COC :/7&/v. I *X ^Lr ^e^ *fe ri «fl ^ ^ Col lected Reprints 1976 Atlantic Oceanographic and Meteorological Laboratories Volume I April 1977 U.S. DEPARTMENT OF COMMERCE National Oceanic and Atmospheric Administration Q. O 0 Collected Reprints 1976 Atlantic Oceanographic and Meteorological Laboratories Miami, Florida 33149 Volume I April 1977 Boulder, Colorado U.S. DEPARTMENT OF COMMERCE Juanita M. Kreps, Secretary g National Oceanic and Atmospheric Administration v Robert M. White, Administrator Environmental Research Laboratories % Wilmot Hess, Director FORWARD This is the tenth consecutive year in which the Collected Reprints of NOAA's Atlantic Oceanographic and Meteorological Laboratories have been published for distribution to scientists, institutions, and libraries here and abroad. This series provides a single reference source for articles by AOML personnel which have appeared in numerous scientific journals and various internal scientific and technical publi- cations . The Atlantic Oceanographic and Meteorological Laboratories conduct research programs in the areas of physical, chemical, and geological oceanography and air-sea interaction. The 1976 edition presents the papers published in that year. They are arranged in alphabetical order by first author within each of five groups: Office of the Director Physical Oceanography Laboratory Marine Geology and Geophysics Laboratory Sea Air Interaction Laboratory Ocean Chemistry Laboratory It is hoped that those recipients with whom we do not already have an exchange arrangement would add the AOML Library to the distribution list for any relevant publications from their institution. Harris B. Stewart, Jr. Atlantic Oceanographic and Director, AOML Meteorological Laboratories NOAA/Environmental Research Laboratories 15 Rickenbacker Causeway Virginia Key Miami, Florida 33149 i n Digitized by the Internet Archive in 2012 with funding from LYRASIS Members and Sloan Foundation http://archive.org/details/collectedreprin1976v1atla CONTENTS VOLUME I General 1. Apel, J. P. Page Ocean Science from Space. EOS, Vol. 57, No. 9, 612-624. 1 2. Sawyer, C. B. High-Speed Streams and Sector Boundaries. Journal of Geophysical Research, Vol. 81, No. 13, 2437-2441. 14 3. Sawyer, C. B. and M. Haurwitz. Geomagnetic Activity at the Passage of High-Speed Stream in the Solar Wind. Journal of Geophysical Research, Vol. 81, No. 13, °435-2436. 19 4. Stewart, H. B. , Jr. Introduction. Proc. of CICAR-I I Symposium: Progress in Marine Research in the Caribbean and Adjacent Regions, Caracas, Venezuela, July 12-16, 1976, p. 126. 21 5. Stewart, H. B. , Jr. Introduction to the CICAR-I I Symposium. Proc. of CICAR-I I Symposium: Progress in Marine Research in the Caribbean and Adjacent Regions, Caracas, Venezuela, July 12-16, 1976, p. 241. 22 6. Stewart, H. B. , Jr. Preliminary Bibliography of Published Results of Marine Research by U.S. Scientists in the CICAR Area, 1968- 1975: Introduction. U.S. Department of Commerce, NOAA/ERL/AOML-National Oceanographic Data Center Publication, Washington, D.C., 50 p.* 23 7. Stewart, H. B. , Jr. Where the Sea and Man Meet: The Coastal Zone. Museum, Vol. 7, No. 11, 19-25, 44-48. 24 * Introduction only. PHYSICAL OCEANOGRAPHY LABORATORY Page 8. Beardsley, R. C, W. C. Boicourt, and D. V. Hansen. Physical Oceanography of the Middle Atlantic Bight. Middle Atlantic Continental Shelf and the New York Bight. ASLO Special Symposia, Volume 2, 20-34. 34 9. Charnell, R. L., M. E. Darnell, G. A. Berberian, B. L. Kolitz, and J. B. Hazel worth. New York Bight Project, Water Column Characterization Cruises 1 and 2 of the NOAA Ship Researcher, 4-15 March, 5-14 May 1974. NOAA Data Report ERL MESA-18 , 220 p.* 49 10. Festa, J. F. , and D. V. Hansen. A Two-Dimensional Numerical Model of Estuarine Cir- culation: The Effects of Altering Depth and River Discharge. Estuarine and Coastal Marine Science , Vol. 4, 309-323. 50 11. Gordon, H. R. Radiative Transfer: A Technique for Simulating the Ocean in Satellite Remote Sensing Calculations. Applied Optics , Vol. 15, No. 8, 1974-1979. 65 12. Hansen, D. V. A Lagrangian Buoy Experiment in the Sargasso Sea. Proc. AIAA Drift Symposium, Hampton, Va., May 22-23, 1974, NASA CP-2003, 175-192. 71 13. Herman, A. Automated Contouring of Vertical Oceanographic Sections Using an Objective Analysis. Proc. of the Third Annual Conference on Computer Graphics, Interactive Techniques, and Image Processing, University of Pennsylvania, Computer Graphics 10, No. 2, 218-223. 89 * Abstract only; Complete text available on microfiche. VI Page 14. Herman, A. and A. C. Campbell. An Automated Solution for Omega Navigation. Proc. of the Fourteenth Annual Southeast Regional ACM Conference, University of Alabama, Birmingham, Alabama, 305-308. 95 15. Leetmaa, A. Some Simple Mechanism for Steady Shelf Circulation. Marine Sediment Transport and Environmental Management , D. J. Stanley and D. J. P. Swift, editors, John Wiley and Son, Inc., Chapter 3, 23-28. 99 16. Leetmaa, A. and M. Cestari. A Comparison of Satellite-Observed Sea Surface Temperatures with Ground Truth in the Indian Ocean. NOAA Technical Report ERL Z76-AOML 22, 10 p. 105 17. Maul, G. A. The Study of Ocean Circulation from Space. Proc. of the Thirteenth Space Congress: Technology for the New Horizon, 3-27—3-36. 118 18. Maul, G. A., H. R. Gordon, S. R. Baig, M. McCaslin, and R. DeVivo. An Experiment to Evaluate SKYLAB Earth Resources Sensors for Detection of the Gulf Stream. NOAA Technical Report ERL 378-AOML 23, 69 p. 128 19. Mofjeld, H. 0. Tidal Currents. Marine Sediment Transport and Environmental Management , D. J. Stanley and D. J. P. Swift, editors, John Wiley and Sons, Inc., Chapter 5, 53-64. 201 20. Molinari , R. L. The Formation of the Yucatan Current Based on Observations of Summer 1971. Journal of Physical Oceanography, Vol. 6, No. 4, 596-602. 213 vn 21. Molinari, R. and A. D. Kirwan. Page Calculations of Differential Kinematic Properties from Lagrangian Observations. Proc. AIAA Drift Buoy Symposium, Hampton, Va., May 22-23, 1974, NASA CP-2003, 193-209. 220 22. Starr, R. B., G. A. Berberian, and M. A. Weiselberg. MESA New York Bight Project, Expanded Water Column Characterization Cruise XWCC-1 of the R/V ADVANCE II. NOAA Data Report ERL MESA- 22, 43 p.* 237 23. Voorhis, A. D., E. H. Schroeder, and A. Leetmaa. The Influence of Deep Mesoscale Eddies on Sea Surface Temperature in the North Atlantic Sub- tropical Convergence. Journal of Physical Oceanography, Vol. 6, No. 6, 953-961. 238 MARINE GEOLOGY AND GEOPHYSICS LABORATORY 24. Bennett, R. H., W. R. Bryant, W. A. Dunlap, and G. H. Keller. Initial Results and Progress of the Mississippi Delta Sediment Pore Water Pressure Experiment. Marine Geotechnology , Vol. 1, No. 4, 327-335. 247 25. Dash, B. P., M. M. Ball, G. A. King, L. W. Butler, and P. A. Rona. Geophysical Investigation of the Cape Verde Archipelago. Journal of Geophysical Research, Vol. 81, No. 29, 5249-5259. 256 26. Dietz, R. S. Iceland: Where the Mid-Ocean Ridge Bares Its Back. Sea Frontiers , Vol. 22, No. 1, 9-15. 267 27. Dietz, R. S. and K. 0. Emery. Early Days of Marine Geology. Oceanus , Vol. 19, No. 4, 19-22. 274 * Abstract only; complete text available on microfiche. viii Page 28. Dietz, R. S. and J. F. McHone. El'gygtgyn: Probably World's Largest Meteorite Crater. Geology, Vol. 4, No. 7, 391-392. 278 29. Freeland, G. L. and G. F. Merrill. Deposition and Erosion in the Dredge Spoil and Other New York Bight Dumping Areas. Proc. American Society of Civil Engineers Specialty Conference on Dredging and Its Environmental Effects, Mobile, Al . , 26-28 January 1976, 936-946. 280 30. Freeland, G. L., D. J. P. Swift, W. L. Stubblefield, and A. E. Cok. Surficial Sediments of the NOAA-MESA Study Areas in the New York Bight. Middle Atlantic Shelf and the New York Bight, ASLO Special Symposia, Volume 2, 90-101. 291 31. Lavelle, J. W. , P. E. Gadd, G. C. Han, D. A. Mayer, W. L. Stubblefield, D. J. P. Swift, R. L. Charnell, H. R. Brashear, F. N. Case, K. W. Haff, and C. W. Kunselman. Preliminary Results of Coincident Current Meter and Sediment Transport Observations for Wintertime Conditions on the Long Island Inner Shelf. Geo- physical Research Letters , Vol. 3, No. 2, 97-100. 303 32. Lowell, R. P. and P. A. Rona. On the Interpretation of Near Bottom Water Temperature Anomalies. Earth and Planetary Science Letters, Vol. 32, No. 1, 18-24. 307 33. Nelsen, T. A. An Automated Rapid Sediment Analyser (ARSA). Sedimentology , Vol. 23, No. 6, 867-872. 314 34. Peter, G. and G. K. Westbrook. Tectonics of Southwestern North Atlantic and Barbados Ridge Complex. American Association of Petroleum Geologists Bulletin, Vol. 60, No. 7, 1078-1106. 320 IX 35. Richardson, E. and C. G. A. Harrison. Page Opening of the Red Sea With Two Poles of Rotation. Earth and Planetary Science Letters, Vol. 30, No. 1, 135-142. 349 36. Richardson, E. and C. G. A. Harrison. Reply: Opening of the Red Sea With Two Poles of Rotation. Earth and Planetary Science Letters , Vol. 30, No. 2, 173-175. 357 37. Rona, P. A. Asymmetric Fracture Zones and Sea-Floor Spreading. Earth and Planetary Science Letters, Vol. 30, No. 1, 109-116. 360 38. Rona, P. A. Book Review: Plate Tectonics and Oil. Earth Science Reviews, Vol. 12, No. 1, 74-75. 368 39. Rona, P. A. , Editor. Mid-Atlantic Ridge. Geological Society of America, Microform Publication, Vol. 5, 490 p.* 369 40. Rona, P. A. Pattern of Hydrothermal Mineral Deposition: Mid- Atlantic Ridge Crest at Latitude 26° N. Marine Geology, Vol. 21, No. 4, M59-M66. 371 41. Rona, P. A. Resource Research and Assessment of Marine Phosphorite and Hard Rock Minerals. Proc. of N0AA Marine Minerals Workshop, March 1976, 111-119. 379 * Abstract only; complete text on microform. 42. Rona, P. A. Page Salt Deposits of the Atlantic. Special Volume of 'Annals of the Brazilian Academy of Sciences. Anais Acad. Brasil Ciencies (Suplemento) , Vol. 48, 265-274. 388 43. Rona, P. A. and L. D. Neuman. Energy and Mineral Resources of the Pacific Region in Light of Plate Tectonics. Journal of Ocean Management, Vol. 3, 57-78. 398 44. Rona, P. A. and L. D. Neuman. Plate Tectonics and Mineral Resources of Circum- Pacific Region. Papers from Circum-Pacif ic Energy and Mineral Resources Conference, Honolulu, Hawaii, August 26-30, 1974, publ . by Amer. Assoc, of Petroleum Geologists, Memoir 25, 48-57. 420 45. Rona, P. A., R. N. Harbison, B. G. Bassinger, R. B. Scott, and A. J. Nalwalk. Tectonic Fabric and Hydrothermal Activity of Mid- Atlantic Ridge Crest (lat 26° N). Geological Society of America Bulletin, Vol. 87, 661-674. 430 46. Scott, R. B., J. Malpas, P. A. Rona and G. Udintsev. Duration of Hydrothermal Activity at an Oceanic Spreading Center, Mid-Atlantic Ridge (lat 26° N). Geology, Vol. 4, No. 4, 233-236. 444 47. Stubblefield, W. L. and D. J. P. Swift. Ridge Development as Revealed by Sub-Bottom Profiles on the Central New Jersey Shelf. Marine Geology, Vol. 20, No. 4, 315-334. 448 48. Swift, D. J. P. Coastal Sedimentation. Marine Sediment Transport and Environmental Management, D. J. Stanley and D. J. P. Swift, editors, John Wiley and Sons, Inc., Chapter 14, 255-310. 468 XI Page 49. Swift, D. J. P. Continental Shelf Sedimentation. Marine Sediment Transport and Environmental Management , D. J. Stanley and D. J. P. Swift, editors, John Wiley and Sons, Inc., Chapter 15, 311-350. 524 50. Swift, D. J. P. and J. C. Ludwick. Substrate Response to Hydraulic Process: Grain- Size Frequency Distributions and Bed Forms. Marine Sediment Transport and Environmental Management , D. J. Stanley and D. J. P. Swift, editors, John Wiley and Sons, Inc., Chapter 10, 159-196. 564 51. Swift, D. J. P., G. L. Freeland, P. E. Gadd, G. Han, J. W. Lavelle, and W. L. Stubblefield. Morphologic Evolution and Coastal Sand Transport, New York-New Jersey Shelf. Middle Atlantic Shelf and the New York Bight, ASLO Special Symposia, Volume 2, 69-89. 602 VOLUME II SEA-AIR INTERACTION LABORATORY 52. Apel , J. R., H. M. Byrne, J. R. Prom", R. Sellers. A Study of Oceanic Internal Waves Using Satellite Imagery and Ship Data. Remote Sensing of Environ- ment 5, No. 2, 125-135. Also appeared in Proc. Thirteenth Space Congress, Technology for the New Horizon, Cocoa Beach, Florida, April 7, 8, 9, 1976, 3-21--3-25. 623 53. Hanson, K. J. A New Estimate of Solar Irradiance at the Earth's Surface on Zonal and Global Scales. Journal of Geophysical Research, Vol. 81, No. 24, 4435-4443. 634 XII 54. Hasselmann, K. , D. B. Ross, P. Muller, and W. Sell Page A Parametric Wave Prediction Model. Journal of Physical Oceanography, Vol. 6, No. 2, 200-228. 643 55. McLeish, W. and S. M. Minton. STD Observations From the R/V COLUMBUS ISELIN During Phase III of GATE. NOAA Technical Re-port ERL 379-AOML 24. 101 p. 672 56. Newman, F. C. Temperature Steps in Lake Kivu: A Bottom Heated Saline Lake. Journal of Physical Oceanography , Vol. 6, No. 2, 157-163. 776 57. Proni, J. R., F. C. Newman, D. C. Rona, D. E. Drake, G. A. Berberian, C. A. Lauter, Jr., and R. L. Sellers. On the Use of Accoustics for Studying Suspended Oceanic Sediment and for Determining the Onset of the Shallow Thermocline. Peep-Sea Research , Vol. 23, No. 9, 831-837. 783 58. Proni, J. R. , F. C. Newman, R. L. Sellers, and C. Parker. Acoustic Tracking of Ocean-Dumped Sewage Sludge. Science, Vol. 193, 1005-1007. 794 59. Thacker, W. C. A Solvable Model of "Shear Dispersion." Journal of Physical Oceanography, Vol. 6, No. 1, 66-75. 797 60. Thacker, W. C. Spatial Growth of Gulf Stream Meanders. Geophysical Fluid Dynamics , Vol. 7, 271-295. 807 61. Webster, W. J., Jr., T. T. Wilheit, D. B. Ross, and P. Gloersen. Spectral Characteristics of the Microwave Emission From A Wind-Driven Foam-Covered Sea. Journal of Geophysical Research, Vol. 81, No. 18, 3095-3099. 832 Xlli OCEAN CHEMISTRY LABORATORY Page 62. Atwood, D. K. Regional Oceanography as it Relates to Present and Future Pollution Problems and Living Resources- Caribbean. IOC/FAO/UNEP International Workshop on Marine Pollution in the Caribbean and Adjacent Regions, Port of Spain, Trinidad, IOC/FAO/UNEP/ IWMPCAR/8, 40 p. 837 63. Gilio, J. L. and D. A. Segar. Biogeochemistry of Trace Elements in Card Sound, Florida Inventory and Annual Turnover. Proc. of the Sea Grant Symposium on Biscayne Bay, April 2-3, 1976, 17 p. 879 64. Hatcher, P. G. and L. E. Keister. Carbohydrates and Organic Carbon in New York Bight Sediments as Possible Indicators of Sewage Contamination. Middle Atlantic Continental Shelf and the New York Bight, ASLO Special Symposia, Volume 2, 240-248. 896 65. Hatcher, P. G. and D. A. Segar. Chemistry and Continental Margin Sedimentation. Marine Transport and Environmental Management , D. J. Stanley and D. J. P. Swift, editors, Chapter 19, 461-477. 905 66. Segar, D. A. and G. A. Berberian. Oxygen Depletion in the New York Bight Apex: Causes and Consequences. Middle Atlantic Continental Shelf and the New York Bight, ASLO Special Symposia, Volume 2, 220-239. 922 67. Segar, D. A. and A. Y. Cantillo. Some Considerations on Monitoring of Trace Metals in Estuaries and Oceans. Proc. of the International Conference on Environmental Sensing and Assessment, IEEE Annuals No. 75CH004-1, 6-5, 1-5. 942 xiv Page 68. Segar, D. A. and A. Y. Cantillo. Trace Metals in the New York Bight. Middle Atlantic Continental Shelf and the New York Bight, ASLO Special Symposia, Volume 2, 171-198. 947 69. Tosteson, T. R,.D. K. Atwood, and R. S. Tsai . Surface Active Organics in the Caribbean Sea. MTS-IEEE Oceans '76, 13C-1-13C-7. 975 xv Reprinted from: EOS, Vol. 57, No. 9, 612-624. Ocean Science From Space John R. Apel Introduction The ocean plays as fundamental a role in the natural scheme of things as does the atmosphere, although its functions, being considerably more varied and diffuse, are probably neither as well appreciated nor as well understood. The sea profoundly affects the weather and climate and in turn is affected by the atmo- sphere, acting as both a heat reser- voir for storing, distributing, and releasing solar energy and as the source for most atmospheric moisture. It interacts with the bounding land and air over times ranging from minutes to millennia. Geological activity on all time and space scales- takes place in and under the seas, which serve as the repository for the detritus of man and nature and, just as important, as practicable sources of petroleum and a few useful minerals. Its cur- rents and dilutant powers are called upon to disperse sewage, poisonous and nonpoisonous wastes, solid trash, and excess heat, while it maintains a role as the aqua viva for an extremely complicated and 612 commercially important food chain and a role as a means of recreation and refreshment for people. In the estuaries and the coastal zones these conflicting demands are especially severe. This article attempts a rather limited review of the types of oceanic information that current experimental results and planning indicate should be available from spacecraft in the near future. To the author's knowledge, plans exist to orbit sensors that will yield measurements or observations of all of the parameters discussed here, albeit often only on an experimental basis. Questions on the usefulness of satellites for ocean science have been raised by oceanographers since the first space-derived imagery was returned to earth. It was not at all obvious what relationships such data might have to the physical oceanographer's usual repertory of salinity, temperature, and depth measurements, the biologist's con- cerns with flora or fauna, or the geologist's interests in rocks or sedi- ments. After nearly two decades of ac- tivity in space it is becoming ob- vious that for several limited but nevertheless important classes of phenomena it is possible to make observations and measurements from spacecraft of considerable usefulness to oceanographers. In a few isolated instances it even ap- pears one may do so with a breadth and accuracy exceeding anything attainable from ships or buoys. For these types of observations, the satellite represents a new tool of great power, and the information on physical and biological processes ob- tained from it will be worthy of in- clusion in the data banks and in the minds of researchers. However, for a sizable percentage of physical and biological ocean sci- entists, much of these data may fall far afield or might be too indirect or perhaps even too esoteric for their tastes. The value of the data to these workers will chiefly be in the concomitant enlargement of the general fund of oceanic knowledge. By and large, satellite oceanogra- phy is confined to surface and near- surface phenomena. This constraint is not as severe as it appears at first glance, hecause data taken from spacecraft will be appended to other, conventionally derived sur- face and subsurface measurements of parameters such as vertical cur- rent or temperature profiles in order to construct a more nearly three-dimensional view of the ocean. In addition, near-surface data are useful in their own right, since the coupled nonlinear interac- tions between ocean and atmo- sphere largely take place in the few tens of meters above and below the sea-air interface, at least for shorter time scales. Man's marine activities are mostly confined to near that surface as well, so that the kind of two-dimensional oceanography that one can pursue from spacecraft is often highly relevant. Uses of Spacecraft Data ble 1 is a listing [LaViolette, 1974; Apel and Siry, 1974; NASA, 1975; Koffler, 19751 of the spacecraft that have been or will be sources of data having oceanographic significance. Of the several listed, the most useful are probably NOAA 3 and 4, ERTS 1/Landsat 2, Geos 3, the SMS/GOES quintuplets, Tiros-N, Seasat-A, and Nimbus-G. The data types are diverse, as is discussed below. The last three satellites, which are to be launched in 1978, are of much interest to oceanogra- phy. Tiros-N is the first of a new generation of operational meteorological and environmental polar-orbiting satellites. Seasat-A is dedicated to oceanography, geodesy, meteorology, and climatology \Apel and Siry, 19741. Nimbus-G is designed to serve experimental ends for both pollution monitoring and oceanography [AVISA, 19751. Data Available From Satellites Spacecraft data presently availa- ble on any basis other than a pri- marily experimental one are quite limited and are effectively confined to low- and medium-resolution visi- ble and infrared imagery (NOAA, GOES), from which sea surface tem- peratures having accuracies of order ±1.5°-2.0°C may be derived, and small amounts of high-resolu- tion Landsat images. However, the near future promises a large in- crease in the quantity, quality, and coverage of oceanic data. The estimates of data accuracy and coverage cited below are thought to be valid for the general 1978-1982 era, when Tiros-N, Seasat-A, Nimbus-G, Landsat 3, and the GOES system are all to be active. In each case the dominant instruments contributing to the The answer as to who needs what information from spacecraft ob- viously depends on the type of infor- mation that is obtainable. In research areas the disciplines served with some degree of useful- ness are marine geodesy and gravi- ty; physical, geological, and biologi- cal oceanography; glaciology; boun- dary layer meteorology; and climatology. Various maritime operations, shipping, offshore min- ing, oil drilling, and fishing, all re- quire an improved and expanded data base and more accurate marine forecasts. The ever-increas- ing fraction of the population living along the seacoasts needs improved forecasting and warning services for protection of life and property. However, because of the great length and breadth of the sea the difficulties in obtaining timely detailed information of sufficient observational density across its ex- panse have prevented an effective monitoring and forecasting system for the oceans. Satellites of Utility to Oceanography The number of satellites carrying sensors that yield data useful to ocean science is large, and the value of the data from them variable. Ta- Fig. 1. Surface isotherms in degrees centigrade of Lake Huron derived from the VHRR sensor on the NOAA 4 environmental satellite, August 7, 1975. Relative ac- curacy is approximately ±1°C (NOAA, National Environmental Satellite Service). 613 TABLE 1. U.S. Satellites of Utility in Oceanography Satellite Launch Date Orbit Utility of Data Character Sensors Oceanic Parameters Mercury Gemini Apollo Apollo-Soyuz Nimbus 4 Nimbus 5 Nimbus 6 Nimbus-G ITOS 1-4 ESS A 1-9 NOAA 1-4 ATS 1-3 SMS/GOES 1-5 Geos 1-3 ERTS 1 Landsat 2 Landsat 3 Skylab 1962- 1975 1970 1973 1975 1978 1966- 1975 1966 1967 1974- 1978 1965 1975 1972 1974 1978 1973 Variable Polar Polar Low to Exploratory Cameras medium Imagery Polar Medium Experimental IR and MW radiometers Temperature, ice cover, progressing and bolometer; color radiation budget, wind, to high scanner color Medium and Operational high Synchronous Medium Prototype Synchronous High Operational Variable High Experimental Medium Prototype progressing to high Medium Experimental progressing to high Visible vidicon; IR scanner Visible, IR scanners; data channel Visible, IR scanners; data channel Laser reflectors; altimeter Visible, near-IR scanner; thermal IR scanner Cameras, visible, IR scanner: spectre radi- ometer; MW radiom- lU'is; altimeter; scatter- ometer Shuttle 1983 Varied M edium to high Varied Varied Tiros- N 1978 Polar H.gh Operational Visible, IR scanners Seasat-A Seasat-B 1978- 1983 Near Po! lar High Experimental Altimeter; imaging radar; scatterometer; MW radiometer; visible/IR scanner Imagery, temperature Imagery, temperature, data relay Imagery, temperature, data relay Geoid, ocean geoid Imagery, temperature Imagery, temperature, wave height, wind speed, geoid Unknown Imagery, temperature Geoid, wave spectra, wind speed, ice, temperature From LaVioh'ttv 119741, Apel and Siry 11974!, and S'ASA 119751. TABLE 2. Sensors of Oceanographic Interest Short Form Sensor Name Wavelength or Frequency Spatial Spacecraft Resolution NOAA 1-4 7 km NOAA 1-4 1 km GOES 1-7 km Tiros-N 1 km ERTS/Landsat 1 and 2 70 km Landsat 3, Landsat 4 25 m, 100 m (IR Nimbus-G 800 m Nimbus 5 15 km Nimbus-G, Seasat-A 15-140 km Skvlab. Geos 3, Seasat-A 2 km Skvlab, Seasat-A 25 km Seasat-A 25 m SR Scanning radiometer VHRR Very high resolution radiometer VISSR Visible and infrared spin scan radiometer AVHRR Advanced very high resolution radiometer MSS Multispectral scanner Thematic mapper CZCS Coastal zone color scanner ESMR Electronically scanned microwave radiometer SMMR Scanning multichannel microwave radiometer Alt Short pulse altimeter Scatt Radar wind scatterometer SAR Synthetic aperture radar Visible and thermal IR Visible and thermal IR Visible and thermal IR Visible and thermal IR Four channels, visible and reflected IR; Thermal IR Six channels, visible, reflected and thermal I R 19 GHz Five channels: 6.6. 10. 18. 21, 35 GHz 13.9 GHz 13.4 GHz 1.4 GHz 614 measurement are listed, although to achieve the precision or accuracy cited, ancillary data will usually be required. There is every reason to blend surface and satellite data together, so that the space-derived information can be calibrated and verified by point surface measure- ments and thus can often extend the surface observations to near-plane- tary scales. The sensors of prime in- terest are also cited and with their shortened forms are listed in Table 2. One finds a diverse list of features, or observables, that enter into oceanic processes. In listing these parameters it is convenient to begin at the level of the action of the atmosphere upon the sea; then follow the ocean's response, waves and currents, and its effects upon the shore. Other land-sea interac- tions are then listed. Identification of water mass properties established by natural and man- made influences is discussed next. Finally, some estimates of the role of the ocean in establishing climatology are given. In many of the parameter values and ranges given below the lack of experimental verification requires that the data be regarded as preliminary estimates only, and the reader is cautioned to remain skep- tical. In most cases they represent compromises between requirements leveled by the ocean scientists and the attempts of the instrument designers to meet those require- ments via remote sensing. Air-Sea Interaction The transport of matter, momen- tum, and energy across the air-sea interface is chiefly due to solar radiation and atmospheric stress. Such parameters as the air-sea tem- perature difference, exchange of la- tent and sensible heat, and the vec- tor surface wind field are important observables for climatological, meteorological, and oceanic pur- poses. For spacecraft the following estimates appear reasonable. Sea surface temperature. For the estimated capability, in cloud-free areas it should be possible to deter- mine absolute temperature ac- curacy to order 1°C and precision or relative accuracy to approximately ±0.5°C. Over coastal waters and lakes, space-time averaging of order 4 km and 1 day is needed \Koffler, 19751; for regional ocean areas, 10- km and few-day averages are re- quired; in the open ocean, 50-km and several-day averages should suffice [Bromer et al., 19761. The sensors to be used are VHRR, VISSR, and AVHRR (Table 2). In cloudy areas or in light rain a tem- perature precision of ±1.5°-2.0°C should obtain with 100-km and few- day averages away from coasts by using SMMR. To the satellite- derived temperatures should be ap- pended ship surface and vertical temperature profiles to the max- imum extent possible. Figure 1 shows isotherms for Great Lakes surface temperatures as an example of the current high- resolution thermal mapping in a limited region, derived from the VHRR sensor on the NOAA 4 satellite [Koffler, 19751. Surface rector wind field. As referenced to a 20-m height, the scatterometer may measure surface wind speed from a very few to perhaps 20 m/s, with a precision of about ±2 m/s or 25% of the actual value (whichever is larger) and wind direction to ±20° through clouds and light rainfall; 25-km resolution over a several hundred kilometer swath width will be the case [Grantham et al., 19751. For higher winds, attempts will be made to determine speed from 5 to perhaps 35 m/s within ±25'/!', of ac- tual speed over a several hundred kilometer swath through clouds and light rain by using the SMMR [Apel and Sirx, 1974, p. 14; NASA, 1975; Baruth and Gloersen , 1975!. Figure 2 shows radar backscatter cross section of the ocean \ A \ 17n15m20s 17n16m00' GMT-TIME-S 17n16m30' Let 198° 160° Fig. 8. (Topi Altimeter geoid heights referenced to a spheroid, as measured across western Puerto Rico by the Skylab S-193 altimeter; precision is about ± 1 m. (Bottom) Bottom topography over a portion of the subsatellite track (NASA Wallops Flight Center). imately 5 days; the resolution is ap- proximately 15 km \Gloersen and Salomonson, 19751. Marine geoid. In a quite separ- ate category from the previous nb- servables is the marine or ocean geoid, defined as the surface assumed by a motionless uniform ocean under the influence of gravi- tational and rotational forces only. Geostrophic currents, tides, storm surges, setup, and waves lead to an ocean surface that departs from the geoid; the latter must then be known on a spatial grid with preci- sion at least as fine as that with which the observable is to be deter- mined. Although only preliminary data have been published, it appears altogether possible to measure rela- tive short-scale vertical variations in the marine geoid to ±20 cm and long-scale to perhaps ±100 cm along the subsatellite track over a grid spacing of order 25 km over all open ocean areas by using the altimeter and precise orbit deter- mination \Apel and Stry, 1974; Kaula, 1970; Apel, 1972; Apel and Byrne, 1974; McGoogan et al., 1975!. Data from Skylab I McGoogan et al., 19751 and Geos 3 (H. R. Stanley, pri- vate communication, 1975) support this view. Some of the data from Skylab are illustrated in Figure 8, which shows the variation in rela- tive geoid height and water depth along the subsatellite track across Puerto Rico. By using the altimeter, whose noise figure was approx- imately ± 1 m, the gravity anomaly associated with the Puerto Rico trench is clearly seen as a geoidal depression of order 15-20 m [McGoogan et al., 19751. Such short wavelength data, taken globally, can be combined with long wavelength geoidal models obtained via satellite tracking and orbit analysis to obtain a precision geoid over the ocean. While this has not yet been done, attempts have been made to combine marine gra- vimetric measurements with satellite geoids to produce a geoidal map such as is shown in Figure 9 for the western Atlantic I Vim cut et al., 1972; Marsh et ui, 1973!. Here heights are given in meters relative to the reference ellipsoid. The problem of measuring geostrophic 620 Fir. 9. Geoid calculated in the western Atlantic from satellite orbit perturbations and marine gravity measurements; elevations are in meters relative to the reference ellipsoid. The track of Sky lab while taking the data of Figure 8 is shown as a stripe off Puer- to Rico (NASA Goddard Space Flight Center). currents is equivalent to discerning the 100-cm setup due to current shown in Figure 5 against the hack- ground of 100-m geoidal undula- tions illustrated in Figure 9 \ Kan la, 1970; A/W, 1972; Apcl an, I Byrne, 19741. Climatology The role of the ocean in climatic change is not completely under- stood, but it it clear that the transformation of absorbed sunlight into thermal energy in the upper layers of the sea is an important one, as is the poleward transport of this heat by western boundary cur- rents. Variations in the positions of major ocean currents in part appear to be induced by changing wind stress, which apparently lead to the El Nino phenomenon, for example. The appearance of anomalous large areas of warm water in the Pacific has been hypothesized as the origin of warm winters in the eastern United States through poorly under- stood processes involving motions of the upper atmosphere [Gates and Mintz, 19751. The contributions which spacecraft can make to ocean climatology therefore appear to be mainly related to the global deter- mination of sea surface tern- 10 621 10 5 d. 5 4.13 mg m~3 9.2 xlCT'nn-'sr-1 GULF STREAM Chl-aC13mg m-J yS i 1.7xl0_sm*"sr-' COASTAL Chi a 0.28mg m-' /3 :6.7xlO"V'sr-' I I 400 500 600 700 WAVELENGTH 800 900 (nm) Fig. 10. the Gulf surface i Upwelling spectral irradiance as measured in three types ofwatei masses in of Mexico; the shift toward the red end of the spectrum is due to increased hlorophyll a iNOAA Environmental Research Laboratories). perature and heat transport \Gates and Mintz, 1975; Stommel, 1974; NACOA, 19741. The five GOES-type synchronous satellites appear capa- ble of delivering the temperature data over mid-latitude regions with the required accuracy of ±0.5°C rel- ative, if special processing is under- taken. Over polar regions the Tiros- N series is more suitable. Programs for optimal extraction of the global temperature fields, averaged over approximately 100 x 100 km- areas and several days, are in the embryonic stages. Water Mass Properties Variations in the physical or chemical composition of a water mass lead to variations in its color or reflectivity, for example. These changes can be natural or man- made; in either case they tend to be more pronounced near continents. The color is determined primarily by molecular scattering and secon- darily by nutrients, chlorophyll a in plankton and algaes, suspended sediment load, pollutants, and, TABLE 3. Summary of Sensors and Observables Imaging Radiometers Short Pulse Imaging Observables Visible Thermal IR Microwave Altimeter Radar Chlorophyll and algaes Current position 1 2 1 1 1 1 Current speed Estuarine circulation Fog Ice cover 1 1 2 1 1 1 3 1 Icebergs Internal waves - 1 1 Marine geoid Oil spills - - 1 1 Pollutant identification Salinitv - 3 • - Sea state and swell Sediment transport Setup 2 1 - 2 1 1 1 Shallow water bathymetrv 1 - - - Storm surges Surface winds 3 3 3 1 1 2 3 2 Temperature Tides 1 1 1 - Tsunamis 1 - Upwellings Water vapor 2 1 1 2 1 . Wave refraction 1 - - - 1 Wave spectrum 2 1 Scatterometer Numbers indicate order or importance in determining the observable with 1 for primary. 2 for secondary, and 3 for tertiary. Hy- phens indicate no utility. 622 11 where water is sufficiently shallow, water depth and bottom type. Other environmental factors such as atmospheric conditions, sun and viewing angles, surface Winds, and waves also influence the measure- ment of ocean color. Figure 10 shows surface measure- ments of upwelling spectra from three types of water masses and il- lustrates the increase in energy in the green and red regimes of the spectrum as the transition "from Gulf Stream to estuarine water is made [Maul and Gordon, 1975]. Figure 11 is a computer-enhanced Landsat image of a 140 x 140 km2 sector of the New York Bight, show- ing suspended sediments from the Hudson River, acid-dumping events, water mass variations, and internal waves, the last being visible because of the sun glint [Apel et al, 1975]. Ocean color. The CZCS on Nim- bus-G will image the ocean surface and near-surface in multiple wavelengths of visible light and reflected and thermal infrared radiation with 800-m spatial resolu- tion over swath widths of 700 km under controlled illumination condi- tions; the observation interval will be 1-6 days. The choice of wavelength bands was dictated by the requirement for making quan- titative measurements relating to chlorophyll and sediment concen- trations (W. Hovis, private com- munication, 1976). Measurement of ocean color from radiometric quality imagery of the desired area in several spectral in- tervals will perhaps allow measure- ment, at least under certain limited conditions, of the following features: suspended near-surface sediment distribution and concentration; chlorophyll distribution and concen- tration between perhaps 0.1 and 20 mg/m3 (W. Hovis, private com- munication, 1976); fish stock loca- tion via relationship to biosignifi- cant observables [Stevenson et al., 1973]; and pollutant distribution and concentration [Wezernak and Fig. 11. Image of the New York Bight madeVith Landsat 1 on July 24, 1973. The 'marbling' effect is due to light winds; internal waves are visible in the southeast section (NOAA Environmen- tal Research Laboratories). Thomson, 1972]. The CZCS sensor may be used to make most of the measurements 17VAS/4, 1975]. Surface reflectivity. By viewing toward rather than away from the sun it is possible to observe surface features in the sun glitter owing to the changes in surface reflectivity. A variable viewing angle is required to measure either color or reflected sunlight; viewing upsun allows determination of oil spills, internal waves via surface slicks, and varia- tions in surface roughness (Figure 11). Table 3 summarizes the various parameters discussed above and lists the sensors and instruments contributing to their determination. The estimates of their usefulness are given by primary (1), secondary (2), and tertiary (3) designations. Surface Data Collection From Spacecraft The United States and France have programs in data collection from unmanned automatic buoys, both anchored and drifting, with methods for data transmission through such satellites as SMS/GOES and Tiros-N. In addi- tion, the United States maintains large archives for surface-derived oceanographic and meteorological data. It is felt that presently planned systems are sufficient to meet the buoy data collection and positioning requirements in the next 5 years. Integrated Global Ocean Station System (IGOSS) A system called IGOSS is an evolving cooperative services system for international exchange of ocean data proceeding under the auspices of the Intergovernmental Oceanographic Commission of Unesco [Junghans and Zachariasont 1974]. The coordination activities needed to amalgamate the quite dis- parate oceanic data sources, includ- ing some of the data coming from the spacecraft systems discussed here, will be undertaken by IGOSS if present plans materialize. However, much of the spacecraft data are experimental, and their reliability and accuracy not yet established, and it is not clear how the archiving will be accomplished. The presently recommended method of utilizing satellite-derived data is to become involved with the ongoing programs as a scientific investigator or similar role. Summary It has almost invariably been the case that the introduction of a sig- nificant new instrument technology has yielded for the science to which it was applied a number of un- suspected and often highly signifi- cant results. Such serendipitous dis- coveries can surely be expected from instruments as advanced as those being orbited on ocean-look- ing satellites. Oceanographers have been hard put to gain the overview of their domain required to under- stand synoptic or planetary scale events in the sea; for a limited but important group of phenomena, satellites promise to provide the vantage point for this vision. References Apel, J. R., Ed., Sea Surface Topography From Space, vol. 1 and 2, Tech. Rep. ERL 228, Nat. Oceanic and Atmos. Admin., Boulder, Colo., May 1972. Apel, J. R., and H. M. Byrne, Oceanogra- phy and the marine geoid, in Applica- tions of Marine Geodesy, p. 59, Marine Technology Society, Washington, D. C, 1974. Apel, J. R., and J. W. Siry, A synopsis of Seasat-A scientific contributions, in Seasat-A Scientific Contributions, NASA, Washington, D. C, July 1974. Apel, J. R„ H. M. Byrne, J. R. Proni, and R. L. Charnell, Observations of oceanic internal and surface waves from the Earth Resources Technology satellite, J. Gcophys. Res.. 80, 865, 1975. Barath, F. T., and P. Gloersen, The scan- ning multichannel microwave radiometer, paper presented at U.S. Annual Meeting, Int. Union of Radio Sci., Boulder, Colo., 1975. Bromer, R. L., W. G. Pichel, T. L. Seg- nore, C. C. Walton, and H. S. Gohr- band, 'Satellite-derived sea-surface temperatures from NOAA spacecraft, NOAA/NESS Tech. Memo., in press, 1976. Brown, W. E„ Jr., C. Elachi, and T. W. Thompson, Radar imaging of ocean surface patterns, J. Gcophys. Res., 81, 2657, 1976. Defant, A., Physical Oceanography, vol. 1, Pergamon, New York, 1961. Gates, W. L., and Y. Mintz, Understand- ing Climatic Change: A Program for Action, National Academy of Sciences, Washington, D. C, 1975. Gloersen, P., and V. V. Salomonson, Satellites: New global observing tech- niques for ice and snow, J. Glaciol., 15, 373, 1975. Grantham, W. L., E. M. Bracalente, W. L. Jones, J. H. Schrader, L. C. Schroeder, and J. L. Mitchell, An operational satellite scatterometer for wind vector measurements over the ocean, NASA Tech. Memo. X-72672, 1975. Hendershott, M. C, W. H. Munk, and B. D. Zetler, Ocean tides from Seasat-A, in Seasat-A Scientific Contributions, p. 54, NASA, Washington, D. C, July 1974. Junghans, R., and R. Zachariason, The integrated global ocean station system (IGOSSl, in Environmental Data Ser- vice, National Oceanic and Atmo- spheric Administration, Government Printing Office, Washington, D. C, July 1974. Kaula, W. M. (Ed.), The Terrestrial En- vironment: Solid Earth and Ocean Physics, MIT Press, Cambridge, Mass., April 1970. Koffler, R., Uses of NOAA environmen- tal satellites to remotely sense ocean phenomena, in Ocean '75 Conference Record, Institute of Electrical and Electronics Engineers and Marine Technology Society, Washington, D. C, 1975. LaViolette, P. E., Remote optical sensing in oceanography utilizing satellite sen- sors, in Optical Aspects of Oceanogra- phy, edited by N. G. Jerlov and E. S. Nielsen, Academic, New York, 1974. Marsh, J. G„ F. J. Lerch, and S. F. Vin- cent, The geoid and free air gravity anomalies corresponding to the Gem-4 earth gravitational model, NASA/GSEC X-592-73-58, Feb. 1973. Maul, G. A., and H. R. Gordon, On the use of the Earth Resources Technology satellite (Landsat-1) in optical oceanography, in Remote Sensing of the Environment, p. 95, Elsevier, New York, 1975. McGoogan, J. T., C. D. Leitao, and W. T. Wells, Summary of Skylab S-193 altimeter altitude results, NASA Tech. Memo. X-69355, Feb. 1975. Molinari, R. L., Buoy tracking of ocean currents, Advan. Astronaut. Sci., 30, 431, 1974. NACOA, Third annual report to the President and Congress, Government Printing Office, Washington, D. C, 1974. NASA, Announcement of opportunity Science support for the Nimbus-G sen- sors, NASA A.O. OA-75-1, Wash- ington, D. C, 1975. Pirie, D. M., and D. D. Steller, California coastal processes study. Third ERTS 1 Symposium I, NASA Spec. Publ. 351, 1413, 1974. Polcyn, F. C, and D. R. Lyzenga, Updat- ing coastal and navigational charts using ERTS-1 data, Third ERTS-1 Symposium I, NASA Spec. Publ. 351, 1333, 1974. NASA, Announcement of opportunity: Science support for the Nimbus-G sen- sors, NASA A.O OA-75-1, Washington, D. C, 1975. Stevenson, W. H., A. J. Kemmerer, B. H. Atwell, and P. M. Maughan, A review of initial investigations to utilize ERTS-1 data in determining the availability and distribution of living marine resources, in Third ERTS-1 Symposium I, NASA Spec. Publ. 351, 1317, 1973. Stommel, H., The Ocean 's Role in Climate Prediction, National Academy of Sci- ences, Washington, D. C, 1974. Vincent, S., W. E. Strange, and J. G. Marsh, A detailed gravimetric geoid of North America, the North Atlantic, Eurasia, and Australia, NASAIGSFC X-553-72-331, September 1972. Walsh, E. J., Analysis of experimental NRL radar altimeter data, Radio Sci., 9, 711, 1974. Wezernak, C. T., and F. J. Thomson, Barge dumping of wastes in the New York Bight, ERTS-1 Symposium Pro- ceedings, NASA Doc. X-650-73-10, 142, 1972. John R. Apel is a supervisory oceanographer and Director of Pacific Marine Environmental Laboratory in Seattle, Washington, a component of the NOAA Environmental Research Laborato- ries. He holds B.S. and M.S. degrees in theoretical physics from the University of Maryland and a Ph.D. in applied physics from Johns Hopkins University. His specialties are in the physics of fluids and in remote sensing. Apel is a consultant to numerous government organizations, and has played a leading role in the development of satellites for oceanography. 624 13 2 Reprinted from: Journal of Geophysical Research, Vol. 81, No. 13, 2437-2441 High-Speed Streams and Sector Boundaries C. Sawyer Ocean Remote Sensing Laboratory, Atlantic Oceanographic and Meteorological Laboratories Environmental Research Laboratories, NOAA. Miami, Florida 33149 High-speed streams in the solar wind are located with respect lo interplanetary magnetic sectors, and t lie i r location in the sector is analyzed. The relation lo the sector taken as a whole is clearer than the rela- tion lo sector boundaries. The streams occur preferentially near the center of the sector. Although high- speed streams, as do sector boundaries, show a clear pattern of recurrence with solar rotation and although they are about equally frequent in the period studied (1965-1471). there is no one-to-one relation between them: sectors w ith no stream or w ith more than one stream are common. Longer sectors contain more streams, showing that streams occur at a certain rate per unit lime rather than at a constant number per sector. Introduction This paper describes an investigation of the relation of high- speed streams in the solar wind to interplanetary magnetic sectors. Earlier studies showed that the average solar wind velocity rises after passage of a sector boundary [ Wilcox and Ae.vv. 1965; SesseiaL, 1971]. Here we find high-speed streams to occur most frequently near the middle of a sector. This is true for sectors of either polarity, although the frequency of streams organized about sector boundaries may depend on the sense of the boundary. D^ta on High-Spfed Streams and on Magnetic Sectors Intriligator [1973. 1974] presents and discusses a com- pilation of data on high-speed streams in the solar wind based on measurements of velocity made at earth-orbiting Vela and sun-orbiting Pioneer spacecraft. The times of peak speed at the spacecraft are given, along with the corotation delay time appropriate to the beginning of the stream. These delay times were used to estimate the times of both beginning and max- imum at earth passage of the stream. In the appendix the error in applying to the maximum the delay time appropriate to the beginning is shown to be relatively small. Times of maximum were used to identify observations of the same stream at differ- ent spacecraft. Grouping of these observations reduced the 349 entries in Intnligator's list to descriptions of 235 separate streams, for 215 of which interplanetary magnetic data are available. A list of "observed and well-defined' sector boundaries given by Wilcox [1973] includes 51 sector boundaries in the period July 1965 to November 1970. These are supplemented by data from charts presented bv Wilcox and Colburn [1969. 1970, 1972]. hxcept for one period of missing data, April to August 1967. a plausible map of the sector structure from July 1965 through 1970 can be completed. This defines 195 sectors, of which 44 are 'well-defined" in the sense that both boundaries appear on the Wilcox list, while the remainder fit into an evolving recurrent pattern that includes all of the well-defined boundaries. Sectors shorter than 4 days are included only when they belong to a recurrent sequence. Rk lrrence of Sectors and High-Speed Streams Both sector boundaries and maxima of high-speed streams are plotted in Bartel's 27-day recurrence scheme in Figure 1. Copyright © I97d b> the \mencan Geophysical Union. This plot is strongly compressed in the vertical direction, and departures of recurrence period from 27 days are thus exagger- ated. Shading connects velocity maxima considered to be members of a sequence of rotational recurrences. Open circles have been added to emphasize gaps in these sequences, i.e., when an expected stream does not appear on the list. In- triligator points out that gaps exist in the data w hen there was no ground tracking of the spacecraft. These data gaps are not seriously detrimental to our purpose. Even if all the gaps in sequences represent data gaps, no more than 10% of the streams were missed. In any case, missing streams are expected to have no systematic effect on the main conclusions of this study. The identification of sequences is of course not certain, and the reader will have to judge to what extent a different identi- fication of sequences is possible and how it might affect the derived recurrence period, noting that this value is unaffected by wiggles but is determined by the mean slope. While one can pick out streams with recurrence patterns that match a nearby sector boundary, there are also many stream sequences that cross from one sector to another, seem- ing to develop quite independently of the sector pattern. Fig- ure 2 shows the distribution of values of recurrence period for the sector boundaries and high-speed streams shown in Figure 1. Sector boundary recurrence periods are distributed more broadly than those of stream maxima, no doubt because boundaries run behind or ahead of the sector center when the sector is waxing or waning. The mean recurrence period of high-speed streams is shorter than that of sector boundaries in the same epoch by 0.45 day, which is more than 4 times the variance of the mean. Gosling [1971] compared daily velocity measurements from close and from distant spacecraft and found that solar wind speeds at one location are almost uncorrelated with speeds measured at another location when the separation corresponds to a rotation delay time of more than 4 days. He concluded that solar wind speed cannot be successfully predicted over a span of more than 4 days. In contrast, the recurrence of high- speed streams in Figure I is remarkably stable and would permit accurate advance prediction of the occurrence and time of passage of a high-speed stream, though perhaps not of the peak velocity. The difference between the two conclusions lies in the fact that here we consider the high velocities, forming the top tenth of all the days. Gosling's results, which show a high mean predicted speed for the highest category of observed speeds, are not in conflict with the present conclusion. 14 2438 Saw. i r: Brih Rkport Apr 24 65 May 21 Jun 17 Jul 14 Aug 10 • Sep 06 Oct 03 Oct 30 Nov 26 Dec 23 Jan 19 66 Feb 15 Mar 14 SApr 10 May 07 Jun 03 Jun 30 Jul 27 Aug 23 Sep 19 Oct 16 Nov 12 Dec 09 Jan 05 67 Feb 01 Feb 28 Mar 27 Apr 23 May 20 Jun 16 Jul 13 Aug 09 Sep 05 Oct 02 Oct 29 Nov 25 Dec 22 Jon 18 68 Feb 14 Mar 12 Apr 08 .May 05 WJun 01 <£jun 28 Jul 25 Aug 21 Sep 17 Oct 14 Nov 10 Dec 07 Jan 03 69 Jon 30 Feb 26 Mar 25 Apr 21 Moy 18 Jun 14 Jul II Aug 07 Sep 03 Sep 30 Oct 27 Nov 23 Dec 20 Jan 16 70 Feb 12 Mor II Apr 07 May 04 May 31 Jun 27 Jul 24 Aug 20 Sep 16 Oct 13 Nov 09 Dec 06 Jan 02 71 Jan 29 Feb 25 Mar 24 Apr 20 May 17 Jun 13 Jul 10 Hg. 1. Twenty-seven-daN recurrence diagram of inlerplanelar) magnetic sectors and high-speed streams in the solar wind. Battel's da> zero is listed at the right and is located at the arrow, shown on the left. Dashed lines are in the toward sector: continuous lines are in the aw a> sector. Time of earth passage of the peak of a high-speed stream is indicated b> solid circles, and absence of a listed stream in a sequence is indicated b> open circles. Note that the strong vertical compression of the chart exaggerates drills due 'o periods longer or shorter than 27 da\s. Location of Strfam in Sfctor In order to make a more quantitative investigation ol the location of streams in magnetic sectors. I classified each stream according to the day in its sector that peak speed occurred. noting also the field direction in the sector and the duration of the sector in days. Then 1 counted the number of streams in each category; e.g.. there are two streams with maxima on day zero (same day as boundary) of 6-day sectors of away (away from sun) polarity. 15 Sawyer: Brief Report 2439 CC => LU O CD CD 50 r- 40 - 30 - 20 - 10 - 0 180 sector boundaries mean L _i_ _i_ 22 23 24 25 26 27 28 29 30 31 32 RECURRENCE PERIOD, DAYS Fig. 2. Frequency distribution of recurrence period values for (a) high-speed streams and (h) sector boundaries, in sequences indicated in Figure I . Number oj streams in sectors of different duration. First, let us examine the density of high-speed streams as a function of sector length. The data for away and toward sectors both show the same trend and are combined in Figure 3, where the quantity (number of high-speed streams in sectors of length /(/(number of sectors of length /) is plotted against /. If high- speed streams were closely associated with sector boundaries, we should expect to see one stream in each sector, since the total number of streams and sectors is approximately equal. In fact, the number of streams per sector increases with sector length, in agreement with the hypothesis that streams occur at a constant rate of 3. 1 per rotation regardless of sector length. Occurrence frequency of streams at each sector day. From the results of Wilcox and Ness [1965] and Ness el al. [1971] we expect the maximum of a high-speed stream to tend to fall 2-4 days after passage of a sector boundary. Figure 4 shows the number of maxima occurring in each day of a sector, summed over all away sectors and separately over all toward sectors. Because a day 0 occurs in every sector but later sector days are less frequent (day 20 occurs only in sectors of a duration of at least 21 days), the number of maxima decreases away from the beginning of the sector. In order to interpret the distribution of stream maxima among sector days we need to know the ex- pected frequency, given the distribution of sector lengths. The smooth continuous curves in Figures 4a and 4b show the Fig. "0 " 2 4 6 8 10 12 14 16 18 20 22 24 26 28 I LENGTH OF SECTOR, DAYS 3. Number of high-speed streams per sector as a function of sector length. Streams occur at the rate of about three per rotation regardless of sector length 1 1 1 1 1 1 1 1 1 r + SECTORS ALL SECTORS 8 10 12 14 16 18 20 22 24 26 28 DAY IN SECTOR Fig. 4. (o) The number of high-speed streams occurring on each day of a sector for away (plus) sectors, (b) The number of high-speed streams occurring on each day of a sector for toward (minus) sectors. The continuous curve shows the expected number (see text) with dashed curves at one standard deviation (square root of counted number), (c ) All data are combined, and the quantity plotted is the reduced number, the difference between the observed and expected numbers, divided by the variance. Error bars show the square root of the observed number for the peak values. Eighty percent of the points are expected to lie between the dashed horizontal lines. expected number as a function of sector length /: (total number of high-speed streams/total number of sector days) X number of sectors with duration greater than /. The general tendency for the frequency of high-velocity streams to fall as distance -12-10-8 -6 -4 -2 DAYS BEFORE 2 4 6 8 10 DAYS AFTER 16 Fig. 5. The reduced number of high-velocity streams is plotted (a) for away/toward (plus/minus) sector boundaries, {b) for to- ward/away (minus/plus) boundaries, and (c ) for all boundaries. One third of the points are expected to lie beyond the horizontal dashed lines. 2440 Sawv ir: Brih Ri fori from sector beginning increases follows the expected trend. The dashed curves differ from the expected number by plus or minus the square root of the expected number, which we take as an estimate of the variance. In a normal error distribution we expect a proportion of 0.32, or about one third of the values, lo diller by this amount or more from the mean or expected value. Of the 21 points of figure 4a or the 22 points of figure Ah we expect 7 to fall beyond the dashed curves and observe 7 and 5. In figure 4c. data from away and toward sectors have been combined. The plotted quantity is the re- duced number: observed number of high-speed streams - expected number (expected number)' 2 Again, deviant points are scattered across the sector, the total number deviating beyond the expected variance about as ex- pected, from this analysis the high-speed streams seem to be randomly located with respect to the leading boundary of the sector, though we shall lind a relation to the sector itself. Location of stream with respect to sector boundary. Next. let us designate the last day of a sector as day —I, the day before as day -2, etc.. so that we can examine the stream frequency on either side of a sector boundary. In Figures 5a and 5A is plotted the reduced number, i.e.. (observed number -expected number)/expected variance, for away/toward and for toward/away sector boundaries, and in Figure 5c it is plotted for all sector boundaries. The difference between this organization and that of the preceding section and Figure 4 can be shown by an example. A stream that falls on dav 8 of a 10-day sector appears at dav 8 in Figure 4 and at dav -2 in Figure 5. Again, difference from a random distribution is difficult to demonstrate. Of interest, however, is the minimum at dav - I, which corresponds to a minimum in geomagnetic index Kp found by Shapiro [1974] to precede aw ay/ toward sector boundaries and to be the outstanding feature of his analv sis. Stream location as Jraciion oj sector. Finally, each stream maximum was characterized as being in the first tenth, second tenth. ■ • ■ , nlh tenth of its sector. Figure 6 show s the plots for SECTOR BOUNDARY SECTOR BOUNDARY I -I - i i — l 1 1 1 1 1 1 1 r— i t— \ t ;„, - \ * A - \ / \ rv> -J v\ ■'/ V ,6 \> ^%>s. / N4 \ y y N i < i i > i i 1 1 1 1 1 1 0 12 3 4 5 6 7 90 12 3 4 5 6 7 STREAM POSITION IN SECTOR (FRACTION OF SECTOR LENGTH) Fig. 6. The reduced number of high-velocity streams is plotted for each tenth ol a sector (a) I'lot lor away sectors, [b) Plot lor toward sectors, (r) Plot lor all sectors, One third of the points arc expected to kill hcvond the hon/ontal lines away sectors, for toward sectors, and for all sectors. Although deviations from expected values are not large, these plots show a consistent trend through the sector, with below-average oc- currence frequency near the beginning and end of the sector and above-average occurrence frequency in the center of the sector for both away and toward sectors. In the combined data we lind 6 of 10 points falling beyond the expected variance where 3 are expected. Although high-speed streams may fall anywhere in a sector, they fall more frequently near the center of the sector and least frequently near the sector boundary. Thus we see that a clearer relation to high-speed streams emerges when we consider the sector as a whole rather than sector boundaries. The majority of sectors do not fit the simple- picture of a stream in midsector, however. Of 195 sectors, only 91 (47^ (contain a single stream, 56(29^ ) have no stream, and 48 (24%) have more than one stream. Conclusion. Although earlier studies showed the average solar wind speed to be organized around sector boundaries, specific high-speed streams are nearly randomly distributed with respect to sector boundaries, with a more obvious pattern of occurrence with respect to the whole sector. The total number of streams is similar to the total number of sectors, but fewer than half the sectors have just one stream. Longer sec- tors have more streams, so that the occurrence rate of streams is nearly constant instead of the number of streams per sector. Individual high-speed streams show a stable and predictable pattern of recurrence, in contrast to Gosling's conclusion about the unpredictability of solar wind speeds in general. Appendix: Effect of Applying Delay Time Appropriate to Beginning Velocity to Locate Maximum of Stream The delay from observation of the beginning of the stream at (he spacecraft to observation at the earth, given by In- triligator [1973], is u 17 where 0 is the longitudinal displacement of the spacecraft from the earth and Ir is the radial displacement. The angular veloc- ity of rotation of the stream is taken as 2.6934 10"" rad s '. corresponding to a synodic rotation period of 27 days. U is the velocity at the beginning of the stream. The mean peak veloc- ity for all measurements is 577 km s"1. The mean number ol 50 km s ' steps of velocity increase is 3.77. giving a mean velocity increase of 189 km s ' and mean beginning velocity of 388 km s" '. Taking a typical value of 0. 1 Au for Sr. we find the error from using beginning rather than peak velocity to be between 3 and 4 hours. In Figure 2. showing the distribution of recurrence period values, the full width at half maximum is 3.0 days, so 1.5 days is#an estimate of the uncertainly in the value of the period. This leads to an uncertainty in r of 9 hours when 0 = 90°. The median difference in time of earth passage of streams observed at dilferenl spacecraft but deemed to be the same stream is 0.9 day. The ditference from the mean is half this value, or about I I hours. We conclude that the error due to using beginning velocity rather than peak velocity is small relative lo the errors due to uncertainty in the time of maximum and in the appropriate rotation period. ■icknowledgments. The Editor thanks D. S. C olburn and R. Sha- piro lor their assistance in evaluating this report S VVVV. IK BKII I Rl I'OKI 2441 Rl I I Rl N( is Gosling. .1 P . Variations in the solar wind speed along the earth's orbit". Solar Phys . 17, 4W, 1971 Intnligaior. I) . High speed streams in the solar wind. Rep L AG--7. World Data Center \ For Solar Terr Phys.. Boulder. Colo., 1973 Iruriiigator. I) . I videnee ol solar-cycle variations in the solar wind, Astrophy* J Leu., lli/i. I 2.1-1 2b. 1474 Ness, V. V Hundhausen. and S. Bame. Observations ol the inter- planetary medium Vela 3 .md Imp 3. 1965-1967.7 Geophvs Re\ 76, 6tv4.1. |47| Shapiro. K . Cieomagnetie activity in the \icinit\ ol" sector boundaries, J. Geophvs Res . 7V. 289, 1474 Wilcox. J.. Solar activity and weather. Rep. 544. In si lor Plasma Res.. Stanford Uni\ . Stanford, C alii".. 