U.S. DEPARTMENT OF COMMERCE/Environmental Science Services Administration

COLLECTED

REPRINTS

ESSA INSTITUTE FOR OCEANOGRAPHY

-<tAENTp^

Sc'tHCE S£R^V

ATLANTIC-PACIFIC
OCEANOGRAPHIC LABORATORIES

1967

U.S. DEPARTMENT OF COMMERCE
Maurice H. Stems, Secretary

ENVIRONMENTAL SCIENCE SERVICES ADMINISTRATION

Robert M. White, Administrator
RESEARCH LABORATORIES
George S. Benton, Director

Collected Reprints
Essa Institute for Oceanography

1967

Atlantic Oceanographic Laboratories
Miami, Florida

Pacific Oceanographic Laboratories
Seattle, Washington

For sale by the Superintendent of Documents, U.S. Government Printing Office, Washington, D.C. 20402

Price $3.00

FOREWORD

The effectiveness of any research organization is indi-
cated not by the size of its budget, nor by the size and
impressiveness of its building, not even by the number of
Ph.D. scientists it includes on its staff. Rather a research
organization should be judged on the contribution it makes to
the growing fund of human knowledge. One indication of this
is the published papers of its scientists.

New knowledge, however, is of relatively little use
unless it is disseminated so those that can benefit from it
know it exists. To these ends the various published papers
of the members of the Institute for Oceanography of the
Environmental Science Services Administration, U. S.
Department of Commerce, have been compiled into this
volume for the year 1967.

In November of 1967, the ESSA Institute for Ocean-
ography, became two separate activities, the Atlantic Ocean-
ographic Laboratories (Miami, Florida) and the Pacific
Oceanographic Laboratories (Seattle, Washington). Collected
reprints for 1968 will reflect this change.

Harris B. Stewart, Jr.

Director

ESSA Institute for Oceanography

11

TABLE OF CONTENTS
METEOROLOGICAL OCEANOGRAPHY

1. Harris, D. Lee,

(a) A critical survey of the storm surge protection problem,

(b) The drag coefficient between wind and water,

Proc. Eleventh Pacific Science Congress in Vol. 2, Oceanography,
Tokyo.

2. Jelsnianski, Chester P. , Numerical computations of storm surges

with bottom stress, Monthly Weather Rev. _9_b, No. 11, 740-75b.

3. Kuhn, Peter M. , and James D. McFadden, Atmospheric water vapor

profiles derived from remote-sensing radiometer measurements,
Monthly Weather Rev. _9_5, No. 8, 5bb-bb8.

4. Lettau, Bernhard, Thermally and frictionally produced wind shear

in the planetary boundary layer at Little America, Antarctica,
Monthly Weather Rev. _9_b, No. 9, b2Y-b34.

5. McFadden, James D. , Sea- surface temperatures in the wake of

Hurricane Betsy (19bb), Monthly Weather Rev. _9_b, No. b,
299-302.

b. McFadden, James D. , and John W. Wilkerson, Compatibility of
aircraft and shipborne instruments used in Air-Sea Interaction
Research, Monthly Weather Rev. 9j>, No. 12, 9ib-94l.

7. McFadden, James D. , and Robert A. Ragotzkie, Climatological

significance of albedo in Central Canada, J. Geophys. Res. 72,
No. 4, Il3b-ll43.

MARINE GEOLOGY AND GEOPHYSICS

8. Anikouchine, William A. , Dissolved chemical substances in com-

pacting marine sediments, J. Geophys. Res. J72, No. 2,
505-509.

9. Anikouchine, William A. , and Hsing-Yi Ling, Evidence of turbidite

accumulation in trenches in the Indo-Pacific region, Marine
Geol. 5, 141-154.

in

10. Burns, Robert E. , and Paul J. Grim, Heat flow in the Pacific off

Central California, J. Geophys. Res T2, No. 24, 6239-6247.

11. Dietz, Roberts., Astroblemes, McGraw-Hill Yearbook of Science

and Technology, 109-111 (McGraw-Hill Book Co. , Inc., New
York, N.Y.).

12. Dietz, Robert S. , Craters and legends, Medical Opinion and Rev. _3,

No. 1, 50-57.

13. Dietz, Robert S. , More about continental drift, Sea Frontiers 1 3 ,

No. 2, 66-82.

14. Dietz, Robert S. , Passive continents, spreading sea floors and

continental rises; A reply, Am, J. Sci. 265, 231-237.

15. Dietz, Robert S. , Shatter cone orientation at Gosses Bluff astro-

bleme, Nature 2l6_, 1082-1084, Dec. 1967.

16. Harbison, R. N. , De Soto Canyon reveals salt trends, Oil and Gas

J. 6_5, No. 8, 124-128.

17. Harrison, W. , Environmental effects of dredging and soil deposition,

Proc. World Dredging Conf. , 535-539.

18. Harrison, W. , and A. M. Richardson, Jr. , Plate-load tests on

sandy marine sediments, Lower Chesapeake Bay, Marine
Geotechnique, 274-290 (Univ. of Illinois Press, Urbana, 111.).

19. Keller, G. H. Pleistocene-recent boundary in the Malacca Strait,

Southeast Asia, Proc. Seventh Internatl. Sedimentological
Congress.

20. Keller, George H. , and Adrian F. Richards, Sediments of the

Malacca Strait, Southwest Asia, J. Sed. Pet. _37, No. 1, 102-127.

21. Ling, Hsing-Yi, and W. A. Anikouchine, Some spumellarian radio-

laria from Java, Philippines, and Mariana trenches, J. Paleon-
tology 41_, No. 6, 1481-1491.

22. Malloy, R. J. , Vertical crustal movement associated with the 1964

Alaskan earthquake, Proc. Eleventh Pacific Science Congress
in Vol. 2 Oceanography, Tokyo.

IV

2 3. Perry, R. B. , and Haven Nichols, Submarine geology of the Aleutian
Arc, Alaska, Proc. Eleventh Pacific Science Congress in Vol. 2
Oceanography, Tokyo.

24. Peter, G. , D. Elvers, and O. Dewald, Results from a geophysical
survey in the North-East Pacific Ocean, Proc. Eleventh Pacific
Science Congress in Vol. 2 Oceanography, Tokyo.

25. Sokolowski, T. J., andG. R. Miller, Automated epicenter locations

from a quadripartite array, Bull. Seismol. Soc. Am. _57, No. 2,
269-275.

26. von Huene, Roland, Richard J. Malloy, George G. Shor, Jr. , and

Pierre St-Amand., Geological structures in the aftershock region
of the 19&4 Alaskan earthquake, J. Geophys. Res. _72, No. 14,
3649-3660.

27. "Weeks, L. A., R. N. Harbison, andG. Peter, Island arc system

in Andaman Sea, Am. Assoc. Petrol. Geol. Bull. 51_, No. 9,
1803-1315.

PHYSICAL OCEANOGRAPHY

28. Butler, J. P. , Double-humped waves on a sloping beach, HIG-67-16,

ESSA/JTRE-2 - Univ. of Hawaii.

29. Chew, Frank, On the cross-stream variation of the k-factor for

geomagnetic electrokinetograph data from the Florida current
off Miami, Limn, and Ocean. 12_, No. 1, 73-78.

30. Gassaway, John D. , New method for Boron determination in sea

water and some preliminary results, Internatl. J. Ocean, and
Limn. !_, No. 2, 85-90.

31. Munk, W. H. , and B. D. Zetler, Deep-sea tides: A program,

Sci. 158_, No. 3003, 884-886.

32. Nelson, Raymond M. , Sensing ocean currents from space: Ocean

Industry 2, No. 5, 40-42.

33. Seelinger, P. E. , R. A. Wallston, B. H. Erickson, J. E.

Master son, and W. E. Hoehne, An oceanographic data collection
system, Trans. Second Internatl. Buoy Technol. Symp. , Marine
Technol. Soc, Washington, D. C.

34. Vitousek, M. J., and G. R. Miller, Low-frequency wave study in

the Meso-deep ocean, HIG-67-11, ESSA/JTRE - Univ. of Hawaii.

35. Zetler, Bernard D. , Tides and other long period waves, U.S.Natl.

Rept. 1963-1967 to 14th Gen. Assembly, IUGG, Trans. A.G.U. ,
591-595.

36. Zetler, B. D. , and G. W. Lennon, Some comparative tests of tidal

analytical processes, Internatl. Hydro. Rev. _44, No. 1, 139-147.

37. Zetler, B. D. , and R. A. Cummings, A harmonic method for pre-

dicting shallow water tides, Proc. Eleventh Pacific Science
Congress in Vol. 2, Oceanography, Tokyo.

38. Zetler, B. D. , and R. A. Cummings, A harmonic method for pre-

dicting shallow water tides, J. Marine Res. _2_5, No. 1, 10 3-114.

GENERAL

39. Schuldt, M. D. , C. E. Cook, and B. W. Hale, Photogrammetry

applied to photography at the bottom, Part of Deep Sea Photo-
graphy, ed. J. B. Her sey, (Johns Hopkins Univ. Press,
Baltimore, Md.).

40. Stewart, Harris B. , Jr. , Seafloor Geology 1968 Yearbook Directory,

Oceanol. Internatl., June, 30-31.

41. Stewart, Harris B. , Jr., Killer at the seashore, ESSA World_2,

No. 3, 30-31, U. S. Dept. of Commerce, Rockville, Md.

42. Stewart, Harris B. , Jr., Sea science ships, The Miamian,

Sept. 1967, p. 27.

PUBLICATIONS NOT INCLUDED

GENERAL

Stewart, Harris B. , Jr. , True-life adventures: dive to a mountain-
top, Summer Senior Weekly Reader 6_, No. 5, 6.

Stewart, Harris B. , Jr., Underwater worlds of tomorrow, Presidents
Assoc. , Inc.

VI

lA

Reprinted from PROCEEDINGS THE ELEVENTH PACIFIC SCIENCE CONGRESS
Volume 2 Oceanography, Tokyo, 1966

-13-
A CRITICAL SURVEY OP THE STORM SURGE PROTECTION PROBLEM

D. Lee Harris Institute for Oceanography Environmental Science Services
Administration, United States of America

The goal of a disaster warning service is to provide the maximum
protection to the public with a minimum of inconvenience. It is necessary
to consider the uncertainties in both basic data and predictions in decid-
ing on the level of protection required. For storm surge protection,
these uncertainties include those related to the movement and development
of the storm, the relation between wind velocity and wind stress, and
those arising from the uncertainties resulting from approximations and
assumptions made in the computation of the response of the sea to
atmospheric stresses.

These uncertainties will be reviewed both from the standpoint of
identifying areas in which more basic research is needed and from the
standpoint of showing how these uncertainties should affect the operation-
al natural disaster warning service.

Reprinted from PROCEEDINGS THE ELEVENTH PACIFIC SCIENCE CONGRESS
Volume 2, Oceanography, Toyko 1966

-49-

THE DRAG COEFFICIENT BETWEEN WIND AND WATER

D. Lee Harris. Institute for Oceanography, Environmental Science Services Administra-
tion, U. S. A.

Two of the most popular techniques for determining the stress between wind and
water are to record the slope of a water surface exposed to the wind and to analyze
the profile of the wind speed above the water according to the Prandtl boundary layer
theory or some extension of it which allows for the effects of buoyancy when the lapse
rate is not adiabatic.

Determinations of the drag coefficient by the first technique are characteristi-
cally larger than those determined by the second. It will be shown that second order
phenomena normally overlooked in the application of these techniques act to produce
an overestimate of the stress coefficients derived by the first technique and to
produce an underestimate of the stress coefficient derived by the second technique.

740

Reprinted from MONTHLY WEATHER REVIEW Vol. 95 No. 11

Vol. 95, No. 11

NUMERICAL COMPUTATIONS OF STORM SURGES WITH BOTTOM STRESS

CHESTER P. JELESNIANSKI

Institute for Oceanography, ESSA, Silver Spring, Md.

ABSTRACT

A linear form of the transport equations of motion is used to compute numerically storm surges generated by
model tropical storms traveling across model basins. The storms move in any fixed direction and speed relative to
a straight line coast and have a restricted number of physical parameters to fix their strength and size. These param-
eters are readily available in most weather stations.

A dissipating mechanism, introduced by Platzman, using only an eddy viscosity coefficient is modified to include
a bottom slip current by means of a bottom slip coefficient. These two coefficients are used to control the amplitude
of resurgences on the sea following the passage of tropical storms. Numerical values for the coefficients are empirically
determined by comparing computed and observed resurgences off Atlantic City.

Nomograms prepared from the computations may have some skill in forecasting future storm surges.

1. INTRODUCTION

The storm surge prediction problem is concerned with
I he rise of coastal waters brought about by meteorological
storms. The rising waters not only inundate coastal areas
but also act as a pathway for short surface or wind
waves to move and break farther inland. It is the purpose
of this paper to provide some further insight into the
mechanics and prediction of storm surges.

The response of the sea, from driving forces generated
by a moving tropical storm, is of such complexity that
practical results are obtained only through bold assump-
tions and empirical tests using numerical computations;
an electronic computer, therefore, is viewed as a laboratory
to compute storm surges using model storms traveling
across model basins. The entire response of the sea,
however, is much too general for storm surge computations
and only portions of the response are considered.

In the natural oceans there is a basic flow composed
of the general circulation, varying seasonally, and the
daily astronomical tide. The present state of knowledge
and data acquisition for hurricane conditions on the
open coast does not permit a direct incorporation of the
basic flow into the storm surge computations, nor provide
the ability to consider nonlinear interactions with storms.
For this reason, and as a great matheinatic convenience,
only linearized forms of the equations of motion are used
in the present study.

The basic flow can be partially accounted for in the
computations by appending the predicted astronomical
tide and the observed, extrapolated, or predicted seasonal
variations of the sea surface to the computed storm surge
via the superposition principle (Harris [3]). This is feasible
il the effects of nonlinear interactions are small; in any
case these corrections can be applied only at shore stations
where data are available and not in the open sea.

Model tropical storms have been used by Jelesnianski
[6] to compute storm surges but without considering
bottom stress in the storm surge equations of motion.
The computed surges were found to be reasonable for fast
moving storms making landfall but had serious deficiencies
for storms moving slowly or traveling parallel to the coast
at any speed. Computations therefore were restricted to
storms traveling at moderate or higher speeds and with
direction of travel at not too acute a crossing angle to the
coast. For convenience, storms moving from land to sea
were omitted even though the computed surges were
reasonable.

A detailed description is given in this paper to surges
generated by storm travel inadmissible in the previous
paper [6]. These particular surges are complicated in
space and time. The techniques developed in [6] to predict
storm surges using a restricted number of meteorological
parameters are extended to consider storms crossing the
coast at any angle and speed, as well as storms traveling
parallel to the coast at any speed and distance from the
coast. To consider this broad spectrum of storm velocity
relative to a coast, methods of applying bottom stress in
the numerical computations are necessary. The methods
used are useful palliatives in the absence of a sound theory
for bottom stress and dissipating mechanisms.

The addition of a bottom stress in the equations of
motion does not significantly change the results of [6] but
does have a commanding effect with storm travel inad-
missible in [6]. Storms traveling parallel to the coast at
any speed, or landfalling at slow speeds, form second
order surge oscillations due to initialization effects and
special wave phenomena, all of significant amplitude;
these are superimposed on the generated surge and can
be controlled by a dissipating mechanism.

Test computations show that certain portions of the
coastal surge profile are almost unaffected when using any

November 1967

Chester P. Jelesnianski

741

bottom friction law, including a no-friction law if the storm
is not moving too slowly. For an observer on sea, facing
land, and watching a storm landf ailing, the coastal pro-
file and peak surge to the right of landfall are not greatly
affected; on the other hand, the profile to the left of land-
fall is sensitive to the type of bottom stress law used.

2. EQUATIONS OF MOTION FOR STORM SURGE
COMPUTATIONS

The model in this study corresponds to that of the
previous report [6], except for the addition of bottom
stress, and consists of an analytically described storm
traveling across a rectangular shaped, variable depth"
basin that is open to the sea on three sides. Initially the
sea in the basin is assumed at rest, and the storm is al-
lowed to grow to maturity from zero strength in a rapid
but continuous manner.

In storm surge computations, we are primarily in-
terested in the height of the sea surface and only casually
in the current field. It is convenient then to transform the
equations of motion to two-dimensional transport fields.
This transformation, however, presents serious problems
with bottom stress.

For future use we shall need a continuity transport
equation which can be written as (Welander [20]) :

dh

where

a:

(U, 10= (w, v)dz', i.e., transport components

ft = storm surge (height of mean sea surface
above equilibrium level)
u, v = horizontal components of current field
7}=depth of the sea
x, ?y, z' — right hand coordinate system [z' in antic-
ipation of scaling) .

The momentum equations of motion (not yet in trans-
port form) with hydrostatic approximation can be written
in linear and complex form (Welander [20]) as:

where

w=u-\-iv, q-

-[

d

a(ft-fto)

dx

(2)

+i

d(ft-fto)'

p = vertical kinematic eddy viscosity
/=Coriolis parameter (constant)
<y = gravity

ft0=inverse barometer effect (hydrostatic height due
to surface pressure).

The last equation considers inertio-gravitational waves
since the Coriolis parameter is not varied. This was
purposely done for the storm snree is small compared to

the scale length of planetary waves. For similar reasons
a map scale factor is not considered. Lateral stresses are
excluded since the vertical stress influenced by the surface
wind is believed to be much larger over most of the area
of interest.

To formulate transport fields, one may directly in-
tegrate (2) in the vertical to obtain

= Q-iJW+T-Tt

(3)

wher.

Ts> Tb=V

dw
ds7

complex form of surface, bottom stress

(4)

W= complex form for transports
Q=Dq.

The surface stress can be formulated as a function of
the wind, but the bottom stress depends on the vertical
gradient of the bottom current. Since only transport
terms are available if (3) is used directly, it has been cus-
tomary to assume the bottom stress as a simple quadratic
function of transport in conformity with experiments from
pipe or channel flow ; corrections to such an empirical law
for a system under the influence of a surface wind stress
has been given by Reid [16]. This type of bottom stress
will not be considered since computational experiments
gave results that were not always satisfactory.

Other systems of representing bottom stress, which are
linear in nature, have been designed by Nomitsu [91, [10],
[11], [12]; Nomitsu and Takegami [13], [14]; and Platzman
[15]. Platzman's scheme is more convenient for nu-
merical computations. In what follows, we will adhere to
the notation given by Platzman whenever possible.

Let the surface boundary condition be v(dw/dz')\z-=o = R,
where 7? is the complex form of the surface wind stress,
taken as

R=^\V.

W

where Fs = complex wind, pa, p = air, water density,
C is assumed to be a constant drag coefficient, and
Cpjp — 3 X 10~6. We formulate the bottom boundary
condition as

dw
b~7

= SW\-' = -D

(5)

where s is a slip coefficient; here we are assuming a
"gliding" current above a very thin boundary or skin
layer, where for practical purposes the depth of the skin
layer is taken as zero.

If only one friction parameter consisting of an eddy
viscosity coefficient is used, then computations show that
the storm surge is somewhat sensitive to small changes of
the parameter. The introduction of a slip coefficient as a
second friction parameter greatly reduces this sensitivity

742

MONTHLY WEATHER REVIEW

Vol. 95, No. 11

and also gives more freedom when working with dependent
data to better fit computed and observed surge profiles.
It is convenient to make the vertical coordinate non-
dimensional by the transformation z=z'/D. If the time
derivative in (2) is treated as an operator, and the resulting
second order differential equation with variable z is solved
with surface and bottom boundary conditions, then

i-, sinh az „ . cosh crz , , „ , . 1 n

wD= R-\ ^-j — (cosh o7f— sw>_iH -2 V

r)„v T],a sum a y\va

where

v>=v/D2

''=«=' 0+1)

a-R+ (sinh a)Q

(6)

W_,=

»[

y\v sinh a-\-y;(j cosh

']

= complex bottom current.

If (6) is now integrated in the vertical (with respect to
z from — 1 to 0) , the result is

where

7l,WJrG{«)]M=QMl + H{a)]R

(7)

M= complex transport

(equivalent to dimensionalized W in (3))

G(*y-

H(<r) =

[{g}+,coth,-l]

1-

sinh a

m

+<x coth— 1

For the purposes of integration with respect to z, and
of algebraic manipulation, the operators, a, 67(a), a2, H(<r),
etc., can be treated as ordinary algebraic quantities and
parameters, save that the operator must remain to the
left of some operand, such as M, Q, or R in equation (7).

The meaning of a compound operator such as G(u),
H(<r), etc. is based on power series expansions in a; e.g.,

a(sinh „)F(t) = [c2+y+~+ . . . ]F(t)

(where the terms of the power series are just the same as
those of a function z sinh z) provided the series converges.
When it does not converge directly, means similar to
analytic continuation can be employed to get the result of
the operator; these means lead to a unique result.

Equation (7) was formed to set Q by itself; it is just as
easy to set <j2M by itself if this is desired. The choice of
which term is set by itself can be governed by the nature
of the numerical scheme to be used.

Note that for s = 0, G and H are zero; (7) is then no
more than (3) without bottom stress, i.e., frictionless flow.
Platzman [15] treated exclusively the special case, s= °o ,
or w-i=0, i.e., no bottom slip. This is equivalent to as-
suming that the horizontal velocity gradually approaches
zero. near the bottom. However, the horizontal velocity
near the bottom is often quite large (excepting a thin
boundary layer which is not included in the present
analysis). In order to recognize the existence of this
boundary layer without being concerned with its detailed
structure we have taken bottom stress (5) as proportional
to a slip velocity assumed to glide over the top of the
bottom boundary layer. Our results with idealized storms
appear to point out that v controls the peak surge on the
coast, whereas s controls the dispersion of the surge es-
pecially to the left of the storm center on the coast.

The difficulty in applying equation (7) lies in represent-
ing the operators 67(a) and H(a) numerically. The func-
tions M, R, and their first derivatives in time can be
readily approximated directly from available information,
but higher derivatives resulting from a Taylor's expansion
of 67(a) and H(a) are more difficult to obtain.

Accordingly, Platzman suggests that G and H be ap-
proximated by truncating their Taylor's expansion about
cr02=ifD2/p to obtain

6?(<r)~670(<ro)+y Qfa) ^ H(a)~H0(c0)+^ Hx{ca) ^ (8)

where the subscripts represent the zeroth and first deriv-
atives with respect to a2, and the derivatives are evaluated
at a2=o02. Using the approximation (8), equation (7) takes
on the simpler form

dM
bt

where

^BQ-ifAM+^C+i^R (9)

1 + 67, '

B

1

1 + 67,

^ 1 + H0 T

c==y+g;' j

GqHi

1 + 67,'

Numerical tests using (9) gave good results for slow-
moving storms, but spurious waves formed, especially
along the storm's track, with fast -moving storms. When
the J term was dropped the spurious waves did not occur.
This situation prompted a closer look at the truncated
forms of (8) to see whether they are sufficiently represent-
ative for storm surge computations, and whether the J
term is important or not. (See Appendix I.)

Equation (9) with the J term omitted has the real and
imaginary parts

W
dt~

dV
dtr

■gD

[*

d(h-h0) B d(h-h0)

—B,

-gD^Bt

dx by J

d(h—h0) B d(h—h0)l

dy ~+Bi dx
-f[ArU-A(V]+C^rs+0^rsJ

y (io)

November 1967

Chester P. Jelesnianski

743

DEPTH [FEETI
0 83 I66 248 33I 4I4 497 580 662 0 83 I66 248 33I 4I4 497 580 662

-I 1 1 1 1 1 1-

Figtjre 1. — Real and imaginary parts of the four coefficients
(A, B, C, J) of equations (9) and (17) as functions of Ekman
number "t" or depth.

These equations involve only first derivatives with respect
to time. The six subscripted functions (A, B, C) are
dependent only on depth when eddy viscosity and bottom
slip coefficients are specified; their form is given in figure 1 .

The numerical scheme for (10) used in this study is
given in Appendix II. An heuristic approach to form
values for the eddy viscosity coefficient v, and slip
coefficient s, is given in Appendix III.

In this study, the parameters describing the model
storms and basins have a range usually less than an order
(if magnitude. Thus, in dealing with the drag coefficient
of the surface wind stress as well as eddy viscosity and
slip coefficients, we have tacitly assumed constant values
as sufficiently serviceable for the range of parameters in
this report. This means that the results of the computa-
tions are restricted mainly to tropical storms.

3. GEOGRAPHICAL ORIENTATION, STORM
PARAMETERS, DEPTH PROFILES, DEFINITIONS

For purposes of orientation in the following sections,
the observer will always be at sea and facing the coast,
The coast to his right will be considered relative north,
to his left relative south. Crossing angles of the storm's
path to this orientated coast will be described in me-
teorological sense; thus a storm on the coast moving from
relative north lias a crossing angle of 0°, moving normal
to the coast from sea, a crossing angle of 90°, etc.

There are five simple parameters to describe the
strength, size, and motion of a model storm; these in turn

0 5 10 15 20 25 30 35 40 45 50 55 60

RADIUS OF MAX WINDS (MILES)

Figure 2. — A nomogram relating three model storm parameters,
stationary storm maximum wind, radius of maximum wind in
statute miles, and pressure drop. The inflow angle occurs 100 mi.
from the storm center. The storm center is at latitude 30°.

determine the driving forces of surge generation, the
pressure gradient, and wind stress. The parameters are:

(1) Latitude. — Normally the latitude of the storm's
landfall; if the storm does not landfall, the latitude of a
point of interest on the coast, The storm surge is only
mildly sensitive to this parameter and varies by less than
10 percent between latitudes 15° and 45°, all other
parameters being the same. For this reason and because
we are interested in transient effects as opposed to general
circulation, latitude is not varied in the equations of
motion.

(2) Radius oj Maximum Winds. — The distance from
the storm center to the maximum wind of the storm.
This distance is not dependent on storm motion, and for
any given time it is assumed to be the same in all direc-
tions. This parameter controls the horizontal extent of
the surge on the coast. If only the value of the peak surge
on the coast is desired then the accuracy of this parameter
becomes unimportant, and for most purposes a rough
estimate of this distance is sufficient.

(3) Pressure Drop oj the Storm. — The pressure difference
from the center to the periphery of the storm. For an
actual storm, this could be the mean of several differences
measured along rays from the storm center to the first
anticyclonically turning isobar. This is the most im-
portant storm parameter; it controls the peak surge on
the coast. For a fixed pressure drop, the peak surge on the
coast is only weakly dependent on the radius of maximum
wind. The pressure drop is not used directly in the model
computations, instead it is used as an argument (fig. 2)

744

MONTHLY WEATHER REVIEW

Vol. 95, No. 11

to arrive at a more convenient measure for computations,
the stationary-storm-?naximum-wind. In the previous report
a simple Newton-Raphson integration method was used
to derive the nomogram; in this report the more precise
Runge-Kutta method was used. The differences are only
minor.

(4) Speed of Storm. — Rate of motion of the storm center.
With all other parameters held fixed, there is a critical
storm speed that gives the highest peak surge on the coast.

(5) Direction of Storm. — Direction of motion of the storm
center. With all other parameters held fixed, there is a
critical direction of storm motion which gives the highest
peak surge on the coast.

The computed surge depends on the depth contours of
the basin as well as the model storm. The continental
shelf of the oceans vary predominantly in one direction,
therefore a one-dimensional depth profile is used in the
model. This profile is used solely for convenience in the
present stage of developing the dynamic model. There are
no essential difficulties in the use of two-dimensional
bottom specifications when such detail is desired or when
its effect is believed to be significant relative to other
terms.

For further reference it is convenient to make the
following definitions :

Standard Storm. — A model storm having a stationary-
storm-maximum-wind of 100 m.p.h. and with storm center
at latitude 30°.

Standard Basin. — A model basin having a linear sloping
depth profile consisting of a 3-ft. drop for each mile length
along the continental shelf, a 15-ft. depth at the coastal
boundary, a shelf length of 60 mi., and a deep water open
boundary depth of 195 ft. This basin has a slightly larger
slope near the coast than the standard basin of the previ-
ous report [6] ; hence computed surges from a standard
storm in this basin are slightly smaller than those in the
previously used basin. In the numerical model, for a given
depth profile, the storm surge is only weakly dependent
on any alteration of the immediate slope at the coast of
the continental shelf; therefore a vertical wall is substituted
at the coast with finite depths at the coastal boundary.

Coastal Surge Profile. — A plot or snapshot picture of the
surge heights along the coastline for a given time.

Directly Generated Surge. — Storms traveling parallel to
the coast can generate traveling and/or standing waves
superimposed on the coastal surge profile. The first crest
and trough of the coastal surge profile associated with the
storm's center, and moving with the storm, is the di-
rectly generated surge. Figure 3 illustrates the motivation
for this definition; notice that fast moving storms travel-
ing parallel to the coast can generate traveling waves
behind the storm's track and these traveling waves am-
plify that portion of the directly generated surge behind
the storm's track, i.e., storms moving to the right amplify
the directly generated trough, storms moving to the left
amplify the directl, generated crest.

Figure 3. — Plots of computed coastal surge profiles generated by a
storm moving to the right, and moving to the left along the
coast; and then compared to a stationary storm, center of storm
remains on coast.

Resurgences. — At any point on the coast, large ampli-
tude oscillations of the surge can occur with time for
storms moving parallel to the coast. These oscillations,
excluding passage of the directly generated surge, will be
called resurgences. The resurgences can be shelf seiches
or edge waves (Munk et al. [8], Reid [17]). Shelf seiches
also occur for slowly moving storms making landfall;
this is a special type of resurgence, generally with higher
harmonics and usually of small amplitude unless the
storm is moving very slowly.

4. GROWTH TIME OF STORM, INITIALIZATION, AND
SPECIAL WAVE PHENOMENA

The growth time to maturity for fast moving storms
making landfall, and traveling at not too small a crossing
angle to the coast, was found empirically to be of trivial
concern. For storms traveling parallel or nearly parallel
to the coast at any speed, initialization phenomena de-
pendent on growth time and of significant amplitude are
generated or superimposed on the surge profile.

In order that the peak surge have only weak dependence
on initial storm placement, it was necessary initially to
place landf ailing storms at least past the continental
shelf (the deep water open boundary) or to place storms
traveling from land to sea at the mirror image point;
storms traveling parallel or at a small angle to the coast
required an initial placement based on several criteria
that will now be discussed.

Figure 4 shows the directly generated crest plotted against
time in the case of a stationary, standard storm with center
on the coast of a standard basin. This peak surge with time
acts similarly to a damped-forced oscillator. The transient
oscillations result from rapid growth to maturity of the
storm; they cannot be completely eliminated for any

November 1967

Chester P. Jelesnianski

745

reasonable growth time in the computations. Over a
wide range of growth time the phase varied but the ampli-
tude of the second crest of the transient oscillations was
nearly constant. Henceforth, we shall always use 100 min.
as a growth time solely for convenience. We adopt a
working criterion that whenever transient oscillations of
this nature occur, the second crest will be chosen as repre-
sentative of the peak surge on the coast; the same type
of transient oscillations will affect the moving directly
generated surge. Thus for a working criterion, the initial
placement of the storm center in the model basin must
be sufficiently distant from the point of interest so that a
second crest has time to form.

Figure 4 also shows the surge with time for the point
on the coast having peak surge in the case of a stationary
storm placed SO mi. from the coast. Here, there is a
continuous growth of the surge with time, i.e., it takes
time for the surge to build on the coast; notice that higher
harmonic oscillations occur for this case. Henceforth, for
a working criterion, we shall adopt the computed peak
surge 8 hr. after initialization as a representative value
for those cases where the coastal surge displays many
variations with time.

The damped-transient oscillations, superimposed on the
directly generated surge from initialization processes,
is not the only phenomenon occurring with storms travel-
ing parallel to the coast. There are also the phenomena of
standing waves, traveling waves, or a combination of
both superimposed on the surge profile behind the storm's
track; there are even further complications if the storm
is varying in strength, size, and speed with time.

As an example of these complexities, consider the
September 1944 storm which moved parallel to the Eastern
Seaboard (this storm is discussed in Appendix III; its
track is given in figure 20 and the generated surge at
Atlantic City in figure 21). Figure 5 pictures the entire
computed coastal surge profile against time. Notice that
the traveling directly generated surge associated with the
storm center has initially a large transient oscillation
thai dies out with time, and the directly generated surge
becomes smaller with time due to decreasing storm
strength with time. In this figure, the oscillations or re-
surgences witli time at Atlantic City, after passage of the
directly generated surge, are readily seen to be shelf
seiches and not traveling edge waves The computed
resurgences may have been affected by reflective prop-
erties inherent in the open boundary conditions of the
model; possibly the use of open boundaries with radiative
properties would be more appropriate. The two lateral
(relative north, south) open boundaries, with normal
transport gradient set to zero, do not strongly affect the
first few resurgences; this was determined by varying the
length of (he basin and repositioning the two lateral
boundaries (in (lie natural coast. No equivalent tests
were made for the deep water open boundary.

In t lie numerical model, traveling edge waves will pre-
dominate over shelf seiches for storms that are fast moving,

STORM CENTER ON COAST
STORM CENTER 80 MILES
SEAWARD

Figure 4. — The computed height of the peak surge against time
generated by a stationary storm.

STORM REACHES
FULL INTENSITY

Figure 5. — The computed coastal surge profile, contoured against
time, for the September 1944 storm modeled in Appendix II.
The line with arrows represents the position of the storm center
normal to the coast (center of storm 40 mi. seaward). This figure
shows the formation of shelf seiches behind the storm's center.

of constant strength, size, and speed, and for basins which
have shallower coastal depths and gentler slopes. Consider
now the above storm but with constant parametric storm
values equivalent to those observed off Atlantic City;
consider also a standard basin whose profile differs sig-
nificantly from that off Atlantic City (fig. 17). Figure 6
pictures the computed coastal surge profile for this hypo-
thetical storm and basin. Notice that the resurgences
which form with time behind the storm's track are now
traveling waves in contrast to those in figure 5. The

746

MONTHLY WEATHER REVIEW

Vol. 95, No. 11

600

500 --

400 --

en

<
o
o

300 --

200 --

I00--

0

STORM REACHES
FULL INTENSITY

0

H h

4 8

HOURS

12

Figure 6. — Same as figure 5 for a standard storm, in a standard
basin, and traveling parallel to the coast at 40 m.p.h. with center
of the storm 40 mi. seaward. This figure shows formation of
traveling edge waves on the coast.

directly generated crest does not change in value, except
for rapidly decaying initialization phenomena at the be-
ginning of the computations; the directly generated trough
is amplified by the traveling edge waves forming behind
the storm's track.

Figure 7. — Contours of distance, in statute miles, from landfall
position to point of peak surge on the coast. Radii are storm
speeds, rays are crossing angles of storm track to the coast
(standard storm in standard basin), (a) Radius of maximum
wind, 15 statute mi. (b) Radius of maximum wind, 30 statute mi.

For a working convenience, we shall henceforth consider
only storms of constant strength and size after reaching
maturity, and traveling with uniform rectilinear velocity.

5. LANDFALLING STORMS

Landfalling storms affect only a segment of the coast-
near the point of landfall ; consequently, we are interested
not only in the peak surges but also in the horizontal
extent or dispersion of the surge along the coast. The
dispersion is strongly dependent on the radius of maximum
winds. These storms, with speed and crossing angle to the
coast not too small, do not generate initialization phe-
nomena and resurgences on the surge profile; therefore the
generated surge is much less complicated than the exam-
ples of the previous section. The peak surge occurs at only
one point on the coast, it is generally larger with storms
traveling from sea to land.

As a preliminary aid to determine the surge dispersion
along the coast, we construct a pre-computed surge pro-
file generated by a standard storm, with fixed radius of
maximum winds, moving at fixed velocity across a stand-
ard basin. In what follows, we assume that the position
of landfall is known and use it as an origin. To construct
this preliminary profile, we use nomograms (figs. 7-15)
which give contours of pre-computed distances and heights
at selected points along the surge profile. These figures, in
polar coordinates, have rays as crossing angles of the
storm to the coast, and radii as storm speed. Figures
marked (a) and (b) are for storms having 15- and 30-mi.
radius of maximum winds respectively; presumably one
can interpolate for other values of the radius. These dia-
grams consider variation of three storm parameters; in a
later section we shall consider corrections to the pre-
computed profile using parameters for any particular
storm and basin.

In figures 8 and 9, we have outlined a region in broken
lines to call attention to edge wave phenomena which can
affect the directly generated crest and trough respectively.
An example of this situation is given in figure 3.

November 1967

Chester P. Jelesnianski

747

Figure 8. — Contours of peak coastal surge values, in feet. Argu-
ments are identical to figure 7. The upper region bounded by
broken lines point out edge wave phenomena that could be
affecting the directly generated crests in the model computations.

Figure 11. — Same as figure 10, but to left of landfall.

Figure 12. — Same as figure 10, but for yt peak surge value.

Figure 9. — Same as figure 8 for the minimum surge, portrayed at
time of the peak surge. The absolute minimum surge does not
necessarily occur at time of peak surge.

Figure 13. — Distance, in statute miles, from peak surge to zero
surge on the coast. Arguments same as figure 7.

Figure 10. — Contours of distance on coast, in statute miles, from
point of peak surge to point on coast having % the peak surge,
to the right of landfall. Arguments same as figure 7.

Very slowly moving and landfalling storms form shelf
seiches with phase angle depending on initial storm
placement. These seiches are superimposed on the surge;
thus, for different storm speeds, there is no a priori way
to relate phase angle of seiches to peak surge on the coast,
Therefore, it was necessary to force continuity of the
various contours about the origin of the polar graphs,

subjectively; the stationary storm at the origin, with
representative value, was used as an anchor or invariant.
The surges thus derived never differed by more than one-
half foot from actual computations. Similar circumstances
occurred for storms crossing the coast at a small angle
and at any speed.

We emphasize that the constructed profile is only for
the time of peak surge on the coast. The absolute mini-
mum surge does not necessarily occur at time of peak
surge; the negative surges on the profile are only transi-
tory and eventually can turn to respectable positive values

748

MONTHLY WEATHER REVIEW

Vol. 95, No. 11

Figure 14. — Distance, in statute miles, from peak surge to minimum
surge on the coast. Arguments same as figure 7.

_

- 14
si?

*"" "

^^

X

JT-^

//■'

^10

^O* ^

//

s

^V ~* -^

8_

UJ

/'yC'

-6£

^V

s s

y

sN. *"" ~-

^

"4

~~" ~- ^-^-^^^

-***''

-2

"""'•-— ^r^ -

__.- "' LANDWARD

5EAWARD -~~JJ^ •

50 7 5 TOO

| i 1 j

1 1

50 25

25 -" "" _

MILES ^-^ __ — -p^~ ~~

--2

/ ^" /

\^\

/ s

—4

/ / '

-6

/ <' /

\

\^.

"V

/

\

--IO

/

\

■-12

/

\

--14

/

\

--16

/

. 0 MPH

\

/ rt£l \1f*'t

\

-18/

40 MPH

^

—

—20

— POSITIVE

RESURGENCE (40 MPH)

Figure 15. — Arrival time, in minutes, of peak surge on the coast
after storm landfall. Arguments same as figure 7.

Figure 16. — Height nomogram of the directly generated crest and
trough, and the resurgence amplitude, for storms (standard storm
and standard basin) moving at different speeds parallel to the
coast. The abscissa is distance of the storm center from the coast
in statute miles.

after passage of the storm. The history of the region of
negative surges has not been documented.

The surge profile from a stationary storm can be depicted
by the nomograms of this section only if the center of the
storm rests on the coast. If the storm center is at a dis-
tance from the coast, then the heights of the crest and
trough can be extracted from figure 16 of the next section
and the surge profile completed with the nomograms of
this section. We are assuming here that the dispersion of
the surge does not depend on distance of the resting
storm center from the coast; separate computations show
this to be a good assumption.

6. STORMS NOT LANDFALUNG

For convenience, we treat all storm motions that stag-
nate, loop, recurve, or in any way fail to landfall, as
storms traveling parallel to the coast. For simplicity, we
restrict description of the surges from these storms to the
moving, directly generated surge and follow the rules set
forth in past sections to form representative surge values.

We shall not consider here the form* of the coastal

/option here since the crest and trough
could he interested in the dispersion on

•Stationary storms could be treated as an e:
remain stationary on the coast. In this care oni
the coast of the stationary surge profile.

**We do not consider travel along the left side since this is a rare occurrence limited to
lower California and to the western Floiida coast.

surge profile at fixed times, nor time of passage of crests
and troughs on points along the coast.

To construct a nomogram, for purposes of forecasting
the directly generated surge from these storms, we con-
sider a standard storm traveling parallel to the coast of a
standard basin at particular storm speeds. Let the radius
of maximum winds be 30 mi.; other values for this storm
parameter need not be considered since we are primarily
interested in peak surge and not seiches and resurgences.
We now focus attention on the directly generated crest
and trough and note that these are traveling with the
storm center along the coast. Figure 16 is a resulting
nomogram from a series of computations which give
representative values of the directly generated crests and
troughs for the storm traveling along the right** side of
the coast at various distances from the coast; above the
abscissa (miles), positive peak surges are shown, below
negative peaks. At large distances from the coast, slowly
moving storms have higher surges (the criteria of the
previous sections give the surge 8 hr. after initialization;
hence there is time for the surge to build), but near the
coast fast moving storms have higher surges.

The standard basin of this study favors traveling (edge)
waves as opposed to seiches for fast moving storms moving
parallel to the coast. One should therefore be careful in
accepting these resurgences, and the directly generated

November 1967

Chester P. Jelesnianski

749

trough, as fully representative since we have not con-
sidered storms that vary in strength, size, and speed, nor
depths varying in two dimensions, nor curvilinear coasts,
nor the effect of various depth profiles on the resurgences.
In the present model these resurgences do not become
important unless the storm is very fast moving with center
slightly seaward of the coast.

Evidence of resurgences (edge waves) does not begin
until the storm travels in excess of 20 m.p.h., and not until
the storm travels about 40 m.p.h. is there significant
amplitude. Figure 16 also shows the amplitude of re-
surgences for storms traveling at 40 m.p.h. and at time
10 hr. after initialization; notice that the maximum re-
surgence amplitude occurs slightly seaward of the coast,
Greenspan [2]. Only the first resurgence behind the
storm's track is shown, and this only after passage of the
directly generated surge since the following resurgences
are dampened.

Corrections to the pre-computed surge of figure 16, for
non-standard storms and basins, are given in the next
section.

7. CORRECTING THE PRE-COMPUTED SURGE FOR
IN-SITU STORMS AND BASINS

In the development of a practical forecasting system
for storm surges, it is desirable to modify the preliminary
constructed profiles and surges of the preceding sections
for particular storms and basins that differ from standard.
The corrections would then be for non-standard values of
stationary-storm-maximum wind, latitude, and basin
depth profile.

Correcting the pre-computed surges for different values
of the parameter, maximum wind, is very easy since the
computed surge is almost proportional to the square of the
maximum wind (as shown in [6]).

For the same pressure drop, the peak surge on the coast
is not unduly sensitive to the parameter, radius of maxi-
mum wind. This can be verified with the nomograms of
figures 2 and 8 and the above correction for the maximum
wind since the figure is used only to determine peak surges
on the coast.

Variations of the latitude parameter alters the computed
surge in only a minor way (see [6]). It was decided as an
added convenience to incorporate corrections for latitude
with corrections for depth profiles.

Figures 17 and IS give correction factors FD for special
points along the Eastern Seaboard and Gulf States of the
United States to correct the precomputed surge for in-situ
depth profiles; corrections for latitude have been incor-
porated. Presumably interpolation can be used between
the special points. The factors were obtained from com-
putations using the given depth profiles at the special
points and various storm conditions; they are somewhat
subjective since they do change with storm conditions,
but in most cases only slightly. The correction factors
are for the peak surge and would differ for other points
on the surge profile, being least reliable for the negative

portion of the surge. We assume the factor, a function
of the depth profile normal to the coast, can be used for
the pre-computed surge profile providing the storms do
not differ greatly from standard. The factors are not
invarient when comparing with other varied storm param-
eters, but they change only slightly for the parametric
range of storm values used in this study.

The depth contours of a natural basin vary in two-
dimensions, but the variation normal to the coast gen-
erally is much greater than the variation parallel to the
coast. In this study we compute only for variation of
depths in one-dimension and assume the depth correction
factors, applied selectively at selected points along the
surge profile, are good approximations for two-dimen-
sional basins.

The dispersion of the surge on the coast does not change
appreciably when varying the parameters maximum wind
and depth profile, but does change some when varying
the latitude. For convenience it will be assumed here that
the dispersion of the surge remains invariant.

There is a unique region along the Florida coast between
Miami and Palm Beach where the bottom depths descend
from the coast with extreme rapidity. The model discussed
here does not cover this case. South of Miami the depths
descend with equal vigor; however, there is a shallow
shelf along the coast in this region. It was subjectively
decided to give this shelf a length of 10 mi. to arrive at
some correction factor for the non-standard depth profile ;
the reliability of the factor in this area is questionable.

To correct the pre-computed surge heights of the
previous sections for non-standard storms and basins
along the United States coast, the following could be
used at selected points on the coast:

hc=hs(VR/l00)2FD

where hc is the corrected surge height, hs is the standard
pre-computed surge height, VR is the stationary-storm-
maximum wind parameter, and FD is the depth profile
correction factor (figs. 17 and 18).

8. SUMMARY AND CONCLUSIONS

It is possible to compute a reasonable storm surge
tide with a numerical model that uses a simple linearized
form of the equations of motion. Bottom stress in these
equations was not found to be significant in surge gener-
ation with fast moving storms making landfall, but a
dissipating mechanism was necessary to control large
amplitude resurgences and/or initialization phenomena
for storms moving parallel to the coast at any speed, as
well as slowly moving storms making landfall.

In this paper, a bottom stress formulation was chosen
with certain desirable properties in the generation of
the coastal surge profile. These properties were most
evident for the range of storm velocity not admissible
in a previous report [6]. Some of the properties were
suppression of large transports at and near the coast,

750

MONTHLY WEATHER REVIEW

Vol. 95, No. 11

JACKSONVILLE

Figure 17 '.• — Correction factors at selected points along the Eastern Seaboard of the United States for depth profiles other than standard;
the factors are used to correct pre-computed surge heights in a standard basin. The inserts are the mean depth profiles of the selected
points.

and damping of special wave phenomena and initialization
effects generated by storms traveling parallel to the coast.
Since non-dimensional analysis shows that the dissipating
term is small compared to the inertial term for rapidly
moving storms traveling across the continental shelf
(Kajiura [7]), values for the friction parameter were
chosen so that the coastal surge profile computed with
or without bottom stress was nearly identical for fast
moving storms traveling at or near normal incidence to
the coast. To demonstrate the usefulness of a bottom
stress formulation with the above properties, observed

surges generated by a fast moving storm at a tide station
that was undergoing special wave phenomena and
initialization effects were compared with computed
surges. To better fit the observed and computed values,
the scheme for bottom stress introduced by Platzman
[15] was modified to include a bottom slip current.

In [6] a proto-type prediction scheme for forecasting
storm surges was introduced. This was done with nomo-
grams that were prepared from pre-computed data using
the parameters of a standard storm and a standard basin.
In the prediction scheme, a preliminary surge profile

November 1967

Chester P. Jelesnianski

751

iS.'o

Figure 18.' — Same as figure 17 for the Gulf States and Florida.

was first constructed from the nomograms; three simple
correction factors for maximum wind, latitude, and basin
depth profile were then used to correct the preliminary
profile for in-situ storms and basins that differed from
standard. In this report the same scheme is adhered to
except that only two simple correction factors were used;
for simplification, corrections necessary for latitude and
basin depth profile were combined as one, at the expense
of limiting the predicting scheme to the Eastern and Gulf
States of the United States.

Our model does not consider curvilinear boundaries,
bays, inlets, etc. ; consequently the method of constructing
surge profiles from pre-computed nomograms in this
study are to be considered only as a preliminary guide
for field forecasting purposes.

The important parameters of the storm model are not
difficult to ascertain or forecast in weather stations
excepting the point of landfall. For storms crossing the
coast, the distance from landfall to peak surge is roughly
equivalent to the radius of maximum winds, and this
sets the horizontal scale of the entire surge profile; there-
fore a high order of accuracy in landfall prediction is
required.

The methods of this study consider a straight line
coast. Further research is desirable to consider curvilinear
boundaries in the model. With this more natural boundary
condition it should be possible in the future to prepare

in-situ surge forecasts by computer, using forecasted
storm parameters and landfall point. The forecasted land-
fall point would determine the basin to be used as well as
the depth contours of the basin.

APPENDIX I

To test the representativeness of equations (8) and
(9), we use an heuristic approach (suggested by Dr. A. D.
Taylor) that only partially resolves the problem.

Let us consider, for simplicity and illustrative purposes,
the case of no bottom slip, i.e., s= °° . Then G and H are

6(a)-

a2 tanh a
a — tanh a

; H{a)-

tanh a— a sech a
a — tanh a

(ID

These operators, regarded as analytic functions of <r,
have removable singularity at a=0, and simple poles
(i.e., zero denominator) on the imaginary axis of a at

approximately o-~±i(2n— 1)-. This implies that if (11)

is expanded in power series about a center in the a plane,
not on the imaginary axis, the series will converge in
some region about that center.

Suppose one of the operands was the function £rei<^<+*)
for some complex values E, /3, cj>. Then

dt

2£eif/K+«)=^3iEei(0«+*)

752

MONTHLY WEATHER REVIEW

Vol. 95, No. 11

so that the effect of operating with d/dt is just the effect
of multiplying by if3. An operator formed as a "function"
J (d/dt) has, on the operand Ee'^'+*), the effect of multi-
plying by /(?/?), for the operand Ee'^'+v the operator
d/dt "takes on the value" i/3.

In general, the operands will not be of the form £V!*<0'+*> ;
however, at any time t for which the operand is not
zero, there is an exponential function which fits most
closely to the operand. If the values of E, /3, <j> of the
approximating exponential do not rapidly change with
time, tl u it i> reasonable to approximate d/dt with the
value i|8.

The values of E, /3, <j> depend on the operands and the
time t for which the "evaluation of d/dt" is performed
and are different for the operands M and R. If /3 is real,
so that the opera nd is neither increasing or decreasing in
value, the values of G and H will be finite and bounded.
Figures 19 a-b give the real and imaginary parts of
G and H against a2, with a2 = iD2(j-\-fi)/v (i.e., replacing

d/dt by i/3) . These figures indicate that a linear approxima-
tion to G and H may be acceptable, provider ji is not large.

We first examine our experimental computations
to determine for which values of /3 the actual oper-
ands approximate exponentials. (If M=E W+*\ then
dM/dl=ipM, so 0=(l/iM)dM/dt where Mis i egarded as a
complex valued function of time and space.) Empirically
it was noticed that the transport field usually consists of
a train of vortices along the storm's path; in general,
these vortices do not travel or increase in strength with any
great rapidity, suggesting that for the trans] ort operand
M, the value of /3 remains small, and a linear approxima-
tion may be acceptable.

The storm model used in this study is a previously
determined analytic function, and moves with a uniform
rectilinear motion. For this case, the time derivative of
the forcing function can be written in the form

dt'

=-v..v

(12)

Figure 19. — (a) Plot of real and imaginary parts of G(<r) against
ia1. (b) Plot of real and imaginary parts of H(cr) against ia1.

where Vs is the storm velocity. The appearance of Vs
suggests that the value of /3 might be too large to admit a
linear approximation for H(a), acting on the storm stress
operand R. This may be why the linear approximation
of H acting on R in (7) gave spurious waves for fast
moving storms.

It appears fioii empirical computations tl at dropping
the J function in (9) [i.e., H(a)^H0(<T0)] gave results in
our numerical computations that could be acceptable for
the present state of art in storm surge computations.
We wish to show that this hoMs for the fast moving
storms through comparisons of computations using an
exact rather than a linear approximation of H(a). One
can determine an exact H(a) from the stoi m model itself by
first forming at any local point:

R=R0e

i(t-t0)(,a-ip)

(13)

then using (12); for uniform rectilinear storm motion,
we have,

dr=«+^=-i(Vs-V)/?. (14)

dt

Thus we can then form

D2

R

[%(f+0)+a].

(15)

H(a), as shown in (11), has poles (zero denominator) and
cannot be used directly in (7) ; there was no problem with
poles when only the linear terms derived from the ex-
pansion of H(a) given by (8) were retained. Instead, we
first reform (9) and approximate the last term as

t+ijdt

l + H(a)

G(a)—G)(a0)

1 +

(16)

D2/v[c2-al]
If H and G are truncated as in (8) we recapture the left

November 1967

Chester P. Jelesnianski

753

side. The right side is now free of poles; its value is de-
termined at each grid point when using (15). Several tests
were made with (9) where in one case the J function was
omitt ed and in another case the right side of ( 16) was used ;
except in a minor sense, there were no significant differ-
ences in the coastal surge profile with these two cases.

APPENDIX II

For numerical computations using finite difference
forms, the following notation (Shuman [19]) will be
employed :

1 2

— / —ixvy

ux

SAs

-1

0

1

_2

0

2

— 1

0

1

U j.kJ <-' V

2 4
1 2

SAs

Uf,k;

(17)

U7.,

The following finite difference form was applied to (10) :
ul = j(Ai)J,,Ur*1-!lDJ.d(Br),,X'/y-(Bl)J,Ji.r}

v)=J(At)},kVfrk1-9DJA(Br)J,ir+(BihJ7V)

(18)

wujri[oyr+orr-2(B.%.B,%)]

where p is atmospheric pressure and the surface pressure
gradients are derived from the model storm. There is little
difference in the numerical results whether the A, B, C
functions are placed inside or outside the operators
given in (17) ; (18) is mixed in this respect.

The closed boundary at the shore was treated by a
numerical scheme given by Harris and Jelesnianski [4]
and Jelesnianski [5], when using (18):

■•-{■

Aom;= 4#..* h?,k -D2.khlk

2As f

(bxA

f(Ar)0.kVS.,

(Br)

+ 3(B,)0.*Di

0(B

'V I z?

Jilo.k

Pa \O.s/0,Jt

To.k

m,-^-Lm"

(19)

The term (dh/dy)™ A. cannot be directly applied since h"'
is not known on the boundary; tests made showed no
significant differences in the coastal surge if the term was
ignored, computed with time value m— 1, or if an iterative
process was used.

These forms are an adaptation of a finite difference
scheme given by Shuman [19]. Note that the time incre-
ment of the dissipating terms were formulated at time
(m— 1) rather than time (m). This procedure w-as necessary
to prevent instability in the finite difference computa-
tions (Richtmyer [18]).

The continuity equation (1) becomes

h

~TTXVy

' LJ x

APPENDIX III

Observed surge data during storm conditions are of
poor quality, awkwardly distributed, and too limited in
quantity to effect a satisfactory comparison between

Figure 20. — Orienting ;i model basin along a portion of the Eastern
Seaboard of the United States; the depths chosen for the onc-
dimensional model basin is shown in the insert, and is the mean
depths about Atlantic City. The model track, simulating the
natural storm track, for the September 1944 hurricane, is shown.

278-472 O - G7 ■

754

8

MONTHLY WEATHER REVIEW

Vol.95, No. 11

H i

OBSERVED \ I I

NO BOTTOM STRESS \ 1 1

NO BOTTOM SLIP, v=25ft2/sec /.-■ /

BOTTOM SLIP; r=.25ftZ/sec; s = 006 ft /sec ^

SEPT 1944 HURRICANE
ATLANTIC CITY

Figure 21. — Observed and computed surges, against time, at Atlantic City for the September 1944 storm. The basin used in the

computations is given in figure 20. Hours are for model time.

observed and computed coastal surge profiles. This
prevents a direct empirical approach to determine values
for eddy and slip coefficients. However, the special phe-
nomenon of observed resurgences at selected tide gages,
generated by storms traveling parallel to the coast, can
be used to determine plausible values for these coefficients,
at least for fast moving storms.

Since actually observed trains of resurgences are ir-
regular, we cannot readily obtain these coefficients by
comparison of computed and observed amplitudes of
resurgences in sequence. Instead, we shall compare the
directly generated crest and following resurgences ob-
served at Atlantic City against computed values; by
appropriate variations of the slip and eddy coefficients,
the amplitudes can be made to agree. Comparison of the
remainder of the surge, observed and computed, against
time will show whether additional modifications are
necessary.

To demonstrate this method consider the path of a
hurricane shown in figure 20. In this figure a rectangular,
one-dimensional depth basin is oriented along the coast;
the depth profile of the basin was derived from a mean
approximation of the seaward depth off Atlantic City.
The observed storm varied in strength, size, and speed
with time (see [1]) ; it had a pressure drop of about 95
mb. off Cape Hatteras that decreased to about 30 mb.
off Rhode Island ; its radius of maximum winds decreased
from 50 to 30 mi.; its speed increased from 25 to 35 m.p.h.
These model parameters give a stationary storm maximum

Figure 22. — Comparison of computed surge profiles, without and
with bottom stress (equation (10)), generated by a fast moving
storm traveling normal to the coast. The eddy viscosity coeffi-
cient v ranges through an order of magnitude; no bottom slip.

wind of about 105 m.p.h. initially and decreasing to
65 m.p.h. off Rhode Island (fig. 2*). In computations
described below, the model storm parameters, excepting
latitude and storm direction, were ohanged at each hour
of natural time as the storm moved across the basin; after
passing Rhode Island, the model storm parameters
remained constant.

This model storm and basin gave a computed surge,
with resurgences, at Atlantic City as shown in figure 21
for various bottom stress conditions. The observed di-
rectly generated crest was translated to the time origin of

•The latitude of Atlantic City is used in the computations, consequently the winds are
slightly different than given in figure 1.

November 1967

Chester P. Jelesnianski

755

Figure 23. — Same as figure 22 with bottom slip s equal to 0.006

ft. /sec.

Figure 24. — Same as figure 21, for hurricane Donna. Eddy and
slip coefficients were used in the computations.

the computed crests, i.e., model time. Notice that the
computed surge without bottom stress has a directly-
generated crest that agrees with the observed crest but
the computed resurgences are too large in amplitude;
this suggests the need of a dissipating mechanism. The
computed crests at time 4 hr. are the result of initialization
phenomena due to rapid storm growth to maturity; a
slower growth would suppress this precursor.

We wish to determine a value for an eddy viscosity
coefficient that suppresses the resurgences computed with-
out bottom stress but at the same time does not affect the
directly generated crest. To do this we first digress to
consider the effect of different eddy coefficient values on
the directly generated surge. We consider at this time the
case of a zero slip coefficient (no bottom stress) and plot
the computed surge profile generated by a fast moving
(30 m.p.h.) standard storm traveling normal to the coast
in a standard basin (fig. 22). We consider further the case
of an infinite bottom slip coefficient (no bottom current)
and bracket the no bottom stress profile in the figure with
profiles computed with eddy coefficients that range
through an order of magnitude. The peak surge decreases
monotonically with increasing eddy values for the range
shown in the figure. Notice that small values of the eddy
coefficient give a directly generated crest larger than com-
putations without bottom stress. We now arbitrarily
choose a middle value for the eddy coefficient of c=0.25
ft.2/sec. and recompute the surge off Atlantic City.

For v=.25 ft.2/sec, figure 21 shows that the directly
generated crest is not significantly affected but the re-
surgences are dampened too strongly; there are also some
changes in the period of the resurgences. Since the two
flow conditions, frictionless flow and vanishing bottom
current, did not adequately portray the resurgences, it
was decided to use a bottom slip condition to better fit the
computed and observed resurgences.

To determine the effects of bottom slip, we return to
figure 22 and note that the profiles can be thought of as
extremes ; we then have a certain freedom in choosing the
bottom slip coefficient between zero and infinity. Figure
23 illustrates how the profiles computed with stress in
figure 22 were changed when incorporating a bottom slip
coefficient of s = 0.006 ft. /sec. The smaller the slip value
the closer thei profiles approach the no-bottom-stress
profile.

Figure 21 shows that when the above values for the
eddy viscosity and slip coefficients are used the directly
generated surge is not significantly affected by the bottom
slip and the agreement between computed and observed
resurgence amplitude is improved.

For an independent check of these coefficients, it was
decided to repeat the computations for another storm,
hurricane Donna, September 1960. Donna passed Atlantic
City with its center about 35 mi. seaward. From data
supplied by the Hydrometeorological Branch of the
Weather Bureau, ESS A, the storm parameters used in the
computations were: stationary-storm-maximum-wind 75
m.p.h. increasing by 0.5 m.p.h. each hour, radius of
maximum wind constant at 40 mi., speed initially at 30
m.p.h. and accelerating at 0.667 mi./hr.2; after passing-
Rhode Island, the storm parameters remained constant.
Figure 24 illustrates the computed versus observed surge
with time at Atlantic City. The observed secondary peak
or spike at time 11% hr. after initialization was not com-
puted by the model. No explanation is given for this
observed spike, whether it be dynamically or locally
generated; it appears to be one of several higher har-
monics, all of small amplitude, excepting the portrayed
spike. It should be mentioned that the observed surges
are best fit-by-eye curves from hourly observations sup-
plied by the Coast and Geodetic Survey, ESSA; the ob-
servations were corrected for astronomical tide with the
methods given by Harris [3]. The computed resurgences
of both storms in this section are predominantly shelf
seiches, but there were indications of edge or traveling
waves but only of small amplitude.

Although the slip and eddy viscosity coefficient values
were derived on the basis of fast moving storms passing-
over a basin equivalent to the seaward depths off Atlantic
City, we use these same values when computing for
slowly moving storms, and for all the basins encountered
on the coastal shelf of the United States. We have done
this for two reasons: in the first place, storms traveling at
more than 10 m.p.h. give computed peak surges which
have only minor differences whether bottom stress is

756

MONTHLY WEATHER REVIEW

Vol. 95, No. 11

used or not, and in the second place there are only few
observations of storms traveling less than 10 m.p.li.
(Harris [3]), and such observations as are available com-
pare as well with the computed surges of this study as
those of fast moving storms.

ACKNOWLEDGMENT

I am indebted to Dr. A. D. Taylor for his unstinting help and
assistance in the preparation of this report. In particular I am
grateful for his suggestions in the preparation of Appendix I where
he introduced many of the mathematical concepts.

REFERENCES

1. H. E. Graham and G. N. Hudson, "Surface Winds Near the

Center of Hurricanes (and Other Cyclones)," National
Hurricane Research Project Report No. 39, U.S. Weather
Bureau, Washington, D.C., Sept. 1960, 200 pp.

2. H. P. Greenspan, "The Generation of Edge Waves by Moving

Pressure Distributions," Journal of Fluid Mechanics, vol. 1,
No. 6, Dec. 1956, pp. 574-592.

3. D. L. Harris, "Characteristics of the Hurricane Storm Surge,"

Technical Paper No. 48, U.S. Weather Bureau, Washington,
D.C., 1963, 139 pp.

4. D. L. Harris and C. P. Jelesnianski, "Some Problems Involved

in the Numerical Solutions of Tidal Hydraulics Equations,"
Monthly Weather Review, vol. 92, No. 9, Sept. 1964, pp.
409-422.

5. C. P. Jelesnianski, "A Numerical Computation of Storm Tides

by a Tropical Storm Impinging on a Continental Shelf,"
Monthly Weather Review, vol. 93, No. 6, June 1965, pp.
343-358.

6. C. P. Jelesnianski, "Numerical Computations of Storm Surges

Without Bottom Stress," Monthly Weather Review, vol. 94,
No. 6, June 1966, pp. 379-394.

7. K. Kajiura, "A Theoretical and Empirical Study of Storm

Induced Water Level Anomalies," Project 202, Reference
59-23F, Texas A. & M. University, Department of Ocean-
ography and Meteorology, Dec. 1956, 97 pp.

8. W. Munk, F. Snodgrcss, and G. Carrier, "Edge Waves on a

Continental Shelf," Science, vol. 123, No. 3187, Jan. 1956,
pp. 127-132.

9. T. Nomitsu, "A Theory of the Rising Stage of Drift Current
in the Ocean: I. The Case of No Bottom Current," Memoirs,
College of Science, Kyoto Imperial University (Series A),
vol. 16, 1933, pp. 161-175.

10. T. Nomitsu, "A Theory of the Rising Stage of Drift Current

in the Ocean: III. The Case of a Finite Bottom-Friction
Depending on the Slip Velocity," Memoirs, College of Science,
Kyoto Imperial University (Series A), vol. 16, 1933, pp.
309-331.

11. T. Nomitsu, "On the Development of the Slope Current and

the Barometric Current in the Ocean: I. The Case of No
Bottom-Current," Memoirs, College of Science, Kyoto Im-
perial University (Series A), vol. 16, 1933, pp. 203-241.

12. T. Nomitsu, "Coast Effect Upon the Ocean Current and the

Sea Level: II. Changing State," Memoirs, College of Science,
Kyoto Imperial University (Series A), vol. 16, 1934, pp.
249-280.

13. T. Nomitsu and T. Takegami, "On the Development of the

Slope Current and the Barometric Current in the Ocean:
II. Different Bottom Conditions Assumed," Memoirs, College
of Science, Kyoto Imperial University (Series A), vol. 16,
1933, pp. 333-351.

14. T. Nomitsu and T. Takegami, "Coast Effect Upon the Ocean

Current and the Sea Level: I. Steady State," Memoirs,
College of Science, Kyoto Imperial University (Series A),
vol. 16, 1934, pp. 93-141.

15. G. W. Platzman, "The Dynamical Prediction of Wind Tides

on Lake Erie," Meteorological Monographs, vol. 4, No. 26,
Sept. 1963, 44 pp.

16. R. O. Reid, "Modification of the Quadratic Bottom-Stress Law

for Turbulent Channel Flow in the Presence of Surface Wind
Stress," Technical Memorandum. 93, U.S. Army Corps of
Engineers, Beach Erosion Board, 1957, 33 pp.

17. R. O. Reid, "Effect of Coriolis Force on Edge Waves (1)

Investigation of Normal Modes," Journal of Marine Research,
vol. 16, No. 2, 1958, pp. 109-144.

18. R. D. Richtmycr, Difference Methods for Initial-Value Problems,

Interscience Publishers, New York, 1957, 238 pp. (see p. 94).

19. F. Shuman, "Numerical Experiments with the Primitive Equa-

tions." The Proceedings of the International Symposium on
Numerical Weather Predictions, Tokyo, .Nov. 7-13, 1060,
Meteorological Society of Japan, Tokyo, Mar. 1962, pp. 85-
108.

20. P. Wclander, "Numerical Prediction of Storm Surges," Ad-

vances in Geophysics, vol. 8, Academic Press, New York,
1961, pp. 315-379.

[Received January 30, 1067 ; revised July 26, 1967}

Ausust 1967

Reprinted from MONTHLY WEATHER REVIEW Vol. 95, No. 8

565

ATMOSPHERIC WATER VAPOR PROFILES DERIVED FROM
REMOTE-SENSING RADIOMETER MEASUREMENTS

PETER M. KUHN

Atmospheric Physics and Chemistry Laboratory, Institute for Atmospheric Sciences, ESSA, Madison, Wis.

JAMES D. McFADDEN

Sea Air Interaction Laboratory, Institute for Oceanography, ESSA, Silver Spring, Md.

ABSTRACT

The feasibility and preliminary testing of a low cost, remote-sensing air-borne, double bolometer technique for
inferring atmospheric water vapor is illustrated. To deduce the water vapor profile with commercially available equip-
ment, the radiative transfer equation is solved for the water vapor transmissivity employing an input data remote
radiometer-measured upward irradiances obtained at aircraft holding levels. Radiometers sensitive in two separate
spectral bands are used. The primary radiometer covers the 4.39, to 20.83^. broad atmospheric radiation band, and
the second, for surface temperature deduction, covers the atmospheric window region, 7.35 to 13.16^.

The transfer solution results are acquired from computer programs developed specifically for this purpose.
Results indicate an accuracy for inferred total troposphcric water vapor and mixing ratio profiles close to that of the
standard sounding electrical hygrometer. The absolute accuracy of the radiosonde hygrometer, considering surface
calibration procedures, and for a single ascent, is not better than ± 12 percent. The absolute accuracy is greatest
for "dry" soundings where the largest changes in irradiance occur for given changes in moisture.

Specifically, tests for a vertical profile averaging 6.00 gm./kg. of water vapor produce an average error of 0.70
gm./kg. in the inferred mixing ratio. The average error in mixing ratio obtained by this technique for profiles aver-
aging 2.3 gm./kg. is 0.05 gm./kg. The implications for use on high-flying aircraft or on rockets with highly sensitive
radiometers are obvious. The primary purpose in reporting this research is to suggest a technique and illustrate its
use. It is clear that with more sensitive bolometer radiometers with selective band pass filters a considerable increase
in accuracy can be achieved.

LIST OF SYMBOLS

Upward irradiance (watts/meter2)

Subscripts employed in indexing variables

Wave number increment (centimeter-1)

Black body irradiance (watts/meter2)

Wave number (centimeter-1)

Temperature (°C.)

Pressure (millibars)

Filter transmissivity (percent)

Transmissivity increment (percent)

Water vapor quantity

Carbon dioxide quantity

Subscript denotes surface value of subscripted

parameter

Mixing ratio (grams/kilogram)

Exponent of a power law equation

Optical depth of atmospheric gas (grams/

centimeter2), pressure and temperature scaled

Calculated upward irradiance (watts/meter2)

Subscript denoting irradiance

Generalized absorption coefficient for water

vapor after Elsasser [9]

Limiting wave number increment identifier

n Number of atmospheric levels between surface.

and a reference level
TeQ Black body equivalent radiating temperature

r Transmissivity (percent)

1. INTRODUCTION

The radiative power transfer equation can be solved
iteratively to infer atmospheric water vapor from remote
radiant power measurements. The purpose of this research
is to propose and describe this technique for deducing
water vapor quantities employing observations of radiant
power as input to the radiative power transfer equation.
With more sensitive bolometer radiometers having selec-
tive band pass filters greater accuracy than is reported
can be obtained. On-hand equipment was used in this
pilot study.

The observations were made in two spectral regions,
4.39 through 20.83u and 7.35 through 13.16/u. Results of a
similar technique employing balloons were given by
Kuhn [1] and Kuhn and Cox [2]. The technique differs from
those of satellite radiometric inferences of atmospheric
water vapor and temperature (Houghton [3] ; Wark [4] ;
Moller [5]; King [6], [7]). The latter observations are

566

MONTHLY WEATHER REVIEW

Vol. 95, No. 8

necessarily limited to observations from a fixed satellite
orbit. They employ highly sensitive sensors receptive to
radiant power in one or more different spectral intervals.
This procedure requires measurements of temperature,
height, and spectral irradiance at a number of aircraft
holding levels. Two "shelf-type" radiometers sensitive
over two different spectral intervals constitute the sensor
capability. One radiometer monitors the air-surface
interface temperature in the relatively transparent posi-
tion of the atmospheric spectrum, while the other measures
the spectral component of upward terrestrial long-wave
irradiance as a function of height in the broad-band,
earth-atmosphere, self-emission region. Ideally, this second
radiometer would be sensitive over the strongly absorbing
atmospheric water vapor band centered at 6.7/x. Funding
and ready availability prevented use of such a special
purpose bolometer radiometer.

2. RADIANT POWER COMPUTATIONS

As mentioned, pressure, temperature, and water vapor
data are the normal input to the radiative transfer equation
solution for spectral irradiance passing upward through the
atmosphere.

For a plane parallel atmosphere, containing no scat-
terers, in local thermodynamic equilibrium and consisting
of gaseous water vapor and carbon dioxide, the upward
irradiance through any reference level above the surface
may be obtained by the following finite difference form of
the radiative transfer equation,

m n

Ft =-EAv,E5^, T(p))$(v)ATt(v, u*(p, w), T)

m n

-S>,S-B*(", T(p))*(v)ATt(v, u*(p, C), T)

;=1 i=\

m

IS^, T(p0))(rt(v, U*(p, w), T)rt{v, u*(p, C), T.) (1)

i = l

The effects of other radiatively active gases such as ozone
and nitrous oxide in the wavelength regions considered are
insignificant, amounting to less than 1 percent of the
upward irradiance, and are not considered.

Equation (1) is employed in deducing the quantity of
atmospheric water vapor from measurements of upward
irradiance for a particular spectral interval. An iteration
procedure resulting in a direct solution of the water vapor
transmissivity, tw{v, p, T), is used. Since the water vapor
transmissivity is a function of the quantity of water vapor,
w, beneath a given reference level, and since the amount of
water vapor is a function of the mixing ratio, the solution
yields the mixing ratio.

The iteration procedure requires calculations of irradi-
ance with a stepped series of trial values of water vapor
(equation (1)) until the difference between calculated and
observed upward irradiance is minimized. Carbon dioxide
emission and transmission are calculated assuming a
constant mixing ratio. A first approximation for the water

vapor profile is described by a power law expression of the
form,

w=wa{piPoy (2)

after Smith [8]. X, the exponent of a power law expression,
changes the water vapor profile as required in the iteration.
For tropical soundings W0 is assumed to be 5.0 gm./kg.,
certainly a lower bound. For mid-latitude summer sound-
ings, W0 is assumed to 0.6 gm./kg. For mid-latitude
winter soundings W0 is assumed to be 0.06 gm./kg. In
other words, a lower bound is chosen to start the compu-
tations.

The iteration procedure, involving repeated solutions of
equation (1), requires minimizing the quantity,

i

Stepwise changes in X, (equation (2)) for subsequent
successive increases in WQ provide the repeated input of
the "trial" water vapor quantities required for equation
(1). The computer evaluates Ft] from an initial "mini-
mum" profile of W0 shaped by an initial X value of 0.5.
The X values are stepped upward to a maximum value of
3.0 and then the process repeats with a new stepped
increase in W0. Negative values of X will allow the mixing
ratio profile to increase with altitude above the surface.
The changing of the entire water vapor profile by the
power function approximation of equation (2) imposes a
stabilizing constraint on the solution of equation (1). This
same procedure has been used by Kuhn and Cox [2] to
infer stratospheric water vapor profiles. It should be noted
that the constraint of equation (2) does not preclude
convergence at every level, within reason, since the step-
wise variation of X allows, literally, almost any charac-
teristic shape to the W profile to obtain. Average computer
solution time is 0.3 min. (CDC-1604) for 10 levels, over
the spectral range 4.39 to 20.83ju (480-2280 cm.-').

3. UNIQUENESS OF THE SOLUTION FOR INFERRED
WATER VAPOR

In view of the rate of change of the water vapor slab
transmissivity, rF, with changes in the amount of atmos-
pheric water vapor, u*, it is necessary to establish the
uniqueness of the radiometrically inferred water vapor
quantity. Figure 1 gives the water vapor slab or irradiance
transmissivity as a function of the sum of the logarithm
of «*, the pressure and temperature scaled optical thick-
ness, and the logarithm of the generalized absorption
coefficient L, (Elsasser [9]). This is expressed by,

tv=tf (logio «*+logio L).

(3)

From this figure it is evident that the greatest change in
transmissivity for a given change in log u*-\-log L occurs
between values of —1.50 and 0.50 for log u*+\og L.
Assuming a mean value for log L of —0.50, this represents

August 1967

Peter M. Kuhn and James D. McFadden

567

WAVENUMBER (CM"')

-2 -I 0
LOG U + LOG L

Figure 1. — Water vapor beam transmissivity vs. (logi0M* + logioi).

an optical thickness range of from approximately 0.1
gm./cm.2 to 10.0 gm./cm.2

Aerosol contamination is clearly evident as a sharp
discontinuity in the observed moisture profile. In fact,
detection of thin clouds at night is possible with these
instruments.

4. IRRADIANCE OBSERVATIONS

The primary instrument employed in this research was
a chopper bolometer radiometer having a 30° optical
field of view, germanium optical flat-front lens with a
spectral bandpass from 4.39 to 20.83ju (480 to 2280 cm.-1).
The mean band transmissivity is 0.50.

The relative resolution and special characteristics of
the primary radiometer used in these experiments are
given in table 1. Various optical companies can provide
bolometer radiometers with resolution of at least 0.025°C,
twice the resolution of the unit used. Thus for a 0.025°C.
change in target temperature at 5.0°C the corresponding

Table 1. — Radiometer specifications

Spectral passband

Average filter transmissivity

Field of view

Maximum temperature resolution

Irradiance change behind germanium flat filter
for 0.05'C. change at 5.0°C.

Temperature range

Recorder output..

4.39 to 20.83 ».

0.58

30°.

0.05°C.

0.7 micro watt/cm.' (.007 watt/m.')

-40.0°C. to + 30.0°C.

0-50 millivolts full scale into 1000 ohms

00

1500

14 13

12

1 IOOO .

300

800

700

, , , |, |

'

'

O 10 gm/cm2 THRU 2000"

80

V j

o

60

,

X

o ,
i 1 gm/cm2 THRU 3000

40

!tf

|\

20

•iily^i!

\

Jpi^

V--x,

,x>

v///////i

'/M&Ss*o.

~*«

8 9 10 II 12 13

WAVELENGTH (MICRONS)

IRW-FILTER TRANSMISSIVITY

Figure 2. — Window radiometer filter transmissivity vs. wavelength.

2500

WAVE NUMBER (CM"')
1000 625

500

•- 4 6 8 10 12 14 16 18 20

WAVELENGTH (MICRONS)

Figure 3. — Broad-band radiometer filter transmissivity vs. wave-
length. \og10L vs. wavelength.

irradiance change behind the optical flat filter is 0.0035
watt/m.2 For an average tropospheric sounding to 20,000
ft., this would correspond to a mean water vapor mixing
ratio resolution of 0.05 gm./kg. or 50 parts per million
for a moderately dry atmosphere. Throughout, we are
considering air-borne radiometers at costs not exceeding
$10,000. The absolute accuracy of the broad bandpass
radiometer is 0.028 watt/m.2 The absolute accuracy of
the "window" radiometer is 2°C. above 0°C.

The curves of the filter transmissivity of the "window"
radiometer and the primary broad-band earth and atmos-
phere self emission radiometer are shown in figures 2
and 3. The solid angle aperture of the surface temperature
monitor "window" radiometer is 3°. The radiometers
were calibrated against a hemispherically symmetrical
black source. We then have,

F\?=^ B(v,

TtMv)dv.

(4)

The equivalent blackbody temperature Teq, can be
determined from equation (4). B(v,Teq) is the hemisphere
blackbody irradiance. The curves of radiometer power
versus equivalent blackbody source temperature are

568

MONTHLY WEATHER REVIEW

Vol. 95, No. 8

10

5

0

_ -5
o

V10

-15
-20

! • '

.•'

CALIBRATION

7.35 -

3 16 u

1.6 1.8 2.0 2.2 2.4 2.6 2.8
(MICRO WATTS/CENTIMETER
BAND IRRADIANCE

30
2*

3.2

Window radiometer calibration curve, equivalent
temperature vs. micro watts/square centimeter.

CALIBRATION

30' FIELD OF VIEW

Figure 4.—
blackbody

6.0 7.0 8.0 9.0 10.0

SPECTRAL IRRADIANCE (WATTS/METER2 )

Figure 5. — Broad-band radiometer calibration curve, equivalent
blackbody temperature vs. watts/square meter.

reproduced in figures 4 and 5 for the window and broad
bandpass radiometers, respectively.

Reference to figure 2 shows the considerable oveilap of
the water vapor absorption band in the bandpass area,
7.1 to 9.5m- To show the closing effects of this band, ir-
radiance calculations for received power were run for the
June mean monthly sounding at Saidt Ste. Marie, Mich.,
from sea level to 18,000 ft. Figure 6 illustrates the influ-
ence of water vapor absorption for this sounding in the
shorter-wave end of the filter used on the window radiom-
eter (7.35-13. 16/u), and the lesser influence in the 10.00
to 12.05-/. bandpass of a curiently available "window"
radiometer. The deduced surface temperature error for
the narrow pass filter is about half that of the window
radiometer used in this experiment when observing the
surface at 700 mb. (or 10,000 ft.). In addition, a tempera-
ture and moisture inversion below 900 mb. has consider-
ably less effect in the narrower bandpass region. These
facts are, of course, not new, but in light of the fact that

500 -

600 -

700 -

SSM

JUNE

© ©

10.00

12.05 MIC

""

1

\

® TARGET

-

» TEMP.

1 © T
\ \

\ \

© ©

\ \

s ©

\ \
\ \

© ©

©<?

1

l

i \

3 ©
1 J

1 o 1

i

o- 800 -

900 -

1000

7 8 9 10 11

TEMPERATURE °C

Figure 6. — Temperature correction vs. pressure at Sault Ste.
Marie, Mich. (June) for radiometers equipped with 7.35-13.16-m
filter and a 10.00-12.05-m filter.

IR radiometers are in such heavy use today in a variety
of disciplines, this point deserves this amplification.

5. AIRCRAFT SOUNDINGS

The NCAR Queenaire Beechcraft was made available
for this research by Dr. D. R. Rex, Director of Flight
Facilities at National Center for Atmospheric Research.
Viewing ports with shock mountings supported the two
radiometers. Simultaneous measurements of the two up-
ward irradiances, altitude, and air temperature were made
coincidentally with visual observations of the surface.

One aircraft sounding was made in visually determined
cloudless conditions on July 14, 1965. The area surveyed
was over Lake Superior, 20 mi. east of Duluth, Minn.,
covering the period 2245 cdt through 2331 cdt. Table 2
gives the observed pressure, air temperature, surface
temperature, mixing ratio, and upward flux. In addition it
displays the calculations of upward irradiance and the in-
ferred mixing ratio vertical profile. The iteration on
mixing ratio converges quite rapidly.

Table 3 gives the results of a second ascent made in a
light aircraft, to approximately 9,000 ft., on May 25, 1966,
just east of Green Bay, Wis., but not over Lake Michigan.
The atmosphere was fairly moist, with an optical thickness
of 2.1 gm./cm.2 through 10,000 ft.

The several tabulations presented enable a comparison
of measured and calculated irradiances and the inferred
water vapor profile in gm./kg. The errors, shown in the
last column, are small enough to make these remotely
sensed inferred water vapor measurements useful.

September 1967

Reprinted from MONTHLY WEATHER REVIEW Vol. 95, No. 9

627

THERMALLY AND FRICTIONALLY PRODUCED WIND SHEAR IN THE PLANETARY
BOUNDARY LAYER AT LITTLE AMERICA, ANTARCTICA

BERNHARD LETTAU

Institute for Atmospheric Sciences, ESSA, Silver Spring, Md.

ABSTRACT

Pilot balloon wind profiles obtained by the first and second Byrd Antarctic Expeditions arc analyzed to show
that the mean observed wind shear between the surface and 1,000 m. can be resolved into a frictional component
which produces a normal boundary layer wind spiral, and a thermal component resulting from the temperature
gradient at the ice edge, which deforms the normal wind spiral. Values of surface stress, surface Rossby number, gco-
strophic drag coefficient, energy dissipation, and roughness length derived from the wind profiles are collectively
sufficiently different from values obtained over land or water surfaces, to suggest that the ice surface produces its
own characteristic wind distribution.

1. INTRODUCTION

Little America Station was first established by the
Byrd Antarctic Expedition at 78°34' S., 163°56' W., near
the seaward edge of the Ross Ice Shelf in January 1929,
and continuous meteorological measurements were ob-
tained through February 1930. The base was reoccupied
in March 1934 by the Second Byrd Antarctic Expedition
and finally dismantled in February 1935 after another
full year. Included in the data were 983 pilot balloon
wind profiles— 414 in 1929-30, and 569 in 1934-35— from
which wind speeds and directions for standard levels at
roughly 200-m. intervals have been tabulated and pub-
lished (Grimminger and Haines [2]). In April 1940, the
West Base of the United States Antarctic Service Expedi-
tion was established as Little America III, 7 mi. north-
northeast of the camp of the Byrd expeditions. This sta-
tion was operated until January 1941, and produced an
additional 233 wind profiles, which, however, have not
been used here.

In this preliminary study the individual mean wind
shears between the surface and the top of the boundary
layer have been separated into thermally and frictionally
produced components, which are classified by season and
by surface wind direction. Representative mean wind
profiles are analyzed for various surface parameters in a
later section.

2. WIND DATA

Within the planetary boundary layer of a barotropic
atmosphere the wind profile is a function of the surface
stress, the Coriolis parameter, and the horizontal pressure
gradient. The resulting hodograph has a spiral form with
the surface wind directed to the left of the free ah geo-

strophic wind in the Northern Hemisphere, and approach-
ing it asymptotically at the top of the boundary layer.
In the Southern Hemisphere the surface wind is to the
right of the geostrophic wind.

In the following discussion a Cartesian coordinate
system will be used whose components are directed parallel
and normal to the surface wind. Components along and
to the right of the surface wind will be defined as positive.
In this system, applied in the Southern Hemisphere, the
wind vector at the top of the planetary boundary layer
(H= 1,000 m.) will generally have a positive parallel and
a negative normal component.

The condition of barotropy is rarely fulfilled in the
boundary layer, particularly not at Little America where
the seaward edge of the Ross Ice Shelf provides a strong-
horizontal temperature contrast throughout the year. The
relatively warm water to the north and the colder ice to
the south produce a thermal wind parallel to the ice edge,
generally toward the east, which will distort the simple
spiral hodographs. Under the given geographical condi-
tions the spiral will be elongated for westerly winds and
foreshortened for east winds.

In preparation for the analysis, the surface and 1,000-m.
wind readings were extracted from the pilot balloon
observations at Little America for both the 1929-30 and
1934-35 seasons, and were grouped by surface wind
direction and by season. The directional resolution was to
16 points, while the seasonal distribution was limited to
summer (November to February), and winter (May to
August). The directional distribution is asymmetric with
a preponderance of observations of southerly winds at
the surface. The least frequent wind direction was north-
northwest with two cases, both occurring in summer;

628

MONTHLY WEATHER REVIEW

Vol. 95, No. 9

the maximum number in each season was 47, for east
winds in summer, and south-southwest winds in winter.
It should be remembered, however, that, northwesterly
onshore winds were generally accompanied by low cloud
or fog, and that consequently the asymmetry is accentu-
ated, if not caused entirely, by the lack of a 1,000-m.
value. It is also true that when drifting snow or low clouds
precluded a reasonably complete sounding the balloon
launching schedule was suspended temporarily.

For each sounding that extended above a nominal
altitude of 1,000 m. (tabulated as an actual 990 in.) the
components of the 1,000-m. wind parallel and normal to
the surface wind were obtained. The component values
were then averaged by class, and the total shear, i.e., the
differences between components at 1,000 m. and the sur-
face, were calculated. The averaged values are given in
table 1 by surface wind direction and season.

The directional distribution of the surface wind speeds
shows that in summer easterly winds are somewhat
stronger than westerly winds, while in winter, winds
with a component from the northeast are stronger than
winds with a component from the southwest. If the mean
wind speed from each direction is considered representa-
tive of that sector, the mean summer and winter wind
speeds are nearly the same at about 4 m.p.s.

The directional distribution of the mean shear com-
ponents shows that the parallel component of the 1,000-m.
wind is less than the surface speed for easterly wind
directions, and exceeds the surface speed for westerly
wind directions in both winter and summer. This would
be expected from regional horizontal density gradients
and from the station location with respect to the open sea.
The normal component of the shear vector is directed to
the left for all surface wind directions in all seasons,
however, its magnitude is greatest for winds generally
from the north, and least for southerly winds. The angular
deviation of the 1,000-m. wind to the left of the surface
wind, when averaged over all wind directions, is 24° in
summer and 28° in winter.

Figure 1. — Geometric constructions to determine the frictional
and thermally produced shear vectors. V and AV are the surface
wind vectors and observed shear vectors, aV|l, and aV2f arc the
frictionally produced wind shear vectors, and T is the thermal
wind vector. Sec text for details of construction.

3. METHOD OF ANALYSIS

A single shear vector, which may represent the sum of
the frictionally and thermally produced shear, may not
uniquely be resolved into these two elements. It becomes
necessary to take at least pairs of observed shear vectors
and to make some assumptions about the structure of the
wind profile within the boundary layer. Suitable assump-
tions are: that for each pair the thermal wind vector
stays the same, that the angles formed by the frictional
shear vectors and the surface wind are equal, and that the
magnitude of the frictional shear vector is proportional
to the surface wind speed. These assumptions correspond
to a fixed geographic orientation of the thermal wind
vector, and a fixed orientation of the frictional shear
vector with respect to the surface wind direction; con-
sequently, any angular difference between two surface
wind vectors will be sufficient to determine uniquely
the thermal and frictional shear vectors.

Table 1. — Averaged surface wind, and component values of (he observed wind shear between the surface and 1,000 m. by surface wind direction
and season. Surface wind speed units are meters per second; shear units arc meters per second per kilometer.

SUMSIKlt

Surface wind

Shear component

parallel...

normal. . .

WINTER
Surface wind. .
Shear component

parallel. . .__ .

normal

(I. 71)
-3.48

0.70
-2. 90

0.34
-3.08

3.48
-0.51

-0.01
-2.50

4.40

-2.82
-1.12

3.17

-0. 00
-1.74

Surface Wind Direction

-2.80
-1.43

-0.03
-1.73

-2.29
-1.37

-0. 11
-1.14

-1.30
-1.78

0.70
-1.04

-1.40
-1.29

0.54
-1. 15

1.03
-1.83

0.99
-1.05

2. 71
-2.02

1.47
-1.79

3.01
-2. 50

2.94
-1.39

4. 11
-2.51

w

WNW

NW

NNW

2. 89

2.72

3.34

4.09

3.79
-1.81

4.09
-3.37

4.27
-3. 49

2.01
-3.50

2.58

2. 73

3. 13

5. 52

4. 31

-■>. 50

2. 93
-4.80

1.87
-6.00

0.72
-4.03

September 1967

Bernhard Lettau

629

NORTH ^ \ ) \ \

^\ \ / \

">^>\ \ \

/ ^~\^

Y- J s —

L-'-^""/-'

\ \ ^-^"

\ s-\ /

\ ^\

---

Wind Seed Imps) ^-^

s. Xs

0

12 3 4 5 x
Wind Sh.or (mp5/km)

Figure 2. — Directional distribution of the surface wind vectors,
frictionally produced shear vectors, and thermal wind vectors.
Summer.

