur. Rech. Geol. Min. No. 79, pp. 233-244.
Stubblefield, VV. L., J. W. Lavelle, T. F. McKinney, and D. J. P.
Swift (1975). Sediment response to the present hydraulic
regime on the Central New Jersey shelf. J. sediment. Petrol., 45:
337-358.
Stubblefield, W. L. and D. J. P. Swift (in press). Ridge develop-
ment as revealed by sub-bottom profiles on the central New
Jersey shelf. Submitted to Mar. Geol.
Swift, D. J. P. (1973). Delaware shelf valley: Estuary retreat path,
not drowned river valley. Geol. Soc. Am. Bull., 84: 2743-2748.
Swift, D.J. P. (1974). Continental shell sedimentation. In C. A.
Burke and C. L. Drake, eds., The Geology oj Continental Mar-
gins. Berlin and New York: Springer-Verlag, pp. 117-133.
Swift, D.J. P. (1975). Tidal sand ridges and shoal retreat massifs.
Mar. Geol., 18: 105-134.
Swift, D. J. P. (in press). Continental shelf sedimentation. In R.
Fairbridge, ed., Encyclopedia of Sedimenlology. Princeton, N.J.:
Van-Nostrand-Reinhold.
Swift, D. J. P., D. B. Duane, and T. F. McKinney (1973). Ridge
and swale topography of the Middle Atlantic Bight, North
America: Secular response to the Holocene hydraulic regime.
Mar. Geol., 15: 227-247.
Swift, D. J. P. and S. D. Heron, Jr. (1969). Stratigraphy of the
Carolina Cretaceous. Southeast. Geol., 10: 201-245.
Swift, D.J. P., J. W. Kofoed, F. P. Saulsbury, and P. Sears (1972).
Holocene evolution of the shelf surface, central and southern
Atlantic coast ot North America. In D. J. P. Swift, D. B.
Duane, and O. H. Pilkey, eds., Shelf Sediment Transport: Process
and Pattern. Slroudsburg, Dowden, Hutchinson & Ross, pp.
499-574.
Swift, D. J. P. and P. Sears (1974). Estuarine and littoral deposi-
tional patterns in the surficial sand sheet, central and southern
Atlantic shelf of North America. In International Symposium on
Interrelationship of Estuarine and Continental Shelf Sedimentation.
Inst. Geol. du Bassin d'Aquitaine, Mem. 7, pp. 171-189.
Swift, D. J. P., D. J. Stanley, and J. R. Curray (1971). Relict
sediments, a reconsideration. J. Geol., 79: 322-346.
Todd, T. W. (1968). Dynamic diversion: influence of longshore
current-tidal floor interaction on Chenier and barrier island
plains. J. Sediment. Petrol., 38: 734-746.
Uchupi, E. (1968). Atlantic continental shelf and slope of the
United States— Physiography. U.S. Geol. Surv. Prof. Pap. 529-C,
30 pp.
Uchupi, E. (1970). Atlantic continental shelf and slope of the
United States: Shallow structure. U.S. Geol. Surv. Prof. Pap.
524-1, 44 pp.
Van Andel, Tj. H. and J. R. Curray (1960). Regional aspects of
modern sedimentation in the northern Gulf of Mexico and
similar basins, and Paleogeographic significance. In F. P.
Shepard, F. B. Phegler and T. H. Van Andel, eds., Recent
sediments, northwest Gulf of Mexico. Tulsa: Am. Assoc. Pet.
Geol., pp. 345-364.
Van Andel, Tj. H. and J. J. Veevers (1967). Morphology and Sediments
of the Timor Sea. Commonwealth of Australia, Bureau of
Mineral Resources, Geology and Geophysics, Bull. 83, 172 pp.
Van Straaten, L. M. J. V. (1965). Sedimentation in the north-
western part of the Adriatic Sea. In VV. F. Whitland and R.
Bradshaw, eds., Submarine Geology and Geophysics. London:
Butterworths, pp. 143-162.
Veenstra, H. J. (1969). Gravels of the southern North Sea. Mar.
Geol., 7: 449-464.
Venkatarathnam, K. (1968). Studies on the sediments of the con-
tinental shelf off the regions Visakhapatnam-Pudimadaka and
Pulicat Lake-Penner River confluence along the east coast
of India. Natl. Inst. Set. India Bull., 38: 472-482.
Visher, G. S. and J. D. Howard (1974). Dynamic relationship
between hydraulic and sedimentation in the Altamaha estu-
ary. J. Sediment. Petrol., 44: 502-521.
Whitmore, F. C, Jr., K. O. Emery, H. B. S. Cooke, and D. J. P.
Swift (1967). Elephant teeth from the Atlantic continental
shelf. Science, 156: 1477-1481.
Wright, L. D. and J. M. Coleman (1973). Variations in morphol-
ogy of major river deltas as functions of ocean wave and river
discharge regimes. Am. Assoc. Pet. Geol. Bull., 57: 320-348.
Wright, L. D. and J. M. Coleman (1974). Mississippi River mouth
processes: Effluent dynamics and morphological development.
J. Geol., 82: 751-778.
563
50
Reprinted from: Marine Sediment Transport and Environmental Management, D. J.
Stanley and D. J. P. Swift, editors, John Wiley and Sons, Inc., Chapter 10,
159-196.
CHAPTER
10
Substrate Response to Hydraulic Process:
Gram-Size Frequency Distributions and Bed Forms
DONALD J. P. SWIFT
Atlantic Oceanographic and Meteorological Laboratories, Miami, Florida
JOHN C. LUDWICK
Institute of Oceanography, Old Dominion University, Xorfolk, Virginia
Chapters 8 and 9 dealt with the entrainment of sand
and mud, respectively, on the continental shelf. In addi-
tion, Chapter 8 discussed the most ubiquitous shelf bed
form, the sand ripple formed by bottom wave surge,
since it plays a critical role in the entrainment and trans-
port of sand on the continental shelf.
This chapter explores further the response of the shelf
floor to the hydraulic climate. Two key responses that
are used to infer regional patterns of sediment transport
are grain-size frequency distributions and substrate bed
forms. The chapter also describes a numerical model for
estimating sediment transport and areas and rates of
erosion and deposition.
GRAIN-SIZE FREQUENCY DISTRIBUTIONS
Krumbein (1934) was the first to bring to popular atten-
tion the concept that the size frequency distribution of
sand samples tends to be log-normally distributed. It has
become a tenet of conventional wisdom that this dis-
tribution, as defined by its mean and standard distribu-
tion, is the signature of the depositional event, and that
deviations from log normality, as measured in terms of
standard deviation, skewness, and kurtosis, reflect both
the provenance and subsequent hydraulic history of the
sediment (see Inman, 1949; Friedman, 1961; Visher,
1969).
Genesis of the Normal Curve
Recent theoretical studies (Middleton, 1968; Swift et al.,
1972b) have attempted to present this hypothesis in a
more rigorous manner by consideration of probability
theory. The reader is referred to these papers for the
mathematical foundation of the following discussion.
The probability model for the genesis of a log-nor-
mally distributed grain population considers a flow over
a sand substrate in which the total load is adjusted to
flow conditions. If deposition is to occur, there must be
a decrease in bottom shear stress ( — dro/dx) and dis-
charge ( — bq/dx) down the transport path. The distri-
bution of grain sizes in the load undergoing transport
down this shear stress gradient and the absolute value
of the gradient is such that for each grain-size class, an
upstream portion of the path is experiencing supercritical
stress, and a downstream portion is experiencing sub-
critical stress. We are concerned with the central portion
564
159
160
SfBSTRATF. RESPONSE TO HYDRAULIC PROCESS
of the transport path, where a series of transition points
for critical shear stress occur, with each successive down-
stream transition point being associated with a succes-
sively finer grain-size class. The grains are assumed to
travel down the transport path in a series of discrete
hops as a consequence of the turbulent structure of
the flow, and as a consequence of a larger scale
cycle of flow events separated by periods of quiescence.
The model is thus a stochastic model, with an inherently
random aspect to its behavior, and the problem may
be dealt with in terms of probability theory.
Under these conditions, it is conceptually possible to
define the grain-size frequency distribution at each point
as the product of two probability vectors, an admittance
vector and a retention vector (Fig. 1). The admittance
vector is the sequence of probabilities of entrance of the
size classes present, ordered in sequence of decreasing
grain size. The retention vector is similarly the sequence
of probabilities of retention of successively finer grain
sizes.
PHI UNITS
(b)
FIGURE 1. Grain-size frequency distributions as a product of a
retention vector and an admittance vector. See text for explanation.
From Swift et al. (1972b).
If P]n is an element in an admittance vector, where j
denotes thej'th station in the transport path and n de-
notes one of n grain-size classes, and if P'}n is a corre-
sponding element in a retention vector for the same
station, then the product of the two probabilities,
Pjn( 1 — P'jn) gives the probability that the particle in the
local input enters but does not leave the station. The
product of the input vector with all corresponding ele-
ments in the admittance and retention vectors for a sta-
tion gives the frequency distribution for that station
(Fig. 1). This is a restatement, in probabilistic terms, of
the intuitively apparent fact that the modal diameter of
a deposit is that grain size most likely to arrive and least
likely to be carried away from the place of deposition
under prevailing flow conditions; progressively coarser
sizes are progressively less frequent because they are less
likely to arrive, and progressively finer sizes are progres-
sively less frequent because they are more likely to be
carried away.
In Fig. la, the two linear numerical filters (admit-
tance and retention vectors) are applied to a local input
frequency distribution that is uniform in nature and a
symmetrical retained frequency distribution results. If,
however, the local input has a skewed distribution (Fig.
lb), then the retained distribution is still skewed, al-
though it has been modified by the station probabilities.
If the filters are not linear, then further modification of
the input vector occurs.
In Fig. 2, various hypothetical input distributions are
subjected to sorting down the stations of a hypothetical
transport path according to the probabilistic algorithm
described above. In column A, an initially rectangular
distribution is seen to evolve into a distribution with a
distinct mode, and the mode is seen to shift toward the
finer end of the distribution at successive stations. The
coarse flank of the mode becomes visibly sigmoid
(S-shaped) as is characteristic of the side of the normal
distribution frequency curve. The increasingly sigmoid
shape is the consequence of the multiplication of succes-
sive admittance vectors in order to obtain the coarse
admixture of the local input frequency distribution. For
instance, if the admittance vector has the form
0.100, 0.200, 0.300, 0.400, 0.500, 0.600,
0.700, 0.800, 0.900, 1.00
and retains this form from station to station, then at the
third station the size frequency distribution of the coarse
admixture will be determined primarily by the third
power of the admittance vector,
0.001, 0.008, 0.027, 0.064, 0.125, 0.216,
0.343, 0.512, 0.729, 1.00
565
RECTANGULAR
INPUT
20
INPUT 0
20
STN 1
•1
20n
STN 2 0J
20
STN 3 0
20
STN 4
iilllll
i 20]
STN 5 > 0
o 1
z
=> 30
o J
STN 6
STN 7
STN 8
40 -i
40-
r
-rr{]\
40-
STN 9 0
Ldrif.
50
STN 10
B
NORMAL
INPUT
30
30
0]i
k.
30
,iiL
•1 ' 4
30-
,:..
30
30
-rl l 1 11 h.
30
m 3"
Jlllk
4
i
30
1
it
30-
[}l
cr62
X-34
_=£fl
30-i
■1 4
PHI UNITS
•1 4
PHI UNITS
EXPONENTIAL EXPONENTIAL
SIZE VALUES DISTANCE VALUES
30n
- JTL 20H m
oimllL oVfllL 0A
30
- ro
30
rfh
30
rliii
30n
\
i-jliilK.
i
20^
20-|
0
m
i
30 -|
L
o
30-i
■J
A
•l
30n
HI
30-
Jm oLJml
l
PHI UNITS
40-
JL. nil 11 Ik
40
40n
40-,
40
40-
40
] m
40
40
60^
J L
■1 4
PHI UNITS
FIGURE 2. Grain-size frequency distributions along sediment transport paths under
different conditions. See text for explanation. From Swift et al. (1972b).
566
161
162
Sl'BSTRATE RESPONSE TO HYDRAULIC PROCESS
The initially small probabilities have decreased more
than the initially large ones, and the resulting curve of
frequency against size class will be exponential in form.
The modal shift is the phenomenon of progressive sorting
(Russell, 1939) whereby the deposit becomes finer down
the transport path, as a consequence of the steady de-
pletion of the transported material in coarser particles.
In physical terms, this means that the coarsest particles
tend to get left behind whenever the bottom is eroded
by a flow event that is weaker than the one that preceded
it.
If it is assumed that the input distribution is normal
to begin with (Fig. 2, column B), and if it is assumed
that the probabilities of admittance and retention vary
linearly with grain size, then the mode shifts toward the
fine end of the distribution as the sediment is traced
down the transport path with no change in the shape of
the normal curve. However, in a more realistic case,
the probabilities of admittance and retention are assumed
to vary exponentially with grain size; in other words,
the transport rate varies exponentially with grain size.
As a consequence, vector multiplication acts on the two
sides of the frequency curve in a dissimilar fashion (col-
umn C). The greater range of transport probabilities as-
signed to the coarser sands results in greater efficiency
of sorting on that side of the curve, and progressive
steepening of that side, as the sediment is traced down
the transport path. The sediment becomes increasingly
enriched in the fine admixture at the expense of the
coarse admixture (becomes fine-skewed), as well as be-
coming finer down the transport path. This effect is
particularly marked where the intensity of the flow field
is made to decrease down the transport path (column D).
Size Frequency Subpopulations and Flow Regimes
Thus there are at least theoretical reasons supporting the
concept that the size frequency distribution of fluid-de-
posited sands constitutes hydraulic signatures of the flow
process. Attempts to interpret these signatures have in
general generated more heat than light (Emery and
Uchupi, 1972, p. 375). However, the analysis of the
subpopulations constituting sand samples has proved
more fruitful. The basic work has been undertaken by
Moss (1962, 1963, 1972). He notes that most grain-size
frequency distributions of sand deposited from fluid flow
do not plot as a straight line on probability paper as
they should if they arc normally distributed. Instead the
curves are Z-shaped (Fig. 3). He has demonstrated that
these Z-shaped curves are composite distributions and
are the consequence of the presence of three or more
log-normally distributed subpopulations, and that these
subpopulations are an outcome of the manner in which
PHI SCALE
067 MM SCALE
FIGURE 3- Cumulative curve of a swash zone sample. In Moss'
terminology, A is framework population, B is interstitial population,
and C is contact population. From Visher (J969).
the bed is built (Fig. 4). A. framework population (A popu-
lation) constitutes the bulk of the sample. Its modal di-
ameter is a function of the average dimensions of the
relatively large spaces between grains on the aggrading
surface. There is a strong feedback in this system between
deposit grain size and bed load grain size; the dimen-
sions of grains selected from bed load for deposition in
such holes depend on the dimensions of grains already
deposited, which in turn depend on the dimensions of
available grains in the bed load, and ultimately on the
dimensions of the hydraulic parameters of the flow.
A fine interstitial subpopulation (B population: fine tail
of the size frequency distribution curve) consists of
grains that are small enough to filter into the interstices
of the grain framework of the deposit. Their average
diameter is not that of the bowl-shaped openings on the
bed surface but the smaller average diameter of the
interstices within the deposit.
A coarse contact subpopulation (C population: coarse tail
of the frequency curve) consists of grains that are too
coarse to fit into or through the surface openings as do
the grains of the A and B populations. Instead they
accumulate as slowly moving to stationary clogs of mu-
tually interfering coarse grains on the bed surface. When
a critical area of these rejected coarse particles has ac-
567
GRAIN-SIZE FREQUENCY DISTRIBUTIONS 163
10-
FIGURE 4. Size-frequency curves of sands from various environments. Curves have been
dissected to indicate subpopulations. From McKinney and Friedman (1970).
cumulated, it will be buried beneath further layers of A
population grains.
It is important to assess the relationship between Moss*
rather sophisticated theory of subpopulation genesis, and
the prevailing equation of transport modes with sub-
population characteristics. It has been generally assumed
[see the review by Yisher ( 1969)] that the contact popu-
lation represents particles moved by dragging; or rolling,
the framework population represents particles moving by
saltation, and the interstitial population represents par-
ticles traveling in suspension. There is a correlation be-
tween these differing modes of transport and the per-
centage of respective subpopulations in the deposit, be-
cause each of these modes is most likely to carry the
appropriate size of material for the subpopulations with
which they have been correlated. The relative percent-
ages of subpopulations, however, are a direct consequence
of mechanisms of bed construction, and only indirectly
reflect modes of transport. Moss has shown, for instance,
that both B and A subpopulations may be generated
from saltative transport alone.
The percentages of these three populations in a given
deposit will vary, within limits set by grain geometry and
grain interaction processes, according to the regional
TABLE 1. Nomenclature and Grain-Size Characteristics of Sediment Flow Regimes
Population
Southard and
Moss (1972)
Boguchwal (197
3)
A
Fine ripple stage
Ripples
Dominant
Coarse ripple stage
Ripples
Dominant
Dune stage
Dunes
Dominant
Rheologic stage
Transition
Upper flat
Antidunes
bed
Dominant
Mean Diameter
(Moss. 1972)
Abundant Scarce
Scarce
Scarce
Scarce
Scarce
Abundant Abundant
0.07-0.25 mm
(3.75-2.00*)
0.25-0.92 mm
(2.00-0.25*)
0.25-2.2 mm
(2.0 to -1.1*)
0.17-4.8 mm
(2.6 to -2.3*)
568
164
SUBSTRATE RESPONSE TO HYDRAULIC PROCESS
availability of the three populations, and also according
to the hydraulic microclimate of the bed. Moss (1972),
on the basis of flume studies and studies of river deposits,
has described five bed regimes. These may be correlated
w ith the flow regimes described by Southard and Boguch-
wal (1973, Fig. 23). Each tends to form a characteristic
admixture of subpopulations (Table 1). Moss (1972)
notes that in the fine ripple stage, grains do not pro-
trude through the laminar sublayer of the bottom bound-
ary layer of the flow and microturbulence is absent from
the bed surface. Fine particles can become concentrated
near the bed. and can pass copiously into the interstices.
Hence the fine ripple regime is characterized by an
abundant B population.
In the coarse ripple stage and dune stage, grains protrude
through the lamina sublayer. Fluid dynamic lift and
bed grain turbulence operate to keep fine particles from
being concentrated near the bed, and the interstitial (B)
population is normally a minor bed constituent.
In the rheologic stage, flow is supercritical, and bed load
particle behavior is dominated by the dispersive pressure
associated with grain collisions (Bagnold, 1954,). These
pressures force the particles against and into the bed.
This effect is evidently dominant over the lift forces
which act at the bed, and the interstitial B population
again passes copiously into the bed. The rheologic stage
is furthermore the only stage in which Moss observed
an abundant contact (C) population.
Moss' theory may thus be used to infer flow regime
from the grain-size distribution. It must be applied with
caution, however, as it was developed for quasi-steady
flows, and the continental margin environment tends to
be subjected to an additional oscillatory flow component
because of wave surge. Grain-size distributions conse-
quently tend to indicate more intense unidirectional
flows than actually exist (Stubblefield et al., 1975).
BED FORMS
In this section.it is necessary to deal with more varied
and larger bed forms than the wave ripples described in
Chapter 7. Sand wave fields and sand ridge fields may
generate bed form spacings of a kilometer or more, and
bed form amplitudes of up to 30 m. Such large-scale
bed form arrays become significant storage elements in
continental margin sediment budgets, and such budgets
cannot be understood without an awareness of bed form
mechanics. Furthermore, large-scale bed forms impact
directly on human usage of the continental margin.