1973. Wilcox, J., and I). Colburn, Interplanetary sector structure in the rising portion ol ihe sunspot cycle. J Geophvs. Res . 74, 23XX. 1969. Wilcox, J., and I). Colburn. Interplanetary sector structure near the maximum of the solar cycle, J Geophvs. Res . 75. 6366. 1470. Wilcox. J., and 1). Colburn. Interplanetary sector structure at solar maximum, J. Geophvs. Res., 77, 751. 1472. Wilcox. J., and N. Ness. Quasi-stationary corolaiing structure in the interplanetary medium. J Geophvs Res., 70. 5793. 1965. (Received October 6. 1975: accepted January 13, 1476.) 18 Reprinted from: \OL SI. NO. 13 Journal of Geophysical Research, Vol. 81, No. 13, JOURNAL OF GEOPHYSICAL RESEARCH 3 2435-2436. MAY I. 1976 Geomagnetic Activity at the Passage of High-Speed Streams in the Solar Wind C. Sawyer Ocean Remote Sensing Laboratory. Atlantic Oceanographic and Meteorological Laboratories Environmental Research Laboratories, SO A A, Miami, Florida 33149 M. Haurwitz Fori Collins, Colorado 8052! The times of maximum velocity of high-speed streams in the solar w ind are used lo organize the analysis of planetary geomagnetic activity index Ap. and this organization of the data is shown to give a clearer pattern than the organization of the data around sector boundaries. Geomagnetic activity is highest on the day preceding peak velocitj in the high-speed stream. The sector boundary analysis confirms the minimum in geomagnetic activity preceding sector boundary crossing found by Shapiro ( 1974) but shows little dependence on the sense of the boundary. Hirshberg and Colburn [1973] discussed the mechanism of geomagnetic disturbance that involves merging of southward- directed interplanetary field with earth's field. They showed that the presence of southward-directed held tends to be short- li\ ed. lasting only about 6 hours before the vertical component ol the interplanetary Held returns to normal. On the other hand, geomagnetic indices Kp (planetary index I and AE (au- roral zone substorm index) remain elevated for a day or longer. They suggested that the disturbance-prolonging factor may be a high-speed stream in the solar wind and showed for the period 1965-1967 that the disturbance index AE increased as clearly following passage of a high-speed stream as it did following passage of an interplanetary magnetic sector bound- ary Patterson [1973] found geomagnetic activity to be markedly higher in away sectors, where the magnetic field is directed predominantly outward from the sun. than in toward sectors, when the sectors are defined by geomagnetic diurnal variation at high latitude, although the difference disappears in space- observed sectors. Shapiro [1974] analyzed Kp about sector boundaries (space-observed) and concluded that the salient feature is a Kp minimum preceding passage of a toward/away boundary. He emphasized the difference between away /to- ward and toward away boundaries, suggesting that it tends to confirm the Hirshberg-Colburn model of disturbance initiated by held line merging and prolonged by a high-speed stream. A list of such high-speed streams measured from earth- orbiting Vela and sun-orbiting Pioneer satellites has been pub- lished and discussed by Intriligaior [1973. 1974]. The data presented there allow determination of the time of earth pas- sage of each observed stream and matching of observations at various satellites to obtain a single description of each of 235 high-speed streams in the period July 1965 through June 1971. The distribution of the observed streams in a recurrent pattern indicates that no more than I0°t of the streams were missed through lack of ground tracking of the satellites. The times of earth passage of peak speed in these streams are used as zero days in a superposed epoch analysis of the geomagnetic index Ap. and this analysis is compared to similar analyses in w hich Ap is organized around magnetic sector boundaries. These include the 'well-defined' sector boundaries determined by Wilcox [1973] as well as less certain boundaries found from charts published by Wilcox and Colburn [1969, 1970, 1972]. The 'well-defined' boundaries are preceded and followed by at least 4 days of consistent polarity. Additional sector bound- aries that do not meet this criterion are included only if they are members of a 27-day recurrent series. In Figure 1 the average Ap at away/toward sector boundaries is compared to that at toward/away boundaries. The bold lines show mean Ap values for all boundaries, and the light lines show those for well-defined boundaries. In order to estimate the significance of departures from the mean value we need an estimate of the variance of the distribution of values of Ap. For each of 20 months in 1972 and 1973 the variance a was computed accord- ing to the definition E Ap2 n Ap' n - 1 where n is the number of days in the month and zip is the mean for the month. A plot of a versus Ap showed that Q. < Photo by Dr. Donald P. deSylva A Mangrove Thicket Thus, even though he would share in the depletion of the re- source, his own net gain would be some 9.9 per cent. So by adding one cow to his herd, he would gain nearly a 10 per cent increase; so he added his one cow. Other herdsmen saw what he had done, and each of them added one cow, and very soon the grass was not able to keep up with the munching, the whole common disappeared, and none ol them was able to benefit from what originally had been a very fine resource. There is a lesson to be learned from this analogy insofar as the Florida coastal zone is con- cerned. When there were 200 people living on Biscayne Bay, they could dump all of their sew- age directly into the bay and the bay could accommodate it. But now with several million people living on the same bay, the pro- blem has become acute: if you want to have boats tied up at your marinas with no system for taking care of their sewage, then you can not expect to swim in the boat slips with impunity — at least esthetic impunity. If you build causeways, you cannot sail through them. If you bulkhead your man- grove areas to put up condomi- niums, you can not expect to have your sport and commercial fisheries as productive as they have been in the past. If you want to use bays like Biscayne Bay as disposal sites for sewage, fine. They are good ones ; they flush themselves twice a day and are relatively effective disposal mechanisms. But if you want to use them for 23 28 Photo by Dr Donald P. deSylva A Baij Squatter that, don't plan to use them for much else. Coastal Zone Program The question then is : how do you avoid totally destroying your "common" and still maintain viable uses for many of those who would use it even though the uses are in conflict? Its solu- tion will, I suspect, be in large measure a political solution. This bothers me a little bit as a scien- tist who is in love with the ocean on the wet side of the coastal zone, but again we have come to learn to live with life as it really is. Through the Coastal Zone Management Act and the inter- action between the federal gov- ernment and the states there has developed a really fine pro- gram whereby the federal gov- ernment will provide funds for the development of individual state plans for management of the coastal zone. There is one other aspect: the need for new ideas, particularly exciting new concepts. Somehow, though, "the system" just does not seem ready for new ideas. There seems to be an allegiance to the status quo which I find quite disconcerting. But let me give you an example of what I mean. I can think of five specific south Florida coastal zone prob- lems which could be solved with one solution : the problem of our incredible beach erosion ; the problem related to storm surge associated with hurricanes; the problem of inadequate beach frontage ; the problem of pro- viding adequate offshore sport fishing areas ; and the esthetic problem or the environmental enhancement problem. My contention is that all five of these could very neatly be solved by the establishment of offshore islands. My proposal is 24 29 (O Q. < that the southeast Florida area look into the engineering and economic feasibility of develop- ing offshore islands. These is- lands would be man-made is- lands situated in 50-60 feet of water. Off Miami Beach, for example, this would probably be in the order of a mile or a mile and a half offshore. These would be linear islands, maybe a mile to a mile and a half long, maybe a hundred yards wide. They would be built up from the sea floor, and once they broke the surface they would be covered with top soil, and one would plant sea grapes and palms that can sur- vive in that fairly rugged coast- al zone environment. But how would these solve the problems? Let me take the points I made in reverse order. First, the esthetics. As you look out of a Miami Beach hotel in- stead of seeing an empty ex- panse of ocean, beautiful though I consider that happens to be, you would see a series of palm- covered islands parallel to the shore, which would help to erase our image as spoilers of the en- vironment and switch it to one of improvers of the environ- ment. Sport Fishing What about the sport fishing? It is well known that the devel- opment of offshore reefs, par- ticularly artificial reefs wheth- er they be made up of rocks or old automobile tires or sunken "My proposal is that the southeast Florida area look into the engineering and economic feasibility of de- veloping offshore islands . . . These would be lineal is- lands, maybe a mile to a mile and a half long, maybe a hundred yards wide." vessels or serpulid worms, pro- vides an ecological niche where fish gradually congregate in in- creasing numbers until you have developed a very good sport fishery. So, my offshore islands would also do this. Within the south Florida area we have relatively few beaches which are open to the public. With the increasing population, there is increasing pressure for recreational use of beaches, and these islands would in fact pro- vide additional frontage for swimming, sunning, surfing and all the other things that people do on beaches. These offshore islands would also reduce the waves and it is on that aspect that my last two benefits from these offshore is- lands rest. With an approach- ing hurricane, there is a build- up of sea level because of the waters being pushed shoreward by the strong winds. The rising water is bad enough, but the real problem comes from the strong storm waves on the sur- face of these rising waters. (Continued on Page 44) 25 30 Coastal Zone Continued from Page 25) Waves in a storm surge begin to attack areas where normally waves do not cause problems. I am thinking in terms of the up- per berm on beaches, hotel lob- bies, the living rooms of beach homes. The development of these offshore islands would very def- initely reduce the amount of wave action and thus would re- duce the damage resulting from these hazardous waves riding on top of a hurricane storm surge. Politically Controversial Perhaps the most politically controversial aspect of these islands relates to beach erosion. The erosion of south Florida's beaches results from two phe- nomena. The first of these is the longshore current which moves sediment in suspension general- ly on this coast from north to south. The second is the waves themselves. As waves break on the beach, they throw sand grains into suspension. The sand grains then start to fall back down to the bottom, but the place where they fall is some- what to the south of the place where they were picked up be- cause of the longshore current. Therefore, if you can reduce the wave action at the beach, you will then reduce the southerly movement of sand; that is, the sand grains can not be thrown into suspension by the smaller (Continued on Page 46) Coastal Zone (Continued From Page 44) waves in the lee of these off- shore islands. What this means is that sand migrating down the coast with the longshore current, once coming in the "shadow" of these islands, will then be de- posited. Gradually you will have a buildup of sand on the beach in the lee of the islands. For something over five years I have been trying to get the federal government, the state government, the Miami Tourist Development Authority, anyone, to sponsor a relatively inexpen- sive engineering study of the feasibility of offshore islands for the south Florida area. Maybe the idea is no good ; if so, it should be discarded. On the other hand it may be a good idea, and I would hope that someday some group could say this is in fact worth investigat- Limit Acreage One possibility that should be considered is the limiting of ac- tual ocean front acreage to those industries or other uses which require their being right on the water. If an activity can be lo- cated equally well in West Palm Beach or in west Dade County or in the western part of Broward County rather than on the waterfront, it should be de- nied access to the waterfront. Virginia Key is an island be- tween Key Biscayne and the mainland connected to both of 31 them by causeways and bridges. Several years ago the Miami City Commission and the Dade County Commission were con- vinced that with a heavily tour- ist-dependent economy, it was important to develop other ac- tivities within the area which could lure new industry and new dollars into the county. Thus we were able to convince both Metro Dade County and the City of Miami to zone 162 acres spe- cifically for marine research work. Presently the complex has the Miami Seaquarium, Planet Ocean of the International Oceanographic Foundation, the world-renowned Rosenstiel School of Marine and Atmo- spheric Science of the Univer- sity of Miami, and two NOAA Laboratories : the Southeast Fisheries Center of NOAA's Na- tional Marine Fisheries Service, and my own Atlantic Oceano- graphic and Meteorological Lab- oratories. Over the past year, we have worked with Dade County, and three acres of county land have just been transferred to Miami- Dade Community College for its marine technician training pro- gram, and five acres have gone to a group from industry known as Palisades Geophysical Insti- tute which does primarily un- derwater acoustic research work for the navy. In ten years I can see Vir- 32 ginia Key being the major ma- stitution itself has had to build I'ine research area in the United a second campus several miles States, if not the entire world, away and inland from their Today one thinks of the Scripps main coastal lab and ship fa- Institution of Oceanography in cility. California and the Woods Hole The sort of thing that is hap- Oceanographic Institution in pening with the growing marine Massachusetts as the major science complex on Virginia Key oceanographic research places in is not something that happens the United States. But neither by itself. It takes concerned and group had enough foresight to dedicated citizens willing to ap- provide space for long term de- proach their local governments velopment. If you wanted to lo- and to lobby, if you will, to see cate near the Scripps Institu- that the things that have to be tion of Oceanography today, you done are in fact accomplished, could get probably no more In conclusion. I would like to closer than eight or ten miles be sure that I leave with you and be way back on the mesa. If only really one major point. you wanted to be at Woods Hole. That is in consideration of the there would be no chance. Even economics of the Southeast Flor- Woods Hole Oceanographic In- ida Coastal Zone, you can not afford to neglect the ocean. We must con-icier it : we must take care of it : we must utilize it ef- fectively. But in order to do this and to assure the continuing eco- nomic growth of the Southeast Florida Coastal Zone, we must continue to consider the ocean as the major aspect that makes our coastal zone and its eco- nomic and industrial develop- ment problems considerably dif- ferent from those of Saint Louis, Kansas City, or Cedar Rapids. Iowa. t-> 33 8 Reprinted from: Middle Atlantic Continental Shelf and the New York Bight, ASLO Special Symposia, Volume 2, 20-34. Section 2 Physical ysical processes Physical oceanography of the Middle Atlantic Bight1,2 R. C. Beardsley Woods Hole Oceanographic Institution, Woods Hole, Massachusetts 02543 W. C. Boicourt Chesapeake Bay Institute, The Johns Hopkins University, Baltimore, Maryland 21218 D. V. Hansen Atlantic Oceanographic and Meteorological Laboratories, NOAA, Miami, Florida 33419 Abstract Kinetic energy spectra from moored current meters in the mid-Atlantic Bight reveal marked differences in current variability between the inner shelf and the outer shelf and slope regions. The nearshore subtidal current variability appears to be dominated by mete- orological forcing. The amplitude of the semidiurnal and diurnal tidal peaks decreases in the offshore direction. Shallow water records show little or no inertial energy, while at the shelf break and over the slope, inertial motion contributes significantly to the current vari- ance. A simple conceptual model is presented to explain how intense winter low pressure systems ( "northeasters" ) drive strong alongshore currents which are coherent over much of the bight. A map of "mean" currents measured in recent moored array experiments demon- strates subsurface water flow along the shore toward the southwest. The average currents generally increase in magnitude offshore and decrease with closeness to bottom. At most sites, the mean current veers toward shore with increasing depth. The alongshore volume transport measured at three transects across the bight shows surprising uniformity, consider- ing die possible sources for discrepancy. This transport (order 2.0 X 105m3s— *) of water within the 100-m isobath implies a mean residence time of the order % year. Much of the shelf water observed flowing westward south of New England must originate in the Gulf of Maine-Georges Bank area. Before 1970, information on the circula- tion of the mid-Atlantic Bight came mostly from temperature and salinity measure- ments and from drift bottles and seabed drifters. Bigelow (1933) and Bigelow and Sears (1935) first described seasonal tem- 1 Contribution 3701 of the Woods Hole Oceano- graphic Institution and 228 of the Chesapeake Bay Institute. 2 The collection of these data has been supported by the NOAA-MESA New York Bight project, the National Science Foundation, and the Office of perature and salinity changes on the con- tinental shelf, where vernal warming and freshwater runoff build a strong stratifica- tion which is subsequently destroyed in the fall by storms and cooling. Iselin ( 1939, 1955) postulated an offshore motion in the Electric and Gas Company and B. Magnell (EG&G) contributed ideas and data. Preparation of this re- port has been supported by the National Science Foundation under grants DES-74-03001 (B.C.B.) and DES-74-03913-A02 (W.C.B.) and the MESA Naval Research. The New Jersey Public Service project (D.V.H. AM. SOC. LIMNOL. OCEANOGR. 20 34 SPEC. SYMP. 2 Physical oceanography 21 upper layers of the shelf water and corre- momentum across it are not well under- sponding shoreward flow in the lower layers stood. because salinitv generally increases with depth. He also' noted that the circulation Current variability, circulation, and obeys the "rule of coastal circulation." water structure whereby the average flow is parallel to the Self-contained current meters, tempera- coast with land on the right-hand side of an hire and pressure gauges, and other instru- observer facing downstream. Bumpus ments deployed in moored arrays in the (1973) in his summary ol a L 0-year pro- mid- Atlantic Bight are now beginning to gram of drift-bottle and seabed drifter re- provide records of sufficient' length to char- leases and occasional drogue and drift-pole aetcri/e the variability of the subsurface measurements, concluded that a mean current field in this region. We will present alongshore flow of order 5 cm s ' occurs here some preliminary results of these field from Cape Cod to Cape llatteras. except programs with an emphasis on describing during periods of strong souther!) winds the "mean" circulation and subtidal current and low runoff (Bumpus 1969). Nantucket variability. Shoals and Diamond Shoals appeal" to be We begin by examining in Fig. 1 several oceanographie "barriers" which limit the kinetic energy spectra computed from 1- alongshore flow. At the Cape llatteras end. month or longer current records obtained the alongshore How turns seaward and be- at several different sites on the middle At- comes entrained in the Cult Stream. Oc- lantic continental shelf and rise. The four casionally strong northeast winds drive a sites, labeled "A" through "D." and the loca- small amount ol mid-Atlantic Bight water tion. water depth, depth at which the cur- southward around Cape llatteras ( Bumpus rent record was taken, and other pertinent and Pierce 1955). ml on nation for each site are given in Table The transition /one between shell water 1. (Sites A through D correspond respee- and warmer, saltier slope water often oc- tivelv to stations IS. 1. 4. and 11 shown in curs during winter as a sharp inclined front lrig. 2. ) located near the shelf break. In summer, the The site A data has been taken and re- front is less distinct but large temperature ported by KG&C (1975) under contract to and salinity gradients still occur in the oft- the Public Service Electric and Cas Corn- shore direction below the seasonal thermo- panv ol New Jersey. Flagg et al. (1976) dine. These gradients are due to a band ol obtained the site B and C data. The low- cold, low-salinity water located near the frequency cutoff for each estimated spec- bottom on the outer shell. Described by trum is inversely proportional to the length Bigelow (1933) as a remnant from the of the particular current record analyzed. previous winter's cooling, these waters can The Woods Hole Oceanographic Institution have temperatures of 6 -S C in August. has maintained moored arrays at site D for Temperatures are around 16 C only 20 km almost a decade and the very long current offshore of the "cold pool." The mechanisms records obtained there allowed Webster governing the movement of this front. d (1969) and Thompson (1971) to make a /one and exchanges of heat. salt, water, and reliable estimate of the kinetic energy spec- Table 1. Location and other pertinent information for the current and wind spectra shown in Fig. 1. Water [nstr. Sta. Nn depth depth Data Site ( Kin. 2 ) Location Time On ) (m ) source A IS 39°2S\\ 75°15\Y Dec73-Feh74 12 5 FG&G (1975) B 1 40°54X. 71 ()4\V Mar 74 58 28 Flagg et al. (1976) C 4 40°1S\. 75=51 W Mar 74 112 30 Flags et al. (1976) 1) 1 1 39°20X, 70°0()\V Se\ cral years 2.640 100 Webster (1969); Thompson ( 1971 ) 35 22 Physical processes trum in the slope water over a seven-decade range frequency. The power density of the wind stress observed at site A is also shown in Fig. 1. Wind stress has been computed using the quadratic drag law t = CD|Wio|W10 where W10 is the observed wind vector at 10-m height and the as- sumed constant drag coefficient is CD = 3.2 x 10- T in c.g.s. units. The spectra have been visually smoothed within the estimated uncertainties to sim- plify graphical presentation and our spec- CL LU >- h; CO -z. LU O >- o LT Id O H LU 10 _ 10" 10' 10" 10* PERIOD T (hours) I03 I02 I01 I0C T_,— i — n — rrm — I — r1 *- 60 30 15 10 54 3 2 I 0.5 PERIOD (days) wind stress power density at site A inst. water SD peak depth depth ■*- site A 5 m 12 m ♦-site B 28m 58m ♦-site C 30m 112m ♦ site D 100 m 2640m site A wind stress at site A • site B site C — site D Tidal frequencies SD = semidiurnal = l/l2.4h =.081 cph D = diurnal = l/24h = .042 cph Inertiol frequency at 40.5°N 1=1/ 18. 5h = .054 cph ■+- .-3 H — 10" D I SD 10' LU Q 10° I o a. V) CO UJ rr \- co o z 10' \& FREQUENCY f (cph) Fig. 1. Spectra of currents and wind in the Middle Atlantic Bight. Locations of current meter ings are listed in Table 1. (See text for explanation of different formats.) 36 Physical oceanography 23 PERIOD T (hours' 4 10" 12 m T3 >- c h- i (/) 13 h_ 7* o o I LU ^_ **" n O rr — F tr CO UJ O 0J CO 6 o CO o> c LL b X (-> ■o >- CO 5 o o — z ll) , — , - — . D O jO o UJ tr c/> 4 Ll c 3 rO CM Code site A — wind stress at site A , site B site C site D — i ' i i i rrn — r 60 30 15 10 5 4 3 2 PERIOD T ( doys) 10' I0L i r I 0 5 I — 26.90(site A) 10" ■—10.75 (site B ) •3.70 (site C) site D- 1.15 (site D) 10 10 FREQUENCY f (cph) 1 1 Tk D I SDI0 D,B,C \o" Fig. 1. Continued 37 24 Physical processes tral characterization of wind and current variability over the continental shelf. The reader should remember that spectra ob- tained from much longer records will pre- sumably show more structure than the smoothed estimates shown in Fig. 1. The spectra have been plotted in both the log £(/) versus log / format (Fig. la) and the area-preserving linear 2.3 X / x £(/) versus log / format (Fig. lb). The first format best displays the functional form of the energy density E as a function of frequency, e.g. £(/) oc l/f'" corresponds to a straight line plotted in Fig. la with a slope of —m. The second format in Fig. lb is used to illustrate how much different frequency bands con- tribute to the total variance of the current record. The total area under the 2.3 X / x £(/) curve is equal to the variance, and the area under the curve between two specific frequencies is the contribution from that frequency range to the variance. The spectra shown in Fig. 1 illustrate sev- eral fundamental features of wind and cur- rent variability on and near the mid-At- lantic continental shelf. Wind stress and current spectra are inherently "red," with the power or kinetic energy density gener- ally decreasing with increasing frequency. Wind stress power density at site A is ap- proximately constant at lower frequencies with a transition occurring at periods be- tween 2 and 4 davs and a higher frequency falloff of about f1:>/2. Most of the fluctuation in the wind stress at site A is caused by rather wideband meteorological transients which have char- acteristic periods between 1 and 8 days. The intense low pressure disturbances or cyclones which generally form over the southeast United States and intensify while propagating up along the eastern seaboard have characteristic periods of 2 to 4 days and cause the peak in the site A wind stress power density curve shown in Fig. lb. In addition to being red at lower fre- quencies, the four current spectra exhibit relatively sharp peaks at the semidiurnal (SD) and diurnal (D) frequencies. Ampli- tude of the semidiurnal peaks generally increases across the shelf toward shallower water; at sites B and C, the kinetic energy density at the semidiurnal frequency crudely follows the relationship E oc h~3/2 as predicted by shallow-water wave theory. The large semidiurnal peak observed at site A is probably caused by the proximity of Little Egg Inlet which can channel and intensify local tidal currents. Semidiurnal and diurnal tidal currents are weakest at site D on the continental rise. While the kinetic energy density at the diurnal fre- quency shows a general increase with de- creasing depth across the shelf, the spatial structure of the diurnal tidal currents is not yet understood. It is important here to note, however, that semidiurnal and diurnal tidal currents on the continental shelf are in part predictable since the astronomical forcing is determin- istic and periodic. The accuracy of this pre- diction depends on the basic accuracy of the initial calibration of local tidal currents with the astronomical forcing, the degree of local nonlinearity (e.g. the phase shifting of the surface tide by strong storms), and the relative importance of baroclinic or "in- ternal" tides, i.e. internal waves of tidal fre- quency. We expect baroclinic effects to be important perhaps all the time in the deeper water near the shelf break and over most of the shelf during the warmer months when a strong seasonal pycnocline has formed. Wunsch and Hendry (1972) observed bot- tom-intensified semidiurnal tidal currents in about 850 m of water on the New En- gland continental slope. They described these observations as a train of internal waves of semidiurnal frequency generated at greater depth on the slope and propagat- ing up the slope toward the shelf. How far these internal tides penetrate onto the shelf and how much mixing is caused by their dissipation is as yet unknown. Tidal flow over topographic features can also generate higher frequency internal waves via nonlinear mechanisms. For ex- ample, trains of large- amplitude internal waves have been observed by remote sens- ing to propagate almost across the shelf during summer stratified conditions. Apel et al. (1975) believed such wave trains are formed near the shelf break by diurnal and semidiurnal tidal currents. The question of 38 Physical oceanography 25 how much energy is really drained from the frequency peaks; hence the large question barotropic tides via topographic generation mark shown in Fig. 1. It is not known at this of internal waxes remains unanswered. time how much of the lower frequency end Current spectra at sites C and D show an of the spectra is caused by local or regional additional kinetic energy density peak near meteorological forcing or by the transmis- the local inertia! frequency. The contribu- sion (or leakage) of lower frequency en- tion of the near-inertial frequency band to ergv onto the continental shelf from the the current variance is considerable at these deeper ocean. We have used the model of two sites and especially so at site C near Xiiler and Kroll (in prep.) to estimate the the shelf break (Fig. 1). The local genera- possible transmission of topographic Rossby Hon of near-inertial currents lw meteoro- wave energy from the rise onto the shelf logical transients has been well documented and find that this flux of energy across the at site D by Pollard and Millard (1970); shelf break is comparable with the direct fast moving fronts or strong veering winds kinetic energy input due to a surface wind which rotate clockwise with near-inertial stress of 1 dyne /cm- acting over the width frequency clearly excite nearly vertically of the continental shelf. Based on this and propagating internal waves. The absence of other preliminary observations, we suggest near-inertial peaks in the kinetic energy that the open ocean causes energetic low- spectra at sites A and B nearer shore is prob- frequency motion on the outer continental ably due to the existence of other "natural" shelf of the mid-Atlantic Bight. Longer cur- modes like edge and shelf waxes (see Reid rent records (8 months or longer) are 1958) which are preferentially excited dm- needed to quantify accurately the impor- ing any transient adjustment period. The tancc of low frequency energy transmission observed lack of strong near-inertial energy onto the shell. in shallow nearshore water should simplify Having shown that much of the current the local current prediction problem. variability in the shallower section of the We now turn to the lower frequency end bight is directly wind driven, we now de- of the current spectra. Long records at site scribe a simple conceptual model for the D show that much of the current variance dynamics of the response of this region to at 100 m in slope water is caused bv low strong wind events. This model, suggested frequency motion with characteristic pe- by Beardsley and Butman (1974), has been riods centered at about 30 days. Propaga- supported by other observations (Boicourt Hon of topographic Rossby waxes up the and Hacker 1976; Beardsley et al. in prep.), continental rise ( perhaps generated by the Intense winter lows, the "northeasters" Gulf Stream), meandering of the Gulf which pass to the east of the mid-Atlantic Stream itself, and formation of antic\ clonic Bight, produce strong wind stress fields (warm core) eddies can all generate strong toward the south and west over the shelf, low frequency currents at site D which generally paralleling the coast from Cape cause the spectral shape shown. In con- Cod to Cape Hatteras. The transient mass trast with site D. kinetic density spectrum flux in the surface Ekman layer has a com- at site A ( Fig. lb) shows that subtidal ire- ponent to the right of the wind stress vector fluency currents in nearshore shallow water and a component parallel to the wind stress, are strongly wind driven and cause most During northeasters, the Ekman component of the total current variance (sec EG&G directed to the right of the wind stress is 1975). We thus suggest that the current onshore, causing sea level to rise along the prediction problem in shallow nearshore coast. Wunsch ( 1972) and Brown et al. water further simplifies to the development ( 1975) have shown that sea level over the of a model which relates the subtidal cur- deep ocean (and presumably the outer rent to measurable meteorological forcings. slope) is nearly constant over time scales of Current records obtained at sites B and several days, so that the coastal rise in sea C are too short for the computed spectra to level creates a large onshore pressure gradi- indicate locations and magnitudes of lower ent that is roughly in geostrophic balance 39 26 Physical processes with the strong alongshore flow. Since the wind stress field tends to parallel the coast- line, the intense northeaster generates strong alongshore currents and cross-shelf pressure gradients which appear to be co- herent over the entire shelf from Cape Cod to Cape Hatteras. Boicourt and Hacker ( 1976 ) observed that the more energetic subtidal current fluctuations (especially those associated with northeasters) ori- ented along the 35-m isobath off Maryland and Delaware are coherent and approxi- mately in phase over distances of 230 km. They report typical maximum daily mean speeds of 40 cm/s at depths of 10 and 20 m, which can produce alongshore fluid particle excursions of 40-80 km during the several days of the storm. Beardsley et al. (in prep. ) found that subsurface pressure gra- dients caused by sea level changes are co- herent over the mid-Atlantic shelf from Cape May to Cape Cod. These observa- tions suggest that the wind-driven com- ponent of the alongshore flow may be pre- dicted from the more easily measured wind-stress and pressure fields and coastal sea level fluctuations. We will now focus on the "mean" or very low frequency current field on the mid- Atlantic Bight. We have plotted in Fig. 2 the average currents which have been mea- sured in recent moored array experiments. Only records of 1 month or longer duration have been used and information on the in- dividual measurements (e.g. local water depth, instrument depth, time of measure- ment, current values, source of data, etc. ) is given in Table 2. The mean currents are plotted as vectors with the magnitude equal to the average speed. The same current meter stations are numbered in Fig. 2 se- quentially starting from the north and the same key is used in Table 2. The depth (in meters) of an individual measurement is indicated in Fig. 2 by a small number lo- cated near the head of the current vector. We have separated the measurements into winter ( unstratif ied ) measurements (de- noted by solid vectors ) and summer ( strati- fied) measurements ( dashed vectors ) . Mea- surements from several sites (5-11) on the continental rise and outer slope are in- cluded to show the mean westward flow of slope water. The mean position of the north- ern edge of the Gulf Stream is also shown, with the reminder that the actual position of the Gulf Stream in this region is highly variable ( Hansen 1970) . These direct measurements of the mean current field on the shelf demonstrate sub- surface water flow along the shore toward the southwest. The mean currents generally increase in magnitude offshore and de- crease with closeness to the bottom. At most sites, the mean current veers toward shore with increasing depth. With the ex- ception of station 21, a net southwestward transport is observed at all sites. Measurements made along the three tran- sects labeled I (New England), II (New York), and III (Norfolk) in Fig. 2 have been used to estimate mean alongshore volume transport. The transects cover the bulk of the continental shelf out to the 100-m isobath. Calculated transport values, cross-sectional area, and mean speeds for each transect are listed in Table 3. Although Wright and Parker (1976) estimated that roughly half of the volume of the shelf water from Cape Cod to Cape Hatteras lies in a thin surface wedge outside the 100-m isobath, there are essentially no direct mea- surements of mean current in the shelf water wedge beyond the 100-m isobath. The estimated volume transports for the three transects are surprisingly consistent, considering that the northern transects (I and II ) are early spring measurements in two different years, while the southern transect (III) value represents summer measurements. In addition, the transects were made at different depths, with differ- ent instruments, and with varying spatial resolutions. For these reasons, we hesitate to speculate about exchange of shelf water and slope water based on continuity argu- ments and assumed stationary flow through the transects. We are uncertain, for ex- ample, whether the higher mean alongshore speed shown in transect III is due to a con- tinuity of transport within the 100-m iso- bath which forces the mean speed to in- crease through the smaller cross-sectional 40 Physical oceanography 27 area, or whether it is clue to a more con- sistency of the transports lead us to specu- sistent southward flow in summertime (for late that there may be little significant sea- which there is some evidence). The con- sonal change in alongshore transport. Only 5? .J / 30 ' ^ ^PE 'J / HATTERAS/ ~$ ^" CAPE / 76= / 75° 74o I 73° 72° 70° Fig. 2. Mean velocities as measured by moored current meters in the Middle Atlantic Bight region. Winter measurements are indicated by solid arrows, summer velocities by dashed arrows. Individual sta- tions are numbered according to Table 2; station numbers are circled. Measurement depths (in meters) are shown near the head of the arrows. 41 28 Physical processes Table 2. Tabulation of the recent direct measurements of sub- and near-surface mean currents shown in Fig. 2. ( NA = not applicable. ) Sta. No. Location Start time Record length ( days ) Water depth (m) Instr. depth (m) E (cm/s ) N (cm/s) Data source* 1 40°54N/71°04W 28 Feb 74 35 58 28 57 -2.1 -0.2 -0.5 0.8 A 3 40°33N,70°56W 28 Feb 74 35 72 24 44 62 71 -5.7 -2.2 -2.2 -0.1 0.5 1.0 1.8 -0.5 A 4 40°18N,70°51W 28 Feb 74 35 110 30 50 70 109 -7.8 -7.4 -5.9 -0.8 0.7 3.0 3.3 0.0 A 7 39°23N,70°59W 20 Aug 70 46 2,527 1,504 -2.8 0.3 B 9 39°35N,70°58W 20 Aug 70 111 2,263 2,163 -6.4 0.2 B 5 39°50N,70°40W 20 Aug 70 104 876 776 -6.4 1.6 B 6a b 39°50N,70°56W 39°50N,70°56W 20 Aug 70 20 Aug 70 45 45 45 451 86 943 993 846 933 941 880 990 -2.0 -2.0 -1.5 -4.6 -3.6 1.0 0.4 0.4 1.8 -0.7 B 8 39°37N,71°15W 20 Aug 70 111 2,150 2,052 -5.1 -0.4 B 10 39°23N,71°01W 20 Aug 70 56 2,509 2,394 -2.4 -0.5 B 2 40°45N,71°03W 8 Mar 73 33 60 42 -6.4 0.0 C 11 39°20N,70°00W NA NA 2,640 10 100 500 1,000 2,000 -13.0 -5.7 -3.7 -3.5 -1.6 -0.6 1.1 -0.6 0.3 -0.1 D 12 40o25N,73°28W 22 Mar 74 59 89 23 2 20 -0.7 -0.6 -3.5 -0.7 E 13 40°16N,73°13W 25 Feb 75 25 Feb 75 29 Apr 75 25 Feb 75 111 37 48 111 38 2 25 23 37 -4.7 -3.6 -5.5 -1.5 -4.8 -2.2 -3.9 -0.3 E 14 40°06N,72°54W 25 Feb 75 25 Feb 75 25 Feb 75 112 62 112 48 10 26 40 -2.4 -2.8 -1.6 -2.6 -1.1 -1.3 E 16 40°03N,72°42W 1 Mar 75 1 Mar 75 1 Mar 75 59 59 108 59 2 27 42 -2.8 -3.1 -2.9 -2.8 -2.8 -1.9 E 17 39°39N,72°38W 24 Feb 75 23 May 75 24 Feb 75 24 Feb 75 30 Apr 75 24 Feb 75 64 25 64 64 42 64 76 2 2 26 41 42 75 -2.6 -5.4 -4.9 -4.0 -6.2 -2.6 -1.7 -9.1 -0.7 -0.6 -7.4 1.0 E 18 39°28N,74°15W 1 Jul 72,73,74 1 Dec 73,74 1 Jul 72,73,74 1 Dec 73,74 60 60 60 60 12 5 5 10 10 -2.1 -2.0 -1.7 -1.3 -2.9 -2.8 -2.3 -1.8 F 15 40°07N,72°51W 18 Jun 74 18Jun74 18 Jun 74 18 Jun 74 35 50 2 13 26 46 -3.8 -6.8 -4.1 -1.7 -6.9 -5.5 -3.0 -1.8 E 42 Physical oceanography 29 Table 2. Continued Record Water Instr. Sta. Start length depth depth E N Data No. Location tone (da\si ( m ) (m) (em's) (cm/s) source* 19 38049N,74o12\V 29 Oct 74 36 43 9 29 Oct 74 23 29 Oct 74 35 20 37°55\,74°39\Y 26 . Tun 74 22 35 24 21 36°50X.75°42\V 21 Jul 74 29 16 4 15 22a 36o50\",75°02\Y 21 Jul 74 37 36 9 21 Tul 71 20 21 Tul 7 1 30 b 36°50X,75002\V 15 Tan 74 29 36 7 15 Tan 74 20 15 .Tan 74 32 23 36°o0X,74°48\V 21 Jul 74 26 70 11 21 Tul 74 30 21 Jul 74 58 24 36°50\.74o40YV 21 Jul 74 18 70 76 21 Jul 7 t 104 6.2 -3.2 G 4.9 -1.0 3.0 2.8 5.7 -7.5 II 1.7 1.5 11 0.2 3.1 2.4 -8.4 II 2.6 -5.5 2.4 -1.3 2.6 -8.6 1.6 -6.9 0.7 -4.7 3.7 -10.7 H 2.3 -16.6 0 -13.8 3.1 -12.6 H 1.4 -6.6 * A-Beardsle> and KlaRg (1976): B-Schmitz ill)74>. C-Beardsle> and Butnvan (1974); D-\Vebster (1969); E- N'OAA-MESA I in prep.); F-EGfcG i in prep.); (• Boicmirt i personal i imitation); H-Boicourt and Hacker (1976). .simultaneous measurements will provide 100-m isobath, while 60 times the river run- conclusive evidence, off, is only about 0.3' < of the northward If the fluxes through the three transects transport of the Gulf Stream. The volume are approximately the same, we postulate of the shelf water within the 100-m isobath that there is little net flow between the is estimated to be VSi, — 6,000 km:! ( Ketchum shelf and slope regions; Wright (1976) esti- and Keen 1955; Wright and Parker 1976) mated that as much as 2,000 knv'Vyr might and the estimated alongshore mean flux of leave the shelf region off New England via shell water within the 100-m isobath is the "calving" process. This number, how- TS|, 8,000 knv'/yr. The mean residence ever, was determined on the basis of a much time is then r = \*si,/TSi, - !4 yr. This im- larger alongshore gradient in transport than plies that the shelf water between Cape we observed. Hatteras and Cape Cod is removed from The various volume fluxes for the mid- the shelf and entrained into the Gulf Stream Atlantic Bight are shown schematically in in less than a year. This estimate is slightly Fig. 3. For comparison, note that the along- less than the 1.3 years estimated by Ket- shore transport of shelf water within the chum and Keen ( 1955 ) who knew the fresh- water inflow and the salinity distribution on the shelf and who assumed no flow Table 3. Alongshore transport to the 100-m iso- t,nterin„ ()r leaving the bight via the Nan- bath estimated through three transects across mid- , , P1 , , S TT ^ Atlantic Bight. Position of individual transects tucket Shoals and Cape Hatteras. shown in Fig. 2. The Gulf of Maine and Georges Bank region must supply the low salinity water observed flowing westward through tran- sect I, i.e. Tsii T,;M + T,;n following the notation in Fig. 3. This conclusion is im- plicit in the hydrographie structure of the shelf water in this region, namely that the shelf-slope water salinity front is a persis- tent and continuous feature from New York to the southern flank of Georges Bank and 43 Cross- sectional area tn Transport 100-m to 100-m Mean isobatli isobath speed A T ii = TA Transect ( km- ) ( knv1 'yr ) ( cm ;s ) Period i 6.4 5.300 2.7 Mar 74 ii 7.6 8.800 3.7 \1 ir— Apr 75 in 3.6 8,200 7.2 J" 1-Aug 74 30 Physical processes **->- *~1 ,pCAPE TGM C ~~~^^ y If J 1 'GEORGES /' BANK rJ Tr ■— 1 , ' • ( "%.(I>\ \ . Job ^y / * \ % jiW ' ■ -V- --*"' J •J- . * r vN •/ S / ^ y ^" r S" s VF/ i ! " JSh(ni) / - Tgs ^ CAPE HATTERA H/ s" / // y ••'■'£' ' Tr = 125 km' /,r / / TSh ■ 8000 km3 /yr VSh ' 6000 km' 'y r = Tsh " 4 y |TSL S 2000 kr nVy, Note 60 TR;TSh; 0 3 % TGS Fig. 3. Schematic diagram of the important vol- ume transports for the Middle Atlantic Bight. Tr is the total annual freshwater runoff ( of which over 50% occurs via the Chesapeake Bay), Tsh is the alongshore transport over the shelf out to the 100-m isobath, Tsl is the net flux of slope water into the shelf water, Tgs is the transport of the Gulf Stream, and Tgm and Tob are the unknown fluxes of shelf water from the Gulf of Maine and southern flank of Georges Bank. The volume of the shelf water mass out to the 100-m isobath is Vsh, and the average residence time r is simply Vsh/Tsh. the "cold pool" is also continuous during spring and summer along this same section of the shelf (see Bumpus 1976). Any sub- stantial flux of more saline slope water oc- curring across the 100-m isobath must be balanced by an increased flux of low sa- linity water (above TSh) from the Gulf of Maine and Georges Bank to maintain a steady salt balance. This conclusion is also dictated by simple continuity arguments which require a northern source region to maintain the observed westward flux of shelf water shown in Fig. 2. The summer current measurements in transects II and III show that the along- shore currents in the cold pool water equals or exceeds the mean southward current of the surrounding warmer water. These mea- surements counteract the traditional im- pression that the cold pool, formed by win- ter cooling, remains stationary throughout the spring and summer seasons (Ketchum and Corwin 1964). There is good evidence (Ford et al. 1952; Boicourt 1973) that the cold pool moves southward and is entrained by the Gulf Stream. High alongshore veloci- ties of the cold water, as measured in tran- sects II and III, imply that the cold water found near Cape Hatteras in August must have formed by winter cooling near Cape Cod or perhaps in the Gulf of Maine. Two large unknowns in the calculation of water and salt budgets in the mid-Atlantic Bight are fluxes of water and salt into the region from the north and amounts of water and salt exchanged across the shelf-slope boundary. Although we cannot yet quantify shelf-slope exchanges, we can describe some processes involved. In summer and winter, much exchange appears to be wind controlled, with onshore-offshore flows in the upper Ekman layer compensated by opposite flows in the lower layer ( Boicourt and Hacker 1976). In winter the cross-shelf flows driven by northeast winds enhance the thermal front at the shelf break and vertically mix the midshelf region. Winds from the south and southwest, on the other band, cause offshore flows in the upper Ekman layer and intrusions of warm salty slope water along the bottom, thereby tend- ing to stratify the outer shelf region. Summertime cross-shelf circulation is larger and has a more complex vertical structure. Boicourt (1973) and Boicourt and Hacker (1976) found that southerly winds can drive an intrusion of high salinity slope water onto the shelf at middepths in the southern mid-Atlantic Bight. Because these intrusions have been commonly ob- served on the outer shelf, they may be an important process in shelf water-slope water exchange. Gordon et al. (1976) ob- served a high salinity layer at middepth in the New York Bight, indicating that such intrusions may occur widely in the bight. The cold pool and strong thermocline are evident in the water temperatures in the southern mid-Atlantic Bight (Fig. 4). The salinity distribution shows an intrusion of high salinity slope waters in the upper ther- 44 Physical oceanography 31 2 3 4 5 6 7 70 6 84 9 94 10 Fig. 4. Distributions of temperature, salinity, and (Tt in a cross-shelf vortical section off Ocean City, Maryland, July 1975. mocline. This intrusion extends about 30 kin inshore of the slielr break, apparently driven by southerly winds. A small parcel of cold (<8°C), low salinity water may have been detached from the cold pool and moved off- shore. Because such parcels are commonly found in this position, however, the amount of water actually detaching is uncertain. South of New England, Bigelow ( 1933 ) and Cresswell (1967) described calving of the cold pool with parcels or bubbles of shelf water moving into slope water. Wright (1976) suggested that significant inter- change of shelf and slope water may occur via this mechanism. This process may be related to the formation of anticyclonic Gulf Stream eddies and their subsequent south- west drift along the edge of the slope. Satel- lite infrared photographs (e.g. Hughes 1975) suggest some exchange of shallow surface water, and Saunders' (1971) aerial temperature survey of one warm-core eddy suggests that some deep shelf water is pulled off the shelf and entrained into the trailing side of the eddy. How much shelf water is exchanged via these processes and 45 32 Physical processes with what frequency (i.e. the intermittency of these processes) is not known. We conclude this section with a brief dis- cussion of the physical processes that gov- ern the mean circulation in the mid-Atlantic Bight. Stommel and Leetmaa (1972) have constructed a theoretical model ( with linear dynamics) for the winter shelf circulation driven by a mean wind stress and a dis- tributed freshwater source at the coast. They then applied this model to the bight and concluded that an alongshore sea level slope of about 10 cm drop from Cape Cod to Cape Hatteras must exist to drive the mean flow toward the southwest (as ob- served! ) against the mean eastward wind stress. This same basic conclusion was also reached by Csanady ( in prep. ) who ex- amined the influence of wind stress vari- ability on the Stommel and Leetmaa model. This inferred alongshore pressure gradient can be either created by a succession of long, shore-trapped waves as suggested by Csanady ( in prep. ) , who showed evidence for this process in Lake Ontario, or main- tained by an upstream source of fresh shelf water, presumably here the St. Lawrence system and inshore Labrador Current. Sut- cliffe et al. (1976) reported evidence that fluctuations in the transport of the St. Law- rence system can be traced down the Scotian Shelf and into the Gulf of Maine. This, together with our early point that most of the fresh shelf water observed flowing westward through transect I (Fig. 2) must be supplied by the Gulf of Maine and outer Georges Bank regions, suggests a continuous freshwater pathway from the St. Lawrence to Cape Hatteras. The along- shore pressure gradient inferred to occur over the mid-Atlantic Bight may be par- tially supported by a northward rise in sea level found by oceanic leveling in the slope water by Sturges ( 1974) . Special features of the New York Bight and adjacent nearshore zone The New York Bight contains several fea- tures of general interest that have been in- tensely studied. Special topographic fea- tures of this region include a relatively deeply incised inner shelf region into which enters one of the major river systems of the region and the Hudson Shelf Valley and Hudson Canyon. The Hudson-Baritan estuary has a char- acteristic circulation consisting of a sea- ward flow of relatively brackish estuarine water in the near-surface water, and a shoreward flow of more saline water near the bottom. The relatively great width and complicated channel system in the Sandy Hook-Bockaway transect allows inertial and Coriolis effects to further modify cur- rents such that seaward flow tends toward the southern side of the entrance, and the inflow occurs mainly in the navigation channels and along the northern side of the entrance (see Parker et al. 1976). This mean flow of a few centimeters per second is a weak residual superimposed on stronger tidal flow but causes most of the material exchange between the estuary and shelf re- gions. The Hudson Shelf Valley is the offshore expression of the Hudson estuary. Current measurements in this valley (30 km off the New Jersey shore ) indicate that the average flow in the valley over intervals as long as a month can be shoreward with an average speed of a few kilometers per day. Such flows are more than ample, if coherent in space, to return suspended materials to the harbor entrance from far out on the con- tinental shelf. The combination of the Hudson estuary, the complex bottom topography, and the nearly right-angle bend in the shoreline produces quite complicated flow patterns over the inner shelf. There is evidence in the water properties that the near-surface flow from the estuary tends to move south- ward along the New Jersey shoreline. Be- covery of seabed drifters suggests the sta- tistical occurrence of a mean clockwise circulation within the inner bight, counter to the flow over the shelf farther offshore. This circulation is sometimes reflected in current measurements (Charnell and Mayer 1975 ) , but the current regime is best described as more dispersive than advective, especially during spring and summer, the seasons of maximum stratification. 46 Physical oceanography 33 An interesting and significant aspect of tiie flow in the inner bight is a shoreward velocity component in the bottom boundary layer. Numerous current measurements have been made for the NOAA-MESA pro- ject at distances of 1-5 m above the bottom. Averages of such measurements over any significant time frequently show a distinctly shoreward component. In 19 out of 21 cases examined in which a clear distinction could be made, there was a shoreward component in the bottom boundary layer. Furthermore, subdividing the data into sets in which the flow is east or west along Long Island, for instance, yields the same result: flow in the bottom boundary layer is shoreward in both cases. It is not yet ascertained whether this shoreward veering is a result of surface winds or whether it may be a manifestation of estuarine circulation generally over the shelf, but in any case it suggests a tendency for near-bottom materials to be carried in- shore. Such a process is a plausible explana- tion for the relatively high and constant rate of return of seabed drifters from bight waters ( Charnell and Hansen 1974) and supports previous reports ( Bumpus 1973). Some remaining problems Although progress has been made in de- termining the current variability and cir- culation pattern over the mid-Atlantic shelf, we are still unable to provide unambiguous answers to many questions of a basic en- gineering sort posed by environmental man- agers. Only a general estimate of the flush- ing rate of the shelf is available, and critical evaluation of the importance of the shelf- break exchange is not yet possible. Al- though a first-order description of flow to be expected can now be given for main parts of the bight, our ability to predict de- tails and events remains poor. The domi- nant forces controlling the circulation are believed known but their relative impor- tance and region of influence are not. Neither conceptual nor observational tools are adequate to the task for modeling of other than tides and tidal currents. Local models have useful applications but must be posed very carefully (especially bound- ary conditions) in the context of what is and what is not known about the physics of water movement over the shelf. It cannot be safely assumed that the way to solve a given management problem will be pointed by a mathematical model in any straight- forward sense. Finally, there remain funda- mental questions related to smaller scale phenomena, especially mixing and other dissipative processes. Smaller scale topo- graphic features like the inner New York Bight embayment, the ridge and swale areas, and the shelf valleys and submarine canyons must exert some steering influ- ences on the local flow. Some of these smaller scale problems will be immediately addressable when the physics of shelf cir- culation are better known; others must await improvement of observational instru- ments and techniques. References \i>el, J. H.. II. M. Byrne, J. R. Prom, and R. L. Charnell. 1975. Observations of oceanic internal and surface waves from the Earth Re- sources Technology Satellite. J. Geophys. Res. 80: 865-881. ' Hi ariislev. R. C, and B. Butman. 1974. Cir- culation on the New England continental shelf: Response to strong winter storm. Geo- phys. Res. Lett. 1: 181-184. . and C. Flago. 1976. The water struc- ture, mean currents, and shelf water/slope water front on the New England continental shelf. Proc. 1975 Liege Colloq. Ocean Hy- drodynam., in press. Bk.elow. 11. B. 1933. Studies of the waters on the continental shelf. Cape Cod to Chesapeake Ray. 1. The cycle or temperature. Pap. Phys. Oc'eanogr. Meteorol. 2(4): 135 p. , and M. Sears. 1935. Studies of the waters on the continental shelf, Cape Cod to Chesapeake Bay. 2. Salinity. Pap. Phys. Oceanogr. Meteorol. 4(1): 94 p. Boicouht, \V, 1973. The circulation of water on the continental shelf from Chesapeake Bay to Cape Hatteras. Ph.D. thesis, The Johns Hop- kins Univ., Baltimore. 183 p. , and P. Hacker. 1976. Circulation on the Atlantic continental shelf of the United States, Cape Mav to Cape Hatteras. Mem. Soc. R. Sci. Liege Ser. 6 10: 187-200. Brown, W„ W. Munk, F. Sxoix.rass, H. Mofjeld, and B. Zetler. 1975. MODE bottom ex- periment. J. Phys. Oceanogr. 5: 75-85. Bumpus, D. F. 1969. Reversals in the surface drift in the Middle Atlantic Bight area. Deep-Sea Res. 16( suppl. ) : 17-23. 47 34 Physical processes — . 1973. A description of the circulation on the continental shelf of the east coast of the United States. Prog. Oceanogr. 6: 1 11— 157. — . 1976. Review of the physical oceanog- raphy of George's Bank. Int. Comm. N. Atl. Fish. Res. Bull. In press. and E. L. Pierce. 1955. The hydrogra- tal shelf south of Long Island, New York. Limnol. Oceanogr. 9: 467-475. and D. J. Keen. 1955. The accumula- phy and the distribution of chaetognaths over the continental shelf off North Carolina. Deep-Sea Res. 3(suppl.): 92-109. Charnell, R. L., and D. V. Hansen. 1974. Summary and analyses of physical oceano- graphic data collected in the New York Bight apex during 1969 and 1970. MESA Rep. 74-3. NOAA-ERL. 44 p. , and D. A. Mayer. 1975. Water move- ment within the apex of the New York Bight during summer and fall of 1973. NOAA Tech. Memo. ERL MESA-3. 29 p. Cresswell, G. M. 1967. Quasi-synoptic monthly hydrography of the transition region between coastal and slope water south of Cape Cod, Massachusetts. Woods Hole Oceanogr. Inst. Tech. Rep. Ref. 67-35. 114 p. EG&G. 1975. Summary of oceanographic obser- vations in New Jersey coastal waters near 39°28'N latitude and 74°15'W longitude dur- ing the period May 1973 through April 1974. Rep. B4424, EG&G Environ. Consultants, Waltham, Mass. Flagg, C, J. Vermersch, and R. Beardsley. 1976. Report on the 1974 MIT New England shelf dynamics experiment, Part 2; the moored array. Dep. Meteorol. Mass. Inst. Techno!., Lab. Rep. (Unpublished.) Ford, W. L., J. R. Longard, and R. E. Banks. 1952. On the nature, occurrence and origin of cold low salinity water along the edge of the Gulf Stream. J. Mar. Res. 11: 281-293. Gordon, A. L., A. F. Amos, and R. D. Gerard. 1976. New York Bight water stratification- October 1974. Am. Soc. Limnol. Oceanogr. Spec. Symp. 2: 45-57. Hansen, D. V. 1970. Gulf Stream meanders be- tween Cape Hatteras and the Grand Banks. Deep-Sea Res. 17: 495-511. Hughes, P. 1975. The view from space. En- vironmental Data Service, January Bull., NOAA. Iselin, C. O'D. 1939. Some physical factors which may influence the productivity of New England's coastal waters. J. Mar. Res. 2: 74-85. . 1955. Coastal currents and the fisheries. Deep-Sea Res. 3(suppl.): 474-478. Ketchum, B. H., and N. Corwin. 1964. The persistence of "winter" water on the continen- tion of river water over the continental shelf between Cape Cod and Chesapeake Bay. Deep-Sea Res. 3(suppl.): 346-357. Parker, J. H., I. W. Duedall, H. B. O'Connors, Jr., and R. E. Wilson. 1976. Raritan Bay as a source of ammonium and chlorophyll a for the New York Bight apex. Am. Soc. Lim- nol. Oceanogr. Spec. Symp. 2: 212-219. Pollard, R. T., and R. C. Millard, Jr. 1970. Comparison between observed and simulated wind-generated inertial oscillations. Deep-Sea Res. 17: 813-821. Reid, R. O. 1958. Effect of Coriolis force on edge waves (I) investigation of the normal modes. J. Mar. Res. 16: 109-141. Saunders, P. M. 1971. Anticyclonic eddies formed from shoreward meanders of Gulf Stream. Deep-Sea Res. 18: 1207-1219. Schmitz, W. J. 1974. Observations of low-fre- quency current fluctuations on the continental slope and rise near site D. J. Mar. Res. 32: 233-251. Stommel, H., and A. Leetmaa. 1972. Circula- tion on the continental shelf. Proc. Natl. Acad. Sci. 69: 3380-3384. Sturges, W. 1974. Sea level slope along conti- nental boundaries. J. Geophys. Res. 79: 825- 830. SUTCLIFFE, W. H., R. LOUCKS, AND K. DRINK- water. 1976. Considerations of ocean cir- culation and fish production on the Scotian Shelf and in the Gulf of Maine, Pt. 1. Ocean circulation and physical oceanography. J. Fish. Res. Bd. Can., in press. Thompson, R. 1971. Topographic Rossby waves at a site north of the Gulf Stream. Deep-Sea Res. 18: 1-19. Webster, F. 1969. Vertical profiles of horizon- tal ocean currents. Deep-Sea Res. 16: 85-98. Wricht, W. R. 1976. The limits of shelf water south of Cape Cod, 1941 to 1972. J. Mar. Res. 34: 1-14. , and R. Hendry. 1972. Array measure- ments of the bottom boundary layer and the internal wave field on the continental slope. Geophys. Fluid Dynam. 4: 101-145. -, and C. E. Parker. 1976. A volumetric temperature/salinity census for the Middle At- lantic Bight. Limnol. Oceanogr. 21: 563- 571. Wunsch, C. 1972. Bermuda sea level in relation to tides, weather, and baroclinic fluctuations. Rev. Geophys. Space Phys. 10(1): 1-49. 