\ \

,

NORTH 1

/

1 k

\

\

\ f"

/ /

y 4

(y

V— --"~ '

\ ^~
\

\ /

w.nd Speed Imps) ^ I

0 12 3 4 5 -yj'
Wind Sheer Imps /km)

~~~~~'—-.. ^

Figuke 3. — Directional distribution of the surface wind vectors
frictionally produced shear vectors, and thermal wind vectors.
Winter.

The separation of the two shear vectors is most straight-
forward if opposed wind directions are paired, as is shown
in the example in figure 1. Vectors of surface wind and
observed shear are drawn so that the heads of the observed
shear vectors coincide. In this hypothetical Southern
Hemisphere case, the east wind VI; turns sharply to the
left, while the west wind V2, turns slowly to the right
with altitude.

The following will explain the method in more detail.
Since the thermal wind vector is assumed to be the same
in both observations, and since the observed shear vectors
were drawn to one point, we may allow the head of the
thermal wind vector to fall on that point. It then follows
that the heads of the frictional shear vectors for the two
surface winds must also fall on one point, which must be
the tail of the thermal wind vector. The second assump-
tion, that the angles between the surface winds and the
frictional shear vectors are the same, requires that this
triple point lie on the line joining the heads of the two
surface wind vectors. The third assumption, that the
magnitude of each frictional shear vector is proportional
to the corresponding surface wind, requires that this
point coincide with the intersection of the lines joining
the heads and the tails of the two surface wind vectors.
In this example the frictional shear vector displaces the
1,000-m. wind vector 16° to the left of the surface wind
vector, while the thermal wind vector displaces it toward
the southeast. This has the effect of augmenting the rate
of turning in the one case, and reversing it in the other.

Figures 2 and 3 show the observed shear vectors for
the Little America data separated into frictional and
thermal components in summer and winter. The frictional
shear invariably has the effect of turning the wind vector
to the left and increasing the speed with height. The
amount of frictional turning of the wind vector from the
surface to 1,000 m. varies from 17° to 28° in summer,
and from 23° to 36° in winter. The seasonal difference may
reflect greater hydrostatic stability in the boundary layer
in winter, since the other effects on which the surface
wind angle depends are either not applicable — variation
with latitude — nor not very pronounced — variation with
surface roughness (cf. Johnson [3]).

The angles themselves are somewhat greater than would
be expected from theory. A representative geostrophic
wind speed of 550 cm. /sec, a Coriolis parameter of
1.42X 10~4 sec."1, and surface roughness of 0.01 cm., which
is typical of an Antarctic snow field, will produce an
angle between the wind at the surface and at the top of
the boundary layer of 15° (cf. H. Lettau [4]).

The effect of the thermal wind is to turn the surface
wind vector toward an azimuth of 91° in summer, and
toward an azimuth of 42° in winter. The effect is less
pronounced for north or south winds than for east and
west winds in both seasons, presumably because the
effect of the temperature gradient at the edge of the ice
near Little America is suppressed within a homogeneous
air mass moving perpendicular to the shore. The largest
thermal shears occur with zonal winds in summer, when
the ice edge is much closer to the station and of nearly
east-west orientation.

The change in direction of the thermal wind from sum-
mer to winter is related to the magnitude of the annual
temperature variation in the area surrounding Little
America. A shift such as that observed requires a much
greater seasonal temperature contrast to the west and
southwest than to the east and northeast of the station.

630

MONTHLY WEATHER REVIEW

Vol. 95, No. 9

WINTER

I I I

SE S SW

Surface Wind Direct!

Figure 4. — Source region temperature as a function of the surface
wind direction. The shaded areas show the directions of minimum
and maximum seasonal temperature contrast.

Although the air temperatures in the vicinity of the
station are not known directly, the seasonal contrast
may be investigated by assuming that the mean temper-
ature observed at the station with each wind direction is
representative of thermal conditions some distance up-
wind. As shown in figure 4, the seasonal temperature
contrast does vary with the wind direction, ranging from
a minimum of about 18° C. for north-northeast winds to
a maximum of about 28° C. for winds generally from the
west. It is suggested therefore that both the orientation
of the thermal wind vectors and the change from summer
to winter are direct results of the local temperature
distribution, rather than spurious geometrical values
introduced by the method of analysis.

4. DETAILED WIND PROFILES

A more complete representation of the boundary layer
may be obtained by a detailed analysis of the observed
wind profiles. Since the thermal wind is apparently
insensitive to changes in wind direction, this section has
been limited to the examination of the mean profiles
observed with north and south winds at the surface for
both the summer and the winter seasons.

The theoretical background for the analysis of boundary
layer wind profiles which include a constant thermal
wind has been given by H. Lettau [5], H. Lettau and

Hoeber [6], and Johnson [3]. The assumptions are made
that the large-scale motion is uniform and unaccelerated
over level terrain of constant surface roughness, that there
are no mean vertical motions, and that there are no
inertial forces, and that the vertical density variation
can be neglected. The wind velocity is then a function
only of the pressure gradient and the vertical derivative
of the shearing stress. If the coordinate system is oriented
with the y-axis parallel and the x-axis normal to the
direction of the surface stress, which is also the direction
of the surface wind, the vertical variation of the geo-
strophic component parallel to the surface wind is con-
strained by the fact that the surface stress has no compo-
nent normal to they-direction, and that the shearing stress
becomes negligible at height H. If v(z) is the observed wind
profile in the y-direction, and V(z) the geostrophic wind

profile in the same direction, then (V— v)dz=0. With

the assumption of a constant thermal wind and geostrophic
ambient conditions at z=H=l,000 m., V{z) is represented
by a straight line tangent to v(z) at z= 1,000 m., such that
the algebraic sum of the differences (V—v) (z) is zero. A
similar line of reasoning will not give the analogous U(z)
since the surface stress parallel to the surface wind is
not zero.

It is now possible to determine the vertical profile
of 7V,

T* = Pf£ (V-V)dz (1)

where p is the air density, and/ is the Coriolis parameter.
A similar expression can be written for t,„

l>-

)dz

(2)

where U(z) is the geostrophic wind profile and u{z) the
observed wind profile in the x-direction, although the
relation is not very useful at the moment since neither
the vertical profile of r„ nor that of U(z) is known. One
may, however, also express the shearing stress at any
level as the product of air density, wind shear, and eddy
diffusivity. Both components of the wind shear are known
and there is no reason to suppose the diffusivity to vary
with direction. Thus for all values of z,

du/bz
'dv/dz

(3)

in which r„ is the only unknown. A convenient value to
use is z=z*, the height at which V(z) and v(z) intersect,
which is the height of maximum rx. Thus U(z) is obtained
by the straight line tangent to ii(z) at 2=1,000 m., such
that

€

= Pf\ (U-u)dz.

(4)

Figures 5 through 8 show the above constructions for
smoothed mean northerly and southerly wind profiles in

September 1967

I000

Bernhard Lettau

6 7 8

Wind Speed Parallel to Surface Wind Direction (mps

—4 —3 —2 —I

Wind Speed Normal to Surface Wind Direction (mps)

Figure 5. — Observed ambient mean wind profile and computed linear geostrophic wind profile separated into components parallel and

normal to the surface wind direction. North winds in winter.

winter and summer. For the most part the differences
among the four cases are minor, and related to directional
rather than seasonal differences, suggesting that the ice
surface produces its own characteristic wind distribution.
The component parallel to the surface wind increases
with height immediately above the surface in all four
cases, but reaches a maximum below 400 m. and decreases
slowly with height above that level. The geostrophic wind
decreases continuously in the boundary layer indicating
that this component of the thermal wind is antiparallel to
the surface wind. Its magnitude however is relatively
small, ranging from 0.3 to 1.5 m. sec."1 km.-1 The height
at which tt reaches a maximum is approximately 160 m.,
with the exception of southerly winds in summer when

tx reaches a maximum at 200 m., and the maximum value
attained ranges from 0.17 and 0.25 dyne/cm.2

The component normal to the surface wind shows a
definite directional difference, caused by the relatively
fixed thermal wind vector. For the southerly winds this
component increases from the surface to roughly 500 m.,
then decreases to the top of the boundary layer, the devia-
tion being to the left of the surface wind. For the northerly
winds the component value in summer increases con-
tinuously to the left of the surface wind through the
boundary layer; in winter the profile is very similar with
the exception of a slight relative maximum at 600 m.
The geostrophic wind increases to the left for the northerly
components, and to the right for the southerly components,
implying an eastward-directed thermal wind for all cases.

Table 2. — Derived boundary layer -parameters at Little America

Height of maximum r?.__

Surface geostrophic wind.. . .

Surface stress..

Surface geostrophic wind angle.
Thermal wind vector

magnitude. . ... . .

azimuth.

Surface Rossby number

Geostrophic drag coefficient..

Energy dissipation

Roughness length

V,o

Ros

C

E

m./sec.

dyne/cm.-

degrees

m. sec.-1 km.-i
degrees

watts/m.-
cm.

North Winds
Winter Summer

155
8.78
0.81

1.72
30

8. 32 x 10°
0.025
0.61
0.8

160
5.81
0.64
24

1.24
75
2. 76 x 106
0.033
0.33
1.6

South Winds
Winter Summer

162
6.23
0.56
23

1.47
108
4. 57 x 10« I
0. 030
0.31
1. 1

202
6.15
0.94
34

1.42
108

0. 045 x 10e

0.041

0.40
105

632

MONTHLY WEATHER REVIEW

Vol. 95, No. 9

000

- 800

-600 Z

- 400 5

- 200

3 4 5 6—4

Wind Speed Parallel to Surface Wind Direction (mps)

—3 —2

Wind Speed Normal to Surface Wind Direction (mps)

Figure 6. — Observed ambient mean wind profile and computed linear geostrophic wind profile separated into components parallel and

normal to the surface wind direction. North winds in summer.

The magnitude of the normal component of the thermal
wind is again relatively small, ranging from 0.8 to 1.4 m.

sec.

km."1

A number of other boundary layer parameters, given
in table 2, may be determined either directly or sequen-
tially from the observed and the geostrophic wind pro-
files. Those determined directly include the surface
geostrophic wind, \gQ, obtained as the vector sum of the
two geostrophic components at the surface, the surface
stress, T0, determined from the relation

T0=f>f( (U-u)dZ

,/n

(5)

and the angle, a0, between the surface geostrophic wind
and the surface stress, determined by the arctangent of
the ratio U0/V0- Derived parameters include the surface
Rossby number, Ro0, which is a unique function of the
angle an, the geostrophic drag coefficient, C, determined by
the relation

r- h l

(6)

the energy dissipated in j the boundary layer, E, which
may be obtained from the geostrophic wind and the
surface stress (cf. H. Lettau [4]), and the roughness
length, z0, from the relation

Zo--

'RooJ

(7)

The tabulated values are internally reasonably con-
sistent with the exception of those parameter values de-
rived from the surface geostrophic wind angle for the mean
south wind profile in summer. The relatively much higher
value for this angle produces a much lower surface Rossby
number and consequently a much higher and quite
spurious roughness length.

Similar analyses of wind profiles in the boundary layer
have been undertaken by Johnson [3] for kite wind data
from four stations in the midwestern United States, and
by H. Lettau and Hoeber [6] for pilot balloon profiles ob-
tained on Helgoland in the North Sea. Although all three
studies are in reasonable agreement with one another, re-
sults of the first study are generally indicative of more
vigorous flow over a rougher surface than that at Little
America, while the second study shows more rapid air
motion over a surface comparable to that at Little
America. The differences in the surface stress and in the
frictional energy dissipation within the boundary layer
specifically emphasize these conclusions. At the inland
stations in the first study the surface stress always exceeds
0.8 dyne/cm.2 and generally ranges from 1.5 to 2.0 dyne/
cm.2, while the energy dissipation generally exceeds 1

Bernhard Letrau

3 4 5 6

Wind Speed Parallel to Surface Wind Direction (mps)

-3 —2 —I 0

Wind Speed Normal to Surface Wind Direction (mps)

Figure 7. — Observed ambient mean wind profile and computed linear geostrophic wind profile separated into components parallel and

normal to the surface wind direction. South winds in winter.

watt/m.2 On the ice shelf at Little America the stress
ranges from about 0.6 to 0.9 dyne/cm.2, and the energy
dissipation from 0.3 to 0.6 watt/m.2, lower by a factor of
roughly three. The Helgoland data, which essentially
represent wind profiles over a water surface, produce
surface stress values of 0.6 and 0.9 dyne/cm.2, and energy
dissipation values of 0.6 and 1.4 watt/m.2 Since the sur-
face stress value can be said to be determined by the shape
of the wind profile components in the boundary layer, it
is evident that these are roughly the same for the Helgo-
land and the Little America data. The energy dissipation
values, on the other hand, also depend on the mean geo-
strophic wind in the boundary layer, which at Helgoland
exceeds that at Little America by a factor of about two.
Thus the observed difference is entirely due to the
observed higher wind speed at Helgoland.

The computed angles between the surface geostrophic
wind and the surface stress in the Little America data
do not follow the similarity pattern described above.
These are more nearly equal to those found for the inland
data, which average about 25°, than to those found for
the littoral data (9.5° and 11.2°). From this point of
view the ice shelf is better described as a land surface
than as a water surface.

A second point of similarity between the midwestern
United States data and the Little America data is that
the observed angles exceed by roughly 7° the values
theoretically predicted by independently derived rough-
ness lengths. If one takes the roughness length obtained
as typical for the snow surface at the South Pole by
Dalrymple et al. [1], 20=0.014 cm., together with the
observed wind speeds, one obtains a surface Rossby
number of 3X108, which corresponds to an angle between
the surface geostrophic wind and the surface stress of 17°.
The difference, as obtained by Johnson [3], was attributed
to a real height variation of the thermal wind which
would become obscured by the method of analysis,
rather than to topographical or other external effects.
A similar real height variation of the thermal wind
should be expected in the Little America data because
of the complex thermal structure of the boundary layer
which would produce a number of abrupt wind velocity
changes rather than the smooth transition that has been
shown here. The diabatic effects which should be con-
sidered on the ice shelf include radiational cooling near
the ice surface, and temperature profiles which sometimes
change from inversion to lapse conditions within the
lowest 1,000 m.

MONTHLY WEATHER REVIEW

3 4 5 6-4

Wind Speed Parallel to Surface Wind Direction (mps)

-3 —2

Wind Speed Normol to Surface Wind Direction (mps)

Figure 8. — Observed ambient mean wind profile and computed linear geostrophic wind profile separated into components parallel and

normal to the surface wind direction. South winds in summer.

5. DISCUSSION

A hypothetical example of precisely such diabatic
influences on the wind spiral near a snow surface has been
prepared by H. Lettau [4]. Here a surface cooling rate
of 26 langleys/day produced a significant reduction in the
surface wind speed, and a correspondingly greater angle
between the surface stress and the surface geostrophic
wind vector than under adiabatic conditions. Although a
surface inversion is in fact one of the major characteristics
of the Antarctic boundary layer, it is not possible to
investigate this diabatic effect in the Little America I
and II data, since almost no free-air temperatures were
obtained by the Byrd Antarctic Expeditions. Subsequent
scientific efforts in the Antarctic have of course obtained
simultaneous temperature and wind profiles, although
none has matched the nearly 1,000 boundary layer
profiles that have been used in this study to provide
reliable mean values.

REFERENCES

P. C. Dalrymple, H. H. Lettau, and S. H. Wollaston, "South
Pole Micromctcorology Program, Part II, Data Analysis,"
Report No. 20, Institute of Polar Studies, Ohio State Univer-
sity, 1963, 94 pp.

G. Grimminger and W. C. Haines, "Meteorological Results of
the Byrd Antarctic Expeditions 1928-30, 1933-35: Tables,"
Monthly Weather Review Supplement No. 41, 1939, 377 pp.

W. B. Johnson, Jr., "Climatology of Atmospheric Boundary
Layer Parameters and Energy Dissipation," Studies of the
Three-Dimensional Structure of the Planetary Boundary Layer,
Dept. of Meteorology, University of Wisconsin, 1962, pp.
125-158.

II. II. Lettau, "Notes on Theoretical Models of Profile Structure
in the Diabatic Surface Layer," Studies of the Three-Dimensional
Structure of the Planetary Boundary Layer, Dept. of Meteor-
ology, University of Wisconsin, 1962, pp. 195-226.

H. II. Lettau, "Windprofil, innere Ucibung, und Energic Umsatz
in den unteren 500 m. iiber dem Mccr," Beitrdge zur Physik
der Almosphdre, vol. 3D, No. 2, 1957, pp. 78-96.

H. II. Lettau and II. Ilocber, "Uber die Bestiinmung der
Hohenvertcilung von Schubspannung und Austauschkocffi-
zienten in der atmospharischen Heibungsschicht," Beitrage
zur Physik der Almosphdre, vol. 37, No. 2, 1964, pp. 105-118.

[Received May SI, 1967; revised June 15, 1967]

May 1967

Reprinted from MONTHLY WEATHER REVIEW Vol. 95, No. 5

299

SEA-SURFACE TEMPERATURES IN THE WAKE OF HURRICANE BETSY (1965)

JAMES D. McFADDEN

Sea-Air Interaction Laboratory, Institute for Oceanography, ESSA, Silver Spring, Md.

ABSTRACT

Following the passage of hurricane Betsy (1965) through the Gulf of Mexico two flights were made five days
apart aboard a research ah craft to collect sea-surface temperatures with an infrared radiometer. The purpose was
to study the effects of a hurricane on the sea-surface temperatures field. Data from the first flight, which occurred one
to two days after the hurricane passage, showed two cores of colder water to the right of the storm's track and very
little structure to the left. The flight made five days later still showed a core of colder water to the right, but by this
time its shape had been badly distorted by the surface current system. These results are compared with the findings
of other investigators, and the value of real-time synoptic coverage with the use of aircraft is pointed out. The
plan for an experiment utilizing aircraft and airborne oceanographic techniques to provide a 3-dimensional picture of
the ocean temperature structure prior to and following a hurricane is also presented.

1. INTRODUCTION

There has been considerable interest shown in recent
years in the effects of the passage of a hurricane on sea-
surface temperature. While it is agreed that the hurri-
cane causes a cooling of the sea surface of up to 5° C,
there appears to be some disagreement as to the mechanics
involved. Fisher [1] noted that pools of cold water were
created behind some hurricanes during parts of their lives,
and that this phenomenon is apparently produced by
upwelling in the ocean where the top layers are thermally
stratified. Jordan [5], working with ship temperature
data obtained prior to and following several typhoons in
the Pacific, concluded that vertical mixing is the primary
factor in the cooling of the surface layers and that mechan-
ical stirring is probably more important than organized
upwelling in this cooling process. He reached these con-
clusions mainly because the cooling was much more
pronounced on the right side of the storm, (relative to
the forward motion) the region of most intense wind and
wave action.

Stevenson and Armstrong [9], by measuring sea tem-
peratures in a zone of low-salinity shallow water near the
coast in the northwestern Gulf of Mexico after the pas-
sage of hurricane Carla (1961), observed that bathy-
thermograph traces revealed temperature inversions as
great as 2.5° C. extending as deep as 83 m. They hy-
pothesized that these inversions were formed in the surface
waters through a lowering of the water temperature by a
loss of heat to the hurricane. Leipper [7] made the most
complete study to date in his detailed oceanographic
investigation of that portion of the western Gulf through
which hurricane Hilda (1964) passed. His observations
indicated that the hurricane caused surface waters to be
transported away from its center, cooling and mixing

them to a slight degree as they moved. Convergence
outside the storm area resulted in downwelling to 80 to
100 m. in that area, while water on the order of 5° C.
colder upwelled from about 60 m. in the central region of
the storm.

Hidaka and Akiba [3] developed a theory to explain
cold water areas observed after hurricane passages which
indicates a considerable amount of upwelling in the center
of the storm. Ichiye [4], basing his results on fairly
rigorous mathematical treatment, shows weak descending
motion ahead of the storm, reaching somewhat larger
though negligible values near the center, followed by
strong vertical ascending motion thereafter. Gutman [2]
and O'Brien [8] have also done some modeling of these
phenomena. Gutman's solutions, which are obtained
using variable stress as a function of time and then com-
puting upwelling using continuity considerations, show
maximum upwelling to occur at the center of the storm.
O'Brien, on the other hand, derived a non-linear, theo-
retical model which describes upwelling and mixing
induced in a stratified, rotating two-layer ocean by mo-
mentum transfer from a stationary, axially symmetric
hurricane, and concluded that maximum upwelling occurs
in the region of maximum turbulent shearing stress.
O'Brien, however worked with a stationary model while
Gutman incorporated forward movement of the system
in his model.

Thus, some question remains as to the origin of these
cold spots observed in the wakes of hurricanes. Leipper's
conclusion that upwelling is responsible is certainly rea-
sonable based on the results of his cruise following the
passage of hurricane Hilda (1964), but his inference that
this upwelling occurred in the central region of the storm
is still not completely proven. The "after Hilda" data

300

MONTHLY WEATHER REVIEW

Vol. 95, No. 5

were collected on a seven-day cruise, which means that
the information obtained was not synoptic. Uncertainties
of interpretation may have resulted from the fact that no
consideration was given to the effects of the Gulf circu-
lation on the thermal structure in the interval between
the passage of the hurricane and the time the observations
were made.

2. OBJECTIVES

During the past 12 yr. there have been only 11 hur-
ricanes that have qualified as "great hurricanes", i.e.,
hurricanes with central pressure less than 950 mb. (Kraft
[6]). Hurricane Betsy (1965) was one of these. It had
qualified for this category before entering the Gulf of
Mexico on September 8 and continued as an intense
storm until shortly after landfall on September 10. Its
maximum surface wind speed averaged about 120 kt.
during the Gulf transect, and the eye diameter varied
between 25 and 80 n. mi. during this period (see fig. 1).

Because of the size and intensity it was immediately
recognized that this storm should have a profound effect
on the thermal structure of the surface of the ocean.
On September 9, the Director of the National Hurricane
Research Laboratory, ESSA, agreed to support the Sea
Air Interaction Laboratory's efforts to study these
effects by making available a research aircraft from
ESSA's Research Flight Facility for two nights to obtain
sea-surface temperature data with an infrared radiometer
in the wake of the storm. These missions were successfully
completed on September 10 and 15.

The objective of this paper is to present the essentially
real-time sea-surface temperature patterns obtained from
these two flights, to discuss the similarities and differences
between these results and those of previous investigators,
and to suggest an investigation that could possibly lead to
a more thorough understanding of the effects of the
storm on the ocean thermal structure.

3. DATA COLLECTION

ESSA's Research Flight Facility is adequately equipped
with multi-engine aircraft (two DC-6's, a DC-4, and a
B-57) suitable for long-range reconnaissance and especially
for hurricane research. The aircraft are outfitted with
a system of meteorological sensors, radars, and photo-
graphic equipment as well as digital tape, analog, and
photo-recording devices. For obtaining sea-surface tem-
peratures from the DC-6 aircraft a Barnes IT-2 radiom-
eter is employed, and the infrared (IR) data are recorded
on an oscillographic recorder. This sensor is shock
mounted vertically on a frame which fits inside the drop-
sonde chute during normal operation but which can
easily be removed at any time during flight in order that
in-flight calibration checks of the radiometer can be
made. Two-point calibration checks are made ap-
proximately every 30 min. during flight using two agitated

FLIGHT TRACK 10-11 SEPT 1965
ALONG PATH OF HURRICANE BETSY

Figure 1. — Flight track of September 10-11, 1965, superimposed on
the path of hurricane Betsy. Dashed lines define width of eye as
determined by radar and reconnaissance aircraft.

SEA SURFACE TEMPERATURE
DISTRIBUTION 10-11 SEPT 1965

+ MERCHANT SHIP DATA
(WithinJ2 hr. of IlighMilt

I

Figure 2. — Sea-surface temperature distribution on September
10-11, 1965.

water baths of different temperatures. This rather
frequent calibration helps to minimize readout errors
resulting from changes in detector bias voltages, detector
responsivity, amplifier gain, and amplifier drift of the
radiometer. Such changes otherwise could lead to errors
in the analysis of the data, which from experience could
be as much as 1.5° to 2° C.

The track of the first flight, superimposed over the path
of Betsy, is shown in figure 1. The times of three turning
points are given for comparison with the hurricane time
coordinates. The dashed lines denote the eye width as

May 1967

James D. McFadden

301

determined by radars at Key West and New Orleans and
by reconnaissance aircraft. Greater emphasis was placed
on the right side of the storm track, although the left
portion was adequately covered for detection of any
colder water zones in that region.

A DC-6 research aircraft departed Miami on September
10 at 2225 gmt and returned after completing the
mission at 0600 gmt on September 11. A flight altitude of
between 800 and 1000 ft. was maintained throughout
the flight. In addition to the IR data, meteorological
information was also obtained throughout the flight at a
sampling frequency of once every 10 sec. This information
included: temperature, pressure, humidity, pressure alti-
tude, radar altitude, and wind direction and wind speed as
determined by the Doppler navigational system. Precise
positioning was provided by means of Loran. The second
flight on September 15, with the exception of the flight
track, was conducted in the same manner as the first
flight,

4. RESULTS

The sea-surface temperature distribution as measured
on September 10-11 is shown in figure 2. The solid
lines are isotherms drawn with a reasonable degree of
certainty while the dashed lines, exclusive of the eye
size indication, are extrapolated isotherms. The four
temperatures at positions denoted by plus signs were
obtained by merchant ships within 12 hr. of the time in
which the IR data were collected. They were not used
in analyzing the data and positioning the isotherms and
are presented only for comparison with the airborne IR
measurements for this flight,

At a glance the cold water zone induced by the hur-
ricane is immediately the outstanding feature of this
figure. It may be noted, however, that all of this cold
water appears to the right of Betsy's path and farther
from the center than the eye wall. Another interesting
feature is the existence of two cores of cold water rather
than one continuous trough. The northward curl of
isotherms in the eastern core is of importance as will be
shown below. The lowest temperature detected by the
radiometer was slightly less than 25° C. This indication
was recorded several miles north of the intersection of
the leg of the flight occurring after the 2331 gmt turning
point with the leg made prior to the 0404 gmt turning
point on both tracks. The temperatures recorded on
both tracks at each of the two intersections were in
agreement with each other, thus indicating relative
validity of all of the IR flight data.

Based on preliminary results of the first flight, a second
plan was drawn up to cover some of the major features
which had been observed. Major emphasis was placed
on the north (right) side of the hurricane's path. At
0500 gmt on September 15, a DC-6 research aircraft
departed Miami on another mission to collect sea-surface
temperature data. Figure 3 shows the track of that

FLIGHT TRACK 15 SEPT 1965

Figure 3. — Flight track of September 15, 1965, superimposed on the
path of hurricane Betsy.

Figure 4. — Sea-surface temperature distribution on September 15,

1965.

flight, The aircraft returned to Miami at 1200 gmt
after completing the second of the two nighttime flights
made during the operation.

The major feature of the second flight shown in figure
4 is an elongation to the northeast of the cold core detected
on the earlier flight as indicated by the northward curl
of the isotherms shown in figure 2. The data also indicate
that during the five-day interval between flights the
cold-core surface temperatures increased about 0.5°
to 1.0° C.

Another interesting feature is the relatively small
cold water area southwest of St. Petersburg and Tampa
Bay. The shape of the isotherms indicates that this

302

MONTHLY WEATHER REVIEW

Vol. 95, No. 5

temperature structure is possibly associated with the
bay circulation and may result from relatively cool rain-
water runoff.

On the latter flight the aircraft did not traverse the
area of the Gulf where the second cold core was detected
on the western legs of the first flight.

5. CONCLUSIONS

The results of this experiment lead to two conclusions.
First, they confirm the observations of the other investi-
gators, as was expected, that hurricanes do cause well
defined areas of cold water to occur at the sea surface
in their wakes. The surface temperature data alone,
however, do not establish whether maximum upwelling
occurs in the central region of the storm, as concluded
by Gutman [2] and Leipper [7], or whether it occurs in
the region of maximum turbulent shearing stress as
concluded by O'Brien [8]. Certainly the positions of
the cold water areas observed on these flights fit more
closely positions of cold water zones described for Pacific
typhoons by Jordan [5], but again it is not obvious from
the IR data that vertical mixing induced by mechanical
stirring is more important than organized upwelling in
this cooling process as he concluded.

Perhaps even more important than the first observation
above, at least from a data collection standpoint, is the
rapid deformation of the eastern cold water core observed
in the interval between the two flights. This temperature
pattern change could have been realized by a northeast-
ward flowing current of less than 1 kt. Such a current,
the Yucatan Current, is a feature of the surface circulation
of the Gulf of Mexico.

The emphasis here is on the sampling time involved in
acquiring data for the solution of this particular problem.
On the one hand a research ship can obtain, in addition
to surface temperatures, sub-surface data which are
essential to answering the questions concerning upwelling
and mixing. A voyage to obtain this information, how-
ever, requires a considerable amount of time — about 10
days to cover such an area as discussed here.

The aircraft, on the other hand, covered the area in
less than 10 hr., but it was not suitably equipped to
obtain sub-surface temperatures. It would be desirable,
then, to combine the speed of the aircraft with the
versatility demonstrated by the ship for success! ully
attacking the problem. Wilkerson [10] points out that

with an airborne infrared sensor and an air-droppable
expendable bathythermograph, observations can now be
made of sea-surface temperatures and temperatures with
depth, thus providing a near-synoptic picture of the
thermal structure of the first few hundred meters of the
ocean over a large area. It would take several ships at
considerably greater expense to duplicate such observa-
tions. The Sea-Air Interaction Laboratory plans in the
near future to employ these techniques of airborne
oceanography in a more detailed synoptic study of the
effects of a hurricane on the thermal structure of the ocean
by measuring surface and sub-surface temperatures both
ahead of and behind the storm.

ACKNOWLEDGMENTS

The author expresses his thanks to the National Hurricane
Research Laboratory for making the flight time available, to
ESSA's Research Flight Facility, in particular to Dr. Gerald
Conrad, for collecting the data, and to Mr. Feodor Ostapoff and
Dr. Donald V. Hansen of the Institute for Oceanography for their
comments and suggestions.

REFERENCES

1. E. L. Fisher, "Hurricanes and Sea-Surface Temperature Field,"
Journal of Meteorology, vol. 15, No. 3, June 1958, pp. 328-333.

2. G. Gutman, "A Preliminary Study of Ocean Conditions as
Affected by a Hurricane Passage," Paper presented at the
Fourth Technical Conference on Hurricanes and Tropical
Meteorology, Miami Beach, Fla., Nov. 22-24, 1965.

3. K. Hidaka and Y. Akiba, "Upwelling Induced by a Circular

Wind System," Records of Oceanographic Works ■ in Japan,
vol. 2, No. 1, Mar. 1955, pp. 7-18.

4. T. Ichiye, "On the Variation of Oceanic Circulation, Part 5,"
Geophysical Magazine, Tokyo, vol. 26, No. 4, Aug. 1955, pp.
283-299.

5. C. L. Jordan, "On the Influence of Tropical Cyclones on the
Sea Surface Temperature Field," Proceedings, Symposium on
Tropical Meteorology, J. W. Hutchins (ed.), New Zealand
Meteorology Service, Wellington, 1964, pp. 614-622.

6. R. H. Kraft, "Great Hurricanes, 1955-1965," Mariners Weather
Log, vol. 10, No. 6, Nov. 1966, pp. 200-202.

7. D. F. Leipper, "Observed Ocean Conditions and Hurricane
Hilda, 1964, "Journal of the Atmospheric Sciences, 1967 (in press).

8. J. J. O'Brien, "The Non-Linear Response of a Two-Layer,

Baroclinic Ocean to a Stationary, Axially-Symmetric Hurri-
cane," Texas A&M University, Department of Oceanography,
Reference 65-34T, Dec. 1965, 98 pp.

9. R. E. Stevenson and R. S. Armstrong, "Heat Loss From the
Waters of the Northwest Gulf of Mexico During Hurricane
Carla," Geofisica Internacional, vol. 5, No. 2, 1965, pp. 49-57.

10. J. W. Wilkerson, "Airborne Oceanography," Geo-Marint
Technology, vol. 2, No. 8, Sept. 1966, pp. 9-16.

[Received February 9, 1967 ; revised March 3, 1967]

936

Reprinted from MONTHLY WEATHER REVIEW Vol. 95, No. 12

Vol. 95, No. 12

COMPATIBILITY OF AIRCRAFT AND SHIPBORNE INSTRUMENTS USED IN AIR-SEA

INTERACTION RESEARCH

JAMES D. McFADDEN

Sea Air Interaction Laboratory, ESSA, Silver Spring, Md.

and

JOHN W. WILKERSON

U.S. Naval Oceanographic Office, Washington, D.C.

ABSTRACT

On June 16, 1966, an experiment was performed off the east coast of Florida that involved two research aircraft,
one from the Naval Oceanographic Office and one from ESSA's Research Flight Facility, and the USCGSS Peirce,
aboard which were two scientists from ESSA's Sea Air Interaction Laboratory, and the Weather Bureau Airport
Station at Jacksonville, Fla. The purpose of this investigation was to determine the comparability of data for air-sea
interaction research as determined by aircraft temperature, humidity, pressure, and wind sensors; airborne IR radiom-
eters; a tethered boundary layer instrument package, radiosondes, rawinsondes, and dropsondes. Results showed
generally good agreement (within listed instrumental accuracies) between comparisons of aircraft and radiosonde
temperature and humidity observations, fair agreement of wind observations, and very poor comparisons between
dropsondes and radiosondes. The sea surface temperature readings obtained by the airborne radiation thermometer
aboard the Navy aircraft were well within ±0.4° C. operational accuracy of the instrument when compared with
bucket temperature measurements taken aboard the Peirce. Whether the accuracies of these presently available
instruments are good enough for mesoscale and macroscale ocean-atmosphere interaction investigations now being
planned will have to await studies of the environments in which these experiments will take place.

1. INTRODUCTION

A major, comprehensive, field investigation is now being
planned by ESSA, the primary focus of which will be on
the problem of ocean-atmosphere interaction, as well as
related topics in physical oceanography and microscale
and mesoscale meteorology. The primary objectives of
this experiment are: 1) to study the total fluid environ-
ment within a limited area, and 2) to provide a realistic
pilot field study for the planning and execution of a suc-
ceeding major Tropical Ocean Area Study within the
framework of the World Weather Watch.

The plans for this experiment call for among other things
the deployment of several ships on fixed stations several
hundred miles apart, a roving ship for making meteorolog-
ical and oceanographic measurements within the study
area, and several research aircraft that will be used in a
variety of measurement programs, such as vertical pro-
filing, making line integral observations, and obtaining
sea surface and subsurface temperature data. Further,
these plans call for the utilization of ship launched
rawinsondes, for obtaining vertical soundings of temper-
ature, pressure, humidity, and winds; tethered boundary
layer instrument packages for obtaining both profiles of
temperature, humidity, pressure, and winds and the time
variation of these quantities at any height up to 2000 m.;
and air released dropsondes for obtaining vertical profiles
of temperature, humidity, and pressure.

Before the investigation can be initiated, however, there
are certain preliminary steps that must be taken. These
include: 1) an evaluation of the compatibility of the
variety of instruments mentioned above, 2) an improve-
ment in our knowledge of the environment in the region
of the experiment, and 3) a test of the major instrument
systems in the experimental area. This paper concerns
itself with the first of these steps.

In June 1966 an experiment was performed to evaluate
the comparability of existing operational instruments of
the types planned for use in future mesoscale and macro-
scale ocean-atmosphere interaction studies. The objectives
of this investigation were to observe the comparability of
the data obtained by the various methods mentioned
above, to ascertain whether observed differences between
any two readings of a given parameter were within the
quoted accuracies of the measuring instruments, and to
arrive at a conclusion regarding the use of these state-of-
the-art instruments for air-sea interaction research.

2. INSTRUMENTATION AND DATA ACQUISITION

Because of the nature of this experiment a variety of
organizations was called upon to pool their efforts toward
accomplishing the mission. The U.S. Naval Oceano-
graphic Office provided the services of its Ocean Aerial
Survey Unit of the ASWEPS program and ESSA was
represented by the Weather Bureau Airport Station at
Jacksonville, Fla., the USCGSS Peirce, the Research

December 1 967

James D. McFadden and J. W. Wilkerson

937

Flight Facility (RFF) of the Institute for Atmospheric
Sciences, and the Sea Air Interaction Laboratory (SAIL) of
the Institute for Oceanography.

The Navy aircraft was a Lockheed Super-G Constella-
tion which, as one of the airborne platforms, was used to
obtain measurements of air temperature, pressure, and
humidity and sea surface temperature. The air tempera-
ture and pressure aboard this aircraft are obtained by a
meteorological set (AN/AMQ 17) which consists in pari
of a platinum wire resistor vortex thermometer and a
bellows mechanically linked to a potentiometer. The
operational accuracies listed for this device are 0.5° C.
for the thermometer and 5 mb. for the pressure sensor.

Humidity data are obtained by an infrared absorption
hygrometer system designed and built by the Weather
Bureau's Equipment Development Laboratory. Basically
this hygrometer is an optical instrument designed to
measure the absolute humidity of the atmosphere by
measuring the absorption of radiant energy over a given
optical path in the spectral region of the infrared water
vapor absorption band. A full description of this system
is given in [1].

A relatively large console model airborne radiation
thermometer (ART) developed by Barnes Engineering
Co. is used to obtain sea surface temperature aboard
the Navy aircraft. Radiation from the sea surface in the
8 to 13-/z region is detected by a thermistor bolometer
and this received energy, which is proportional to the
fourth power of the sea surface temperature, is translated
into temperature readings by electronic processing. The
operational accuracy of this system is listed as ±0.4° C.

These instruments represent just three of a large array
of devices designed for oceanographic research from the
Navy aircraft. A complete description of this aerial
platform and its capabilities is given in [2].

ESSA's Research Flight Facility provided this project
with a DC-6 aircraft instrumented primarily for hurricane
research. This aircraft was used as a platform from which
air temperature, pressure, and humidity; wind speed
and direction; and sea surface temperatures were obtained
and from which dropsondes to measure temperature,
pressure, and humidity were released.

An AMQ-8 vortex thermometer system is used aboard
the RFF aircraft to measure the free-air temperature
in flight. System accuracy is quoted to be 0.5° C. Ambient
pressure is measured by a pressure transducer operating
on an independent static source.

Humidity is measured by a system identical to that
described for the Navy aircraft. An interesting secondary
objective of this experiment was to obtain data from which
a thorough evaluation of the in-flight capabilities of these
two absorption hygrometers could be made. The infrared
hygrometer, while used on both aircraft as the primary
source of humidity information, is still classed as a
special purpose device. It has three outstanding features
that make it desirable for use aboard aircraft, high

sensitivity at low water-vapor concentrations, fast speed
of response for all water-vapor concentrations, and ability
to effect a humidity measurement without altering the
sample concentration by either adding or subtracting
water or changing the state of any part of the sample.

Wind speed and direction, along with latitude and
longitude information, are determined by a Doppler
navigation and wind-computing system (APN-82) manu-
factured by General Precision Laboratories. The Research
Flight Facility, from operational experience quotes for
this system an accuracy of ± 3 kt. for the wind speed
and an error function in degrees of roughly (150 -h- wind
speed) for wind direction.

The dropsonde system used aboard the RFF DC-6
is a military type (AN/AMT-3) and is used to obtain
soundings of temperature, pressure, and humidity from
aircraft flight level to the surface. Following launch from
the aircraft the instrument descends by parachute at a
rate of approximately 1,200 ft./min. During this descent
the package transmits measurements of temperature,
humidity, and pressure in International Morse Code
approximately 12 times a minute, and these transmissions
are hand copied aboard the aircraft.

The temperature element consists of a bimetallic strip,
the humidity element of several strips of hair; and the
pressure element is a double-bellows aneroid cell. RFF
experience with this system indicates that pressures
determined by this instrument are accurate to within plus
or minus 2 mb., and that temperature and humidity
data are comparable in accuracy to conventional radio-
sondes.

To obtain sea surface temperatures the RFF employs
a Barnes IT-2 infrared thermometer. In principle this
device operates similarly to the ART on the Navy aircraft,
being a thermistor bolometer with a spectral bandpass
of 8 to 13 /x. Physically, however, the unit is much smaller
and can be hand held. The sensing head of the radiometer
is mounted in the dropsonde chute of the DC-6. A more
detailed description of its operation is given in [3]. The
manufacturer lists a resolution of 1° F. and an absolute
accuracy of 2° F. for this instrument.

The USCGSS Peirce was used as a platform from which
scientists from SAIL launched conventional radiosondes,
obtained sea surface temperatures for comparison with
aircraft infrared-sensed temperatures, and gathered air
temperature and humidity data at 1,000-ft. and 500-ft.
heights with a boundary layer instrument package being-
developed in-house. This device, which was in the early
stages of development at that time, consisted of a stand-
ard 403-mHz. radiosonde package supported at the given
heights by a pair of kytoons tethered to the ship. The
aneroid switch had been replaced with a small clock
switch and this permitted alternate transmission of tem-
perature and relative humidity data to a standard radio-
sonde receiver-recorder located on the ship. (Subsequent
to this experiment a more sophisticated package has been
developed by SAIL which can be used to measure and

938

MONTHLY WEATHER REVIEW

Vol. 95, No. 12

transmit temperature, humidity, and wind speed data
simultaneously to receivers on the ship from any height
from the surface to 2000 m. The development of this sys-
tem is the subject of a forthcoming report now in prepara-
tion within the laboratory.)

The Weather Bureau participation in this experiment
consisted of a rawinsonde launch from the Airport Station
at Jacksonville, Fla. A 1200-gm. balloon was released at
this station shortly before the arrival of the aircraft over
Jacksonville and was tracked by GMD radar to an
altitude of about 32,000 ft. Air temperature, pressure, and
humidity, of course, were also obtained on this flight.

On June 16, 1966, under ideal weather conditions the
instrument comparison test was performed. The operation,
as shown in figure 1, proceeded in the following manner.
The two aircraft departed Miami, Fla., about 1345 gmt
and proceeded at an altitude of 5,000 ft. to a point north
of Cape Kennedy staying a few miles inland at all times.
This route was necessitated by a launch from the Cape
which had been rescheduled from the prior day. After
reaching this point the research aircraft turned eastward,
descended to 1,000-ft. altitude and continued on the
course shown to the USCGSS Peirce while maintaining a
separation between them of 2 to 3 mi. Upon arriving at
the ship and making an initial pass past the boundary
layer instrument package tethered at 1,000 ft. the RFF
aircraft began its climb to the 500-mb. level while the
Navy aircraft made three additional passes over the
Peirce before beginning its climb.

After completing the ascent to the desired altitude,
personnel aboard the DC-6 released a dropsonde while
simultaneously a radiosonde was released from the ship.
By the completion of the radiosonde and dropsonde
soundings the Navy plane had arrived at the same level
as the RFF aircraft, and both planes began a 500-ft./min.
spiral descent to 500 ft. where four additional passes were
made over the ship. A southwesterly flight path was then
followed from the Peirce to Jacksonville, Fla. Enroute,
the Navy plane flew at an altitude of 1,000 ft. collecting
sea surface temperature data while the RFF plane
climbed to 15,000 ft. for a second dropsonde release about
halfway between the ship and the airport station. After
reaching Jacksonville the DC-6 made a short ascent to
18,000 ft. and released a third dropsonde, as indicated in
figure 1, while the Constellation was completing its climb
from 1,000 ft. to the same height. Both aircraft then made
a spiral descent as described before to 1,000 ft. just
seaward of the coast and east of Jacksonville. After
completing the spiral descent, the Navy Constellation
terminated its data acquisition operation and returned to
its home base at Patuxent River, Md. The RFF DC-6
proceeded on a southerly course toward its home base at
Miami, breaking off data acquisition at Daytona Beach.

3. RESULTS

Figures 2-8 and table 1 present the results of this
experiment. A quick glance at these graphs makes it

CLIMB & DESCEND
US C&GS
PIERCE

Figure 1. — Flight plan of Navy and Research Flight Facility
aircraft on June 16, 1966.

Table 1. — Comparison of aircraft sensors with boundary layer
instrument package and sea surface temperature thermometer

Air temperature

(°C.)

Humidity

(gm./m.3)

Sea surface
temperature (°C.)

Altitude(ft.)

Navy

RFF

Peirce

RFF

Peirce

Navv
(ART)

Peirce
(Bucket)

1000

24.0
23.8
23.7
23.7
24.6
24.6
24.7
24.7

24.0

a

e

5

24.8
24.8
24.8
24.6

24. 1
24.3
24.0
25.1
24.7
25.0
25.0

IS. 3
No data

27.3
27.3
27.3
27.3
27.4
27.4
27.4
27.5

27.2

27.4

27.4

27.3

500

18.5
18.5
18.5
18.5

16.8
16.1
15.0
16.8

27.3
27.2
27.3

27.3

December 1967

James D. McFadden and J. W. Wilkerson

939

immediately apparent that there are no data for some of
the instruments originally planned to be used and de-
scribed in the previous section. Unfortunately, there were
equipment malfunctions during the flight and even more
unfortunate no back-up systems for substitution. This
will be discussed further, later.

Figure 2 depicts sea surface temperatures measured by
the ART on the Navy plane and a plane-to-plane compari-
son of air temperatures on the Cape Kennedy-to-ship leg
of the flight. As mentioned earlier the lateral separation
between the two aircraft was 2 to 3 mi. which possibly
could account for some of the observed differences
(maximum 0.5° C.) in the air temperature record.

Perhaps the greatest interest, particularly in the Navy
Oceanographic Office and the Research Flight Facility,
lay in the comparison of the infrared radiometers and the
absorption hygrometers of the two aircraft. It is sad to
say that neither of these comparisons could be made

AIR TEMPERATURES

NAVY AIRCRAFT
RFF AIRCRAFT

1530 1540

1550 1600

Figure 2. — Sea surface temperatures as measured by Navy ART
and air temperatures as measured by both aircraft on Cape
Kennedy to USCGSS Peirce leg of the flight.

\J

\

RFF

NA\

AIRCRAFT
r AIRCRAFT

\

^

SHIP

sf

\

I

\

AIR TEMPERATURE C

v,

1

*L

\\

RFF AIRCRAFT

\

\

RADIOSONDE

600

\

\^

*v

^

\

\

JAX

^

SHIP

a oo

\ r

\

11'

v

\\

\\

900
1000

\

\\

V

\

t-

\

I TtMPEHAIUBt

Figure 4.

-Comparison of radiosonde temperature profile
RFF aircraft profile.

with

500

\

OROPSONDE

\

x

HA

3IOSONDE

\\

\L

\
\\

V

=

JAX

\ \

SHIP

\

|jk

^

\

V

V

900

\ \

1000

vx

^ Is

p1 N

20 30

Figure 5. — Comparison of radiosonde and dropsondc temperature
sounding.

500

1

If
1

1 \
1 J
\f

i

!V JAX

\ SHIP

RADIOS

ONDE

\ \

/ /

\

\ \
\
\l
\\

1

\

\ \
\ ^
\

\l
\\

V

s

\ I

\
\

\

V 1

\ \
\ \
\

\

v\
\\
s
\

\ \
\ \

N

s

HUMIDITY (G/MJ

Figure 3. — Comparison of air temperature profiles measured by Figure 6. — Comparison of RFF aircraft and radiosonde humidity
both aircraft. sounding.

940

MONTHLY WEATHER REVIEW

Vol. 95, No. 12

I I

RFF AIRCRAFT

0 0 JAX RAWINSONDE

500

VP*

i

5>"~p

U

~" -jrv

-- -°

__LZ<r

6/

"S iX

700

y

\l?i

V}

0

800

u

-° <^

£^L_

? ~~~x>

900

0 *~~*~^

<f *

Y^\

I d"

6

90 180 270

WIND DIRECTION (DEG )

WIND SPEED (M/SEC)

Figure 7. — Doppler determined wind direction and speed compared
with rawinsonde data at Jacksonville, Fla.

20 30

TEMPERATURE °C

Figure 8. — Aircraft temperature profiles for the two stations com-
pared with the radiosonde profiles.

because of the malfunction inflight of one of these systems
on each aircraft. The Barnes IT-2 radiometer failed only
moments after crossing the coastline just north of the Cape
and could not be repaired aboard the plane. On the other
hand, the infrared hygrometer on the Navy plane mal-
functioned shortly after takeoff and its ailment was not
diagnosed until the Constellation returned to its home
base at Patuxent River, Md. The Navy plane has a
backup system for humidity measurements — part of the
AMQ-17 system described earlier — but it was not opera-
tional that day. Thus, no humidity measurements were
obtained by this group. The RFF had no backup IR unit
for its aircraft and thus sea surface temperatures were not
obtained by that group.

Comparisons of certain aircraft sensors with an early
version of SAIL's boundary layer instrument package and

of the Navy ART and Peirce bucket sea surface tempera-
tures are shown in the table. Here, plane-to-plane compari-
sons of air temperature are excellent and plane-to-boundary
layer package comparisons are reasonably good for the
Navy's 1,000-ft. passes and very good for the 500-ft.
passes of both airplanes. The humidity data are not too
comparable, and the fluctuations by the amounts shown for
the boundary layer package for the 1-min. intervals
between measurements indicate that part of the difference
may have been a result of relative humidity measurement
errors in the package sensor. The IR sea surface tempera-
ture readings certainly are within the ±0.4° C. accuracy
figure quoted by the Navy.

Figures 3 through 5 show the results of the vertical
profiles of air temperature obtained by the aircraft, radio-
sondes, and dropsondes over the ship and at Jacksonville.
Certainly one must agree that the aircraft comparisons
are the best of the lot, although a large part of the differ-
ences observed in the aircraft-radiosonde comparisons can
be rationalized away. The fact that the radiosonde read-
ings over the ship are consistently higher than the aircraft
readings might be attributed to trouble in obtaining a
baseline check aboard the ship, as was indicated by one
of the scientists making the launch. The excellent com-
parison at Jacksonville was obtained even though the
radiosonde was launched at the airport and the aircraft
profile was made several miles offshore or a total distance
of about 20 n.mi.

The dropsonde-radiosonde comparisons of air tem-
perature are so poor that figure 5 really should not receive
further comment. The large observed differences may be
attributed to calibration errors or improper baseline
checks for the dropsonde, although it seems unreasonable
that these errors would have caused the odd shapes of the
dropsonde soundings. Whatever the cause, the fact that
one unit was 12 yr. old and the other 14 yr. old probably
did not help the situation. At any rate this comparison
should be made again with newer equipment.

The humidity profiles shown in figure 6 are given as
absolute humidity (gm./m.3). For the radiosonde, absolute
humidity was computed from the temperature and relative
humidity data, thus presenting a possible double source
of error. The 2- to 3-gm./m.3 differences observed in the
soundings could be a result of temperature and relative
humidity errors in the radiosonde and/or some error in
the infrared hygrometer system on the aircraft. Again,
it was unfortunate that the second hygrometer system
was not available.

Because of the large differences between the radiosonde
and dropsonde air temperatures and the use of this para-
meter along with relative humidity to compute absolute
humidity, no comparison was made of the humidity data
from these two sources. Possibly a future experiment
will permit such a comparison.

The comparison of the Doppler wind system aboard
the RFF DC-6 with the winds-aloft data from the Jackson-
ville rawinsonde are shown in figure 7. Generally, for wind

December 1967

James D. McFadden and J. W. Wilkerson

941

direction the agreement above the 920-mb. level is reason-
ably good, the observed differences at most points being
within the accuracy limits quoted for the Doppler system.
This system, however, shows rapid fluctuations in wind
speed throughout the profile whereas the rawinsonde
data have a smoothed appearance. This, of course, is
due to the fact that individual point readings for rawin-
sonde wind data represent layer averages while Doppler
readings are instantaneous values.

4. CONCLUSIONS

What do these results mean in terms of air-sea in-
teraction studies? Surely, one recognizes that a more
uniform region, in terms of atmospheric parameters and
oceanic thermal conditions, places much more stringent
requirements on instrumental accuracies than an area
where there are large horizontal gradients of temperature
and moisture. One also realizes that accuracy requirements
are much greater where single samples are used as op-
posed to the cases where many samples are averaged.
With respect to this experiment these considerations
can be applied to the comparison shown in figure 8.
Here is shown a comparison of the vertical profiles of air
temperature at the ship and at Jacksonville, Fla., as
obtained in one case by the RFF aircraft and in the
other case by radiosonde measurements. Such a com-
parison could be made because the ship and Jacksonville
happened to fall on a line essentially parallel with the
wind trajectory (see fig. 7).

There are a couple of things worth noting in this
figure. First, for practically the whole of both soundings
the observed differences between the Jacksonville station
and the ship station are so small that they might rep-
resent instrumental inaccuracies rather than air tem-
perature changes. Second, note that for the layer between
660 and 830 mb. the aircraft soundings show almost
no net change in temperature while the radiosonde

soundings show a significant temperature increase. Thus
for an environment such as was found east of Jacksonville
in June, instrumental accuracies would be of major
concern.

It is fair to conclude from the results, then, that aircraft
and radiosonde measured air temperatures, and possibly
humidities also, were within the accuracies stated for the
sensors. Another look at the dropsonde should be taken
before it is accepted for air-sea interaction studies. A
further development of SAIL's boundary layer instrument
package should provide more accurate instrumentation
for boundary layer research.

Again, it is emphasized that this experiment was pri-
marily intended to show what the available instruments
could do. Whether they will be suitable for the ocean-
atmosphere interaction investigations now being planned
must await studies of the environment in which the ex-
periments are planned.

ACKNOWLEDGMENTS

The authors wish to acknowledge the cooperation and help of
the officers, crew, and scientific personnel of the Navy aircraft
El Coyote, of the RFF aircraft 39-Charlie, of the USCGSS Peirce
and the Weather Bureau Airport Station at Jacksonville, Fla.
A special note of thanks is given to Michael Bratnick of the Naval
Oceanographic Office and to Monte Poindextcr and Frank Shibuya
of the Sea Air Interaction Laboratory for their help in processing
the data.

REFERENCES

1. W. F. Staats, W. W. Foskett, and H. P. Jensen, "Infrared

Absorption Hygrometer," Humidity and Moisture, Measure-
ment, and Control in Science and Industry, vol. 1, Reinhold
Publishing Corporation, New York, 1965, pp. 465-480.

2. J. W. Wilkerson, "Airborne Oceanography," Geo-Marine

Technology, vol. 2, No. 8, Sept. 1966, pp. 9-16.

3. J. D. McFadden, "Sea Surface Temperatures in the Wake of

Hurricane Betsy (1965)," Monthly Weather Review, vol. 95,
No. 5, May 1967, pp. 299-302.

[Received July 11, 1967]

279-151 O - 67 - 9

Reprinted from JOURNAL OF GEOPHYSICAL RESEARCH
Vol. 72 No. k The American Geophysical Union

Journal of

Geophysical Research

Volume 72

February 15, 1967

No. 4

Climatological Significance of Albedo in Central Canada

James D. McFadden1 and Robert A. Ragotzkie

Department of Meteorology, University of Wisconsin, Madison

Between 1961 and 1965, short wavelength albedo data were collected on a number of aerial
reconnaissance flights made over the north-central portion of the United States and central
Canada from Wisconsin to the Arctic Ocean during periods of lake freezing and thawing. These
data show that there is very little horizontal variation in the albedo of the tundra in the
summer, when lakes are free of ice, and in the winter, when they are frozen and the region is
snow covered. Large horizontal variations of albedo occur in the tundra when lakes are frozen
but there is no snow, and in the boreal forest region when the lakes are frozen and snow
covered. The albedo values obtained over the tundra region were used to estimate the change
in the heat balance for the land surface for the period of rapid snow disappearance in June
1963. This estimate indicates that the sensible heat transfer to the atmosphere was sufficient
to heat the lower 1000 meters of air at a rate of 1.6°C per day for this period, a value that
agrees quite well with actual observations.

Introduction

Surface albedo, which may be denned as the
ratio of reflected to incident solar radiation at
the earth's surface, is important in any climatic
study, because the portion of the incident solar
radiation that is not reflected by the surface is
absorbed as heat that becomes available for
heating the lower atmosphere or for evaporat-
ing water. The horizontal variations and the
seasonal changes of albedo over large areas are
important to air mass modification and regional
climate studies. These variations and changes can
be estimated from aerial measurements of al-
bedo. Adequate studies on the feasibility and
suitability of using aircraft for surface albedo
measurements have been made by such investi-
gators as Fritz [1948], Bauer and Dutton [1960,
1962], Dutton [1962], Rung et al. [1964], and
Hanson and Viebrock [1964]. The reader is di-
rected to these papers for detailed discussions
of this subject.

1 Now at the Sea Air Interaction Laboratory,
ESSA, Silver Spring, Maryland.

Between 1961 and 1965, albedo data were
collected on a number of aerial reconnaissance
flights made over north-central United States
and central Canada from Wisconsin to the Arctic
Ocean during periods of lake freezing and thaw-
ing. These data were collected in conjunction
with a study on the interrelationship of lake ice
and climate [McFadden, 1965], and therefore
the main emphasis of this paper will be on the
effects of lakes and snow cover on albedo in the
tundra and boreal forest regions of central
Canada, and the climatological significance of
these effects for these particular regions. Al-
bedo values from the northern hardwood-conifer
forest and farm land region of the United States
are included for purposes of comparison.

Method

A U. S. Navy P2V Neptune patrol aircraft
was used as the aerial platform from which
incident and reflected short-wave radiation used
to compute albedo were measured. The plane
and its crew were provided by the Service Test
Division of the Naval Air Test Center, Patuxent

1135

1136

McFADDEN AND RAGOTZKIE

River, Maryland, through the Office of Naval
Research. Ideally, the data were collected at
1000 feet above the terrain, although on occa-
sion low stratus clouds forced the aircraft below
this level in order for the pilot to maintain
visual contact with the ground.

Two Kipp and Zonen solarimeters were
mounted level for flight on the upper and
lower surfaces of the fuselage of the aft section
of the aircraft. The construction of the upper
and lower surfaces of the fuselage is essentially
uninterrupted except for the vertical stabilizer,
thus providing both sensors almost completely
unobstructed fields of view. The solarimeters
give an output proportional to the incident
short-wave radiation of about 8 mv/ly/min.
They have a time constant of about 2.0 sec and
are accurate within an error of 3% [Bener,
1951]. The outputs were recorded on a Minne-
apolis-Honeywell Brown 12-point recorder. Each
of the outputs from the top and bottom sensors
was recorded alternately on this recorder every
2 sec.

The albedo records were analyzed in the fol-
lowing manner. The record of each flight was
first stratified according to terrain type, snow
cover, time of day, and cloud cover as recorded
by 16 mm movies, still pictures, and notes of
existing conditions made at frequent intervals

by the observer during each flight. After the
sectioning of the record was complete, values
of the upper and lower solarimeters were read
at %-min intervals within each section. Albedo
in per cent, which is the ratio of reflected to
incident solar radiation times 100, was com-
puted for the individual readings. The mean,
standard deviation, and relative variance (co-
efficient of variability) for the albedo of each
section were also computed.

It should be pointed out at this time that the
values given in this paper are hemispheric al-
bedo and somewhat different from the beam
albedo values presented in the paper by Kung
et al. [1964]. The significance of this difference,
as far as the climatological aspects of this paper
are concerned will be examined later, but dif-
ference in the mechanics of the two methods of
obtaining albedo from aircraft will be discussed
here.

Beam albedo is calculated from the output of
an upward-facing hemisphere radiometer and a
radiometer with a downward-facing parabolic
mirror that intercepts energy from a small area
within a 4° beam width and focuses this energy
on the radiometer. Hemispheric albedo, on the
other hand, is determined from the records from
the same type of upward-facing radiometer and
a down-facing solarimeter that has a full 2-ir ster.

TABLE 1. Representative Albedo Values for Various Regions

Terrain

Lakes

Standard Relative

Mean, Deviation, Range, Variance,
Snow Cover % % % %

Sky

Tundra

Frozen

Snow

89.0

2.98

84.0-92.0

3.4

Clear

Tundra

Frozen

Snow

77.7

3.45

70.5-83.2

4.4

Stratus

Tundra

Partly frozen

No snow

24.9

8.82

15.2-47.1

35.4

Clear

Tundra

Unfrozen

No snow

10.8

1.14

7.9-12.4

10.6

Clear

Tundra

None

No snow

14.8

0.95

13.5-15.6

6.4

Clear

Tundra

None

No snow

14.9

1.14

13.8-16.5

7.7

Clear

Forests

Frozen

Snow

48.0

1.14

30.3-63.2

23.7

Clear

Forests

Frozen

Snow

46.9

0.97

37.9-68.2

20.7

Clear

Forests

Partly frozen

Snow

34.5

8.00

22.7-54.5

23.2

As

Forests

Partly frozen

No snow

26.7

3.80

11.1-29.5

14.2

Clear-Hazy

Forests

Partly frozen

No snow

16.9

2.00

10.7-20.2

11.9

As

Forests

None

Light snow

18.2

3.00

13.0-26.0

16.7

Clear-Hazy

Forests

None

No snow

14.9

2.00

12.7-19.0

13.5

Clear-Hazy

Fields, woods

None

Snow

43.8

4.00

31.2-48.3

9.2

Clear-Hazy

Farms

None

None

22.1

3.8

13.4-26.4

17.2

Clear-Hazy

Farms

None

None

18.9

0.8

17.7-19.7

4.5

Clear-Hazy

Farms, woods,

and bogs

None

None

18.6

2.5

13.5-22.7

13.7

Clear-Hazy

Plowed fields

None

None

12.3

1.6

11.1-14.2

13.3

Clear-Hazy

CLIMATOLOGICAL SIGNIFICANCE OF ALBEDO

1137

100

r '<

-,

I

1

80

-

}

-

„

1

o

60

-

-

<

0£
<

I

>

40

<

1

„ <

<

20

EA

CLE

o

-

u S
1

£

LAKE
CONDITION

FROZEN

PARTLY
FROZEN

UNFROZEN

NO LAKES

COVER

SNOW

NO SNOW

Fig. 1. Albedo means and ranges for tundra
region. Sky condition for each set of measure-
ments is also shown.

field of view. At an altitude of 500 feet the beam
albedo system will instantaneously sample an
area 35 feet in diameter, while the hemisphere
system samples over the entire solid angle.

In a general way, values from the two dif-
ferent systems can be compared. Bauer and
Dutton [1962] and Dutton [1962] observed
that for a fresh layer of snow on a lake, which
they assumed to be a homogeneous and isotropic
surface, the measured hemispheric and beam
albedo values were in agreement and that a
calibration factor of 1.294 for the beam reflector
incorporated all deviations from the ideal para-
bolic reflector. Further, they showed that for
various terrain surfaces, excluding water, with-
out snow cover the ratio of hemispheric albedo
to (beam albedo X 1.294) was very close to
unity. One should expect, then, that the albedo
values reported herein for areas with no snow
and for a snow covered tundra would be in
agreement with beam albedo values, incorporat-
ing the calibration factor above. A comparison
of beam albedo values obtained by Kung et al.
[1964] and hemispheric values obtained by the
authors for similar type terrain in southern
Manitoba, under like conditions and at approxi-
mately the same date, showed that the values
were almost equal after the calibration correc-
tion was made. The beam albedo averages for
swamps, fields, and wooded farms in this area
ranged from 12 to 14%, whereas the corrected

'00

80

oe

< </i

J Z>

e*T _i

< U <

c£

t—

U

o

60

o

_

m

< >.

<

<

>

1 t

X

t—

IM
<

X

ea

I—

40

at
<

<

as

-

i—

<

T

u

o

i

1—

<

T

13
as

20

i

!

-L

LAKE
CONDITION

FROZEN

PARTLY FROZEN

NO LAKES

SNOW
COVER

SNOW

NO SNOW

LIGHT
SNOW

NO
SNOW

Fig. 2. Albedo means and ranges for boreal
forest region. Sky condition for each set of meas-
urements is also shown.

hemispheric averages ranged from 11.8 to 14.4%.
The uncorrected values ranged from 15.2 to
18.6%.

Results

Representative sections from the albedo rec-
ords obtained on various flights and covering
horizontal distances of 15 to 50 miles over the

100

80

O

2 60

CO

<

40

20

(

1

SXY CONDITION

1 - I

- CLEAR
BUT HAZY

5

TERRAIN

FARMS
WOODS

FARMS

FARMS,
WOODS,
AND BOGS

PLOWED
FIELDS

SNOW
COVER

SNOW

NO SNOW

Fig. 3. Albedo means and ranges for northern
hardwood-conifer forest and farmland mixture.

1138

McFADDEN AND RAGOTZKIE

TABLE 2. Albedo Means, Standard Deviations, and Ranges
in Per Cent for Various Surfaces

Date

Time,
CST1

Area Description

Standard
Mean Deviation

Range

July 13, 1962 1714 Kazan River, tundra, clear sky, altitude

2200 ft2
1800 Tundra and lakes, clear sky
1830 Tundra and lakes, clear sky, altitude 2000 ft
1845 Tundra and lakes, altitude 2000 ft
July 15, 1962 1103 Hudson Bay water (no ice)

1120 Fog banks over Hudson Bay

1130 Fog banks over Hudson Bay

1130 Ice flow on Hudson Bay
1130-1200 Marshy tundra and lakes, coastal lowlands,
high percentage of lakes
1200 Tundra and small lakes, stratus3
1210 Tundra and lakes, 70% lakes mostly ice

covered; thin stratus
1220 Tundra and lakes, some ice on lakes,

40-50% lakes, 63°30'N, thin stratus
1330 White Hills Lake, ice covered 10/10 cirrus
1330 Tundra and few lakes, north of Rossby Lake
1445 Chantrey inlet, floating ice

1450 Chantrey inlet, light colored water
1530 Sherman inlet, ice; Alto-stratus clouds

3500 ft
1530 Sherman inlet, ice, 3500 ft Alto-stratus
1730 Tundra and lakes, 3500 ft
1830 Dabawnt Lake, ice; stratus
1920 Tundra and lakes, near Ennadai
1950 Tundra east of Ennadai
May 22, 1963 1100-1200 Spruce trees, no snow, frozen lakes, darkish

ice, bogs, clear sky
1300-1345 Spruce forest, lakes, no ice, bogs, altitude
8000 ft
June 9, 1963 1215 Kasba Lake, ice covered, stratus, thinning,

heavy snow cover
1230 Ennadai Lake, ice cover, stratus, thinning,

heavy snow cover
1325 Large lake surrounded by rocky tundra,

ice, no snow, clear
1420 Spruce forest, heavy snow cover, lakes open,
stratus thinning
June 12, 1963 0835-0922 North of Winnipeg, farms and marshes, no

snow, stratus, var. thickness
0922-0941 Lake Winnipeg, stratus-var. thickness
1015-1030 55°40'N, 98°W, spruce forest, no snow
1225-1245 Mostly trees, some bogs (may be tundra),

lakes open, stratus
1250-1308 (Mostly trees) mixed forest-tundra, lakes

partly frozen, stratus
1308-1508 (Mostly trees) mixed forest-tundra, lakes

partly frozen, thinning stratus
1508-1528 Marsh and bogs, east of Lake Winnipeg
stratus
June 14, 1963 0835-0853 Bog and marshy area NNW of Winnipeg,

some scattered lakes, clear
0855-0915 Bog and marshy area NNW of Winnipeg,

some scattered lakes, clear
0933-0937 Lake Winnipegosis, clear
0941-0945 Cedar Lake 53°15'N, 100°W, clear

17.1

1.82

11.8-21.3

19.9

1.24

18.0-21.8

19.1

1.00

18.3-19.6

20.4

1.26

17.4-23.7

4.9

0.616

4.6-5.1

9.0

0.436

7.7-10.1

10.2

0.995

8.5-11.4

28.7

3.94

24.5-33.1

10.8

1.14

7.9-12.4

16.1

2.28

12.3-19.3

16.7

2.42

13.5-21.4

15.4

2.24

11.5-20.0

27.5

0.773

26.8-28.8

14.9

1.14

13.8-16.5

21.7

1.60

18.3-23.8

13.4

0.894

11.3-15.3

25.8

2.10

23.4-29.0

25.1

2.09

20.6-27.7

16.0

0.949

12.6-19.1

31.7

0.671

29.7-33.4

20.0

2.561

16.4-25.6

18.5

1.76

13.6-22.7

12.3

1.31

10.2-18.9

10.4

1.23

7.2-12.2

60.1

3.16

57.5-63.0

63.7

3.85

58.6-70.8

29.4

7.5

17.7-38.7

23.0

3.38

16.7-28.4

17.8

1.55

16.2-20.3

11.1

1.18

9.7-12.4

10.5

1.45

7.6-11.8

14.2

1.38

11.3-17.0

13.3

2.09

8.6-16.1

18.2

2.72

14.5-29.4

17.2

2.72

12.5-27.2

15.2

1.23

11.0-16.9

14.2

1.38

12.6-15.7

8.0

0.77

7.8-8.5

7.3

0.84

6.8-8.1

CLIMATOLOGICAL SIGNIFICANCE OF ALBEDO
TABLE 2 (Continued)

1139

Date

Time,
CSTi

Area Description

Standard
Mean Deviation Range

1023-1029

1334-1344

June 15, 1963 0905-0945

June 29, 1963 0900-0920

0940

0950

1035

1145-1200

1200-1245

1245-1255
1325-1410
1420
July 1, 1963 1108-1112

1115-1121

1150-1210

1225-1238

1245
1630

Oct. 15, 1963 1407-1447

Forests, lakes, and some open areas near

Cranberry Portage
Dubawnt Lake, ice, light colored, heavy

stratus
Cultivated fields South of Winnipeg,

prairies, no lakes, clear sky
Tundra, open lakes, clear sky
Tundra, open lakes, clear sky
Tundra, open lakes, clear sky
Lakes and Tundra
McLeod Bay, open but with some small ice

floes, clear
Rocky tundra and lakes north of Great

Slave Lake, most lakes open, some

partly frozen
Rocky tundra north of Great Slave Lake,

some lakes open, more frozen, clear
Mostly tundra, some small lakes and some

ice, clear
Lakes with fairly dark ice, tundra small

lakes open, stratus
Mostly light brown and green tundra, very

little water, just south of Aberdeen Lake,

clear
Mostly light brown and green tundra, but

more lakes and ice. Just north of Aber-
deen Lake, clear
Greenish tundra with some sandy areas,

lakes, 0.5 ice north of Gary L., clear

(some cumulus)
67°25'N, 97°-98°W. Dark brown tundra,

lakes, partly open, some light ice, some

snow banks
Sea ice in Chantrey Inlet, clear
Hudson Bay, open waters, under thin

stratus near Churchill
63°30' to 64°30'N, tundra, frozen lakes,

snow cover, under stratus

15.2
53.1

17.7

12.7
12.7
13.1
11.6

6.6

1.95

5.93

1.32
1.34
1.70
1.72
1.79

0.028

12.7-18.0
41.8-64.2

15.2-21.1
9.4-14.8
8.2-15.2

10.1-15.3
8.0-14.8

6.1-8.7

14.6

2.57

8.9-28.7

24.9

8.82

15.2-47.1

16.2

1.10

14.1-20.4

32.7

9.54

20.1-48.7

14.8

0.95

13.5-15.6

14.9

2.38

13.0-20.2

16.4

4.29

12.0-31.6

16.8

2.72

13.2-25.7

47.5

1.84

45.5-49.8

10.9

0.539

9.7-11.6

77.7

3.45

70.5-83.2

ay be estimated by (3 X time

1 Where time intervals are given distance of transect in nautical miles may be estimated by
interval in minutes).

2 All values obtained at aircraft height of 1000 ft or lower except where indicated.

3 Aircraft below cloud level in all cases.

thrqe regions mentioned above were selected.
The mean, standard deviation, relative variance,
and range of the albedo for each of these sec-
tions are given in Table 1 and shown graphically
in Figures 1, 2, and 3. Additional values of
albedo for various regions are presented in
Table 2. Relative variance, or coefficient of vari-
ability (see Steel and Tome [1960] and Kung
et al. [1964]), which is the ratio of standard
deviation to mean in per cent, expresses gen-

erally the variation of the reflectivity caused by
the heterogeneity of the surface.

The tundra region of Canada, because of the
general absence of trees, exhibits the highest
seasonal range of albedo of the three major
areas studied (see Figure 1). The numerous
lakes in this region do not, however, have a
large effect on the horizontal variation of albedo
in the winter or the summer seasons. During the
winter when all lakes are frozen, the surface

1140

McFADDEN AND RAGOTZKIE

can be considered continental, and, with a snow
cover, it has a high albedo with a low coefficient
of variability. During the summer the albedo of
the tundra is low enough so that the addition
of lakes causes only a slight decrease in the
mean reflectivity of the surface. However, in
spring and autumn, when the lakes are partially
frozen with no snow cover on the ground, there
is a pronounced lake effect on the albedo. During
the spring and autumn, the albedo is extremely
variable, and the standard deviation and rela-
tive variance are very high.

In the boreal forest region (Figure 2) the
seasonal changes of albedo are not as great as
in the tundra, but the lake effect is larger as
indicated by wider ranges of values and gen-
erally higher coefficients of variability. Consid-
ering the nature of the terrain this is under-
standable. The crowns of the trees do not
support the generally dry and powdery snow, and
the forest appears dark during both winter and
summer, particularly when viewed from the di-
rection parallel with the incident solar radia-
tion. Interspersed throughout the forest are
numerous lakes and bogs. In the winter when

these are frozen and snow covered, they appear
as light areas, raising the average albedo of the
region by a factor of 2 or more compared with
forest areas without lakes and bogs (Figure 4).
As mentioned earlier, the method of obtain-
ing albedo information for an essentially uni-
form or isotropic surface is not too critical. For
a snow covered forest that is clearly anisotropic,
as Figure 4 shows, careful consideration must be
given to whether the values of reflected radia-
tion were obtained with a beam or a hemi-
speric radiometer. The reflectivity of a snow
covered forest of the type observed in central
Canada is much higher when viewed from above
than from an angle, because more of the lighter
snow covered ground is visible. A beam solarim-
eter, then, would probably give higher reflected
values over this region than a hemisphere de-
vice, because the hemisphere instrument also
'sees' the darker ground at an angle. Actually,
the surface might absorb more energy than
either albedo method would indicate because
the hemispheric solarimeter also receives most
of its energy from the vertical. When consider-
ing the amount of energy absorbed by the boreal

Fig. 4. View of snow covered terrain in boreal forest region.

CLIMATOLOGICAL SIGNIFICANCE OF ALBEDO

1141

forest and by the tundra in the autumn and
spring, the albedo measurements obtained by
either method, and in particular by the beam
method, might tend to indicate smaller absorbed
energy differences between the two regions than
actually exist.

Using values of direct and diffuse solar radia-
tion from Bernhardt and Phillips [1958], an
observed mean surface albedo value under a
stratus cloud condition for diffuse radiation, and
a computed estimate of direct radiation, based
on a knowledge of the albedo of the trees and
snow covered lakes and bogs, it is possible to
conclude what the true albedo of an area in a
boreal forest region similar to that shown in
Figure 4 would possibly be in early November
at 55°N latitude under a cloudless sky. An ob-
served value of 34.5% (Table 1) was used in
the estimate for the albedo of this area for
diffuse radiation. For the direct radiation a
value of about 28% was computed, using albedo
values of 12% for the conifers and 77% for the
snow covered open areas, and a land cover
ratio of 3 to 1 in favor of the trees. At that
time of year and for that location, half the
radiation incident on the earth's surface is direct
and half is diffuse. Thus, the estimated true
albedo would be approximately 31%, indicating
that the measured mean values (Table 1) are
possibly too high by a factor of about 35%.

Albedo values obtained from the northern
hardwood-conifer forest and farm land mixture
region (Figure 3) are included primarily for
comparison with values obtained over similar
areas by Bauer and Dutton [1960] and Kung
et al.. [1964]. Kung et al., for example, give
November albedo values (converted by means
of the ^eam-to-hemispheric calibration factor of
1.294) for Wisconsin farmland of approximately
18 to 22%, which compares favorably with
values given in Table 1, also obtained during
November.

Climatological Significance

Of significance to the meteorologist are the
effects that seasonal changes of albedo have on
regional climate and general circulation patterns.
For example, Bryson and Lahey [1958] have
suggested that a rapid and drastic change of the
albedo of the tundra in June might trigger the
change from one natural season to another. By
using albedo values given above, snow observa-

TABLE 3. Effect of Observed Albedo on Effective
Solar Radiation for Tundra Region

June 1

June 21

Albedo of lnnd, %

83

15

Albedo of lakes (15% of

surface), %

83

40

Solar radiation at surface

under average cloud con-

ditions (estimated from

Bernhardt and Philipps

[1958]), ly/day

352.8

446.4

Radiation absorbed per unit

area, ly/day

Land

60.0

379.4

Lakes

60.0

267.8

Land-lake combination

60.0

362.6

tions from Operation Breakup 1963 \McFadden,
1965], and solar radiation values estimated from
the results of Bernhardt and Phillips [1958],
it is possible to estimate the change in absorbed
solar radiation between snow and no snow con-
ditions. The results shown in Table 3 were com-
puted for the tundra area of northern Canada
lying south of the sixty-eighth parallel and west
of Hudson Bay.

On the flight of May 22, 1963, the snow line
was observed some 250 miles south of the forest-
tundra border, and on the flight of June 14, no
snow was observed as far north as 64°N, or 450
miles north of the May 22 snow line position.
No snow was observed on the July 1 flight to
the Arctic Ocean. It is reasonable to assume
from these three observations that the tundra
was still completely snow covered as late as
June 1 and probably snow free no later than
June 21. During this 3-week period the albedo
of the tundra dropped from an average of
approximately 83% to about 20% (15% for
the land surface and about 40% for the partly
open lakes that constitute up to 15% of the
tundra surface). This decrease in the reflectivity
of the surface plus the increase in incident radia-
tion at the surface between June 1 and 21 re-
sulted in a 600% increase in the absorbed radia-
tion of from 60 to 363 ly/day during this period.

Whether this sudden and drastic change of
albedo actually served as a triggering mecha-
nism for the atmosphere will have to await
further study, but it seems plausible that this
tremendous increase of energy occurring sud-
denly and at about the same time in all the

1142

McFADDEN AND RAGOTZKIE

TABLE 4. Heat Balance Estimate for
Tundra Region 1963, ly/day

June 1 June 21

Incoming solar radiation 352 . 8 446 . 4
Reflected solar radiation —292.8 —66.9
Effective long wave radiation — 131 . 0 — 132 . 5
Net radiation -71.0 247.0
Heat storage (land) 2.5
Heat required to melt perma-
frost (0.5 cm/day) 36.0
Heat for evaporation (latent)* —160.4
Sensible heat* -48.1

* Using Bryson and Kuhn's Bowen ratio value
(0.30).

tundra regions of the northern hemisphere
should produce a noticeable effect.

To obtain some idea of the partitioning of
this energy absorbed by the tundra and the
amount available as sensible and latent heat,
the heat balance for the area was estimated
(Table 4).

The equation for estimating the net radiation
at the tundra surface is given by

RN = Rr + Rr + Ri

where

RN = net radiation.
Ri = incoming short-wave radiation.
RR = reflected short-wave radiation.
RLelt = effective long-wave radiation.

The values of incoming solar radiation and
albedo are the same as given in Table 3, and the
formula for computing effective long-wave ra-
diation is that of Budyko [1958].

The heat balance equation for estimating
sensible and latent heat transfer over the land
surface is given by

S+M = RV+Q + E

where

S — heat stored in the soil.
M = permafrost melt (estimated at 0.5 cm/

day from in situ measurements).
Q = sensible heat.
E = heat for evaporation (latent heat).

For estimating the storage and melting terms
of the heat budget equation, Larsen's [1965]
values for permafrost depths and soil tempera-

tures in tussock muskeg material for the period
June 1-21, 1963, near Ennadai, Northwest
Territories, were used. The July Bowen ratio
estimate (0.30) of Bryson and Kuhn [1962]
for the region from Norman Wells, Northwest
Territories, to Fort Smith, Alberta, was used for
determining the sensible and latent heat terms
because a ratio estimate for June was not avail-
able.

Applying the estimated value for sensible
heat transfer to the atmosphere, 48.1 ly/day,
to a column of air 1000 meters in height and
with a mean temperature of 0°C gives a heating
rate of approximately 1.6°C per day. This heat-
ing rate does not appear to be unrealistic. From
radiosonde data obtained at Baker Lake on
June 20, 1963, the average temperature increase
between 0600 and 1800 hours CST in the lower
1000 meters was computed to be 1.3°C. An
average temperature increase of 1.85°C per day
for this layer was also computed for the period
between 1800 hours on June 17 and 1800 hours
on June 21. This increase agrees quite well with
the results in Table 4. Winds at this station
during this period were very light to calm and
more northerly in frequency.

Conclusion

Whereas the major change in the albedo of
central Canada was a result of the presence or
absence of snow, the effects of lakes were strik-
ingly different in the tundra and boreal forest
regions. Lakes caused very little horizontal vari-
ation in either the summer or winter in the
tundra, but they produced a large effect during
the transition months of fall and spring. In the
boreal forest region lakes also had little effect
on the range of albedo in the summer. Small
horizontal range was noted during the fall as
lakes were freezing and prior to the first snow.
The largest variation of albedo was obtained
over this region in the winter, when the lakes
were frozen and there was a good snow cover
on the ground. A large range was also noted in
the spring after the snow had melted but before
the lakes had thawed.

The rapid disappearance of snow from the
tundra observed in June 1963 caused an esti-
mated 600% increase in the absorbed radiation
at the surface. Whether this rather sudden and
drastic change in absorbed radiation occurring
over the entire tundra region had any effect on

CLIMATOLOGICAL SIGNIFICANCE OF ALBEDO

1143

the general circulation pattern must await fur-
ther study. The estimated heating of 1000
meters of atmosphere by an average of 1.6°C
per day during this period agrees fairly well
with radiosonde observations at Baker Lake,
Northwest Territories.

Acknowledgments. The authors are grateful to
Professors Bryson and Horn for their advice and
criticism, to the U. S. Navy, in particular the
officers and crew of P2V 128362, and to Messrs.
Bernhard Lettau, David Drury, James Peterson,
Wayne Wendland, Dennis Finke, Ben Bullock, and
Robert Knollenberg for their efforts toward col-
lecting and reducing the data.

This research was supported by the Geography
Branch, Office of Naval Research Contract Nonr
1202 (07) with the Department of Meteorology,
University of Wisconsin.

References

Bauer, K. G., and J. A. Dutton, Flight investiga-
tions of surface albedo, Tech. Rept. 2, Contract
DA-36-039-SC-S0282, Department of Meteorol-
ogy, University of Wisconsin, Madison, 1960.

Bauer, K. G., and J. A. Dutton, Albedo variations
measured from an airplane over several types of
surface, J. Geophys. Res., 67(6), 2367-2376, 1962.

Bener, P., Untersuchung iiber die Virkungsweise
des Solarigraphen Moll-Gorezynski, Arch. Mete-
orol., Geophys., Bioklimatol., Ser. B, 2, 188-249,
1951.

Bernhardt, F., and H. Phillips, Die Raumliche
und Zeitliche Verteilung der Einstrahlung, der
Ausstrahlung und der Strahlungsbilanz im
Meeresniveau, 1, Die Einstrahlung, Abhandl.
des Meteor ologischen Hydrologischen, Dienstes
der Deutschen Demokratischen Republik, NR.
45, Akademie-Verlag, Berlin, 1958.

Bryson, R. A., and P. M. Kuhn, Some regional
heat budget values for northern Canada, Geo-
graph. Bull. 17, 57-66, 1962.

Bryson, R. A., and J. F. Lahey, The march of the
seasons, Final Rept. Department of Meteorol-
ogy, University of Wisconsin, Madison, Contract
AF 19-(604)-922, 1958.

Budyko, M. I., The heat balance of the earth's
surface, (translated from Russian), Office of
Technical Services, U. S. Department of Com-
merce, Washington, D. C, 1958.

Dutton, J. A., An addition to the paper 'Albedo
variations measured from an airplane over sev-
eral types of surface,' J. Geophys. Res., 67(13),
5365-5366, 1962.

Fritz, S., The albedo of the ground and the at-
mosphere, Bull. Am. Meteorol. Soc, 29, 303,
1948.

Hanson, K. J., and H. J. Viebrock, Albedo meas-
urements over the northeastern United States,
Monthly Weather Rev., 92(b), 223-234, 1964.

Kung, E. C, R. A. Bryson, and D. H. Lenschow,
Study of a continental surface albedo on the
basis of flight measurements and structure of the
earth's surface cover over North America,
Monthly Weather Rev., 92(12), 543-564, 1964.

Larsen, J. A'., The vegetation of the Ennadai Lake
area, N.W.T., Studies in subarctic and arctic
bioclimatology, Ecological Monographs, 35, 37-
39, 1965.

McFadden, J. D., The interrelationship of lake
ice and climate in central Canada, Tech. Rept.
20, ONR Contract Nonr 1202(07), Department
of Meteorology, University of Wisconsin, Madi-
son, 1965.

Steel, R. G. D., and J. T. Torrie, Principles and
Procedures of Statistics, McGraw-Hill Book Co.,
New York, 1960.

(Received August 25, 1966.)

Reprinted from JOURNAL OF GEOPHYSICAL RESEARCH
Vol. 72, No. 2 The American Geophysical Union

8

Journal of Geophysical Research

Vol. 72, No. 2

January 15, 1967

Dissolved Chemical Substances in Compacting Marine Sediments1

William A. Anikouchine

Joint Oceano graphic Research Group
Institute for Oceanography, ESSA
University of Washington, Seattle

A mathematical description of the distribution of a dissolved chemical species in inter-
stitial water of clayey marine sediments is obtained by applying the equation of the distribu-
tion of a scalar variable to a constantly accumulating column of marine sediments. An equa-
tion describing compaction, and hence interstitial water velocity, is obtained empirically
from porosity data. When the space coordinates are transformed to the moving sediment-
water interface, the advection term vanishes and a simple equation involving diffusion, re-
action, and local change describes the distribution of chemical species dissolved in interstitial
water. This equation is solved to obtain a steady-state distribution of dissolved silicate and
the diffusive flux of dissolved manganese across the sediment-water interface, and agree-
ment between theoretical predictions and empirical data is found. The validity of assump-
tions used in developing the mathematical model is discussed.

Introduction

Several mathematical models have been de-
vised to describe the depth distribution of con-
centration of various chemical species dissolved
in the interstitial water of marine sediments
[Koczy, 1961; Goldberg and Koide, 1963;
Berner, 1964, 1966]. These models include the
effects of diffusion, reaction, and moving space
coordinates, but no provision has been made
for the movement of interstitial water caused
by compaction of the sediments. This provision
is included in the model described in this
paper.

The procedure used is (1) to state a dis-
tribution equation in terms of coordinates fixed
with respect to a fixed basement, (2) to derive
an advection term from an empirical porosity-
depth relationship, (3) to transform to moving
coordinates to include the effects of sediment
accumulation, (4) to solve the resulting equa-
tion under steady-state conditions, and (5) to
compare the solution with data on interstitial
water.

Development of the Model

The distribution of dissolved chemical spe-
cies in interstitial water is dependent upon

1 Contribution 398 from the Department of
Oceanography, University of Washington; JORG
contribution 6.

diffusion, advection, and reaction within the
interstitial water (or between the interstitial
solution and sediment grains). The dependence
is expressed by the equation for local change
dc/dt of concentration:

The second-order term describes Fickian dif-
fusion with a constant coefficient of diffusivity,
D. Advection at a point h above the fixed
basement is described by the product of- vK, the
velocity of expelled interstitial water, and dc/
dh, the concentration gradient. The reaction
term is the product of the first-order reaction
rate constant fo and the extent-of-completion
parameter (c — cr), which is the difference
between c, the concentration at h, and the
final or saturation concentration c,.

The combined effects of sediment accumu-
lation and compaction of the sediments cause
the sediment-water interface to move upward
from the fixed basement. In a hypothetical case
(Figure 1) the interface moves upward at a
constant rate K after sufficient time has
elapsed. Experimental evidence [Long, 1961]
suggests that clayey sediments behave simi-
larly. Because sediment properties are normally
measured relative to the sediment-water inter-
face, the space coordinate of (1) is transformed
to move with the interface by substituting

505

506

WILLIAM A. ANIKOUCHINE

ZONE OF
.COMPACTION
MOVING UPWARD
THROUGH SPACE
AT A CONSTANT
RATE, K

FIXED BASEMENT
AT h = 0

V

COMPACTION IS
ESSENTIALLY
COMPLETE IN
THIS REGION

' i > —t £

* '•" I"" '■'

_l I I I I I I L

U AFTER THIS TIME, THE

r*^ INTERFACE BUILDS UPWARD AT
ALMOST A CONSTANT RATE, K

Fig. 1. Compaction behavior of a hypothetical column of sediment. Solid lines are tra-
jectories of individual sediment particles. Vertical distances between adjacent dotted curves
represent the sediment porosity at that time and place.

z = h - Kt (2)

t = T (3)

The change with time T of concentration c at
a depth z below the interface is described by
the transformed equation :

dc dc d2c dc . s ( .

w-KTz=D-&-v7z-k^-Cf) (4)

The interstitial water velocity (with respect
to the interface v) is determined by consider-
ing that the sediment compacts so that the
profile of porosity n with depth z remains un-
changed as indicated in Figure 1. Porosity is

piaue.

assumed to vary with depth as follows [Athy,
1930] :

n = n0 exp —(mz) (5)

The porosity at the interface n* and the com-
paction factor m are derived empirically.
The continuity equation

dn
dt

ds
dz

(6)

relates the time rate of change of porosity and
the gradient of superficial velocity s. The su-
perficial velocity is the flow of interstitial water
per unit cross-sectional area of sediment. The

DISSOLVED CHEMICALS IN MARINE SEDIMENTS

507

rate of change is found, from (2) and (5), to
be

dn
dt

— mKn0 exp (mz)

(7)

Substituting (6) and integrating yields

s = Kn0 exp (mz) + d (8)

The constant of integration Cj is evaluated by
assuming that the sediments are infinitely thick
and are totally compacted at depth, so that

s — > 0 z — > - oo (9)

Under these conditions, Ci is zero and

s = Kn0 exp (mz) (10)

The interstitial water velocity v and the super-
ficial velocity s are related by the equation

v = s/n

(11)

Hence (10) becomes

v = K (12)

and the distribution equation (4) becomes

£-*§-**-«) 03)

The advection term disappears because move-
ment of water toward the interface is count-
ered by interface movement during sediment
accumulation.

Equation 13 is solved by assuming that a
steady state exists and that diffusion balances
reaction. The boundary conditions are

C—*Ct g — > — oo

(14)

C = Co 2=0

(15)

and the solution is

f^f- = exp [z(h/D)1/2]

C0 — Cf

(16)

Discussion

Several assumptions are made in obtaining
equation 16:

1. The rate of accumulation of sediments is
constant.

2. Sediments are homogeneous and have
uniform mineral and chemical composition.

3. Interstices are filled with liquid, and no
gas phase is present.

4. There is lateral homogeneity in the sys-
tem, and a one-dimensional treatment is ade-
quate.

5. The effects of temperature and pressure
are negligible.

6. A steady state is attained ultimately.

7. Chemical reactions obey first-order ki-
netics.

8. The diffusivity coefficient is constant.

9. Interstitial water flow is laminar.

10. Sediment grains and interstitial water
are incompressible.

11. After sufficient time, the sediment-water
interface moves upward at a constant rate, and
the depth profile of porosity remains station-
ary with respect to the interface.

12. The empirical porosity-depth relation-
ship applies to clayey marine sediments.

13. The sediment column can be approxi-
mated by an infinitely thick column.

Assumptions about the physical nature of the
system are most nearly valid for sediments in
an environment free of disruptive effects such
as variable sediment supply and changes in
thermal conditions. Assumptions about the
kinetics of chemical reactions are valid for
reactions involving substances that are present
in excess. Although the reactions in the system
modeled here are heterogeneous, empirical first-
order kinetics are suggested by the apparent
success of the model in the case of dissolved
silicate considered later in this paper. The re-
maining assumptions are necessary to provide
a mathematically tractable model.

A dimensionless form of (13) is

aZ dr

It is obtained by applying the transforms

(17)

r = ktT

(18)

Z ~ KZ

(19)

c-c~Cf

Co Cf

(20)

The dimensionless constant,

represents the ratio of advective effects to dif-

508

WILLIAM A. ANIKOUCHINE

Fig. 2. Theoretical distribution of silicate in
interstitial water fitted to data of Siever et al.
[1965].

fusion and reaction. This ratio has a value of
about 3 X 10"5 in the case to be considered,
which indicates that a simple diffusion-reaction
model should suffice to describe the distribution
of dissolved substances in interstitial water.

Comparison of Theoretical Result
and Empirical Data

The validity of (16) can be examined by-
comparing results based on (16) with empirical
data from the literature. Measurements of sili-
cate concentrations in interstitial water in sev-
eral sediment cores were reported by Siever
et al. [1965]. Two of these cores that exhibit
uniform lithology and low degree of scatter of
silicate concentrations were compared with the
theoretically derived silicate concentration dis-
tribution. The cores are L-139, a diatomaceous
clay from the Gulf of California, and 258-5, a
gray clay (upper part only) from the con-
tinental shelf off Cape Cod, Massachusetts.
The silicate concentrations in these two cores
are fitted (Figure 2) by the curve

c = 63 - 26 exp (1.9 X 10~2z) (21)

If we assume a diffusivity coefficient of 0.3 X
10"* cm3 sec"1, the reaction rate given by (16)
is 10 X 10"10 sec"1. This is smaller than the re-
sults of Grill [1961] and Lewin [1961], who
obtained constants of between 10"7 and 10"8
sec"1 in vitro silica dissolution experiments. It
appears that the dissolution of silica in in-
terstitial water is suppressed, perhaps by large

concentrations of dissolved iron and aluminum
[Lewin, 1961].

The model can be used in calculating the
diffusive flux of a substance dissolved in in-
terstitial water. The appropriate expression is

-Z)|= -{kxDY%0 -C/)

•expKV-D)"2] (22)
To illustrate the use of (22), we consider the
distribution of dissolved manganese. If ex-
tremely slow dissolution (k\ = 10"12 sec"1) is
assumed to control the distribution of dissolved
manganese, the diffusive flux of manganese at
the sediment-water interface is 1.7 X 10"" g
cm"2 sec"1. This means that each year about
0.5 gram of manganese is transferred from the
sediments to each square meter of bottom. The
values for c0 and cf used in this calculation,
10"* g cm"3 and 10"3 g cm"3, are the concen-
trations observed in seawater [Harvey, 1963]
and in argillaceous rocks [Rankama and Sa-
hama, 1950]. The diffusivity coefficient is as-
sumed to be 0.3 x 10"5 cm2 sec"1.

If the manganese in the interstitial water
precipitates as nodules upon contact with the
seawater and if it is assumed that the diameter
of the nodules is 6 cm, their density is 2.8
g/cm3, and the manganese content is 19%,
about 14 x 103 years would be required for
the accumulation of 120 nodules/ma. The
growth rate of these nodules is about 2 mm/
1000 yr, a little larger than the rates (0.01 to
1 mm/1000 yr) reported for rapidly growing,
shallow-water nodules [Manheim, 1965]. These
results suggest that, within the limitations of
the model developed here, the source of man-
ganese nodules is the interstitial water of the
underlying sediments.

Conclusions

If compaction is included in a model of the
interstitial water system, a simplification of
the mathematics results even though advection
caused by compaction is a negligibly small ef-
fect. The diffusion-reaction, steady-state model
yields distribution profiles that agree with
empirical data on dissolved silicates and growth
of manganese nodules. If assumptions made in
developing the model are valid, empirical con-
centration profiles can be examined for anoma-

DISSOLVED CHEMICALS IN MARINE SEDIMENTS

509

lies. In this way the model can be used in
interpreting the geochemistry of marine sedi-
ments.

Acknowledgments. I should like to thank Dr.
M. G. Gross for his critical reading of the manu-
script and for his helpful suggestions.

References

Athy, L. F., Density, porosity and compaction of
sedimentary rocks, Bull. Am. Assoc. Petrol.
Geologists, 14, 1-24, 1930.

Berner, R. A., An idealized model of dissolved
sulfate distribution in recent sediments, Geo-
chim. Cosmochim. Acta, 28, 1497-1503, 1964.

Berner, R. A., Chemical diagenesis of some mod-
ern carbonate sediments, Am. J. Sci., 264, 1-36,
1966.

Goldberg, E. C, and M. Koide, Rates of sediment
accumulation in the Indian Ocean, in Earth
Science and Meteoritics, compiled by J. Geiss
and E. D. Goldberg, pp. 98-100, North Holland
Publishing Company, Amsterdam, 1963.

Grill, E. V., A chemical study of nutrient regen-
eration from phytoplankton decomposing in
sea water, M.S. thesis, Department of Oceanog-
raphy, University of Washington, Seattle, 1961.

Harvey, H. W., The Chemistry and Fertility of
Sea Waters, 2nd ed., pp. 146-147, Cambridge
University Press, 1963.

Koczy, F. F., Radioactive tracers in oceanogra-
phy, Intern. Union Geodesy and Geophys.
Monograph 20, 1961.

Lewin, J. C, The dissolution of silica from diatom
walls, Geochim. Cosmochim. Acta, 21, 182-198,
1961.

Long, D. V., Mechanics of consolidation with
reference to experimentally sedimented clays,
Ph.D. thesis, California Institute of Technology,
Pasadena, 1961.

Manheim, F. T., Manganese-iron accumulations
in the shallow marine environment, Occasional
Publ. 30, pp. 217-276, Narragansett Marine Lab-
oratory, University of Rhode Island, 1965.

Rankama, K., and Th. G. Sahama, Geochemis-
try, p. 652, The University of Chicago Press,
1950.

Siever, R., K. C. Beck, and R. A. Berner, Com-
position of interstitial waters of modern sedi-
ments, J. Geol., 78, 39-73, 1965.

(Received July 25, 1966.)

Reprinted from MARINE GEOLOGY Vol. 5
Marine Geology - Elsevier Publishing Company, Amsterdam - Printed in The Netherlands

EVIDENCE FOR TURBIDITE ACCUMULATION IN TRENCHES IN THE
INDO-PACIFIC REGION1

WILLIAM A. ANIKOUCHINE AND HSIN-YI LING

Joint Oeeanographic Research Group, University of Washington, Seattle, Wash. (U.S.A.)
Department of Oceanography, University of Washington, Seattle, Wash. (U.S.A.)

(Received June 29, 1966)

SUMMARY

Sedimentary evidence in three cores taken from the Java, Mindanao and
Mariana Trenches indicates that the Java and Mindanao Trenches are receiving
sediments that are mostly turbidites. The Mariana Trench is receiving mostly pelagic
muds with minor, but significant additions of sediments transported by turbidity flows.

INTRODUCTION

In 1964, cores (Fig.l) from the Java, Mindanao and Mariana Trenches, were
taken from the ship U.S.C. and G.S. "pioneer" as part of the International Indian
Ocean Expedition. The sediments in these eores exhibit unexpected textural and
compositional features that indicate they are turbidites.

METHODS

The cores were taken with a large-diameter (3-inch) piston corer operated
without a piston deactivator. The lower portions of the cores were disturbed, and
were not analyzed. Analyses were performed by the senior author at the Pacific
Oeeanographic Laboratory (POL), Institute for Oceanography, Environmental
Science Services Administration, Seattle, Washington. The upper portions were
logged (U. S. Department of Commerce, 1965), analyzed granulometrically, and
examined under a stereomicroscope ( X 60). A summary of these data is presented
in Fig.2, 3 and 4. The cores, both reference and working halves, are stored at POL.

1 Contribution No. 2, Joint Oeeanographic Research Group and Contribution No. 389, Department
of Oceanography, University of Washington, Seattle, Wash. (U.S.A.).

Marine Geo!., 5 (1967) 141-154

142

W. A. ANIKOUCHINE AND H. Y. LING

140°

CHALLENGER

SNELLIUS EXPEDITION CORES

SWEDISH DEEP SEA

EXPEDITION CORES
PIONEER (USCaGS)

TRENCH CORE POSITIONS

^MINDANAO
'/, TRENCH

/264
■ 108

-271

225 ,/M

MARIANA -
TRENCH

<o '&*§

*^_

Fig.l. Location of cores taken in the Java, Mindanao and Mariana Trenches.
Features diagnostic of turbidites

Similarity between the sediment properties described on the following pages
and the properties of turbidites listed by Kuenen (1964a) is presented as evidence
of turbidites in the Indo-Pacific Trench cores.

Maximum grain sizes

The coarsest particles in the trench sediments fall within Kuenen's (1964)
size limits. Grain sizes at the first percentile are coarser than very fine sand (4.5 phi)
in Java and Mariana Trench sediments, and fine sand (3.5 phi) in Mindanao Trench
sediments.

Grading

Pronounced grading occurs in the Java Trench core. A bed grades 2.3 phi
units over 10 cm in one case. Less well-defined graded beds occur throughout the
core. Size grading was not observed in the sediments in the Mindanao Trench core;
however, the contacts between beds of coarse and of fine texture are often transitional
over several centimeters. This grading of texture can be considered as graded bedding
in this core. In the Mariana Trench core, the size and abundance of mud lumps grades
uniformly in beds up to 20 cm thick. There is no corresponding gradation in mean
grain size in these sediments.

Marine Geo!., 5 (1967) 141-154

INDO-PAOTIC TURBIDITES

143

/<?/&/
&&%&'

A//

'$/s3

WiS£.

ffr/x.

W*M&ffito&,.

#1

<9

w4m/£MmMZmw/f

0

20

40

60

80

E
o

100

2

120

I

Q.
UJ
Q

140
160

180

200

220

240

260

5Y6/I

5GY6/I

> - ,

2 —

29 —

59 _
63 —
69 —
83 —

91 —

119 —

151 —

178 —

210 —

249 —

7.58

8.05

7.72
6.61
5.35
4.69
7.48

7.43

7.84

7.89

623

7.64

□

LUTITE

SILT

HEAVY FRACTION

LIGHT FRACTION

TOTAL FRACTION

PRESENT

ABUNDANT

FLOOD

TRACE

|2T
3H
/3L
4T

t

X

X

A

28.5

?

A

X

F

F

t

t

i2T
/3T

I/4T

A

X

t

36.2

F

///3H

//3L

f//4T

t

t

X

30.0

X

t

t

///3H
//3L
I/4H
i 4L

X

2.16 'I

X

X

I7.0

A

t

2.04 ']

A

A

X

/3H
/3L
' 4T

X

X

t

t

8.2

2.27 %

F

t

t

1.63 '

1.04*;

/3H
' 3L
\4T

F

F

t

2.6

0.57 C

F

t

X

3H
\ 3L