Large tankers navigate the Thames estuary channels
(Langhorne. 1973) with scant meters of clearance over
sand wave crestv Sewage outfalls and nuclear power
plants are planned or are being constructed in the'inner
shelf ridge fields of the Atlantic shelf. Seafloor well heads
are subject to burial by migrating bed forms.
concepts. A bed form is an irregularity in the par-
ticulate substrate of a fluid flow. This definition includes
the subaqueous sand wave and sand ridge fields of the
earth's shelves, the subaerial dune fields of the earth's
deserts and those photographed on Mars, and the bed
forms of the base surge deposits surrounding the lunar
craters, sedimented out of a transient fluid of gas, dust,
and debris generated by the impact of meteors. Bed forms
are not independent phenomena; they are equilibrium
configurations of the interface between a mobile, usually
cohesionless substrate, and an overlying flow field, and
tend to occur in repetitious arrays rather than alone.
They are the product of feedback between flow structure
and substrate structure. The three-dimensional pattern
of flow does not "cause" the bed form to arise, nor does
the bed form "cause" the deformation of the boundary
layer of the flow field; instead, strictly speaking, these
two elements of a flow-substrate system interact to cause
each other.
Wilson (1972, p. 204) notes that when a fluid is
sheared, either against another fluid, against itself, or
against a rigid boundary, there are many situations in
which secondary flows develop. Secondary flows are
regularly repeated patterns of velocity variation super-
imposed on the mean flow. The primary flows satisfy
the three continuity laws (of mass, energy, and momen-
tum), but in such a way that any small disturbance is
initially self-aggravating; in other words, the flow is an
unstable system. In sheared flows, this usually involves
the development of any combination of such secondary
flows as transverse internal waves, or transverse or flow-
parallel vortices. Such secondary flows may occur simul-
taneously at several scales.
Wilson further notes that sheared fluids may become
unstable in response to almost any sort of strong gradient
in velocity, pressure, viscosity, temperature, or density
in the direction normal to the shear force. These may
arise over completely plane beds. Eventually, however,
as the perturbed flow and the bed deform in response to
each other, a new stable state is attained.
The theory of fluid instability has been outlined by
Lin (1955), Chandrasekhar (1961), Rosenhead (1963),
and Yih ( 1965), and these authors have discussed many-
cases to which it has been applied. Allen ( 1968a, p. 50)
has summarized their computational approach. The al-
gorithm requires that equations of motion be set up to
describe the fluid motion of interest. These equations
are solved to discover whether a small sinusoidal dis-
turbance of one variable will be damped or amplified
569
BED FORMS
165
under the chosen limits for other variables. The motion
is stable if the disturbance is damped, but unstable if it
is amplified. In nature the unstable disturbance is ampli-
fied until the other variables of the system set some lim-
iting condition on the amplification and a new stale of
quasi equilibrium is attained. Stability analysis has been
successfully applied to the problem of ripple and sand
wave formation (e.g., Smith, 1969) and it seems likely
that all bed forms will ultimately prove susceptible to
this mode of attack.
BASIC MODES OF BED FORM BEHAVIOR. Most bed forms
fall into two basic categories: those that are oriented
across the flow direction, such as sand waves and ripples,
and those that are oriented parallel to the flow direction,
such as sand ribbons. These two basic patterns must
correspond to two basic patterns within the flow field
itself, a transverse pattern in which zones of scour and
aggradation alternate down the flow path, and one in
which zones of scour and deposition alternate across the
flow path (Figs. 5 and 6). There is considerable evidence
to indicate that this is the case, although the basic mech-
anisms are far from clear.
(a)
(b)
DEVELOPMENT
gj^ 2223
^r
lA
|K7
FIGURE 6. The development of a longitudinal bed form.
(a) The pattern of secondary flow over longitudinal bed form
elements: PP, flow attachment lines along ridge trough; QQ, flow
separation lines along crests, (b) Development of longitudinal
elements in vertical cross section perpendicular to mean flow
direction. (i) Form and flow components; (ii) components in z
direction; (Hi) components in x direction. Numbers as in Fig. 4-
Note alternate notation of coordinate axes. From Wilson (1972).
FIGURE 5. The development of a transverse bed form. (A)
Initiation; (B) growth; (C) equilibrium. (I) Sand transport rate;
(2) shear velocity at bed; (3) erosion rate; (4) streamlines. From
Wilson (1972).
Transverse Bed Forms
mode of formation. As noted by Wilson (1972),
most transverse bed forms are probably caused by trans-
verse wave perturbations in the flow. The problem is a
complex one, and the solutions offered to date have not
been altogether satisfactory. Summaries are presented by
Allen (1968a, pp. 130-149), Kennedy (1969, p. 151),
and Smith (1970, p. 5928).
Smith points out that many of these studies are un-
necessarily restrictive; they assume an eddy viscous mean
flow but neglect the inertial terms in the equation of
motion (Exner, 1925, in Raudkivi, 1967) or assume in-
viscid irrotational flow (Kennedy, 1969). These assump-
tions require an a priori phase shift in the velocity field
relative to the interface disturbance in order for insta-
bility to occur. Smith (1970) has undertaken a stability
analysis employing inertial terms in the equations of
motion. His results indicate that the interface is unstable
with respect to infinitesimal perturbations of wavelength
greater than the wavelength for which the inertia of the
grains is important (wavelengths less than 10 times mean
570
166
SUBSTRATE RESPONSE TO HYDRAULIC PROCESS
grain diameter). Smith utilizes the sediment continuity
equation, which may be presented in its simplest two-
dimensional form as
drj dq
dt
dx
(1)
where 77 is height of the interface above a datum, / is
time, k is a constant, q is sediment discharge at a level
near the bed, and x is horizontal distance. In physical
terms, the time rate of change of the height of the inter-
face at a point above the datum is proportional to the
horizontal discharge gradient at that point, assuming sat-
uration of the boundary flow with sediment; a decrease
in discharge across the point ( — bq/dx) must result in
aggradation, while an increase (dq/dx) must result in
erosion. Smith has rewritten the equation in terms of
boundary shear stress and discharge:
(2)
dt
\to dro/ dx
where c0 is the boundary concentration of sediment, q is
the mean volume flux of sediment per unit width, and
To is the local mean shear stress on the bed.
Smith's analysis divides the nonuniform horizontal ve-
locity along the waveform interface into an in-phase
component and an out-of-phase component. The in-
phase component consists of accelerating flow over crests
with maximum shear stress at those points, as required
by flow continuity. Since boundary erosion varies di-
rectly with dro/dx, this in-phase component simply
causes upstream erosion of the interface perturbation
and downstream deposition; the perturbation moves
downstream with neither growth nor decay. However,
the inertia of the high-velocity water of the upper part
of the water column causes it to converge with the rising
bed on the upstream side of an interface perturbation,
and there is as a consequence an out-of-phase component
of velocity and bottom shear stress which attains its
maximum value at this zone. This maximum persists
when the components are added; hence, since deposition
is proportional to — dro ''dx, some sand must always be
deposited on the crest of the perturbation.
Smith (1970) cites Exner's (1925) earlier stability
analysis as qualitatively correct, despite neglect of the
inertial terms in the momentum equations. Exner had
shown that when downcurrent spacing is wide, the
crests of perturbations move faster than the troughs be-
tween them, resulting in oversteeping of the downcur-
rent slope to the angle of repose (30° underwater), and
consequently, in the formation of a horizontal roller
eddy (wake, flow separation bubble) downstream of the
crest. The perturbation is now a mature ripple or sand
wave.
The generation of a wake behind a growing bed form
results in propagation of interface instability in the down-
stream direction. Smith (1970) cites Schlichting (1962,
p. 200) who has studied the development of a turbulent
wake. Behind a negative step such as the avalanche
slope of a growing transverse bed form, flow accelerates
downstream of the attachment line (Fig. 7) and at the
same time a boundary layer is initiated that grows in
height downstream. Shear increases downstream because
of flow acceleration, then decreases as the effects of
boundary layer growth dominate over the effects of flow
acceleration. Here, in a zone where dr0/i>x < 0, sand is
deposited and a new ripple grows, which in turn de-
forms, develops a wake, and triggers a third. Smith's
stability analysis does not specify wavelength for growing
bed form perturbations, and it is apparent that this
parameter must be defined by spatial adjustments in the
turbulent velocity field. As downstream ripples grow in
height and their separation bubbles in width, they must
grow in length, which is accomplished by the smaller
ripples moving faster, and stretching out the ripple field.
Saporotlon line
Separation line
Surface of
separation
Surface of
separation
Attachment line
FIGURE 7. Three-dimensional separated flows, (a) Roller; (b) vortex. Note alternative notation
of coordinate axes. From Allen (1970).
571
BED FORMS
167
Smith's scheme of transverse bed form formation by
the spontaneous deformation of the interface into a
moundlike perturbation, its increasing asymmetry, the
formation of a separation bubble, and the downstream
propagation of the instability, has been strikingly con-
firmed in experimental work by Southard and Dingier
(1971). Their work suggests that if the critical flow
threshold is approached slowly, preexisting bed irregu-
larities may trigger downstream ripples in the interval
of metastability before the threshold is attained. How-
ever, if the threshold is passed rapidly, or if marked
preexisting irregularities do not occur, mounds will spon-
taneously appear and transform themselves into regular
ripples.
Other schemes for the formation of transverse bed
forms have been proposed, in which the wavelike per-
turbation of flow precedes bed deformation, rather than
arising from interaction with the bed. Cartwright (1959)
has proposed that the shelf-edge sandwave field of
La Chapelle Bank in the Celtic Sea are responses to
stationary internal waves (tidal lee waves) in the strati-
fied water column. Furnes (1974) has analyzed the for-
mation of sand waves in response to internal waves of a
fluid whose density stratification is a consequence of its
suspended load. While compatible with the field evi-
dence, these models for sand wave formation remain un-
confirmed. They are important contributions, however,
if only in that they reduce the bias toward the results of
experimental laboratory work. The space and time scales
and the internal structure of shelf flows are qualitatively
different than those of laboratory flumes and there is no
reason to assume that such further modes of sand wave
formation do not exist in nature.
types of transverse bed forms. Field and labora-
tory observations show that there tend to be two over-
lapping populations of transverse bed forms: ripples, with
wavelengths up to 0.6 m, and sand waves, with wave-
lengths in excess of 0.6 m (Fig. 8). Sand waves com-
monly bear ripples on their backs. The two populations
appear to be responses to two distinct genetic mech-
anisms. As small forms grow up through the velocity
gradient of the boundary layer, the zone of maximum
stress on the upcurrent flank shifts to the crest, at which
point the entire upcurrent slope is erosional and the lee
slope depositional; upward growth is stopped, and the
ripple migrates at constant speed (Wilson, 1972, p. 200).
Transverse bed forms of larger wavelengths are insensi-
tive to the boundary layer velocity gradient and their
upflank zone of maximum shear stress shifts to the crest
only when the whole flow is significantly deformed by
their upward growth. As a consequence, the equilibrium
height of sand waves in shallow flows is proportional to
flow depth, while the equilibrium height of ripples is
10"
8
6
A,
I03
8
6
2 -
10'
1
1 —
— 1 1 1
•
i -i T - 1 —
•
o „0. o
-
o 8o°
0
•' 8/1
• 0 •/
• o •
0
8 *
•
t
/ .
O j
• •/
0
o/
•
*s
•
o
•
0
o y
.•»
• ." K
=10000
. _jC_
- /
0
» •:.
-
0
* • • • •
•
oa
••
-
• *»
••
■
• Ripples] 0.19,0.27,0.28,0.45
l
• Dunes j & 093 mm sonde
■ ■ i i 1 J — i —
10'
6 8
6 8.0«
FIGURE 8. Wavelength parallel to flow of experimental ripples
and sand waves in relation to flow depth. Data of Guy etal. (1966).
From Allen (1970).
depth independent for all flow depths (Allen, 1967); see
Fig. 8. Expressions for the equilibrium heights of ripples
and sand waves have been considered by Kennedy and
used by McCave (1971).
Stride (1970) has plotted measurements of height
versus depth for sand wave fields of the North Sea at
depths of 90 to 60 m, and found no correlation. Large
bed forms grow slowly, and equilibrium heights may be
rarely obtained in such shallow tidal seas subject to
strong periodic storm surges. Deep-sea sand waves (Lons-
dale and Malfait, 1974) can obviously never equilibrate
with total water depth, although the significant flow
depth may be only a small fraction of total depth,
because of density stratification.
The distinction between small- and large-scale bed
forms may be due to more than interaction of wavelength
with the velocity gradient. Kennedy ( 1 964) has suggested
that small transverse bed forms represent perturbations
of the traction and saltation loads that move very near
to or on the bed, and hence must react quickly to
changes of flow speed. Larger transverse bed forms, on
the other hand, could reflect a perturbation of the sus-
pended load, which will tend to respond slowly, and
therefore over a large distance to a change in flow speed.
572
168 SUBSTRATE RESPONSE TO HYDRAULIC PROCESS
cm 10
S - Separotion point or line
A - Attachment point or lins
FIGURE 9. Skin friction lines and streamlines associated with a portion of a bed of
experimental ripples in fine-grained quartz sand. Mean flow velocity 22 cm /sec from
left to right. Mean flow depth 4.5 cm. Note alternative notation of coordinate axes.
From Allen (1970).
In continental margin sand wave fields, there are often
three orders of transverse bed forms: current ripples,
sand waves, and larger sand waves. McCave (1971) sug-
gests that the two classes of sand waves may be the con-
sequence of Kennedy's two categories of substrate re-
sponse.
Because of the turbulent diffusion of sand normal to
the flow direction, an initially equant interface pertur-
bation will tend to extend itself normal to flow, hence
the quasi-two-dimensional nature of ripples and sand
waves (ripple profile does not change down the length
of the ripple crest). However, at increasing values of
mean velocity, and therefore of turbulent instantaneous
velocity component, transverse bed forms tend to become
three-dimensional (Znamenskaya, 1965). As crests be-
come locally inclined to the mean flow direction, the
horizontal "roller eddy" of the separation bubble be-
comes a horizontal helical vortex (Fig. 7) and irregular
patterns of skin flow result (Fig. 9). Under yet more in-
tense flows, the irregularities may take on ordered pat-
terns (Fig. 10). Bagnold (1956) attributes one particu-
larly common pattern, that of the lingoid ripple, to
"... the partial diversion of grain flow . . . and its fun-
neling into channels between existing ripples; deposition
(of a new lingoid ripple) would take place immediately
downstream of such a funnel." A diagonal or diamond-
like pattern of lingoid ripples results.
TRANSVERSE BED FORMS AND FLOW REGIMES. It has
long been known that as a shallow flow over a nonco-
hesive substrate intensifies, a sequence of bed configura-
tions transpires (Simons et al., 1961; Simons and Rich-
ardson, 1963; Guy et al., 1966). The flow variables gov-
erning this sequence are h, depth of flow; u, mean velocity
of flow; p, density of fluid; ps, density of sediment; /i,
viscosity of fluid; and D, mean diameter of sediment.
The critical parameters are fluid power (proportional
to «3; see Chapter 8) and grain size (Fig. 11). Grain
density is variable to the extent that heavy minerals
may be present; and fluid density and viscosity vary
somewhat with temperature and salinity. Flow depth
determines whether or not the flow is subcritical or
supercritical as expressed by the dimensionless Froude
number F = u/(gk)112, where (gh)112 is the celerity of a
shallow water wave. In supercritical flows (F > 1),
surface waves couple with substrate perturbations (anti-
dunes) that tend to migrate upcurrent. On the continen-
tal margin supercritical flows are confined to the swash
and breakpoint zones of the surf, and to tidal flats; and
the antidunes and rhomboid ripples that form in these
zones are ephemeral.
Southard (1971) and Southard and Boguchwal (1973)
have argued that bed configuration diagrams such as
Fig. 1 1 should be presented in terms of dimensionless
depth, velocity, and grain size to eliminate the overlap-
573
BED FORMS
169
FIGURE 10. (a) Lingoid ripple pattern on shelf floor off Cape
Hatteras, Sorth Carolina, (b) Lingoid ripples on back oj sand
ivave, straight ripples in trough, same area.
ping of fields that occurs in diagrams utilizing fluid
power or bed shear stress.
Figure 1 1 shows that dunes (sand waves) occur at
higher values of fluid power than do ripples by them-
selves. This fact is consonant with Kennedy's suggestion
that sand wave formation involves suspended load trans-
port, which requires higher values of fluid power than
does bed load transport.
TRANSVERSE BED FORMS AND TIDAL FLOWS. Tidal flows,
which reverse every 6 hours, generate transverse bed
forms in a cohesionless substrate. Tidal current ripples
are no different than ripples generated by unidirectional
currents, except that their sense of asymmetry is reversed
as the tide changes. Small sand waves (height of 1 m or
less) may have their asymmetries partly or wholly re-
versed by strong reversing tidal currents (Klein, 1970).
Larger sand waves tend to display a time-integrated
response to reversing tidal flows, maintaining an ebb or
flood asymmetry in accord with the dominant flow com-
ponent residual to the semidiurnal cycle. "Cat-backed"
sand waves are large sand waves that have a sloping
upcurrent side, a flat top, and (in profile) an •■ear"
perched on the edge of the downcurrent slope (Van Veen,
1936). The ear is a response to the subordinate portion
of the tidal cycle. Tide-formed sand waves in areas of
equal ebb and flood flow are commonly symmetrical.
As distance from shore increases, the tidal current is
no longer reversing but rotary (Chapter 5). The advent
of midtide cross flow tends to inhibit the formation of
sand waves large enough to survive through the tidal
cycle (McCave, 1971). Under such circumstances longi-
tudinal bed forms are favored (Smith, 1969).
Longitudinal Bed Forms
Wilson (1972) comments that practically all longitudinal
bed form elements, whether formed in wind or water,
are initiated by regular helical vortices with axes parallel
to flow. His reasons for his admittedly sweeping assertion
are as follows:
1 . Longitudinal helical flow cells occur in many dif-
ferent kinds of situations. They are the only kind of flow
perturbations known to fluid mechanics whose wave-
length is measured normal to the mean flow direction.
2. With the exception of alternating parallel lanes of
fast and slow flow, the double helical pattern is the only
one that meets the theoretical requirements, namely bi-
lateral symmetry parallel to flow, regular repetition nor-
mal to flow, and conformity with the law of continuity.
3. Many investigations of flow over longitudinal bed
forms resulted in some evidence for the occurrence of
helical flow over the longitudinal elements, for instance,
model ripples and dunes (Allen, 1968a); in riv«r chan-
nels (Gibson, 1909); over tidal sand ridges (Houbolt,
1968); and over desert dunes (Hanna, 1969).
The theory of longitudinal flow perturbation is less
well developed than the theory of transverse flow per-
turbation. Such perturbations are not as obvious in lab-
oratory flumes as transverse (streamwise) flow pertur-
bations, and many occur at scales far beyond those of
laboratory flumes. As in the case of transverse bed forms,
longitudinal bed forms appear to be able to form in re-
sponse to perturbations of boundary flow, or in response
to perturbations of the whole flow field. As in the case
of transverse bed forms, they appear to form during the
course of flow-substrate interaction, and also in response
to the preexisting internal structure of a sheared flow.
Preexisting flow structures appear to be more impor-
tant than in the case of transverse perturbations. Perhaps
the most general statement that can be made is that in
a sheared flow that is wide relative to its depth, a sig-
nificant portion of flow energy must be diverted to an
ordered secondary flow component, in order to maintain
lateral flow continuity. At least three basic varieties of
such secondary flow structure exist.