48 9 Reprinted from: NOAA Data Report ERL MESA-18, 220 p. ABSTRACT During April 1974, two oceanoqraphic cruises were made by the NOAA Ship Researcher in the New York Bight. The cruises were used for deployment and recovery of three bottom-mounted pressure gauges and to collect physical and chemical oceanographic data from the water column. Thirty-one oceanographic stations were occupied on a seg- ment of the continental shelf bounded on the east by Block Island, on the south by Cape May, and extending outward to the edge of the continental shelf. This report presents the corrected water column data from these two cruises and describes the measurement methods and corrections applied to the data. 49 10 Reprinted from: Estuarine and Coastal and Marine Science^ Vol. 4, 309-323, A Two-dimensional Numerical Model of Estuarine Circulation: The Effects of Altering Depth and River Discharge John F. Festa and Donald V. Hansen Atlantic Oceanographic and Meteorological Laboratories, 15 Rickenbacker Causeway, Virginia Key, Miami, Fla 33149, U.S.A. Received 22 December 1974 and in revised form 2 s June 1975 Steady-state numerical solutions are obtained for a two-dimensional, vertically stratified model of a partially mixed estuary. The boundary at the seaward end of the estuary is considered to be open, with the profiles of salinity, vorticity and streamfunction obtained by extrapolating interior dynamics out to the boundary. A salinity source is maintained at the bottom at the mouth. Zero salt flux is required at a free-slip top and no-slip bottom boundary. Zero salinity and a parabolic velocity profile are maintained at the head of the estuary. A number of cases are run for various estuarine parameters; the river transport and Rayleigh number being the two parameters that have the most pronounced effect. The river transport is varied by adjusting the mean freshwater velocity, Uf. Decreasing Us allows salt as well as the stagnation or null point to penetrate upstream. The estuarine circulation weakens, but expands over a larger portion of the estuary. The position of the stagnation point, with respect to the seaward boundary, varies as £/f_6/8 for Uf>i cm/s and as U( ~5/6 for Uto(I+^)' 's assumed and the Boussinesq approximation (Spiegel & Veronis, i960) is employed. The horizontal and vertical momentum balances, continuity of flow and conservation of salinity are: ut+uux+wuz = -/?0~1Px+(^h^)x+(^vwz)z, (la) wt+iavx+zewz = -p0"1PzMAh^x)x+(Avwz)z-/]gS, (ib) ux+wz = o, (ic) St +uSx+toS, = (KhSx)x+(KyS:)z. (id) where u and w are the horizontal and vertical components of velocity respectively, P is the hydrostatically reduced pressure, S is the salinity field, /? is the coefficient of 'salt contraction', p0 is the density of fresh water, Ah, Ay, and Kh, KY are the horizontal and vertical exchange coefficients of momentum and salt, respectively, and g is the gravitational acceleration. 51 Two-dimensional circulation model 311 Tidal fluctuations have been averaged out; however, the tides are considered to be the primary source of energy for turbulent mixing. The exchange coefficients therefore repre- sent a measure of the strength of tidal mixing. For simplicity, these coefficients are chosen to be constant. The vorticity and salt equations corresponding to (1) are: rjt = -J(V, l)+AriXx+Ayr!zz—PgSx, (2a) St = -j(ys,S)+KhSxx-{-KvSzz. (2b) where y/ is a streamfunction with 11 = pVxj'O being a unit vector in the -\-y direction), rj = A22y/ is the vorticity, J is the Jacobian and A22 is the two-dimensional Laplacian operator. Non-dimensional equations, corresponding to equation (2), are obtained by scaling t and 77 by rd = H2/Kv and ra~\ respectively, x and z by H, y/ by Ky and 5 by ASh. rd is the vertical diffusive time scale, H is the depth of the estuary, and ASh is the horizontal salinity difference between the river and mouth of the estuary. In non-dimensional form the vorticity and salt equations are: It = -%¥, ri)+o{Atixx+rjzl-RaSx), (3a) St = -J(W,S)+KSXX+SZZ, (3b) where // = A2y/ = Vxx^Wzz, Ra ~ PgdSh H3/(AVKW) is the estuarine Rayleigh number, a = Av/Kv is the Prandtl number, A = AhjAy and K = KJKV. Non-dimensionalizing both horizontal and vertical distances by the estuarine depth, H, while arbitrary, forces an aspect ratio, s = H/L, to enter only through the boundary conditions. Here, L is defined as the computational length of the model estuary, that is, the location of the upstream bound- ary. This length should not be confused with the dynamical length of the estuary, Ld, roughly equivalent to the extent of salinity intrusion. The determination of the dynamical length is a major object of analysis. Boundary conditions The boundary conditions to be satisfied at the river end are zero salinity and a parabolic velocity profile (consistent with constant density and viscosity) having a transport per unit width TT = UtH, where Ut is the vertically averaged river flow per unit width. At the bottom boundary, a no-slip condition and zero vertical flux of salt are specified. At the top boundary, a free-slip condition and zero vertical flux of salt are specified. These are expressed by: 5=o, y/{z) = i-^R(z2—z3/2), and t]{z) = t,R{i—z) at x = e_1 Sz = o, if/ — o and y/z = o at z = o, (4) Sz = o, y/ = R and 77 = 0 at z = 1, where R = TT/Ky is the non-dimensional river transport. Inclusion of non-zero wind stress at the surface is an easy modification to the model, but is not pursued herein. The remaining boundary conditions to be considered are those at the mouth of the estuary, x = o. These are perhaps the most difficult part of the model and will be discussed at some length. Estuaries empty either into a larger bay or directly onto a continental shelf (see Figure 1). They usually widen abruptly, allowing geometrical and rotational effects to become import- ant. A two-dimensional model is no longer appropriate. To investigate estuarine dynamics in its simplest form, attention must be focussed landward of this outer legion. The inshore limit of this region is herein considered to be the mouth of the estuary. Salinity and velocity distributions at the mouth are functionally dependent upon river flow, depth, horizontal density difference and other parameters. The surface layers become 52 312 J. F. Festa & D. V. Hansen Figure i. An idealized estuarine system. fresher as the river transport increases. The estuarine circulation becomes stronger for increasing Rayleigh numbeis. Consequently internal dynamics determine seaward boundary profiles as well as those within the estuary. The boundary conditions given at the seaward end of the model must be consistent with these internal dynamics. Thus, salinity and velocity profiles cannot be specified as boundary conditions. Preliminary numerical experi- mentations support this result, since unrealistic seaward boundary layers occur where salinity and velocity profiles are specified as seaward boundary conditions. Experimentation also showed that unless a source of salt in the form of a definite salinity value is specified somewhere in the region, the solution S = o is obtained. Observations suggest that, although the salinity distribution everywhere within estuarine regions is strongly influenced by varia- tion of river discharge and other parameters, the salinity of the deep water near the seaward boundary is least influenced. We have therefore made salinity at the bottom of the seaward boundary invariant, >S(0,0) = i, to assure estuarine behavior. In order to obtain the seaward boundary conditions, attention is focused on the dynamics near the estuarine mouth. In the vicinity of the seaward boundary for the model it is expected that the estuarine circulation is relatively well developed. Pritchard (1954, 1956) has shown that in this situation the salt balance is maintained primarily by a dynamic balance between horizontal advection and vertical diffusion of salt and a vorticity balance is maintained primarily by a balance between buoyancy forces due to horizontal density gradients and vertical diffusion of vorticity. Horizontal diffusion of salt, especially, while shown by Hansen & Rattray (1965) to be essential to the overall estuarine regime, does not appear to be locally important where the gravitational circulation is well developed. In addition, horizontal diffusion of vorticity and horizontal shear in the vertical velocity field are also assumed to be locally unimportant. These conditions are: and t]xx = o, at x = o ¥xx - o. 53 (5) Tzvo-dimensional circulation model 313 Thus, horizontal diffusive fluxes of salt and vorticity are required to be constant, but unspecified, at the open boundary. Although we are unable to provide a completely rigorous justification of these conditions, they do provide a means of completing the mathematical specification of the problem, without inducing boundary layer behavior near the seaward boundary. Numerical formulation and procedures A finite-difference grid is chosen to be uniform in x and z, such that xt = iAx, i = o, 1, / ] Zj = jAz, j = 0, i, J (6) t" = nAt, n — o, 1, j where Ax = (sl)~1, Az =J~X and At is the time step whose magnitude depends upon the stability of the differencing scheme that is chosen. The Laplacian operator is approximated by the usual five-point difference scheme. The advection of salt and vorticity, expressed in terms of the Jacobian, J, is approximated by using the % and J3 forms of Arakawa (1966), respectively. These conserve salinity and salinity squared and vorticity and kinetic energy. Diffusion is approximated by the time-centered scheme of DuFort-Frankel (1953). The resulting finite difference analogs to (3) are: faf ' - tffl)l(zAt) = -J\(y,,n)l(4A xAz) + aAAx~%rj1+ „ + tft_ „ - tfi+ ' - rftf «) -aRa(S1+li-S^j)l(2Ax), (7) (Sf ' - Slr')l{2At)= -j\{¥,S)!{±AxAz) + KAx-\S1+Xj + SUj-Sl^- S«fl) + Az-\S"ini+S1]_x~ S$+i- STf1). (8) The streamfunction at the latest time step, y/"j+i, is calculated by means of a direct solution (Buzbee et ah, 1970) to the finite difference Poisson equation, nnu+i = ^x-^i+xi + w7-ii-2Vu+l) + **-%Vu+\ +rt±\-Wij+1)- (9) In finite difference schemes, the size of the time increment, At, is often limited by two stability requirements. The first of these is the advective stability condition, the Courant- Freidricks-Lewy criterion, which requires that for a given grid spacing, S, AtVm/3 vRa(—4zf+$Zj- 1)/8, vRa(— 2Zj5-\-6zJi— 5Zj3)j 240 +R(-zj*+iz/-zf)18 (") and where v is a function of Ra, R and K. Finite difference boundary conditions The computational grid is extended one grid point beyond the bottom, top and mouth of the estuary. This allows the finite difference representation of derivatives at the boundary to be consistent with interior computations. Values of S, if/ and tj outside the boundaries are defined as (1) Bottom (j = o) 5?,. y/"j+l; {u = o). Wi)-\ (2) Top(j = J) Wu+i = 2W"j - Wv-u ("z = °)- (3) Mouth (z = o) S'i-\j — 2S(j — S"+\j'i(Sxx = °)> Wi-u = 2¥ij ~ Vt+u\ (Vxx = °). fJi-lj = 2*1 u - >/< + l/> (f!xX = o). (12) 55 Two-dimensional circulation model 315 Solutions at the boundaries are obtained by substituting these values into the difference equations, (7), (8) and (9). The bottom vorticity is evaluated using the first order Taylor series approximation developed by Bryan (1963): 7?o+1 = ayii+IM*a- (13) At the mouth of the estuary, a three point forward difference scheme for the salinity gradient, S"x\ ,_„ = (2AV)-1 (-355, + 4S7,. - S"2j), (14) is used in equation (7) to calculate vorticity boundary values. We double integrate the boundary vorticity field by means of a Gaussian-elimination procedure to calculate y/ at the seaward end of the estuary. Given y/ on all boundaries, we invert Poisson's equation to obtain the interior streamfunctions. Salinity and velocity profiles obtained by this method are slightly closer to the similarity solution than are those obtained by using the central difference representation of the salinity gradient. A third method, in which the boundary vorticity is simply extrapolated out from the interior, «nof1 = 2T,l+1-TI°2tl, (15) also produces acceptable solutions. Simple extrapolation, however, does not work well when calculating the salinity dis- tribution at the mouth. We have found that simple smoothing produces significantly lower values for the salinity throughout the estuary. Smaller horizontal salinity gradients occur and as a consequence lower values of the streamfunction and velocity fields result through- out the interior. The full salinity equation (8) must therefore be used for the calculation at the open boundary. Discussion of results Steady-state solutions to the model equation were obtained using a 33X33-point finite difference grid. Initially, a 17-point vertical resolution seemed adequate; however, tests with similarity solutions indicated that truncation errors may produce significant differences between analytical and numerical values of y/ and 77. Results of computations not included here showed that variations of A between 1 and io6, and ct from 1 to io2 have a negligible effect upon the results. Prandtl number independence is common in thermal convection problems (Beardsley & Festa, 1972). Variation of K from 1 to io7 does produce significant change in the solutions, but this effect will not be fully explored here. The model was run for a range of parameters characteristic of, or centered on, nominal values typical of coastal plain estuaries for which data have been published. These nominal values are ASh = 30%0, K = io6 (Ky = 1 cm2/s), Ra = 3 X io9 (H = 10 m), and R = 2 X io3 (Us = 2 cm/s). Contours of the streamfunction and salinity distribution obtained are presented in Figures 3 and 8. The salinity distribution shows a stratified intrusion into the estuary, and the streamfunction shows the typically estuarine pattern of seaward flow of near surface water and a landward flow of deeper water. The vertical component of flow is of considerable interest in connection with estuarine problems, but is not directly, or in many cases indirectly, measurable. It is of interest there- fore, to explore the magnitude and structure of the vertical flow associated with the 56 316 J. F. Festa & D. V. Hansen conditions and parameter ranges of the model. The longitudinal variation of vertical velocity at three levels for H = 10 m and Uf = 2 cm/s is shown in Figure 2. The boundary conditions require the vertical flow to be identically zero on the top and bottom boundaries. Both the order of magnitude and the vertical structure of the vertical flow are consistent with the determination of vertical velocity in the James River estuary made by Pritchard (1954). The principal feature of this longitudinal variation is that at all levels a maximum occurs somewhat seaward of the stagnation point (intersection of the internal zero of the stream- function with the bottom) and a rapid falloff across the position of the stagnation point. 70 1 1 1 1 ! 1 1 1 1 60 - - 50 (/I '*' S^ \ \ £ 40 : -.-"" ^^^ \ \ O 30 — \ \\ 20 - \ V 10 1 1 1 1 1 1 1 ^^-^ 3a_„ 1 0 10 20 30 70 HO 90 100 40 50 60 *(km) Figure 2. Longitudinal variations of the vertical velocity field at z = 0-25, 0-5 and 0-75 for Ut = 2 cm/s. Ra = 3X io9 (H = 10 m), \0-8> I0'6 1 0-4 \ 0-2 \ 0-01 : : 1 1 1 ,1 , , 1, J ,1,1 1 • . . li 1 1 A i._ i , ! 120 km ry-r. : \ \ \ : K. N. \ \ Ui = 4 cm/s : ^s\\ \ \ : S\\\ ^k \ \ ■ V\\\V\\\ \ \ j \os\o-^0-4\ 0-2 ' 1 0-01 ! ill .L 1. 1.1 .1.1.. 60 km Figure 3. Salinity and streamfunction fields as a function of river flow, Uc. Salinity fields are contoured from o to 1 in intervals of o-i. The streamfunction fields are scaled by io3 and contoured from o in intervals of 0-4. Ra = 3 X io9 (H = 10 m), a = 10, K = io6 (Kv = 1 cm2/s). 58 3i8 J. F. Festa & D. V. Hansen 5(%„) 12 15 18 21 24 27 30 I ' ■ V T 1 \ h \ \ \ - \ \ \ \ \ \ N^ \ \ \ \ - \\\ " - X^V - ^V - 1 - (b) i i i 1 -20 -15 -10 -5 0 5 10 0-3 0-4 0-5 0-6 0-7 0-8 0-9 1-0 i/(cm/s) S Figure 4. Variations of (a) velocity and (b) salinity profiles at the seaward boundary, x = o, with river flow, Ut. Ut (cm/s) : , 1 ; , 2 ; , 4. nevertheless, they are quite different overall. The total landward transport into the lower layer is almost independent of the amount of freshwater discharged. This is a somewhat surprising result, but may be due in part to making the bottom salinity at the seaward entrance independent of river flow. Upstream attenuation of the landward flow is consider- ably more rapid for larger river flows. This model varies qualitatively from the similarity solutions in that the vertical profiles are not constrained to be similar throughout the estuary. Two features of particular interest are the length of the salinity intrusion into the estuary and the position of the stagnation point. The horizontal variations of salinity at the surface and at the bottom are shown in Figure 5. The longitudinal patterns are similar to the exponential forms given by Hansen & Rattray (1965), except that here we obtain what is not available as an analytic similarity solution: a complete transition to zero salinity. The length of the salinity intrusion, defined by the _! L 0 10 20 30 40 50 60 70 80 90 100 Xlkm) Figure 5. Longitudinal variations of surface and bottom salinity as a function of river flow, U,. Uf (cm/s): , 1 ; — , 2; , 4. 59 Two-dimensional circulation model 319 distance over which 5^o-i5%0 at the bottom of the estuary, is a strong function of fresh- water discharge. It increases from 44 km for Uf = 4 cm/s to 170 km for U( = 0-5 cm/s as shown in Figures 5 and 6. This behavior has been empirically, if qualitatively, known to hydraulic engineers for many years. The functional form of the dependence of salinity intrusion upon freshwater discharge is of special interest because a characteristic value of Uf in coastal plain estuaries is approximately 1 cm/s. It is apparent from Figure 6 that for the values of other parameters used, the length of salinity intrusion changes behavior in the 100 120 140 X(km) 200 220 240 Figure 6. Influence of river runoff on salinity intrusion and stagnation point location. , S = 0-05 ; — , stagnation point; , S = 0005. vicinity of Uf = 1 cm/s. Increases of U{ from 1 cm/s lead to modest retreat of salinity intrusion, but reductions of Uf result in greatly increased salinity intrusion. The implica- tion of this non-linear relationship for reduction of freshwater discharge into already critical estuaries is fairly obvious. Within the range of values explored, the salinity intrusion varies approximately as £/f~4/7 for C/f>i cm/s and as t/f_5/6 for Ufi cm/s, the position of the stagnation point varies approximately as C/f_5/8, and as C/f_5/6 for Uf ' ft\ 0-9 i\\ \\ v^ 0-8 A \ , \ V 0-7 • \ \ \ , \ \ \ 0-6 V^ \ \ ->-_ 0 10 20 30 40 50 60 70 80 90 100 110 120 * (km) Figure 10. Longitudinal variations of surface and bottom salinity as a function of depth, H. H: , 75 m; , 10 m; , 12-5. 15-0 12-5 3 10-0 ~ 7-5 - 5-0 - S ** 1 1 - - 4* - _J 1 1 1 i_ 20 40 60 80 100 120 140 *(km) Figure 11. Influence of estuarine depth on salinity intrusion and stagnation point location. , S1 = 0-05 ; , stagnation point; , S = 0-005. 63 Two-dimensional circulation model 323 t 10 X(Wm\ Figure 12. Vertical velocity contours ats = Vertical velocity contours are scaled by 10" 0-5 as a function of estuary depth, H. ' cm/s with a contour interval of 20. increase in the magnitude of the vertical velocity as well as a seaward displacement of its spatial maximum. The caveat regarding dependence of particular values upon the choice of values for the exchange coefficients used in the model must also be accepted here, but the general behavior will be unchanged. Acknowledgements The progress of this research has benefitted from numerous discussions and criticism of the manuscript by Drs H. Mofjeld, A. Leetmaa and C. Thacker. The manuscript was prepared by Ms K. Phlips. References Arakawa, A. 1966 Computational design for long-term numerical integration of the equations of fluid motion. Two-dimensional incompressible flow. Part 1. Journal of Computational Physics I, 1 19-143. Beardsley, R. C. & Festa, J. F. 1972 A numerical model of convection driven by a surface stress and non-uniform horizontal heating. Journal of Physical Oceanography 2, 444-455. Bryan, K. 1963 A numerical investigation of a non-linear model of a wind-driven ocean. Journal of Atmospheric Science 20, 594-606. Buzbee, B. L., Golub, G. H. & Nielson, C. W. 1970 On direct methods for solving Poisson's equations. SIAM. Journal of Numerical Analysis 7, 627-656. DuFort, E. C. & Frankel, S. P. 1953 Stability conditions in the numerical treatment of parabolic differential equations. Mathematical Tables and other Aids to Computation 7, 135. Hamilton, P. 1975 A numerical model of the vertical circulation of tidal estuaries and its application to the Rotterdam Waterway. Geophysical Journal of the Royal Astronomical Society 40, 1-22. Hansen, D. V. 1964 Salt balance and circulation in partially mixed estuaries. Proceedings from the Conference on Estuaries, Jekyll Island. Hansen, D. V. & Rattray, M. Jr 1965 Gravitational circulation in straits and estuaries. Journal of Marine Research 23, 104-122. Harleman, D. R. F., Fisher, J. S. & Thatcher, M. L. 1974 Unsteady salinity intrusion in estuaries. Technical Bulletin No. 20, U.S. Army Corps of Engineers. Pritchard, D. W. 1954 A study of the salt balance in a coastal plain estuary. Journal of Marine Research I3» I33-I44- Pritchard, D. W. 1956 The dynamic structure of a coastal plain estuary . Journal of Marine Research 15, 33-42. Rattray, M. Jr & Hansen, D. V. 1962 A similarity solution for circulation in an estuary. Journal of Marine Research 20, 1 21-123. Richtmyer, R. D. & Morton, K. W. 1967 Difference Methods for Initial-value Problems. 405 pp. Inter- science, New York. Spiegel, E. A. & Veronis, G. i960 On the Boussinesq approximation for a compressible fluid. Astro- physics Journal 131, 442-447. Printed in Great Britain by Henry Ling Ltd., at the Dorset Press, Dorchester, Dorset 64 Reprinted from: Applied Optios, Vol. 15, No. 8, 1974-1979. Reprinted from Applied Optics, Vol. 15, page 1974, August 1976 Copyright 1976 by the Optical Society of America and reprinted hy permission of the copyright owner 11 Radiative transfer: a technique for simulating the ocean in satellite remote sensing calculations Howard R. Gordon A method is presented for computing the radiative transfer in the ocean-atmosphere system which does not require detailed knowledge of the optical properties of the ocean. The calculation scheme is based on the observation that the upwelling radiance just beneath the sea surface is approximately uniform, which implies that the effect of the ocean can be simulated by a lambertian reflector just beneath the sea surface. It is further shown that for aerosol concentrations up to ten times the normal concentration, the radiative transfer in homogeneous and vertically stratified atmospheres (of the same optical thickness) is nearly iden- tical. Examples indicating the applicability of these results to the remote sensing of ocean color from space are discussed in detail. Introduction It is now well established that the upwelling light field above the oceans can contain significant infor- mation concerning the oceanic concentration of sedi- ments and organic material. However, when the oceans are viewed from satellite altitudes, the increased ra- diance due to the intervening atmosphere has made quantitative interpretation (in terms of oceanic prop- erties) of the radiance observed by the satellite difficult. We can realize the full potential of oceanic remote sensing from space in the visible portions of the spec- trum only if we can learn to relate the radiance which reaches the top of the atmosphere to the optical prop- erties of the ocean. To effect this, the radiative transfer equation must be solved for the ocean-atmosphere system with collimated flux incident at the top of the atmosphere. In such calculations, the optical properties of the ocean which must be varied are the scattering phase function Pn(#) and the single scattering albedo o)(j (defined as the ratio of the scattering coefficient to the total attenuation coefficient). Furthermore, unless the ocean is assumed to be homogeneous, the influence of vertical structure in these properties must be con- sidered. To describe the cloud-free atmosphere, we must know the optical properties of the aerosols and their variation with wavelength and altitude as well as the ozone concentration. Considering the ocean for the present to be homogeneous, we can relate the radiance The author is with University of Miami, Physics Department, and Rosenstiel School of Marine & Atmospheric Sciences. Coral Cables. Florida 33124. Received 20 December 1975. at the satellite to the ocean's properties by choosing an atmospheric model and solving the transfer equation for several oceanic phase functions and u>o's at each wavelength of interest. The number of separate com- putational cases required is then the product of the number of phase functions, the number of values of u>o, and the number of wavelengths. Even if the multiphase Monte Carlo method (MPMC)1 is used, the wo resolu- tion of Gordon and Brown2 would require a number of simulations equal to ten times the number of wave- lengths for each atmospheric model considered. In this paper an alternate method of computation that does not require detailed knowledge of the ocean's optical properties is presented. Calculations From the work of Plass and Kattawar3-4 on radiative transfer in the ocean atmosphere system, it is seen that when the solar zenith angle is small, the upwelling ra- diance just beneath the sea surface is approximately uniform (i.e., not strongly dependent on viewing angle) and hence determined by the upwelling irradiance. It is possible for remote sensing purposes to utilize this observation in simulations of the transfer of radiation in the ocean-atmosphere by assuming that a fraction R of the downwelling photons just beneath the sea surface are reflected back toward the surface with a uniform radiance distribution, while the rest of the downwelling photons are absorbed. The ocean is then treated as if there were a lambertian reflecting surface of albedo R just beneath the sea surface. In this case, Gordon and Brown ' have shown that any radiometric quantity Q can be written Q = Qt + \Q,R/<\ - rR\\. (1) where Q] is the contribution to Q from photons that 1974 APPLIED OPTICS / Vol. 15. No. 8 / August 1976 65 Table I. Three Ocean Scattering Phase Functions 0 KA KB AT Cde.n) ( • 10") (\ 10) (V H)J) 0 1092 1 10171 9521 1 4916 4 577 4285 o 57 3.5 53 1.0 499.9 10 169.3 157.7 147.6 20 29.5 29.39 29.31 30 12.56 1 1.95 11.42 4 5 3.059 3.661 4. 1 89 60 1.092 1.57 7 1.999 i 0 0.546 0.915 1.190 90 0.34 ! 0.661 0.952 1 05 0. 3 1 1 0.611 0.928 120 0. 3 1 7 0.7 32 1.094 135 0.410 0.8 29 1.309 150 0.4 92 1.017 1.618 165 0.5 79 1.261 1.856 ISO 0.617 1.357 1.999 never penetrate the sea surface (hut may he specularly reflected from the surface). Q> is the contribution to Q from photons that interact with the hypothetical 1am- hertian surface once for the case R = 1, and r is the ratio of the number of photons interacting with the lamber- tian surface twice to the number of interacting once, again fori? = 1. I 'sing Eq. ( 1 ) any radiometric quantity can then be computed as a function of R. Physically the quantity R is the ratio of upwelling to downwelling ir- radiance just beneath the sea surface and is known as the reflectance function'5 [/?(().— )| in the ocean optics literature. Spectral measurements of the reflectance function i?(,\) have been presented for various oceanic areas by Tyler and Smith.' Henceforth in this paper R(\) will be referred to as the ocean color spectrum. A series of Monte Carlo computations has been car- ried out to see if an approximate simulation ( ASl ) using this assumption of uniform upwelling radiance beneath the sea surface yields results that agree with computa- tions carried out using an exact simulation (ES) in which the photons are accurately followed in the ocean as well as the atmosphere. The Monte Carlo codes used in Rets. 2 and 5 were modified by the addition of an atmosphere. The atmosphere consisted of fifty layers and included the effects of aerosols, ozone, and Rayleigh scattering, using data taken from the work of Klterman.8 The aerosol scattering phase functions were computed by Fraser9 from Mie theory assuming an index of re- traction of 1.5 and DeirmendjianV" haze C size distri- bution. Also, to determine the extent to which the vertical structure of the atmosphere influences the ap- proximate simulation, a second approximate simulation (AS2) was carried out in which the atmosphere was considered to be homogeneous, i.e., the aerosol scat- tering, Rayleigh scattering, and ozone absorption were independent of altitude. The oceanic phase functions in the ES are based on Kullenberg's11 observations in the Sargasso Sea and are given in Table I. KA is roughly an average of Kullenberg's phase function at 632.8 nm and 655 nm, and KC is his phase function at 460 nm. KB is an average of KA and KC. These phase functions show considerably less scattering at verv small angles 0 < 1° than observed by Petzold12 in other clear water areas; however, the exact form of the oceanic phase function is not very important, since it has been shown '•'' to influence the diffuse reflectance and /?(0,— ) only through the backscattering probability (B) R= 2; £ P„(tf) sinHdf). In all the computations reported here, the solar beam is incident on the top of the atmosphere from the zenith with unit flux. At visible wavelengths, the variable atmospheric constituent that will most strongly influ- ence the radiance at the top of the atmosphere is the aerosol concentration. (Plass and Kattawar14 have shown that for the range of expected variation of the ozone concentration, the radiance at 700 nm is essen- tially independent of the concentration.) Thus, we have carried out computations for aerosol concentra- tions in each layer of one, three, and ten times the nor- mal concentration given by Elterman. These aerosol models are henceforth labeled N, 'A X N and 10 X N. All computations presented here are carried out at 400 nm, the wavelength in the visible portion of the spec- trum where the atmospheric effects are expected to be most severe. Results A sample of the results is given in Table II where the upward flux at the top of the atmosphere for the AS cases is compared with the ES case for oceanic phase function KC and ujo = 0.8. The values of i? used to ef- fect the AC computations were taken from the EC computation of this quantity; however, if R is taken from H = 0.0(H)] + 0.3244.x + 0.1425*'- + 0.1308* ■'. (2) where x = u.-nft/[l - u.-o(] - B)\ which, according to Cordon rt o/.,1, reproduces the in-water reflection function for the corresponding case, but with no at mo- sphere present, the results of the AS computations agree with those listed to within 0.2%. The numbers in the parenthesis next to each flux value represent the sta- tistical error in the flux based on the actual number of photons collected in each case. It is seen that ES and AS simulations generally agree to within the accuracy of the computations. Notice also the excellent agree- ment between the ASl and AS2 fluxes. In Fig. 1 the comparison between the ES, ASl, and AS2 upward radiances at the top of the atmosphere is presented for the three aerosol models. The steplike curve in the figure is for ES, the solid circles for ASl, and the open circles for AS2, and n is the cosine of the Table II. Comparison of the Flux at the Top of the Atmosphere for the ES, ASl, and AS2 Simulations Aerosol concen- tration Model ES ASl AS2 .V 0. 22 2( •0.002) 0. 224 ( '0.001) 0. 226 0 0.001) x N 0. 27 1 ( •0.003) 0. 273 ( ■0.001) 0. 275 ( 0.001) ■ .V 0. 42 3( •0.004) 0. 126 ( • 0.002) 0. 4 25 ( 0.002) August 1976 / Vol. 15. No. 8 / APPLIED OPTICS 1975 66 015- 010 o(M and R(X). Application: Minimum Detectable Change in R As an example of the application of ASl to oceanic remote sensing, we compute the minimum change AR in R at 400 nm, which can be detected with a sensor of given sensitivity, or conversely specify the sensor sen- sitivity required to detect a given change in R. Applying Eq. ( 1 ) to the radiance I(^) at the top of the atmosphere with the sun at the zenith, we have l(n) = h(n) + \[RI2(n)]/d -rR)\. (3) /i(m) and I-z(n) are presented in Figs. 2 and 3, respec- tively, for the three aerosol models discussed above as well as an aerosol-free model (0 X N) and a model with seven times the normal aerosol concentration (7 X N). Now I\in) and 1 2(11) depend only on the direction of the incident solar beam, the properties of the atmosphere, and ocean surface, but not on R, so if we assume these latter properties remain essentially constant over hor- izontal distances large compared to those over which R changes significantly, we can directly relate changes in I(n) to changes in R. Noting that in general R < 0.1, we have [a/(/i)]/(afl) * hin). (4) Figure 3 shows that nI/aR is not an extremely strong function of the aerosol concentration for concentrations 10 0 8 0 6 0 4 0 2 H Fig. '2. /](<;) as a function of h for various aerosol concentrations. 1976 APPLIED OPTICS / Vol. 15, No. 8 / August 1976 67 12 i. i i i i i — r — i r OxN UN 1 10 3xN ' 1 - 8 _ 2 _J«N i - (xlOO) 6 _J0xN 1 1 1 " 4 1 1 1 - 2 - \=400nm 0 _l 1 1 1 L . 1 i 1.0 0 8 0 6 0 4 0 2- H Fiji, :f. /;j(m' as a function of m for various aero ncent r.il ions up to three times normal and viewing angles up to 35° from nadir. This suggests that horizontal gradients in R can he estimated without knowing the aerosol optical thickness with great accuracy. We can now use Eq. (4) to relate changes in radiance A/(/i) to changes in R(AR). i.e.. Equation (f>a) enables one to determine the minimum radiance change the sensor must be able to detect for a given AR. For example, suppose that observing at ^ = O.K.") it is desired to detect a 5"<> change in R lor clear ocean water at 400 nm (ft * 0.1 ) through an atmosphere with three times the normal aerosol concentration. Figure 3 shows that /;>((). 85) is about 0.092. and noting that the extra terrestrial flux at 400 nm is about 140 juW/cm- nm, we find from Eq. (4) that A/(0.85) is 0.064 (iW/cm- nm sr. In a similar way we can relate radiance changes to AR for a nadir viewing sensor and any solar zenith angle. As mentioned previously from the reci- procity principle. /nadir = /(^o)//o, where ju.i = cosfln, "n is the solar zenith angle, and /(/uu) is the radiance at the top of the atmosphere seen by a sensor viewing at j*n when the sun is at the zenith. Following through with the same arguments that led to Fq. (5a). we find A/ni,H,r = Htl\:iHlln)/ilR\±R * *Jn/j(K,,>A/(\ (5b) Clearly, for a given AR, A/nHriir decreases substantially with increasing solar zenith angle due to the presence of the mo factor in Eq. (5b). For example, with a three times normal aerosol concentration, a nadir viewing sensor would have to have about 2.5 times more sensi- tivity at Ik) = 60° as compared to 0t) * 0 to detect the same A/?. The above examples indicate how ASl can be used in the design of a satellite sensor system for estimating some ocean property such as the concentration of sus- pended sediments or organic material. Specifically, one must first determine the effect of the property to be investigated on R, then, based on the sensitivity desired, find A/?, and, finally, use Eqs. (5a) or (5b) to find the minimum radiance change the sensor must be capable of detecting. If the sensor has a limited dynamic range, Eq. (3) can be used with Eqs. (5a) or (5b) to aid in the sensor performance design tradeoffs. Unfortunately at this time relationships between R(X) and sea water constituents are not well established . Conclusions It has been shown that the upward radiance at the top of the atmosphere can be accurately computed by as- suming the radiation entering the ocean is diffusely reflected from a hypothetical lambertian surface (be- neath the ocean surface) of albedo r?(0\— ). This leads to the natural definition of R(\) [/?(0,— ) as a function of wavelength) as the ocean color spectrum. The de- termination of subsurface oceanic properties from space can thus be divided into two problems: (1) the deter- mination of R[\) from satellite radiance measurements and (2) the establishment of relationships between R(\) and the desired ocean properties. Since the method of computation conveniently separates the radiance into a component that interacts with the ocean (I2) and a component due to reflection from the atmosphere and sea surface (/]), it is easy to relate changes in radiance to changes in R(X). It was found that for viewing angles up to 35° from nadir, I> is a relatively weak function of the aerosol concentration for concentrations up to three times normal. This suggests that spatial gradients of R(\) can be determined with only a rough estimate of the aerosol concentration. Presently, computations are being extended to sev- eral wavelengths in the visible and near ir portions of the spectrum. When these are complete, it will be possible to determine the radiance for any R(\) and perhaps point the way toward recovering R(\) from satellite radiances. The author is also affiliated with N0AA Atlantic Oceanographic and Meteorological Laboratories, Physical Oceanography Laboratory, Miami, Florida 33149. Appendix: Influence of Aerosol Phase Function on /, and l2 It is natural to inquire how strongly the computations of I\(n) and /_>(/j) presented in Figs. 2 and 3 depend on the shape of the aerosol phase function. To effect a qualitative understanding of the influence of the aerosol phase function, computations of I\ and 1% have been carried out using the well known Henyey-Cireenstein (HCi) phase function, II -A'-)/4tt ' h<;(«> = ; 7^. (1 + H- - 2(! rns/M''- where the asymmetry parameter g is defined according '-"J>" ) cos«sin«rf«. August 1976 / Vol. 15, No 8 / APPLIED OPTICS 1977 68 I I I I I I I I I I I I I I I ' I • •• "HAZE C" _ HENYEY- GREENSTEIN 2 co"1 i ' I ' I I L_l I l__l I — I— I — I — 1—1- 120 140 160 180 0 (degrees) Fig. 4. Comparison between the haze (' and various Henvey- ('■reenslein phase functions characterized by asymmetry parameters 0.6. 0.7. and 0.8. and II is the scattering angle. Since i,' for the haze C phase function used in the computations described in the text is 0.690, computations have been made with Phi\W for i,' values of 0.6, 0.7. and 0.8. Figure 4 com- pares these PhCiWs with the haze C phase function. The HG phase function for# = 0.7 clearly fits the haze C phase function quite well in the range of 5° < 0 < 140°: however, as is well known, the HG formula is in- capable of reproducing phase functions computed from Mie theory in the extreme forward and backward di- rections. The HG phase functions with asymmetry parameter 0.6 and 0.8 are seen to be substantially dif- ferent from the haze C distribution at nearly all scat- tering angles. On the basis of Fig. 4. it should be ex- pected that / 1 and /_> computed with Phc.C) will be in close agreement with the haze C computations only for ii close to 0.7. Figures 5 and 6, which compare the re- sults of computations of I\ and /_., respectively, for Phc.UH with a' = 0.6, 0.7. 0.8 (steplike lines) and the haze C phase function (solid circles) for the normal aerosol concentration, show that this is indeed the case. It is seen that except for apparent statistical fluctuations, the HG phase function for# = 0.7 yields values of /j and 1 2 in good agreement with the haze C computations. This suggests that the detailed structure of the phase function is not of primary importance in determining 1 \ and 1 2, and it may be sufficient for remote sensing purposes to parameterize the phase function by i,'. To get a feeling for the importance of variations in the phase function in the remote sensing of ocean color, consider the effect of changing the aerosol phase func- tion from a HG with # = 0.6 to g = 0.8 over an ocean with R =0.1. From Figs. 5 and 6, it is found that the normalized radiance at n = 0.85 (the assumed obser- vation angle) decreases by 4.9 X 10-:'; this decrease in radiance would be interpreted under the assumption of no atmospheric change as a decrease in R from 0.10 to 0.056. This clearly indicates then that variations in the aerosol phase function in the horizontal direction could be erroneously interpreted as horizontal variations in the optical properties of the ocean. It is, however, probably unlikely that, except in extreme cases, the clear atmosphere oceanic aerosol phase function will exhibit variations as large as considered in this example. o O 6 X 1 1 r "HAZE C" HENYEY-GREENSTEI X = 400nm 9=0.7 J I I I I I l_ 1.0 09 08 07 06 05 04 03 0.2 01 • H- Fitj. :"). Comparison between /i(^i) computed for the haze C and Henvev-Creenstein phase functions for an atmosphere with a normal aerosol concentration. 1 1 T i i "HAZE C" 1 1? HENYEY-GREENSTEIN X =400nm - II =w-Ti~l . 1 9=08 - -^ • |9=C 7 10 IVa 361 9 8 o Q 7 X • h-7 6 5 4 3 2 l i i 1 1 1 1.0 09 08 07 06 05 04 03 02 01 H- Fig. 6. Comparison between /^(m! computed for the haze C and Henyey-Oreenstein phase functions for an atmosphere with a normal aerosol concentration. 1978 APPLIED OPTICS / Vol 15, No. 8 / August 1976 69 Assuming that the aerosol concentration of the atmo- sphere can be determined, the uncertainty in the aerosol phase function will still of course provide a limit to the accuracy with which the ocean color spectrum can be retrieved from satellite radiance measurements. Quant. Spectrosc. Radiat. References 1. H R Gordon and 0. B Brown, Transfer 15,419 (1975). 2. H. R. Gordon and 0. R. Brown. Appl. Opt. 12, 1544 (1973). 3. G. N. Plass and G. W. Kattawar. Appl. Opt. 8, 455J1969). 4. G. W. Kattawar and G. N. Plass. J. Phys. Ocean 2, 146 (1972). 5. H. R. Gordon and 0. B. Brown, Appl. Opt. 13, 2153 (1974). 6. R. VV. Preisendorter, C.G.G.I. Monogr. 10, 1 1 (1961). 7. .1. E. Tyler and R. C. Smith. Measurement.-, of Spectral Irra- diance I'nderwater (Gordon and Breach. New York, 1970). 8. L. Klterman, VV, Visible, and IR Attenuation for Altitudes to 50km. 1968. Air Force Cambridge Research Laboratories, Report AFCRL-68-0153U968). 9. R. S. Fraser. Goddard Space Flight Center, Greenbelt, Md., Personal Communication. 10. D. Diermendjian, Appl. Opt. 3, 187 (1964). 11. G. Kullonberg, Deep Sea Res. 15, 423(1968). Note that all the phase functions in the present paper are normalized according to «/0 P(8)sin6d6 = 1. 12. T. J. Petzold, Volume Scattering Functions for Selected Waters (Scripps Institution of Oceanography, University of California at San Diego. 1972), SIO Ref. 72-78. 13. H. R. Gordon, Appl. Opt. 12, 2803 (1973). 14. G. N. Plass and G. W. Kattawar, Appl. Opt. 11, 1598(1972). 15. H. R. Gordon, O. B. Brown, and M. M. Jacobs, Appl. Opt. 14,417 (1975). 16. S. Chandrasekhar. Radiative Transfer (Clarendon, Oxford, 1950). 17. R. Curran, Appl. Opt. 11, 1857 (1972). 18. J. L. Mueller, "The Influence of Phytoplankton on Ocean Color Spectra," Ph.D. Thesis, Oregon State University ( 1973). 70 12 Reprinted from: Proc. AIAA Drift Symposium, Hampton, Va., May 22-23, 1974, NASA CP-2003, 175-192. A LAGRANGIAN BUOY EXPERIMENT IN THE SARGASSO SEA by Dr. Donald V. Hansen Atlantic Oceanographic and Meteorological Laboratories Environmental Research Laboratories National Oceanic and Atmospheric Administration Miami , Florida As indicated, we'll hear from a group of distinguished drifters this morning. In order to be sure we don't run out of time for me, I'll say my piece first. I can make mine a little bit shorter -than I'd planned because a number of comments that have already been given set the stage for it. The genesis of my story begins back about 1970 when a number of people in the physical oceanographic community in this country and abroad began thinking and talking about a project to be called the Mid-Ocean Dyna- mics Experiment (MODE). It was referred to yesterday by Doug Webb and others as the M0DE-1 Project. About that time, I began talking to Sam Stevens about the possibility of hitching a free ride, or at least an inexpensive ride, on the French EOLE satellite system, and through the yery good offices of Sam and his crack team, we were indeed able to do that. The engineering for the project was done by the Miami Branch of the Engineering Development The Author: Dr. Hansen received his Ph.D. in Oceanography from the University of Washington in 1964. He worked as a Research Assistant Professor at the University for 1 year before becoming a Research Oceanographer with the Department of Commerce in 1965. He is presently Director, Physical Oceanography Laboratory, Atlantic Oceanographic and Meteorological Laboratories, NOAA, Miami, Florida. 71 Laboratory of NOAA's National Ocean Survey in Miami. Charlie Kearse described yesterday some of the shipboard procedures and arrangements that were developed by them for us to get these buoys in the water, but what he did not mention was that they also were entirely in charge of the engineering and fitting out of these buoys, and in getting them into the water on what turned out to be extremely short notice. As the project developed, it really didn't go quite as we had planned to have it no, because, due to changes in the scheduling of the MODE Project and of the EOLE Satellite Project, it appeared at a critical time that the two after all were not going to be coincident in time. The EOLE Project was to terminate before the MODE Project went to sea. However, it seemed an interesting and important enough experiment to do in its own right, so we pressed on and did it anyway, almost totally independent of MODE. There was about a 1 month overlap between the termination of this project and the initiation of MODE and, in fact, the buoy that we initially had deployed farthest from the MODE area passed within 30 miles of the central mooring of MODE during the second month of that project. I want to show you a few slides first to indicate some of the motivation fcy~ having done the experiment in the way we did it, and to set the stage to address the question of interpretation which Dean Bumpus raised yesterday w^th some vigor. If I can see the first slide now, please. This is an example of a publication that is put out by the Navy. They're called Pilot Charts and show currents and wind to be expected in this reqion of the Sargasso Sea, what mariners and, in fact, what the rest of us know about surface currents in the Sargasso Sea. I might mention in passing, that all of the data that you can find anywhere on such atlases or cherts are, in fact, derived by Lagrangian means. These currents summarized '•<■■ atlases are about 99 44/100% pure ship drift calculation. They're currents inferred from the deviation of ships from their navigational calculations. The major feature I want to point out here is the fact that all of these current vectors show a very smooth steady flow to the west at 72 speeds ranging from about a knot to speeds on the order of 1/2 a knot. The MODE Project which you saw illustrated in one of Doug Webb's slides; I believe, was conducted in a circle of about 200 kilometer radius. "Figure 2 is a copy of a slide taken from some Soviet work in this region. The Soviets have an active interest in the oceanography of the low latitude Atlantic because they conduct vigorous fisheries activities out there and they have conducted intensive research cruises in this region in 1969 and again in 1971. Figure 2 shows their interpretation of those obser- vations. They're a rather intensive set of observations. Soviet literature is a bit hard to interpret as many of you know, in that they don't document their conclusions by Western standards, but as best one can determine, the observations themselves are good. The interpretation is that the solid dark vectors represent the conventional wisdom about the Antilles Current - the northward and westward flow. Imbedded within them are open vectors which are directed to the southeast, which they interpret as a major countercurrent within the Antilles Current and flowing from someplace just off Florida, all the way down, as a con- tinuous feature, joining the complicated equatorial current system and then flowing off to the east. The light lines you see are where they have intensive sets of observations. The observations consist of moored current meter measurements and shipboard measurements of temperature and salinity, from which are computed the velocity field by classical methods. This is the interpretation of what looks like a rather good set of conventional measurements in the region. When I first saw it, I was a little skeptical to say the least - if it's true, it certainly is rather exciting news to the oceanographic community in general and, in fact, rather embarrassing news to the American oceanographic community: that the Soviets should discover right on our doorstep a very major oceano- graphic feature about which we have no knowledge. This is a very major current. It is a surface current which, however, extends to about a kilometer deep in the ocean and it has a volume transport approximately equivalent to that of the Gulf Stream or Florida Current as it issues from the Florida Strait and heads up the east coast, which all of you are aware, I am sure, is the major oceanographic feature off the U.S. east coast. 73 So to try to serve two purposes here--one, we recognized before we went to sea that we would not be able to conduct an experiment in close coordination with the rest of the MODE operations; nontheless, it seemed worthwhile to try to obtain a direct measure of the near surface current structure and its variability in the MODE region. Hence we deployed our buoys along 67°W, immediately to the east of the MODE area, presuming that with the northward and westward drift they would sweep through the MODE area and probably be gone, along the lines of the rather imaginative sketch that Vukovich showed us yesterday, before MODE-I operations began. That was my preliminary guess as to what we might expect in the way of a trajectory development of these buoys when they were deployed, but as you will see, it didn't go quite that way. The idea then was to deploy the buoys so that they would sweep through the surface water in the MODE area before MODE ships came out for that project, except for the southernmost buoy. We learned of the Russian work fairly late in the game and modified the plan to some extent. The buoys were deployed 1° of latitude, 60 miles apart, between 28 north and 25 north. We placed the last one an additional 30 miles south, to place it in the middle of the region where the Soviets claimed to have discovered the countercurrent , to test that particular hypothesis. Figure 3 shows one of our buoys in the water, using the EOLE satellite tracking system which is exhibited in the side room. The next slide is of some interest because I think there probably will be additional discussion of this EOLE system today. Figure 4 shows the dis- tribution of position fixes in time for the No. 5 buoy. It shows the hour of the day from midnight to midnight versus day of drift, so the points show the hour and day from time 0 that positions were obtained through the satellite system. They have a quasi-random pattern providing generally 2-5 fixes per day which round the clock slowly. The satellite "day" turns out to be something on the order of 23 1/2 hours. This is not a p, y i/ #- o to -t-> c >- CD - -o »— ■a co c <: 0) u .c •!- +-> -o c cr>T- o a> CO $- CO < r- cn i q; lu < Q CO o LiJ CD •i- o to to Z3 i— to >> CD O X- .O E C7> 0) C +J •r- tO 4J >> M- to CO CO T3 C 2 CO UJ <£ en _j a: i— c_> UJ i— ( o < 2: U. 3 cc o id 3: CO CO S- Z3 CTl ir> +j o o to c c o o +-> +-> to t/1 O O a. a. 83 Figure 2 THE ANTILLES COUNTERCURRENT AS HYPOTHESIZED BY V.G. KORT 84 Figure 3 DRIFTING BUOY USING EOLE TRANSPONDER DEPLOYED AT SEA 85 >- I > ±> CD ■~1 1 5 J -I Is —i o <\l O O O- -I CM 0 „ i ° o — { «r o ? o 10 0) o o — o 05 o o I o u. ) 3 .}<> u. ! S a IS o o 0 o ' o o * o -I ,'J 1 0 I .'J O o o O O o v I 0J u to avq iO co HDC H 86 >- o ZD co CJ O 1/0 X O0 o o o I — I I— co C£ I— oo t — I Q 0J S- en 15c - j I i ■ ' » \ ■ ■ i » I i i 15c 85° 80° 75< 70° 65< April 22, 1973 25( V 1>^ 20c CO Sept. 22, 1972 J i — 25' 20c 75< 70° 65< Figure 5 DRIFT TRAJECTORY FOR BUOY No. 4 87 240.1 200. 160. oo cc o 1— o 120. LU ■> >- 1- t— 1 o o 80. _1 LxJ > I/O I u_ 40. o s: o ►— 1 \- ■:>! ;;/\y.S 280. 320 rioutp 6 ! Af ;PA';r, I, AN MMf CORRELATION riJNCTIOM FOR BUOY Mo. 4 88 13 Reprinted from: Proc. of the Third Annual Conference on Computer Graphics, Interactive Techniques, and Image Processing, University of Pennsylvania, Computer Graphics 10, No. 2, 218-223. AUTOMATED CONTOURING OF VERTICAL OCEANOGRAPHIC SECTIONS USING AN OBJECTIVE ANALYSIS ' A. HERMAN National Oceanic and Atmospheric Administration Atlantic Oceanographic and Meteorological Laboratories 15 Rickenbacker Causeway Miami , Florida This paper describes a group of computer programs developed for contouring vertical sections of oceanographic parameters. The vertical profiles can be constructed from data collected in a variety of ways. The input data for the driver subroutine need not be equally spaced horizontally or vertically. The routines are written in Fortran for a UNIVAC 1 10 S with an offline Gould Plotter, but can easily be adapted to any computer with a Fortran compiler and a plotter which accepts Calcomp-like commands. The routines are of modular construction. 1 INTRODUCTION This paper describes an automated tech- nique for producing vertical profiles of oceanographic data using quasi-objective analysis. The computer program based on the method is also described. Though obj- ective analysis is well established in meteorology, it has seldom been used in oceanography, and when used, it has been restricted to a specific geographical area [Bretherton (2)]. This program is not restricted to a specific geographical area and is currently being used by oceano- graphers at the Atlantic Oceanographic and Meteorological Laboratories (AOML) of the National Oceanic and Atmospheric Adminis- tration (NOAA) for studying profiles and creating contour naps in the Gulf of Mexico and the New York Bight areas. The advantages of this technique over others generally available to oceanographers are: 1. The sampling points for the input data need not be uniformly spaced in either the vertical or the horizontal direction. 2. Contours are not extrapolated beyond the input data. 3. Interpolation of data is done using statistically based correlation coeffi- cients . The program is based on the assumptions that all stations in a single profile lie close enough to a straight line connecting the two farthest stations that no signi- ficant errors will be introduced bv assuming all stations to be on that line, and that time differences in the collec- tion of dna points may be ignored. These assumptions arc needed when data do not exist for accurate spatial and time corrections. Khen studies are made with adequate resolution in time and space to make such corrections the program can be modified to include them. 2. PROGRAM FLOW PLAN The driver subroutine Versex, calls many subroutines which together accomplish the fol lowing : 1. The conversion of latitude and longi- tude to rectangular coordinates on a Mercator projection. 2. The projection perpendicularly of all stations onto a line connecting the end stations, and the computation of the distance of each station from an end of the line. The distances are in inches at a scale of four inches equal to one degree of longitude (Figures 1 and 3) . 3. The determination of the depth scaling is such that the deepest depth is equiva- lent to the distance between the farthest stations. This will produce a square chart with a fixed horizontal scale and the ver- tical scale being a function of depth and horizontal size. 4. The fitting of data to a matrix with user controlled smoothing with matrix values below deepest depths being logged. 5. The construction of a contour matrix such that a chart is left blank below deepest data, all data points are marked, and the vertical scale is labeled (Figures 2 and 4) . 218 89 o o m o >- o 2 ro 'Z CD u o L_ r-i <+, o C 0 ^r 1— t n 0) 219 90 o o o O o o o o CJ u G> •H X V <4-l o 3 b ro c ■H o o UJ ■s, c tr o ZD ID u (3 o U- >H 4-1 O o a. o o m Cf> 0> >~ o o o CO en 3 220 91 09- O o o ro O O if) I O O O h cvj \- O o in L o o o I- LjJ UJ X »- a. UJ Q UJ cr Z) e> Ll. 3 o. • *-> n! 3 ^ O ctt Q 3 +-> O JZ C 'H O PQ <_> O O >- 2 0) 0> U 2 3 ■m C aj o (-. 0) T3 e h) u = 0, -(f) This is known as Couette flow. Rotational effects are no longer important and all the flow is parallel to the coast in the direction of the stress. The velocities close to the bottom are small. In the examples presented so far, the magnitude of the eddy coellicient plays an important role in determining the nature of the solution. This is another reason for dilhculty in formulating satisfactory models of shelf circulations. The criterion which determines the nature of the solution is the ratio of Del, the Ekman depth, to k, the depth of water. When Del is less than h the flows tend to be rotationally dominated. When Del is com- parable to or larger than h, the effects of rotation diminish or disappear. Consequently, the nature of the solutions for a given value of Av can depend also on the depth of the water. In deep water offshore, the solutions may be rotationally dominated, whereas inshore, where the water is shallower, they might become more Couette or estuarine in nature. The simplest possible effects of two types of forcing in a simple model have now been briefly examined. As should be obvious the problems can become extremely complex even for this simple model, when the depth varies, when sx is to be determined, and when rotational and viscous effects are equally important. 101 26 MECHANISMS FOR STEADY SHELF CIRCULATION A MODEL OF CONTINENTAL SHELF CIRCULATION The ideas exploit d in the preceding sections can be used to form a model of shelf circulations driven by freshwater runoff from land and by wind stress (Stommel and Leetmaa, 1972). As before a shelf of infinite length () direction) and a semiinfinite width (extending from the deep ocean at x = 0 to negative x— infinity) is considered. The depth of the shelf is /;. A mean flux of freshwater, Tr per unit length of coastline, flows toward the sea, due to the cumulative effects of river discharge along the coast. The steady wind stress com- ponents at the surface z = h are tx and ry. The salinity and density are related by p = p„( 1 + 0s). Assume linear dynamics: A ,fc™ - (3gsx - fr: = 0 A,V„ +f/f, = 0 Krs,. + \p:sx — \pxs: = 0 where the motion is independent of y; x derivatives of diffusion terms are neglected because of the large ratio of horizontal to vertical scales; and the stream function \p defines the velocity components u = — i/-.-, ic = if/x. The boundary conditions in * are that at c = 0, \p = \p: = sz = v = 0 at z — h, \p = — TR\ —puA,\p:: = rx; p{,A ,.!'- = Ty\ st = 0 This problem models wintertime conditions on the east coast continental shelf of North America. Winter is attractive because ( 1 ) density is primarily controlled by the salinity distribution and (2) the weak vertical density gradient in winter permits a simplification in the treat- ment of the third equation above, which is nonlinear. Even for this simple model, a complete solution is dillicult. Instead the model is used to estimate the natural horizontal scale length L = SqVx Vj where Vx is the observed width of the shelf and Vs is the decrease in salinity over that distance from the ocean value, sn. The details of the solution are given by Stommel and Leetmaa ( 1972). The solution for L for various values of the wind stress and the eddy coefficient A ,, is shown in Fig. 3, where E (the Ekman number) is equal to A , fir. For these solutions, values of the parameters which are appropriate for the eastern L'.S. continental shelf from Nantucket Shoals to Cape Hatteras have been chosen: / = 0.7 X 10"4 sec"1, /( = 5 X 10:t cm, (3gs0 = 30 cm' sec2, 'I'u = 50 cm2 sec, A, A',, = 1. The upper curves of the diagram correspond to the purely wind-driven regime. The lowest curve, which is convex upward, is the pure density-driven model. For IU' \\T, * Ty = ' \ Tx - 0. Ty ' 2 \> \ V \ \> \ >> \ V \ L cm T. °\ \ T„ ■ ' \SN \ V ¥ \ * \ X \ X \ \ \ x \ \ \ * \ \ \ * \ v\ \\\ 108 T. ■ 0.5 \. \ \ \ \. \* \ \. \* \ T> ■ 1. T„ ■ 0,_»\ \\ \ "~ ~" ^'vNaX ^^^ ^\ xX>\ ^^ ^^+*S*\ x v ^^^. 107 0.01 0.1 E FIGURE 3. Solutions fori* for different values of the wind stress and the Ekman number (E = Av fH2). large values of the mixing coefficient all curves coalesce and the motion is basically density driven and of the estuarine nature that was described earlier. For small values of the vertical mixing coefficient, with v stress predominant, the Ekman transports which convect salt onshore and offshore are independent of Av (as was pointed out earlier). However, vertical mixing, A',., "short-circuits" these transports. This is propor- tional to A „ since we assumed that the Prandtl number, A , A',, was unity. Thus small values of Av correspond to small mixing between the upper and lower Ekman layers and large penetration of salt occurs (i.e., large L). When there is no applied wind stress, the Ekman trans- ports are driven by stress associated with the shear produced by the horizontal salinity gradient, i.e., Avv2 = (Avg(S j)sx. This diminishes as A,, becomes smaller, and despite a partial compensation because k is also smaller, the salt penetration diminishes. This accounts for the different behavior for L (A ,.) for density as compared to wind forcing. For large values of mixing, as pointed out earlier, the dynamics of the flow become nonrotational and also vertical mixing is enhanced. Thus all the curves coalesce in the purely salinity-driven case. Attempts to compare this theory with the observations are dillicult because a priori the appropriate values of Av and k are unknown. It is assumed that Av/Kv = 1. The observations in this area indicate that L is about 3.2 X 108 cm. Wintertime mean wind stress in this area can be estimated from Hellermann's (1967) world charts. These indicate that the magnitude of the x and^> components of stress is about 1 dyne cm2. Thus with tx = tv = 1 and L = 3.2 X 10s, Fig. 3 indicates that Av is about 37 cm2, sec. This is consistent with the observations. It 102 SUMMARY 27 should also be noted that values of /. as large as those observed imply that the shelf circulation, at least for this model, is basically wind-driven. As another test of the model the difference between the salinity at the top and bottom can be computed. This turns out to be 0.14r/f. Again this' is of the right order of magnitude according to the observations. Despite these limited successes of the model, there is a serious discrepancy between the predicted v component of velocity and the observed value. All the observations indicate a negative v velocity of an order of 5 cm sec. The theoretical v component is positive, about 20 cm sec. If the observations arc correct, this indicates that this simple model is not adequate to describe the observed shelf circulation. There Is some observational evidence to indicate that there is a northward rise in sea level along the coast (Sturges, 1974). If this feature is introduced into the model, the discrepancy in the direction of flow parallel to the coast can be resolved (Stommel and Leetmaa. 