X

t

2.23 k

X

X

A

1

t

32.3

X

X

X

X

A

\

t

t

t

I 2T
\ 3H
\3L
\\4T

X

X

X

X

t

A

28.6

2.41 k,

X

X

A

t

A

A

X

A

X

3H

\ 3L
\\4H

\\4L

X

2.08v

X

A

t

t

X

t

39.0

X

X

X

X

A

t

A

A

l.70v

\3H
\\3L

A

X

X

X

t

t

X

X

X

35.5

X

t

t

F

X

t

t

2.21 .

\3H
,\3L
\*T

X

X

t

t

I4.9

A

t

X

\ 3H
\ 3L
\4T

X

X

A

30.0

A

A

Fig.2. Java Trench core (PI-442-64-33) sediment analysis data. Numerical values in mean grain size,
sorting and size fraction columns are in phi units.

Marine Geol, 5 (1967) 141-154

144

W. A. ANIKOUCHINE AND H. Y. LING

/3H

X

3L

X

t

A

35.8

\ 4H

X

\4L

X

X

A

t

^3H

X

X

3L

X

X

1 1.9

\ 4H

X

X

X

A

X

\4L

t

A

A

^2T

F

X

v 3T

A

X

X

26.4

\4T

A

A

t

t

t

/2T

X

t

X

/3T

A

A

X

t

X

35.0

/ 4T

X

t

/

2T

7.4

\4T

X

A

A

t

X

A

/IT

F

X

' 2T

A

F

30.1

\ 3T

iA_

A

\4T

A

t

t

/2T

A

X

\ 3T

A

F

t

46.7

\4T

A

X

X

X

/it

/2T

A

X

t

X

40.1

/ 3T

^A

X

t

X

4T

A

X

X

X

'/3H

X

1

/ 3L

t

F

X

t

t

35.6

4H

t

X

X

X

^4L

X

A

/IT

X

F

X

<* 2T

F

A

47.4

> 3T

X

F

X

X

\4T

X

F

X

X

t

I-2T

X

OT

X

X

IT

t

A

X

33.8

2T

A

X

\ 3T

X

X

A

t

"--\41

F

F

\ IT

A

X

\\ 2T

A

t

t

X

35.8

\\ 3T

A

X

X

t

X

X

t

\\4T

A

X

X

X

X

X

X

X

t

\ 3T

F

X

X

41.5

\4T

F

X

X

X

X

X

I I LUTITE
IH^ MUD LUMPS
EE3 SILT

E551 DIATOM OOZE
P77I WOOD

H HEAVY FRACTION

L LIGHT FRACTION

T TOTAL FRACTION

X PRESENT

A ABUNDANT

F FLOOD

t TRACE

Fig.3. Mindanao Trench core (PI-442-64-35) sediment analysis data. Numerical values in mean size,
sorting ana size fraction columns are in phi units.

Marine Geol., 5 (1967) 141-154

INDO-PACIFIC TURBIDITES

145

0

20

S 40

? 60

I 80
o_

LU 100
Q

120-

140

I0YR6/2
I0YR4/2

I0YR5/2

I0YR4/2

7YR4/2

5YR5/4

13 —
30 —

60-

105 ■

139 —

5.53
4.04

7.74

5.99

6 55
6.01
6.10

4.10

2.95'

4.10,

□

LUTITE

m

MUD LUMPS

m

SILT

H

HEAVY FRACTION

L

LIGHT FRACTION

T

TOTAL FRACTION

X

PRESENT

A

ABUNDANT

F

FLOOD

t

TRACE

f3H

X

A

[3L

X

X

t

t

13.8

4H

X

t

A

t

4L

A

X

X

X

/3H

t

X

X

t

X

t

t

t

t

3L

X

X

t

A

X

t

X

t

t

t

19.4

4T

X

X

t

X

X

A

X

X

/3H

X

X

X

t

;3L

t

A

t

A

t

30.7

4H

X

A

X

A

X

41

1

t

t

A

X

X

X

,'2T

t

X

t

X

X

3H

X

X

X

X

X

22.7

3L

J^

X

X

X

X

X

t

A

X

X

t

4T

X

X

X

X

X

X

t

F

t

t

Ah

A

A

X

A

X

t

3L

X

X

X

X

X

X

X

t

37.7

4H

X

X

X

t

4L

t

X

X

A

X

t

X

X

X

t

3H

F

A

X

X

A

3L

A

X

X

X

A

t

t

t

X

X

29.2

4H

t

A

A

t

\4I

X

X

X

t

X

X

3H

F

A

t

3L

t

A

X

A

t

X

35.1

L4H

1

X

X

X

X

A

t

\4L

X

X

X

t

X

X

3H

X

A

X

t

3L

t

A

X

19.1

i4H

F

X

A

A

X

4L

A

A

A

t

X

A

Fig.4. Mariana Trench core (PI-442-64-36) sediment analysis data. Numerical values in mean grain
size, sorting and size fraction columns are in phi units.

Sorting

The generally poor sorting in these cores does not agree with the moderate to
good sorting cited by Kuenen (1964a, p. 10) as typifying deep-sea sands. The disparity
probably is caused by excessive amounts of lutite in the trench core sediments.

Inorganic constituents

Beside clay minerals, the most abundant component, trench sediments contain
angular quartz, mica, feldspar and pyrite. These are all cited (Kuenen, 1964) as
existing in turbidites. These materials are not diagnostic of their depth of original

Marine Geol , 5 (1967) 141-154

146 W. A. ANIKOUCHINE AND H. Y. LING

deposition but the presence of rock fragments and feldspar suggests that the materials
have undergone few sedimentary cycles.

Origin and distribution of organic constituents

Organic materials in the trench sediments are mostly the remains of siliceous
planktonic organisms. Foraminifera in the Java Trench core occur principally in
fine-textured layers. In the Mindanao Trench core, the diatom, Ethmodiscus rex
(Plate IB, C), occurs only in certain coarse-textured layers (Plate 1A). All other
organic remains (Radiolaria, sponge spicules, .fish teeth and wood occur in both
fine and coarse-textured layers. Fish teeth and fish bone fragments occur throughout
the Mariana Trench core, but most of the Radiolaria occur in layers containing
abundant mud lumps.

Of all organic constituents in these sediments, the wood is unquestionably
of terrestrial origin. Fish bones and sponge spicules might be of neritic origin. It
is difficult to interpret the origin of the few tubular, arenaceous Foraminifera found
in Java Trench core or of the other organic constituents in the Mindanao and Mariana
Trench cores.

Lamination

The sediments from the Java Trench and Mindanao Trench cores are strongly
and irregularly laminated. Laminae range in thickness from a few millimeters to as
much as 20 cm. Alterations in sediment laminae are marked by changes in color and
texture or, in silt streaks, by changes in mean grain size.

Layering in the Mariana Trench core is marked by changes in texture and by
subtle changes in color. Textural changes reflect changes in the amounts of mud
lumps present in the laminae. The thickness of lamina of lutite is similar to the other
trench cores. Silt occurs in streaks about 1 mm thick or layers as much as 20 cm thick.

Distorted beds

These are found in the sediments from the Java Trench and Mariana Trench
cores. Sediments with contorted bedding lie between layers having normal bedding
in the Java Trench core. In the Mariana Trench core there are irregular silty lenses
lying about 50° to the core axis. These lenses lie between layers having bedding normal
to the core axis.

Ripple lamination

A structure resembling cross-bedding occurs in the core from the Java Trench.
Streaks of fine sand, 1-2 cm apart, accentuate bedding surfaces tangential to the
underlying layer and truncated by the overlying layer. This structure is about 5 cm
thick and may be a portion of a large-scale ripple mark. No ripple marks were found
in the other two trench cores.

Marine Geol., 5 (1967) 141-154

PLATE I

H*

A

3*?$t?52E

'

icSt. t?

'•■tf-^w

B

5wSSSBt'j*sC?^fiS8»S!a<<vi'"-

Photograph of a portion of Mindanao Trench sediment core (PI-442-64-35) showing the alter-
nation of lithology. Occurrence of diatom, Ethmodiscus rex, in seidment with coarse texture is
indicated by white arrow. Mud lump layers occur at 56 cm and at 87 cm.

Photomicrographs of central area of Ethmodiscus rex from Mindanao Trench core sample,
137-138 cm below sediment surface; x 250 (phase contrast).

Photomicrographs of central area of Ethmodiscus rex from Mindanao Trench core sample,
137-138 cm below sediment surface; x 770 (phase contrast).

148 W. A. ANIKOUCHINE AND H. Y. LING

Mud lumps

Mud lumps occur in all three cores. They are about 1 mm in diameter in the
Mariana Trench core and about 1 cm in diameter in the Mindanao Trench core
(Plate IA). Only a few large (up to 3 cm in diameter) lumps occur in the Java Trench
core. The lumps have the same color as the surrounding sediment.

An analysis of mud lumps indicated that they are not aggregations of the sur-
rounding fine-textured sediment. Sand-silt-clay ratios for disaggregated mud lumps
are plotted onFig.5, 6 and 7 as open circles. Ratios for the sediments including mud
lumps are plotted as solid circles.

The sediments from the Mindanao Trench core have sand-silt-clay ratios
that fall along the dashed line in Fig.6. Because mud lumps fall in the sand size,
their removal from the sediment should not affect the silt-clay ratio of the remaining
material and the ratio should remain that indicated by the dashed line in Fig.6.
Ratios for disaggregated mud lumps do not fall near the dashed line, indicating a
silt-clay ratio different from that of the complete sediment.

In the sediments from the Mariana Trench core (Fig. 7), there is no linear trend
in sand-silt-clay ratios. However, a line drawn through the sand apex and a solid
circle intersects the silt-clay axis at a point representing a silt-clay ratio lower than
that of the mud lumps removed from that sediment. It appears that these mud lumps
are not formed in situ.

CLAY

SAND^ *SILT

Fig.5. Java Trench core triangular diagram of size components.

Marine Geol., 5 (1967) 141-154

INDO-PACIFIC TURBIDITES

149

CLAY

SAND

SILT

Fig.6. Mindanao Trench core triangular diagram of size components.

CLAY

SAND

SILT

Fig.7. Mariana Trench core triangular diagram of size components.

Marine Geol., 5 (1967) 141-154

150 W. A. AN1KOUCH1NE AND H. Y. LING

Thickness

The thickness of sediment layers interpreted as turbidites varies from a few
millimeters to about 18 cm. If the entire series of alternating layers in the cores from
the Java and Mindanao Trenches represents deposition from a single turbidity flow,
then the thickness of the deposits exceeds the lengths of these cores; i.e., 2.5 and 4 m,
respectively. In the Mariana Trench core, turbidites are a graded mud lump layer 17
cm thick and possibly streaks of silt less than a millimeter thick.

Contacts

In the Java Trench core most contacts are sharp. In only two cases gradational
upper boundaries were observed. The Mindanao Trench core shows sharp contacts
with several gradational upper contacts and a few gradational lower contacts. Both
contacts are gradational in a few layers. Few distinct contacts appear in the Mariana
Trench core; these are sharp rather than gradational.

Burrowing

Evidence of burrowing is restricted to the upper few centimeters of the Java
and Mindanao Trench cores and the upper 70 cm of the Mariana Trench core. The
Java Trench core shows thin zones of oxidized coloration at 40, 70 and 88 cm. These
zones of oxidized sediments represent either pelagic sediments deposited at the top
of turbidite accumulations or the oxidized upper portions of individual turbidite
layers.

Erosion at lower contacts

Many of the thin layers of sediment in the Java Trench core have thin streaks
of coarse mica at their lower contact. In the Mindanao Trench core a few layers
have a unit thickness of mud lumps, diatom frustules, or silt streaks at their lower con-
tact. No such feature occurs in the Mariana Trench core. Kuenen (1964b) interprets
this feature as a lag deposit resulting from erosion of the underlying layer. Although
he refers to erosion of a pelagic layer, one could expect erosion of an underlying
turbidite layer if turbidity flows occurred in rapid succession. The only condition is
that the underlying layer be uncompacted.

Alternations of layers

There is a contrast in the amount of microorganisms in the layers in the Java
Trench core. Layers of fine-textured sediments contain Foraminifera and Radiolaria,
whereas coarse-silt layers with coarse texture and containing considerably more mica
contain only a trace of these organisms.

Marine Geol., 5 (1967) 141-154

INDO-PACIFIC TURBIDITES 1 5 1

The Mindanao Trench core has alternating layering that shows a contrast
among three sediment types. Fine-textured sediment layers are rich in Radiolaria,
the diatom Ethmodiscus rex, and organisms tentatively identified as Foraminifera.
Layers rich in clay lumps are free of diatoms and contain only a few Radiolaria.
Ethmodiscus ooze layers contain a rich microfauna. These ooze layers contain a
large amount of Frustules, have a coarse texture, and lack the rock fragments and
mineral grains present in fine-textured layers. The sediments in the Mariana Trench
core are mainly uniform and finely textured with interbedded, graded layers of mud
lumps and streaks and lenses of silt.

Color bands and variegation up to 1 cm thick in the Java and Mindanao Trench
cores present striking evidence that a nonuniform environment of deposition exists
in these trenches.

Influence of source materials

This influence is striking if the three trench cores are compared. The Java
Trench receives sediments rich in mica. The Mindanao Trench receives sediments
either rich in clay lumps, silt and terrestrial plant debris, or rich in Ethmodiscus rex.
Besides pelagic brown mud, the Mariana Trench receives silt and mud lumps. The
existence of two ore more sediments in two of these cores indicates that more than
one source may supply a single trench even though the diversity of materials may be
transported to the site of deposition by a single mechanism; e.g., a turbidity flow.

Influence of distance to source

Of the three trenches studied, the Mariana Trench is the farthest from a major
land mass. This is reflected by the preponderance of pelagic sediments and the thin-
ness of the silt layers in the Mariana Trench core. The coarseness of the mud lumps
and thickness of the graded mud lump layer in this core suggest that the source of
the mud lumps is much closer.

The Mindanao and Java Trenches lie close to major land masses. In the Java
Trench core, the layers of coarse mica are thicker (up to 10 cm) and closer together.
The interbedded lutite comprises about 80% of the core. The Mindanao Trench
core contains still less fine-textured lutite. The Java and Mindanao Trenches lie about
the same distance from nearest major land masses, therefore other factors such as
frequency or degree of tectonic activity (Bezrukov and Petelin, 1959) and amount
of relief and type of drainage of the land masses must exert influences upon the
nature of turbidite accumulation that override the influence of distance to source.

Pelagic interbeds

The pelagic sediments in the Mariana Trench core are brown mud (clay)
typical of open ocean abyssal regions. The only organic materials found in this

Marine Geo/., 5 (1967) 141-154

152 W. A. ANIKOUCHINE AND H. Y. LING

sediment are fish teeth and bones, and possible faecal pellets. The pelagic sediments
in the Java Trench core are fine-grained, micaceous, gray-green lutites. The Mindanao
Trench core appears to contain few pelagic interbeds. These are fine-grained lutites
containing Radiolaria but few or no diatoms.

Slope of bottom

Slopes were not measured in the areas of the three trench cores. It is likely
that the slope of some of the bottom of the trenches is less than a degree so that
turbidites can accumulate there.

DISCUSSION

The sediments in all three trench cores contain several features typifying tur-
bidites. Although it is not reasonable to identify the sediments in Mariana Trench
core as turbidites, the effects of turbidity flows can be recognized in these sediments.
The abundance of evidence leaves little doubt that the sediments in the Java and
Mindanao Trench cores are turbidite accumulations.

The sediments in the Java Trench core are interpreted as hemipelagic micaceous
lutites alternating with graded and ungraded micaceous silts that are introduced by
turbidity flows. There is only a single source of sediments, presumably the island of
Sumatra. Of the three cores studied, the Java Trench core presents the least complex
example of turbidite accumulation.

The turbidite accumulations in the Mindanao Trench core are the most com-
plex. The sediments in this core are interpreted to be a turbidite sequence of lutites
containing Radiolaria and diatoms interbedded with terrigenous silts, mud lump
accumulations and Ethmodiscus ooze. It is possible that the lutite layers are hemipe-
lagic. However, the occurrence of wood fragments in these layers argues against this
interpretation. It is more likely that lutite is introduced to the core site by turbidity
flows originating close to the coast of Mindanao Island. Mud lumps probably result
from mass movement of preexisting sediments from higher portions of the trench.
Bezrukov (1957) reached the same conclusion regarding the Izdu-Bonin and Mariana
Trenches.

Several observations indicate that Ethmodiscus frustule fragments are trans-
ported to the core site, presumably by turbidity flows. Although the three trenches
are located within the biogeographic distribution region of Ethmodiscus rex (Semina,
1959), the occurrence of Ethmodiscus ooze in these cores and in others from the Indo-
Pacific region (Neeb, 1943; Olausson, 1960; Zhuze et al., 1959) is erratic. The ooze
occurs in the Mindanao Trench core in irregularly spaced layers between 75 and 328
cm. It is rare in the Java Trench core and is absent from the Mariana Trench core.
Neeb (1943) reports no significant amounts of diatoms in cores from the Mindanao
Trench taken during the "Snellius" Expedition (Fig.l), but Kolbe (1954) states that

Marine Geol., 5 (1967) 141-154

INDO-PACIFIC TURBIDITES 153

diatom fragments thought to be Ethmodiscus rex are found in "nearly all cores and
levels of the Equatorial Pacific."

Semina (1959) found no close relationship between the presence of Ethmodiscus
rex in sediments and in the water above. She is inclined to agree with Kolbe (1957),
Hendey (1958) and Zhuze et al. (1959) that there are other factors, e.g., bottom
current, flow of the ooze, or irregular submarine topography, that control the distri-
bution of this diatom ooze on the present-day ocean bottom. Riedel (1954) reexa-
mined the Radiolaria assemblage from the original "Challenger" Station 225
(11°24.0'N— 143°16.0'E; depth: 4,475 m) and concluded that "There is little doubt
that the sediment obtained at "Challenger" Station 225 is an outcrop of Middle
Tertiary Radiolarian ooze. The occurrence of Ethmodiscus rex in this sediment has
no connection with its existence in the overlying water at the present day, but rather
means that the diatom lived in that region during Middle Tertiary time." These
observations are in accord with the concept that Ethmodiscus ooze accumulates
within the limits of its biogeographical region but is transported by occasional
turbidity flows, so that its distribution on the bottom is patchy and its distribution
within bottom sediments is irregular.

The sediments in the Mariana Trench core contain fewer features typical of
turbidites than do the sediments in the Java and Mindanao Trench cores. Neverthe-
less, the features found in the Mariana Trench core, e.g., graded mud lump layers
and silt streaks, are just as important because they show that nonpelagic sediments
with properties characteristic of turbidites can occur in a trench remote from land
and separated from land by several ridges and basins.

It is likely that mud lumps are introduced into the Mariana Trench by mass
movement such as sliding or slumping; the dislocation of slightly compacted clay is
probably the mechanism whereby the mud lumps are formed. The silt in layers and
streaks has been transported from distant terrestrial sources. The mechanisms for
the transportation of this material cannot be specified at present.

acknowledgements

The authors wish to thank D. A. McManus for reading the manuscript. The
work of the junior author was supported by the Office of Naval Research Contract
No.477(37), Project NR 083-012.

REFERENCES

Bezrukov, P. L., 1957. The sediments of the Izdu-Bonin, Marianas and Ryudu deep-sea oceanic

depressions. DokL: Earth Sci. Sect. (English TransL), 114 : 393-396.
Bezrukov, P. L. and Petelin, B. P., 1959. The bottom sediments of the trenches of the western

Pacific Ocean. Intern. Oceanog. Congr., 1, Preprints Abstr. Papers, pp.451-452.
Hendey, N. I., 1958. Diatoms from equatorial Indian Ocean cores. Nature, 181 (4614) : 953-954.

Marine Geol., 5 (1967) 141-154

154 W. A. ANIKOUCHINE AND H. Y. LING

Kolbe, R. W., 1954. Diatoms from equatorial Pacific cores. Rept. Swed. Deep-Sea Expedition,

1947-1948,6(1): 1-49.
Kolbe, R. W., 1957. Diatoms from equatorial Indian Ocean cores. Rept. Swed. Deep-Sea Expedition,

1947-1948, 9 (1) : 1-50.
Kuenen, Ph. H., 1950. Marine Geology. Wiley, New York, N.Y., 568 pp.
Kuenen, Ph. H., 1964a. Deep-sea sands and ancient turbidites. In: A. H. Bouma and A. Brouwer

(Editors), Turbidites. 3. Developments in Sedimentology. Elsevier, Amsterdam, pp. 1-33.
Kuenen, Ph. H., 1964b. The shell pavement below oceanic turbidites. Marine Geol., 2 : 236-246.
Neeb, G. A., 1943. The composition and distribution of the samples. Snellius Expedition, Sci. Results

Snellius Expedition Eastern Pt. East-Indian Geol. Archipelago, 1929-1930. 5. Results. 3. Bottom

Samples, 268 pp.
Olausson, E., 1960. Description of sediment cores from the central and the western Pacific with the

adjacent Indonesian region. Rept. Swed. Deep-Sea Expedition, 1947-1948, 6 (5/8) : 161-214.
Riedel, W. R., 1954. The age of the sediment collected at "Challenger" (1875) Station 225 and the

distribution of Ethmodiscus rex (Rattray). Deep-Sea Res., 1 (3) : 170-175.
Semtna, G. I., 1959. The distribution of the diatom Ethmodiscus rex (Wallich) Hendey among the

plankton. Dokl: Earth Sci. Sect. (English Transl.), 124 (6) : 1309-1312.
U. S. Department of Commerce, 1965. International Indian Ocean Expedition, U.S.C. and G.S.-Ship

Pioneer — 1964. 1. Cruise Narrative and Scientific Results. U. S. Govt. Printing Office, Washing-
ton, D. C, 139 pp.
Zhuze, A. P., Petelin, V. P. and Udintsev, G. B., 1959. Concerning the origin of diatomaceous

muds containing Ethmodiscus rex (Wallich) Hendey. Dokl. Earth Sci. Sect. (English

Transl). 124(6) : 1301-1304.

Marine Geol, 5 (1957) 141-154

10

Reprinted from JOURNAL OF GEOPHYSICAL RESEARCH
Vol. ?2, No. If The American Geophysical Society

Journal of Geophysical Research

Vol. 72, No. 24

December 15, 1967

Heat Flow in the Pacific Ocean off Central California

Robert E. Burns and Paul J. Grim
Institute jor Oceanography, ESSA, Seattle, Washington 9S102

As part of the Upper Mantle Project, a marine geophysical survey has been conducted off
central California, between 35° and 39°N, extending about 800 km offshore. Heat flow measure-
ments were made at thirty-one stations including twelve along a special section between the
highest and the lowest values previously reported. The heat flow is relatively constant (1.7 to
1.8 Meal cm"2 sec"1) over the eastern part of the continental rise; farther offshore a banding of
high and low heat flow on a regional scale (wavelengths of several hundred kilometers) is
present. Shorter wavelength fluctuations (about 50 km) in the heat flow are detected by the
closely spaced stations along the special section.

Introduction

As a part of the Upper Mantle Project
(UMP), the seaward extension of the transcon-
tinental geophysical survey in the Pacific has
been surveyed by the U.S. Coast and Geodetic
Survey. An initial set of measurements was
made from the Pioneer in 1965, and the Sur-
veyor completed the work in the spring of 1966.
Data were collected in an area located between
35° and 39°N extending about 800 km offshore
into the Pacific. The basic series of measure-
ments included measurements of bathymetry,
gravity, total magnetic intensity, and sea floor
heat flow. This report is principally concerned
with the heat flow measurements, but it includes
tentative interpretations as to their implications
in view of the other geophysical parameters.

The area covered by the investigation (Fig-
ure 1) includes several important physiographic
units [Menard, 1964] including the continental
shelf, slope, and the two large sedimentary fans
(Delgada and Monterey) that encroach on the
area of abyssal hills to the west. Also within
the area is the eastern part of the Pioneer ridge,
extending westward from the Delgada fan be-
tween 38° and 39°N. Although there has been
some geophysical work done in this area in the
past, it has not been as comprehensive as the
work on the East Pacific rise to the southeast
or the work done in the general area of the
Mendocino fracture zone immediately to the
north.

The program of heat flow measurements
was motivated by several considerations. Von
Herzen [1964] reported a banding of high,

normal, and low heat flow trending generally
north-south in the region north of the Mendo-
cino fracture zone and suggested that the trends
might cross to the south without being offset.
From the area to the south, a banding of high
heat flow had been reported along the East
Pacific rise [Von Herzen and Uyeda, 1963;
Langseth et al., 1965], and several scattered
measurements (Figure 1) of heat flow have been
reported from within the study area [Foster,
1962; Von Herzen, 1964]. Of particular interest
before the UMP work was the apparent con-
trast between the relatively consistent values
associated with the eastern part of the con-
tinental rise and the presence of scattered high
and low values over most of the offshore area.
In this study, specific interest was directed to
the high (3.45 ^cal cm"2 sec"1) reported by Von
Herzen [1964] at 38°25'N, 126°09/W and the
low (0.6 /xcal cm"2 sec"1) reported by Foster
[1962] at 38°35'N, 127°45'W. In view of these
data and the constraints of available ship time,
the basic objectives of the heat flow phase of
the UMP investigation were (1) examination
of the eastern part of the continental rise to
ascertain if it has the relatively uniform heat
flow that the earlier data implied, (2) verifica-
tion of the high and low values reported by
Von Herzen and Foster, and (3) examination of
the area between these stations to determine the
spatial extent of the high and low values (if
they were verified).

Method and Results

The instrumentation used for the heat flow
measurements was basically the Ewing thermo-

6239

G240

BURNS AND GRIM

(i)

• SURVEYOR -CONFIDENCE HIGH
o 5<y/?l//f>W-C0NFIDENCE "LOW
▲ WW£"£7?-C0NFIDENCE "HIGH"
A ^/O/Vff/P-CONFIDENCE "LOW"
D PREVIOUS REPORTS

(VH)-VON HER2EN. 1964

(F)-FOSTER, 1962

Fig. 1. (Top) Map of Upper Mantle Project area off California. Fourteen heat flow stations,
shown below, lie along A-A'. Bathymetry and the total magnetic field were obtained along
A-A' , B-B', C-C, and D-D'. The hatched areas represent areas having relief of over 100
meters. Physiographic provinces are based on the map of Menard [1964]. (Bottom) Heat
flow values to the nearest tenth in cal cm"2 sec"1 X 10"6. Previously published values by
Foster [1962] and Von Herzen [1964] are indicated by F and VH. Pioneer stations are shown
by triangles; Surveyor stations, by circles. Solid circles or triangles indicate high-confidence
values and open circles or triangles indicate low-confidence values, as explained in the text.
(The remaining figures in this paper, which show heat flow, use this method of identification.)
Station numbers from Pioneer and Surveyor are in parentheses and correspond to numbers in
Table 1. Contours, in meters, are from the map by Menard [1964].

grad [Gerard et al., 1962; Langseth, 1965]. The
equipment and the many uncertainties inherent
in the general method of estimating heat flow
have been discussed at length by many previous
investigators (see for example, Lachenbruch
and Marshall [1966] and Langseth [1965]), and
these aspects will not be discussed in this paper.
The thermal conductivities were determined

from thermal resistivity estimates based on
water content of the sediments collected by the
coring device, using the linear relationship re-
ported by Bullard and Day [1961]. The results
of the measurements are tabulated in Table 1.
The values of heat flow reported in Table 1
are simply the products of the conductivity and
the gradient. The indicated confidence has been

HEAT FLOW IN PACIFIC OCEAN

6241

estimated by alternative methods, depending on
how many thermistors actually penetrated the
bottom.

When three thermistor probes were in the
sediment, independent estimates of heat flow
were made over the two intervals between the
thermistors. Using the notation of Lachenbruch
and Marshall [1966], where (q) is the average
and Aq is the difference of the two heat flow
values, 'high' confidence is indicated for values
of Aq/2(q) less than 10% and 'low' confidence
is indicated for values greater than 10%.

When only two probes were in the sediment,
the estimate of confidence is based on the un-
certainty in the determination of both the
thermal gradient and the thermal conductivity.
When temperature differences between the
two probes were read from several places on the
record and were equal or differed by less than
0.01 °C, the temperature gradient was consid-

ered high confidence; a greater range was con-
sidered low. For conductivity, where AK is the
range of values and (K) the average, values of
AK/2(K) of less than 10% were considered
high confidence; a value greater than 10% (or
if there was only one determination of K) was
considered low. The confidence estimate in
Table 1, for the stations in which only two
probes penetrated the sediment, is reported high
only if both the temperature gradient and the
thermal conductivity were accepted with high
confidence.

In the cases (stations P-10, P-15, P-22) where
no sediment sample was obtained, the reported
(K) is the regional mean, and the listed heat
flow is considered a low confidence estimate.

If the eight measurements of heat flow pre-
viously reported from the area (see Figure 1)
are included, a total of thirty-nine heat flow
measurements are now available from the UMP

TABLE 1. Heat Flow Measurements Taken during Upper Mantle Project Marine Geophysical Survey

Station*

North
Latitude

West
Longitude

Depth, m

No. Conductivity

Probes Conductivity X 10' cal/cm
in Mud Samples sec °C

Heat Flow
Gradient X 10«,

X 108, °C/cm cal/cmJ sec

Confidence

S-l

39°01.6'

129 "59 .7'

4456

3

S-2

35°29 0'

132 "00. 4'

5167

3

S-3

35 "29. 5'

130 "58. 6'

5090

3

S-4

35 "29. 7'

130 °01 7'

4858

3

S-5

35°29.2'

128°59.0'

4796

2

s-e

37°28 2'

123°50.5'

3514

3

S-7

37°59.7'

124°53.2'

3959

3

S-8

38 "10. 2'

125°28.0'

3937

3

S-9

38°24.1'

126 "02. 2'

4264

3

S-10

38°29.3'

126°43.0'

4467

2

S-12

38 "35 3'

127 "30. 9'

4613

2

S-14

38°45.1'

128°59.0'

4533

2

S-16

38 "31. 4'

127°08.3'

4551

3

S-19

38°29.1'

126 "43.0'

4494

2

S-20

38 "27. 4'

126 "28. 9'

4403

3

S-21

35 "48. 3'

122 "59. 7'

3710

3

S-22

36°59.9'

123 "38. 0'

3425

3

S-23

36 "59. 5'

124°22.4'

4052

2

S-24

36°49.7'

124 "45. 9'

4290

2

S-25

36 "39.0'

125 "10. 5'

4194

2

P-3

36 "56. 9'

127 "55. 3'

4635

4

P-10

38°35.1'

127 "15. 9'

4462

3

P-12

38 "19 1'

124 "01 .8'

3273

2

P-13

38 "30.1'

124 "15. 3'

3438

3

P-15

37 "55. 5'

123 "49. 9'

3310

3

P-16

37 "47. 5'

124 "16. 9'

3465

3

P-17

38°25 2'

126 "08. 2'

4370

3

P-18

38°25.6'

126 "19 .8'

4224

3

P-20

38°29.3'

126 "47. 6'

4370

2

P-21

38°31.9'

127 "03. 6'

4416

2

P-22

38 "34. 3'

127 "12 7'

4452

3

1.94

0.30

0.58

High

1.95

0.98

1.91

High

1.91

2.40

4.58

High

1.86

0.84

1.56

High

1.82

1.18

2.14

High

2.00

0.98

1.96

High

1.95

0.86

1.68

High

1.83

0.97

1.77

High

1.88

0.97

1.82

High

1.95

0.73

1.42

High

2.00

0.15

0 30

High

1.88

0.25

0.47

Low

1.99

0.64

1.27

Low

1.98

0.72

1.42

Low

1.93

0.93

1.79

Low

1.98

0.86

1.70

High

1.82

0.70

1.27

High

1.98

0.74

1.46

Low

1.78

0.19

0.33

High

1.92

0.98

1.88

High

1.96

1.32

2.58

High

(1.91)

0.40

0.76

Low

1.73

1.01

1.74

High

1.87

0.90

1.68

High

(191)

0.99

1.89

Low

1.89

1.06

2.00

Low

1.86

1.50

2.79

High

1.74

1.18

2.05

Low

2.42

1.07

2.58

High

1.81

0.37

0.66

High

(191)

0.46

0.87

Low

* S prefix for Surveyor stations; P prefix for Pioneer stations.

6242

BURNS AND GRIM

area. These measurements are not evenly dis-
tributed and can be discussed by considering a
subdivision based on spatial grouping.

Eastern Part of Continental Rise

This subdivision covers the eastern parts of
the Delgada and Monterey fans and is arbi-
trarily defined as being east of the 4,000-meter
isobath (see Figure 1). Earlier measurements
from this region indicated a relatively high heat
flow ((Q) = 2.2 jucal cm"2 sec"1) with low spatial
variability. These earlier stations are clustered
in a relatively small portion of the area, and one
of the objectives of the present investigation
was to determine whether they are representa-
tive of this part of the continental rise. Conse-
quently, additional heat flow measurements
were made at stations P-12, P-13, P-15, P-16,
S-6, S-7, S-8, S-21, S-22, and S-23. The mean
regional heat flow estimated from the data ob-
tained at these ten stations is 1.7 iical cm"2 sec"1
(s = 0.2 tical cm"2 sec"1), and, if the previous
data are included, the mean is 1.8 /teal cm-2 sec"1
(s = 0.3 jucal cm-2 sec"1) .

From these estimates, heat flow in the eastern
part of the continental rise, as defined here,
appears to be slightly high (<Q) = 1.7 to 1.8
/xcal cm"2 sec"1), and the spatial variability is
low enough to permit acceptance of the mean
as a representative estimate of the regional flow
through the present sea floor. The low vari-
ability in this part of the UMP area is likely
to be caused, at least in part, by the thick
accumulation of sediment and consequent lateral
near-homogeneity of thermal conductivity in the
Delgada and Monterey fans.

Considering the area, volume, and age (107 to
4 X 107 years) of the fan deposits [Menard,
1964], we can estimate an average sedimenta-
tion rate of 1 to 7 X 10"3 cm/yr. At this rate,
the pre-existing (subsedimentary) regional heat
flow would be from 2 to 10% higher than the
presently measured heat flow, which is already
significantly higher than the world average of
about 1.5 /xcal cm"2 sec"1 [Lee and Uyeda, 1965].

Western Part of Delgada and

Monterey Fans

Prior to the UMP, five scattered measure-
ments (see Figure 1) had been taken from the
western part of the Delgada and Monterey fans.
These measurements have a mean of 1.7 /xcal

cm"2 sec"1 but are extremely variable (s = 1.2
/teal cm"2 sec"1). One of the basic objectives of
the UMP work was to determine if this vari-
ability was representative of this part of the
area (as contrasted with the small variability
over the eastern part of the continental rise)
and to attempt a relatively detailed investi-
gation of a profile {A- A' in Figure 1) between
the two previously published strongly contrast-
ing high and low heat flow values (3.45 and
0.6 ^cal cm"2 sec"1) mentioned above.

Of the fifteen additional stations taken over
this part of the fan deposits, twelve were along
A-A' (P-10, P-17, P-18, P-20, P-21, P-22, S-9,
S-10, S-12, S-16, S-19, and S-20) and three
(P-3, S-24, and S-25) were taken as scattered
stations off the profile. In aggregate, these
clearly indicate the absence of any consistent
regional heat flow over the western part of the
fan deposits.

Geophysical profile. Although some previous
studies (see, for example, Lister [1963] and
Lachenbruch and Marshall [1966]) have ex-
amined closely spaced heat flow measurements,
the measurements along profile A-A' are more
closely spaced than most and have additional
geophysical measurements associated with them.
Although the data along A-A' were not obtained
concurrently, the use of satellite navigation dur-
ing the UMP program reduces the positional
uncertainty to well within the limits of the
positions indicated in Figure 1.

Figure 2 shows heat flow, total magnetic in-
tensity, and topography for line A-A', and
magnetic intensity and topography obtained
during the UMP survey for lines B-B', C-C,
and D-D' (see Figure 1 for locations).

Topography. The western part of A-A' and
B-B' are over the sediment-filled trough south
of the Pioneer ridge [Menard, 1960], in what
appears to be a narrow part of the abyssal plain
extending in an east-west direction parallel to
the Pioneer ridge [Menard, 1964]. The western
part of D-D' is clearly south of the abyssal
plain, as shown by its relatively rough topog-
raphy. In general, the topography does not have
any clearly defined continuity of north-south
ridges or troughs in the area of A-A', which may
be characterized as over abyssal plain at the
west and on the fan deposits to the east. It
should be noted that the western end of A-A' is
clearly south of the Pioneer fracture zone.

HEAT FLOW IN PACIFIC OCEAN

6243

Fig. 2. Heat flow along profile A-A' is shown together with the topography and total mag-
netic field along profiles A-A', B-B', C-C, and D-D'.

Magnetics. Two relatively prominent mag-
netic anomaly bands can be seen in Figure 3.
To the west, an anomaly band (indicated as
la, 16, Ic, Id) characterized by a complex peak
has continuity in a north-south direction. At the
east, a second anomaly band (indicated as Ila,
116, and lid) also appears to have continuity
through the area. Figure 3 shows the apparent
correlation of magnetic peaks and troughs
across the sections A-A', B-B', C-C, and D-D'.

Each of these anomaly bands is clearly shown
by Mason and Raff [1961]. They indicate the
north-south trend of anomaly I and show

anomaly II trending northwest-southeast over
a distance of 80 to 100 km. In addition to these
two prominent anomaly bands, a less promi-
nent one appears to be situated at Ilia, III6,
IIIc, and Ma', but this band is not clearly de-
fined on the Mason and Raff map.

Heat flow. The section along A-A' indicates
a general increase in heat flow from west to
east. Considering the large variability char-
acterizing this section, the UMP measurements
agree well with the basic trend implied by the
two earlier values reported by Von Herzen
[1964] and Foster [1962]. In addition to the

6244

BURNS AND GRIM

Fig. 3. Correlation of magnetic anomalies along profiles A-A', B-B', C-C, and D-D'. They
are identical to those shown in Figure 2 but are arranged differently. Each profile is north of
the one below it. Since A-A' crosses over C-C, as shown in Figure 1, it is split. There is a
vertical offset of 200 gammas between profiles. The scales for the eastern part of A-A', C-C,
the western part of A-A', and B-B' are obtained by subtracting 200, 400, 600, and 800 gammas,
respectively, from the scale shown, which is for D-D'.

general trend, there appear to be smaller scale
variations that have sharp peaks and shallow
troughs.

One of the first questions to be asked is
whether the implied regional gradient is real.
The alternative to a regional gradient would
be to assume a regional mean and to examine
the observed values in terms of such effects as
erosion-sedimentation, lateral variations of ther-
mal conductivity, and secondary heat sources.
The mean value of the heat flow measurements
along the section (1.6 /xcal cm"2 sec"1) is com-
parable to that estimated for the eastern part
of the continental rise (1.7-1.8 jucal cm"2 sec"1),
but the contrasts along the section are such that
only an extremely arbitrary assignment of the
effects mentioned above can force a regional
mean into any approximation of the observed
values.

On the other hand, if the regional gradient
exists, the observed values of heat flow may be
examined in terms of departure from the re-
gional field. The basic constraints on any ap-
proximation of a regional gradient along section
A -A' are (1) the observed values of heat flow
along the section, (2) that the eastern end
should be comparable to the regional mean
observed over the eastern part of the con-
tinental rise (1.7-1.8 jucal cm"2 sec"1), and (3)
that the western end should not be unrealisti-
cally less than a representative oceanic mean
(about 1.0 jucal cm"2 sec"1).

The observed heat flow values from section
A-A' have been subjected to least-square fits for
first- through twelfth-order polynomials. A re-
gional smooth field is apparent through the
fourth order but is increasingly obscured by
shorter wavelength influences at fifth and
higher orders. In view of constraints 2 and 3,
both the third and the fourth order indicate a
physically acceptable approximation of a re-
gional smooth field, and there is no basic differ-
ence between them in terms of local departure
of the individual station values of heat flow.

The regional field determined by the fourth-
order polynomial (Figure 4) indicates the pos-
sibility of systematic distribution of heat flow
repeating over a 'wavelength' of several hun-
dred kilometers. This appears to be compatible
(in both sense and magnitude) with the physical
constraints (2 and 3, above) and the conclu-
sions reached by Von Herzen [1964] in regard
to apparent 'banding' of high and low heat flow
in this part of the Pacific. If such a regional
gradient is present over the distance covered by
section A-A', the heat flow 'anomalies' (Figure
4) have an absolute contrast of a magnitude
that can be explained by physically realistic
assumptions in regard to modifications in the
smooth regional field by surface and near sur-
face effects.

A fourth-harmonic Fourier fit has been made
to the heat flow data from section A-A', assum-
ing a fundamental wavelength in the regional

(0

5'o

3 8

4.0

3.0

\- e 2.0

< o
UJ v.

1.0

HEAT FLOW IN PACIFIC OCEAN

6245

A'

•

1

1 1

1 1

!

1

VH

D

A

A

—

REGIONAL HEAT
FLOW CURVE "A

0 Jr"

-^ •

o

A

t

F

a

•
1

1 1

1 1

1

1

20

40 60 80 100 120

DISTANCE IN KILOMETERS

140 160

Fig. 4. Heat flow values along A-A' are shown at the top with a regional heat flow curve
obtained with a fourth-order least-squares polynomial fit. The bottom section shows the
'anomalies' obtained by subtracting the regional curve from the observed values.

field of 260 km (estimated from Figure 4) . The
computed and observed heat flow values along
section A-A' are shown in Figure 5. In addition
to providing a closer fit to the basic data than
the smooth field of the fourth-order polynomial
(Figure 4), the Fourier fit strengthens the con-
clusion that shorter length variations of heat
flow are actually superimposed on the regional
field. The wavelength of these variations may be
estimated from either Figure 4 or Figure 5 and
is approximately 50 km.

In general, both the heat flow anomalies and
the Fourier analyses indicate that the heat flow
at stations P-17, VH, and P-20 is clearly above
the regional; stations S-9, P-18, S-20, S-10, 19,
P-21, and S-12 are low. Both analyses also
indicate a slight localized high value associated
with station S-16, which is neither strong nor
well defined, and an indication of a slightly high
value at station F with further increase im-
plied to the west.

Discussion. One of the more striking geo-
physical characteristics in this region of the
Pacific is the north-south trend of magnetic
anomaly bands [Mason and Raff, 1961;
Vacquier et al., 1961; Raff and Mason, 1961].
Although possible local associations of heat

flow variability with these magnetic lineations
has been considered by Von Herzen [1964], his
data spacing did not permit identification of
variations in heat flow with wavelengths com-
parable to the magnetic anomalies. If the 50-km
wavelength implied by the heat flow data along
section A-A' is real, however, the wavelengths
of the magnetic and the heat flow anomalies are
of comparable magnitude in this general area.
Because of their size, the heat flow anomalies
will have to be examined with more closely
spaced measurements to define the size and
depth of their sources. However, even the exist-

20 40 60 80 100 120 140 160

DISTANCE IN KILOMETERS

Fig. 5. A fourth-harmonic Fourier fit of heat flow
data along A-A'.

6246

BURNS AND GRIM

ing data indicate the presence of relatively
shallow sources for the high heat flow anomalies,
and the range of sizes and of depths of sources
that can be fitted to these data is within the
size and depth ranges proposed for the sources
of the positive magnetic anomalies (see, for
example, Mason and Raff [1961]).

Other UMP Heat Flow Measurements

Stations S-l and S-14 were taken along the
general trend of section A-A' extended to the
west. Both indicate relatively low heat flow
(0.6 and 0.5 /xcal cm"2 sec"1) but cannot be
considered with the data from along A-A', since
they are not south of the Pioneer fracture zone.

Stations S-2, S-3, S-4, and S-5 were taken to
fill the open part of the UMP area along lati-
tude 35°30/N. They form a reconnaissance sec-
tion along the parallel, are generally high, and
have very high variability. One extremely high
value (4.6 /teal cm"3 sec"1) was measured at
S-3, and an average value (1.6 jucal cm"2 sec"1)
was measured at S-4. As in the case of section
A-A', the absolute contrasts here are too large
to explain in terms of a simple regional mean,
but the data available are too few for any
pertinent analyses for regional gradients and
heat flow anomalies.

Summary and Conclusions

Heat flow data from the eastern part of the
continental rise before the Upper Mantle
Project indicated a relatively constant heat flow
over this part of the UMP area, but the data
were spotty. The additional measurements re-
ported here indicate that there is very little
departure from the regional mean of 1.7 to 1.8
jttcal cm"2 sec"1 over this part of the continental
rise.

Very strongly contrasting high and low heat
flow values and an implied banding over a
wavelength of several hundred kilometers were
suggested by the pre-UMP data over the off-
shore parts of the UMP area. These general
characteristics have been verified by the addi-
tional measurements.

Section A-A', which is the only section con-
taining reasonably close-spaced data, appears to
be on the flank of longer wavelength of high
and low heat flow which have a fundamental
wavelength of about 260 km. The regional
gradient along the section is from high heat

flow at the east (comparable with the regional
mean of the eastern part of the continental
rise) to low heat flow at the west.

In addition to a regional field, the data from
section A-A' indicate a shorter wavelength varia-
tion in heat flow with a wavelength of about
50 km. This appears to be attributable to local-
ized high and low heat flow superimposed on
the regional trend. By treating the observed
heat flow as anomalies from the regional field,
the absolute contrast of the data is reduced to
a magnitude that may be explained by physi-
cally realistic models of such effects on the re-
gional field as are caused by nonhomogeneous
thermal conductivity, erosion-sedimentation,
and supplementary shallow heat sources.

Acknowledgments. We wish to thank the offi-
cers and crew of the U. S. Coast and Geodetic
Survey ships Pioneer and Surveyor and especially
William Lucas, who assisted in making the heat
flow measurements. We have benefited from dis-
cussions with C. R. B. Lister and J. C. Kelley, and
received help and advice on various aspects of this
study from T. V. Ryan and W. Anikouchine.

References

Bullard, E. C, and A. Day, The flow of heat
through the floor of the Atlantic Ocean, Geo-
phys. J., 4, 282-292, 1961.

Foster, T. D., Heat-flow measurements in the
northeast Pacific and in the Bering Sea, J.
Geophys. Res., 67, 2991-2993, 1962.

Gerard, R., M. G. Langseth, Jr., and M. Ewing,
Thermal gradient measurements in the water
and bottom sediment of the western Atlantic,
J. Geophys. Res., 67, 785-803, 1962.

Lachenbruch, A. H., and B. V. Marshall, Heat flow
through the Arctic Ocean floor: The Canada
basin-Alpha rise boundary, J. Geophys. Res.,
71, 1223-1248, 1966.

Langseth, M. G., Techniques of measuring heat
flow through the ocean floor, in Terrestrial Heat
Flow, Geophys. Monograph 8, edited by W. H.
K. Lee, pp. 58-77, American Geophysical Union,
Washington, D. C, 1965.

Langseth, M. G., P. J. Grim, and M. Ewing, Heat-
flow measurements in the east Pacific Ocean,
J. Geophys. Res., 70, 367-380, 1965.

Lee, W. H. K., and S. Uyeda, Review of heat-flow
data, in Terrestrial Heat Flow, Geophys. Mono-
graph 8, edited by W. H. K. Lee, pp. 87-190,
American Geophysical Union, Washington,
D. C, 1965.

Lister, C. R. B., A close group of heat-flow sta-
tions, J. Geophys. Res., 68, 5569-5573, 1963.

Mason, R. G., and A. D. Raff, Magnetic survey off
the west coast of North America, 32°N latitude
to 42°N latitude, Bull. Geol. Soc. Am., 72, 1259-
1266, 1961.

HEAT FLOW IN PACIFIC OCEAN

6247

Menard, H. W., Possible pre-Pleistocene deep-sea
fans off central California, Bull. Geol. Soc. Am.,
71, 1271-1278, 1960.

Menard, H. W., Marine Geology of the Pacific,
McGraw-Hill Book Company, New York, 1964.

Raff, A. D., and R. G. Mason, Magnetic survey
off the west coast of North America, 40 °N lati-
tude to 52°N latitude, Bull. Geol. Soc. Am., 72,
1267-1270, 1961.

Vacquier, V., A. D. Raff, and R. E. Warren,
Horizontal displacements on the floor of the

northeastern Pacific Ocean, Bull. Geol. Soc. Am.,
72, 1251-1258, 1961.

Von Herzen, R. P., Ocean-floor heat-flow measure-
ments west of the United States and Baja Cali-
fornia, Marine Geol., 1, 225-239, 1964.

Von Herzen, R. P., and S. Uyeda, Heat flow
through the eastern Pacific Ocean floor, J. Geo-
phys. Res., 68, 4219-4250, 1963.

(Received June 27, 1967.)

11

Reprinted from YEARBOOK OF SCIENCE AND TECHNOLOGY, McGraw-Hill, New York

Astrobleme

The search for astroblemes, ancient meteorite-
impact scars on the Earth, has intensified in re-
cent years. Interest is keen because the Moon is
presumably covered with such scars, and several
large Quaternary meteorite craters have been
found on the Earth's surface. About 50 probable
examples of terrestrial astroblemes have been
identified to date, but at none of them has meteor-
itic material been found, except in connection
with the meteorite craters of geologically recent
date. Even so, the geological record of pre-Quater-
nary time might be expected to include numerous
impact scars even though erosion has occurred
since they were formed.

The larger Quaternary meteorite craters, both
definite and probable, are the following, arranged
in order of size from about 1 km to 10 km in di-
ameter: Pretoria Salt Pan Crater, South Africa;
Meteor (Barringer) Crater, Arizona; Talemzane,
North Africa; Lonar Crater, India; Roter Kamm,
Southwest Africa; New Quebec Crater, northern
Canada; and Bosumtwi Crater, Ghana. The
15,000,000-year-old (latest Miocene) Ries Basin
structure in West Germany may be regarded as a
feature transitional between a meteorite crater
and an astrobleme because its craterlike appear-
ance is not fully erased. All the basic shock cri-
teria for astroblemes are found at the Ries Basin.

Fig. 1 . A shatter cone in limestone from the Kent-
land, Ind., astrobleme.

110 Astrobleme

Fig. 2. Photomicrographs, magnified about 40
times, of (a) normal unfractured Coconino sandstone
and (b) its shock-fractured equivalent from Meteor
(Barringer) Crater, Ariz. Rhombohedral sets of planar

In addition, meteoritic spherules in silica glass,
but not ponderable meteorites, are present.

Criteria. The term cryptoexplosion structure
pertains to a highly disturbed geological feature
resulting from a natural explosion of uncertain
cause. Such structures are fairly common, but
only some of them appear to be astroblemes.
Others are generally believed to be cryptovolcanic
structures caused by phreatic (steam) or volcanic
explosions but without the extrusion of volcanic
products or hydrothermal alteration.

The distinctive hallmarks of astroblemes result
from intensive shock-wave damage to the rocks.
At least four types of criteria are now generally
recognized. First is the presence in the rock for-
mations of a distinctive type of conical fracture,
called shatter cones, which display striated mark-
ings like "horsetails" (Fig. 1). These are apparently
the result of shock fracturing and therefore seem
to be a useful criterion. Second is the presence of
shock-induced polymorphs of silica or feldspar,
such as coesite, stishovite, and maskelynite. Third
is the occurrence of shock-induced microfractur-
ing of crystals whereby, for example, certain planar
sets of fracture appear in quartz (Fig. 2) and biotite
becomes kink-banded. The fourth criterion is the
presence of lechatelierite, high-temperature silica
glass.

Discoveries. Putative astroblemes contain one
or more of the above criteria. Seventeen are now
known from shatter coning alone. This is often
the first hallmark to be identified as it is a field
criterion rather than a laboratory one. The Stein-
heim Basin, a twin structure to the Ries Basin
formed at the same time, is the prototype shatter-
coned structure. The Wells Creek Basin structure
in Tennessee (Fig. 3) is an example of shatter-
coned structure in the United States. The geolog-
ical plan shows that the typical damped-wave
form is created. A central dome is surrounded
by a ring syncline and a ring anticline. Other
shattcr-coned structures in the United States
are Sierra Madera in Texas, Kentland in Indiana,
Serpent Mound in Ohio, Decaturville and Crooked
Creek in Missouri, and Flynn Creek in Tennessee.

fractures are seen in the quartz grains, between
which is silica glass containing coesite and stishovite,
all created by shock waves from a meteorite impact.
(NASA)

uncertain
fault extension

scale in miles
J St. Louis formation

J Warsaw-Fort Payne formation
2 Hermitage to Chattanooga formations
_] Stones River formation
h| Knox formation

Fig. 3. Geologic map of Wells Creek Basin struc-
ture in Tennessee, a putative astrobleme. (R. Stearns
and H. Tiedeman, Vanderbilt University)

During 1965, a new shatter-coned structure was
identified, the Middlesboro Basin structure in
Kentucky.

A highly interesting probable new astrobleme,
although regarded by Australian geologists T.
Quinlan and C. E. Prichard as a diapir, is Gosses
Bluff, 150 mi west of Alice Springs in central
Australia in the Amadeus Basin geosyncline.
Lower Paleozoic and Proterozoic strata are in-
volved in the deformation of this 2-mi-diameter
circular uplift. The structure must have about a
6-mi diameter overall, but the outer portion is

Astrogeology 111

hidden beneath desert alluvium. The astrobleme
identification is based upon the definitive presence
of shatter cones. The structure promises to do
much to unravel the nature of cryptoexplosions
as many types of beds are affected rather than
only carbonates, which rarely show shock effects
other than shatter cones. The structure has
been test-drilled to more than 3000 ft, and the
incompetent Bitter Springs limestone, supposedly
the active bed which flowed into a plug to form the
supposed diapir, was not encountered.

Recently shatter cones have been discovered
by E. L. Krinov in the main crater of the Kaalijarv
meteorite crater group in Estonia.

For background information see ASTEROID;
METEORITE in the McGraw-Hill Encyclopedia of
Science and Technology.

[r. s. dietz)

Bibliography: N. L. Carter, Am. J. Sci., 263:
786-806, 1965; R. S. Dietz, Meteoritics, 3(1):
27-29, 1966.

12

Reprinted from MEDICAL OPINION AND REVIEW Vol. 3, No, 1

':>^k,'i^V-'Tr f: ' % \ ..,.-... .^

Science:

Craters and Legends

Roberts. Dietz, Ph.D.

Meteor Crater, Arizona

On four mornings last November
I spent the early predawn twilight
on the rim of Arizona's Meteor
Crater, observing the Ikeda-Seki
sun-grazing comet through a Ques-
tar telescope. I had traveled there
from the annual Metcoritical So-
ciety meeting at Odessa, Texas, with
Richard Barringer, son of D. Mor-
eau Barringer. It was the elder Bar-
ringer who, at the turn of the cen-
tury, first recognized Meteor Crater
as the site of a giant meteorite ex-
plosion in the recent geologic past.
He was so intrigued by this yawn-

Robeht S. Diktz is a Research
Occanographer witJi the US Coast
and Geodetic Survey in Washing-
ton, DC

ing hole in the ground that he spent
the last twenty-five years of his life
(from 1903 to 1927) attempting to
mine the great meteorite he as-
sumed lay buried beneath the crater
floor— a heroic venture fraught with
difficulties and disappointment, and
unique in mining annals. The Bar-
ringer family still maintains the
crater as a public trust and fosters
research in the field of meteorities.
It seemed to me singularly appro-
priate to be observing this spectacu-
lar comet from the rim of a crater
created in a flaming instant 25,000
years ago by the impact of another
celestial object. Meteorite craters
are rare geomorphic forms; only
about a score are known from all
over the world, so we are fortunate

50

in having the finest example on na-
tive soil. There is only one other
meteorite crater here in the United
States. It is near Odessa, Texas, and
it was the reason for the convening
of the Meteoritical Society meeting
there. The Odessa site is small and
it has been badly despoiled Since its
discovery in 1932. But recently a
group of public-spirited citizens of
Odessa have restored the site, and
its future as a fascinating natural
feature is now assured.

The list of large (more than 3,000
feet in diameter) meteorite craters
of the world is short, even including
probable as well as definite exam-
ples. They are, in order of increas-
ing size: Pretoria Salt Pan in South
Africa, 3,400 feet across; Meteor

Medical Opinion & Review

Crater in Arizona, 4,000 feet across;
Talemzane in North Africa, 5,600
feet across; Lonar Crater in India,
6,000 feet across; Roter Kamm Cra-
ter in Southwest Africa, 8,000 feet
across; New Quebec ( Chubb ) Cra-
ter in northern Canada, 11,300 feet
across; and the Ashanti Crater in
Ghana, 33,000 feet across. Some
smaller meteorite craters that are of
exceptional interest because of as-
sociated mcteoritic debris are the
Ilenbury, Boxhole, and Wolf Creek
Craters in Australia; the Wabar
Craters in Arabia; and the just-dis-
covered Monturaqui Crater in
northern Chile.

Moon, Mars

Scientific interest in meteorite
craters has, recently been stimulated
by the prospect of man landing on
the Moon. The controversy on the
volcanic versus meteoritic origin of
lunar craters has now run its course,
with a clear victory for the theory
of creation by impact. Mars, too, is
covered with impact craters, as was
recently shown by photos from the
Mariner IV fly-by. This greatly en-
hances interest in the few meteorite
craters on Earth. Astronauts already
receive some of their training at
Meteor Crater, and we can expect
intensive scientific study of all sites
that provide clues to the lunar ter-
rain.

Meteor Crater, seen from the ap-
proaching highway, looms across
the desert as a flat-topped mesa,
giving no hint of the great hole in-
side. This silhouette is actually the
crater's upturned rim, rising 170
feet above the surrounding dry
russet plain and containing 300 mil-
lion tons of fragmented rock. The
crater it circumscribes is nearly a
mile across and 600 feet deep.
About 25,000 years ago, an errant
cosmic body, presumably from the
asteroid belt between Mars and
Jupiter, zeroed in on the Earth.
Coming in from the north, a huge
nickel-iron mass of perhaps a quar-
ter-million tons traveling at about

ten miles per second, so that each
ounce of meteorite had 400 times
the energy of a high-speed bullet,
struck the ground. In a blinding
flash of light, the meteorite bored
into the solid rock, causing a ten-
megaton explosion. Under the enor-
mous percussion, the meteorite be-
haved as though it were fluid. It
flattened, turned inside out, and
swept rock debris outward and up-
ward—leaving a paraboloid crater
like those on the moon. Rocks as
big as cottages were hurled for a
quarter of a mile. Part of the mete-
orite was converted into a vapor
that mixed with Earth debris, rose
into the stratosphere as a mushroom
cloud, and rained down minute me-
teoritic droplets that can still be
found in the surrounding country-
side.

White men first learned about
Meteor Crater in 1871, but they
thought it was just another volcanic
(rater. It was already familiar to the
Navajo and Ilopi Indians of the re-
gion. The Ilopi tribe used to gather
the finely powdered white silica—
Coconino sandstone crushed into
"rock flour" by the force of the im-
pact—to use in their ceremonies.

In 1891, a prospector in the region
discovered "nuggets of silver" that

on examination proved to be nickel-
iron meteorites. They were chris-
tened Canyon Diablos, after the
principal river in the area, but their
direct association with the crater
went unrecognized at first. Never-
theless, they achieved immediate
international fame when they were
found to contain diamonds, even
though the stones were carbonados
or black diamonds, not the precious
variety. Diamonds are, of course,
the hardest known natural sub-
stance in existence. Their presence
was detected when an attempt was
made to cut a nugget. The cutting
disk abruptly stopped working, and
its teeth were stripped down to a
nubbin when they struck a nest of
minute diamonds.

We now believe that these dia-
monds were formed instantaneous-
ly as the shock of impact on the
Earth converted the mineral cohen-
ite ( an iron carbide ) into the high-
pressure polymorph of carbon we
call diamond. Canyon Diablos are
the only diamondiferous nickel-iron
meteorites known, although there
are at least three other stony mete-
orites that contain diamonds. These,
too, are thought to be shock-created,
but by impacts in space, not on
the Earth. Ordinary terrestrial dia-

Meteorite Crater, Odessa, Texas

January. 1967

51

monds are of quite a different gene-
sis, formed by static high pressures
in the Earth's upper mantle. They
are carried to the surface through
kimberlite pipes that, in places like
South Africa, have bored their way
through the Earth's granitic crust.

All meteorites are highly prized
by scientists. They are rare samples
of the cosmos that give us some ink-
ling of the nature of the sun and the
other planets. Curiously, it was by
determining the age of Canyon
Diablo meteorites from the ratio of
their lead isotopes, which provide
a radiometric "clock," that Califor-
nia Institute of Technology scien-
tists first inferred the Earth is 4.5
billion years old. They found the
meteorites to be older than any
rocks known on Earth and, as the
Earth rocks have been subjected to
many changes, they reasoned that
the Canyon Diablos are primitive
substances representing the true age
of both the solar system and the
Earth.

The attempt to mine the Canyon
Diablo meteorite commercially for
its riches failed. People did not real-
ize that giant meteorites vaporize
almost completely on impact. Find-
ing meteoritic fragments at a mete-
orite crater is about as difficult as
finding TNT in a bomb crater; only
about thirty tons of Canyon Diablos
have ever been recovered. Before
his death, however, D. Moreau Bar-
ringer did succeed in establishing
the meteorite impact origin of Me-
teor Crater, over the objections of
the two renowned United States
Geological Survey geologists G. K.
Gilbert and N. H. Darton, who con-
tended that the crater was blasted
out by a steam explosion, a sort of
incipient volcanism.

Of all the world's meteorite cra-
ters, Meteor Crater is the best pre-
served, the most elegant, and the
most accessible. It is visited by
150,000 tourists a year and is one of
the world's most remarkable natural
wonders. Svante Arrhenius, a Swed-
ish Nobel Laureate in chemistry,

Meteorite in situ

once described it as "the most fas-
cinating spot on Earth."

Another meteorite crater is the
twin Wabar ( actually there are two
meteor craters) discovered by the
explorer H. St. J. Philby, who tra-
versed Arabia's Empty Quarter
(Rub 'al Khali) in 1932. He was
searching for the legendary ruined
city of Wabar, supposedly de-
stroyed by fire and brimstone like
Sodom and Gommorah, and for a
legendary chunk of native iron
called Al Hadida ( the lump of iron )
allegedly "as big as a camel." But let
Philby tell the story:

"We marched on till the ruins
came into sight, a long black wall,
as it seemed, riding on the sands.
We drew nearer and halted at a
suitable camping spot, while I im-
mediately hastened to the top of a
low sand-hill nearby to get a good

first view of the site before dark.
The great moment had at last ar-
rived, the moment I had longed for
for fourteen years, and 1 found my-
self looking down on the ruins of
what appeared to be a volcano! So
that was Wabar. the city of a wicked
king, destroyed by fire from heav-
en and thenceforward inhabited
only by semi-human monomem-
brous monsters. A volcano in the
midst of the Bub 'al Khali! And be-
low me, as I stood on that hill-top
transfixed, lay the twin craters,
whose black walls stood up gauntly
above the encroaching sand like the
l>attlements and bastions of some
great castle. These craters were re-
spectively about 100 and 50 yards in
diameter, sunken in the middle but
half choked with sand, while inside
and outside their walls lay what I
took to be lava in great circles where
it seemed to have flowed out from
the fiery furnace. Further examina-
tion revealed the fact that there
were three similar craters close by,
though these were surmounted by
hills of sand and recognizable only
by reason of the fringe of blackened
slag round their edges.

"My companions were soon busily
engaged in digging into the ruins
for treasure, and nothing I said
could stop them. Their saddle-bags
were soon filled to bursting with the
little jet-black shining pellets which

■-■
Wabar Crater, Arabia

January, 1967

53

strewed the place and which they
took for pearls, the pearls of the
numerous ladies who, according to
the tradition, had graced the court
of King 'Ad and perished in the
flames. Two months later these
relics were going around the mar-
ket at Mecca leaving disappoint-
ment in their train, but to this day
my companions believe that the
ruins we saw were indeed the ruins
of a great city. Yet the truth is that,
after all we had already seen of flint
implements and other evidences of
ancient man, this locality failed to
produce the slightest vestige of
human occupation, temporary or
otherwise, modern or ancient. The
nearest water was ten miles away
at Faraja, and the river I had hoped
to find was nowhere to be seen. The
secret of Wabar was out at last, and
all that remained to do was to search
for the iron 'as big as a camel.' It
was like searching for a needle in a
haystack, and I thought it best to
leave my companions to it.

Fragment

"The party had to confess failure
in their search. They had found a
silly little fragment of iron, but it
was not as big as a camel, so they
had not worried about it. I sent
somebody in search of it, and he
duly brought it along. It was about
the si/e of a rabbit, but I have very
little doubt that it was the very
thing we had come to find. Legend
had magnified it as we still magnify

fish The camel had gone through

the eye of a needle with a ven-
geance, but the meteorite ( for that
is what it was ) will doubtless be of
some interest to experts. It was im-
mensely heavy, as the camels had
good reason to know for the next
month or more. So, for all practical
purposes, we had found the iron
and we had found the ruins. The
two localities had merged into one."
( From Geographical Journal, 1933. )

Subsequent history has proved
Philby somewhat mistaken. Recent-
ly Thomas Abcrcrombie of the Na-

1

tional Geographic Society found a
5,000-pound nickel-iron meteorite at
Wabar (National Geographic, Jan-
uary, 1966). This is certainly Al
Hadida— the lump of iron "as big as
a camel." One is reminded of an old
Arab proverb: "There are three
things impossible to hide: smoke,
a man riding a camel, and love be-
tween two people." The proverb
seems not to apply to a meteorite
as big as a camel amidst the shifting
sands of the trackless Empty Quar-
ter.

Lonar Crater in India is still an-
other crater most likely caused by
a meteor. It has all the character-
istic features except for any known
associated meteoritic debris. A
splendid, perfectly circular natural
feature 6,000 feet across and 600
feet deep, it is rich with wild fowl
and archeologic sites. A spring-fed
brackish lake lies in the middle,
while around the shore are fourteen
ancient temples. The lake has slow-
ly risen over the past 100 years, until
some of the houses of worship are
knee-deep in water.

Prior to my study of the crater,
Indian geologists had simply con-
sidered it to be another volcanic ex-
plosion crater. But plateau lavas of
this type, like the Columbia basalts
of western Washington, issued qui-
etly from fissures without ever pro-
ducing craters. And the Deccan
traps ceased erupting 60 million

Lonar Crater, India'

years ago, whereas Lonar is obvi-
ously young, not more than a few
hundred thousand years old. Fur-
thermore, it is regionally unique;
the entire subcontinent of India, in-
cluding even the lofty Himalayan
Mountains, completely lacks any
contemporary volcanism. Finally,
vvc don't have to invoke rare chance
to account for a meteorite striking
the Deccan traps. These ancient
lavas cover 250,000 square miles, so
they offer a sizable target for a cos-
mic bolide.

Any thought of either volcanism
or meteorite impact is quite foreign
to the natives in the village of Lonar.
According to local legend, this great
crater was once the covered and
m eluded den of a mad giant who
ravaged the surrounding country-
side. To counter the giant, the god
Vishnu came down from Heaven
and courted the giant's sister until,
eventually, she revealed his hiding
place. Vishnu forthwith tore the top
off the den and threw it five miles
away, where the lid may still be
seen, preserved as a dome-shaped
mountain. Then, in a bloody strug-
gle, he slew the giant. The crater
remains a remnant of this heroic epi-
sode in Indian folklore.

Lonar Lake is fed largely by a
single gushing spring of cool, clear,
and potable water that is also the
sole water supply for the village of
Lonar. Lithe women swarm to the

January, 1967

57

spring all day long like a stream of
ants, filling their water jugs and
toting them back on their heads.
The spring is regarded as holy. It is
told that a fakir threw a stick into
the sacred waters of the Ganges at
Banaras, 500 miles to the northwest.
He watched it disappear in a whirl-
pool, then ran all the way back to
Lonar, arriving just in time to see
the very same stick issue from the
spring.

From my own studies, I am con-
vinced that Lonar is a bona fide
meteorite impact crater. It has many
of the usual hallmarks by which we
have learned to identify such cra-
ters, including a raised rim with
upturned, and locally flipped over,
country rock, and drilling has re-
vealed a nest of breccia (highly
fractured and crushed rock) under-
lying the crater floor.

From the vast expanse of gently
undulating bushveld located about
twenty-five miles north of Pretoria,
the capital of the Republic of South
Africa, a low, tree-clad ridge of
broken outline rises. To the surprise
of the stranger approaching the
ridge, he finds it encloses a perfectly
circular depression that has a flat
bottom covered with a dazzling
white saline encrustation. Over-all,
this remarkable crater is 3,500 feet

across, 400 feet deep, and about two
miles in circumference.

One's immediate thought is that
it is an ancient and extinct volcanic
crater, but the entire bowl is im-
pressed in granite and there is abso-
lutely no sign of lava or any other
products of volcanism. A dark,
placid pool of soda-salt brine is in
the center of the depression, mirror-
ing the many-hued greenery of the
inner slopes. This quiet pool and its
white border present a striking sight
not easily effaced from memory. It
is especially impressive by moon-
light, when somber shadows are
cast by the bush-clad rim. Some
abandoned mineworkings, about
fifty years old, are scattered over the
crater floor, relics of a brief episode
during which common salt, trona,
and other salts were recovered.

Until quite recently this pan was
the most important salt lick in the
central Transvaal, and was visited
by herds of game, including ele-
phant. Before the advent of the
white man it was a hunting ground
for Bushmen, and remnants of their
implements may still be found.

Geologists originally thought the
Pretoria Salt Pan was created by an
explosion that had somehow es-
caped from a deep-seated pool of
molten rock. It is now widely ac-

Pretoria Salt Pan

ceptcd as meteoritic; it has all the
hallmarks, although no meteoritic
debris has ever been found. Drilling
exposed a nest of crushed and frag-
mented rock underlying the pan.
The strata are upturned in a way
that suggests a violent explosion of
such an intensity that only a great
meteorite; could have caused it.

Residents

A native family "occupies" the
former manager's residence over-
looking the pan. I say "occupies"
advisedly. The members of the fam-
ily actually live around a fire out in
front, except when nightfall or rare
inclement weather drives them in-
doors. Even the term "indoors" is
not quite appropriate; the doors,
along with the windows and all
other pieces of available wood, have
been removed to feed the eampfire.
So the natives have not entirely
taken to the ways of the white man.

Farther around the periphery I
ran into a single native man with
half a dozen boys living in a rude
lean-to. At first they resented my
presence, but eventually they ac-
cepted it. The man finally told me
he was running a coming-of-age
school to teach the boys tribal mores
and rites. When I told him of my
meteorite impact theory for the
crater's origin, he agreed that it was
certainly correct— and he volun-
teered the information that his kudu
horn also came irom heaven.

So there are perhaps a score of
large meteorite craters on the Earth,
which is not many compared to the
Moon's 30,000 or so. But the Moon's
surface is old and erosion there is
nil, so its crust is a museum of an-
cient geomorphic forms preserved
in pristine beauty. Its surface is a
counting plate recording all the cos-
mic bolides that have struck over
the last billion or more years. Me-
teorite craters on Earth are short-
lived. They don't last more than a
million years, but they serve to re-
mind us that Earth, too, is subject
to cosmic bombardment. end

58

Medical Opinion & Review

13

Reprinted from SEA FRONTIERS Vol. 13 , No . 1

Mid-ocean ridge. Slow convection currents in the mantle below are set in motion
by the internal heat of the earth. It is these currents which some believe to have
caused separation of the continents, leaving the mid-ocean ridge as a "scar."

More A bout Continental Drift

By Robert S. Dietz

Institute for Oceanography
Environmental Science Services Administration

(Illustrated by John Holden)

Because of readers' comments and interest in recent articles on continental
drift, Sea Frontiers is pleased to present a fuller explanation of this contro-
versial theory.

These are times of scientific fer-
ment. One of the great geologic
debates now simmering argues the
question whether the continents have
been fixed in position throughout their
history or whether they have drifted
to their present locations. The cham-

pions of the drift theory will eventual-
ly be judged as either an avant garde
or an idiotic fringe, depending on how
the question is finally resolved. Fun-
damental aspects of the earth's his-
tory are involved, including the size
and whereabouts of the ancient ocean

66

basins. So opinions are strongly op-
posed and statements likely to be in-
temperate. Reams of paper are being
expended and gallons of ink spilled in
this confrontation but the argument is
now just warming up.

Briefly, the debate is drawn along
the following lines. The continental
fixist believes that continents and
ocean basins have always remained
much as we see them today, anchored
in position and with but minor mod-
ifications of outline. This is the con-
servative philosophy and American
geologists have been raised in this
tradition. Continental drifters, while
disagreeing among themselves as to
details, believe that all the continents

were once either joined together into
one universal continent called Pangaea
or, more likely, organized into the
northern hemisphere supercontinent of
Laurasia and the southern hemisphere
supercontinent of Gondwana. These
great landmasses then broke into
pieces and drifted apart 150 to 100
million years ago. The case for Gond-
wana is particularly strong. It is sup-
posed to have been a single mass in-
corporating the present continents of
South America, Africa, Antarctica,
and Australia; the subcontinents of
India and Madagascar; and some
"microcontinent" pieces such as the
Seychelle Islands Bank and possibly
the submerged Kerguelen Plateau,

Reams of paper are being expended and gallons of ink spilled . . . but the argument
is only just warming up.

67

both in the Indian Ocean, and perhaps
other as yet undefined pieces.

Compasses in Congealed Rocks

Interest in drift has been aroused
by rock magnetism or paleomagnetic
studies which offered a new and inde-
pendent approach to the drift problem
(See K. M. Creer, "Continents on the
Move," Sea Frontiers, Vol. 12, No. 3,
May-June, 1966). When certain
rocks, having magnetic properties, as
in the case of basaltic lava flows, cool
from the molten state at about 575°
(through the Curie point), they retain
the magnetism induced in them at that
time by the earth's magnetic field. Thus
the crystals in them become, as it were,
frozen compasses, showing the earth's
magnetic orientation and magnetic dip
at the time of freezing. The horizontal
direction of the magnetism in the rock
points in the general direction of the
earth's magnetic pole, just as in the
case of a compass needle. In addition,
the magnetic axis of the rock crystal
dips downward towards the earth's
magnetic pole, being vertical in rocks
near the poles and decreasing its dip
towards the equator, where the axis
becomes horizontal. The "rock com-
pass" gives information as to both
latitude and orientation of the rock at
the time its magnetism was frozen.
Yatchsmen and sailors will see that the
problem of determining the original
longitude of those rocks remains un-
solved— precisely as with marine navi-
gation before John Harrison invented
his chronometer. Only a line-of-
position is obtained, not a fix. Never-
theless,the changes in position of rocks
shown by paleomagnetism provide
rather convincing evidence that the

continents today are not where they
were 150 million years ago. In fact,
most of these rock formations are not
magnetized along the present-day di-
rection of the earth's field, indicating
they were somewhere else when they
became magnetized.

Thin Ocean Floor

There are other reasons for believing
that continents have been on the move.
As a marine geologist, my own con-
version to continental drift resulted
from the numerous surprising discov-
eries about the sea floor. The blanket
of sediment on the sea floor, if it has
thickened at a rate similar to that of
present day sedimentation, would have
a thickness related to the length of life
of the ocean floor. If the ocean floors
have always existed undisturbed, there
would be a very thick layer of sedi-
ment. Actually, it is surprisingly thin,
indicating that the ancient ocean floor
has been destroyed or reworked in
some way.

Another surprising fact is that, al-
though the layer of granitic rock (sial)
beneath the continents is very thick,
it is entirely absent under nearly all
of the deep sea, suggesting that the
lighter granite of the continents, float-
ing on the heavier mantle rocks below,
was pulled apart when the continents
separated, leaving the ocean crust bare
of granite. Also, the rocks of the sea
floor, as dated by their fossils or radio-
metrically, are surprisingly young.
Further, in the absence of erosion,
mountains on the sea floor should last
forever as they do on the moon, but
we can find no very ancient seamounts
beneath the sea.

With the continental drift concept,

68

these and other surprises become the
natural expectations.

Fossil Evidence

The fossils of Africa and South
America indicate that, prior to 150
million years ago, the same kinds of
creatures existed on both continents.
Subsequently, the evolution of life had
already proceeded along quite differ-
ent lines on the two continents. This
suggests that the continents were orig-
inally united and that free exchange
of living creatures could take place,
but that later they separated with both
the continents and the evolution of life
going their separate ways. Yet, most
geologists remain skeptical of conti-
nental drift. For example, to explain
this faunal dispersion most fossil ex-
perts find drift a sufficient explanation
but not a necessary one. They argue,
with unquestionable logic, that it is
easier to drift a species than it is to
drift a continent. Rafting long dis-
tances at sea upon a floating log is
one mode for such dispersion, and al-
though such rafting would be a rare
event, it is argued that, over the long
sweep of geologic time, anything that
might happen will happen. And, also,
supposedly simpler than continental

Most paleontologists claim that it is
easier to explain the occurrence of sim-
ilar species on continents widely sep-
arated by the ocean on the basis of (top)
transfer of living creatures by rafting on
flotsam, by easy moves along ancient
island chains, or by long subsided land
bridges, rather than by continental drift
(bottom).

Rafting

Island Stepping Stones

Continental Drift

69

drift, would be dispersion of living
creatures over isthmian links (like the
Isthmus of Panama today), which in-
cluded narrow land bridges between
Africa and South America, for
instance.

On Continents, Ocean Basins
and Icebergs

The drift concept was first formal-
ized by the Austrian meteorologist
Alfred Wegener. He made rather bold,
even bizarre, assumptions about the
nature of continents and ocean basins
such as were required by his hypothe-
sis. Actually, most of his suppositions
have turned out to be true. We have
subsequently learned that continents
are not deeply rooted masses of rock
but instead are separated from the
earth's mantle by a sharp boundary
called the Mohorovicic Discontinuity,
or the Moho. We also now know that the
continental blocks around most of the
Atlantic and Indian Oceans have "raw
edges" — mountain ranges intersect the
continental margin and abruptly stop.
Since World War II we have learned
that the ocean floor is crustless in the
sense of having no liner of continental-
type granitic rock or sial. And there is
no question but that continents actual-
ly float high because of their low den-
sity. Continents all in all very much
resemble icebergs. The plaguing ques-
tion remains: Like icebergs, do they
also drift?

In our everyday experience, very
large things are very hard to budge,
so it defies our instincts that continents
can be moved. But our instincts also
deny other things — such as the entire
earth moving around the sun sup-
ported by nothing at all. Galileo was

forced to recant this "illogical notion"
before the Inquisition, but upon leav-
ing is said to have muttered, "Eppur
si muove" ( "But still it moves" ) . Late-
ly the mechanical objection to drift has
been removed by the suggestion that
the earth's mantle undergoes a slow
convection movement. In each convec-
tion cell, the mantle material, rising as
a semi-liquid current in the middle,
curves outwards to the sides and de-
scends along the edges. This is the
simple, well-known convection move-
ment of water when heated in a beaker.
The mantle movement sluggishly flows
merely a few centimeters per year. This
would be ten million times slower than
the flow within the earth's molten core
whose liquid turbulence accounts for
the magnetic field. If such convection
does occur, the continents may be en-
trained by the flow of the mantle and
carried along as though on a conveyor
belt. This would be fully analagous to
the drift of icebergs which are trans-
ported, not by any inherent or im-
pressed force, but passively by the
drift of the ocean.

This mode of transport is independ-
ent of scale. A huge iceberg is moved as
readily as a bergy bit. In 1927 a 10,-
000-square-mile iceberg, larger in area
than the state of Rhode Island, drifted
past the Faulkland Islands, off Argen-
tina. So size per se is of no conse-
quence. So far as the continents are
concerned, we already know that sub-
continental-size pieces drift. The
northward motion of southern Califor-
nia, relative to the remainder of North
America, is a case in point. It brings
Los Angeles two inches closer to San
Francisco every year.

Wegener erred in departing from

70

the iceberg analogy. He assumed that
continents actively ploughed their way
through the mantle like great ships at
sea, using forces he called Westwand-
erung (westward drift) and Poleflucht-
krajt (pole fleeing force). These
forces, although real, were shown to
be negligibly small and served to dis-
credit the continental drift concept for
a few decades.

A Jigsaw Puzzle

The reconstruction of the ancient
supercontinents envisioned by conti-
nental drifters is a jigsaw puzzle. The
jigsaw puzzle player has two clues:
matching the picture, using color, tone
or graphics; or fitting, using the shape
of the pieces. The continental drift
puzzle has the same clues. The "pic-
ture" is sketched by the bedrock geolo-

gy, the foldbelt trendlines, distribution
of glacial formations (tillites), etc. But
there are difficulties. The geological
picture resembles an impressionistic
painting. No two observers seem to
agree whether or not any particular
assembly is really pictorially in sharp
focus. Also full rendering is hindered
by the contiguous parts being sub-
merged continental shelves which are
zones of no picture information, or
"white paper." Furthermore the prop-
er pieces are the continental outlines
as delineated by the 1000-fathom line
rather than by the shoreline or the
shelf edge. Our marine surveys are
not sufficiently good as yet to delineate
the position of this contour.

Following Wegener's lead, many
drifters have fitted the continents to-
gether by forcing the fit, i.e., by taking

The reconstruction of the ancient supercontinents envisioned by drifters is a
jigsaw puzzle.

71

~^~- Tnassic fold belts

□ Karroo
Catarin

and Santa
o Systems

Undifferentiated

>k.(n^&^_

A jigsaw player has two clues. He may proceed by fitting the pieces by shape or
by matching the design of the picture. Top: the pieces fit and the picture matches.
Bottom: the pieces fit but the picture does not match.

72

great liberties with the shapes of the
pieces as though the continents were
made of rubber. The reconstructions
of former supercontinents are easy to
make by such sketch-map methods,
but are they meaningful? Using conti-
nental cutouts or templates for achiev-
ing fit on a globe is another favorite
device of drifters. Unfortunately some
of these have been done on a globe of
12-inch diameter or less. And one
recent paper records the use of a
6-inch globe! Good "fits" are easy to
achieve on a ping-pong-ball-sized
globe.

Nonetheless, the case for the former
existence of the southern hemisphere
supercontinent of Gondwana, at least,
is particularly strong. Attempting its

reconstruction is a fascinating jigsaw
puzzle — a game for the Gods, as some
of the pieces are missing.

Tibetan Plateau: Two Continents Thick

India is universally recognized by
drifters as a part of Gondwana, al-
though it is physically attached to Asia.
Geologists agree that the rocks of India
are wholly unlike those of Asia. India's
present position is ascribed to its hav-
ing impinged against the underbelly of
Asia on its journey north, forcing up
the Himalaya Mountains and the Ti-
betan Plateau in the process. This may
seem too much to believe, but it is ac-
tually reasonable. Tibet is 14,000 feet
high, or equivalent to the relief of a
continental slope. Looking north at

issoe ?

Following Wegener's lead, many drifters have made forced reconstruction by
considering the continents as pliable as rubber . . . but are such fits meaningful?

73

Good fits are easy to achieve on a ping-pong-ball-sized globe.

the Himalayan rampart from the resort
city of Darjeeling, one sees the only
land scarp in the world as grand as the
continental slopes which surround all
continents. The reason is precisely the
same — -the Indian continental plate
has underridden the Asian plate, pro-
ducing a double continent 70 km thick.
Since continents float isostatically (i.e.,
are floating hydrostatically), the Ti-
betan Plateau is two continents high
relative to the deep-ocean floor and
one continent high relative to the land
surface of India.

Almost Perfect Fit

Of course, the fitting of South
America into Africa is a straightfor-
ward proposition. The parallelism of

these opposing shorelines across the
South Atlantic has been an inspiration
to all drifters since the time of Wege-
ner. He claimed that the geologic rec-
ord could be read between South
America and Africa as well as if those
continents were pieces of a torn sheet
of newspaper — a bit of an overstate-
ment. The sketch maps of drift invari-
ably show this fit but it apparently
was thought to be only approximate.
Only recently have geologists made the
transposition with cartographic pre-
cision and, surprisingly, discovered that
it dovetails beyond even the fondest
hopes of the drift enthusiasts. Not only
does the bulge of South America fit
into the bight of Africa but the major
bumps fit as well, and to some extent

74

so do the bumps on the bumps. Some
overlaps and gaps remain, but consid-
ering the vicissitudes of geologic his-
tory, the fit is too remarkable to be
fortuitous. It must mean that these two
continents were once juxtaposed. Our
maps show a great ocean between Af-
rica and South America, but geologi-
cally these two continents should be
contiguous and conterminous.

Walter Sproll and I have recently
obtained two other rather good Gond-
wana fits using precise and computer-
ized methods. One of these places

India, including Ceylon, against west-
ern Australia. Another wraps the south
coast of Australia around Antarctica
from the region of the Ross Sea west-
ward. Many problems remain, how-
ever, before entire Gondwana can be
pieced together using strict morpho-
logic criteria. And we must stick to
the rules, too, for the continental jig-
saw game is not without constraints.
It is not cricket, for example, to play
continental leapfrog, although one
South African geologist alleges that
some jump-fits are as good as any of

Continental drift is a game for the Gods as some of the pieces are still missing.

75

India's present position is ascribed to its having impinged against the underbelly
of Asia on its journey north, forcing up the Himalaya Mountains in the process.

those usually promoted by drifters.

The Mesozoic Break-Up

My tentative speculation as to how
the continents dispersed would begin
with the breakup and dispersal of the
two supercontinents Laurasia and
Gondwana commencing late in the
Jurassic Period, in the middle of the
age of dinosaurs, about 150 million
years ago. Most likely this was a unique
event in earth history. Certainly the
continents have not been repeatedly
fragmented through geologic time.
There clearly had been a long period
of geologic calm before this abrupt
dispersal. In the northern hemisphere,

North America swung away from Eur-
asia, rotating on a hinge point in south-
eastern Alaska. Previously Canada had
been tied to Europe, and the eastern
United States faced an Atlantic Ocean
differently configured than now.

The southern continents were prob-
ably clustered together to form Gond-
wana and centered in the mid-Indian
Ocean. An X-shaped rift opened up
which dispersed India, South America-
Africa, Antarctica, and Australia to
the four directions of the compass.
Sometime later, South America split
off from Africa and moved still farther
westward. The main stage of this scat-
tering of crustal pieces was largely

76

Alfred Wegener claimed that the geologic record could be read between South
America and Africa as well as if they were pieces of a torn sheet of newspaper — a
bit of an overstatement.

11

78

Our maps show a great ocean between Africa and South America but because of
geological features and similarities these two continents should be contiguous and
conterminous.

79

s&

The continental drift jigsaw game is not without constraints. It is not cricket, for
example, to play continental leapfrog, although one South African geologist claims
that some jump fits are as good as any of those usually promoted by drifters.

over by the end of the Mesozoic Era,
60 million years ago.

Drift was initiated by the earth's
mantle becoming suddenly "restless"
and undergoing some sort of overturn
in response to having built up too much
internal heat by the slow decay of

radioactive uranium, thorium and po-
tassium. The great world-wide sinu-
ous mid-ocean ridge with its central
rift was the scar resulting from this.
But this seems only approximately
true ; for the continents cannot be fitted
back into this linear birth scar. The

80

mid-ocean rift virtually encircles
Africa (except in the Mediterranean)
forming a "ghost of Africa," an en-
larged outline of the continent, which
apparently expanded outward in all
directions.

The effect of continental breakup
can thus be seen around Africa, for it
was the heartland of Gondwana. In
pre-drift time, only the north coast of
Africa now facing the Mediterranean
was open to the world ocean. The rift-
ing presumably began along the north-
east coastline of India, detaching India
and then Madagascar. The splitting
then extended southward, along Mo-
zambique to Cape Horn, detaching
Antartica. Next the rifting turned
northward up the Atlantic side of

Africa, separating South America.
Thus nearly all of the coastline of
Africa was blocked out and a modern
shoreline established 150-100 million
years ago (in late-Jurassic to mid-
Cretaceous time), the heroic period of
continental drift.

The post-Mesozoic period of qui-
escence was temporary, ending about
15 million years ago. It now appears
that the unstable ends of the snake-
like mid-ocean rift have shifted. One
tail, formerly in the central Pacific,
has swung under western North Amer-
ica opening up the Gulf of California
and arching the entire western limb of
our continent. Another tail, in the
Indian Ocean, has circled into the Red
Sea and down through eastern Africa,

In Australia, the belief is popular that the continents have been dispersed by an
expanding earth.

creating the African Rift Valleys. The
Arabian block has recently been trans-
lated 200 km and rotated counter-
clockwise 8°.

A Final Comment

In Australia the belief in an expand-
ing earth is a popular explanation for
the apparent continental dispersion. It
is proposed that the ocean basins alone
have essentially doubled their surface
area in the past 1 50 million years while
the continents have remained un-
changed in size and shape. I can hardly
accept this interpretation, for such an
expansion would have resulted in a
great lowering of sea level. A funda-
mental relationship exists between the
depth of the sea and the thickness of
the continental plate so that any draw-
down of sea level would have caused

catastrophic denudation of the conti-
nental areas. The resulting erosional
havoc to the continents would be clear-
ly visible in the geologic record, but it
is not. The continents would have been
stripped of their Paleozoic veneers of
marine sediments, but they have not.
Although controversy continues un-
abated, the theory of continental drift
bolstered by the new evidence of pale-
omagnetism, after a period of rejec-
tion, is now very much alive. Project
Mohole is dead, but a new program
jointly sponsored by a consortium of
oceanographic institutions for drilling
many shallow holes in the deep sea
is now in the offing. When the results
of this effort are finally assessed, I am
quite sure that we will find that our
shifty continents are, and have been,
on the move.

82

14

Reprinted from AMERICAN JOURNAL OF SCIENCE Vol. 265
[Amfrican Journal of Scikncf., Vol. 265, March 1967, P. 231-237]

PASSIVE CONTINENTS, SPREADING SEA FLOORS
AND CONTINENTAL RISES: A REPLY

ROBERT S. DIETZ

Environmental Science Services Administration,

Institute for Oceanography,

Silver Spring, Maryland 20910

This reply pertains to a discussion in this journal, by Grant M.
Young, of. my paper "Passive continents, spreading sea doors, and con-
tinental rises" (Dietz, 1966) and earlier publications (Dietz, 1963a; Dietz
and Holden, 1966a). It is a pleasure to attempt to bring into sharper
focus our differing viewpoints. I would like to thank Walter Sproll and
Rose Hudson for assistance in preparing this manuscript.

Multiple orogenies— -There are, of course, many examples of a
single region being affected by a later orogeny; Young cites a few for
Canada. I don't think this adds much to the argument as to whether
foldbelts have been repeatedly tectonized or not. I doubt, however, his
Belcher Island orogeny. Future studies probably will show that this
foldbelt is related to the Hudson Bay (Nastapoka Arc) astrobleme, an
impact scar imposed on the craton and not a true otogenic foldbelt
(Beals, Innes, and Roltenberg, 1963). I argued for more tectonic stabil-
ity in Earth history than is found, for example, in the textbook view-
point that: "We find in the geologic record of all .continents evidence of
ancient movements on a gigantic scale, which repeatedly changed the
geography of the globe. The bedrock has been fractured, bent, or
mashed, and in general the oldest rocks give the most eloquent testi-
mony of crustal unrest, because they have been subject to deformation
repeatedly during geologic time" (Longwell, Knopf, and Flint, 1939, p.
313). This belief in such high otogenic mobility has been modified in
their later editions. Certainly the great antiquity of many little deformed
and metamorphosed sedimentary strata, formerly regarded as Eo-Cam-
brian, argues for more cratonic stability than previously recognized. The
great calm of geologic history is equally as impressive, if not so dramatic,
as the orogenies.

Certainly the Precambrian age of many little-deformed and well-
preserved sedimentary sequences is one of the surprises of modern
geology. We may cite, as Canadian examples, the probably Precambrian
Athabaska formation regarded as between 1400 and 1800 million years
and especially the remarkably similar Dubawnt group regarded as about
1500 million years (Fahrig, 1961). This all seems to be most understand-
able if eugeosynclinal Archean-type beds become tectonized by virtue of
their geologic position mostly as collapsed continental rise prisms. The
Ep-Archean unconformity then is "profound" only because ensialic plat-
form beds laid down on the craton tend not to become tectonized.

231

232 Reply

My principal disagreement with Stockwell's (1964) interpretation
of the Grenville foldbelt is that I doubt'that a preexisting continental
strip was orogenized. Sialic material, yes, for a continental rise is com-
posed of sialic detritus, and juvenile magmas would also be sialic by my
concept. Regarding the number of orogenies involved, I presume that
there may have been several, but these probably were pulses in the one
grand Grenville orogeny sensu lato—as with the Appalachian orogeny
which, in the broad sense, occupied the entire mid- and late-Paleozoic.
It seems to me that the abrupt transition across the Grenville Front is
most readily accounted for by a continental rise.

Young cites some Canadian opinion favoring multiple orogenies.
On the other hand, Wilson (1949) states that he "knows of no evidence
which proves that any large area of the Canadian Shield has been sub-
ject to two periods of folding and mountain building differing widely
in time". Without doubt the final word on this subject remains to be
spoken.

Sialic or simatic basements to eugeosynclines?.— Young quotes Dott
(1964) to the effect that simatic basements are wholly unknown beneath
eugeosynclines. It should be pointed out, in the first place, that geologists
tend to restrict the term basement to granitoid rocks and so by definition
rule out any simatic rock found as qualifying as basement. And we can
probably generalize Dott's statement to say that no basements under
eugeosynclines are known with certainty. For example, the basement
of the crystalline Appalachian orogen is unknown; but we may argue,
with some logic, that if it were sialic it should appear, as granitoid rocks
are low density masses. Simatic rocks, on the other hand, are heavy and
would tend not to rise. It is also not clear what Dott would admit as
oceanic rocks. For my part, spilites, serpentinites, and ultramafics (most-
ly peridotite) all qualify (Dietz, 1963b). These do occur in eugeosyn-
clines and, I suppose, constitute fragmented and tectonically incorporated
pieces of the basement. By my geosynclinal concept, one would expect
such tectonic incorporation of simatic rocks rather than any continuous
basement (as with the sialic basement under a miogeosyncline) because
the foldbelt is as thick as the continental plate.

Young refers to the Grenville as Proterozoic rocks laid down on a
continental plate that was not passive; but my point was that such highly
orogenized rock could not have been laid down on any continent at all
but rather ensimatically. Some confusion is introduced as Young follows
the new Precambrian classification of Stockwell (1964) which introduces
a new tongue-twisting Greek-root terminology. It is pertinent here to
note that such new subdivisions of the Precambrian based on radio-
metric dating differ markedly from the classical classification based
largely upon degree of deformation and metamorphism. Although now
considered Proterozoic, the Grenville rocks are of the Archean litho-
facies in the classical sense. We can understand such basement complexes

Reply 233

as being of any age if they are collapsed and intruded continental rises
deformed by virtue of their ensimatic position. Such prisms must be
deformed in order to be thickened vviiich then provides the bouyancy
needed to achieve a subaerial position. To achieve continental accre-
tion, rock thicknesses of about 20 miles must be attained; collapsing con-
tinental rises uniquely offer a simple orogenic theory capable of doing
this.

Young asks why the fact that orogenic zones are clean-cut (for
example, llie contact between the crystalline Appalachians and the
folded Appalachians), rather than vaguely merging, favors the collapsing
continental rise interpretation of eugeosynclines. The other principal
orogenic theories are the tectogene and the tectonic-borderland/island-
arc concepts. Both of these involve the orogenizing of a previously exist-
ing continental sialic plate. Such a plate is a uniform realm so it is not
clear why a tcctonized foldbelt should cease abruptly. On the other hand,
by the collapsing continental rise concept, the line between an accre-
tionary orogen and the older craton is the continental margin (the con-
tinental slope) , a realm of abrupt topographic, geologic, and petrologic
contrast. The Grenville Front may be a zone of some faulting, especially
thrusting. Stockwell's (1965) tectonic map of Canada shows about one-
third of the front to be a fault contact, but more fundamentally I sup-
pose that it may be the trace of a former continental slope.

Young cites some evidence that the Sudbury lopolith could not have
been emplaced near the pre-Grenville continental margin contra my
belief that its deformation (a squashing from circular to oval) may only
provide evidence for deformation in proximity to a continental margin
rather than cratonic interior orogenesis. May I suggest that the wedge
of Huronian sediments thickening toward the Grenville Front around
this region may provide a Precambrian example of a continental-margin
terrace wedge or miogeosyncline (Dietz and Holden, 1966b). In their classic
paper on the Dissappearance of the Huronian, Quirke and Collins (1930)
were faced with the problem of a sedimentary wedge which became ever
thicker and then abruptly terminated at what is now called the Gren-
ville Front. They "solved" this problem by presuming granitization of
sediment on a vast scale within the Grenville zone. It is not necessary,
however, to presume that a sedimentary wedge that thickens must again
pinch out, implying that all thick sedimentary deposits are laid down in
basins. Continental terrace wedges laid down on a continental shelf
characteristically thicken out if modern examples are taken as diagnostic
(Dietz and Holden, 1966b). Accordingly, the disappearance of the Huron-
ian sediments may be an entirely normal termination by thickening out
at the pre-Grenville continental margin.

Accretion of North America— It is better to consider the case for
accretion in reference to the younger crystalline Appalachian orogen
than to the Grenville. To disprove accretion, Young shows some outliers

234 Reply

of Precambrian near the continental nfargin in his figure 1; and they
are also referred to by Gastil (1960) . Three examples exist in the north-
ern Appalachian belt— in Massachusetts, Nova Scotia, and Newfoundland.
But these are all in the age range from 500 to 800 million years, and so
are Cambrian to Eo-Cambrian, and thus are not outliers of the Gren-
ville rock dated as around 1000 million years (P. King, personal com-
munication; tectonic map of North America, in preparation). Thus their
inclusion in the Appalachian bell is consonant with accretion and sug-
gests only that some Appalachian thermal events date into the late Pre-
cambrian.

An alternate suggestion is Wilson's (1966) who regards them as
fragments of Europe and Africa left after the lower Paleozoic impinge-
ment and subsequent separation of Eur-Africa. I personally am not in-
clined to accept this speculation, but it is an example of the type of
transformation that may be needed before the problem of accretion
is finally worked out.

The existence of the Baltimore gneiss within the crystalline Ap-
palachians is not so readily explained, as these are rocks of Grenville age.
The term gneiss suggests sialic basement, but Hopson (1964) considers
that this gneiss most likely has been metamorphosed from volcanic rocks
similar to the Eocene-Oligocene spillitic suit of the Pacific Northwest.
This suggests that they are the metamorphosed equivalents of oceanic
rocks (layer 2) rather than continental sialic basement.

The San Gabriel norite-anorthosite complex of California is cited
by Young as evidence for no accretion to North America. I have com-
mented upon this elsewhere (Dietz, 1965) noting other possible ways to
account, from the accretionary viewpoint, for this anomalous outlier of
ancient rock. One may wonder, in the first place, if this largely mafic
mass qualifies as a true cratonic block; it may be an entrapped sea floor
fragment. Silver and others (1963) emphasize its anomalous nature by
stating: "These age determinations provide the first compelling evidence
of Precambrian rocks on the west coast of North America". The San
Gabriel mass is probably an allocthonous block rather than an outcrop
of a Precambrian terrane underlying the California coast.

Accretion and continental drift.— Stockwell's tectonic map of Canada
does, in my opinion, support continental accretion if we look for what is
usual rather than anomalous. Considering the vicissitudes of geologic
history, we cannot, of course, expect accreted continents to have a perfect
onion-like structure— only to be more onionlike than layer cake-like
(Dietz, 1966). Furthermore, accretion and continental drift (rifting and
drifting apart) are complementary processes under my sea-lloor spreading
concept, so that we cannot expect the final accretionary pattern to appear
until we solve the drift problem.

Fitch's (1965) chelozonic map for showing the structural unity of
North America and Europe by radiometric age zonation strikes me as

Reply 235

reasonably consistent with continental accretion by collapsing contin-
ental vises if Ave exclude the bulge of Africa as ever being against the
United Slates. The main discrepancy lies in the north-trending Varisco-
Caledonian gcosyncline extending through Norway and marginal to
Greenland. If Laurasia still existed in the Paleozoic, as seems likely, this
would have to lip ronslrnod to be an inn acralonic but not necessarily
ensialic geosynclinc.

The foldhelts off western North America are concordant as should
cluirai lei i/c ace rolionaiy bells. Oil eastern Canada llic fold bell appears
to become discordant, striking toward the deep sea. Here the continental
slopes are probably modified rift scars remaining from the break-up of
Laurasia in (he Mcsozoic. Regarding the continental-drift fit of con-
tinents around the Atlantic Ocean, Walter Sproll and I (ms) have
studied the continental drift jigsaw problem using precise cartographic
methods and a computer program. Our present view is that the bulge
of Africa juxtaposition of the United States is really a misfit; some others
seem valid. Thus we prefer the Laurasia-Gondwana option for the drift
reconstruction of continents. Accordingly the Appalachian foldbelt
would have formed marginal to a continent and not intracratonically.
An alternate suggestion by Wilson, which also is consistent with the
Appalachian geosyncline being extra-cratonic, has already been noted
above.

Universal sialic crust and intracratonic geosynclines?.—rTo avoid