574
170
SUESTRATE RESPONSE TO HYDRAULIC PROCESS
20,000
10.000
8000
_ 6000
u
S 4000
£ 2000
3t 1000
o. 800
§ 600
w 400
200
40
PLANE BED
(Even lominotion)
RIPPLES
(Cross-lamination)
PLANE BED
(Even lamination)
NO SEDIMENT MOVEMENT
0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.1
D (cm)
FIGURE 1 1. Bed forms in relation to stream power and grain size. Data of
G. P. Williams, H. P. Guy, D. B. Simons, and E. V. Richardson. From Allen
(1970).
MICROSCALE LONGITUDINAL BED FORMS (PARTING
lineation). It has been repeatedly suggested that the
logarithmic boundary layer tends to be so patterned,
although an adequate analytic model has not yet been
devised (Schlichting, 1962, pp. 500-509). Kline (1967)
and Kline et al. (1967) have conducted dye experiments
in flumes which suggest that the laminar sublayer and
the lower part of the buffer sublayer of the turbulent
boundary layer have a structure characterized by vig-
orous transverse components of flow (see discussion,
p. 94). Dye introduced into the boundary layer forms
into bands that are more slowly moving than those in
the intervening water zones. Although the streaks are
randomly generated, they have a mean transverse spac-
ing of Xr = \00 i>u* in which v is the kinematic viscosity
and u> is the shear velocity (Kline, 1967). The response
to helically structured boundary flow over a cohesionless
particulate substrate is, however, well known; it is the
ubiquitous parting lineation (Sorby, 1859), so named
for the tendency of flagstones (silty sandstones with
strong bedding fissility) to exhibit lineations on bedding
planes. Closer examination reveals a waveform bedding
surface whose undulations parallel flow direction; ridges
are a few grain diameters high and are up to several
centimeters apart (Allen, 1964; 1968a, pp. 31-32); see
Fig. 12. There is clear evidence for the divergence and
convergence of bottom flow in that the azimuths of long
grains are bimodal, although this evidence does not re-
solve the secondary flow pattern. A similar structure has
been reported from mud beds (Allen, 1969). Here the
notches are frequently narrower than the ridges.
Coupling probably occurs between bed and flow struc-
ture, in that the grain ridges localize flow cells. Also, the
sand of the ridges is coarser (Allen, 1964) and the result-
ing roughness would tend to slow crestal flow. This
feature would cause downstream growth in the retarded
wake of the grain ridges, and would perhaps induce
upward ridge growth until ridge crests reach a level
whose flow is rapid enough to counteract growth.
MESOSCALE LONGITUDINAL BED FORMS (CURRENT
lineations). "Current lineations" (McKinney et al.
1974) is a generic term for low-amplitude strips of sand
resting on a coarser substrate (sand ribbons) and for
strips of coarse sand or gravel flooring and elongate de-
pression of slight depth (longitudinal furrows). Current
lineations are a larger scale of longitudinal bed form,
with spacings ranging from a few meters to many hun-
575
BED FORMS
171
15 20 25 30 35 40
M«an flow velocity (cm/sec)
FIGURE 12. The mean transverse spacing of parting lineations
as a junction of mean flow velocity and flow depth. From Allen
(1970).
dreds of meters (Allen, 1968a). They are best observed
by means of sidescan sonar (Figs. 13 and 14). The large-
scale patterns are characteristic of shelves with strong
tidal flows (Kenyon, 1970); see Fig. 15. Sand ribbons
and longitudinal furrows are probably the most common
mesoscale bed form on the continental shelf, being widely
distributed on both shelves dominated by tidal flows
(Kenyon, 1970; Belderson et al., 1972) and storm-domi-
nated shelves (Newton et al., 1973; McKinney et al.,
1974); see Fig. 14. Unpublished data of the Atlantic
Oceanographic and Meteorological Laboratories, Mi-
ami, Florida, show them to be characteristic of large
sectors of the Middle Atlantic Bight. Relief is negligible
relative to width. Kenyon would restrict the term "sand
ribbon" to features having length-to-width ratios of 1 : 40
and refers to shorter, broader features as elongate sand
patches, but the distinction seems arbitrary. Unlike
parting lineation ridges, sand ribbons tend to consist of
streamers of finer sand in transit over a coarser substrate
which may, in fact, be a gravel. A continuum may exist
between a sand ribbon pattern of sand and gravel streets
of equal width, to a "longitudinal furrow" pattern
(Stride et al., 1972; Newton et al., 1973) in which
widely spaced, elongate erosional windows in a thin
sand sheet reveal a coarser substrate. Ribbon width
relative to the width of the interribbon zone does not
appear to be simply a function of the height of a sinus-
oidal surface of sand layer over a coarser substrate, since
the windows in profile are notchlike affairs separated by
flat plateaulike zones (Fig. 16). Furthermore, the ribbons
are commonly rather asymmetrical, as though the sand
sheet occurs at minimum thickness on one side, and in-
creases very slowly to maximum thickness on the other
side. Such asymmetrical ribbons could be interpreted as
degraded sand waves, but the sharpness of the contacts
plus the lack of relief suggests instead asymmetrical
helical flow cells (Fig. 17).
Small sand ribbons may be large-scale analogs of the
responses described in the preceding section that involve
the entire logarithmic boundary layer. However, most
shelf sand ribbon patterns have spacings of tens or hun-
dreds of meters, and as noted by Allen (1970, p. 69),
can only be responses to the entire depth of flow.
Theoretical studies (Faller, 1971; Faller and Kaylor,
1966; Brown, 1971; Lilly, 1966) and experimental studies
(Faller, 1963) show that there is a mechanism by which
a helical flow pattern may be induced in the large-scale
flows of the continental margin. When such flows are
in geostrophic balance (pressure term balanced by Cori-
olis term in the equation of motion; see p. 25). The
lower portion of the flow is an Ekman boundary layer
(see p. 97). The basal meter behaves as a logarithmic
boundary layer in that flow speed decreases rapidly to
a zero value or nearly so at the seafloor. Flow direction
(in the northern hemisphere) is to the left of the free-
stream direction, however, since the Coriolis term is re-
duced along with mean velocity; the equation of motion
more nearly constitutes a balance between friction and
pressure terms. With increasing height off the bottom,
flow is more nearly geostrophic and its direction is more
nearly parallel to the isobars, until free stream condi-
tions are reached. Thus velocity vectors at successively
higher levels constitute a left-handed spiral. On the con-
tinental shelf, this lower Ekman boundary layer may
extend to the base of the mixed layer, if it exists, or to
the surface, where it is overprinted with a right-handed
Ekman spiral (upper Ekman boundary layer) because
of direct wind stress (Ekman, 1905, Plate 1).
Above a critical Reynolds number, this Ekman layer
is unstable. However, because the instability transpires
in an Ekman field subject to the Coriolis effect, the in-
stability does not result in random turbulence, but in-
stead in a regular pattern of secondary flow (Faller,
1971, pp. 223-225). In this pattern zones of surface
convergence, downwelling, and bottom divergence alter-
nate with zones of surface divergence, upwelling, and
bottom convergence. The resulting flow structure con-
sists of horizontal helical cells with alternating right-
and left-hand senses of rotation (Fig. 6). Angles of con-
vergence and divergence (pitch) are generally a few
degrees; in other words, the secondary component of
flow is weak, relative to the main geostrophic component.
The flow cells may occur at several scales (Faller and
Kaylor, 1966). In laboratory studies (Faller, 1963),
smaller scale cells have a spacing of approximately 1 ID,
where D is a characteristic depth of the Ekman layer,
and tend to be oriented up to 14° to the left of the mean
flow. They occur at Reynolds numbers above 125. Larger
576
172
SUBSTRATE RESPONSE TO HYDRAULIC PROCESS
A\ 111 UMIN.'.TlON DIRECTION
'ERO RANGE °-^— -*■
PRO' IE 0. SEA FLOOR ^^R DE
BENEATH The Ship pL \ I
VID UNES OF ECHOES 2
FROM SIDE LOBES
NEAR EDGE OF MA
DISTANT EDGE OF MAIN BEAM
FIGURE 13. (A) Sidescan sonar. (B) The resulting record. A,
Bottom of seafloor; B, turbulence in water column due to ship's
wake; C, zigzag pattern is due to refraction of sound in density-
stratified water; D, main lobe (see above); E, side lobe. From
Belderson et al. (2972).
scale cells have wavelengths much greater than 1 \D and
are oriented to the right of geostrophic flow. They occur
at much lower Reynolds numbers.
Helical flow structure may occur in the upper Ekman
layer where its wind-driven stirring creates the mixed
layer above the thermocline (Faller, 1971), or may occur
in the lower layer (Faller, 1963). In surface helical flow
the downwelling zones that collect the high-velocity
wind-driven surface water are more sharply defined than
the upwelling zones (Langmuir, 1925). In bottom helical
flow, downwelling zones deliver higher velocity water to
the seafloor, and may also be more sharply defined than
the upwelling zones. During intense flows, when strati-
fication breaks down and the layers partly or completely
overlap, a compound top-to-bottom helical flow struc-
ture might be expected.
Observational and theoretical studies required to link
this scheme to the observed shelf sand ribbon patterns
have not been undertaken; however, there are obvious
points of compatibility. The ribbons tend to be parallel,
or oriented at a small angle to the regional trend of shelf
contours, and presumably to the mean geostrophic flow
direction. The greater intensity of downwelling zones
would explain the dissimilar width of ribbons and inter-
vening erosional windows. The Reynolds numbers re-
quired are not excessive for either tide- or wind-driven
shelf flows.
10 m and spacings measured in kilometers, are called
sand ridges (Off, 1963; Swift et al., 1974); see Fig. 18.
They are comparable in scale to the seif dunes, and the
yet larger "draas" of the sand seas of the world's deserts
(Wilson, 1972), except that as befits submarine sand
bodies, their side slopes are much lower, usually being
measured in fractions of a degree.
Sand ridges appear to form in two basic types of situ-
ations. They are characteristic of the reversing flows of
tidal estuaries and bays, where they tend to form in
complex arrays parallel to the estuary axis (Figs. \8A,B).
They also appear on inner shelves of coasts undergoing
erosional retreat (Figs. 18C,Z)), where they appear to be
specific responses to the coastal boundary of the shelf
flow field (Duane et al., 1972; Robinson, 1966; Swift
et al., 1972a); the mechanism is discussed in detail in
Chapter 14. On the inner shelf, either tidal or storm
flows may be the forcing mechanism (Duane et al.,
1972; Swift, in press). The ridges tend to extend obliquely
seaward from the shoreface. Like sand ribbon patterns,
the generally larger scale sand ridge fields tend to com-
prise discontinuous sheets of finer sand over a coarser
substrate. However, where sand ridges build up into the
wave-agitated zone on open coasts, their crests tend to
be coarser than their flanks although generally not as
coarse as the substrate exposed in trough axes (Houbolt,
1968; Swift et al., 1972b; Stubblefield et al., 1975).
longitudinal sand ridges. Large-scale longitudinal
bed forms of the continental margin, with relief of up to
SAND RIDGES AS RESPONSES TO WIND-DRIVEN FLOWS.
Sand ridges are found on continental shelves seaward
577
BED FORMS 173
SEA FLOOR
NORTH OF
SHIP
WATER
COLUMN {
WATER {
COLUMN
SEA FLOOR
SOUTH OF
SHIP
FIGURE 13— Continued.
of active inner shelf-generating zones, and occur as
well on some shelves whose inner margins are not
actively forming them (Swift et al., 1974). It appears
that shelf flow fields can continue to maintain these
ridges of coastal origin after the retreating shoreline has
abandoned them, and can even- locally generate them
afresh (see discussion in Chapter 14). Without a main-
taining mechanism, shelf flows might be expected to de-
grade sand ridges by leveling crests and filling in troughs.
In fact, however, the ridges of the Atlantic continental
shelf tend to expose compact clays or lag gravels in their
troughs, indicating continuing trough scour (Swift et al.,
1972a; McKinney et al., 1974).
There are a variety of competing hydraulic mech-
anisms that may serve to explain the formation and
maintenance of large-scale sand ridges on the continental
margin, none of which is clearly understood. On the
open shelf, cellular flow structure in storm flows, as de-
scribed in the preceding section, may couple with the
shelf floor. Such cellular flow structure might generate
ridges along the coastal boundary (see discussion in
Chapter 14) where wind-driven flows are frequent and
intense and there is an abundant supply of sand. As
these ridges have been left behind by the retreating
shoreline during the Holocene transgression, the same
cellular flow structure may be continuing to maintain
them on the outer shelf.
If this analysis is correct, then sand ribbons and sand
ridges may differ in that sand ribbons represent responses
to one flow event or a flow season while sand ridges rep-
resent time-averaged responses to repeated flow events,
whose emerging relief tends to localize the position of
large-scale flow cells. Events capable of forming such
large-scale flow cells would presumably be peak storm
578
01*2
WW
kW-f
wmmv*
ittr
ridges of asymmetrical ribbons; 3: pinna (pele-
cypod) bed. (b) 1: Symmetrical ribbon; 2; sand
waves. From Newton et al. (1973).
i\1±
FIGURE 14. Sand ribbon patterns from the
Spanish Sahara shelf. Light is sand; dark is
coarser sand and gravel. Distortion ellipse with
scales on first record, (a) 1, 2: Sharply defined
174
579
BED FORMS
175
TYPE A
appron horizontal sraie
FIGURE 1 5. Categories of sand ribbon from the shelf around the British Isles, and associated current velocities.
From Kenyon (1970).
or tidal flows, in which secondary circulation involves
the entire water column.
The most problematic aspect of shelf ridge fields is
the depth-to-width ratios of the troughs, which range
from 1 : 10 to 1 : 150. The smaller ratios are compatible
with the "type I" flow cells of Faller's (1963) experi-
mental work, whereas the large ratios may derive from
Faller's "type II" cells which have "much greater" di-
ui)
FIGL'RE 16. (a) Wide sand ribbons alternating uith narrow
streets of coarse sand (erosional uindous) due to intersection of
sinusoidal surface of sand sheet uith horizontal substrate, (b) Same
pattern due to notchlike incision of erosional uindous. The latter
pattern is a common one.
mensions. It is perhaps easier to conceive of such flat-
tened cells if it is remembered that the central down-
welling zone is the only sharply defined portion of a
double helical flow cell; the marginal zones of diffuse
upwelling may take up much of the "stretch," serving
to complete flow continuity in a fashion analogous to
the role of "ground" in electrical circuitry.
The advent of appreciable relief in a growing system
of sand ridges may bring other hydraulic mechanisms
into play. Secondary flow cells appear to be an innate
response to channeled flow. It has long been known
(Gibson, 1909; Jeffreys, 1929; Einstein and Li, 1958;
Leopold et al., 1964, pp. 251-284; Wilson, 1972) that
driftwood or ice in a river tends to move toward the
center, whose surface is elevated slightly above that of
the margins, and that the thread of maximum velocity
tends to be depressed below the surface. The result is a
double helical flow cell, in which bottom water spreads,
rises along the margins, converges over the center, and
sinks there. Flume and theoretical work (Kennedy and
Fulton, 1961; Gessner, 1973) indicates that in flumes of
580
176
SUBSTRATE RESPONSE TO HYDRAULIC PROCESS
(A) AS SAND WAVE
(PROFILE)
PLAN)
PLAN)
FIGURE 17. Interpretation oj asymmetrical sand ribbons; B is
more probable.
square cross section the unequal distribution of turbulent
(Reynolds) stresses will result in secondary flow from
the center toward the corners. The resulting multiple
flow cells do not form the double helical pattern postu-
lated for natural channels, however, and their applica-
bility to the natural situation is uncertain. Bagnold
(1966, pp. 112-115) offers an independent explanation.
He suggests that there is asymmetrical exchange of mo-
mentum between the bottom boundary layer of a river
and the overlying flow in that tongues of boundary water
abruptly penetrate the overlying flow, to be compensated
by a general sinking of the latter (see Chapter 7, p. 98).
This results in elevation of the water surface over the
channel axis where this exchange is most intense. The
ensuing pressure head, he suggests, drives the secondary
component of flow.
The preceding discussion has dwelt on double helical
flow cells as mechanism for generating a large-scale sand
ridge topography. An attempt has been made to match
(a)
(c)
(b)
FIGURE 18. Patterns of sand ridges on tide-dominated shelves. From Off (1965).
(d)
581
BED FORMS
177
theoretical and experimental studies with characteristics
of shelf ridge fields. However, such large-scale coupling
of flow with substrate has not yet been observed in the
field. It is worth noting that there is an independent
mechanism that is theoretically capable of maintaining
a ridge topography, either by itself or in conjunction
with other mechanisms. The mechanism described by
Smith (1969) requires that ridges be aligned with mean
flow direction and that the variance in flow direction be
high, either because the flow is a rotary tidal flow; or
because it is storm-driven, and the direction of flow
varies during a storm and also among storms (see Chap-
ter 4). As a consequence, most flows intense enough to
entrain sand will be aligned at an oblique angle with
the ridges during most of their duration. Flow across
the ridge can be treated two-dimensionally according to
slender body theory (Smith, 1969) and the stability
analysis of Smith (1970) applies (see the preceding sec-
tion). First one flank then the other flank of the ridge
will be eroded, with sand transferred to the crest and
far flank each time.
SAND RIDGES IN RESPONSE TO TIDAL FLOWS. The
reversing nature of nearshorc tidal flows adds another
mechanism capable of maintaining a ridged topography.
The velocity of the tidal wave is a function of water
depth, and flow over a step or across a sill in a cohesion-
less substrate will result in a phase discontinuity between
the behavior of the tidal wave on either side of the sill
(Fig. 19). Thus, when the tide is in the last stages of ebb
on one side, it may be already beginning to flood on the
other, so that there is an opposing sense of flow over the
crest of the sill. If the flow is broad relative to its depth,
and if the sill is a relatively large-scale feature, then this
is an inherently unstable situation. Slight irregularities
in the seafloor on either side of the sill will result in in-
equalities in the rate of propagation of the tidal wave,
and during the brief period of opposing flow the two
water masses on either side of the sill will tend to inter-
penetrate along a zigzag front. The tongues of flow on
either side of the sill crest will tend to scour its channels
until the crestline of the sill has also become zigzag
(Fig. 19). A channel that is on the side of the sill facing
the oncoming tidal wave and opens in that direction is
called a. flood sinus (Ludwick, 1973). It experiences an
excess of flood over ebb discharge (is flood-dominated).
A channel on the other side of the sill is called an ebb
sinus, and is ebb dominated.
Scour in the interdigitating channels of such an ebb-
flood channel complex is matched by aggradation of the
interchannel shoals. This transfer is perhaps aided by
the secondary circulation mechanisms described in the
preceding section. As a consequence of the residual cur-
rent pattern, net vectors of bottom flow integrated over
the tidal cycle meet obliquely head-on over crests (Fig.
20), with the result that each ridge becomes a sand cir-
culation cell, or closed loop in the sand transport pattern.
Mean dro dx is negative along these vectors toward the
m
A
m
®
\
V/y
'J'
'///A
^
®
PRE-EXISTING HIGH
TIDAL CURRENT
NEAR SLACK WATER
SAND RIDGE
FIGURE 19. Hypothetical scheme of development of an ebb-flood channel
topography as a consequence of the phase lag, experienced by the tidal wave in
its passage across a submarine sill.
582
178
SUBSTRATE RESPONSE TO HYDRAULIC PROCESS
■J VM v'/>
ilAA Of •
1
"^\MEAN BOTTOM FLOW ~~^x BOUNDARIES FOR MEAN FLOW SHEARS
mm SHOAL
FIGURE 20. Nearshore and offshore patterns of tidal flow about sand ridges.