1972). However, the observations are not conclusive on this point. The theoretical flow close to the bottom then is in the right direction and is on the order of a centi- meter or two per second. CONCLUSIONS As contradictions occur between model results and observations, more details can lie added to the models. At some point, however, the question has to be asked as to how applicable steady state models arc to shelf circulations and in particular to sediment transport. Examination of daily wind records at Nantucket Shoals light vessel shows that the wintertime root-mean-squarc wind stress is 5 to 10 times larger than the mean. Thus the transient fluxes are possibly an order of magnitude larger than the mean ones. For sediment transport this could be the dominant factor since the steady models give rather low near-bottom velocities. Better observations are needed to indicate the direc- tion that modeling should go. Long-term series of current and density measurements are needed to obtain an observational verification of the mean fields and their vertical structure. Time series of currents and density as functions of depth are needed; without these the more complex transient theories of shelf circulations cannot be adequately attained. Finally, for the results of the physical oceanographer to be of relevance to those interested in sediment transport we need to know whether the means or the transients are important in sediment transport. In this chapter the reader is introduced to some of the problems facing a shelf modeler. Other more compli- cated models exist that were not examined. Two of these are the models by Csanady (1974) and Pietrafesa (1973). Csanady discussed the barotropic (depth independent) response of a shelf to an imposed wind stress or external pressure gradient. Pietrafesa considers a steady state, nonlinear, wind-driven model of an eastern meridional coastal circulation. Both are considerably more complex analyses than the one presented here. A more extensive list of references can also be found in them. SUMMARY This chapter provides an introduction to steady state models of the oceanic circulation on the continental margin. Horizontal salinity gradients comprise a major forcing mechanism for shelf circulation. A gradient of seaward-increasing salinity will result in a seaward net transport of surface water and a larger landward net transport of bottom water, if the flow is relatively viscous (low values for the eddy coefficient and depth). With decreasing viscosity, the earth's rotation plays an increasingly important role in determining the nature of the flow . The primary flow tends to parallel the coast, while onshore and offshore transport is confined to the surface and bottom layers. Wind forcing is effected by the application of wind stress to the sea surface. For small values of the eddy coefficient, the solution for the horizontal components of motion occurs in two parts. There is a coast-parallel "geostrophic" component of flow in the interior of the fluid. Additional flow components are experienced at the upper and lower boundaries, which die away exponentially toward the interior of the flow. Net transport in these Ekman layers is given by 7"el = tvJ, where ry is the component of shear stress parallel to the coast and / is the Coriolis parameter. Net transport is independent of the eddy coefficient. The thickness of each Ekman layer is proportional to the square root of the eddy coefficient. As wind-driven flow becomes more viscous (because of increasing eddy coefficient or decreasing depth), the upper and lower Ekman layers merge and the vertical velocity gradient becomes linear in nature (Couette flow). Rotational effects are no longer important, and all flow is parallel to the coast in the direction of stress. These relationships may be combined into a single steady state model for shelf circulation. When applied to the Middle Atlantic Bight of North America, the model predicts an eddy coefficient of 37 cm2 sec, and for the observed horizontal length scale, a primarily wind- driven circulation. However, it is necessary to postulate a northward rise in sea level along the coast, in order for the model to predict net flow to the south, as observed. 103 28 MECHANISMS FOR STEADY SHELF CIRCULATION SYMBOLS .4,, vertical edd>- mixing coefHcient for momentum Del thickness of the Ekman layer E Ekman number f Coriolis parameter:/ = 2Q sin d h depth of water A',, vertical eddy mixing coelhcient for salt L natural horizontal length scale p pressure s salinity T h river transport u x component of velocity v y component of velocity w z component of velocity x horizontal distance from origin perpendicular to y coast horizontal distance from origin parallel to coast vertical distance upward from origin coeilicient of contraction for salt 0 latitude r shear stress p density \p stream function U angular velocity of the earth REFERENCES Bumpus, D. F. (1973). A description of the circulation on the continental shelf of the East Coast of the United States. Prog. Oceanogr.,(t: 111-158. Csanady, G. T. (1974). Barotropic currents over the continental shelf. J. Phys. Oceanogr. 4(3): 357-371. Hellermann, S. (1967). An update estimate of the wind stress on the world ocean, hlon. Weather Rev., 95: 607 -626. Pietrafesa, L. J. (1973). Steady baroclinic circulation on a con- tinental shelf. Ph.D. Dissertation, Dept. of Oceanography and Geophysics Group, University of Washington, Seattle. Stommcl, H, and A. Leetmaa (1972). Circulation on the conti- nental shelf. Proc. Sail. Acad. Sci. U.S.A., 69(11): 3380-3384. Sturges, \V. (1974). Sea level slope along continental boundaries. J. Geophys. Res., 79(6): 825-830. 104 16 Reprinted from: NOAA Technical Report ERL 376-AOML 22, 10 p, NOAA Technical Report ERL 376-AOML 22 ^wsr^ A Comparison of Satellite-Observed Sea-Surface Temperatures With Ground Truth in the Indian Ocean Ants Leetmaa Matthew Cestari Atlantic Oceanographic and Meteorological Laboratories Miami, Florida August 1976 U.S. DEPARTMENT OF COMMERCE Elliot Richardson, Secretary National Oceanic and Atmospheric Administration Robert M. White, Administrator Environmental Research Laboratories Wilmot Hess, Director .CA-UT/Ov , Boulder, Colorado 105 NOTICE The Environmental Research Laboratories do not approve, recommend, or endorse any proprietary product or proprietary material mentioned in this publication. No reference shall be made to the Environmental Research Laboratories or to this publication furnished by the Environmental Research Labora- tories in any advertising or sales promotion which would in- dicate or imply that the Environmental Research Laboratories approve, recommend, or endorse any proprietary product or proprietary material mentioned herein, or which has as its purpose an intent to cause directly or indirectly the adver- tised product to be used or purchased because of this Envi- ronmental Research Laboratories publication. 106 CONTENTS Page 1. INTRODUCTION 1 2. THE SATELLITE-OBSERVED SEA-SURFACE TEMPERATURE MAPS 2 3. SEA-SURFACE TEMPERATURE VARIATIONS ACCORDING TO SATELLITE DATA 3 4. COMPARISON OF SATELLITE DATA WITH SHIP REPORTS 5 5. COMPARISON OF SATELLITE DATA WITH 1963 SURFACE OBSERVATIONS 8 6. SUMMARY 10 7. REFERENCES 10 107 A COMPARISON OF SATELLITE -OBSERVED SEA-SURFACE TEMPERATURES WITH GROUND TRUTH IN THE INDIAN OCEAN Ants Leetmaa Matthew Cestari Daily worldwide sea-surface temperature maps are produced by the National Environmental Satellite Service. For the first half of 1975, sea-surface temperatures recorded on these maps were com- pared with concurrent ship observations in the Indian Ocean. Addi- tional comparisons were made with historical data. These show sys- tematic differences between the satellite and sea-surface observa- tions. The satellite-derived temperatures appear to be too low along the equator and along the East African coast in the vicinity of the equator. Furthermore, in April, May, and June the areas off the equator (and not along the coast) appear to have temperatures that are too high. Although the mean differences are not large (1°-2°C), the fact that the errors vary in time and space made it difficult to apply the satellite data for oceanographic interpre- tations. 1. INTRODUCTION Numerical experimentation has shown that the tropics are an important area for interactions and feedbacks between the ocean and the atmosphere. From present planning, it is clear that during the First GARP Global Experi- ment (FGGE) equatorial regions will receive special attention in the ocean as well as in the atmosphere. The Indian Ocean, because of the monsoons, will also have a special observing period during FGGE, the Monsoon Experi- ment (MONEX). Because of the importance of equatorial regions to climatic studies, and because FGGE will provide relatively complete meteorological coverage, a group of oceanographers has started planning an Indian Ocean Experiment (INDEX). The primary goal of INDEX will be to study the transient reponse of a low latitude ocean to a strong regular forcing by the atmosphere. Pilot experiments, whose results will aid in the design of the final experiment, are now taking place. Sea-surface temperature maps from satellite data could be a valuable tool to study the onset of the Somali Current, upwelling along the Arabian coast, and heat budgets in the Arabian Sea. At the present time such maps are available from the National Environmental Satellite Service. However, as with e\/ery new product or technique, they have to be examined carefully to ascertain their limits of accuracy and applicability. This study reports on a number of intercomparisons between the satellite-observed sea-surface temperatures and "ground truth" in the Indian Ocean during the first half of 1975. The results suggest that more work has to be done before reliable sea-surface temperatures can be obtained from satellites. 108 2. THE SATELLITE-OBSERVED SEA-SURFACE TEMPERATURE MAPS The National Environmental Satellite Service provides daily worldwide satellite sea-surface temperature (SSST) maps. This product is known as the Global Sea-Surface Temperature Computation (GOSSTCOMP). One form of this is an uncontoured computer printout with sea-surface temperature values for each one-half degree of latitude and longitude. With each numerical value for temperature is a code that indicates the estimated reliability of the data. If the code is "+4" , then the last reading had been taken four days before the date of the map, etc. If the number of days exceeds nine, the code space is blank, and the temperature value given is from historical data. If data are available for the day of the map, a letter appears in the code space. An "+A" indicates that the temperature listed is an average of five readings. A "+B" indicates an average of five to eight values and so on up to "+H" which indicates that over 25 values were averaged. The better maps in our analysis had mostly D's through H's associated with the temperature readings. For this study, the daily map with the highest code letter was selected to represent an entire week. One day was chosen to be representative of a whole week because changes from day to day were observed to be small, and weekly representations were more readily compared than daily maps. They start with the week of January 3-9 and end with June 1-7, 1975. Each map selected was contoured in the area of the Indian Ocean off the coast of Africa from 6°S to 15°N and 35°W to 65°E in latitude and longitude. From the collection of maps, one was selected from the early portion of each month to illustrate any monthly differences (Figs. 1-3). 50° E IO°N IO°N FlguA.2. 1. SatdJUUitz Ana-AuAfiace. Ftau/i<£ 2. SatdUUXz. A&a-Au/i&ace. tmpeAotuAe. data ^on. January 1975. tempeAotuAn data fati V nbtiuaAy 1975. 109 3. SEA-SURFACE TEMPERATURE VARIATIONS ACCORDING TO SATELLITE DATA The seasonal variations of sea-surface temperatures in the Indian Ocean is strongly related to the NE and SW monsoons, the transition periods be- tween them, and the ocean current systems established by the winds. The features observed on the maps must be interpreted in the context of these phenomena. Figure 1 shows that there was not a wide range of temperatures in January. Most of the readings were either slightly greater or less than 26°C. On either side of the equator the temperatures are somewhat warmer than at the equator. During February (fig. 2), the sea surface immediately north and south of the equator warms, while temperatures at the equator re- main cool, as in January. North of approximately 8°N the temperatures begin to decline with areas containing temperatures lower than 24°C. This is colder than in January. In March (fig 3) the same pattern persists, but a warming trend is evident. The area of the equator continues to remain cool, and the areas immediately north and south of the equator (5°S-8°N) are warmer. Larger areas of 28°C and higher temperatures are visible, with 30°C tempera- tures reported for some locations. Temperatures for April (fig. 4) show an increased warming trend, with many areas containing temperatures of 30°C and higher. 50° E 10° N 10° N VlaixAd 3. ScuteUUte. 4ea--6uAtfa.ee tern- F-tguAe 4. Sattttitz *ea-4uAtface tern- pojia&viz. data Ion MaAch 1975. poAatuAn data tfoA kpnJX 7975. 110 Temperatures in the area of the equator, as noted in all previous months, remain between 26°C and 28°C. Areas immediately north and south of the equator have become considerably warmer. In the north there are isolated areas with temperatures higher than 31°C. The warmest month from January to June, 1975, is May (fig. 5). Again the area at the equator remains cool. Large areas of 30°C and higher temperatures are visible north and south of the equator. The north exhibits a slight cooling trend in June (fig. 6). The equatorial band remains cool, and areas to the north and south become cooler. Fewer and smaller areas of 30°C and higher temperatures are still present, and the major portion of the entire region contains temperatures of 28°C or slightly higher. The areas of 30°C and higher temperatures seem to have moved toward the north and south, away from the area immediately north and south of the equator. In all months, temperatures along the East African coast were cooler than those offshore. On first glance the seasonal variation in the sea-surface temperature pattern as seen from these maps appears to be reasonable. The transition period between the northeast and the southwest monsoon occurs during March. April, and through the middle of May. A major factor in the heat budget of the surface layers is evaporation, which is proportional to wind speed. Dur- ing the transition, the evaporation decreases and the sea-surface temperature increases. The cooler coastal areas could be related to upwelling or to north-south transport of cooler water by the Somali Current along the coast. A feature that is anomalous, however, is the cool band of water along the equator. In the Pacific and Atlantic Oceans, such a cool band is indicative of equatorial upwelling. However, in the Indian Ocean the winds are not favorable for upwelling, and this feature is rarely present. To examine the validity of this indication and others in more detail, a comparison was made of these satellite data with data from a number of other sources. 50° E --. IO°N IO°N VIqvjul 5. SaJtoJULLtd texi-Au/iiaae. torn- TIquJiz 6. SatoJUUte. beja-buxiaao, tem- peAatusiz data ion Hay J 975. peAatu/ie. data, ion June J 975. Ill 4. COMPARISON OF SATELLITE DATA WITH SHIP REPORTS We can compare the satellite data with actual ship observations obtained at the same time in the same area. For February through May 1975 data are available from a chartered research vessel, La CutUzuaz, in the vicinity of the equator. Bucket thermometer readings (estimated accuracy ±0.2°C) were taken periodically along 55°40'E from 3°S to 2°N. The National Weather Ser- vice also provides information on air temperature, dew point, and sea surface temperature at ship positions through its twice-daily surface-weather maps. There were 106 cases in which bucket thermometer readings from La CuAi&uAe. could be compared with data from the satellite. The mean differ- ence for the whole data set was +0.4°C. Temperatures recorded from the ships were, on the average, higher. The standard deviation was 0.9°C. This indicates that the scatter was quite large. The mean difference actu- ally is rather small. However, if the data are studied in more detail, it becomes clear that there are obvious trends. The satellite temperature data for the equator are always lower than the surface observations and the differ- ence becomes greater as time goes on. For example, for all intercomparisons (32) in the region from 0.5°S to 0.5°N the mean difference is +0.9°C. The standard deviation is 0.7°C. Clearly the equator is systematically colder in the satellite data. This conclusion supports our previous speculations. The satellite data were also compared with merchant ship reports. Un- fortunately, there is a yery limited amount of ship data available in real time. Also, frequently only the air temperatures are available rather than the sea-surface temperatures. In the tropics this is not a serious problem because the differences between these are usually small. Thus, to maximize the data set, the ship reports were compared three ways. First the Satellite Sea Surface Temperatures (SSSTs) were compared with the reported air tempera- tures. The SSSTs were then compared with the sea-surface temperatures, and finally with air and sea temperatures that were within one degree of each other. Approximately 270 ship reports were available in the period from January to Jiine 1975. The difference between air temperature from ship reports and satellite sea-surface temperature was determined from three 2-month groups. For Janu- ary-February, the mean difference was +0.85. For March-April it was -0.59, and for May-June it was -0.97. This indicates that the satellite tempera- tures are lower than actual temperature measurements for January and February, but higher than actual for March-April and May-June. These differences also appear to have a geographic dependence. Figure 7 shows the geographical distribution of these differences. For May-June the satellite reads low in the vicinity of the equator and along the Somali coast, and high else- where. March-April shows the same trend. In January-February the satellite reads systematically low almost everywhere. The difference between sea-surface temperatures from merchant vessels and SSST was investigated for only May-June because not enough data were available for other months. The geographic distribution of the differences is shown in figure 8. Again the satellite appears to read low in the vicin- ity of the coast and the equator and high elsewhere. 112 40° E 50' 60* \-LU 1 + 1.4 3 + 1.23 1 -0.1 1 + 1.0 2 + 1.08 4 +0.7 1 +1.1 2 +27 +s.» 1 +1.3 2 +0.1 1 + 1.0 i + !• t +2.05 1 -o.e l +1.4 I + 14 1 -o.2 1 + 10 i +3.0 1 + 15 1 + 1.0 2 + 1 « 1 -1.5 t ♦s.os 1 -t.O 13 +0.13 II fO.«4 IO°N 1 ) -S.i 1 -3.3 -l 4 S+0.& 1 * -i 0 1 + 1.3 i + 1.2 2 + 0.2 1 +0.2 1 -0 i 45 -0 24 i -0.73 « -0.0« IO°N January-February March-April \L_UW — I.I J , 1 -2.6 2 -1.6 Si ^1.78 4 -1 OS 1 -1.2 ^1.7 1 + 1.1 2 +0»5 1 +0.3 1 +0.6 t +2.7 1 + 1.3 1 +0* 1 -l.« 1 -1.3 1 + 3.6 1 +2.9 t -IJB 1 -o.e a -i.i* 2 -2.1 1 -0.3 1 -1.0 i -P. 5 4 -0.3 1 + 0* + 1.2 4» -1.7 2 -103 1 -3.4 IO°N May-June Figune. 7. GioQKa.pkid dLUtAibutlon oi di{i{i2Ae.nczA between oJji tmp&icutivite and batdUUXd. i>ojx-i>uJi{cui the. numbeA o& componi> . 113 40° E 50< 60* \|_UJ | 1 -2 2 J ; 1 + 0« 2 -2 1 &0.7S * -i 2a 1 -2 2 f+0.7 2 +045 1 + 1.3 1 + 1 « 1 + 3.1 1 -0.7 1 + 0.7 1 +oe 2 -1.2 3 -223 4 -I.I 1 -0.3 1 + 10 -05 3 -0.17 tot 1 -0 2 i -2.4 IO°N F-iguAe. 8. GnognapKLc diA&UbLutLon o& di^&iznceA be- tuxizn Ada-iuA^ace tempeAcutuAU {nam mzAchant i>hipi> and tatoJUUXi 6 to ileal 6 kip data. Table. 1. CompaA-Uon o& Average SSST' 6 with HlAtontcal Skip Vata. SSST SHIPS Apri 1 SSST SHIPS May SSST SHIPS June REGION 28.49 28.52 29-64 29-51 29-32 27-90 Mean Temp, 0.48 0.66 0.86 Standard Deviation REGION 29.25 28.38 30.34 29.50 29.74 27.17 Mean Temp, 0.60 0.78 0.90 Standard Deviation REGION III 29.93 28.84 30.82 29.43 29.49 26.65 Mean Temp 0.56 1.06 1.20 Standard Deviation 5. COMPARISON OF SATELLITE DATA WITH 1963 SURFACE OBSERVATIONS From the International sea-surface temperature for surface observations from s fig. 10) may *be compared wi obvious difference is that temperature minimum. Also along the coast sea-surface are in the satellite data, that the International Indi Indian Ocean Expedition, data are available on 1963 (Wyrtki , 1973). These were accumulated from hips. Maps for January through June 1963 (see th satellite temperature maps from 1975. One in January-June, 1963, there was no equatorial in the 1963 data, there is no indication that temperatures are lower for January-April as they Another difference between the two data sets is an Ocean Expedition maps for April and May contain 115 50° E 60° 50° E 60° ^~yr>25 ^r C **\ I0°N January February I0°N March April *v O A ^ ^ May I0°N June VZguAz 7 0. S£a.-.6uAf). _ » »■ ^^ ti • _ _• r __ r-rif* <:„„ = - i. i lu • sPace potential). The intersection of these lit Lf «! : "nder.Cfcrtam conditions, densitv surfaces with the sea surface !»v k. , /,"" lnaT ^f th° fol]cwjl>e marks the so-called cyclonic boundary of „ ^ed }° ldentify the current's thc current, which is the left-hand side ;°„a{;, changes m sea surface tempera- facing downstream in the northern hemisphere, ture, salinity, color (diffuse), sea state ifSffJJ^'J* Se* sur£ace topography, wave Figure 1 schematically represents a Jhl lnwiS -? EnS* a"d.r,odifi"tions to cross-section of a geostrophically adjusted the lower atmosphere. Infrared sensors , T, . ? , ' '. .r l_,,_ u-*.,, j R" sensois current. The view is downstream in the have been used most extensively to study northern hemisphere. Several important ocean circulation; however, new instruments features should be noted: The mean density s^n.o^c PasSlvc and actlvc microwave (p) in the current is slightly less than in "" 5°[ tCan SV'se temperature, salinity, the juxtaposed water to left. Typically , ?t ' e; a"d surface topography, and the surface temperature (TJ in the current J"!?"*?""1 vlsIble scanners and spectro- ranges fr0I- 2oC to 10°C warmer and salinity raaioneters are providing new information rr n/ \ ■ ■, i mi ->n i _ ■ \ __ _____ „_-, ' j AUX"i> llL* i nunndiion (s o/00) 1S usually lo/oo to 2°/oo higher; on ocean color and sea state Man ' s ml? *u ■ ■ i i ■ _ • t .. ■ j _c ._ _!,___, . . ;>Lc»LC' 'wn s roie this is due to high insolation and evapora- IL m Z ana pnotographer provides tion in the tropi2al source region of the f f .P.J"1 resolution to dale for Stream. In terms of density, thermal ex- describing visible changes across boundaries pansion is largeT than saline effects, and as rfexl as sea and swell patterns. ~he average density of Gulf Stream waters is less than the slope waters along the INTRODUCTION left-hand side looking north. From the hydrostatic equation, Ocean currents have several sea surface 0W manifestations that can be used singularly u-. t °P or in concert, to locate their, boundaries • «V> PQ C1) Coastal currents typically have significant energy at the tidal(l-2 cycles per day; cpd) where g is gravitv and p is pressure, it is or local inertia] frequencies (0.5 - 1.5 cpd seen that the height of the sea surface for mid-latitudes) . These frequencies are (H) is larger in the current when the in- too high for the ocean's density field to tegration is to some deep base pressure, adjust to the motion. Adjustment of the say p=2000 db (i.e. approximately 2000 m) . density iicld provides the conditions by That is, there is a physical rise in the winch many satellite 'sensing techniques sea surface of thc order of 1 m when cress- can be employed in studying the ocean. ing into the Gulf Stream. The last feature on this figure to be noticed is the hori- lhc Gulf Stream off the east coast of North r.ontal velocity profile drawn at the top. America is an example of a quasi-stationary In a geostrophically balanced system, the current system that is well described by surface velocity (Vc) is given by its density field alone. The density _ distribution defines that portion of the V„ - -- -°". pressure field which is used to measure S f dX <>'•■' the flow called a geostiophic current. where f is thc Coriolis parameter, and x Frequencies associated with the boundary of is the cross-stream horizontal dimension, this current arc approximately 0.25-0.1 cpd The horizontal velocity shear, 3V./3.X, in the Straits of Florida (1) and 0.03 - becomes a valuable feature in the study 0.01 cpd in thc meander region off New of ocean circulation from space, because tnglandU). Geostrophic adjustment associ- it has surface manifestations. More ated with low frequencies requires that thc importantly, there is some prospect of density surfaces are inclined with respect determining V's directly from remote sensing; 118 hrl- G o in r\ SURFACE VELOCITY Ts= 21°c Ss= 357^ ^$=10245 o U Ps--1.02'l3 I-*— 10&m — ►- I GULF STREAM SEA SURFACE GEOID ISOPYCNALS Figure I. Schematic cros a- section of a western boundary current in the northern hemisphere . Hove the exaggeration of the sea surface as compared to the sub- surface feazurcs, where a 102 scale change is used for clarity. one example of that will be discussed later. Asso ther whic Deta full cont euph Thes thet bear blue tion the asso prop ter i sedi dele wh i c the riated e appe h is a i 1 s of y unde inuous otic z e nutr i c org ing pi light , shif green, ciated c r t i e s iig oi ments ct ed a h also water . with t a r s to 1 s o s k e this i! rstood ; 1 y b r i n one a ] o i e n t s a ani s;;is ants . , and w t the c The p o r g a n i of the light. from th long th. modify l-i) he h o r i be a v e t c h e d o pwel 1 in howeve g n u t r i n g the re u t ] 1 w h i c h a Pigment hen in olor o f hytopla sms cha water Freque c coast e curre the op zon rt i n F g m r , erit eye izc re ed suf th n k t n^c bv n'tl al nt 1 1C the to motion , c i rcu ] at ion , e 1. (■>) n are not effect is to the c edge. photos yn - r o p h >■ 1 1 - culcs absorb ent concentra- a towards nd other opt ica 1 eased scat- nt rained ons can be a he c r e undaries , properties of When wind and waves run in opposition to the current, the local sea builds higher than when they run in the same direction. Thus the sea "ay be higher or lower in the current depending on the relative wind/cur- rent directions. The latter feature often translates into changes in white cap and foam distributions, changes in glitter patterns, and changes in surface wave refraction patterns. To reiterate, when crossing into the Gulf Stream from the west, one typically encount- ers an increase in temperature, salinity, and perhaps sea state; the color shifts from the green to deep blue, and the parti- culate scattering decreases; there is a rise in sea level due to steric conditions, and a sudden increase in horizontal velocity The detection of these features of the edge of the currents from spacecraft is dis- cussed in the following sections. INFRARED SENSING The ocean surface acts very nearly as a blackbody. Its radiative behavior closely 3-28 119 follows Plank's Law with e?s;ii ssi v i t.i cs (c) greater than 0.99. As j- consequence ol' Kirchoff's law, this requires that the source of the radiation he fro::, the upper millir.eter. Evaporation and condensation can thus play ;. rok in the radiative temperature of water v.hich is frequently 0.5"C or so less than its thermodynamic' temperature. I ° J . This is of fundamental importance in det err.:ini:.p, '!' fro;:; space, but for ocean current beu:- Jary determina- tion, the observe! is looking for thermal gradients rather than absolute temperatures Thermal radiation leaving the earth is modi' ficd by atmuspheri c absorption and emis- sion. Clouds arc opaque to the earth's radiation, which peak.- at about H)..:i. The radiative transfer equation for spectral thermal radiation is Nn=N, 1-or tran subs pher top the pher iii I c but i ncgl is v thi smi cr i c , of los ic gra on cc t cry tt; pt: re: th< s ri- ot ed s formula , p and t i v o i a r. s L e i" t;\ i n mal tmc 5 I! 1 m i 1 t i on , N is radiance, i s p res sure , and t h a are s u r f a c e a n d a el)'. The radiance a phere iNc, ) is the su ace radiance due to ance (\'s . -, ) plus i cii i o i iiiu 1 a t es I he c n o s o h c r c . S c a 1 1 c r i n (3) T 1 P e trios - t the !il of atmos- the ontri • o ] 5 se it this formulation, he can 1 at there wave] cnglhs ( > ) . One cons is that lessened moist at vapor ah in T ' c a Thi i" is red re no a r i e s . diffcrcn and the infrared equence oi ocean surf a t s a t e 1 1 mo sphere i s o r p t i o n w a be rcduc a fundamen tc sensing In 1 o \\ 1 a t cos across a tmospheri tcchni que at ace itc n t and ed tal o f 1 tu 00 c r.i When thro view to t liqu red dept clos hi nd sate coas ance wave ocea that and iy, c leu rad ugh s c 1 he f id w phot h. e to cran Hit tal s . long n i s oi i n f r then d-fr lation a c 1 o u 0 U d ~ t O act t h a t cr , ons ar I.ow cl the o cc to c d a t a lands 1 1 o w e v e ths (- an or clouds a red d the p e e can t rom a j he to re oc dc s c a;i oi s re c r i ma 1 tud . 5 - :; S C l r ic n c her fr re m i t tran sun s? ion radients are . lor a very . 5 _n wa tor difference on in infra- rent bound - the the r in a 1 t s a re sma 1 1 high, the p te at c and e em ouds ccan prop S can r , i 0 . 0 u der and at a i c tu be a ' mper loud. as n ittc oft 1C V ei l ii,;i 1 have n t ii m) l o f in i a :i are re e id en surf a c c nd the u r e s . T a r e m a d e d earl from a have t u e s a n d e r d r c t a ly', lots e a r - o c e longer r e f 1 e c n i t udc I !.' bo qui red ::.ents IV ficd. p a s s space his i e up icr, v e r >' emper bene t i o n 1 >■ in anic v l s i b t a n c e less th vi simul li l c h Dual es craft s due of infra- s ha 1 low a t u l e s e are a cf r a d i - le of the than s i b 1 e t a n e o u s ■ a r e channel visible and craft are pr int crprct a t i Figure 2 . The left -ban infrared I 1 u Atlantic P. i g Cape Matters hand panel i visible (0.6 1975, at 145 positive pri are lighter; print with 1 tops) 1 i giit e Scotia is sc cloud, but i clear where comparing th seen to be c Stream no a rid surface t h e r hints of the image as v.el the section d i s c r i m i n a t i visible clian (approximate highly absor cloud deterin synchronous earth in the 30 minutes i day. C9. 10) 3 0 - minute i notion p i c t u rates sevcrj than ocean f b e r e a d i ly i t hernia] back interval is all f requeue cpd can be 5 spatial r e s o and the mid- geometry are sampling sys infrared scanners aboard space oviding the data for such on; an example is given in (8) d panel of Figure 2 is an .S - I2.5nmj image of the Mid- ht showing the coastline from s to Nova Scotia. The right - 'he simultaneously scanned 'c.7:im) image taken on 11 May 0 GMT. The visible image is a nt , that is, higher radiances the infrared is a negative over radiances (i. e. cloud r. The large feature off \ova en in the visible panel to be n the infrared alone it is not the cloud-sea boundary is. By e images, radiance patterns are loud-free expressions of Gulf ers, rings, eddies, and other mal features. There arc some se patterns in the visible 1. Tli is will be discussed in on visible imagery. The cloud on would be improved if the ael were in the near infrared ly lam) where the water is more bing. An alternate method of i nation involves use of the satellites that observe the same infrared band, but every nstead of several times a Using only infrared data, the mages arc made into time-lapse res. Clouds have advection 1 orders of m a g n i t u d e faster catures, and therefore they can dcntificd against the ocean ground. Since the sample 3 0 minutes, ocean features of ies down to the Nyquist of 24 tudied if detectable. The lution of approximately 10 km, latitude limits of the viewing the limiting features of this t em . Inf pri Man to fil Ian res pro occ ran don app 2 bla syn i s for abl rare m a r i y oc the m an d, a olu t cess an ic ge- c , m car the ck. chro be in the c . d sat ly fo can f match d the nd cl ion c ing t r a n g Mi en os t c whi te land A sp nou s g con ocea lite d meteor tures n the a d i a n c ds. I be o b d a t a is mat his t y uds sa and, a d vcrv ially tel 1 i t r u c t c d commun a ar ogic e no ay s rang rove ined ch t ed t of rate in t a r m sign to so f y i.'i e pro a 1 pu t emp cale e of d oce by s ha t o o the proce the h e c a water ed de accom ilra 1 11 be cess rpos ha s i rang the' an r peci nly gra ss in film se o app vice pi 1 s oo p come ed es . ;ed due e of the sea , ad i ance ally the y scale g is and f I" i g u r e ca r for the h this product s avai 1 - 3-29 120 Figure 2. Infrared Cleft-panel) and visible (right panel) image pair of Gulf Stream off [Jew England. These aata are from the very high resolu- tion •scanning radiometer of the NOAA polar orbiting satellite (8). VISIBLE SEN'S l.\G N0=NQ + r Ns + °Nd (4) Detection of ocean wavelengths (0.4-0. currents in the visible ';ini) depends on the change of ocean color and the change of sea state associated with the boundary"; The -spectra! radiance at the top of the atmos- phere by in the visible region (Nn ) is given O' where N is the contribution of the atmos- phere alone (most ly . Rayleigh scattering), Ns is the contribution at the surface due to reflection from the surface, N, is the diffuse radiance at the surface due to photons that have penetrated the surface;Y and a arc atmospheric t ransmi ttance factors 3-30 121 for Ns and N'j respectively. Changes in sea complex patterns in the slope water off New state, by which is meant changes in white York, are observable in both the visible caps, foam, glitter, etc., enter the equation and infrared images. In the visible they only through the N*? term. Similarly arc regions of higher radiance embedded in changes in the optical properties of the a zone of low and probably uniform specular water itself, winch is information on the return. This suggests that what is seen is absorption and scattering of light by due to patterns in the diffuse radiance. particles in the water, is represented by Thus the interpretation is that these d' circulation features are being detected due to variations in the optical properties of The spectrum of diffuse radiation is a the water and are variations in biochromes complex function oi scattering and absorption. and particulate scatterers . ( 13) In a simple single- scatter ing model, the independent variables were shown to be the The interpretation is difficult here total attenuation coefficient, the total because only two channels of data are scattering coefficient and the fraction of available. The 0.6-0.7 Mm channel is a backscattercd light. ^ UJ Each variable is good -choice for visible radiance and was also wavelength dependent, so that an the most useful in applying LAN'DSAT (Earth infinite variety of optical conditions can Resources Technology Satellite) data to the combine to produce the same Nd . In the marine environment . >-4J However, since it case of current boundary determinations, is a valuable channel in the study of the highly productive water along-side the optical oceanography, it is not the best Gulf Stream cyclonic front is high in spectral interval for cloud and sea state pigmented molecules as well as in particir descr imination . Experience with LANDSAT late natter. The net result is to shift and the experimental scanner on SK'YLAB have the peak of the upwelling radiance spectrum shown that 0 . 95- 1 . 05 vm is the optimum band toward longer wavelengths, that is towards f0r this purpose. Incorporation of this green colors. At the sane time, the opti- wavelength interval in future multispectral cal intensity increases due to increased imagers is strongly recommended in order to scattering. When these conditions hold, overcome the ambiguities of oceanic inter- the water along the cyclonic side of the pretation in the present system. stream will have a higher radiance in a multispectral image such as on the NOAA MICROWAVE SENSING polar orbiting satellite. In Figure 2 the edge of the current can be seen in the Microwave sensing may be considered in two visible imagery as well. as the infrared. ways: Active systems, by which is meant The Gulf Stream appears to have lower radar type devices such as altimeters, radiance than the slope waters in agreement scat terometers , and imaging radars; and with the above explanation. passive or radiometer- type devices that sense the emitted microwave energy in the However, the S's term also contributs to same sense as infrared or visible radio- NQ, and its behavior requires discussion. meters do. Many features of the edge of a Reflection from the surface at these wave- current (see Figure 1) can be identified in lengths changes the radiance spectrum in a microwave data, including sea state changes, wavelength dependent fashion. In the 0.6- temperature and sal inity" changes , the 0.7um interval of Figure 2, a higher sea physical shape of the sea surface, and state in the Gulf Stream could raise the actual current speeds. t 14 ) radiance to the sane level as the N'j influence on the slope water (and thus Passive microwave energy is sensitive to there would be no visible signature), or it changes in surface temperature, salinity, could exceed the radiance and once again a roughness, and foam coverage, and to the signature would exist. Kinds for this day presence of sea ice.1- J The transfer of were from the southwest at less than 2 m s"1 radiation follows Equation 3, except that due to a high-pressure ridge lying parallel the N«. term is a function of polarization to the east ccast. Under conditions of as well as nadir angle and wavelength, such weak and variable winds no foam or Since foam transmits energy from below, white caps arc ant icipated ll 2 ' , and the transmission (t ), emissivity (e ) and contribution to N0 is dominated by N'j, with reflectivity (p ) all contribute to N as Ns contribution at a uniform value over the follows: s whole image. Some effects of what is a probable lack of Ns= T\ Nj "•" £\ NBB+ P\Nfl ("1 specular return can be seen in Long Island Sound and south of Nova Scotia. These dark areas arc interpreted to be zones of calm seas where glitter does not reflect the morning sun (P"?>0 F.ST). The major features The subscripts in this expression denote of the Gulf Stream front, the large eddy- the blackbody radiance (BB), the incident like structure south of Cape Cod, and the 3-31 122 radiation from tl inc i dent rad iat ic the foam. '! microwave en due to water 100 cm (50-0 r.i i s s i v i t y is ic c V( a tine splie re (a ) , ant! f rom t he ocean ( i ) cm atmosphere is opaque tc y at wave I ei:;; t lis he low ] cm por, -but between 1 c;:i an J Gil.; the atmospheric trans- r v " c 1 o s c to 1.0. Measurements oft lennodynam i c temperature relate to N.,„ by ' lank ' s 1 aw. N..., can be approxir.at e J by t io Raylcigh-JcanS Law at microwave wave 1 en ',ths, and this is proper - tional to '!_. -V-l . This s impl i f ies the c a ?. c u I a t i o n s in t iat the radiance is dir- ectly proper Lion a t o the first po w e r o f the temperature. Salinity affects ' , and absolute measure!'! :nts by aircraft of the 'oo arc be ins reported. (■■'■"J order of l°/oo-2fJ Salinitv deter."', in . ny techniques measure enissivity at 21 cm and require the ancil- lary measurement of temperature. Salinity effects on the dielectric constant of sea water are small at 8 cm. Thus a two- channel microwave sensor can provide salin- ity and temperature measurements in the absence of foam coverage (p=0 ) . Sea state effects on an imaging radar are shown in Figure 5. These7data from a side- looking airborne radar1- ' show two narrow lineatiens which were shown by simultaneous infrared thermometers to be associated with a large temperature grad- ient in the Gulf Stream front region off Cape llatteras. These lineations are changes in the radar backscattering cross- section which suggest changes in sea state asso- ciated with the thermal change. This is ^ St rear: •'.•• oj tit j u:rJ/ • into rent a plo on of on per (K) . throyg lterna by an ' , a r t h e r m iting cksoattcr ing appl i cat ion oce allograph resting appl peed determi t of current radar backsc unit area f If the wind h ^ne or sev 10) .then Vs tivcly, if V altimeter or efinement on i c r o w a v e tec horir. on for from radars to many y(18) .A i cat ion is nat i on (14) speed (Vs) a 1 1 e r i n g or several speed can be eral independ- c a n be can be s . a L a g r a n g l a n W is possible, hninues are a remote sensing, Figure 4. Plot of current speed versus radar backscattering as a function of different wind speeds (12). Gulf Stream current speeds are typically 1-2 -meters per second. Photographic Sensing Man ' s appre none for t servi from the r hopef futur such other obser Thesc the d resol syste An ex Figur turbu of th South south Curre At Ian the a horde sea . m i x i n role ciated so ful h e fir ng and space . eccnt u 1 1 y w e. Ma as con pstte v e d an data etails ution ms . as an o from t 1 y as S st t ime p h o t o g This Apollo- ill be n >■ o c e a f luence rns tha d photo are pro of the unavail bserv lie ea KYLAB give raplii progr Soyuz s t r c n n ci r s , up t ref graph v i d i n ocea able er in sp rlier mi Astro n traini ng ocean am was c test p r gthened culation welling , 1 c c t cur ed from g new in n at a s by other to ace was ssions , but nauts were ng in ob- f eatures arried on oject and in the features eddies, and rents were space L20? sight into patial imaging ample c 5 w lent e Fal Amer crn h nt/Gu t ic stron r and The g acr of hich mixi klan ica emis If S The auts was colo oss such d is a ng in d and phere trcam c on f 1 u near folio r boun the fr etai hand the Braz The anal syst ence the w e d dary ont , 1 is -hel conf il c sec og o em i was Braz for sho unl give d pho luenc urren urren f the n the firs il/Ur over wed v ike t n in tograph of e region ts off ts are a Labrador North t seen by uguay 1000 km to ery little he eddies 3-33 124 Figure 5. snrf- ,ij of c si otis cf the bozo:.: Current taken fron have beet: : ho to.j ra by laboratory tc.ah es tima tea O DC shown in isure The edd be seen because plankton o materials cause color v.iri sea. The eddy appears to and is poss'.ibly caused by shear alon" the Falkland C Photographic derails like this arc usual chance happenings. In the case of SKYLAH irbi: I ■ , >: t r :■:: r c . 'if i ', :,.. 1 .<' (• < I a. :• LA B ' ' ~'i) . Pciai cr. '": ; >:? cd : e s . y'r c eddy i n a c i c r . y i n i"i; ure 5 r s u s p e : ulcd a t i o r s n the be a n t i c y c 1 o n i the v eh ) C 1 t >• u rrt n t . Ic ly 'i ii o w c v c r , 1 1 1 e s several times on sever and features such as t not on I \ v i oil ed sevi i a ecraft transited the area were s t iu! i i hart h t vo sou ice con s ecu t i \e d a y se in f i <: u re 5 w times, but they t he m.my device s i n t h pe r i men t i'ac ka go . ere An important reason for the successful observations was the real-time communica- tions bet w eon the crew and occanog rap hers at mission control. Future manned earth observing missions must exploit this com- munications feed-back even further, so that ships and aircraft can conduct detailed studies into these interesting new- found features of the ocean's circulation. Conclusions Several techniques have been described that offer useful application to the study of ocean circulation from space. Each has 3-3-1 125 advant ag es and disadvantages. The very high spatial resolution of visible and infrared sensors is not practical at microwave frequencies because of limita- tions in antenna design. On the other hand, the microwave devices offer all- weather sensing capability but at reduced radiant resolution. The sens hip, of ocean currents should be approached with multi- spectral techniques using visible, infra- red, and microwave observations. This will not only provide the best opportunity to o b s o r ve the feature, but will also increase the degrees - of - freedom in an automated identification scheme which is an important goal for studying ocean physics from space Acknowledgement: The author wishes to express his appreciation to his colleagues at AOMI, for assistance in preparing this manuscript, and particularly to D.V. Hansen with whom there have been many fruitful discussions. References (1) Duing, W. (1975). Synoptic Studies of Transients in the Florida Current, J. Mar. Res. , 33(1) , pp. 53-73. (2) Hansen, D.V. (1970). Gulf Stream Meanders between Cape Hatteras and the Grand Banks. Deep Sea Res . ,17, pp. 495- 511. (3) Neumann, G. and W. (1966) Pierson, Jr. iles of Phvsical Oceano- graphy , Prentice-Hall, Fnglewood Lli'rts, N.J., pp. 2 24-228. (4) Maul, G.A. and H.R. Gordon (1975). On the Use of the Earth. Resources Technology Satellite (LANDSAT-1) in Optical Oceano- graphy, Remote Sensing of Environ . , 4_, pp 95-128. (5) Teague, W.J. (1974). Refraction of Surface Gravity Waves in an Eddy. Scienti- fic Report, U. of Miami, Coral Gables, Fla'. , UM-RSMAS-No. 74034, 94 pgs. (6) Ewing, G. and E.D. McAnster (1960). On the thermal Boundary Layer of the Ocean, Sc i once , 1_31_, pp. 1574-1576. (7) Maul, G.A. and M. Sidran (1973). Atmospheric Effects on Ocean Surface Temperature Sensing from the NOAA Satellite Scanning Radiometer, J. Geophys . Res . , 28(12), pp. 1909-1916. (8) Data from the NOAA-4 meteorological satellite provided by H.M. Byrne (NOAA- AOML) . (9) Maul, G.A. and S.R. Baig (1975). A new Technique for Observing Mid-latitude Ocean Currents from Space, Proceedings, Amer. Soc. Photogram. , Wash. D.C., pp. 713-716. (10) Legeckis, R. (1975). Application of Synchronous Meteorological Satellite Data to the Study of Time Dependent Sea Surface Temperature Changes Along the Boundary of the Gulf Stream, Geophys. Res . Ltrs. , 2(10), pp. 455 - 438. (11) Gordon, H.R. (1973). A Simple Cal- culation of the Diffuse Reflectance of the Ocean., AgpJ^ Opt. , 12, pp. 2804 - 2805. (12) Ross, D.B. and V. Cordone (1974). Observations of Oceanic Whitecaps and their Relation to Remote Measurements of Surface Wind Speed., J. Geophys . Res . , 79(5) , pp. 444-452. (13) Compare with the interpretation under different wind conditions given by Strong, A.E. and R.J. DeRycke, Ocean Current Moni- toring Employing a New Satellite Sensing Technique, Science , 181 , pp 482-484. (14) Parsons, C. and G.S. Brown (1976). Remote Sensing of Currents Using Back- scattering Cross-Section Measurements by a Satellite Altimeter. (Submitted to J . Geophys . Res . ) (15) Hanson, K.J. (1972). Remote Sensing of the Ocean. In: Remote Sensing of the Troposphere , V.E. Derr, ed., U.S. Gov't. Printing Office, Washington, D.C., pp. 22-1 to 22-56. (16) Thomann, C.G. (1975). Remote Sensing of Salinity. Proceedings of the NASA Earth Resources Survey Symposium, NASA TM X- 5816S, pp. 2099-2126. (17) Data provided by D.B. Ross, NOAA- AOML . (18) Eisenbcrg, R. P. (1974). Practical Considerations to the Use of Microwave Sensing from Space Platforms. In: Remote Sensing Applied to Energy-Related Problems , T.N. Veziroglu, FJ. , U. ot Miami, Coral Gables, Fla., pp. S3-29 to S3-41. (19) Molinari, R.L. (1973). Buoy Tracking of Ocean Currents. In: Advances in _the Astronaut ical Sciences , F.S. Johnson, ed . , Amer. Astron. Soc., Tarzana Calif., pp. 431-444. (20) Kaltenbach, J.L., W.B. Lenoir, M.C. McEwen, R.A. Weitcnhagen, and V.R. Wilmarth, eds. (1974). SKYLAB-4 Visual Observations Project Report, NASA, JSC- 09055, TM X- 58142, Houston, Texas, 250 pgs. 3-35 126 (21) Johnson, W.K. A Mul t i spcct ral Ai: bet Keen the Bra : j 1 from SKYLAB. Subm of Knvi ron. and U.K. Nor is (1976) . .'.lysis of the Interface and i'a 3 k] and Curv< nts itted to Remote Sens inn 3-3G 127 18 Reprinted from: NOAA Technical Report ERL 378-AOML 23, 69 p, NOAA Technical Report ERL 378-AOML 23 .uOMM»S^ NOflfl An Experiment to Evaluate WM SKYLAB Earth Resources Sensors for Detection of the Gulf Stream *>%£!?* & George A. Maul Howard R. Gordon Stephen R. Baig Michael McCaslin Roger DeVivo Atlantic Oceanographic and Meteorological Laboratories Miami, Florida August 1976 U.S. DEPARTMENT OF COMMERCE Elliot Richardson, Secretary , o * 0 CO rtf id Mh U <+H k 0 -P O -H o O • Q) • - CC < or CD a: < UJ o < < 300- 200- 100 - 400 450 500 550 600 650 700 WAVELENGTH (nm) Figure 2.2 Example of upwelling spectral irradiance before (broken line) and after (solid line) filtering to eliminate the effect of ocean, surface glitter variations due to surface waves. The spectra were digitized off the strip chart at intervals of 20 points per inch; the points were sufficiently close to retain the shape of the original traces. The trace in Fig. 2.2, provided 248 data points. As different wavelength drive speeds were used on different stations, this number varied from scan to scan . In order to choos frequency energy with is important to identi signals. Fig. 2.3 con relative strengths of scan shown in Fig. 2.2 me try scan using Tukey One can see strong per 30 data points per eye signals that give the appearance . e a filter that successfully removes high a minimum loss of significant trends, it fy the periods of the high frequency tains a plot (light, broken line) of the various periodicities of the spectrometry This is a power spectrum of the spectro- 's method (see Herman and Jacobson, 1975). iodicities at approximately 6,7,9,20, and le. These are the dominant high frequency original scan in Fig. 2.2 its sawtooth 140 cr < CD cc < o o. ■ 100 0.80 o < or '■' 060 UJ 7 o Q. (f> 0.40 UJ cc or UJ i- 020 u. 90| 30 20 1513|ll| 9 8 7 6 5 4°°° 45 12 10 SAMPLE INTERVAL Figure 2. Z Tower spectrum of upwelling spectral irradiance shown in Figure 2.2 The narrow line is the high frequency motion caused by surface waves reflecting specularly; the heavy line is the response of the Fourier filter to low-pass the data. The response of the filter chosen to remove these high frequency signals is shown by the heavy, solid curve in Fig. 2.3. This response is in terms of a ratio of the contribution various frequencies make to the form of the original trace, to the con- tribution the same frequencies are allowed to make to the form of the filtered result. Thus in the example chosen, no contri- bution is allowed for periods smaller than about 9 data points per cycle (10 nm) ; full contribution is allowed for periods greater than about 9 0 points per cycle (10 0 nm) , and half contributions are allowed at about 25 points per cycle (30 nm) , which was the longest period of the major peak in Fig. 2.3. In general, all filters were chosen to remove the short period (high frequency) signals in the same way. Power spectra of different spectrometry scans did not always closely resemble each other however, and each filter had to be chosen on the basis of an individual inspection of each scan. c) Wavelength Calibration Irradiance and wavelength are indicated by separate voltage outputs. The wavelength voltage is produced by a potentiometer directly connected to the diffraction grating, voltage varying with angle. In addition, an inscribed wavelength scale is connected directly to the potentiometer. Thus it is possible to compare the voltage of the wavelength output as recorded by whatever strip chart recorder is used with the wavelength indicated by the scale. Such a comparison was performed on the strip chart recorder used during the cruise, and is the basis of 141 the wavelength calibration employed for these data. The drift- free nature of the grating-scale design permits this type of calibration. Recorder outputs were compared with scale readings at 5 nm intervals over the entire wavelength range ased in this experi- ment. A third-order polynomial was fitted to the resulting numbers to obtain an equation giving wavelength in terms of the position within the spectra on the time axis of the strip chart. This form of equation was chosen because it does not require a fixed number of data points for all spectra; the only require- ment is that the digitizing interval remain constant. Comparison of this calibration with known spectral lines indicates that the calibration is within 3 nm of true wavelength over the whole range of visible light. 3. PHOTOGRAPHY The use of color photographs to measure phytoplankton con- centrations in natural waters is based on the argument that the photographic material responds in a quantitative, reproducible manner to the variance in the light field of the water. The vari- ance is associated in a fixed manner with the concentration of the phytoplankton, its distribution with depth, and its species and nutritive history. For a number of years there have been appli- cations of these variance techniques to remote sensing of the ocean Ce.g. Baig and Yentsch, 1969; Mueller, 19 7 3). The SKYLAB experiments provided the opportunity to extend some of the tech- niques developed from laboratory tanks and low-level aircraft flights to synoptic mesoscale coverage. The field program (section 2.1) was to provide surface truth for the variance analysis; the SKYLAB photography was to provide the photographic products . 3.1 Measurements A photograph is merely the record of the integral of the intensity of the illuminant falling on a subject and the re- flectivity of the subject. In a color photograph a third variable is introduced in the spectral properties of both the illuminant and the subject. A color photograph is satisfactory if it produces, in a viewer's eye, a response similar to that produced by the actual subject. The satisfactory spectral response brought about by the color photograph is a result of the eye's inability to distinguish between a pure spectral source of light and a mixture of such sources. A color photo- graph does does not produce in each pixel (picture element) the exact spectral reflectivity of the corresponding spot of the subject (with the exception of a subject that is itself a color photograph) . Instead the photograph produces in each pixel a mixture of three colors which the eye perceives as a single color, and it produces only these three c&lors; this is called a "metameric" match. 142 A color can be thought of as a vector in n-space, with pure light as the origin. An infinity of coordinate systems may be created around this vector, but once one of the coordinate axes is fixed the others also become fixed. Common varieties of color films need only three colors to reproduce the variety of colors seen in the real world. To the eye these colors look like yellow, magenta, and cyan (blue-green). These colors are the coordinate axes of the color space . Every other color is an unique combination of these three primary colors. If the subject is composed of two different colors then the eye will perceive it as if it were a single color. The eye in this case performs the metameric match. A color photograph will do the same. If the color of a subject is changing then the change may be noted as differing quantities of the three dyes in a color photograph of the subject. The utility of monitoring the changes in dye concentration of a color photograph can be carried a step further. It has been shown that provided the continuous spectrum of the subject has previously been measured, the dye concentrations can be used to generate a new spectrum without re-measuring the spectrum (Baig and Yentsch, 1969). The new spectrum must have been part of a "training set" of spectra for which the color photographs exist, or must be any combination of the original- training set spectra. Then, through a multivariate analysis and regression technique, a synthetic spectrum can be generated using only the dye concentrations. The technique is especially useful when the concentration of one of the components of a mixture is changing. Tank (Baig and Atwell 19 75) and low- level aircraft flights (Baig, 19 73) have amply demonstrated that phytoplankton concentration in natural waters can be easily and accurately measured with the technique. If the spectra of the phytoplankton are not of immediate interest, then the multi- variate reduction is not necessary. The problem then reduces to a correlation between the concentration of phytoplankton in a training sample and the variation in the dye concentrations in the color photograph. 3.2 Data Analysis The first photographs to be analysed were the 9-inch color transparancies from the aircraft aerial cameras . On a number of frames there were subjective color differences. However, similar differences were noted on frames in which the color of the subject area would have been expected to be uniform. A densitometric analysis of one of these frames revealed a variable blue/green ratio as a function of radial distance from the principal point. These data are presented graphically in Fig. 3.1. Attempts to use these data to "correct" data from other frames of aircraft transparencies were not successful. This vignetting problem was so severe that the images were displayed on the non-linear portion of the film's D-log e curve. This introduced an unknown non-linear error which could not be 143 corrected without an in-flight calibration with targets known spectral response. of 432101 2345 CENTIMETERS FROM CENTER OF PHOTOGRAPH Figure 5,1 Example of the. effect of vignetting in aircraft film data on the -percent transmittance of blue C450nm) and green (550 nm) light > and on the blue /green ratio. Aerial color- positive film was used in the RC-8 camera. Because the scale of the satellite photos is so much smaller than that of the aircraft photos , discontinuities such as fronts and eddies are recorded in only a small area of the photo. By comparison, similar elements have to be recorded over large areas of single aircraft photos, or even in a sequence of such aircraft photos. The practical effect of this scale difference is that variations in film density related to position on the photo can be ignored where the area of interest covers only a small area of the photo. Of course comparisons between areas widely scat- tered over the photo are still subject to the problem of spatial density in the photo. All of the pertinent SL-M- duplicate films were analysed on a hybrid transmissometer. The light table and associated aper- tures, filters, and diffuse acceptor are from a Welch Densichron densitometer. The light sensor and associated electronics are a Gamma Digital Photometer. The transmission of each of the three color filters and of the white light setting was calibrated with a non-silver standard step wedge which is traceable to N.B.S. standards. Such a step-wedge is a better approximation to the actual attenuation characteristics of color film than the usual silver grain step wedges. This is because color films do not have any silver in the final image, depending instead on dye 144 *£ 83d *TS 08P V8VN ZIC-B8 Figure Z. 2 S-190B panchromatic (SO-022) photograph of the Straits of Florida and the northern coast of Cuba. The box in the lower left brackets the anticy clonic front of the Gulf Stream, and defines the area where densitometric measurements were made. 145 rm oat vsvn ♦IE-68 Figure 3.3 S-190B panchromatic (S)-022) photographs of the Straits of Florida and the western Florida Keys. The box in the lower left brackets a plume of water from Florida Bay3 and defines the area where densitometrio measurements were made. 146 densities for attenuation of the illuminant . Data in the follow- ing paragraph are reported as the percent fraction of transmitted light rather than as density, since transmission ratios will have more meaningful interpretation than density ratios. Stand- ard deviations (a) follow each ratio. The first area analysed was a plume of water off the coast of Cuba, in frame 97, roll 64 SL-4, near reseau #8 (Fig. 3.2). The average Blue/Green transmission (B/G) ratio is 5.9, cr* 0.1. Just to the left of the interface of this plume with the Gulf Stream, the B/G ratio changes to 7.0, a+_ 0.1 in Gulf Stream water. In frame 98 a similar plume on the Florida Keys side of the Gulf Stream has a B/G ratio of 6.0, o +_ 0.1, while the Gulf Stream water immediately to the right of the plume has a B/G ratio of 6.8, a + 0.1 (Fig. 3.3). This particular plume shows in frames 97, 98, and 99. The B/G ratios for the plume are 5.5, 6.0, and 6 . 2 Respectively ; the B/G ratios for the Gulf Stream water adjacent to the plume are 6.3, 6.8, and 7.1, respectively. Thus, while the absolute values of the ratio are changing, the difference between the plume ratio and the Gulf Stream water ratio is nearly constant from frame to frame. Both filtered panchromatic films SO-0 2 2 showed some apparent density changes in the same areas as those in roll 64. Roll 65, filtered to pass 0.6 to 0.7 micron light showed a transmission change from 40.0x10"! to 34.0x10"-'- on going from the plume off Cuba to adjacent Gulf Stream water. Roll 66, filtered for 0.5 to 0.6 micron light showed no transmission change between these two areas, both noted as 34xl0~l. Neither of the two b/w IR films showed any transmission differences among the areas analyzed. The color IR film did show some differences that are considered to be statistically significant. The plume off of Cuba showed a B/G ratio of 7.0, while adjacent Gulf Stream water showed a ratio of 7.5. The plume off the Florida Keys showed a B/G ratio of 8.5, while across the interface of the Gulf Stream the water showed a ratio of 7.2. It should be noted that the B/G ratio of the color infrared film is really a ratio of the green to the visible red radiation. The conclusions that can be drawn from this limited data set are that a significant variation in ocean color can be observed by changes in dye concentrations in color photographs of the scene. When the surface truth is considered the evidence tends to favor the variation in suspended chlorophyll as the most probable cause of the color variation. 