totally the continent-edge problem, one must assume a universal sialic
crust in Precambrian time. Creer (1965), for example, has supposedly
demonstrated a remarkable fit of the present continents on a globe of a
radius approximately 0.55 times that of the present Earth, so that the
continents once would have formed a continuous "skin" around the
Earth. With universal sial, we would have no ocean basins and all the
continents would have been covered to a depth of 2440 meters even for
an Earth of the present diameter— an absurdity. To believe that nearly
all the hydrosphere is of recent generation is equally untenable

The small scale and distorted projection used by Creer make it im-
possible for the reader to evaluate his universal fit. But how poor the
fit must be is indicated by Creer's initial assumption that the contin-
ents occupy only 30 percent of the present Earth. While the Earth's sur-
face is 29 percent land, it is correct to include also as continents the
continental shelves and the slopes down to the 1000-fathom isobath.
The Earth is covered by 40 percent continental crust and 60 percent
oceanic crust down to the 1000-meter isobath (Hess, 1962). Menard (1964,
p. 235) gives the area of continental crust as 2.1 x 108km2 and the area of
oceanic crust as 3.0 x 108km2, or 41 percent continental crust. This alone
means that on an Earth with a radius of only 55 percent that of the mod-
ern Earth, the sialic skin would have a one-third overlap. So how can we
accept Creer's fit?

236 Reply

Young prefers to believe that all geosynclines are intracratonic.
Some geosynclines are, of course, intracratonic, but my view is, briefly,
that those that consist of the mio-eugeosyncline couplet are formed mar-
ginal to continents with the former being ensialic and the latter ensima-
tic as evidenced by entrapped presumed sea-floor rocks (spilites, serpen-
tines, and peridotites). Some intracratonic geosynclines were nevertheless
ensimatic, for example the Ural geosyncline as evidenced by its ser-
pentine belt (Dietz, 1963b). Delving further into this discussion would
be beyond the scope of this reply.

Young suggests that the present positioning of deep oceans over
simatic crust is atypical of geologic history. But, for reasons of isostasy,
the division of the world into sialic and simatic realms is necessary for
the very existence of ocean basins. And these are clearly needed to con-
fine the 1350 x 10° km3 of seawater present on Earth, a volume which
cannot have grown appreciably since Precambrian time.

To fragment an initial universal sialic crust is another way out, but
I cannot take the expanding Earth hypothesis of Egyed (1961) and others
seriously. Egyed supposed that the Earth had an original density of 35
and enlarged at a rate of 7 millimeters per annum. His sole geologic
argument concerns the withdrawal of water from the continents during
the Phanerozoic, using the paleogeographic maps of the Termiers and of
Strakov. For several reasons this evidence is specious, for example: (1)
Egyed considers the continents to be now 100 percent emergent but they
are only about 88 percent if we include the continental shelves to the
shelf-break found usually at about 65 fathoms. Also not taken into con-
sideration is the atypical withdrawal of water due to glacial ice. The
return of this water to the ocean would cause North America to be about
25 percent submerged. (2) The extent of transgression and regression of
shallow platform seas across the continents is largely a measure of the de-
gree of ultrapeneplanation and nothing else. (3) It is generally recog-
nized that continents have maintained high relief during the Geno/oic
for tectonic reasons and not because the deep ocean has become shal-
lower. (4) If continents actually have gained "freeboard" since the Pre-
cambrian, they should be largely stripped of their extensive Paleozoic
cover, but they have not been.

Concluding remarks.— I, of course, do adhere to actualism and uni-
formitarianism, but is this "unwarranted"? It seems to me that the
burden of proof falls upon those who would downgrade this principle
of geology by arguing, for example, that there are no modern examples
of the ancient orthogeosynclines or, conversely, that the modern para-
liageogynclines (like the Atlantic and Gulf modern accumulations)
belong to an entirely new species. If so, where are the ancient continental
terrace deposits and the continental-rise prisms? The usual ensialic
geosynclinal theories ignore the basic dichotomy of the Earth into sialic
and simatic segments, and the presence of the continental slope (the great
scarp connecting these two realms) is ignored.

Reply 237

References

Beds. C. S., Inncs, M. J. S., and Rottenbcrg, J. A., 1003, Fossil meteorite craters, in

Middlchursi, B. M., and Knipcr, G. P., The Solar System, v. 1: Chicago, Illinois,

Univ. Chicago Press, p. 235-281.
Crccr, K. M., 1905, An expanding cartli?: Nature, v. 20"), p. 539-544.
Diet/, R. S., 1903a, Collapsing continental rises: An actualist ic concept of geosynclines

and mountain building: Jour. Geology, v. 71, p. 311-333.
, 11)631), Alpine serpentines as oceanic rind fragments: Gcol. Soc. America

Bull., v. 74, p. 917-952.

-, 1905, Collapsing continental rises: an actualisiic concept of geosynclines

and mountain building: a reply: Jour. Geology, v. 73, p. 901-906.

1900, Passive continents, spreading sea floors and collapsing continental

rises: Am. Jour. Sci., v. 204. p. 177-193.
Dictz, R. S., and Holder), J. C, 1900a, Deep sea deposits in but not on the continents:

Am. Assoc. Petroleum Geologists Bull., v. 50, p. 351-302.
1900b, Miogcoclines (miogeosynclines) in space and time: Jour. Geology,

v. 7-1. p. 566-583.
Dictz, R. S., and Sproll, W, P., 1960, Marine geologic aspects of continental drift and

Gondwana [abs.]: Intcrnat. Occanog. Cong., 2nd, Moscow 1906, p. 97-98.
Dott. R. II., Jr., 1901, Mobile belts, sedimentation and orogenesis: New York Acad.

Sii. Trans., sci. 2, v. 27, p. 135-114.
Egvcd. I,.. 1901, The expanding earth: New York Acad. Sci. Trans., ser. 2, v. 23.

' p. 424-132.
Fahrig, \V. F., 1901. The gcologv of the Athabaska formation: Canada Geol. Survey

Bull., v. OS, 41 [).

Fitch. F. J., 1965, The structural unity of the reconsiructed North Atlantic continent:

Royal Soc. [London] Philos. Trans., v. 258, p. 191-194.
Gastil, R. G, 1900, The distribution of mineral dates in time and space: Am. Jour.

Sci., v. 258. p. 1-35.
Hess. II. II., 1902, History of ocean basins, in F.ngel, A. F. J.. James, H. L., and

Leonard, B. F. eds. Pctrologic Studies: a volume in honor of A. F. Buddington:

New York, Gcol. Soc. America, p. 599-020.
Hopson, C. A., 1964, The crystalline rocks of Howard and Montgomery Counties

[Maryland], in I lie geology of Howard and Montgomery Counties: Baltimore,

M. inland, Mankind Gcol. Survey, p. 27-215.
Longwcll, C. R.. Knopf, Adolph, and Flint, R. F., 1939, A tcxtlwok of geology: New

York, John Wiley & Sons, Inc., 543 p.
Menard, H. W., 1901. Marine geology of the Pacific: McGraw-Hill Hook Co., New York,

271 p.
Quirke, T. T., and Collins, W. H., 1930, The disappearance of the Huronian: Canada

Gcol. Survey M< m. 100, 129 p.
Silver, L. T., "M< Kinney, C. R., Dcutsch, S., and Bolinger, J., 1903, Prccambrian age

determinations in the western San Gabriel Mountains, California: Jour. Gcologv,

v. 71, p. 190-214.
Siockwell, C. II., 1901, Age determinations and geological studies: Canada Geol.

Survey, Dcpt. Mines Tech. Surveys Paper 64-117, pt. 2, 30 p.
■ . 1905, Tectonic map of the Canadian Shield: Canada Gcol. Survey, Dcpt.

Mines Tech. Surveys, Map 4-1965.
Wilson, J. T., 1949, Some major structures of the Canadian Shield: Canadian Inst.

Mining Metallurgy Trans., v. 52, p. 231-242.
, 1966, Did the Atlantic close and then re-open?: Nature, v. 211, p. 676-681.

15

Shatter Cone Orientation at Gosses Bluff Astrobleme

by
ROBERT S.

DIETZ

Institute for Oceanography, ESSA,
Miami, Florida

Shatter cones caused by shock fracturing are widely developed at the
Gosses Bluff cryptoexplosion ring structure in central Australia. The
force field can be reconstructed whereby the applied shock arrived
centrally and from above, which is consistent with a cosmic impact.
For this and other reasons, Gosses Bluff is an astrobleme.

Gosses Bluff is an isolated, ring structure of upturned
Lower Palaeozoic rock nearly 3 miles across, situated
in the Amadeus Basin of central Australia, about 200 km
west of Alice Springs. It rises about 700 ft. above the
surrounding plain and is breached by an intermittent
stream, on the north-east side. The term "bluff" is
something of a misnomer, or at least is an inadequate
description.

The interior bowl shaped basin might well be misinter-
preted as a modern meteorite crater (Fig. 1). It is, how-
ever, at the same level of erosion as the surrounding plain
and is not depressed, and there are no remnants of lake
beds which would indicate that a former closed basin exis-
ted here. Furthermore, geological inspection shows that it
is the upturned ring of a great dome, the soft centre (Lara-
pinta) of which has been erosionally gutted, forming the
central bowl. Lining the bowl, the cliff-forming Meerenie
sandstone now stands etched out in high bold relief and is
in turn surrounded by the Pertnjara formation.

If Gosses Bluff is to be interpreted asametoorite structure,
then it can only be regarded as an astrobleme — an ancient
cosmic impact scar, with the bluff per se being the uplifted
central dome of a much larger circular structure. Although
the strata are largely hidden from view outside the bluff,
this interpretation is quite permissible. Gravity and
magnetic surveys at the Bureau of Mineral Resources
show that the bluff marks the centre of a circular deforma-
tion 12 miles wide. A structure of this size is also apparent

as a double "ghost ring" on a Gemini IV photograph
from near space (Fig. 2).

On the basis of much evidence of explosive force, Crook
and Cook1 rejected an earlier view which was widely held,
that Gosses Bluff was a salt diapir: while agreeing that it
was of cryptoexplosive origin, Crook regarded it as an
astrobleme and Cook thought it was crypto volcanic.

Fig. 1. Aerial oblique view of Gosses Bluff lookin" south, showing the
ring of upturned strata.

NATURE. VOL. 216. DECEMBER 16. 1967

1083

with the possible exception of the Vredefort Ring of South
Africa. The coning is least developed in the southern
quadrant and most strongly developed in the northern
quadrant where, in places, nearly all rock talus revealed
shatter coned faces. There is also a strongly preferred
orientation of the cone apexes. In the vertically dipping
Larapinta formation of the central basin and in the
Meerenie sandstone marking the interior "lining'' of the
uplifted ring, tho shatter cones are oriented radially
outward — that is, normal and upward with respect to
bedding. Further out from ground zero (4,000-7,000 ft.)
in the Pertnjara formation, the shatter cones point
generally outward but only at angles of about 60° upward
with respect to bedding. Still further out, at two sites,
shatter coning is oriented upward and parallel to bedding.
The first of these two sites occurs close to the spot where
the road track enters the breach in the bluff ring at a
position about 10,000 ft. north-east of ground zero. The
other site is a hogback, fully detached from the ring
proper and lying about 13,000 ft. north-west of ground

Fig. 2. Gemini photo of Gosses Bluff from near space. Dark circle in

centre marks the upturned ring of rocks 2-5 miles across. Note "ghost

ring" surrounding Gosses Bluff which identifies the full extent of the

structure, 12 miles across.

Cook2 has since revised his opinion and concurs in the
astrobleme interpretation. Daniel Milton and Robin
Brett, who are at present mapping this structure in
detail as a joint project of the US Geological Survey
and Australian Bureau of Mineral Resources, also regard
it as an astrobleme. Their results will soon be published.
This article is limited to a discussion of shatter coning at
Gosses Bluff which further supports this interpretation.

Shatter cones (Fig. 3) wore first noted by Crook and
Cook, who, however, regarded them merely as an inciden-
tal aspect of the structure with the orientation of the
shatter cone being random. Actually, Gosses Bluff is
the most intensely shatter coned of the eighteen sites I
know around the world4 and the cones show a high degree
of preferred orientation which reveals the shock force
field impressed on the structure. Crook and Cook also
doubted the validity of shatter cones as a criterion for
astroblemes, but I have attempted to answer their objec-
tions elsewhere5.

In order to study the shatter cone orientation and
distribution at Gosses Bluff, I made several spoke-like
traverses of the structure commencing at the centre
(ground zero) and working radially outward. Shatter
coning was found to be extensively developed throughout
the structure on a scale which I have never before observed.

(TH mm m Tin in iiri'l IM tn Htxk

K ^<

^F

~t— --'

V

__™m/5i'V% K /'^0v>^_^^,T^/^\/7 7\_

/ * A / ! v v A

/ P/ / -M7 / L \ \M;\

s P

MILES

Fig. 3. A group of shatter cones from the Pertnjara formation showing
orientation oblique to bedding.

Fig. 4. Diagrammatic representation of shatter cone orientations at
Gosses Bluff revealing the shock wave force Held which was impressed
on the structure. A, Impact of bolide and radial spreading of shock
wave instantaneously prior to upheaval. L, Larapinta group; M,
Meerenie sandstone; P, Pertnjara formation. B, Presently observed
orientation of shatter cones (arrows) in upturned strata. Vertical
exaggeration. 2:1.

If one can envision a return of the rocks to their
presumed pre-event, essentially horizontal position, wo
observe a pattern of shatter coning which reconstructs
the shock wave force field which acted on the structure
(Fig. 4). Before upheaval the shatter cones in the central
region pointed directly skyward, those farther out pointed
inward and upward, while those farthest out pointed
radially inward. Because the apices of shatter cones
point toward the impressed shock wave front, this indi-
cates a shock originating from above as, most likely,
the impact of a cosmic body with an apparent or
effective diameter (that is, for purpose of impressing a
parallel shock front) of about 7,000 ft. This derived
diameter may well be beyond the limits of resolution of the

1084

NATURE. VOL. 216, DECEMBER 16, 1967

method and the object could have been much smaller
because we are now examining an inverted mass of rock.
The strata now at ground zero were originally positioned
several thousands of feet deeper. This effect tends to
enhance the normal and upward-to-bedding aspect of
shatter cones. Nevertheless, the apparent diameter
of the bolide would seem to have been at least a few
thousand feet.

As elsewhere, thero are exceptions to the preferred
orientation noted here, but those are sufficiently few to be
regarded as uncommon. Most often, fully inverted cone
orientations are found ; more rarely the cones point
randomly. It is necessary, of course, to distinguish
positive cone faces from negative ones and to measure the
apical direction of a full 360° cone rather than the direc-
tions of various cone segments ; otherwise confusion
results, giving a false appearance of randomness. According
to the impact interpretation, random and inverted orienta-
tions may be ascribed to the reflexion of shock waves
which were still sufficiently intense to cause shatter
coning.

Outcrops in the outer annulus of the structure, and
peripheral to Gosses Bluff proper, are few and far between.
Those outcrops which were found were searched; no
shatter cones were found. Possibly at this range of from
2 to 6 miles from ground zero, shock pressure had already
dropped below the level required to produce shatter
coning (20 to 80 kbars). It would not be surprising, how-
ever, if a further search revealed some shatter coning
on a reduced scale. This places Gosses Bluff in an inter-
mediate position among known shatter coned structures,
in most of which shatter coning is confined to the central
eye (for example, Serpent Mound, Ohio) and the very
large astroblemes where shatter coning extends outwara,
15 miles or more (Sudbury, Vredefort Ring).

At seven sites around the bluff, I found nests of breccia
containing shatter coned fragments. In no case does the
shatter coning extend into the breccia matrix. This is
expected for an astrobleme because, in a cosmic impact.
the target is first engulfed by the shock wave, followed

almost instantaneously, but clearly separated in time, by
brecciation and upheaval.

The bluff is circular in its external limits : the interior
plan, however, as outlined by the Larapinta and Meerenie
formation, has the form of an equilateral triangle with
truncated basal angles and with a sharp apical angle to
the north, giving the internal structure an overall penta-
gonal form. An imaginary line running from north to
south through the apex provides an axis of bilateral
symmetry. Such bilateral symmetry is a well known
aspect of some other astroblemes3 and it could reflect
the arrival of the bolide at an oblique angle from the
south and towards the north. Shatter coning is clearly
more intensely developed at the northern apex of the
structure than elsewhere.

While I remain convinced of the astrobleme interpreta-
tion, numerous questions remain unanswered. First,
what was the original form of the crater ? Second, why
is tho central dome so large (about 3 miles) and uplift
so great (about 10,000 ft.) with respect to the overall
12 mile diameter of the structure ? One suggestion is
that the cosmic bolide was a comet head combining
sufficient excessive hypervelocity and low density to pro-
vide for a surface or very shallow detonation which
enhances tho central uplift effect6. Third, what is the
age of the structure ? As Crook and Cook point out,
Gosses Bluff appears to have been exhumed after early
Tertiary peneplanation so that a Mesozoic age seems
reasonable.

I thank Milton, Brett and George Berryman for
hospitality and the Bureau of Mineral Resources of
Australia for logistical support.

Received November 9, 1967.

1 Crook, K.,and Cook, P., J. Geol. Soc. Austral. ,13, 495 (1966).

- Cook, P., J. Geol. (in the press).

3 Boon, J., and Albritton, C, Fid. Lab., 5, 53 (1937).

4 Dietz, R., Shatter Cones and Astroblemes., Trans. Lunar Geol. Field Con/.,

Bend, Oregon, 1965, 25 (Oregon Dept. Geol. and Min. Ind.) (1966).
' Dietz, R., NASA Symposium on Shock Metamorphism in Natural Material"

(in the press).
6 Sun, J.. Trans. Amer. Geophys. Un.AS. 186 (1967).

Reprinted from THE OIL AND GAS JOURNAL Vol. 65, No. 8

16

De Soto Canyon reveals salt trends

Seismic-reflection profiles in Gulf of Mexico discover
domal structures and a buried eroded slope.

A 1963 reconnaissance seismic-
reflection profile study by the
USC&GSS Hydrographer, using a
1,000-joule arcer, was made in the
De Soto Canyon area of the north-
eastern Gulf of Mexico.

An apparent domal closure and
shallow faulting were discovered de-
spite poor penetration caused by
equipment failure.

During 1965, approximately 950
miles of seismic reflection profiles
were made by the Hydrographer in
the De Soto Canyon area (Fig. 2)
using the 1963 results as a guide.
The 1965 study located five domes
in the De Soto Canyon area and two
more domes to the southwest near
the 1,000-fathom depth curve. No
major faulting was observed.

A buried eroded slope, apparent-
ly an old eroded shoreline, was dis-
covered. Although the seismic-re-
flection profiler studies were started
primarily to determine the geology
of De Soto Canyon, only the domal
structures and the eroded slope will
be discussed in this report, which
will use the 1965 results exclusively.

A 3,000-joule arcer, activated
every 2 sec, supplied a broad-band
sound source whose energy was con-
centrated between 150 and 300 cps.
The reflected returns were not fil-
tered.

A 14-ft hydrophone array with
10 variable inductance hydrophones
was towed 300 ft aft on the port
side, and the arcer was towed 250
ft aft on the starboard side of the
ship at speeds varying between 6
and 10 knots. Maximum penetra-
tion was approximately 0.7 sec, or
almost 2,000 ft assuming a velocity
of 5,500 fps. Loran A provided
navigation control.

Geological age postulations.

Groups of strata with a visible dif-
ference in character are labeled A,
B, C, D, and E in Figs. 3 through 7.
No direct evidence is available for
the age of the strata in the study.
Project of the geology of Florida
(Puri and Vernon, 1964) seaward
and the extension of refraction data

SHADED AREA shows salt trends and their possible extension into areas
where domes have been located by seismic reflection study. Fig. 1.

R. N. Harbison

Environmental Science Services

Administration

Institute for Oceanography

Silver Spring, Md.

reported by Antoine and Harding
(1965) for the northeastern Gulf of
Mexico were used in the following
age postulations:

Group A strata — Pleistocene and
Recent

Group B strata — Pliocene-Pleisto-
cene

Group C strata — Upper Tertiary

Group D strata — Lower or Mid-
dle Tertiary

Group E strata — Lower Tertiary
or Upper Cretaceous

Eroded slope. An eroded slope,
complicating the strata correlations
somewhat, crosses the area of study
approximately between 29°27' and
29°28'N latitudes and 86°31' to
87°13'W longitudes.

The slope probably extends east
and west of the area of track-line
coverage. This slope is buried be-
neath the sea floor except for a
short portion between 86° 5 3' and
86°59'W longitudes which forms
the anomalous east-west portion of

northern De Soto Canyon. A report
by Jordan (1951) contains a de-
tailed bathymetric chart showing
this east-west portion as the north-
ern limits of De Soto Canyon.

This east-west-trending slope is
interpreted as a transgressing Ter-
tiary shoreline which was eroded
into Group D strata. Group E strata
continue unbroken beneath this old
shoreline indicating that the result-
ing slope is not a fault scarp (Figs.
4 and 6).

Following the erosion of the Ter-
tiary shoreline, sea level rose with
respect to the land elevation and
group C strata were draped over
the slope east of dome 1. Group C
strata thickened to the south and
west forming the prominent reflec-
tors seen above and on the flanks
of domes 2, 3, and 4 (Figs. 4, 5,
and 6).

Group B strata were deposited
under deeper water than group C
strata completely masking the ex-
pression of the old shoreline slope
beneath the sea floor east of dome 1 .

The seismic reflection profiles
show the old buried shoreline to be
approximately 500 ft deeper near
87° 13' W longitude than at 86°31'
W longitude suggesting that the

124

THE OIL AND GAS JOURNAL • FEBRUARY 20, 1967

northeastern Gulf of Mexico has
been subsiding more rapidly on the
west than on the east.

This east-west-trending Tertiary
shoreline, appearing on or near the
flanks of domes 1 and 2, will be
elaborated on in the descriptions of
Figs. 3, 4, and 6.

Domal features. A family of five
domes and six more features inter-
preted as possible domes were lo-
cated within the vicinity of De Soto
Canyon. Two other domes were lo-
cated near the 1,000-fathom con-
tour west of De Soto Canyon
(Fig. 2).

The five domal features in the
De Soto Canyon area were deter-
mined by intersecting profiles. The
six possible domes were indicated
usually on a single trackline by one
or more of the following: anomalous
dip, apparent closure, shallow fault-
ing, small localized graben. The cri-
teria listed for possible domes are
also present in the known domal
structures found by intersecting pro-
files.

A north-south profile through
dome 1 (Fig. 3) reveals a localized
north-south fault extending north-
ward from dome 1 through group
D strata within the plane of the
profile. A sea-floor expression of a
rise approximately 200 ft high and
2.000 ft wide above dome 1 ap-
pears to be an erosional remanent
of outcropping strata or caprock.

If it is an outcrop, it is probably
resistant group E strata. The ero-
sional slope interpreted as an old
shoreline lies on the south flank
of dome 1. Most of D strata have
been eroded away south of dome 1.
Another cycle of erosion occurred
following the deposition of group
B strata eroding and reworking the
groups B and C strata south of
dome 1 . This same erosional cycle

LOCATION of seismic reflection profile tracklines are shown here with
domes 1 through 7 and possible domes 8 through 13. Contours of sea
floor are in fathoms. Fig. 2.

eroded B and C strata to form the
east slope of the north-south por-
tion of De Soto Canyon. Group A
strata were deposited unconform-
ably over this eroded surface.

South of dome 1, group E strata
appear to have been pierced by the
dome, but have not been breached
by the erosion which removed most
of the younger group D strata.

Fig. 4 is a northwest-southeast
profile through domes 1 and 2. The
group D strata have been pushed up
almost to the sea floor on the north-
west and southwest sides of dome 1 ,
truncated by erosion and covered by
a thin layer of group A sediments.

A slope extending from near the
top of the southeastern portion of
dome 1 down to a point approxi-
mately halfway to dome 2 has been
eroded into the group D strata.
Southeast of this slope, only the
lower one-third of group D strata

SOUTH

NORTH-SOUTH profile through dome 1. Fig. 3.

survived erosion. The slope is the
east-west trending Tertiary shore-
line mentioned previously.

Groups B and C thin rapidly on
the northwest and southeast flanks
of dome 2, suggesting that they were
draped over the dome when it was
a topographic high or that they
were deposited contemporaneously
with growth of the dome.

Notice the comparison with the
dip and thinning of groups B and
C strata on the flanks of dome 2
with the dip and thinning of group
D strata on the northwest flank of
dome 1 .

It appears that group D strata on
the northwest side of dome 1 were
pushed up, while groups B and C
strata were deposited during or after
the uplift of dome 2. The center
of dome 2 seems to have collapsed,
causing a block approximately one-
third the diameter of the dome to be
dropped down into the center of
the domal uplift.

The strata on both sides of the
down - dropped block have been
bowed down causing apparent clo-
sures in the shape of arches.

Small local faults occur over
dome 2 and on its flanks. A rather
flat-topped expression shown by the
deepest reflector may be the re-
flection of a salt-sediment contact.

A north-south profile (Fig. 6)
through dome 2 shows the Tertiary
shoreline on the north side of domal
uplift. Group C strata are built out

THE OIL AND GAS JOURNAL • FEBRUARY 20, 1967

125

SEC.
0.4

SOUTHEAST

< 3.3 NAUTICAL MILES >

NORTHWEST

A NORTHWEST-southeast profile from dome 1 to dome 2. Fig. 4.

over the slope eroded into group D
strata and up on the north flank of
dome 2, giving a false impression
of a rim syncline.

Group E strata seem to have been
pushed up almost to the sea floor
on the north side of dome 2 and
sag into the center of the domal up-
lift. Group C strata dip steeply
away from the south flank of dome
2 and probably overlie group E
strata which is possibly too steep
to give a reflected return.

The same flat - topped reflector
mentioned previously as a possible
salt contact in the northwest-south-
east profile is also present here, but
with another reflector between it
and E strata on the north, which
may be caprock.

An east southeast-west northwest
profile through dome 3 (Fig. 5a)
shows group C strata sagging down
into the down-dropped center of the
domal uplift causing arch-like struc-
tures in group C strata over both
sides of dome 3. The arch-like struc-
ture on the west-northwest side of
dome 3 has suffered faulting on
both sides of its apparently closed
portion. This closed portion has
dropped down between these faults.
The localized faults over dome 3
reach up to very near the sea floor.

Fig. 5b shows a north-south
profile through dome 3. An expres-
sion of a graben with localized fault-
ing extending nearly to the sea floor,
can be seen over the center of the
dome. The group C strata are bowed
down into graben as they were in
Fig. 5a, but in a less pronounced
manner. A possible salt-sediment
contact in the shape of a dome lies
approximately 0.6 sec below the
sea floor in Fig. 5b.

An east-west profile through
dome 4 (Fig. 6) shows numerous
small faults over the domal uplift.
Group C strata have collapsed into
the center of the dome especially
on the east side, causing an appar-
ent closure similar to those men-
tioned in the descriptions of domes
2 and 3.

From near the center of dome 4
westward, groups B and C strata
have been truncated by erosion
which formed the east slope of De
Soto Canyon. A thin layer of group
A strata unconformably overlie the
truncated strata of groups B and C.
A possible salt-sediment contact
lies approximately 0.5 sec (approx-
imately 1,400 ft) below the sea
floor.

Fig. 7 is an east northeast-west
southwest profile from dome 1

WEST-NW

<30 NAUT. Ml-^

EAST-SE S^C6- [SOUTH

SEAJLOOR
BAT

AN EAST southeast-west northwest profile through dome 3 (Fig. 5a, left),
and a north-south profile through dome 3 (Fig. 5b, right).

through the east-west portion of
northern De Soto Canyon to dome
5. Group E strata can be seen as
an unbroken horizon between these
two domes, indicating the absence
of faulting along the north-south
portion of De Soto Canyon at this
point.

A north northeast - south south-
west profile through dome 5 shows
the east-west erosional slope, which
is interpreted as a Tertiary shore-
line, north of the domal uplift. A
graben and associated faulting ex-
tends up from group E strata near
the center of dome 5 to the vicinity
of the sea floor. Strata have been
bowed down adjacent to the graben.

Fig. 8 is a profile through domes
6 and 7 which protrude over 1,000
ft above the surrounding sea floor.
The direction of the profile changes
from east-west to northeast-south-
west over dome 6. Extremely rough
weather caused the poor record
quality.

Dome 6 appears as a closed high
on the USC&GSS chart 1115 at
28°42' N latitude and 88°05' W
longitude. It resembles an unburied
version of the northwest-southwest
profile through dome 2.

Dome 7 appears as a closed high
on a detailed bathymetric chart by
Jordan (1951). It is located at ap-
proximately 28°52' N latitude, and
87°58' W longitude. A sea-floor de-
pression southwest of dome 7 has
a very steep slope on its southwest
side in the first shallow reflector.
The direct return from arcer to
the hydrophone array occurs every
2 sec and partially obscures the top
of dome 7.

Domes 1 through 5 in the De
Soto Canyon area are interpreted
as salt domes because of their size,
shape, associated localized faulting,
and graben over the crests. They

126

THE OIL AND GAS JOURNAL • FEBRUARY 20, 1967

AN EAST-west profile through dome 4, left, and a north-south profile through dome 2, right. Fig. 6.

resemble very closely the published
data on known salt domes.

The collapsing of sediments into
the centers of domes 2, 3, 4, and
5, along with the associated fault-
ing which nearly reaches the sea
floor, perhaps explains the depres-
sions which form circular lakes over
some salt domes in the marsh areas
of South Louisiana.

These closed depressions over
shallow domes may provide the
proper environment for the gener-
ation of caprock and native sulfur.

Domes 6 and 7, which protrude
through the sea floor, are probably
salt domes which are growing con-
temporaneously with deposition of
sediments.

The possible-domes (8 through
13) in the De Soto Canyon area
may be salt domes, relict salt struc-

tures, shale masses or fault struc-
tures. It is likely that these fea-
tures 8 through 13 are deep-seated
salt structures.

Possible extensions of known
salt trends. Fig. 1 illustrates pro-
jections of known salt trends into
the areas of domal occurrence in the
seismic reflection profile study area.

The domes in the De Soto Can-
yon area are in line with the East
Texas, North Louisiana, Mississip-
pi, and Alabama salt trend. Antoine
(1965) reported a dome in the De
Soto Canyon area, and another pos-
sible salt structure aligned with the
Pickens, Gilberton, and Pollard
fault zone (the North Louisiana,
Mississippi, and Alabama salt
trend).

The two domal features 6 and 7

west of the De Soto Canyon area
suggest that the South Louisiana
salt trend may also extend into the
De Soto Canyon area, forming a
juncture with the northern salt
trend.

Bibliography

Antoine, J. W. and J. L. Harding, "Struc-
ture Beneath Continental Shelf, Northeastern
Gulf of Mexico": Bulletin of the AAPG,
Vol. 49, No. 2, 1965. pp. 157-171.

Antoine, J. W., "Structural Features Under
Continental Shelf off the Florida Panhandle
Revealed by Seismic Reflection Measure-
ments": Abstract; Geophysics, Vol. 30, No.
6, 1965, p. 1,228.

Puri, H. S. and R. O. Vernon, "Summary
of the Geology of Florida and a Guidebook
to the Classic Exposures": Florida Geological
Survey, Special Publication No. 5, revised,
1964.

Jordan, G. F., "Continental Slope off
Appalachicola, Florida": Bulletin of the
AAPG, Vol. 35, No. 9, 1951, pp. 1,978-
1,993.

JlSrf;

DOME NO. 1

AN EAST northeast-west southwest profile from dome 1 to dome 5. Fig. 7.

.,'■;; , =■- i . j:'f v ; if Li' 1> i*§

< 5.5 NAUTICAL MJLES*> ' '" jj I '■' rJl'{ i{ ' ' 1 1

L

A PROFILE over domes 6 and 7. Fig. 8.

lSI5JBiaJ3ISJ^(3JSMafaMSI3J2J3J2KJ^^

17 a

Reprinted from PROCEEDINGS OF THE WORLD DREDGING CONFERENCE, 196^

a

ENVIRONMENTAL EFFECTS OF DREDGING
AND SPOIL DEPOSITION

by

W. HARRISON

Land and Sea Interaction Laboratory

Institute for Oceanography, ESSA

Atlantic Marine Center

Norfolk, Virginia

GlfSJiiMa/SJEJeMe^^

535

ABSTRACT

The effects of dredging or associated spoil disposal
are reviewed for two dredge operations in the lower
Chesapeake Bay area. In one operation 1, 260, 000 cu.
yds. of spoil were dumped from a hopper dredge on a
target rectangle measuring 0. 5 by 1.0 naut. mi. , in
water depths of 75 to 96 ft. Owing to the strong simi-
larity in composition and size gradation of the spoil and
natural bottom sediments, it was necessary to make
an exhaustive sedimentological study of the dump area
in order to identify concentrations of spoil. The most
promising method for delineation of spoil utilized ano-
malous profiles of sediment strength and void ratio, as
determined from piston-core samples. Spoil disposal
appeared to have only a transitory effect on the popula-
tions of infauna and epifauna. Resettlement in both the
areas of dredging and spoil deposition was very rapid,
occurring by active migration of the animals and by the
hydrodynamic distribution .of juvenile or larval stages.
It was found important to differentiate between transitory
high populations of juveniles at certain seasons and nor-
mal faunal distributional patterns when attempting to
assess the effects of dredging or spoil deposition on
benthic organisms by means of "before" and "after" fau-
nal surveys.

Precise monitoring of possible spoil buildup on an
oyster ground near a spoil outfall was accomplished by
repeated diver measurements of the distance from
the water-sediment interface to the tops of steel refer-
ence rods located along the perimeter of the oyster
ground. Details of the field procedures are presented.
Because the observed bottom fluctuations can be attri-
buted exclusively to natural variations in erosion and
deposition, the data presented constitute a useful refer-
ence on the variability in bottom level to be expected in
the absence of spoil-induced sedimentation. The varia-
tions are of the order of a few millimeters per week.

537

INTRODUCTION

The questions most frequently asked by those con-
cerned with environmental pollution by dredging activity
have to do with the effects of dredging or spoil deposi-
tion on the local marine life. Most concern is usually
directed toward the marine organisms found upon or
within the bottom in the affected areas, rather than to-
ward the organisms in the water column itself. The
problem of documenting such effects usually resolves
into obtaining adequate data on changes in local bottom
elevations, suspended sediment concentrations, and
animal populations. What may be considered adequate
data for one study will quite often be irrelevant for an-
other, due basically to differing requirements or envi-
ronmental conditions. It is the purpose of this paper to
review two recent studies that were designed to deter-
mine the effects of dredging or spoil disposal on benthic
organisms in lower Chesapeake Bay. At two of the sites
(fig. 1, A and B) there was significant damage to faunal
populations. At sites C and D damage was anticipated,
but failed to materialize. The monitoring techniques
developed for the last two sites should prove useful, how-
ever, and the data on natural variability in bottom ele-
vations may provide helpful reference material for
studies in which variations in spoil-induced bottom
changes approach variations due to natural causes.. As
the body of such case histories grows, it becomes pro-
gressively easier to develop rational design criteria for
each successive monitoring study. Eventually, those
involved in specifying restrictions on the activities of
dredging or spoil deposition will have a reasonably ade-
quate base upon which to work.

DREDGING AND SPOIL DEPOSITION IN OPEN BAY

WATERS

In connection with a 1961 dredging project on the
Rappahannock Shoal (fig. 2), the Corps of Engineers,
U. S. Army, dumped 1, 260, 000 cu. yds. of spoil on a
target area six to ten nautical miles from the dredged
area, over a two-week period. This target rectangle
measured 0. 5 by 1. 0 naut. mi. , and water depths were
between 75 and 96 feet. A contract between the U. S.

538

Army, Corps of Engineers, and the Virginia Institute of
Marine Science (VIMS) provided an instrument for in-
vestigating the effects of the spoil dumping in the target
area, and numerous investigations were carried out by
VIMS personnel in "before" and "after" surveys of the
sediments and animal populations in the area. Many of
the important results and conclusions of the sedimento-
logical and faunal surveys are described elsewhere
(Harrison, and others, 1964).

The most difficult problem encountered in the study
was that of locating concentrations of spoil. The spoil
material consisted largely of sandy silt and it was
dumped in an area also composed dominantly of sandy
silts. Sediment mineralogy was of little help in identi-
fying spoil, owing to its uniformity within the entire
area of investigation. Bulk chemical composition like-
wise proved inadequate for spoil differentiation.

In seeking a way to delineate areas of spoil bottom
it was noted that the dumping process might load the
bottom sediments and cause shear failure in them if the
stresses were of sufficient magnitude. Resulting slumps,
flows, or settlings of the bottom materials could be dis-
cerned by strength and void- ratio anomalies or, in cer-
tain cases, by bottom contours.

As commonly reported in the literature of soil
mechanics, concerning natural, normally-consolidated
silty sediments, the relation between undrained shear
strength and effective stress is constant for a given soil.
Thus, the strength must increase with depth if an equi-
librium exists in the sediment. Examination of shear
strength data (Harrison, et al. , 1964, table 3) from a
suite of piston cores from the area of spoil dumping re-
vealed that strength did increase with depth inmost
cores. One core, however, displayed a distinctly ano-
malous strength profile.

A parallel examination of void- ratio data for the
core samples revealed that certain cores displayed
anomalies in the expected reduction of void- ratio with
depth. These anomalies are shown in figure 3 as curves
1, 2, 3, and 4. The minimum thickness of the spoil, at
the two stations where positive identification was made,

539

was probably 10 cm. Various data suggested that the
maximum thickness was only about 40 cm. Hydrogra-
phic surveys of the spoil disposal area did not reveal
any obvious highs or lows on the bottom. This was pro-
bably because the bulk of the spoil was spread thinly by
the prevailing currents over a relatively wide area.

Once the areas of spoil had been established by the
use of the strength and void- ratio data, it was possible
to examine the effects of the deposition on the infaunal
populations. Variations in the numbers of animals and
species present at the two spoil-disposal stations, be-
fore and after dredging, are shown in figure 3. The
animal collections taken one month after dredging show
a marked decrease in numbers of animals and species.
Harrison, et al. (1964, p. 752) did not believe the
decrease was related to seasonality in the animal popu-
lations. Nephtys incisa, a worm, was the most preva-
lent animal of the species found at the two stations one
month after dumping of the spoil. Because of the very-
active nature of this worm, it probably migrated to the
area, immediately after spoil deposition. Juvenile
mollusks, particularly Ensis directus (692 individuals)
made up a large part of the population in the faunal
sampling made six months after spoil dumping, indica-
ting that hydrodynamic factors were effective in reset-
tling the spoil-disposal area. The large numbers of
animals present six months after dredging and spoil
dumping in part reflects seasonal effects in faunal fluc-
tuations. A final sampling made 15 months after spoil
deposition indicated successful recovery of animal
populations in the spoil area.

The effects of dredging on a population of infauna
can be examined by reference to the curves on figure 4
for a station located within the dredged channel (fig. 2).
In the sampling done one month after dredging a marked
decrease in both total numbers of animals and species
was found. This decrease was much greater than could
be accounted for by seasonality alone. Nephtys incisa
made up 7 5 percent (6 individuals) of the population.
Recovery of the infaunal population was relatively com-
plete after a lapse of only six months (fig. 3).

540

These and other data indicate that while dredging
and spoil disposal in the two areas temporarily destroy-
ed the infaunal populations, resettlement and recovery
were fairly rapid. The resettlement occurred by mi-
gration of active species and by hydrodynamic distribu-
tion of many species while in the free- swimming or
free-floating larval stages.

SPOIL DISPOSAL IN PROXIMITY TO AN ESTUARINE
OYSTER GROUND

The author was requested to design a study to moni-
tor the buildup of sediment on an "oyster ground" in the
York River estuary, in response to anticipated spoil
deposition from an outfall (fig. 5, inset) located 0. 8 to
2. 0 miles in a down-estuary direction. Also requested
was an analysis of the effects of the changes in bottom
elevation on oysters living within the leased area (fig. 5).
Because a suit over damage to the grounds could have
been involved, three considerations seemed of impor-
tance:

1) The measurements would have to be pre-
cise. (Fathometer surveys, accurate to
± 0. 2 ft. , at best, would be inadequate).

2) The measurements would need to be nu-
merous enough both temporally and

spatially to permit adequate descrip-
tion of variations in the location and
extent of spoil buildup.

3) The biological and sedimentological inter-
pretations of the data would have to stand
up in court.

In regard to the last consideration, it was necessary
to employ the services of a diver who was also a marine
biologist familiar with oyster ecology.

541

Materials and Methods

Steel reference rods were placed at 80 stations
(fig. 5) around the perimeter of the lease property.
Horizontal control was provided by the U. S. Army,
Corps of Engineers, and was based upon the official
plat of the oyster lease. The Corps survey team placed
markers at the seven corners of the property and at
certain intermediate stations along the lease boundary
lines. A field crew then placed pound poles at 300-foot
intervals around the perimeter of the property, but five
feet to the outside of the surveyed lease boundaries.

The steel reference rods were positioned by means
of a graduated line stretched between successive pairs
of pound poles. Seventy-three reference rods were
equipped with an 18-inch- square metal plate located
several inches from the top of the rod (see inset, fig. 5).
The plates prevented the rods from sinking into the bot-
tom at those few stations where the bottom was very
soft. (As explained below, it was later found that the
plates were unnecessary at most stations. ) Rods were
"planted" in the bottom by removing an 18-inch square
of sediment to a shallow depth. Once the plate was in
the pocket, it was covered again with the sediment that
had been removed originally. Thus, the tops of the rods
projected above the bottom some one to five inches after
planting.

Measurements of changes in bottom level were
taken by a two-man team consisting of a diver and a
helper in a skiff. (A gasoline-powered air compressor
supplied air to a reserve tank from which the diver drew
air via a hose). A steel carpenter's rule with an ad-
justable right-angle arm containing a level bubble was
used to measure the distance from the top of each re-
ference rod to the sediment surface (fig. 6). The
distance was locked by means of a tightening screw, and
the rule was passed to the helper in the skiff who im-
mediately recorded the distance, station number, and
any of the diver's observations.

Nine rods were planted without plates, in order to
obtain an estimate of the vertical stability of such re-
ference rods in a relatively firm bottom. Such rods

542

proved stable, as indicated by later statistical analysis
of the elevational data for adjacent rods with and without
the stabilizer plates. At stations 39 and 57, two refer-
ence rods were emplaced to obtain an estimate of the
variability in sedimentation to either side of a pound-
pole marker stake. Differences noted between rod 57
and 57A (table 1) are more apparent then real because
the plate on rod 57 limited erosion of the bottom during
the period December 4-19, 1965.

The condition of the bottom in the lease area was
determined before and after the dredging and spoil dis-
posal by means of a visual inspection. Those areas of
the bottom visible from the surface were examined from
a skiff either drifting with the current or driven slowly
over the area. Occasional dives were made to determine
bottom texture and consistency. Deeper areas, not
visible from the surface, were examined by the diver
while drifting with the current or fastened to a drifting
skiff. The helper in the skiff kept track of the position
of both the boat and the diver.

In addition to the visual inspections, 650 oysters
obtained commercially from York River beds were
planted at station 45 (fig. 5), six days before the dredg-
ing and spoil pumping commenced. This was done in
order to have a measure of any deleterious effects that
dredge spoil might have on living oysters, should the
spoil material move upstream from the area of spoil
outfall (fig. 5, inset). The oysters were recovered and
examined a few days after final measurements of bottom
level.

Results

The results of the field program are given in table
1. Three sets of measurements of the elevation of the
bottom were made before commencement of dredging
(December 10, 1965), two complete sets and part of a
third were made during the dredging period, and three
sets were made after dredging ceased (January 4, 1966),

A severe winter storm occurred in Virginia during
the latter part of January and early February, 1966,

543

r

(1

c

o

rrj

rrj

>

>

0)

l-

X)

V

(0

in

XI

*j

0

6-5

E-3

2 o
> -

-O C
It
9)

rrl c

co o

6

c

X>

-

(M

m

■*

m

~o

r~

CO

O^ O — < fM

m ^<

in nD r- co on o

(M

(M
fM

rn f
rM fM

m no
cm rvj

01

o

<*> ro

_; 3 *H

0 J3 -
u £ o

■— ' u

01

u
C

in

o

o

o

rri

r-

o

CO

fM

(M

00 f- >o CO

O H ff| fl

fM o

■^ oo co rn ^ on
po in ro — i m o

fM

<*

o ~o

CO on

Tf CM

o
1

o
+

O
l

o
1

o

o

o
1

o

o o o o
1 + + 1

o o
■ ■

o o O — i O -H
■ i i i ■ i

1

O

— o
1 1

O O
1 I

V

o
u a
dj "J
2 '*

cd
>

00
00 S —

.S «f 0
3 U U
X) ^ 01

-a a

^r

o

o

o
o

r-

o

o
o

o

o

fj^ CM ■* 00

o o o o

On O

o o

O O O O O On

o o o o o o

^r

o

(M

vO fM

—i o

in o

-H o

o

o

o
i

o
1

o

o
1

o

o
1

o o o o
+ 111

o o

+

o o o o o o

+

o

+

O
+

o o

+ 1

o o

+

o
en
0

u
m

£

0

4-)

■*->

0
in

e £

oi rS

s .a

01 T>

h c

» dj

s t
II

01 00

a c

Hi a)
o
S e"

m O

c u

™ cj

a) o)
> £

o

■o
c

01

s
>

>

0

a
<

(0

Xj

Ul

01

o
c

rM

m

oo
c

no
-0
01

u

T>
en

o

Dh

in
o

nj xi
2^>

rM

o

fM

O
I

sO CJ

O C

o
1

XI

rM

o

n) -Xi
2^>

~

u

c

J3

in

o
+

o

o
+

o

o

1

XI

o
1
XI

<7»

o
1

0)

u
>.

Xl

a

V
XI

■v
o

CO

o

fM — i 00 U

o o o c

(J o

c o
o

fM — i U <J <J V

o o c c c c

o o
+ 1

XI XI XI XI

~v
o

o
1

o
+

o o o
1 1 +

XI

r—
N

X>
01 XI

O

O

c

XI

o

o
+

CO

o
o

t-

o

o
+

o

in
o

o
+

CO

o

^ in r^ in
O P0 <*") o

"tf 00

o m

— 1 t~ o u o u
o o c c c c

NO

o
+

CM — i
fM O

o o

1 1

XI

o
+

o o o o
1 + + 1

o o

i +

o o
+ +

X) XI XI XI

fM

xi

0) nD

In •*>

rj>

in

1-1 "O

00

o

o
i

u

c

42

o

o

CO

O

o

o

in

o

^i * O N

o o ^ o

CO u

o c

U (J U (J ON

c c c c o

in

o

o

rM

•XI ■*

O O

o
1
XI

o
1
XI

o
1

x>

o
1

Xi

o
1
XI

o
t

o o o o
+ + 1 +

o
1

XI

o

XI XI XI XI XI

o
i

O
1

X)

o o

1 +

XI

O O
I +

XI

0

11

D.

oo
c

(Nl
O

S^5

r>

o

o
i

U

q

XI

rj

o

o
+

fM

o

o
+

o

o

+

o

o
-t-

u

c

r-

IM T — i CT-
(M — I O fM

ON CJ

O C

U U IJ U U fJN

c c c a a o

o

+

XI XI XI XI XI

o

fM

o

+

o

(M

O
+

— i o

o o
+ +

—i cj
— C

o

o

o o o o

+ 1 + +

o
+

XI

ON

Q ^

^o

m

o

o

+

u

c

XI

o^

o

<N1

o

00

o

in

o

a-

o

m fM in r>

— * -4 o ro

u u

a c

XI XI

(J u u u u o

c c c c c c

XI XI XI XI Xl XI

m

o
XI

o

c

XI

CJ O

c o
o

XI

O £3

O

+
X!

o
1

o

o
+

o
XI

o

o
1

o o o o
i + i i

T3

01

Q

-

in

o

c

o

o
+

u

c

XI

u

c

XI

O

o

1

00

c

00

■V

01

-o

1

01

u

o

o

oi in
Q ^>

<*

0

c

XI

c

XI

o

c

o
o

CO

o

1

Xi

o

c

u

C

o

<M (M r^ »h

fM o O — i

(J Tt<

C O

o

XI Xl

U U U U U -h

C C C C C o

o

1

XI XI XI XI XI XI

CO

o

o

u

c

XI

CJ CO
C CM

O
XI

o o

o o
i i

XI

o
+

o o o o
i i + i

o
cm

>

0 in

2 -o

^o

o

P0

in

o

o

CO

o

CO

m

o

r- r~ fM *o

O O O -i

on m

fM r*">

r^ o co m ip co

ro O rn i-H in o

fM

O

o

in co
CM rM

o o

1 1

o

o
1

Xi

o
f

o
1

o
+

o

O

o
+

o o o o
+ 111

o o

1 1

XI

O — i O -< O —i
1 t 1 1 1 1

XI XI XI XI XI

""*

o
1
XI

1

Xi

00

C oi
i-1 h

in

£ 5. >

<* 0 o

,g -< <m

IT)
vD

fM

00

in

IN)

rM

-o

o
m

00

in

fM

vO in r* n
m rM o o

fM a-
o — •

in o no -h o -^
in no in o o ~*

ON

fM ^D

fM ro

00 ^o

ro ro

m

m

m

m

m

ro

en

ro

PO ■* ■* rfl

m ^

fO N ^ Ul Tt* fO

m

■*

^< m

m ^c

to o

w *3

6

c

-

-

fM

ro

■*

in

•XI

r^

CO

ON o - fM

m ^

in vO r^ oo on o

fM

fM
fM

CM (NJ

m vO
cm rM

544

^

<-

<f

x>

r-

00

O

O

rg

co

-*

in

n£>

r-

00

0s

o

rg

co

■*

m

g3

r-

aj

0^

o

rg

P0

•*•

HI

■u

r-

00

CT-

o

~

rg

rg

oj

CO

co

CO

co

co

co

CO

m

CO

ro

rO

■*

■*

■*

<^

1"

*

^

■*

■*

•*

ui

m

in

in

in

ai

m

in

in

m

m

X)

vO

on

X>

rg

v0

nO

co

o

in

r-

CO

CO

in

CO

Cf

CO

in

CO

vO

rg

I—

(T

0>

in

in

00

CO

o~-

X)

m

m

■3-

en

no

— i

•*

CO

rg

o

rg

in

in

o

O

in

Tf

•*

in

■*

x>

in

^r

o

^

•^

rg

in

o

r-

in

in

rg

rn

CO

CO

rg

rg

rg

*

rg

o
1

+

o
i

O
1

o
1

o

1

o

o

o
i

■

1

o
t

o

o
i

o
i

o

o

o

o
i

o
+

o

+

o
+

o
+

O
1

o
+

o
i

l

o

o
+

o
+

o
+

o
+

o
i

o
1

o
i

o

1

o
+

^f

o

O

o

O

o

in

o

en

r-

a

o

r-

o

o

o

o

vO

o

r-

r>

CO

o

1"

CO

in

■*

o

X)

o

o

o

rg

0>

in

o

— <

o

o

o

~

o

o

rg

o

o

o

o

o

o

o

o

CO

o

— i

rg

—*

—

vD

o

^r

o

rg

~i

CO

o

— '

<*

— '

rg

o

o

o

+

o

o

o

o
+

o

o

o
i

o

o

o
+

o

o

o
i

o

o
i

O
1

o

o
+

o
+

o
1

o
+

O
1

o
1

o

o
+

o

+

o
1

o
1

o
+

o

o
1

o
i

o

o
i

on

u

rvl

u

o

u

u

u

u

o

u

o

00

rg

-o

*

■*

u

-r

^

o-

00

_

rg

rt

^

rg

c

o

C

c

c

C

c

c

c

c

c

— 1

o

—i

^

o

c

**

~-*

co

o

o

CO

o

l_)

CO

n

o

o

o

o

o

o

o

o

o

o

o

o

o

o

+

+

+

+

i

+

+

+

i

1

1

1

1

+

+

XI

XI

XI

XI

XI

Xl

XI

XI

XI

XI

XI

O

CO

rg

o-*

o

c

(M

<-J

rg

— '

c

— ,

o

O

o

XI

+

1

+

1

o

<*

rg

u

r-

u

o

in

■*

_,

in

t-

in

in

r-

CO

^

u

~~

c

o

o

a

o

c

c

— <

o

o

—<

o

^

-*

^~

o

o

o

c

o

o

o

o

o

o

o

o

o

o

V

o

o

o

XI

+

+

XI

+

XI

+

+

1

i

1

+

+

u

>v

X3

1

+

+

u

-H

u

o

o

o

u

u

tj

o

u

o

o

o

u

u

o

e

o

r.

a

c

c

C

el

c

1-

c

a

D

c

a

a

c

o

c

X>

Xi

XI

XI

XI

XI

XI

XI

X)

XI

XI

Xi

XI

XI

XI

XI

Xi

■a

0
u

rg

o

o

u

(1

0

u

(J

r-

O

<;

^

u

u

r-

u

^o

u

o

rg

u

in

CO

CO

vO

rg

■gs

a*

r~

„

c

c

c

c

c

o

c

c

o

c

c

o

c

—

c

c

o

d

"i

o

o

—

-?

^

rg

co

o

o

rn

CO

o

o

o

o

o

o

o

_

o

o

-H

o

o

o

o

o

o

XI

XI

XI

XI

XI

XI

Xi

XI

1

XI

XI

XI

1

XI

Xi

+

XI

i

XI

i

1

XI

1

'

+

+

+

+

+

+

+

o

c

„,

u

0

u

r-

rg

c

c

c

o

00

o

o

X)

+

XI

XI

XI

'

U

rr-

u

U

0

in

(J

()

^H

o

o

r-

o

u

^

y

u

rvl

o

m

NO

o

in

o

o

CO

IT-

in

M

r-

_

o

rt

f

X)

o

CO

c

co

c

c

c

— 1

c

c

o

c

c

o

c

c

o

c

c

in

-o

— '

—

c

o

in

o

o

o

— I

o

U)

o

—

o

o

— *

c

o

_

o

o

O

o

o

o

o

o

o

o

o

,-i

o

o

o

O

o

o

o

o

o

o

X>

+

XI

XI

XI

+

XI

XI

1

XI

X)

XI

+

XI

XI

+

XI

XI

X)

+

+

XI

1

■

X)

1

1

X)

+

+

1

+

+

+

+

1

1

u

m

O

u

u

o

(J

fn

43

u

u

u

u

o

IN]

u

o

vO

m

rg

CO

ro

f

_,

m

Tf

"tf

0s

0-

r-

r-

u

CO

*

XJ

en

rg

c

rn

c

c

c

c

c

o

rg

c

c

c

c

C

o

e

c

-^

o

O

— '

■^

rg

—

o

rg

—

o

o

rg

o

c

rg

*-"

o

o

o

X>

O

o

o

o

o

o

o

o

o

o

o

o

o

o

o

o

o

o

o

o

o

o

o

XI

+

x>

XI

XI

XI

1

X)

1

XI

XI

XI

XI

XI

XI

XI

XI

+

+

+

+

1

XI

+

i

+

+

+

+

+

+

1

XI

1

1

1

1

<J

o

o

u

o

00

c

c

c

c

c

—■

in

o

XI

XI

XI

XI

Xi

on

■r

u

*f

r-

rg

o

o

u

o

^r

o

^

0>

in

o

r-

in

rg

o

<Jv

rg

CO

nO

rg

in

^

_

_

r-

m

_

CO

c

o

c

c

c

o

c

<tf

o

rg

c

c

c

c

o

c

■*

— '

— '

c

o

— '

o

— »

o

rg

o

■-1

(->

o

o

o

o

o

o

o

— '

■*

o

o

n

o

o

o

o

o

o

o

o

o

o

o

o

o

o

o

o

o

o

o

o

o

o

o

+

XI

XI

X>

i

XI

1

1

1

XI

XI

XI

Xl

XI

+

XI

1

1

i

X!

+

+

1

+

+

+

1

+

i

1

XI

i

+

+

1

+

en

rsl

M

43

g?

r-

00

rg

CO

rr>

m

CO

rO

in

o

r>

rg

r-

r-

CO

rg

■*

^

00

CO

^

^

oo

i-

Tf

in

CO

in

■*

o

x>

o

rO

rg

■*

CO

rd

o

■^

o

rg

■*

x>

m

"t

rr

■o

^r

CM

o

rj

CO

o

o

o

o

o

rg

rg

CO

~*

o

o

rg

— '

1-1

rg

Tt<

~*

co

O

O

n

n

o

r>

o

o

o

o

o

o

o

o

o

o

o

o

o

o

o

o

o

o

o

o

o

o

o

o

o

o

o

o

o

o

o

1

+

X)

Xi

Xi

i

1

X)

+

1

i

1

XI

i

XI

X)

i

XI

1

i

XI

1

+

+

I

X)

+

1

+

1

1

+

1

1

1

1

1

+

1

1

.

on

-0

CO

o

■*

o

-0

rg

CO

r-

rn

x>

t>

^0

in

in

GO

vO

in

r-

vO

a-

<*

o

o

rg

-*

o

m

■^,

00

•tf

rn

rg

rg

— i

rg

rO

rt-

x>

*

co

\f

o

rg

■*

rO

rg

*

m

o

rg

in

— i

^r

o

o

m

•*

■*

o

rg

— '

m

o

rn

m

rg

—■

^r

*

■*

■*

■*

■*

-*

rg

<*

tT

■*

ro

■*

^

<*

■*

•*<

>*

■*

in

rO

rg

rn

ro

CO

■*

m

m

to

■*■

T

"!>

rg

^f

rg

m

<

<

■ '

r-

on

0*

O

H

rg

co

■*

in

X>

r-

oo

CT-

a-

o

_-t

rg

rn

■*

in

sD

r-

00

o-

o

— i

rg

CO

^f

in

^o

r-

r-

00

en

o

— i

—

rg

(M

rg

m

CO

rn

co

rO

CO

CO

ro

CO

r^

CO

<}>

•*

Tf

<?<

■*

■*

<*

■*

■*

■*

in

m

in

in

m

in

in

m

m

m

m

~L>

gj

545

m

•*

m

in

r-

TT

a-

in

(*-

co

CO

■3-

■o

■*

CO

o

o

IM

M

o

CM

— 1

r\l

— •

— '

o

fi

fM

o

— '

o

rn

INJ

o

m

o

•

o
■

O
1

o
+

o

o
+

O

o
1

o
+

o
i

o
+

o
+

o
+

o

o
+

o

o

o

o
+

r~ r- r- ao

— » rs) ^ cm
o o o o

ooooooooooooooooooo

ooooooooooooooooo

ooooooooooooooooo
+ I++II+ + +II+I+I +

U Tj» Tf Ifl ^

C N NO o

ooooooo — o

o o o o o o

iA >0 r- f* ^
o o o o o

oooooooooo

oooooooo

-O ^ — I

o o
o o o

oooooooo

o o o o o o

ooooooooooo

o o o o o o o

■^

<^t

■*

CO

I-

m

0^

CO

t

r\)

o

1—

PO

o

■~o

rr

o

o

o

o

o

o

o

o

o

—

o

— 1

— '

o

o —

r\l

— '

— <

INJ

m

o

o

o

o

o

o

o

o

o

o

o

o

o

o o

o

o

o

O

i

1

1

1

+

+

1

+

+

t

+

+

1

1 +

'

'

1

+

_- -Xi nO -i> x> sD o -o 'O r* r— r*- t*-

m so

^ 4J

in _

546

which resulted in the formation of extensive ice floes in
the York River. These ice floes removed several of the
pound poles along the northern boundary of the study
area. Reference rods adjacent to these poles could not
be located after the storm. In addition, the action of the
ice disturbed a few of the shallow reference rods. As
a result, only one or two "after dredging!' measurements
were made at some of the reference rods. These are
indicated in table 1.

The visual inspection of the study area made before
the commencement of dredging revealed that the shal-
lower areas were covered with extensive beds of eel
grass (Zostera marina) and associated epiphytic algae.
No oyster reefs were found in the shallow areas.

The deeper areas were generally free of vegetation.
The bottom materials consisted of silt and the bottom
was soft. There was a one-half to three-inch layer of
extremely fine, loose sediment overlying either soft,
slightly firmer sediment or accumulations of dead shell.
There was a fair-sized population of the hard clam
(Venus mercenaria) in the top few inches of sediment.
These clams averaged two to three inches in width.

Oyster reefs were found only in the northwest corner
of the lease area and at scattered intervals along the
line between stations 21 and 25 (fig. 5). These reefs
were of very limited size and consisted of both living
oysters and dead shell. None of the material on the
reefs was handled by the diver.

The appearance of the bottom at the time of the
final swim over (12-13 March 1966) was identical to that
at the time of the initial swimover. This was expectable,
however, in view of the fact that dredging downstream
was unexpectedly cut short and only a very small amount
of spoil entered the York River at the outfall. None of
the spoil-polluted estuarine water was observed moving
as far upstream as the oyster lease.

A total of 582 of the 650 oysters planted in the vi-
cinity of station 45 were recovered on 12-13 March 1966.
(The remaining 68 oysters could not be located because
of limited visibility). Of these, 26 (4.4 percent) were

547

boxes (no oyster meat within shell; shell sometimes
filled with sediment), six (1. 0 percent) were gapers
(animals within the shell were too weak to keep the shell
closed in air), 12 (2. 1 percent) were strong enough to
keep the shell tightly closed but did not appear healthy
enough to eat. The 538 remaining oysters (92. 5 per-
cent) appeared to be in excellent condition and were
eaten by several persons after being prepared in a
variety of ways. A summary of the oyster data is
presented in table 2. Although no quantitive estimate

Table 2. Summary of data obtained from oysters
maintained at the periphery of the study
area (in the vicinity of station 45) during
the period 4 December 1965 to 13 March
1966.

Numbe r

Percent

total

Percent

planted

total

recovered

Oysters planted

650

100

Oysters recovered:

a. Healthy oysters

538

82.9

92. 5

b. Gapers

6

0.9

1. 0

c. Other*

12

1.8

2. 1

d. Total live

556

85. 6

95. 6

e. Boxes

26

4. 0

4.4

f. Total recovered

582

89. 6

100

Unrecovered: 68 10.4

* Animals strong enough to close shell in air, but did
not appear healthy enough to eat.

was made, many of the oysters recovered had healthy
spat attached to them, measuring about one inch across,
Spat this size had not been noticed during the planting
operations.

Information such as the above, for oysters planted

to determine the effects of spoil deposition or lack

of it, is valuable for the determination of damages and

548

awards in courts. This is true if the information re-
presents the expert evaluation of a marine biologist
knowledgeable in oyster ecology.

Discussion

A study of the net changes in the bottom elevations
(table 1) during the entire period of observation reveals
that the perimeter of the oyster lease was undergoing
erosion almost everywhere between stations 1 and 44
(fig. 5). The following stations along the perimeter
nearest the dredge spoil outfall (fig. 5) showed net
accretion during the period 27 November 1965 through
12 March 1966: stations 45, 46, 47, 48, 50, 54, 55, 56,
57, 61, 65, 67, 70, 72, and 73. Of these stations, the
following ones exhibited no change or net erosion during
the period of dredging: 47, 50, 55, 56, 61, 70, 72, and
73. This indicates that the slight amount of deposition
observed over the entire period at stations along the
southern perimeter was not directly related to deposi-
tion of dredge spoil, but to natural sedimentation in the
area.

Review of the net changes in bottom level during the
period of dredging (table 1) reveals that at 59 measure-
ment points there was no net change or that the bottom
was being eroded. At 23 stations the bottom was build-
ing up. The amounts of buildup and erosion were in
almost all cases negligible (less than one -half inch).
At only one station, number 28*, was the net or the
weekly deposition greater than one-half inch. It is only
logical to presume that similar conditions of bottom
erosion, bottom stability, and very slight and sporadic

When the stake and plate at station 28 were removed,
it was found that amphipod tubes extended from the sur-
face of the sediment to the plate. These tubes were very
closely packed together. During the swimover on the
following weekend, it was noticed that patches of the
bottom containing amphipod tubes were raised slightly,
relative to adjacent bottom that did not have these tubes.
The parts of these tubes that extend above the surface
probably act as a sediment trap and, as the sediment
accumulates, the animal builds its tube higher to keep
pace with the rising sediment surface.

549

bottom deposition existed over the area of the entire
oyster lease during the period of dredging and spoil
outfalL The changes in bottom level observed during
the period of dredging were in all probability natural
changes, the effect of dredge-spoil outfall, if any, being
too minute to be discerned.

Conclusions

The overall trend in sedimentation along the peri-
meter of the oyster lease before, during, and after the
dredging was one of slight erosion. There was only
very slight deposition over limited areas within the
leased property during any of these three periods and
such deposition was confined to MLW depths of less
than three feet. The changes in bottom level would in
no way have been detrimental to live oysters on the
bottom within the lease area. Mortality within a group
of planted oysters held just outside the study area, on
the perimeter closest to the dredging operation, was
very low. This mortality was probably entirely natural
and not due to the dredging operations. Thus, it was
concluded that dredging and spoil disposal had no ob-
servable effect on the character of the river bottom or
the natural animal population within the survey area.

BUCKET DREDGING IN PROXIMITY TO AN ESTUARINE

OYSTER GROUND

The author and a marine-biologist diver conducted
a monitoring program similar to that described above
in Hampton Roads, Virginia (area D of figure 1). The
concern in this study was that the activity of a bucket
dredge, operating in the entrance to Hampton Creek,
would produce a deleterious buildup of silt on an adja-
cent oyster ground (fig. 7). It was important to know
the extent of damage to the natural "rock" (shell sub-
strate on which oyster larvae "set"), and to the living
oysters present, as a result of silt deposition from the
disturbed bottom in the dredged channel.

As in the York River study, it was concluded that
the oyster ground was not damaged in any way by the
dredging. Data on the changes in bottom elevation due

550

to natural causes will be released to interested research-
ers upon request.

SUMMARY

The studies reviewed show that if the design and ex-
ecution of a monitoring program are carefully done, it
is possible to make precise statements about the effects
of dredging or spoil deposition on changes in bottom
elevation or on the populations of bottom -dwelling organ-
isms in the affected areas. Where litigation over
damages to the bottom organisms or the substrate upon
which they live may be involved, it is desirable to have
monitoring programs designed and executed by the same
marine scientists who will evaluate and interpret the
data. Reviews of other studies, such as the two pre-
sented here, should be published in order to aid the
advancement of environmental studies of pollution by
dredging and related spoil disposal.

ACKNOWLEDGMENT S

Mr. M. P. Lynch, Virginia Institute of Marine Science,
acted as the marine-biologist diver for the three studies
and Dr. M. L. Wass, also of VIMS^provided the faunal
lists used in the reference cited below.

REFERENCE

HARRISON, W. ; LYNCH, M. P. ; ALTSCHAEFFL, A. G. ,
1964, Sediments of lower Chesapeake Bay,
with emphasis on mass properties: Jour.
Sedimentary Petrology, v. 34, p. 727-755.

551

Figure List

Map showing sites in lower Chesapeake Bay where
environmental effects of dredging and spoil disposal
were studied.

2. Map showing dredged area on Rappahannock Shoal
and site of spoil dumping from hopper dredge.

3. Void-ratio reduction with depth in five piston cores
from the area of spoil dumping. Curves 1-4 show
anomalous profiles. Curve 5 shows an expectable
curve for bottom sediments that have not been sub-
jected to a load of spoil. ( Data from Harrison, et al.,
1964, fig. 8, cores 17a, 17b, 20a, 20b, and 22. j~~

4. Variation in numbers of animals and numbers of

species present in grab samples taken in the dredge
area ( one station) and in the spoil disposal area (two
stations) before and after dredging.

5. Layout of stations for determination of changes in
sediment level along perimeter of an oyster lease,
lower York River estuary.

6. Method of making underwater measurement of depth
to water-sediment interface, using top of steel rod
as reference level.

7. Location of a study of the effects of dredge-induced
siltation on an oyster ground near an area of bucket
dredging.

552

553

JL?

c - „

-OL

— O
K

<

O
a.
<A
Q <

UJ

, C£.

-*&.

554

Fig. 3,

2 3 4

VOID RATIO (SATURATED)

555

lOOOr

500-

DREDGED AREA

Station

SPOIL DISPOSAL AREA

a.

<

E
m

s

Z

a.
t/>

to

_j

<

Z
<

CO

r>
z

100

so-

io

9
8

7
6

5

4

Fig. 4.

6 MONTHS
BEFORE

1 MONTH
AFTER

6 MONTHS
AFTER

15 MONTHS
AFTER

TIME RELATIVE TO DREDGING AND SPOIL DISPOSAL

556

HHWIWWP

1

lie

\ m

w

%r

la

S2

1

u

z
Hi

fa

■o

i

s

557

558

559

18

Reprinted from MARINE GEOTECHNiaUE, University of Illinois Press, Urbana

PLATE-LOAD TESTS ON SANDY MARINE SEDIMENTS, LOWER CHESAPEAKE BAY *

W. HARRISON** AND A. M. RICHARDSON, JR.***

ABSTRACT

Two plate-bearing tests (similar to ASTM D 1196-57) were made 1.8 nautical
miles west of Little Creek Harbor entrance in 16 to 20 ft (5 to 6 m) of water.
A Ul-ton (3T»000 kg) load frame was constructed to ensure adequate reaction for
each jack load. The load frame was emplaced by the USS Salvager , a U.S. Navy
rescue-salvage vessel, and divers from the ship conducted the tests according
to telephoned instructions from the deck.

Bearing capacity factors (N ) for use in Terzaghi's bearing capacity equa-
tion were estimated at 62 for site A and 195 for site B. These factors are
somewhat higher than would have been estimated on the basis of triaxial tests on
the sand. Average coefficients for subgrade reactions of 50 tons per cu ft and
110 tons per cu ft were obtained at site A and site B, respectively. These
values compare favorably with values proposed for medium to dense submerged sands

It is concluded that future research should be directed toward the develop-
ment of a standard penetration sampling test for use in the marine environment.
This test should be calibrated against further in-place load tests.

INTRODUCTION

Many marine engineering projects
that will be undertaken in the immedi-
ate future will involve placement of
structures on surficial sediments of
the continental shelves. Although the
configurations and functions of the
structures can be expected to vary
greatly, their placement for the most
part will be limited to shelf areas
having depths less than 180 m
and covered by sandy marine sediments.
Consequently, the design and fabrica-
tion of suitable submarine structures ,
and the ultimate success of many proj-
ects, will depend largely on accurate

Contribution number 5 of the Land and

Sea Interaction Laboratory.

**Land and Sea Interaction Laboratory,

Institute for Oceanography, ESSA,

Norfolk, Virginia.

***Department of Civil Engineering,

University of Pittsburgh, Pittsburgh,

Pennsylvania.

foreknowledge of the load-carrying
capacity of sandy marine sediments.

A better understanding of the
ultimate bearing capacity and settlement
behavior of submarine sands is required
to proportion properly the base-plate
foundation elements of submarine struc-
tures, to predict the relationships of
structural load and rate of settling,
and to make allowance for dynamic re-
sponse behavior (such as natural
frequency) of completed structures.
This knowledge can be gained only by
in-place investigations of marine sedi-
ments. These investigations are limited
further by the need to conduct them
easily and reliably from shipboard in
water depths encountered over the conti-
nental shelves. This paper provides
background information about foundation
investigations in sandy marine sediments
and presents the results of preliminary
research directed toward the design of
a simple method for submarine foundation-
site investigation.

274

Plate-Load Tests on Sandy Marine Sediments, Lower Chesapeake Bay 275

LOAD-CARRYING BEHAVIOR OF SANDS

For convenience, the determination
of the load-carrying behavior of sur-
face foundations on sands is usually
considered as two separate problems:
(l) the determination of the load at
which ultimate failure will occur
and (2) the determination of the
settlement of the foundation under any
load less than failure.

Both aspects of foundation behavior
are influenced by the structural design
of the base and the properties of the
soil. The load-carrying behavior of
bases on sand is a function of the base
size, the base geometry, the distribu-
tion of the load on the base, and the
flexibility of the base. It is also
a function of the angle of internal
friction of the sand, $ , the relative
density of the sand, Dr , and whether
the sand is dry, moist, or submerged.
The friction angle of the sand is
influenced by the void ratio, grain
size distribution, grain geometry, and
mineral composition of the sand; and to
a lesser degree it is influenced by
the wet or dry condition of the sand
and rate of load application. In
addition, the method of testing
(whether triaxial, direct shear, or
plane strain) has been shown to have
a marked effect upon the friction angle.

DETERMINATION OF ULTIMATE CARRYING
CAPACITY

Terzaghi (19^3) presents a solution
for the ultimate carrying capacity of
foundations on sand. This solution is
based on limiting equilibrium along
two failure surfaces approximating
logarithmic spirals, and yields, for
uniformly loaded circular areas,

where

Lult

Lult

= 0.6yRN
Y

= ultimate bearing capacity

= total unit weight of sand
above water or the
buoyant unit weight of
submerged sand

R = radius of loaded area

N = dimensionles s bearing
Y

capacity factor.

Terzaghi (19^3, p. 125) has plotted
N versus the friction angle of the
sand for two extreme conditions, general

shear and local shear. For local shear
it is assumed that a failure of the load
area occurs before full mobilization of
the shear strength on the failure sur-
faces has occurred. Sowers (1962) has
suggested that the general shear analy-
sis be used for relative densities over
TO percent and the local shear analysis
for relative densities less than 20
percent. Interpolation between these
extremes on the basis of relative den-
sity is suggested.

Values of <\> can be determined by
laboratory shear tests at field void
ratios and eonfining pressures. This
requires a knowledge of field void
ratios. Special and costly techniques
are necessary for obtaining undisturbed
samples of sands. More commonly, both
<(> and the relative density are estimated
from the results of the Standard
Penetration Test, in which a lUO-lb
weight is dropped 30 in. onto a 2-in.
(outside diameter) split-spoon sampler.
A plot of Ny versus the results of
standard penetration tests is given in
Peck and others (1953, p. 222). This
plot is commonly used in designing
foundations for sands on land.

DETERMINATION OF LOAD-SETTLEMENT
BEHAVIOR

The load versus settlement behavior
of loaded areas on sands requires
considerable study. Methods for deter-
mining or estimating this behavior have
been suggested as follows:

(1) results of the Standard Pene-
tration Test ;

(2) results of field loading tests;

(3) results of consolidation tests
(one-dimensional compression tests) and
elastic stress analyses; and

(k) triaxial test results and an
elastic stress analysis akin to the
method for estimating settlements of
foundations on clays.

Method 1 is the easiest to apply.
Method 2 must rely upon an incomplete
knowledge of the relationship between
load settlement behavior and the dimen-
sions of the loaded area. Method 3 can
be used to assess the effects of sub-
mergence on the sands but neglects
settlement resulting from lateral strain
of the soil element. Method h, although
it requires considerable detailed testing,
accounts for settling caused by lateral
strains and can be used for either sub-
merged or dry sands. It is deficient in
that it places too great a reliance
upon an elastic stress analysis.

276

Marine Geotechnique

The common procedure for land-
based foundations on sands is to use
the results of the Standard Penetration
Test and a graph prepared by Terzaghi
and Peck (19^8), reproduced in Figure
1. The suggestion is made by Terzaghi
and Peck that the pressure for any
given settlement be halved to consider
submergence. Sowers and others (1966)
questioned the validity of this
reduction .

LOAD-CARRYING BEHAVIOR OF MARINE SANDS

From the above discussion it seems
apparent that a reliable and reasonably
convenient method for determining the
load-carrying behavior of ocean bottom
sands does not exist. Among the prob-
lems encountered are the following.
First, it is difficult to obtain un-
disturbed samples of submarine sands.
Second, the Standard Penetration Test
might be performed in shallow water,
but its use in moderately deep water
would be almost impossible. Finally,
underwater load tests are generally
too cumbersome for practical applica-
tion. In an attempt to solve some of
these problems, the authors designed
load tests similar to those used in
subaerial environments and conducted
in-place tests on sandy marine sedi-
ments in the lower Chesapeake Bay.
Differences were studied in the loading
characteristics in submarine and sub-
aerial environments and standard load
test data were made available for two
sites. The standard data can be used
to compare the results of future in-
place load tests in the same area and
to calibrate the instruments used in
such test s .

LOAD TESTS

LOAD FRAME

A Ul-ton (37,195 kg) load frame
(Figure 2) was constructed using steel
and concrete. Six U.S. Coast Guard
concrete buoy sinkers, each weighing
^,737 kg (12,700 lbs) provided most of
the mass of the frame. The design was
dictated in part by the necessity for
a crane transfer from the dock where
the frame was fabricated to a pick-up
point (in air) where the frame could
be secured (Figure 2) to the open
shackles of the cables of the USS
Salvager . These cables were 5-^ m
apart, center to center.

EQUIPMENT

Three 2 . 5-cm- (l.0-in.-) thick
steel plates, having diameters of 30.0,
1+7.5, and 60.0 cm (12, 19, and 2U in.)
were used. Small handles (Figure 3)
were welded near the perimeters of each
plate to aid in handling by divers. A
calibrated hydraulic jack having a
maximum capacity of 10 tons and an
excursion of 20 cm was used. The datum
beams for the dial gauges were 255 cm
long and stood on legs 20 cm high.
Standard dial gauges with 3.5-cm-diam-
eter faces were mounted in plexiglass
chambers (Figure 3) filled with mineral
oil. Thin rubber sheets protected the
dial pin arms. A schematic diagram
and photograph of the test setup are
shown in Figure 3.

TEST SITES

The two submarine sites of Figure
h were chosen for the plate tests be-
cause of the uniformly smooth sandy
bottom over short distances in the area,
and because the waters of this part of
Chesapeake Bay have relatively low con-
centrations of suspended solids. Phys-
ical properties of the sediments at
test sites A and B are given in Table 1,
together with water temperature, salin-
ity , and mean water depth. The tidal
range in the area is 0.76 m.

Sediments at depth at site A were
sampled with a Chesapeake Bay Insti-
tute piston corer (Silverman and Whaley,
1952); as modified by Harrison (Harrison,
Lynch, and Altschaeffl, 196U , p. 730).
The results of size analyses and sedi-
ment profiles are shown in Figure 5.
Failure to recover sediment after full
penetration with the 2.0-m core barrel
at site B makes us certain that sand
is present to a depth of at least 2.0
meters below the bottom. A special
piston-core sample of the upper 20 cm
was taken to determine the void ratio.

The mechanical properties of
cohesionless sediments depend almost
entirely on their relative density, Dr.
For this reason the description of
such a sediment is inadequate unless it
contains reliable information concerning
its relative density. The D values of
Table 1 required the determination of
the in-place void ratio. Void ratios
were approximated by measurements made
on cores taken by the modified Chesapeake
Bay Institute piston corer and retained
in plastic core liners for transport to
the laboratory. Soil profiles (Figure

Plate-Load Tests on Sandy Marine Sediments, Lower Chesapeake Bay 277

>>

-p ^

•H o

CO

CO

C o

•

•

■H -^

ro

m

Ho

CM

CM

cd -^

CO

u

CO

p

ft'-'

6 o

ro

J

Q

cd

0) o

LTN

IT\

Z

S

H — '

CM

CM

=t

<

*

t/j

a

_p ^-~

_*

j-

UJ

cd

ft E

•

•

t-

0)

<D ^_-

NO

vo

i — i

S

Tj

CO

Q

z
\-

Ul

s:
■ — i

Q
UJ
1/1

o

D.
O

00

I'-

ll CM

> >>

•H -p

.p -H

cd to

H C

<D <U

d O

3 "^ -P CU

-P P Cd — -

n) ? Jh
S

0)

•H

CD

6

hO

a;

cd

O

a

S

H

•H

O

hfl

(L)

c

bO

cu cd

a

3 P

•H

H U

-P

cd 0)

m

> >

o

—

CO

Cm

c

•H <D

a!

cd tsi

e

a;

U -H

e

2

bO to

0) ,C

0

H P

e

>

o

ft ft

o

o

-P

S <u

H

-P

cd th

Pi

(1)

o

CO --'

•H

P

p

CM
ON

CM
CO

t—

ON

o

CO

t3 tiJ tz! t3

C C C C

cd cd cd cd

CO co co to

T3

CU

>>

-P

H

m

0) ^~-

*— N

CM

o

[—

-p "O

MO <d

-d"

10

CO

cd a;

h -p

MO >> CD
• H -P

O

H

o

0) fc

H f< h

H

t3 O

O O

0)

O co

O CO

£

e

ft

l/N

MO

MD

ON

CO

CO

O

-=»■

H

O

O CM

r-i i-i CM ro

< I < I

O LTN

o
on co
< I

o

MD

CO

CM

Ox

ON

t—
o

— 1 X!

-P (0 -p

H cd -H

•H H >

CO ft

>>■— ■

^

>> !» Cd TJ

tJ

■d

11 H r-H C

C

c

>3 p o cd

cd

cd

cd hfl co

CO

CO

H -H O

—

i — »

CJ5 -ri

--" ' -P

O

Ol

-d

oo

>

cu
■P

o

>>

U

u

o

0)

co

>

*-^

t—

CO

CM

J
ro

-=j-

MO

ON

H

-e-

»

V£>

CO

1

•

H

VO

cd

UA

•H

ON

^l

-e-

UJ

-P

+

cd

S

MO
H

Tj

-e-

c

cd

1

.*

hO

^t

C

CO

•|H

-©•

■P

CO

II

CU

Eh

H

U

..

o

u^

Cm

ON
-G-

>>

+

P1

J-

CU

CO

•H

-e-

O

CM

o

+

CO

o

o

l^

H

c

-e-

cd

-*

o

+

•H

MD

M

H

(U

H3-

B

CM

<

+

LC\

-e-

EH

-cr

ll

VO

On

S

-=t

O

*— ^

CM

t—

R

ON

a

H

o

—

. — .

■H

_*

-P

Tt

LTN

cd

U

ON

a

cd

H

bO
■H

5

CO

TH

■d

cu

C

h

p

cd

cd

ft

2

M

0)

EH

H

P

CO

O

CO

<

C

t<

278 Marine Geotechnique

5) at sites A and B give the Dr values the 30.0 cm plate resulted in severe

where determined. disturbance. The divers attempted to

jack to an excessively high initial
load. After completion of the loading

TEST PROCEDURE of the two larger plates, the load

frame was moved 6 m north of its

After the ship had made a three- original position and the 30.0 cm plate

point anchor at each site, the load was loaded.
frame was slowly lowered to the bottom,
and as soon as it was placed the

lowering cables were slackened. This RESULTS
procedure was followed to prevent any

disturbance to the frame from heaving The results of the load tests at

of the ship. sites A and B are plotted in Figures

Two SCUBA divers examined the 6 and 7. The plotted penetration

bottom and the attitude of the frame usually is the average of the readings

at each site. At each test site the taken at the three points on each

bottom was roughly horizontal and small plate. In a few instances one of the

ripple marks were noted. At site A readings given by the diver would be

the concrete base blocks sank k . 0 and obviously in error. This reading was

7.6 cm into the sand. At site B, both discarded and the average of the other

blocks sank 2.5 cm. two was plotted. Tilting of a plate

After the pressurized dial gauges, was not significant until the highest

jack,i load plates, underwater floodlight, load increments.

and the beams for mounting the dial Estimates of resultant ultimate
gauges had been lowered to the SCUBA bearing capacities are tabulated in
divers, a hard-hat diver in telephone Table 2. These estimates are somewhat
communication with the deck was put arbitrary and were obtained by inter-
over. The SCUBA divers positioned the secting s t rai ght -1 i ne-f i t s to the low-
30-cm-diameter plate, jack, reaction pressure and high-pressure portions
beams for the dial gauges, and a shim of each lead-deflection curve (Figures
between the jack and the jacking member 6 and 7).

of the load frame. Dial gauges were Average values of coefficients of

then attached to mounting rods. These subgrade reaction as suggested by

rods were screwed into the load plates Terzaghi (1955) were estimated (Table

near their perimeters and spaced 120 de- 3) assuming linear load versus settle-

grees apart. As the plate was jacked ment behavior to 1/2 ton per sq ft

into the sediment the gauge-pin arms (1000 lbs per sq ft) at site A and 1 ton

bore upon the rigid reaction beams whose per sq ft (2000 lbs per sq ft) at site B.

feet extended well beyond the area of Average values of this coefficient were

disturbance. Each gauge was numbered. 50 tons per cu ft at site A and 110

After the hard-hat diver had jacked the tons per cu ft at site B.
plate to a given load, and upon instruc-
tion from the deck, he would read one

gauge and signal the two SCUBA divers DISCUSSION
to write the dial readings of the two
other gauges on their slates. The three

gauge values were then read to the ULTIMATE BEARING CAPACITIES
scientist on deck by the hard-hat diver.

Additional shims were added during the Values of ultimate bearing capacity

tests as necessary, but this in no way were calculated according to Terzaghi

influenced the test results. and Peck (19U8, p. 172) and appear in

The small and medium load plates Table 2. These values were calculated

were always positioned beneath jacking assuming 6l lbs per cu ft as the buoyant

members at either end of the load frame. unit weight of the sand, for <$> values

This left the greatest reaction for the of 35 degrees and Uo degrees and for

large load plate, when it was positioned the local shear and general shear cases,

beneath the central jacking member. No Specimens of the sand from site B were

difficulties were encountered in the tested in drained triaxial compression,

overall procedure. Wave action was The results of these tests appear in

practically nil during the tests and Figure 8. From these tests, <j> at

tidal currents always less than about site B was determined to be 38 degrees.

0.5 m/sec. Thus it can be seen that, at site B,

At site A, the initial setup of ultimate bearing capacities in excess

Plate-Load Tests on Sandy Marine Sediments, Lower Chesapeake Bay 279

TABLE 2.
ULTIMATE BEARING CAPACITIES (LB/FT2)

Plate
Dia.
(ft)

General
Shear

<J> = 35'

Local
Shear

4 = k0° $ = 35'

Estimated from
Load Tests

<t> = U0° Site A

Site B

1.0 815 2330
1.5 1220 3500
2.0 1630 U66O

185 70U IU50
277 1050 2100+
369 1^00 1800

^250
5300
6000 +

Note: calculated values from:

q n , = 0.6 RN ; R = radius of
ult y Y

plate; y = 6l . 5 lb/ft3; N from Terzaghi (19^3, p. 125) solution,

TABLE 3.

COEFFICIENTS OF SUBGRADE REACTION (TON/FT3)
CORRECTED TO 1 -FT PLATE (SITE B). (See correction note below.)

Plate

Dia. (D

(ft)

Site A Site B Proposed Values*

q=l/2 ton/ft2 q=l ton/ft2 Loose Medium Dense

1
1.5

2
Avg.

12 (?)
86
U3
50

Uo

25

80

300

111

25

80

300

177

25

80

300

110

25

80

300

^Terzaghi (1955) for submerged sands

Correction = kn _, = k . x / 2D
1 ft size

Vl + D )

of those predicted by the Terzaghi equa-
tion were obtained.

One explanation for the higher
bearing capacities is that the bearing
capacity formula used here is actually
a shortening of the following:

qult - 1/2T (2R) Vy
As the "shape factor" sY(~0.6 for a

circle )

as suggest

ed by

Terzaghi (19^8)

has been determine

d empirically by com-

par is on

between circular

and strip

f oundat

ions , Ny sh
ane-strain"

ould correspond to

the "pi

frict

ion angle (for

a strip

footing ) ,

which

is probably 2

to 5 de

grees higher than

the triaxial

friction angle (J.

Brine

h Hansen, 1966,

writt en

communi cat

ion ) .

According to

more re

fined calcu

lations of N

280 Marine-Geotechnique

(Hansen, 1961, Figure l), the Ny values of 0.5 ton per sq ft for site A and

of 62 and 195 correspond to $ values of 1 ton per sq ft at site B.
37.5 degrees and U 3 - 5 degrees, respec-
tively. Tbrus , for site B, there is

actually good agreement between a tri- FUTURE STUDIES
axial angle of 38 degrees and a plane
one of U3.5 degrees. As interest in placement of struc-

In Figure 9, the ultimate bearing tures on the continental shelves in-
capacities are plotted against the plate .creases, it will be necessary to develop
diameters. On the basis of a buoyant a reliable in-place method to determine
unit weight of 6l.5 lbs per cu ft, it the load-carrying behavior of shelf
would appear that bearing capacity sediments. Only one approach is pre-
factors N of 62 for site A and 195 for sented in this paper. There are two
site B would be appropriate for use in major limitations to this procedure.
Terzaghi's bearing capacity equation. First, conventional diving techniques

limit the depth at which tests can be
performed. Second, a special type of
vessel and specially trained personnel
must be used .

Disallowing the possibility that
tests up to 200 m or more water depth
The relationship between the settle- might be made practicable by divers

INFLUENCE OF DIAMETER OF LOADED AREA ON
LOAD VS. SETTLEMENT BEHAVIOR

ment an.d the plate diameter for site B is operating out of deep-submerged systems,

shown in Figure 10 for loads of 1000 lbs per ^ f.rgt objection can be overcome by

sq ft, 2000 lbs per s ft, U000 lbs per sq ft , automation and instrumentation of the

and 6000 lbs per sqf.. Because the tests equipment described in this paper. A

at site A were not performed in close hydraulic pump could be incorporated

proximity, a similar analysis of these in thg load f rame ? and be controlled

data was not undertaken. from thg surface. The preSsure in the

Taylor (19U.8) postulated that the jack and the plate deflections could be

settlement under a given pressure would raonitored electrically from the surface,

be more or less independent of the size Th thg diver actlvit would be

of the loaded area, falling off slightly tQ position the Jack and plate, set the

as the size of the loaded area increased. -.,.,„+„„„-;„ j„fi„„+„„«4.„„ ^ „,,,• „„„ „„»

electronic deflect omet er devices, and

Such a relationship is shown m Figure survey the site. The vessel require-

10. Terzaghi and Peck (19H8, p. ^26) ments could be subst ant ially simplified

suggest the relationship by attaching floatation chambers to the

2 load frame and towing it to the test

— *- = (— — — ) sites. It could then be sunk in position

^1 and refloated by pumping out the

floatation chambers,
between settlement and the size of the Even the be st -des igned in-place
loaded area. This relationship is also load tests will be expensive and
plotted in Figure 10. The data from cumbersome. Therefore, the authors feel
site B indicate that there is less that the development of a standard ship-
settlement of the plate, under any load, board technique, such as the Standard
as the diameter of the plate increases. Penetration Test used on land, is most
This behavior pattern is consistent for desirable. The suggested device is a
loads of 2000, U000, and 6000 1bs per sq ft. weighted penetrometer, built around a
At 1000 lbs per sq ft, the settlement was convenient sampling device. This device
almost identical for the 1.0-ft- and could record the depth of penetration,
1 . 5-f t-diameter plates, but was consid- and possibly the resistance to penetra-
erably less for the 2 . 0-ft -diamete r plate. tion versus depth, and obtain a disturbed

Terzaghi (1955) has suggested values sediment sample. The representative

of moduli of subgrade reaction (k) rep- grain-size distribution and degree of

resentative of average values for sub- grain angularity of the sample could be

merged sands. Values proposed by determined. The device should be

Terzaghi are compared with those obtained calibrated in the laboratory first, and

in this series of tests in Table 3. then in the field at sites of plate-load

These values were corrected to a 1.0-ft- tests. Multiple regression analyses of

diameter plate by the correction suggested the data could be undertaken and adequate

by Terzaghi, and appear in Table 3. predictor equations relating ultimate

Average values of 50 tons per cu ft and bearing capacity and load-settlement

110 tons per cu ft for sites A and B, behavior to penetration, grain-size dis-

respectively , were obtained under loads tribution, and grain angularity might be

Plate-Load Tests on Sandy Marine Sediments, Lower Chesapeake Bay 281

obtained. Thus, for certain rapid es-
timates of the bearing capacity of
surficial sandy sediments, all that
would be required would be the data
easily obtained by the penetr omet er-
cor ing devi ce .

CONCLUSIONS

The most useful assessment of the Folk, R. L., and Ward, W. C. (1957
load-carrying behavior of continental Brazos River bar: a study in the
shelf sediments will be that gained by significance of grain size parameters,
in-place load tests. Such tests generally Journal of Sedimentary Petrology ,
will be reliable because the sandy vol. 27, pp. 3-26.
shelf sediments are relatively homo-
geneous in the upper several meters. Hansen, J. B. (1961). A general
However, load tests usually are so formula for bearing capacity, The
cumbersome that a simply executed Danish Geotechnical Institute
standard technique must be developed. Bulletin, 11, pp. 3 8 — U 6 .

At the sites tested in this study,

higher ultimate carrying capacities (1966). Improved

were obtained than would have been pre- settlement calculation for sand.

dieted by Terzaghi's bearing capacity The Danish Geotechnical Institute

equation, based on the angle of shearing Bulletin, 20, pp. 15-20.

resistance, <f> , of the soil. An analysis

of load versus settlement behavior at Harrison, W., Lynch, M. P., and

site B indicates less penetration under Altschaeffl, A. G. (I96U). Sediments

less ultimate loading than would have of Lower Chesapeake Bay, with

been expected from a conventional emphasis on mass properties,

analysis (such as Terzaghi's pressure Journal of Sedimentary Petrology ,

versus width chart for 1-inch settlement). vol. 3U, pp. 727-755.

In addition, the settlement under any

given load decreased as the plate diam- Peck, R. B., Hanson, W. E., and

eter increased. Thornburn, T. H. (1953). Foundation

ACKNOWLEDGMENTS

Engineering . New York, John Wiley
& Sons. 1+10 pp.

Shepard, F. P. (1951*). Nomenclature

Commander Service Squadron 8, U.S. based on sand-silt-clay ratios,

Navy, Norfolk, Virginia, made available Journal of Sedimentary Petrology ,

the USS Salvager (ARSD-3) for the field vol. 2U, pp. 151-158.
tests. The Portsmouth Buoy Base, U.S.

Coast Guard, Portsmouth, Va. , provided Silverman, M., and Whaley, R. C. (1952).

the concrete buoy sinkers for the load Adaptation of the piston coring

frame. The authors are grateful for device to shallow water sampling,

the assistance rendered by the following: Journal of Sedimentary Petrology s

Professor J. Brinch Hansen, The Danish vol. 22, pp. 11-16.
Geotechnical Institute, Copenhagen,

Denmark, for critically reading the Sowers, G. F. (1962). Shallow founda-

paper; Mr. Earl W. Rayfield, Land and tions, Foundation Engineering ,

Sea Interaction Laboratory, Norfolk, Leonards, G. A., ed. New York,

Va. , for helping in all phases of the McGraw-Hill Book Company, pp. 525-632.
study; Capt. James Charles Rowland,
Lt. E. E. McNiff, and Lt . J. C. Furr,

U.S. Navy, for aid in designing the Sowers, G. F., Peck, W., and Gooding, P.

load frame and guiding the field opera- (1966). Settlement of foundations

tions aboard the Salvager; CPO H. R. on loose sand. Unpublished paper

Carrol and CPO J. V. Kesser, U.S. Navy, given at the American Society of

for directing the diving and coring work; Civil Engineers Structural Engineering

and. Lt. E. E. McNiff, Lt . (j.g.) E. T. Conference, Miami Beach, Florida,

Beckett, Ens. S. Kaye , and Messrs. January, 1966.
D. R. Oney (EN2), N. L. R. Tetrault

(HM2), B. S. Kepesky (BMCA), and J. D. Taylor, D. W. ( 19U8 ) . Fundamentals of

Bitscheneider (DC2), U.S. Navy, for Soil Mechanics. New York, John

making the dives. Wiley & Sons. 700 pp.

Plate-Load Tests on Sandy Marine Sediments, Lower Chesapeake Bay 283

Terzaghi, K. (19^3) . Theoretical Soil
Mechanics . New York, John Wiley &
Sons. 510 pp.

(1955). Influence of

geologic factors on the engineering
properties of sediments, Economic
Geology (50th Anniversary Volume),
pp. 557-618.

(1955). Evaluations of

coefficients of subgrade reaction,
G4otechnique t vol. 5, pp. 297-326.

Terzaghi, K., and Peck, R. B. (19U8).

Soil Mechanics in Engineering Practice
New York, John Wiley & Sons. 566 pp.

s

<=>

'6

o
■ — * o

«-» If)

uu

CO
CO

Qi L
Q- <=»

CO

o

OH o

LU O

> ~
<

s

#

#

#

WIDTH OF FOOTING (Feet)
5 10

1.0

*.

# SITE A
+ SITE B

41

0

38<

39'

N=50

N=40

*

36

N=30

N=20

N=10

N=5

4.0

WIDTH OF FOOTING (m)

15

o

CO

00

g QC

S co

CO
UJ

o
<

8

— s <

Note : Diameter values from Peck, Hanson, and Thornburn , 1963, on

basis of blow counts.

FIGURE 1. PRESSURES AND WIDTHS OF FOOTINGS FOR 1-INCH SETTLEMENT ON SUBMERGED

C0HESI0NLESS SOILS.

284 Marine Geotechnique

« SALVAGER ICLEVIS)

PADS
b CRANE PADS
c SNUBBING PADS
d JACKING PLATE
e BARS 4"x'/2"x4'4"
f I BEAM 10WF60
g I BEAM 10WF33
h 1 BEAM 6WF15 5

CROSS BARS (On
concrete blocks )

3V 8"-
1,9.5* ta)

:. .

FIGURE 2. LOAD FRAME. TOP: DRAWING AND SPECIFICATIONS. BOTTOM: PHOTOGRAPH
OF FRAME BEING TRANSFERRED - BY CRANE - FROM WHARF TO SUSPENSION

CABLES OF USS SALVAGER.

Plate-Load Tests on Sandy Marine Sediments, Lower Chesapeake Bay 285

\WIVB.

XWWWWWWX 10 ¥t J3\\\V

FIGURE 3. SCHEMATIC REPRESENTATION OF TEST SETUP AND PHOTOGRAPHS OF COMPONENTS,

286 Marine Geotechnique

76° 13' 12'

FIGURE 4. LOCATION OF TEST SITES.

11' LONGITUDE

Plate-Load Tests on Sandy Marine Sediments, Lower Chesapeake Bay 287

100

80

O

CO

60

oc

uj 40

U

£ 20

■■^^ '"-^

\A3

^~

\ » \A2

A1\\B1
NX

k.

E
I

h-

a.

LU

a

20

40

60

80

SITE A

Dr = 50

Dr = 67

0.5 0.1 0.05

GRAIN SIZE (mm)

SOIL PROFILES

0.01 0.005

SITE B

WELL SORTED
SAND

MODERATELY
SORTED SAND

POORLY SORTED
CLAYEY SILT

Or -57

0.001

WELL TO
MODERATELY
SORTED
SAND

TO 200 cm

FIGURE 5. GRAIN-SIZE DISTRIBUTION CURVES AND SOIL PROFILES FOR SITES A AND B

288 Marine Geotechnique

Z

o

8.0-

90 —
10.0 —
11.0

AVERAGE PRESSURE ( lb/ft')
.1,000 2.000 3.000 4 000 5.000

- PLATE DIAMETER

o 1.0 ft. (30.5cm)
■ 1.5 ft. (45.7cm)
a 2.0 ft. (61.0cm)

J_

1

5.000 10,000 15.000 20.000

AVERAGE PRESSURE! kg /m')

25.000

FIGURE 6. LOAD VERSUS PENETRATION, SITE A.

z
o

<

9 3.0

z

LU

a- 40

AVERAGE PRESSURE (lb/ft2)
2,000 4JJ00 fiOOO 8,000

PLATE DIAMETER

o 1.0 ft. (30.5cm)
s 1.5 ft. (45.7cm)
a 2.0 ft. (61.0cm)

10,000 20,000 30,000 40,000

AVERAGE PRESSURE (kg/m2)

FIGURE 7. LOAD VERSUS PENETRATION, SITE B

Plate-Load Tests on Sandy Marine Sediments, Lower Chesapeake Bay

289

b"

50

40

30

20

10

10,000
— r

20,000
— i —

((T, +C73)/2kg/m2
30,000 40,000

50,000

60,000

TEST
NO.

1
2
3

<p= 38°

INITIAL

VOID RATIO

0.64

066

057

FAILURE

VOID RATIO

0.67

0.68

0.59

TEST NO. 3

-X

b"

I

90

FIGURE 8. RESULTS OF DRAINED TRIAXIAL TESTS, SITE B [D = 0.55 (TEST 1);

0.50 (TEST 2) : 0.72 (TEST 3)].

290 Marine Geotechnique

fc

<

a.

<

U

O

z

of
<

<
2

0

.2

METERS
.4 .6 .8

6000

i

I V 1
0

4000

—

G

A— Ny = 195

2000

7

n

—

o,

1

*^Ny=62

1

- 30,000

20,000

E

H 10,000

0 12 3

PLATE DIAMETER (ft.)

FIGURE 9. ULTIMATE BEARING CAPACITIES AND PLATE DIAMETERS AT SITES A AND B

METERS
.4 .6

Q.

1 i r~

P = SETTLEMENT UNDER GIVEN LOAD

P, = SETTLEMENT OF l' DIAMETER PLATE UNDER
SAME LOAD

1-

SITE B

TAYLOR

%$^'^0 ^

'*- -T-^rrrnr.4000 lb/ft'

O 1 2

PLATE DIAMETER (ft.)

FIGURE 10. RELATIONSHIP OF SETTLEMENT AND PLATE DIAMETER AT SITE B

19

©This paper is published solely for the information of members of the 7th International Sedimenjolojlcal
Congress and may not be reproduced, as a whole or in part, without the prior permission of the author.

"PLEISTOCENE- RECENT BOUNDARY IN THE MALACCA STRAIT, SOUTHEAST ASIA

G. H. Keller
Institute for Oceanography, ESSA

INTRODUCTION

Several studies have been presented describing the Pleistocene drainage on the
Sunda Shelf in the vicinity of Sumatra, Java and Borneo (Molengraaf f , G.A.F. 1919-
Proc. Roy. Acad. Amsterdam. 23, 395; Dlckerson, R.E. 1941. Univ. Pennsylvania.
Bicentennial Conf., Univ. Pennsylvania Press, Phila; Wyrtki, K. 1961. Naga Report, 2,
Univ. Calif. Scrippslnst. Oceanography). Recent investigations to the north of this
area in the Malacca Strait, reveal a marked change in direction of the Pleistocene
drainage from that reported by the earlier studies of the shelf. Until now, little has
been known of the depositional environment that existed on the Sunda Shelf during the
Pleistocene epoch. However, an indication of the environmental conditions during the
latter part of this epoch has been gained from the sediment cores collected during the
investigations of the Malacca Strait. The following discussion presents the findings
of these recent studies.

The Malacca Strait Is a shallow passage between the Malay Peninsula and Sumatra
connecting the Indian Ocean with the South China Sea. It is approximately 805 km long
and varies in width from 64 km in the south to 257 km in the north. Water depths seldom
exceed 45 m and more commonly are about 27m. Currents are generally strong, often reaching
velocities of 2-1/2 knots in many areas. The strait is a part of the Sunda Shelf, which
is one of the major continental shelf areas of the world. Unlike most others, this shelf
is believed to be a peneplain formed in the late Pliocene (Molengraaf f, G.A.F. 1919. ibid).

The area now occupied by the Malacca Strait was a topographic low in the peneplain prior
to the Quaternary as a result of late Tertiary folding. This depression has been termed the
"backdeep" as opposed to the foredeep presently located west of Sumatra (Bemmelan, R.W. Van,
1949. The Geology of Indonesia, IA, Govt. Printing Office, The Hague). It was along this
low that later regional runoff was channeled.

During 1961 and 1964, oceanographic cruises by the U.S. Naval Oceanographic Office
and the U.S. Coast and Geodetic Survey respectively, collected 76 sediment cores and
conducted hydrographic surveys in the Malacca Strait. This study is based on these data.

LOWERING OF SEA LEVEL

As the Pleistocene icecaps grew, sea level was lowered and long periods of erosion
followed. Stream and river channels were cut deeper and extended across the exposed
continental shelf as a result of the lowering of their base-level. One Pleistocene river
system to the north of Borneo, was approximately three times larger than that of the
present Mississippi Rlver(Dickerson, R.E. 1941. ibid). The lowering of sea level, estimated
to be of the order of 73 to 91 m (Dickerson, R.E. 1941. Ibid) resulted in an extensive area
being laid dry and uniting the present islands of Sumatra, Borneo and Java with the
Malay Peninsula. The extent of this exposed land area is approximated by the 80 m isobath
(Fig. 1).

It is generally accepted that the submergence of the Sunda Shelf was caused by the
melting of the Pleistocene ice sheets rather than any by crustal movement (Molengraaf f )
G.A.F. 1921. Geographical Jour. 57 , 95 ; Bemmelan, R.W. Van 19^9, ibid)

Ihe uniform depth of the shelf, between 55 and 64 m, and the numerous drowned river valleys
tend to support this hypothesis. Carbon-l4 analysis of peat found in a sediment core taken from
the strait at a depth of 26.5 m below sea level indicated an age of 10,000 ± 200 years
B.P. (U.S. Geological Survey No. W-1675). This date agrees with similar dates on the
curve established by Curray (i960. In Recent Sediments, Northwest Gulf of Mexico.
Shepant, F.P, Phleger, F.B. , Andel, T.H. van (eds.). Sediments and history of the Holocene
transgression, continental shelf northwest Gulf of Mexico, p 221) for the rise of sea level
since the late Quaternary. It is evident from this correlation that the Malacca Strait has
remained stable during the last 10,000 years.

ERAINAGE PATTERN

Further evidence of the Pleistocene-Recent history was found in other core samples
as well as in the bathymetry of the strait. Previous discussions pertaining to the
Pleistocene on the Sunda Shelf have usually dealt only with the drowned river channels
extending eastward from southern Sumatra into the South China Sea (Molengraaff , G.A.F. 1921.
ibid); Kuenen, Ph.H., 1950 Marine Geology, John Wiley and Sons, New York). As a result of
the hydrographic data obtained from the recent oceanographic cruises the bathymetry of the
strait can be delineated much more accurately than was possible in the past. A study of these
data clearly reveals the presence of several drowned river channels extending from Sumatra and
the Malay Peninsula into the Malacca Strait. These Pleistocene rivers comprised a major
river system draining to the northwest into the Indian Ocean (Fig. l). No evidence is found
to indicate that there was any major drainage to the southeast into the South China Sea.
although this possibility should not be excluded without further study.

Several completely buried channels, approximately 1 km wide and 9 to 10 m deep, have
been observed on the echo-sounding records. Some channels are reflected in the present
morphology of the strait floor which appear to have been even larger than those now
completely buried.

SEDIMENTS

Approximately one-third of the cores taken in the strait penetrated what is believed
to be the Pleistocene and exhibit two distinct layers: an upper layer of Recent sand with
shell debris and an underlying Pleistocene stratum of stiff, gray, silty clay. The sand is
characteristically dirty and composed of medium to fine sand-size particles of quartz, shell
debris, glauconite and rock fragments. The underlying silty clay is homogenous and contains
only minor or traceable amounts of shell and rock fragments. Peat is usually disseminated
throughout the silty clay both as fine particles and as large fragments of wood or plant roots.
The lower 15 cm of one core was entirely composed of peat.

The pronounced change in sedimentary characteristics across this contact is quite
evident (Figs. 2 and 3) . Texturally, the underlying silty clay is essentially uniform, but
considerable compositional variations are found.

A distinct Pleistocene-Recent boundary can be traced over large portions of the strait
by means of the sediment cores and echo-sounding records because of the difference between the
Recent and Pleistocene sediments mentioned above. This boundary commonly has an irregular
profile - relief of 1 to 5 m - but often exhibits a flat surface. The Pleistocene sediments
appear to crop out on the strait floor in many areas. The presence of Pleistocene submarine
terraces similar to those described in the neighboring South China Sea by Krempf and Chevey

(193^ Proc. Fifth Pac.Scl. Congress, 2, 849), were not observed in any of the echo-sounding
records from the strait. If the terraces exist they are probably masked by Recent sediments.

CONCLUSIONS

The two sediment types found below the strait floor are indicative of two distinctly
different environments of deposition. The silty clay has a relatively high organic and
low calcium carbonate content corresponding to the relationship found in an estuarine
environment (Emery, K.O., Stevenson, R.E. 1957- In Treatise Marine Ecology and Paleoecology,
J.W. Hedgpeth (ed), v. 1, Estuaries and Lagoons, p. 673). The overlying coarse sediments
represent the present high energy environment of the strait. I believe that the stiff clay
represents a late Pleistocene estuarine deposit, which is in agreement with the Pleistocene
geography proposed by Molengraaff (1921, ibid) for this area. The stiffness of the clay
indicates it may have been partially indurated as a result of sub-aerial exposure during
a period prior to submergence. This clay was eroded during the period of exposure and the
drainage pattern now observed in the present bottom topography is a result of that erosion.
Based on the radiocarbon date mentioned earlier, from a peat sample taken 20 cm below the sand
- sllty clay contact, it is reasonable to assume the stiff, silty clay is of late
Pleistocene age.

The depth of this Pleistocene surface below the present sea floor differs considerably,
depending on the rates of deposition in the various portions of the strait. Sediment cores
collected along the axis of the strait commonly penetrate this boundary at depths of 20 cm or
less (Fig 3). However, in many areas cores 3 in In length do not penetrate to the Pleistocene.
This may indicate a greater thickness of Recent sediment, or that the surface may have been
obliterated by burrowing organisms such as seen in core W5 (Fig. 2).

As sea level rose during the various interglacial periods and particularly at the end
of the last glacial period, the rivers entering the strait were drowned and an estuarine
environment evolved. This environment remained until the sea rose to a level submerging the
entire Sunda Shelf. At this time the present strait environment came into existence.

ILLUSTRIATIONS

Fig. 1. Pleistocene drainage pattern.

Fig. 2. Pleistocene - Recent contact as shown in sediment cores.

Fig. 3- Vertical variation of textural and chemical properties in cores penetrating the
Pleistocene sediments.

0 100 200

KILOMETERS

BS-85 BS-96

BS-II2

fiq. 2

i k, \ ijte

STATION
96

Sand.Silt
Clay

.1, . , ,i. . . .i I L^J i

Sorting
1 2 3

Ca Coq

(%) 6

Organic Carbon

Nitrogen

0 20 40 .4 .8 1.2 1.6 2.0 .04 .06 .08

J_l l_J I ^J I L_l 1 I I I I I I I I I I

i x

RECENT

/ /

PLEISTOCENE iy

T
i

V

i

STATION Md

130 (0)

3456789 10

LU

Or

CtC

o

10-

u

20-

z

E

30-

X

40-

1—

Q_
LU

50-

a

60i

I I I I I

Sand
Silt
Clay
0 (%) 100

Sorting

Ca Coo

(0)

234

0 20

1

\

Organic
Carbon

(%)
1.2 1.6 2.0

Nitrogen

(%)
.05 .07

i

~T"

/ /•

RECENT
'JPLEISfOCENEY

\

\

\T

i

\

T
i -

fig. 3

I SUBSAMPLE INTERVAL)