Based on Luduick (1970b) and Caston and Stride (1970).
zone of residual current shear at the ridge crest. The
ridges therefore tend to aggrade toward the intertidal
zone where they become "drying shoals" or swash plat-
forms dominated by wave processes (Oertel, 1972). On
open coasts, however, wave surge erosion may balance
tidal current construction when the crests are still sub-
tidal.
Tidal sand ridges that partition ebb- and flood-domi-
nated flows usually experience a stronger residual cur-
rent on one side than on the other, and tend to migrate
away from that side (Fig. 21). In such cases, where the
cross-ridge component of flow is strong, a ridge may itself
deform into a sigmoid pattern, and eventually into two
or three separate ridges (Caston, 1972); see Fig. 21.
Sills with interdigitating ebb and flood channel sys-
tems occur at the mouths of most tidal estuaries (Fig. 22),
as a consequence of frictional retardation of the tidal
wave within the estuary, and the resultant phase lag.
On the Bahama Banks, they occur on the inner sides of
islands, where the two wings of the tidal wave meet as
they refract around the island (Fig. 23). The evolution
of such a system portrayed in Fig. 19 probably rarely
occurs in nature; the channel systems form simultane-
ously with such sills, not afterward. For instance, the
ebb-flood channel system of the Chesapeake Bay mouth
shoal appears to have formed during the Late Holocene
reduction in the rate of sea-level rise (Ludwick, 1973).
It can be inferred from the present morphology that the
sill prograded south across the bay mouth, fed by the
littoral drift discharge of the Delmarva coastal com-
partment. The ridges would have developed in zig-
zag fashion, alternately and progressively segregating the
flow into ebb-dominated and flood-dominated channels
(Fig. 24).
Tidal flows often occur in the presence of salinity
stratification so intense as to persist for part or all of the
tidal cycle despite the powerful mixing effect of flow
turbulence. Flow structure may be yet more complicated
as a result. In Fig. 25, the residual circulation over the
Hudson estuary mouth shoal (New York Harbor en-
trance) is seen to be a resultant response to flow inter-
digitation due to the phase lag effect (Fig. 25C) and to
estuarine (two-layer) circulation (Fig. 255).
£Z> SAND WAVE
0 GRAIN ORIENTATION
— BANK CRESTLINE
ff DOMINANT CURRENT; SAND STREAM
// MAJOR, MINOR BANK MOVEMENTS
FIGURE 21. Above: Anatomy of a tidal sand ridge. From
Houbolt (1968). Below: Evolution oj a tidal sand ridge. From
Caston (1972).
583
FIGURE 2 2. A hydraulic and geomorphic interpretation of the
net nontidal {residual) flow pattern at the bottom in the entrance
to Chesapeake Bay. Numbers are measured flood and ebb durations
at the bottom in hours; small arrows show measured direction of
near-bottom currents. Stippled areas are shoaler than 18 ft.
Ruled areas show where there is an ebb or a flood flow predomi-
nance. From Ludwick (1970a).
584
179
180 SUBSTRATE RESPONSE TO HYDRAULIC PROCESS
FIGURE 23.
Charles True.
Ebb-flood channel pattern on the Great Bahama Bank. Altitude 3000 ft. Photo:
Stratification may also play a role in the formation of
ridge topography within the estuary. Weil et al. (in
press) describe the formation of subtidal levees in Dela-
ware Bay as the consequence of the penetration of sub-
surface saline tongues up the channels during flood tide,
resulting in an internal pressure head that can drive
channel axis downwelling (Fig. 26), and as a conse-
quence of the overriding of the tongues by fresher water
during the ebb tide, with similar effects. One of us
(Ludwick, in press) has mapped near-bottom con-
vergences and divergences of flow in the Chesapeake
Bay mouth during flood tide. These are absent during
the more thoroughly stratified ebb. Here stratification
appears to inhibit channel axis downwelling and
bottom current divergence (see Fig. 31). Velocity
profiles of Chesapeake Bay mouth tidal flows tend to be
parabolic but with markedly sigmoidal perturbations
(Ludwick, 1973), and may imply the presence of
standing internal waves or wakes from shoals.
Tidal flows in confined estuary mouths thus tend to
develop an interdigitating pattern of ebb- and flood-
dominated channels, whose sequence of partitioning
ridges tends to alternate between clockwise and counter-
clockwise current flows (Fig. 20). On the offshore shelf,
however, the tide becomes rotary rather than reversing
and a different pattern tends to appear (Caston and
Stride, 1970). Ridges appear in free-standing sets rather
than in continuous zigzag arrays. Residual current shears
occur in channel axes as well as on ridge crests, and
successive ridges experience residual flows with the same
sense of rotation.
Huthnance (1972) attributes this open shelf flow pat-
tern to interaction of the ridges with the shelf tide. His
model considers a rectilinear reversing tide whose flow
directions make an oblique angle with the ridge axis.
The cross-ridge component of flow must accelerate over
the ridge crest for continuity reasons. The ridge-parallel
component of flow must decrease up the upcurrent flank
as the water column shoals, and influence of friction be-
comes proportionately greater. However, because high-
velocity fluid is being transported into the shoal region,
the decrease in the ridge-parallel flow component lags
behind the decrease in depth. On the downcurrent
flank, the restoration of the ridge-parallel flow to am-
bient velocity is similarly lagged. When the tide changes,
upcurrent and downcurrent flanks reverse roles. When
flow is averaged over the tidal cycle, a clockwise pattern
of residual flow around the ridge results (or counter-
585
BED FORMS
181
FIGURE 24. Evolution oj "submarine zigzag spit" across Chesapeake Bay mouth. Based
on Luduick (1972).
clockwise, depending on whether the oblique, reversing
tidal stream is sinistral or dextral with respect to the
ridge). Huthnance proposes a second mechanism
whereby in the northern hemisphere, Coriolis force also
results in clockwise circulation.
Huthnance's mechanisms are interesting, but the re-
quirement that there be a significant angle between the
axis of the tidal stream and that of the ridge presents a
problem. The ridges are a response element within the
flow field-substrate system, not an independent forcing
element. It seems doubtful that ridges of cohesionless sand
could maintain a significant angle with the tidal stream
for any length of time, unless it were somehow an equi-
librium response to flow. Smith (1969) notes that tidal
sand ridges might be expected to orient themselves
parallel to the long axis of the tidal ellipse, as the sand
body would then be at a small angle of attack through-
out most of the high-velocity part of the tidal cycle.
According to slender body theory the cross-shoal com-
ponent of flow during this period can be considered to
be two-dimensional and driven by the cross-shoal pres-
sure gradient. It would thus sweep sand first up one
side and then up the other as the tide rotated.
Possibly the dilemma is resolved by the lag effect
cited by Postma (1967) and Stride (1974); see Fig. 27.
Because of a lag in the entrainment of sand, the period
of maximum sand transport is believed to lag behind
maximum flood flow, and again behind maximum ebb
flow. The result should be to align the response element
(sand ridge) obliquely across the major axis of the tidal
ellipse. It also seems likely that the large-scale, un-
bounded tidal flow field of the open shelf might at least
locally generate Ekman flow structure during midtide,
and couple with inner shelf ridge fields in the fashion
that has been suggested for wind-driven flows.
Limiting Conditions of Bed Form Formation on the
Continental Margin
In attempting to apply the elements of bed form theory
presented on the preceding pages to analyses of conti-
586
182
SUBSTRATE RESPONSE TO HYDRAULIC PROCESS
SANDY HOOK
0 2
0 t . i '
ROCKAWAY
8
e
o
©
FIGURE 2 5. (A) Profile across the Hudson estuary mouth (mouth oj New York
Harbor) contoured for velocity residual to the tidal cycle. The flow pattern is a
resultant response to component flow patterns shown in (B) and (C). (B)'Schematic
diagram oj two-layered, estuarine flow pattern. (C) Schematic diagram oj com-
ponent oj flow pattern resulting jrom phase lag oj tidal wave. From Duedall et al.
(in press), after Kao (1975).
nental margin sedimentation patterns, it is useful to keep
in mind some generalizations presented by Allen (1966;
1968a, pp. 50-53; 1968b). Allen, following Bagnold
(1956), notes that the grain-fluid system is a decidedly
multivariate one, and that we should expect to find co-
existing instabilities of several different modes and scales.
Flows that experience both transverse and streamwise
perturbations may develop bed form associations consisting
of two different bed form types, for instance a reticulate
pattern with sand waves overprinted on sand ridges.
Likewise, flows tend to experience one or more instability
modes at several different spatial scales, resulting in a
bed form hierarchy as, for instance, in the case of the
Diamond Shoals sand wave field (Hunt et al., in press),
where photos show that current ripples are superimposed
on sand waves (Fig. 10) and sidescan sonar records show
in turn that sand waves are superimposed on giant sand
waves (Fig. 28). Elaborate hierarchical associations of
bed forms occur over vast areas of the earth's surface,
in subaerial sand seas (Wilson, 1972), and also in widely
disparate environments on the continental margin (com-
pare Fig. 29 with Fig. 30).
The physical scale at which bed forms occur affects
their response characteristics, and in turn the flow fre-
quency to which they are tuned. For instance, on the
crests of the drying sand ridges of the Minas Basin,
current ripples reflect radial drainage at the last stages
of ebb, sand waves are oriented with slip faces seaward
as responses to peak ebb flow, while larger dunes locally
are landward facing, reflecting a stronger flood than
ebb flow (Swift and McMullen, 1968; Klein, 1970;
Dalrymple, 1973).
The largest scale transverse and longitudinal bed
forms have had to readjust to continuous environmental
change associated with Holocene deglaciation and the
accompanying transgression of the continental margins.
In some cases, they appear to have taken nearly the
duration of the Holocene to form. Sand ridges on the
central New Jersey shelf have basal strata containing
11,000-year-old shells (Stubblefield et al., 1975). These
features and many other shelf ridge fields appear to have
been formed by shoreface ridge formation and detach-
ment (Swift, in press) during the Holocene transgression;
see Fig. 28, Chapter 14. Plan geometry and internal
structure of Atlantic Shelf ridge fields suggest that
ridge spacing has increased by ridge migration or
coalescence as the water column deepened (Swift et al.,
1974).
587
BED FORMS
183
A - FLOOD
ISOVELS, cm/ttc
DENSITY ISOPLETHS
B-EBB
C -TRANSPORT DOMINANCE
FIGURE 26. Tidal sand ridge as a submarine levee, formed in
response to stratified flow. From Weil et al. (in press).
FIGURE 27. Lag effects in a rotating tide. Radial arrows are
vectors of tidal current velocity at intervals through the tidal
cycle. Sand entrainment starts at velocity Vi and continues to
velocity V2. Net sand transport is to right and onshore. From
Postma (1967).
The response of larger bed forms tends to lag beyond
the peak flow event or may comprise an average re-
sponse to repeated events. Allen (1973) notes that maxi-
mum sand wave height in the Fraser River and Gironde
estuary is lagged behind peak tidal flow by as much as
a quarter of the tidal period. Ludwick (1972) notes that
tidal sand waves are symmetrical over portions of
the Chesapeake Bay mouth where the tidal cycle
is symmetrical, but are asymmetrical when there is
flood or ebb asymmetry in the tidal cycle. Thus their
response to reversing tidal flow is time-averaged in a
manner entirely analogous to the response of oscillation
ripples to wave surge (Chapter 8). Tidal sand waves in
Chesapeake Bay mouth attain their greatest height and
slopes during the summer months when wave activity is
FIGURE 28. Sidescan sonar record of sand waves on the
back of giant sand waves, Cape Hatteras, North Carolina.
Sand waves are larger in coarse sand of trough than on finer
sand of giant waves. Giant sand waves are 120 m apart,
7 m high. Unpublished data of Swift and Hunt.
588
184
SUBSTRATE RESPONSE TO HYDRAULIC PROCESS
28"20'N
28" 15' N -
74°25'W
74°20'W
FIGURE 29. Erosional jurrows and large-scale silt ridges on the Blake-Bahamas outer ridge, 4700 m depth. From Hollister et al. (1974)-
at a minimum; they are degraded by the more intense
wave activity of winter months (Ludwick, 1970a,b,
1972). The Piatt Shoals sand wave field on the open
Virginia shelf appears to be induced by storm flows,
hence it may have the opposite behavior pattern; sand
waves would be highest in the winter months and would
tend to be degraded by fair-weather wave surge and
burrowing organisms during the summer (Swift et al.,
1974).
ESTIMATES ON SEDIMENT TRANSPORT
A Numerical Model
port systems, and to determine the rates of erosion, trans-
port, and sedimentation associated with these elements.
Much of the material in the following chapters is devoted
to available information of sediment sources, pathways,
and sinks on the continental margin. However, there
have been very few attempts to estimate rates of sedi-
ment transport. It should be possible to measure the
time history of a marine flow by means of a current-
meter array, then employ the empirical relationships
developed by hydraulic engineers to estimate the time
history of sediment transport. The difficulties however,
are formidable. In situ recording current meters are ex-
pensive and difficult to maintain. Data processing is com-
589
1 I :.
70 15
70 00
69 45
FIGURE 30. Pattern of sand waves (dark lines) and sand ridges appear to have initially formed as sboref ace-connected ridges
at Nantucket Shoals. Significant highs are stippled and ebb-flood similar to those attached to the shoreface of modern Nantucket
channel couplets are indicated by arrows. Ebb and flood sinuses as Island, and to have been stranded on the shelf floor as the shoreface
inferred from morphology, are indicated by arrows. The ridges underwent erosional retreat. From Swift (2975).
185
590
186
Sl'BSTRATE RESPONSE TO HYDRAULIC PROCESS
wave surge. There is no general agreement on the most
satisfactory transport equation, or on the applicability
of equations developed under laboratory conditions to
the complex deep-water flow fields of the continental
margin.
A simple numerical model for estimating rates and
patterns of sediment transport in areas of tidal flow has
been devised by one of us (Ludwick, in press). It is
summarized below.
structure of the model. The model requires deter-
mination of the distribution across the study area of a
sediment transport index ro«ioo over a tidal cycle. The
index is derived from Bagnold's (1956) work (see p.
1 13), in which sediment discharge q is set proportional
to fluid power to defined as
so that
0> = Toll
q = KtqU
where t» is bottom shear stress and u is the depth-aver-
aged flow velocity. For convenience of measurement,
Ludwick substitutes Mioo, the velocity measured 100 cm
off the bottom. With this information, it is possible to
use the sediment continuity equation (p. 166) to deter-
mine the distribution and relative rates of aggradation
and erosion along streamlines of sediment transport.
determination of Tn. In order to determine the
distribution of to, current velocities were measured over
27 hour intervals at 24 stations in the mouth of Chesa-
peake Bay. A Kelvin Hughes direct reading current
meter was employed from an anchored ship. At each
station the current meter was used successively through
1 1 different depth levels. Hourly profiles with 4 minute
observation periods were obtained at each level.
These speed values were then reduced to pseudo-
synoptic data sets for standard times and depths at each
station (see Ludwick, in press). Each data set was fitted
to Hama's (1954) parabolic velocity defect law (see the
discussion in Chapter 7, p. 96). This empirical func-
tion pertains to outer boundary flow, at distances greater
than 0. 1 5/?, where h is the thickness of a turbulent bound-
ary layer, or water depth in the case of fully developed
flow in a uniform channel. The equation is
^ -"(•-!)'
where ux is the free stream velocity, u is velocity at dis-
tance z above the bed, and u* is the friction or shear
velocity.
An estimate of u* on the bottom is then obtained by
least squares curve fitting. The value can be converted
to an estimate of ro, the boundary shear stress, through
the relationship u* = (to p)1'2, where p is fluid density.
This measurement, obtained by observation of the entire
water column, provides a far more reliable estimate for
ro«ioo, the fluid power, than does ro determined simply
from the product Cioop(«ioo)2, due to uncertainties in
determining Cioo (see Chapter 7, p. 99).
maps of bed sediment transport. Values of the
sediment transport index obtained for 24 stations must
be converted to maps of near-bottom streamlines of sedi-
ment transport. The values are adjusted to the mean
tidal range, a process described by Ludwick (1973).
They are further corrected by subtracting 150 dyne-cm,'
sec cm'-', a threshold value for the initiation of sediment
movement (Fig. 31). The value at each station is inte-
grated separately over each flood and each ebb half-
cycle, and the results are averaged for ebb and flood.
After averaging and integrating, the units of the sedi-
ment transport index are dyne -cm/cm2 per average ebb
(or flood) half-cycle.
The values obtained at points on the field grid of 24
irregularly placed stations must then be redistributed
over a systematic grid. This is a problem in vector inter-
polation. The first step is to prepare separate maps of
the north-south and east-west components of ro«ioo for
the flood cycle. Each map is contoured. The flood com-
ponent maps are superimposed. Resultant vectors may
now be calculated at any point, if the contour interval
is sufficiently small. The density of resultant vectors may
be increased in areas of complex flow. Finally, stream-
line maps may be prepared by drawing lines that are
everywhere tangent to the vectors (Fig. 32-4). The proc-
ess is repeated for the ebb half-cycle and the vector sum
of ebb and flood (Figs. 33.4 and 34.4).
FIGURE 31. Tidal current speed and bottom shear stress at a
flood channel station, Chesapeake Bay mouth. Speed values are for
a distance of 18.5 ft off the bottom. Total depth, 56 ft. Observed
speeds were corrected from mean tidal range and averaged over
six cycles, zo is the roughness length estimated from vertical velocity
profiles, k. is the height of bottom roughness elements, and rc is
the critical shear stress, calculated from the Shields entrainment
diagram. From Luduick (1970a).
591
76*1, W
FIGURE 3 2. Ebb-directed sediment transport at the
bed. (a) Streamlines of the sediment transport vector
toUioo; depths are in meters; vertically ruled areas are
shoaler than 5.5 m. (b) Erosion-deposition chart on
which erosion is positive ( + ) and deposition is negative
75°|S5
(— ),• units are dyne-cm /cm2 per ebb half-tidal cycle
per 463 m of transport X 10'*; cross-hatched areas
indicate erosion intensity greater than —400 units;
stippled areas indicate deposition intensity greater than
+ 400 units. From Ludwick {in press).
187
592
FIGURE 33. Flood-directed sediment transport at
the bed. (a) Streamlines of the sediment transport
vector T0Uioo." depths are in meters; vertically ruled
areas are shoaler than 5.5 m. (b) Erosion-deposition
chart on which erosion is positive ( + ) and deposition
is negative ( — ); units are dyne-cm I cm1 per flood halj-
tidal cycle per 46J m of transport X 10'*; cross-hatched
areas indicate erosion intensity greater than — 4OO
units; stippled areas indicate deposition intensity
greater than +400 units. From Ludwick (in press).
593
FIGURE 34. Vector sum of ebb and flood sediment
transport at the bed. (a) Streamlines of the resultant
sediment transport vector roUioo," depths are in meters;
vertically ruled areas are shoaler than 5.5 m. (b)
Erosion-deposition chart on which erosion is positive (+)
and deposition is negative ( — ); units are dyne-cm I cm1
per tidal cycle per 463 m oj resultant transport X 10*;
cross-hatched areas indicate erosion intensity greater
than — 400 units; stippled areas indicate deposition inten-
sity greater than + 400 units. From Ludwick (in press) .
594
189
190
SUBSTRATE RESPONSE TO HYDRAULIC PROCESS
It is important to realize the limitations of the stream-
lines of bottom sediment transport so determined. The
redistribution of information has not in any way in-
creased the accuracy or resolution of the original data.