3 . 3 Discussion It is immediately apparent from the data that the transmission ratio technique is a useful means of analyzing variations in color of satellite-derived photography. At the same time the data 147 might have been more useful had certain precautions been taken. Reference is specifically made to the aircraft-derived photography. To achieve a flat spectral response across the film the associated optics should have been fitted with anti-vignetting filters. The space craft cameras suffered to a lesser extent with the same problem. In the latter case each of the associated optics had been calibrated so that the error in transmission was known. There is however, no indication that such care was taken in prep- aration of the subsequent duplicate images. While care was taken to ensure that duplicate grey scales were reproduced at the same levels as those on the on-board films, apparently no account was taken of the variation in illumination across the print head of the printer. All of these problems taken together substantially reduce the possible intercomparisons that might have been attempt- ed. The photographs in Figs. 3.2 and 3.3 are both S-190B products that have been enhanced by printing on high-contrast film. Ex- posure levels were set to saturate the details in the non-oceanic features. This is a trial and error technique that extracts markedly more low radiance level information. The change in texture marking differing sea states in Fig. 3.2 is not measurable by the densitometer technique, but it is clearly noticeable to the eye. The boundary between the two levels of radiance is probably the anticyclonic edge of the Florida Current. The de- tection of the features in Figs. 3.2 and 3 . 3 by the S-192 scanner is discussed in section 5. 4. SPECTROMETER EXPERIMENT The SKYLAB S-191 steerable spectroradiometer was to be used in this experiment to study changes in the visible (0.4-0. 7ym) and infrared (7.0-14.0 ym) spectra of the ocean across the current's cyclonic boundary. The plan was for the crew to acquire a cloud- free oceanic area with the S-191 looking 45° forward of nadir, and to track that site until 0°. Thereafter, the spectro- radiometer was to be locked into nadir viewing across the cyclonic front and up into the waters of Florida Bay. The experiment proved to be not very successful for several reasons: although the crew did as the plan said, the data acqui- sition camera (DAC) was turned off and the exact tracking data (angles, times, locations) were never recorded; the calibration of the S-191 infrared detectors is not known; the visible region data radiance values do not agree with theoretical or observed values reported by other investigators . 4.1 Tracking Data Location of the data was made difficult by the lack of DAC output. The voice log was the best clue to what actually was done by the crew. According to the transcript of the voice tape, at 148 16:29:33 GMT the pilot had the S-191 set at 35° looking forward along the track, although, all :rew instructions were to set the S-191 target acquisition at 45°. The word "thirty-five" was not clearly audible however, and the pilot may have followed the in- structions sent up to SKY LAB just prior to the pass. At 16:29:35 GMT, the pilot reported tracking a clear area of water; the assigned start time was 16:29:33. The exact time of reaching nadir is difficult to tell from the voice log however, 16:30:45 is the approximate time. The location and time of the nadir point were calculated from geometrical considerations assuming a spherical Earth with a ra- dius of 6 378 km and a satellite altitude of 443 km. If the nadir angle was 45° at 16:29:33 GMT, the position of the point tracked was 2 3°53'.4 N, 81°5 8'.0 W; the time of arrival of the spacecraft over this point from the best available positioning data was 16:30:41.1 GMT, which is in good agreement with the voice log estimate. The message sent up to the crew had the finish of the tracking at 16:31:05. The S-19 2 line-straightened data show that the position given above was in the middle of a clear ocean area and it appears reasonable to have tracked this as the site. The vehicle was over Florida Bay at 16:30:55 according to the S-192, but the pilot commented at 16:31:10 that they were going across the Keys. This discrepancy cannot be accounted for unless the Florida Keys were observed well after the spacecraft transit. If the above analysis is correct, then according to the sur- face truth data in Fig. 2.1, the S-191 probably never acquired data from the Gulf Stream. The position of the 22°C isotherm at 10 0 meters depth indicator was 2 5 km SW of the point where the pilot tracked a clear area. Although the exact location of the front cannot be identified in the ship track data it appears that it was also SW of the nadir tracking point. Maul (19 75) reported the mean separation between the indicator isotherm and the front to be 11 km in this area, and that further supports the contention that the S-191 did not obtain spectra in the Gulf Stream. The objective of analyzing the change in spectra across the front can- not be accomplished with these data. 4.2 Infrared Radiance The infrared experiment was designed to study the accuracy of atmospheric transmission models. This objective could not be accomplished because the calibration of the infrared detector is an unknown function of wavelength (Barnett ,NASA-JSC personal communication; Anding, and Walker, 1976). Several relative tests were made however, which provide some information on the atmos- pheric transmission model dependency, and these will be discussed below. 149 4.2.1 Theoretical calculations Emitted infrared radiation (7 ym<_A<_14y) leaving the Earth passes through the atmosphere before detection at the S-191 sensor. The atmosphere modifies the infrared radiation by ab- sorption and, to a very minor degree, by scattering. Details of the theory are given by Chandrasekar (1960) and recent reviews on its application to oceanography are given by Hanson (19 72) and Maul (1973). The radiative transfer equation through an absorbing but non-scattering atmosphere is: r */ N(6,A) = e(9r , A) L(T,A) t(6,A) P s L(Ta, P,A) 9-r(p,6,A) p 9P o p(6' ,A) N (G",A) t(6,X) (4.1) as where 9,0', A" are the nadir angle, angle of reflectance, and angle of incidence, respectively. Radiance (N) at the satellite is wavelength-dependent, and is a function of the surface black- body radiance (L), the emissivity of the surface (e) and the transmittance of the atmosphere (t); these three _ parameters describe the absorption of emitted blackbody radiation by the atmosphere. The second term in the equation, the integral term, describes the atmospheric (a) modification of the radiance as a function of pressure (P). The third term describes the contribu- tion of the reflected (p) atmospheric radiance at the surface (N ) , again as modified by transmittance. as The theoretical calculations discussed herein are an ex- tension of the model used by Maul and Sidran (19 73) which uses the transmissivity data of Davis and Viezee (1964). The area of interest is the 10.5 - 12.5 ym band that is used on many space-- craft including the SKYLAB S-19 2 multispectral imager. In this spectral interval e>0.99 at low nadir aggies; hence p(=l-e) is very small and equation (4.1) may be written N(6) = 4>(A) L (Ts,A) x(9,A)dX 00 ps U (A) L (Ta, p,X) 3r_(p,6,A) dpdA (4.2) 'o /0 ap The filter function () is zero outside the interval discussed above. The radiance may be converted to equivalent ^blackbody temperature by inverting the Planck equation (L) which has been integrated over the same 10 . 5<_<_12 . 5 interval. The calculations were carried out on the AOML computer. A special radiosonde was released by the Key West office of the 150 National Weather Service at tha time of SKYLAB transit Csee Fig. 4.1). Before the radiative transfer from a radiosonde is computed the data must be inspected to insure that no clouds are in the path of ascent, in order to compute a cloud-free radiance. Clouds are ^ readily identified by their characteristically high relative humidity and _ isothermal temperature. There is evidence of clouds in the data in Fig. 4.1, so calculations were made to test the effect of clouds. RELATIVE HUMIDITY (%) 0 20 40 60 80 AIR TEMPERATURE 1000 100 -I — -80 -60 -40 -20 0 *20 AIR TEMPERATURE (°C) ►40 Figure 4.1 Vertical ■profiles of atmospheric pressure and rela- tive humidity taken at the times of SKYLAB transit. The dotted lines on the relative humidity profile are the oloud-free estimate of atmospheric moisture. Two cloud layers are in evidence, one centered at 744 mb and one centered at 6 71 mb . Clouds are characterized by a sudden increase in relative humidity and a small (near zero) lapse rate. An equivalent clear sky estimate is made by assuming the clouds are absent; the estimated relative humidity profile in the clouds region is given by the dotted curve. The calculated equivalent blackbody temperatures for TQ 2 9 8.15°1C are: Wavelength Observed (Appendix B) Cloud-free Equivalent llym 12. 5ym 293.22° 290.46° K K 293.85° K 291.28° K The differences in this case are small, 0.6 3°K at llym, and 0.8 2°K integrated over the 10.5-12.5ym region where the S-19 2, NOAA-4/5, and SMS-1/2 observe. Other experience with this type of cloud- free equivalence has been as high as 5°K over the Gulf of Mexico. 151 4.2.2 Comparison of S-191 and models As stated in section 1.3, the wavelengths chosen for the two-channel technique, CAnding and Kauth., 1970) of atmospheric correction depend on the radiative transfer model. SKYLAB was to be used to study that question but since the calibration of the S-191 infrared detector is unknown, the problem cannot be investigated. The mean sea surface temperature along the trackline was 2 5.0°C. This value has been used in the calculations shown in Fig. 4.2. The Davis and Viezee (1964) model does not include absorption due to the ozone molecules which show up as a maximum at 9.6ym in the S-191 observation. The comparison shows that ozone does not affect the 10.5-12.5um window and hence is not a factor in the S-19 2 infrared scanner data. At 11.0ym, the ap- parent difference between the observed and calculated equivalent blackbody temperature CTgg) is 3.5°C. This seems to be the approximate error estimate of other SKYLAB investigators (personal communications), but no conclusions can be drawn. CALCULATED 7.0 8.0 9.0 100 1 1.0 12.0 13.0 14.0 15.0 WAVELENGTH (/im) Figure 4.2 Spectral infrared radiance observed by the S-191 spectroradiometer (dashed line) 3 and calculated by the ozone excluding model of Davis and Viezee (fine solid line). Heavy solid lines are blackbody curves. Since the calibration uncertainity is wavelength-dependent the data at llym were studied to determine the shape of the nadir angle dependence curve. In the lowe^ half of Fig. 4.3 is a least squares fourth-order polynomial fit to all the observed radiances as a function of time (dashed line). Since the same ocean spot was to be tracked, radiance should be a function of nadir angle up to 16:30:45 GMT. The maximum on this curve (arrow) is at 152 16:30:19. The upper curve is the theoretical calculation using equation 4.2 for the same atmosphere in Fig. 4.1 and for T=25°C. Nadir angles were computed using a start time of 16:29:35 GMT (45°) and a stop time of 16; 30: 45 GMT (0° ) following the discus- sion in section 4.0. The match in the curves maxima would be approximately coincident if the tracking started with a 3 7° nadir angle, which is in agreement with the voice tape transcript. NADIR ANGLE 45* 40* 35* 30° 25* 20° 15° 10* 5» 0° E — I- 1 — i i i i i > 1 l (§j 860 -CALCULATED _ E 850 _ - T* S -» ■i_ i .,. . 830 7E o 820 - ■ . • ."*"•• • ~ $ -*'u> • m\ 4. I1* . >'^^' * • *~*^^- • "' — 810 LlI ~ 0|0 • -re...- •♦( ... MAXI IMAX • . \^7 O 1 • Z 800 - A ' - < / t \ O «,„.. • r < 790 t * ~ or i ' i 1 1 1 1 i i I6'29'30 40 50 l&30'00 10 20 30 40 50 I&3I-00 GMT Figure 4.3 Radianoe at 11 \im as a function of nadir angle of the same ocean spot as that tracked on the S-191 (dots). The dashed vertical line separates channel Al and channel A6 data. Dashed curve is fourth-order polynomial fit to all data; fine solid line is fourth-order polynomial fit to A6 data only. Heavy solid line is the calculated nadir angle dependence. The S-191 spectroradiometer uses a series of detectors that cover a segment of the spectrum. In some regions these overlap and ambiguity often exists as to which detector to use. In the sections to follow, those detectors (channels) chosen were as recommended in the NASA reports on instrument performance. It is suggested that only those data that are well calibrated be reported to non-instrument engineering investigators in the future . Barnett (personal communication) cautioned against the use of radiometer channel Al in the S-191 (see again Fig. 4.3). Accordingly a second fourth-order polynomial Csolid line) was fitted to the channel A6 data only. The rms spread of the ra- diance about this polynomial is 4.48yW cm~2 sr~i. This corresponds to a noise-equivalent temperature difference (NEAT) of +0.3°K at 2 89. 6 5° K. Since the atmospheric attenuation tends to diminish surface gradients, this results in a calculated NEA.T of 153 approximately +_0.6°C in T for this model at 11pm on this day. The equivalent blackbody temperature at the maximum in the. poly- nomial is 16. 5° C at the top of the atmosphere. This implies a temperature correction of 8.5°C which is not unreasonable for a tropical winter atmosphere whose precipitable water vapor is 3.6 cm (cf. Maul and Sidran, 1973). 4. 3 VISIBLE RADIANCE The visible radiance experiment was designed to study the accuracy with which spectral changes across oceanic fronts can be observed and interpreted from satellite altitudes. Unfortunately, the strongest front expected in the experiment area was missed by the S-191 so this objective, as discussed before, could not be accomplished. However, during the course of this work, a theore- tical technique for recovering the "ocean color spectrum" through the atmosphere was developed. This is discussed in detail below and an attempt is made to compare the predictions of the theory with the S-191 data and the associated ground truth. 4.3.1 Theoretical calculations It is clear that the full potential of oceanic remote sens- ing from space in the visible portions of the spectrum can be realized only if the radiance that reaches the top of the atmos- phere can be related to the optical properties of the ocean. To effect this , the radiative transfer equation must be solved for the ocean-atmosphere system with collimated flux incident at the top of the atmosphere. In such calculations the optical proper- ties of the ocean that must be varied are the scattering phase function (PQ(6)) and the single scattering albedo (wQ; defined as the ratio of the scattering coefficient to the total attenua- tion coefficient). Furthermore, unless the ocean is assumed to be homogeneous , the influence of vertical structure in these properties must be considered. To describe the cloud-free at- mosphere, the optical properties of the aerosols and their variation with wavelength and altitude as well as the ozone concentration must be known. Considering the ocean for the pres- ent to be homogeneous, the radiance at the satellite can be related to the ocean's properties by choosing an atmospheric model and solving, the transfer equation for several oceanic phase functions and oo0's at each wavelength of interest. The number of separate computational cases required is then the product of the number of phase functions, the number of values of w , and the number of wavelengths. Even if the multi-phase Monte Carlo method (MPMC) (Gordon and Brown, 1975) is used, the co resolution of Gordon and Brown C19 7 3) would require a number or simulations equal to ten times the number of wavelengths for each atmospheric model considered. It is possible, however, to obtain the necessary information without modeling the ocean's optical properties in such detail. 154 Th_e model is based on an observation evident in results of computations given by Plass and Kattawar C19 69) and by Kattawar and Plass (197 2) on radiative transfer in the ocean-atmosphere system, namely, that when the solar zenith angle is small, the upwelling radiance just beneath the sea surface is approximately uniform, (i.e., not strongly dependent on viewing angle) and hence determined by the upwelling irradiance . This observation is utilized in simulations of oceanic remote sensing situations by assuming that a fraction R of the downwelling photons are absorbed. The ocean is then treated as if there is a Lambertian reflecting surface of albedo R just beneath the sea surface. In this case Gordon arid Brown (19 74) have shown that any radiometric quantity Q-, can be writte n Qo R Q = Q, + J_ (4.3) 1-rR Qj_ is the contribution to Q from photons that never penetrate the sea surface (but may be specularly reflected from the surface) . Q2 is the contribution to Q from photons that interact with the hypothetical "Lambertian surface" once for the case R=l. r is the ratio of the number of photons interacting with the "Lambertian surface" twice, to the number of photons interacting once, again for R=l. By use of equation 4.3, any radiometric quantity can then be computed as a function of R. Physically the quantity R is the ratio of upwelling to downwelling irradiance just beneath the sea surface and is known as the reflectance function [R(0,-)] in the ocean optics literature (Preisendorfer , 1961). Spectral measurements of the reflectance function R(A) have been pre- sented for various oceanic areas by Tyler and Smith (19 70). Henceforth, R(A) will be referred to as the "ocean color spectrum" A series of Monte Carlo computations have been carried out to see if an approximate simulation (AS1) , using this assumption of uniform upwelling radiance beneath the sea surface, yields results that agree with computations carried out using an exact simulation (ES) , in which the photons are accurately followed in the ocean as well as the atmosphere. The Monte Carlo codes used in Gordon and Brown (19 73, 19 74) were modified by the addition of an atmosphere. The atmosphere consisted of 50 layers and includes the effects of aerosols, ozone, and Rayleigh scattering, using data taken from the work of Elterman (1968). The aerosol scat- tering phase functions were computed by Fraser (NASA-GSFC, personal communication) from Mie theory assuming an index of re- fraction of 1.5 and Deirmendjian ' s (1964) "haze-C" size distribution. Also, to determine the extent to which the vertical structure of the atmosphere influences the approximate simulation, a second approximate simulation (AS2) was carried out in which the atmosphere was considered to be homogeneous; i.e., the aerosol scattering, Rayleigh scattering, and ozone absorption were inde- pendent of altitude. The oceanic phase functions in the ES 155 are based on Kullenherg's C19 6 8) observations in the Sargasso Sea, and are given in Table 4.1 CNote that all the phase functions in the present paper are normalized according to 27T/71" PCB) sin d9 = l). o Table 4.1 The Three Ocean Scattering Phase Functions e KA KB KC (deg) (xlO2) (xlO2) (xlO2) 0 10924 10171 9521 1 4916 4577 4285 5 573.5 534.0 499.9 10 169 .3 157.7 147.6 20 29.5 29. 39 29.31 30 12.56 11.9 5 11.42 45 3.059 3.661 4.189 60 1.092 1.577 1.999 75 0.546 0.915 1.190 90 0.344 0.661 0.952 105 0.311 0.641 0.928 120 0. 317 0.732 1.094 135 0.410 0. 829 1.309 150 0.492 1.017 1.618 16 5 0.579 1.261 1.856 180 0.617 1.357 1.999 KA is roughly an average of Kullenberg's phase function at 632.8 nm and 655 nm, and KC is his phase function at 460 nm. KB is an average of KA and KC . These phase functions show con- siderably less scattering at very small angles (8<1°) than was observed by Petzold (19 72) in other clear-water areas; however, the exact form of the oceanic phase function is not very important, since it has been shown (Gordon, 1973) to influence the diffuse reflectance and R(0,-) only through the back-scattering probability (B) J TT. = 2tt f P (6) sinede. 72 In all of the computations reported here the solar beam incident on the top of the atmosphere is from the zenith, and _ with unit flux. At visible wavelengths the variable atmospheric constit- uent that will most strongly influence the radiance at the top of the atmosphere is the aerosol concentration, so the computations have all been carried out as a function of the aerosol computa- tion . Table 4.2 gives a sample comparison of upward fluxes at the top of the atmosphere at 40 0 nm in the three simulation models (ES, AS1, and AS2) as a function of the aerosol concentration. N, 3xN, and lOxN refer to aerosol concentrations in each layer 156 of 1, 3, and 10 times the normal concentration given by Elterman. 400 nm is chosen because in the visible portion of the spectrum it is the wavelength at which the atmospheric effects are ex- pected to be most severe. The ES case uses w = 0.8 and phase function KC. The values of R used to effect the AC computations were taken from the EC computation of this quantity. However, if R is taken from 3 R = 0.0001 + 0.3244x + 0.1425x + 0.1308x (4.4) where x = u> B/Cl-u (1-B) which, according to Gordon, Brown and Jacobs (1975), reproduces the in-water reflection function for the corresponding case but with no atmosphere present, the results of the AS model computations agree with those listed to within 0.2%. The numbers in the parenthesis next to each flux value represent the statistical error in the flux based on the actual number of photons collected in each case. It is seen that ES and AS simulations generally agree to within the accuracy of the computations. Notice also the excellent agreement between the AS1 and AS2 fluxes. Table 4.2: Comparison of the flux at the top of the atmosphere for the ES, AS1, and AS2 simulations . Aerosol Concentration ES AS1 AS2 N 0.222 (+.002) 0.224 (+.001) 0.226 (+.001) 3xN 0.274 (+.003) 0.273 (+.001) 0.275 (+.001) lOxN -.423 (+.004) -.426 (+.002) 0.425 (+.002) Fig. 4A presents a comparison between the ES, AS1, and AS 2 upward radiances at the top of the atmosphere. The step-like curve in the figure is for ES , the solid circles for AS1, and the open circles for AS2 , and u is the cosine of the angle be- tween the nadir and the direction toward which the sensor is viewing. The radiances in Fig. 4.4 for the ES cases are accurate to about 3% in the range y=l to about 0.4, while for the AS cases the accuracy is about 1%. To within the accuracy of the computa- tions , the three simulations again agree for all the aerosol concentrations except within the range y=0 to about 0.3; i.e. viewing near the horizon. These computations appear to demons strate that the transfer of the ocean color spectrum through the atmosphere can be studied with either the AS1 or AS2 model as long as radiances close to the horizon are not of in+^rest. Furthermore, from the reciprocity principle (Chandrasekhar , 1960) 157 the nadir radiance, when the aolar heam makes an angle 6Q with the zenith, can be found by multiplying the radiance I Cy) in Fig. 4.4 by p where p is taken to be cos 9 . This implies that as long as the Sun is not too near the horizon, the AS1 and AS 2 methods of computation can be used to determine the nadir radiance at the tip of the atmosphere as a function of the ocean's prop- erties through equation 4.4. The fact that the AS2 model (homogeneous atmosphere) yields accurate radiances is very im- portant in remote sensing since it implies that only the total concentration (or equivalently the total optical thickness) of the aerosol need be determined to recover the ocean color spec- trum from satellite spectral radiometric data. 015- » Figure 4.4 Comparison between ES (step-like curve) AS1 (solid circles) and AS2 (open circles) upward radiances at the top of the atmosphere for an ocean with w0 * 0.8 and phase function KC and an atmosphere with a normal (lxN)3 three times normal (3xN) and ten times normal (lOxN) aerosol concentration. 158 It should be noted that these results also strongly suggest that R(A) is the quantity relating to the subsurface conditions that can be determined from space, and hence, is. the most natural definition of the "ocean color spectrum". Moreover, it has been shown (Gordon, Brown, and Jacobs, 1975) that R(A) is not a strong function of the solar zenith angle Cthe maximum variation in R(0,-) with 60 is of the order of 15% for 0 R^ y0I0(yn)AR. (4.8b) nadir ° ° -020 Clearly, for a given AR, Ina^ip decreases substantially with increasing solar zenith angle because of the presence of the y0 factor in equation (4.8b). For example, with a three-times normal aerosol concentration, a nadir-viewing sensor would need about 2.5 times more sensitivity at 8o = 60° as compared with 6o~0 to detect the same R. The above examples indicate how the theory (AS1) can be used in the design of a satellite sensor system for estimating some ocean property such as the concentration of suspended sediments or organic material. Specifically, one must first determine the effect of the property on R. Then, on the basis of the sensi- tivity desired, find AR, and finally, use equation (4.8a) or (4.8b) to find the minimum radiance change the sensor must be capable of detecting. If the sensor has a limited dynamic range, then equation (4.5) can be used with equation (4.8a) or (4.8b) to aid in the sensor performance design trade-offs. EUnfortunately at this time, relationships between R(X) and sea- water constituents are not well established.] Considering the fact that we have used only the "haze-CM aerosol phase function (which is clearly only approximately characteristic of the actual aerosol scattering) it is natural to inquire how strongly the computations of Ii(y) and I;?(y) presented in Figs. 4.5 and 4.6 depend on the shape of the aerosol phase function. To effect a qualitative understanding of the influence of the aerosol phase function, computations of I]_ and 1 2 have been carried out using the well known Henyey-Greenstein (HG) phase function P ( 6) = (l-g2)/4TT HG (l+g2-2g cos 6)3/2 , where the asymmetry parameter g is defined according to g = 2-rr / PC6) cos 6 sin 6 d 6, and 6 is the scattering angle. Since g for the haze-C phase 162 function is 0.69Q, computations haye been made with P^p C 6) for figure 4 . 7 compares these P„p ( 8) ' s g values of 0.6, 0.7, and 0 . with the haze-C phase function. The HG phase function for g=0.-7 clearly fits the haze-C phase function quite well in the range 5°<0^140°; however, as is well known, the HG formula is incapable 10 i - S 10"* 10'- 1 i i i — r i i i i i i i i i i i i i ••• "HAZE C" • \ HENYEY- GREENSTEIN \ \ •v • ^^^9 = 0.7 g=0.8 iiii • i • i i 'I' I0"3 0 20 40 60 80 100 120 140 160 180 9 (degrees) Figure 4. 7 Comparison between the "haze-C" and various Eenyey- Greenstein phase functions characterized by asymmetry parameters 0.63 0.7 j and 0.83 as a function of scattering angle ( 6 ). of reproducing phase functions computed from Mie theory in the extreme forward and backward directions. The HG phase functions with asymmetry parameter 0.6 and 0.8 are seen to be substantially different from the haze-C distribution at nearly all scattering angles. On the basis of Fig. 4.7 it should be expected that I]_ and I2 computed with PhG^6^ will be in close agreement with the haze-C computations only for g close to 0.7. Figures 4.8 and 4.9, which compare the results of computations of Iq_ and I2 for P C e) for the normal aerosol concentration, show that this is indeed the case. It is seen that except for apparent statistical fluctuations, the HG phase function for g=0.7 yields values of 163 II 10 i 1 1 1 1 1 1 ••• "HAZE C HENYEY - i .■I GREEI 1 1 OSTEIN r— — - r _• X = 400nm 9 — r-= 8 • — r — r i i i i i i ■ 7 g=o.6 • o g=o.7 O 6 • • X • g»o.8 h-r 5 J 4 3 2 1 - " i i i ' 1.0 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 Figure 4.8 Comparison between Ij(\x) computed for the "haze-C" and Henyey-Greenstein phase functions for an atmosphere with a normal aerosol concentration, as a function of cosine Q(\i) . wavelength of calculations is 400 nanometers. I]_ and 1 2 in good agreement with the haze-C computations. This suggests that the detailed structure of the phase function is not of primary importance in determining I]_ and l2» and it may be sufficient for remote sensing purposes to parameterize the phase function by g. To get a feeling for the importance of variations in the phase function in the remote sensing of ocean color, consider the effect of changing the aerosol phase function from an HG with g = 0.6 to one with g = 0.8 over an ocean with R = 0.1. From Figs. 4.8 and 4.9 it is found that the normalized radiance at \x - 0.85 (the assumed observation angle) decreases by 4. 9x10 3. This decrease in radiance would be interpreted under the assump- tion of no atmospheric change as a decrease in R from 0.10 to 0.056. This clearly indicates then that variations in the aerosol phase function in the horizontal direction could be erroneously interpreted as horizontal variations in the optical properties of the ocean. However, it is probably unlikely that the clear 164 12 II 10 9 8 O Q 7 X i— i 6 5 4 3 2 ■fc-TTL T 1 1 1 1 ••• "HAZE C" HENYEY-GREENSTEIN X = 400nm 9=0.8 am. glofi 1.0 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 Figure 4.9 Comparison between IgfyJ computed for the "haze-C" Eenyey-Greenstein phase functions for an atmosphere with a normal aerosol concentration 3 as a function of cosine 6 (v-) ; wavelength of calculations is 400 nanometers. atmospheric oceanic aerosol phase function will exhibit varia- tions as large as that considered in this example, except in extreme cases. Assuming that the aerosol concentration of the atmosphere can be determined, the uncertainity in the aerosol phase function will still of course provide a limit to the ac- curacy with which the ocean color-spectrum can be retrieved from satellite radiance measurements. In summary then, the theory CAS1) leads to the natural definition of RC*) [RC0,-)J as a function of wavelength]] as the "ocean color- spectrum" . The determination of subsurface oceanic properties from space can thus be divided into two problems: 165 1) the determination of RCA) from satellite radiance measure ments , and 2) th_e establishment of relationships between RO) and the desired ocean properties. Since the method of computa- tion conveniently separates the radiance into a component that interacts with the ocean CI?) and a component due to reflection from the atmosphere and sea surface CI-, ) , it is easy to relate changes in radiance to changes in R(A). It is found that for viewing angles up to 35° from nadir, In is a relatively weak function of the aerosol concentration for concentrations up to three times normal. This suggests that spatial gradients of R(A) can be determined with only a rough estimate of the aerosol concentration. It is further found that variations in the aerosol phase function can strongly influence the interpretation of the radiance at the satellite. Clearly then, it is vital to under- stand the magnitude of aerosol phase function variations. 4.3.2 Technique for Atmospheric Correction As discussed in section 4.3.1 it is necessary to know the aerosol concentration in order to recover the ocean color- spectrum R(A) from the nadir radiance spectrum observed at the satellite. In this section a method based on Curran's suggestion of using the near-infrared radiance to determine the concentra- tion is developed and applied to the S-191 data. The method involves finding a band of wavelengths in the near infrared for which the absorption by ozone and water vapor is negligible. Since the Rayleigh scattering by air is very small in the near-infrared, the greatest contributor to the optical thickness at the wavelength in question is the aerosol. It is found that at 7 80 nm ozone and water vapor do not absorb significantly, and the Rayleigh scattering contributes only about 0.023 to the total optical thickness of the atmosphere. This implies that aerosols play the dominant role in the radiative transfer here with the normal aerosol concentration yielding an optical thickness of about 0.2. Also since R(0,-) for wave- lengths greater than about 70 0 nm is essentially zero, the upward radiance at the top of the atmosphere at 7 80 nm simply becomes I(y) = I-,(y) F° I where FQ is the solar irradiance (mW cm"2ym" ) at the top of the atmosphere CKondrat'ev 19 73). By use of the reciprocity principle for nadir viewing and any solar zenith angle ( 6Q ) , Ina(^-Lr can be written I . = u _F I-i Cp « ) • nada_r o lv-po/ IjCjj) and I2GO have been computed for aerosol concentrations OxN, lxN, 2xN, and 3xN at 400 nm, 500 nm, 600 nm, and 780 nm. 166 The results for the OxN and lxN computations are presented in Appendix C. Using I-j_Cy) for 780 nm and noting that the S-191 nadir radiance was recorded with 8 40° the upward radiance at the top of the atmosphere for nadir viewing is found to be 0.32 3 and 0.9 38 mWcm-2 sr~lym~l for aerosol concentrations of OxN and lxN respectively. Using the S-191 radiances at 780 nm for the nadir viewing spectra taken on Jan. 8, 19 75 at 16:30:45.75 GMT (spec- trum A) and 16:30:52.2 GMT (spectrum B) which respectively were 0.64 and 0.72 mW cm-2 sr-1, it is found that the theory suggests the aerosol concentration was 0.51xN for spectrum A and 0.64xN for spectrum B. In order to compute R(A) from the S-191 data the assumption is made that the variation of the aerosol extinc- tion coefficient with wavelength is exactly as given by Elterman. 4.3.3 Recovery of R(A) from the S-191 data As mentioned above, in order to recover R(A) from the S-191 data it is necessary to assume that the variation of the aerosol extinction coefficient with wavelength is identical to that given by Elterman. Also, since I-j_(y) and T^Cy) at 7 80 nm were derived using the haze-C phase function for the aerosols, the assumption is implicit that this phase function is correct. With these assumptions the nadir radiance at the top of the atmosphere has been computed at 400, 500, 600, and 7 80 nm for aerosol concentrations OxN and lxN , assuming that R(A) is zero. These radiances are presented in Table 4.3 along with those from spectra A and B. Since the actual aerosol concentration is known to be be- tween OxN and lxN, it appears that the S-191 data at 400 nm are in error. It is virtually impossible for the nadir radiance to be less than that for a OxN atmosphere. (It should be noted that the discrepancy here is great, i.e., the S-191 radiances at 400 nm appear to be too small by more than a factor of 2.) The radiances at the other wavelengths listed in Table 4.3 seem to be reasonable and were used in equation (4.7) to estimate R(\) . The results are shown in Table 4.4. It is seen that R(A) is negative except in the spectral region 500-550 nm where the values shown compare well with the Tyler and Smith Gulf Stream data for R(0,-). As discussed above, the 400 nm data are apparently in error. However, the data at other wavelengths appear realistic, so the negative R(0,-) values are probably due to the assumptions that the haze-C phase func- tion characterizes the aerosol, and that the spectral variation of the aerosol scattering coefficient is correctly described by Elterman' s data. It is clear that considerably more experimental work is needed to test the ability of the theory discussed in 4.3.1 4o obtain an accurate RCA) from the satellite radiance. 167 Table 4.3 Wavelength F nadir (R=0) (nm) lmW cm" • 2 -1 mW cm" 2 ym sr~l 157 OxN lxN Spect A Spect B 400 5.61 6.58 2.30 2.50 500 201 2.93 4.10 4.13 4.39 600 184 1.25 2.13 1.55 1.67 780 125 0. 323 0.938 0.64 0. 72 Table 4.4 Wavelength R(0,-) (nm) Spectrum A Spectrum B 400 -0.274 -0.268 450 -0.0242 -0.0235 500 0.0318 0.0370 550 0.0295 0.0303 600 -0.00715 -0.00547 780 0 0 5. MULTISPECTRAL SCANNER EXPERIMENT SKYLAB's multispectral scanner was a unique design that had 13 spectral channels of data spread over the visible and infra- red bands. The system used a conical scan which had the ^ advantage of keeping the atmospheric path length the same at all times. The visible region of the spectrum (0.4 - 0.75 ym) was divided into 6 channels, each about 0.05 ymwide. Two reflected infra- red (0.75 - 1.0 ym) channels and one in the emitted infrared (10.2 - 12.5 ym) were also provided. The channels useful to Table 5.1. LANDSAT-1 has been shown to have several useful applications of visible region imagery to marine science (Maul, 1974). The much finer spectral resolution of the S-19 2 provided an opportun- ity to expand those results to ocean current boundary determination and to test if the lower wavelength (0.0 5ym) inter- vals were useful through the intervening atmosphere. 168 Tahle 5 . 1 Spectral Channels Useful for Oceanography BAND DESCRIPTION RANGE (ym) 1 Violet 0.41 - 0.46 2 Violet-Blue 0.46 - 0.51 3 Blue-Green 0.52 - 0.56 4 Green-Yellow 0.56 - 0.61 5 Orange-Red 0.62 - 0.67 6 Red 0.68 - 0.76 7 Reflected Infrare d 0. 78 - 0.88 8 Reflected Infrare d 0.98 - 1.03 13 Thermal Infrared 10.2 - 12.5 5.1 S-19 2 Data S-192 data were collected from 16:29:22 GMT (over the open sea just north of the Cuban coastline) to 16:31:04 GMT (over the mainland Florida coast north of Florida Bay) . All channels listed in Table 5.2 were carefully examined in the analog format pro- vided by NASA to the principal investigator. The data in the images were compared with the S-19 0A and S-190B photographs to see if what is interpreted in section 3.2 as the anticyclonic edge of the current could be detected. This feature was not observable in the standard data product. The cyclonic edge of the stream appears to be obscured try clouds. This is often a useful means of locating the edge of the current but unfortunately made the objective of directly sensing the edge an impossibility. However, an unexpected opportunity to evaluate the S-19 2 developed by the photographic detection ( section 3) of a mass of water from Florida Bay flowing south into the Straits of Florida just west of Key West. This water is milky in appearance and somewhat greener in color. No ocean surface spectra were ob- served inside or outside of the plume of Florida Bay water, although it could have been easily accomplished if the SKYLAB crew had observed the feature and notified the ship of its presence. Upwelling spectral irradiance reported by Maul and Gordon (1975) probably describes the essential features of the plume and water in the straits. An intensive effort was made by Norris (NASA-JSC) , Johnson (Lockheed-JSC) , and Maul (NOAA-AOML) to identify from S-192 data the plume and the anticyclonic edge, using the computer enhance- ment facilities at NASA-JSC. After approximately 10 hours of 169 machine time on both, conical and line straightened data, the feature described as the anticyclonic edge could not be identi- fied, although it is clearly brought out in the photographic enhancements Csee Fig. 3.2). Further effort to bring out the anticyclonic edge was judged to be unwarranted and attention was turned to the plume feature which is visible in Fig. 3.3, and which preliminary computer enhancement showed to be a useful area in which to work. 20 15 12 10 9 8 7 6 DATA SAMPLE INTERVAL Figure 5.1 Power spectitwi of the radiance in the unfiltered S-192 ooniaal format. Significant noise is noted every 15, 8-9 } and 6 data points. Before a general computer enhancement technique was develop- ed, the data were examined for periodic features in a spectrum. Figure 5.1 is a spectrum of data specially provided for this experiment that was to be high-pass filtered only; the calibrat- ion of the S-19 2 data is considered a high-pass filter. Significant periods at about 15 data sample interval? are noted in these conical data as has been reported (Schell, Pnilco-JSC, personal communication, 1975). The line-straightened data (see Figure 5.2) have been band-pass filtered to remove this 15-data- sample periodicity. The wavy patterns near the edge of clouds are the result of filter ringing. 5 . 2 Computer Enhancement Computer enhancement of S-19 2 data was an objective of the experiment. The technique described below is a step toward automatic detection of clouds in multispectral data. The goal is to use a near infrared channel Cchannel 8 in this case) to specify where cloud-free areas are, for analysis of sea surface temperature or ocean color. Channel 8 CO. 9 8 -1.0 3 jim) is selected as the cloud discri- mination channel because there is a maximum in the atmospheric transmissivity at this wavelength, and a maximum in the 170 Figure 5.2 S-192 Line -straightened, filtered, scanner data over the Straits of Florida near the western Florida Keys. The appropriate S-192 channel number is at the top of each panel. 171 absorption coefficient of water. The high absorption coefficient of water at 1 urn causes the ocean surface to have a very low radiance when compared with, land or clouds. Thus there should be two modes in the frequency distribution of radiance: one mode for the clear ocean and another mode for land and/or clouds. An example of such a bimodal distribution is given by the histogram in Fig. 5.3. N - 2.15 /*W cm-«tr-» a* ±4.00fiW cm-'sr"' CLASS INTERVAL -0.5 01 23456789 10 II RADIANCE (^.W cm^sr"1) Figure 5. 3 Histogram (normalized to unity) of the radiance over the area shown for channel 8 in figure 5.2. The primary peak at the left is clear ocean; the broad peak centered at 7 vm' cm~2sr~l is due to clouds and land. In this figure, the low ocean radiances are clustered at the mode centered at N = 0.2 yW cm"2 sr'1. The other mode, centered at N = 5 . 7 pW cm" 2 sr"l is a contribution of the clouds. (There is no land in this example.) If these modes can be identified and separated, a statistical identification of cloud- free ocean pixels can be made. Cox and Munk (.19 5 4) observed that the radiance reflected from the ocean is essentially Gaussian in character. The problem then is to fit a curve of the form y = ni exp E-CN - N)2/sa2] 172 (5.1) to the data at the. lower valued mode. In this equation, the normal frequency curve Cy) is a function of the total number of observations Cn) , the class interval CD , and the standard de- viation (a); the overbar on the dependent variable CN) denotes ensemble average. Fitting equation C5.1) to the data is done in an iteration scheme that uses the lower valued mode as a first estimate of N. (.Only the values M +_ 2a from the original ensemble are used In this first Iteration; this eliminates many of the cloud contaminated data. ) After the first fit using a predescribed N, the scheme is_to iterate, the data using only +_2a of each new fit. When a (or N) changes less than 0.1% between iterations, the fit is considered acceptable and the cloud- free pixels are defined as those between 0_N + Ka X, - M RN+ko) - N] for (N- 24035'N 788 011.2 72 A 81°42 'W 744 009.9 66 740 008.3 72 722 008.5 30 700 007.8 31 671 005.2 55 640 003.3 42 621 003.1 22 530 -06.3 21 500 -09.4 13 470 -12.4 10 384 -24.9 14 300 -39.1 14 250 -48.9 224 -50.3 212 -48.9 200 -50. 3 150 -62.6 100 -77.1 070 -75.7 066 -75.2 061 -71.5 058 -72.7 054 -67.6 050 -67.8 045 -64.7 043 -58.1 030 -51.2 023 -46. 3 020 -47.4 017 -48.0 016 -48.2 13.5 -43.9 191 APPENDIX C - Monte Carlo Simulations This appendix lists the Monte Carlo simulations of radiances 1-^ and Io as described in Section 4.3.2. Wave' lengths at which the OxN and lxN atmospheric aerosol concentrations were computed are 400, 500, 600, and 780 nm. The cosines of ten zenith angles (m) were the in- dependent variables. 1^ and I2 are normalized to unit solar flux on a surface normal to the solar beam. 192 Wavelength = 400 nm Aerosol = OxN y !]_( y) I2 ( y) 0.00 0.10 0.20 0.30 0.40 0.50 0.60 0.70 0.80 0 .90 0 .95 1.00 .10089+00 .53255-01 .88850-01 .72697-01 .78011-01 .87278-01 .67433-01 .98883-01 .57839-01 .10661+00 .52975-01 .11159+00 .48109-01 .11406+00 .46607-01 .11634+00 .44244-01 .11615+00 .42872-01 .11701+00 .75049-01 .11782+00 193 Wavelength = 4-00 nm v 0.00 0.10 0.20 0.30 0.40 0.50 0.60 0.70 0.80 0 .90 0 .95 1.00 IxCy) .93753-01 .90653-01 .83232-01 .76449-01 .67122-01 .61046-01 .57118-01 .54662-01 .52434-01 .52278-01 .71513-01 Aerosol = lxN I2U) .51310-01 .63140-01 .77826-01 .87640-01 .96479-01 .10118+00 .10457+00 .10757+00 .10919+00 .10916+00 .10906+00 194 Wavelength = 500 nm Aerosol = OxN 0.00 0.10 0.20 0.30 0.40 0.50 0-60 0.70 0. 80 0.90 0.95 1.00 .63253-01 .59700-01 .52971-01 .91621-01 .37429-01 .10902+00 .29710-01 .12474+00 .24246-01 .13138+00 .22112-01 .13641+00 .19925-01 .13898+00 .19015-01 .14171+00 .18002-01 .14225+00 .17044-01 .14326+00 .67308-01 .14203+00 195 Wavelength = 500 nm v 0.00 0.10 0.20 0.30 0.40 0.5«0 0.60 0.70 0 .80 0 .90 0 .95 1000 i-l(p) .62318-01 .59056-01 .49220-01 .41716-01 .35218-01 .31928-01 .27678-01 .26638-01 .25573-01 .26419-01 .57385-01 Aerosol = lxN I2(u) 56038-01 75227-01 96058-01 11022+00 11781+00 12576+00 12875+00 13063+00 13228+00 13315+00 13155+00 196 Wavelength = 600 nm V 0.00 0.10 0.20 0.30 0.40 0 .50 0 .60 0 -70 0 .80 0 .90 0 .95 1.00 I^v) .27632-01 .36531-01 .29628-01 .25526-01 .21423-01 .17841-01 .16190-01 .15107-01 .14509-01 .16199-01 .52310-01 Aerosol = lxN I2(.v) .36991-01 .67189-01 .92357-01 .10823+00 .11883+00 .12614+00 .13059+00 .13303+00 .13448+00 .13524+00 .13560+00 197 Wavelength = 600 nm y 0.00 O.ffiO 0.20 0.30 0.1*0 0.50 0.60 0.70 0.80 0.90 0.95 1.00 Ix( u) .30110-01 .24508-01 .17292-01 .13224-01 .10725-01 .98865-02 .90777-02 .88360-02 .80934-02 .77006-02 .61865-01 Aerosol = OxN I2(y) .42002-01 .80947-01 .10497+00 .12254+00 .13037+00 .13672+00 .14052+00 .14304+00 .14358+00 .14664+00 .14694+00 198 Wavelength = 780 nm V 0.00 0.10 0.20 0.30 0 .40 0 .50 0 .60 0 .70 0 ,80 0 .90 0 .95 1.00 i-lCh) .27838-01 .13026-01 .76052-02 .51505-02 .42167-02 .37095-02 .34865-02 .33889-02 .30901-02 .30811-02 .65909-01 Aerosol = OxN I2(v) .69923-01 .10854+00 .13086+00 .14849+00 .15502+00 .16096+00 .16342+00 .16269+00 .16464+00 .16649+00 .16607+00 199 Wavelength = 780 rim V 0.00 0.10 0.20 0.30 0.40 0.50 0.60 0.70 0.80 0.90 0.95 1.00 i-lCv) .40506-01 .33748-01 .24905-01 .18224-01 .14512-01 .11931-01 .10491-01 .98166-02 .98476-02 .11295-01 .57149-01 Aerosol = lxN i2C u> .67910-01 .95346-01 .11832+00 .13459+00 .14180+00 .14856+00 .15236+00 .15399+00 .15395+00 .15613+00 .15511+00 200 19 Reprinted from: Marine Sediment Transport and Environmental Management, D. J. Stanley and D. J. P. Swift, editors, John Wiley and Sons, Inc., Chapter 5, 53-64. Tidal Currents HAROLD O. MOFJELD Atlantic Oceanographic and Meteorological Laboratories, Miami, Florida CHAPTER In regions where they are sufficiently strong, tidal currents constantly rework bottom sediment. Weaker currents combine with storm-generated wave motion and currents to move sediment both at the water- bottom interface and in suspension. Tidal currents are especially effective agents of sediment transport because they persist throughout the year, whereas other types of water motion, particularly storm events, tend to be seasonal. They are the background upon which are superimposed other kinds of currents causing sediment transport. The tides typically rise and fall twice a day (semi- daily tides), once a day (daily tides), or occur as a com- bination of daily and semidaily components. Figure 1 illustrates tides at different locations along the east coast of the United States and in the Gulf of Mexico. The tide is semidaily along the eastern seaboard and dominantly daily in the Gulf of Mexico (Pensacola and Galveston). As the earth rotates about its axis, the forces producing the tides move across the earth's surface from east to west. The motion of the moon around the earth and the earth-moon system around the sun produce variations of the tides and tidal currents with periods of about two weeks, a month, six months, a year, and longer. Approximately twice monthly the range of the tide is a maximum; that is, the difference in sea level between successive high tide and low tide is largest. These spring tides occur for both the daily and semidaily tides, although not necessarily on the same day. The term neap tides refers to the tides with minimum range. If the ocean covered the entire earth to a constant depth, the pattern of sea-level changes caused by the tides would be simple. However, since the oceans have complicated shapes, these patterns in the real oceans are also complicated. The global distributions of the daily and semidaily tides are given in Figs. 2 and 3. The numbers in small type along the coast and at islands in Figs. 2 and 3 are the spring ranges averaged over a year. The lines traversing the oceans in these figures give a general idea of the stage of the tide, i.e., when high, mean, and low water occur. For example, assume that in a particular region of the North Atlantic the semi- daily high water is occurring along the cotidal line marked 0 hour. Then along the 3 and 9 hour lines, the semidaily tide is passing through mean sea level; and along the 6 hour line, the tide has reached low water. About 3 hours later, high water for the semidaily tide will occur along the 3 hour line, low water along the 9 hour line, and so forth. The pattern rotates counter- clockwise around a point in the North Atlantic where the range of the semidaily tide is zero. Such a rotating pattern is called an amphidromic system, and the point is called an amphidromic point. Amphidromic systems are a general feature of both the daily and semidaily tides in the deep ocean. Amphidromic systems also occur in shallow seas, such as the North Sea, as shown in Fig. 4. The tides and tidal currents on the continental shelves and seas bordering the open seas are propagated as waves from the open oceans. These waves are partially 201 53 54 TIDAL. CURRENTS TYPICAL TIDE CURVES FOR UNITED STATES PORTS 12 13 14 15 16 17 19 19 20 Luni' dltj mil S dBChnit.on 9th ipogw 10!h lilt qulrtti I3ttv on fouilo' !6ln rww moon 20th panftt 22d mil N dtcnnttion 23d FIG! RE 1 . Predicted tides at selected ports along the I S. east coast and in the Gulf of Mexico. The range and character of tides differ significantly along coasts and between distinct regions. From the I .S. Department of Commerce Tide Tables (1974). reflected back out to sea by the shoaling bottom which rises toward the beach. The combination of the in- coming, or incident, wave and the reflected wave is called a standing tide; the semidaily tide on the conti- nental shelf between Cape Hatteras and Long Island is an example of a standing tide. The wave may also propagate along the coast, in which case it is called a progressive tide; that is, the stages of the tide progress down the coastline. An example of a progressive tide is the daily tide along the U.S. cast coast. Both of these examples can be seen in Figs. 2 and 3. The technical term for a tide that is generated elsewhere and propa- gates into a given region is cooscillation. The water on the continental shelf off the U.S. east coast cooscillates with the North Atlantic. Coastal lagoons, bays, and estuaries cooscillate with the water on the continental shelves. The tides often enter these coastal bodies of water through inlets and over bars, both of which can significantly attenuate the tides so that the tides inside the embayment are smaller than the tides along the open coast. EQUATIONS OF MOTION To compute the tides and tidal currents within a region having a complicated shape and realistic bathymetry, computer programs have been developed which use information at a number of points to compute the motion at those or other points. How well the results of the calculations describe the motion depends on the spacing between points (how well the grid of points resolves depth variations), approximations to the fundamental equations, knowledge of the motions at the boundaries of the region, and estimates of the bottom stress coefficient determining the drag of the sediment on the water. When idealized depth variations are assumed and when less significant forces are neglected, the tides and tidal currents can sometimes be described by simple formulas from which considerable insight can be gained into tidal phenomena. Historically, intensive research was done on the behavior ol tides in channels having constant depth and vertical side boundaries. The channel theory of tides is used in the present discussion and then extended to open regions such as the conti- nental shelves. In a channel where bottom stress and the Coriolis eflect due to the earth's rotation can be neglected, a tide causes the water to accelerate through the downchanncl slope of the sea surface. Horizontal differences in the resulting tidal currents in turn cause sea level to change. The interplay between these two effects produces a wave that propagates down the channel, away from the source of the tide. A general theory of waves has been presented by Mooers in Chapter 2; tides are waves whose wavelengths are long compared with the water depth but whose amplitudes are small compared with the depth. With these assumptions, the tidal motion in a channel is described by the pair of equations dt -g dx dit dx (1) (2) 202 80 •5 8 fig .» p .g -3 St !* "S ** HI 1 >-> ■* 3 x ^ (N s ^1 LU -» — * — COMPONENT OF CURRENT IN WAVE DIRECTION 4 5 6 7 6 HOURS AFTER HICH TIDE TIDAL HOURS AfTCR HW 6 5 4 3 2 10 VELOCITIES AT TIDAL.HOURS HW 777777777777/7777777 16 8 0 3 16 Cl^/SEC DISTRIBUTION OF VELOCITY COMPONENT ABOVE BOTTOM TRAJECTORY OF A WATER PARTICLE FIGURE 6. Typical tidal currents on a continental shelf where bottom stress forces the currents to zero speed at the bottom. From Fleming and Revelle (1939, p. 130); after Sverdrup (1927). With turbulent stresses, a more realistic profile of u is shown in Fig. 6, in which u decreases within the bottom boundary layer to essentially zero at the water-sediment interface. Since the tide propagates down the channel, it is a progressive tide. In this case, the maximum horizontal currents occur at higrrwater and low water where the current is in the direction and opposite to the direction of propagation, respectively. As the sea level passes through mean sea level, t; = 0, the current is momen- tarily zero. As a water parcel moves in the channel, its position X is the integral in time of the horizontal velocity: X u dt + A'0 (5) where A'0 is the position of the parcel at the initial time / = 0. Using (4), the horizontal displacement of the parcel is given by ^W'^M'-t)] + A'0 (6) Water subject to the tidal motion oscillates about an average position A'n with an amplitude equal to half the total excursion. For a semidaily tide (7" = 12.5 hours) and a tidal range 2a = 1 m, the excursion for open sea depths (h = 4000 m) is 0.35 km, whereas for shelf depths {h = 100 m) it is 2.25 km. In (6), tidal currents would produce no net displace- ment of water or suspended particles. That is, if a water parcel were tagged using dye and observed throughout a tidal cycle, the parcel would return to the same location at the end of each tidal cycle. STANDING TIDES In bays and many estuaries, the incident progressive tide is reflected. The tide in this region is the combina- tion of the incident and reflected tides. An analogous tide can be created in a channel by inserting a reflecting barrier. The sea surface displacement tj above mean sea level and horizontal water parcel velocity for the standing tide are 17 = a cos &Hf) uma\v sin(^jsinlvj (7) (8) where a is the tidal amplitude at the reflecting barrier, x is the distance seaward from the barrier, and X = (gh)lt2T is the wavelength of the tide. The tide and tidal currents are out of phase. At mean water, the strongest flood tide current occurs. It brings water into the embayment whose sea level must then rise. The incoming current finally ceases when the tide outside the embayment has reached high tide. As sea level begins to drop outside, an ebb current develops which continues until low tide outside the embayment. On the next rising tide, the flood tide again refills the embayment. An important parameter determining the characters of tides in bays, sounds (large bays), and estuaries is the ratio R = L/X, of the distance L between the reflecting barrier and the mouth of the embayment to the wave- length X of the tide. Where the ratio is small, such as a deep bay or a small indentation in the coastline, the tide as expressed in (7) has the same range as the tide outside the embayment. This is because the factor cos(27rx/X) determining the variation of the sea surface displacement t\ over the embayment is equal to unity if 27rx/X is always much less than unity. The tidal currents are small since the factor sin(27rx/X) is equal to 2irx/X, a small number. For example, a semidaily tide (T ~ 12.5 hours) in a fjord with a depth h of 500 m and a length L of 100 km has a wavelength X = (ghyi2T of 3180 km and hence a maximum tidal current of 2.8 cm/sec at the mouth of 206 EFFECTS OF THE EARTHS ROTATION 59 the fjord (x = L) if the amplitude a of the tide is 1 in. The current decreases linearly from the mouth to the head where the reflecting barrier forces the horizontal tidal current to be zero. Usually, fjords arc separated from the outside by a shallow sill which can strongly inhibit the tidal penetration into the fjord. The tidal currents over sills can reach several meters per second. The example above treats currents inside the sill. For large, shallow embayments, areal variations in tides can occur, if the length L is comparable to one- fourth of the semidaily or daily tidal wavelengths. One measure of this variation is the ratio between ampli- tudes of the tide at the head of the embayment and at the mouth: ■q{x = 0) 1 v(x = L) cos(2ttL X) (9) If L is close to X 4, the ratio is large so that the tidal range inside the embayment is much larger than the tide outside. The embayment is near resonance with the tide. The Gulf of Mexico is near resonance with the daily tides. In shallow bays, such as the Bay of Fundy, the Straits of Georgia, and Long Island Sound which are near resonance, bottom stress and other effects limit the motion. By assuming that the tides consist of progressive waves that diminish exponentially with the distance of propagation, Redfield (1950) has been able to repro- duce most of the tidal characteristics in these bays. As the incident wave propagates up a bay, its amplitude diminishes; after reflection and return to the mouth of the bay, the wave is significantly smaller in amplitude than when it entered the bay. Near the mouth the tide is progressive, whereas near the closed end of the bay it is standing. The maximum amplification of the tide in naturally occurring bays is about four times the incident tide. undistorted tides is a shallow water tide. The flood tidal current is also greater than the ebbing current. Ebb and flood channels occur in shallow water in which the tidal currents in one direction are largely confined to one set of channels and the currents in the opposite direction are confined to other channels. Ebb-flood channel systems are described in detail in Chapter 10. A shallow bar separating the shelf from an embayment offers considerable resistance to tidal flow. Large differences in sea level develop during the tidal cycles, which generate strong tidal currents. While the velocity of the water is large across the bar, the total amount of water that can flow in and out of the embayment is severely limited by the constriction. As a result, the tide in the embayment is less than it would be without the bar. In an embayment having a complicated bathymetry, a water parcel wanders into a variety of tidal regimes, such as tidal flats and channels, bars, and shoals; the simple theory predicting a return of the parcel to its original position at the end of each tidal cycle does not apply in this case. The Stokes drift induced by the distinct tidal regimes is a net drift of the parcel which may be thought of as a steady current. Under some circumstances, tides should cause a net transport of sediment through the Stokes drift. However, this phenomenon has not been adequately documented by field study. In the frictional boundary layer near the bottom where the horizontal tidal currents increase with height above the bottom, the increase in wave tidal momentum with height produces a steady current. This steady current, which can advect suspended sediment, is driven by variations in tidal momentum and is limited by turbulent friction. TIDES IN SHALLOW WATER Where the tidal range is a significant fraction of the depth, processes that cause the waveform to propagate act in the deeper water under the wave crest to move that part of the waveform more rapidly than the wave- form near the trough. The tide becomes distorted, with the slope of the sea surface greater on the leading side of the crest. Where this distortion is large, there is significantly more landward water discharge associated with the crest of the tidal wave than there is seaward discharge associated with the trough of the tidal wave. This landward net transport of water is known as Stokes drift. The difference between the distorted and EFFECTS OF THE EARTH'S ROTATION In larger bodies of water, the tides and tidal currents are subject to the Coriolis effect, caused by the earth's rotation. A moving water parcel experiences a force proportional to its speed which, looking down on the sea surface, is to the right in the northern hemisphere and to the left in the southern hemisphere. This Coriolis effect, when not counteracted by another force, drives the water in an elliptical path: the direction of the tidal current rotates clockwise in the northern hemi- sphere and counterclockwise in the southern hemisphere; the speed of the current is never zero. The semidaily tidal currents in the Middle Atlantic Bight are an example of this type of motion, shown schematically in 207 60 TIDAL CURRENTS COT10E UNES 40° 33° 30° 70° FIGURE 7. Theoretical corange chart jor the M: semidaily tide off the L'.S. east coast. Ranges are in feet. From Redfield (1958). Fig. 6; the corange and cotidal charts are given in Figs. 7 and 8. In regions where bottom stress can be neglected, the motion is determined approximately by the following equations: du -fo = ~g dt bv dt dr, dt ^1 dx + /« = ~g dr, dy = -h /du dv\ \dx+ dy) (10) (ID (12) The second terms, — fv and +/«, in (10) and (11) represent effects of the earth's rotation. There are two ways in which the Coriolis effect can alter tides and tidal current. In the case of a Poincare wave, the water parcel trajectories are ellipses whose major (larger) axis is in the direction of propagation; the ratio of the major to minor axis is the inertial period Te divided by the period T of the tidal con- stituent. This type of tide can occur only where Te > T and is generally found in exposed regions such as conti- nental shelves. The semidaily tide in the Middle Atlantic Bight (Fig. 7) is a standing Poincare wave (Redfield, 1958). In restricted cmbayments such as the North Sea (Fig. 4) or on continental shelves where the direction of propagation of the tide is parallel to the coastline, a slope in the sea surface set up against the shore can balance the Coriolis effect. The result is a Kelvin wave. For a coastline parallel to the x direction and located at y = 0, a Kelvin wave has the form (13) (14) (15) H> ~ 7 )] v = 0 208 IOTTOM STRESS 61 40° 35< 30c CORANGE UNES 70° FIGURE 8. Theoretical cotidal chart for the Mj semidaily tide of] the U.S. east coast. The cotidal lines are in hours after the Greenwich transit of the M2 moon. From Redfield (1958). The tidal currents are parallel to shore (v = 0). At a latitude of 45° and with a depth of 50 m, a Kelvin wave decays to e~x (36.8%) of its magnitude at the coast in a distance y = c/f of 286 km. Conversely, a Kelvin wave propagating at 45°N along a continental shelf 150 km wide with a depth of 50 m has a tidal amplitude 59% of the amplitude at the coast. A Kelvin wave propagating around a sea or ocean produces an amphidromic system. When a Kelvin wave enters an embayment in the northern hemisphere, such as the North Sea, it propagates counterclockwise around the embayment with the maximum tides and currents nearshore. Because the Kelvin waves do not decay rapidly away from their respective coasts, the motion at any given location is a combination of Kelvin waves. As a result, the tidal currents may not be colinear with the bathymetry. The sense of rotation of the tidal current direction is counterclockwise in this case, which is opposite to the direction for a Poincare wave on a continental shelf. 