Reprinted from THE JOURNAL OF SEDIMENTARY PETROLOGY Vol. 2>1 , No • 1
Journal of Sedimentary Petrology, Vol. 37, No. 1, pp. 102-127
Figs. 1-17, March, 1967

The Society of Economic Paleontologists and Mineralogists

SEDIMENTS OF THE MALACCA STRAIT, SOUTHEAST ASIA1

20

GEORGE H. KELLER2 and ADRIAN F. RICHARDS
Department of Geology, University of Illinois, Urbana, Illinois

ABSTRACT

The Malacca Strait is a shallow passage between the Malay Peninsula and Sumatra with oceanographic and
bottom-sediment characteristics closely related to strong currents, debouching rivers, climatic variation, and
the close proximity of bordering land masses. The strait assumed its present configuration as a result of the
postglacial rise of sea level which drowned the Sunda Shelf. An essentially tidal northwest current flow prevail-
ing in the strait throughout the year is largely responsible for the hydrographic and oceanographic conditions
in the region. Surface salinities and temperatures are generally found to be lower than in the surrounding seas.
A prominent wedge of cool-temperature, high-salinity bottom water is found to extend from the Andaman Sea
into the strait.

Bottom sediments primarily consist of muddy sands, with large areas cf mud occurring in the vicinity of
river mouths and in the Andaman Sea. Calcium carbonate, primarily composed of mollusk shelL and fora-
miniferal tests, and organic carbon are generally found only in minor amounts in the strait. Higher concentra-
tions of calcium carbonate are confined to local shell deposits and larger percentages of organic matter are
found in the vicinity of river mouths and in the fine sediments of the Andaman Sea. The non-calcareous detrital
fraction is dominated by quartz with minor amounts of orthoclase and plagicclase feldspars. The heavy mineral
suite is complex because of the varied geology of the bordering land areas and consists primarily of leucoxene,
ilmenite, staurolite, biotite and amphiboles. Kaolinite and mixed-layer minerals consitute the dominant clay
minerals; lesser amounts of illite and montmorillonite also are present. A slight decrease in illite and an increase
in mixed-layer minerals with depth was observed in some cores. Volcanic ash of an andesitic origin is found
throughout much of the area. Many cores penetrate what appears to be a Late Pleistocene surface consisting
of stiff, slightly indurated silty clay. The clay contains much peat, some of which has given a radiocarbon date
of 10,000 years B. P.; it appears to represent a former tidal flat or estuarine deposit.

INTRODUCTION

The Malacca Strait, a shallow passage be-
tween the Malay Peninsula and Sumatra, con-
nects the Indian Ocean with the South China
Sea (fig. 1). It is approximately 805 km long and
varies in width from 64 km in the south to 257
km in the north. The strait provides a deposi-
tional environment of considerable complexity
owing to its physiography, oceanographic and
climatological conditions, and the varied sedi-
ment source areas provided by the adjacent land
masses. This study was initiated to investigate
the importance of these variables upon the sedi-
ments of the strait.

Previous studies were largely made to support
fisheries research or as a minor portion of a more
extensive study of the Sunda Shelf. The earliest
recorded observations were those of the Dana
Expedition between 1928 and 1930, from which
plankton and oceanographic data were obtained
from two stations in the strait. Surface salinity
measurements over varying periods of time were
reported by Soeriaatmadja (1956) and Selvara-
jah (1962). Additional oceanographic stations
were occupied by the Indonesian research vessel

1 Manuscript received June 29, 1966.

2 Present address: Institute for Oceanography,
ESSA, Silver Spring, Maryland 20910.

SAMUDERA (Anonymous, 1957) and the
French vessel JEANNE D'ARC (Anonymous,
1963). Wyrtki (1961) discussed the climatology
and oceanography in the region of the strait as
part of his study of the physical oceanography of
the southeast Asia waters. Van Baren and Kiel
(1950) include 18 samples from the strait in a
sedimentary petrology study of the Sunda Shelf.
According to the files of World Oceanographic
Data Center A, five sediment samples were col-
lected by the Russian Oceanographic vessel
VITYAZ (cruise 35), but as yet these data are
not available (B. S. Richmond, personal com-
munication).

In the spring of 1961, the U. S. Naval Oceano-
graphic Office began an investigation of the
general sediment distribution and hydrographic
conditions within the Malacca Strait; a more de-
tailed survey and sampling program was under-
taken later in 1961. During March and April
1964, the senior author, while aboard the U. S.
Coast and Geodetic Survey Research Vessel
PIONEER, obtained additional hydrographic
data as well as bottom samples to supplement
those collected by the Naval Oceanographic Office.
Data collected during these three cruises include
observations obtained from 24 oceanographic
stations (temperature and salinity), 155 bottom
sediment stations, and five 26- to 33-hour anchor

SEDIMENTS OF THE MALACCA STRAIT, SOUTHEAST ASIA

103

stations for the determination of the spatial and
temporal variations of currents (fig. 2). It is up-
on these data that this study is based. The loca-
tions of all stations plus the sedimentological and
oceanographic data obtained at each site are on
file at the National Oceanographic Data Center,
Washington, D. C. (NODC geology acquisition
number 051866-1 and cruise numbers 00545 and
00639).

PHYSIOGRAPHY AND AREAL GEOLOGY

Bordering Lands and Islands
The regional physiography and geology of the
adjacent lands is germane to this investigation
because sediments brought to the strait are con-
trolled by the nature of these bordering pro-
vinces, as will be demonstrated later.

Sumatra can be divided topographically into
three parallel provinces extending along the is-
land from northwest to southeast; the Barisan
Mountains, a transition zone, and a wide coastal
plain. The volcanic Barisan Mountain chain
largely consists of Late Paleozoic crystalline
schists with some shallow depressions containing
Mesozoic, Tertiary and Quaternary sediments.
Tertiary-Quaternary volcanics, including rhyo-
lite and pumice tuffs as well as andesite cones,
cover wide portions of the province. Some iso-

lated intrusives of plutonic rock of possible
Quaternary age constitute the remaining out-
crops (fig. 2). The mountains descend gradually
eastward to a transition zone of rolling hills com-
posed of Paleozoic schists and Mesozoic and
Tertiary sediments. Minor amounts of Tertiary-
Quaternary rhyolithic tuff and plutonic intru-
sions also occur in the transition zone. The third
zone is an extensive flat lowland of Quaternary
and Recent sediments that borders the strait; it
is 48 to 193 km wide. Most of this coastal zone is
swampy and covered by jungle. Although nu-
merous rivers drain both sides of Sumatra, the
majority have their headwaters in the Barisan
Mountains and drain into the strait. The east
coast of Sumatra has generally shifted consider-
ably toward the strait owing to the deposition of
the large sediment load of these streams (Van
Bemmelen, 1949, p. 702). A recent mariner's re-
port based on actual observations compared with
existing charts for the area by H. P. Hansen
(personal communication) indicates, however,
that the shoreline along northeastern Sumatra in
the vicinity of Belawan has receded as much as 2
km. Hansen has also observed drowned tree
stumps 300 m off shore in this same area.

The Malay Peninsula consists of a hilly and
mountainous interior bordered on both sides by

100

110

120

10

0

10

CO ^

10"

ou

10C

100 110" 120^

Fig. 1. — -Index map, Malacca Strait and adjacent seas.

104

GEORGE II. KELLER AND ADRIAN F. RICHARDS

96 9 8 100 102 104 106

Fig. 2. — General geology of the bordering land areas and station locations in the Malacca Strait.

coastal lowlands. Several north-south trending
mountain ranges extend from the Thailand bor-
der across approximately two-thirds of the pen-
insula, becoming discontinuous in the southern
third. Granitic intrusives of Jurassic age crop out
along the entire length of these ranges as well as
in the south, where little of the high relief re-
mains. A. hilly topography exists between the
ranges, which consists mainly of Permo-Carbonif-
erous limestones and shales and Triassic arenace-
ous deposits (Van Bemmelen, 1949, p. 360).
Effusive rocks of unknown composition are
common throughout much of this region. Rolling
hills and lowland plains of Paleozoic and Qua-
ternary sediments extend out from the moun-
tains toward the coast where the coastal low-
lands bordering the strait have an average
width of 64 km. The extensive volcanism that
occurred on Sumatra in the Quaternary also re-
sulted in the widely distributed ash deposits on
the Malay Peninsula (United Nations Economic
Commission for Asia and the Far East, 1961, p.
36).

The islands in the vicinity of Singapore consist
mainly of sedimentary rocks and granites of
Mesozoic age. To the north, Penang Island is also
composed of Mesozoic granites, and Langkawi
Island near the border of Thailand consists pre-
dominantly of an Ordovician-Silurian calcareous

sequence (Alexander, 1959). Islands along the
coast of Sumatra in the narrow portion of the
strait are formed of Recent sediments.

Determination of the sedimentary provenance
in the strait is difficult as a result of the numer-
ous streams and rivers draining the complex and
varied geology of the adjacent lands. Sumatra
has the greatest influence on the sediments of the
Malacca Strait although Malayan drainage con-
tributes large quantities of material. Much of the
sediment entering the strait from Sumatra ap-
pears to be finer than that from streams drain-
ing the Malay Peninsula. This is attributed to
the extensive drainage system from western
Sumatra eastward across the lowlands and fi-
nally through the wide coastal plains covered with
thick vegetation. There is a tendency in the
coastal region for much of the coarser part of the
sediment load, that otherwise reaches the strait,
to be intercepted. Islands south of Singapore
contribute only minor amounts of material to
the southern portion of the strait, as indicated by
the concentration of various iron minerals be-
lieved to have come from the laterite deposits
on these islands. Wind-derived sediment is
present in the form of volcanic ash and pumice,
but it is of minor importance relative to the
quantity of stream-borne material. Much of the
ash found in the strait is considered to be detri-

SEDIMENTS OF THE MALACCA STRAIT, SOUTHEAST ASIA

105

tal rather than wind-borne. Sediment entering
from the Andaman Sea consists mainly of plank-
tonic foraminifera carried in by southeast-flow-
ing bottom currents. Organic production in the
strait is the only remaining source of sediment,
and in some regions it is a significant source.

Strait Floor

The topography of the strait floor prior to
1961 was only generally known because of the
lack of adequate bathymetric surveys. Data from
the recent cruises of the Naval Oceanographic
Office and the Coast and Geodetic Survey have
provided much more detail. Contours at 10
fathom (18.3 m) intervals in the strait and 100
fathom (183 m) intervals beyond the continental
shelf have been drawn (fig. 3) based on these re-
cent soundings that are corrected for sound
velocity and those from the U. S. Hydrographic
Office Chart 1595 that are mainly wire-line
soundings.

Numerous islands clutter the southern en-
trance to the strait where water depths seldom

exceed 20 fathoms (37 m) and more commonly
are about 15 fathoms (27 m). Within the narrow
portion of the strait, from Singapore northwest-
ward to approximately Port Swettenham, water
depths increase slightly toward the middle por-
tion with shallower sill-like depths at both ends.
The shallow areas that occur at either end may
be due to relict deposits which are expressions of
the topography prior to an eustatic rise of sea
level.

The broad shallow areas bordering Sumatra
along the southern portion of the strait result
from the large quantity of sediment supplied by
local rivers. In this region, prominent fine sand
and mud ridges trend parallel to the axis of the
strait. Some of these ridges rise as much as 21.9
m above the bottom and are up to 48 km long
(fig. 4, profiles A & B). Present orientation of
these ridges appears to be maintained by bottom
currents, which may reach velocities of 77 to 103
cm/sec. Umbgrove (1949, p. 5) reports the pres-
ence of similar features to the south in Banka
Strait, between Sumatra and Banka Island,

Fig. 3. — Bathymetry of the Malacca Strait with a 10-fathom (18.3 m) contour interval from
10 to 100 fathoms and a 100-fathom (183 m) interval beyond 100 fathoms.

106

GEORGE H. KELLER AND ADRIAN F. RICHARDS

0

«2 40
E 80
9 120
i3 160
,« 200
240-

vert X100

Fig. 4. — Depth profiles.

which he describes as remanent features result-
ing from currents eroding numerous channels in
the floor of the strait and leaving ridges between
the channels. The term "channel," as used by
Umbgrove is a misnomer in that the bathymetry
shows no indication of trenches or channels hav-
ing been cut in the strait floor. Ridges occur
above the sea floor in the Malacca Strait and the
depressions one commonly associates with chan-
nels are not found. If these ridges are the result
of erosion by sea currents, as Umbgrove (1949,
p. 6) suggests, one would expect the ridge tops
to be at the same depth as the floor of the strait
surrounding the area where these features are
found. This is not the case. We believe the ridges
in both straits are depositional rather than ero-
sional features. Once the sediments from the
Kampar River and its tributaries enter the strait
they are carried northwestward by the currents.
Farther out in the strait some of these sediments
are deposited to form linear ridges parallel to the
current flow and extending away from the river
mouths. Some ridges show a gradual tapering in
width away from the rivers. Many of the ridge

tops are composed of mud, but sand is generally
found in the troughs. Shepard, Emery, and
Gould (1949, p. 28) suggested that the presence
of sand between the ridges may possibly be
attributed to erosion of the ridges. Since the
ridges result from the deposition of fine sedi-
ments entering the strait from nearby rivers, the
possibility that enough sand would be available
from erosion of the ridges seems doubtful. It is
more likely that the sand is transported into the
area (and concentrated in the troughs) by cur-
rents moving through the strait. Van Veen
(1936, p. 198) believes that these features are
similar to the sand ridges he has observed in the
Straits of Dover. The only similarity between the
ridges in these two areas is that they are deposi-
tional features formed parallel to the current
flow. The ridges in the Strait of Dover are at-
tributed to strong currents and a small supply of
sand, whereas those in the Malacca Strait are
the result of strong currents and large quantities
of fine sediment.

The general physiography of the strait floor
was probably established during a period of

SEDIMENTS OF THE MALACCA STRAIT, SOUTHEAST ASIA

107

lower sea level in the Late Pleistocene. Present
sea currents tend to maintain the bottom con-
figuration except in areas where large quantities
of sediment are deposited from nearby rivers
such as in the southern portion of the strait.

Sand waves occur in the region consisting en-
tirely of coarse to medium sand in the area of
maximum current velocity between Malacca
and Port Swettenham. These waves are oriented
perpendicular to the axis of the strait and have
wave lengths of approximately 241 to 900 m and
heights of 4.6 to 15.3 m (fig. 5). The larger waves
show a definite asymmetry with steep slopes to
the northwest. Smaller waves tend to show a
greater degree of symmetry. Small sand waves
also occur near One Fathom Bank, west of Port
Swettenham (Wyrtki, 1961, p. 9).

In the central portion of the strait (lat. 4° 42'
N., Long. 99° 07' E), underwater photographs
revealed the presence of small ripple marks
having heights of 3 to 4 cm that trend in a north-
west-southeast direction. This unusual trend,
parallel to the general current flow, is attributed
to local variation in current direction.

The strait widens to the northwest from a
width of 64 km at Port Swettenham to about 257
km where it enters the Andaman Sea. Over this
distance the axis of the strait gradually deepens
(fig. 4, profile E). The slope off the coast of the

Malayan Peninsula decreases to the north, but
increases along the Sumatra coast (fig. 4, pro-
files C and D). The 50 fathom (91.4 m) "hole"
west of Port Swettenham may be a relict feature
of Pleistocene drainage or possibly structurally
controlled. Little can be said about the latter
possibility without subbottom information. The
presence of Pleistocene submarine terraces,
similar to those found in the neighboring South
China Sea (Krempf and Chevey, 1934, p. 849),
were not observed in any of the echo-sounding
records from the strait. If present, they may
have been masked by Recent sediments. Bottom
contours indicate the presence of a drowned
river channel extending northward from the
mouth of the Rokan River. Numerous sub-
merged river valleys have been reported and de-
scribed on the Sunda Shelf by Molengraaff
(1921), Umbgrove (1949, p. 2), and Kuenen
(1950, p. 482).

DEPOSITIONAL ENVIRONMENT

Climate
Climatological variation directly influences
depositional conditions in the strait in two
major ways: runoff from bordering land prov-
inces and general circulation in the strait. Since
both factors are significant to an understanding
of the sediments in the Malacca Strait a brief

1 f •^W^WW^\b T^p^f

-90

400-
(ft)

0

km

-120
(m)

Fig. 5. — Sand waves. Series of nearly continuous echo-sounding records taken from an area
extending south-east from Port Swettenham.

108

GEORGE II. KELLER AND ADRIAN F. RICHARDS

96 98'

100 102 104 9 6 9 8

SURFACE
CURRENTS&WINDS

(n.e. monsoon) -

0 10 20

i , , , 1 1

wind°/9 _

100 102 104

96" 98v

100 102 104

10

— i 1

SURFACE
CURRENTS&WINDS
(s.w. monsoon) Jq

0 10 20

i .... i , ... i

wind % -

96 98

100 102 104

Fig. 6. — Surface currents and winds during the northeast and southwest monsoons. Wind roses fly with the
wind, and the number of barbs on the arrow indicate the Beaufort scale of average wind force. The length of
the arrow measured from the center of the circle gives the average percent of time the wind blows in a particular
direction. The number in the circle indicates the percent time of calms. Modified from the U. S. Navy Hydro-
graphic Office (1944).

discussion of the climatological aspects is per-
tinent. The climate of the region is typically
equatorial — hot and humid. Monsoonal effects
are not as severe in the strait as in the more open
surrounding seas because of the sheltering effect
offered by the Malay Peninsula and the island of
Sumatra. Indirectly, however, the monsoon
seasons greatly influence the circulation in the
strait. As a result of the monsoons, two rainy
seasons of unequal magnitude occur without any
really dry period. Average annual rainfall over
most of the area ranges from 1961 to 2649 mm,
with the greater rainfall occurring in the lower
latitudes (U. S. Navy Hydrographic Office,
1951, p. 18).

Variation in surface wind direction and force
follow the monsoonal seasons. Northeast mon-
soon conditions prevail from December to
March. Near the equator during this time the
variable winds have an average Beaufort force of
2 or 3 and originate mainly from the north. The
southwest monsoon season extends from May to
October and is preceded by squally weather,
strong winds and considerable rain. Winds are
mainly from the south and southwest and are of
about the same force as those occurring during
the northeast monsoon (Royal Netherlands
Meteorological Institute, 1935). Severe squalls
with gale force winds often accompany the
southwest monsoon, but have little influence on
the circulation in the strait because of their

limited fetch (Great Britain Hydrographic De-
partment, 1958, p. 31). Wind roses and the
associated surface currents for the two monsoon
seasons are shown in figure 6.

Oceanography

An understanding of tides, currents, salinity
and water temperature is important in the
interpretation of a shallow-water marine sedi-
mentary environment. Currents normally ac-
count for the grain-size distribution. Tempera-
ture and salinity measurements delineate various
water masses, which usually have considerable
influence on the type of sediments deposited,
particularly those of organic origin that are de-
pendent on the properties of the overlying water.

Oceanographic conditions in the Malacca
Strait differ considerably from those of the sur-
rounding seas yet they are greatly influenced by
the currents and water masses driven into the
strait fr;om these seas by the monsoons. The
most significant effect of the monsoons is their
regulation of the circulation in the South China
and Java Seas as well as the Indian Ocean, which
in turn control conditions in the strait. Within
the strait only a minor effect is noted directly
from the monsoons because of the shielding pro-
vided by the bordering land masses. Heavy
rainfall in the higher elevations of the bordering
lands, however, results in a large runoff into the
strait which does influence salinity conditions.

SEDIMENTS OF THE MALACCA STRAIT, SOUTHEAST ASIA

109

Tides and Currents. — Currents flow through the
strait in a general northwest direction through-
out the year as a result of monsoonal effects on
the neighboring seas. This flow is also strongly
influenced by tides. It is a continuation of the
south-flowing current during the northeast mon-
soon which rounds the tip of the Malay Penin-
sula from the South China Sea and passes into
the strait (fig. 6). During the southwest mon-
soon, part of the northwestward-flowing water
moving from the Java Sea into the South China
Sea passes directly into the Malacca Strait
(U. S. Navy Hydrographic Office, 1945). Cur-
rents are weak in June, July and August, and
often a southeast flow is recorded as a result of
the Indian Ocean current pattern that causes a
pile up of water in the Andaman Sea. The differ-
ence in sea level during this period from the
southern to the northern entrance of the strait is
of the order of 10 to 15 cm with the slope toward
the Andaman Sea. At the time of maximum flow
to the northwest in January to March a differ-
ence in sea level of 50 cm has been reported by
Wyrtki (1961, p. 126). Measurements during
similar periods on both sides of the strait at vari-
ous points indicate that there is little if any
difference in sea level across the strait (Wyrtki,
1961, p. 125).

Tides are of the semidiurnal type throughout
the strait except in the southern portion near
Singapore where they are of a "mixed type."
Flood currents set to the southeast and ebb to
the northwest in the strait. The combination of
tide and the general northwesterly flow causes
the ebb to last longer and run slightly stronger
than the flood. Current velocities average be-
tween 55 and 103 cm/sec. in the center of the
strait, but closer to shore and in restricted chan-
nels they may reach 180 cm/sec. during spring
tides.

Five current stations were occupied for periods
of 26 to 33 hours during November 1961 (fig. 2).
All measurements were made with a Roberts
current meter. Surface and bottom currents for
four of these stations as well as the respective
tides are shown in figure 7. Tidal conditions at
stations A and B are more uniform and the
curves show less inequality than at stations C
and D, which would be anticipated in an area of
semidiurnal tides. North of Penang Island at
station B, bottom currents are of about the
same or slightly higher velocity than surface
currents during times of flood, but appear to be
slower during ebb (fig. 7). Bottom currents tend
to be stronger than surface currents during
periods of flood along the western margin of the
strait at station A. The general surface current
pattern during the southwest monsoon indicates
that currents turn away from the vicinity of
Penang Island and flow westward toward Suma-
tra (fig. 6). This is believed to account for the
slightly higher current velocities observed near
Sumatra at station A. Tidal currents are com-
plex in the narrow part of the strait, particularly
just north of Singapore. At station C, between
the Kampar River mouth and Singapore, sur-
face currents flowing towards the northwest are
strongest during ebb flow while bottom currents
in the opposite direction are strongest during
flood. A surface velocity of 108 cm/sec. at station
C was the highest measured in the strait during
this series of current observations. Station D,
southeast of Singapore, is in an area affected by
tides from both the Malacca Strait and the
South China Sea. Surface and bottom currents
at this station have essentially the same velocity.
These currents are variable in direction and
little information can be gained from this single
station when compared with the local tide data.

Sediments within the strait are predominantly

CURRENT
N W

(knots)

CURRENT

NW I SE

(knots)

a O

Fig. 7. — Tide and current observations from various portions of the Malacca Strait.
The locations of stations A to D are shown in figure 2.

110

GEORGE H. KELLER AND ADRIAN F. RICHARDS

transported to the northwest, because of the
general current pattern. This is particularly
evident in the narrow portion of the strait be-
tween Malacca and Port Swettenham where the
steep slopes of the large asymmetrical sand
waves are inclined to the northwest. Local varia-
tions in currents and bottom topography in the
northern and central parts of the strait also ac-
count for the movement of sediment in directions
other than to the northwest as indicated by the
small ripple marks mentioned earlier. The rela-
tively short period of current observations pre-
sented here gives only an indication of the gen-
eral current conditions with depth during the
beginning of the northeast monsoon. Longer and
more frequent observations are required to
interpret more adequately the current-sediment
relationship. While the above-mentioned cur-
rents are of great importance in transporting
sediment throughout much of the strait, streams
draining the adjacent land areas are also signifi-
cant with respect to sediment distribution along
the coastal margins. Stream currents in these re-
gions appear to be stronger than the currents in
the strait and account for much of the sediment
transported into these areas.

Salinity. — A south-flowing current rounds the
Malay Peninsula from the South China Sea and
causes water of salinity greater than 32.0%
to enter the Malacca Strait, displacing the low
salinity waters in the southern portion of the
strait during the northeast monsoon (Soeriaat-
madja, 1956, p. 31). This low salinity water
moves northward mixing with the higher salinity
water of the central and northern portions of the
strait. The result of this movement is a salinity
maximum in the southern part of the strait dur-
ing the northeast monsoon in spite of the heavy
rainfall.

A decrease in salinity in the southern portion
of the strait is thought to be caused by runoff
from the large rivers on the east coast of Suma-
tra and the low salinity water entering from the
Java Sea during the southwest monsoon (Sel-
varajah, 1962, p. 2). The annual variation of sur-
face salinities in the northern and central parts
of the strait shows two maxima and two minima,
essentially a mirror image of the amount of rain-
fall (Soeriaatmadja, 1956, p. 30). Rainfall and
salinity are not interdependent factors to the
south.

Surface salinity measurements made during
late March and early April 1961 do not agree
with the general pattern described by Soeriaat-
madja (1956, p. 40) for that time of the year.
Soeriaatmadja shows an almost uniform surface
salinity throughout the eastern half of the strait
and lower salinities along much of the Sumatra

coast. Data used in the present investigation
indicate an increase in salinity from Port Swet-
tenham to the Andaman Sea and a distribution
pattern clearly reflecting the general current flow
to the northwest. Additional measurements ob-
tained by the Research Vessel JEANNE D'ARC
during late March 1963 (Anonymous, 1963) are
in concurrence with the pattern shown in figure
8. The salinity distribution pattern shown by
Soeriaatmadja (1956) represents an averaging of
data over a five year period which may account
for the discrepancy mentioned here. The pattern
of bottom salinity shows an even more pro-
nounced high salinity wedge extending south-
eastward into the strait. This wedge is more
evenly distributed across the strait at depth than
at the surface. The bottom currents flowing to
the southeast during periods of flood may help to
maintain the position of this wedge. Salinity
profiles also show this wedge to be thicker along
the western margin of the strait and sloping to-
wards the east (fig. 9). The penetration of higher
salinity water from both the Indian Ocean and
the South China Sea is readily apparent in the
axial profile. The influence of large rivers is also
clearly evident from the profiles and areal pat-
terns, particularly in the southern entrance to
the strait. Within the narrow, shallow part of the
strait very little variation is found both laterally
and vertically as a result of mixing.

Temperature. — The average annual variation of
sea surface temperature is approximately 1.5° C.
and is closely related to changes in the air tem-
perature (Wyrtki, 1957, p. 21). Measurements
made during March and April 1961, indicate
that surface temperatures increase from off the
Sumatra coast northeastward across the strait
(fig. 8). This gradient, transverse to the axis of
the strait, probably results from less mixing and
weaker currents along the eastern margin of the
strait.

Bottom temperature measurements during
this same period show a cooler wedge of water
extending into the Malacca Strait from the
Andaman Sea and a general increase in tempera-
ture to the southeast. Isothermal conditions
exist in the narrow portion of the strait because
of vertical mixing.

Temperature profiles across the strait show a
distinct tongue of warmer water along the east-
ern margin of the strait. The southeastward ex-
tension of cooler Indian Ocean water along the
bottom is clearly shown in the axial profile (fig.
9).

An important influence on the sedimentary
environemnt of the north and central portions of
the strait is the large wedge of cool, high salinity
water that extends into the area from the Anda-

SEDIMENTS OF THE MALACCA STRAIT, SOUTHEAST ASIA 111

9 6° 98° 100° 102° 104° 9 6° 98° 100° 102° 104°

96 98 100 102 104 96 98 100 102 104

Fig. 8. — Water temperature and salinity distribution for April 1961.

STATIONS 214 24 20

^40- ---'
o- 2 60- -340

LU D

Q £ 80

~100

A A

STATIONS 25 24 23

20-

Q-«j 60-

Luaj - ~"3<0-.

QE 80-

~100-

SALINITY

(%o)

TEMPERATURE

(°C)

200

— I

Fig. 9. — Water temperature and salinity sections.

112

GEORGE H. KELLER AND ADRIAN F. RICHARDS

man Sea. We believe this wedge is more or less
stationary and is not an actively moving water
mass. The fluctuation of bottom currents with
changes in the tide is probably largely responsi-
ble for the location of this wedge. The signifi-
cance of this water mass with respect to the
sedimentary constituents is discussed in a later
section.

Sediments

The complexity of the depositional environ-
ment in the Malacca Strait is mainly attributed
to the varied sediment source areas provided by
the bordering land provinces and to the oceano-
graphic conditions in and around the strait.
Tectonic activity in the adjacent land masses
has resulted in extensive areas of igneous in-
trusion and associated zones of metamorphic
rock which contribute numerous mineral suites
to the rivers entering both sides of the strait. In
addition, the remains of both planktonic and
benthonic faunas are found in the bottom sedi-
ments. Factors other than source material tend
to further complicate the environment. The
magnitude of river discharge increases con-
siderably during the winter and summer months
as the result of monsoonal rains. Sediment enter-
ing the strait is subjected to a strong annual
northwest current that has superimposed upon it
a semidiurnal tidal current having a flow of
nearly 55 cm/sec. to the southeast, then later
reverses direction to flow northwest at velocities
up to 100 cm/sec. Bottom irregularities account
for isolated patches of coarse or fine material
within areas of nearly uniform sediment type. As
a result of these many factors constituting the
depositional environment, the various sediment
parameters of size, sorting, mineral content and
other constituents undergo short term spatial
and temporal variations of considerable magni-
tude.

One hundred-forty-one grab samples and 76
sediment cores were collected at 155 stations
(fig. 2). An orange-peel bucket sampler (Sumner,
and others, 1914, p. 7) and the Shipek grab
sampler (manufactured by Hydro Products, San
Diego, California) were used to obtain the sur-
face samples. All cores with the exceptions of
those at stations 184, 185, 195, 199, 202, 204 and
209 were collected with a Phleger corer (U. S.
Hydrographic Office, 1955, p. 54-57), and the
remainder was taken with a modified Hydro-
plastic corer using a piston (Richards and Keller,
1961). Core lengths vary from 10 to 349cm,
averaging about 46 cm. Only relatively short
cores were obtained from much of the area be-
cause of the coarse surface sediment and an
underlying stiff clay.

Cores taken in the vicinity of large rivers indi-

cate that the thickness of silt and lutite extends
throughout the sampled interval, while farther
away from the rivers sand overlies finer sedi-
ments. Sediment grain size and calcium car-
bonate content vary considerably with depth
from station to station. In some areas only
minor variations of the sedimentary parameters
are noted with depth, while in others a pro-
nounced change is seen where the core pene-
trates a prominent silty clay layer. The influence
of local conditions is clearly evident from the
sediment cores.

Textures

Prior to this investigation the only available
textural information on the bottom sediments in
the Malacca Strait came from chart notations
which were based on brief visual examinations of
sediment adhering to the tallowed lead used in
early hydrographic surveys. The validity of
many of the notations is questionable because of
the manner in which samples were collected and
the crude field descriptions usually logged by
mariners. Shepard, Emery, and Gould (1949)
compiled these notations into a series of sedi-
ment distribution charts for the east Asiatic
shelf areas. Their study of the Malacca Strait
shows that along the axis the bottom is com-
posed mainly of firm sand and mud and irregular
patches of sand and rock. Mud was found to
predominate in the north-central and north-
eastern sections of the strait.

Based on the surface distribution of various
textural parameters determined during this in-
vestigation, the depositional environment of the
Malacca Strait can be divided into four distinct
regions (fig. 10), if local irregularities are ignored:
north of Penang Island (Northern); Penang
Island to Port Swettenham (Central); Port
Swettenham to Malacca ("Narrows"); and
Malacca to Singapore (Southern).

Sediment samples are assigned to one of four
types based on grain-size analysis data: sand (at
least 90 percent of the sample consisting of mate-
rial coarser than 4 phi); muddy sand (50 to 90
percent sand grains); sandy mud (10 to 50 per-
cent sand) ; and mud (at least 90 percent consists
of material finer than 4 phi). While the distri-
bution of these textures shows a vague agree-
ment with the bottom notations mentioned
above, the more detailed classification used here
affords a better delineation of the sediment
types. Mud deposits are less extensive than indi-
cated by Shepard, Emery, and Gould (1949),
particularly in the Northern region; they are
found mainly in the vicinity of river mouths and
in the Andaman Basin. The present study finds
that muddy sand is the dominant sediment type
in the strait. Sand occurs throughout large

SEDIMENTS OF THE MALACCA STRAIT, SOUTHEAST ASIA

113

areas, but mainly in association with gravel and
large shell fragments along the axis of the strait
as well as along the continental shelf break where
currents tend to winnow out the finer material
(fig. 11). An attempt was made to verify the
presence of rock areas reported by Shepard,
Emery, and Gould (1949), but none were found.
Vertical variation of the textural properties
appears to be minor. Most cores penetrating
only the Recent surface sediments show little
variation whereas those reaching the underlying
stiff clay reveal a distinct change in textural
properties. Grain size decreases markedly and
sorting generally is poorer below this contact.
Within the clay the properties again are re-
markably uniform. A 3.84 m core from the
southern portion of the Andaman Basin shows
little change in the textural parameters with
depth.

Color. — Color of the grab samples was deter-
mined at the time of their collection by com-
parison with the Rock-Color Chart devised by
Goddard, et al., (1948). Color determination of
the wet core samples was made later in the
laboratory. Nearly all samples can be approxi-
mated by five main colors if minor transitions
are ignored: grayish olive (10Y4/2), olive gray
(5Y3/2), light olive gray (5Y5/2), dusky yellow
green (5GY5/2), and medium gray (N5). Sedi-
ment colors often indicate different depositional
environments and are commonly related to such

parameters as grain size, sorting, or carbonate
content (Niino and Emery, 1961, p. 749). This
relationship does not appear to hold true for the
Malacca Strait. The grayish-olive material
found over much of the Central and Northern
regions includes sediments varying in texture
from mud to sand. This variation of texture is
also true for the other colors. It appears that in
the Malacca Strait sediment color is not signifi-
cantly indicative of the conditions in the de-
positional environment, but rather of the source
areas from which the sediments are derived. Al-
though there is a lack of samples from the vicin-
ity of river mouths, there is a strong indication
that the observed color patterns are related to
the sediments coming from the rivers (fig. 11).
Each area delineated by a different color is
found to be associated with one or more streams.
An indication of this relationship is noted in
samples collected off the mouth of one river in
the southern portion of the strait, where the
freshly collected river sediment has the same
color as that found further away from shore.

The large extent of grayish-olive material
across the northern portion of the strait is prob-
ably related to the current system in the area.
Examination of samples from north of the strait
and in the Andaman Sea appears to indicate a
less complex pattern as one might expect where
local drainage is not significant.

Median Diameter. — Grain-size analyses were

CLAY

REGIONS OF THE STRAIT

x NORTHERN
o CENTRAL
A "NARROWS"
• SOUTHERN

SAND 7 5 50

SILT

5-
O 4-

z

o

V) 2-

T

Po°° 00000000°7''/'x<, , /

rV§o s0 y~r x \ J

*1 +2 +3 +4 +5 +6 +7
MEAN (0)

Fig. 10. — Delineation of regions in the Malacca Strait based on textural properties as shown by the sand-silt-
clay content and the relationship of sorting to mean grain size.

114

GEORGE H. KELLER AND ADRIAN F. RICHARDS

Oo 96° 9 8 100 102 104 96 98 100 102 104 Qc

o| 1 t-i «t n 1 r 1 | 1 n ^ r\ 1 1 1 o

I \ 1

\

\ COLOR

CZH GRAYISH OLIVE

V EEE3 OLIVE GRAY
\ E3 LT OLIVE GRAY -
.1 1113 DUSKY YELLOW
M GREEN

||\HI MEDIUM GRAY

■ v \vz^r

A

C^^LyT

"'*^N

1 - 1 :

^N. 1

-6

-4"

-2

100* 102°

FlG. 11. — General description of total sediment, color, median diameter and sorting of surface sediments
from the Malacca Strait and adjacent Andaman Sea. Median diameter and sorting are the descriptive measures
of Inman (19521.

made on the whole sample with the coarse and
fine fractions separated by wet-sieving through a
0.062 mm screen. The dried coarse fractions were
further separated by sieving, and the fine frac-
tions pipetted (Krumbein and Pettijohn, 1938,
p. 165). The various measures used in this study
to describe size distribution such as central
tendency, dispersion and skewness follow the
procedure recommended by Inman (1952).
Median diameters and quartiles were picked
from cumulative curves drawn on arithmetic
probability paper.

The effects of currents and debouching rivers
on the depositional environment in the strait are
clearly shown by the grain-size distribution
pattern (fig. 11). Water depth appears to have
little influence on the sediment distribution. Al-
though the southern portion of the strait —
Singapore to Port Swettenham — is a high-
energy environment, it is nearly choked with

fine material from the numerous rivers of Suma-
tra that drain extensive low-lying areas of allu-
vium. Clean coarse to very coarse sands are only
found in the "Narrows," where a constriction
occurs in the strait. In this region the currents
often reach velocities of 140 cm/sec. which is
sufficient for the removal of finer sediment.

The large area of fine-grained sediment north-
west of Penang Island, which extends westward
beyond the axis of the strait, is a clear reflection
of the general current pattern. The northwest
current emerging from the narrow part of the
strait tends to veer away from the Malayan
Peninsula as it approaches the Andaman Sea.
The southeast current from the Indian Ocean
begins to enter the northern portion of the strait
and then turns westward just north of Penang
Island during the southwest monsoon. The re-
sulting large eddy forms an area of lower current
velocities and transport energy between the two

SEDIMENTS OF THE MALACCA STRAIT, SOUTHEAST ASIA

115

westward moving currents (fig. 6). Clay-size
sediments, with minor amounts of fine sand-size
foraminiferal tests, predominate in the deep
Andaman Basin.

The percentages of sand-silt-clay clearly
show the sediment variation in the major regions
of the strait (fig. 10). In the Southern region, the
finer size fraction is largely made up of lutite
rather than silt-size material, indicating that the
rivers in this section of the strait are carrying
more clay- than silt-size particles. Silt, rather
than clay-size material, comprises the bulk of the
fine sediment in the Northern region. The South-
ern region shows the effect of local drainage,
while the Northern is more indicative of an
environment farther from land in which the
textural properties are related to the energy
level rather than the sediment supply. The
greater concentration of sand in the "Narrows,"
relative to all other regions, is indicative of the
strong currents found there. The fine detrital
sediment from local drainage into the "Narrows"
apparently moves northward along the coast and
does not reach the axial portion of the strait.

Sorting. — Sorting is poor off the coast of Suma-
tra particularly in the Southern region as a
result of the sediment supply from rivers (fig.
11) Low sorting values characterize the sedi-
ments along the axis of the strait where the
strong current and less fluvial material result in
better sorting. Isolated patches of poorer or
better sorted sediments can be related to topo-
graphic depressions and elevations, respectively.
Poor sorting in the vicinity of Penang Island is
probably related to the current pattern dis-
cussed above.

Sorting related to mean grain-size shows the
distinct characteristics of the four major en-
vironments in the strait. Coarse sands are the
best sorted whereas the fine silts are the poorest
Sorting isopleths reveal a marked similarity to
those of the median diameter, especially in the
Southern region.

Skewness. — A study of the areal distribution of
skewness shows very little relationship to other
measured parameters. Coarse and fine (negative
and positive) skewness is found in both areas of
sand- and clay-size material. The sands of the
"Narrows" and the Central region of the strait
are largely near-symmetrical to strongly coarse-
skewed while in the Northern and Southern
regions finely-skewed sediments predominate as
the quantity of fine silt- and clay-size material
increases. Sediments are fine-skewed, in the
vicinity of some rivers although near others they
are coarse-skewed even though the median di-
ameters are similar. Skewness plotted against

mean grain size indicates a general correlation
between these two parameters and does suggest
a sinusoidal relationship as has been reported by
Folk and Ward (1957, p. 19) and Hubert (1964,
p. 777). However, since this correlation is not
more diagnostic than median diameter or sort-
ing, it is neither illustrated nor further discussed.

Organic Constituents

Calcium Carbonate.— The fraction of the whole
sample that is soluble in dilute hydrochloric
acid is primarily organic material in the form of
shell fragments and foraminiferal tests. Values
are reported here as calcium carbonate, al-
though minor amounts of magnesium carbonate
may also be present. Acid-soluble fractions range
from 0 to 69 percent. Low concentrations (trace
to 17 percent) occur throughout most of the
strait. Two exceptions occur in the Central re-
gion, where shells tend to concentrate on banks
and to the north, where there are more shell
fragments and foraminiferal tests (fig. 12).

Areal distribution of calcium carbonate bears
little or no relationship to any of the textural
parameters because of its biogenic origin. The
lowest percentages are found in the narrow por-
tion of the strait where calcium carbonate may
be completely absent, possibly because of strong
currents that break-up and winnow-out shell
fragments. Low calcium carbonate percentages
in the vicinity of large rivers results from the
high sedimentation rate of fine material that is
detrimental to the living habits of shell-con-
tributing organisms. The carbonate concentra-
tion increases both to the south of the strait and
to the north into the Andaman Basin where the
calcium carbonate fraction consists mainly of
foraminiferal tests.

Although shells and foraminiferal tests are
present in most samples, they constitute only a
small percent of the entire sample and are often
masked by large quantities of fine detrital mate-
rial.

Shells. — The areal distribution of both shell
fragments and foraminiferal tests (fig. 12) re-
sembles that of total calcium carbonate: each is
absent in the "Narrows," most abundant in the
Northern and Southern regions, and of inter-
mediate abundance in the Central portion except
for two isolated areas with relatively high con-
centrations. Mollusks contribute the major por-
tion of the shell material; pelecypods are the
predominant forms in shallow water and gastro-
pods are more commonly found in deeper water.

Foraminifera. — Foraminiferal tests are particu-
larly abundant in the Andaman Sea and in the
Northern region; their numbers decrease to the

116

GEORGE H. KELLER AND ADRIAN F. RICHARDS

96 98 100 102 104

96 98 100 102 104

Fig. 12. — Surface distribution of calcium carbonate, shell, foraminifera, and organic
carbon, Malacca Strait and adjacent Andaman Sea.

southeast. The shelf break is a distinct divide
between the high-test concentrations in the
Andaman Sea and the relatively low abundance
in the strait. The tongue-like area of forami-
niferal tests projecting into the strait from
the Andaman Sea is attributed to the environ-
ment caused by the wedge of cool, high salinity
bottom water that also extends from the Anda-
man Sea into this region (fig. 8).

Organic Matter. — Duplicate organic carbon de-
terminations were made using the potassium
dichromate-ferrous ammonium sulfate titration
method of Allison (1935). The values obtained
are only approximate because this method is
based on the assumption that all organic matter
in sediments is at the same state of oxidation.
Percentages of organic carbon range from 0.01 to
2.71. The lowest concentrations are found in the
area of maximum current velocity between Port
Swettenham and Malacca (fig. 12). A fair cor-

relation exists between the areal distribution of
grain size and organic carbon; low organic car-
bon normally is associated with the coarser
sediments. High organic carbon content along
portions of the coastal margin reflects the in-
fluence of river discharge. This is particularly
evident in the Southern region and north of
Penang Island. Isolated areas of relatively high
organic carbon also occur where organic matter
has collected in topographic depressions. A pro-
nounced deficiency of organic carbon occurs
along the shelf break where the currents tend to
winnow away the finer material. Organic carbon
percentage increases beyond the continental
shelf towards the Andaman Basin.

Nitrogen content was determined for 77 sur-
face samples using a modified micro-Kjeldahl
method (Henwood and Garey, 1955). Values
ranging from a trace to 0.235 percent, all of
which is considered to be organic nitrogen. Areal
distribution of nitrogen content throughout

SEDIMENTS OF THE MALACCA STRAIT, SOUTHEAST ASIA

117

8

0

100

102

Fig. 13. — Surface distribution of nitrogen and carbon to nitrogen ratio,
Malacca Strait and adjacent Andaman Sea.

104

ou

most of the strait is usually less than 0.050 per-
cent (fig. 13). Lowest values occur in the coarse
to medium sands in the "Narrows" and in the
sands along the shelf break. A noticeable in-
crease in nitrogen is observed in the Southern
region along the Sumatra coast where organic-
rich sediments are deposited from nearby rivers.
Values to the north increase in areas of finer
sediment with the highest percentages occurring
in the Andaman Basin.

In the Malacca Strait, C/N ratios vary be-
tween 0.4 and 36.9; smaller values occur in the
"Narrows" (fig. 13). High ratios are found along
the Sumatra coast and in the vicinity of Penang
Island indicating abundant drainage from
nearby streams of terrestrial humus, such as
peat, which is often found disseminated through-
out the finer sediments. Carbon-nitrogen ratios
of 56 samples give a mean of 15.4, which is
slightly higher than the mean of 13 found for the
Gulf of Thailand sediments (Emery and Niino,
1963, p. 548) and considerably higher than the
8.4 reported by Trask (1932, p. 21) for all Re-
cent marine sediments. The large mean found in
the Malacca Strait is attributed to the high in-
flux of high-carbon organic matter from rivers
draining areas of heavy vegetation.

Although organic carbon and nitrogen content
is indicative of the amount of total organic
matter present in the sediment, there is no con-
sensus as to the proportion of these constituents
in organic matter (Bader, 1954; Kaplan and
Rittenberg, 1963, p. 593). For this reason, no
attempt is made here to determine the total
organic content, but only to show its relative
concentration by means of the organic carbon
and nitrogen content.

Sponge spicules are the dominant siliceous
organic remains in these samples. They are most
abundant in the Central and Northern regions
where, however, they constitute only a very
small percentage of the whole sample. Minor
amounts of radiolaria are found in most samples;
larger numbers occur in samples from the Anda-
man Sea and the Northern region.

Vertical variation of organic constituents is
only significant in core samples that penetrate
the stiff, silty clay underlying the coarser Recent
sediments. A noticeably lower calcium carbonate
content is almost always found in the underlying
clay. Organic carbon and nitrogen content is
generally higher in the clay than in the overlying
sediment in most areas, except in the Southern
region where the reverse is true. Variation of the
C/N ratio between the sediment above the clay
and in the clay may amount to a factor of four.

Variation of the sedimentary parameters with
depth is often masked or completely altered by
burrowing organisms. Burrow holes or passages
are readily apparent in many cores owing to the
sharp contacts between the coarse material fill-
ing the passages and the finer surrounding sedi-
ment. Organisms appear to have dragged the
coarser surface sediment downward in some
areas. In others, they have apparently con-
centrated the quartz and volcanic glass shards
from the finer sediment through which they are
burrowed. Mottling, which is commonly an indi-
cation of organism activity is quite noticeable in
the finer sediments of the Andaman Basin, but is
not observed in cores taken from the strait. The
presence or absence of mottling is attributed to
the different forms of burrowing organisms and
the types of sediment found in the two areas.

118

GEORGE H. KELLER AND ADRIAN F. RICHARDS

Table 1

. — Mineralogy

of the

coarse

frac

ion from various

regions

9/ </*e Malaccc

-SI

raj7

a

nd the

southern part of the Andaman B

isin

Percent Light Mineral Fraction

Percent Heavy

Mineral Fraction

rt

e
'3

E

c

0

<

c

c

u

rt

(5

°|

a

.- CD

a;

c

*rt

CD

g

V

<D

<u

*

t,' *~"

o

rt

o
c/5

rt

a

o

6

o
rt

o

S

rt

CO

O

6

a Jj

J to

£8

o

c e

CD CO

E «

2
a
E

<

E
o

c
o

N

c
rt

o

O

s

'2

rt

(/}

15

CD
O
O

O

£8

c

1

44

tr

4

38

14

99.6

60

17

7

3

2

1

3

5

2

0.4

<u c:

3

tr

tr

2

98

99.7

74

19

4

1

2

0.3

J3 O

6

28

tr

tr

18

36

18

98.2

1

98

1.8

i'8

9

64

tr

17

17

2

97.9

100

2.1

°£

179

93

6

1

tr

99.8

35

37

4

6

18

tr

0.2

6

11

100

tr

97.8

3

9

1

1

86

2.2

5

47

98

tr

2

98.8

5

10

2

tr

83

1.2

12

88

2

1

8

1

99.6

32

31

3

7

6

19

2

0.4

16

47

2

tr

7

16

28

96.0

100

4.0

18

1

tr

tr

20

36

43

99.7

17

4

39

7

6

2

1

16

5

5

0.3

ri ^

20

3

tr

1

93

2

99.7

20

12

22

3

7

15

12

5

4

0.3

Sh o

184

86

8

2

2

2

tr

98.9

44

33

4

6

3

7

1

2

1 .1

c'wi

188

73

5

13

7

tr

2

99.0

19

42

5

4

2

20

7

1

1 .0

ua

196

43

1

1

IS

32

7

99.7

tr

tr

tr

99

tr

0.3

c

23

77

3

tr

20

99.7

30

11

3

11

39

6

0.3

5.2

24

66

3

tr

tr

21

99.4

31

2

24

9

6

24

4

0.6

25

92

S

tr

tr

3

99.4

26

16

24

6

6

18

4

0.6

1 •»

214

94

3

tr

3

99.7

32

19

6

6

4

1

27

5

0.3

Is

211

5

S

90

99.7

31

15

23

4

2

23

2

0.3

26

5

2

93

98.9

7

24

25

8

1

30

5

1.1

•O c/i

208

31

3

tr

1

65

97.6

5

8

50

22

9

2

1

4

2.4

<m

209

5

95

96.6

4

89

3

1

3

3.4

* Glauconite, volcanic glass

** Allanite, Andalusite, Cassiterite, Corundum, Epidote. Monazite, Muscovite, Pyrite. Sillimanite, Titanite, Topaz

Disturbances due to organisms are observed to a
depth of 84 cm in cores from the strait, while in
the soft sediments of the Andaman Basin, mottl-
ing attributed to organisms is found as deep as 2
m below the sea floor.

Detrital Constituents

Separation of the sand-size particles into light
and heavy mineral fractions was made by a
heavy liquid technique using tetrabromoethane
(specific gravity 2.96). The light minerals were
identified by the staining technique described by
Laniz, Stevens, and Norman (1964), and ex-
amination under a binocular microscope. Heavy
mineral fractions were weighed and their weight
percent in the total sand sample determined.
Identification of the mineral constituents was
made using a petrographic microscope. The re-
sults of this examination are given in table 1.

Light Minerals. — Light minerals dominate the
coarse sediment fraction (75 to 100 percent by
weight) and are generally concentrated in the
sands along the axis of the strait. Quartz com-
prises over 50 percent of the light fraction in all
but a few samples and makes up the entire
sample in the high energy zone between Port

Swettenham and Malacca. An area of low quartz
content (1 to 35 percent) and high glauconite,
volcanic glass and shell fragments extends across
the strait from Belawan to Penang Island.
Isolated areas of low quartz content sediments
are observed in the Northern region in topo-
graphic depressions where shell material pre-
dominates. Quartz content is also low in the
southern entrance to the strait (trace to 44 per-
cent).

Orthoclase is generally more abundant than
plagioclase; concentrations of a trace to 9 per-
cent are found in the Central and Northern re-
gions, and a trace in the Southern region. Trace-
able amounts of plagioclase occur throughout
the strait with a few samples along the Malay-
sian coast, north of Port Swettenham, containing
as much as 5 to 13 percent. The other few sam-
ples in which plagioclase dominates over ortho-
clase occur in low energy environments either
close to shore or in topographic depressions. No
distinct distribution pattern can be detected for
either of the feldspars as a result of the numerous
rivers and the different areas drained by these
rivers.

Heavy Minerals. — With a few exceptions, the

SEDIMENTS OF THE MALACCA STRAIT, SOUTHEAST ASIA

119

heavy mineral assemblages present are composed
of stable to moderately stable minerals derived
from intermediate to silicic igneous rocks and
from associated crystalline metamorphic rocks.
These rocks are abundant in both the bordering
land provinces and the islands at the southern
entrance to the strait. However, in view of the
lack of information concerning coastal detritals,
only generalizations can be made about prove-
nance.

A few minerals dominate the assemblages in
most instances. Opaques consisting of leucoxene,
ilmenite and magnetite constitute large percent-
ages of the majority of the samples. It is be-
lieved that most of the black opaques are ilmen-
ite rather than magnetite because of the ob-
served partial alteration to leucoxene of many
grains of similar appearance. The percent of
leucoxene compared to ilmenite is high in two
areas: the extreme southern end of the strait and
a large area occupying the north-central axial
portion of the strait and extending beyond the
shelf into the Andaman Sea. This relationship is
reversed towards the coastal margins and south-
ward from the Northern region of the strait.

Biotite is concentrated in an elongate area
paralleling the northern Sumatra coast and in
another area beyond the shelf in the Andaman
Sea. Large hexagonal books of biotite are found
to make up an entire heavy fraction of a surface
sample collected in this area off the Sumatra
coast. These concentrations indicate an adjacent
source, possibly entering from the nearby Asa-
han River of Sumatra. The northwesterly cur-
rent in the strait may account for the northward
distribution of the biotite along the coast and
into the Andaman Basin.

Amphiboles and staurolite occur in high per-
centages in most samples from the northern and
central portions of the strait. The largest quan-
tity of staurolite is found in the north while
amphiboles are concentrated in the west-central
strait area; a source for both minerals in the
crystalline schists of Sumatra is suggested.
Tourmaline is usually found in relatively high
concentrations (3 to 13 percent) in association
with staurolite. Both minerals are also abundant
in the vicinity of the Rokan River, indicating a
possible source in the metamorphics of the
Barisan Range of Sumatra.

Heavy fractions from the "Narrows" are
composed almost entirely of siderite or goethite.
A few of these grains contain an olive-brown
rounded aggregate of minute crystals, while
others have a similar oval shape with colors
varying from yellow-brown and brownish-red to
brick-red, and display an earthy surface appear-
ance. Chemical tests on selected grains of the
olive-brown material show it to be siderite. X-

ray analysis of these various grains indicate
siderite in some cases and goethite in others al-
though the outward appearance of the grains is
similar. Alteration of the siderite appears to be
the answer, with oxidation of the ferrous iron
and leaching of carbonate. Partial alteration has
masked the surface of some grains and gone to
completion in the crystallization of goethite in
the other samples.

Heavy-mineral weight percentages of the total
sand-size fraction are low, generally less than
one percent, and correlate with the relative
abundance of medium to fine sand. Heavy min-
eral concentrations greater than 1 percent are
confined, with few exceptions, to those samples
in which a single mineral comprises all or nearly
all of the heavy-mineral fraction.

Van Baren and Kiel (1950) divided the sedi-
ments of the Malacca Strait into three distinct
groups based on the mineral suites from 18
samples. Their results are not corroborated by
the present study (table 1).

In an environment like the Malacca Strait,
with its large number of debouching rivers and
varied sediment source areas, it is not believed
that distinct mineral suites can be delineated for
the major portion of the area. Some distinction is
possible, however, in the Northern region where
the effect of reduced local drainage is shown by
the distribution of the amphibole-staurolite-
tourmaline suite that does appear to have a def-
inite distribution.

Clay Minerals. — The clay minerals were identi-
fied in 54 surface samples as well as in selected
intervals from four cores by means of X-ray
diffraction.

The less-than-two micron fraction was taken
from samples which had been treated with
sodium hexametaphosphate and used for the
grain-size analyses. Two glass slides of oriented
aggregate were prepared from each sample. Each
set of slides was X-rayed (Cu Ka radiation) after
drying at room temperature and again after
solvation in ethylene glycol. Additional X-ray
analyses were obtained from selected samples
treated in warm 8 N HC1.

The clay minerals observed in this study —
montmorillonite, kaolinite, undifferentiated
mixed-layer minerals, illite and chlorite — were
identified by their characteristic X-ray diffrac-
tion peaks.

Clay minerals of the montmorillonite group
were identified by a (001) peak reflection which
shifts to 17A after ethylene glycol treatment.
However, this could be a true montmorillonite or
an illite and/or chlorite which has been stripped
of its inter-layer cations and now absorbs two
layers of ethylene glycol (Griffin and Goldberg,

120

GEORGE H. KELLER AND ADRIAN F. RICHARDS

UNTREATED

3.6 5 7 A

CENTRAL

ETHYLENE GLYCOL

3.6 5 7 A

Sta 195
(0-5 cm)

Sta 204
(40-43c

Sta 209
(1-3cm

Sta 179
(0-13 cm)

Sta 184
(0-6 cm)

Fig. 14. — Typical X-ray diffraction patterns of clays from various parts of the Malacca Strait

and adjacent Andaman Sea.

1963). Diffraction peaks at 7 and 3.5k are
attributed to kaolinite. Using a modification of
the procedure described by Brindley (1961, p.
264), a representative number of the samples
were placed in warm 8 N HC1 for two hours and
then X-rayed to distinguish between kaolinite
and chlorite. Broad, asymmetrical diffraction
peaks varying between 15.5 and 16. 8A after
ethylene glycol solvation characterized the un-
differentiated mixed-layer minerals. These con-
sist of a complex grouping of three-layered min-
erals interlayered with montmorillonite. It ap-
pears that illite may be the significant mineral

mixing with montmorillonite, although chlorite
is undoubtedly also present. Diffraction peaks at
10, 5 and 3.3A which are not affected by ethyl-
ene glycol treatment are defined here as belong-
ing to minerals of the illite group. A diffraction
peak at 4.7A as well as a slight reduction of the
3.5A peak after acid treatment is used as evi-
dence of chlorite. Typical X-ray patterns for
samples from different regions of the Malacca
Strait and Andaman Sea are shown in figure 14.
A semi-quantitative determination was made
utilizing the intensities of the X-ray peaks to
study the clay-mineral distribution. The pro-

SEDIMENTS OF THE MALACCA STRAIT, SOUTHEAST ASIA

121

cedure used is similar to that described by
Johns, Grim, and Bradley (1954), with the ex-
ception that the area of the 1 7 A peak was di-
vided by three rather than four. This procedure
was used by Ferrell (1965, p. 21-22) and found
to be satisfactory. Values are reported in percent
rather than in parts in 10 as a matter of con-
venience. This method, although not truly
quantitative, is suitable for comparison of rela-
tive abundance.

Kaolinite, mixed-layer minerals, illite, mont-
morillonite and chlorite comprise the clay-
mineral assemblage found in the Malacca Strait
in order of decreasing abundance. Kaolinite and
mixed-layer minerals constitute from 60 to 85
percent of the clay minerals present in most
samples. Illite is present in all samples, but is
never found to be the dominant clay mineral.
Montmorillonite is the predominant clay min-
eral at only a few locations and combines with
other minerals to form the mixed-layer minerals

in all other samples. Chlorite was only observed
in trace amounts although it is present in many
samples.

The highest concentrations of kaolinite are
observed in the shallow areas along the Sumatra
coast and in the southern portion of the strait
(fig. 15). Kaolinite is common in all samples as
might be expected of sediments collected rela-
tively close to land in a tropical climate. The
kaolinite content bears a close relationship to
the isobaths; a decrease in kaolinite occurs with
increasing water depth. Minor concentrations
are found in the Andaman Basin. Undifferenti-
ated mixed-layer minerals constitute the domi-
nant clay minerals in the Northern and deeper
portions of the strait as well as in the southern
part of the Andaman Basin (fig. 15). However,
mixed-layer minerals comprise a large portion of
the clay minerals in most of the samples ana-
lyzed. Montmorillonite is the predominant clay
mineral at only six stations, all of which are

96° 98° 100° 102° 104° 9 6° 98° 100° 102° 104°

\ PREDOMINANT
CLAY MINERALS

i*\ WW KAOLINITE

I | MIXED-LAYER
MONT

\ MIXED -LAYER

i: C/o CLAY FRACTION)

96° 98° 100° 102

104C

100° 102C

Fig. 15. — Surface distribution of the predominant clay minerals, kaolinite, mixed-layer minerals
and illite in the Malacca Strait and adjacent Andaman Sea.

122

GEORGE H. KELLER AND ADRIAN F. RICHARDS

close to shore and generally in the vicinity of
river mouths. The concentration of montmoril-
lonite (46 to 53 percent of the clay minerals) in
the southern portion of the strait decreases
away from the rivers. Drainage through the
Tertiary tuffs on Sumatra provides the probable
source material for these clays. Streams drain-
ing Permian shales on the Malay Peninsula in
the vicinity of Penang and Langkawi Islands may
be responsible for the montmorillonite in this
area. Although illite is present in all the samples
it seldom makes up more than 25 percent of the
clay minerals and usually is much less. The
largest concentrations of illite are found east of
northern Sumatra in an area where sediments
also contain large amounts of biotite. This as-
sociation plus the presence of glauconite, which
is mentioned later, may indicate a diagenic
relationship between these minerals (Burst,
1958, p. 493). The illite and biotite could also
have been transported to the strait from a source
area in Sumatra.

The montmorillonite peak in samples col-
lected progressively farther from land in the
Southern region is observed to lose its character
and assume that of a mixed-layer peak. This
change may be caused by the occurrence of de-
graded illite with the montmorillonite. The pres-
ence of degraded illite can be explained by the
loss of potassium ions during transport in
slightly acid river waters. As the rivers empty
into the strait these ions are normally regained
(Bradley, 1953). Although degraded illite may
fix potassium from the sea water in a matter of
days as suggested by Weaver (1958), it might be
a long time before the illite has completely
achieved its characteristic 10A peak. In such a
situation, the presence of a certain amount of
degraded and crystalline illite would cause the
montmorillonite peak to broaden and assume
the characteristics of a mixed-layer mineral.
The amount of magnesium ion fixation by
chlorite could have a similar effect on the mont-
morillonite peak (Grim and Johns, 1954, p. 100).

Four cores ranging in length from 0.8 to 3.5 m
were examined for their clay mineral content,
and, of these, only three showed any discernible
variation with depth. The predominant minerals
are the same as those found in the surface sedi-
ments.

The percentage of mixed-layer minerals in-
creases slightly and illite decreases with depth in
the three cores. It is generally reported in the
literature that there is little or no variation in
clay mineral composition with depth for cores in
Recent marine sediments (Emery, 1960, p. 231;
Griffin and Goldberg, 1963, p. 739; Heezen, and
others, 1965, p. 5821). This unusual occurrence
found in the Malacca Strait may possibly be ex-

plained by an upward migration of slightly acidic
water that strips potassium ions from the illite.
This would result in a decrease in 10A illite and
an increase in the mixed-layer minerals. Each of
the cores displaying this change with depth con-
tains some finely disseminated peat which would
reduce the pH of the interstitial water.

Volcanic Fraction. — Volcanic material found in
the sand-size fraction consists mainly of shards of
volcanic glass and pieces of pumice. Shards are
colorless and triangular, plate-like or irregular in
shape and often have inclusions of oval-shaped
bubbles. The refractive index is 1.512 + 0.002.
Glass of the same refractive index consists ap-
proximately of 67 percent silica and is derived
from an andesitic magma according to George
(1924).

Glass shards occur in almost all the samples in
amounts varying from a trace to 25 percent of
the light fraction. The highest concentration is
found along the Sumatra margin of the strait
between the Asahan River and Belawan, where
percentages range from 10 to 25.

Volcanic glass is also quite common in many
cores where its presence is clearly noted by the
numerous layers of glass shards. These layers are
very distinct owing to their light gray color in
contrast to the darker surrounding material. Ash
layers vary in thickness from 0.1 to 4 cm, but are
usually less than 0.25 cm thick (fig. 16).

There appears to be a correlation of an ash
layer, occurring at a depth of 30 cm below the
sea floor, among three adjacent cores from the
Central region of the strait. This may be a coin-
cidence because it is not seen in nearby cores,
and other ash layers cannot be traced from one
core to another. A 4 cm-thick ash layer found in
the bottom of core 103 probably represents a
local deposit in a former topographic depression
because it is not observed in adjacent cores less
than 3.2 km away. Shards are also concentrated
in irregular bands or pockets that are probably
the result of burrowing organisms.

Light gray layers similar in appearance to the
ash are observed in certain parts of the strait.
These layers, however, are largely composed of
clean quartz and glass shards of a single grain
size (4 phi). Such is the case west of Singapore
where, upon initial examination, the cores look
much the same as those containing layers of
shards, but on closer scrutiny clean quartz
rather than volcanic glass is found to be the
dominant constituent. Similar occurrences of
such layers are found in a few cores from the
Central region of the strait.

It is believed that the ash layers do not rep-
resent wind-borne tephra erupted from nearby
volcanoes, but rather detrital deposits of ash

SEDIMENTS OF THE MALACCA STRAIT, SOUTHEAST ASIA

123

eroded from the neighboring land areas. Large
ash concentrations off the Sumatra coast in the
vicinity of rivers draining tuff deposits on
Sumatra lends support to this conclusion. The
numerous ash layers in relatively short intervals
of the core section, 25 laminae in a 10 cm in-
terval, may be more indicative of heavy runoff
during the monsoons than volcanic eruptions.
Reworking by currents may also account for the
ash distribution.

Authigenic Constituents

Glauconite is found extensively throughout
the strait occurring mainly as pellets of dark-
green grains and to a lesser extent occupying
the cavities of foraminiferal tests. Percentages of
glauconite range from a trace to 80 percent of
the light fraction with a mean of about 6 percent.
The highest concentrations occur in the Central
region along the Sumatra margin. Only traceable
amounts are found in the coarser sediments be-
tween Malacca and Port Swettenham. No ap-
parent correlation is observed between the occur-
rence of glauconite and grain size because of the
way in which the pellets are deposited. Glau-
conite pellets become part of the sediment as the
foraminiferal tests surrounding the pellets are
broken. Most of the broken tests are winnowed
away leaving the pellets behind. Excluding the
narrow portion of the strait in the vicinity of
Port Swettenham, there appears to be a general
inverse relationship between the presence of
foraminifera and glauconite as a result of the
removal of the broken tests.

Pyrite occurs in the form of internal molds of
foraminifera and radiolaria and is frequently
found in the surface sediments. The possibility
that reducing conditions may exist in the micro-
environment within cavities of various tests and
shells may be significant in explaining the origin
of such authigenic minerals as pyrite and glauco-
nite in an environment often oversaturated with
oxygen. Emery (1960, p. 267) reported similar
occurrences of pyrite off the coast of southern
California. No indication of phosphorite or
manganese oxide coating of gravel was observed
in any of the samples.

PLEISTOCENE-RECENT HISTORY

The Malacca Strait is a part of the Sunda
Shelf, which is one of the major continental
shelf areas of the world. This shelf, unlike most
others, is believed to have been a peneplain
formed in the Late Tertiary or Early Pleistocene.
The peneplain was further denuded during the
Pleistocene, when sea level was lowered between
73 and 91 m (Dickerson, 1941, p. 30). The
present Sunda Shelf was drowned and the
Malacca Strait, as it is essentially found today,
came into existence as a result of the postglacial
eustatic rise of sea level. Molengraaff (1921, p.
100) suggests that no crustal movement was in-
volved in the submergence. The rather uniform
depth of the shelf, between 55 and 64 m, and the
numerous drowned river valleys tend to support
this hypothesis. Additional evidence from a
study of Pleistocene geology on Sumatra and
Java by Smit Sibinga (1952) lends further cre-

FiG. 16. — Volcanic ash layers in the cores from the Central region of the Malacca Strait.

124

GEORGE H. KELLER AND ADRIAN F. RICHARDS

CaCO-
1?3P.qp

Organic Carbon
4 . .8 . 12. 1.6,2,0.

Nitrogen

(%)

.03 ,.0,5,

RECENT i

PLEISTOCENE

STATION 112

Md

(0)
34.5 6 78 910

Sand

Silt
Clay

Sorting

(%)

(0)

0

10

_20

40-
50-
60-

STATION 130

CaC03

0 100 2 3 4 0 20

i i ,i

Organic
Carbon

1.2 16 2.0

"i

tffiijjj

FfHIJIJ
tipJJJJl
fc=iJH/J

/

T

i

~T

RECENT

T ■ T

l\PLEISTOCENEi

\ \

( 1 SUBSAMPLE INTERVAL)

Nitrogen

(%)
05 07

i i I

Fig. 17.- — Vertical variation of textural and chemical properties in cores penetrating the Pleistocene sediments.

dence to the hypothesis. A carbon-14 date for
peat found in a core approximately 24 km north
of Port Swettenham at a depth of 26.5 m below
sea level gave an age of 10,000 + 200 years B. P.
(U. S. Geological Survey No. VV-1675). This date
is on the curve established by Curray (1960) and
Shepard (1960) for the rise of sea level since the
Late Quaternary. It is evident that the Malacca
Strait has remained stable during the last 10,000
years.

Evidence of the Pleistocene-Recent history is
seen in many core samples as well as in the
bathymetry of the strait. Numerous cores ex-
hibit two distinct sediment layers: an upper layer
of sand with shell debris and an underlying stra-
tum of stiff, silty clay. The sand is characteristi-
cally dirty and composed of medium to fine sand-
size particles of quartz, shell debris, glauconite
and rock fragments. The underlying gray silty
clay is relatively clean and contains only minor
or traceable amounts of shell fragments, glau-
conite and rock fragments. Peat is usually dis-
seminated throughout the silty clay both as fine
particles and as large fragments of wood or plant
roots. The lower 15 cm of core 185, from just

north of Port Swettenham, was entirely com-
posed of peat. In many cores the silty clay sec-
tion is relatively stiff and appears to have be-
come indurated before the overlying sand was
deposited. The pronounced change in sedimen-
tary characteristics across this contact is quite
evident from figure 17.

These two types of sediment are indicative of
two distinctly different environments of deposi-
tion. The silty clay with its abundant peat and
few shell fragments is similar to that found in
some tidal flat or estuarine areas today (Hantz
schel, 1939, p. 197; Guilcher, 1963, p. 624; Van
Straaten, 1963, p. 164). The overlying coarse
sediments represent the present high energy
environment of the strait. We believe that the
stiff clay represents a Late Pleistocene tidal flat
or estuarine deposit, which is in agreement with
the Pleistocene paleogeography proposed by
Molengraaff (1921) for this area. The stiffness of
the clay indicates that it may have been partially
indurated as a result of being subaerially exposed
for some period prior to its submergence. This
material was probably eroded during the period
of exposure and the drainage pattern now ob-

SEDIMENTS OF THE MALACCA STRAIT, SOUTHEAST ASIA

125

served in the present bottom topography is a re-
sult of that erosion. As a result of the radiocar-
bon date mentioned earlier, from a peat sample
taken 23.5 cm below the sand-clay contact, it
seems feasible that the stiff, silty clay is of Late
Pleistocene age.

The depth at which this Pleistocene surface oc-
curs below the present strait floor differs con-
siderably depending on the local rates of de-
position in the strait. The surface is often found
at depths of less than 20 cm in the axial portion
of the strait, while between Langkawi and
Penang Islands it occurs at a depth of 41 cm.
Cores that do not display any indication of this
surface may not have penetrated deep enough,
or the surface may have been obliterated by
burrowing organisms.

CONCLUSIONS

The sediments in the Malacca Strait are
largely derived from the adjacent land provinces
of Sumatra and the Malay Peninsula where the
geologic condition varies considerably. Minor
amounts of sediment are also transported into
the strait from the islands south of Singapore.

Although the Malacca Strait is in a part of the
world where monsoons strongly influence clima-
tological and oceanographic conditions, little
direct effect is observed because of the protection
provided by the bordering land masses. The cur-
rent circulation is mainly caused by the move-
ment of water into the strait from the South
China and Java Seas and to a lesser extent from
the Andaman Sea. The depositional environment
is primarily influenced by the currents and the
numerous streams draining into the strait. Al-
though rivers account for the movement of most
sediment into the strait, as well as along its
coastal margins in some sections, a northwest
current provides the mode of transport for the
sediments within the strait. Deposition of bio-
genic constituents, however, shows a greater
dependence on the physical and chemical char-
acteristics of the overlying water mass. This is
evident from the similarity of the foraminifera
distribution with that of a pronounced wedge of
cool-temperature, high-salinity water extending
into the strait from the Andaman Sea.

The depositional environment of the Malacca
Strait can be divided into four regions based on
the textural parameters of the surface sediments:
Northern, Central, "Narrows" and Southern.
Each region also shows a difference in oceano-
graphic conditions as well as in the influence of
streams on the environment. Sediment distribu-
tion in the Northern region is almost entirely
controlled by currents; in the Southern region
both currents and local streams are the con-
trolling factors.

The significance of sediment color in the
Malacca Strait is not as an indication of the con-
ditions in the depositional environment, as is
often reported in the literature (Emery, 1960,
p. 234; Niino and Emery, 1961, p. 749), but
rather as a property of the sediments entering
the strait. Various streams are depositing sedi-
ments of different colors, which is clearly shown
by the sediment color distribution in the strait.

Organic matter is generally concentrated in
the vicinity of river mouths or in the deep
Andaman Sea. High carbon to nitrogen ratios
near debouching rivers are an indication of the
greater abundance of high-carbon, terrestrial
organic matter being deposited in these areas.
Peat is believed to be the source of the organics.
The lower ratios in the Andaman Sea sediments
are attributed to the partly decomposed remains
of marine plankton and benthos.

The mineralogy of sediments in the strait is
complex because of the numerous streams drain-
ing the varied geology on the Malay Peninsula
and Sumatra. Delineation of mineral suites is
particularly difficult because of the interfinger-
ing of the sediments from these streams. Only
along the western margin of the Northern and
Central regions of the strait can any inference be
made as to sediment provenance based on the
mineral suites. The concentration of amphiboles,
staurolite, tourmaline and biotite in these re-
gions suggests a possible source in the metamor-
phics of the Barisan Range of Sumatra. Sumatra
rather than the Malay Peninsula appears to be
the predominant source area for the detrital
sediments in the strait.

Kaolinite is the predominant clay mineral
found in the strait, as would be expected in
marine sediments derived from land masses in a
tropical climate. Undifferentiated mixed-layer
clay minerals are abundant in the deeper areas
as the result of what appears to be the combin-
ing of montmorillonite with degraded illite and
possibly chlorite. Most of the clay minerals are
detrital although some of the illite may be a
diagenic product of glauconite.

Slight increases in the percent of mixed-layer
minerals and decreases in the illite content with
depth are believed to be due to leaching of
potassium ions from the illite in an inferred acid-
ic environment. The degraded illite then com-
bines with the montmorillonite to form the
mixed-layer minerals.

Although volcanic glass is found throughout
most of the strait no ash layers can be traced
over any great distance. The layers are normally
very thin (0.25 cm) and are repeated frequently
in a cored interval. This evidence, together with
an increase in ash concentration off some of the
rivers, indicates that most of these sediments are

126 GEORGE H. KELLER AND ADRIAN F. RICHARDS

carried into the strait by streams draining tuff acknowledgements
deposits rather than being wind-borne material

from nearby volcanoes. The numerous layers in We gratefully acknowledge the assistance of

a short core interval may possibly reflect sea- Mr. Richard Stewart and others of the Oceano-

sonal changes in runoff. The lack of eolian ash graphic Division of the U. S. Naval Oceano-

deposits is attributed to the strong water cur- graphic Office who collected much of the field data

rents that do not allow the ash to settle into dis- and performed many of the textural and chemi-

tinct layers as is found in some deep-sea sedi- cal analyses. Thanks are also extended to Dr.

ments. Kelvin Rodolfo for carrying out the nitrogen

Authigenic minerals such as glauconite and analyses and to Dr. Meyer Rubin for the car-

pyrite are common constituents of the sediments bon-14 determination. Discussions with Dr. J. B.

in the strait. These minerals are all believed to Alexander of the Malaysian Geological Survey

have formed within the tests of foraminifera and proved most helpful during this investigation,

radiolaria and do not, by their presence, suggest We acknowledge with appreciation the assis-

a reducing environment. tance of Professors A. V. Carozzi, R. E. Olson,

The Sunda Shelf, of which the Malacca M. F. Wahl and H. R. Wanless of the University

Strait is a part, is, believed by many investiga- of Illinois who critically read and made helpful

tors to represent a Tertiary-Quaternary pene- suggestions to this manuscript

plain drowned as a result of an eustatic rise of sea The U. S. Naval Oceanographic Office, the

level at the close of the last glacial period. Cores U. S. Coast and Geodetic Survey and the Insti-

examined during this study support this hy- tute for Oceanography of the Environmental

pothesis and further indicate that the area now Science Services Administration provided both

occupied by the strait was a tidal flat or estuarine the oceanographic vessels and funds for this

environment prior to submergence. study.

REFERENCES

Alexander, J. B., 1959, Geology and palaeontology in Malaya: Nature, v. 183, p. 230-232.

Allison, L. E., 1935, Organic carbon by reduction of chromic acid: Soil Sci., v. 40, p. 311-320.

Anonymous, 1957, Oceanographic station list, 1957: Marine Res. Indonesia, no. 4, p. 26-29.

Anonymous, 1963, Observations Oceanographique, Croiseur-Ecole d'Application "Jeanne d'Arc," Cahiers

Oceanographiques, XV Annee no. 9, p. 666-671.
Bader, R. G., 1954, Use of factors for converting carbon or nitrogen to total sedimentary organics: Science, v.

120, p. 709-910.
Baren, F. A. Van, and Kiel, H., 1950, Contribution to the sedimentary petrology of the Sunda Shelf: Jour.

Sedimentary Petrology, v. 20, p. 185-213.
Bemmelen, R. W. Van, 1949, The geology of Indonesia: Vol. I A, General geology of Indonesia and adjacent

archipelagoes. The Hague, Gov't Printing Office, 732 p.
Bradley, W. F., 1953, Analysis of mixed-layer clay mineral structures: Analytical Chem., v. 25, no. 5, p. 727-

730.
Brindley, G. W., 1961, Chlorite minerals, in The X-ray Identification and Crystal Structures of Clay Min-
erals: Brown, G, ed., Mineralogical Society, London, p. 242-296.
Burst, J. F., 1958, Mineral heterogeneity in "glauconite" pellets: Am. Mineralogist, v. 43, p. 481-497.
Curray, J. R., 1960 Sediments and history of Holocene transgression continental shelf, northwest Gulf of

Mexico, in Recent Sediments, Northwest Gulf of Mexico: Shepaid, F. P., Phleger, F. B. and Andel, Tj. H.

van, eds., Am. Assoc. Petroleum Geologists, Tulsa, Okla., p. 221-266.
Dickerson, R. E., 1941, Molengraaff River: A drowned Pleistocene stream and other Asian evidences bearing

upon the lowering of sea level during the ice age, in Shifting of Sea Floors and Coast Lines. University of

Pennsylvania, Bicentennial Conference, University of Pennsylvania Press, Philadelphia, 30 p.
Emery, K. O., 1960, The sea off Southern California: a modern habitat of petroleum. John Wiley & Sons,

New York, 366 p.
Emery, K. O., and Niino, H., 1963, Sediments of the Gulf of Thailand and adjacent continental shelf: Geol.

Soc. America Bull., v. 74, p. 541-554.
Ferrell, R. E., 1965, Paleoenvironmental significance of clay minerals in the Shell Creek Shale, Muddy

Sandstone, and Thermopolis Shale on the East Flank of the Bighorn Mountains, Wyoming, Unpub. M. S.

Thesis, University of Illinois, 66 p.
Folk R. L., and Ward, W. C, 1957, Brazos River bar: a study in the significance of Grain size parameters:

Jour. Sedimentary Petrology, v. 27, p. 3-26.
George, W. O., 1924, The relation of the physical properties of natural glasses to their chemical composition:

Jour. Geology, v. 32, p. 353-372.
Goddard, E. N., Trask, P. D., De Ford, R. K, Rove, O. N., Singewai.d, J. T., Jr., andOverbeck, R. M.,

1948, Rock-color chart: Washington, D. C, Nat. Research Council Comm. on Rock-color chart, 6 p.

(distributed by The Geological Society of America).
Great Britain Hydrographic Department, 1958, Malacca Strait Pilot, 4th Ed. London, 491 p.
Griffin, J. J., and Goldberg, E. D., 1963, Clay mineral distributions in the Pacific Ocean, in The Sea: Ideas

and Observations on Progress in the Seas. Hill, M. N. ed., v. 3, p. 728-741, Interscience Publishers, New

York.

SEDIMENTS OF THE MALACCA STRAIT, SOUTHEAST ASIA 127

Grim, R. E., and Johns, W. D., 1954, Clay mineral investigation of sediments in the northern Gulf of Mexico,
in Clays and Clay Minerals: Swineford, A., and Plummer, N. V. eds., Nat'l Res. Council Pub. 327, p.
81-103.

Guilcher, A., 1963, Estuaries, deltas, shelf, slope, in The Sea: Ideas and Observations on Progress in the Seas:
Hill, M. N. ed., v. 3., p. 620-654, Interscience Publishers, New York.

Hantzschel, W., 1939, Tidal flat deposits (Wattenschlick), in Recent Marine Sediments: Trask, P. D. ed.,
Am. Assoc. Petroleum Geologists, Tulsa, Okla, p. 195-206.

Heezen, B. C, Xesteroff, W. D., Oberlin, A., and Sabatier, G., 1965, Decouverte d'attapulgite dans les
sediments profonds du golfe dAden et de la mer Rouge: C. R. Acad. Sc. Paris, t. 260, p. 5819-5821.

Hen wood, A., and Garey, R. M., 1955, A modified technique for the Kjeldahl procedure: Hengar Company,
Phila., Pa., 4 p.

Hubert, J. F., 1964, Textural evidence for deposition of many western North Atlantic deep-sea sands by
ocean-bottom currents rather than turbidity currents: Jour. Geology, v. 72, p. 757-785.

Inman, D. L., 1952, Measures for describing the size distribution of sediments: Jour Sedimentary Petrology,
v. 22, p. 125-145.

Johns, W. D., Grim, R. E., and Bradley, W. F., 1954, Quantitative estimations of clay minerals by diffrac-
tion methods: Jour. Sedimentary Petrology, v. 24, p. 242-251.

Kaplan, I. R., and Rittenberg, S. C, 1963, Basin sedimentation and diagenesis, in The Sea: Ideas and
Observations on Progress in the Seas: Hill, M. N. ed., v. 3, p. 583-619. Interscience Publishers, New York.

Krempf, A., and Chevey, P., 1934, The continental shelf of PVench Indo-China and the relationship which
formerly existed between Indo-China and the East Indies: Proc. Fifth Pac. Scien. Congress, Canada 1933,
v. 2, p. '849-852.

Krumbein, W. C, and Pettijohn, F. J., 1938, Manual of sedimentary petrography: Appleton-Century-
Crofts, New York, 549 p.

Kuenen, PH. H., 1950, Marine Geology: John Wiley & Sons, New York, 568 p.

Laniz, R. V., Stevens, R. E., and Normon, M. B., 1964, Staining of plagioclase feldspar and other minerals
with F. D. and C. Red. No. 2: U.S. Geol. Survey Prof. Paper 501-B, p. B152-B153.

Molengraaff, G. A. F., 1921, Modern deep-sea research in the East Indian Archipelago: The Geographical
Jour., v. LVII, no. 2, p. 95-121.

Niino, H., and Emery, K. O., 1961, Sediments of shallow portions of East China Sea and South China Sea:
Geol. Soc. America Bull., v. 72, p. 731-762.

Richards, A. F., and Keller, G. H., 1961, A plastic-barrel sediment corer: Deep-Sea Res., v. 8, p. 306-312.

Royal Netherlands Meteorological Institute, 1935, Oceanographic and meteorological observations in
the China Seas and in the western part of the North Pacific Ocean: The Hague, Gov't Printing Office folio
115, 80 p.

Selvarajah, \\, 1962, A study in the North Malacca Strait from salinity determinations: Indo-Pacific Fisheries
Council, 9th Session Proceedings, Section II (1961), p. 1-6.

Shepard, F. P., 1960, Rise of sea level along northwest Gulf of Mexico, in Recent Sediments, Northwest Gulf
of Mexico: Shepard, F. P., Phleger, F. B., and Andel, Tj. H. van, eds., Amer. Assoc, of Petroleum Geolo-
gists, Tulsa, Okla., p. 338-344.

Shepard, F. P., Emery, K. O., and Gould, H. R., 1949, Distribution of sediments on East Asiatic continental
shelf: Univ. of Southern Calif., Allan Hancock Found., Occasional paper 9, 64 p.

Smit Sibinga, G. L., 1952, Interference of glacial eustasy with crustal movements and rhythmic sedimentation
in Java and Sumatra: Geol. en Mijnb., New series, v. 14, p. 220-226.

Soeriaatmadja, R. E., 1956, Surface salinities in the Strait of Malacca: Marine Res. Indonesia, no. 2, p. 27-48.

Straaten, L. M. J. V. Van, 1963, Aspects of the Holocene sedimentation in the Netherlands: Verhand. Kon.
Ned. Geol. Mijnb. Gen., Geol. Ser., v. 21, p. 149-172.

Sumner, F. B., Louderback, G. D., Schmitt, W. L., and Johnson, C. E., 1914, A report upon the physical
conditions in San Francisco Bay, based upon the operations of the United States Fisheries Steamer "Alba-
tross" 1912-1913: Univ. of California Publ. Zool., v. 14, p. 1-198.

Trask, P. D., 1932, Origin and environment of source sediments of petroleum: Gulf Publ. Co., Houston, 323 p.

Umgrove, J. H. F., 1949, Structural history of the East Indies: University Press, Cambridge, England, 63 p.

United Nations Economic Commission for Asia and the Far East, 1961, Geological map of Asia and the
Far East: Calcutta, India, 6 sheets.

U. S. Navy Hvdrographic Office, 1944, Atlas of surface currents, Indian Ocean: Hydrographic Office Pub-
lication No. 566, Washington, 12 charts.

U. S. Navy Hydrographic Office, 1945, Currents in the South China, Java, Celebes, and Sulu Seas: Hydro-
graphic Office Publication No. 236, Washington, 14 p.

U. S. Navy Hydrographic Office, 1951, Sailing directions for the Malacca Strait and Sumatra, 4th ed.:
Hydrographic Office Publication No. 70, Washington, 485 p.

U. S. Navy Hydrographic Office, 1955, Instruction manual for oceanographic observations: Hydrographic
Office Publication No. 607, Washington, 210 p.

Veen, J. Van, 1936, Onderzoekingen in de Hoofden in verband met de gesteldheid der Nederlandsche kust:
'S Gravenhage-Alegemeene Landsdrukkerij, 252 p.

Weaver, C. F., 1958, The effects and geological significance of potsasium "fixation" by expandable clay min-
erals derived from muscovite, biotite, chlorite, and volcanic materials: Am. Mineralogist, v. 43, p. 838-
861.

Wyrtki, K, 1957, Precipitation evapoiation, and energy exchange on the surface of the southeast Asian
waters: Marine Res. Indonesia, no. 3, p. 7-40.

1961, Naga Report v. 2: The University of California, Scripps Institution of Oceanography, La Jolla,

195 p.

Reprinted from JOURNAL OF PALEONTOLOGY Vol. kl, No. 6
The Society of Economic Paleontologists and Mineralogists

21

SOME SPUMELLARIAN RADIOLARIA FROM THE JAVA, PHILIPPINE,

AND MARIANA TRENCHES

HSIN-YI LING and WILLIAM A. ANIKOUCHINE

University of Washington, Seattle

Abstract — Eight species belonging to five genera of patagium-bearing and morphologically
closely related spumellarian Radiolaria were found in three sediment cores from the Java,
Philippine, and Mariana Trenches in the I ndo- Pacific region. These various forms are illus-
trated and discussed with special emphasis on their intraspecific variation in the degree of
patagium development or preservation.

INTRODUCTION

IN 1964, while returning from the International
Indian Ocean Expedition, the USC&GS
Pioneer collected sediment cores from three
trenches in the Indo-Pacific region (Text-fig. 1).
The core locations are (names approved by the
United States Board on Geographic Names) :

Java Trench, lat 6°00' S., long 101°17' E. ; water

depth, 3,380 m.
Philippine Trench, lat 5°25' N., long 127°40'

E. ; water depth, 8,010 m.
Mariana Trench, lat 11°18' N., long 141°57'

E.; water depth, 10,170 m.

During the preliminary examination of the
core samples, the presence of uncommonly rich
radiolarian faunas and abundant fragments of a
diatom, Ethmodiscus rex (Rattray) Hendey,
were noticed. Some Radiolaria in the assemblage
possess highly diversified degrees of patagium.

In a previous paper, Ling (1966) pointed out
an inconsistent degree of patagium found in some
spumellarian Radiolaria genera and discussed in
detail its relationship to ontogenetic develop-
ment and taxonomy. At that time, his observa-
tion was based only on the radiolarian assem-
blage in bottom sediments from the northeast
Pacific Ocean. The finding of a similar feature in

20°

100° E

20°

100° E

Text-fig. / — Geographic locations of three trench cores from the Indo-Pacific region: 33, Java Trench; 35,
Mindanao Trench; 36, Mariana Trench; A — Ehrenberg (1861a), location of Enhreberg's sample from the
Philippine Sea.

1481

1482

HSIN-YI LING AND WILLIAM A. ANIKOUCHINE

50

- 100

X

o
o

150

200

250 «"

JAVA TRENCH CORE

] LUTITE
FS| SILTY STREAKS

- ABSENT

+ PRESENT (SINGLE SPECIMEN)

r RARE (2-5 SPECIMENS)

f FREQUENT (6-10 SPECIMENS)

C COMMON (11-25 SPECIMENS)

Q ABUNDANT FRAGMENTS

Text-fig. 2 — Distribution of some Radiolaria in the
Java Trench core.

the samples from the Indo-Pacific region fur-
nishes additional evidence that such variation in
degree of patagium seems to be neither a rare
phenomenon nor biogeographically significant.
In the present investigation, only those pata-
gium-bearing Radiolaria and the forms that are
morphologically closely related are treated.

ACKNOWLEDGMENTS

The authors are indebted to Captain Harold J.
Seaborg, Seattle Regional Office, U.S. Coast and
Geodetic Survey, and Mr. Ted V. Ryan, Pacific
Oceanographic Laboratory, Environmental Sci-
ence Services Administration, Seattle, Washing-
ton, for their kind permission to use samples
from the trench-sediment cores for the present
study. Special thanks are due Mr. Frederick J.
Collier, Smithsonian Institution, for kindly ar-
ranging the loan of Martin's (1904) type speci-
mens, and to Mr. William R. Riedel, Scripps
Institution of Oceanography, University of
California at San Diego, and Dr. Catherine Anne
Clark Nigrini, Department of Geology, North-
western University, for their constructive advice
and critical reading of the manuscript. The

Plates 189 & 190
are transposed.

authors, however, bear the sole responsibility
for the contents of the paper.

Technical assistance received from Mr.
Donald R. Doyle and Mrs. Shirley J. Patterson
during the preparation of the manuscript was
most helpful.

The investigation was supported financially
by Office of Naval Research contract Nonr 477
(37), project NR 083 012, and National Science
Foundation grant GA-297. The latter grant also
allowed the senior author to examine the radio-

100

150

200

250

300

350

400

450

PHILIPPINE TRENCH CORE

□

LUTITE

|Oc=>0|

MUD LUMPS

-

ABSENT

H

SILTY STREAKS

+
r

PRESENT (SINGLE SPECIMEN)
RARE (2-5 SPECIMENS)

tM

DIATOM

C

COMMON (11-25 SPECIMENS)

S

WOOD FRAGMENTS

a

ABUNDANT FRAGMENTS

Text-Fig. 3 — Distribution of some Radiolaria in the
Philippine Trench core.

SPUM ELLA MAN RADIOL ART A FROM THE INDO-PACIFIC

1483

larian specimens deposited at the Museum of
Paleontology, University of California, Berkeley,
and at the United States National Museum,
Washington, D.C.

This is Contribution no. 403 from the Depart-
ment of Oceanography and Contribution no. 3
from the Joint Oceanographic Research Group,
University of Washington, Seattle.

LITHOLOGY

The detailed account on the lithology of the
core sediments has already been presented
(Anikouchine& Ling, in press). Therefore, only a
brief description is given here. The positions of
the samples studied in these cores are shown in
Text-figures 2 to 4.

The Java Trench core contains olive-gray to
greenish-gray clayey silt interbedded with silt
layers about 1 cm. thick and consisting of finely
divided mica. The silt layers are spaced about
every 25 cm. throughout the core. Cross bedding
was observed at several horizons in the core. The
otherwise uniform clayey silt is streaked and
banded olive brown, moderate yellowish brown,
dark yellowish brown, and lighter and darker
grayish green.

The Philippine Trench core also contains an
alternating series of sediments. Dark greenish-
gray clayey silt or uniform texture in layers 2 to
8 cm. thick are randomly intercalated with
similarly colored irregular layers, lenses, and

cr
o

o

50

100

150

1ARIANA TRENCH CORE

] LUTITE
HH MUD LUMPS
F?i SILTY STREAKS

- ABSENT

+ PRESENT (SINGLE SPECIMEN)

r RARE (2-5 SPECIMENS)

Text fig. 4 — Distribution of some Radiolaria in the
Mariana Trench core.

streaks of rough-textured clayey silt that is rich
in diatom valves in most samples and rich in clay
lumps in other samples. These layers and lenses
range in thickness from 5 mm. to 13 cm. and are
similarly variable in color. The core contains an
abundance of carbonaceous smears, pyrite
nodules, and wood fragments.

It is likely that both the Java and the Philip-
pine Trench cores represent pelagic sedimenta-
tion punctuated with accumulations from tur-
bidity flows. Silt-sized mica is the turbidity-flow
contribution in the Java Trench, whereas in the
Philippine Trench diatom valves and, to a lesser
degree, terrestrial sediments and clay lumps are
probably contributed by turbidity flows.

In contrast, the Mariana Trench core has a much
more uniform composition of light and dark
yellowish-brown clayey silt, except for some scat-
tered intercalations of silty layers of a few milli-
meters thick and layers up to 40 cm. thick con-
taining clay lumps. The clay lumps contain less
silt than does the surrounding sediment and
were probably introduced into the trench from
areas outside. The mineralogy of the silty layers
is that of a graywacke — rock fragments, quartz,
feldspar, serpentine, biotite, and a variety of
accessory minerals — and thus is indicative of
sudden changes in sedimentation. The layers
probably represent the winnowed remains of a
quantity of sediment that was suddenly intro-
duced to the ocean bottom from a continental
source. In this sense, the three trench cores show
that a similar mechanism of sedimentation
operates in these trenches. The differences ob-
served in the cores probably reflect closeness to
source of sediment and nature of the provenance.

The authors should like to emphasize the
following concerning the presence of ooze of the
diatom Ethmodiscus rex (Rattray) Hendey in
these three trench cores — for further discussion
of this subject, see Anikouchine & Ling, in press:
First, the three studied cores are located within
or approximately within the biogeographic dis-
tribution region of this diatom (Semina, 1959,
fig. 1), yet the diatom is completely absent from
the top or the uppermost samples of the cores.
Second, the distribution of this diatom ooze in
the cores is quite erratic; the ooze is frequently
found in the middle part of Philippine Trench
sediments, between depths of 75 and 328 cm.
from the sediment surface, but is rare in Java
Trench samples and is completely absent from
the Mariana Trench subsurface section.

SYSTEMATIC PALEONTOLOGY

After the grain-size analysis was made, it was
found that most of the radiolarian specimens are
quite free from detritus or organic substances.

1484

HSIN-YI LING AND WILLIAM A. ANIKOUCHINE

Thus, no further chemical treatment was neces-
sary. The specimens were picked up under a low-
power stereomicroscope and both single-speci-
men and strewn slides were made with Canada
balsam.

In the tables of measurements, the location of
the illustrated specimen is recorded in the follow-
ing way: The first two numerals in the codes, (for
example, 33-250) refer to the cores and corre-
spond to the suffix of the core serial numbers
(for example, PI-442-64-33) assigned at the
Pacific Oceanographic Laboratory. The number
33 signifies the Java, 35 the Philippine, and 36 the
Mariana Trench cores. The following numerals
indicate the position of the sample in the core,
that is, the depth in centimeters below the
surface sediments. If the specimen is in a strewn
slide, its location is next recorded with the aid of
the England finder (Riedel & Foreman, 1961).
During the present investigation, the finder was
always placed on the mechanical stage of a Zeiss
photomicroscope in such a way that grid A 1
was located at the upper left-hand corner. Text-
figure 5 illustrates the measurements (W, \V,
L, L') made on the specimens. All the slides will

be deposited permanently in the Micropaleon-
tology Collection at the Department of Ocean-
ography, University of Washington, Seattle,
Washington.

Genus Euchitonia Ehrenberg, 1861

Euchitonia furcata Ehrenberg

Pis. 189, 190, figs. 1-2,5-7

Euchitonia furcata Ehrenberg, 1861a, p. 767; .

18616, ibid., p. 823; 1873a, p. 308; ; 18736,

p. 288-289, PI, 6 (III), fig. 6; Haeckel, 1887,
p. 532-533.

Euchitonia midleri Haeckel, 1862, p. 508-510, PI.
30, figs. 5-10; Stohr, 1880, p. 110, PI. 5, fig. 5;
Haeckel, 1887, p. 533; Cleve, 1901 (incl. E.
ypsiloides), p. 1 1 ; Cocco, 1905, p. 11-12 [question-
able]; Pofofsky, 1912, p. 137-138, Text-fig. 54
only [52 and 53 questionable]; Clark, p. 81-85, PL
4, figs. la,b.

Euchitonia aequipondata Popofsky, 1912, p. 139-140,
PI. 7, figs. 3,4.

not Astromma yelvertoni Macdonald, Haeckel,
1887, p. clxxix (see discussion).

not Euchitonia midleri Haeckel, Bachmann and
others, 1963, p. 136, PI. 11, fig. 62.

Discussion. — The reason for considering E.
muelleri to be a senior synonym of E. furcata is
that E. furcata is the first and only nominal

Explanation of Plate 189
All figures ca. X159 unless otherwise indicated; bright field.

Figs. 1,2 — Euchitonia furcata Ehrenberg. /, Sample 35-328; 2, sample 35-267.
3,4 — E. elcgans Ehrenberg. 3, Sample 33-2; 4, sample 35-328.

5-7— E. furcata Ehrenberg. 5, Sample 35-138; 6, sample 35-405 (033/1); 7, sample 35-328 (S40/0).
X,9— E. cf. E. trianglulum (Ehrenberg) 8, Sample 35-300 (J31/0); 9, sample 35-300 (Q36/0), X200.

Journal of Paleontology, V. 41 Plate 189

Ling & Anikouchine

Journal of Paleontology, V. 41 Plate 190

Ling & Anikouchine

SPUMELLARIAN RADIOLARIA FROM THE INDO-PACIFIC

1485

species of the present genus mentioned by Ehren-
berg (1861«, p. 767) in a table and again in a
table (18616, p. 823) as part of an article in
which he gave the generic diagnosis. In both
citations, however, the name E. jurcata was ac-
companied by two asterisks indicating a new
genus. For new species, only one asterisk was
given. (See Ehrenberg, 1858, p. 553.) It is true
that the description and illustration of E. Jurcata
were given by Ehrenberg (1873o, p. 308; 18736,
p. 288-289, PI. 6 (III), fig. 6, respectively) later
than those of E. muelleri by Haeckel (1862, p.
508-510, PI. 30, figs. 5-10). Because only the spe-
cies name was mentioned, however, and because
the two asterisks were included, Ehrenberg's
action is here presumed to constitute an indica-
tion as specified in Article 16 (a) (v) of the Inter-
national Code of Zoological Nomenclature (Stoll
and others, 1961). Therefore E. jurcata is consid-
ered to be a valid name and consequently has
priority over E. muelleri.

Campbell (1954, p. 86), in his compilation for
the Treatise on Invertebrate Paleontology, desig-
nated E. jurcata as the type of the genus and

Text-fig. 5 — Schematic diagram of patagium-bear-
ing Radiolaria showing the measurements (in mi-
crons) made in this paper. W = the maximum width
of the arm; W' = the minimum width of the arm;
L = the length of the arm, measured from the pre-
sumed geometrical center or center of the central
structure to the distal end of the arm; L' = the
length of the arm.

Explanation of Plate 190
All figures ca. X159 unless otherwise indicated; phase contrast.

Figs. 1—7 — The same as corresponding figures in Plate 189.

8,9— E. cf. E. trianguhim (Ehrenberg). 8, Sample 35-300 (N37/1), X200; 9, Sample 35-300 (Q36/0),
X200.

1486

HSIN-YI LING AND WILLIAM A. ANIKOUCHINE

marked the date 1872 with an asterisk, indicat-
ing that the species was fixed by the original
author on the date of the original publication.
It is clear from the discussion above that the
valid date of the present species should be 1861.
Macdonald (1871, p. 226, figs. 1,2) proposed
and illustrated a new species, Astromma yelver-
toni, but did not describe it. Haeckel (1887, p.
clxxix) considered the species to be synonymous

are found between the two showing a continuous
variation, and no clear separation is possible.
Although we cannot completely support Clark's
(1965, p. 85) view that "the three specimens de-
scribed by Popofsky (1912) as E. aequipondata
are apparently only particularly well-developed
examples [italics ours] of E. mitlleri" (E. furcata
in this paper), we nevertheless agree here that it
is best to consider Popofsky's species as E.

MEASUREMENTS IN

MICRONS

Sample

W

W

L

L'

Plate

fig-

35-

-328

33

75

265

190

189,

190

1

35-

-267

30

63

210

185

189,

190

2

35

-138

38

60

150

120

189,

190

5

35

-405(033/1)

40

85

155

125

189,

190

6

35-

-328(S40/0)

30

50

140

120

189,

190

7

Observed range

30-50

50-90

140-265

100-190

based on 20 specimens

with his species, E. muelleri, but completely
neglected to mention this in his Challenger text