Sediment input in a stream tube does not necessarily
equal sediment output, since deposition or erosion may
occur. There is no underlying stream function in the
method, and the spacing of the streamlines is not a
measure of transport rates. It is assumed, however, that
transport is confined to a path of unit width that con-
forms to the bathymetry of the seafloor, and that the
streamline is the center line of this path. It is further
assumed that conditions are steady and nonuniform for
the entire pattern.
net sedimentation maps. As a separate and ensuing
procedure, it is possible to estimate the extent of areas
of erosion and deposition, and also the rates at which
these processes occur. The estimate utilizes the sediment
continuity equation written in terms of discharge:
dr) dq
dt dx
where r\ is bed elevation relative to a datum plane, / is
time, 6 is a dimensional constant related to sediment
porosity, q is the weight rate of bed sediment transport
per unit width of streamline path, and x is distance along
the streamline.
Discharge (q) may be taken as proportional to to"ioo
and the right-hand partial derivative may be approxi-
mated by a finite difference:
dq
dx
Aq (t0Uu
— - — K -
Ax
(TqUioo)]
*•> — Xi
The term x-> — *i is held constant arbitrarily at a value
of 465 m, hence
dt
oc — AroUioo
Thus a decrease in transport rate along a transport path-
way induces deposition; an increase causes erosion.
The resultant vector map for a half-cycle is super-
imposed on the equivalent streamline map. The mag-
nitude of ro«ioo is determined at equispaced points along
each streamline; AtoMioo is determined as a positive or
negative value, and mapped over the area of study as an
estimate of relative erosion and deposition intensity. In
Figs. 325, 33B, and 345, net sedimentation maps have
been prepared for the ebb and flood half-cycles and for
the vector sum of ebb and flood.
utility of the model. Such a manipulation of the
data from 24 current-meter stations extracts a surprising
amount of information from them. Streamlines of bed
sediment transport associated with the ebb tidal jet are
seen to pass over the bay mouth shoal in parallel fashion.
Flood streamlines, however, form a pattern in which
flow divergence and flow convergence alternate across
the flow in sympathy with the topographic pattern of
interdigitating ebb and flood channels.
The vector sum map shows a complex pattern of flow
dominance that is also correlated with bottom morphol-
ogy. Patterns of net sedimentation do not correlate as
closely with the topography, probably because they do
not indicate the areas of maximum relief, but instead
areas undergoing maximum change. In particular, the
parabolic shoals that envelop each ebb or flood sinus are
seen to be subject to a systematic pattern of sedimenta-
tion. The sides of shoal segments that face the dominant
flow, however obliquely, are eroding. The crest and
downcurrent sides, however, are undergoing aggrada-
tion. Thus, the processes that Smith (1970) has inferred
to cause sand waves (see p. 165) appear also to be
applicable to ebb-flood channel topography.
The model can be generalized for portions of the shelf
dominated by storm flows if each flow event is treated
in the same fashion as Ludwick treated a tidal half-cycle,
or sediment transport can be integrated over an arbi-
trary period of observation.
Transport Estimates from Tracer Dispersal Studies
One of the main stumbling blocks in divising quantita-
tive estimates of sediment transport has been the limited
applicability of empirical relationships based on labora-
tory observations to the complex flows of the marine
environment. The model partially circumvents this prob-
lem by recourse to the sediment transport index, based
on Bagnold's generalized evaluation of fluid power
(Chapter 8, p. 113). In doing so, it provides only a
relative answer. Sediment transport is proportional to
fluid power, and the proportionality constant remains
unevaluated. Despite this sacrifice, the model has not
resolved the problem of adequately treating the complex
time and space scales of marine flows. In particular, it
fails to deal with the vexing problem of the role of bot-
tom wave surge in "lubricating" bottom sediment trans-
port by reducing the effective transport threshold for a
unidirectional flow component (see discussion, Chapter 8,
p. 115). This wave surge factor becomes part of the pro-
portionality constant. Wave-surge-amplified transport is
not that critical a problem in the analysis of a primarily
tide-built topography. It becomes critical, however, in
open shelf transport, where wind-driven unidirectional
flows attain their maximum intensities just as the wave
regime does.
595
ESTIMATES ON SEDIMENT TRANSPORT
191
It is clear that the best resolution of a marine sediment
transport system will be obtained when a model such as
the one presented above is employed together with an
independent method for evaluating the proportionality
constant. The most promising method to date is the de-
ployment of radioisotope tracers. Fluorescent tracers
have been widely used (see Ingle and Gorsline, 1973;
Inman and Chamberlain, 1959). However, since count-
ing of labeled particles must be done in the laboratory,
the analysis is tedious, and it is generally not possible to
watch the development of dispersal patterns in real time.
Furthermore, fluorescent tracers have a very limited ap-
plicability seaward of the surf, as a consequence of the
limited sensitivity of the method and the difficulties of
hand sampling. Tracer dispersal can usually be observed
in an area 50 m in diameter or less, under fair-weather
conditions. After a storm, when a major displacement
of sediment has occurred, the tracer grains are liable
not to be there at all.
Radioisotope tracers avoid much of this difhculty.
The RIST (Radio Isotope Sand Tracer) system, devel-
oped by Oak Ridge National Laboratories and the
Coastal Engineering Research Center (Duane, 1970) de-
tects radioisotope-labeled tracers by means of a towed
scintillometer. The data logging system provides for real
time readout, which greatly aids mapping of the dis-
persal pattern. A relatively long-lived isotope such as
ruthenium- 103, with a half-life of 40 days, permits
effective tracing for three times that duration, or an
entire storm season.
A numerical estimate of sediment transport may be
fine-tuned by quantitative analysis of radioisotope tracer
dispersal patterns. The procedure requires not only the
mapping of successive outlines of the tracer pattern but
the ability to account for all of the labeled particles at
each stage. In order to establish such a mass balance,
it is necessary to know the depth of reworking, which is
the depth to which labeled particles have penetrated
during dispersal. This depth can be calculated from the
known ability of the sediment to absorb radiation, if it
is assumed that the tagged particles are mixed into the
reworked layer in a homogeneous fashion. If tracer
particles can be accounted for through successive map-
pings of the dispersal pattern, then the rate of sediment
transport as indicated by dispersal of tracers may be
checked against the rate of transport as estimated from
current-meter records in one of two ways. Transport
rates may be determined directly from the dispersal
pattern and compared with estimates based on current-
meter records. Or current-meter records may be used
to simulate tracer dispersal patterns, and these ideal
patterns may be compared with observed dispersal pat-
terns.
Figure 35 shows a series of radioisotope dispersal pat-
terns from an experiment conducted by J. W. Lavelle
and his associates, Atlantic Oceanographic and Meteor-
ological Laboratories, Miami, Florida. Water-soluble
bags of labeled sand were released along a line in 20 m
of water on the south shore of Long Island during April
and May of 1974. Over a period of 69 days, a typical
fair-weather dispersal pattern formed (panels B-F). The
data in these panels have been corrected for decay, but
in the last panel, the corrected values on the margins
of the pattern are so much weaker than background,
that they were lost when smoothed background values
were subtracted. The mild summer storms during this
period only briefly generated flows strong enough to
transport sand, and much of the labeled sand remained
in close proximity to the drop line, where it was not
readily resolved by the towed scintillometer. It will be
necessary to apply a statistical smoothing function to
the data, in order to undertake a mass balance calcu-
lation for the dispersed tracer sand. If continued experi-
mentation leads to improved field techniques and data
processing, then radioisotope tracers should prove a
fruitful method for calibrating numerical models of sedi-
ment transport.
SUMMARY
The size frequency distribution of marine sand samples
tends to be log-normally distributed. This distribution,
as defined by its mean and standard deviation, is the
"signature" of the depositional event, and deviations
from log normality, as measured in terms of skewness
and kurtosis, may be taken to reflect both the provenance
and the hydraulic history of the sediment.
The modal diameter of a sand deposit is that grain
size most likely to arrive and least likely to be carried
away from the place of deposition; progressively coarser
sizes are progressively less frequent because they are
progressively less likely to arrive, and progressively finer
grain sizes are progressively less frequent because they
are more likely to be carried away. This intuitively ap-
parent concept can be explained in terms of probability
theory.
As sand progresses down a transport path by inter-
mittent hops, it tends to leave its coarser grains behind,
and the deposits are progressively finer in the direction
of transport (have undergone progressive sorting). They
also will tend to be fine-skewed, particularly if the in-
tensity of hydraulic activity also declines down the
transport path.
Moss has shown that the size distributions of marine
sands tend to be made up of three log-normal popula-
596
192
SUBSTRATE RESPONSE TO HYDRAULIC PROCESS
* GRAB SAMPLES
40°29'50"
40°29'45"
73°42'00'
73°41 30
B
DAY 0
-
PRELIMINARY
i
D
f/A,
DAY 22
-
4
PRELIMINARY
DAY 48
% TRANSPORT
W
20
% TRANSPORT
0 5
J">VJV>W.
° n
DAY 3
RIST DROP
i — i — i — r
10
DAY 69
% TRANSPORT
0 10
% TRANSPORT
0 20
THRESHOLD VELOCITY =20 OCM/SEC
~r~~i — i — "i — r — i — i — r^ — n — r — i — i — i — i — i — i — ^^n — i — i- "i — i — i — i — ' — i — i —
20 J
DAY 22
i i
30
% TRANSPORT
0 20
40
"i — r
50
~i 1 r
60
DAY 48
o
30 £
O
0 O
FIGURE 3 5. Time sequence oj dispersal oj radioisotope-
labeled sand, south shore of Long Island, April 22-July 2,
1974- (A) Background radioactivity in arbitrary units.
Heavy black line is line oj emplacement oj sand labeled
with ruthenium-105. (B-F) Maximum extent oj detectable
signal ajter removal oj background on successive mapping
days. Data have been corrected jor decay. Bottom panel:
time-velocity record {jagged line). Height oj vertical bars
indicates percentage oj total bed load transported, as
determined from a normalized sediment transport index,
Vioo — Vr- Width oj vertical bars indicates duration oj
transport event. Rose diagrams indicate direction oj
transport. Length oj radial bar indicates percentage oj
transport during that event; width is proportional to direc-
tion and intensity. Velocity was sampled every 10 minutes.
Unpublished data oj Lavelle et al., Atlantic Oceanographic
and Meteorologic Laboratories, Miami, Florida.
tions as a consequence of the fashion in which the bed
is built; the main subpopulation (A population) com-
prises the framework of the deposit. A fine B population
is interstitial; a coarse C population is the consequence
of "traction clogs." The A:B:C ratio varies with the
flow regime.
Bed forms are irregularities in the particulate sub-
strate of a fluid flow. Sheared flow is innately unstable,
and tends to develop repeated patterns of velocity vari-
ation, either parallel or normal to the flow direction.
Such instabilities tend to interact with the bed so as to
cause rhythmic variations in elevation. Flow and bed
597
REFERENCES
193
perturbation amplify each o'ther until equilibrium is
attained.
Bed forms occur in associations (more than one genetic
type present), in hierarchies (successive scales of bed forms
of similar genesis), and in hierarchical associations. Trans-
verse bed forms interact with wavelike perturbations of
flow transverse to the flow direction. Current ripples are
small-scale transverse bed forms that appear to result
from boundary layer instability; their wavelength is in-
dependent of depth. Sand waves result from perturbations
of the whole flow field, or a density-homogeneous por-
tion of it. Several scales of sand waves may occur to-
gether; the smaller scale may perhaps be a response to
primarily tractive transport, whereas the larger scale
may be a response to primarily suspensive transport.
Longitudinal bed forms are caused by velocity perturba-
tions parallel to flow. In some cases, the perturbation
takes the form of horizontal, flow-parallel vortices whose
sense of rotation is alternately right and left handed,
and this may be true for all cases. Parting lineations are
small-scale longitudinal bed forms. They are sand ridges
a few grain diameters high and a few centimeters apart.
Current lineations have wavelengths ranging from a few
meters to hundreds of meters; their heights are negligible
relative to width.
In a characteristic pattern, sand ribbons occur on a
gravel substrate. In longitudinal furrow patterns, the lows
are more sharply defined than the intervening highs.
Sand ridges may have wavelengths of hundreds of
meters to several kilometers, and amptitudes of 10 m or
more. They are induced by tidal flows at estuary mouths,
by tidal or wind-driven flows on the shelf, and perhaps
by boundary undercurrents on the continental rise. They
appear to be time-averaged responses to intermittent
flow, and in many cases have survived successive en-
vironmental transitions associated with the Holocene
transgression.
A simple numerical model for estimating bed load
transport on the continental margin requires as input
current-meter measurements. Streamlines of bottom sedi-
ment transport may be based on the sediment transport
index. The index is derived from Bagnold's energetics,
in which sediment discharge is set proportional to fluid
power, equal to bottom shear stress times the depth-
averaged velocity. The sediment continuity equation is
used to predict areas and relative intensities of erosion
and deposition. In this equation, the discharge gradient
along a streamline is related to the time rate of change
of bottom height above a datum by a dimensional
constant.
It may be possible to evaluate Bagnold's proportion-
ality constant for sediment transport by means of mass
balance assessments of radioisotope dispersal patterns.
However, improvements in field tracer techniques and
data processing are required before such an evaluation
is possible.
ACKNOWLEDGMENTS
We thank J. R. L. Allen and R. L. Miller for critical review
of the manuscript.
SYMBOLS
Cioo drag coefficient determined from measurements
100 cm above the bottom
D grain diameter
h water depth
K a constant
ks height of bottom roughness elements
q sediment discharge
t time
u velocity
u depth-averaged velocity
u^ time-averaged free stream velocity
Kioo velocity 100 cm off the bottom
u* shear velocity; shear stress with velocity units
x distance downstream
y distance above the bed
Z distance transverse to flow
Zo roughness length
« dimensional constant related to sediment porosity
X wavelength
p fluid density
p., sediment density
rj elevation of a surface above a datum
t shear stress
ro shear stress at the bed
rc critical bed shear stress
v kinematic viscosity
a; fluid power
REFERENCES
Allen, J. R. L. (1964). Primary current lineation in the lower old
red sandstone (Devonian) Anglo-Welsh Basin. Sedimentology,
3: 89-108.
Allen, J. R. L. (196b). On bedforms and paleocurrents. Sedimen-
tology, 6: 153-190.
598
194
SUBSTRATE RESPONSE TO HYDRAULIC PROCESS
Allen, J. R. L. (1967). Depth indicators of clastic sequences. Mar.
Geol., 5: 429-446.
Allen, J. R. L. (1968a). Current Ripples. Amsterdam: North-Holland
Publ., 433 pp.
Allen, J. R. L. (1968b). The nature and origin of bedload hier-
archies. Sedimentology, 10: 161-182.
Allen, J. R. L. (1969). Erosional current markings of weakly co-
hesive mud beds. J. Sediment. Petrol., 34: 607-623.
Allen, J. R. L. (1970). Physical Processes of Sedimentation. New York:
American Elsevier, 248 pp.
Allen, J. R. L. (1973). Phase difference between bed and flow in
natural environments and their geologic relevance. Sedimen-
tology, 20: 323-329.
Bagnold, R. A. (1954). Experiments on a gravity-free dispersion
of large solid in a Newtonian fluid under shear. Proc. Roy.
Soc. London, A225: 49-63.
Bagnold, R. A. (1956). The flow of cohesionless grains in fluids.
Proc. Roy. Soc. London, A249: 235-250.
Bagnold, R. A. (1966). An approach to the sediment transport
problem from general physics. U.S. Geol. Surv. Prof. Paper
4 22- 7, 37 pp.
Belderson, K. H., N. A. Kenyon, A. H. Stride, and A. R. Stubbs
(1972). Sonographs of the Sea Floor. Amsterdam: Elsevier, 185 pp.
Brown, R. A. (1971). A secondary flow model for the planetary
boundary layer. J. Atmos. Sci., 27: 742-757.
Cartwright, D. E. (1959). On submarine sand waves and tidal lee
waves. Proc. Roy. Soc. London, A253: 218-241.
Caston, V. N. D. (1972). Linear sand banks in the southern North
Sea. Sedimentology, 18: 63-78.
Caston, V. N. D. and A. H. Stride (1970). Tidal sand movement
between some linear sand banks in the North Sea off north-
east Norfolk. Mar. Geol., 9: M38-M42.
Chandrasekhar, B. (1961). Hydi adynamic and Hydromagnelic Stability.
New York and London: Clarendon Press, 654 pp.
Dalrymple, R. W. (1973). Preliminary investigations of an inter-
tidal sand body, Cobequid Bay, Bay of Fundy. Marit. Sedi-
ments, 4: 21-28.
Duane, D. B. (1970). Synoptic observations of sand movement.
Proc. 12th Coastal Eng. Con/., Washington, DC, 799-813.
Duane, D. B., M. E. Field, E. P. Meisburger, D. J. P. Swift, and
S. J. Williams (1972). Linear shoals on the Atlantic inner
shelf, Florida to Long Island. In D. J. P. Swift, D. B. Duane,
and O. H. Pilkey, eds., Shelf Sediment Transport: Process and
Pattern. Stroudsburg, Pa.: Dowden, Hutchinson & Ross,
pp. 447-498.
Duedall, I., H. B. O'Connors, J. B. Parker, R. Wilson, W. Miloski,
and G. Hulse (in press). The seasonal and tidal variation of
nutrients and chlorophyl in the New York Bight Apex.
Marine Sciences Research Center, State University of New-
York.
Einstein, H. A., and H. Li (1958). Secondary currents in straight
channels. Trans. Am. Geophys. Union, 39: 1085-1088.
Ekman, U. W. (1905). On the influence of the earth's rotation on
ocean currents. Ark. Mat. Astron. Fysik, 2: 1-53.
Emery, K. O. and E. Uchupi (1972). Western North Atlantic
Ocean: topographs, rocks, structure, water, life, and sedi-
ments. Am. Assoc. Pet. Geol. Mem., 17: 532.
Exner, F. M. (1925). L'ber die wechselw irking zwischen wasser
und geschiebe in flussen. Set-itngsbenchte der Academic der W'issen-
schaften. Vienna Heft 3-4. pp. 165-180.
Faller, A. J. (1963). An experimental study of the instability of
the laminar Ekman boundary layer. J. Fluid Mech., 15: 560-
576.
Faller, A. J. (1971). Oceanic turbulence and the Langmuir cir
culations. Ann. Rev. Ecol. System., 2: 201-233.
Faller, A. J. and R. B. Kaylor (1966). A numerical study of the
instability of the laminar Ekman boundary layer. J. Atmos.
Sci., 23: 466-480.
Friedman, G. M. (1961). Distinction between dune, beach, and
river sands from their textural characteristics. J. Sediment.
Petrol., SI: 514-529.
Furnes, G. K. (1974). Formation of sand waves on unconsolidated
sediments. Mar. Geol., 16: 145-160.
Gessner, F. B. (1973). The origin of turbulent secondary flow
along a corner. J. Fluid Mech., 58: 1-25.
Gibson, A. H. (1909). On the depression of the filament of maxi-
mum velocity in a stream flowing through an open channel.
Proc. Roy. Soc. London, A82: 149-159.
Guy, H. P., D. B. Simons, and E. V. Richardson (1966). Summary
of alluvial channel data from flume experiments 1956-1961.
U.S. Geol. Sun: Prof. Pap. 462- J , 96 pp.
Hama, F. R. (1954). Boundary characteristics for smooth and
rough surfaces. Trans. Soc. Naval Architects and Marine Engi-
neers, 62: 333-358.
Hanna, S. R. (1969). The formation of longitudinal sand dunes
by large helical eddies in the atmosphere. J. Appl. Meteorol.,
88: 874-883.
Hollister, C. D., R. D. Flood,. D. A. Johnson, P. Lonsdale, and
J. B. Southard (1974). Abyssal furrows and hyperbolic echo
traces on the Bahama outer ridge. Geology, 2: 395-400.