209 BOTTOM STRESS Bottom stress modifies tides and tidal currents;^ its effect is greatest where strong tidal currents occur in shallow water. To model quantitatively the stress applied by the sediment on the water above, the flow is assumed to consist of a slowly varying tidal current superimposed on turbulence. The distribution of turbulent stress within the water determines the varia- tion of tidal currents with distance above the bottom (velocity profile) and the dissipation of tidal energy. The details of flow near the bottom and estimates of bottom stress are central to the study of sediment trans- port. The turbulent stress r is often modeled as proportional to the rate of change of the current with increasing distance z above the bottom: - = A, P du (16) 62 TIDAL CURRENTS where the stress vector t is that part of the horizontal stress caused by vertical changes in the horizontal current u and Av is the vertical eddy or turbulent viscosity. There are other terms caused by horizontal variations in u which could be added to the stress, but the term in (16) dominates turbulent processes in shallow water. A layer of water will produce a force opposite to the relative motion of the water just above the layer. The slower moving water near the bottom therefore acts as a drag on the water above. In general, Av is determined by the spatial variations of currents, distance from boundaries, stratification of the water density, and the past history of the motion. In turbulent boundary layers, a sublayer near the boundary layer exists where the stress is constant and the current speed increases logarithmically with distance from the boundary: 1 ( t,, \ - 30c « = -■•(- In-— (17) k\ p ) Co where c<> is the roughness length of the boundary which is determined by bottom irregularities, t>, is the magni- tude of the bottom stress, and /, is von Kai man's constant (Ci;0.4). The effects of turbulence generally diminish with height above the water; the inertia of the water and the Coriolis effect become more important in balancing the pressure force due to the sea surface slope. In an oscillating tidal flow, the water farther from the bottom is moving faster and therefore has more inertia than tin- water near the bottom. In Fig. (> the water farther from the bottom takes longer to respond to the pressure force and lags in time the motion near the bottom. To model the attenuation of progressive tides due to bottom stress, an empirical formula is often used which relates the bottom stress t<, p to the vertically averaged tidal current I ': *b = -C/pL'2 (18) The bottom stress is proportional to the square of the tidal current and opposite in direction to the current. The stress depends on the depth // through the current I ', which is inversely proportional to y/'h. The stress is inversely proportional to // and hence is greater in shallower regions. The constant of proportionality ('., lias been found from field studies to be about 0.0025. A number of such studies are described in Proudman (1952) for shallow regions around England. INTERNAL TIDES An internal tide is a wave with tidal period, associated with displacements within the water column and with very little displacement of the sea surface. Where there are two layers, the currents are in opposite directions in the two layers. The speed of propagation in this case is / Ap h,h2 V" c = \g 7 ' aTT^] (l9) where Ap p is the fractional change in water density between the lower and upper layers, g is the acceleration of gravity, and Ai and ht are the thicknesses of the upper and lower layers. On a typical shelf with Ap/p ~ 0.002, g = 980 cm sec2, h = 10 m ,and h2 = 50 m, an internal wave would propagate with a speed c of 40.4 cm sec, which is about 60 times slower than the surface tide's speed of propagation. On the continental slope and at the shelf break, tidal currents interact with bathymetry to produce vertical displacements of density layers within the water column. The resulting undulations propagate both shoreward and seaward as internal tides. As internal tides propagate inshore, the shoaling bottom thins the lower layer and hence slows the wave. Since the wave energy then becomes more concentrated, the amplitude of the currents increases as does the dissi- pation into turbulence. Sufficiently strong currents produce an internal bore in analogy to tidal bores in rivers. The internal tide becomes a series of pulses of waves with periods of several minutes, the pulses separated in time by the tidal period. The formation of internal bores occurs when the internal tidal currents equal the speed of propagation of internal waves. Internal waves in a two-layered fluid cannot propa- gate shoreward of the intersection of the density interface with the bottom. Any internal waves that have not dissipated will lose the remainder of their energy to turbulence at the location where the water becomes unstratified. On some narrow shelves with strong stratification intercepting sharply rising bottom topog- raphy, internal tides are reflected back to sea, producing an internal standing tide. A more realistic description of internal tides requires a continuously stratified water column and the Coriolis effect. The tides then propagate in the vertical as well as the horizontal direction. Whether an internal wave as it reflects off the bottom continues to propagate shoreward, or whether it is reflected seaward, is determined by the slope of the bottom and the direction of wave energy propagation (slope of the wave characteristic). A bottom slope steeper than the wave characteristic produces reflection seaward. Smaller slopes allow the wave to continue in the incident direction. A discussion of the reflection process may be found in Cacchione and Wunsch (1974). Since the water density structure depends on the time of year, the existence and behavior of internal waves are 210 SUMMARY 63 also seasonal. In summer when the shelf water is strati- fied, internal waves ean exist over most of the shelf regions; in winter, lbs lack of stratification precludes occurrence of internal waves. ADDITIONAL READING This chapter was written to provide a qualitative intro- duction to the study of tidal currents. There is a large literature on tidal phenomena; as in any scientific field, the recent research is presented in succinct journal articles which presuppose a knowledge of the field. There are a number of texts which treat tides and tidal currents in much more detail and more quantitatively than was possible in this chapter. The general texts by Sverdrup et al. (1942), Proudman (1952), and Dietrich (1963) provide such treatments. The text by Neumann and Picrson (1966) is more recent and more advanced. SUMMARY Equations may be written to describe the propagation of an idealized tidal wave down a straight-walled channel. If bottom stress and the Coriolis effect are neglected, the wave is seen to propagate as a result of the interaction between water level displacement and the flow of water induced by this displacement. The speed of the tidal wave form (c) is equal to (gf/)] '"', where g is the acceleration of gravity and h is water depth, while the speed of the associated current («) is propor- tional to this value, In very shallow water, u is reduced by turbulent dissipation of energy and frictional loss of energy to the bottom. In nature, tides arc propagated onto the continental margin as waves from the open ocean. Such marginal tides are said to cooscillate with the oceanic tide. Since the incoming wave is rarely parallel to the coast, it appears to propagate along the coast. Tidal waves behaving in this fashion are referred to as progressive tidal waves. The tidal wave may be partially reflected back out to sea by the shoaling bottom and interact with the next incoming wave so as to produce a standing tidal wave. In a progressive tidal wave, maximum flood velocity occurs at high water, while maximum ebb velocity occurs at low water; in a standing tidal wave the tide and tidal currents are out of phase, so that maximum flood velocity occurs during the rising tide, and maxi- mum ebb velocity occurs during the falling tide. An important parameter determining the character of tides in bays and estuaries is the ratio R = Z./X, where L is the distance between the reflecting barrier and the mouth of the embayment, and X is the wavelength of the tide. When the ratio is small, the tide within the bay has the same range as outside, and tidal currents are small. However, if L is comparable to one-fourth of the scmidaily or daily tidal wavelength, the embayment resonates with the outside tide. Ranges are up to four times higher, and currents are more intense. When the tide range is a significant fraction of the depth, the wave form becomes distorted, with the slope of the sea surface becoming greater on the leading side of the crest. The difference between the time-water height curves of the undistorted and distorted tides is called a shallow water tide. Where this distortion is large, the velocity and discharge associated with the crest are greater than those associated with the trough. The resulting net transport of water is known as Stokes drift. In larger bodies of water, the tides and tidal currents are subject to the Coriolis effect, caused by the earth's rotation. A moving parcel of water experiences a force proportional to its speed, which looking down at the sea .surface, is to the right in the northern hemisphere, and to the left in the southern hemisphere. On open conti- nental margins, the pressure force associated with the passage of the tidal wave, together with the apparent Coriolis force, results in a water parcel following an elliptical trajectory with right-hand sense of rotation. A tidal wave behaving in this fashion is a Poincare wave. It occurs where the inertial period 7 ',. is greater than the period T of the tidal constituent. In restricted embay- ments such as the North Sea, or on continental shelves where the tidal wave propagates parallel to the coastline, coastward water flow induced by the Coriolis effect is blocked by the coast, and there results a slope of the sea surface up toward the coast. A tidal wave thus modified is a Kelvin wave. A Kelvin wave propagating around a sea or ocean is known as an amphidromic system. The sense of rotation is counterclockwise. The turbulent stress r is often modeled as proportional to the rate of change of the current with increasing distance above bottom. The proportionality constant Av is the vertical eddy viscosity. It is determined by the spatial variation of the currents, distance from bound- aries, stratification of the water density, and the past history of the motion. In turbulent boundary layers, a sublayer near the boundary exists where stress is con- stant and the current speed increases logarithmically with distance from the boundary. The slope of velocity profile is determined in part by the degree of roughness of the bottom, as measured by a bottom roughness length Z0. An internal tide is a wave with a tidal period, asso- 211 64 TIDAL CURRENTS ciated with displacements within the water column, and with very little displacement of the sea surface. The wave may occur at the interface between fluids of two densities, or may occur in a continuously stratified fluid. On a typical shelf, an internal wave would propagate with a speed about 60 times slower than the surface tide's speed of propagation. As the internal tide propagates inshore, the shoaling bottom thins the lower layer and hence slows the wave. Amplitude increases as does dissipation into turbulence; eventually the wave becomes a bore. Internal waves in a two-layered fluid cannot propagate shoreward of the intersection of the density interface with the bottom. At this point the waves lose their energy to turbulence, or if the bottom slope is steep enough, are reflected. SYMBOLS A, vertical eddy coelhcient a amplitude Cf drag coefficient c phase velocity of tidal wave g acceleration of gravity h water depth A' a constant; von Karman's constant (~0.4) L horizontal length scale T period of tidal wave / time IS vertically averaged tidal current u current velocity x horizontal distance Z0 roughness length Z vertical distance X wavelength r\ vertical displacement of sea surface with respect to mean water level p density REFERENCES Cacchione, D. and C. I. Wunsch (1974). Experimental study of internal waves over a slope. J . Fluid Mech., 66: 233-239. Dietrich, G. (1963). General Oceanography. New York: Wiley-Inter- science, 588 pp. Doodson, A. T. and H. D. Warburg (1941). Admiralty Manual of Tides, London: HM Stationery Office, 270 pp. Fleming, R. H. and R. Revelle (1939). Physical processes in the ocean. In P. D. Trask, ed., Recent Marine Sediments. New York: Dover, pp. 48-141. Neumann, G. and \V. J. Pierson (1966). Principles of Physical Oceanography. Englewood Cliffs, N.J.: Prentice-Hall, 545 pp. Proudman, J. (1952). Dynamical Oceanography. New York: Dover, 409 pp. Rcdfield, A. C. (1950). The analysis of tidal phenomena in nar- row cmbayments. Pap. Phys. Oceanogr. Meteorol., 11(4): 1-36. Rcdfield, A. C. (1958). The influence of the continental shelf on the tides of the Atlantic coast of the United States. J. Mar. Res., IT: 432-448. Sverdrup, H. V. (1927). Dynamics of tides on the North Siberian shelf, results from the Maud Expedition. Ceofys. Publ., 4: 5. Sverdrup, H. V., M. VV. Johnson, and R. H. Fleming (1942). The Oceans. Englewood Cliffs, N.J.: Prentice-Hall, 1087 pp. U.S. Department of Commerce, National Oceanic and Atmos- pheric Administration, National Ocean Survey (1974). Tide Tables, East Coast of North and South Americas, 7973. National Ocean Survey, Rockville, Maryland, 288 pp. 212 20 Reprinted from: Journal of Physical Oceanography, Vol. 6, No. 4, 596-602. 596 JOURNAL OF PHYSICAL OCEANOGRAPHY Volume 6 The Formation of the Yucatan Current Based on Observations of Summer 1971 Robert L. Molinari NO A A Atlantic Oceanographic and Meteorological Laboratories, Miami, Fla. 33149 2 January 1975 and 15 January 1976 ABSTRACT Temperature, salinity and Lagrangian current data collected during the summer of 1971 in the western Caribbean Sea are employed to evaluate the ageostrophic components of the flow in the formation region of the Yucatan Current. The ratio of tangential and centripetal accelerations to Coriolis acceleration for data averaged over 24 h periods remain less than 10% except in two areas. An anticyclonic turn, centered at 19°30'N, 86°W, has the largest centripetal accelerations, and in the region of Cozumel Island significant tan- gential accelerations occur. The large-scale accelerations and additional evidence support the hy pothesis that inertia! effects dominate in the formation of the Yucatan Current. 1. Introduction The Yucatan Current is considered the first segment of the Gulf Stream system, in the sense that current speeds similar to those measured further downstream are first observed in this region. Recent studies by Molinari (1975) and Molinari and Kirwan (1975) suggest that inertial effects dominate in the formation of the Yucatan Current. Fig. 1. The depth (m) of the 10°C isotherm, during the interval A, 14 to 23 July, 1971. The station locations are indicated by solid circles. Trajectories for the center of mass of buoy triads are also shown (see text). 213 July 1976 NOTES AND CORRESPONDENCE 597 90° 88° 86c 84° 82° i 80° DEPTH OF THE 10°C ISOTHERM IN METERS 24 JULY TO 1 AUGUST CICAR SURVEY MONTH I 20 m CONTOUR INTERVAL TRAJECTORY %£ca 24° 22° 20° 90c 88° 86° 84° 82° Fig. 2. As in Fig. 1, except for the interval B, 24 July to 1 August, 1971. 18° 16° 80fl A two-ship operation was conducted in the summer of 1971 to investigate the formation of the Yucatan Current, and in particular to determine the nature and extent of the ageostrophic component in the western Caribbean Sea (Fig. 1) where the Current forms. In this field study, the ageostrophic component of the flow was determined by measuring the acceleration of water parcels tagged by drift buoys. The surface buoy was described by Molinari (1973); nominally it senses the motion of the water at 40 m by a parachute drogue. Data reduction techniques, and the reduced surface drifter, temperature and salinity data collected by the NOAA ship Researcher were also discussed by Molinari (1973). These data, supplemented by additional temperature data collected by Discoverer (Hazelworth and Starr, 1975) are used in the following sections to describe the formation of the Yucatan Current in the summer of 1971. 2. Temperature-salinity data Certain isotherm topographies are useful surrogates for the density distribution in this area. In particular, Molinari (1975) found that the 10°C topography closely maps the geostrophic current regime in the western portion of the Caribbean Sea. The study period, 14 July to 22 August, 1971, is divided into three time intervals. The intervals are selected to provide spatial resolution of the temperature field over the shortest time span. Figs. 1-3 show the 10°C topographies during these intervals (identified as A, B, C). The 10°C topography at 83°\V (Fig. 1) has a slope at 18°N consistent with a narrow and fast geostrophic current. The maximum slope occurs along a band bounded bv the 460 m and 520 m contours. The band extends continuously from 83°\V, 18°N to the Yucatan Strait. The acceleration along this band is not mono- tonic. Rather, the direction of and gradients along the band varv, suggesting accelerations, decelerations and meanderings of the current. The temperature structure indicates the presence of eddies on either side of the maximum slope. The flows around these eddies are both cyclonic and anticyclonic. Figs. 1 and 2 suggest that there is a south to north 214 598 JOURNAL OF PHYSICAL OCEANOGRAPHY 88° 86° 84° 82° Volume 6 88° 86° 84° 82° Fig. 3. As in Fig. 1, except for the interval C, 8 to 22 August, 1971. 80e decrease in eddy size, with no resolvable eddies found at the Yucatan Strait. The migrations of the 460 to 520 m band show the temporal changes which occur in the main current. The only significant temporal variability occurs in the southern basin where the large eddy initially centered at 17°30'N, 86°W (Fig. 1) apparently drifts to the west. There is no significant movement of the band at the Yucatan Strait during the six-week time period. 3. Drifter data The technique for reducing the drifter data was described by Molinari (1973) and Molinari and Kirwan (1975). This procedure provides geographic positions every 2h. The individual drogue trajectories are shown in Fig. 4. Four sets of trajectories were obtained, and with at least three drifters being deployed in each set. Legs 1 and 2 employed the same buoys which drifted un- attended from points 4 to 5 during an emergency port call. The buoys were retrieved at the end of leg 2, and redeployed in an unsuccessful attempt to sample the cyclonic flank of the current (leg 3). After a schedulep port call leg 4 was initiated, and then prematurely terminated when a storm threatened the operations area. The results of the drifter analysis are presented below in increasing order of the derivatives of the buoy coordinate-vs-time functions, i.e., trajectories, speeds and accelerations. Finally, the ageostrophic components of the flow, as evaluated from the drifter data, are discussed. a. Trajectories Trajectories are computed for the center of mass of the buoy triads, and are given in Figs. 1-3. The current fields inferred from these trajectories are very similar to the circulation fields inferred from the 10°C topog- raphies. For instance, in the Yucatan Strait region the trajectories closely follow the depth contours of the 10°C isothermal surface during all three time intervals. However, in the southern basin the trajectories parallel only those contours obtained concurrently, i.e., the leg 1 trajectory parallels the phase 1 temperature field 215 July 1976 NOTES AND CORRESPONDENCE 599 (Fig. 1), and the leg 4 trajectory parallels the phase 3 temperature field (Fig. 3). This result is consistent with the temporal variability observed in the tempera ture field and discussed above. Visual inspection of the trajectories in Fig. 4 indicates that large-amplitude meanders did not occur in the area. The largest curvature in the trajectories occurs in the anticyclonic turn indicated in the temperature fields at 19°30'N, 86°W (Figs. 1-3). The average anticyclonic radius of curvature from points 7 to 10 shown in Fig. 4 is 75 km. b. Speeds and accelerations Buoy speeds have been computed from the position data by using a centered difference approximation to the differential. The accelerations have been computed in natural coordinates, i.e., downstream s, crossstream n (positive to the left of the downstream axis), and vertical z (positive up). The accelerations are tangential (dV/dt), centripetal (KV2) and Coriolis (fV), where d( )/dt = d( )/dt+Vd( ) ds, V is the measured speed, K the horizontal curvature (positive for downstream cyclonic turning), and / the Coriolis parameter. The trends in the individual buoy speeds have been determined by fitting in a least-squares sense a cubic polynomial in time to the speed values. The fitted polynomial trend has been subtracted from the observed values to arrive at residual speed curves. The speed and residual speed curves for each drifter, and the poly- nomial fit curve for one drifter, are given in Fig. 5. The large-scale accelerations inferred from the temperature data are apparent in the fitted polynomial curves. For instance, the temperature gradients of the anticyclonic turn centered at 19°30'N, 86°\V suggest a deceleration in the flow. This deceleration occurs at 206/1930 (Fig. 5), as the drifters enter this turn. The temperature gradients increase as do the buoy speeds (leg 3, 210/0315 to 213 1315), as the drifters approach the Yucatan Strait. The largest average downstream accelerations occur in the vicinity of and north of Cozumel Island, where the speeds approach those observed in the Gulf Stream. These large-scale accelera- tions occur on time scales of days and/or space scales of hundreds of kilometers. Smaller time and or space scale disturbances are superimposed on these lar^e-scale features. In Fig. 5, the speed curves with the trends removed show that the amplitude of these oscillations are relatively constant Fig. 4. Drogue trajectories as determined from 2 h positions. The trajectory of buoy 2 is continuous during legs 1 and 2 (identified by circled numbers), although the buoy was not continuously tracked from interval 4 to 5. The first and last position times (Julian day/hour) are: leg 1, 200/1830 to 204/ 1630; leg 2, 205/1930 to 209/0730; leg 3, 210/0315 to 213/1315; and leg 4, 228/1600 to 232/1000. 216 600 JOURNAL OF PHYSICAL OCEANOGRAPHY Volume 6 I.Oi- Ll) . a. in. o Ul Ul Q. W _l <. 3 Q. C/5 UJ q: a 8 2.0 16 1.2 - J L i LkJ I L 200/1830 201/1830 202/1830 203/1830 204/1830 2P 1 1 1 1 1 1 1 1 ^M I I I L O UJ UJ CL in BUOY I BUOY 2 BUOY 3 BUOY 4 POLYNOMIAL FIT CURVE I.Or o 8 - 4 210/0315 211/0315 2I2/03I5- 2,1 ri ' ' ' 4 I 213/0315 1 1 205/1930 206/:930 207/1930 208/1930 p — r^ 5 hlA- 3 TMy* 1 _i_. V i i 8 P ^^Vy* 228/1600 229/1600 230/1600 231/1600 2P 1 1 1 1 1 1 1 1 FlG. 5. The observed and residual speed curves for the trajectories of Fig. 4 as a function of time (Julian day/hour). The residual speeds are determined by sub- tracting the third-degree polynomial fit curve (a representative curve is shown in the upper panels of each time interval) from the observed curve. throughout the basin, although their primary period appears to vary. A visual inspection of the records indicates that the principal period in the southern speed data is 24 h, and in the northern data, 12 h. As indicated, no large-amplitude meanders, similar to Gulf Stream meanders, appear in the trajectory data (Fig. 4). Thus, the velocity oscillations discussed above occur along the axis of the flow, rather than normal to it. The downstream spatial extent of the oscillations vary from approximately 25 km in the low-speed regions to 75 km in the high-speed region of the basin. c. Ageostrophic components The horizontal equations of motion can be written in natural coordinates as dV dD —+—=Ru dt ds dD KV2+JV+ — =i?2) dn (1) (2) where D is the dynamic height relative to a level of no motion, and Ri and /?•> include all the forcing and retarding functions. In a frictionless system the ageostrophic components are the centripetal accelera- tion KV'2 and the tangential acceleration dV /dt. The terms on the left-hand side of (1) and (2) are evaluated for those portions of the trajectories where density data are available. Dynamic heights are computed relative to 600 m, since the majority of the hydrographic casts were to this depth. Accelerations for the center of mass trajectories and the density gradients are averaged over 24 h periods to eliminate the small- scale oscillations shown in Fig. 5. Table 1 lists these average properties at 12 h intervals. If i?> = 0, Eq. (2) becomes the gradient equation. The gradient balance expressed in percentages as [_Ri/ (A'V2+/V)]X100 was maintained on the average to within 10% durings legs 1, 2 and 3. The gradient balance computed relative to 600 m was not maintained during leg 4 for undetermined reasons (although internal wave aliasing of the density field may be a cause). The terms in (1) are an order of magnitude less than the terms in (2) and therefore more difficult to evaluate realistically. However, the large-scale tangential accel- erations are consistent ■• with downstream pressure gradients observed durirtg legs 2 and 3 (Fig. 4). 217 July 1976 NOTES AND CORRESPONDENCE 601 Table 1. Ageostrophic and geostrophic components averaged over 24 h periods. \KV*\/fV \dV/dt\//V Time Julian V fv dD/dn xioo X100 Fig. 4 point day/hour (cm s"1) (cm s~2X 104) (cm s-2X104) (percent) (percent) 1 202/1330 0.69 0.31 0.27 8 2 2 203/0130 0.68 0.30 0.25 11 2 3 203/1330 0.71 0.32 0.31 6 4 4 204/0130 0.71 0.33 0.36 7 3 5 206/0830 0.61 0.28 10 4 6 206/2030 0.58 0.27 9 4 7 207/0830 0.58 0.28 0.20 12 3 8 207/2030 0.61 0.30 0.22 17 10 9 208/0830 0.69 0.34 0.28 12 8 10 208/2030 0.76 0.38 7 3 11 211/0215 0.66 0.32 4 3 12 211/1415 0.75 0.37 0.29 13 4 13 212/0215 0.94 0.47 0.34 8 13 14 212/1415 1.15 0.59 0.50 1 8 15 213/0215 1.49 0.70 1 7 16 229/0500 0.32 0.14 2 5 17 229/1700 0.33 0.15 9 4 18 230/0500 0.30 0.14 3 1 19 230/1700 0.29 0.13 7 3 20 231/0500 0.28 0.13 9 8 21 231/1700 0.26 0.12 0 6 The data listed in Table 1 verify the results described qualitatively in previous sections ; that is, the large- scale flow undergoes little acceleration in the mid-basin (points 1-4, Fig. 4), the largest centripetal accelerations occur in the anticyclonic turn centered at 19°30'N, 86°W (point 8, Fig. 4), and the largest accelerations occur in the vicinity of Cozumel Island (point 13, Fig. 4). If Ri = R2 = 0, the large-scale flow is geostrophic to within 20%. In particular, although the average velocity more than doubles from points 1 1 to 15 (Fig. 4), the maximum value of the ratio [(f/t/ 202 0530 time: a'0 3 0 5 30 TIME 204 0530 Figure 4 229 2275 2250- 2225- 2200 o -z. (/) o I 2175- 2150 - 2125 2100 550 575 600 X-DISTANCE (KM) 625 650 Figure 5 230 0 J30 TIMH Figure 6 231 2 Li O 2 < H co o i >- 2375 2350 2325 2300 - 2275 - 2250 - 2225 - 2200 - 2175 - 2150 - 500 525 550 575 600 X- DISTANCE (KM) Figure 7 232 u <^ 2 U 2 ID * vc hi > s o^ -2-- -4 H 1 1 I TIME t- q: a > -2- 210 2015 201 5 TIME! 212 2015 213 201S Figure 8 233 to m o lO lO ^^ S *: uj o I) 33NV1SIQ-A 234 219 no J3I I 100 1100 2 100 TIME Figure 10 235 o CO O ^° O C\i cc o :r> CO o o o X v. oo \ o CM >- f— i Ck ^ ) computed from observed wind speed and direction, assuming a drag coefficient of !.2X10~:'; mean mixed layer depth (//) from CTI) observations; and menn surface wind drift speed (I) and direction (0) computed from Gonella (1971), assuming an eddy viscositv of 102 cm2 s""1. T 4> // r e Period (d; in cm"2)* , (oT) (m) (ki ii day-1 ) (°T) 9 Mar-30 Mar 0.16 68 38 1.7 115 31 Mar-14 Apr 0.30 37 34 3.1 81 15 Apr-29 Apr 0.82 260 48 8.5 305 30 Apr-14 Mav 0.40 260 28 4.1 305 15 Mav-29 Mav 0.26 353 11 i.2 68 30 Mav-13 June 0.50 261 13 6.1 335 14 Jun-28 Jun 0.21 329 12 3.1 36 29 Jun-13 Jul 0-20 i2i 11 3.0 ii Mode mean 0.20 293 25 2.2 356 * 1 dvn cm-2 = 0.1 N m"2. completely confined to the surface mixed layer (zero stress at the bottom of the layer). The surface stress was determined from the usual relation GpA | Vir j W, where Vw is the reported wind velocity, pA is the standard air density (1.2X10~3 g cm-3), and C is a drag coefficient taken to be 1.2X10-3. From Gonella (1971) we assumed a constant and conservative value of 102 cm'2 s_1 for the eddy coefficient of viscosity in the layer. A larger viscosity will reduce the currents, but not less than about 25% of the values computed. A smaller value will increase the current (by a factor roughly proportional to the inverse square root of the viscosity). In Table 1 we have listed the spatially averaged net vector wind stress and surface drift current for each of the mapping intervals in Figs. 3 and 4. Also shown is the average depth of the mixed layer, which clearly shows the effect of decreasing wind stress and increasing surface heating over the duration of MODE. The computed currents are all less than 10 km day-1 and except for the two periods 15 April-29 April and 30 May-13 June, when there were periods of persistent wind direction, less than 5 km day-1. It is apparent, therefore, from the maps of dynamic height (Figs. 3 and 4) that advection by the eddy surface currents (of order 20 km day-') around the periphery of the eddies dominates the surface wind drift. The latter, however, is significant over large areas where there is little relief in the dynamic topography. The average surface displacement due to the wind drift was approximately 60 km per mapping interval. Also shown in Table 1 is the net vector wind stress and surface current over the entire 127 days of MODE. The stress is quite comparable in magnitude and direc- tion to the typical long-term stress in the MODE area computed by Saunders (1976) for the years 1959-1971. The overall surface drift is predominantly northward and the total northward volume transport, computed from the surface stress, is 2.6X102 m3 s_1 across each kilometer in an east-west direction. The spatial variation of surface wind drift con- tributes, of course, to the distortion of surface tem- perature structure. Superficially it appeared small because the reported daily wind speed and direction were remarkably similar from ship to ship. Never- theless, there were differences. By comparing wind data from one ship to another we found over an average horizontal scale of 100 km that there were variations of net surface drift speed of about l the spatially averaged value in Table 1, and variations of drift direction of about ±20°. The distortion of surface temperature features by such variations is an order of magnitude less important than that due to the spatial variations of the eddy surface currents. We tentatively conclude, therefore, that the dominant effect of the surface wind drift is to shift but not distort the temperature pattern shown in Figs. 3 and 4 according to the drift currents in Table 1. 6. Conclusions and discussion In the MODI-: area we conclude that the surface temperature field, on time scales less than about one month and over space scales from 400 to about 40 km, tags primarily the surface currents associated with the baroclinic mesoscale eddy field of the main ther- mocline. Surface currents induced by wind stress appear to be of secondary importance in generating spatial structure in this scale range. Relatively little can be said about scales less than 40 km. However, there is some evidence from MODE but mostly from previous measurements in the same area that much of the spatial structure on scales less than about 10 km 4 tags not only the mesoscale current field but also a relatively shallow field of currents which are in geostrophic equilibrium with horizontal density gradients in the near surface layers. This new field of currents is often jet-like and is associated with surface frontogenesis. Mesoscale eddies appear to be an effective mecha- nism for stirring the large-scale thermal (and haline) field imposed on the near surface layers by the atmo- sphere. One can speculate that this surface process on an eddy time scale may generate a net meridional heat transport in the surface layer on a longer time scale. For example, if a single anticyclonic eddy develops in the convergence zone one would expect warm water to move initially northward on its western side and cool water southward on its eastern side. (The flows will change sides if the eddy is cyclonic.) In time both flows will simply circulate in a complex manner around the eddy with a great deal of stirring but no net heat transport if there is no heat exchange between the warm and cool water. However, if there 4 It is significant that this scale is of the order of the internal radius of deformation of the near-surface pycnocline. 245 November 1976 A. D. VOORH1S H. SCHROEDER AND A. L E E T M A A 961 are many eddies, which are evolving, moving and decaying, it is highly likely that surface water is ex- changed5 from eddy to eddy and one might expect to observe at times long tongues of warm and cool surface water running north and south. This is very similar to what one sees in Fig. 2. The result would be a mean meridional heat transport northward in the MODE area of the order of Nil per eddy, where V is the geostrophic advecting surface velocity, and // is the anomalous heat carried by each tongue. The latter can be approximated by pwCT,DL AT, where AT is the temperature difference between north and south flowing tongues, L is the zonal width of the tongue, and D is the depth of the heat anomaly. Representative values for these parameters are F=20 cm s-1, pw=l g cur3, C'p=4.18 J g"1 K~\ D=50 m, L=100 km, Ar=2°C. Using these values one computes a transport of 8.2 X1012 W per eddy. Taking 200 km as a mean zonal spacing between eddies one finds a northward eddy heat transport of 4.2X1010 W across each kilometer in an east-west direction. Assuming the northern Sargasso Sea to be bounded on the north and west by the Gulf Stream, on the east by 50°W longitude, and on the south by 30°N latitude, one computes an annual heat input of 32X1020 J across its southern boundary (length 2400 km) by the eddy mechanism. This is of the same order as the annual heat loss to the atmosphere across its surface area (2.2X106 km2) computed from Bunker and Worthington (1976), using an average net heat surface flux of 66 W m-2 (50 kcal cm-2 year-1). Speculating on a still larger scale and assuming that the observed mesoscale eddy activity extends across both the northern Atlantic and Pacific Oceans at mid-latitudes, a total distance of the order of 1.6 X 104 km. one finds an annual poleward heat transport by the eddies at these latitudes of the order of 6.8 X 1014 W. This can be compared with the annual oceanic poleward energy transport of about 22.6 X1014 W (1.7X1022 cal year-1) estimated by Vonder Haar and Oort (1973). Considering the uncertainties in all of these estimates one concludes that the meso- scale eddy heat transport may not be inconsequential. Finally, our results can have important implications for oceanographers and meteorologists interested in annual or longer term changes in sea surface tem- perature and their effect on world climate. The fluc- tuating mesoscale temperature field is unwanted noise from their point of view and introduces an uncertainty to estimates of mean temperatures. For data collected from a fixed point (or within an eddy radius of this point) this uncertainty is of the order of (ATe)/yjn, where (ATe) is the rms temperature change due to a typical eddy, and n is the number of eddy events 5 This may be greatly enhanced by the unusually strong surface currents associated with surface frontogenesis. in the averaging time. Assuming (ATe)«0.5°C and no other sources of noise, one would have to average over 25 eddy events in the MODE area in order to resolve a climatic 0.1 °C change in mean surface tem- perature. If the eddy residence time is of the order of 2 months this would take 4 to 5 years. Acknowledgments. This work was supported by the Office of Naval Research under Contract N00014-74- C-0262, XR 083-004 and by the Office of the Inter- national Decade for Ocean Exploration of the National Science Foundation under Funding Agreement AO-385. The data used in this paper were collected and processed by many people in the MODE program and the authors wish to acknowledge all of this work and to express their gratitude. We would also like to thank N. Fofonoff of the Woods Hole Oceanographic Institution, Woods Hole, Mass., who programmed and computed the objective maps of dynamic height in Figs. 3 and 4. REFERENCES Bryden, H. L., 1974: Geostrophic comparisons using moored measurements of current and temperature. Nature, 251, 409-410. Bunker, A. I'., and I,. V. Worthington, 1976: Energy exchange charts of the North Atlantic Ocean. Bull. Amer. Meteor. Soc, 57, 670-678. Crease, J., 1962: Velocity measurements in the deep water of the western North Atlantic. J. Geopliys. Res., 67, 3173-3176. Dynamics and the- Analysis of MODE-1, March 1975: Report of the MODE-1 dynamics group (unpublished manuscript). [The MODE Executive Office, 54-1417, M.I.T., Cambridge, Mass. 02139.] Gonella, }., 1971: The drift current from observations made on the Bouee-Laboratoire. Call. Oceanogr., 23, 1-15. Katz, E. J., 1969: further study of a front in the Sargasso Sea. Tell us, 21, 259-269. McWilliams, J. C, 1976: Maps from the Mid-Ocean Dynamics Experiment. I. Geostrophic streamfunction. /. Pliys. Oceanogr. (accepted for publication). Robinson, A. R, 1975: The variability of ocean currents. Rev. Geopliys. Space Pliys., 13, 598-601. Saunders, P. M., 1976: On the uncertainty of wind stress curl calculations. J. Mar. Res. (submitted for publication). Schmitz, W. J., J. R. Luyten, R. E. Payne, R. H. Heinmiller, G. H. Volkmann, G. 11. Tupper, J. P. Dean and R. G. Walden, 1976: A description of recent exploration of the eddy field in the western North Atlantic with a discussion of Knorr Cruise 49. W1IOI Tech. Re]), (to be published). Schroeder, E. H., 1966: Average surface temperatures of the western North Atlantic. Bull. Mar. Set., 16, 302-323. Vonder Haar, T. H., and A. H. Oort, 1973: New estimate of annual poleward energy transport by Northern Hemisphere oceans. J. Pliys. Oceanogr., 3, 169-172. Voorhis, A. D., 1969: The horizontal extent and persistence of thermal fronts in the Sargasso Sea. Deep-Sea Res., 16, 331-337. and J. B. Hersey, 1964: Oceanic thermal fronts in the Sargasso Sea. Deep-Sea Res., 69, 3809-3814. Wiist, G., 1928: Der Ursprung der atlantischen Tiefenwasser. Z. Ges. Erdk. Berl., Sonderband zur Hundertjahrfeier, 506-534. 246 24 Reprinted from: Marine Geoteoknology , Vol. 1, No. 4, 327-335. Initial Results and Progress of the Mississippi Delta Sediment Pore Water Pressure Experiment RICHARD H. BENNETT,* WILLIAM R. BRYANT,t WAYNE A. DUNLAP,t AND GEORGE H. KELLERtt Abstract This report describes the instrumentation, initial results, and progress of an experiment designed to measure and monitor submarine sediment pore water and hydrostatic pressures in a selected area of the Mississippi Delta. The experiment also is intended to monitor significant pressure perturbations during active storm periods. Initial analysis of the data revealed excess pore water pressures in the silty clay sediment at selected depths below the mudline. Continuous monitoring of the pore water and hydrostatic pressures was expected to reveal important information regarding sediment pore water pressure variations as a function of the geo- logical processes active in the Mississippi Delta. Introduction The NOAA-Atlantic Oceanographic and Meteorological Laboratories is presently engaged in a NOAA program directed toward the delineation and understanding of important processes and mechanisms related to submarine sediment stability. A unique situation arose to test some of the equipment and concepts being developed in this program on Project SEASWAB (Shallow Experiment to Assess Storm Waves Effecting Ztottom), which is part of a larger study of the Mississippi Delta being conducted by the U.S. Geological Survey. NOAA-Atlantic Oceanographic and Meteorological Laboratories, Miami, Florida. 'Texas A&M University, College Station, Texas. ft School of Oceanography, Oregon State University, Corvallis, Oregon. (Received December 4, 1975; Revised February 13, 1976.) Marine Geotechnology, Volume 1 , Number 4 Copyright © 1976 Crane, Russak & Company, Inc. 327 247 328 RICHARD H. BENNETT ETAL. 29°30' - Mississippi Birdfoot Delta 29°20' - 29°10' - 29°00- 28°50' - 89°30' 89°20' 89°10' 89°00' Figure 1. General study area. 88°50' The purpose here is to describe the instrumentation, progress, and initial results of an experiment designed to measure and continuously monitor, over a period of several months, submarine sediment pore water pressures in a selected area of the Mississippi Delta (Figure 1). Objectives of the experiment are to measure not only pore water and hydrostatic pressures at various depths below the mudline, but also to monitor significant pressure perturbations during active storm wave periods. It is interesting to note that while engineers have known for decades that pore water pressures are an important geotechnical consideration, the first reported attempt to measure pore water pressures in submarine sedi- ments was made by Lai et al. (1968) and Richards et al. (1975). The Mississippi Delta is well known for being a very dynamic region charac- terized by the interaction of riverine and marine processes and the large dis- charge of bedload and suspended sediment. Large plumes of sediment extend considerable distances beyond the subaerially exposed delta and deposit vast 248 MISSISSIPPI DELTA SEDIMENT PORE WATER PRESSURE EXPERIMENT 329 quantities of silt and clay in the prodelta environment. This environment is characterized not only by the rapid deposition of fine-grained sediment having very high water contents, but also by the accumulation of organic material (Coleman et al., 1974). Methane and carbon dioxide gases, intimately related to decomposition of the organic material, influence substantial portions of the Mississippi Delta submarine sediments (Whelan et al., 1975). Knowledge of the sediment geotechnical properties in this complex and dynamic environment is of great importance to engineers faced with the design and construction of offshore structures, and to geologists investigating sedimentological processes relating to submarine diagenesis, environments of deposition, mass movement, and sedi- ment stability (Morelock and Bryant, 1966; Keller and Bennett, 1968; Bennett and Bryant, 1973). Not only will the measurement of pore water pressures in the Mississippi Delta sediment aid in understanding and interpreting the sediment geotechnical properties, but it also should provide an insight into the behavior of these sediments in response to dynamic and static loads. Instrumentation The NOAA sediment pore water pressure probe (piezometer) system consists of the following components: 1 . Probe and Sensing Units (Transducers) 2. Signal Conducting Cables 3. Signal Conditioner Units 4. Voltage and Frequency Regulator Units 5. Recording Unit. The probe enclosing the pressure sensing devices is a 0.10-m O.D. steel pipe having a total length of 17.12 m. A weight stand mounts to the top of the probe on four steel gusset plates. The weight stand is fastened to the probe with two steel pins and the probe assembly is lowered into the seafloor by a steel cable fastened to the top of the weight stand. Four sensing units (variable reluctance pressure transducers) were placed in the probe at selected intervals; two sensors measured pore water pressure and two measured hydrostatic pressure (Figure 2). The pore water was transmitted through 0.05-m diameter porous corundum stones, which were inset in the pipe and ground to the pipe radius. Hydrostatic pressure was transmitted from the mudline through the seawater-filled steel pipe to the sensors installed inside the probe. Sensors were mounted inside oil-filled capsules in the probe and connected to the appropriate pressure ports with short tubing. Separate measurement of the pore water and hydrostatic pressures was necessary since one objective of the experiment was to determine the effect of 249 330 RICHARD H. BENNETT ETAL. bottom pressures from storm waves on the pore water pressures. This approach, however, is feasible only in shallow water. A sensing unit which is robust enough for use in deeper water will generally not have sufficient resolution for the purposes desired. For measurement of static pressures, this limitation is largely overcome by a differential piezometer of the type described by Hirst and Richards (1976). The signals from each transducer are transmitted through a conducting cable to signal conditioners and filtering systems before being recorded on a strip chart recorder. Electronic units and pressure transducers were tested and calibrated prior to the assembly of the probe. Calibration was' carried out using a fused quartz Bourdon tube pressure gage having a sensitivity of 1 part in 200,000. The transudcers have a maximum working range of 689.5 kPa (1 psi =6.89 kPa) and a reproducibility of ± 3.5 kPa. The electronic units were checked frequently during various phases of the probe assembly. Installation of the Probe The probe and electronics system we.e assembled in the field aboard the Texas A&M University Ship R/V 1yr°. during 18-19 September, 1975. The total weight of the probe and weight stanrl loaded with four train wheels was 1.115 Mg in sea water; this weight having been calculated as adequate to implant the probe to the desired depth of penetration. On the afternoon of 19 September, 1975, the probe was lowered from the Gyre in the Mississippi Delta sediment (Block 28, Soutn Pass Area, slightly south of 29°00'N, 89°15'W) at a preselected site 145 m from an offsnore production platform where the recorder and signal conditioner units were installed later. The water depth was approximately 19 m at the site. After installation, divers removed the steel pins and made a general inspection of the exposed portion of the instrument. The weight stand and weights were returned to the ship. Pore water and hydrostatic pressures were monitored from the ship during installation and for 40 min afterward. In the ensuing 4 h period, no readings were made while the electronic units were transferred to the platform. During this time divers installed the signal conducting cables along the sea floor to the platform. After reconnection, all systems appeared to be functioning properly. The probe was implanted only a few days before the passage of Hurricane Eloise near the site. Discussion Sediment pore water pressures, uw, were measured at depths of approxi- mately 8 and 15 m below the mudline. Hydrostatic pressures, us, were measured simultaneously at depths of approximately (actual mudline difficult to de- 250 MISSISSIPPI DELTA SEDIMENT PORE WATER PRESSURE EXPERIMENT 331 -A Steel Cable 1.98m m ^ 'T7 Pressure Transducer #1 (Hydrostatic Pressure) 1.41m 6.95m Pressure Transducer #2. (Pore Water Pressure ) Po r o u s St o ne 6.88m Pressure Transducer #4 (Pore Water Pressure) Steel Pins Weights (Train Wheels) 0.91m Dia. Welded Steel Plates Cable Clamps Multiconductor Armored Cables 0.10m O.D. Steel Pipe 6.40m Sections Fastened With Inner Couplings Pressure Transducer #3 1 Or — ■"" (Hydrostatic Pressure) 7 Figure 2. Sediment pore water pressure probe and weight stand. Drawing not to scale. termine) 1 and 15 m below the mudline (Figure 2). A value of 68.9 kPa, equivalent to 7 m of sea water, has been added to Pressure Transducer 1 data for direct comparison with the sediment pore water pressures recorded by Trans- ducer 2. Comparison of the hydrostatic pressure and sediment pore water pressure for a given depth below the mudline may reveal one of three possible conditions: Condition 1. Sediment pore water pressure equals the hydrostatic pressure (uw ~ us). Condition 2. Sediment pore water pressure exceeds the hydrostatic pressure ("w > "*)• Condition 3. Hydrostatic pressure exceeds the sediment pore water pressure (uw o u Q CO o o o o o o m o m •"■ CO ^- T3 H 0) -H 4J n CO ca 01 S c I; r A z "■ < z I C/3 o is .c y) n ■iH o C c Hi CD .C 4J (J « o T3 C c HI 3 O o l_> « V) (1) (U X. c 4-1 by the Pillsbury and Discoverer (Figure 1) . In order to estimate the most likely replacement density the terrain-corrected Bouguer anomalies were computed for a range of reasonable density contrasts. If the gravity field varies smoothly, it can be represented by a low-order polynomial. Various low-order polynomials were calculated for each density contrast. The root mean square deviations between the predicated values of the field and the actual values were computed for each surface fitted. A fifth-order polynomial was accepted for which a least squares best fit was obtained. Comparison of the results of this polynomial fitting method with those of the density profiling method of Nettleton [19 39] revealed that a value of 2.3 g/cm for the subaerial parts and 2.4 g/cm for the submarine portions should be accepted for the replacement bulk densities o be used for detailed investigations in tne Cape Verde region. These values are all in accord with those predicted from the Nafe and Drake [1963] curve. The detailed terrain-corrected Bouguer anomaly map is shown in Figure 9. There are numerous positive anomaly closures which are assumed to be caused by high-density plutonic bodies. To the southeast of the island of Maio there occurs a sharply angled bend in the isogals coupled with extremely high gradients perpendicular to both arms of the bend. The northward continuation of the bend passes through major deviations in the isogals to the southeast of Boa Vista and emerges to the east of Sal, where another steep gradient is observed. A major roughly N-S fault is postulated to the east of the islands of Maio and Sal. Bebiano [1932] postulated a major fault through the four northern islands extending to Boa Vista. There is no extreme anomaly gradient perpendicular to this proposed fault trace. It is possible that the observed minor and major deviations of the isogals could indicate a large tectonic feature whose field has been partially obscured by the interfering effects of extruded magma occupying a large area. Analysis of the data corrected with the regional replacement density of 2.58 g/cm indicates that the Moho ranges between 16 and 18 km beneath the island block. These values are in excellent agreement with those inferred from the seismic refraction data. Magnetic Anomalies The magnetic quiet zone boundary in the eastern North Atlantic lies between the Cape Verde archipelago and the African continent and trends basically N-S [Heirtzler and Hayes, 1967; Rona et al. , 1970]. A sequence of oceanic magnetic anomalies, the Keathley sequence or J anomalies, forms a 350-km-wide band seaward of the quiet zone boundary and extends up to the eastern margin of the Cape Verde archipelago (Figure 10). Vogt et al. [1970] along with Rona et al. [1970] have shown that this J anomaly band occurs with almost mirror image similarity on the east and west sides of the central North Atlantic. 263 Dash et al . : Geophysical Investigation of the Cape Verde Archipelago 5257 18 w TERRAIN-CORRECTED BOUCUER ANOMALY CAPE VERDE ISLANDS CONTOUR INTERVAL 10 MGAL 26 25 24 23 22 Fig. 9. Terrain-corrected Bouguer anomaly map around the Cape Verde archipelago with a contour interval of 10 mGal. Tentative correlation of oceanic magnetic anomalies in the Cape Verde area based on our new data and added on to those published by Rona et al. [1970] substantially confirms the general N-S trends established by previous work. There is some indication of minor offsets of anomalies, although no fracture zones can be mapped except the possible eastern extension of the Kane fracture zone at about 21°N latitude. Where deviations from the general N-S anomaly trend occur, the preferred orientation is NNW-SSE, parallel with structural features and morpholo- gical lineations within the island group as well as with the postulated deep fault trends derived from the seismic and gravity data. Conclusions The present crustal structure beneath the Cape Verde archipelago determined by our seismic refraction and gravity measurements is transi- tional in that the Moho lies at a crustal depth between 16 and 17 km, midway between dimensions typical for continental and oceanic crust. It is interesting to compare the crustal structure of the Canary Islands, lying off the coast of Spanish Sahara with that of the Cape Verde archipelago. Dash and Bosshard [1969] postulated that the five western islands of the Canaries group are not related structurally to the African continent. The crust is of oceanic thickness on the west. In the central part of the island group the thickness of the crust is transitional. The depth to Moho varies from 12 to 14 km. Roeser et al. [1971] , from their refraction seismic and gravity studies of the area between Africa and Gran Canaria, suggest a Moho depth of 21 km with a crust originally of oceanic character having presumably been depressed to the depth of 21 km with 10 km of differentially meta- morphosed sediments deposited on it. The transitional zone between the oceanic and continental crust is characterized by major faults of NE-SW strike. The original composition of the crust underlying the Cape Verde archipelago was probably oceanic as is evidenced by (1) the 264 S2S8 Dash et al.: Geophysical Investigation of the Cape Verde Archipelago 16'| ■J\ v j s.iA ■ ^» ■ ,v. > CAP BLANC - ^ *y\ v v v^vu >",/. V. ^ . / , t 1/ 'v;- •*,;_, "'•:, ^"v^. CAPE VERDE ISLANDS DAKAR Fig. 10. Magnetic anomaly map around Cape Verde Islands and continental shelf of Africa to 24°N. The Keathley sequences, a possible fracture zone, and the magnetic quiet zone boundary are indicated. No major bottom topographic features occur along tne magnetic track lines. occurence of a typically oceanic layer 3 with a velocity of 6.3 km/s^ (2) the presence of the magnetic quiet zone boundary landward of the islands [Rona et al. , 1970]; and (3) the presence of open water aptychus limestone (Upper Jurassic-Lower Cretaceous) exposed on Maio similar to limestones recovered from analogous locations of the western North Atlantic deep ocean basin [Hollister et al. , 1972]. The origin of the Canaries is closely linked with the faults through the islands [Dash and Bosshard , 1969]. The Cape Verde archipelago also originated as igneous intrusions and extru- sions along a fault system. The similarities between the Canary and Cape Verde islands end at the comparison of their crustal structures. Owing to a lack of any detailed seismic refrac- tion data in the area between the t^o island groups it is not possible to draw any extension of similarities between them. The magnetic signature of each of the two island groups is singularly distinctive. The Cape Verde islands lie west of the magnetic quiet zone, whereas in the Canaries the magnetic quiet zone passes through the island of Tenerife. The magnetic quiet zone almost follows the transitional crustal zone on the west coast of Africa. To reconcile the present transitional thickness and the inferred former oceanic composition of the crust beneath the Cape Verde archipelago, we postulate that these islands originated as igneous intrusions and extrusions along the N-NW, NW, and W-SW trending fault systems. The Moho was depressed from oceanic to transitional depth as a result of the loading of accumulated materials of the islands. It is suggested that the building of the Cape Verde archipelago began at an early stage in the Mesozoic opening of the central North Atlantic and at no time did the islands belong to the African continent. The position of the archipelago was apparently controlled by the convergence of fracture zones as a consequence of the marked change in the trend of the adjacent mid-Atlantic ridge. Acknowledgements . The authors wish to thank A. Richardson, F. Machado , P. Hubral , B. Buttkus, and C. J. M. Hewlett for their 265 Dash et al.: Geophysical Investigation of the Cape Verde Archipelago 5259 help during the collection of the data. Thanks are also due to the officers and crew of the R/V John Elliot Pillsbury and the NOAA ship Discoverer. Financial support for the project was kindly made available by the Natural Environment Research Council of Great Britian, U.S. ONR grant N00014-67-A- 0201-0013, NSF grants GA-39744, GA-19471, GB-27252, and GA-27465, and NOAA. The authors gratefully acknowledge this support. References Assuncao, C.F.T. de , F. Machado and A. Serralheiro, New investigations on the geology and volcanism of the Cape Verde Islands, Int. Geol. Congr. Rep. Sess. 2 3rd, 2, 9-16, 1968. Ballard, J. A., and L.G. Hemler, Structure of the Cape Verde rise (abstract), Eos Trans. AGU, 50, 210, 1969. Bebiano, J.B., A geologia do arquipelago de Cabo Verde, Comun. Serv. Geol. Port. , 18, 1-275, 1932. Dash, B.P., and E. Bosshard, Seismic and gravity investigations around the western Canary Islands, Earth Plan et. Sci. Lett. , 72, 169-177, 1969. Dash, B.P. , K.O. Ahmed, and P. Hubral, Seismic investigation in the region of Poulo Panjang, offshore from southwestern Vietnam, ECAFF, Tech. Bull. , 3, pp. 37-54, 1970. Egloff, J., Morphology of the ocean basin sea- ward of northwest Africa: Canary Islands to Monrovia, Liberia, Artier. Ass. Petrol. Geol. Bull. , 56, 694-706, 1972. Heezen, B.C., and M. Tharp, Physiographic diagram of the North Atlantic Ocean, Geol. Soc. Amer. Spec. Pap. , 65, 1968. Heirtzler, J.R. , and D.E. Hayes, Magnetic boundaries in the North Atlantic Ocean, .Science, 157, 185-187, 1967. Hollister, CD., et al . , Initial Reports of the Deep Sea Drilling Project, vol. 11, p. 1077, U.S. Government Printing Office, Washington, D.C. , 1972. LeBas, M.J., Per-alkaline volcanism, crustal swells and rifing, Nature Phys . Sci. , 230, 1971. Lowrie, A., and E. Escowitz, Eds. Kane 9, in Global Ocean Floor Analysis and Research Data Series, 971 pp. , U.S. Naval Oceano- graphic Office, Washington, D.C, 1969. Machado, F. , Vulcanismo das Ilhas de Cabo Verde e das outras ilhas Atlantidas , Junta Invest. Ultramar Port. , Estud. Ensaios Doc. , 117, 1-83, 1965. Machado, F. , J. Azeredo Leme , and J. Monjardino, 0 complexo sienitocarbonatitico dal Ilha Bravo, Cabo Verde, Garcia de Orta, 15, (1) 9 3-98, 1967. Nafe, J.E., and C.L. Drake, Physical properties of marine sediments, in The Sea, vol. 3, The Earth Beneath the Sea, edited by M.N. Hill, pp. 794-813, Interscience , New York, 1963. Nettleton, L.L. , Gravity and magnetic calcula- tion, Geophysics, 4, 176-185, 1939. Part, CM., Volcanic rocks from the Cape Verde Islands, Bull. Brit. Mus . Natur. Hist. Geol. , 1(2) , 27, 1950. Roeser, H.A., K. Einz, and S. Plaumann, Continental margin structure in the Canaries, The Geology of the East Atlantic Continental Margin, 2, Africa, edited by F.M. Delany , rep. 70/16, pp. 27-36, Inst. Geol. Sci., Rons, P. A., Bathymetry off central northwest Africa, Deep Sea Res. , 18, 321-327, 1971. Rona, P. A., J. Brakl, and J.R. Heirtzler, Magnetic anomalies in the northeast Atlantic between the Canary and Cape Verde islands, J. Geophys. Res. , 75(35), 7412-7420, 1970. Serralheiro, A. , Geologia da Ilha de Maio (Cabo Verde) , Junta Invest. Ultramar Port. , Estud. Ensaios Doc. , 103, 1970. Talwani , M. , and M. Ewing , Rapid computation of gravitaional attraction of three- dimensional bodies of arbitrary shape, Geophysics, 25, 203-205, 1960. Turner, F.J., and J. Verhoogen, Igneous and Metamorphic Petrology, p. 694, McGraw-Hill, New York, 1960. Vogt, P.R., C.N. Anderson, D.R. Bracey , and E.D. Schneider, North Atlantic magnetic smooth zones, J. Geophys . Res . , 75(20), 3955-3968, 1970. (RSceived July 31, 1975; revised February 13, 19 76; accepted February 29, 1976.) 