because he evaluated Macdonald' s paper as one
of six listed pieces of " . . . absolutely worthless
literature, which contains either only long
known facts or false statements, and may hence
be entirely neglected with advantage." Whether
or not Macdonald's publication is worthless is
not the subject of the present discussion, how-
ever. Because all the angles between the arms
of his figure 1 are the same, the specimen can-
not be classified in the genus Euchitonia, as
Haeckel presumed, according to the current
classification.

Haeckel (1862, p. 508) once considered his E.
ypsiloides (originally as Histiastrum ypsiloides
Haeckel, 1861, p. 843) to be a synonym of E.
muelleri, but in his Challenger report he then
treated them as separate species. The descrip-
tion of E. ypsiloides is quite brief and no illustra-
tion is given; therefore, the authors could not
reach any conclusion at this time, and the name
of E. ypsiloides is not included in the synonymy
above.

Numerous specimens of E. furcata recovered
from the trench samples show a wide range of
variations in the extent of patagium, the size of
the specimen, and in the form of the arms, par-
ticularly at the distal end. For example, three
specimens here illustrated (Pis. 189, 190, figs.
5-7) are considered as near typical for the pres-
ent species, whereas another specimen (Pis. 189,
190, fig. 1) is comparable to Popofsky's new
form, E. aequipondata (1912, p. 139-140; PI. 7,
figs. 3,4). Numerous transitional forms, however,

Euchitonia elegans (Ehrenberg)
Pis. 189, 190, figs. 3,4

Pteractis elegans Ehrenberg, 1873a, p. 319; ,

1873b, p. 298-299, PI. 8, fig. 3.
Euchitonia elegans (Ehrenberg). Haeckel, 1887,

p. 535; Cleve, 1901, p. 11; P.iedel, 1952, p. 1-18

(particularly p. 11-12 and 14); Clark, p. 85-88,

PI. 4, figs. 2a,b; fig. 18.
Non? Euchitonia elegans (Ehrenberg), Popofsky,

1912, p. 138-139, text-figs. 55-57, PI. 7, fig. 2.

Discussion. — Ehrenberg described (1873a, p.
319) and illustrated (1873b, PI. 8, fig. 3) the
present species from a bottom sediment
(18°03'N, 129°11' E; depth, 6,040 m.) in the
Philippine Sea. Although he added: "cfr.
Monatsbericht 1860, p. 767." at the end of his
original description, the name of this species was
not found on the mentioned page nor on his
faunal list in the paper, and it is our understand-
ing that he referred only to the sample.

His original illustration (1873b) clearly shows
that the distal end of each of the three arms nar-
rows to a point, but actually only one arm, an
odd one, possess a spine at its distal end. We,
therefore, agree with Popofsky's (1912, p. 138)
comments that Haeckel (1887, p. 535) misinter-
preted Ehrenberg's original figure and stated
that there are terminal spines at the end of the
arms. However, judging from the illustration and
particularly his Text-figure 57, which he con-
sidered as the only mature as well as the complete
form, it is doubtful that the specimens Popofsky
found actually belong to E. elegans.

The extent of patagium found in the trench
samples ranges from considerably developed, but
not as complete as Ehrenberg's specimen, to
entirely absent.

SPUMELLARIAN RADIOLARIA FROM THE INDO-PACIFIC

1487

MEASUREMENTS IN MICRONS

Sample

W

W

L

U

Plate fig.

33-2
35-328

40
40

55
50

170
210

135
175

189, 190 3
189, 190 4

Observed range

33-40

35-55

165-265

115-2

30

Based on 25 specimens

Euchitonia cf. E. Triangulum (Ehrenberg)
Pis. 189, 190, figs. 8,9

Stylactis triangulnm Ehrenberg, 1873a, p. 320;

, 1873b, p. 298-299, PI. 8, fig. 9; Stohr, 1880,

p. 113, PI. 6, fig. 2
Euchitonia triangulnm (Ehrenberg), Haeckel, 1887,

p. 533.

Discussion. — As in the example of E. elegans,
Ehrenberg (1873a, p. 320; 1873b, p. 298-299)
twice indicated that the name of the species was
proposed during the year of 1860. However, the
search on this account is without success.

The original figure illustrated by Ehrenberg,
based on a specimen from the Philippine Sea, has
a nearly complete patagium, but the nature of
the arms is completely obscured. Stohr's (1880,
PI. 6, fig. 2) figure does not show the patagium,
and he considered the specimen to be a broken
one. Both figures, however, illustrate clearly
that E. triangulnm has a relatively large con-
centric central structure.

The specimens assigned here to this species,
despite the stated discrepancies, seem to agree
in general with the above-mentioned illustrations
and with the description subsequently given by
Haeckel (1887, p. 533).

The intraspecific variations observed during
the present study are (1) the inconsistency in the
extent of the patagium and (2) the shape of the
arms, particularly the distal ends.

Dictyastrum angulatum by Ehrenberg (1861a,
p. 767; 1873a, p. 306; 1873b, p. 288-289, PI. 8, fig.
18), which Haeckel (1887, p. 589, 590) considered
to be Rhopalodictyum truncatum of Ehrenberg
(1862, p. 301), closely resembles the E. triangu-
lum, except that all arms are equidistant and
the surface is completely spongy.

Genus Cyclastrum Rust, 1898

Cyclastrum? sp.

Pis. 191, 192, figs. 1-2

Discussion. — The genus Cyclastrum was es-
tablished by Rust (1898, p. 28, PI. 9, fig. 5) from
the Jurassic samples from Italy, C. injundibuli-
forme being both genotypic and monotypic. The
generic characteristic given by him is: "Die Dis-
talenden der drei Arme durch einen spongiosen
Patagialgiirtel verbunden."

It is important to point out that Campbell's
(1954, p. 86) diagnosis for this genus is mislead-
ing, inasmuch as he described the genus as "like
Chitonastrum but has patagium," whereas for
Chitonastrum he indicated, "three distally forked
[italics ours] arms; no patagium." It seems quite
clear from the description and illustration given
by Rust that he never intended to include the
forked-arms form in his genus.

The two illustrations here show a specimen
having the complete patagium and a specimen
having only a partial patagium located only in
one of the three interbrachial areas. If the
patagium is completely absent from them, they
would be classified in different genera, such as
Dictyastrum cr Rhopalastrum, according to the
current generally accepted scheme.

The concentric nature of the central structure
is clearly shown in our samples, whereas such a
feature was not mentioned nor illustrated by
Rust.

In view of these discussions and the fact that
Rust's genus is the closest to the forms that can
be found, the authors tentatively assign them to
Cyclastrum.

MEASUREMENTS IN MICRONS

Sample

W

W

L

L'

Plate fig.

15-300 (J34/0)
35-300 (N37/1)
35-300 (Q36/0)

60
60
50

100
80
90

130
110
110

80

75
70

189 8

190 8
189, 190 9

Observed range

50-60

80-110

110-140

70-95

based on 15 specimens

1488

HSIN-YI LING AND WILLIAM A. ANIKOUCHINE

MEASUREMENTS IN MICRONS

Sample

W

W

L

L'

Plate

fig-

35-201 (R41/4)
35-405 (K4 1/0)

40

35

95
100

220
200

190
165

191,192
191,192

1
2

Genus Hymeniastrum Ehrenberg, 1847

Hymeniastrum euclidis Haeckel

Pis. 191, 192, fig. 3

Hymeniastrum euclidis Haeckel, 1887, p. 531, fig. 13.
Cleve, 1901, p. 11; Popofsky, 1912, p. 136-137,
Text-fig. 51.

Discussion. — On the basis of the fauna from
bottom sediments of the northeast Pacific

Ocean, the degree of patagium shown in the
specimen of the present species and its relation
to taxonomy and ontogenetic considerations
have been discussed by Ling (1966). The oc-
currence of similar phenomena in the trench core
samples indicates that such variation actually
occurs beyond any biogeographic limitation and
quite possibly is a common phenomenon in these
patagium-bearing Radiolaria.

MEASUREMENTS IN MICRONS

Sample

\V

W

L

V

Plate fig

35-227 (J36/0)
Observed range

35
30-40

85
75-100

180
150-210

145
110-175

191,192 3
based on 10 specimens

Explanation of Plate 191
All figures ca. X159 unless otherwise indicated; bright field.

Figs. 1,2—Cyclastrum? sp. 1, Sample 35-201 (R41/1); 2, sample 35-405 (K41/0).
3 — Hymeniastrum euclidis Haeckel. Sample 35-227 (J 36/0).
4,5—Dictyocoryne sp. 4, Sample 33-29 (N39/3); 5, sample 33-250.
6 — D. profunda Ehrenberg. Sample 35-267.
7 — Rhopalodictyum abyssorum Ehrenberg. Sample 36-120 (M42/0), X200.

Journal of Paleontology, V. 41 Plate 191

Ling & Anikouchinc

' • 'w ~*"-'fr?pcff> * ' . :* V» ******

Journal of Paleontology, V. 41 Plate 192

Ling & Anikouchine

•■•^Bfe,':,.

SPUMELLARIAN RADIOLARIA FROM THE IN DO-PACIFIC

14X9

Genus Dictyocoryne Ehrenberg, 1861

DlCTYOCORYNE Sp.

Pis. 191, 192, figs. 4,5

Discussion. — The rapidly expanded distal half
and flatly truncated nature of the distal end,
forming chalice-shaped arms, are quite charac-
teristic and clearly distinguish this form from
any other published species.

Rare specimens observed in the present study
prevent this from being considered as new. A
highly variable extent of patagium in this form
is, nevertheless, clearly demonstrated. The form
is found only in the Java Trench sediments.

MEASUREMENTS IN MICRONS

Sample

W W L Plate fig.

33-29 (N39/3)
33-250

39 114 156 191,192
38 120 144 191,192

Dictyocoryne profunda Ehrenberg
Pis. 191, 192, fig. 6

Dictyocoryne profunda Ehrenberg, 1861a, p. 767;

, 1873a, p. 307; , 1873b, p. 288-289, PI. 7,

fig. 23; Haeckel, 1887, p. 592; Martin, 1904, p.
454, PI. 80, figs. 11-13.

Discussion — The type locality of this species
is in sediment at a depth of 6,040 m. in the
Philippine Sea. Ehrenberg (1861a, p. 767) found
only one specimen in the sample. Two additional
occurrences of this species were reported by
Haeckel (1887, p. 592) from his Challenger
samples: Station 198, 3,930 m. in the Celebes
Sea, and Station 274, 3,200 m. in the eastern
central Pacific Ocean.

Martin (1904, p. 454) noticed considerable in-
traspecific variation in the specimens and stated
"it may be seen from the figures that the arms
are not absolutely equidistant as they are sup-
posed to always be in this genus."

Through the kind arrangement of Mr.
Frederick J. Collier of the Smithsonian Institu-
tion we were allowed to examine Martin's pre-
sumed Miocene slides, deposited at the U.S.
National Museum (USNM). Four specimens are
identified by Martin as D. profunda, and among
them three forms figured by him are found in the
original marked area of the slides. In all cases,
the measurements calculated from his figures are
slightly smaller than those measured directly
under the microscope. The reexamination pre-
sents the following results:

Explanation of Plate 192
All figures and magnification the same as corresponding
in Plate 191, except phase contrast.

ligures

1490

HSIN-YI LING AND WILLIAM A. ANIKOUCHINE

Martin (1904)
PI. 80

USNM Colin. No.
and position

Remarks

Figure 1 1

No. 649655 T38/0

The concentric pattern is present in the patagium cf the specimen
but is not illustrated in the figure.

Figure 12

No. 649656 N38/3

The concentric nature of the central structure is observed in the
specimen but is quite obscure in the figure .

Figure 13

No. 649657 N38/0

As illustrated by Martin.

Not figured

No. 649673 N40/1

The patagium in this specimen is completely absent, and only a
trace of the structure can be detected in interbrachial areas if
observed carefully. The reason he did not figure the specimen is
unknown.

Although Martin (1904) did not mention it in
his text, his illustrations show clearly that differ-
ent degrees of patagium are present in his Mio-
cene materials. Furthermore, it seems quite clear
that he already had recognized that, as a part of
a continuous intraspecific variation, the pata-
gium of this species could be completely absent.

In the trench core specimens, the extent of
patagium observed ranges from nearly complete
to entirely absent.

von Frattingsdorf N. O.: Geol. Gesell. Wien, Mitt.,

v. 56, pt. 1, p. 117-162, Pis. 1-24.
Campbell, A. S., 1954, Radiolaria, in Campbell,

A. S., & Moore, R. C, Protista 3: Lawrence,

Kansas, Geol. Soc. America and Kansas Univ.

Press, Treatise on Invertebrate Paleontology, pt.

D, p. 11-163, figs. 6-86.
Clark, C. A., 1965, Radiolaria in Recent pelagic

sediments from the Indian and Atlantic Oceans

(Ph.D. thesis): Cambridge, England, Cambridge

Univ., 234 p.

MEASUREMENTS IN MICRONS

Sample

W

W

Plate

fig.

35-267
Observed range

45
35-50

85
80-120

150
130-150

191,192 6

based on 10 specimens

Genus Rhopalodictyum Ehrenberg, 1861

Rhopalodictyum abyssorum Ehrenberg

Pis. 191, 192, fig. 7

Rhopalodictyum abvssorum Ehrenberg, 1861a, p. 769;

, 1873a, p. 319; . p. 298-299, PI. 8, fig. 17;

Haeckel, 1887, p. 589

Discussion. — There seems little doubt that
forms illustrated here belong to the species de-
scribed and illustrated by Ehrenberg (1873«, p.
319; 1873b, PI. 8, fig. 17). He found two specimens
in the sediments from the Philippine Sea.

Cleve, P. T., 1901, Plankton from the Indian
Ocean and the Malay Archipelago: K. Sv. Vetensk.-
Akad., Handl., v. 35, no. 5, p. 1-58, Pis. 1-8.

Cocco, L., 1904, I Radiolaria fossili del Tripoli di
Condro (Sicilia): R. Accad. Zelanti, ser. 3a, v. 3,
p. 1-14.

Ehrenberg, C. G., 1847, Ueber die Mikroskopischen
kieselschaligen Polycystinen als machtige Gebirgs-
masse von Barbados: K. preuss. Akad. Wiss. Ber-
lin, Monatsber., p. 40-60.

, 1858, Ueber die Organischen Lebensformen in

MEASUREMENTS IN MICRONS

Sample

W

W

L

Plate fig.

36-120 (M42/0)
Observed range

50
40-50

65
65-80

130
115-145

191,192 7
based on 10 specimens

REFERENCES

Anikouchine, W. A., & Ling, H. Y., in press, Evi-
dence for turbidite accumulation in trenches in the
Indo-Pacific region: Marine Geol.

Bachmann, A., Papp, A., & Stradner, H., 1963,
Mikropalaontologische Studien im "Badener Tegel"

unerwartet grossen Tiefen des Mittel-meeres:
Ibid., Jahrg. 1857, p. 538-570.

, 1861a, Ueber die Organischen und unorganischen

Mischungsverhaltnisse des Meeresgrundes in 19800
Fuss Tiefe nach Lieut. Brookes Messung: Ibid.,
Jahrg. 1860, p. 765-774.

SPUMELLARIAN RADIOL ARIA FROM THE INDO-PACIFIC

1491

, 1861b, Ueber den Tiefgrund des stillen Oceans

zwischen Californien und den Sandwich- Inseln aus
bis 15600' Tiefe nach Lieut. Brooke: Ibid., Jahrg.
1860, p. 819-833.

, 1862, Ueber die Tiefgrund- Verhaltnisse des

Oceans am Eingange der Davisstrasse und bei
Island: Ibid., Jahrg. 1861, p. 275-315.

, 1873a, Mikrogeologische Studien als Zusam-

menfassung seiner Beobachtungen des kleinsten
Lebens der Meeres-Tiefgrunde aller Zonen und
dessen geologischen Einfluss: Ibid., Jahrg. 1872, p.
265-322.

, 1873b, Mikrogeologische Studien iiber das

kleinste Leben der Meeres-Tiefgrunde aller Zonen
und dessen geologischen Einfluss: K. Akad. Wiss.
Berlin, Abh., Jahrg. 1872, p. 131-399, Pis. 1-12.

Haeckel, E., 1861, Fernere Abbildungen und Diag-
nosen neuer Gattungen und Arten von lebenden
Radiolarien des Mittelmeeres: K. preuss. Akad.
Wiss. Berlin, Monatsber., Jahrg. 1860, p. 835-845.

, 1862, Die Radiolarien, eine Monographic:

Berlin, Georg Reimer, 572 p., 35 pis.

, 1887, Report on the Radiolaria collected by

H.M.S. Challenger during the year 1873-1876: Rept.
Voy. Challenger, Zool., v. 18, clxxxvii + 1893 p.,
140 pis.

Ling, H. Y., 1966, Notes on patagium in the radiolar-
ian genera Hymeniastrum and Dictyastrum: Micro-
paleontology, v. 12, no. 4, p. 489-492, PI. 1.

Macdonald, J. D., 1871, Examination of deep-sea
soundings, with remarks on the habit and structure
of the Polycvstina: Annals Mag. Nat. History, ser.
4, v. 8, p. 224-226, figs. 1-2.

Martin, G. C., 1904, Radiolaria: Maryland Geol.
Survey, Miocene, p. 447-459, PI. 130.

Popofsky, A., 1912, Die Sphaerellarien des Warm-
wassergebietes der Deutschen Stidpolar-Expedition
1901-1903: Deutsche Sudpolar-Exped., v. 13 (Zool.
v. 5), pt. 2, p. 73-159, Pis. 1-8, 77 text-figs.

Riedel, W. R., 1952, Tertiary Radiolaria in western
Pacific sediments: Goteborgs K. Vet. Vitterh.
Samh. Handl., ser. B, v. 6, no. 3 (Medd. Oceanogr.
Inst. Goteborg, 19), p. 1-21, Pis. 1, 2.

Riedel, W. R., & Foreman, H. P., 1961, Type
specimens of North American Paleozoic Radio-
laria: Jour. Paleontology, v. 35, p. 628-632, 7 text-
figs.

Rust, D., 1898, Neue Beitnige zur Kenntnis der
fossilen Radiolarien aus Gesteinen des Jura and
Kreide: Palaeontographica, v. 45, p. 1- 67, Pis.
1-19.

Semina, H. J., 1959, The Distribution of the diatom
Ethmodiscus rex (Wall.) Hendey among the plank-
ton: Akad. Nauk USSR, Doklady, v.' 124, no. 6,
p. 1309-1312, 3 figs

Stohr, E., 1880, Die Radiolarien fauna der Tripoli
von Grotte Provinz Girgenti in Sicilien: Palaeon-
tographica, v. 26, p. 69-124, Pis. 1-7.

Stoi.l, N. E. and others, 1961, International Code
of Zoological Nomenclature adopted by the XV
International Congress of Zoology: London, In-
ternational Trust for Zoological Nomenclature,
176 p.

Manuscript received November 7, 1966

22

Reprinted from PROCEEDINGS THE ELEVENTH PACIFIC SCIENCE CONGRESS
Volume 2 Oceanography, Tokyo, 1966

-71-

VERTICAL CRUSTAL MOVEMENT ASSOCIATED WITH THE 1964 ALASKAN EARTHQUAKE
R. J. Malloy. Institute for Oceanography, Environmental Science Services Administra-
tion, Rockville, Maryland, U. S. A.

Geophysical reconnaissance including gravity, magnetics, seismic reflection pro-
filing, and hydrography was carried out in the Spring and Summer of 1964 aboard the
U. S. Coast and Geodetic Survey ship SURVEYOR in the epicentral area of the March 27
Alaskan earthquake. Portions of the sea floor southwest of Montague Island, Alaska
wexe calculated to have been uplifted in excess of 50 feet on the basis of these recon-
naissance data. This, and adjacent areas of tectonic focus were revisited by the
SURVEYOR during 1965 to conduct combined hydrographic and ocean bottom scanning sonar
studies. This work was controlled by precision navigation. A conventional echo-
sounder using a conical 21 KC sound beam was used to collect hydrographic data, and a
bottom scanning sonor using two fan shaped sound beams of 150 KC and 160 KC covering a
total of 2400 feet of the bottom was used to yield a pictorial record of the sea floor.

The submarine extensions of the fresh fault scarps exposed on Montague Island
were traced to sea a distance of 12 nautical miles by the combination of the two tools.
Other lineaments brought out by the bottom scanning sonar are interpreted as strata
ridges and joints. Areas of bedrock outcrop responded to the high frequency sound used
in the bottom scanner, whereas areas of the sea floor with a fine grained sediment cover
did not, in general, return geologically meaningful data. Presumably, this contrast
in performance is a function of the reflection coefficient of these sea floor environ-
ments and the penetrability of the water by these frequencies. Large sand waves
recorded by the bottom scanner in the entrance to Cook Inlet showed, upon subsequent
coring to be composed of a large percentage of very coarse sand and pebbles.

23

Reprinted from PROCEEDINGS, THE ELEVENTH PACIFIC SCIENCE CONGRESS
Volume 2 Oceanography, Toyko, 1966

-75-

SUBMARINE GEOLOGY OF THE ALEUTIAN ARC, ALASKA

Richard B. Perry and Haven Nichols. ESSA, Institute for Oceanography, U. S. Department

of Commerce, Rockville, U. S. A.

Six new maps show the bathymetry of the Aleutian Arc (170 E. to 160 W. , 50 N. to
55 N. ) at a scale of 1 : 400, 000 and a contour interval of 50 fathoms. The geologic
history of this tectonically-act ive area can be inferred from the bathymetry as well as
from observations of the seismology, sub-bottom acoustic reflections, gravity, geomag-
netism and island geology.

The arcuate Aleutian Ridge was formed in originally oceanic crust by magma
pushing up from below along a line where the northward-moving floor of the Pacific
Ocean moves under a more stable section of the ocean floor near the continental blocks
of Asia and North America. The later thrusting-up of Bowers Ridge, which extends
northward perpendicular to the Aleutian Ridge at approximately 180 longitude, caused
extensive transverse faulting and rotation of major blocks on the Aleutian Ridge from
Adak Canyon west to the Near Islands. The formation of relatively recent volcanoes
along the northern portions of the Aleutian Islands and erosion by glacial ice and
water have produced many of the land forms that are found down to 200 meters below sea
level. The Aleutian Terrace and Trench on the south side of the Aleutian Ridge show a
relatively youthful form that is still being shaped by tectonic forces, while the
more stable area to the north of the Ridge is an area of thick, flat-lying sediments.

24

Reprinted from PROCEEDINGS THE ELEVENTH PACIFIC SCIENCE CONGRESS
Volume 2 Oceanography, Tokyo, 1966

-70-

RESULTS FROM A GEOPHYSICAL SURVEY IN THE NORTH-EAST PACIFIC OCEAN

G. Peter, D. Elvers, and 0. Dewald. Office of Oceanography, Environmental Science

Services Administration, Washington , D. C, U. S. A.

Data on sea floor topography and the magnetic and gravity fields have been obtained

2
through a closely spaced grid network over an area larger than 650,000 km in the North-
East Pacific Ocean.

Magnetic anomalies form bands of highs and lows aligned North-Northwest over the
surveyed area. These bands are disrupted not only by the well known Mendocino fracture
zone, but by smaller hitherto unreported fracture zones as well. The fracture zones
revealed by the detailed survey of the ocean floor and those which may be inferred from
the discontinuities of the magnetic anomaly bands indicate the dominance of an approx-
imately dast-west fracture pattern of the North-East Pacific Basin.

The anorialy bands, running nearly perpendicular to the fractures, further complicate
the tectonic matrix of the area; these still await an adequate explanation as to their
geological origin. Such an explanation would need to include the lack of correlations
that have been found between the gravity and magnetic anomalies in the area.

25

Reprinted from BULLETIN OF THE SEISMOLOGICAL
SOCIETY OF AMERICA Vol. 57, No. 2

AUTOMATED EPICENTER LOCATIONS FROM A
QUADRIPARTITE ARRAY

By T. J. Sokolowski and G. R. Miller

ABSTRACT

A quadripartite seismic array has been installed on the island of Oahu, Hawaii,
to supply additional data for the Tsunami Warning System of the USCGS. The
nuclear explosion LONGSHOT on October 29, 1965, on Amchitka Island served
in calibrating the epicenter location system. A moving cross-correlation de-
termined the P wave time delays between the various station pairs. Time delays
were then used in a least-squares method to obtain azimuth and emergence
angles. The emergence angle is used with the Jeffreys-Bullen travel-time curve
to obtain the distance to the source. The emergence angles obtained by using
the observed station-pair time delays should be corrected to the emergence
angles obtained by using the station-pair time delays as computed from the
Jeffreys-Bullen travel-time curve. This comparison will then compensate for the
anomalies of the Pacific area. With the values for azimuth and distance, one can
then arrive at an approximate epicenter location for earthquakes of normal
depth. This method can be used for those situations which require an epicenter
location within minutes after the earthquake has been recorded. Methods
developed here are designed for an on-line computer.

Introduction

In April 1965, the Coast and Geodetic Survey installed a quadripartite seismic
array on Oahu, Hawaii. This system consists of four seismic detecting stations and
a central control system. The four seismic stations are located at a maximum dis-
tance apart on the island, and in places of minimum background noise. Location
points for the four seismic stations are: Kaena Pt., Ft. Barrette, Mokapu, and
Pupukea. The central control system is located at the Honolulu Observatory (see
Figure 1). Each station is equipped with a short-period vertical, Benioff seismometer
(35 lb. mass transducer) and a console unit. The operating magnification of the
seismometers is between 5000 and 15000. The console which accompanies the
seismometers contains a calibration mechanism, control unit, phototube amplifier,
frequency modulator, batteries, and a charger. The ground motion detected by the
seismometer is telemetered via telephone cables as an FM signal to the central
control system at the Observatory. Figure 2 is a block diagram of the system.

The central control system consists of three major units; a main console, Develo-
corder, and a Dynograph recorder. The main console, which is the first unit of the
central control system, contains discriminators (which remove the telemetry fre-
quency), control modules, batteries, charger, timing system, power amplifier, power
supply, and radio. The main console functions as a sensing, directing link, supplying
both the Dynograph recorder and Develocorder with the seismic signals from the
distant seismic stations. This link is also used as a check point to any of the major
components in the system or for calibration of the instruments.

269

270

BULLETIN OF THE SEISMOLOGICAL SOCIETY OF AMERICA

The Develocorder is the second unit of the central control system. It contains
five enclosed components: recording, processing, film transport, projection, and
power. The recording component can receive a maximum of 16 channels of seismic
data in electrical signal form and convert these signals into rotation of mirror
galvanometers. Light reflected from the mirror galvanometers then exposes the 16

21° 40'

1 1

■" 1 ~

J ©PI

ipukeaS

<^AENA~P~L

21° 30'

- 1 0 A

h u r

MOKAPU

-

0 J

21° 20*

FT.\BARRETTE^-v

L J? ? 0B]

ONOLULUt'-^L

5ERVAT0RY >. ^- \S

i

SCALE IN MILES

158° 10'

158°

157° 50

157° 40

Fig. 1. Map showing location of quadripartite seismic array and central control system on

Oahu.

STATI 0 IN SITES

HONOLULU OBSERVATORY

IY1 0 K A PU
| S eismometer

PU PUK EA
Seismometer

KAEIMA PT.

| Seismometer |-

FT BARRETTE
| Seismometer ~J-

Remote
Console

Remote
Console

Remote
Console

Remote
Console

Transmission

Cable 35 km
Transmission

Cable 34km
Transmission

Cable 34km
Transmission

Cable 2 kr

MAIM
CONSOLE

DEVELOCORDER

DYIMOGRAPH
R EC0RDER

Fig. 2. Block diagram of quadripartite array.

mm film. The processing component, using X-ray chemicals, develops, fixes, and
washes the exposed film. The film transport component assures a correct recording
rate in conjunction with the frequency-regulated power supply. Forward or reverse
film travel can be obtained for viewing or roll storage. The projection component
then magnifies and focuses the recorded images on a small-view screen approxi-
mately ten minutes after the earthquake has been sensed at seismic stations.
The Dynograph recorder, which is the third unit, is a direct writing recorder. Its

AUTOMATED EPICENTER LOCATIONS

271

main function is as a standby unit. This recorder operates only at those times when
the Develocorder is inoperative for reasons of maintenance or malfunction.

Time Delays

Time delays obtained from the six station pairs are used in determining the
azimuth from the detecting stations to the event. Since the maximum delay from
any of the pairs is approximately 4 seconds, these delays should then be determined
as accurately as possible. The recorded LONGSHOT event was digitized with
intervals of 0.12 second. This included values of the recording both before and after
the initial onset of the P-phase pulse. The four stations, therefore, gave four sepa-
rate time series, which we grouped into six station pairs. Figure 3 shows the four-

FT. BARRETTE

o
o

KAENA PT.

MOKAPU

_L

10

20

25

time (sec)
Fig. 3. Quadripartite array recording of the LONGSHOT event.

station recording of the LONGSHOT event. Three methods were applied in attempt-
ing to obtain the station-pair time delays; first, visual inspection of initial first
motion; second, moving cross-correlation; and third, Fourier spectrum analysis. In
general, in the visual inspection method, all four traces of the initial P-motion were
not well defined. In most cases only two or three of the traces of the initial P-motion
were well defined. An attempt was then made to obtain the time delays from the six
station pairs using the phase-change computed from the cross spectrum (Blackmail
and Tukey, 1959). For each of the six station pairs, the spectrums, coherence, and
phase were plotted as functions of frequency. In most cases an expected gradual
phase-change did not occur although the coherence was relatively high. This rapid
phase-change resulted in inconsistent time delays. Power spectrum, coherence, and
phase-change plots are shown in Figure 4 for one of the station pairs. The time delay
was computed using (Blackman and Tukey, 1959)

272

BULLETIN OF THE SEISMOLOGICAL SOCIETY OF AMERICA

TK =

1

phase- change

/ L 360 deg.
where T K is the time delay for the Kth station pair, at frequency/.

180°

<D

-180°
1.0

<D
O

d

0.5 £

frequency (cps)

frequency (cps)

Fig. 4. Plot showing power spectrum, coherence, and phase-change for two of the seismic

stations.

The moving cross-correlation technique yielded the most accurate and consistent
time delays between station pairs. A data window was applied to each section of the
time series. The cross correlation is given by

Li

TO i=\

Ui+j Vi

■I,

-1,0,1,

•• I

where :

Lj is the cross correlation

Wj is a weighting function

Ui and iu are the series being cross-correlated
m is the number of terms in «,- and yt-
and

I is the maximum number of lags.
For each station pair, the maximum positive peaks in the cross-correlation were
accepted as the representative maximum correlation. Series segments of various
lengths were used which included at least the first five to six cycles after the onset of
initial P-motion. Figure 5 shows the cross-correlation of one of the station pairs
from which the time delay was obtained.

AUTOMATED EPICENTER LOCATIONS

273

1000

1000

time (sec)

Fig. 5. Plot of moving cross-correlation between recordings at two of the seismic stations.
Maximum positive peak occurs at —0.8 second.

Azimuth and Distance

Once the time delays were obtained for all the station pairs, a least-squares com-
puter program was developed to obtain the azimuth and distance from the net
stations to the event. The function H which was minimized is given by

H = £ {Bk - [DKcos (6 - aK) cos<p]/V}'-

K = 1,

6 (1)

where :
BK is the observed time delay for the Kth station pair
DK is the distance between the stations of any Kth station pair
d is the azimuth to the event measured clockwise from north at the center of the

array
aK is the angle measured clockwise between the Kth station pair and the parallel

through the station with the highest latitude
<P is the emergence angle
and

V is the assumed velocity oi 8.0 km/sec.
The azimuth obtained from (1) is the result of a scanning process. The scanning-
process starts at 0° north and computes the 360° of 6 for each value of <p. <p is limited
from 0° to 90°. The incremental steps are F° for both <p and 6. When a minimum of H
is found, it steps back F° in both <p and 6 values and then proceeds forward in smaller
incremental steps. These steps being G°, where G° < F°. The process stops after a
length of 2F° is computed. Incrementation can be made as small as desired. This
scanning process is rapid enough so that there would be small gain from a more
sophisticated procedure. Figure 6 shows a graph of the least-squares values versus
azimuth for <p = 0. This curve is a smooth curve which has, in this case, one maxi-
mum and one minimum. Values of H versus <p and 6 yield a smooth surface which
leads to no difficulty in obtaining a minimum H value. An attempt was made to ob-
tain the minimum value of H by considering the simultaneous solution of the equa-
tions dH/dd = 0 and dH/d<p = 0. No simple analytic solution was obtained for
these equations.

274

BULLETIN OF THE SEISMOLOGICAL SOCIETY OF AMERICA

0.6 1-

180
azimuth (9)

360

Fig. 6. Graph of least-squares solution versus 6 as 6 varies from 0° to 360° for assumed

emergence angle <p = 0°.

Arrival times to the net stations from various earthquake sources were computed
from the Jeffreys-Bullen normal depth travel-time curve. The various time delays
were then computed by (2) for various distances from each of the station pairs:

Cm

— ) DK COS (0 — aK)

K = 1.

,6 (2)

where :

CK is the expected J-B time delay for Kth station pair for any azimuth m, and

dA/dt is the reciprocal of slope obtained from J-B travel-time curve ( Jeffreys,

1962).

These CK values, for distances from 30° to 95°, substituted into ( 1) relate various

<p's for this distance range. Figure 7 shows a graph of <p versus distance. The curve is

60
distance (degrees)

Fig. 7. Graph of <p for various distances from Oahu array. Arrival times, computed from
the Jeffreys-Bullen travel-time curve, were used in the least-squares method to obtain the
angles. Assumed velocity of 8.0 km/sec. Point <p = 57° was obtained for the LONGSHOT event.

AUTOMATED EPICENTER LOCATIONS 275

monotonically increasing and is only applicable to the CK which are computed from
the Jeffreys-Bullen curve for normal depth. In order to relate these <p's to the net
system on Oahu, corrections must be applied to account for the travel-time anoma-
lies. The corrections are obtained in the following manner. Station-pair time delays
(BK) may be obtained from the moving cross-correlation technique. These BK, for
distances from 30° to 95° and various azimuths, substituted into ( 1 ) yield <p's for
the various distances and azimuths; in general the <p's will differ from those ob-
tained by the substitution of the CK in ( 1 ) . Applying the corrections ( the difference
between the BK values and CK values), a final <p, distance curve can be obtained. The
corrections imply corrections in velocities for various azimuths which deviate from
the J-B velocities. To be of practical use, many earthquakes must be analyzed to
obtain <p for distances between 30° and 95° as a function of azimuth. As of the pres-
ent, only one event has been analyzed from which the methods have been devised.

Once the azimuth and distance have been obtained, the geocentric coordinates
can be easily determined from spherical trigonometry.

Conclusion

The quadripartite seismic net on Oahu, Hawaii, can be used to rapidly obtain the
location of an earthquake epicenter. The earthquake must be large enough to be
recorded by the array and must fall within the P -range. All tsunamigenic earth-
quakes meet these criteria. Methods developed here are for those situations which
require a determination of the epicenter within minutes after the earthquake has
been recorded. A small computer will do this adequately.

Acknowledgments

The authors wish to express their thanks to Dr. W. M. Adams and Dr. A. S. Furumoto for
their constructive criticism of this paper.

References

Blackman, R. B. and J. W. Tukey (1959). The Measurement of Power Spectra, Dover Inc., New

York.
Jeffreys, H. (1962). The Earth, Fourth ed., Cambridge University Press, London.

Joint ESSA-U of H Tsunami Research Effort Environmental Science Services Administration

University of Hawaii

Honolulu

Hawaii Institute of Geophysics Contribution No. 168

Manuscript received October 7, 1966.

Reprinted from JOURNAL OF GEOPHYSICAL RESEARCH
Vol. 72, No. lh, The American Geophysical Union

Journal of Geophysical Research

Vol. 72, No. 14

July 15, 1967

26

Geologic Structures in the Aftershock Region of
the 1964 Alaskan Earthquake1

Roland von Huene,2 Richard J. Malloy,3 George G. Shor, Jr.,4
and Pierre St.-Amand2

Seismic and echo sounder profiles in the aftershock region of the 1964 Alaskan earthquake
define a pre-existing zone of discontinuous faults in the area of maximum aftershock strain
release. Steep reverse faults and possibly other types of steep faults occur in the zone; sur-
face rupture has probably not been continuous along its full length during any Recent earth-
quake. The fault zone may represent a narrow area of maximum flexure and uplift in the
broader area deformed during the 1964 earthquake. An anticline at the continental margin
with local large structural relief was also uplifted during the earthquake.

Introduction

After the 1964 Alaskan earthquake three
organizations made geophysical observations to
find structures associated with the earthquake
and also to determine the structure of the con-
tinental shelf, the Aleutian trench, and the
deep Gulf of Alaska seafloor. The U. S. Coast
and Geodetic Survey ship Surveyor made gravi-
metric, magnetometric, and continuous seismic
reflection profiles in the summer of 1964 be-
tween the Hinchinbrook Entrance and the
Trinity Islands (Figure 1). The gravity and
magnetic observations have been combined with
previous observations at sea and with observa-
tions on land by the U. S. Geological Survey
and have been reported by Barnes et al. [1966]
(also Barnes, in preparation). Closely spaced
seismic profiles immediately surrounding Mon-
tague Island have been reported by Malloy
[1964, 1965]; they show a seaward continua-
tion of the faulting on Montague Island which
was active during the 1964 earthquake [Plajker,
1965]. Malloy used a 1,000-joule EG&G-Alpine
spark sound system aboard the Surveyor; he
also made a series of profiles across the con-
tinental shelf (Figure 1). A 20,000-joule Ray-
flex spark sound system was employed during

1 Contribution from the Scripps Institution
of Oceanography, University of California, San
Diego, California.

2 Research Department, U. S. Naval Ordnance
Test Station, China Lake, California 93555.

3 Institute for Oceanography, ESSA, Silver
Spring, Maryland 20910.

4 Scripps Institution of Oceanography, La Jolla,
California 92152.

the same summer by Shor and Reimnitz on
the Scripps Institution of Oceanography ship
Oconostota. The Scripps profiles (Figure 1)
extend from the Gulf of Alaska to Kodiak
Island, where Shor previously had made seismic
refraction profiles [Shor, 1962, 1965]. The third
Scripps profile extends past Middleton Island,
through the Hinchinbrook Entrance, and into
Prince William Sound; a study of structures
and post glacial sedimentation of Prince Wil-
liam Sound has been reported [von Huene et al.,
1967].

Scripps also made seismic profiles off the
Copper River delta [Reimnitz, 1966]. In the
summer of 1966 von Huene made seven seismic
profiles across the continental shelf and the
Aleutian trench, using combined Rayflex and
EG&G spark sound systems to obtain as much
as 38,000 joules of energy. A Rayflex signal-
processing and recording system completed the
instrumentation. Additional profiles made on
the continental shelf during this cruise have
been reported in part [von Huene, 1966] .

In this paper we summarize and consider
data, gathered by the U. S. Coast and Geodetic
Survey, Scripps Institution of Oceanography,
and the Naval Ordnance Test Station, that bear
directly on the structure of the continental shelf
and the 1964 Alaskan earthquake, combining
the Coast and Geodetic Survey data with those
reported by von Huene et al. [1966].

Nature of Seismic Reflection Data

Seismic reflection profiling instruments use
sound sources that are essentially omnidirec-
tional. The linear receiving hydrophone array

3649

3650

von HUENE, MALLOY, SHOR, AND St.-AMAND

9 CA o

3

<

5o*

C*5

-

<r>

s

<

£">»

_i

= T5 5

<

* 1-

>s

trt-O •

li-

o

S1" £

Li.

o0> tj

oOOO

s

3 U) >

r

a.

c — Z

UJ

in o<"

»

T

<3 » =

h-

511 §

o

o

UJ

o

— J=

-I

• So
0» c o»

a.

o. ° «

o

o-o

*l

q-

1

i-

>■ s °

°y

LlI

i- 8 <=

" 1

2
>

°l

1

H

3<£

<

o"

00

r

H

c

O

fe

M

3

a

03

o

03

od

-n

-tj

(/J

a>

a

Cfl

M

111

03

a

>

u

X)

U
03

T3

a

<u T3

CI

~

O

3
O

03

AFTERSHOCK REGION OF ALASKAN EARTHQUAKE

3651

is normally one to three wavelengths long and
therefore has some fore-and-aft directivity, but
none to the sides. Thus strong reflections from
either side of the ship are received along with
the reflections from directly below. The result is
an addition of reflections from the area insoni-
fied by the seismic source, and especially in
deeper water failure to record geologic irregu-
larities of small horizontal extent.

Another effect to consider in interpreting
seismic profiler records is the relative inclina-
tion of planar geologic features. An inclined
reflector returns energy to the source area only
by scattering, so that steeply inclined faults,
bedding, or other acoustic discontinuities may
not be seen.

In the records on which this study is based,
we have considered sharp folding to be the
result of deeper faulting. No fault planes were
directly observed in the reflection records. The
geometry of beds adjacent to faults, as well as
seafloor scarps, are good fault indicators. Oc-
casionally faults are represented by 'blank areas
or by a series of noncoherent reflections. This
may be caused by severe deformation adjacent
to and within a fault zone. It is possible that
many smaller structures were not recognizable
on the seismic profiler records, and that not
only the larger tectonic features but also only
particular types of features were seen.

The records obtained are of two distinct
types. The EG&G-Alpine system produced a
signal rich in higher frequencies, which brought
out details not seen in the Rayflex records. The
lower-frequency signals from the Rayflex system
gave greater penetration, and under some cir-
cumstances these records showed distinct struc-
tures far below the first seafloor multiple. In
general, both types of records were good when
sea conditions were favorable; under adverse
conditions the records deteriorated significantly.
Our information is therefore not uniform in its
presentation of various geologic features, and
smaller faults that are clearly recognizable on
one profile may not be recognized on others.
The individual interpretation of records some-
times differed.

Distribution and Character of Tectonic
Structures

The locations of folds and faults found in
seismic records and bathymetric profiles are

shown in Figure 1. Bathymetric profiles made
without concurrent seismic profiles are so indi-
cated, and the sense of vertical separation
along faults is determined with less certainty
where seismic information is absent. Only se-
lected ship tracks south of Montague Island
are used in the compilation of Figure 1; more
detailed information can be found in Malloy
[1965].

Structures that appear to have formed re-
cently and that may have been active at the
seafloor during the 1964 earthquake are denoted
by special symbols in Figure 2. The somewhat
subjective criteria for judging the recency of a
fault scarp are a sharp bathymetric expression
of the structure on the seafloor and the absence
of erosion or burial. Such judgment was usually
made from fathograms recorded concurrently
with the seismic profiles. Activation or rupture
at the seafloor in 1964 can be demonstrated
only for the faults on, and adjacent to, Mon-
tague Island [Malloy, 1964, 1965; Plafker,
1965]. Because these faults cut hard, well-
lithified rocks, they cannot be used as general
examples for the durability of other scarps on
the seafloor, which all occur in presumably
soft sediments. Since there are suggestions of
both rapid and slow scarp destruction, sharp
bathymetric feaaures were probably formed at
various times before 1964.

The sequence of structures in adjacent profiles
is often similar in the order of occurrence and
in the configuration of constituent reflections.
Lines connecting these structures are approxi-
mately parallel to the regional trend, but such
lines would imply that structures are generally
continuous; it is equally possible that many
small and discontinuous en echelon faults occur.

Faults on land and the few areas where faults
are crossed by closely spaced seismic profiles
comprise the only available basis for judging
the general length of structures. Condon and
Cass [1958] and Condon [1965] have mapped
a number of probable faults in Cretaceous and
Eocene rocks that extend for 20 km. However,
only segments of two of these faults were active
during the 1964 Alaskan earthquake [Plafker,
1965; Malloy, 1965], and many of these faults
show no evidence of strong activity since the
last major glacial recession [von Huene et at.,
1967]. In middle Tertiary strata north of Kayak
Island, later Tertiary structures are generally

3652

von HUENE, MALLOY, SHOR, AND St .-AM AND

continuous for a few kilometers but are rarely
continuous for distances longer than 15 km
[Miller, 1961]. Structures in upper Tertiary
and Recent strata do not appear to have so
great an extent as structures in Mesozoic and
older Tertiary rocks.

Where structures can be identified with some
confidence in successive closely spaced seismic
profiles, they show significant variations along
strike. The longest continuous fault observed in
sediments at the seafloor has a 20-km-long
landward-facing scarp. Sometimes well devel-
oped faults change character and die out be-
tween profiles only 4 km apart. One large
broad fold southwest of Middleton Island
changes character significantly in an adjacent
profile made several kilometers away. Continuity
between structures in adjacent profiles is indi-
cated in Figures 1 and 2 only where outstand-
ing similarity or very close profile spacing
establishes such a continuity with a good degree
of certainty. It is possible that similar sequences
of structures on some adjacent profiles may be
continuous, because structures continuous at
depth may deform only short segments of the
shallow strata clearly displayed in the seismic
records. The present evidence suggests a few
long tectonic features and many relatively short
ones, but a much more detailed survey would
be required to establish the length and strike of
individual faults and folds.

Where the strike of structures can be estab-
lished, it seems to follow the regional structural
trend. In Prince William Sound, along Hinchin-
brook and Montague islands, between Montague
Island and Kodiak Island, as well as along
Kodiak Island, tectonic structures appear to
strike approximately parallel to the continental
margin and the Aleutian trench (Figure 1).
Off the Copper River delta, however, where the
regional trend changes direction fairly rapidly,
subparallel structures striking nearly east and
also northeast are separated from each other
by only short distances, indicating an abrupt
change in direction rather than a continuity
of two structural trends. The seismic reflection
and bathymetric data indicate extreme struc-
tural complexity in this area, similar to that
established in the Katella district by Miller
[1961].

Tectonic structures plotted in Figure 1 sug-
gest a moderately deformed continental shelf.

What Figure 1 fails to show is the relative size
of each structure and any differences in age.
Also, because a greater concentration of data
occurs in some areas, these areas appear to
contain the greater number of faults. A few
critical areas, including the ones off Pt. Elring-
ton and east of Afognak Island, have little or
no coverage; however, Plajker [1965] records
no sharp changes in uplift near Pt. Elrington
to suggest vertical fault displacement during the
1964 earthquake.

The largest structures found are closely
grouped near the landward ends of most tran-
sects across the continental shelf. Large struc-
tures sometimes occur at the edge of the con-
tinental shelf also. When the positions of the
structures near land are seen in map view they
appear roughly aligned parallel to the regional
trend. Two possible large tectonic features are
suggested by the assembled information, espe-
cially when the relative quality of each record
and other conditions are taken into account:
(1) a zone of faults, well developed from south-
ern Montague Island to Portlock Bank, but less
clearly defined south of Portlock Bank and
north of Montague Island, and (2) a discon-
tinuous anticlinal structure with attendant
faulting at the seaward edge of the continental
shelf. The locations of these features are indi-
cated by shaded areas in Figures 2 and 3.

Selected simplified reflection records (Figure
3) show how groups of faults are separated
from the anticlinal structure at the continental
margin by an area with much less deformation.
Between southern Montague Island and Port-
lock Bank there is little doubt that the groups
of faults belong to a single fault zone. Present
data suggest that most faults are shorter than
30 km, the distance between most profiles in
this area. Along northern Montague Island,
where seismic transects are not so numerous,
faulting parallel to the shore was found, espe-
cially in the Hinchinbrook Entrance (see pro-
file A, Figure 3). Along northern Hinchinbrook
Island, Reimnitz [1966] inferred a large fault
near the shore where there is a steep contact
between upper Tertiary sediments of the shelf
and the lithified rocks of the island. Because
these faults are along the strike of the fault
zone off southern Montague, they are probably
an extension of this zone. Some faults of the
zone cross a platform of the harder rock that

AFTERSHOCK REGION OF ALASKAN EARTHQUAKE

3653

Fig. 2. Map of strain release and tectonic features. Strain release summed in 0.4° radius
circles. Contours are based on preliminary data from the U. S. Coast and Geodetic Survey
and are smoothed graphically. Dark shaded area is the zone of faults inferred from marine
geophysical data.

projects around southern Montague Island.
Along Hinchinbrook and Montague islands the
fault zone separates uplifted lithified and slightly
metamorphosed rocks of Prince William Sound
from a subsiding area of younger Tertiary and
Recent strata on the continental shelf. This
relation cannot be established from about 25
km southwest of Montague to just northeast
of the Kodiak group of islands, because only
shallow strata are shown clearly in the records,
but seismic refraction profiles off the northern
Kodiak Island [Shor, 1965] show the depression
of the lithified sedimentary rocks under the con-
tinental shelf. Along Kodiak Island the reflec-
tion seismic transects are very sparse, but two
features suggest that a fault zone similar to the
one along Hinchinbrook and Montague islands
follows the southeastern Kodiak coast. First, a
large fault system along the coast separates a
thick section of upper Tertiary sedimentary
rocks on its seaward side from the metamor-
phosed and lithified rocks that make up most
of the island. Secondly, it appears likely that a

thick sequence of sediments has accumulated
on the continental shelf, not only from the re-
fraction work [Shor, 1965] but also from the
profile south of Kodiak (Figure 3F), which
shows strata with an apparent thickness of
more than 4 km on the shelf. Thus, there is
evidence of a fault zone that may extend from
off Hinchinbrook Island southwest partly across
the platform off southern Montague Island, be-
coming a well developed series of faults at the
seafloor from southern Montague to Portlock
Bank where it again becomes more questionable
because of sparse data, and then joins the
Kodiak fault system, possibly extending into
the Trinity Islands. One edge of the zone may
come ashore at southern Montague Island and
also on Kodiak Island, from Narrow Cape south-
ward along the coast. Although the zone's
southern terminus is somewhat ambiguous, it
does not appear in one of the profiles between
the Trinity and Chirikof islands (Figure 3F).
The zone's northern terminus is probably in
the complexly deformed region around the

3654

von HUENE, MALLOY, SHOR, AND St.-AMAND

001 X S83J.3W Hld3Q U31VM

AFTERSHOCK REGION OF ALASKAN EARTHQUAKE

3655

Copper River, where two major structural
trends intersect.

Bathymetric expression of the fault zone is
generally discontinuous and weak. Along the
coastlines of Hinchinbrook, Montague, and
Kodiak islands an irregular seafloor suggests
faulting parallel to the regional tectonic trend,
but between Montague and Kodiak most bathy-
metric features trend across the fault zone.
Some of these are possible glacio-genetic troughs
trending south and southeast that begin in the
fiords of the Kenai Peninsula and extend across
the continental shelf [von Huene, 1966]. How-
ever, two major bathymetric trends of unde-
termined origin are athwart the fault zone. The
first is a large east-west embayment outlined
by the 100-fathom contour (Figure 1). Closely
spaced ship tracks and good records establish
the fault zone through this embayment at a
point where contours diverge from the embay-
ment's general trend. The second feature is
Portlock Bank, where the ship tracks are sparse
and deep reflections were obscured by multiple
reflections. A few sharp faults occur along the
profiles northeast and southwest of the bank.
The only profile to cross the bank does so
parallel to the regional trend. It is difficult to
distinguish tectonic features in the complex
sequence of poor reflections recorded here.

Without better control it is not possible to
establish with certainty whether structures
strike parallel to the bank's bathymetric trend,
and whether the fault zone crosses it. Close con-
touring of bathymetry shows, however, an in-
flection in the trend of the bank's crestal ridge
where the fault zone would project across it.
In PDR profiles of this area a rough bottom is
indicated. Thus, some evidence on and near the
seafloor suggests a continuity of the fault zone
through this feature, but it is probably best to
consider this continuity with reserve.

During the earthquake the fault zone did not
rupture in a single or continuous series of breaks
at the surface; in some of the transects across
the zone no surface scarps were found. This
has probably been true for any earthquake of
Recent age along it, because in some areas the
faults are buried by probable glacial and post-
glacial sediment (Figure 3C; and von Huene
[1966]). The profiles suggest that faults in the
zone are discontinuous from 200 to 500 meters
deep, and thus the fault zone may have behaved

in the past just as it did in the 1964 earth-
quake.

Seismic profiles across the fault zone exhibit
deformation of Quaternary and possibly upper
Tertiary strata. The most frequently encoun-
tered deformational geometry is exemplified by
Figure 4A. The northwest flank of this structure
is composed of reflectors whose dip increases
with depth, indicating progressive continuing
deformation over a long period of time. Local
unconformities show that the crest of the fold
was sometimes elevated to wave base or above
sea level. The steepest limb of the asymmetric
fold is always on the southeast side; associated
fault scarps always face seaward. The structure
shown in Figure \A changes amplitude signifi-
cantly in 4 km (profile B, Figure 4). No sharp
surface expression occurs in profile A (Figure
4), but a 10-meter scarp on the seafloor is
found 4 km away in profile B. In general, this
type of structure appears to have a limited
extent along strike. Profiles in the fault zone
frequently cross two or three of these struc-
tures at relatively close intervals, suggesting
overlapping shorter faults.

In no instance have we been able to deter-
mine by direct observation the fault plane asso-
ciated with this type of structure ; however, the
data indicate a 60° to vertical attitude of the
fault plane. The fold geometry would be diffi-
cult to produce without a reverse component
of slip along a steeply dipping fault that is
inclined landward.

Parkin [1966] has probably the best evidence
of strike slip along the fault zone in the re-
established quadrilateral across the zone be-
tween Montague and Middleton islands. Middle-
ton Island appears to have moved southwest
about 10 meters relative to Montague Island
and the mainland in the period between 1933
and 1965. The discontinuous nature of faults
in the zone suggests strike slip, but discon-
tinuity is not a feature unique to such faults.
Although there are indications of a strike com-
ponent of slip, especially for the 1964 earth-
quake, it does not appear to be the dominant
slip component.

Another type of structure that always has a
sharp surface expression and a narrower zone
of deformed reflectors on either side of the fault
occurs (profiles E and F, Figure 4). It is ex-
pressed in the records as a sharp fold that must

3656

von HUENE, MALLOY, SHOR, AND St.-AMAND

N W

SE

SECONDS
00 0

E

***]

5^==

===^— ^

VERTICAL EXAGGERATION
12 X

VERTICAL EXAGGERATION
6 X

Fig. 4. Tracing of reflection seismic profiles across structures in the fault zone and at the
continental shelf edge. A to C are near 58°50'N latitude and 149° 10' W longitude; B is 6 km
southwest of A; C is 4 km southwest of B. Inset D was recorded at 57°55'N latitude and
149°20'W longitude; inset E. at 60°05'N latitude and 146°20'W longitude; and inset F at
59°35'N latitude and 147°30'W longitude.

be fault controlled. The scarps, some of which
have a 40-meter relief, face both northwest and
southeast, or both landward and seaward. These
structures are sometimes longer (at least 20 km
in one instance) and apparently younger than
the first type, because the deepest strata are
deformed about as much as the seafloor. In
other words, the faulting has occurred in a short
period of time, and no erosion or sedimentation
has modified the seafloor since faulting occurred.
The sharp flexure and continuity along strike
may indicate a thinner sequence of strata
through which stresses from the basement are
transmitted to the seafloor. Because they have
only been found in shallower water, the seismic
records of these structures are obscured by
multiple reflections that generally begin 120
meters below the seafloor. The bedding below
the multiple is probably, however, also parallel

to the seafloor, because enough seismic energy
was used in making some of the profiles to give
strong reflections from beds below the first sea-
floor multiple; if inclined beds were present,
their reflections should have been seen on the
records as divergent reflections. Since only the
shallow parts of the records are clear, it is
difficult to determine the fault plane attitude.
Both the first and second type of structures
have approximately the same strike, and they
are found 10 km or less from each other. It
may be that both types of structures are pro-
duced by reverse faults, and that some of these
fault planes dip seaward. One cannot, however,
exclude the possibility of shallow normal fault-
ing ; normal faults are not reported from studies
on land but occur only 30 km away, near the
continental margin. The absence of deformation
in the presumed incompetent strata immedi-

AFTERSHOCK REGION OF ALASKAN EARTHQUAKE

3657

■ately adjacent to the fault plane may be evi-
dence of extension, but some reverse faults show
little deformation of adjacent strata.

Faults with landward or northwest-facing
scarps are too numerous to be dismissed as
local variants, but their significance must re-
main in question until further investigation.

Relation of the fault zone and the zone of
maximum aftershock-strain release is shown in
Figure 2, where strain is represented in terms
of the equivalent number of magnitude 3.0
earthquakes, as was done by Allen [Allen et al.,
1965]. Each end of the aftershock strain re-
lease field has two distinct maxima. The largest
occurs off Hinchinbrook Island and spreads into
the area of structural complexity off the Copper
River. West of this maximum is a smaller one
which occurs in an area of few data and of no
known, large, active tectonic feature. At the
southern end the eastern maximum is near
Albatross Bank, where an anticlinal structure
occurs at the continental margin. This sug-
gests deformation during the earthquake along
the southern part of the continental margin.
Other evidence suggests that uplift occurred
along the continental margin, such as the up-
lift of Middleton Island [Plafker, 1965] and the
apparent origin of the tsunami at the edge of
the continental shelf [Van Born, 1965] ; how-
ever, significant strain release is indicated only
at Albatross Bank.

A linear area of greater strain release be-
tween these termini corresponds very well to
the fault zone between Kodiak and Montague
islands, suggesting that the zone defined here
was deformed during the earthquake.

Well developed segments of the fault zone and
the zone of maximum after-shock-strain release
are located in the same area where an axis of
maximum uplift and an axis of most severe
flexure in the crustal surface may occur [Plafker,
1965]. The positions of these axes in the uplifted
area can only be inferred because most of the
uplifted area is under water. It is reasonable to
propose that surface rupture and maximum
flexure are coincident in deforming a homo-
geneous plate; however, the fault zone existed
before the 1964 earthquake, and it may have
been a convenient zone of weakness for surface
rupture. Although no exclusive case can be
argued at present, the hypothesis that maximum
uplift, strain release, surface rupture, and

flexure occurred along or near the fault zone
in 1964 is probably the most acceptable at this
time.

Profiles A and F (Figure 3) show the con-
tinental margin structural high. Generally this
structure is a broad deformed anticline which
occasionally affects much of the continental
shelf. Although it is absent in some profiles and
thus appears to be discontinuous at moderate
depths, it may be a more continuous feature
deeper in the section, since a fairly continuous
gravity high occurs along the whole continental
margin [Barnes et al., 1966].

Discussion and Summary

The marine data indicate a zone of steep
faults located where the greatest aftershock
strain release occurred after the Alaskan earth-
quake. Surface fault rupture, as well as inferred
maximum flexure and uplift along this zone, are
further evidence linking the fault zone and the
earthquake. The regional tectonic setting and
the geometry of the fault zone therefore indi-
cate the type of faults that may have been
active near the surface in 1964.

The steep faults suggested by marine data
are also seen along subaerial parts of the Kodiak
fault system. Capps [1937] observed that faults
along the southeast coast of Kodiak Island are
nearly vertical and are accompanied by local
severe folding, distortion, and alteration. Moore
(in Grantz [1964] and personal communica-
tion) describes the Kodiak fault system as a
high-angle fault system separating Tertiary
from metamorphosed Mesozoic rocks with fault
planes dipping either landward or seaward. On
southern Montague Island Plafker [1965] (and
Plafker in preparation, 1967) reports a 55°
northwest to vertical dip on faults that rup-
tured at the surface in 1964. Therefore the
proposed fault zone is expressed (1) by a
system of faults dipping steeply northwest and
southeast that separate Tertiary from meta-
morphosed Mesozoic rocks along Kodiak Island,
(2) by reverse faults dipping between 60° and
90° northwest, as well as by southeast-dipping
faults that deform Recent and probably upper
Tertiary strata on the seafloor northeast of
Kodiak Island, (3) by scarps in the hard sedi-
mentary rocks of the platform around and on
southern Montague Island, and (4) by de-
formed Recent strata and by steep faults

3658

von HUENE, MALLOY, SHOR, AND St.-AMAND

separating Recent strata from hard rocks off
Montague and Hinchinbrook islands. Along
Hinchinbrook, Montague, and the Kodiak group
of islands this fault zone occurs at the separa-
tion of uplifted Mesozoic and lower Tertiary-
rocks, and a subsiding thick sequence of upper
Tertiary to Recent sediments. This relation be-
tween rocks of different age cannot be estab-
lished between Montague and Kodiak islands but
at present there is no evidence to suggest that it
does not occur. Elsewhere in the Prince William
Sound area the many shear zones that dip
steeply northwest and southeast are apparently
inactive [Condon and Cass, 1958; Condon,
1965; Richter, 1965; von Huene et al., 1967].
The marine and land data show that an inter-
section of two fault systems occurs at the north-
eastern area deformed by the 1964 earthquake.
East of Kayak Island and Controller Bay an
east-west-trending system of faults is dominant,
and from Hinchinbrook Island to Unimak Pass
a northeast-trending system is dominant. The
data at sea do not delineate this intersection
as clearly as does Condon's mapping in Prince
William Sound [Condon, 1965], where the
northwest- and east-west-trending Chugach
thrust faults apparently terminate the south-
west-trending Hinchinbrook set of traces. The
east-west-trending faults are generally exempli-
fied by the Chugach-St. Elias fault along which
Mesozoic and older (?) crystalline rocks are
thrust over Tertiary sedimentary rocks along
30° to 55° north-dipping fault planes. Appar-
ently thrust faults characterize the east-west-
trending system, and steep dipping faults char-
acterize the northeast-trending system.

It is important to maintain the distinction
between these two systems when considering
deformation accompanying the 1964 earth-
quake, because an interaction between the two
fault systems could occur with one system re-
maining essentially passive. Plafker's contours
of elevation change swing sharply and extend
east for 160 km at the northeast end of the
Kodiak zone, indicating limited and possibly
passive deformation along the east-west-trend-
ing system. Press and Jackson [1965] and
Savage and Hastie [1966] apparently used some
elevation changes from the east-trending lobe
of the deformed surface and projected them on
a plane normal to the northeast-trending fault
system to approximate the deformation they

used in applying dislocation theory to model the
fault plane on which the 1964 earthquake oc-
curred. Thus, the assumed surface they used
probably descends too gradually in its southeast
segment. In a more advanced fault model based
on dislocation theory, Stauder and Bollinger
[1966] have used a revised profile determined
by Plafker across Prince William Sound, which
is more correct than profiles used in previous
dislocation theory solutions.

The deformed surface has been well estab-
lished by Plafker across Prince William Sound.
From the maximum uplift along Montague
Island across 80 km of water to Middleton
Island, however, there are no available data.
If deformation during the 1964 earthquake was
similar to a summation of late Tertiary defor-
mation, which is one of subsidence, the de-
formed surface may have subsided seaward of
Montague Island. Conceivably, the surface
could bow gradually downward over a mildly
deformed subsiding area between Montague and
Middleton islands and then arch again over the
tectonically active anticline at the shelf edge,
which rises briefly out of the water at Middle-
ton Island. The validity of fault models using
dislocation theory will remain at least as ques-
tionable as the certainty of the surface of
deformation seaward of Montague Island.

Two possible fault planes have been deter-
mined from the main shock bodywave data
(Algermissen, 1964, in Plafker [1965] ; Stauder
and Bollinger [1966]). One fault plane is verti-
cal or dips 80° southeast, and the auxiliary
plane is horizontal or dips 10° northwest.
Whether the earthquake occurred along the
steep or the gently dipping plane cannot be
determined from the marine seismic profiles
because the attitude of the deep zone (prob-
ably more than 10 km) on which initial de-
formation occurred is not necessarily the same
attitude as that of the steep faults in the upper
1 km of the crust. Press [1965], Plafker [1965],
Savage and Hastie [1966], and Stauder and
Bollinger [1966] have discussed the merits of
both planes. Observations from the work re-
ported here that are relevant to the Alaskan
earthquake are as follows :

1. The discontinuous surface rupture on, and
south of, Montague Island and the possible
short seafloor ruptures elsewhere along the fault

AFTERSHOCK REGION OF ALASKAN EARTHQUAKE

3659

zone are probably typical of surface deforma-
tion during past earthquakes along the zone.
Parts of the fault zone have little, if any,
bathymetric expression. This may suggest in-
frequent events such as the one in 1964 or
possibly suggest local rapid sediment accumula-
tion. There is a Tertiary and Quaternary his-
tory of uplift at the edge of the continental
shelf, and at least part of this area around
Middleton Island was elevated during the 1964
earthquake.

2. The proposed fault zone contains struc-
tures of various ages, not all of which may have
been active in Recent time. In some areas of
sparse data the continuity of the fault zone is
not definitely established. Along at least more
than half of its proposed trace, however, the
zone occurs between uplifted older highly de-
formed rock and a probable subsiding area of
the continental shelf. Thus it is probably an
important tectonic element of the continental
shelf.

3. There is a rather close correspondence in
location gf the fault zone and the area of maxi-
mum aftershock-strain release between Mon-
tague and Kodiak islands. Aftershock-strain
release is rather evenly distributed around the
surface trace of the fault zone, suggesting a
major readjustment of the crustal stress field
in the volume of rock immediately below its
surface trace.

4. Between shore and the continental margin
is an area that has been subsiding during the
upper Tertiary. But in models of the causative
fault based on dislocation theory, uplift of this
area is inferred. This may cast an element of
doubt on the basic assumptions for the model,
but it should be noted that during the earth-
quake a reversal of the geologic history of net
uplift in the Kenai and Kodiak mountains
occurred.

Probably the most difficult feature to explain,
if the continent is overriding the ocean basin,
occurs off the continental margin. It has been
found (von Huene and Shor [1966], and in*
preparation, 1967; Hamilton, in preparation,
1967) that a thick sequence of horizontal strata
in the Aleutian trench as well as the strata of
the adjacent abyssal plains, are impressively
undeformed by compressional folds or faults.

Acknowledgments. We wish to thank David

Barnes and Erk Reimnitz for letting us use some
of their unpublished data. George Plafker and
George Moore reviewed the manuscript and their
suggestions greatly improved the paper.

References

Allen, C. R., P. St.-Amand, C. F. Richter, and
J. M. Nordquist, Relationship between seis-
micity and geologic structure in the southern
California region, Bull. Seismol. Soc. Am., 55(4),
753-797, 1965.

Barnes, David F., W. H. Lucas, E. V. Mace, and
R. J. Malloy, Reconnaissance gravity and other
geophysical data from the continental end of
the Aleutian arc (abstract), Proc. Am. Assoc.
Petrol. Geol., 1966.

Capps, S. R., Kodiak and adjacent islands, Alaska,
U. S. Geol. Surv. Bull, 880C, 111-184, 1937.

Condon, William H., Map of eastern Prince Wil-
liam Sound area, Alaska, showing fracture
traces inferred from aerial photographs, U. S.
Geol. Surv. Misc. Geol. Investigations Map
1-458, 1965.

Condon, William H., and John T. Cass, Map of
a part of the Prince William Sound area,
Alaska, showing linear geologic features as
shown on aerial photographs, U. S. Geol. Sur-
vey Misc. Geol. Investigations Map 1-278, 1958.

Grantz, Arthur, in Mineral and Water Resources
of Alaska, Committee Report, 88th Congress,
2nd session, p. 59, 1964.

Malloy, R. J., Crustal uplift southwest of Mon-
tague Island, Alaska, Science, .7.46(3647), 1048-
1049, 1964.

Malloy, R. J., Seafloor upheaval, Geo-Marine
Tech., 1(6), 22-26, 1965.

Miller, Don J., Geology of the Katalla District,
Gulf of Alaska Tertiary Province, Alaska, U. S.
Geol. Surv. Open File Rept., 1961.

Parkin, Ernest J., Alaskan surveys to determine
crustal movement, 2, Horizontal displacement,
U. S. Coast and Geodetic Surv. Rept., 1966.

Plafker, George, Tectonic deformation associated
with the 1964 Alaska earthquake, Science,
148(3678), 1675-1687, 1965.

Press, Frank, Displacements, strains, and tilts at
teleseismic distances, J. Geophys. Res., 70(10),
2395-2412, 1965.

Press, Frank, and David Jackson, Alaskan earth-
quake, 27 March 1964: vertical extent of fault-
ing and elastic strain energy release, Science,
147(3660), 867-868, 1965.

Reimnitz, Erk, Late Quaternary history and sedi-
mentation of the Copper River delta and vi-
cinity, Ph.D. dissertation, University of Cali-
fornia, San Diego, 1966.

Richter, D. H., Geology and mineral deposits of
Central Knight Island, Prince William Sound,
Alaska, Alaska Div. Mines Minerals, Rept. 16,
1965.

Savage, J. C, and L. M. Hastie, Surface deforma-
tion associated with dip-slip faulting, J. Geo-
phys. Res., 71(20), 4892-4904, 1966.

3660 von HUENE, MALLOY, SHOR, AND St.-AMAND

Shor, George G., Jr., Seismic refraction studies off von Huene, Roland, and George G. Shor, Struc-
the Coast of Alaska 1956-1957, Bull. Seismol. ture of the continental to oceanic transition in
Soc. Am., 62(1), 37-57, 1962. the Gulf of Alaska, Proceedings oj the 11th Pa-
Shor, George G., Jr., Structure of the Aleutian cific Sci. Congress, Tokyo, 1966.
ridge, Aleutian trench, and Bering Sea (ab- von Huene, Roland, George G. Shor, and Erk
stract), Trans. Am. Geophys. Union, 4^(1), 106, Reimnitz, A geologic interpretation of seismic
1965. profiles in Prince William Sound, Alaska, Bull.
Stauder, William, and G. A. Bollinger, The focal Geol. Soc. Am., 78, 4, 1967.
mechanism of the Alaskan earthquake of March von Huene, Roland, George G. Shor, and Pierre
28, 1964, and its aftershock sequence, J. Geo- St.-Amand, Active faults and structures of the
phys. Res., 71 (22), 5283-5296, 1966. continental margin in the 1964 Alaskan after-
Van Dorn, W. G., Tsunamis, Advan. Hydrosci- shock sequence area (abstract), Trans. Am.

ence,2, 1-47, 1965. Geophys. Union, 47(1), 176, 1966.
von Huene, Roland, Glacial marine geology of
Nuka Bay, Alaska, and the adjacent conti-
nental shelf, Marine Geology, 4, 4, 1966. (Received December 20, 1966.)

Reprinted from THE AMERICAN ASSOCIATION OF PETROLEUM
GEOLOGISTS BULLETIN Vol. 51, No. 9

27

ISLAND ARC SYSTEM IN ANDAMAN SEA1

L. AUSTIN WEEKS,2 R. N. HARBISON,2 and G. PETER2

Silver Spring, Maryland

ABSTRACT

A sub-bottom profiler survey of the Andaman Sea was conducted as part of the marine geo-
physical program of the U. S. Coast and Geodetic Survey during the International Indian Ocean
Expedition. The survey lines were run at right angles to the predominantly north-south tectonic
lineations of the island arc system.

A 20,000-joule sparker, energized every 4 sec, was used as a sound source and was towed about
100 m behind the ship. A 20-hydrophone array received the reflected signals, which were recorded
both on a paper strip chart and magnetic tape.

The sub-bottom profiler sections and the marine gravity and magnetic measurements augmented
knowledge of the geology of the island arc system; the linear structural belts on land were traced
through the Andaman Sea by geophysical methods. The structural development of the island arc
system from east to west can be traced, based on available continental and marine data.

It was possible to delineate the major segments of the island arc system through a distance of
600 nautical mi (1,110 km) in the Andaman Sea, specifically, the foredeep, outer sedimentary island
arc, interdeep, inner volcanic arc (and rift valley), and backdeep.

Introduction

As part of the International Indian Ocean Expe-
dition the U. S. Coast and Geodetic Survey ship
Pioneer conducted continuous sub-bottom profiler
surveys in the Andaman Sea. These surveys were
designed to study the nature of the great Indone-
sian island arc system between Sumatra and
Burma, and to show the possible interrelations of
these areas as common members of a single great
structural geologic province.

Geologic History and Tectonic Development

The Indonesian arc developed on the landward
side of its associated submarine trench, a feature
which is typical of all arc-shaped island chains.
For this reason the origin of island arc trenches
is believed to be closely related to crustal move-
ment. This view is substantiated further by the
facts that earthquakes commonly occur along
trenches, and their foci deepen markedly land-
ward to depths greater than 200 mi. Volcanoes
also occur in parallel zones along many of the
trenches and lie approximately above the zone of
intermediate-focus earthquakes (landward of the
trenches). In the Andaman Sea volcanoes are
present on the landward side of the outer sedi-
mentary island arc.

The tectonic development and patterns of the

'Manuscript received, May 27, 1966; accepted, No-
vember 18, 1966.

2 Institute for Oceanography, Environmental Science
Service Administration; formerly with U. S. Coast
and Geodetic Survey, Rockville, Maryland.

Andaman Sea region are discussed in an east-to-
west direction (Fig. 1).

The Malay Peninsula is the tectonic continua-
tion of the eastern Burma north-south fold-moun-
tain system, which, at the southern end, swings
eastward, parallel with the island arc system, into
Borneo. The main fold axes in the southern pen-
insula trend slightly west of north, changing per-
ceptibly to east of north at the Thailand-Burma
border. These structural trends are at an acute
angle to those of the island arc and are nowhere
precisely parallel with them. The Mergui Archi-
pelago along the west coast of Burma is moder-
ately faulted and has been submerged slightly in
late geological time (Chhibber, 1934). The Ma-
lacca Strait and adjoining Sunda Shelf also were
submerged during Recent time.

A large fault, striking north-south through cen-
tral Burma, extends seaward into the Gulf of
Martaban. In the 1964 Pioneer survey the sub-
bottom profiler sections did not extend far
enough east to detect this fault under the Anda-
man Sea. However, the bathymetry of the eastern
Andaman Sea shelf and the magnetic observa-
tions suggest that it is present. Differences in
structural grain between the Malay Peninsula and
trends of the island arc on the west could be ex-
plained by such a fault, downthrown toward the
west.

The Malay peninsula came into existence dur-
ing the Mesozoic as a result of a series of di-
astrophic cycles during Triassic-Jurassic time.
The cycles radiated from an older center of

1803

1804

L. AUSTIN WEEKS, R. N. HARBISON, AND G. PETER

Fig. 1. — Map showing major structural trends in Andaman Sea area, southeast Asia.

ISLAND ARC SYSTEM IN ANDAMAN SEA

1805

orogeny on the east (Van Bemmelen, 1949). Evi-
dence for this Triassic-Jurassic orogeny includes
the folding of older sediments and intrusions of
granitic rocks during Triassic-Jurassic time.

The sedimentary oil-producing basins of Suma-
tra and Burma (backdeep) are interconnected
through the Andaman Sea, and are terminated on
the east by a fault or by the abruptly sloping
"basement rocks" of the Malay Peninsula. Ac-
cording to Krishnan (1960) the Andaman Sea
probably acquired its present shape at the end of
the Cretaceous.

West of the peninsula and the backdeep basin
zone is the Cretaceous folded belt or inner vol-
canic arc. This belt can be traced from central
Burma across the Irrawaddy delta, through Nar-
condam and Barren Islands and Invisible Bank,
into the volcanic Barisan Range of Sumatra,
thence through Krakatoa and the Indonesian is-
lands. The main orogeny, at the end of the Creta-
ceous, folded and thrusted the pre-Tertiary sedi-
ments toward the southwest. Uplift and emplace-
ment of batholiths followed. In Sumatra, the
whole length of the Barisan Range was elevated
during the Plio-Pleistocene (Van Bemmelen,
1949). The Semangko graben (rift) zone de-
veloped as a post-elevation collapse feature. The
whole inner volcanic arc comprises a positive
isostatic anomaly.

Across an intervening inner sedimentary trough
or interdeep, the next belt on the west is the non-
volcanic outer island arc. It can be traced from
the eastern Himalayan arc southward through east-
ern India, Burma, the Andaman and Nicobar Is-
lands, and the islands west of Sumatra. This is a
Tertiary fold belt, and forms the present-day outer
island arc. The rocks of this belt are predominantly
marine sediments which have been folded, faulted,
and uplifted. The belt is isostatically negative, in-
dicating a deficiency of mass — in direct contrast
to the inner volcanic arc.

West of the outer island arc is the foredeep or
trench. On the south, this feature is called the
"Java trench." As a morphological feature the
trench does not extend north of Simalur Island
(3°N.), nor is it present off the Andaman and
Nicobar Islands (Van Bemmelen, 1949). How-
eveif the 1964 Pioneer survey work indicates that
the trench is present as a buried structural fea-
ture off these islands. Vertical and (or) horizon-
tal movements in the northern part of the Anda-

mans are believed to have occurred earlier than
in the southern part. There is a definite gradation
from coarse to fine sediments in Eocene beds
from north, to south. Therefore, movements in
the Andamans and Nicobars are believed to pre-
date the equivalent belt farther south along the
Indonesian chain. Westward thrusting of the
outer island arc geanticline had more or less
ceased prior to the Miocene — as indicated by
Miocene sediments that rest unconformably on
older rocks and are hardly folded (Van Bemme-
len, 1949). Subsequent elevation of the outer is-
land arc during the Quaternary has resulted from
vertical uplift combined with a post-glacial rise in
sea-level.

Van Bemmelen (1949), in his classic synthesis
of the geology of Indonesia, ascribed variations
along various parts of the same structural belt to
the fact that different orogenic centers or foci
were involved. He believed that the Andaman Sea
belts developed from a different orogenic focus
than did the areas south and north. Van Bemme-
len also concluded that northern Sumatra (Atjeh
section), near which the Pioneer ran several track
lines, belonged to the same orogenic system as the
Andaman Sea.

Sub-Bottom Profiling Results

Sub-bottom profiling was done along five sec-
tions that cross the structural belts of the island
arc system in the Andaman Sea (Fig. 2). Figure
3 shows sections 1 and 2. Figure 4 shows sections
3, 4, and 5. Section 5, the southernmost, is
confined to the backdeep. Section 4, just north of
Sumatra, extends from the backdeep to the axis
of the outer island arc. Section 1, a discontinuous
profile, extends from the backdeep to the fore-
deep along the Ten Degree Channel just north of
Car Nicobar Island. Section 3, the northernmost,
extends from the backdeep in the vicinity of the
Tenasserim coast of Burma across the submerged
Irrawaddy delta to a point just south of Preparis
Island, and then northwest across the outer island
arc.

The sections, as interpreted, show form lines of
the structure and approximate thicknesses of
sedimentary layers. Some of the faults indicated
on the sections are clearly observed on the sub-
bottom profiler records. Others are inferred be-
cause the attitude of the beds and the structural-
ly complex geology of the area make their iden-

1806

L. AUSTIN WEEKS, R. N. HARBISON, AND G. PETER

ANDAMAN SEA

15

10'

B E

5°

Compiled By

90c

95°

Fig. 2. — Island arc elements of Andaman Sea. Locations of sections 1-S (Figs. 3, 4) shown. Locations of
Figures 5-8 also can be determined from this figure.

ISLAND ARC SYSTEM IN ANDAMAN SEA

1807

ANDAMAN SEA - SECTION 1

METERS
800

1400

2400

3200

4000

NW

284 283 261 260 276 276 274 272 27f 270

Z£^^FSS?^^-

269 268 267 266 265 264 263 262 261 260 258 257 256 255 253 252 251 250 248 246 245

tfl 212 210 208 206 205 204 202 200 199 198 196 195 194 193 192

METERS
1600

2400

3200

4000

4800

800
1600
2400
3200

122 120 118 117 116 115 113 112 111 110 109 108 107 106

FAULT ZONE

SE

92 91

90

89

67 66

85 84 63

*r*

\^_^y^^_

yOr^V-

qT

^7—

WFv

NW

SECTION 2

SE

INNER VOLCANIC ARC

Fig. 3. — Sections 1-2, Andaman Sea island arc system. Locations shown on Figure 2.

tification less certain. Fault planes and the direc-
tion of movement along faults commonly are
indefinite.

The horizontal scale of the interpreted sections
is based on the ship's position fixes. These were
made at 30-min intervals while traveling at a
speed of about 5 knots. Hence, each fix point
marked on the sections is separated from the ad-
jacent fix by a 30-min time interval — regardless
of the number assigned to it — and the horizontal
scale is approximately 5 nautical mi between
every second fix mark. Analysis and interpreta-

tion of the results preceded the receipt of adjust-
ed fix locations by many months, but the
differences between the true or corrected track
line (as on Fig. 2) and the section track lines
(Figs. 3, 4) are relatively slight.

Penetration depths are uncorrected for sound
velocity. The scale of the sections is based on a
velocity of 1,600 m/sec, or 5,248 ft/sec. Where
"lower velocity" sediments overlie or are adja-
cent to "high-velocity" rocks, any particular ve-
locity assumption is incorrect. A higher estimated
velocity would increase the thicknesses of sedi-

1808

L. AUSTIN WEEKS, R. N. HARBISON, AND G. PETER

ANDAMAN SEA - SECTION 3

NW

41 939 937 935 933 931 929 927 925 923 920 9)9

SECTION 4

sw

620 62 1 622

626 627 629 630 631

INNER VOLCANIC ARC

657 656 659 660 661 662 663 6«4 665 666 667

669 670 671 672 673 674 675 676

1600
2400

BACKDEEP (EXTENSION OF SUMATRA OIL BASIN)

=¥

3^i

FAULT ZONE

SECTION 5

sw

20 119 118 117 116 114 112 111 110 109 107 106 105 104

NE

1600
2400

FAULT ZONE

Fig. 4. — Sections 3-5, Andaman Sea island arc system. Locations shown on Figure 2.

ISLAND ARC SYSTEM IN ANDAMAN SEA

1809

?•*'<.«£' 5v->;^i£r ,;"'•;;;

! • ,; ■ w; :•{•

Fig. 5. — Section 4, fix 639, southwest-northeast profile. Sediments on southwest (left) are seen to
disappear against hard bedrock or volcanics toward northeast. Mass on northeast is western block of inner
volcanic arc, supposedly a horst block. A fault is postulated at right side of picture, upthrown toward north-
east. For location, see Figures 2 and 4.

merits penetrated, perhaps as much as SO per cent
in some cases. Bottom depths change greatly
throughout the Andaman Sea; therefore instru-
mental scale shifts were frequent. The results of
the sub-bottom profiler surveys are discussed
from south to north.

NORTHERN SUMATRA— SECTIONS 4 AND 5

The two profiles just north of Sumatra were
made in order to cross known trends of major
proportions. Section 4 (Fig. 4) begins on the east
flank of the outer island arc, which plunges steep-
ly into the interdeep. Clear evidence of sedimen-
tation is lacking, possibly because of the rugged-
ness and steepness of the slope. However, bed-
ding and some indications of faulting can be seen
at the top of the arc. Simalur Island, on the
south, is a geanticline cut by north-south-trending
transverse faults (Van Bemmelen, 1949). From
the interdeep to fix 639 several unconformities
are detectable. Sediments Southwest of fix 639
(section 4, Fig. 4) disappear against hard bed-
rock or volcanics on the northeast (Fig. 5). The
mass on the northeast is the western block of the
inner volcanic arc — a horst. A fault is postulated
at fix 639, section 4 (Fig. 4), upthrown toward
the northeast.

The part of section 4 from fix 639 to 648 is

the offshore continuation of the pre-Tertiary and
lower Tertiary block-mountain system of north-
ern Sumatra. The block is represented by the
western side of the Atjeh graben and the islands
west of the Bengal Passage. There is a lack of
obvious bedding on the records, probably because
the rocks are dense. On Sumatra, these rocks are
primarily Permo-Carboniferous sediments (un-
doubtedly metamorphosed), diabase, and serpen-
tinites (Geologic Maps of Netherlands Indies,
1927). The Atjeh graben, part of the Semangko
fault zone which can be traced the entire length
of Sumatra, shows up very clearly. It is consid-
ered to be a relaxation feature after the Plio-
Pleistocene uplift in the Barisan (Van Bemmelen,
1949). On Sumatra the graben is filled with Neo-
gene (Pliocene and Miocene) sediments which
are overlain by Quaternary and alluvial deposits,
and extends into the Bengal Passage. East of the
graben in northern Sumatra, the rocks are primar-
ily volcanics (post-lower Tertiary?), identified as
andesite effusives, or are block-mountain struc-
tures similar to those on the west side. The whole
belt from fix 639 to fix 656 represents the inner
volcanic arc, including the Semangko fault zone.

Northeast of the inner volcanic arc is the back-
deep. This is the Andaman Sea extension of the
Sumatra oil basin. Folding and faulting are evi-

1810

L. AUSTIN WEEKS, R. N. HARBISON, AND G. PETER

Fig. 6. — Section 4, fix 671, southwest-northeast profile. Unconformity, showing bottom beds rising toward
northeast and overlying beds wedging out. Bottom beds reflect one side of an arch whose peak reaches bottom
surface beyond picture (right). For location, see Figures 2 and 4.

dent. In this area and also in Sumatra, structural
complexity tends to decrease away from the inner
volcanic arc. The zone between fixed 660 and 665
is more contorted than are zones on the north-
east. Fixes 665 to 673 are in a broad synclinal
trough, although the bottom rises gradually to-

ward the northeast (Fig. 6). Beyond fix 673 and
northeastward to the large fault at fix 675 (Figs.
4, 7), sub-bottom reflections disappear (Fig. 6).
The writers interpret this to mean that the "base-
ment" or volcanic rocks are faulted against ap-
proximately 1,600 m of sediments. The fault does

Fig. 7. — Section 4, fix 675, southwest-northeast profile. Large fault downthrown toward northeast (right).
Southwestern block consists of "basement" or volcanic rocks faulted against approximately 1,600 m of sedi-
ments. Note how basal sediments dip into fault, whereas youngest beds drape across fault and onto "basement"
or volcanic block. Small buried channel can be seen above fault plane, as well as larger buried channel on
right. For location, see Figures 2 and 4.

ISLAND ARC SYSTEM IN ANDAMAN SEA 1811

"■'W^&rB^i .:- ':■ -■■ v* ,,■ ' "• ■'' ^^rn^./'rU '■' J yi"f$W:- ■:"■$■'*'■■ ■.■•"•w-?,

fklij lli^iU.^'t'^&Tj" il -' .*. *K.U i»L I

Fig. 8. — Section S, fix 100, northeast-southwest profile. Fault with about 200 m of displacement, down-
thrown toward southwest (right). About 1,000 m of penetration can be seen here. Fault plane does not quite
reach ocean floor. For location, see Figures 2 and 4.

not reach the sea floor, which commonly is the
case within the areas surveyed in the Andaman
Sea. Northeast of the fault, the basal sediments
dip into the fault, whereas the upper beds overlap
the fault and the "basement" or volcanic rocks.
The records show an old channel above the fault
plane; this channel probably was cut into the
softer materials at that place.

Section 5 (Fig. 4) is about 50 mi (92 km)
southeast of section 4. The purpose of this sec-
tion was to investigate the backdeep, to include
trends found on section 4, and to search for the
fault which occurs at the end of section 4 on the
north (fix 675). The southwestern part of the
section shows sub-bottom folding and faulting in
rocks that lie unconformably beneath a thin ve-
neer of surface deposits. The surface veneer is
relatively flat, except at fix 118. Between fixes 99
and 100 a large fault zone was observed (Fig. 8).
The large fault found on section 4 (fix 675) does
not appear to be present. A similar relationship
of "basement" or volcanic rocks to sediments was
not found on section 5; consequently its extent
and true significance are not known.

Northeast of fix 99 the remainder of the back-
deep shows a folded sub-bottom section, in part

unconformably overlain by surface deposits. At
the northeast end of the section, the bottom and
sub-bottom deposits are conformable and rise
onto the Andaman Sea shelf.

TEN DEGREE CHANNEL — SECTION 1

Section 1 (Fig. 3), the first profile run in the
area, was interrupted frequently in order to make
oceanographic observations at various stations.
Consequently, this profile is not continuous and
some important features were missed. The sec-
tion from fixes 245 to 284 includes structural fea-
tures of the western outer island arc (Nicobar
Islands), west of the Ten Degree Channel. From
fix 272 to fix 284, the records indicate sedimenta-
ry blocks that are thrust westward against the an-
cient foredeep. The bathymetry west of fix 284
indicates gradual shoaling. The greatest depths
are adjacent to the thrust blocks. The Java
trench (foredeep) terminates considerably south
of this area and a deep trench does not appear on
any crossings of the foredeep. However, the
deepest water is everywhere adjacent to the west-
ern limit of the outer island arc. It seems proba-
ble that later sedimentation filled the foredeep
west of the Andaman and Nicobar Islands.

1812

L. AUSTIN WEEKS, R. N. HARBISON, AND G. PETER

From fix 245 to the thrusted blocks at fix 278
the nature of the western flank of the outer is-
land arc is evident. The shelf has a rugged and
youthful appearance, both structurally and topo-
graphically, which one would expect from an is-
land mass which has emerged so recently. Bottom
depth increases approximately 2,400 m between
fixes 245 and 284, in a distance of 92 nautical mi
(170 km).

There is a 16>^-mi gap (31 km) in the record
between fixes 245 and 212. Both fixes lie in the
same structural trend along the outer island arc
and can be considered to be in structurally equiv-
alent positions. The youthful and structurally
complex area from fix 192 to fix 212 is the east-
ern flank of the outer island arc. Bathymetric
data show its continuation a short distance south-
east of fix 192. The total width of the outer is-
land arc in this sector is about 100 mi (185 km).
The maximum width of Car Nicobar Island, near
which the Pioneer passed, is 7 mi (13 km). The
sea-floor base of the outer island arc is many
times this width. East of the outer island arc the
bottom is rough and the sub-bottom structure has
a youthful appearance. Just south of this traverse
the Nicobar Islands are broken up into several
groups with trends that apparently extend north-
ward.

Fixes 192 and 122 are 30 mi (55 km) apart.
Sub-bottom profiles were not obtained between
these fixes, but continuous depth soundings were
made.

Bathymetric data indicate that the bottom rises
markedly to less than 180 m depth just southeast
of fix 192. The bottom then plunges abruptly to-
ward the interdeep where depths of about 4,100
m were observed. It rises again toward the west-
ern block of the inner volcanic arc west of fix
120. The bottom of the graben valley (Semangko
rift) is between fixes 116 and 117 at about 4,800
m depth. This is deeper than the interdeep by ap-
proximately 700 m. Lack of sub-bottom penetra-
tion prevented delimiting the graben valley be-
tween the faults. Also, the survey line did not
cross structural trends at right angles. The result
of this is that the graben has a greater apparent
width in this section. The peak of the eastern
block of the inner volcanic arc appears at fixes
84-86. The western block lies between fixes 120
and 192, as is the interdeep. The backdeep lies a
short distance east of fix 83.

Ritchie's archipelago — section 2
Section 2 (Fig. 4) was run south of Barren Is-
land and extends west to Ritchie's Archipelago,
just east of the Andaman Islands. This section
and section 4 show very well the structure of the
inner volcanic arc and associated rift valley. Fix
680 is above the east flank of the outer island
arc. The interdeep is evident, with about 800 m
of sediment in the structural trough. The peak at
fix 672 lies on the western part of the inner vol-
canic arc and is approximately half way between
Invisible Bank and Barren Island. Both the bank
and the island are part of the inner volcanic arc
(western flank). The area between fix 672 and
658 is a continuation of-the Semangko rift valley
which occurs both on the island of Sumatra and
offshore (section 4). A large fault may be present
at fix 660 along the steep slope, but the records
are inconclusive. Tipper (1911) believes that Bar-
ren and Narcondam Islands have emerged along a
master fault zone east of the Andaman Islands.
However, both Barren and Narcondam Islands,
as well as Invisible Bank, form only the west side
of the inner volcanic arc.

Southeast of fix 658, the profiler and bathymet-
ric data indicate the presence of another ridge-
like area. Sediments are more abundant than on
the flanks of the volcanic arc. Except for a possi-
ble small intrusive at fix 651, volcanic or base-
ment-type rocks appear to be lacking. This ridge
does not extend very far either north or south of
this section.

IRRAWADDY DELTA — SECTION 3

Section 3 (Fig. 4), the northernmost traverse
made in the Andaman Sea, crossed the submerged
extension of the Irrawaddy delta from the Tenas-
serim coast of Burma. Just south of Preparis Is-
land the east-west traverse was turned northwest
to its termination.

The western part of section 3 shows the west-
ern slope of the outer island arc. At the western
limit, the slope plunges steeply seaward in the
foredeep area. Faulting and folding are not no-
ticeable along the smooth sea floor of this slope.
A smooth sea floor is typical of the northern tra-
verse of the Andaman Sea, even where the under-
lying structure is complex. A small channel at fix
933 is the only indication of sub-bottom structure
controlling bottom topography between fixes 923
and 935. Sub-bottom profiling shows the structur-

ISLAND ARC SYSTEM IN ANDAMAN SEA

1813

al complexity of the wide belt from fix 939 to fix
909, a distance of 38 mi (70 km). This is the
outer island arc, whose apex apparently is at fix
920. East of fix 919 the sub-bottom features are
even more complex. Here, younger sediments lie
unconformably on the older folded rocks of the
outer island arc. Perhaps this is the pre-Miocene
unconformity, described by Van Bemmelen
(1949). Thrusting presumably ended before Mio-
cene time in the outer island arc. The younger
beds undoubtedly are part of the massive Irra-
waddy delta. The Irrawaddy River deposits
670,000 tons of silt per day on its rapidly build-
ing delta (Chhibber, 1934). Protection from the
open sea and lack of effective longshore currents
favor the advance of the delta toward the south.

Fixes 908 to 888 show the gradual descent of
the bedrock to the interdeep at about fix 895, and
gradual rise of the bedrock to the inner volcanic
arc at fix 888. The deepest or basal reflection is a
composite, and should not be interpreted as a
continuous reflection. Single beds cannot be
traced for long distances. The largest free-air
gravity value along the entire traverse occurs at
fix 888 (plus 40 mgals). The extent of the tra-
verse is covered by deltaic deposits, and lack of
sub-bottom penetration beneath the bedrock pre-
vents interpretation of the underlying features.

East of fix 886 equipment failure prevented
obtaining data for 16 mi (30 km). Hence it was
impossible to identify the rift valley (Semangko
fault zone), which may be buried by overlying
sediments. The inner volcanic arc does not appear
to be as well developed here as it is farther south
(sections 2 and 4). Attenuation of sparker energy
through the soft overburden masks some fea-
tures, but the main reason for poor development
of island arc structure may be that the volcanic
arc has more mature topography at the northern
end. There are at least two indications that the
island arc becomes younger toward the south:

(1) the infilling of the former trench west of the
outer island arc to such an extent that it is only
barely indicated by the bathymetry, even though
structurally apparent in the sub-bottom; and

(2) the southward gradation of Eocene sediments
on the Andaman and Nicobar Islands from
coarse terrestrial to fine marine (Van Bemmelen,
1949).

East of fix 887 the beds appear to dip gently
east with a few faults. The surface beds lie un-

conformably on older sediments. Whether all the
reflections are from delta deposits is uncertain. If
the basal beds are of deltaic origin, the faults, al-
though older than the surface sediments, must be
very recent. Between fixes 857 and 862 an inter-
esting sub-bottom structure was developed. Its
interpretation is not clear. The structure may be
a folder section of the backdeep or the eastern
segment of the inner volcanic arc, whose twin na-
ture was noted on the south. Free-air gravity and
magnetic intensity values increase above this fea-
ture. Thin surface beds unconformably overlie
the folded sediments.

If this folded section is the eastern block of
the inner volcanic arc, then there is considerable
divergence of structural trends north from the
Ritchie's Archipelago traverse (section 2, Fig. 3).
Off Sumatra and Ritchie's Archipelago, the dis-
tance between segments is about the same, about
20 mi (37 km). On the Irrawaddy delta traverse
the segments may be separated by as much as 61
mi (113 km), or three times the separation far-
ther south. However, it appears more likely that
the eastern segment, if still in existence, is in the
16-mi (30-km) area of no records between fixes
876 and 886.

The remainder or eastern part of the traverse
has no distinctive features, and contains gently
folded beds, presumed to be delta deposits. The
records contain many multiple reflections which
may mask some structural features.

Description of Structural Belts

The study of sub-bottom profiles, bathymetry,
gravity, and magnetic measurements in the Anda-
man Sea makes it possible to describe the various
structural belts of the island arc system. Region-
ally, these are, from east to west, the following.

BACKDEEP

The typical structure of the backdeep is best
shown in sections 4 and 5 (Fig. 4). In section 3
(Fig. 4), which also traverses the backdeep, the
sub-bottom features are masked by the cover of
deltaic sediments. In general, the structural fea-
tures of the backdeep are less complex than those
of the belts on the west, but some large anti-
clines, synclines, and fault zones are present. The
largest anticline was recorded off the Sumatra
coast, between fixes 100 and 105 of section 5. Its
width is about 14 mi (26 km). This feature could

1814

L. AUSTIN WEEKS, R. N. HARBISON, AND G. PETER

Table I. Similarities of Andaman Sea Inner

Volcanic Arc and Central Part of

Mdd- Atlantic Ridge

A rtdaman Sea Inner
Volcanic Arc

Mid-Atlantic Ridge

Volcanic rock types: serpentine,
diabase, andesite

Earthquake belt: zone of epicen-
ters (200 km)

Prominent rift valley

Adjacent high mountains

Peaks above sea-level: Narcon-
dam and Barren Islands

Considerable relief between rift
valley and adjacent peaks

Comparable width of rift valley:
20-25 mi, peak to peak

Depth to floor of valley ranges
from 5 000 to 15,000 ft where
crossed by pro61er

High length to width ratio

Volcanic rock types: serpentine,

diabase, basalt
Earthquake belt: zone of epicen-
ters (25 km or less) (G. H.

Sutton, pers. comm.)
Prominent rift valley
Adjacent high mountains
Peaks above sea-level: Iceland to

Bouvet Island
Considerable reliei between rift

valley and adjacent peaks
Comparable width of rift valley:

15-30 mi, peak to peak
Depth to floor of valley averages

12,000 ft

High length to width ratio

be of impressive size — depending on its extent
normal to the traverse. Numerous faults were re-
corded across this anticline, which appears to ter-
minate against a fault on the northeast, at fix
100. A smaller anticline is present between fixes
111 and 115. This feature is about 7 mi (13 km)
wide and also is faulted. In the backdeep part of
section 3, two anticlinal structures were observed,
one \2y2 mi (23 km) wide between fixes 845 and
851 and the other 8 mi (15 km) wide between
fixes 858 and 861. The rocks in both are uncon-
formably overlain by flat-lying sediments.

The backdeep consists primarily of sedimenta-
ry rocks. In most places, the sedimentary sections
are thicker than 800 m. Suspected volcanic or
"basement" rocks were found only between fixes
673 and 675, section 4. The greatest water depths
at which the backdeep complex was observed
were at fix 658 of section 4 (about 1,900 m)
and fix 656 of section 2 (about 2,400 m), both
adjacent to the inner volcanic arc.

INNER VOLCANIC ARC AND SEMANGKO RIFT VALLEY

The inner volcanic arc (with its associated
Semangko rift valley) was the most interesting
structural feature observed in the Andaman Sea.
On land, the rift valley can be traced approxi-
mately 1,100 mi (2,040 km), the entire length of
Sumatra. During the 1964 Pioneer cruise, by
means of sub-bottom profiling, it was traced for
the first time about 600 mi (1,110 km) farther
north through the Andaman Sea. Free-air gravity
values also indicated the presence of the volcanic
arc beneath the sediments of the Irrawaddy delta
(Peter et al., 1966), but the sub-bottom structur-

al detail along the traverse in this area (section
3) could not be resolved beneath some 400 m of
sediments.

The width of the rift valley, as determined by
the east-west traverses between northern Sumatra
and Ritchie's Archipelago, is 5-10 mi (9^-18^
km) between the bounding ridge crests. The val-
ley relief, between adjacent ridge crest and valley
floor, is 1,200 m off Sumatra (section 4) and
2,000 m at section 2. The western ridge of the
inner volcanic arc rises higher above the sea floor
and is a more massive feature than the eastern
ridge. Narcondam and Barren Islands and Invisi-
ble Bank are part of the western structural ele-
ments of the inner arc system.

The magnetic and gravity results in the Anda-
man Sea area provide further evidence of the
submarine continuation of the inner volcanic arc
(Peter et al., 1966). A broad magnetic high belt
extends from the tip of Sumatra approximately
600 mi north to the Irrawaddy delta traverse
(section 3). This belt broadens northward to
Narcondam Island and then narrows before it
crosses section 3. Gravity highs in excess of 50
mgal occur off the tip of Sumatra and at Invisi-
ble Bank, Barren Island, and Narcondam Island.
The largest free-air values, in excess of 100 mgal,
were observed at Invisible Bank and Barren and
Narcondam Islands — a value of more than 150
mgal being present at the south end of Invisible
Bank. Interesting similarities between the Anda-
man Sea inner volcanic arc and the Mid-Atlantic
Ridge are listed in Table I. The Mid-Atlantic
Ridge is a much longer structural feature and is
not part of an island arc development, but the
rift valley and many related tectonic features
are common to both the Mid-Atlantic Ridge and
the Andaman Sea inner volcanic arc.

INTERDEEP

The interdeep is the structurally depressed belt
between the two major uplifts of the island arc
system. Its appearance is similar along the north-
south extent of the arc except in the extreme
north where deltaic deposits have masked many
features. Off Sumatra (section 4) approximately
700 m of flat-lying sediments fill the depression
(interdeep) between the two arcs. The eastern
slope of the outer island arc is steeper than the
western slope of the inner volcanic arc and may
have been the source of most of the sediment. On

ISLAND ARC SYSTEM IN ANDAMAN SEA

1815

the south, the interdeep separates Sumatra and
the offshore Mentawai Islands. The interdeep was
not profiled in the Ten Degree Channel (section
1), but is indicated by the bathymetry on several
other traverses made by the ship.

Off Ritchie's Archipelago on the east flank of
the outer island arc in section 2, the narrow and
downwarped interdeep is filled with sediments
that appear to have been deposited from both
flanks and folded in the form of a syncline as the
downwarping continued. In section 3 off the Irra-
waddy delta, the broad interdeep lies between
flanks with very slight slope, and is filled with
sediments that mask its topographic expression on
the sea floor.

OUTER ISLAND ARC

The outer island arc is composed predominant-
ly of sedimentary rocks. It is the youngest arc of
the whole Andaman Sea structural system. It
continues north of section 3 as the Arakan Yoma
(mountains) of western Burma and finally abuts
against the eastern Himalayan arc along the
Burma-India border. Toward the south, the outer
arc includes the Mentawai Islands off Sumatra
and the submarine ridge on the landward side of
the Java trench.

The outer arc is structurally complex. It has
the overall features of a large anticline. Its width
in places exceeds 100 mi (185 km). Near the Ir-
rawaddy delta the outer arc narrows before en-
tering Burma, but toward the south, between
Narcondam and Barren Islands, a distinct east-
ward bulge is indicated by the bathymetric data
obtained by the Pioneer. The sub-bottom profiles
of sections 1 and 3 suggest that westward thrust-
ing of the outer arc took place. The apparent
thrust ridges of section 1 can be traced north-