Houbolt, J. J. H. C. (1968). Recent sediments in the southern
bight of the North Sea. Geol. Mijnbouw, 47: 245-273.
Huthnance, J. M. (1972). Tidal current asymmetries over the
Norfolk sandbanks. Estuarine Coastal Mar. Sci., 1: 89-99.
Ingle, J. C. and D. S. Gorsline (1973). Use of fluorescent tracers
in the nearshore environment. In Tracer Techniques in Sediment
Transport. Vienna: Int. At. Energy Agency, pp. 125-148.
Inman, D. L. (1949). Sorting of sediments in the light of fluid
dynamics. ./. Sediment. Petrol., 19: 51-70.
Inman, D. L. and T. K. Chamberlain (1959). Tracing beach
sand movement with irradiated quartz. J. Geophys. Res., 64:
41-47.
Jeffreys, H. (1929). On the transverse circulation in streams.
Proc. Cambridge Phil. Soc, 25: 20-25.
Kao, A. (1975). A study of the current structure in the Sandy
Hook-Rockaway Point transect. M.S. Research Paper, Ma-
rine Sciences Research Center, State University of New York,
Stony Brook.
Kennedv.J. F. (1964). The formation ol sediment ripples in closed
rectangular conduits and in the desert. J. Geophys. Res., 69:
1517-1524.
Kennedy, J. F. (1969). The formation of sediment ripples, dunes
and antidunes. Ann. Rev. Fluid Mech., 1: 147-168.
Kennedv, R. J. and J. F. Fulton (1961). The effect of secondary
currents upon the capacity of a straight open channel. Trans.
EIC, 5: 12-18.
Kenyon, N. H. (1970). Sand ribbons of European tidal seas. Mar.
Geol., 9: 25-39.
599
REFERENCES
195
Kiein, G. D. (1970). Depositional and dispersal dynamics of inter-
tidal sand bars. J. Sediment. Petrol., 40: 1095-1127.
Kline, S.J. (1967). Observed structure features in turbulent and
transitional boundary layers. In G. Sovran, ed., Fluid Me-
chanics of Internal Flow. Amsterdam: Elsevier, pp. 27-68.
Kline, S. J., VV. C. Reynolds, F. A. Schraub, and P. W. Runstler
(1967). The structure of turbulent boundary layers. J. Fluid
Mech., 30: 741-773.
Krumbein, W. C. (1934). Size-frequency distributions of sedi-
ments. J. Sediment. Petrol., 4: 65-77.
Langhorne, D. N. (1973). A sandwave field in the outer Thames
estuary. Mar. Geol., 14: 129-143.
Langmuir, I. (1925). Surface motion of water induced by the
wind. Science, 87: 119-123.
Leopold, L. B., M. G. Wolman, and J. P. Miller (1964). Fluvial
Processes in Geomorphology. San Francisco: Freeman, 522 pp.
Lilly, D. K. (1966). On the instability of Ekman boundary flow.
J. Atmos. Set., 23: 481^94.
Lin, C. C. (1955). The Theory oj Hydrodynamu Stability. London
and New York: Cambridge University Press, 155 pp.
Lonsdale, P. and B. Malfait (1974). Abyssal dunes of foraminiferal
sand on the Carnegie Ridge. Geol. Soc. Am. Bull., 85: 1697-
1712.
Ludwick, J. C. (1970a). Sand waves in the tidal entrance to
Chesapeake Bay: Preliminary observations. Chesapeake Sci., 11:
98-1 10.
Ludwick, J. C. (1970b). Sandwaves and Tidal Channels in the Entrance
to Chesapeake Bay. Technical Report 1, Inst, of Oceanography,
Old Dominion University, Norfolk, Va., 7 pp.
Ludwick, J. C. (1972). Migration of tidal sandwaves in Chesa-
peake Bay entrance. In D. J. P. Swift, D. B. Duane, and O. H.
Pilkey, eds., Shelf Sediment Transport: Process and Pattern.
Stroudsburg, Pa.: Dowden, Hutchinson & Ross, pp. 377-410.
Ludwick, J. C. (1973). Tidal Currents and Zig-Zag Sand Shoals in a
Wide Estuary Entrance. Technical Report 7, Inst, of Oceanog-
raphy, Old Dominion University, Norfolk, Va., 23 pp.
Ludwick, J. C. (in press). Tidal currents, sediment transport, and
sand banks in Chesapeake Bay entrance, Virginia. In M. O.
Hayes, ed., Second International Estuarine Conf. Proc. Myrtle
Beach, South Carolina, Oct. 15-18, 1973.
McCave, I. N. (1971). Sandwaves in the North Sea off the coast
of Holland. Mar. Geol., 16: 199-225.
McKinney, T. F. and G. M. Friedman (1970). Continental shelf
sediments of Long Island, New York. J. Sediment. Petrol., 40:
213-248.
McKinney, T. F., VV. L. Stubblefield, and D. J. P. Swift (1974).
Large scale current lineations on the central New Jersey
shelf: Investigations by sidescan sonar. Mar. Geol., 17: 79-102.
Middleton, G. V. (1968). The generation of log normal size fre-
quency distributions in sediments. In Problems of Mathematical
Geology. Leningrad: Science Press, pp. 37-46.
Moss, A. J. (1962). The physical nature of common sandy and
pebbly deposits (Part I). Am. J. Set., 260: 337-373.
Moss, A. J. (1963). The physical nature of common sandy and
pebbly deposits (Part II). Am. J. Sci., 261: 297-343.
Moss, A. J. (1972). Bedload sediments. Sedimentology, 18: 159-219.
Newton, R. S., E. Seibold, and F. Werner (1973). Facies distribu-
tion patterns on the Spanish Sahara continental shelf mapped
with sidescan sonar. Meteor. Forsch. Ergebnisse, 15: 55-77.
Oertel, G. F. (1972). Sediment transport of estuary entrance
shoals and the formation of swash platforms. J. Sediment.
Petrol., 42: 857-863.
Off, T. (1963). Rhythmic linear sand bodies caused by tidal cur-
rents. Am. Assoc. Pet. Geol. Bull., 47: 324-341.
Postma, H. (1967). Sediment transport and sedimentation in the
marine environment. In G. H. LaufF, ed., Estuaries. Washing-
ton, D.C.: Am. Assoc. Adv. Sci., pp. 158-180.
Raudkivi, A. J. (1967). Loose Boundary Hydrodynamics. Oxford:
Pergamon, 331 pp.
Robinson, A. H. W. (1966). Residual currents in relation to
shoreline evolution of the East Anglian coast. Mar. Geol., 4:
57-84.
Rosenhead, L. (1963). Laminar Boundary Layers. London and New
York: Oxford University Press (Clarendon), 687 pp.
Russell, R. D. (1939). Effects of transportation on sedimentary
particles. In P. P. Trask, ed., Recent Marine Sediments. London:
Thomas Murphy, pp. 32-47.
Schlichting, H. (1962). Boundary Layer Theory. New York: McGraw-
Hill, 647 pp.
Simons, D. B. and E. V. Richardson (1963). Forms of bed rough-
ness in alluvial channels. ASCE Trans., 128: 284-302.
Simons, D. B., E. V. Richardson, and M. L. Albertson (1961).
Flume studies using medium sand (0.45 mm). U.S. Geol. Surv.
Water Supply Pap. 1498A, 76 pp.
Smith, J. D. (1969). Geomorphology of a sand ridge. J. Geol.,
77: 39-55.
Smith, J. D. (1970). Stability of a sand bed subjected to a shear
flow of low Froude number. J. Geophys. Res., 30: 5928-5940.
Sorby, H. C. (1859). On the structures produced by the current
present during the deposition of stratified rocks. Geologist, 2:
137-147.
Southard, J. B. (1971). Representation of bed configurations in
depth-velocity-size diagram. J. Sediment. Petrol., 41: 903-915.
Southard, J. B. and L. A. Boguchwal (1973). Flume experiments
on the transition from ripples to lower flat bed with increasing
grain size. J. Sediment. Petrol., 43: 1114-1121.
Southard, J. B. and J. R. Dingier (1971). Flume study of ripple
propagation behind mounds on flat sand beds. Sedimentology,
16: 257-263.
Stride, A. H. (1970). Shape and size trends for sand waves in a
depositional zone of the North Sea. Geol. Mag., 469-477.
Stride, A. H. (1973). Interchange of sand between coast and
shelf in European tidal seas (abstract). In Abstracts Symposium
on Estuarine and Shelf Sedimentation, Bordeaux, France, July 1972,
p. 97.
Stride, A. H. (1974). Indication of long term tidal control of net
sand loss or gain by European coasts. Estuarine Coastal Mar.
Sci., 2: 27-36.
Stride, A. H., R. H. Belderson, and N. H. Kenyon (1972). Longi-
tudinal Furrows and Depositional Sand Bodies of the English Channel.
Memoire Bureau Recherches Geologique et Minieres, No.
79, pp. 233-244.
Stubblefield, W. L„ J. W. Lavelle, T. F. McKinney, and D. J. P.
Swift (1975). Sediment response to the present hydraulic
regime on the central New Jersey shelf. J. Sediment. Petrol.,
15: 227-247.
600
196
SUBSTRATE RESPONSE TO HYDRAULIC PROCESS
Swift, D.J. P. (1975). Tidal sand ridges and shoal retreat mas-
sifs. Mar. Geol., 18: 105-133.
Swift, D J. P., D. B. Duane, and T. F. McKinney (1974). Ridge
and swale topography of the Middle Atlantic Bight, North
America: Secular response to the Holocene hydraulic regime.
Mar. Geol.
Swift, D.J. P., B. Holliday, N. Avignone, and G. Shideler (1972a).
Anatomy of a shoreface ridge system, False Cape, Virginia.
Mar. Geol., 12: 59-84.
Swift, D. J. P., J. C. Ludwick, and W. R. Boehmer (1972b).
Shelf sediment transport: A probability model. In D. J. P.
Swift, D. B. Duane, and O. H. Pilkey, eds., Shelf Sediment
Transport: Process and Pattern. Stroudsburg, Pa.: Dowden,
Hutchinson & Ross, pp. 195-223.
Swift, D. J. P. and R. M. McMullen (1968). Preliminary studies
of intertidal sand bodies in the Minas Basin, Bay of Fundy,
Nova Scotia. Can. J. Earth Sci., 5: 175-183.
Van Veen, J. (1936). Under zoekingen in der Hookden. The" Hague:
Algemene Landsdrukkerij, 252 pp.
Visher, G. S. (1969). Grain size distribution and depositional
processes. J. Sediment. Petrol., 39: 1074-1106.
Weil, C. B., R. D. Moose, and R. E. Sheridan (in press). A model
for the evolution of linear tidal built sand ridges in Delaware
Bay, U.S.A. In G. Allen and A. Klingbiel, eds., Symposium
International: Relations Sedimentaires entre estuaires el plateaux con-
tinentaux, University of Bordeaux, July 9-14, 1973.
Wilson, I. G. (1972). Aeolian bedforms — their development and
origins. Sedimentology, 19: 173-210.
Wilson, I. G. (1973). Equilibrium cross-section of braided and
meandering rivers. Nature, 241: 393-394.
Yih, C. S. (1965). Dynamics of Non-homogeneous Fluids. New York:
MacMillan, 235 pp.
Znamenskaya, N. S. (1965). The use of the laws of sediment dune
formation in computing channel .formation. Soviet Hydrology
Selected Papers (Am. Geophys. Union 1966), 5: 415-432.
601
51
Reprinted from: Middle Atlantic Shelf and the New York Bight, ASLO
Special Symposia, Volume 2, 69-89.
Section 3
Geological processes
Morphologic evolution and coastal sand transport, New York— New Jersey
shelf1
Donald J. P. Swift, George L. Freeland, Peter E. Gadd, Gregory Han,
J. William Lavelle, and William L. Stubblefield
Atlantic Oceanographic and Meteorological Laboratories, NOAA, Virginia Key, Miami, Florida
Abstract
The surface of the New York-New Jersey shelf has been extensively modified by land-
ward passage of nearshore sedimentary environments during the postglacial rise of sea level.
The retreat of estuary mouths across the shelf surface has resulted in shelf valley complexes.
Constituent elements include shelf valleys largely molded by estuary mouth scour, shoal
retreat massifs left by the retreat of estuary mouth shoals, and midshelf or shelf-edge deltas.
The erosional retreat of the straight coast between estuary mouths has left a discontinuous
sheet of clean sand 0-10 m thick. During the retreat process, a sequence of oblique-trending,
shoreface-connected sand ridges formed at the foot of the shoreface. As a consequence, the
surficial sand sheet of the shelf floor bears a ridge and swale topography of sand ridges up
to 10 m high and 2-4 m apart.
The mechanics of sedimentation in these two nearshore environments ( estuary mouth
and interestuarine coast ) are now being investigated for purposes of environmental man-
agement as well as for further understanding of shelf history. In late fall and winter 1974,
current meters were deployed on the Long Island coast and a radioisotope tracer dispersal
pattern was traced over an 11-week period. Eastward or westward pulses of water were
generated during this period of successive weather systems. Flows in excess of the computed
threshold velocity of substrate materials were sustained for hours or days and were separated
by days and weeks of subthreshold velocities, and the sand tracer pattern expanded accord-
ingly. A single intense westward flow transported more sand than all the other events com-
bined. The storm was anomalous with respect to the short term observation period, but it
may in fact have been representative of the type of peak flow event that shapes the inner
shelf surface.
Systematic observations of sedimentation in New York Harbor mouth have not yet been
initiated. However, reconnaissance data reveal a complex pattern of ebb- and flood-domi-
nated zones that control the pattern of sand storage.
We review in this paper our knowledge
of the surface of the continental shelf off
New York and New Jersey by considering
two distinct topics: the geological history of
this surface and the nature of sand transport
across it. Our knowledge of the New York-
1 Contribution of the New York Bight Project of
the NOAA Marine EcoSystems Analysis (MESA)
Program.
AM. SOC. LIMNOL. OCEANOGR.
New Jersey shelf surface is primarily the
result of a decade of work by K. O. Emery
and his colleagues at the Woods Hole
Oceanographic Institution. A summary of
this information and much more has re-
cently been provided by Emery and Uchupi
(1972). As the work of the Woods Hole
group drew to a close, we attempted to con-
sider in greater detail some aspects of the
morphologic evolution of the Middle At-
gg SPEC. SYMP. 2
602
70
Geological processes
lantic Bight surface (Swift et al. 1972, 1974;
Swift 1973; Swift and Sears 1974; Stubble-
field et al. 1974). A summary of this work
constitutes the first section of this paper.
As participants in NOAA's MESA (Ma-
rine EcoSystems Analysis) program, we
have been asked not only to evaluate the
geological history of the New York Bight,
but also to provide quantitative estimates of
sediment transport that will be of direct use
to environmental managers. It turns out that
these two goals are closely related. Our sur-
veys of the shelf surface have led us to in-
fer that it has been shaped by the landward
retreat of two basic sedimentary regimes
during the Holocene transgression: tide-
dominated sedimentation at estuary mouths,
and the sand transport pattern of the ad-
jacent shoreface and adjacent inner shelf.
Environmental engineers and managers
must deal with these same regimes.
To satisfy their needs, we have initiated
real-time studies of fluid motion and sub-
strate response. State-of-the-art techniques
for such studies are inadequate and progress
has been slow. We report in the second por-
tion of this paper fragments of our studies
of sand transport to encourage colleagues
engaged in similar studies. Our own initial
experiments have raised more questions
than they have answered.
Evolution of the continental shelf surface
Evolution of shelf valley complexes — The
New York Bight is a pentagonal sector of
the North American Atlantic shelf, extend-
ing 800 km from Cape May, New Jersey, to
Montauk Point, Long Island. Off New York,
the shelf is 180 km wide ( Fig. 1 ) .
The sandy shelf floor is divided into com-
partments by shejf valley complexes extend-
ing from the shoreline toward shelf edge
canyons ( Fig. 1 ) . The most obvious ele-
ments of these complexes are the shelf val-
leys themselves which may appear as nar-
row, well defined channels ( Delaware Shelf
Valley; Hudson Shelf Valley) or as broad,
shallow depressions which barely perturb
the isobaths defining the shelf surface
( Block Shelf Valley, Long Island Shelf Val-
ley, North New Jersey Shelf Valley, Great
Egg Shelf Valley). Shelf valley complexes
generally contain other morphologic ele-
ments. The north rims of the shelf valleys
76°39°
75° 40°
74° 41°
73° 42°
74c
--■SURFACE CHANNEL
• ••• SUBSURFACE CHANNEL
' SCARP
-200m-
L\X| CUESTAS
SHELF EDGE, MID-SHELF DELTAS
ftS SHOAL RETREAT MASSIFS S-= SAND RIDGES
-X.
/\
^L
73° 38° 72° 39° 71° 40°
Fig. 1. Morphologic framework of the New York-New Jersey shelf. (Modified from Swift et al.
1972.)
603
New York-New Jersey shelf
71
tend to be elevated like levees above the
adjacent shelf. Seaward ends of shelf valleys
often terminate in delta terraces. Shelf val-
ley complexes tend to be broken into seg-
ments by coast-parallel scarps, which may
have been formed by temporary stillstands
of the returning Holocene sea.
The origin of the shelf valley complexes
is best inferred from the configuration of
the Delaware Shelf Valley (Fig. 2), which
can be traced without interruption into its
modern estuary. The Delaware estuary
mouth has a sill of sand nourished by littoral
drift from the New Jersey coast ( Swift
1973). The sill is stabilized by an inter-
digitating system of ebb- and flood-domi-
nated channels, whose discharge inequali-
ties are a consequence of the phase lag of
the tidal wave in its passage across the sill
( Swift and Ludwick in press ) . The Dela-
ware Shelf Valley may be traced directly
into the flood channel of the main ebb
channel-flood channel couplet. Its leveelike
north rim may be traced directly into the
complex of smaller ebb channels, flood
channels, and sand banks on the north side
of the estuary mouth. This shoal area serves
as the depositional center for the littoral
sand discharge of the New Jersey coast.
The shelf valley complex, then, is not a
drowned river valley, but is rather the track
left by the retreat of the Delaware estuary
mouth across the shelf during the Holocene
sea-level rise. The shelf valley is the retreat
path of a flood channel. The north flank
levee is the retreat path of the estuary
mouth shoal or is a shoal retreat massif —
massif in the sense of a compound topo-
graphic high consisting of smaller scale
highs (Swift 1973). The surface channel
does not directly overlie the buried river-
cut channel but is offset to the south ( Sheri-
dan et al. 1974). As the estuary retreated
up the river valley, it not only tended to fill
the river valley but in the final, estuary-
mouth stage decoupled from it altogether
by migrating to the south.
The largely constructional nature of the
Delaware Shelf Valley complex is also char-
acteristic of the Great Egg Shelf Valley
complex (Fig. 1), although the associated
massif has been heavily dissected by the
posttransgressional regime of southerly
storm flows. To the north, however, the
Hudson and Block Shelf Valleys occur on a
terrain of innately greater relief. There are
cuestalike highs, and the estuarine deposits
only partly fill the shelf valleys. The deeply
incised nature of the Hudson Shelf Valley
may reflect the era when it received Great
Lakes meltwater ( Veatch and Smith 1939 ) .
Evolution of interfluves — Plateaulike in-
terfluves between the shelf valleys have
likewise been intensively modified by pas-
sage of the shoreline. Interfluve surfaces
range from exceedingly flat plains (slopes
of 1:2,000) to irregularly undulating sand
ridge topography (Fig. 3). Sand ridges ex-
hibit up to 10 m of relief, are spaced 2 to 4
km apart, and their crestlines may be traced
for tens of kilometers. Side slopes are gener-
ally less than a degree. Crestlines are not
quite parallel to the regional trend of the
isobaths but tend to converge to the south-
west with the shoreline (Fig. 1). Ridges at-
tain their maximum development on the
northeast sides of shoal retreat massifs.