266 26 Reprinted from: Sea Frontiers, Vol. 22, No. 1, 9-15. Iceland Where the Mid-Ocean Ridge Bares Its Back By Robert S. Dietz NOAA. Atlantic Oceanographic and Meteorological Laboratories Miami, Florida. Only in ICELANO can man walk on the Mid-Atlantic Ridge. This is die one place where pan of the 45.0<)(l-kilonieter-/ong ocean rift is exposed above sea level. Robert S Dietz ICELAND, a bleak, windswept island in the far North Atlantic, touching on the Arctic Circle, lies on rock hotter than lands at the equator. It is not entirely a foolish joke to say that an inhabitant of this island who runs short of hot water in his bathroom has only to drive a pipe down through the floor toget plenty for his hot bath. But the interest of geologists runs deeper and concerns more funda- mental aspects of the earth's history than hot water. A Grand Scheme Only at Iceland does the 45.000- kilometer-Iong mid-ocean ridge, a rift marking the pulling apart of the earth's crustal plates, breach the sur- face of the ocean. This island is. there- 45.000 kilometers - ?7'I00 miles LOOKING north along Iceland's central nft. one can see where the earth's crust is slowly being pulled apart. To the left of the rift, the western Atlantic Ocean and North America to as far as the San Andreas Fault in California are drifting west at a rate of I centimeter each year. To the right of the rift, the eastern Allan- tic Ocean and all of Tunisia are drifting eastward to as far as the Pacific trenches off Kamchatka and Japan. In Iceland, the rifting is strongh overprinted h\ com- panion effects — the formation of vol- canoes and the effusion of lava above a vast ascending plume of magma, rising from deep within the earth's mantle. 10 fore, crucial to the revolutionary new concept of plate tectonics, or struc- tural geology of the earth's crust. {Also see "A Magnificent Revolu- tion." Sea Frontiers. Vol. 18. No. 6, November-December. 1972.) According to plate tectonics, the earth's crust is a mosaic of about eight 100-kilometer-thick rigid plates, or shells, which slowly drifts over a 100 kilometers = 62 miles 268 plastic upper mantle. The plates do not collide with one another: instead, one edge subducts, or descends, into the earth's mantle while the opposite edge accretes new ocean floor to its margin. The latter process occurs at the mid-ocean ridge and is called sea- floor spreading. Along still other boundries of a crustal plate are giant zones of shear, or transform faults, where a plate slides past its neighbor. Geologists only recently have come to understand this grand scheme of earth tectonics because the evidence is largely not on land but out of sight beneath the sea. Tectonism. or per- manent displacement of the earth's crust, is confined to the plate bound- aries which, although over 100.000 kilometers in total length, are nearly all in oceanic crust. Major exceptions 100 000 Hometers = 62 000 miles Roberts Dietz 269 are California's San Andreas Fault (a transform fault); Africa's Afar tri- angle (a triple junction where three plates join) at the nexus of the Red Sea and Gulf of Aden: and Iceland, the only place on earth where the mid- ocean ridge is above sea level. Three Types of Volcanism Geologists recognize three distinct- ly different types of volcanism. or lava production, on the earth. The first is suhduction (calc-alkalic) vol- canism associated with the oceanic trenches and island arcs. It is caused by the return of molten rock to the surface from crustal plates that are being subducted, or carried down, into the earth's mantle. The lavas are charged with steam and thus are high- ly explosive. They create the classic volcanoes around the Pacific "ring of fire" such as Mount Rainier in the United States and Mount Fuji in Japan. The second type of volcanism (tholeiitic) is that which injects the dikes and pillow lavas that fill in the mid-ocean ridge as it spreads apart. This process, which generates new ocean floor by symmetrical accretion to the plates that are moving apart, is called sea-floor spreading. This vol- canism is effusive and quiet, produc- ing dikes and flows, but not a single volcanic cone. Although never di- rectly observed, this type of volcan- ism adds more to the earth's crust than either of the others. It repaves the ocean floor along the mid-ocean ridges around the world at the rate of 2 square kilometers per year- enough to renew the entire ocean floor in only 1 50 million years. A third type of volcanism is plume (alkalic) volcanism, caused by lavas that rise as ascending columns from deep within the earth's mantle. Upon breaking through the earth's crust, they create volcanoes that may be compared to the thunderheads that form over ascending columns of air. Plume volcanoes usually form in a row as the magma rises from the fixed deep mantle over which the earth's outer crust is drifting. The Hawaiian Chain is a good example. The Pacific plate is moving northwest at about 10 centimeters each year so that, as the old volcanoes drift away, a new one is created over the fixed plume site. The only modern active volcanism is on the big island of Hawaii, at the southeast end of the chain. Iceland has been built by the last two named types of volcanism: tholeiitic (rift injection) lavas and plume lavas. Among the chief centers of plume volcanism on earth, Iceland probably ranks first, spewing out about 20 percent of all surface lavas. (Other major centers of plume vol- canism are Hawaii, the Galapagos Islands, and the Azores.) In Iceland, the rift lavas are abun- dantly augmented by plume lavas, which have built more than 200 vol- canoes, many of them active. This volcanic pile that straddles the Mid- Atlantic Ridge is thus of a composite nature. The process of sea-floor spreading (rifting) observable in Ice- land is strongly "overprinted" by plume lavas. Accordingly, the spread- ing process within Iceland is more complex than the beautiful simplicity 2 square kilometers = 0 77 square mile 10 centimeters = 39 inches 12 Sea Frontiers 270 ASTHENOSPHERE This mcim simim n n i> sketch illustrates the process of marginal plate accretion, or sea-floor spreading, which lakes place at ihc mid-Atlantic rift. As the America plate 1A1 and the Eurasia plan- 1D1 are pulled apart, a dike <>l hot lava is injected into the earth s lithosphere. or crust. Partial melting in the soft asthenosphere (the region below the crust I provides cm ever-present source of new magma. I he hot dike I speckled I cools ihluci against the adjacent plates. With renewed extension tinsel I. the dike breaks svtnmeiricullv along us warm and. hence, weak axis. L'pon congealing and passing through the so-called Curie point at ^~5°C .. the upper portion of the dike takes on the ambient sense of the earth's magnetic field. The while bands are intervals when the earths magnetic field has been normal: the black ones indicate intervals of reversed magnetic fields. observed farther to the south along the Reykjanes Ridge, and along most other portions of the world-wide mid- ocean ridge system. While it is true that the mid-ocean ridge does hare its back at Iceland, this exposure is somewhat anomalous, complex, and atypical. The geologic structure of Iceland is dominated by two giant rifts which trend, generally, north-south. These rifts are clearly discernible from the air. The eastern rift is now compara- tively inactive, anil it is believed that active rifting, or spreading apart, is lamely confined to the western rift. January-Februdfy 1 ')/<» 13 271 The 0' :f theOcea- : R c^es 3> Ej:- 1959 b> S: e-r ( c Arr,er can Inc. A;: '.ghts reserved Flown bi na\ 'i aircraft over the cresi of the Reykjanes Ridge south of Iceland, this magnetic survey shows, in red. the present period of normal magnetism (during which the north magnetic pole has been near the north geographic pole). This period extends back to "DO. 111)1) years ago. and it overlies and flanks the axis of the mid-ocean ridge. Other rainbow colors mark earlier periods of normal magnetism back to anomaly five which occurred about in million vears ago. Intervening periods of reversed magnetism i w hen the north magnetic pole was near the south geographic pole) are shown in white. The anomaly patterns are symmetric, as each injected dike eventually split into two equal parts which accreted to opposite plates. This sun ey thus provides "back-to-back tape recorders" ot ocean-floor growth. Each limb of anomaly patterns is lot) kilometers wide which means, since anomaly five is 10 million years old. a growth, or spreading, rate of ID kilometers per million vears or 1 centimeter per year. This is a separation rate for the two plates of 2 centimeters per year. The Reykianes Ridge is. therefore, splitting apart at the rate at which a fingernail grows. 14 Sea F-cn: ers 272 Down-dropped blocks of basalt, a dark lava rock, reveal that there has been extension within the earth's crust. The earth's outercrust. or litho- sphere, is computed to be pulling apart at a rate of about 2 centimeters per year. Attempts have been made by scientists from Imperial College in London to actually measure this rifting, using a laser beam. The re- sults, thus far. are not conclusive but are said to be consistent with the theoretically computed 2 centimeter- per-year spreading rate. No clear pat- tern of magnetic anomalies are ob- servable along the Icelandic rift, but this seems certainly related to two factors: the confusion created by the plume lavas and the fact that strong magnetic imprinting occurs only in the quickly quenched pillow lavas, which must be erupted beneath water. Magnetic Anomalies Reference to the Reykjanes Ridge laying athwart the Mid-Atlantic Ridge, immediately to the south of Iceland, is convincing evidence that the Icelandic rift was created by sea- floor spreading. In fact, this process, although inferred earlier by geologic considerations, was first demonstrat- ed by an aerial magnetic survey flown across this ridge. This survey revealed a succession of stripes or bands of strongly magnetized rock with their magnetic signal being alternated, i.e.. normal and then reversed in sign. The banding, or stripes, of this survey quantitatively measured the growth of the ocean floor in a manner some- what analogous to the growth of a tree by its annual rings. The anomalies revealed that the earth's magnetic field switches its polarity, so that the north pole becomes the south pole (and vice versa), about once every one half million years. The ambient direction of the earth's magnetic field is frozen into the mid-ocean ridge lavas as they pass through the so- called Curie point at 575°C. when solidifying. A central band running along the axis of the mid-ocean rift was found to be normally magnetized. With re- spect to this central band, the others on either side of the ridge lay in mirror image, so that, if the survey map was folded into a V along the axis, the anomalies on opposite sides of the fold would be juxtaposed. Clearly these were vertical growth lines dem- onstrating that the ocean floor had grown bv some process whereby new sea floor was being slowly accreted to crustal plates moving apart from the mid-ocean rift locus. Although these remarkable magnetic anomalies could not be traced through Iceland, lava ages showed that a similar pat- tern existed. The strips of lava are progressively older on both wings of Iceland as one moves away from the central rifts. Iceland is thus not only a remote island of vivid contrasts in the far North Atlantic touching on the Arctic Circle. Its mountains, volcanoes. geysers, and thermal springs have a deeper significance. Its rugged youth, with no portion being older than 15 million years, can now be understood. It is the only place on earth where one can actually observe the earth's crust being pulled apart. 2 centimeters = 0 78 inches 575°C = 1.035°F January-February 1976 15 273 27 Reprinted from: Oaeanus, Vol. 19, No. 4, 19-22. EARLY DAYS OF MARINE GEOLOGY BY R.S. DIETZ AND K.O. EMERY We hold no brief for the "good old days" but perhaps it adds to the perspective of marine sciences, and certainly to humor, to recall something of the beginnings of marine geology in the United States by citing some of our early experiences. This stibdiscipline of geology commenced almost simultaneously in the mid- 1930s on the East Coast at the Woods Hole Oceanographic Institution with the research of Henry C. Stetson and on the West Coast with the studies of Francis P. Shepard. Stetson died at sea aboard Atlantis off Chile in 1955, while Shepard is still actively working at the Scripps Institution of Oceanography in La Jolla, California. We were the first of Shepard's sixty or so marine geology students, shuttling with him between the University of Illinois and Scripps. We met at the University of Illinois, where we arrived via modes of transportation that were the norm for those Depression days. Dietz arrived by hitchhiking from the East Coast; Emery came by train, riding boxcars from San Diego. In 1936 Shepard received a grant from the Penrose Fund of the Geological Society of America for studying submarine canyons and the sea floor generally off the coast of California. The amount was SI 0.000, which was a handsome grant for those days— in fact, the largest ever given by the GSA in prewar years. With the money he was able to charter the 96-foot schooner E. IV. Scripps of the Scripps Institution of Oceanography for six one- month cruises, build the necessary scientific equipment, employ us as his assistants at a salary of S30 per month, and support the abortive development (to the tune of S 1 ,000) of the Varney- Redwine hydrostatic corer. It was hoped that this latter device would outperform the famous C. S. Piggot gun corer, which shot the barrel into the ocean bottom. We should add "in principle," because the Piggot device, when used from Atlantis, seemed to obtain cores of equivalent length whether or not the gun actually fired. A subsequent grant provided for three more months in the Gulf of California during the fall of 1940. Since bed and board was provided aboard ship, we both signed on for $1 for the three months to make us official expedition members (but Scripps never paid the Si -perhaps fearing that we'd spend it unwisely). An interesting guideline also was that students should not receive any pay for research that pertained to their own thesis projects. The low funding at least required us to develop some ingenuity in devising simple, inexpensive instrumentation. For example, we used the 2-meter-long Roger Revelle, later director of Scripps Institution of Oceanography, as a wave staff by having him stand at various distances from shore in the buffeting surf. This rather absent- minded wave staff also was noted for having stepped into a bucket while measuring cores aboard ship and wearing it for a couple of hours. As another example, we organized a rock preparation and sedimentation laboratory for which a budget of S50 per year was arranged. This was considered a reasonable proportion of the Scripps' overall budget of SI 25,000 per year. Notably also. Woods Hole Oceanographic Institution was founded in 1930 with a gift of S3, 500,000 from the Rockefeller Foundation received over a period of several years; this was sufficient to construct the large brick Bigelow Building and the ketch Atlantis, and to cover all operations for ten years. The annual budgets for 274 19 The 96-foot schooner E. W. Scnpps. principle research vessel of the Scripps Institution of Oceanography from 1937 to 1 950. (Courtesy of SIOj the two institutions have remained about equal, nowadays almost S22 million for Scripps and S20 million for Woods Hole. Life aboard E. W. Scripps was somewhat different from shipboard duty today. The ship's crew consisted of only four persons— captain, engineer, deck hand, and cook; the scientific party was seven-the number of bunks available. We generally worked around the clock, six hours on and six off. The scientific party was expected to be sailors to run the ship and technicians to operate oceanographic winches, assemble and use the water and bottom samplers, and do various shipboard analyses for water chemistry. Among our duties while steering the ship was to tabulate by hand the water depth every two minutes. We did this with great enthusiasm since we had installed aboard the latest Submarine Signal Co. fathometer, which indicated the depth on a revolving red-flashing neon light. Graphic recorders had not yet been invented, so this instrument represented to us a remarkable advance over the sounding lead. And. in fact, we continued to use the hand-powered wire-and-lead sounding winch installed on a rowboat for making hydrographic surveys of the inner heads of several submarine canyons. Rather remarkably. it was possible to demonstrate that canyon heads were repeatedly filling with sediment and then emptying out. Prior to the cruises we built dredges, grab samplers, sediment traps, and corers. The best corer that we constructed was a 600-pound open- barrel gravity model that increased the weight of such devices over earlier models by a factor of ten. We purchased junk load at 34 per pound, used scrap 2^-inch pipe, and built two corers for about S50 each. It was not until after the war that we heard about Kullenberg's invention of the piston corer. Nevertheless, we commonly obtained cores 12 feet long, and in one instance, a diatomaceous ooze core in the Gulf of California 1 7 feet long for a new record. Of course, things were considerably cheaper in those times. By way of example, an apparently wealthy American tourist at the local swinging bistro named El Tecolote (The Owl) in Guaymas. Mexico, generously offered to buy beer for our ship's staff. When he discovered that the bartender could not change his U.S. S10 bill, he gallantly said, "Set up the whole amount in beer." One hundred and twenty bottles of Carta Blanca were lined up along the bar, and as was customary then in Guaymas, bowls of unshelled shrimp were thrown in like the free peanuts of today. A side advantage was that the long row of beers immediately stopped the girls' pestering us for drinks. After the cruises, when the ship was unloaded so that the biologists or physical oceanographers could take their turns, we were able to study the samples and other results. Since all was new, both to us and to others, we had no difficulty in finding problems. Our masters' theses were on mechanics of coring and on the extensive phosphorite deposits we discovered covering many of the offshore banks. Our doctoral dissertations were on clay minerals of the deep areas ;md on rocks of the shallow banks. We recognized that the offshore basement geology of the Southern California borderland belonged to the Franciscan province. Articles on terraces, currents, barite concretions, and transport of rocks by kelp and sea lions were by-products. The overall results were incorporated into Special Paper 31 of the Geological Society of America by Shepard and Emery, a monograph treating submarine canyons and the general bathymetry of the sea floor off California. These must have been our most productive years in terms of variety and number of investigations, because of newness of the field and, probably, the aid of funds too small to permit much diversion of time and energy. 275 Local transportation was provided by a succession of old cars, starting with a 1928 Chevy that Shepard bought for us for $50. By the time We drove it 1 5 miles, cork in the transmission wore out and serious noises developed. Replacement by junk gears extended the life of the Chevy for a year or so. After tiring of having to tie a rope around the car to keep the doors closed, we swapped it for a 1928 Reo that had a good engine but bad tires. Eventually, this was swapped for Walter Munk's 1028 Buick (The Queen Mary). The state of the Reo's tires is illustrated by a blowout of the spare tire in the hot California sunshine when he drove northward too long. In time the differential of the Buick disentegrated, and a 1928 Ford was next. The total cost of these four cars was $200-nothing •? Emery with hollow giant worm or animal tube of enigmatic origin dredged from the wall of Dume Canyon off California, May 1 938. Dietz with gravity coring device at rail of E. W. Scripps, about June J 938. compared to their present value as antiques if they had been stored until now. The four years of cross-country commuting, cruising (at least I 2 months aboard E. W. Scripps), and study came to an end in 1941. Just before receiving his doctorate, Emery wrote 135 individual letters, blanketing the entire country, seeking employment. Dietz, being congenitally lazier (or possibly more efficient), trusted that this blizzard of inquiries would produce several plums of which he might select one after Emery made his acceptance. But the market for marine geologists, like the job market for poets then and now, was bleak. Not a single position was tendered. As with many products, there is commonly no demand for the first ones off the line. Even the U.S. Navy saw no particular need to know anything about oceanography; in fact, its interest, when it did develop, probably stemmed from the initiative shown by the Army Air Corps in setting up a group of officers and civilians to predict the paths of downed airmen in their rubber life rafts carried by surface currents of the western Pacific. In retrospect, the "good old days" were both the best of times and the worst of times. Happily, one tends to recall the ups rather than the downs— and there is no substitute for the bouyancy 276 21 of youth. Oceanography of today is, of course, much more sophisticated and the results ever more quantitative. But there was a certain enjoyable simplicity, and even beauty, in working with instruments that had less than one vacuum tube, let alone one transistor. The need to do all kinds of work gave us a broad view of the ocean such that we were oceanographers and not just marine geologists. We even thought we understood physical chemical, and biological oceanography. Working on the low-freeboard E. W. Scripps with decks awash gave an intimate feel for the oceans such as experienced today only by scuba divers. As we write this note Charley Hollister is putting out to sea with his "Super Straw," the giant 4'/i-inch coring device, and a new generation of marine sedimentologists. They will study complex seabed forms, subbottom acoustically reflecting layers, and mass physical properties of muds. In this work they will be guided by the multisensor MPL Deep-Tow, a real-life dream machine. All in all, a million-dollar effort. Yes, times have changed-and for the better. R. S. Dietz is a research oceanographer at the National Oceanographic and Atmospheric Administration, Miami, Florida. K. O. Emery is Henry Bryant Bigelow Oceanographer at Woods Hole Oceanographic Institution. Dietz with sediment trap and Emery and Shepard with wire sounding machine setting one for the survey of La Jolla Submarine Canyon, November 1 938. Ill 28 Reprinted from: Geoloqy , Vol. 4, No. 7, 391-392. El'gygtgyn: Probably world's largest meteorite crater +<*' »*^^* o KM e 10 15 20 IE EE Figure 1. LANDSAT images of the Siberian meteorite crater El'gygytgyn. Remarkable circularity of feature is revealed in upper snow- covered winter view (dark spot near center is cloud shadow, not island). Lower scene shows central lake in ice-free summer conditions. GEOLOGY, v. 4, p. 391-392 Robert S. Dietz National Oceanic and Atmospheric Administration Miami, Florida 33149 John F. McHone Department of Geology, University of Illinois Urbana, Illinois 61801 ABSTRACT LANDSAT imagery indicates that El'gygytgyn in northern Siberia is probably a giant meteorite crater, the largest Quater- nary impact structure on Earth, and not a tectonic depression. This probability is supported by the remarkable circularity of the crater, as outlined by (he ring-mountain rim, remoteness from modern volcanic sites, and lack of collapse scalloping of the margin. Siberia appears to have an unusual attraction for cosmic bodies. The Tunguska comet head exploded above central Siberia in 1908, and the Sikhote-Alin nickel-iron meteorite struck eastern Siberia in 1947. A million or so years earlier, Siberia probably was the site of the largest crater-forming meteorite impact to strike the continents in modern times. We refer to the El'gygytgyn crater (sometimes transliterated El'gytkhyn; lat 67°30'N, long 172°00'E), in the remote Anadyr Mountains of eastern Siberia; this crater, by morphologic criteria, appears to be a meteorite impact site. Its 18-km diameter would make it by far the largest meteorite crater on Earth, far exceeding both the Lake Bosumtwe crater in Ghana, 10.5 km across, and the New Quebec crater of Canada, 3.2 km across. There are, of course, larger ancient impact sites or astroblemes, but these are now so deeply eroded that they are the roots of craters that are no longer craterform. El'gygytgyn has been previously listed as a possible impact site by Dence (1972), citing Zotkin and Tsvetkov (1970), who listed the diameter as only 12 km, which is approximately that of the rudely circular lake occupying the center of this depression. LANDSAT imagery (Fig. 1) reveals the remarkable circularity, symmetry, and elevated rim of the overall crater, 18 km across. This circularity is easily overlooked on maps because of carto- graphic emphasis of the lake shoreline, as large sediment aprons have filled parts of the crater (Fig. 1, lower) and produced an irregular form. 391 278 KM 0 5 10 3_ a: Figure 2. LANDSAT image of Crater Lake, Oregon, showing scalloped walls, central volcanic island, and asymmetric perimeter typical of calderas. El'gygytgyn is a unique feature of the maturely dissected and nonglaciated Anadyr mountainland. The nearest active volcanoes lie 600 km away on the Kamchatka Peninsula. This craterform depression is asymmetrically filled with a 170-m-deep lake about 12 km across and of squarish outline. The crater is outlined by a ring mountain that attains its greatest relief to the west. The rim is breached by a river in the southeast quadrant, the outflow of which eventually reaches the Belaya River of the Pacific watershed. The highest elevations along the rim are about 1,060 m, according to the U.S. Air Force operational navigational chart. This rim thus stands about 450 m above lake level, or 620 m above the lake bottom. Extensive talus aprons along the western half of the crater are now being entrenched, suggesting that the lake level was once higher than at present. The almost perfect circularity of the depression is enhanced in the winter LANDSAT image because snow masks the outline of the lake shore. The mature degree of erosion suggests that the crater was created one to a few million years ago. El'gygytgyn was discovered in 1933 by S. V. Obruchev from an aircraft, according to Nekrasov and Raudonis (1973). The lake- filled crater immediately attracted attention because of its unusual shape. Obruchev expressed the opinion that it was a volcanic crater or caldera of vast dimensions, yet there are no young vol- canic rocks associated with the feature. Further negative evidence is provided by LANDSAT images of calderas that are quite unlike El'gygytgyn. Figure 2, for example, is a LANDSAT image of Crater Lake, Oregon, one of the world's best examples of a caldera. Crater Lake is situated on top of a large volcanic dome, and although it is rudely round, its circularity is spoiled by its scalloped margin. This was created by land slippage, subsidence associated with magma withdrawal, and subsequent internal evisceration by ash eruptions. A caldera has been aptly described as "a volcanic crater whose head has fallen in when its insides were blown out." El'gygytgyn does not have the geomorphic aspect of a caldera, which is invariably situated atop a large vol- canic come and which occurs in chains or groups and not as a solitary feature. El'gygytgyn is also larger than most calderas and more symmetrical. Nekrasov and Raudonis (1973) have described El'gygytgyn as a collapse feature of unspecified origin. One must assume that a subsidence of this magnitude would necessarily be tectonic. Such an origin, however, would not account for the ring of moun- tains unless the structure underwent domal uplift by injection of magma prior to collapse. In this event, the remarkable circu- larity would remain unexplained. Nekrasov and Raudonis (1973) studied eight rock samples collected from the north and northeast parts of the ring mountain and found them to be an assortment of silicic, intermediate, and mafic igneous rocks, including both intrusive and extrusive types, of probable Mesozoic age. These specimens appear to be country rocks rather than products of the crater-forming event. Nekrasov and Raudonis concluded that the crater could not be an impact site because they detected no coesite in thin sections. This conclusion is unjustified, as coesite is virtually unrecognizable in thin section, and, in any event, shock overpressures at an impact-crater rim are already far below those needed to create this high-pressure silica polymorph. In general, shock metamorphism and shatter coning are never found in situ beyond one-half of the radius of an impact crater from ground zero. We conclude that El'gygytgyn is probably the world's largest modern impact crater. REFERENCES CITED Dence, M., 1 972, The nature and significances of terrestrial impact structures: Internat. Geol. Cong., 24th, Montreal 1972, Hroc, sec. 15, 1'lanetology, p. 77-89. Nekrasov, I., and Raudonis, P., 1973, Meteorite craters (translation from Russian ms.): Ottawa, Canada. Canadian Translation Bureau. Zotkin, I. T., and Tsvetkov, V. I.. 1970. Searches for meteorite craters on earth: Astron. Vestnik, v. 4, p. 55-65. ACKNOWLEDGMENTS Reviewed by Peter Rona. The work of John Mcllone was supported by a grant-in-aid for meteoritic research from the Barringer Crater Company. MANUSCRIPT RECEIVED MARCH 22, 1976 MANUSCRIPT ACCEPTED APRIL 27, 1976 392 JULY 1976 279 29 Reprinted from: Proa. American Society of Civil Engineers Specialty Conference on Dredging and Its Environmental Effects, Mobile, Al . , 26-28 January 1976, 936-946. DEPOSITION AND EROSION IN THE DREDGE SPGI- AND OTHER NEW YORK BIGHT DUMPING AREAS By: Ge&rge L. Freeland' and George F. Merrill' INTRODUCTION The disposal of solid wastes from the New York C'ty metropolitan area is the cause of considerable environmental conce'"", as most of these wastes are dumped in marine waters outside of the harrcr mouth (Table 1). Dredge spoil and sewage sludge constitute over 94" of the volume of material dumped containing solids. In 1973 the National Oceanic and Atmospheric Administration (NOAA)-, under the Marine Eco Systems Analysis (MFISA) Project, initiated research to determine the effect of dumping in the New York Bight. A new hydrographic survey of the Bight apex was immediately started to determine what changes had occurred in bottom topography since the last previous survey in 193&. Some results from this survey are presented here. HYDROGRAPHIC SURVEYS Hydrographic surveys have been made in the New York Bight since 1845 by the U.S. Coast and Geodetic survey (now the National Ocean Survey, NOS, part of NOAA). Trie last U.S. C.&G.S. survey to cover the Bight apex, the area immediately adjacent to the harbor mouth where dumping is most intense, was in 1936. Comparison with the 1S45 revealed the development of several knolls due to early dumping (4). Our 1973 survey had depth sounding lines spaced 1000 ft (305 m) apart over an area approximately 15 nautical miles (28 km) square (Fig. I). Data from this survey were then compared with data from the 1935 survey to produce a net-change map. NET BATHYMETRIC CHANGE Examination of the boat sheets (detailed maps showing final data plots) from the 1S36 survey (NOS No. H-6193) revealed trackline spacing of approx- imately 0.5 nautical miles (900 m) versus 1000 ft (305 m) spacing for the 1973 survey, and divergence of trackline directions. In order to compare the two surveys, boat sheets from both surveys were cqntoured on a 3 ft. (0.92 m) contour interval (see Fig. 1 for the 1973 map). A 1000 ft. (305 m) grid was then prepared for the entire area and overlaid on both naps. From TT National Oceanic and Atmospheric Administration, Atlantic Oceanographic and Meteorological Laboratories, 15 Rickenbacker Causeway, Miami, Florida 33149 l>36 280 Nl W YORK I'.iCHl 93 Fig. 1. 1973 Bathymetric map of the Mew York' Bight apex. From a NOAA survey. Contour interval one meter. 281 938 DREDGING EFFECTS plotted data on the boat sheets and interpolations between contour lines, a value was picked for the center of each 1000 ft (305 m) square for each survey. These numbers were added algebraically to produce a positive or negative number for each square representing erosion or deposition "'n that square for the 37 years between surveys. The values were corrected for a sea level rise of 0.62 ft (0.189 m) from N0S mca monthly sea levels at Sandy Hook, N.J. They 'were then contoured to produce the net-change map (see Fig. 3). Volumes of erosion and deposition were calculated by pi animetering all contours and multiplying these areas by-the appropriate contour inter- val. Appropriate voluir.es were added for slope sediment between contours. Areas, volumes of erosion and deposition, and net changes for Bight apex features are listed in Table 2. DISCUSSION: ANTHROPOGENIC SEDIMENTS Sediments are introduced into the Bight apex almost entirely in the form of fine-g-ained matter. Dredge spoil constitutes the most important source of sol'ds brought in by man (anthropogenic sediment) (Table 1 ) .■ Estimates of the total amounts of dredgings ba-ged from 1936 to 1973 (records are unreliable prior to 1954) indicate that about 136 x 10° cu. yd. (142 x 10D m3) were dumped, compared to 162 x. 106 cu yd (124 x 106 r:3) calculated on the basis of net bathyinc-tric change for the dredne sooil duinpsite and the dumping areas near Ambrose and Sandy Hook Channels (Table 2). This indicates that approximately 87% of the material barged is still in place on the bottom. Detailed mapping of the dredge spoil dumpsite shown that shoaling of up to 34 ft. (10.36 m) has occured over an area of 11 square nautical miles (36. Km?) south of a knoll which itself was formed by earlier dumping (Figs. 4-6). udge is appa, easily resuspended and dispersed, as only traces of sludge can be found at the designated site. An unknown fraction settles in the Cnristiaensen Basin and the upper Hudson Shelf Valley where it mixes with natural muds, the remainder beifj dispersed by the water column. For this reason, the sludge dumpsite is not listed in Table 2. Differentiation of sludge from natural muds, mostly by chemical means, is currently undergoing intensive study by NOAA. Although the volume of sludge barged is considerable, and will increase in the future as more plants come on line in the New York area and percentage treatment improves, solids by weight do not constitute an important addition to sediment volume on the bottom (Table 1). Contami- nation of bottom sediment and sediment suspended in the water column by 282 NLW YORK BIGHT 43V Source of Solids Transported into Marine «ate of the New York Bight SOURCE , VOLUME 10 ,Cu,yd/yr1 .. of (10 m) J barqod c WEIGHT 10 .-short tons/yr (10 metric tons) of barged . ' of total input Dredge spoil 8.35 (6.38) 62.4 5.21 (4.73) 85.8 1 Sewage sludge (3.27) 32 0.2C3 (0.184) 3.3 2.1 Cellar dirt O./fc (0.5?) b. ,' 0.55;' (0.r,S, l,.8 5.S Total Earned 13. 39 (10.24) 100. i (hh Atirosoheric 0.403 1 e.i 1.0.447: Wastewater* Municipal 0.39 i (0.35) 4.0 Industrial | U.2 (0.2) 0.2 Runoff* Gaged 1.5 (1.4) 1 16 Urban 1.2 12 (i.D ! ; Total input 9.6C (8.76) 100. 2 From Ref. 3 * 98' of these coastal zone inputs come through the Pockaway - Sandy Hook Figur1:. do not include shelf-derived sediment from outside the Bight. TABLE 2 Volumes of Erosion and Deposition in the New York Bight Apex between 1936 and 1973 Area nm^ (Km?) Volume 106 cu yd. (106 m3) Erosion Deposition 1 Net Change 1. Entire Apex 209 (718) 182 (140) 212 29 D (162) | (?7) i. TJredge Spoil Oumpsite 11 (36) 122 122 0 (93) , (<)<) T Cellar Dirt Dumps i te (8) 6 6 0 77 Ambrose 4 Sandy Hook Channels 25 (86) 63 (48) 40 ,_ E Ml) ;i-' 5 . I Anthropogenic 38 .(130) 63 (48) ies io6 ■.' (1?1) (8!) 57 Christiaensen Basin (83) 13 (im 8 5 E (61 ; (4i /. Hudson Shelf Valley' 7 (23) 12 (10) 3( IC E ii. v. non-anthropoqenic 171 (507) 120 (9!) 43 76 C- 1. Area between 14 and 20 fathoms (26-37 m) north of 4024'N. 2. Area deeper than 20 fathoms (37 m) north of 40'19.22'N. 3. Equal to a layer 0.106 in. (2.7 mm) thick per yen'. Some figures may not agree due tc rounJing off. 283 940 DREDGING EFFECTS 73°45' 40°35' 40°30' 40°25' 40°20' - 73°40' _L _L l I _l_ 74°00' 73855" 73°50' 73°45' 73°40' 40°35' 40° 30' 40°25 40820' Fig. 2. Tracklines of the 1973 bathymetric survey. Light lines show track- lines for bathymetry only. On heavy lines both bathymetric and geo- physical data were collected. 284 M W YORK BIGHT 94 i BATHYMETRIC NET CHANGE 1936-1973 74°00' 73°55' 73°50' 73°45' 73°40' 74°00' 73°55' 73°50 73°45 73°40' DEPOSITION 0-6 FT DEPOSITION >6FT | EROSION 0-2 FT I EROSION >2FT Fig. 3. Net bsthymetric change, N.Y. Bight Apex, from 1936 to 1973. 235 9a: DREDGING EFFECTS Fig. 4. N.Y. Bight dredge spoil dumpsite. 1936 Bathymetry. Line marked 198°T and hachured area show the designated dumpsite (Figures 4-6) based on 1936 soundings to lie within the 90 ft. isobath. 286 NEW YORK BIGHT 94? JJJ^tJ »»"«««]|; i j j (aa 8 otitis J» nS5» J*^** gO „„„» ,.„,r-K..,-. ..S8'- ...•■•v^:t[;.;; '*$$« : * c e I i nmiimiiiliiii'M"'""6, 1 50' IMftitiggfS NEW YORK BIGHT APEX DREDGE SPOIL DUMPSITE 1973 BATHYMETRY CONTOUR INTERVAL 5 FEET 0.L. FREELAND NOAA AOML 2-74 Fig. 5. N.Y. Bight dredge spoil dumpsite. 1973 Bathymetry. 287 944 DREDGING EFFECTS I 73*52' 73°51' 73" 'so' 40«26" DREDGE SPOIL DUMPSITE, N.Y BIGHT NET CHANGE MAP 1936-1973 CONTOUR INTERVAL 5 FEET G.L. FREELAND 2/74 NOAA-AOML 73°50' I 40"25- 40«24- 3 40«23'— 40°22 • Fig. 6. N.Y. Bight dredge spoil dumpsite. Net change in depth from 1936 to 1973. Note that the 47 ft. knoll in the 1936 map is essentially un- changed, and that a large volume of material has been dumped north of the start (northeast end) of the arrow designating the minimum distance (4 nm) from Ambrose Light that dumping should be initiated. 288 N1W YORK MICH ! 94! organic pesticides and heavy metals in sludge is, however, a serio r, concern. Cellar dirt, the third anthropogenic sediment, consists of c ■ r.tion rubble from demolition, foundation rock and dirt, and slag. Brie- . norphic rocks, and red sandstone are commonly recovered in grab S3 ;• - ; . Cellar dirt, while making a recognizable spoil mound, is not considered an important pollutant because of low volumes and the absence of toxic chemicals. YATURAl SEDIMENTS Natural sediment input from land sources comes mainly from stream runoff from the Hudson River drainage basin and urban runoff from the New York mecropol itan area (Table 1). These sources are relatively easy to -easure compared to sediment transported from other areas of the shelf. Various estimates of sediment transport indicate that, for the eastern U.S. continental nwgin, a) 90% of the sediment from land sources is deposited in estuaries and wetlands; b) net suspended fine sediment transport on the shelf is probably landward, with possibly much of the material finally settling in estuaries; and c) recycling (resuspension and settling) of sediinent on the shelf may transport orders of magnitude more sediment than either enters or leaves the shelf (1, 2). From our net change map, we have calculated volumes of natural sediment eroded and deposited in the Bight apex (Table 2). After subtracting the anthropogenic naterial in the Ambrose - Sandy Hook Channels area and in the dredge spoil and cellar dirt dumpsites, the volume of material eroded exceeds deposition by 76 x 106 cu yd (58 x 1 0& m^), equivalent to a layer 0.106 in. (2.7 mm) thick over the non-anthropogenic areas. From other ongoing studies, it appears that this erosion occurs primarily during storms which pass most frequently in winter months. ERROR SOURCES Tidal corrections from the Sandy Hook tidegage were made on all records and are not considered to be a significant error source. Bar checks for fathometer error were made before and after daily operations at sea. Surveying was not done or was rerun if fathometer readings were incorrect. After contouring W3S completed, data from some small areas in the northern Christjaensen Basin became suspect because of sinusoidal wiggles in contour- lines which varied systematically with tracklines, amounting to a maximum of about 1.5 ft. (0.457 m) difference between adjacent tracklines. After reruns of bathymetry lines in an east and west direction (perpendicular to the original tracklines), it was determined that the "waves" originally mapped in the bottom were real, although of lower amplitude, in one area, and non- existant in other areas where the original waves were lower than maximum amplitude. Further analysis indicated the source of error to be "squat" of the 289 9-56 DR! DGING 1ITK TS survey boat (level cf ride of the boat in the water) on geophysical tracklir.es in certain areas due to towed instruments. These errors were corrected in the final contour nap and at least partially compensated for in the net chancie map. Maximum error is estimated to be + 0.5 ft. (0.1524 m). ACKNOWLEDGEMENTS Grateful appreciation is given to the Corps of Engineers Operations Section of the new York District, contractor for the 1973 survey: Mr. Lewis Pinata, Acting Chief, Mr. Dennis Suszkowski , Oceanographer, Mr. Herbert w--':- and Mr. Bill Musak, Chief and Assistant Chief of the Survey Branch, and :■ -:• crew of the survey vessel HATTO'I , which did tne northern half of the survey. Thanks are also given to Mr. Robert Spies, and Mr. Robert Wagner, Chief and Assistant Chief of the Survey Branch of the Corps Philadelphia District, and to the crew of the survey vessel SHUMAN, which did the southern part of the survey. Mr. George Lapiene, electronic technician at AOML in 1973, was aboard both survey vessels during the three months of work on geophysical track! ines Dr. Anthony E. Cok, Professor of Geology at Ade 1 phi University, Garden City, New York, was aboard the HATTON during geophysical trac'kline work. Dr. John J. Dowling, Associate Professor, Marine Sciences Institute, University of Connecticut, Groton, Connecticut, made the net ciiange calculations. The Tidal Datum Planes Section, Oceanographic Division, of the national Ocean Survey (NOAA), Rockville, Maryland, supplied monthly mean sea level data for the Sand Hook Station for 1936 and 1973. Cdr. R. L. Swanson, Project Manager, MESA New York Bight Project, Stony Brook, New York, provided additional tidal correction data. Finally, Drs. H. B. Stewart, Jr., and D. J. P. Swift of AOML reviewed the manuscript. REFERENCES CITED 1. Meade, R.H., Sachs, P.L., Manheim, F.T., Hathaway, J.C., and Spencer, D.W., "Sources of Suspended Matter in Waters of the Middle Atlantic Bight", Journal of Sedimentary Petrology, Vol. 45, 1975, pp. 171-188. 2. Milliman, J.D., Pilkey, O.H., and Ross, D.A., "Sediments of the Continental Margin of the Eastern U.S.", Bulletin of the Geological Society of America, Vol. 83, 1972, pp. 1315-1334. 3. Mueller, J. A., Anderson, A.R., and Jen's, J.S.,( "Contaminants Entering the New York Bight - Sources, Mass Loads, Significance", Report sent to NOAA, MESA N.Y. Bight Project Office, Stony Brook, N.Y., 11794, 1975. 4. Williams, S.J., and Duane, D.B., "Geomorphology and Sediments of the Inner New York Bight Continental Shelf", Technical Memorandum 45, U.S. Army, Corps, of Engineers, Coastal Engineerina Research Center, July, 1974. 290 30 Reprinted from: Middle Atlantic Shelf and the New York Bight, ASLO Special Symposia, Volume 2, 90-101. Surficial sediments of the NOAA-MESA study areas in the New York Bight George L. Freeland, Donald J. P. Swift, and William L. Stubblefield Atlantic Oceanographic and Meteorological Laboratories, NOAA, 15 Rickenbacker Causeway, Miami, Florida 33149 Anthony E. Cok Department of Earth Sciences, Adelphi University, Garden City, New York 11530 Abstract In the New York Bight apex, extensive sedimentological studies and a 1973 bathymetric survey reveal that the only significant change in bottom topography since 1936 occurred at the dredge spoil dumpsite where the dumping of 98 X 10fi m3 of dredged material has caused up to 10 m of shoaling. The center of the Christiaensen Basin, a natural collecting area for fine-grained sediment, is no doubt contaminated with sludge but shows no apparent sediment buildup during the intervening 37 years. The apex outside of the Christiaensen Basin is floored primarily by sand ranging from silty fine to coarse, with small areas of sandy gravel, artifact (anthropogenic) gravel, and mud. Nearshore mud patches appear to be covered at times with sand and occasionally scoured out. Sidescan sonar records show linear bedforms, indicative of sand movement, over most of the apex area. Two midshelf areas have been proposed as interim alternative dumping areas. The northern area is in a tributary valley of the ancestral Long Island river system. Fine sands cover the northeast part and medium sands predominate to the west and south. Bottom photographs show a smooth, slightly undulatory, mounded or rippled sea floor. In the southern alternative dumping area coarse sand and gravel deposits lie on the crest and east flank of the Hudson divide, while medium and fine sand occurs in the ridge and swale topography to the west. These distributions suggest fine sediment is winnowed from the crest and east flank of the divide and deposited to the west. Veatch and Smith Trough contains a veneer of shelly, pebble sand with large, angular clay pebbles and occasional oyster shells derived from exposed early Holocene lagoonal clay. These studies suggest that if sewage sludge were dumped, widespread dispersion, mostly to the southwest, could be expected, with winter resuspension and transport of fine material on the bottom. Possible permanent buildup on the bottom could be expected if dredged material were dumped. The nature of bottom sediments and sed- maps at 1 -fathom ( Stearns and Garrison iment particles suspended in the water col- 1967 ) and 4-m intervals on the shelf and umn becomes of interest to environmental 200-m intervals on the continental slope managers when man's activities in the ocean (Fig. 1; LIchupi 1970) were made from disturb the sea floor or the near-bottom 1936 survey data. A new survey of the bight water column. In addition to the immediate was made in 1975; results should be avail- results, one must also consider the effect on able in 1977. long term natural phenomena. How are Surficial morphology of the New York these processes affected by what man has Bight, and sediment distribution across this done, or perhaps more importantly, how do surface, may be explained by sea level flue- natural processes modify what man has tuations caused by continental glaciation done to disturb the natural environment? during the past several million years. The Here we report work done at the Atlantic last glacial stage ended 15,000 years ago Oceanographic and Meteorological Labora- (Milliman and Emery 1968) when the tory as part of the NOAA-MESA New York eastern North American ice sheet extended Bight Project. as far as Long Island and northern New Hydrographic surveys of the New York Jersey. During maximum glacial advance Bight were initiated in 1936 by the Coast sea level was lowered about 160 m ( Veatch and Geodetic Survey (now the National and Smith 1939) so that the shoreline of the Ocean Survey) in nearshore areas and have bight was in the vicinity of Hudson Canyon been repeated periodically. Bathymetric ( see Fig. 1 ) . Since the ice melted, the shore- AM. SOC. LIMNOL. OCEANOGR. 90 SPEC. SYMP. 2 291 Surficial sediments 91 Fig. 1, Index to detailed study areas and topo- graphic features in the New York Bight. ( Bathym- etry from Uchupi 1970. ) Contour intervals 4 and 200 m. 1A — New Jersey nearshore ridge and swale study area and the Atlantic generating station site; IB — -New Jersey central shelf ridge and swale study area; LINS — Long Island nearshore study area; SCOA— Suffolk County outfall area; 2D1, 2D2— proposed interim alternative dumpsites. line advanced to its present position; many- features of the shelf are the result of sev- eral sea level fluctuations. Morphologic fea- tures are discussed in our companion paper in this volume (Swift et al. 1976) and else- where (e.g. McKinney and Friedman 1970; McKinnev et al. 1974; Stubblefield et al. 1974, 1975; Knott and Hoskins 1968; Duane etal. 1972; Williams 1976). Surficial sediments A comprehensive sampling program for the outer shelf was conducted by the Woods Hole Oceanographic Institution and the U.S. Geological Survey, who sampled on an 18-km spacing. The Corps of Engineers Coastal Engineering Research Center has collected about 4,200 km of geophysical data and over 300 cores as a part of its stud- ies on the inner shelf of the bight ( Duane 1969; Williams and Duane 1974; Williams 1976). MESA work had been conducted primarily in New Jersey nearshore and cen- tral shelf areas, the bight apex, the near- shore of Long Island eastward to Fire Is- land, two central shelf alternative dumping areas, and the Hudson Shelf Valley ( Fig. 1 ) . Emphasis here is on the bight apex and the central shelf alternative dumping areas. Source and age of sediments — Sediments covering the floor of the bight were mostly deposited during lowered sea level and were reworked during the landward-seaward migrations of the shoreline. As transgression progressed, fluvial and older sediments were covered by estuarine and lagoonal sediments behind barrier islands or directly reworked by littoral processes associated with the advancing shoreline. During a transgression, bottom currents of the inner shelf interact with the shelf floor to form a concave surface whose profile resembles an exponential curve, with the steep limb com- prising the shoreface (Swift et al. 1972). With a loose, sandy substrate, the inner shelf shoreface tends to extend itself later- ally across the mouths of bays, closing them, except for inlets, by the deposition of sand in the form of spits and barrier islands. Estuaries and lagoons behind these spits and islands then trap suspended fine sedi- ment (mud), while the barrier islands are nourished by littoral drift from eroding headlands and by sand moving landward from the shelf. As sea level rose during the Holocene transgression, the inner shelf profile moved shoreward by means of shoreface erosion. Some eroded sand was swept onto the bar- rier islands by storm overwash and buried, only to be re-exposed again at the eroding shoreface. Most material from shoreface erosion, has, however, been washed down- coast and seaward to form a discontinuous sand blanket 0 to 10 m thick (Stahl et al. 1974 ) . Thus, the New York Bight shelf floor is dominantly sand-sized sediment (Schlee 1973). Fine-grained sediments are gener- ally absent, having been transported either into the Hudson-Raritan estuary, behind barrier islands, or off the shelf edge. Lo- cally, underlying strata of transgressed la- goonal and estuarine semiconsolidated mud deposits or resistant coastal plain strata are exposed on the sea floor (Swift et al. 1972; Stahl et al. 1974; Sheridan et al. 1974). Sediment types — Sediment types have been mapped in the New York Bight pri- marily by dominant grain size (Fig. 2). 292 92 Geological processes Generally, the shelf is covered by sand- sized sediment with isolated gravel patches (Schlee 1973, 1975; Williams and Duane 1974; Williams 1976). In deeper water, gen- erally seaward of the 60-m isobath, in the Hudson Shelf Valley, and in lagoons and estuaries where wave action is less pro- nounced, silt is the dominant sediment ( Freeland and Swift in press ) . In the Long Island nearshore zone west of Fire Island, small mud patches, some of which are sea- sonal, are of considerable environmental concern owing to contamination of the fines by pollutants. Suspended sediments — Meade ( 1972a, /;) noted the following: Pleistocene glacia- tions and sea level fluctuations drastically altered the composition and distribution of sediments on continental margins; it is not always immediately evident whether pres- ent shelf deposits reflect modern or Pleisto- cene conditions. Fine sediment transport studies are hindered by the fact that de- posited sediments may reflect processes act- ing over thousands of years, whereas our Fig. 2. Sediment types in the bight area ( depth in meters). Hatching — gravel, sandy gravel, and gravelly sand; speckling — sand; stippling — silty sand, sandy silt, and clayey silt; dappling — glau- conitic sand, silty sand, and sandy silt. ■■■ — Pyrite- filled foraminiferal tests. 1 — Zone of rounded quartz grains; 2 — zone of limonitic pellets. ( From Uchupi 1963.) Table 1. Source of suspended solids in the New York Bight.* XlO0 tonnes /yr Direct bight ( 68% ) Dredged (54%) 4.73 Sludge (2.1%) 0.18 Cellar dirt (6.8%) 0.60 Total barged (62.9%) 5.51 Atmospheric (5% ) 0.45 Coastal zone (32%) (98% of coastal zone input is through the Rock- away-Sandy Hook transect) Municipal wastewater (4% ) 0.35 Industrial wastewater ( 0.2 % ) 0.02 Gauged runoff (16%) 1.4 Urban runoff (12%) 1.1 Total coastal zone Total input 2.87 8.83 * From Mueller et al. 1976. studies of suspended sediment are com- monly limited to a few days or months of observations. Natural processes may be im- possible to separate from the changes pro- duced by human activities, particularly in estuaries ( and at the present dumpsites ) . Fine sediment sources to estuaries and the shelf — Fine sediment discharged into the bight is shown in Table 1 ( Mueller et al. 1976). Fluvial sediment is comprised of roughly 85% inorganic and 15% combust- ible organic material (Table 2). The fine inorganic fraction is mostly illite, chlorite, feldspar, and hornblende from the Hudson River ( Hathaway 1972 ) . Shelf erosion and coast-parallel transport appear to be significant but unmeasured sources of suspended material and were probably major sources during the Holo- cene transgression. Hathaway (1972) showed that fine sediments near the mouths Table 2. Composition of suspended matter. Rivers 80-90% minerals Estuaries 60-80% minerals + biogenic shells Shelf 10-70% minerals + biogenic shells 10-20% combustible organics 20-40% combustible organics 30-90% combustible organics 293 Surficial sediments 93 of coastal plain estuaries differ significantly from the composition of overborne sedi- ments. It is probable that much estuary- mouth sediment is being eroded from shelf deposits and returned to and trapped in estuaries (Meade 1969). The fact that the sediments from modern rivers have not obscured this conclusion implies that either the modern sediment is bypassing the lower portions of the estuary, or it is trapped al- most completely near the river mouths. Along the east coast, the heads of the Ches- apeake and Delaware estuaries are far up- stream from the estuary mouth, therefore, most river sediment is deposited far inland from the sea. Although saline tidal water is present in the Hudson River up to Albany, fine fluvial sediment is carried by low- salinity surface water to Upper and Lower New York Bays where some fines settle out ( Folger 1972/; ) and the remainder is car- ried with estuarine sediment into the bight apex and mixed with recirculated shelf sed- iment. In the northeast United States, most of the fluvial suspended sediment is effec- tively trapped in estuaries and coastal wet- lands (Millimanl972). At the present, the annual suspended sediment discharge of Atlantic coastal rivers is about equal to the annual deposition on marsh surfaces (Meade 1972a). However, much of the deposited material re-enters the shelf water column after the shoreline has passed over the marsh, through the process of shoreface erosion ( Fischer 1961 ) . Particles derived from biologic processes are also a significant component of sus- pended matter in estuaries and on the shelf (Table 2), ranging from 20-90% in surface waters ( Manheim et al. 1970). However, concentrations of combustible biogenic mat- ter decrease rapidly with depth, and little of this material is preserved in sediment de- posits (Folger 1972a; Gross 1972). Atmospheric fallout over the New York Bight is small relative to other sediment sources (Table 1), but it may be a signifi- cant transport path for specific pollutants (e.g. lead from vehicular exhaust emis- sions ) . Highest concentrations of organic and in- organic suspended materials in the water occur within 10 km of the coastline and de- crease nearly exponentially seaward ( Man- heim et al. 1970). Mineral grains larger than 4 fim (silt-size) comprise 10-25% of near- shore suspended sediment and only 2-5% of offshore samples; the remainder is or- ganic matter. The zone of strong terrige- nous influence is restricted to nearshore waters and, specifically, to the inner shelf zone of turbid water drifting away from the estuary mouth. The coarser grains in this zone are effectively trapped in the "estua- rine" circulation (which serves to reinforce the surface concentrations) and are trans- ferred from one estuary to the next along the path of the longshore current. Studies of other areas (Postma 1967) sug- gest that volumes of suspended sediment transported on the many feedback loops in the bight are probably orders of magnitude greater than both the net volume from the Hudson River that is transported across the shelf and the much larger amounts intro- duced by dumping. Although the factors which influence sus- pended sediment dispersal can be readily defined, many large gaps in our knowledge must be closed before quantitative sediment transport budgets can be constructed on a regional scale. The most important of these are: shelf circulation patterns and mecha- nisms, particularly during storms; hydraulic properties of suspended sediments, particu- larly resuspension and settling properties; and the influence of flocculation and bio- logic aggregation on settling. Detailed studies in the New York Bi^ht apex A 1973 bathymetric map (Fig. 3) of the bight apex was made as the result of a NOAA-Corps of Engineers survey. The principal topographic features are the northern end of the Hudson Shelf Valley, Cholera Bank, and the Christiaensen Basin, an amphitheaterlike feature terminating the Hudson Shelf Valley (Veatch and Smith 1939). Dumpsites for dredge spoils (the mud dump ) , cellar dirt, sewage sludge, and acid wastes are shown. Knolls immediately northwest of Ambrose Light and north and northwest of the dredge spoil dumpsite 294 94 Geological processes 74*00' Fig. 3. 73* 55' 73* 50' 73* 45' 73* 40' Bathymetric map of the New York Bight apex. Contour interval, 1 m. Data ( in meters ) from 1973 NOAA-Corps of Engineers survey. were formed from early 20th century dump- ing of assorted building excavation material and sand and gravel from the dredging of Ambrose and Sandy Hook Channels (Wil- liams 1975). Comparison of the 1973 bathymetric sur- vey results with data from the 1936 survey reveals that only the anthropogenic areas have changed significantly. Figure 4 shows the 1973 and 1936 bathymetry of the dredge spoil site, as well as the net change between the two surveys. The 50-ft knoll on the 1936 map ( relatively unchanged in 1973 ) is itself the result of earlier dumping (Williams 1975). The amount of anthropogenic ma- terial accumulated during these years (1936-1973) has been calculated to be about 124 xlO6 m3. This compares with about 142 xlO6 m3 dumped. The difference easily can be accounted for by settling alone. Surficial sediments have been mapped by analyzing over 700 bottom grab samples collected at 1-km spacing (Fig. 5). The topographically low Hudson Shelf Valley and the Christiaensen Basin are floored 295 Surficial sediment.'} 95 SOlflS 90 95 —22' « i « '>^->80 " DREDGE SPOIL DUMPSITE 1936 BATHYMETRY j 95 isaa DREDGE SPOIL DUMPSITE 1973 BATHYMETRY -:,;;:■ 9* 100 115 vV»o ■;..;;.;;;:;: li.lll!i:B'i!;;-l:..