ward by means of the bathymetric data. The
presence of other ridges with long, straight, east-
dipping slopes suggests westward thrusting.

Gravity measurements show a negative free-air
anomaly approximating the peak axis (line of
highest elevations) of the outer island arc (Peter
et al., 1966). The peak axis of the arc and the
negative gravity axis coincide exactly at the south
end of the index map (Fig. 2). At 8°00' N. and
93°30' E. the negative gravity axis veers east of
the Nicobar Islands. It continues east of the islands

and rejoins the peak axis of the outer islands
where the negative gravity axis enters Burma. The
trend labeled "subsidiary or main mass axis" on
Figure 2 is the negative gravity axis north of 8°
N. A belt as wide as the outer island arc and
tilted westward would be expected to have its
main mass axis lying east of its surface peaks
(Nicobar and Andaman Islands). The zone of
westward thrusts on section 3, Ritchie's Archipel-
ago (southwest of Barren Island), and the clus-
ter of islands at 8°00' N. and 93°30' E. are con-
sidered to be parts of the subsidiary trend, coin-
ciding with the negative free-air anomaly.

Results of Survey

The geologic-geophysical investigations of the
1964 Pioneer Indian Ocean Expedition provided
much new data and information about the geolog-
ical features of the Andaman Sea, particularly the
submarine tectonic patterns and crustal develop-
ment of the area. The sub-bottom profiles made
it possible to delineate the major segments of the
island arc system through a distance of more
than 600 mi (1,110 km) and provided informa-
tion about the structural relations of the base-
ment and overlying rock complexes in the major
structural belts — foredeep, outer sedimentary is-
land arc, interdeep, inner volcanic arc and associ-
ated rift valley, and backdeep. Continued detailed
analysis of the geophysical data obtained in the
Andaman Sea area and correlation of the 1964
results with results of other surveys can be ex-
pected to yield additional discoveries of scientific
significance.

References Cited

Chhibber, H. L., 1934, The geology of Burma: Lon-
don, Macmillan and Co., 538 p.

Geologic Maps of Netherlands Indies, 1927, Scale
1 :1, 000,000: compiled by J. V. Scrivenor.

Krishnan, M. S., 1960, Geology of India and Burma,
4th ed. : Madras, Higginbothams (Private), 604 p.

Peter, G., L. A. Weeks, and R. E. Burns, 1966, A re-
connaissance geophysical survey in the Andaman
Sea and across the Andaman-Nicobar island arc :
Jour. Geophys. Research, v. 71, no. 2 (Jan.), p.
495-509.

Tipper, G. H., 1911, Geology of Andaman Islands,
with references to Nicobars: Mem. Geol. Survey
Ind.v. 35, pt. 4, p. 195-213.

Van Bemmelen, R. W., 1949, Geology of Indonesia, v.
1A: The Hague, Govt. Printing Office, 732 p.

28

ESSA-JTRE-2

HIG-67-16

DOUBLE-HUMPED WAVES ON A SLOPING BEACH

By
JAMES P. BUTLER

UH-3SA iOiNT TSUMAMI RESEAJtOf EHOW
HAWAII WSTITOT! OF CEOfflYSKS
2525 COBBEA ROAD
HONOLULU, HAWAII 96822

AUGUST 1967

Prepared for

ENVIRONMENTAL SCIENCE SERVICES ADMINISTRATION

INSTITUTE FOR OCEANOGRAPHY

JOINT TSUNAMI RESEARCH EFFORT

with the
HAWAII INSTITUTE OF GEOPHYSICS

HAWAII INSTITUTE OF GEOPHYSICS

UNIVERSITY OF HAWAII

ESSA

ESSA-JTRE-2 HIG-67-16

DOUBLE-HUMPED WAVES ON A SLOPING BEACH

By
James P. Butler

Prepared for

Environmental Science Services Administration
Institute for Oceanography
Joint Tsunami Research Effort

with the
Hawaii Institute of Geophysics

Approved: August 17, 1967

Chief, Director,

Joint Tsunami Research Effort Hawaii Institute of Geophysics

CONTENTS

Page

Abstract .

1. Introduction

2. Preliminary Equations

3. Specific Solution .

4. Propagation of the Wave

5. A Numerical Example

6. Acknowledgements

7. References ,

1

2

2

7

19

26

27

28

iii

ABSTRACT

The nonlinear shallow water theory in two dimensions has been
used to compute the behavior of selected wave shapes on sloping beaohes.
These waves all are initially at rest with two humps of specified
heights. An analytic solution is found for the behavior of the wave
height and velocity at the shoreline. Particularly interesting are the
amplifications of the respective humps 3 when related to the wave shape.
Numerical techniques are used to find the general behavior of the wave
off the shoreline.

-1-

-2-

1. INTRODUCTION

In recent years interest in the behavior of water waves on sloping
beaches has increased sharply. As is known, however, the conventional
linearized approach to the study of water waves in the deep ocean is
invalid in the coastline regions and one must turn to the nonlinear
theory for treatment of the problem. Some general two-dimensional
treatments have appeared in the literature (Carrier and Greenspan,
1958; Carrier, 1965). With a uniform sloping beach, explicit solutions
exist from which arbitrary wave shapes can be constructed. In this
paper we consider two-dimensional double -humped wave shapes in the non-
linear theory and their behavior on the beach.

2. PRELIMINARY EQUATIONS

The notation and development given in this section is taken from
Carrier and Greenspan (1958). The conservation equations in the non-
linear shallow-water theory are

[v*(Tf* + h*)]^ = - Tt*t* (1)

and

v*.* + v*v*x# = - gTl*^ (2)

where the asterisks denote dimensional quantities; v* is the
horizontal velocity, Tl* is the height of the surface above mean sea
level, h* the height of mean sea level above the bottom, x* the

-3-

the horizontal distance with origin at the mean shoreline. For our
case, we have a uniform sloping beach, or

h* = - ox*

where a is the slope of the beach. Following the convention,
dimensionless quantities are introduced:

v = v*/v x = x*/A

t = t*/T T) = T\*/a&

c2 = (h* + T\*)/otl

where

1
v = (g£ «)*

T = (&Q/ctg)'

and & is a characteristic length for a specific problem. With
these substitutions, (l) and (2) become

[vcn - x)]x + \ = o O)

Vt + Wx + \ = ° (U)

-4-

and the situation applies to a beach of unit slope. By Stoker (19U8),
these two hyperbolic equations can be transformed into four
equations in the parametric variables a and P :

Xp - (v + c)t =0 vp + 2c& + tp = 0 (5,6)

x - (v - c)t = 0 v - 2c + t = 0 . (7,8)

From (6) and (8), after integration, we have (noting the convenient
choice of integration constants):

v + t = (a - p)/2 (9)

c ■ (o + p)/U . (10)

We define two final independent variables a , X as

X = a - P = 2(v + t) (11)

a = a + p = Uc (12)

and (5) and (7) become

x - vt + ct. = 0 (13)

a a A.

*X + Cta " vtX = ° (lU)

-5-

or by eliminating x

°(\X ~ toa) - 3ta = ° • <15)

Also, by (11)

o"K, - v ) - 3v = 0 . (16)

v XX aa a

We now introduce a potential $(a,\) such that by definition

v =^$a(a,X) , (17)

and after some manipulation the expressions for the other variables become

x = $ /k - a2/ 16 - v2/2 (18)

1) =$,/!+ - v2/2 (19)

t = | - v , (20)

and $ itself satisfies

KA -CT*u = 0 • ('21>

-6-

We now have a linear set of equations in a and X , and T| and v
are implicit functions of x and t . The shoreline is now the
line a = 0 . The waves do not break as long as the implicitly-
defined functions T](x,t) and v(x,t) are single -valued, i.e.,

Ifc4\ ^ 0 for a > 0

The solution of (2l) is given by

$ = AJ (u)ct ) sin(uuX + 6) (22)

o

and since equation (2l) for $ is linear, we car superimpose
solutions as

CO

$ = [ dk k A(k) J (ka ) sin(kX + 6) (23)

J O

o

whence

CO

= 1$ _ ll [ dk k2 A(k) Jn(ka) sin(kX + 6) . (2U)

a a a J 1

v = — $

But we will require our boundary conditions to include an initial wave
shape with zero velocity at zero time. For v = t = 0 , (2) implies
that X = 0 ; the condition on v is satisfied if in (2k) we set
6=0. So we have finally

-7-

$ = J dk k A(k) JQ(ka) sin(kX) (25)

where A(k) must be found by the boundary conditions. Now at
t=0,v=X=0,so Tl and x become:

Tl(o,0) =1 [dkk2 A(k) Jo(ka) (26)

x(a,0) = Tl(a,0) - a2/l6 . (27)

From (26) we obtain

00

^k A(k) = J da a T)(a,0) J (ko) (28)

o

and by specifying 1](x) parametrically in a , one can solve the
problem (at least in theory).

3. SPECIFIC SOLUTION

We now turn our attention to a class of double-humped wave shapes
given initially by (t = v = 0):

2
11(0,0) = eaU(2a - o2f e_pa (29)

x(a,0) = 11(0,0) - a2/l6 . (30)

-8-

2
This form was chosen since for small e , x goes like - o and

T) behaves like ex (a + 8x) e . In this formulation, the

product (ap) is the critical number. It approximately determines the

relative heights of the two humps. Crest-spacing can be changed with

a and also with H for a specific problem. Some examples of T|

versus a and T\ versus x are shown in Fig. 1. The equation (28)

has a closed-form expression:

00

2

A*(k) = I k A(k) = e f dao5(2a - a2)2 e~pa J (ka)
*+ j o

o

= , [bQ + b2k2 + b^ + b6k6 + b8k8]e"k2/UP (3D

where

b m -% (12 - 12ap + Ua2p2)
o 5
P

b2 = ^ (- 12 + 9ap - 2a2p2)
P

2 2
1^29 , a p x

P

b6^(-s + i>

"8-^9 ' (32)

-9-

Now that we have an explicit solution for A(k) , the expressions for
1) , x , v , and t are all integrals involving this A(k) . For
0=0 (shoreline) these integrals have closed-form expression. It
is significant to investigate fully the shoreline characteristics of
run-up position (amplification) and velocity. For a = 0 , equation
(2k) becomes

00

v = - I J dk k3 A(k) sin k\

o

CD

= - 2 J dk k2 A*(k) sin kX
o

= - 2 ( lQ + i2 + ih + i6 + i8) (33)

where the I*s correspond to the five terms in A* (k) . By
Erdelyi, et al. (195*0, we obtain, for n = 0,1,2,3,U:

ha = e I e"PX ^n 1F1 (" I " n> 2> ^ (3h)

where

^n = b2n(Up)n+ (n + 1)1 (35)

or, written out completely:

-10-

Fig. 1. Initial wave height T] versus a ; a = .5 , e = .001.

-11-

Fig. 1. (con't)

-12-

v =

e\e-pX~[bo(UP)2 1 '. lFl(-| , | , p\2)

+ b2(Up)3 2 I XF (- | , | , p\2)

+ \(4P)U 3 ! ^ (-| , | , PX2)

+ b6(iip)5 i+ • lFl(-| , | , pX2)

+ b8(Up)6 5 I ^(-f , | , p\2)] , (36)

and this gives v as a function X for the shoreline a = 0 . A
similar treatment of (19) yields T| :

00
2

11 = - \ + ^ | dk k2 A(k) cos kX
o

00

= - ~ + f dk k A (k) cos k\
o

= - \ + \ ee"p^[bo(Up) 0 1 lFl(- \ , \ , Px2)

+ b2(i+P)2 1 • ^C- 1 , I , Px2)

+ bu(UP)3 2 i lFl(- 1 , § , pX2)

+ b6(Up)U 3 ! -^(-f , \ , PX2)

+ b8(l+P)5 1» • lFl(-| , § , pX2)] ,

(37)

-13-

and this gives T) as a function of X for the shoreline a = 0 .
Some typical plots of T\ and v against X are shown in Fig. 2.
By (20) we can find the time corresponding to each value of X .
Such a table of values is the implicit representation of T|(t) and
v(t) . The same cases as shown in Fig. 2. are shown in Fig. 3«»
plotting T| and v against t . It is easily seen by (20) that time
can run "backwards" if v(X) increases faster than X/2 at any point
X . Such a case would correspond to the wave breaking, and the
solution must be discarded. However, an inflection point in the function
t(X) is a useful criterion for determining values of e , a , and p
such that the wave will almost, but not quite, break.

We now turn to the aspect of amplification. The initial heights
of the waves are known from the boundary conditions. The maximum run-up
heights of the respective humps can be read off from Figs. 2. and 3«
Their ratios determine the amplification factor of each of the two humps.
By numerical investigation it was found that the amplification is essen-
tially independent of a and p for a constant product ap . As long
as e is small enough so that the wave does not break, the amplification
is strictly independent of e , since T|(a,0) is proportional to e
and Tl (0,X) is proportional to e (the v /2 term is zero at Tl ) .

For values of T)(0,X) other than extremum values, T| is practically

2 2\

proportional to e , as v /2 is 0(e). The ratio of the initial

heights is similar to the amplification, in that this ratio is

essentially independent of a and p for a constant ap . However,

both the amplifications and the initial height ratios do vary with ap ,

as shown in Fig. U.

-14-

Fig. 2. Shoreline behavior (a = 0) of wave height Tl (solid line)

and velocity v (dotted line) versus X. ; a = .5 , c = .001.

-15-

Fig. 2. (con't)

-16-

Fig. 3. Shoreline behavior (a = 0) of wave height T\ (solid line)

and velocity v (dotted line) versus t ; a = .5 , e = .001.

-17-

Fig. 3. (con't)

-18-

10

1.5

2j0-

OjO

fication of Maward hump

smptificotiofi of thot% hump

-I I— I !_

1.0

0.5.

00

0 0

25

Fig. h. Amplification factors of the two humps, and the ratio of
initial heights of the two humps (expressed as a number
between 0 and l) versus the product ap .

We conclude that the amplification of the first hump (shoreward)
is essentially constant, but that the amplification of the second
hump (seaward) is strongly dependent on the height of the first hump.
The maximum amplification occurs when the seaward hump is 0.625 the
height of the shoreward hump. For ap between about 1.32 and 1.4l,
the second hump washes up higher even though it was smaller to begin
with. From (20) we see that the maximum value of e without breaking

can be determined by making t(\) have an inflection point at the

dv

(This can be found from Fig. 2.) This maximum

maximum value of

o\

e is shown in Fig. 5. plotted against ap . For values ap and e
below the curve, the wave does not break, and for values above the

curve , the wave breaks ,

-19-

03

02

e
E

Vf

0.1

00

00

Wove breaks in thit region.
Solutions arc invalid.

Wave does not break in
this region.

■ '

1.0

15

2.0

Fig. 5. The maximum value of e for which the wave will not
break versus the product ap .

k. PROPAGATION OF THE WAVE

Our discussion of the wave behavior in the (o",\) plane has been
confined to determining the (0,X) behavior given the (a,0) shape.
This solves the problem for a particular shape with zero initial velocity.
The amplification figures may be questioned, as the stationary initial
shape might be represented as a superposition of two traveling waves,
each with half height. For that part rf the wave which is incident on
the beach this might imply amplification factors of twice the previously
quoted values. In order to investigate these questions, one must solve

-20-

for the general (a,X) behavior, particularly for large values of

X . Unfortunately, for our expression A(k) , the integral (25)

and (18, 19) with the integral form of $ have no closed-form solution

for arbitrary (o,X) , and must be solved numerically. Further, the

o

approximate expression for T](a,X) is (neglecting the v /2 term which

is 0(e2)):

00

fl(a,X) = 5 J <*k k2 A(k) J0(to) cos *&
o

00
= e J dk P(k)e"k /kp JQ(ka) cos kX (38)

o

where

P(k) = k(b + bgk2 + b^k + b,k + bQk )

For large values of X , the integrand oscillates rapidly, making
numerical integration difficult. The method of stationary phase or
steepest descents for approximating this integral is impractical
since we suspect the integral to be largest when a is also large;
the wave being propagated in time (represented by X) travels out to
sea (represented by increasing the values of a). For large a and X
the integral oscillates rapidly with two frequencies, thus malting these
methods inapplicable and numerical integration even more difficult.

That the wave is propagated out to sea can be demonstrated
mathematically in two ways. By Whittaker and Watson (1962), we can
represent J as

-21-

00

J (ka) ■ - | dt sin(ka cosh t) . (39)

O TT J
C

Putting this in (38) and reversing the order of integration we have

00 00

2

T)(a,\) = e - f dt [ dk P(k)e~k ' p sin(ka cosh t) cos kX . (Uo)

O O

Setting t = a cosh t , we can rewrite (Uo) as

00 00

2

Tl(a,X) = e |J dt J dk P(k)e"k /Up •§ [sin k(x+X) + sin k (t-X) ] . (Ul)

o o

The k integral, as we have seen previously, is a sum of confluent

2
hypergeometric functions which will decay like exp(-p(T+X) ) and

2
exp(-p(T-X) ) . For large values of X , only the terms with

2
exp(-p(T-X) ) behavior will contribute significantly to the t integral,

and then only when (t-X) is small. Now (t-X) is small only when

a cosh t is near X . This tells us three things. First, for a

much larger than X , (a cosh(t) - X) ^ (a-X) is large and the integral

is small. For a near X , (t-X) is small for small t and the

integral has a significant non-zero value. For a small, (t-X) is

small only when t is large, but since t ~ ae /2 , not too much time

is spent by the t integral in this region where (t-X) is small. The

integral then has a significant non-zero value, but is much smaller

than for a near X . These qualitative arguments are intended to

demonstrate that for increasing values of time and therefore of X ,

the large values of T] occur for increasing a and therefore for

increasing distance seaward.

-22-

The second method of demonstrating propagation is the one that is
feasible for computerized numerical integration. One writes J as

J (to) = J (kcr ) - / -|- cos(ka - ? ) + J~T~ cos (ka - ?) (U2)

ox ov V TTka k ' Vnka h' v '

and the integral (38) for Tj is written as

7l(a ,X) = e j dk P(k)e_k /Up [jQ(ka) - J J~a cos(ka - £) ]cos kX

- e

LT1 + I2J

+ e J dk P(k)e"k ' P J-fj cos(ka - J) cos kX

(U3)

Now I. has had the oscillating character of J (ka ) taken out ror
1 o '

large values of a , and the result is an integral of a relatively
smoothly varying function times the rapidly oscillating cos kA . We
can integrate this numerically choosing the Ak such that nAkX = 2rc
where n is some integer. We expect this integral to be small. I„
on the other hand can be integrated analytically by writing

cos(ka - J ) cos k\ = |Jcos(k(a + X) - \) + cos(k(a - X) - J ) j

-i- I cos k(a + X) + sin k(cr + X)
2/2 L

+ cos k(a - X) + sin k(a - X) ] . W

-23-

We obtain (again) a stun of .F_ functions that behave like

2

2

e'pT± P (a,P,prJ) (U5)

where

t+ = a ± X (U6)

and a and (3 are characteristic of the specific term involved. Once
again we note that the t terms are small and that the most significant
contributions to H come from the t terms when a is near X
(meaning the wave is being propagated out to sea). The result of such

numerical and analytic integration of (38) is shown in Fig. 6. for a

p
specific case. Here the v /2 term is not ignored, the same arguments

as above being used to compute v(a,X).

We note that v(a,X) is strictly anti- symmetric in X and that
therefore T|(a,X) is symmetric. By (20),t(a,X) is strictly anti-
symmetric in X . Fig. 6. can therefore be used running time backwards
by replacing X by -X to obtain prior times, noting that Tl(a,X)
remains the same and v is replaced by -v , as one would expect.

We conclude this section with a remark on amplification. By the
series shown in Fig. 6. we can see the results of the superposition of
an incoming and outgoing wave, especially as it applies to the seaward
hump. Since the effects of amplification have already begun by the
time the two parts separate, it is difficult to assess what the separate
"heights" of the incoming and outgoing waves are in relation to the

-24-

Fig. 6. The wave height T) versus x for the wave shape given by
a = .5 , ap = 1.25 , e = .001. Note that t = X/2 + 0(e) ,
so that the above graphs for \ = 0., .2, ... , 3.0 are
approximately also for t = 0., .1 , ... , 1.5 •

-25-

X«1.6

X-1.8

X»2.0

X«2.2

X»2.4

X'2.6

2
0
:2
2
0
-2
.2
0
2 ><

2 5

0
t2
.2
0
-.2
2
0
2
2
0
■2
2
0
2

Fig. 6. (con't)

-26-

total height of the wave at t = 0 . The number l/2 seems plausible
enough by an inspection of Fig. 6., so we conclude that the amplifica-
tions recorded in Fig. h. should perhaps really be double their value.

5 . A NUMERICAL EXAMPLE

Suppose for example that one wished to consider two mounds of
water with equal heights of 5 feet. Further suppose that their separation
distance (crest to crest) is 10 feet, and that the bottom has slope a .
Unfortunately, in this formulation, we are not permitted to specify the
shore-to-trough distance independently of the crest-to-crest spacing
and the height ratio. For our shape here, the shore -to- trough distance

o

equals .5*+ the crest-to-crest spacing or 5.*+ x 10 feet. From Fig. h.
we see that ap = 1.33 and that the amplification of the first and
second crests will be, respectively, 1.7 and 2.3. If the wave does
not break, T\* has the same amplification characteristics as Tl since
T\* = at& Tl . If the wave is well behaved, then, we will have T|*
ultimately reaching 8.5 feet and 11.5 feet run-up for the first and
second humps, respectively. For a = .1 , maximum inundations are
85 feet and 115 feet. For ap = 1.3 and the admittedly arbitrary
choice of a = .5 , the initial l\/e is about 2.1 x 10 and their
x separation is about .116. Then from the definitions of the dimension-
less variables,

k

SL = 8.6 x 10 ft.
o

-27-

and

e = 2.76 x 10"3/a

From Fig. 5. the maximum value of e for our case is .13, so for
slopes greater than ,021 the wave will not break.

6. ACKNOWLEDGEMENTS

The author wishes to thank Harold G. Loomis, Gay lord R. Miller,
and Jimmy C. Larsen for their kind encouragement and helpful advice
in this work. The numerical computations were carried out at the
Statistical and Computing Center at the University of Hawaii.

-28-

7 • REFERENCES

1. Carrier, G. F., and H. P. Greenspan (1958), 'Water waves of
finite amplitude on a sloping beach," J. Fluid Mech. k, 97-

2. Carrier, G. F. (1966), "Gravity waves on water of variable
depth," J. Fluid Mech. 2h, 6hl.

3. Stoker, J. J. (1948), "The formation of breakers and bores,"
Commun. Pure Appl . Math. 1, 1.

h. Erdelyi, A., W. Magnus, F. Oberhettinger , F. G. Tricomi, Eds.

(195*0, 'Tables of integral transforms,'' Bateman Manuscript

Project, California Institute of Technology, 1j (McGraw-Hill

Book Company, Inc., New York).
5. Whittaker, E. T. , and G. N. Watson (1962), "A course of modern

analysis ," (Cambridge University Press, London).

29

ON THE CROSS-STREAM VARIATION OF THE fc-FACTOR

FOR GEOMAGNETIC ELECTROKINETOGRAPH DATA

FROM THE FLORIDA CURRENT OFF MIAMI

Frank Chew x

The Bissett-Berman Corporation, Santa Monica, California

ABSTRACT

A factor, k, is used to convert Geomagnetic Electrokinetograph (GEK) observations of
ocean currents to the current speeds. It is usually approximated in terms of the kinematic
structure of a current and is generally assumed to be constant across the width of a stream.
Recent direct transport and surface current measurements across the Florida Current off
Miami were compared with GEK profiles. There was a cross-stream change in the fc-factor
that, in location and magnitude, could account for double speed maxima often observed
in GEK profiles across this stream.

INTRODUCTION

The measurement of ocean currents by
the Geomagnetic Electrokinetograph (GEK)
is subject to many possible errors ( Longuet-
Higgins, Stern, and Stommel 1954). To
correct for these errors, von Arx ( 1950 )
suggested the use of an empirical Zc-factor,
a ratio of the magnitude of the GEK output
to the actual current speed. For most cur-
rents, the major error is usually thought to
stem from the effect of the depth of the
current: The error increases as the current
depth increases relative to the depth of the
total water column. In terms of the fc-factor,
this is equivalent to its approximation when
one considers only the subsurface kinematic
structure relative to the surface current.
The fc-factor varies from region to region,
but for a given section of a stream it has
been generally assumed to be constant (see
Webster 1965). Because of practical diffi-
culties, it has not been possible to test either
the adequacy of the approximation or the
validity of the assumption. The purpose
here is to discuss a test of both in a related
but restricted context.

Many investigators have constructed
cross-stream profiles of the Florida Current
off Miami from uncorrected GEK observa-
tions. These profiles of the downstream
surface flow often show a double speed

1 Present address : Physical Oceanography Lab-
oratory, Institute for Oceanography, Environmental
Science Services Administration, Silver Spring,
Maryland.

maximum. An example of this feature is
shown in Fig. 1, along with samples from
the current in the Yucatan Channel and off
Onslow Bay as illustrations of its rather
widespread occurrence. When underway,
the GEK records the signal from the sea
continuously. However, the magnitude of
the signal takes its numerical origin from a
zero point determined, in practice, only at
discrete intervals. Because of the uncer-
tainty of the zero points between such de-
terminations or "fixes," the record of the
signal between fixes is usually not consid-
ered in profile construction. Further, the
spacing of the GEK fixes is often of the
order of the distance between the maxima.
Hence, good resolution of the feature can-
not be expected in all profiles. Similarly,
mean profiles constructed from GEK fixes
averaged over distances comparable to the
distance between the double speed maxi-
ma will tend to suppress it. This can be
demonstrated by using a method suggested
by Dr. A. Court and applied by Chew
(1965). Of the 623 GEK observations con-
sidered, the three most easterly observations
were discarded. The remaining 620 were
then divided into 20 groups of 31 each. The
pertinent GEK data and computation are
tabulated (Table 1) and feature a pro-
nounced midstream speed minimum in the
uncorrected GEK profile at zone 8. Elemen-
tary statistical tests of the difference be-
tween the means for zones 7 and 8 indicate
the difference is highly significant. This un-

73

74

FRANK CHEW

YUCATAN

MIAMI

ONSLOW BAY

200

160

120

Cm/sec

Fig. 1. Uncorrected, GEK cross-stream profiles of the surface current at three areas (the respective
sources are Texas A&M University, University of M iami, and Woods Hole Oceanographic Institution ) .

corrected GEK feature is suppressed if the
same data are grouped into 10 zones of 62
observations each. This appears to be the
case for the mean profile reported by Web-
ster (1965) for the Florida Current off
Miami. It seems, therefore, that the feature
is fairly persistent and is used as a conve-
nient focus to conduct a limited testing of
the adequacy of the approximation that the
fc-factor is primarily a function of the surface
current relative to the subsurface kinematic
structure. The validity of the assumption
that the fc-factor has no significant cross-
stream variation is also tested by asking
whether the cross-stream variation in the
approximate A>factor can account for the
biaxial feature in the downstream compo-
nent. The basic data for the test are those
reported by Richardson and Schmitz (1965),
who also generously supplied additional
unpublished observations. I am indebted
to Dr. J. A. Knauss for valuable suggestions
and to Dr. F. F. Koczy for his encourage-
ment.

THE fc-FACTOR

An approximation

In the absence of geomagnetic storms and
away from the edges of a current, Long-
uet-Higgins et al. (1954) show that errors
in GEK readings are due principally to
vertical variations of velocity, horizontal
velocity shear, and conductivity of the
underlying sea bed. The electrical effect of
horizontal shear is proportional to the com-
ponent of the earth's magnetic field trans-
verse to the direction of the component of
velocity of interest. In the Florida Current
off Miami, this is the east-west component
which, at this location, is fortuitously small
relative to the vertical earth component.
Thus, error from this source can be ne-
glected. The effect of sea bed conductivity
depends on the ratio of conductivity of
seawater to that of the sea bed. The effec-
tive conductivity of the sea bed underlying
the Florida Current has not been deter-
mined; it will be assumed that it is of the

CROSS-STREAM VARIATION OF THE fc-F ACTOR

75

same order of magnitude as that of the
English Channel; this gives a ratio of the
order of 103. For this ratio, the sea bed
conductivity is a critical factor if the width
of the current is 103 times its total depth
(see Longuet-Higgins et al. 1954). In the
Florida Current, the corresponding dimen-
sion ratio is an order less, and hence the
sea bed conductivity does not appear to be
critical. If this error is considered to be
negligible relative to that due to the vertical
kinematic structure, then for a given cur-
rent component, V, we have the approxi-
mation

Table 1. GEK data, grouped in 20 zones, with
equal number of observations in each zone (\c = 1)

k = Vs/(Vs-V)

(1)

where V (when the electrical conductivity
of the water is uniform ) is the mean speed
taken throughout the total water column;
Vs ( the speed at the surface ) is the quantity
the GEK is intended to measure; and
(Vg-V) is what the GEK does measure,
given the approximation. The GEK deter-
mines the velocity vector of the surface
current by measuring its components. Thus,
the fc-factor for each component is generally
different; an exception occurs only when
the velocity remains constant in direction
with depth. In applying equation ( 1 ) , care
must be taken to avoid situations where the
magnitude of V approaches that of Vs.
Since k can then no longer be reliably de-
termined, the effect of small errors in V
and Vs is greatly magnified. More impor-
tantly, the GEK in this situation would not
measure the electrical effect of ( V8 - V )
but instead, those effects that are neglected
in arriving at the approximation. Also, sit-
uations where V is greater than Vs must be
avoided, because equation ( 1 ) is then no
longer applicable; an example is seen in
the next section.

An evaluation

Richardson and Schmitz ( 1965 ) have
made direct current measurements of the
Florida Current off Miami; for the north-
ward component, their measurements are
accurate to within 1-2% of the absolute
magnitude. The use of their data to eval-
uate equation ( 1 ) gives the numbers tabu-

Zone

Longitudinal
interval

Mean
north
velocity
compo-
nent

(cm/
sec)

n

(cm/ sec)

1

80°05.4'-80

°04.3'

42.5

6.8

31

2

04.3'-

03.0'

69.0

8.6

31

3

03.0'-

00.8'

62.0

9.3

31

4

00.6'-79°58.8'

82.7

6.3

31

5

79°58.8'-

56.4'

90.2

8.0

31

6

56.3'-

54.5'

105.1

6.5

31

7

54.5'-

52.5'

106.1

4.9

31

8

52.5'-

52.4'

85.2

3.9

31

9

52.4'-

49.3'

117.0

5.5

31

10

49.2'-

47.1'

117.2

5.5

31

11

47.1'-

45.9'

103.4

3.3

31

12

45.9'-

44.0'

106.3

4.3

31

13

44.0'-

41.0'

98.8

5.9

31

14

41.0'-

38.0'

84.1

4.4

31

15

38.0'-

35.1'

63.8

7.4

31

16

35.0'-

32.0'

63.7

5.3

31

17

31.1'-

28.0'

49.8

5.1

31

18

28.0'-

25.5'

37.8

5.0

31

19

25.4'-

22.0'

31.4

5.6

31

20

21.8'-

19.5'

23.1

5.4

31

lated in column 7, Table 2, where identify-
ing information is also included. At all
locations Vs is greater than V, except at
location 83.32 km on 16 August 1964. The
factor k, as given by equation (1), is not
computed for this case. On this occasion,
a GEK measurement would indicate a
countercurrent; it is possible that past GEK
indications of countercurrents off Miami
were due to this cause and are spurious.
Omitting this case, the computed k values
range from 1.2 to 3.0 and average 1.9. Their
cross-stream distribution is plotted in Fig. 2.
It is apparent from Fig. 2 that k varies
across the stream, but its statistical signifi-
cance has not been determined, partly be-
cause the data are too few for a meaningful
test, and partly because of the neglect of
temperature and salinity effects in evalu-
ating V. Consideration of a least-squares
fitting to the plotted data leads to at least
two possibilities. One is a single line slop-
ing up eastward and spanning the entire
range corresponding to the width of the
current. Another is the two more or less
parallel lines that have been fitted as shown,
with the implication that they are joined

76

FRANK CHEW

Table 2. A computation for k

Date

Distance east
of Virginia
Key (km)

Depth

of sea

bed (m)

North— south
component
of transport

(m2/sec)

Surface cur-
rent north
component

(cm/sec)

Mean current

through total

depth, north

component

( cm/sec )

k*

Afcf

16 Aug 1964

8.55

54

1

13

2

1.2

0.03

16.35

289

157

215

54

1.3

0.04

26.28

349

352

253

101

1.7

0.11

35.47

688

533

250

78

1.5

0.07

44.26

774

701

218

91

1.7

0.12

53.55

807

735

177

91

2.1

0.22

62.90

839

615

119

73

2.6

0.41

72.51

763

397

80

52

2.9

0.54

83.32

382

269

31

70

17 Aug 1964

10.39

141

90

189

64

1.5

0.08

12.25

241

124

186

52

1.4

0.05

14.31

270

146

200

54

1.4

0.05

21.28

356

282

209

79

1.6

0.10

30.83

375

434

233

116

2.0

0.20

39.69

838

639

215

76

1.5

0.08

48.71

775

705

174

91

2.1

0.23

57.65

800

700

132

88

3.0

0.60

67.09

824

607

120

74

2.6

0.42

19 May 1965

32.4*

350

340

183

97

2.1

0.24

15 Mar 1965

38.6*

860

410

140

48

1.5

0.08

6 Apr 1965

64.4*

830

633

134

76

2.3

0.30

* From equation ( 1 ) .
t From equation ( 4 ) .
% Unpublished data supplied by Dr. E. W. Richardson.

by a third with a reverse slope in the cross-
stream interval between 32.5 and 36.0 km.
This fitting has three noteworthy features.
First, the overall scatter is smaller than a
single line fit. Second, between 32.5 and
36.0 km (the vicinity where the sea bed
drops from 400- to 800-m depth ) , the trend
of k reverses, losing in magnitude rather
abruptly almost all it gained in the preced-
ing interval. And, third, although the two
separate rising trends are unmistakable, the
scatter is least in the vicinity of the ob-
served current axis, possibly implying a
more variable kinematic structure on both
flanks of the axis. This second possibility
constitutes the hypothesis in the test con-
sidered in the section following.

In evaluating equation (1), we have
assumed a constant electrical resistivity for
the water. Had corresponding temperature
and salinity information been available, we
should have evaluated V according to the
equation

where r is the electrical resistivity of the
water and the integration of depth extends
from the sea surface, 0, to the sea bed, h.
Neglecting the small pressure effect, the
resistivity is least where temperature and
salinity are highest; for the Florida Current
low resistivity corresponds approximately
to the layers where V is large. Thus, the
magnitude of V in column 6, Table 2, is
an underestimate in all cases where Vs is
greater than V. To compensate, a positive
quantity, AV, representing the underesti-
mate, must be added to the computed V.
For a given Vs, the effect of AV on k is,
from equation ( 1 ) , to the same approxima-
tion

fc(AV)

Ak

(3)

— dz,
o r

(2;

(Vs-V)

From temperature and salinity sections, A V

is estimated to be Via of V. Using this,

equation (3) is found from

kV

Afc = =-, (4)

10(VS-V)

CROSS-STREAM VARIATION OF THE k-F ACTOR

77

a 16/8/64
x 17/8/64

\\ 19/5/64
• < 2 15/3/65

1 3 6/4/65

-a — FLORIDA CURRENT
OFF MIAMI 16/8/64
FROM RICHARDSON

GEK PROFILE FOR

k FACTOR ON LEFT

200

160

Cm/s

120

Fig. 2. Left, a plot of computed values of k across the Florida Current off Miami. Right, observed
(solid) and its implied, uncorrected GEK (dashed ) profiles of the current off Miami.

column 8, Table 2; in all cases Afc amounts,
at most, to 207" of k. However, with Afc
proportional to k and to be added to the
latter, this correction will accentuate the
slope of the least-squares lines. Thus, the
k pattern suggested in Fig. 2 is conserva-
tive, actually minimizing the cross-stream
gradient of k, especially in the interval be-
tween 32.5 and 36.0 km.

A test

When the profile of Vs corresponding to
a pattern of k is known, the two together
imply a specific uncorrected GEK profile,
obtainable by simply dividing k into Vs at
corresponding points in accordance with
equation (1). For the data from which k
has been computed, Richardson and Schmitz
( 1965 ) give two profiles of Vs obtained on
successive days. The two differ in some
respects and averaging them modifies the
profile of 16 August 1964, broadening
slightly the current axis and changing

slightly the lateral shear, but otherwise
leaving the overall pattern unchanged. For
simplicity, only the profile for 16 August
has been used and is reproduced as the
top curve on the right side of Fig. 2. Use
of the 17 August profile or one obtained
by averaging the two gives substantially
the same result, as the critical feature of
a broad, relatively flat current axis remains.
The corresponding, implied, uncorrected
GEK profile for 16 August is shown by the
broken curve ( Fig. 2 ) . Given the relatively
flat and broad current axis of the solid
curve, the midstream minimum in the
broken curve results largely from the high
fc-value there; similarly, the primary maxi-
mum in the broken curve results from a
low fc-value, and so on.

DISCUSSION AND CONCLUSION

The implied, uncorrected GEK profile
preserves the extent and the shape of the
original anticyclonic shear reasonably well;

78

FRANK CHEW

although the original cyclonic shear is less
well preserved, it is still not unreasonable.
On the other hand, where originally there
is a single current axis some 10 to 15 km
wide, there are now two narrow axes sep-
arated by an unmistakable minimum in the
implied profile. In general shape and in
location and magnitude of the double
maxima, the implied profile bears consider-
able resemblance to the observed GEK pro-
files.

The value of k computed according to
equation ( 1 ) and averaged across j.he
stream may be compared to the correspond-
ing observed fc-value reported by Wagner
( 1955 ) . For each complete crossing of the
current, he compared the total drift indi-
cated by the GEK to the corresponding
drift given by navigation to obtain an
average value of k. For 14 such crossings
completed in 1952-1953, he obtained a
mean of 1.7 and a range of 1.2 to 2.2. This
compares favorably with the 1.9 for con-
stant electrical resistivity and 2.1 for vari-
able resistivity found in this study.

For the Florida Current off Miami, the

approximate equation ( 1 ) appears ade-
quate, and the feature of double speed
maxima is probably spurious.

REFERENCES

Chew, F. 1965. On correcting geomagnetic elec-
trokinetograph data from the Florida Current.
Lockheed Report 19194, Burbank, California.
( Unpublished manuscript. ) 16 p.

LoNGUET-HlGGINS, M. S., M. I. STERN, AND H.

Stommel. 1954. The electrical field in-
duced by ocean currents and waves, with ap-
plications to the method of towed electrodes.
Papers Phys. Oceanog. Meteorol., 13: 1-37.

Richardson, W. S., and W. J. Schmitz. 1965.
A technique for the direct measurement of
transport with application to the Straits of
Florida. ]. Marine Res., 23: 172-185.

von Arx, W. S. 1950. An electromagnetic
method for measuring the velocities of ocean
currents from a ship underway. Papers Phys.
Oceanog. Meteorol., 11: 1-62.

Wagner, L. P. 1955. Graphic representations
and comments on transects across the Florida
Current, p. 24-53. In Some results of the
Florida Current study. Tech. Rept. ML 6787,
54-7. University of Miami, Miami, Florida.

Webster, F. 1965. Measurements of eddy fluxes
of momentum in the surface layer of the Gulf
Stream. Tellus, 17: 239-245.

Reprinted from: THE INTERNATIONAL JOURNAL
OF OCEANOGRAPHY AND LIMNOLOGY Vol. 1, No. 2

New Method for Boron Determination in Sea Water
and Some Preliminary Results

30

ABSTRACT

INTRODUCTION

John D. Gassaway

Environmental Science Services Administration, Institute for Oceanography,
Silver Spring, Maryland 20910

The 1 ,1 ' -dianthrimide method for the determination of total boron in
sea water is described. With the outlined procedure and apparatus, a
standard deviation of approximately ± 0.30 mg per liter, or less, is ob-
tainable. Standard sea water Ps,( was found to have a boron content of
5A9 ± 0.17 mg per liter and a boron-chlorinity ratio of 0.0259 ± 0.0005.
No single, definite relation was observed between oxygen, salinity, and
total boron at station BB 355-033, and lack of data prevent any conclu-
sions concerning the identification of water masses with total boron.

The mannitol method tor determining borate-boron in sea water is based
on the work of Thompson (1893). Various investigators, i.e., Igelsrud,
Thompson, and Zwicker (1938), Gast and Thompson (1958), and
Noakes and Hood (1961) have modified this method for determining
boron in sea water. The method is pH sensitive and directly measures
the borate concentration and not the boron concentration. Also, the end
point is difficult to judge with accuracy.

Parker and Barnes (1960) described a fluorimetric method for the de-
termination of boron in sea water and obtained a precision of ± 0.1 ppm
for a single determination. This method requires the construction of a
fiuorimeter, the de-oxygenation of the alkaline borate and the ethanolic
benzoin solutions, and the removal of cations with an ion exchange resin
before the determination.

Kawaguchi (1955) used carmen red for the photometric determination
of boron in sea water, but first had to separate boron as methyl borate.
Reynolds and Wilson (1961) employed a spectrochemical method in-
volving spark excitation for determining boron in saline waters.

Greenhalgh and Riley (1962) investigated a curcumin method for the
determination of boron in sea water. This is a simple procedure and does
not require separation of boron from sea water. Humidity affects the pre-
cision of this method; therefore, it must be carried out in a completely
homogenous liquid organic atmosphere. The salts in sea water reduce the
sensitivity of the method and correction factors for the evaluation of bor-

Int. J. Oceanol. & Limnol. Vol. 1, No. 2. p. 85-90.

BORON DETERMINATION IN SEA WATER

85

on are necessary. The correction factor for sea water with a chlorinity of

20o/00is8.9o/0.

A number of factors, including sensitivity, ease of use, interferences
from other elements, and knowledge of optimum conditions for color de-
velopment determines the choice of a particular reagent for the spectro-
photometric determination of an element. Ellis, Zook, and Baudisch
(1949) investigated a number of organic compounds for their suitability
as colorimetric reagents for boron and found l,l'-dianthrimide to be the
best colorimetric reagent.
PROCEDURE Four hundred milligrams of l,l'-dianthrimide are dissolved in 100 ml of
concentrated sulfuric acid. This reagent is diluted to a 1:20 concentration
with concentrated sulfuric acid (v/v) immediately before the l,l'-dian-
thrimide is added to the sample. For satisfactory results a freshly diluted
solution must be prepared daily.

A 50-ml aliquot of sea water is diluted to 250 ml with distilled water
(v/v) and 2- to 4-ml aliquot of this solution is placed in a teflon beaker to
which one milliliter of 0.1 N calcium hydroxide is added. The sample is
placed in the oven, evaporated to dryness at 100C, and cooled to room
temperature. Ten milliliters of l,l'-dianthrimide solution are added, and
the beaker containing the sample is placed in an oven for 90 min at 100 C.
It is then cooled in a desiccator to room temperature, and the absorbance
is read in a spectrophotometer at 625 nnx The color of the reagent ranges
from green to blue.
Discussion Dilutions of 1:5, 1:10, and 1:20 may be used, with the slope of the stand-
ard curve approximately the same for all dilutions. The absorption co-
efficient of 1:20 is less than a 1:5 or 1:10 dilution, allowing more concen-
trated solutions to be analyzed.

As there are no energy limitations at 625 m/i, the choice of the sample
volume used depends on the instrument. The limiting factors vary from
one instrument to another, and it is necessary for the analyst to determine
the best values for his own particular spectrophotometer. It has been
found that samples with an absorbance of 0.215 to 0.500 exhibit minimum
variance.

The sample may be incubated at 90C for 3 hr, 100C for 1-1/2 hr, or 1 hr
at HOC in boron-free glass or plastic containers. At 90C, the time for anal-
ysis is increased with no increase in accuracy, while at HOC the temper-
ature of the oven must be carefully regulated, because temperatures ex-
ceeding 1 10C may cause the reagent to break down rapidly, giving a red-
dish-brown color.

The samples are cooled in a desiccator as traces of moisture may inter-
fere with the reagent. A one-centimeter silica cell is used for the absorb-
ance measurement.

High concentrations of oxidizing or reducing material and dissolved
organic matter give anomalous results. If the technique as outlined above

OCEANOLOGY AND LIMNOLOGY

86

TABLE 1. The boron content and chlorinity ratio for standard sea water P34.

Chlorinity

19.360

Number of
samples

Boron
content

5.49 -+■ 0.17

B/Cl

0.0259 ± 0.0005

DISSOLVED OXYGEN

(ml/1)

0.00 2.00 4.00 6.00

~i — i — i — i — i — r

TOTAL BORON B/Cl

(mg/l) (xlO"2)

4.20 4.60 5.00 2.00 2.20 2.40 2.60

FIGURE 1. The variation of the boron content, B/Cl, and oxygen with depth at sta-
tion BB 355-033. The horizontal lines represent the standard deviation
of the B content and the B/Cl ratio.

BORON DETERMINATION IN SEA WATER

87

SOME PRELIMINARY
RESULTS

Discussion

is followed, the concentration of the interfering ions and organic com-
pounds is too small to affect the results. Buffering solutions and solutions
with a high salt content do not affect the results significantly (Powell and
Poindexter, 1957).

This method yields total boron. The temperature at which the sample
is heated and the sulphuric acid decompose the organic compounds which
form complexes with boron (D. G. White, personal communication).
Also, the complex boron concentration reported by other workers (Noakes
and Hood, 1961) is usually less than the standard deviation of the method.

Using the outlined procedure and apparatus, the standard deviation
obtainable is approximately ± 0.30 mg per liter, or less.
The results of the analysis of standard sea water P34, I.A.P.O. Standard
Sea Water Service, Chalottenlund Slot, Denmark, are listed in Table 1
and those from station BB 355-033 are shown in Fig. 1. The location of
the station from where the samples were taken is 46° 23.9 N and 125°
22.3 W, approximately 100 miles west of the mouth of the Columbia Riv-
er. The sea water samples were collected with Nansen bottles which were
drained immediately upon retrieval into polyethylene containers, frozen,
and returned to the laboratory where they were thawed and analyzed. The
boron content is measured in milligrams of boron per liter and the chlor-
mity ratio (B/Cl) is milligram-atoms boron per chlorinity (X 102) (Igels-
rud, Thompson, and Zwicker, 1938). Oxygen is reported in milliliters per
liter, and the depth is in meters.

Standard sea water P34 has a boron content of 5.49 ± 0.17 mg per liter
and a boron-chlorinity ratio of 0.0259 ± 0.0005. This sample was taken
from one ampule. The boron chlorinity ratio is extremely high, and if
the ampule is a boron containing glass, this high value might result from
the solution of the glass by the contained water.

In Table 2 the maximum and minimum concentration of the boron-
chlorinity ratio in sea water, as reported by other authors, is tabulated.
These results agree favorably with those reported in this paper.

There is an over-all increase of total boron with depth and this rate of
increase is greater than that of the chlorinity (Fig. 1). At this particular
station, there is no linear relationship between the total boron and chlor-
inity. Instead, the total boron content varies with depth as in Noakes and
Hood's (1961) stations and in one Atlantic Ocean station examined by
Glebovich (1946).

At station BB 355-033, below 300 m, there is a direct relationship be-
tween the total boron and oxygen (Fig. 1). This is the opposite of the
findings of Igelsrud et al. (1938), but agrees with Glebovich (1946). The
three stations examined by Noakes and Hood (1961), together, do not
show any single, definite relationship. Miyake and Sakuri (1952) observed
no relation between oxygen and boron with depth in the water column,
but they demonstrated the possibility of correlating the boron content

OCEANOLOGY AND LIMNOLOGY

88

with different water masses. There is, apparently, no single, definite re-
lationship between the oxygen and the total boron content, and the lack
of data prevents any comment on tracing water masses with boron.
CONCLUSIONS A precise and accurate method for determining total boron in sea water
was developed. This technique is sufficiently simple to be used for routine
work and does not entail separation of the element. A standard deviation
of ± 0.30 mg per liter is readily obtainable.

Standard sea water P34 has a boron content of 5.49 ± 0.17 mg per liter
and a chlorinity ratio of 0.0259 ± 0.0005.

From the one station examined in this paper and those cited in the
text, there is, apparently, no single, definite relationship between total
boron, oxygen, and salinity, or chlorinity.

REFERENCES Ellis, G. H., E. G. Zook, and O. Baudisch. 1949. Colorimetric methods for the

determination of boron. Anal. Chem., 21, 1345-3148.
Gast, J. A., and T. G. Thompson. 1958. Determination of the alkalinity and borate

concentration of sea water. Anal. Chem., 30, 1549-1551.
Glebovich, T. A. 1946. Boron in the sea. Trudy Biogeokhim. Lab. Acad. Nauk.

SSSR., 8, 227-252.
Greenhalgh, R., and J. P. Riley. 1962. The development of a reproducible spectro-

photometric curcumin method for determining boron, and its application to

sea water. Analyst, 87, 970-976.
Igelsrud, I., T. G. Thompson, and B. M. G. Zwicker. 1938. The Boron Content of

Sea Water and of Marine Organisms. Amer. J. Sci., 35, 47-63.

TABLE 2. The maximum and minimum concentration of the boron chlorinity ratio
in sea water as reported by other authors and this paper.

Maximum B/Cl Minimum BjCl

Method

Author

0.0250

0.0202

Mannitol

Igelsrud et al. (1938)

0.0254

0.0229

Mannitol

Glebovich (1946)

0.0225

0.0219

Mannitol

Gast and Thompson

(1958)

0.0232

0.0221

Fluorimetric

Parker and Barnes

(1960)

0.0226

0.0205

Mannitol

Noakes and Hood (1961)

0.02112

(analyzed

Curcumin

Greenhalgh and Riley

one sample)

(1962)

0.0246 ± 0.0009

0.0211 ±0.0008

l.l'-dianthrimide

(This paper-excluding

standard sea water

P34>

BORON DETERMINATION IN SEA WATER

89

Kawaguchi, H. 1955. Colorimetric determination of traces of boron. Japan

Analyst, 4, 307-310.
Miyake, Y. and S. Sakuri. 1952. Boron in sea water as an indicator for the water

mass analysis. Sea and Sky, 30, 1-4.
Noakes, J. E., and D. W. Hood, 1961. Boron-boric acid complexes in sea water.

Beep-Sea Res., 8, 121-129.
Parker, C. A., and W. J. Barnes. 1960. Fluorimetric determination of boron.

Analyst, 85, 828-838.
Powell, W. A., and E. H. Poindexter. 1957. 1 ,l'-dianthrimide method for the

spectrophotometric determination of boron. Univ. Richmond, Richmond,

Va. 41 p.
Reynolds, R. C. and J. Wilson. 1961. Spectrochemical determination of boron in

saline waters. Anal. Chem., 33, 247-249.
Thompson, R. T. 1893. New aspects of phenolphthalein as an indicator. /. Soc.

Chem. Ind., 12, 120-134.

ACKNOWLEDGMENT The author would like to thank M. Grant Gross and T. J. Conomos for a critical
and helpful review of this paper. I want to thank the Department of Oceanog-
raphy, University of Washington and the Department of Geology, George Wash-
ington University for making the necessary laboratory space available. I would
also like to thank Janie Lochte and Doris Verhunce for typing this manuscript.

OCEANOLOGY AND LIMNOLOGY 90

31

Reprinted from SCIENCE Vol. 158, No. 3803

Deep-Sea Tides: A Program

Walter H. Munk and Bernard D. Zetler

Tidal mathematicians have tradition-
ally stood along the coastlines and
looked longingly to sea, speculating on
what the tides are offshore. They have
seized upon the tide records obtained
at a few island outposts and have ex-
pended great effort in producing co-
tidal charts of the oceans, with lines
connecting points at sea at which the
time of high water is thought to be
simultaneous. The patterns on these
charts have been quite complex, the
most interesting feature being the loca-
tions of amphidromic points. These
are the geographic positions where the-
oretically there is no tide, the cotidal
lines radiating about them in various
directions, with the tidal amplitude pre-
sumably increasing with distance from
the amphidromic point. Inasmuch as
the response of the ocean basins to
the tide-producing forces is frequency-
dependent, the more ambitious mathe-
maticians have drawn their charts for
each of the larae tidal constituents, both

Dr. Munk is with the Institute of Geophysics
and Planetary Physics, University of California
at San Diego, La Jolla; Dr. Zetler, with the
Institute for Oceanography, Environmental Sci-
ence Services Administration, Silver Spring,
Maryland.

diurnal and semidiurnal. Despite the
application of the best techniques and
all available data in their efforts, specu-
lation has been an important ingredient
in the completed charts, and one is
forced to the conclusion that the best
is none too good.

It appears as though modern technol-
ogy has caught up with the problem
and that there is hope of obtaining
within a relatively few years objective
measurements of the tide at positions
on grids spanning the oceans. This
is exciting in itself, but even more ex-
citing is the anticipated solution of
many other geophysical problems as a
by-product of the program.

Several different engineering groups
are developing tide gages for deep-sea
measurements. Snodgrass (/) has built
a self-contained capsule that is dropped
to the sea floor, records absolute pres-
sure in situ, and is subsequently re-
called by acoustic signals from a sur-
face vessel. Some success has already
been achieved with the instrument, but
many failures have demonstrated need
for greater reliability — which entails
duplication of critical circuits, quality
control of individual components, long

periods of pretesting, and pretesting
under severe environmental conditions.
The U.S. Coast and Geodetic Sur-
vey has made some successful meas-
urements of the tide at depths slightly
less than 300 meters over the Atlantic
continental shelf; it has an active de-
velopmental program underway to in-
crease the depth potential of the gage
and to improve its reliability. Martin
Vitousek, University of Hawaii, has
been involved in related research — the
recording of tsunamis; although some
problems are different in that he needs
more frequent observations over a
shorter period, there is significant simi-
larity in the development program.

The effort has not been confined
to the United States. Eyries, Service
Hydrographique, Paris, has successfully
tested a differential gage that is con-
nected by wire to a surface buoy that
transmits an analog signal to a nearby
vessel. The Snodgrass and Eyries gages
also record temperature because it is
required for correction of pressure read-
ings, but of course the temperature
records are of great interest in them-
selves. Other foreign tidal authorities
have shown marked interest in the pro-
gram and stand ready to cooperate in
an international program when the re-
quired instruments become available.

At a tide symposium (2) in Paris in
1965, an international working group
on deep-sea tides was formed to or-
ganize systematic measurements and
analyses of deep-sea tides. When the
group met again in Moscow in 1966
to review the program, representatives
of 1 1 nations attended and expressed
some interest in the program; an as-
sociated committee (3) was formed to
deal with theoretical problems related

to deep-sea measurements of tides. The
working group met again in January
1967 (4) to participate in sea trials of
the Snodgrass gage. Eyries organized
the symposium on deep-sea tides
held in conjunction with the meetings
of the International Union of Geodesy
and Geophysics in Switzerland during
the fall of 1967. In short, the grow-
ing interest in the program promises
much for the future.

Field tests of the Snodgrass gage so
far have been oriented toward develop-
ment rather than scientific results. The
first field program for the latter purpose,
with plans to occupy the North Pacific
semidaily lunar (M.2) amphidrome with
three instruments, will stay on the
bottom for 1 month. Each capsule will
record fluctuations in the bottom pres-
sure to the nearest millimeter of wa-
ter; fluctuations in temperature to the
nearest 10-0°C, and horizontal cur-
rent with a resolution of 1 millimeter
per second. If a gage is properly posi-
tioned at the M, amphidrome, the
M2 amplitude will yield the continuum
of energy rather than the tidal line;
furthermore, the lunar semidaily tide,
observed on either side of the amphi-
drome, will be opposite in phase. Inas-
much as the ocean's response is fre-
quency-dependent, the solar semidaily
tide (So) should exceed the A/, tide
at the amphidrome, and the various am-
plitudes and phases for S2 at the
four locations should point toward the
probable location of the S2 amphi-
drome.

Previous tests have shown an increase
in temperature of the order of 0.05°C
in the bottom few meters of some
basins off the coast of California; it is
not clear how this unstable gradient
can persist. Furthermore, the records
contained intermittent quiescent and ac-
tive periods, with a time scale that is
roughly tidal. For these reasons a
special effort will be made to study
bottom temperature gradients and in-
termittent turbulence during the tests.
One capsule is being modified to shorten
intervals in the recording system for
data sampling, and there will be special
movable arms on the transducer mount-
ing to permit careful examination of
the bottom meter of water. A special-
purpose winch is planned, so that the
capsule can be lifted off the bottom
in a predetermined program to obtain
the vertical temperature profile to a very
high degree of precision. Finally, a
camera system will be incorporated to
monitor the installation and, if pos-

sible, observe the motions of the water
by stereo photographs of particles drift-
ing near the sea floor. Plans for 1968
include an equatorial station for mea-
suring both tides and planetary waves,
and measurements between Australia
and Antarctica to obtain some infor-
mation on Antarctic tides.

The tentative plans of the working
group for coverage of the world's
oceans include a 1000 by 1000-kilo-
meter grid (10 degrees by 10 degrees
at the equator) involving a total of
300 stations. Measurements are to be
made at least hourly for 1 month (cer-
tainly for no less than 2 weeks) to
the nearest centimeter of depth of wa-
ter or better. Nevertheless, results from
programs such as the North Pacific
amphidrome test will be scrutinized
carefully for the possibility of estab-
lishing a variable grid, based on prelimi-
nary theoretical results, with a maxi-
mum of information content per unit.
We visualize two types of expeditions
for obtaining the tidal records: (i) ships
of opportunity, primarily devoted to
other purposes, dropping tide instru-
ments on the way out and retrieving
them on the way back; and (ii) tidal
expeditions primarily devoted to the
program but serving also other needs
and opportunities. A ship should be
able to plant eight instruments ap-
proximately 3 days apart and then start
to retrieve the first almost as soon as
the planting is completed. On this basis
eight stations are established and re-
trieved within 2 months; since most of
the time is spent in cruising, there is
ample opportunity for other measure-
ments.

The scientific objects of the program
include both increased understanding
of global tides and other related geo-
physical phenomena; the former are
listed first:

1) Theoretical calculations of global
tides are now being performed by Per-
keris and Hendershott for the true depths
and boundaries of the world's oceans.
So far both investigators have assumed
perfectly reflective boundaries, but it
appears that ultimately one must al-
low for absorption over the continental
shelf. The availability of an ocean-
wide grid of tidal observations for com-
parison with theoretical calculations will
provide some evaluation of the need
for a modified boundary premise; it
will also permit reliable estimates of
the proportion of total tidal energy
(3 X 1019 ergs per second) dissipated
within the oceans.

2) From knowledge of the global
tides (beyond the continental shelves)
it may be possible to infer tide predic-
tions on a shelf for which land-based
tide records are not available. In the
case of highly nonlinear tides, existing
predictions may be improved by knowl-
edge of offshore tides; thus the non-
linear modification may be separated
from the astronomic tide. The dy-
namic numerical computations will use
depth and friction as parameters; for
tides in rivers and bays the effective
width will be an added parameter. Ulti-
mately, coastal tides will be predicted
from an initial two-dimensional tide
"field" in the open sea.

Although at the moment one cannot
envisage all related geophysical objects
of the program, the following appear
to be possible:

1 ) For geophysical measurements on
land, the tidal frequencies cannot be
interpreted without allowance being
made for the effect of oceanic tides
on a global basis; this point applies to
measurements of gravity, magnetic
field, and so on.

2) The previous item includes cal-
culation of the dissipation of tidal ener-
gy in the world's ocean. The residual
energy, 3 X 1019 ergs per second minus
ocean dissipation, would give important
information concerning the plastic prop-
erties of the solid Earth.

3) The fluctuating tidal currents flow-
ing in Earth's magnetic field generate
electric potentials that can be measured
with suitable electrodes on the sea bot-
tom. The generated potential depends
also on the conductivity within Earth;
with the tides known, the effective con-
ductivity can be estimated for various
tidal frequencies (and hence effective
depth). These estimates in turn give in-
formation about the distribution of tem-
perature in Earth's upper mantle. Hori-
zontal temperature gradients within
Earth, particularly beneath the ocean
boundaries, are associated with a stress
field that may be responsible for the
principal belts of volcanic and seismic
activity.

4) Barotropic signatures of passing
storms, which will be superimposed on
tidal fluctuations as measured with bot-
tom-pressure recorders, represent an in-
teresting problem of air-sea dynamics.
It would be especially interesting to
have long series of simultaneous ob-
servations of horizontal currents at
great depth (away from surface noise).
Planetary waves, which are so promi-
nent in the atmosphere, have not been

convincingly demonstrated in the
oceans, perhaps because they are poor-
ly generated or poorly transmitted, or
because the observations have not yet
measured the proper variables. Such
measurements are simply not available
at this time. Well-documented current
flow near the bottom would provide a
far better reference for geostrophic
computations than do theoretical levels
of no motion; estimates of mass trans-
port could thus be significantly im-
proved.

5) We have already mentioned a sharp
increase in temperature in the bottom
few meters of the oceans. The existence
of a warm bottom layer had been re-
ported earlier by Van Herzen and co-
workers from measurements with a

geothermal probe. Is this warm layer
maintained by greater density because
of higher content of salt or of sedi-
mentary particles? In fact, in the meas-
urements of temperature gradients in
sediments, what is. the role of tidal
"pumping" of interstitial water? For
these studies as well as of the intermit-
tent turbulence, long, reliable records
of temperature are necessary; we hope
they will be supplemented by measure-
ments of heat flow.

6) Finally the observations may be
helpful in explaining the origin of in-
ternal waves of tidal frequencies, and,
if tsunamis are generated during periods
of tide observations, they may be well
documented if the time period of the
observations is short enough. In any

case, the instrumental development can
be used in a tsunami-measurement pro-
gram.

Thus the proposed program has ap-
plications in much of Earth's environ-
ment, air-sea dynamics, various ocean-
wave phenomena, and the anelasticity
and stress fields of the solid Earth.
The bottom of the sea is perhaps the
least explored of the "accessible bound-
ary layers" on this planet, and we may
be in for some surprises.

References and Notes

1. F. Snodgrass, Scripps Inst, of Oceanography,
Univ. of California at San Diego, La Jolla.

2. Sponsored by UNESCO and the Intern. Assoc.
for Physical Oceanography.

3. W. Hansen (Germany) and S. S. Voit (U.S.S.R.)
are joint chairmen.

4. At La Jolla, California.

32

Fig. 1 — Aircraft photo of eddy in the Caribbean, north of St. John, Virgin Islands.
Such eddies aid in fishing and in weather and dynamic current studies.

Reprinted from OCEAN INDUSTRY Vol. 2, No. 5

OCEAN INDUSTRY DIGEST

Fig. 2 — Daylight ESSA-1 TV photo off

Sensing ocean currents from space

By Raymond M. Nelson

Institute for Oceanography, ESSA

As part of its mission to monitor the oceans and
study the related physical environment, the Department
of Commerce's ESSA (Environmental Science Service
Administration) is deeply involved in gathering basic
data related to ocean circulation.

Until recently, sources of such data were quite lim-
ited, but new developments in electronics and advances
in space technology are providing scientists with exciting
new tools which hold great promise in this field.

High Resolution Infrared (HRIR) Scan Imagery. An

examination of the HRIR line scan imagery taken from
the meteorological satellite Nimbus 2 shows that high
thermal contrasts on water surfaces are discernible. Alli-
son, Foshee and Warnecke of NASA and Goddard and
Wilkerson of the Naval Oceanographic Office have col-

laborated on the determination of water surface temper-
atures from such data."' From the accompanying HRIR
scan imagery and the ESSA-1 television photographs, it
is readily apparent that surface boundaries of major
currents can be detected and point temperature distribu-
tions could have been measured between clouds. Time-
sequence analyses of such quasi-synoptic coverage over
large areas may reveal previously undisclosed surface
circulation effects.

EXAMPLES: Fig. 1 is an ESSA-television picture taken
off the east coast of the U.S. on June 2, 1966 and Fig.
3 is a Nimbus 2 HRIR photograph exposed 11 hours
later (at night). A comparison shows the front line
appears to have moved farther east on the infrared pho-
tograph, and some cloud patterns, imaged near the coast
on the infrared photograph, are not discernible on the
Nimbus exposure. Aircraft coverage at the time of this
Nimbus orbit proved that the local sky was free of clouds.

40

OCEAN INDUSTRY

U.S. east coast, taken June 2, 1966

Nimbus 2 HRIR photo taken of part of same area 11 hours later at night.

An additional night-time infrared photograph (Fig.
4 was taken September 18 and 19 off the southern tip
of Cape of Good Hope, Africa. The white arrow points
to the high thermal contrast of the western boundary of
the Agulhas Current.

What's needed to make the
infrared system more effective?

The system has inherent needs. The primary one is for
surface monitoring and checking to verify the imagery.
For example, Fisr. 5 — a Nimbus 2 HRIR photograph of

us 2 HRIR photo off Cape of Good Hope, Africa.

Fig. 6 — Infrared line scan photography
of an eddy about four miles in diam-
eter, 200 miles east of Cape Cod taken
from aircraft. Spectral range 8-14 mi-
crons; length of area shown is 11
nautical miles. White spots are mal-
functions of scanner. Light tones are
warmer than dark tones.

transparent convection cells over the North Atlantic
and Gulf Stream also have been recorded. The apparent
temperature discontinuities of such cells are only about
1° C and the horizontal dimensions range from 1 to 10
miles.

What's needed for current studies?

We know little about the meanders of the Gulf Stream.
We are not able to predict the size, location or likelihood
of its meanders; we do not know whether the meander
dynamics are the same for both sides of the Gulf
Stream. We need a multi-sensor surveillance system,
including microwave components with cloud penetra-
tion capability to supply raw data and surface-recorded
data to aid us in interpretation.

A time-sequence charting of the boundaries of sur-
face currents as to shape size and location will contrib-
ute much toward understanding of total dynamics and

at the same time provide valuable data for the shipping
industry.

The capability of aircraft systems to detect an ocean
eddy is illustrated in Fig. 6. Reasonably, such detection
should be followed up by sending a ship to make a sur-
face and subsurface profile as the aircraft is collecting
such imagery. Themal profiles generated from micronsi-
tometer scans across the image can be correlated with
a ship-towed thermistor for calibration and comparison.
Time-sequencing by the aircraft at 15 to 20-minute in-
tervals could reveal motion within the eddy, etc.

Exciting programs lie ahead

The present aircraft-ship experimentation will con-
tinue; then in the summer of 1969, the ESSA Barbados
Oceanographic and Meteorological Experiment will be
carried out in the eastern Caribbean vicinity. This ex-
periment will serve as a pilot field study for the Global
Atmospheric Research Program (GARP) of the World
Weather Watch. It will utilize sensors ranging in heights
from satellites down to the sea floor.

The Tropical Meterlogical Experiment (TROMEX^

is scheduled for the early 1970s and some thought is
being given to expanding it to a Tropical Environmen-
tal Experiment. In CY 1971, the satellite systems Im-
proved TOS, N and O and Advanced Polar Orbiting
Satellites (APOS) A and B may carry laser altimeters
for determining variations of sea level for geodetic
purposes.

An immediate measure which could contribute a
great deal to these necessary studies would be to utilize
"aircraft of opportunity" to fly instrument systems.
Many airlines run over the oceans on daily schedules,
both night and day, and the possibility of equipping
them with sensors has a certain amount of merit.

Basically, the ocean is an international problem, and
the exploitation of the global surveillance capability of
spacecraft will require international cooperation, possi-
bly under a central cooperative control, where, as auto-
mated data are displayed and critical changes in
meteorological and oceanographic effects or patterns
are detected, aircraft and ships in the area can be sent
to monitor the region in greater detail.

Although the needs in ocean current study technology
are quite pronounced, a tremendous start has already
been made.

BIBLIOGRAPHY

1. Stommel, H. M., W. S. Von Arx, D. Parson, and W. S. Richardson,
"Rapid Aerial Survey of Gulf Stream with Camera and Radiation Ther-
mometer," Science, 1953, 117:639-640.

2. McAlister, E. D., "Infrared-Optical Techniques Applied to Oceanog-
raphy: 1 Measurement of Total Heat Flow from the Sea Surface," Applied
Optics. 3, pp. 6U9-612, 1964.

McAlister, E. D. and McLeish, W. L., "Oceanographic Measurements
with Airborne Infrared Equipment and Their Limitations," Oceanography
from Space, Woods Hole Oceanographic Institution, Woods Hole, Mass.,
pp. 189-214. April 1965.

3. Ewing, Dr. G. C., "The Use of Film as a Sensor," Minutes of the
Fourth Meeting Ad Hoc Spacecraft Oceanography Advisory Group, Space-
craft Oceanography Project, U. S. Naval Oceanographic Office, 18-19 Oc-
tober 1966, p. 12.

4. NASA, Goddard Space Flight Center, (Private Communication).

5. Allison, L. J. L. L. Foshee, G. Warnecke, and J. C. Wilkerson, "An
Analysis of the North Wall of the Gulf Stream Utilizing Nimbus 2 High
Resolution Infrared Measurements," Gulf Stream Symposium, American
Geophysical Union, April 17-21, 1966.

6. Oshiver, A. H., G. A. Berberian, J. R. Clark and R. B. Stone. "Far-
tors in Measurement of Absolute Sea Surface Temperature by Infrared
Radiometry," University of Michigan, Ann Arbor, pp. 737-762, February
1965.

7. Clark. H. (Private Communication), U. S. Naval Research Laboratory.
Applied Oceanography Branch, Washington, D. C, "Clear Air Convection
Cells," March 1967.

8. Smith, J., "Color — A New Dimension in Photogrammetry," Photo-
grammetric Engineering, Nov. 1963. Fig. 7. pp. 999-1013. ■

42

OCEAN INDUSTRY

33

Reprinted from TRANSACTIONS-SECOND INTERNATIONAL BUOY
TECHNOLOGY SYMPOSIUM, The Marine Technology Sooiety

-Jt AN OCEANOGRAPHIC DATA COLLECTION SYSTEM

P. E. Seelinger, R. A. Walls ton,
B. H. Erickson,1 J. E. Masterson,2 W. E. Hoehne3

Pacific Missile Range
Point Mugu, California

ABSTRACT

A buoy system was designed and installed to provide meteorological and
oceanographic data to support surface, aircraft, rocket and missile test
operations off the southern California coast. This system measures wind
speed and direction, air temperature, barometric pressure, sea surface tem-
perature, and wave height. The system also measures subsurface temperature,
current speed , and current direction at various depths . The data are
stored in the buoy and transmitted on command to a shore station, where
they are recorded for reduction by a computer.

The buoy assembly consists of a surface float connected to a subsurface
buoy at a depth of 50 feet. The subsurface buoy is taut -moored to the bot-
tom. Meteorological sensors are mounted on the mast of the surface float.
The recording, digitizing and telemetry equipment are within the surface
float. A wave height sensor is mounted on top of the subsurface buoy. Cur-
rent meters and water temperature sensors are attached to the taut mooring
cable.

One buoy assembly was moored on 3 September 1966, and gathered and trans-
mitted data for about seven months . The reliability and accuracy of the
sensors during this period were judged good. The overall performance of
the buoy assembly demonstrated its potential usefulness in providing envi-
ronmental data affecting Sea Test Range operations.

INTRODUCTION

In testing and evaluating bomb, missile and rocket systems, it is necessary
to have accurate knowledge about the environment in which the tests are
conducted. Sea state and currents, for example, affect water entry tests
and the recovery of missile components from the sea. Wave height spectra
provide information on radar sea clutter for evaluating radar-guided mis-
sile performance. In addition, accurate data are essential in order to
forecast oceanographic and meteorological conditions in and over the test
range.

Required oceanographic information is normally obtained from subjective
estimates made by shipboard and airborne observers, and from analyses of
weather charts . Some information is extrapolated from measurements

1. Present affiliation: Pacific Oceanographic Laboratory, Institute for
Oceanography, ESSA, Seattle, Washington.

2. Present affiliation: National Center for Atmospheric Research, Boulder,
Colorado .

3. Present affiliation: Weather Bureau, System Development Office, ESSA,
Sterling, Virginia.

311

provided by snore-based equipment — e.g. , wave staffs and oottom-mounted
wave meters. A smaller amount of data is obtained irregularly from portable
instruments — e.g. , buoy -mounted wave staffs and bathythermographs.

The buoy system described in this paper was designed to obtain continuous
oceanographic and meteorological data at sites where water depth does not
exceed 3,600 feet (600 fathoms) and the data transmission distance is not
greater than 60 nautical miles. Sea surface and subsurface data, particu-
larly of wave height and period , were the primary requirements . Specifi-
cations for meteorological data were included to make fullest possible use
of the surface float platform and assembly. Table 1 lists the variables to
be measured.

Table 1
ENVIRONMENTAL VARIABLES TO BE MEASURED

Category

Measurement

Meteorological

Wind speed and direction
Air temperature
Atmospheric pressure

Sea surface

Water temperature
Wave height and period

Subsurface

Current speed
Current direction
Water temperature

A preliminary survey of existing buoy assemblies, sensors, and instrumenta-
tion had revealed that equipment and techniques were sufficiently developed
to permit their integration into a system. A number of contractors judged
qualified on the basis of past experience in the design and installation of
buoy systems were asked to submit two-step technical proposals . Five out
of nine completed proposals ; and the contract for design and delivery of the
system was awarded to the Bissett-Berman Corporation of Santa Monica, Cali-
fornia (1). A separate contract covered installation of the first buoy
assembly. The government provided most of the standard components of the
shore station, as well as programming and computer facilities at the PMR
Weather Center.

The buoy assemblies are instrumented to take measurements of the environ-
mental variables shown in Table 1, and transmit them on command to a shore
station for computer processing. These synoptic data are then analyzed
every 6 hours to provide information about sea and weather conditions in
specific test areas, and in detail not obtainable by other means.

The area considered for installation of the buoy system includes the PMR
Sea Test Range and is shown in Figure 1. It lies among the Channel Islands,
east of Santa Cruz Island and north of San Nicolas Island, within the
continental borderland — a region of basins , troughs , shallow banks ,
islands, and minor shelves off the islands. Five sites (Figure 1) were
surveyed as possible locations for buoy assemblies; and Site #2 was selected
for installation of the first buoy assembly.

This paper will first describe the buoy system design and the data-handling
subsystem. It will then cover the survey and installation procedures.
Finally, it will evaluate the data obtained during feasibility tests of one
buoy assembly.

312

SYSTEM DESIGN

The Oceanographic Data Collection System (ODCS) was designed to consist of
four buoy assemblies, installed at four of the five proposed sites (Fig-
ure 1). Each buoy assembly was to be equipped with appropriate sensors,
a data-handling subsystem, a telemetry receiver and transmitter, power
supply and antenna. Data would be received at the shore station and stored
for later processing by computer. One of the four buoy assemblies was in-
stalled as a "shakedown" in order to test overall performance and solve
working system problems before implementing the total system.

Buoy Assembly

The buoy assembly shown in Figure 2 consists of a surface float slack-
moored to a subsurface buoy at a depth of 50 feet. The subsurface buoy is
taut -moored to the bottom by a 3/8-inch steel cable strain member attached
to a shaped anchor clump weighing about 2,500 pounds. A separate Danforth
anchor gives further assurance against movement along the bottom. Two
mechanical swivels, one just above the anchor and the other below the lowest
sensing instrument, keep the assembly from twisting. Separate data trans-
mission cables are attached to the strain cable, but carry no strain. A
system of watertight connectors in the electrical cables allows for replace-
ment of individual sensor pods (by divers) at levels down to 150 feet. The
entire assembly with the exception of the anchor is designed to be recovered
by activating an acoustic release device, located just above the anchor.

The surface float shown in Figure 3 is a steel-covered discoid 10 feet in
diameter. Compartments in the center contain the electronic equipment ,
packaged individually in modular watertight containers , and are surrounded
by compartments filled with polyurethane foam. A 13-foot quadrapod mast
carries the antenna, the meteorological instruments, and the radar reflec-
tors , navigation light and bell .

The subsurface buoy shown in Figure M- is a cylindrical tank 8 feet long and
4 feet in diameter. It is constructed of 12-gauge steel, weighs approxi-
mately 800 pounds, and has a net buoyancy of approximately 2,000 pounds.
The wave height and water temperature sensors are housed together in a steel
case on top of the buoy. Mounted near the sensors is a battery-operated
pinger, which turns on automatically if the electrical cable to the surface
buoy is disconnected or severed. The device thus assists in locating the
subsurface buoy if the surface float breaks free.

Sensors

Figure 5 shows the meteorological sensors mounted on the quadrapod mast
approximately 10 feet (3 meters) above the mean water line. The sensors
include instruments to measure wind speed, wind direction, air temperature
and barometric pressure.

The oceanographic sensors are mounted at various positions along the taut
mooring cable as well as on the subsurface buoy and surface float. Water
temperature, current speed and current direction sensors (Figure 6) are
attached to the mooring cable at 75, 150, and 500 feet below the surface.
Another water temperature sensor is moftnted, with the wave height sensor,
on the subsurface buoy, at a depth of 50 feet (Figure 7). The sea-surface
temperature sensor protrudes through an opening in the bottom of the surface
float at a depth of about 3 feet.

Table 2 lists the meteorological and oceanographic variables and the instru-
ments used to measure them. It also describes the range and accuracy of
each sensor.

313

Table 2

INSTRUMENTS USED TO MEASURE
METEOROLOGICAL At© OCEANOGRAPHIC VARIABLES

Variable

Instrument

Range

Accuracy

Wind speed

3-Cup anemometer

2 to 99 kn

+ 5% true

Wind direction

Wind vane

0 to 360°

+ 10° true

Air temperature

Thermistor

0 to 40°c

+ 0.2°C

Barometric

Aneroid barometer

950 to 1049 mb

+ 1.0 mb

pressure

Wave height and

Pressure transducer

0.5 to 16 ft

+0.1 psi

period

5 to 25 sec

Sea water

Thermistor

0 to 30°C

+ 0.2°C

temperature

Current speed

Savonius rotor
current meter

0.05 to 6.0 kn

± 3% true

Current

Current meter vane

0 to 360°

+ 10° true

Direction

The sensors were not intended for use in research and development; and in
most cases, off-the-shelf instruments were adequate to system needs. The
wind sensors, however, were modified considerably to withstand the marine
environment. All the sensors were calibrated before installation. The
accuracies indicated in Table 2 were met , and calibration curves for each
sensor were made and provided to the Navy at Point Mugu. No further cali-
bration tests on the sensors were conducted after the buoy assembly was
installed in the open sea. However, some comparative observations were
made when possible; and these are discussed in a later section of the paper.

The Data -Handling Subsystem (DHS). The DHS on board the surface float
controls the acquisition, storage, and transmission of data from all the
sensors . Figure 8 is a block diagram of how the total system functions .
Interrogation of the sensors is initiated at 20-minute intervals by the
system timing and control logic circuits. All the sensors are scanned
sequentially within a period of a half-second by the commutator. The
analog voltages from the sensors are converted into digital information by
the A/D converter. These data are then stored in the core memory, which
has a capacity of about 15,000 bits. The wave height data consist of
samples taken at 2-second intervals for a period of 20 or 40 minutes once
every 6 or 12 hours respectively.

The upper portion of the block diagram (Figure 8) represents the shore
station system. A clock timer may be pre-set to interrogate the buoy
automatically at specific times — or the buoy may be interrogated manu-
ally. In either case, a two- tone command is sent to the buoy to start
transmission of the data stored in the core memory. Information from the
buoy is transmitted in binary code, using standard 100 word-per-minute
teletype format. The signals are received at the shore station, where the
data are recorded in 5- level teletype code on punched paper tape. The
data on the tape are later processed by a Control Data Corporation 3100
computer. Specifications for the EHS are shown in Table 3.

314

Table 3
SPECIFICATIONS FOR DATA-HANDLING SUBSYSTEM

Data input

16 analog channels

Range

0-511 millivolts

Data sampling period:
Channels 1-15
Channel 16

600 milliseconds total
180 milliseconds

Data sampling rate:
Channels 1-15
Channel 16

every 20 minutes

every 2 seconds ( Channel
16 is activated either
every 6 hours for a 20-
minute period or every
12 hours for a 40 -minute
period, depending on
the switch position se-
lected. )

Data storage capacity

15,232 bits

Storage requirements :
Channels 1-15

Channel 16 — 20-minute sample
40-minute sample

170 bits (including buoy

ID and time code)
6,000 bits
12,000 bits

Digitizer accuracy

± 2 millivolts

Data output

teletype format

Data rate

10 characters per second

Data readout period

5 minutes and 6 seconds

Figure 9 shows the actual shore installation at Point Mugu, which is con-
tained in essentially one rack of equipment. It includes the receiver,
clock timer, encoder, demodulator, and paper tape punch. The command
transmitter, located at another site, is remotely controlled.

Data Reduction. The data sto»ed on the punched tape are reduced by the
computer. The processing accomplishes the following:

1. Edits the data for proper sensor-scan format, and locates
and separates wave height data

2. Converts the binary numbers representing millivolt values
to engineering units

3 . Determines 20-minute average current speeds and wind speeds

4. Converts time code to Greenwich mean times (time of interro-
gation is an input for each tape) and determines starting
times of wave height samples

5. Lists both raw and processed data.

315

A separd /ave height program edits the individual pre e measurements
and then determines a power spectrum to find the contribution of each wave
period to the pressure variations measured at depth. The contribution of
each period is then multiplied by the factor which describes its attenu-
ation with depth. Finally, the pressure variations are converted to feet,
the contributions of all periods summed, and the wave heights calculated
in accordance with Pierson, Neumann and James (2). In actual practice, the
data are separated into four groups and an average taken.

Figure 10 shows a listing of processed data in the format produced by the
computer. These data were received from the buoy on 21 February 1967 at
1120Z.

SITE SURVEY AND INSTALLATION OF BUOY ASSEMBLY

Before installing a taut -moored buoy system, it is necessary that rather
accurate surveys of prospective sites be made. Five sites within the Sea
Test Range were selected, varying in depth from 1,680 to 2,880 feet.
Bathymetric surveys of each of the five sites were conducted to measure
depths and determine the configuration of the sea floor. Two-inch gravity
cores were taken to determine bearing properties of the sea floor. Current
profiles were also measured at each of the sites . LORAC B was used for
navigation during both the survey and installation.

The four sites best suited for installing the four buoy assemblies of the
complete system were specified. Site #2 (Figure 1) was selected for in-
stalling the first buoy assembly and conducting the feasibility tests.
Water depth at Site #2 was 2,562 feet. Bottom topography insured that the
buoy would not be located on a steep incline; and the sediment sample indi-
cated that anchor slippage would be minimal.

The M. V. Ceylon and a buoy boat were used to install the buoy assembly.
The complete buoy assembly was laid out on the deck of the Ceylon . The
surface float, subsurface buoy, and electrical and mechanical cabling were
arranged in such a way that the entire assembly could be checked out and
installed as a complete unit.

Figure 11 shows how the assembly was distributed at the mooring site. The
surface float was lowered into the water. Rubber floats were attached at
various positions along the assembly so that as the surface float was towed
away from the Ceylon by the buoy boat, the entire assembly was distributed
on the surface of the water. All sensors, transmitters and system compo-
nents were checked out and monitored as they floated on the surface. Loca-
tion position was maintained by the LORAC B. The rubber floats were
released and the anchor finally lowered to swing the taut -moor assembly
into place. Divers made final checks on all equipment and sensors to a
depth of 150 feet; and a final transmission check was conducted with the
shore station at Point Mugu.

DATA EVALUATION

The ODCS buoy assembly was installed on 3 September 1966 and remained
moored for about 7 months. After 3 months of service, an improved mechani-
cal connector was installed at the surface float. Despite interruptions
for various equipment failures , sufficient data of reasonable accuracy were
obtained during the 7-month period to demonstrate the feasibility and value
of the system.

When possible, the processed buoy sensor data were compared with measure-
ments obtained by on-site observations and from meteorological instruments
on San Nicolas Island, Santa Cruz Island, and at Point Mugu. When these
comparisons were made, the processed data were generally found to be

316

reliable and ao_ ate. Examination of large discrepancies in v data
usually disclosed problems that could be attributed to the cabling rather
than to the sensors themselves. Table 4 shows some typical comparisons of
observer and buoy measurements for a few variables.

Table 4
COMPARISON OF OBSERVER AND BUOY MEASUREMENTS

1 October

5 November

18 December

Observer

Buoy

Observer

Buoy

Observer

Buoy

Wind speed
Air temperature
Water temperature

6-8 kn
65.8°F
66.2°F

5 kn
66.4°F
67.8°F

8 kn
60.0°F
63.0°F

9 kn
59.4°F
62.9°F

2 kn
59.3°F
59.3°F

4 kn
59.6°F
59.1°F

Comparative data on wind speed and direction are often difficult to evalu-
ate. Synoptic surface weather charts permit only broad comparisons with
data obtained at a particular buoy location. Hand-held anemometers are
not very accurate; and when an observer is standing on the deck of a small
boat , measurement accuracy further deteriorates . Nevertheless , measure-
ments of wind direction and speed from boats near the installation site
compared well with the data transmitted by the buoy sensors.

Air temperature telemetered by the buoy was within one degree of the tem-
perature as measured from service boats visiting the site. Throughout the
operational test period, buoy measurements of air temperature were reason-
ably accurate, generally within one degree of the reported temperature in
the area.

Continuous comparisons were made of mean barometric pressures at Point Mugu
and San Nicolas Island with the pressures telemetered from the buoy. At
first, large and erratic differences were noted. The sensor was replaced,
and subsequent buoy readings showed minimal differences from the mean pres-
sures at San Nicolas and Point Mugu. The small consistent bias in the buoy
data could easily be adjusted. Hence, the measurements were considered
reliable and accurate.

A few wave height measurements from the buoy were compared with subjective
observations made on service boat visits to the buoy and with synoptic maps.
The original circuit for the wave height sensor on the subsurface buoy did
not permit production of usable data. Once this circuit was modified, how-
ever, buoy measurements compared well with subjective observations from
the boat. Wave height measurements from the buoy also compared well with
calculations based on wind velocity and fetch as shown on the synoptic
weather maps of the area.

On one occasion, wave heights were measured at the buoy site with a free-
floating buoy -mounted wave staff, called an Ocean Wave Profile Recorder
(OWPR). As shown in Table 5, measurements from the buoy generally under-
estimated the wave heights. However, a recent correction of a procedural
error in the computer program is expected to eliminate this bias.

317

Table S

COMPARISON OF COMPUTED SPECTRA OF WAVE HEIGHT DATA

(8 FEBRUARY

1967, 1115 PST)

Buoy

OWPR

Most frequent

0.5 ft

0.6 ft

Average

0.6 ft

0.7 ft

Average of 1/3 highest

1.0 ft

1.2 ft

Average of 1/10 highest

1.2 ft

1.6 ft

Sea-surface water temperatures were compared with bucket observations made
on service visits to the buoy. Differences were within the allowed accu-
racy range.

Subsurface temperatures measured by the buoy sensors were compared with
readings from expendable bathythermographs. On 20 October 1966, a record was
taken from a bathythermograph (BT) located within one-half mile of the buoy.
The purpose of visiting the buoy at that time was to inspect the faulty tem-
perature sensor at the 150-foot depth. The buoy readings at this depth, as
shown in Table 6, reflect the instrument failure. The temperatures at the
other depths , however , were within the accuracy requirements when compared
with the BT values.

Table 6

COMPARISON OF SUBSURFACE TEMPERATURES

(20 OCTOBER 1966)

Temperature

in Degrees Centigrade

Depth in feet

Buoy

1120 PDT

Buoy
1140 PDT

Expendable BT
1130 PDT

50

17.22

17.22

17.1

75

16.74

16.74

16.6

150

18.00

18.00

13.7

500

9.66

9.66

9.5

No comparisons of current speed or current direction data were made during
the operational period. However, measurements at the 75- and 150-foot
depths seemed reasonable when compared with earlier current observations
made during the survey runs.

318

d^

SUMMARY AND CON^ JONS

During the 7 months that the buoy assembly was moored, the processed sensor
data were generally reliable. The meteorological data were periodically
compared with independent observations, and were found usually to be within
the accuracy limits specified for the sensors. Initially, the barometric
pressure sensor caused problems ; but after it was replaced , pressure data
from the buoy compared well with barometric pressures recorded at San Nico-
las Island and Point Mugu. The oceanographic measurements were also gener-
ally reliable and accurate. The wave height sensor as originally installed
was not sufficiently sensitive; however, the problem was corrected, and
subsequent data transmissions were judged reasonable.

During this operational test or "shakedown" period, transmission of data
was not continuous . Interruptions were caused , for example , by radio
interference and by cabling and connector problems. When equipment prob-
lems occurred, the buoy often could not be serviced for days because of
missile test operations in the area. Figure 12 shows the number of trans-
missions possible, the number of transmissions actually received, and the
number of good or reasonable measurements recorded. A good or reasonable
measurement is defined as one that is consistent with previous and subse-
quent measurements and with the synoptic situation, and also compares well
with measurements made by other means .

The ratio of transmissions received to the number possible was rather low.
The ratio of good measurements to the number of transmissions received,
however, was much higher — and improved towards the end of the 7 -month
period. Of the total number of transmissions possible (assuming no system
failures or radio interference), approximately 33 per cent were received.
Of the number received, approximately 80 per cent were considered reason-
able or good data.

Transmission failures were due to causes which can be grouped into four
general categories: problems with buoy assembly components (53 per cent);
radio interference and atmospheric noise ( 30 per cent ) ; failures in shore
station components (13 per cent); and miscellaneous problems (4 per cent).

Most of the difficulties with the buoy assembly involved the cabling and
connectors , especially when they affected a power supply to several sen-
sors. A problem with one cable would then cause distortions in several
other sensor readings. A circuit modification was later incorporated into
the system to overcome these difficulties .

The problem of radio interference was also substantial. Heavy interference
with the buoy signal at night and in the early morning resulted in the
loss of most nighttime and early morning data. This emphasizes the impor-
tance of thorough preliminary examination of frequencies and propagation
conditions to insure good reception of data and minimize interference.
Loss of data, then, was caused primarily by cabling problems and interfer-
ence , rather than by difficulties with the sensors . The meteorological
and oceanographic sensors generally were reliable and provided good measure-
ments.

Some anticipated problems never materialized. It was expected that diffi-
culties might occur with the mooring and with deterioration of the sensors
and buoy components. The mooring, especailly the taut portion, performed
better than anticipated. During the operational test period, at least two
major storms passed over the area: the surface float sustained winds of 30
to HO knots for 12 hours or longer; and the entire assembly withstood the
attendant sea conditions. It was only after 7 months of operation that
the slack moor parted near the lower end; radar surveillance showed that
the buoy had gone adrift on 2 April 1967. The surface float was recovered

319

on 7 Aprf The sensors, surface float, and navigation/ juipment shewed
only modfe. _Le deterioration from exposure to the sea ana ..rather.

Another anticipated problem was that of vandalism and wanton destruction
of the buoy or its components. The buoy assembly was in an area frequented
by fishing boats and pleasure craft, but no such problem developed.

Although the buoy assembly as tested has some problems, none of them is
severe enough to change the basic system or invalidate its use for obtain-
ing oceanographic and meteorological data. The design and construction
of the assembly are basically sound. Its performance during 7 months of
operation and testing showed that taut -moored buoys can function well over
an extended period of time and provide reliable data not obtainable by
other means.

For a buoy assembly more remote than the ODCS, the use of an environmental
data collection and/or communication satellite system should be considered.
A low-orbiting satellite, for example, could interrogate, store, and trans-
mit data from numerous buoy arrays to a data collection and processing
center. Or, a synchronous satellite system could serve for collection and
communication of data from buoy arrays to such a central facility. In
these and other ways , the basic buoy system described in this paper might
be further developed or modified to meet the needs for more extensive col-
lection of measurements of meteorological and oceanographic variables.

ACKNOWLEDGMENTS

The authors gratefully acknowledge the assistance and support of personnel
of the Geophysics Division during the buoy evaluation and operational test
periods. We are particularly indebted to Mr. T. R. Carr and to LCDR
John D. Hague, USN, for their constant interest and valuable support.
Mr. Frederick G. Roehler II was most helpful in electronic and mechanical
analysis, as was Mr. Stevens K. Okura in the data analysis and programming
phase of the tests.

REFERENCES

1. Denton, R. F. , and P. C. Stahl, 1967: "Concept and Design of the SEAS
Buoy System, Trans. 3rd Annual Marine Technical Society Conference and
Exhibit, 5-7 June 1967, 1+93-501.

2. Pierson, W. J., Jr., G. Neumann, and R. W. James, 1955: "Practical
Methods for Observing and Forecasting Ocean Waves by Means of Wave Spectra
and Statistics," U.S. Navy Hydrographic Office Publication No. 603, 284 pp.

320

0

321

FIGURE 2. The buoy assembly consists of the surface float, the sub-
surface buoy, the anchor, the sensors, and the strain and
electrical cables.

322

■i

ANTENNA

BELL

LIGHT

METEOROLOGICAL
SENSORS

RADAR
REFLECTORS

FIGURE 3. The surface float contains the electronic
equipment and power supply — and supports
the mast carrying the meteorological and
navigational instruments .

323

WAVE HE5GHT & WATER
TEMPERATURE SENSORS

FIGURE l+. The subsurface buoy, a cylindrical tank, carries
the wave height and water temperature sensors.

824

AIR TEMPERATURE

WIND DIRECTION

FIGURE 5. The meteorological sensors are mounted on the quadrapod mast.

325

FIGURE 6. The oceanographic sensors measure water temperature,
current speed, and current direction.

326

327

r

L

i

I*
55

5 =
81

1

0 0

O 5

w 2

0

*S

0

p 0

<

Mgy

•S3 3
§££

O P +J
01

(0

5 si

0 s

o C

M 0) TD

S 9J *
o .c

o -P "

w

3*1

<4H 5 0)
4) Mh Q)

Id IT) +3 •

<4H +J fj

fn Jg CO 01

w "2 3

2 o y B

■m c § o
o

&f

»3h m

6 fi <D C

Q) +J +J -H

■P 0)

CO T) 0) 0)

o) ftj +j Q

co m +->

bO-H

M 5) C

•5cTS

■a oi 5J cu
* 0) nj

S •>>+J

CO

328

CLOCK

TELETYPE

ENCODER

RECEIVER

DEMODULATOR

TAPE PUNCH

POWER SUPPLY

FIGURE 9. The shore installation consists essentially
of one rack of equipment.

329

Ai-IONOJIINIfVIOIOtVJlNMI

- — o xoooxxox— ©xooxooxx
jcrzi • • «ii • x ♦ • x • • x . • x x

u.»- XJ1j,*jl*(\j»>f\JB*<^»»r-o.*

•<— O I(VJN(nNm(\JIM-t\l-IMMrt--ft|-

J I Z X •••••••••••••••••

J O £ »

z in *-

o wwNN(vi>jnr)i,n(\ir»ifloflni\io-(\io

■m *\j * r

it »-

uj >
> <
< ■

' X T ^^CO-Jir^Bi^lfl

u.f\jj'aDr\j-«-«ijiff'a'ff,at)aif*-*»'J3*i

kT UJ

mj r

I .3

U.lfl-C-*."^-*J3*^-»J^*U*l-

— X U-

— u. ?

u. ■» -j -j .\ ry - < ry r\, ■* ru ry -* ■* .

*j r t r

> 3 J 3

* Jj Jj Jj

JJ > > >

; a x «

2 X * *

1 iJ J J

j. d r d

« < <x

- r» © « © i*- o o f

uj u"i *j*> if. iou^jitPi/*iPiOi0tOinipiPiiit/i*n

'\ M»* t I

«3<sn9iri<

— ' 25J-TT'3,07,0a-0C>t)*f-*rt — — f\J

oj o ,3u_e©eo— -o— o~ — e.— — (\jr\i*»">©o

C ^

3 I

(jo uj ax— — x— ■ — — ■ © c =• — i— — — e -« © ©

— © £ £.\j-\jl\J\j(\J;\j<\|.\,ry'\j-\jf\j;V)i'\j'\|(\;;\J(,U

TV ^3r©o = ©~o = ©©© = ©;r©©

z-« '-■^■cjo-^^'^rrij'TT'Ooc — — -• m

flAi — ^©c. = ©©©© ceo — — ■-«-• — — —

M C

u —

O (VJ

330

QJ IT)

2 S

3 3

331

CO

J"

CM C7> lO

<-H c~ r~

lO lO lO

oo cm en oo co

lo lo oo oo

oo co co
en en lo

OO 1/5 LO

a)

tJL,

oo oo oo co
co co co co

oo co co oo

LO LO CO LO L/>

CO

en

CO

C

0

•H •

co x)

00

CO

LO 0)

CO

co
co

CO O

4->

(0

4n -M

r^

r-

o c

LO

lO

o

O

^ %

rH

rH

a) S

lO CO (.O

oo ro co

a

LO CO CO

en
co

m

co

en

-r

co

iO

m

rH

rH

a

>

0

2

co lo co

co
co

co co co

+->
CJ

o

co co o o
co co co co
co oo o o

J-

CO

o

co

oo

j-

oo

c~

CO

LO

d-

en

o

CD

o

en

o

en

-a

ex

Q)

GO

U0 LO oo o

CO 00 O rH

rH rH

CO LO
OO CO

o o

•p -a

3

o a) a> <B i>

£ 8. | i |

•H CO x> to xi

o o

+j +j Ch lh lh ch
CM CH

O O O O

lO lO lO lO O O

f~ C^ rH rH lO lO

I I I I I I

g

I

•H

332

LIST OF FIGURES

FIGURE 1. Five sites were surveyed. Site #2 was selected for installa-
tion of the first ODCS buoy assembly.

FIGURE 2. The buoy assembly consists of the surface float, the sub-
surface buoy, the anchor, the sensors, and the strain and
electrical cables .

FIGURE 3 . The surface float contains the electronic equipment and power
supply — and supports the mast carrying the meteorological
and navigational instruments.

FIGURE 4. The subsurface buoy, a cylindrical tank, carries the wave
height and water temperature sensors .

FIGURE 5. The meteorological sensors are mounted on the quadrapod mast.

FIGURE 6 . The oceanographic sensors include instruments to measure water
temperature, current speed, and current direction.

FIGURE 7. A water temperature sensor and the wave height sensor are
mounted on the subsurface buoy at a depth of 50 feet.

FIGURE 8. The Data-Handling Subsystem, within the surface float, con-
trols the scanning, conversion, storage, and transmission
of data from all the sensors . The shore station system ini-
tiates telemetry, encodes the signals, and punches the data
on paper tape for processing by the computer.

FIGURE 9 . The shore installation consists essentially of one rack of
equipment .

FIGURE 10 . A typical data printout shows the format in which data are
produced from the computer.

FIGURE 11. The buoy assembly was distributed on the surface of the water
at Site #2. Location position was maintained by
the LORAC B. The anchor was finally lowered to swing the
taut -moor assembly into place.

FIGURE 12. The chart shows the number of transmissions possible, the

number of transmissions actually received, and the number of
good or reasonable measurements recorded.

333

NVO-235-2

HIG-67-11

34

LOW-FREQUENCY WAVE STUDY
IN THE MESO-DEEP OCEAN

By

MARTIN J. VITOUSEK

Hawaii Institute of Geophysics

and

GAYLORD R. MILLER

Environmental Science Services Administration

Prepared for

U.S. ATOMIC ENERGY COMMISSION

NEVADA OPERATIONS OFFICE

SUPPLEMENTARY PROGRAM

AEC CONTRACT AT(26-1)-235

FINAL REPORT
JULY 1967

HAWAII INSTITUTE OF GEOPHYSICS

UNIVERSITY OF HAWAII

NVO-235-2

HIG-67-11

LOW-FREQUENCY WAVE STUDY
IN THE MESO-DEEP OCEAN

by

Martin J. Vitousek
Hawaii Institute of Geophysics

and

Gaylord R. Miller
Environmental Science Services Administration

Prepared for

U. S- Atomic Energy Commission
Nevada Operations Office
Supplementary Program
AEC Contract AT(26-l)-235

FINAL REPORT
July 1967

Approved by Director
Date: 12 July 1967

LEGAL NOTICE

This report was prepared as an account of Government
sponsored work. Neither the United States, nor the
Commission, nor any person acting on behalf of the Commission:

A. Makes any warranty or representation, ex-
pressed or implied, with respect to the accuracy,
completeness, or usefulness of the information con-
tained in this report, or that the use of any information,
apparatus, method, or process disclosed in this report
may not infringe privately owned rights; or

B. Assumes any liabilities with respect to the use
of, or for damages resulting from the use of any infor-
mation, apparatus, method, or process disclosed in this
report.

As used in the above, "person acting on behalf of
the Commission" includes any employee or contractor of
the Commission, or employee of such contractor, to the
extent that such employee or contractor of the Commission,
or employee of such contractor prepares, disseminates,
or provides access to, any information pursuant to his
employment or contract with the Commission, or his employment
with such contractor.

TABLE OF CONTENTS

Page

INTRODUCTION 1

THE SENSORS 2

THE CABLE 2

SHORE ELECTRONICS 3

INSTALLATION DETAILS 3

ENVIRONMENT AND EXPOSURE 4

TRANSDUCER CALIBRATION 5

CONCLUSION 5

REFERENCES 6

ill

INTRODUCTION

The objective of this research was to install and maintain instru-
mentation for recording low-frequency waves in deep water off the Kona
coast of the island of Hawaii. The signals were to be presented to
the existing AEC data-acquisition system for recording. To the date of
this report, this objective has been achieved.

Timetable

The factor governing the installation timetable was cable delivery.
The specially manufactured cable arrived in Hilo on January 31, 1967.
On the next day, February 1, the Keauhou cable and transducer were
laid at a depth of 90 fathoms at 19°33.5fN, 156°00.0'W. On February 2
the Airport cable and transducer were laid at a depth of 288 fathoms at
19°37.9'N, 156°02.3'W. The following week was spent securing the cable
shore ends and constructing suitable hetrodyne frequency sources to
beat with the final measured signals from the installed Vibrotrons
that would meet the input frequency requirements of the AEC system.
Precision crystal sources at the appropriate frequency were also ordered
at this time. The Vibrotron center frequency of the Keauhou gage is
10,055 Hz and that for the Airport gage is 18,765 Hz.

On February 11 the Airport system became operational employing as
a hetrodyne frequency source the 192nd harmonic of the precision 100 cps
output of a Hewlett-Packard Model 4243L Counter. A Hewlett-Packard
Model 302A Wave Analyzer was employed to filter out this harmonic.

On February 14 the Keauhou system became operational employing the
fourth harmonic of a tuning fork as a hetrodyne frequency source. This
source was lower than the Vibrotron frequency causing an inversion of the
output, i.e., increasing frequency indicated decreasing water level.

During the period February 11 to May 7, two interruptions of signal
were experienced at the airport. The cause, discovered after the second
interruption, was improper output from the wave analyzer resulting from
general power failures. After a power failure the wave analyzer would
lock onto the 193rd harmonic, causing the mixer output to be too high to
pass through the 400 Hz bandpass filter. No interruptions were experienced
at Keauhou.

-1-

-2-

On April 7 the two permanent precision frequency sources were in-
stalled. The outputs were now centered at about 400 Hz and, for both
gages, increasing frequency indicated increasing water depth. No further
interruptions have occurred to date.

THE SENSORS