The ridges are molded into a surficial
sheet of relatively homogeneous, well sorted
sand, 0-10 m thick. In trough axes the sheet
thins to a basal shelly, gravelly sand several
decimeters or less thick, and a more hetero-
geneous older substrate is locally exposed
(Donahue et al. 1966; Stubblefield et al.
1974). This is commonly a muddy sand or
mud deposited behind the retreating Holo-
ce.ie barrier system (Stahl et al. 1974; Sheri-
dan et al. 1974), but it is locally absent due
to erosion or nondeposition, so that the
Holocene sands rest directly on older Pleis-
tocene sands.
To understand the genesis of this post-
glacial stratigraphy, it is necessary to con-
sider the dynamics of a transgressing shore-
line. We are indebted in this regard to
Bruun (1962) and Fischer (1961) who ap-
pear to have independently appreciated the
role of the landward translation of the
wave- and current-maintained coastal pro-
file in generating transgressive stratigraphy.
In the New York Bight, as along most low,
unconsolidated coasts, the coastal profile
consists of a steeply sloping nearshore sec-
tor ( the shoref ace ) and a gently dipping in-
604
72
Geological processes
MODERN ESTUARY
MOUTH SHOAL,
TIDAL CHANNELS
PAIRED FLOOD
CHANNEL RETREAT
TRACK, ESTUARINE
SHOAL-RETREAT MASSIF
40M SCARP
TRANSGRESSED
CUSPATE DELTA;
(CAPE SHOAL-
RETREAT MASSIF)
60M SCARP
605
New York-New Jersey shelf
73
39*I0'N
39"05'
74*00'
73*45' W
Fig. 3. Simplified bathymetry and distribution of grain sizes on a portion of the central New Jersey
shelf. Medium to fine sand occurs on ridge crests. Fine to very fine sand occurs on ridge flanks and in
troughs. Locally, erosion in troughs has exposed a thin lag of coarse, shelly, pebbly sand over lagoonal
clay. ( Reprinted from Stubblefield et al. 1974 by permission of the Journal of Sedimentary Petrology. )
ner shelf floor. The break in slope, which
may be well defined or gently rounded,
generally occurs at depths of 12 to 18 m,
some few kilometers from the shoreline.
Bruun (1962) pointed out that if this
profile is in fact an equilibrium response
of the seafloor to the coastal hydraulic cli-
mate, then a rise in sea level must result in
a landward and upward translation of the
profile (Fig. 4A). Such a translation neces-
sitates erosion of the shoreface. Much of
the resulting debris will presumably be
entrained in the littoral current and move
downcoast, but during periods of onshore
storm winds, the littoral drift may leak sea-
ward, due to an offshore component of bot-
tom flow, to be deposited beneath the rising
seaward limb of the equilibrium profile on
the adjacent inner shelf floor.
Evidence for such seaward bottom trans-
port is varied. Murray (1975) described
periods of offshore bottom flow on the gulf
coast, when winds are onshore and the wa-
ter column is not stratified. Sonu and Van
Beek (1971) noted that sand loss from
North Carolina beaches correlates poorly
with periods of high waves but correlates
well with high waves generated by onshore
northeast winds. On the Long Island inner
shelf, we used sidescan sonar to map inner
Fig. 2. Delaware Shelf Valley complex. Southward littoral drift along the New Jersey coast is injected
into the reversing tidal stream of Delaware Bay mouth. The resulting sand shoal is stabilized as a system
of interdigitating ebb- and flood-dominated channels. The shelf valley complex seaward of the bay mouth
was formed by the retreat of the coastal sedimentary regime through Holocene time. Retreat of the main
flood channel has excavated the Delaware Shelf Valley; retreat of the bay mouth shoal has left a levee-
like high on the shelf valley's north flank. ( Reprinted from Swift 1973 by permission of the Geological
Society of America Bulletin.)
606
Geological processes
VECTOR RESOLUTION OF
PROFILE TRANSLATION
WASHOVER CYCLE
OF BARRIER SANDS
Fig. 4. Models tor a coast undergoing erosional shoreface retreat during a rise in sea level. A — Rise in
sea level results in landward and upward translation of coastal profile ( Bruun 1962). B — Translation is
accomplished. Wind and storm washover transport on the barrier surface and erosion of the shoreface
and seaward transport of the resulting debris (Fischer 1961).
shelf current lineations that form an east-
ward-opening angle with the beach (Fig.
5). A poorly defined asymmetry is appar-
ent: the western sides of the lineations are
gradational, whereas the eastern sides are
sharply defined. The origin of this pattern is
not clear. The dark bands are strips of
coarse or gravelly sand that may either be
troughs between low amplitude, current-
transverse sand waves or troughs between
current-parallel sand ribbons. However,
considering the angle that the lineations
make with the beach, sand ribbons seem
unlikely for reasons of flow continuity. If
sand waves, the lineations are responses to
strong bottom flows trending westerly and
offshore.
Fischer (1961), Stahl et al. (1974), and
Sanders and Kumar (1975) described the
stratigraphic consequences of erosional
shoreface retreat, based on their observa-
tions of the New Jersey and New York
coasts. The barrier superstructure will re-
treat over the lagoonal deposits by a cyclic
process of storm washover, burial, and re-
emergence at the shoreface ( Fig. 4B ) .
Lower shoreface sands will tend to be trans-
ported seaward to accumulate over the
607
New York-New Jersey shelf
75
40° 2 5' i
74°00' 73-55' 73°50' 73°45'
100 METERS
Fig. 5. Sidescan sonar records of current lineations on the Long Island coast, collected at three dif-
ferent periods. Positioning by Raydist. Current lineation pattern (bands A-F) expands to south during
observation period. Apparent change in orientation in last panel is due to ship maneuvering. ( From
Stubblefield et al. in prep. )
eroded surface of the lagoonal deposits as
the leading edge of a marine sand sheet.
Bruun's hypothesis is compatible with the
stratigraphic evidence and with our limited
knowledge of coastal hydraulics. A more
rigorous test requires bathymetric time se-
ries to document changes in the coastal pro-
file. Limited data of this sort are becoming
available. Harris (1954) undertook a study
of the Long Branch, New Jersey, dredge
spoil dumpsite to determine if dumping was
nourishing the beach (Fig. 6). In fact, dur-
ing a 4-year period, the shoreface under-
went between 5 and 26 cm of erosion, while
an irregular pattern of deposition prevailed
on the inner shelf floor. A somewhat longer
time series has been prepared by Kim and
Gardner (Woodward-Clyde Assoc.) during
study of proposed sewage outfall routes for
the Ocean County, New Jersey, sewerage
authority ( Fig. 7 ) . Two out of three profiles
taken indicate 1.5-2.0 m of erosion over 20
years. The third profile is immediately south
of a shoreface-connected sand ridge; here
comparable aggradation has occurred as a
consequence of southward ridge migration.
Growth of ridges — Erosional shoreface
retreat on the Atlantic cannot be adequately
described by a two-dimensional model such
as Fig. 4 because the shoreface appears to
be the formative zone for sand ridge topog-
raphy as well as for the sand sheet into
which it is impressed. Clusters of shoreface-
sand ridges occur on the New Jersey coast
between Brigantine and Barnegat Inlets, on
the north New Jersey coast between Mana-
squan and Sea Bright, and on the Long Is-
land coast from Long Beach to the shore-
face of eastern Fire Island.
The shoreface-connected ridges are
named for their oblique, fingerlike exten-
sions of the shoreface, causing seaward de-
flections of isobaths as shoal as 5 m. The
ridges tend to be asymmetric in cross-sec-
tion, with steep seaward flanks, although
this relationship may be reversed at the
base of the ridge where it joins the shore-
face. Seaward flanks tend to be notably
608
76
Geological processes
Fig. 6. Erosion and deposition near Long Branch, New Jersey, dredge spoil dumpsite during a 4-
yr period. Recorded changes are 0.4-1.4 ft. Shoreface lias undergone erosion; adjacent seafloor primarily
has undergone aggradation. ( From Harris 1954. )
6
4
2
en. 0
oc
u -?
h-
uj -4
-6
-8
-10
200 400 600 800
METERS
c
, , J 1-
NN^ 1973
1953 ^""~:u^zr--
i i . j i i i — i — i ^^r* — i
200 400 600 800
METERS
1000
200 400 600 800 1000 1200 75°00' 45' 30'
METERS
15
74°00
Fig. 7. Profiles of proposed sewage outfall sites on the New Jersey coast. Sites A and B have eroded
over a 20-yr period. Site C, immediately downcoast of a shoreface-connected sand ridge, has aggraded.
( Reprinted from Kim and Gardner 1974 with permission of Woodward-Clyde Assoc. )
609
New York-New Jersey shelf
77
finer than landward flanks. Off Brigantine
Inlet and off the New Jersey coast, shore-
face-eonnected ridges are associated with
free-standing inner shelf ridges that can be
traced seaward for tens of kilometers in ap-
parent genetic sequence. The ridges form
on the shoreface in response to south-trend-
ing coastal storm currents ( Duane et al.
1972 ) and become detached from the shore-
face as it retreats. They tend to migrate
downcoast (to the south or west) and off-
shore, extending their crestlines so as to
maintain contact with the shoreline (Fig.
8). Eventually, however, contact is broken,
and they are stranded on the deepening
shelf floor. Downcoast ridge migration is
part of a general pattern of southwesterly
sand transport on the Atlantic shelf. In the
offshore ridge topography, this pattern is
indicated by the tendency of both ridge
crests and trough talwegs to rise toward
the southwest. Locally, it is indicated by
patterns of erosion and deposition near
wrecks ( Fig. 9).
Sand transport on the inner shelf
The preceding description of the mor-
phologic evolution of the New York shelf
surface is based primarily on the interpre-
tation of bathymetric maps, aided by local
substrate inventories in which the bottom is
39°29'N
39°28N
74°I6'W 74°I5'W 74°I4'
Fig. 8. Patterns of erosion and deposition on
Beach Haven Ridge, New Jersey, between 1935
and 1954, superimposed on 1954 bathymetry. Pat-
tern of north flank erosion and south flank deposi-
tion indicates downcoast migration of ridge. (From
DeAlteris et al. in press. )
examined by grab sampling, photography,
Vibracoring, and seismic profiling. The con-
clusions are qualitative but nonetheless
valid. However, fuller understanding of the
behavior of the shelf surface requires a
different approach.
We must directly measure fluid and sedi-
ment transports involved in the two basic
mechanisms that have shaped the shelf sur-
face: tidal flow and sand storage at estuary
mouths, and erosional retreat of the shore-
face between estuary mouths. Environ-
mental managers who must make decisions
about dredged channels, sewage outfalls,
sewage and dredge spoil dumpsites, deep-
water tanker terminals, and offshore power
plants need to understand these processes
before they can evaluate the stability of the
inner shelf surface.
The nature of coastal sand transport dur-
ing storms is the first major problem we
will consider. Fluid motions in the surf
zone have been studied for decades, and the
role of longshore currents driven by shoal-
ing and breaking waves has been described
(e.g. Bowen 1969). In the New York area,
®
■1 WRECK
r~l ACCRETION
E3 SCOUR
®
MN<M
1
\
NE \
/ENE
i
wsw —
/
SSE1
-— sw
0 100 200
FEET
Fig. 9. Accretion and scour by a wreck near
Beach Haven Ridge, New Jersey. ( From DeAlteris
et al. in press. )
610
78
Geological processes
massive discharges of sand in the surf zones
of the Long Island and New Jersey coasts
move toward the New York harbor mouth;
these discharges have built Sandy Hook
and Rockaway spits within subhistoric to
historic times. However, we know almost
nothing about fluid motions over the ad-
jacent inner shelf, although the geologic
data presented above show that currents
seaward of the surf play a major role in the
coastal sand budget. We must specifically
ask what time and space scales of inner
shelf flows are intense enough to entrain
sand? Is their velocity field so structured
that there are periods of significant offshore
bottom flow and sand transport?
Equally important is the problem of the
inner shelf sand ridges, which seem to occur
wherever a sewage outfall or power plant is
to be located. If we wish to predict the
probable behavior of these features through
the design life of the structure, we must un-
derstand their genesis and how they are
maintained by flow. It is a truism of loose
boundary hydraulics that sheared boundary-
flows are innately unstable, and that these
instabilities tend to interact with the sub-
strate to generate sand ripples, sand ribbons,
sand waves, and sand dunes. The circum-
stantial evidence that inner shelf sand ridges
are similarly responses to flow is strong.
How are they formed and maintained?
As a first attempt to investigate these
questions, Lavelle et al. (in press) placed
40 Aandaraa current meters at 19 stations
over the Tobay Beach sand ridges of the
Long Island inner shelf (Figs. 10 and 11).
The meters were in place for 6 weeks dur-
ing late November and December 1974; a
single meter recorded for an additional 5
weeks. All meters averaged speed over 10
min and took an instantaneous direction
reading during each sampling period.
During the observation period, a series of
moderate storms induced easterly and west-
erly flows parallel to the coast. A final storm
on 1-4 December was very intense, causing
more beach erosion than any storm since
the Ash Wednesday storm of 1962 (C. Gal-
vin personal communication ) .
In Fig. 12, vector averages for all near-
bottom, middepth, and near-surface meters
are presented for periods of both westward
and eastward flows. A wind-controlled pat-
tern of coastal flow emerges. There is a top
to bottom speed shear as well as a direc-
tional shear. Prevailing fall and winter
winds blow out of the northwest, across the
east-west Long Island shoreline; the result
is a tendency toward coastal upwelling.
Surface flows have an offshore component
for both eastward and westward flow direc-
tions. The response is less symmetrical at
depth; westward bottom flows parallel the
isobaths, whereas eastward bottom flows
have an onshore component. Net water
transport during the observation period was
eastward.
During the early December storm there
was a small offshore component to the water
flow near the bottom. Figure 13 shows the
winds during the storm and the associated
current velocities from a near-bottom cur-
rent meter, which have been filtered with
a 40-h and a 3-h low-pass filter. The 40-h
low-pass filtered record, which is a segment
• VERTICAL CURRENT METER STATIONS
® RIST DROP
Fig. 10. Bathymetry, current meter stations,
and tracer release point ( RIST drop) for the Long
Island nearshore (LINS) experiment. (From La-
velle et al. in press. )
611
New York-Neiv Jersey shelf
79
40°
35'
73°28'
73°2I'
ES3>'0 CZ! 1-0-1.5 I |lJS-gJ0 V7A 20-2-5 iH<2.5 -GRAB SAMPLE "BOX CORE
Fig. 11. Distribution of grain sizes over the Tobay Beach ridges, LINS area. Size classes in phi units.
taken from Fig. 12, obscures the brief time-
scale flow associated with the storm. The
3-h low-pass record, which is only slightly
smoothed and still contains the tidal signal,
shows a period of offshore flow more
clearly. These results must be viewed cau-
tiously. The Aandaraa current meters which
were used have large direction and speed
errors when used in shallow water with sur-
face wave amplitudes as large as were pres-
ent during the event described here.
During the November-December exper-
iment on the Long Island inner shelf, esti-
mates of sand transport were made from
calculations from current meter records
(Lavelle et al. in press) and also from radio-
isotope tracer dispersal patterns ( Lavelle et
al. unpublished data). To generate the pat-
terns, about 500 cm3 of indigenous fine to
very fine sand was surface-coated with 10
Ci of ,,,:,Ru (half-life, 39.6 d). On 12 No-
vember, equal portions of tagged sand were
released in water soluble bags at three
points at the east end of the main trough
(Fig. 14). The injection points formed an
equilateral triangle with sides roughly 100
m long. The developing dispersal pattern of
labeled sand was surveyed at intervals by
scintillation detectors mounted in a cylinder
towed across the bottom. Navigation was
by a Raydist system with 10-m resolution.
Four postinjection surveys were made dur-
ing the 11-week tracer experiment. Disper-
sal patterns mapped 2 and 8 weeks after in-
jection are shown in Fig. 14. After 2 weeks
(25 November) roughly ellipsoidal smears
trended east from each of the three injection
points (Fig. 14A). Each smear could be
traced for about 200 m before the signal
was lost in the background radiation. After
8 weeks (10 January) the three eastward
smears had been replaced by a single, more
extensive pattern extending 700 m to the
west (Fig. 14B). Partially processed data
from an intermediate survey ( 17-19 Decem-
ber) indicate that the reversal in fact had
612
80
Geological processes
12 NOV 74
13
14
15
16
17
18 N0V74
19
20
21
22
23
24
25 NOV 74
26
27
28
29
30
®
i5*otf-
*~E
W —
DIR
SPD
B
64°
47
M
68°
69
S
81°
120
(b) long term bottom, middepth.and
near-surface current means —
eastward flow
5cm/s
DIR
SPD
B
25 3°
36
M
241°
69
S
239°
96
LONG TERM BOTTOM, MIDDEPTH, AND
NEAR-SURFACE CURRENT MEANS-
WESTWARD FLOW
ENT DIRECTION AND SPEED
Fig. 12. Summary of flow data for the LINS experiment. A — Vector time series of representative
near-bottom flow. Data have been subjected to a 40-h low-pass filter. B, C — Long term velocity averages
of eastward and westward flow for meters grouped by depth in water column. Bottom, middepth, and
near-surface groupings are labeled B, M, and S. (From Lavelle et al. in press.)
613
Netc York-New Jersey shelf
81
A
(BOTTOM FLOW)
ipnisiiininiiungii'iiii^
0 20
cm/s
B
WIND)
COASTLINE
V />.
cm/s
X
29
30 DEC 1
Fig. 13. Vector time series for bottom current
and wind velocities during the 1-4 December
storm. A — 40-h low-pass filtered record ( Lanczos
filter with response. —6 db at 36 h and —20 db at
40 h ) . B — Wind record from Ambrose Tower. C —
3-h low-pass filtered record (Lanczos filter with
response. —6 db at 2.5 h and —20 db at 3 h).
occurred before this and that it initially had
been at least 1,200 m long.
The temporal pattern or sediment trans-
port over 60 days may be inferred from
Fig. 14C. Current speed, measured 1.5 m
from the bed, is plotted against time. The
horizontal line at 18 cm/s is an estimated
threshold for the fine to very fine sand
(mean diameter, 3.0 (f>) found at the site.
It is based on the work of Shields and sub-
sequent workers (Graf 1971: p. 90) and on
a choice of 3.0 X 10 ~3 for the drag coefficient
( Sternberg 1972 ) . This choice of threshold
velocity was supported by empirical evi-
dence obtained during the course of the
experiment (Lavelle et al. in press). Esti-
mates have been prepared for the relative
role each transport event played in the
overall transport record, based on the con-
cept of factional energy expenditure pro-
portional to the transport volume (Bagnold
1963). For each event where velocities ex-
ceeding threshold were recorded, a trans-
port volume was calculated:
Qi = a \ (|«| - lutnD-'dt,
[
where |u| is measured current speed, \utj,\
is threshold speed, a is a constant of pro-
portionality, and f; is the duration of the
transport event (Lavelle et al, in press).
Expression of sand transport as a power
of the difference of measured and threshold
velocity is supported by Kennedy's ( 1969 )
analysis of stream transport data. Without
assigning a value to a, we can calculate the
rate of transport of one flow event relative
to the next or in relation to the sand dis-
charge that occurred over the entire dura-
tion of the current meter record. The sec-
ond of these options has been used in Fig.
14C, where relative sand transport as per-
cent of total transport has been represented
as solid bars superimposed on the current
meter record. Bar height is a measure of
volume percent of transport; bar width is a
measure of duration of the transport event.
Despite the exceedence of the sediment
transport threshold at many points in the
record, only the solid bars centered on 2
and 16-17 December are visible in the fig-
ure. Thus sand transport during observation
614
82
Geological processes
consisted of periods of quiescence separated
by brief, intense transport events. Further-
more, since discharge is calculated as a
power function of excess velocity, intense
storms are far more efficient transporters
of sand than mild ones. Although the trans-
port index calculated for the 1^4 December
storm may be biased by the choice of
threshold speed as well as by the functional
dependence on velocity, it seems probable
that any reasonable parameterization would
lead to the same general conclusion: the
storm event of 1-4 December moved more
sand at 20-m water depth than the com-
bination of all other transport events.
Attempts have also been made to calcu-
late sediment transport indices over longer
periods of time in the New York Bight apex.
The following computation is based on 30-
80-day Aandaraa current meter records
( Fig. 15 ) . Data in each current meter rec-
ord consist of an average speed, u, and an
instantaneous direction, 6, taken for each
10-min sampling interval. For each inter-
val in which an assigned threshold speed,
|u,J, is exceeded, a sediment transport in-
dex, Q, has been computed, as follows:
Q =
u —
\utn\y, (|w| - \uth\) >o.
For each current meter, the set of vectors of
flow direction, 6 (0°^ 0^359°), and of
sediment transport index, Q, is sorted into
10-degree classes. The results are plotted as
400
V)
K
UJ
W200
25 NOV 74
mm^M
400-
B
u200
2
1000
800
600
400
200
200
400 METERS
10 JAN 75
— i 1 1 1 1 r
1000 800 600
400
200
,r *■■:;. :^> <-■■■■ ■
" 1
II
II ■
II
|l
1
1 1
0
200
400 METERS
S ioo
£ 80-
2 60-
g 40
<
<E 20
* I2N0V'74 20
BM I Threshold Velocity = -*»U
a*T P* — i 1 f""r »^*iipin|imi|^^ i 1 i i 1 1 — ^ 1 1 1 1 1 1 1 1 1 1 1— —i (-0
V'74 20 28 6DEC'74 14 22 30 7JAN'75
80
60
40
7JAN'75
Fig. 14. Sand transport data. A, B — Dispersion patterns measured 13 and 59 days after injection of
tagged sand. Point sources are represented by dots. Broken line is the survey trackline. Dots, coarse dots,
and Xs indicate increasing intensity of radiation. C — Near-bottom current speed record over the duration
of the experiment and calculated sediment transport information. (From Lavelle et al. in press.)
615
Neic York-New Jersey shelf
APRIL - JUNE 1974
83
7.73 x 103
74°00' 7:
^MUD
lllll SILTY-FINE SANDS
□ fine-med. sands
%?m
73°40'
COARSE SANDS
SANDY GRAVEL
ARTIFACT GRAVEL
Fig. 15. Bathymetry, bottom sediment character, and calculated patterns of sediment transport for
April-June 1974 in the New York Bight apex (see text). Depth in fathoms.
rose diagrams in Fig. 15. The length of each For each current meter station in Fig. 15,
radial bar is proportional to the mean sedi- the normalized resultant of all sediment
ment transport index, while the width is transport vectors is indicated by a single ar-
proportional to the duration of flow above row. The resulting magnitude has been di-
threshold, hence the bars may overlap. vided by the total number of days that the
616
84
Geological processes
current meter was in operation ( TP ) to de-
rive a daily average, QD :
T
Qd - -f~ J ( lul _ \Uth\ Vudt,
where u is a unit vector with direction 6,
and T is the total number of days. The in-
tegrand is zero when the velocity is less
than threshold.
Figure 15 suggests that during April-June
1974, sand transport was westward off the
Long Island shore and southward off the
New Jersey shore. Nearshore stations reveal
a strong onshore component of the sand
transport index, perhaps because of wind-
induced upwelling or because of the land-
ward directed asymmetry of bottom wave
surge, or both. The magnitude of the sand
transport index generally decreases seaward
but is anomalously large within the Hudson
Shelf Valley. The easterly transport revealed
by a single station off the Long Island coast
is probably due to instrument problems.
Some unsolved problems
The inner shelf sand budget — Our studies
of sand transport on the New York inner
shelf have resolved some questions but
raised others. It is clear that sand transport
occurs seaward of the surf zone. Transport
is episodic in nature. Sand is entrained and
transported by brief, intense, wind-driven
coast-parallel flows lasting for hours or days
and separated by days or weeks of
quiescence. Our measurements suggest that
inner shelf bottom flows are more likely to
transport shelf sands shoreward than sea-
ward. This appears to be due to intermittent
coastal upwelling induced by northwesterly
winds and perhaps also to the landward-
oriented asymmetry of near-bottom wave
surge. Baylor (1973) has also noted this
pattern of wind-induced coastal upwelling
off Long Island, and R. Scarlet (EG&G,
Waltham, Mass., unpublished) reported a
similar regime of coastal upwelling for the
Beach Haven Ridge site ( Fig. 16 ) .
However, our observations indicate that
the 1-4 December storm was the only event
that caused massive sand transport. It stands
out within our two periods of current moni-
toring not only in its duration, intensity, and
westward direction of net transport but in
the offshore component of bottom flow. We
must consider the hypothesis that the 1-4
December storm, anomalous within the con-
text of our short term winter observation
period, is in fact the kind of peak flow
event that shapes the inner shelf surface and
controls its sand budget. We have noted
that Atlantic shelf stratigraphy is best ex-
plained by erosional shoreface retreat and
seaward transport of the eroded material.
We have described the southwest migration
of shoreface-connected ridges off New Jer-
sey and have cited evidence for the net
southwest transport of sand (Fig. 8). We
note that the Tobay Beach ridges (Fig. 11)
are, like other Atlantic Bight ridge fields,
asymmetrical in both grain-size distribution
and morphology; the seaward-facing south-
west slopes are steeper and finer grained,
implying that westward flows scour the
upcurrent flanks and deposit fine sand on
the seaward-facing downcurrent flanks.
Recent studies by physical oceanog-
raphers also suggest that southwestward
currents generated by "northeaster" storms
have the greatest potential for shaping the
shelf surface. Beardsley and Butman ( 1974)
have described a scale-matching phenome-
non, in which the Middle Atlantic Bight
tends to interact with "northeasters" of the
appropriate size and trajectory so that in-
tensive southwestward flows result (Fig.
17). Their observations indicate that if low
pressure cells cross the bight on a trajectory
such that the isobars of atmospheric pres-
sure cross the isobaths of the shelf surface at
a high angle, then oscillations of the water
column may result, but there is little net
displacement of water. However, when the
trajectory and scale of the storm are such
that for a period it rests in the Middle At-
lantic Bight so that the isobars parallel the
isobaths, then strong sustained coupling of
wind and water flow results. The winds
blow along the isobars, down the arc of the
Middle Atlantic Bight. Landward Ekman
transport of surface water causes 40 to 60
cm of coastal setup and results in a south-
617
Neic York-New Jersey shelf
UPCOAST
85
DOWNCOAST
Fig. 16. Polar histograms of hourly averaged, de-tided summer currents in cm/s in the vicinity of
Beach Haven Ridge, New Jersey. Only flows associated with winds over 5 m/s are shown. Prevailing
wind is indicated by location of histogram on page with wind direction shown in center. Inner ring of
histograms is for near-surface measurements, outer ring is for near-bottom measurements. Directions of
winds and currents are indicated with top of page representing upcoast motion (036° true). Histo-
grams are omitted if fewer than 35 h of data were found for specified wind condition. Solid contours
enclose 50% and dashed contours enclose 90% of data. (Adapted from EG&G Environ. Consult. 1975.)
ward geostrophic transport of the shelf wa-
ter column that is coherent and slablike.
Boicourt and Hacker (1976) described
a similar period of southward storm flow on
the Virginia coast with sustained middepth
velocities of 30-50 cm/s. Both sets of inves-
tigators noted a marked asymmetry in the
hydraulic climate, whereby southwest storm
flows tend to be noticeably more intense
than northeast flows.
It is clear from the preceding discussion
that the role of storm-driven currents in
mediating the coastal sand budget requires
additional study. We need to know more
about the frequency of southwestward
storm flows and their velocity structure. We
618
86
Geological processes
Fig. 17. Surface weather maps for 18 and 22 March 1974. Only the second storm produced sus-
tained coupling between wind and water flow. ( From Beardsley and Butman 1974. )
must also learn to design experiments that
will resolve perturbations of flow that build
and maintain ridge systems.
Sand transport and storage at New York
Harbor mouth — Major sections of the New
York-New Jersey shelf have been shaped by
the tidal regimes associated with estuary
mouths during the postglacial rise of sea
level. Sand budgets of estuary mouths are
also of great interest to environmental man-
agers; the Atlantic coast estuaries are the
approaches to the major coast ports and
require repeated costly dredging. At pres-
ent, the only estuary mouth subjected to
systematic study is that of Chesapeake Bay
(Ludwick 1972, 1974, in press). However,
reconnaissance data are available for the
Hudson estuary mouth, which suggest di-
rections for further study.
New York Harbor mouth is clearly a sink
for the littoral drift of the Long Island and
New Jersey coasts. Within the past century,
much of the deposition has occurred on
the ends of Rockaway and Sandy Hook
spits; these features have grown rapidly,
nearly closing off the harbor mouth within
historic times ( Shepard and Wanless 1971 ) .
However, it appears that much sand has
bypassed the spits; a complex system of
sand banks separated by interdigitating ebb
and flood channels lies between them ( Fig.
18). A profile of velocity residual to the
semidiurnal tidal cycle gi-ves some indica-
tion of the flow structure responsible for
74°00'
73°55'
Fig. 18. Bathmetry of the New York Harbor
mouth, from a 1973 NOAA/AOML survey. Depth
in meters. Dashed line indicates profile of Fig. 19.
619
New York-New Jersey shelf
87
bank-channel topography (Fig. 19). The
characteristic estuarine two-layer flow is
present as indicated schematically in Fig.
19B. The less saline upper water has a re-
sidual seaward flow, and the more saline
lower water has a residual landward flow.
As a consequence of the Coriolis effect, the
interface is tilted so that the east side of the
harbor mouth is flood dominated while the
upper level of the west side is ebb domi-
nated. The distribution of isovels in Fig.
19A suggests that this basic pattern has been
modified by the frictional retardation of
the tidal wave in the shallow estuary and
the resulting phase lag (Swift and Ludwick
in press). Because of retardation, there is
a brief period during the tidal cycle when
the estuary tide is still ebbing through the
central channel while the shelf tide has al-
ready turned and is flooding on either side
of the ebb tidal jet. This flow pattern, in-
tegrated over the tidal cycle, results in
greater ebb than flood discharge in the
central channel (ebb dominance) and
greater flood than ebb discharge in the
marginal zones (flood dominance; Fig.
19C). It is probably because of this lag-
induced flow interpenetration that the
Sandy Hook Channel is not completely ebb
dominated as required by the two-layer,
SANDY HOOK
0
o-k
KILOMETERS
2 4
■ } i.
ROCKAWAY
8
Fig. 19. A — Profile across the Hudson estuary mouth (mouth of New York Harbor), contoured for
velocity residual to the semidiurnal cycle. Pattern is interpreted as a resultant response to component
patterns shown in B and C. B — Schematic diagram of two-layered, density-driven estuary flow. C —
Schematic diagram of pattern resulting from phase lag of the tidal wave. (Modified from data of Kao
1975; reprinted from Duedall et al. in press by permission of Estuarine and Coastal Marine Science.)
620
88
Geological processes
estuarine component of flow but is flood
dominated near the channel floor ( Fig.
19A). The two sand ridges that separate the
three channels are presumably built by this
pattern of flow dominance. Residual flow
on the opposite sides of a given sand ridge
will have the opposite sense; each ridge is
therefore a sand circulation cell or closed
loop in the sand transport pattern.
Here perhaps are the ultimate sinks in
the littoral sand transport pattern of the
New York Bight. Efficient maintenance of
the dredged shipping channels demands
verification of this inferred pattern of flow
dominance and careful analysis of the re-
sulting sand budget.
References
Bagnold, R. A. 1963. Mechanism of marine
sedimentation, p. 507-528. In M. N. Hill
[ed.], The sea, v. 3. Interscience.
Baylor, E. R. 1973. Final report of the oceano-
graphic and biological study for Southwest
Sewer District No. 3, Suffolk County, New
York. Rep. for Bowe Walsh Assoc, by Mar.
Sci. Res. Center. SUNY, Stony Brook.
Beardsley, R. C, and B. Butman. 1974. Circu-
lation on the New England continental shelf:
Response to strong winter storms. Geophys.
Res. Lett. 1: 181-184.
Boicourt, W. C, and P. W. Hacker. 1976. Cir-
culation on the Atlantic continental shelf of
the United States, Cape May to Cape Hatteras.
Mem. Soc. R. Sci. Liege Ser. 6 10: 187-200.
Bowen, A. J. 1969. The generation of longshore
currents on a plane beach. J. Mar. Res. 27:
206-215.
Bruun, P. 1962. Sealevel rise as a cause of shore
erosion. J. Waterways Harbors Div. Am. Soc.
Civ. Eng. Proc. 88: 117-130.
DeAlteris, J. T., J. R. Roney, L. E. Stahl, and
C. Carr. In press. A sediment transport
study, offshore New Jersey. Am. Soc. Civ.
Eng. Spec. Conf., Jun 1975.
Donahue, J. C, R. C. Allen, and B. C. Heezen.
1966. Sediment size distribution profile on
the continental shelf off New Jersey. Sedi-
mentology 7: 155-159.
Duane, D. B., M. E. Field, E. P. Miesberger,
D. J. Swift, and S. J. Williams. 1972. Lin-
ear shoals on the Atlantic continental shelf,
Florida to Long Island, p. 447-498. In D. J.
Swift et al. [eds.], Shelf sediment transport:
Process and pattern. Dowden, Hutchinson &
Ross.
Duedall, I., H. B. O'Connors, J. B. Parker,
R. Wilson, W. Miloski, and G. Hulse. In
press. The abundance, distribution, and flux
of nutrients and chlorophyll a in the New York
Bight apex. Estuarine Coastal Mar. Sci.
EG&G Environmental Consultants. 1975.
Forecasting power plant effects on the coastal
zone. EG&G, Waltham, Mass. 335 p.
Emery, K. O., and E. Uchupi. 1972. Western
North Atlantic Ocean: Topography, rocks,
structure, water, life, and sediments. Am.
Assoc. Pet. Geol. Mem. 17.
Fischer, A. G. 1961. Stratigraphic record of
transgressing seas in the light of sedimentation
on the Atlantic coast of New Jersey. Am.
Assoc. Pet. Geol. Bull. 45: 1656-1660.
Graf, W. H. 1971. Hydraulics of sediment
transport. McGraw-Hill.
Harris, R. L. 1954. Restudy of test shore
nourishment by offshore deposition of sand,
Long Branch, New Jersey. Beach Erosion
Board Tech. Memo. 62.
Kao, A. 1975. A study of the current structure
in the Sandy Hook-Rockaway Point transect.
M.S. Res. Pap., Mar. Sci. Res. Center. SUNY,
Stony Brook.
Kennedy, J. F. 1969. The formation of sediment
ripples, dunes, and antidunes. Annu. Rev.
Fluid Mech. 1: 147-168.
Kim, J. H., and W. S. Gardner. 1974. Geo-
marine investigation. Final design study waste-
water ocean outfalls. Ocean Co. Sewerage
Authority, Ocean Co., N.J.
Lavelle, J. W., and others. In press. Prelimi-
nary results of coincident current meter and
sediment transport observations for wintertime
conditions on the Long Island inner shelf.
Geophys. Res. Lett.
Ludwick, J. C. 1972. Migration of tidal sand-
waves in Chesapeake Bay entrance, p. 377-
410. In D. J. Swift et al. [eds.], Shelf sedi-
ment transport: Process and pattern. Dowden,
Hutchinson & Ross.
. 1974. Tidal currents and zig-zag sand
shoals in a wide estuary entrance. Geol. Soc.
Am. Bull. 85: 717-726.
In press. Tidal currents, sediment trans-
port, and sandbanks in Chesapeake Bay en-
trance, Virginia. Proc. Int. Estuarine Conf.,
2nd, Myrtle Beach, S.C., Oct 1973.
Murray, S. P. 1975. Trajectories and speeds of
wind-driven currents near the coast. J. Phys.
Oceanogr. 5: 347-360.
Sanders, J. E., and N. Kumar. 1975. Evidence
of shoreface retreat and in-place "drowning"
during Holocene submergence of barriers,
shelf off Fire Island, New York. Geol. Soc.
Am. Bull. 86: 65-76.
Shepard, F. P., and H. R. Wanless. 1971. Our
changing shorelines. McGraw-Hill.
Sheridan, R. E., C. E. Dill, Jr., and J. C. Kraft.
1974. Holocene sedimentary environment of
the Atlantic inner shelf off Delaware. Geol.
Soc. Am. Bull. 85: 1319-1328.
Sonu, C. J., and J. L. Van Beek. 1971. System-
621
New York-New Jersey shelf
atic beach changes on the Outer Banks, North
Carolina. J. GeoK 79: 416-425.
Stahl, L., J. Koczan, AND D. J. Swift. 1974.
Anatomy of a shoreface-connected sand ridge
on the New Jersey shelf: Implications for the
genesis of the shelf surficial sand sheet.
Geology 2: 117-120.
Sternberg, R. W. 1972. Predicting initial mo-
tion and bedload transport of sediment in the
marine environment, p. 61-82. In D. J. Swift
et al. [eds.], Shelf sediment transport: Process
and pattern. Dowden, Hutchinson & Ross.
Stubblefield, W. L., J. W. Lavelle, T. F.
McKinney, and D. J. Swift. 1974. Sedi-
ment response to the present hydraulic re-
gime on the centrar New Jersey shelf. J. Sedi-
ment. Petrol. 45: 337-358.
Swift, D. J. 1973. Delaware Shelf Valley: Es-
tuary retreat path, not drowned river valley.
Geol. Soc. Am. Bull. 84: 2743-2748.
, D. B. Duane, and T. F. McKinney.
1974. Ridge and swale topography of the
Middle Atlantic Bight: Secular response to
Holocene hydraulic regime. Mar. Geol. 15:
227-247.
— , J. W. Kofoed, F. P. Saulsbury, and P.
Sears. 1972. Holocene evolution of the
shelf surface, central and southern shelf of
North America, p. 499-574. In D. J. Swift
et al. [eds.], Shelf sediment transport: Process
and pattern. Dowden, Hutchinson & Ross.
, and J. C. Ludwick. In press. Substrate
response to hydraulic process: Grain size
frequency distributions and bedforms. In
D. J. Stanley and D. J. Swift [eds.], Marine
sediment transport and environmental man-
agement. Wiley-Interscience.
— , and P. Sears. 1974. Estuarine and lit-
toral depositional patterns in the surficial sand
sheet, central and southern Atlantic shelf of
North America. Inst. Geol. Bassin Aquitaine
Mem. 7, p. 171-189.
Veatch, A. G., and P. A. Smith. 1939. Atlantic
submarine valleys of the United States and the
Congo submarine valley. Geol. Soc. Am.
Spec. Pap. 7. 101 p.
622
* U.S. Government Printing Office: 1977-777-045/1249 Region 8
\»
PENN STATE UNIVERSITY LIBRARIES
IIIIIIIIIIIIIIIIIIIIIM
A00007afl357El