The sensing units are Vibrotron pressure transducers (Vitousek, 1965)
The diaphragms of these transducers are exposed directly to sea pressure
through an oil-sea interface. They are isolated from environmental
temperature changes by a 5-inch layer of bees wax which has a thermal
time constant of over one hour (Fig. 1). Two Vibrotron amplifiers are
provided for each Vibrotron so that should one fail, the other is
activated by reversing polarity on the cable (Fig. 2). The Vibrotrons
and associated amplifiers are enclosed in a stainless steel pressure case
two inches in diameter and twelve inches long. The case is in turn
imbedded in axle grease in the center of a cylinder of bees wax twelve
inches in diameter and three feet long, which is encased in a twelve-
inch PVC pipe (Fig. 3). The pipe is equipped with three cable clamps
on one side and a twenty-five-pound lead weight on the other. The lead
is attached to the pipe with PVC bolts so that nowhere are dissimilar
metals in electrical contact. A three-hundred PSIA Vibrotron is used
for the ninety-fathom installation and a seven-hundred and fifty PSIA
Vibrotron for the two hundred and eighty-eight fathom installation. The
Vibrotron pressure ports are connected to the sea through oil-filled
nylon tubing which terminates in copper sleeves to retard fouling.

THE CABLE

A special cable was ordered for this installation based on a
successful experience with similar cable on another project. As the
electrical requirements were minimal, a steel conductor was employed to
increase reliability and decrease cost. The conductor consists of a
3/8-inch, 7-wire galvanized steel strand, which had an asphalt-base
flooding compound applied during the stranding operation to prevent water
entering the interstices should a leak occur. Two passes of black high-
density polyethylene insulation were then applied, the second pass was
used to cover any possible pin holes in the first pass. The insulation
was then covered by a 16-strand spiral-wrap layer of galvanized steel wire
to furnish approximately a 95 percent coverage. This armor was then
covered by an outer jacket of black high-density polyethylene. The
cable has a finished diameter of 11/16 inch, a weight of 0.5 lb/foot,

U
o

CO

a

H

00
•H
(i4

CM
-WW-

-5Z- *

p — WW

N
U>

"* az.

a
o
u

h- 1

— WW-

CO

N
<0

-6 o 6 —

-Ift

6 Q <>

§A

CM
<©

z

i

T.

CM
«0

-|<J WVr-

CM

-fcj WW-

S

— 18

UJULUJl)

J

(O

z
o

2
X

o
u

o

O Q.

>

8S

m o

si •*■

CO o

H < S3

gsgi

<o = «J o

3 = 2 K
-» z o u

O - I 2

< ai h *

Ill o =
w li-
h.
o

co <2

o or g

Id uN <

N CO * °

0)

M
0)
•H
«4-l

O

M

o

J3

■a

o

4-1

o

Hi

CM

•H

fa

Mb

-3-

and an electrical loop resistance of five ohms per thousand feet. The
cable was supplied on drums containing 20,000 feet of cable at a cost of
25 cents per foot. At the shore end of the 90-fathom transducer the
sea cable is connected directly to the shore electronics at the head of
Keauhou Bay (Fig. 4). For the deeper transducer the cable emerges
from the ocean approximately one mile from the recording hut (Fig. 5).
A small line amplifier is installed at this point and standard RG 59 U
cable is employed for the remainder of the run to the recording hut,
which is located at the Kona Airport.

SHORE ELECTRONICS

On shore the Vibrotron signal is amplified and hetrodyned against
a highly stable signal, 200 Hz or 400 Hz above the Vibrotron signal
(Fig. 6). For the 90-fathom transducer a hetrodyne frequency of 200 Hz
is used, and the output of the mixer is doubled to 400 Hz. For the
288-fathom transducer the hetrodyne frequency is 400 Hz higher than
the Vibrotron frequency. The hetrodyne frequency is derived from a
crystal oscillator in a proportionally controlled oven. After passing
through a lowpass filter the 400-Hz signals are delivered to the AEC
telemetry system.

INSTALLATION DETAILS

The 90-foot University of Hawaii research vessel TERITU was used
to lay the cable. The transducers were attached to the cable using
the three 3/4-inch cable clamps. The center cable conductor was
swaged to a short length of bare 3/8-inch #16 copper wire using a
3/8-inch Nicopress sleeve. A 1/4-inch Nicopress sleeve was used to
swage half of the outer conductors (armor wire) to another piece of
#16 copper wire. The leads from the Mecca plug were then soldered to
the copper wires, and the splice was sealed using Scotch Fill, Scotch
tape, and Scotch coat. For protection against mechanical damage, the
splice was located in a small cutout about two inches from the end of
the PVC pipe casing and under a steel guard that was secured by the
last of the cable clamps.

For the laying operation, the cable was mounted on a Standard
Reelmaster HRD-3 cable trailer that was borrowed from the local
telephone company. The trailer was secured close to the stern of the
ship, obviating the need for a sheave for the laying operation. The

mm

-4-

cable reel was braked by forcing 4 by 4 timbers against the reel flanges.

Before lowering the instruments, 50-pound kedge anchors were
secured to the cable with 20 feet of 1/2-inch polypropylene line so
that they were just above the transducers. Cable clamps were used
to keep the rope from sliding on the cable. The purpose of the anchors
was to keep the transducers from being pulled along the rough ocean bottom
during the cable-laying operation.

When the ship was on location, the cable was run out a distance
equal to the water depth. Measurement of this cable length was
accomplished by counting turns of the cable spool. During the last
part of the operation, the cable was frequently stopped and manually
felt for bottom contact. On bottom contact, the ship then moved toward
shore very slowly and cable was payed out at a rate slightly greater than
the foreward speed of the vessel. When tension indicated that the
anchors were holding, the ship's speed was increased to about 3 knots
for the remainder of the run, and as the ship neared shore, small
boats were used to pull a bight of cable to the beach and the cable was
cut at the ship.

Where the cables entered the water, they were protected by passing
them through lengths of 1-1/2-inch PVC pipe which were joined by heavy
rubber hoses. This added protection extended from the highest splash
level to a point below the surf action. The pipe sections were secured
firmly to rock on shore and to weights on the bottom to inhibit motion
of the pipe during periods of heavy surf.

ENVIRONMENT AND EXPOSURE

The section of coast line where the installation is located is on the
western or lee side of the island of Hawaii (Fig. 7). The bottom slopes
off very rapidly from this coast making deep-water installations possible
near shore (Fig. 8). Normally the sea is relatively calm and the
weather clear, affording excellent working conditions. The 90-fathom
gage was installed 2.2 nautical miles seaward from Keauhou Bay at the
outer edge of a gradually sloping shelf of 40 fathoms mean depth. The
transducer itself is just over the lip of a very steep drop of the ocean
floor. The 288-fathom gage is 1.4 nautical miles from shore, near the
Kona Airport. It too is on the slope of a very steep drop-off of the
ocean bottom. The two transducers are 5 miles apart. Because the
transducers are located on very steep slopes, the classical laws of
pressure signal attenuation vs. depth for ocean swell are not directly
applicable.

<u

•rl

to

J2

U-l

o

•H

H

00

a

o

to

PC

tO
pq

3
O

tO

01

00
•H

K
UJ

<

O <E

»-
Z
3

<

u.

00
03

O

X

x w or

< UJ

<

u.

o ~ —

o

* 3 c

A

? OT «-

x «t «•

2^

> <c or

?3-

< UJ UJ

FULL VK
RECTIFI
0OUBL

,

1

or

UJ

X

a

10

in

M

_J

v> in
>- o

or

o

g

O 3
0. W

> z

u

00

o
o

M
O

vO

00
•H
fa

,

\ ^KEAHOLE PT.

\ Nil)-.

\ V '.
\ r.*.

1 I KONA COAST OF HAWAII

* ^f I:-'
* •--

* I'*

1 t-v.

J h

V ^L^- r-INSTRUMENT HUT
' /^^^^ K0NA AIRPORT
i / ^% :S%KAILUA

\ V-'.
\ fcr.
\ /*;

1 ^

\ \y,:

'•£

>
\

\

i

V

\
>

INSTRUMENT
p BOX AT
% PIER

y

I

JKEAUHOU

1

<

1

1

1
\
\
\

1 NAUTICAL MILE \

I i '

V1-

1

• 1 \

\

V

\
\

\

\

\

\

\
1
\

Fig. 7. Transducer locations.

in

ro

UJ

g

x
co

s

u.

CO

UJ

CM

-1

a

<

o

M-l
O

h-

4J

3

(0

<
Z

ccj

o
o

o
o

CM

O
O

O
O
(0

O

o

CO

o
o
o

o
o

CM

SWOHlVd Nl Hid30

c
o

o

ex,

s

o

4J
4J
O

cd
u

00

00
•H

-5-

TRANSDUCER CALIBRATION

The Vibrotron selected for the 90-fathom installation has an excep-
tionally small temperature coefficient of 0.04 Hz per degree centigrade.
The sensitivity of the Vibrotron is 4.63 cps/psi. Since this sensitivity
is doubled in the mixer, the sensitivity at the 400-Hz output is 9.26 Hz/psi.
Speaking in terms of water level, the sensitivity is 4.08 cps per foot of
ocean-surface vertical movement, and a one-degree temperature change of
the transducer would produce a signal change corresponding to about a
quarter of an inch of ocean surface change. This excellent temperature
stability makes the 90-fathom transducer useful as an offshore tide gage.
The actual temperature variations at the gage location will be the subject
of a future research project.

The 288-fathom gage has a sensitivity of 8.074 Hz per psi and a rather
poor temperature coefficient of 6.24 Hz per degree centigrade. The bees
wax insulation, however, does prevent any environmental temperature varia-
tion of tsunami period from reaching the transducer. Also, as the gage is
at a greater depth, the environmental temperature changes are likely to
be smaller.

Under constant input pressure, both gages indicated an equivalent rms
noise level of less than 1/32 inch of water during laboratory tests. For
on-sight checking of the system, a one MHz multiple-period counter, a
digital to analog converter, and a strip chart recorder were employed.
Figures 9 and 10 show typical analog conversions of the signals from the
transducers.

CONCLUSION

One of the design criteria for the instrument system was that it
should be able to record the actual noise level of the ocean in the period
range where this noise level is known to be very low (Munk e_t al . , 1963;
Snodgrass e_t al. , 1966). The performance has met this criterion.

The installation was designed to be semi -permanent. It is hoped that
long-term records can be made from these gages which will, upon analysis,
yield significant information on the nature of this noise level.

-6-

REFERENCES

Munk, W. H., G. R. Miller, F. E. Snodgrass, and N. F. Barber, 1963,
Directional recording of swell from distant storms, Roy. Soc.
London Phil. Trans. A, 255, 505-584.

Snodgrass, F. E., G. W. Groves, K. F. Hasselmann, G. R. Miller, W. H.
Munk, and W. H. Powers, 1966, Propagation of ocean swell across
the Pacific, Roy. Soc. London Phil. Trans. A, 259, 431-497.

Vitousek, M. J., 1965, An Evaluation of the Vibrotron Pressure

Transducer as a Mid-ocean Tsunami Gage, Hawaii Inst. Geophys.
Report 65-13, 12 pp., 7 figs.

X

o

z

ELAPSED TIME IN MINUTES

Fig. 9. Typical record, 90-fathom gage (Keauhou)

x

15 20 25 30

ELAPSED TIME IN MINUTES

35

40

45

Fig. 10. Typical record, 288-fathom gage (Kona Airport).

NO. OF COPIES ADDRESSEE

15 Defense Documentation Center (DDC)

Cameron Station
Alexandria, Virginia 22314
ATTN: TISIA-21

4 U. S. Atomic Energy Commission

Nevada Operations Office
P. 0. Box 1676
Las Vegas, Nevada 89101
ATTN: Robert H. Thalgott

Assistant Manager for Operations

3 U. S. Atomic Energy Commission

Division of Technical Information

Extension
P. 0. Box 62

Oak Ridge, Tennessee 37831
ATTN: T. W. Laughlin

1 U. S. Atomic Energy Commission

Albuquerque Operations Office
P. 0. Box 5400

Albuquerque, New Mexico 87115
ATTN: D. W. King, Patent Attorney

1 Chief of Research and Development

Department of the Army
Washington, D. C. 20310
ATTN: Atomic Division

3 Chief of Engineers

Department of the Army
Washington, D. C.
ATTN: ENGGW-NE (1 copy)
ENGTE-E (1 copy)
ENGMC-E (1 copy)

1 Commanding Officer

U. S. Army Corps of Engineers
Coastal Eng. Res. Center
Washington, D. C. 20315

2 Director, Waterways Experiment Station

U. S. Army Corps of Engineers
Vicksburg, Mississippi

ATTN: Library (1 copy)
Mr. J. Strange (1 copy)

NO. OF COPIES ADDRESSEE

2 Chief of Naval Operations

Navy Department
Washington, D. C. 20350
ATTN: OP-75 (1 copy)
OP-03EG (1 copy)

2 Chief of Naval Research

Navy Department
Washington, D. C. 20390
ATTN: Code 811 (1 copy)
Code 418 (1 copy)

2 Commanding Officer and Director

U. S. Naval Civil Engineering Lab.
Port Hueneme, California
ATTN: Code L31 (1 copy)
Mr. T. O'Brien (1 copy)

1 Commanding Officer

U. S. Naval Radiological Defense Lab.
San Francisco, California 94129
ATTN: Technical Information Div.

1 Commander

U. S. Naval Oceanographic Office
Suitland, Maryland

1 Commanding Officer and Director

David Taylor Model Basin
Washington, D. C. 20007

ATTN: Library

2 Commander Joint Task Force 8

Sandia Base

Albuquerque, New Mexico 87115
ATTN: J- 3 (1 copy)
JTG 8.3 (1 copy)

1 Underwater Explosions Res. Div.

DTMB, Norfolk Naval Shipyard
Portsmouth, Virginia

1 Director, U.S.A.F. Project RAND

via: U.S. Air Force Liaison Office

The RAND Corporation

1700 Main Street

Santa Monica, California

NO. OF COPIES ADDRESSEE

1 Director of Defense Research and Engr.

Washington, D. C. 20301
ATTN: Technical Library

4 Lawrence Radiation Laboratory

Box 808

Livermore, California

ATTN: Mr. Paul Fackler (1 copy)

Mr. R. J. Stinner (1 copy)

Dr. C. J. Leith (1 copy)

Dr. M. Knapp (1 copy)

3 Los Alamos Scientific Laboratory

P. 0. Box 1663

Los Alamos, New Mexico 87544
ATTN: Dr. W. E. Ogle (1 copy)
Dr. A. Cox (1 copy)
Dr. Kenneth Olsen (1 copy)

1 Chief, Classified Technical Library

Technical Information Service
U. S. Atomic Energy Commission
Washington, D. C. 20545

1 University of California

Scripps Institute of Oceanography
8602 La Jolla Shores Drive
La Jolla, California
ATTN: Dr. W. G. Van Dorn

1 United Research Services

1811 Trousdale Drive
Burlingame, California
ATTN: Mr. Kenneth Kaplan

1 Oceanics Division of Interstate

Electronics Corporation
707 E. Vermont Avenue
Anaheim, California

1 New York University

New York City, New York
ATTN: Dr. H. B. Keller

1 Oceanographic Services, Inc.

5373 Overpass Road
Santa Barbara, California
ATTN: Dr. R. Kent

NO. OF COPIES ADDRESSEE

3 Director, Defense Atomic Support Agency

Washington, D. C. 20301
ATTN: STBS

1 Hawaiian Institute of Geophysics

University of Hawaii
2525 Correa Road
Honolulu, Hawaii
ATTN: Dr. W. M. Adams

1 Director

U. S. Army Corps of Engineers
Nuclear Cratering Group
Livermore, California

1 Commander

U. S. Naval Ordnance Test Station
China Lake, California

1 Commanding Officer and Director

Naval Electronics Laboratory
San Diego 52, California

2 Air Force Weapons Laboratory

Kirtland Air Force Base

New Mexico 87117

2 Commander

Field Command
DAS A

Sandia Base

Albuquerque, New Mexico 87115
ATTN: GCWT (1 copy)
FCTG (1 copy)

1 Sandia Corporation

Sandia Base

Albuquerque, New Mexico 98115
ATTN: Classified Document Division
Dr. M. L. Merritt

1 Dr. B. LeMehaute

Tetra Tech
630 N. Rosemead
Pasadena, California 91105

NO. OF COPIES ADDRESSEE

1 Cambridge University

Department of Geodesy and Geophysics
Cambridge, England

1 Canadian Hydrographic Office

Marine Sciences Branch
615 Booth Street
Ottawa, Canada
ATTN: Dr. G. Dohler

1 Carnegie Library of Pittsburgh

4400 Forbes Avenue
Pittsburgh, Pennsylvania 15213
ATTN: Head, Book Order Office

1 College of Marin

Library
Kentfield, California

1 Colorado State University

Libraries

Fort Collins, Colorado 80521
ATTN: Miss Lorene M. Ashton

1 Navy Department

Washington, D. C. 20360
ATTN: Code 1622-B

1 Dartmouth College

Baker Memorial Library
Hanover, New Hampshire 03755
ATTN: Acquisitions Department
Mrs. Barbara B. Brown

1 U. S. Naval Branch

Oceanographic Office

FPO San Francisco, California 96618

ATTN: R. P. Higgins

1 The Library of Congress

Exchange Section
Exchange and Gift Division
Washington, D. C. 20540

1 Dr. R. F. McAllister

Department of Ocean Engineering
Fla. Atl. Univ.
Boca Raton, Florida

NO. OF COPIES ADDRESSEE

1 William B. Merselis

184 South Hotel Street, Room 203
Honolulu, Hawaii 96813

1 The New York Public Library

Preparation Division, Acquisition Branch
5th Avenue and 42nd Street
New York, New York

1 Pacific Oceanographic Laboratory

ESSA, Institute for Oceanography
1801 Fairview Avenue East
Seattle, Washington 98102

1 Mr. Robert Q. Palmer

P. 0. Box 996
Kailua, Hawaii 96734

1 Mr. Dale Rhee

233 Merchant Street

R. M. Towill Corporation

Honolulu, Hawaii 96813

1 Seattle Public Library

Seattle, Washington 98104
ATTN: Margarete E. Chalfant

Reference Document Assistant

1 University of Hawaii

Hawaiian and Pacific Collection
Gregg Sinclair Library
Honolulu, Hawaii 96822
ATTN: Miss Janet Bell, Librarian

1 University of Hawaii

Dr. Wytze Gorter
Director of Research
Spaulding 357
Honolulu, Hawaii 96822

1 University of Illinois

Urbana, Illinois 61801
ATTN: Library, Serials Department

1 U. S. Weather Bureau

Emergency Warning Branch
Washington, D. C. 20235
ATTN: Mr. A. Pearson

35

Reprinted from TRANSACTIONS AMERICAN GEOPHYSICAL UNION Vol. 48, No. 2

Reprinted from

U. S. National Report, 1963-1967

Fourteenth General Assembly

Transactions, American Geophysical Union

Vol. 48, No. 2, June 1967

Tides and Other Long Period, Waves
Bernard D. Zetler

Institute for Oceanography

U. S. Environmental Science Services Administration

Silver Spring, Maryland

Until recent years, the subject of tides rested
comfortably on a plateau of achievement dat-
ing back almost to the beginning of this cen-
tury. Within the past few years, however, the
electronic computer has suddenly transformed
all aspects of the field: observations, reductions,
analyses, predictions, and research. It is ironic,
in a sense, that significant improvements in
predicting techniques have been made in tides,
the only phenomenon in physical oceanography
for which there was a reasonably satisfactory
level of achievement already. Some progress
has been made in studies of tsunamis, but the
major need, a real-time method of forecasting
tsunami heights, has not yet been developed.

Tide predictions. Munk and Cartwright
[1966] developed a response method of predic-
tion, in which the input functions are the time-
variable spherical harmonics of the gravita-
tional potential and of radiant flux on the
Earth's surface. This method is the first suc-
cessful major departure from the traditional
solutions, in which the tide oscillations are de-
scribed "by the amplitudes and phase lags for a
finite set of cosine curves of predetermined fre-
quencies. Previously, the only nongravitational
constituents had been the first and second har-
monics of the seasonal variations; the response
method uses a radiational input in all species.

The proposal to substitute the Munk-Cart-
wright method for the traditional prediction
procedure warrants careful consideration, for
there is much more at stake than just the sta-
tistical comparisons of residual variance [Zetler
and Lennon, 1967]. If the new procedure is ac-
cepted, the tidal harmonic constants will no
longer be available. These constants are used
not only for predictions but for co-tidal and
co-range charts and tidal characteristics as
well. The complex admittances obtained from
the weights assigned to the various input series
are proposed to replace these constants as a
measure of response of the oceans to the tide-
producing forces. If one plots the phase angles
for the traditional semidiurnal harmonic con-
stants as a function of frequency, one usually
finds a siiarp bend or even a discontinuity at
30° per solar hour. It is not reasonable that
the oceans should exhibit so sharp a variation
in response to a particular frequency. By way
of contrast, the Munk-Cartwright admittances
in species 2 are far smoother (undoubtedly
related to the use of a radiational input), and
even the jitter described in the paper has since
been eliminated by adding third-order (largely
frictional) interaction series to the input. The
conclusion seems to be that the traditional S2
harmonic constants are a trigonometric com-

592

IUGG QUADRENNIAL REPORT (U.S.A.)

bination of gravitational and radiational re-
sponses of identical frequencies. If so, then in-
ferences made for smaller constituents through
equilibrium considerations only are incorrect.
This appears to explain demonstrated inconsist-
encies in the inference of some semidiurnal
solar constituents.

Thus, the Munk-Cartwright method offers
not only the simplicity of a single set of input
predictions that can be used in preparing pre-
dictions for all places on the surface of the
Earth but a more satisfactory documentation of
the response of the oceans to the tide-produc-
ing forces as well. Nevertheless, an argument
against adopting this method is the rather ap-
palling prospect, even in these days of electronic
computers, of calculating the new type of re-
sponses for much of the tidal data in the
archives and of re-educating tidal authorities
to think in terms of complex admittances rather
than harmonic constants. It will be a difficult
decision.

Zetler and Cummings [in press, 1967] de-
veloped a method for identifying and solving
for the significant frequencies in the tide spec-
trum. In working with shallow water tides at
Anchorage, Alaska, they used a total of 114
constituents in every species up to 12 cycles
per day. Inasmuch as the Munk-Cartwright
method can use any order of non-linear inter-
action in the input functions, the method can be
extended to shallow water tides, such as at
Anchorage. This has not yet been done.

Harris et al. [1965] described the program-
ming effort that retired the U. S. Coast and
Geodetic Survey mechanical tide predictor after
more than fifty years of service. The change to
electronic computers produced a saving in man-
power and greater security, since the Coast
and Geodetic Survey no longer has to rely on
one unique piece of equipment. More important
scientifically, however, the computer permits
greater flexibility in predictions by removing
the restriction of the 37 specific constituent
speeds on the tide predicting machine.

Tide observations and analysis. Barbee
[1967] described three new tide gages: a digital-
recording device driven by a float in the stand-
ard stilling well, a pressure-sensing instrument
suitable for installation in harsh environments
or where no coastal structures exist, and a pres-
sure-sensing gage capable of operating at great

depth [also see Goodheart et al., 1965]. The
Coast and Geodetic Survey uses the digital
(one-tenth hour) data in a computer program
that outputs times and heights of high and
low waters, mean ranges, lunitidal intervals,
various mean planes, and monthly extremes.
Snodgrass [1967] developed a free falling in-
strument capsule with a Vibroton absolute-
pressure transducer for measuring pressure at
a depth of several kilometers with very high
resolution.

Hicks et al. [1965] described tide observa-
tions about 130 km offshore on the Atlantic
continental shelf at about 210 and 260 meters.
Alldredge and Fitz [1964] described observa-
tions at 30 meters on a submerged stabilized
platform in 1280 meters of water. Munk was
named chairman of an international working
group on deep sea tides at the 1965 IAPO meet-
ing in Paris.

Harris et al. [1963] programmed a least-
square analysis that solves for the harmonic
constants of all constituents simultaneously, as
opposed to the traditional method of a modified
Fourier analysis for individual constituents.
After an objective evaluation of five analytical
processes was completed [Zetler and Lennon,
1967], the method was accepted by the Coast
and Geodetic Survey for series of at least one
year. Twenty-nine day analyses are still done by
traditional formulas on a computer. Munk and
Hasselmann [1964], in a study of super-resolu-
tion, pointed out that the precision of tide pre-
diction is ultimately limited by the underlying
noise spectrum. Zetler et al. [1965] extended the
least-square analysis approach to data observed
in random time.

Supported by an IUGG resolution at Berk-
eley in 1963, Munk has collected, edited, and
prepared (in computer-compatible format) long
series of tide data for numerous places through-
out the world. One method of tide data editing
used at IGPP of the University of California
at San Diego was described by Zetler and
Groves [1964].

The U. S. Naval Oceanographic Office [1965]
published a very useful atlas of tides and cur-
rents in the North Atlantic. The engineering
and legal aspects of tidal datum planes were de-
scribed in considerable detail by Shalowitz
[1962, 1964].

Studies of the continuum. Munk and Bull-

ASSOCIATION OF PHYSICAL OCEANOGRAPHY

593

ard [1963] used narrow bandpass filters to ex-
amine the level of the continuum between spe-
cies (cycles per day). Munk et al. [1965] then
extended the study by looking at the level be-
tween the groups and even between the con-
stituents within the same group. In the vicinity
of strong tidal lines, they found that the con-
tinuum rises into cusps, presumably because of
nonlinear interaction of the lines with the peak
of the continuum near zero frequency. The
shallow water predictions by Zetle'r and Cum-
mings [1967] indicated that nonlinear inter-
actions of tidal lines probably contribute to
the cusps; it has also been suggested that the
cusps are associated with a modulation of the
internal tides by the slowly varying thermal
structure. Groves and Zetler [1964} used 50
years of tide observations at San Francisco and
Honolulu to find a reasonably uniform phase lag
of maximum coherence over a wide range in the
low frequencies, but they failed to find any sig-
nificant peaks at frequencies other than those in
tidal theory. Wunsch [1966] found that the
biweekly and monthly tides are not purely
equilibrium tides but forced planetary waves
with a scale of the order of 1000 miles, much
smaller than that of the tide-producing poten-
tial or even of the oceans. Rattray and Charnell
[1966] obtained planetary wave solutions for
quasi-geostrophic free oscillations in enclosed
basins of dimensions comparable to the Pacific,
Atlantic, and Indian oceans. The baroclinic
oscillations (restricted to regions adjacent to
the equator or the ocean boundaries) may ex-
plain observed fluctuations with periods of one
or a few months; the barotropic oscillations
that fill the complete basins may explain ob-
served periods of three or more days.

Sea level. Donn et al. [1964] prepared a
very comprehensive survey of recent sea level
studies. They included wind waves as well as
long waves, pointing out that all sea waves are
sea level variations. The complete range of
wavelengths (centimeters to global) and wave
periods (less than a minute to more than a
century) was presented by Stommel [1963] in
a schematic three-dimensional power spectrum
of sea level variations. Inasmuch as wind waves
are covered in a parallel report, sea level is used
here in a smoothed sense, averaged over a pe-
riod of a month or more. Patullo [1963] de-
scribed seasonal variations on a regional basis

and discussed the relative contributions of as-
sociated phenomena. Rodeh [1966] cross-cor-
related monthly means of sea level, atmospheric
pressure, and sea temperature in a comprehen-
sive study of sea level variations in the Pacific.
Hicks and Shofnos [1965a] contoured the rate
of land emergence in southeast Alaska from
sea level observations and they [Hicks and
Shofnos, 19656] made new determinations of
regional sea level trends in the United States.

Tidal energy. Miller [1966] has used recent
tide and tidal current analyses to arrive at
modified estimates for frictional dissipation of
tidal energy. His estimate for the Bering Sea
is considerably lower than previous estimates
by Jeffreys and Heiskanen, but his total is com-
parable.

Tsunami. The tsunami generated by the
Alaskan earthquake on March 28, 1964, de-
vastated many coastal zones in Alaska and in-
flicted severe damage on the west coast of the
United States [Spaeth and Berkman, 1965].
The wave at Crescent City, California, reached
21 feet above chart datum; this record eleva-
tion was clearly related to the orientation of
energy at the source. The Crescent City in-
undations changed some opinions as to the
vulnerability of the west coast of the United
States to tsunamis of distant origin.

Van Dorn [1964a] interpreted the available
evidence to indicate a dipolar movement of the
Earth's crust, centered along a line running
from Hinchinbrook Island, Prince William
Sound, southwesterly to the Trinity Islands.
The U. S. Coast and Geodetic Survey [1964]
listed the amount of vertical land movement at
various places, the largest movement being a
change of 31.5 feet on Montague Island. Donn
[1964] suggested that the waves of up to 6
feet that occurred at about the same time along
the coasts of Louisiana and Texas were gen-
erated by the seismic waves.

The International Union of Geodesy and
Geophysics [1964] issued a comprehensive bib-
liography of tsunamis compiled and edited by
the U. S. Coast and Geodetic Survey. Berning-
hausen [1964, 1966] has compiled data on
tsunamis in the Atlantic and Indian oceans.
Weigel's [1964] Oceanographic Engineering,
and Wilson's [1965] report are fine references
for many aspects of tsunami research. Miller
[1964] and Loomis [1966] showed that spectra

594

IUGG QUADRENNIAL REPORT (U.SA.)

for various tsunamis at any one recording sta-
tion are similar in appearance, whereas spectra
for the same tsunami at different recording sta-
tions are ordinarily quite different. Van Dorn
[19646] demonstrated that the theory of geo-
metric optics adequately predicts the behavior
of waves radiating from a large explosion.
Wong et al. [1964] used linearized shallow wa-
ter wave theory in analyzing the interaction of
a tsunami with an isolated oceanic island.

Tsunami warning systems. The U. S. Coast
and Geodetic Survey [1965] was host to the
Intergovernmental Oceanographic Commission
Conference on the International Aspects of the
Tsunami Warning System at Honolulu in 1965.
The twelve countries represented at the meeting
discussed possible improvement in the existing
warning systems and made a series of recom-
mendations to the Intergovernmental Oceano-
graphic Commission. A local tsunami warning
service was developed for Alaska, centered at
Anchorage. The Hilo model tests were com-
pleted by the U. S. Army Corps of Engineers;
the results were summarized in a report of the
Hilo Technical Tsunami Advisory Council
[1965]. A report by Cox [1964] described the
state of the art in tsunami forecasting. Vitousek
[1965] continued his efforts to develop a mid-
ocean tsunami gage, a development that was
urgently needed not only to improve the warn-
ing service but for research purposes as well.

Joint tsunami research effort. The U. S.
Coast and Geodetic Survey and the Institute of
Geophysics, University of Hawaii, established
the Joint Tsunami Research Effort on the
campus of the University. Under a subsequent
reorganization, the Environmental Science Serv-
ices Administration took responsibility for the
government participation. Several very capable
scientists have joined the ESSA and University
of Hawaii contingents, and some promising re-
search studies are now under way.

References

Alldredge, L. R., and J. C. Fitz, Submerged
stabilized platform, Deep-Sea Res., 11, 935-942,
1964.

Barbee, W. D., Tide gages in the U. S. Coast and
Geodetic Survey, Report, Unesco and IAPO
Symp. Tide Gage Instrumentation, Paris, 1965,
in press, 1967.

Berninghausen, W. H., Tsunamis and seismic
seiches reported from the eastern Atlantic

south of the Bay of Biscay, Bull. Seismolopical
Soc. Am., 64, 439-142, 1964.

Berninghausen, W. H., Tsunamis and seismic
seiches reported from regions adjacent to the
Indian Ocean, Bull. Seismological Soc. Am., 66,
69-74, 1966.

Cox, D. C, Tsunami forecasting, Hawaii Inst.
Geophys., Univ. Hawaii, H1G-64-15, 22 pp., 1964.

Donn, W. L., Alaskan earthquake of 27 March
1964 : Remote seiche stimulation, Science, 145,
261, 1964.

Donn, W. L., J. G. Pattullo, and D. M. Shaw, Sea-
level fluctuations and long waves, in Research
in Geophysics, II, Solid Earth and Interface
Phenomena, pp. 243-267, Massachusetts Insti-
tute of Technology Press, Cambridge, Mass.,
1964.

Goodheart, A. J., C. W. Iseley, and S. D. Hicks,
Deep sea tide gage, in Ocean Science and Ocean
Engineering, vol. 1, Marine Technology Society
and American Society of Limnology and Ocea-
nography, 1965.

Groves, G. W., and B. D. Zetler, The cross spec-
trum of sea level at San Francisco and Hono-
lulu, J. Marine Res., 22, 269-275, 1964.

Harris, D. L., N. A. Pore, and R. Cummings, The
application of high speed computers to practi-
cal tidal problems, Abstracts of Papers, VI,
IAPO, XIII General Assembly, IUGG, Berke-
ley, 1963.

Harris, D. L., N. A. Pore, and R. A. Cummings,
Tide and tidal current prediction by high
speed digital computer, Intern. Hydrographic
Rev., 42, 95-103, 1965.

Hicks, S. D., A. J. Goodheart, and C. W. Iseley,
Observations of the tide on the Atlantic con-
tinental shelf. J. Geophys. Res., 70, 1827-1830,
1965.

Hicks, S. D., and W. Shofnos, Yearly sea level
variations for the United States, J. Hydraulics
Div., Proc. Am. Soc. Civil Engrs., HY 5, 23-32,
1965a.

Hicks, S. D., and W. Shofnos, The determination
of land emergence from sea level observations
in southeast Alaska, J. Geophys. Res., 70, 3315-
3320, 1965b.

Hilo Technical Tsunami Advisory Council, Physi-
cally feasible means for protecting Hilo from
tsunamis (unpublished), Third Report to the
Board of Supervisors, Hawaii County, through
Its Tsunami Advisory Committee, 38 pp., 1965.

International Union of Geodesy and Geophysics,
Annotated bibliography on tsunamis, Mono-
graph 27, 249 pp., 1964.

Loomis, H. G., Spectral analysis of tsunami rec-
ords from stations in the Hawaiian Islands,
Bull. Seismological Soc. Am., 56, 697-713, 1966.

Miller, G. R., Tsunamis and tides, PhX). disserta-
tion, University of California, San Diego, 1964.

Miller, G. R., The flux of tidal energy out of deep
oceans, J. Geophys. Res., 71, 2485-2490, 1966.

Munk, W. H., and E. C. Bullard, Patching the
long- wave spectrum across the tides, J. Geophys.
Res., 68, 3627-3634, 1963.

ASSOCIATION OF PHYSICAL OCEANOGRAPHY

595

Munk, W. H., and D. E. Cartwright, Tidal spec-
troscopy and prediction, Phil. Trans. Roy. Soc.
London, A, 259, 533-581, 1966.

Munk, W., and K. Hasselmann, Super-resolution
of tides, in Studies on Oceanography (Hidaka
Volume), pp. 339-344, University of Washington
press, Seattle, Washington, 1964.

Munk, W. H, B. Zetler, and G. W. Groves,
Tidal cusps, Geophys. J., #), 211-219, 1965.

Patullo, J. G., Seasonal changes in sea level, in
The Sea, vol. 2, edited by M. N. Hill, pp. 485-
496, Interscience, New York and London, 1963.

Rattray, M., Jr., and R. L. Charnell, Quasi-
geostrophic free oscillations in enclosed basins,
J. Marine Res., 24, 82-103, 1966.

Roden, G. I., Low-frequency sea level oscillations
along the Pacific coast of North America, J.
Geophys. Res., 71, 4755-4776, 1966.

Shalowitz, A. L., Shore and sea boundaries, U. S.
Coast Geodetic Surv. Publ. 10-1, vol. 1, 420 pp.,
1962.

Shalowitz, A. L., Shore and sea boundaries, U. S.
Coast Geodetic Surv. Publ. 10-1, vol. 2, 749 pp.,
1964.

Snodgrass, F. E., Digital recording of tides in the
deep sea, Report, Unescb and IAPO Symp. Tide
Gage Instrumentation, Paris, 1965, in press, 1967.

Spaeth, M. G., and S. C. Berkman, The tsunami
of March 28, 1964, as recorded at tide stations
(unpublished), 59 pp., U. S. Coast and Geodetic
Survey, 1965.

Stommel, H., Varieties of oceanographic experi-
ence, Science, 139, 572-576, 1963.

U. S. Coast and Geodetic Survey, Preliminary re-
port, tide datum plane changes, Prince William
Sound, Alaskan earthquake, March-April 1964
(unpublished), 4 pp., 1964.

U. S. Coast and Geodetic Survey, Intergovern-
mental Oceanographic Commission report of
the working group meeting on the international
aspects of the tsunami warning system in the
Pacific, Honolulu, Hawaii, April 27-30, 1965
(unpublished), 33 pp., 1965.

U. S. Naval Oceanographic Office, Oceanographic
atlas of the North Atlantic Ocean, sect. 1, Tides
and currents, Publ. 700, 1965.

Van Dorn, Wm. G., Source mechanism of the
tsunami of March 28, 1964, in Alaska, Proc.
Ninth Conf. Coastal Eng., Am. Soc. Civil Engrs.,
10, 166-190, 1964a.

Van Dorn, Wm. G., Explosion-generated waves in
water of variable depth, J. Marine Res., 22, 123-
138, 1964b.

Vitousek, M. J., An evaluation of the vibroton
pressure transducer as a mid-ocean tsunami
gage, Hawaii Inst. Geophys., Univ. Hawaii, 65-
13, 12 pp., 1965.

Wiegel, R. L., Oceanographical Engineering, 532
pp., Prentice-Hall, Jnc, Englewood Cliffs, New
Jersey, 1964.

Wilson, B. W., Generation and dispersion char-
acteristics of tsunamis, in Studies on Oceanog-
raphy (Hidaka Volume), pp. 413-444, Univer-
sity of Washington Press, Seattle, Washington,
1965.

Wong, K. K., A. T. Ippen, and D. R. F. Harleman,
On the interaction of a tsunami with an isolated
oceanic island, Trans. Am. Geophys. Union, 45,
69-70, 1964.

Wunsch, C. I., On the scale of the long-period
tides, Ph.D. dissertation, Massachusetts Insti-
tute of Technology, 1966.

Zetler, B. D., and R. A. Cummings, A harmonic
method for predicting shallow water tides, J.
Marine Res., 25, 103-114, 1967.

Zetler, B. D., and G. W. Groves, A program for
detecting and correcting errors in long series of
tidal heights, Intern. Hydrographic Rev., 41,
103-107, 1964.

Zetler, B. D., and G. W. Lennon, Some compara-
tive tests of tidal analytical processes, Intern.
Hydrographic Rev., 44, 139-147, 1967.

Zetler, B. D., M. D. Schuldt, R. W. Whipple, and
S. D. Hicks, Harmonic analysis of tides from
data randomly spaced in time, /. Geophys. Res.,
70, 2805-2811, 1965.

36

Reprinted from INTERNATIONAL HYDROGRAPHIC REVIEW Vol. kk , No. 1

SOME COMPARATIVE TESTS
OF TIDAL ANALYTICAL PROCESSES

by B. D. Zetler
Physical Oceanography Laboratory, Institute for Oceanography, ESSA

and G.W. Lennon
University of Liverpool Tidal Institute and Observatory

Abstract

Tests have been conducted on five analytical processes and the results
have been examined by statistical and power spectral techniques. Tide
observations at Atlantic City (small range and large low-frequency noise),
Swansea (large range and noise primarily in tidal frequencies), and San
Francisco (small range and moderate low frequency noise) were used in the
study. Residuals were obtained by subtracting the predicted hourly heights
from observations and were evaluated for total energy (variance) and energy
per frequency band. The latter calculation was found to be useful in compar-
ing residual energy for particular portions of the frequency spectrum.

Introduction

The advent of electronic computers has made possible new approaches
to tidal analysis. Numerous organizations have initiated least squares ana-
lytical approaches in which a set of fixed frequencies (speeds of tidal
constituents) is proposed to fit the data so that the sum of the squares of
the residuals is a minimum. This modified the traditional approach, as
exemplified by Doodson (1928) and Schureman (1941), in which only one
constituent is examined at a time and the results are modified for the inter-
ference effects of other constituents in the same species. Munk and
Gartwright (1966) have recently developed a response method of prediction
in which the input functions are the time-variable spherical harmonics of
the gravitational potential and of radiant flux on the Earth's surface.

This study was initiated to permit comparative evaluations of some of
the proposed analytical processes. Inasmuch as Munk and Hasslemann
(1964) have shown that the ability to separate two frequencies depends on
the noise level, it was decided to conduct the tests with real data. The tests
have been conducted on the Coast and Geodetic Survey method (Schureman,
1941), the Harris-Pore-Cummings (1963) least squares method that has

140 INTERNATIONAL HYDROGRAPHIC REVIEW

been accepted by the Coast and Geodetic Survey for series of one year, the
Doodson method as treated by Lennon (1965), the Murray (1963) least
squares method, and the Munk-Cartwright (1966) method. The data,
hourly heights of tide, included Atlantic City (1939), Swansea (1961-1962),
and San Francisco (1931 and 1939).

Atlantic City Comparisons

This is the only test in the study that involves all five methods. Table 1
shows the results obtained by analyzing the data, predicting for the same
period by using the derived constants, subtracting the predictions from the
observations, and examining the residuals. At first, only the total variance
of the residuals was obtained, but it quickly became obvious that a more
refined procedure was needed because the small differences between total
variances appeared relatively insignificant. Therefore, the power spectra
were calculated to permit comparisons of residual energy at various fre-
quencies. A peak of energy in the low frequencies shows up clearly on
figure 1 (the spectrum of residuals from the Doodson method), even though

10'

10"

SPECTRAL ANALYSIS OF ATLANTIC CITY RESIDUALS
DOODSON ANALYSIS

1 2 3 4 5 6 7 8 9 10 11 12

Frequency in cpd

Fig. 1. — Spectrum of residuals from Doodson method for Atlantic City, 1939.

the energy scale (vertical) is logarithmic. Atlantic City is directly on the
open coast and the shallow water fetch is short. Hence, the non-linear
interaction terms (ordinarily manifested as compound tides in the high
frequencies) are small. However, the station is wide open to the effect of
all storms on the North Atlantic and the results confirm that the noise is
due to storms having effective periods of greater than a day. Figure 2 shows
the large daily fluctuations in mean sea level for the year analyzed.

COMPARATIVE TESTS OF TIDAL ANALYSES

141

0)

CO

0)
I— I

><

o
o

Eh

•J

<

O
Cm

w

>

W
i-3

w

CO

w
S

CM

o

CO

w

D

>
Q

el
Q

142 INTERNATIONAL HYDROGRAPHIC REVIEW

The degree of resolution in the power spectrum analysis was chosen
to adequately separate the species, to minimize the computer effort, and to
have sufficient degrees of freedom to provide reasonable confidence in the
results. Inasmuch as the data are hourly, the Nyquist frequency is 12
cycles per day. Using 100 lags in the autocorrelations, we get a " delta f "
of .12 cpd which essentially separates the species by about eight values,
a satisfactory separation. The number of degrees of freedom in a Tukey
analysis is roughly 2 N/M, where N is the number of data points and M
is the number of lags used in the autocorrelations (Munk et ah, 1959). With
a year of data, there are about 170 degrees of freedom, a rather conservative
and satisfactory number.

The results shown in table 1 show the Munk-Cartwright method to
have a small advantage over the two least square methods. The small
differences between the Murray and Harris-Pore-Cummings methods may
be caused by other factors, such as the number of significant digits used in
the calculation. Both least squares methods appear to do slightly better
than the Doodson method. Considering the limitations on accuracy imposed
by the minimizing of correction processes when hand computations were
the order of the day, the Doodson method does remarkably well. The fact
that the Harris-Pore-Cummings does as well as the Murray with 19 fewer
constituents relates to the particular station being analyzed, the additional
constituents (primarily compound tides) are essentially trivial (all are less
than .02 foot) and therefore do not give the Murray (or Doodson) method
any significant advantage for this particular station. The Harris-Pore-
Cummings computer program is dimensioned for 41 constituents and this
was considered a limiting characteristic at the time the computations in
this study were made. Since that time the program has been used with as
many as 114 constituents for Anchorage (Zetler and Cummings, 1966)
simply by separating the constituents into groups and using the restriction
that all constituents within any species had to be included in the same
group. As a matter of fact, this approach reduces the computing effort. For
example, it is far simpler to reduce three 40 x 40 matrices than one 120
x 120 matrix. The poorer results for the traditional C&GS method are
clearly related to the small number of constituents in the solution and
there is no doubt that it would have moved into the general area of the
other results had a greater effort been considered warranted. The C&GS
harmonic constants for Atlantic City were taken from the archives and
were typical of the methods that were used twenty years ago when manual
stencil summation for each constituent limited strongly the number of
constituents sought. In more recent years, summations have been made
on an electronic computer but the remainder of the process is continued
by hand (Schureman, 1941). There is no question that the remainder of the
process can be programmed for an electronic computer and it has recently
been done for 15- and 29- day series. It has not been programmed for
a year by the Coast and Geodetic Survey because the least square approach
has been shown to be acceptable and to have greater flexibility in the choice
of series length and by virtue of not necessarily requiring data equally
spaced in time (Zetler et al., 1965).

Consideration was given to plotting in one illustration the spectra for

COMPARATIVE TESTS OF TIDAL ANALYSES

143

all five sets of residuals. In general they are so close together that the
composite graph would be confusing. Therefore only one is shown in
figure 1 to more or less typify the station characteristics.

Swansea Comparisons

Although the test conditions for the Swansea data were comparable
to those for Atlantic City, it is clear that the results in table 1 cannot be
interpreted in terms of comparative evaluations of the Harris-Pore-
Cummings and Doodson methods. Unlike the conclusions for Atlantic City,
it is obvious that some of 23 additional constituents (60 compared to 37)
improve the prediction and thereby reduce the residual variance.

In species 2 (two cycles per day) there are six constituents in the
Doodson analysis that are not included in the standard Coast and Geodetic
Survey analysis. The total energy in these six (using energy equals one
half amplitude squared) is .0256, somewhat greater than the .0173 difference
obtained from the table 1 values for residual variance in species 2. Inasmuch
as the calculated values for one constituent can change if another consti-
tuent in the same species is added to the model in the least square analysis,
no rigorous comparisons can be made from the results. On the other hand,
it seems obvious that if the same model (set of constituents) were used with
both methods, it is unlikely the results would differ significantly. Therefore
the principal interest in these results is in Figure 3 where the contribution
of the additional constituents is demonstrated in the various species. Even
here, in considering these comparisons, it is important to keep in mind
that the vertical scale is logarithmic.

10'

10-4

SPECTRAL ANALYSIS OF SWANSEA RESIDUALS

\ * HARRIS-PORE-CUMMINGS

DOODSON

Fig. 3.

5 6 7

Frequency in cpd

10

12

Spectra of residuals from Doodson and Harris-Pore-Cummings methods
for Swansea, one year (1961-1962).

144

INTERNATIONAL HYDROGRAPHIC REVIEW

W
-J

CQ
<

-a

s a

k 4)

8

as

^

4) ^

> OC

t- >-

41 4>

09 C

•a m

T3
O

eo r-- th ^as

o ooo o
ooooo

sooo
to 00

ON

* •

coco

t^CM

oo
oo

lOOOWOt

CM CM

oo

CO CM

s

co irtiH eo cm

o t->

00 OS

•^00

HrHrtilfq

•fOO

O Tf

OO

CO CO CO CO CO

"* ■<*

OO

O 1H

a

0

ooooo

CO CO CO CO CO
CM CM CM CM CM

CO o o o o

CM CM' Cm' CM Cm'

OO 00 CJC^

COO ■f''*

OSOS COO

0000 CM CM

oo
oo

CO CO

oo
coo

oo

OSOS
CM CM'

OS

CO

t/3

OS

C

fl

. 4>

U o

as

•fl—1

b"»

o

3 c
£ O

o

CM

CO

os

co
os

CO

OS

CO
OS

OS OS th Tf

IftlflfH

" OI>

. COM

a

cu
to

05

•~ 00O5

to

o
cj
.°°

Cj

B
a
t,

fc,

c

c
03

00 OS
i-HUO

03

3

er
00

cs

CJ

J
c

03 ««

U "O

fa °

3 O

4)

t. _

ga

C/3 C

4)

- fa
to 3

03 J2
4) 4) o

2 «« >>

8 5 >

■M S^

.3 3 G

■n 1 4)

■C >_ o
t. o«

S u «

3 « o

3

O

1

a>
t-,
o

B 1

o

to to

O t,
O 03

o
>
t-i

3
Vi

.as

■S.SP
SI

^o3

+2 £i

03 S

o3

4>
►
S-

3
C/3

£

4-

4J

T3

i~

O

?

0)

O

01

«a

U

en

c

0

a

OS

O in
l«

4>

M OS
M 9

.2 cd

.CJ 00

"O -*

» T2

<- c
p. a

.2 rt

o ..

* 9
b 9 °

t; * os

S u °
a§.2

p o 3

■a e 1

3 v <a

as-"

.Sao
0 .2

T3 , 'J

4) "B 1)

en O ^

3 » 13

a* 0.

.S? a it

u .2 t*

o 4) T-

Eo g

3=«3 s

COMPARATIVE TESTS OF TIDAL ANALYSES

145

San Francisco Comparisons

The results shown in table 1 for the Coast and Geodetic Survey and
the Munk-Cartwright methods for San Francisco are derived from different
basic data and therefore they are not directly comparable. Nevertheless,
some of the results are quite interesting.

The C&GS predictions for San Francisco for both 1931 and 1939 are
based on mean harmonic constants from 11 one-year analyses and the
annual and semi-annual constituents (Sa and Ssa) are based on 19 years
of monthly mean sea level. The 1931 Munk-Cartwright predictions were
based on an analysis of the same year and therefore it was to be expected
that the residual variance would be less. Spectra were not computed for
the 1931 residuals.

The 1939 Munk-Cartwright predictions were made with weights

derived from just one year, 1931. Table 1 shows the C&GS ahead on total

variance but behind in species 2. Figure 4 shows comparisons in each
species.

The big difference is in the low frequency portion of the spectrum.
Although Sa and Ssa fall within the first frequency band (0 to .06 cpd), one
effect of tapering the autocorrelations is to widen the main pass band to
include the two bands on either side, in this case to 0 to .30 cpd (Munk
et al., 1959). Inasmuch as Sa and Ssa are due to thermal variations (the
astronomical input to Ssa is small), they are only quasi-stationary and a
mean of 19 years of data can be expected to furnish a more reliable
predictor than just one year. Therefore, it is not surprising to find the sum

10'

10l

SPECTRAL ANALYSIS OF SAN FRANCISCO RESIDUALS

5 6 7

Frequency in cpd

10

12

Fig. 4. — Spectra of residuals from Munk-Cartwright and Coast and Geodetic Survey

methods for San Francisco, 1939.

146 INTERNATIONAL HYDROGRAPHIC REVIEW

of residual energy from 0 to .30 cpd is .0898 for the Munk-Cartwright
method and .0683 for the C&GS method.

The Munk-Cartwright method does remarkably well with stationary
phenomena in species 1, 2, and 3. Part of its apparent advantage is tempered
by the smaller number of variables in the C&GS method (2 x 18 = 36
compared to 59) but nevertheless it is quite an achievement to do so much
better with the one year of input compared to 11 years for the C&GS
method.

The C&GS method appears to have done better in species 4 but actually
there was no contest. The Munk-Cartwright prediction did not include an
input for bilinear interaction of species 2 with itself which would have
produced a species 4 prediction. A bilinear input for the interaction of
species 1 with species 3 could have been used also but this would have
been considerably less important.

Summary

Objective test procedures have been developed for comparing tidal
analytical processes. The use of power spectral techniques for evaluating
residual energy for particular species has been found to be a useful tool.
As a direct result of these tests, the Coast and Geodetic Survey has accepted
the Harris-Pore-Cummings least squares procedure for the routine analysis
of series of one year in length.

In retrospect, it is unfortunate that the only direct comparison of the
five analytic methods was done with the least complicated set of data. This
deprives the more sophisticated methods of a chance to demonstrate
competence to cope with additional complications. Perhaps the study can
be extended in the future to include various methods for predicting shallow
water tides.

During the tests, it was found that the residual variances from the
Harris-Pore-Cummings method did not match the claimed reduction in
variance computed as part of the program. The resulting investigation
found a small error in the program and this was corrected. The data shown
in this paper are from the corrected program and therefore some values do
not match the data in a preliminary report (Zetler and Lennon, 1966). The
detection of the discrepancy was an unanticipated by-product of the study.

An initial objective of this study was to compare computer cost as
well as accuracy of analysis. This becomes confusing for various reasons,
particularly since the same computers at different installations may have
significantly different costs. With one exception, the Munk-Cartwright
method, the computer costs are so low that differences are not very signi-
ficant. Unlike the other methods which were developed for routine produc-
tion usage, the Munk-Cartwright method was devised for research purposes.
It would be unfair to compare it on the basis of cost until some effort has
been made to make the program more efficient in a cost sense.

Possibly the greatest importance of the study is that representatives of
tidal offices in two countries have agreed to use the same data to evaluate
their analytical processes and have proposed criteria for the comparisons.

COMPARATIVE TESTS OF TIDAL ANALYSES 147

The authors hope that this is the beginning of a broader international
program along similar lines.

Acknowledgement

The authors thank Dr. Walter Munk for providing the training and
computing facilities necessary for the tests involving the Munk-Cartwright
response method. The availability of BOMM (Bullard et al., 1966), a
system of programs for the analysis of time series, reduced significantly the
programming effort in this study.

REFERENCES

Bullard, E.G., Oglebay, F. E., Munk, W. H. and Miller, G. B. (1966) :
A User's Guide to BOMM, Institute of Geophysics and Planetary
Physics, University of California, San Diego.

Doodson, A. T. (1928) : The Analysis of Tidal Observations, Phil. Trans,
of the Royal Society of London, Series A, Vol. 227.

Harris, D. Lee, Pore, N. A. and Cummings, Bobert (1963) : The Application
of High Speed Computers to Practical Tidal Problems, Abstracts of
Papers, Vol. VI, IAPO, XIII General Assembly, IUGG, Berkeley,
VI-16.

Lennon, G. W. (1965) : The Treatment of Hourly Elevations of the Tide
using an IBM 1620, International Hydrographic Review, Vol. XLII, No.
2, 125-148.

Munk, W. H. and Cartwright, D. E., in press, Tidal Spectroscopy and
Prediction, Royal Society Transactions A.

Munk, W. H. and Hasselmann, K. F. (1964) : Super-Resolution of Tides,
Studies on Oceanography (Hidaka Volume), 339-344.

Munk, W. H., Snodgrass, F. E. and Tucker, M.J. (1959) : Spectra of Low-
Frequency Ocean Waves, Bull, of the Scripps Institution of Oceano-
graphy, Vol. 7, No. 4, 283-362, University of California Press.

Murray, M. T. (1963) : Tidal Analysis with an Electronic Digital Computer,
Cahiers Oceanog., 699-711.

Schureman, Paul (1941) : Manual of Harmonic Analysis and Prediction
of Tides, U.S. Coast and Geodetic Survey Spec. Pub. No. 98.

Zetler, B. D. and Cummings, B. A. (1966) : An Objective Method for Identify-
ing Hidden Frequencies in Shallow Water Tides, Abstract in Transac-
tions, American Geophysical Union, Vol. 47, No. 1, 117-118.

Zetler, B. D. and Lennon, G. W, in press, Evaluation Tests of Tidal
Analytical Processes, UNESCO Publication on Tide Symposium at
Paris, 1965.

Zetler, B. D., Schuldt, M. D., Whipple, B. W. and Hicks, S. D. (1965) :
Harmonic Analysis of Tides from Data Bandomly Spaced in Time,
Journal of Geophysical Research, Vol. 70, No. 12, 2805-2811.

37

Reprinted from PROCEEDINGS THE ELEVENTH PACIFIC SCIENCE CONGRESS
Volume 2, Oceanography , Toyko 1966

A HARMONIC METHOD OF PREDICTING SHALLOW WATER TIDES

Bernard D. Zetler* and Robert A. Cummings**. * Institute for Oceanography, Environmen-
tal Science Services Administration. ** Coast and Geodetic Survey, Environmental
Science Services Administration

The development of an objective technique for identifying significant hidden
frequencies in the spectrum makes it possible to accurately predict shallow water
tides by harmonic methods. For Anchorage, Alaska, the 114 constituents that are used
include frequencies in every species (cycles per day) from 0 to 12. This large set
of constituents improves the predictions (times of high and low waters, range of tide,
and shape of curve). The technique was also applied to Philadelphia data. Two addi-
tional years at Philadelphia were then analyzed for the same constituents. The com-
parison of harmonic constants for the three years permits an evaluation of the station-
ary characteristics of the added constituents.

38

Reprinted from JOURNAL OF MARINE RESEARCH Vol. 25, No. 1

A Harmonic Method

for Predicting Shallow-water Tides

Bernard D. Zetler and Robert A. Cummings

Physical Oceanography Laboratory
Institute for Oceanography, ESS A
Division of Oceanography
U.S. Coast and Geodetic Survey, ESS A

ABSTRACT

The development of an objective technique for identifying significant hidden frequencies
in the spectrum makes it possible to accurately predict shallow-water tides by harmonic
methods. For Anchorage, Alaska, the 114 constituents used include frequencies in every
species (cycles per day) from o to 12. The larger set of constituents improved the predictions
in times of high and low waters, range of tide, and shape of curve. The stationary character-
istics of some of the added constituents have been tested with three years of Philadelphia data.

This study was initially designed for a specific purpose — to improve tidal
predictions at Anchorage, Alaska. Anchorage is located at the northern end
of Cook Inlet near the western end of the Gulf of Alaska. The mean range
of tide at Anchorage is about 25 feet. After oil was found in the area, huge
deep-draft tankers required more accurate tidal predictions than were available
by using the standard U.S. Coast and Geodetic Survey procedures for tidal
analysis and prediction (9).

There was no question as to why the problem existed. Those harmonic
constants that were determined in the Coast and Geodetic Survey analysis
showed large amplitudes for some shallow-water constituents — clear indic-
ations of additional significant compound tides that are not included in the
routine analysis. The compound tides are associated with the distortion of the
sinusoidal shape of the tidal curve as the tidal wave travels over shallow depths (5).

British authorities have had to cope with shallow-water tides to a much
greater extent than the U.S. Coast and Geodetic Survey, and therefore they
have traditionally used 60 tidal constituents (3) compared with the 37 used
by the Coast and Geodetic Survey. In some extreme cases, the 60 constituents
do not adequately describe the shape of the curve, and the British have devel-
oped a non harmonic modification of their procedures to cope with these tides (4).

1. Accepted for publication and submitted to press 15 October 1966.

IO3

REPRINT FROM JOURNAL OF MARINE RESEARCH, VOLUME 25, 1, 1967

104 ^Journal of Marine Research [25,1

A logical approach, therefore, was to send the Anchorage data to the Tidal
Institute and Observatory, University of Liverpool, for analysis and prediction
until such time as the Coast and Geodetic Survey could learn the British
techniques. However, this did not solve the problem. We were informed that
the Tidal Institute routine analysis obviously would not match the shape of
the curve and that a continuous record of hourly heights for one year would
be required for the nonharmonic method. Inasmuch as the harbor at Anchor-
age freezes every winter, the required length of record was not available.
Although an effort to get a continuous year of record was initiated by installing
a pressure gauge on the bottom, there remained an element of doubt as to
whether the tidal characteristics remain unchanged during the winter freeze.

Fortunately, recent technical changes in tidal analysis and prediction made
possible another approach. Essentially, the principal change provided greater
flexibility in both analysis and prediction in that additional constitutents can
now be included. Until now there was a constraint to work only with a fixed
set of constituents; no others could be readily analyzed for, nor could they be
included in, the prediction.

Traditional analysis included (i) a modified Fourier analysis for particular
frequencies, (ii) a modification of the results for the interference effects of
nearby frequencies, and (iii) an elimination of sideband contributions of fre-
quencies in the same species — the same number of cycles per day (9,3). A least-
square analysis in which any combination of frequencies is inserted as a model
is now readily accomplished on a computer. A recent study showed that the
harmonic constants for the same set of constituents are slightly more accurate
than those obtained by the previous Coast and Geodetic Survey or British
(Doodson) methods (10).

The Coast and Geodetic Survey tidal predictions until 1964 were made on
a mechanical analog computer having gears designed for a fixed set of frequencies.
The predictions now made on an electronic computer do not have the above
restriction; in this particular study, 114 constituents have been included.

In shallow water, rhe nonlinear interaction among large-amplitude con-
stituents generates additional constituents whose frequencies are integral
sums or differences of the frequencies of known constituents (8). It is necessary
to identify the important additional compound tides, include them in a new
analysis, and then predict, using the enlarged set of harmonic constants. The
availability of the BOMM (1) programs for time-series analysis makes some
of these steps relatively easy.

First, a routine 37-constituent analysis was made of 192 days of Anchorage
hourly data for the middle of 1964. Using the derived constants, hourly heights
were predicted and subtracted from the observations, and then a spectral anal-
ysis was made of the residuals. This identified frequency bands of greatest
energy in the residuals, and a Fourier analysis, using maximum resolution,
was made for these bands. Wherever large values stood out above the continuum

1967] Zetler and Cummings : Predicting Shallow-water Tides

105

10"'

'\ 37 CONSTITUENTS
I \

114 CONSTITUENTS

5 6 7

Frequency in cpd

Figure 1. Spectral analysis of Anchorage residuals.

in a plot of the Fourier amplitudes, an effort was made to identify an integral
combination of frequencies of constituents that were known to be important
and that closely matched the frequencies of these peaks. A new least-square
analysis was performed, adding these new frequencies to the original 37. As
a check, a new total prediction was prepared, new residuals were determined,
and Fourier- and power-spectrum analyses were conducted with these data.

Fig. 1 shows the comparative power spectra with residuals from 37 and
114 constituents, respectively. The solid line below the dotted line shows
that a significant improvement has been made in every species.

Fig. 2 shows the results of Fourier analysis near two cycles per day for the

uj 4

LA.

?.3

UJ

a

z

z

2

3 c.2.5
z" ^JSz

* osSS

z
5

2.2

#

G_

•

•

.

s 1

•

• ■*■••

<
n

. .

,...,....:••.

••••!.""*•• f"

'.•

.' .1

J

l.'"".-"

•1 •••'•••.

••7.-...V

.-'•r-...

25 26 27 28 29 30

CONSTITUENT SPEEDS IN °/hr.

25 26 27 28 29 30

CONSTITUENT SPEEDS IN °/hr.

32

33

»—

uj 4

Ll_

Z o

— .3

UJ

O

B.2

-

Q-

S 1

<

• . •

#,

n

•.•!...•

■..VI.""" "•• -f" ' - •

■'•.J ••'. • •■!••

• *• I..*"" '.'"•L •••

.••••.. r.-.. •

.'.:.t„ ,.

32

33

Figure 2. High-resolution Fourier analysis near 2 cycles per day; Anchorage. Top: Residuals from
37 constituents. Bottom: Residuals from 114 constituents.

io6

Journal of Marine Research

[25,1

.41-
.3
.2
.1
.0

-

0

z
S

to

0

s

CM

of* *T

CXtN <

'•

........ .,...;*..

.• !"••."

•

70 71 72 73

CONSTITUENT SPEEDS IN °/hr.

74

.J

z

0
2.2

_i

5 j

n

i

...... •..••*». ...•*.

...!•..••••. ■

U..."\...*1.«.. .-..

70 71 72 73

CONSTITUENT SPEEDS IN °/hr.

74

Figure 3. High-resolution Fourier analysis near 5 cycles per day: Anchorage. Top: Residuals from
37 constituents. Bottom: Residuals from 114. constituents.

two sets of residuals. Although a major portion of the energy has been removed,
it is obvious that some remains. In a recent paper on tidal cusps (7), it was
shown that the continuum rises in cusps near the large tidal lines. There are
indications of a systematic residual midway between the groups (identified
by the first two digits of the Doodson number), including speeds of about
28. 2°, 28. 70, 29.2°, and 29. 8° per hour. Combinations of constituents sig-
nificantly more contrived than those used would be required to approximate
these speeds, and this has not been done because the cusps are now believed
to include both line-line and line-noise interactions (W. H. Munk, personal
communication).

Fig. 3. shows a comparable Fourier analysis of the residuals near five cycles
per day. No other study has been found in which this portion of the spectrum
has been examined for tidal lines. There is significantly less evidence of a cusp
in the final residuals than in Fig. 2; this could have been anticipated from the
spectral analysis in Fig. 1. Furthermore, cusps are found adjacent to large
tidal lines; these do not occur at five cycles per day.

Table I identifies all of the 114 constituents and shows their amplitudes.
The subscript at the right of the name indicates the number of cycles per day.

1967] Zetler and Cummings : Predicting Shallow-water Tides

107

Table I. Anchorage Tidal Constituents.

Name Source Doodson No.

Sa 056.555

Ssa 057.555

Mm 065.455

MSf 073.555

Mf 075.555

2Qj 125.755

ff, 127.555

Qt 135.655

Ql 137.455

O, 145.555

MPj M„-P, 147.555

Mj 155.655

Xl 157.455

Pj 163.555

S, 164.555*

Kt 165.555

J, 175.455

2POx 2P!-©! 181.555

SO! S2-0, 183.555

OOj 185.555

2NS2 2N2-S2 217.755

2NK2S2 2N2 + K2-2S2 219.755

MNS2 M2+N2-S2 227.655

MNK2S2 M2 + N2 + K2-2S2 229.655

2MS 2K2 2M2 + S2-2K2 233.555

2N2 235.755

H2 237.555

N2 245.655

v2 247.455

2KN2S2 2K2 + N2-2S2 249.655

OP2 O, + Pt 253.555

M2 255.555

MKS2 M2 + K2 - S2 257.555

M2(KS), M2 + 2K2-2S2 259.555

2SN(MK)2 2S2+N2-M2-K2 261.655

A2 263.655

L2 265.455

T2 272.556

S2 273.555

R2 274.554

K2 275.555

MSN2 M2 + S2 - N2 283.455

2KM(SN)2 2K2 + M2-S2-N2 287.455

2SM2 2S2-M2 291.555

SKM2 S2 + K2-M2 293.555

NOs N2 + 0! 335.655

* Doodson uses 164.556 with a speed of 15.0000020.

Speed

Amplitude

(°M

(feet)

.0410686

.519

.0821373

.182

.5443747

.139

1.0158958

.347

1.0980331

.101

12.8542862

.041

12.9271398

.129

13.3986609

.210

13.4715145

.030

13.9430356

1.197

14.0251729

.236

14.4966939

.169

14.5695476

.026

14.9589314

.588

15.0000000*

.119

15.0410686

2.238

15.5854433

.088

15.9748272

.060

16.0569644

.225

16.1391017

.084

26.8794590

.072

26.9615963

.136

27.4238337

.162

27.5059710

.107

27.8039338

.066

27.8953548

.368

27.9682084

.683

28.4397295

1.830

28.5125831

.493

28.6040041

.072

28.9019669

.145

28.9841042

11.039

29.0662415

.108

29.1483788

.087

29.3734880

.095

29.4556253

.274

29.5284789

.573

29.9589333

.128

30.0000000

2.937

30.0410667

.200

30.0821373

.922

30.5443747

.208

30.7086493

.054

31.0158958

.181

31.0980331

.073

42.3827651

.101

io8

yournal of Marine Research

[25,1

Table I. Anchorage Tidal Constituents (continued).

Name Source Doodson No. Speed Amplitude

(°/hr) (feet)

2MK3f 2M2-K, 345.555 42.9271398 .311

M3 355.555 43.4761563 .084

S03 Sj + O! 363.555 43.9430356 .157

MKS M2 + K! 365.555 44.0251729 .218

SK3 S2 + K! 383.555 45.0410686 .141

N4 2 N2 435.755 56.8794590 .093

3MS4 3M2-S2 437.555 56.9523127 .117

MN4 M2 + N2 445.655 57.4238337 .280

MNKS4 M2 + N2 + K2-S2 447.655 57.5059710 .085

M4 455.555 57.9682084 .839

SN4 S2 + N2 463.655 58.4397295 .079

KN4 K2+N2 465.655 58.5218668 .134

MS4 M2+S2 473.555 58.9841042 .538

MK4 M2 + K2 475.555 59.0662415 .123

SL4 S2 + L2 483.455 59.5284789 .056

S4 491.555 60.0000000 .061

MN05 M2+N2 + 0! 535.655 71.3668693 .077

2M05 2M2 + 0! 545.555 71.9112440 .141

3MP5 3M2-PX 547.555 71.9933813 .064

MNK6 Mj + Njj + K! 555.655 72.4649023 .061

2MP6 2JVL. + P! 563.555 72.9271398 .134

2MK5 2M2 + Kj 565.555 73.0092770 .198

MSK5 M2+S2 + K, 583.555 74.0251728 .097

3KM5 K2 + K! + M2 585.555 74.1073100 .042

3NKS6 3N2 + K2-S2 627.855 85.4013258 .081

2 NM, 2 N2 + M2 635.755 85.8635632 .090

2NMKS6 2N2 + M2 + K2-S2 637.755 85.9457005 .102

2MNS 2M2 + Na 645.655 86.4079380 .281

2MNKS6 2M2+N2 + K2-S2 647.655 86.4900752 .098

M6 655.555 86.9523127 .507

MSN6 M2 + S2+N2 663.655 87.4238337 .093

MKN6 M2 + K2+N2 665.655 87.5059710 .127

2MS6 2M2 + S2 673.555 87.9682084 .483

2MK, 2M2 + K2 675.555 88.0503457 .146

NSK6 N2+S2 + K2 683.655 88.5218668 .104

2SM„ 2S2 + M2 691.555 88.9841042 .090

MSK6 M2 + S2 + K2 693.555 89.0662415 .057

S6 6E1.555** 90.0000000 .006

2MNO, 2M2+N2 + 0! 735.655 100.3509735 .059

2NMK, 2N2 + M2 + K! 745.755 100.9046318 .079

2MSO, 2M2 + S2 + Ot 763.555 101.9112440 .095

MSKO, Mi+S2 + Yiz + Ol 783.555 103.0092771 .067

2(MN)8 2M2 + 2N2 835.755 114.8476674 .053

3MN8 3M2 + N2 845.655 115.3920422 .125

3MNKS8 3M2+N2 + K2-S2 847.655 115.4741795 .062

Me 855.555 115.9364169 .160

f 2MK3 is named MO3 in Admiralty Manual of Tides, p. 68.

** According to Doodson code, E = +6 and 1 = — 6 above and below base 5.

I967J

Zetler and Cummings : Predicting Shallow-water Tides

109

Table I. Anchorage Tidal Constituents (continued).
Name

2MSN8 2M2 +

2MNK8 2M2 +

3MS8 3M2 +

3MK8 3M2 +

MSNK8 M2 + S

2(MS)8 2M2 +

2MSK8 2M2 +

2M2NK, 2M2 +

3MNK, 3M2 +

4MK, 4M2 +

3MSK9 3M2 +

4MN10 4M2 +

M10 5M2

3MNS10 3M2 +

4MS10 4M2 +

2MNSK10 2M2 +

3M2S,0 3M2 +

4MSKU 4M2 +

4MNS12 4M2 +

5MS12 5M2 +

3MNKS12 3M2 +

4M2S,„ 4M.+

Durce

Doodson No.

Speed

Amplitude

(°/hr)

(feet)

S2+N2

863.655

116.4079380

.088

Na + K2

865.655

116.4900752

.055

s2

873.555

116.9523127

.230

K2

875.555

117.0344500

.056

jj+Nj + Kj

883.655

117.5059710

.083

2S2

891.555

117.9682084

.087

S2 + K2

893.555

1 18.0503457

.046

2N2 + K,

945.755

129.8887360

.036

N2 + Kt

955.655

130.4331108

.015

K,

965.555

130.9774855

.023

S2+Kt

983.555

131.9933813

.039

N2

1045.655

144.3761464

.051

1055.555

144.9205211

.055

N2 + S2

1063.655

145.3920422

.065

s2

1073.555

145.9364169

.103

N2 + S2 + K2

1085.655

146.4900752

.052

2S2

1091.555

146.9523127

.060

S2 + Kx

1183.555

160.9774855

.033

N2 + S2

1263.655

174.3761464

.051

s2

1273.555

174.9205211

.056

N2 + K2 + S2

1283.655

175.4741795

.042

2S2

1291.555

175.9364169

.045

In considering the amplitudes, it is important to remember that, although a
tidal prediction is a summation of cosine curves, the high and low waters oc-
cur when the sum of the first derivatives is equal to zero. In the first derivative,
each constituent is weighted according to its speed. Therefore, 2MS6, with
an amplitude of about one-half foot, contributes as much to the times of high
and low waters as a diurnal constituent having an amplitude of three feet.
There are two problems concerning these results that need to be explored.
First, as the number of interactions to obtain a particular frequency increases,
the number of combinations that add up to this frequency also increases. For
example, the Coast and Geodetic Survey uses 2 MK3 for the frequency that
is named MO 3 by the British. Although the speeds are the same, the node
factors and phase corrections are not. In seeking to identify a peak in the high
resolution Fourier plot, the tendency is to accept the first combination of
constituents that satisfies the data; there may be a more logical one that has
not been discovered and, in any case, the multiplicity of satisfactory combin-
ations is bound to introduce some error due to the different corrections for
the longitude of the moon's node. Inasmuch as equilibrium relationships are
not necessarily valid, there does not seem to be any way to resolve'the problem
except, possibly, by analyzing a large number of consecutive years of data and
empirically determining node corrections.

no

"Journal of Marine Research

[25,1

Furthermore, are these constituents part of a stationary process? That is,
if the harmonic constants are used for future predictions, will they fit the
phenomena at that time? This is the characteristic that made reliable tidal
predictions possible many years ago. The easiest way to check this point is to
compare the harmonic constants derived from different series of data. Data
were not available to do this for Anchorage, but the same procedure was used
with three years of Philadelphia data (1946, 1952, and 1957) for a smaller
set of constituents (Table II). The set of constituents was determined from
only the 1957 data.

Using subjective criteria, 16 of the original 37 constituents were un-
satisfactory as compared with 9 of the added 24 constituents (roughly a similar
ratio). Traditionally, the Coast and Geodetic Survey omits from its predictions
analyzed constituents that have amplitudes of less than .03 foot because the
phase tends to be unreliable for such small constituents. Of the original 16
poor values, six are larger than .03 foot. However, five of these are long-period
constituents (Sa, Ssa, Mm, Mf, and MSf), and it has been shown (6, 7) that
the continuum rises sharply in the low frequencies, making the .03-foot limit
too low in this portion of the frequency spectrum. This left one unsatisfactory
routine constituent, 2N2, and two unsatisfactory new constituents, KN4
and MKN6, that were greater than .03 foot. A study of all unsatisfactory
new constituents showed that the sum of K2 and N2 appeared in six of the
nine. Furthermore, there appeared to be for these six constituents a consistent
pattern in the phase relationships that indicated that a very small change in

Table II. Tidal Constants; Philadelphia.

^ Amplitude ^

1946 1952 1957

(feet) (feet) (feet)

Mm 0.187 0.067 0.048

MSf 0.166 0.056 0.146

Mf 0.045 0.123 0.108

Ssa 0.129 0.123 0.153

Sa 0.370 0.153 0.267

MPif 0.027 0.042 0.039

Xi\ 0.022 0.020 0.036

KP,f 0.024 0.011 0.037

2Qt 0.013 0.003 0.031

Qt 0.030 0.051 0.031

e, 0.018 0.013 0.030

O, 0.264 0.289 0.266

M, 0.019 0.008 0.018

Px 0.067 0.061 0.096

Kx 0.316 0.331 0.342

Sj 0.089 0.065 0.081

J, 0.030 0.006 0.010

r

- Phase lag*-

N

1946

1952

1957

16.3

72.2

335.0

19.8

67.9

348.8

80.8

292.6

304.6

44.6

115.3

1.0

104.5

55.6

91.2

308.2

308.2

294.0

119.5

146.3

155.4

123.1

355.2

76.8

343.0

100.0

140.1

230.3

225.2

209.3

225.8

81.3

199.8

199.0

198.6

202.0

342.0

15.2

252.3

222.2

207.7

195.6

212.7

209.2

211.1

170.4

175.5

151.5

257.5

130.7

108.2

1 967 J Zetler and Cummings : Predicting Shallow-water Tides 1 1 1

Table II. Tidal Constants, Philadelphia (continued).

s Amplitude ^

1946 1952 1957

(feet) (feet) (feet)

OOj 0.010 0.010 0.003

MSN2f 0.020 0.020 0.038

2 N2 0.084 0.080 0.041

Hi 0.158 0.168 0.138

N2 0.403 0.425 0.417

v., 0.143 0.145 0.120

M2 2.583 2.676 2.506

A2 0.090 0.095 0.061

L2 0.221 0.238 0.283

T2 0.032 0.026 0.015

S2 0.322 0.346 0.328

R2 0.019 0.005 0.016

K2 0.108 0.094 0.078

2SM2 0.014 0.019 0.022

2NP3| 0.015 0.015 0.016

N03t 0.022 0.019 0.024

S03t 0.026 0.026 0.030

M3 0.025 0.029 0.014

MK3 0.081 0.083 0.068

2MK3 0.080 0.088 0.068

ML4f 0.046 0.061 0.056

SL4f 0.014 0.019 0.028

MK4t 0.032 0.034 0.026

MN4 0.121 0.122 0.114

MS4 0.098 0.115 0.095

M4 0.345 0.378 0.311

S4 0.007 0.006 0.010

MN05t 0.022 0.022 0.017

2M05f 0.044 0.053 0.039

2MP6f 0.028 0.031 0.021

2MK5f 0.051 0.055 0.045

MNK5f 0.018 0.023 0.020

2MN6f 0.076 0.071 0.075

2ML6f 0.046 0.060 0.061

2MS„t 0.063 0.068 0.059

MSL6f 0.018 0.024 0.033

Se 0.004 0.000 0.005

M6 0.148 0.159 0.138

2(MN)„t 0.018 0.022 0.015

3MN8| 0.048 0.045 0.041

3ML8f 0.024 0.033 0.025

3MS8f 0.038 0.041 0.033

2MSL8f 0.012 0.020 0.021

M8 0.059 0.067 0.052

* Referred to 75°W (g).

t Not included in original 37 constituents.

r

- Phase lag*-

N

1946

1952

1957

210.2

195.9

205.2

281.9

267.4

282.9

332.4

59.3

48.5

175.5

164.6

174.5

29.8

31.9

31.5

33.9

21.8

22.4

47.1

44.8

45.3

52.8

58.2

55.7

52.2

70.0

56.4

68.7

74.9

227.2

83.9

79.8

78.2

29.2

191.4

237.4

73.9

75.4

60.1

316.6

326.7

266.1

143.4

337.5

8.2

77.0

100.8

77.8

145.4

125.9

131.0

194.7

170.3

37.9

124.7

120.9

125.6

97.1

95.8

95.7

343.5

348.4

337.3

26.6

51.8

26.2

17.5

16.4

6.2

336.0

334.3

330.9

33.6

26.3

25.3

350.3

346.1

346.1

359.4

328.8

73.5

349.0

342.0

336.7

355.3

342.0

346.7

28.1

26.6

7.9

7.8

0.5

358.1

49.3

313.9

326.0

195.5

188.4

179.2

205.1

209.0

189.0

243.4

239.0

229.5

235.7

272.2

245.7

324.7

297.0

33.6

206.4

205.0

196.4

108.4

124.2

118.1

131.7

129.0

120.4

153.5

130.2

117.1

187.1

174.3

171.9

183.7

195.9

178.4

146.3

146.8

140.9

1 1 2 Journal of Marine Research [25, 1

speed could make the harmonic constants acceptable. The sum of M2 and L2
varies from the sum of K2 and N2 by .ooo.°/hr (one cycle in about 4.5 years).
These speeds are too close together to be separated with only one year of data.
When the change in speed was made on the six constituents and the data were
reanalyzed, the constants for all became satisfactory, leaving only 3 of the
added 24 constituents unsatisfactory; all three are less than .03 foot. Further-
more, a new spectral analysis of the residuals showed a lower level of energy
in species 4, 6, and 8 — the species containing the six modified constituents —
thereby indicating an improvement in fitting the data by virtue of the changed
speeds.

In retrospect, it is now obvious why the combination of M2 and L2 should
have been given preference over N2 and K2. In equilibrium theory, the four
largest semidiurnal constituents are M2, S2, N2, and K2, in that order. L2
is only 22°/0 of K2. Hence, the choice of species- 2 constituents was limited
to the first four in trying various combinations to match the frequencies where
high resolution Fourier analysis showed large peaks. However, the analysis
for Philadelphia shows L2 to be roughly three times larger than K2. There-
fore, it is obvious that larger interactions should be expected from the sum
of M2 and L2 than from N2 and K2.

Essentially then, 3 of 24 new constituents are unsatisfactory compared with

16 of the original 37 — an amazing improvement considering that the 37

include the major constituents that are not subject to question. These results

raise questions as to whether some corrections in the speeds of the smaller

standard constituents may be indicated. At one time the manpower requirements

for successive analyses of a large number of years of data made such a study

virtually impossible. The present availability of both computer programs for

analysis and data in a format compatible with computers makes such a study

now only a modest effort ;

„, TTT . - . , plans for future tidal re-

1 able 111. Comparisons or observations and pre- , . t-cc a mi •

... r . / T , . _ , , r search in LbbA will in-

dict.ons for May, July, and October 1964; dude ^ sfud

Anchora§e- Tables III and IV
No. of constituents in predictions 37 114 show comparisons of ob-
Time differences (pred.-obs.) servations and predictions
High water (hours) -0.13 -0.03 for Anchorage and Phi-
Low water (hours) -0.35 -0.19 fcdelphia, respectively. If

Ratio, obs./pred. the time differences were

Mean ranSe L05 101 smaller and if the ratio

Tide level minus sea level of observed range to pre-

Observed (feet) -1.09 -1.09 ,. , 1

Predicted (feet) - .70 -1.00 dlCted ran§e Were doSel"

r. j 1 • /1 00 a ,nc« to 1. 00, the fit would be

Residual variance (192 days, 1964) ' .

in feet2 9767 .2954 better- Tlde level mlnus

% of original variance 1.33 0.40 sea level is a measure of

1967] Zetler and Cummings : Predicting Shallow-water Tides 113

the distortion of the Table IV. Comparisons of observations and pre-
curve from a pure cosine dictions for January, March, July, and October

curve, primarily due to 1 957 ; Philadelphia.

the contributions of com- _, . .... „_ _,

No. or constituents in predictions 37 61

pound constituents. An

. r 111 Time differences (pred.-obs.)

optimum ht would show High water (hours) -0.18 -0.14

identical values for ob- Low water (hours) -0.33 -0.28

served and predicted val- Ratio> obs./pred.

ues. The residual var- Mean range 1.06 1.04

iance, which should be as Tide level minus sea level

small as possible, is sig- Observed (feet) -0.23 -0.23

nificant only as a com- Predicted (feet) -0.12 -0.14

parative number. The Residual variance (355 days, 1957)

absolute numbers depend in feet* 3467 -3299

., . °/0 of original variance 8.49 8.08

primarily on the range

of tide and the energy in

the meteorological continuum. The improvement in practical predictions by

using more constituents is clearly shown in the results presented in Tables

III and IV.

The use of "harmonic" in the title warrants some explanation. Even though

annual or seasonal node corrections (amplitude factors and phase corrections)

are applied to lunar constituents by all organizations preparing tidal tables,

the predictions are ordinarily regarded as harmonic if subsequent corrections

are not added to the derived times, the heights of high and low waters, or both.

It is used in the title of this paper in this sense to differentiate the procedure

from Doodson's shallow-water technique (4). Doodson (2) developed a purely

harmonic set of constituents (calling the usual method, "quasi-harmonic"),

but he never used the method for practical tidal predictions.

REFERENCES

1. Bullard, E. C, F. E. Oglebay, W. H. Munk, and G. R. Miller

1964. A user's guide to BOMM, a system of programs for the analysis of time series.
Inst, of Geophysics and Planetary Physics, U. of Calif., La Jolla. 108 pp.

2. Doodson, A. T.

1921. The harmonic development of the tide-generating potential. Proc. R. Soc, (A) 100:
305-329.
3.1928. The analysis of tidal observations. Phil. Trans., A 227: 223-279.

4. 1957. The analysis and prediction of tides in shallow water. Int. Hydrogr. Rev. 34(1) : 5-46.

5. Doodson, A. T. and H. D. Warburg

1941. Admiralty manual of tides. Hydrographic Department, Admiralty, London. 270 pp.

6. Groves, G. W., and B. D. Zetler

1964. The cross spectrum of sea level at San Francisco and Honolulu. J. Mar. Res., 22 (3):
269-275.

114 Journal of Marine Research [25, 1

7. Munk, W. H., B. D. Zetler, and G. W. Groves

1965. Tidal cusps. Geophys. J. R. Astr. Soc, 10: 211-219.

8. Rauschelbach, H.

1924. Harmonische Analyse der Gezeiter des Meeres. Arch, dtsch. Seewarte, Hamburg,
42 (r): 114 pp.

9. SCHUREMAN, P.

1941. Manual of harmonic analysis and prediction of tides. Spec. Publ. H. S. est. geodet.
Surv., g8 : 317 pp.
ic. Zetler, B. D., and G. W. Lennon

Evaluation tests of tidal analytical processes. UNESCO Publication on Tide Sym-
posium at Paris. In press.

Printed in Denmark for the Sears Foundation for Marine Research,

Yale University, New Haven, Connecticut, U.S.A.

Bianco Lunos Bogtrykkeri A/S, Copenhagen, Denmark

39

Reprinted from DEEP SEA PHOTOGRAPHY, The Johns Hopkins Univ. Press

J. Photogrammetry applied to photography at the
bottom

M. D. Schuldt, C. E. Cook, and B. W. Hale U.S. Coast and Geodetic Survey, Washington, DC.

Abstract

Photogrammetric techniques used in stereographic aerial
surveys are applied to deep-sea stereophotography, per-
mitting the photoanalyst to describe the subject quanti-
tatively. The technique and conditional requirements are
discussed and a few contoured and cross-section examples
are shown. Internal accuracy is approximately one part
in 2,000.

5-1. Introduction

It is the purpose of this chapter to describe photogram-
metric measurements of the relief and the size of objects
appearing in deep-sea photographs. This technique
requires stereographic pairs of photographs and is
basically similar to that of aerial photogrammetry. Its
use eliminates the uncertainties arising from comparison
with an object of known dimensions appearing in the
field of view and/or reliance on shadow effects. The
application of this method does, however, impose a few
conditions and requires some specialized equipment.

5-2. Conditions

Varying degrees of success in photogrammetric contour-
ing can be achieved with varying picture quality, but for
best results the image must be sharp and evenly lighted,
with good contrast. These conditions are basically a
function of equipment, system configuration, and film;
however, deficiencies in lighting and contrast can, within
reason, be improved upon in the photo laboratory. For
deep-sea work, the conditions of stereophotography must
be met by a synchronized system of two cameras and a
light source suitably mounted on a supporting vehicle.
This procedure cannot be replaced by the forming of a
stereo pair from two successive photographs taken by the
same camera, since the motion of the camera cannot be
controlled as precisely as is done in aerial photographic
mapping.

The following condition is imposed on the ratio of
base-line length to lens-subject distance. Let b be the
distance between the lens centers of the two cameras, and
h be the distance from the subject to the lens plane. Then,
with all values expressed in the same units,

fel < — </C2

~ h ~

(1)

where k\ is the lower limit imposed by the stereoscopic
photograph measuring instrument and kj the upper limit
imposed by the angular field of the camera lens; /; is
limited by the transparency of the water and its actual
value can be derived from the system geometry and the
amount of overlap in the stereopair; b is preset — assum-
ing a value for h — so that the ratio is within the pre-
scribed limits. The value of h then becomes the working
distance of the cameras from the subject. (Typical values

are b/h =

1

0.25 and k2 = 0.4.)

5-3. Procedure

To enable the dimensioning of any subject appearing in
underwater stereophotographs, each frame of a stereo-
pair is projected and viewed so that points common to
each frame are superimposed, and the left and right
frames are seen with the left and right eye respectively.
There are four ways of doing this, one of which is illus-
trated in fig. 5-1 and will be discussed; the others will
be mentioned briefly.

The dichromatic method

The method illustrated in fig. 5-1 requires that each
frame of a stereopair, whether in color or black and white,
be printed on black-and-white, glass dipositive plates for
use in a dichromatic multiplexing projection stereoplotter.
One frame is then projected in red, the other in blue-
green, and, by wearing spectacles with matching red and
blue-green lenses, the operator can view the image in three

69

70

JOHNS HOPKINS OCEANOGRAPH1C STUDIES NO. 3 1967

Figure 5-1. Principle of the Multiplexing Stereoplotter.
(Photo courtesy of the U.S. Geological Survey.)

dimensions. This projection-viewing system permits,
through filtering, the viewing of one frame with one eye
and the other frame with the other eye. A table serves as
the x-y plane and is perpendicular to the axis of pro-
jection, which is the z axis. The plane of focus is parallel
to and above this table. The scales of both the x-y plane
and the z axis are functions of the ratio b/h. However,
the z scale is also a function of the focal lengths of the
photographing and projecting cameras, and hence the
two scales can be, but are not necessarily, the same. Fig.
5-1 shows a small circular projection screen mounted on
a suitable fixture and aligned parallel to the image plane.
The screen's center serves as the reference point for either
tracking the image contours or measuring elevations.
Directly below this point, and in contact with the table,
is a pencil which traces the horizontal path of the reference
point. The fixture can be moved freely in the x-y plane
and the screen can be moved vertically, through gearing,
by a graduated dial. The dial readings are convertible to
true height differences from an arbitrary zero reference
plane. Since the vertical scale can be adjusted to fit the
vertical range of the screen, the reference point can be
made to coincide with any desired point in the image. By
moving the reference point along the curves of coinci-
dence, those image elevations are sketched, and thus
the entire relief can be contoured at any desired interval.

H ^immm^&^jjm&iix mam

Figure 5-2a. Pair of stereoscopic photographs, showing sand ripples. Depth, 18.3 m. Location, 35°26.7'N 75°24.0'W. July, 1963.

DEEP-SEA PHOTOGRAPHY

71

Figure 5-2b. Contours drawn on one of the stereopair in fig.
5-2a. Contour interval, 15 mm.

Also, by moving the reference point along any horizontal
line and adjusting it vertically — noting dial readings
and x-y coordinates — any desired cross section of the
image can be drawn. Fig. 5-2 shows contour map and
section profiles constructed from the respective stereo-
scopic pairs using the above method. Fig. 5-3a shows a
stereoscopic pair from which the contours of fig. 5~3b
were constructed.

The size of the photographs and illustrations was
altered to fit these pages. Actually, contouring is done on
a much larger scale than shown here; somewhere in the
neighborhood of one square meter. The contour intervals
and cross sections shown were chosen for no other reason
than to demonstrate the technique. Intervals could just
as well have been larger, smaller, or irregular depending
on the purpose. Likewise, the cross sections could have
been located anywhere and oriented in any direction.
Furthermore, any object, regardless of its orientation
with respect to the coordinate system, can be dimensioned
provided that it has sufficient definition in the projected
image.

The absolute accuracy of measurement of the instru-
mentation used for preparing the accompanying drawings
is better than one part in 2,000. Level datum or coordi-
nate axis orientation is dependent on the assumption that
the plane of the photograph is perpendicular to the
gravity vector. Level data could be incorporated into the

Figure 5-2c. Enlarged contour map of the sand ripples in fig. 5-2a. Contour interval, 15 mm

72

JOHNS HOPKINS OCEANOGRAPHIC STUDIES NO. 3 1967

Figure 5-2d. Cross section along the line A-B in fig. 5-2c.

system to verify this assumption or provide the necessary
data to compute the angular correction.

Other methods

The projection-viewing method discussed above is
referred to as the dichromatic method; there are also
the flicker, optical, and polarized-light methods. The
dichromatic method requires black-and-white transparen-
cies whereas either color or black-and-white positive
transparencies may be used in the other three. Glass
transparencies are preferred because of their stability,
but ordinary film may be used.

The flicker method is a synchronized projection-viewing
system which utilizes the image retention capabilities of
the eyes. Alternately, the left frame is seen with the left
eye and the right frame with the right eye, and the images
alternate or flicker at such a rate that both frames appear
to be viewed simultaneously, thus giving depth to the
image. In the optical method the frames are viewed
through an optical system to achieve image depth. The
polarized-light method projects one frame with light
polarized in one direction and the other frame with light
polarized in a direction at right angles to the first. Then,
when viewed with spectacles fitted with appropriate
lenses, the image is seen to have depth. Contouring and

profiling with these methods is analogous to the pro-
cedure described above, but with modification according
to the particular method.

5-4. Applications

Although any sampling system has limitations, the tech-
niques described can be applied to the measurement of:
ripple characteristics for interpretation of the water
movements they reflect; bottom roughness for sonar
studies; grading of sediments into fine, medium, and
coarse; and the linear extent of such grades. Geological
features can be measured, such as width and height of
outcrops, size of fracture patterns, inclusions in con-
glomerates, size of minor structures like nodules for
mineral-resource evaluation studies, pillow lava, etc.
The dimensioning of organisms for quantitative descrip-
tion and taxonomic studies, and the measurement of
burrows, mounds, tracks, and trails for biological studies
are also subject to this technique. A single stereopair or
a mosaic with stereo between adjacent pairs can be
analyzed. The mosaic allows examination of large fea-
tures not definable in a single pair. Further refinement
of the system and development of operating techniques
will extend the applications.

Figure 5-3a. Pair of stereoscopic photographs, showing a rock. Depth 2,750 m. Location, 38°15.0'N 71°21.0'W. October, 1961.

DEEP-SEA PHOTOGRAPHY

73

5-5. Summary

A method has been described to utilize stereophotography
beyond the usual visual sensation of three dimensions.
Within its limitations the system permits the scientist to
measure what he sees without being there himself or tak-
ing samples. The method elevates deep-sea stereophotog-
raphy to the status of a sampling device and can be used
alone or correlated with other sampling means, permitting
the extension of "point" information to "area" informa-
tion. This technique allows an accurate quantitative
description of objects and features that otherwise could
only be described in nondimensional terms.

To quote Lord Kelvin (1824-1907): "I often say that
when you can measure what you are speaking about, and
express it in numbers, you know something about it; but
when you cannot express it in numbers, your knowledge
is of a meagre and unsatisfactory kind; it may be the
beginning of knowledge, but you have scarcely, in your
thoughts, advanced to the stage of science, whatever the
matter may be."

Figure 5-lb. Contours drawn on one of the stereopair in
fig. 5-3«. Contour interval, 10 mm.

40

Reprinted from 0CEAN0L0GY YEARBOOK, 1968, Oceanology International

SEA-FLOOR GEOLOGY

hi/ Dr. Harris B. Stewart jr., director.
Institute for Oceanography, Environ-
mental Science Services Administration

If the world ocean is considered as a
big flat pan of water, physical oceanog-
raphy is concerned with the water and
marine geology with the pan. Studies of
this "pan" include the origin of its major
features— trenches, mid-ocean ridges, sea-
mounts, and submarine canyons.

Included also are the geophysical
studies of the earth beneath the sea,
the big-picture research leading to more
complete knowledge of the origin of the
"pan" itself and of its edges— the con-
tinental margins.

Even the most minute characteristics
of the sediments covering the bottom
and sides of the ocean basin come un-
der the scrutiny of the marine geologist.
His tools run the gamut from expensive
research ships and drilling rigs to scuba
gear and a geology hammer; from ship-
board gravity meters, towed magnetom-
eters, and deep-sea dredging equipment
to an underwater Brunton compass, a
bottle of hydrochloric acid, and a pair
of dividers. His most effective tools, how-
ever, are an insatiable curiosity and a
basic and sound background in geologv.

During the last year or two, the field
of marine geology has seen some bril-
liant thinking leading toward the syn-
thesis of exciting new hypotheses. We
also have witnessed the development
of highly complex but workable equip-
ment that gives the marine geologist the
means to get the data necessary to put
these new hypotheses to the test and
to develop what primary school teachers
term a "basic understanding."

Among the new hypotheses is the con-
cept of a spreading sea floor and the
use of the history of reversals in the
earth's magnetic field as a means of
verification. The long-in-disrepute Wege-
ner hypothesis of continental drift has
in the last few years enjoyed a renewed
popularity based on new data and en-
lightened thinking. Men like Tuzo Wil-
son, Robert Dietz, and Fred Vine have
proposed controversial ideas, and more
marine geologists have added more think-
ing to their traditional role of describing.

The death of Project Mohole will be
ascribed, when its history is written, pri-
marily to the inability of marine geol-
ogists to agree among themselves— a
healthy state— and their lack of under-
standing of the economic and political
factors involved in major scientific projects.

However, some significant byproducts
resulted from Mohole— mainly the Guad-
alupe test drilling results, and the forma-
tion of the Joint Oceanographic Institu-
tions Deep Earth Sampling project.

New equipment also has given impe-
tus to the field of marine geology. Most
important here is the satellite naviga-
tion system, bv which research and sur-
vey ships can determine position with
remarkable accuracy.

When supplemented by other long-
range navigation methods, the system
will make it possible to obtain highly-
accurate, tight-grid, detailed survevs of
such features as the mid-ocean ridge-
and-rift systems.

Seismic reflection profilers have pro-
vided the third dimension to geological
investigations at sea, and the next few
years should see improved models in-
stalled as standard equipment on all
research and survey vessels. Similarly,
the boom in manned submersibles for
marine geological research will produce
new knowledge as more geologists are
able to observe the sea floor and its
processes at close range.

The report, "Effective Use of the
Sea," by the President's Science Advis-
ors Committee's Panel on Oceanogra-
phy (PSAC-POO), downgraded the sys-
tematic survey as a research technique.
This, in effect, has halted for the next
year or so the highly productive surveys
of the Pacific sea floor.

Such results can come only from the
accurately controlled systematic surveys
that government vessels do so well.

Hopefully, the systematic survey will be
reinstated as one of the most valuable
techniques available to the marine geol-
ogist who is interested in learning of the
origin of the ocean basins.

The push for the economic recovery
of sea-floor minerals and the need for
data on the mass physical properties
of marine sediments should result in
increased work in these areas over the
next few years.

We can be sure that new tools and
the increasing ratio of tKnking to de-
scribing in marine geology should result
in important contributions to our knowl-
edge of "the pan" that holds the waters
of the ocean. □

This article is offered to members
of the ESSA family in the hope that
it will make their summer vaca-
tions safer and happier.

Reprinted from ESSA WORLD July, 1967

This summer, make
sure you outwit the

KILLER AT THE
SEASHORE

I his summer your child could drown
— needlessly — at the seashore.

He may be an able swimmer. He may
be in very shallow water perhaps only up
to his chin. It is a warm sunny day, with
a good surf running. Lots of other chil-
dren are playing in the shallows, and
tragedy seems far removed.

But suddenly, and quite unexpectedly,
he may feel the bottom moving fast
beneath his feet and realize he is being
swept out to sea. Knowing that he can
swim well, he may strike out hard against
the current for shore. But after a few
minutes, he will find that he is not mak-
ing any headway, that the water around
him is over his head, that he is almost
out to the surf zone.

He may call for help, but no one will
hear him above the surf's roar. Panic
will take over, along with exhaustion, and
maybe a bad cramp. And in another
moment, he will be dead by drowning.

Every year, many persons unfamiliar

BY HARRIS B. STEWART, JR.

Acting Director, ESSA Institute

for Oceanography

with rip currents lose their lives un-
necessarily.

They may be excellent swimmers, but
they may not know what to do when
caught in a rip current. And they will
die, as victims of killer currents, in fact,
but as victims of the exhaustion and
panic that would never have occurred
if more swimmers know how to recognize
a rip current and how relatively easy it
is to swim out of one.

The Institute for Oceanography has
gathered and facts about rip currents
that endanger the lives of swimmers
and developed some simple rules of cop-
ing with them. It would be useful, and
even life-saving, for
every swimmer to
know them as he
prepares to take a
summer vacation at
the beach.
I What is a rip cur-
JJ rent? Technically
I speaking, it is a
strong narrow cur-
rent flowing out to
£__ sea perpendicular to

30

the shore and carrying back to sea the
water brought in by waves and longshore
currents. It is part of a generally-circular
pattern of water movement found off
most long, gently-sloping sand beaches.
It can travel at speeds up to two or even
three miles an hour, and change its posi-
tion from day to day and even during
the same day. The same beach may have
several rip currents operating at one
time, and then go weeks with none at all.
Once outside the surf zone, the rip
current dies rapidly, spreads out, and
often forms a big sluggish eddy which
oceanographers call a "rip head".
How do you recognize a rip current?
Rip currents are usually easy to see once
you know what to look for. In general,
the pattern of the sea surface between
the beach and the area where the waves
are breaking offshore, is one of long lines
that run parallel to the beach. A rip
current makes a break in this pattern by
providing a cross-pattern line running
perpendicular to the beach. Sometimes
small choppy waves form a line out to
the surf zone indicating a rip current.
Often, a foam line will show where it is.

Photo courtesy American Airlines

At other times, when there is suspended
sediment in the water, a rip current may
be marked by a long brownish band of
darker water.

Usually the surf is lower where a rip
current passes through the surf zone,
and there will be a break in the line of
breakers. Or if a rip current has been
stabilized in one place for a while, there
is often a short sand spit built out from
the beach at the base of the rip.

If, as you are swimming, you notice
that you tend to move faster in one direc-
tion along the shore, there are probably
strong longshore currents, and you should
expect rip currents to be developing. Or
if you are walking from the beach into
shallow water, and you feel a longshore
current pulling at your legs, you may be
able to see a spot down-current where a
rip is moving water seaward. Or look at
the outer end of any jettv. groin, or other
solid obstruction to the longshore move-
ment of water, and there likely will be
a rip current where the water has been
deflected seaward.

How do you get out of a rip current?
Fortunately, you will know when you
are in one. Your first indication, if your
feet touch bottom occasionally, will be a
feeling that the bottom is moving fast
toward shore. It's a strange feeling, for
you have no nearby frame of reference
to show that you yourself are moving,
and you feel as though the sand beneath
your feet is moving. When your feet
aren't occasionally touching bottom, you
will soon notice that you are much fur-
ther out to sea than you expected to be.
or moving out faster than other swim-
mers near you. or that the area where
the waves are breaking seems to be ap-
proaching fast.

This is the point where most swimmers
who lose their lives start swimming their
hardest toward the beach and where they
make a fatal mistake. Since the rip cur-
rent is seldom more than ten or twenty
feet wide, swimmers should swim parallel
to tlie beach, and they can very soon be
out of it.

An alternative is to relax and let the
current carry you seaward through the
surf zone and into the rip head where
the current slows down, and from where
you can then have a leisurely swim back
to the beach on a course parallel to the
rip current. Surfers along the California
coast actually search out rip currents and
ride them on their boards back out
through the surf zone to the place where
the big ones are humping up.

The rules are simple. Knowing them
may save your life: (1) Learn to recog-
nize a rip current: (2) Look for them
every time you go to the beach: (3)
Point them out to the children and tell
them about them: (4) Avoid them if
possible: but (5) If you do get caught in
one. swim parallel to the beach, and you
will soon be out of it. □

•&Z£r£

*£

Photo courtesy Mia

-Metro News Bureau

31

Reprinted from THE MIAMIAN, September 196?

42

ea
cience

hips

Dr. Stewart is the director of the ESSA
(Environmental Sciences Services Ad-
ministration) east coast oceanography
laboratory which will soon be built on
Virginia Key. Berthing facilities for
ESSA's ships and other necessary build-
ings will be constructed at the new Port
of Miami on Dodge Island. Dr. Stewart
tells of their plans.

The MIAMIAN • September, 1967

By Dr. Harris B. Stewart, Jr.

Even as the New Port of Miami on Dodge Island it-
self is taking shape, so too are the plans for its use as a
major staging area for oceanographic research ships. In
mid-August Admiral John Bull and Captain Allen
Powell of ESSA's Coast and Geodetic Survey came to
Miami to work out construction schedules for the Dodge
Island facility with Port Director Irvin Stephens. Their
plans are the result of several months of discussions with
officials of the Institute of Marine Science of the Uni-
versity of Miami and of the Bureau of Commercial Fish-
eries' Tropical Atlantic Biological Laboratory. It is
hoped that the oceanographic ship facility on Dodge Is-
land will meet the combined requirements of these three
major oceanographic research organizations.

The details of the plans are not yet sufficiently worked
out to warrant publication. However, the university, the
federal agencies and the Port of Miami are working
together to come up with a facility that will not only
meet the requirements for the ships to base there but
will also fit in architecturally with the other structures
already built or planned for the new Port of Miami.

The Coast and Geodetic Survey ship DISCOVERER,
which visited Miami the end of July, officially trans-
ferred from Jacksonville, where she was constructed, to
Miami on August 28th. Until the oceanographic ship
facility on the south side of Dodge Island is completed,
she will berth on the north side of the island. This ship is
the major research ship of the ESSA Institute for Ocea-
nography which last winter chose Virginia Key as the
site for its East Coast Laboratory. Located temporarily
at 901 South Miami Avenue, scientists of the Institute for
Oceanography will start using the DISCOVERER this

The MIAMIAN • September, 1967

fall for their research work at sea. In a bit over a year,
the DISCOVERER will be joined by a second ESSA
ship, the RESEARCHER, now being constructed at
Lorraine, Ohio. This newer ship will be a slightly scaled
down version of the DISCO, as we call her, and will also
have the Port of Miami on Dodge Island as her home
port. From time to time, other ESSA ships will be based
out of the Port of Miami as a logistic support port while
they are doing research or survey work in Florida waters.

The UNDAUNTED of the Tropical Atlantic Bio-
logical Laboratory on Virginia Key will also base at
Dodge Island. Hopefully, too, the ships PILLSBURY
and GERDA of the Institute of Marine Science of U.M.
will be part of the Port of Miami Oceanography fleet.
These three marine research activities, ESSA, TABL,
and IMS, together with the Miami Seaquarium form the
nucleus of the Virginia Key Marine Science Complex.
The Seaquarium's vessels are berthed there at the Vir-
ginia Key site, but there is insufficient depth of water for
the larger research ships.

The Port of Miami facility has the water for these
larger ships. When the south side of the island is bulk-
headed and the power, fresh water, and other services
are in, the oceanographic ships will have the place to tie
up they so badly need. By then the shoreside construc-
tion on Dodge Island will be completed, and Miami will
have the finest oceanographic ship facility on the east-
ern seaboard.

Dodge Island is gradually rising from the floor of
Biscayne Bay. The ships are here or under construction.
Plans for the buildings, bulkheads, and support services
are well along. The next few years will be exciting ones
for Dodge Island, for Miami, and for oceanography.

27

GPO 849 - 483

PENN STATE UNIVERSITY LIBRARIES