-1v..i;ii-l|- NET CHANGE MAP 1936-1973 with fine-grained sediment, whereas the rest or the area contains assorted sizes of sand and both anthropogenic (artifact) and natural gravel deposits. Artifact gravels consist of recognizable construction rubble ■ — brick, schist, concrete, etc. Geophysical data taken during the 1973 survey consisted of 3.5-kHz shallow-pene- tration seismic reflection records and side- Fig. 4. Bathymetric maps ( 5-ft contour inter- vals ) of the dredge spoil dumpsite, New York Bight apex. The 198"T azimuth (minimum distance 4 nmi from Ambrose Light ) and the 90-ft isobath define the designated site (hatched). Upper left — 1936; upper right — 1973; left — net change from 1936-1973. scan sonar records with 150-m range on each side of 610-m-spaced tracklines. Al- though data interpretation is incomplete, bottom roughness patterns and trends of linear bedforms (sand ribbons and de- graded sand waves) have been mapped from sidescan data (Fig. 6). These bed- forms appear as alternating light and dark bands corresponding to fine- and coarse- 296 96 Geological processes 40°30'N - 40°20'N 74°00'W CONTOUR INTERVAL: 5fm = MUD Mill SILTY-FINE SANDS 73°50'W 73°40'W FINE-MED. SANDS Xvl SANDY GRAVEL COARSE SANDS jgggg ARTIFACT GRAVEL Fig. 5. Distribution of surficial sediment based on visual sample examination. Bathymetry from 1936 data. grained sediment or as isolated dark bands. Streaky, patchy, and rough textures are as- sociated with the dredge spoil and cellar dirt dumpsites and may be related to indi- vidual dumps. Preliminary analysis of seismic data shows filling of the Hudson Shelf Valley from Cholera Bank. Suspended sediment studies are particu- larly important in the bight apex because of the large amounts of fine particles dispersed in the water by waste disposal operations. These particles are in addition to the fine sediments discharged from the Hudson River, other river mouths, and tidal inlets connected to coastal wetlands. Fine-grained sediment is also eroded from the sea floor during storms. Of immediate concern is sewage sludge which contains bacterial, viral, and heavy metal contaminants that adhere to fine sediment particles in the water column. The suspended fraction of dredge spoils is also probably similarly con- taminated. All of these fines are largely re- 297 Surficial sediments 97 PATCHES OF DEGRADED SAND WAVES LARGE, IRREGULAR SAND RIBBONS - STREAKY TEXTURE PATCHY TEXTURE £& ROUGH TEXTURE Fig. 6. Distribution of bottom roughness pat- terns from sidescan sonograplis. Blank area NW and SE of Ambrose Light (A) shows no bedforms. M — Dredge spoil dumpsite; CD — cellar dirt site; SS — sewage sludge site. tained in the nearshore water column as a consequence of the bight circulation pat- tern. Suspended sediment studies were initi- ated in the bight apex during 1973 when sample stations were occupied to collect chemical and physical oceanographic data. Water samples were collected, filtered, and examined from the surface, 10-m depth, and the bottom at 25 stations. Preliminary 10 METEKSTOIAl SUSPENDED LOAD-mg/L results for data taken in fall 1973 (Drake 1974; Figs. 7-10) indicate the existence of a fair-weather, clockwise current-circula- tion gyre, driven in part by the southwest drift of offshore shelf water. This has been verified by current meter studies in the T^W sm^ Fig. 7. Total suspended sediment load in waters at 10-m depth, late November 1973. ( From Drake 1974.) Fig. 8. Distribution of ferric hydroxide particles in the water column in late November 1973 ( grains X lOVliter). A — Surface; B — midwater; C — bot- tom water. ( From Drake 1974. ) 298 98 Geological processes FERREL 2 SEP 16-20, 1973 Fig. 9. Vertical distribution of total suspended load (in mg/liter) seaward of Long Beach, Long Island. ( From D. E. Drake unpublished. ) apex (Charnell and Hansen 1974). Part of the total suspended load in the bight apex is easily identifiable, red-orange ferric hydrox- ide particles. These particles are formed by precipitation of iron in seawater as the re- sult of acid waste dumping. They consti- tute an excellent tracer of suspended sedi- ment circulation. The vertical distribution of suspended sediment shows high values ( 1.0 mg/liter) near the surface, and 2.0 mg/ liter in the near-bottom "nepheloid" layer, typical of shelf areas (Fig. 9). It is expected that this layer will transport much of the suspended particulate matter and its asso- ciated contaminants. CtNEDALIZEO FINE SEDIMtNl T8ANSPQ8T FALL 1973 — 40*20' N Fig. 10. Fine sediment transport system as in- ferred from distribution of suspended sediments during fall 1973. Dashed line is mean position of boundary between more turbid coastal water and less turbid offshore water. Clockwise gyre is ap- parently driven by southwesterly drift of offshore shelf water, and, on the bottom, by influx of saline water into New York Harbor. Regional currents which appear to be persistent are indicated by solid arrows. ( From Drake 1974. ) Preliminary results show there is a con- centration of fine-grained sediment in en- closed lows in the Hudson Valley axis, sandy mud in the remainder of the valley axis, and coarser sediment up the flanks of the valley and onto the shelf. Alternative dumping area studies Two midshelf areas have been designated as possible interim alternative dumping areas for sewage sludge and dredge spoils from the New York metropolitan area (see Fig. 1). The northern area is to be a mini- mum of 46 km from the Long Island shore- line, 18 km from the axis of the Hudson Shelf Valley, and 120 km from the entrance to New York Harbor. The southern area is seaward of the 36-m isobath and the same distances from the Hudson Valley axis and the New York Harbor entrance as the northern area (areas 2D1 and 2D2 on Fig. 1 ) . Each area is 18.5x18.5 km. Northern area — In the northern area ( Fig. 1, 2D1 ), the sampling grid was placed seaward of the center of the location-criteria triangle to investigate, in part, a shallow tributary valley of the ancestral Long Island drainage system. The surficial sediments consist of sand with some areas of over 5% gravel ( Fig. 11 ). Fine sands lie in the north- eastern part of the area, medium sands cover the western and southern parts, with a gravel deposit ( —39% gravel ) at one sta- tion associated with an area of coarser med- ium sand in the southern part of the area. Only two stations contained >5% mud. Bottom photographs indicate that the area is characterized by a smooth, slightly un- dulatory, mounded or rippled bottom. Side- scan sonar records reveal elongate dark areas which may be erosional windows in the Holocene sand sheet that expose the basal Holocene pebbly sand or may be areas of abundant large shell fragments. Grab samples were spaced too far apart to be definitive. Bottom photo and submers- ible-observation data support the existence of windrows of shell fragments. Southern area — The southern study area in Fig. 1 (2D2) is centered over the broad, flat high of the Hudson divide ( Fig. 12). To 299 Surficial sediments 99 72°50 72"45 MEDIUM SAND .".•. 100-124*1 :':■'.'■ 1.23-1.49*/ I I 130-1.74 ♦ ( •£S: 175-199 * I Mil > 200+ FINE SAND Fig. 11. Northern proposed interim alternative dumping area (2D1 on Fig. 1). Grain-size distri- bution of sand-sized fraction. Large dots — sample stations. ( Bathymetry from Stearns and Garrison 1967; 1-fm contour intervals.) ing ridge and swale topography. Geophysi- cal data, sediment samples, and two dives in submersibles showed that grain-size pat- terns appear to be related to bottom topog- raphy; coarser sand and gravel deposits lie on the crest and east flank of the Hudson divide, while medium- and fine-grained sand occur in the ridge and swale topog- raphy (Fig. 13). These distributions sug- gest that fine sediment is winnowed from the crest and east flank of the divide and deposited to the west. Observations from a submersible in Veatch and Smith Trough reveal a veneer of shelly, pebbly sand with large, angular clay pebbles and occasional oyster shells derived from the underlying early Holocene lagoonal clay. Seismic data also reveal that the reflector associated with this surface, outcrops on the ridge flank. It appears that storm-generated currents from the northeast have winnowed the east flank of the Hudson divide and formed or main- tained the ridge and swale topography on the west side of the divide. the northeast the bottom grades gently into the Hudson Shelf Valley, while the western section is characterized by northeast-trend- Fig. 12. Southern proposed interim alternative dumping area (2D2 on Fig. 1). (Bathymetry from Stearns and Garrison 1967, 1-fm contour intervals. ) Solid lines — geophysical tracklines; bars — sites of dives by submersibles. C 73°30' 73*25' 73*20' 73°15' ;.?S <100 + COARSE SAND ;.;•;■ 1.00-1.24+ ••':;■'.• 125-1.49 + I I 150-174 + iSS 175-199 + Mill >200+ FINE SAND MEDIUM SAND Fig. 13. Southern proposed interim alternative dumping area, (2D2 on Fig. 1). Grain-size distri- bution of sand-sized fraction. Large dots — sample stations. (Only the 20-fm isobath is shown.) 300 100 Geological processes Suspended sediment — As previously men- tioned, most fluvial suspended sediment is effectively trapped in estuaries and coastal wetlands. Consequently, the terrigenous fraction of the suspended matter decreases rapidly seaward. Suspended solids through- out the water column in the alternative dumping areas were predominately plank- ton and their noneomhustible remains. Total suspended matter concentration in surface water is from 100-500 fig/ liter, comprised of 5% or less terrigenous matter, 80% com- bustible matter, and 15% siliceous and cal- careous noncombustible planktonic remains (D. E. Drake personal communication). Subsurface water-suspended matter con- centration is similar or somewhat less, ex- cept in the nepheloid layer 5-10 m above bottom. There, suspended matter concen- trations are 500-2,000 fig/ liter, consisting of 30-60% combustible matter and 50-80% noncombustible matter which includes 10- 20% terrigenous matter. Textural proper- ties of sediment deposits in the alternative dumping areas show that very little sedi- ment finer than 62 microns is present. References Charnell, R. L., and D. V. Hansen. 1974. Summary and analysis of physical oceanogra- phy data collected in the New York Bight apex during 1969-1970. NOAA-MESA Rep. 74-3. 74 p. Dhake, D. E. 1974. Suspended particulate mat- ter in the New York Bight apex: September- November 1973. NOAA Tech. Rep. ERL 318-MESA 1. Duane, D. B. 1969. Sand inventory program. A study of New Jersey and northern New En- gland coastal waters. Shore Beach October. , M. E. Field, E. P. Meisburber, 1). J. Swift, and S. J. Williams. 1972. Linear shoals on the Atlantic inner continental shelf, Florida to Long Island, p. 447-498. In D. J. Swift et al. [eds.], Shelf sediment transport: Process and pattern. Dowden, Hutchinson & Ross. Fischer, A. G. 1961. Stratigraphic record of transgressing seas in light of sedimentation on Atlantic coast of New Jersey. Bull. Am. Assoc. Pet. Geol. 45: 1656-1666. Folger, D. W. 1972a. Texture and organic carbon content of bottom sediment in some estuaries of the United States, p. 391-408. In B. W. Nelson [ed.], Environmental framework of coastal plain estuaries. Geol. Soc. Am. Mem. 133. . 1972i>. Characteristics of estuarine sedi- ments of the United States. U.S. Geol. Surv. Prof. Pap. 742. 94 p. Freeland, G. L., and D. J. Swift. In press. Sur- ficial sediments. NOAA-MESA New York Bight Atlas Monogr. 10. Gross, M. G. 1972. Geologic aspects of waste solids and marine waste deposits, New York metropolitan region. Geol. Soc. Am. Bull. 83: 3163-3176. Hathaway, J. C. 1972. Regional clay mineral facies in estuaries and continental margin of the U.S. East Coast, p. 293-316. In B. W. Nelson [ed.], Environmental framework of coastal plain estuaries. Geol. Soc. Am. Mem. 133. Knott, S. T., and H. Hoskins. 1968. Evidence of Pleistocene events in the structure of the continental shelf of the northeastern U.S. Mar. Geol. 6: 5-43. McKinney, T. F., and G. M. Friedman. 1970. Continental shelf sediments of Long Island, N.Y. J. Sediment. Petrol. 40: 213-218. , W. L. Stubblefield, and D. J. Swift. 1974. Large-scale current lineations on the central New Jersey shelf: Investigations by side-scan sonar. Mar. Geol. 17: 79-102. Manheim, F. T., R. II. Meade, and G. C. Bond. 1970. Suspended matter in surface waters of the Atlantic continental margin from Cape Cod to the Florida Keys. Science 167: 371- 376. Meade, R. H. 1969. Landward transport of bottom sediments in estuaries of the Atlantic Coastal plain. J. Sediment. Petrol. 39: 222- 234. . 1972a. Transport and deposition of sedi- ments in estuaries, p. 91-120. In B. W. Nel- son [ed.], Environmental framework of coastal plain estuaries. Geol Soc. Am. Mem. 133. 1972i>. Sources and sinks of suspended matter on continental shelves, p. 249-262. In D. J. Swift et al. [eds.], Shelf sediment trans- port: Process and pattern. Dowden, Hutchin- son & Ross. Milliman, J. D. 1972. Marine geology, p. 10-1 to 10-91. In Coastal and offshore environ- mental inventory, Cape Hatteras to Nantucket Shoals. Mar. Publ. Ser. 6. Univ. R.I. , and K. O. Emery. 1968. Sea levels during the past 35,000 years. Science 162: 1121-1123. Mueller, J A., A. R. Anderson, and J. S. Jeris. 1976. Contaminants entering the New York Bight: Sources, mass loads, significance. Am. Soc. Limnol. Oceanogr. Spec. Symp. 2: 162- 170.- Postma, H. 1967. Sediment transport and sedi- mentation in the estuarine environment, p. 158-179. In G. H. Lauff [ed.], Estuaries. Publ. Am. Assoc. Adv. Sci. 83. Schlee, J. 1973. Atlantic continental shelf and slope of the U.S. Sediment texture of the 301 Surficial sediments 101 northeast part. U.S. Geol. Surv. Prof. Pap. 529-L. 64 p. 1975. Sand and gravel. MESA New York Bight Atlas Monogr. 21. 26 p. Sheridan, R. E., C. E. Dill, Jr., and J. C. Kraft. 1974. Holocene sedimentary environment of the Atlantic inner shelf off Delaware. Geol. Soc. Am. Bull. 85: 1319-1328. Stahl, L., J. Koczan, AND D. J. Swift. 1974. Anatomy of a shoreface-connected sand ridge on the New Jersey shelf: Implications for the genesis of the surficial sand sheet. Geologv 2: 117-120. Stearns, F., and L. E. Garrison. 1967. Bathy- metric maps, middle Atlantic U.S. continental shelf, 1:125,000. NOAA, Natl. Ocean Surv. Stubblefield, W. L., M. Dicken, and D. J. Swift. 1974. Reconnaissance of bottom sediment on the inner and central New Jersey shelf. NOAA-MESA Rep. 1. , J. W. Lavelle, T. F. McKinney, and D. J. Swift. 1975. Sediment response to the present hydraulic regime on the central New Jersey shelf. J. Sediment. Petrol. 45: 337- 358. Swift, D. J., G. L. Freeland, P. E. Gadd, G. I Ian, J. W. Lavelle, and W. L. Stubblefield. 1976. Morphologic evolution and coastal sand transport, New York-New Jersey shelf. Am. Soc. Limnol. Oceanogr. Spec. Symp. 2: 69-89. , J. W. Kofoed, F. P. Saulsbury, and P. Sears. 1972. Holocene evolution of the shelf surface, central and southern Atlantic shelf of North America, p. 499-574. In D. J. Swift et al. [eds.], Shelf sediment transport: Process and pattern. Dowden, Hutchinson & Ross. Uchupi, E. 1963. Sediments on the continental margin off eastern U.S. U.S. Geol. Surv. Prof. Pap. 475-C, p. C132-C137. . 1970. Atlantic continental shelf and slope of the U.S. — shallow structure. U.S. Geol. Surv. Prof. Pap. 529-1. Veatch, A. C., and P. A. Smith. 1939. Atlantic submarine valleys of the United States and the Congo Submarine Valley. Geol. Soc. Am. Spec. Pap. 7. Williams, S. J. 1975. Anthropogenic filling of the Hudson River shelf channel. Geology 10: 597-600. . 1976. Geomorphology, shallow subbot- tom structure, and sediments of the Atlantic intercontinental shelf off Long Island, New York. U.S. Army Corps Eng. Coastal Eng. Res. Center Tech. Pap. 76-2. 123 p. , and D. B. Duane. 1974. Geomorphology and sediments of the inner New York Bight continental shelf. Tech. Memo. 45. U.S. Army Corps Eng. Coastal Eng. Res. Center. 81 p. 302 Reprinted from: Geophysical Research Letters, Vol. 3, No. 2, 97-100, 31 Vol. 3, No. 2 Geophysical Research Letters February 1976 PRELIMINARY RESULTS OF COINCIDENT CURRENT METER AND SEDIMENT TRANSPORT OBSERVATIONS FOR WINTERTIME CONDITIONS ON THE LONC ISLAND INNER SHELF J.W. Lavelle, P.E. Gadd , C.C. Han, D.A. Mayer, W.L. Stubblef ield , and D.J. P. Swift NOAA/Atlantic Oceanographic and Meteorological Laboratories, 15 Rickenbacker Causeway, Miami, Florida 33149 R.L. Charnell NOAA/Pacific Marine Environmental Laboratory, 3711 15th Avenue N.E., Seattle, Washington 98105 H.R. Brashear, F.N. Case, K.W. Haff, and C.W. Kunselman Oak Ridge National Laboratory, P.O. Box X, Oak Ridge, Tennessee 37830 Abstract . We have observed late fall and winter bedload sediment transport and the overlying cur- rent field in ridge and swale topography on the inner continental shelf south of Long Island, and can report movement of bed material at a water depth of 20 m to a distance of approximately 1500 m after several storm events. Movement over an 11-week observation period was longshore and oblique to the ridge crest at the experimental site. Currents were also predominately longshore, but long term averages demonstrate that a vertical shear existed in the fluid motion. Although the number of sediment transport "events" suggested by the current meter data is nearly balanced in east- ward and westward directions, both estimates of transport from current speeds and sand tracer dis- persion patterns show that several westward flow- ing events dominated the transport during a two and one-half month period. A quantitative upper bound of 31 cm/sec on the threshold velocity for sediment movement in this size range is also set by the data. Introduction Increasingly widespread interest in the charac- terization and quantification of shelf sediment transport stems from the requirements of the growing number of shelf and nearshore users to understand the dynamics of an area on which they may have potential impact. The uses and interests are myriad, but more common expressions of concern are phrased in terms of recovery rates of contami- nated sediments by replacement, the stability of the substrate for offshore structures, the in- fluence of offshore work on beach and nearshore features, and the temporal variability of sediment transport as an influence on faunal habitats. While considerable efforts have been made in ob- serving and describing water-sediment coupling in the laboratory, under riverine flow, and in the nearshore area, few direct measurements of off- shore sediment movement and the associated near- bottom water velocity field have been made. For these reasons, we are reporting preliminary re- sults of an experiment recently completed in the New York Bight to directlv measure offshore co- hesionless sediment movement and its immediate forcing mechanism, the overlying water velocity field. The Long Island Near-Shore (LINS) Study (Figure 1) was centered at 40°33'N and 73°25'W, halfway between Jones and Fire Island Inlets, Long Island, New York, some 9 km offshore. The study area, an 8 x 10 km rectangle, was located in an area of undulating morphological features described in Duane et ai. (1972) as ridge and swale topo- graphy. Bedforms at the study site have wave- lengths of approximately 1 km with wave heights of 4-7 m, intersect the shoreline obliquely, and are composed of relatively clean, medium to fine sands; the ridges are asymmetrical with steeper southwest facing flanks. The experimental design was twofold: to gather sediment dispersion and current meter data which could be used to aid in quantifying sediment transport; and to gather qualitative data on the construction and/or the maintenance mechanism of ridge and swale features which are widespread on the Atlantic continental VERTICAL CUK«[NT Will STATIONS HIST DBOP Copyright 1976 by the American Geophysical Union. Fig. 1. Bathymetry and current meter station locations for the Long Island Nearshore Study (LINS). Stations CI, C6, C7, and C8 returned no usable data. Q7 303 98 Lavelle et al.: Results of Meter and Sediment Observations shelf (Swift et al. , 1973). Field work was divi- ded into two concurrent operations: a sediment tracer experiment and a current meter array of high spatial resolution. We present here a quali- tative, preliminary view of the data collected in those efforts. Current Meter Observations During the first six weeks of the current meter operation (October 16 to December 4, 1974), nine- teen stations (Figure 1) were occupied. A single current meter string was retained in the area during the remainder of the experiment. Aandaraa RCM-4 Savonius rotor current meters which record instantaneous direction and integrated average speed at 10-minute intervals were used throughout. Measurement emphasis was placed on a well-defined ridge and trough; meters were located on a crest, flank, and trough on each of three transects (B, C, and D of Figure 1) as well as the adjacent flank of transect C. Additional meters were set outside the central study area to measure far- field velocities. Flow during the observation period trended both east and west, parallel to the coast. Figure 2 is a vector time series of velocities at station 2C (1.5 m above the bottom) and is representative of near-bottom water movement during one of the most active periods of flow. The data presented here have been subjected to a 40-hr low pass filter and then resampled at hourly intervals. Although east is the dominant flow direction during this sampling interval, the most intense flow was westward during a three day period near the end of this period. Predominance of eastward flow is consistent with 15 - 16- ,*os*- 17 T ia ( 19 m 20 f 21 L 22 Wr 23 i 24 V 25 r 26 1 27 k 28 f 29 j 30 !.! DEC 74 " 2 <^k\ 3 ™d» LONG TERM BOTTOM, MIODEPTH, AND NEAR -SURFACE CURRENT MEANS - EASTWARD FLOW (c) LONG TERM BOTTOM, MIODEPTH, AND NEAR-SURFACE CURRENT MEANS- WESTWARD FLOW CURRENT SPEED AND DIRECTION Fig. 2. Near bottom current vector time series and velocity averages. Shoreline direction represents the trend of the 5 fathom isobath between Jones and Fire Island Inlets. the observation of Charnell and Mayer (1975) who reported the existence, in the statistical sense, of a clockwise gyre in the long term mean flow within the New York Bight apex during the fall and winter of 1973. The strong westward flow (Figure 2A) occurred during the storm of December 1 - December 4, 1974, an event which was reported to have been the most damaging northeaster since the Ash Wednesday storm of 1962 (C. Galvin, Coastal Engineering Research Center, personal communica- tion). Winds from the east-northeast up to 16 m/s were recorded at John F. Kennedy Inter- national Airport during the initial 36 hrs of this period; winds from the northwest at an aver- age speed of 10 m/s followed on December 3 and 4. The second most important sustained flow during the observation period, that which began during December 16, also followed east winds. Periods of high speed winds from the west and northwest cause less intense near-bottom water movement. The asymmetry of the fluid response to easterly and westerly winds in this area has been noted by Beardsley and Butman (1974). Vertical shear in current velocities was unmasked in the data (Figures 23 and 2C) when long term velocity averages were made on data from meters grouped by position in the water column. Flow recorded by meters 1.5 to 4 m from the bottom (B) , 5 and 6 m from the bottom (M) , and 6 to 11 m from the surface (S), were averaged separately in time over periods when flow had eastward and westward components. Water depths at the stations varied from 15 to 22 m. These data show that near sur- ; face water flow had an offshore component for both eastward and westward flow, while bottom flow tended to be more inshore, parallel to the iso- baths during westward flows and more strongly in- shore during eastward flows. Speeds decreased in a relatively uniform fashion from the upper to the bottom meters. Wind records document a northerly wind component throughout much of the observation period; the observed shore-normal components may be an indication of upwelling contributions to fluid motion. Sand Tracer Measurements In order to directly assess the flow response of the sediment to the observed water movement, we employed the Radioisotope Sand Tracer (RIST) system developed at Oak Ridge National Laboratory (Duane, 1970; Case et al. , 1971). Indigenous sand was sorted to produce a fraction whose size dis- tribution was roughly Gaussian, with a mean dia- meter of .15 mm (fine to very fine sand), a standard deviation of .03 mm, and no material larger than .25 mm or smaller than .06 mm. Approx- imately 500 cm3 of this material was surface coat- ed with 10 Curies of the isotope 103Ru (T*j = 39.6 days). On November 12, equal portions of the tagged sand in water soluble bags were released at three points at the east end of the main trough (Figures 1 and 3). The injection points formed an equilateral triangle with sides roughly 100 m in length. The ensuing dispersal pattern of labeled sand was surveyed at intervals by scintillation detectors mounted in a cylindrical vehicle which was towed across the bottom. Raydist precision navigation with 10 m resolution was used. Four post-injection surveys were made during the 11- week tracer experiment. 304 Lavelle et al. Results of Meter and Sediment Observations 99 400 A £ UJ200 z 25 NOV 74 $0 -L 1000 800 600 400 200 200 400 METERS B 400- 10 JAN 75 : j M V) ■\\\\\\vM'- "• '". '■J&K K • ' ; *,' . l • bi f- LJ200- 2 n. 1000 800 600 400 200 200 400 METERS S 100 £ 80 2 60 z 40 £ 20 , JM I Threshold Velocity = OV'74 20 28 6DEC'74 14 22 30 7JAN'75 r80 o ■60 w to ■40 v. 20 o * I2N0V'74 20 6DEC'74 i — i — i — p* 7JAN'75 Fig. 3. a and b) Dispersion patterns measured 13 and 59 days after injection of tagged sand. Point sources are represented by dots. Broken line is the survey trackline; stippling represents radiation intensity. c) Near bottom current speed record over the length of the experiment, and calculated sediment transport infor- mation (see text) . Dispersion patterns mapped two and eight weeks after injection are shown in Figure 3; each has been corrected for background radiation and decay. After two weeks (November 25) roughly ellipsoidal smears trended east from each of the three in- jection points (Figure 3a). Each smear could be traced for about 200 m before the signal was lost in the background radiation. After eight weeks (January 10) , the three eastward smears had been replaced by a single, more extensive pattern extending 700 m to the west (Figure 3b) . The reversal of the patterns from east to west was more markedly demonstrated by preliminary data from a survey made during mid-December (December 17-19) . Although those data have not yet satis- factorily been processed, the data at the time of that survey were sufficient to indicate that the reversal of the dispersion pattern of Figure 3a had already occurred, and in fact extended approx- imately 1,500 m to the west. We should point out that the patterns of Figure 3 must be regarded as minimum transport patterns, in the sense that tagged material which has been buried or has dif- fused downward into the bed is attenuated in sig- nal strength by the overburden. For this reason, the signal measured is an underestimate of the true signal and the observations must be regarded as a lower bound to the true transport. The temporal pattern of sediment transport over a 60-day period may be inferred from Figure 3c. The basic record is current speed, measured 1.5 m from the bed, versus time. The horizontal line at 18 cm/sec is an estimated threshold, based on the work of Shields and subsequent workers for shear velocity (Graf , 1971) and a choice of 3.0 x 10-3 for the drag coefficient (Sternberg, 1972). We believe that this choice of threshold velocity is in part verified by empirical evidence obtained during the course of the experiment (see below). We have made estimates of the relative role each transport event played in the overall transport record, based on the concept of the proportionality of frictional energy expenditure to the transport volume (Bagnold , 1963) . For each event where velocities exceeding threshold were recorded, we have calculated a transport volume: ^H (|v| - |vTH|) dt (I) v TH1 where |V| is measured current speed, threshold speed, a is a constant of proportional- ity, and T. is the duration of the transport event. Expression of sediment transport as a power of the difference of measured and threshold velocity is supported by analysis of stream transport data (Kennedy , 1969) . Without assigning a value to a, one may calculate relative rates of transport, one event to the next, or one event to the total trans- port evidenced by the current meter record. We have taken the second of these options, and have represented relative sand transport by solid bars superimposed on the current record (Figure 3c). Despite the exceedence of the sediment transport threshold at many points in the record, only the solid bars centered on December 2 and December 16- 305 100 Lavelle et al. Results of Meter and Sediment Observations 17 arc visible in the figure, bearing witness to the dominance of the calculated transport by these two events. Furthermore, the figure also shows that most of the calculated transport occurred during the early December storm. While this cal- culated transport index may be biased by the choice of threshold speed as well as the functional de- pendence on velocity, we believe any other reason- able parameterization is likely to lead to the same general conclusion: the storm event of December 1 - December U moved more sand at 20 m water depth than the combination of all other transport events. The reversing nature of sediment flow during the observation period provides a constraint on the entrainment velocity. A threshold speed greater than approximately 31 cm/sec at 150 cm off the bottom would eliminate transport during the first 14 days of the record, in contradiction to the observation of eastward transport (Figure 3a). Setting the threshold much below 12 cm/sec would result in more eastward transport during the entire tracer experiment than was the case. Based on the relative extent of the dispersion patterns in Figures 3a and 3b, we believe that the calcu- lated threshold velocity of 18 cm/sec is realistic. Summary Water movement on the Long Island Inner Shelf at depths of 10 to 20 m and at frequencies below 1/40 hr was predominately alongshore with a net flow over the observation period to the east. The non-tidal flow reversals at these depths suggest domination by winds associated with frontal pas- sages; the net eastward flow likely reflects the average winds from the north and west through the fall and winter months. Vertical shear of the flow is observable in long term averages of the current records; small offshore mid-depth flows and some onshore bottom flow may reflect as an upwelling circulation the net offshore component of the wind. The most intense water movements recorded during the experimental period followed high northeasterly and easterly winds. Sediment is transported both eastward and west- ward parallel to the shoreline, and oblique to the ridge and trough system. Current speeds recorded 150 cm from the bed show that the sediment en- trainment threshold is exceeded only intermit- tently; sand transport occurs only during storm events, separated by periods of quiescence. Mean water movement was to the east over the observa- tion period in sharp contrast to the observed mean westward sediment transport. Some eastward sediment transport was observed, but the most intense water movement and resultant sand move- ment were associated with several "northeaster" storm events. Asymmetry of the ridges (steeper southwest facing flanks) suggests that westward flows associated with such storms constitute the primary sediment flow mechanism in this ridge and swale topography. Acknowledgements . Support for this work has come from NOAA's New York Bight Marine Ecosystems Analysis (MESA) Project, NOAA's Environmental Research Labora- tories, and ERDA's Division of Biomedical and Environmental Research. Oak Ridge National Laboratory is operated by Union Carbide Corpora- tion for the U.S. Energy Research and Development Administration. References Bagnold, R.A., Beach and near-shore processes, part I, mechanics of marine sedimentation, In: The Sea, vol. 3, pp. 507-528, Interscience Pub. , New York, 1963. Beardsley, R. , and B. Butman, Conditions on the New England continental shelf: response to strong winter storms, Geophys. Res. Letters, 1, 181-184, 1974. Case, F.N., E.H. Acree, and H.R. Brashear, Detection system for tracing radionuclide- labeled sediment in the marine environment, Isotopes and Radiation Technology, 8, 412-414, 1971. Charnell, R.L., and D.A. Mayer, Water movement within the apex of the New York Bight during summer and fall of 1973, Tech ■ Memo. , National Oceanic and Atmospheric Administration, Boulder, Co. (in press). Duane, D.B., Tracing sand movement in the littoral zone: progress in the Radio Isotopic Sand Tracers (RIST) study, July 1968-February 1969, Coastal Eng. Res. Center Misc. Paper, Washing- ton, D.C. , 1970. Duane, D.B., M.E. Field, E.P. Meisburger, D.J. P. Swift, and S.J. Williams, Linear shoal on the Atlantic inner continental shelf, Florida to Long Island, I_n: Shelf Sediment Transport: Process and Pattern, pp. 447-498, Dowden, Hutchinson and Ross, Stroudsburg, Pa., 1972. Graf, W.H., Hydraulics of Sediment Transport, p. 96, McGraw Hill, New York, 1971. Kennedy, J.F., The formation of sediment ripples, dunes, and antidunes, _In: Annual Review of Fluid Mechanics, vol. 1, pp. 147-168, Annual Reviews, Inc., Palo Alto, Calif., 1969. Sternberg, R.W., Predicting initial motion and bedload transport of sediment particles in the shallow marine environment, In: Shelf Sediment Transport: Process and Pattern, pp. 61-82, Dowden, Hutchinson and Ross, Stroudsburg, Pa. , 1972. Swift, D.J. P., D.B. Duane, and T.F. McKinney, Ridge and swale topography of the Middle Atlantic Bight, North America: secular re- sponse to the Holocene hydraulic regime, Mar. Geol., 15, 227-247, 1973. (Received October 14, 1975; accepted November 21, 1975.) 306 32 Reprinted from: Earth and Planetary Science Letters , Vol. 32, No. 1, 18-24. 18 Earth and Planetary Science Letters, 32 ( 1976) 1 8-24 © Elsevier Scientific Publishing Company, Amsterdam - Printed in The Netherlands [5| ON THE INTERPRETATION OF NEAR-BOTTOM WATER TEMPERATURE ANOMALIES R.P. LOWELL School of Geophysical Sciences, Georgia Institute of Technology, A tlanta, Ga. 30332 (USA) and P.A. RONA National Oceanic and A tmospheric Administration, A tlantic Oceanographic and Meteorological Laboratories, Miami, Fla. 33149 (USA) Received November 16, 1975 Final revised version received June 21, 1976 A positive water temperature anomaly of 0.1 LC and an inverse gradient of potential ternperature ot 1 .5 X jq-2 aQjm nus bCL,n mcasu^d llt the TAG hydrothermal field in the rift valley of the Mid-Atlantic Ridge at latitude 26°N by means of a thermistor array towed between 2 and 20 m above the seafloor. This anomaly appears to be as- sociated with hydrothermal discharge from the oceanic crust. Thectemperature data are interpreted in terms of (1) a steady, turbulent thermal plume rising in a homogeneous, neutrally buoyant medium, and (2) turbulent diffusion in the ocean-bottom boundary layer. The calculations indicate that the thermal output of the TAG anomaly area is of the order of several megawatts, which is of the same order of magnitude as some continental geothermal systems. The thermal output from the TAG anomaly area represents a significant fraction of the total heat loss resulting from the generation of new lithosphere at the Mid-Atlantic Ridge at 26°N. 1. Introduction Conductive heat flow measurements in the erestal zone of various sections of the ocean ridge system exhibit a high degree of scatter [1 ,2], in a manner opposite to that expected from heat flow refraction effects [3]. Moreover, the mean conductive heat flow in the erestal zone is frequently less than the conduc- tive heat flow from a uniformly spreading lithospheric plate generated at the ridge axis [4,5]. This discrepancy between the observed heat flow and conductive heat flow models based on a uniformly spreading litho- sphere is usually attributed to convective heat losses due to hydrothermal circulation in the newly created oceanic crustal rocks. Additional evidence for hydro- thermal circulation comes from the occurrence of hydrothermally altered rocks [6] and deposits of hydrothermal origin [7,8]. Lastly, thermistor probes towed over segments of the axial zone within 10-20 m of the seafloor have measured temperature anom- alies which appear to be associated with hydrothermal discharge [2,9,10]. Theoretical models for hydrothermal circulation in the oceanic crust have been based both on models of convection in porous rock [3,1 1 ], as well as on models of convection in fractured rock [12,13]. Ocean ridge hydrothermal systems are exceptionally complicated and the theoretical modelling is still in its initial stages of development. Near-bottom water temperature data, however, may provide useful, quan- titative information with regard to the thermal regime in the oceanic crust. Such information may, for ex- ample, give estimates of the heat flux through the ocean floor and place some constraints on acceptable hydrothermal convection models. This is of particular importance in regions of young crust, where the sedi- 307 ment layer is too thin for standard conductive heat flow measurements to be made. The purpose of this paper is to examine the im- plications of near-bottom temperature anomalies mea- sured with towed thermistors. Since the crestal zone of the Mid-Atlantic Ridge at 26°N has been studied in some detail [8-10,14-16], the data from this area will be used in discussing the theoretical results. 2. The TAG hydrothermal field During the 1972 and succeeding cruises of the NOAA Trans-Atlantic Geotraverse (TAG) project, anomalously thick manganese oxide crusts were re- peatedly dredged from the southeast wall of the rift valley at 26°N (Fig. 1). Radiogenic dating of the manganese crusts, which attain thickness of 42 mm only 5 km from the axis of the rift valley, show them to be accumulating at about 200 mm per 106 years, about two orders of magnitude faster than hydrog- enous ferromanganese [8]. The crusts are almost pure manganese (40%), with only trace quantities of Fe, Cu, Ni, and Co [8]. The rapid accumulation rate and pure composition evidence a hydrothermal origin for crusts. Fig. 1. Bathymetric map contoured in hundreds of meters [161 showing locations of profiles A and B, along which water tem- perature measurements and bottom photographs were concur- rently made at the southeast wall of the rift valley of the Mid- Atlantic Ridge at 26° N. A water temperature anomaly (AT) was measured between 2950 and 3000 m along profile A [10]. No water temperature anomaly was present along pro- file B. The floor of the rift valley is shaded. The TAG hydro- thermal field is outlined (dashed lines). TABLE 1 Temperature profiles at the TAG hydrothermal field (Fig. 2) Tempera- Cumulative Depth (m) Potential tempe rature (°C). Thermistor position Vertical ture profile distance along in vertical array (m above lowermost thermistor) gradient (n Mfn^Ar^ no Pin hottn m rc/m) u v. t. a 1 1 uuuuni (m) 4 3 0 i 0- 115 3080-3068 2.461 2.449 2.434 +0.007 2 115- 230 3068-3055 2.460 2.454 2.433 +0.007 3 230- 345 3055-3043 2.464 2.453 2.434 +0.008 4 345- 460 3043-3030 2.464 2.455 2.440 +0.006 5 460- 575 3030-3015 2.491 2.482 2.462 +0.007 6 575- 690 3015-2997 2.484 2.470 2.464 +0.005 7 690- 805 2997-2990 2.529 2.518 2.561 -0.014 8 805- 920 2990-2975 2.542 2.510 2.559 -0.016 9 920-1035 2975-2965 2.603 2.599 2.574 +0.007 10 1035-1150 2965-2950 2.617 2.610 2.598 +0.005 11 1150-1265 2950-2935 2.615 2.599 2.469 +0.037 12 1265-1380 2935-2915 2.482 2.481 2.470 +0.003 13 1380-1495 2915-2910 2.508 2.499 2.476 +0.008 14 1495-1610 2910-2905 2.514 2.501 2.482 +0.008 Based on Rona et al. [ 10]. 308 20 CO > -*— o -Q o < E fl) 1- o c o o ^E4 2 3 E 03 Q 0 3 4 PROFILE NUMBER 6 7 8 9 10 246 246 246 246 249 248 2 56 2 56 2 60 262 " , " r-J r 243 243 243 244 246 11 12 13 14 246 252 2 51 2 57 2 60 2 47 POTENTIAL TEMPERATURE (°C) 2 61 248 2 51 247 247 248 251 Pig. 2. A plot of the potential temperature vs. height above the lowest thermistor based on the data in Table 1 from Rona et al. [10] The hydrothermal manganese oxide occurs both as crusts on basalt talus and as veins filling fractures in the talus along the inner margins of steps on the southeast wall of the rift valley. Bottom photographs [15] and narrow-beam bathymetry [16] reveal that the steps range from tens to hundreds of meters in width, are tens of meters in height, and kilometers to tens of kilometers in length. The steps are interpreted as fault scarps. The manganese oxide is hypothesized to have been deposited by a sub-seafloor hydrothermal convection system involving the circulation of seawater through basalt [16], driven by intrusive heat sources beneath the rift valley. The discharge is thought to be focused by fractures in the rift valley wall which are overlain by a porous and permeable body of talus that may act to diffuse the fracture focused flow [16]. An abrupt temperature anomaly was measured in the water column over one of the steps on the southeast wall of the rift valley within the area of hydrothermal deposits, suggesting persistence of hydrothermal ac- tivity [9,10]. The temperature anomaly of +0.1 1°C associated with an inverse gradient .of 1.5 X 10"2 °C/m was measured within 20 m of the bottom, along a hor- izontal distance of 250 m, between water depths of 3000 and 2950 m using three thermistors mounted in a 4 m long towed vertical array (see Table 1 and Fig. 2). A second temperature profile made 5 km away on the southeast wall of the rift valley, showed no tempera- ture anomaly [10] (see Fig. 1). The evidence for past and present activity led to designation of this area, the TAG hydrothermal field [14]. 3. Thermal models In order to simplify the interpretation of the water temperature data, we will assume that the tempera- ture anomaly and superadiabatic gradient results from steady-state heat transfer to the seafioor. This is a reasonable assumption because the thickness of the hydrothermal deposits suggest a time scale for the dis- charge of 2.5 X 105 years [8], which should be long enough for a steady state to be achieved [12]. More- over, the existing data is insufficient to develop a meaningful time dependent model. The water temperature data will be interpreted in two ways. First we will assume that the temperature anomaly is due to a steady, turbulent thermal plume discharging at the seafioor. The plume model will give an estimate of the heat transfer due to the as- sumed hydrothermal discharge. Secondly, we will estimate the heat transfer to the seafioor on the basis of a simple turbulent diffusion model. The heat trans- fer estimates based on these models provide useful in- formation with regard to the hydrothermal circula- tion in the oceanic crust. 3. 1. Thermal plumes We will first assume that the positive water temper- ature anomaly of 0.1 1°C measured over the south- east wall of the rift valley at the TAG area is due to hydrothermal discharge. Following Williams et al. [2], we will assume that this anomaly is due to a steady, turbulent, thermal plume rising by free convection in 309 21 a neutrally buoyant medium. We will assume that near- bottom currents are negligible. Expressions for the heat transfer in such a plume have been derived by Batchelor [18] based on the experimental data of Rouse et al. [19]. They may be written as Q = (ps/ag)F where: „ i=sl"ag(AT) ,_, 2/2 )3/ F = (— — exp(71r \z ), 11 for a three-dimensional plume and: F = fe^exp(41*W3/2 (1) (2) for a two-dimensional plume. In the above equations a, p, s andg represent the thermal expansion coeffi- cient, the fluid density, specific heat and accelaration of gravity respectively. AT represents the magnitude of the temperature anomaly at height z above the bottom, and at a distance from the axis of the plume given by r and x for the three-dimensional and two- dimensional plume respectively. Since the temperature anomaly was measured with the lowest thermistor ranging from 2 to 20 m above the bottom, we will for convenience choose z- 10 m. We will also let p = 1 03 kg/m3 , s = 4.2 X 1 03 J/kg °C, a= 1.6 X 10-4/°C,£= 10 m/s2. Lastly, we will as- sume that the temperature measurement was made at the axis of the plume. The results are: Q=53X 104 watts for a three-dimensional plume, and: <2 = 4.6X 104 watts/m for a iwo-dimensional plume. These results are approx- imately an order of magnitude greater than those given by Williams et al. [2] for a similar temperature anomaly at a similar height measured on the Galapa- gos spreading center. The reason for the difference is that Williams et al. [2] used eqs. 1 and 2 directly to obtain their estimate for the heat transfer. These equations must be multiplied by ps/ag in order to be made dimensionally correct (see Batchelor [18]). We must admit, here, that there is a significant dif- ference between the rather narrow temperature anom- alies observed over the Galapagos spreading center [2, fig. 8] and the anomaly observed over the TAG area. The anomaly feature over the TAG area is quite broad (250 m at a height of approximately 10 m from the bottom). At a height of 10 rn, however, the plume given by Batchelor [18] would be less than 10 m wide. On the plume model, it may be possible to partially explain the width of the temperature anomaly be means of (1) advection by currents, (2) plume dis- charge into a stably stratified surrounding medium, or (3) discharge from several vents along the profile and mixing of the plumes rising from each. Any or all of these mechanisms may be operating in the TAG area. Bottom photographs were made concurrently with the thermistor profiles [15]. When the camera com- pass suspended 5 m below the camera hit pockets of sediment, sediment plumes rose vertically, indicating that currents were negligible at the time of the tem- perature measurements. The presence of ripple marks in the sediment indicated the existence of intermittent currents. Sizeable near-bottom currents (<25 cm/s) have been measured elsewhere in the median valley of the Mid- Atlantic Ridge [20] Plumes discharging into a stable environment and forced plumes have not been as well investigated as the free plume model which has been used here. However, Morton et al. [21 ] have shown that plumes discharging in a stable environment reach a finite height and tend to spread laterally near the maximum height. An STD profile near the TAG field shows that the water column there is stable [10] and this may also be the case in the region of the thermistor profile. Since numerous faults were en- countered along the thermistor profile, the tempera- ture anomaly may in part be due to discharge from several vents with mixing of the individual plumes. These are all ad hoc hypotheses, however, and there is no reliable data available to either substan- tiate or disprove them. Moieover, the width of the anomaly is of the same order of magnitude as the distance traversed by typical semi-diurnal tidal cur- rents. Therefore we will examine an alternative mod- el by which to interpret the water temperature anom- aly. This model is based on the theory of turbulent diffusion. 3.2. Turbulent diffusion Wimbush and Munk [22] have recently reviewed the structure of the ocean-bottom boundary layer. Only the essential points will be stated here. 310 22 (1) In nearly all cases the boundary layer is turbu- lent. (2) There is a "constant stress layer" near the boundary such that a "friction" velocity may be de- fined by: U* = (T0/p)112 where t0 is the stress on the boundary. (3) Within the constant stress layer there often exists a viscous sublayer in which heat is transferred by conduction. This layer has a thickness of the order of a few centimeters. (4) Above the viscous sublayer, if it exists, there is a "logarithmic layer" in which the mean velocity and temperature increase as the natural logarithm of the height. This layer usually extends above the constant stress layer. Within the logarithmic layer we can de- fine an "eddy diffusion" coefficient: K=kU*z (3) where k — 0.4 is von Karmen's constant. Thus we may write for the heat flux in this part of the boundary layer: H=~psK bz (4) where p and s are the density and specific heat of sea- water, respectively,// is the heat flux from the earth's interior, and dT/dz is the gradient of the potential temperature. For the TAG area we will assume ps = 4.2 X 106 J/m3 °C and z = 10 m. From Table 1 , the gradient of the potential temperature is -1.5 X 10-2 C/m. The value U* is uncertain, but observations in the ocean bottom boundary layer have yielded values from 2 X 10"4 to 2 X 10"3 m/s [22]. Table 2 gives the heat flux per unit area, the heat flux per unit length of the ridge axis assuming a width of 250 m fur the TAG temperature anomaly, and the total heat output from the anomaly region, assuming lengths of 1 km and 10 km. The heat flux values in Table 2 appear to be rather large. This may partially be due to the fact that the logarithmic layer is expected to extend to a height of the order of 1 m above the seafloor, whereas the tem- perature gradient was measured at a height of the order of 10 m. Wimbush and Sclater [23] suggest that application of eqs. 3 and 4 at heights above the log- arithmic layer may lead to overestimates of the heat flux. Nevertheless, for a "typical" value of U* = 0.1 cm/s [22], and a length of 10 km, the total heat out- put from the TAG anomaly area is of the same order of magnitude as for the Wairakei high-temperature area in New Zealand [24], Furthermore, in high-tem- perature continental geothermal areas, the heat flow per unit area is often found to be of the order of sev- eral tens of watts per square meter [25,26]. It is pos- sible that such heat transport could be achieved by hydrothermal circulation in the upper few kilometers of the oceanic crust. The absence of an observed temperature anomaly a few kilometers away (Fig. 1) suggests that the 1 km length may be more appropri- ate for the TAG area. Thus the results in Table 2 do not appear to be too unreasonable. In any case, it would appear that the heat flux in the region of the TAG anomaly is a significant fraction of the heat loss due to the creation of new lithosphere at the ridge axis. 4. Conclusions In order to estimate the heat flux through the ocean floor from temperature anomaly data, much TABLE 2 Heat transfer estimates based on turbulent diffusion u* H Heat output per meter of Total heat out- Total heat out- X10-2 (W/m2) ridge axis assuming 250 m put assuming 1 put assuming 10 (m/s) anomaly width (kW/m) km length (MW) km length (MW) 0.02 5.0 1.26 1.26 12.6 0.05 12.6 3.15 3.15 31.5 0.1 25.2 6.30 6.30 63. 0.2 50.4 12.60 12.60 126. 311 23 better measurements are needed. Towed thermistor data can give only semi-quantitative results. Thermis- tors should be separated by no more than 1 m, and the array should be towed as close to the bottom as possible. Since it is generally not feasible to tow the thermistor array within the logarithmic layer (z < 1 m) because of the irregular topography on ridge crests, we recommend that ocean floor heat flux mea- surements in the crestal zone be made by the tech- niques described by Wimbush and Sclater [23]. This would involve determination of the velocity and tem- perature spectra within the logarithmic layer by means of a bottom mounted device. Such measure- ments have not been made in regions of young oceanic crest, and they would be especially useful in regions where temperature anomalies have been measured with towed thermistors. Measurements of this type would show (1) whether the large superadiabatic temperature gradients measured by towed thermistors are real, (2) whether strongly unstable layers persist in the ocean-bottom boundary layer, at least on a time scale of a tidal cycle, and (3) whether the turbu- lence in the boundary layer is shear generated or buoyancy generated. It may also be useful to measure temperatures in the upper 10-20 cm of sediment in regions of suspect- ed hydrothermal activity. Dawson [26] has used soil temperatures to infer heat flow in convection domi- nated regions of the Wairakei area. The technological problems are, of course, somewhat more difficult for seafloor measurements. It may be difficult to correct for variations in sediment temperature due to periodic variations in bottom water temperature. The two models which we have presented here for interpreting ocean-bottom water temperature anom- alies have rather apparent limitations. This is espe- cially true in view of the qualtiy of the existing data. The results presented here, however, do sugge that small, localized near-bottom water temperature anomalies may be associated with a convective heat transfer through the seafloor of a significant magni- tude. This suggests that water temperature anomalies may not be steady-state phenomena, but rather are indicative of transient cooling of very young oceanic crust by episodic hydrothermal circulation. Measur- able water temperature anomalies may therefore be somewhat rare. Acknowledgements We thank the reviewers for their valuable com- ments with regard to the original manuscript. In par- ticular, we thank Dr. G. Bodvarsson for suggesting that the water temperature anomaly be interpreted on the basis of turbulent diffusion theory. This work is part of the NOAA Trans-Atlantic Geotraverse (TAG) project. This work was supported by NOAA and the Oceanography Section of the National Science Foundation under NSF Grant DES 74-00513 A01. References 1 M. Talwani, C.C. Windisch and M.G. Langseth, Jr., Rcykjanes Ridge Crest: a detailed geophysical study, J. Geophys. Res. 76(1971)473. 2 D.L. Williams, R.P. von Herzen, J.G. Sclater and R.N. Anderson, Galapagos spreading center: lithospheric cool- ing and hydrothermal circulation, Geophys. J. R. Astron. Soc. 38(1974)587. 3 C.R.B. Lister, On the thermal balance of a mid-ocean ridge, Geophys. J. R. Astron. Soc. 26 (1972) 515. 4 J.G. Sclater and J. Francheteau, The implications of ter- restrial heat flow observations on current tectonic and geochemical models of the crust and upper mantle of the earth, Geophys. J. R. Astron. Soc. 20 (1970) 509. 5 R.L. Parker and D.W. Oldenburg, Thermal model of ocean ridges, Nature 242(1973) 137. 6 F. Aumento, B.D. Loncarevic and D.I Ross, Hudson geotraverse: geology of the Mid-Atlantic Ridge at 45°N, Philos. Trans. R. Soc. Lond., Ser. A, 268 (1971) 623. 7 J.B. Corliss, The origin of metal-bearing submarine hydro- thermal solutions, J. Geophys. Res. 76 (1971) 8128. 8 M.R. Scott, R.B. Scott, P.A. Rona, L.W. Butler and A.J. Nalwalk, Rapidly accumulating manganese deposit from the median valley of the Mid-Atlantic Ridge, Geophys. Res. Lett. 1 (1974) 355. 9 P.A. Rona, B.A. McGregor, P.R. Betzer and D.C. Krause, Anomalous water temperatures over the Mid-Atlantic Ridge crest at 26°N, EOS 55 (1974) 293. 10 P.A. Rona, B.A. McGregor, P.R. Betzer, G.W. Bolger and D.C. Krause, Anomalous water temperatures over Mid- Atlantic Ridge crest at 26°N latitude, Deep-Sea Res. 22 (1975)611. 11 E.R. Lapwood, Convection of a fluid in a porous medium, Proa Cambridge Philos. Soc. 44 (1948) 508. 12 G. Bodvarsson and R.P. Lowell, Ocean-floor heat flow and the circulation of interstitial waters, J. Geophys. Res. 77 (1972)4472. 13 R.P. Lowell, Circulation in fractures, hot springs and convective heat transport on mid-ocean ridge crests, 312 24 14 15 Geophys. J. Astron. Soc. 40 (1975) 351. R.B. Scott, P.A. Rona, B.A. McGregor and M.R. Scott, The TAG hydrothermal field, Nature 251 (1974) 301. B.A. McGregor and P.A. Rona, Crest of the Mid-Atlantic Ridge at 26°N, J. Geophys. Res. 80 (1975) 3307. 16 P.A. Rona, R.H. Harbison, B.G. Bassinger, R.B. Scott and A.J. Nalwalk, Tectonic fabric and hydrothermal activity of Mid-Atlantic Ridge crest (26° N) Geol. Soc. Am. Bull. 87(1976)661. 17 E.T.C. Spooner and W.S. 1'yfe, Sub-sea-floor metamor- phism heat and mass transfer, Contrib. Mineral. Petrol. 42(1973) 287. 18 G.K. Batchelor, Heat convection and buoyancy effects in fluids, Q. J. R. Meteorol. Soc. 80 (1954) 339. H. Rouse, C.-S. Yih and H.W. Humphreys, Gravitational convection from a boundary source, Tellus 4 (1952) 201. G.H. Keller, S.H. Anderson and J.W. Lavelle, Near-bottom currents in the Mid-Atlantic Ridge rift valley, Can. J. Earth Sci. 12(1975) 703. B.R. Morton, Sir G. Taylor and J.S. Turner, Turbulent 19 20 gravitational convection from maintained and instanta- neous sources, Proc. R. Soc. Lond., Ser. A, 234 (1956) 1. 22 M. Wimbush and W. Munk, The benthic boundary layer, in: The Sea, Vol. 4, A.E. Maxwell, ed. (Interscience, New York, N.Y., 1970) 731. 23 M. Wimbush and J.G. Sclater, Geothermal heat flux eval- uated from turbulent fluctuations above the sea floor, J. Geophys., Res. 76 (1971)529. 24 G.E.K. Thompson, C.J. Banwell, G.B. Dawson and D.J. Dickenson, Prospecting of hydrothermal areas by surface thermal survey, in: Proceedings of United Nations 1961 Conference of New Sources of Energy, Geothermal Energy 2, No. 1 (1964) 386. 25 D.E. White, Rapid heat flow surveying of geothermal areas, utilizing individual snowfalls as calorimeters, J. Geophys. Res. 74 (1969) 5191. 26 G.B. Dawson, The nature and assessment of heat flow from hydrothermal areas, N.Z. J. Geol. Geophys. 7 (1964) 155. 21 313 33 Reprinted from: Sedimentology , Vol. 23, No. 6, 867-872. Sedimentology (1976) 23, 867-872 An automated rapid sediment analyser (ARSA) TERRY A. NELSEN NO A A, Atlantic Oceanographic and Meteorological Laboratories, 15 Rickenbacker Causeway, Miami, Florida 33149, U.S.A. ABSTRACT The automated rapid sediment analyser (ARSA) is a pressure-transducer grain- size analysis system. This basic Woods Hole-type fall tube was automated by the addition of a digital voltmeter, Hewlett-Packard 9810A calculator, and an x-y plotter. Eight min after sample introduction, the system automatically produces size distribution data in 025-9 intervals, distribution statistics, and a plotted frequency histogram. INTRODUCTION As early as 1938, Emery (1938) turned to settling tubes as an alternative to tradi- tional sieves for a more rapid method of sediment textural analysis. Since then others have modified the original sand accumulation vs. time technique (Emery, 1938; Poole, 1957) by measuring pressure changes in the water column with the transit of falling grains (Zeigler, Whitney & Hays, 1960; Schlee, 1966; Bascomb, 1968) or by weight accumulation on a balance pan (Felix, 1969) similar to earlier Dutch work. Although the settling tubes achieve a significant time savings over sieving, the reduction of the analogue data produced still requires operator time for interpretation, statistical analysis, and graphic display. Only one early attempt at automated data acquisition from a settling tube is in the literature (Zeigler, Hayes & Webb, 1964), but it does not provide for real time data reduction, statistical analysis, and graphic dis- play. For laboratories analysing hundreds of samples, it is desirable to have a rapid sediment analyser (RSA) which is as fast as those previously built and also yields highly accurate and precise data while eliminating the human element from the time of sample introduction to final statistical treatment of the data. Although the concept of the settling tube does not limit the analysis range to sand size particles, the long fall times required for silt and clay sized particles would negate the benefits of rapid 867 314 868 Terry A. Nelsen analysis gained by the settling tube. Hence the analysis of fines (< 62 urn) is best undertaken by alternative methods (pipette or electronic particle counters), and the rapid sediment analyser is most efficiently employed for the textural analysis of sand. The instrument described here was therefore developed for only the size analysis of sand. It is a computer based data acquisition system coupled to a Woods Hole type (Schlee, 1966) rapid sediment analyser and is hereafter referred to as an automated rapid sediment analyser (ARSA). ARSA COMPONENT HARDWARE The fall tube used in this system is clear plastic and the inside diameter measures 10 cm by 200 cm in total length. Previous work (Gibbs, 1972) on the accuracy of particle-size analysis by settling tube indicated that tubes 7-5 cm and 12-7 cm in dia- meter were burdened with fall time inaccuracies of up to 34-8% respectively. It should be noted that these inaccuracies cited by Gibbs (1972) were the result of comparing the differences in fall velocities of a given size particle for a single sphere against samples of up to 4 g. Although the ARSA system described here is 10 cm in diameter, the calibration of the system was conducted relative to sieve analysis (Sanford & Swift, 1971), and the processes accounting for fall velocity errors were compensated for in the calibration technique. Pressure ports are located at 0-5 and 133 cm below the water level in the tube. This separation is necessary to insure that all introduced particles are in the sensing zone after the time required to damp surface oscillations resulting from sample introduction. Pressure and pressure changes within the water column are detected by a Hewlett- Packard Model # 270 differential gas pressure transducer and are interpreted as voltage changes resulting from the displacement of the transducer diaphragm. The rate of change of pressure represents the size distribution of the sample being analysed. This analogue voltage signal is conditioned by a Sanborn (Hewlett-Packard) Model 350-1 100C carrier preamplifier before it is sent to a Hewlett-Packard Model 3480B digital voltmeter (with Model 3482A DC range unit) where it is transformed into a digital voltage signal. A Hewlett-Packard 2570A coupler/controller with crystal clock provides a reference time base for the calculator's predetermined 0-25- fall times. Initial fall times were derived from Schlee's (1966) work and adjusted for the longer tube length of this system. The coupler/controller also provides electronic compati- bility between the digital voltmeter and the calculator memory. The memory-calcula- tion function of this system is provided by a Hewlett-Packard Model 9810A calculator. Final histogram display is generated on a Hewlett-Packard Model 9862A plotter. Total system compatibility dictated the exclusive use of a single electronics system. It should also be noted that line voltage fluctuations can introduce spurious transient signals into the system which cause erroneous voltmeter readings. Therefore it is necessary to supply power through a voltage regulator. The ARSA system is pictured in Fig. 1. The fall tube is suspended from a wooden frame by metal turnbuckles with foam rubber separation pads. The entire system is shock mounted from the floor by additional foam pads. This minimizes vibration transmission to the tube mounted transducer. Spirit levels secured to the fall tube at right angles insure a perfectly vertical tube orientation through turnbuckle adjustments. 315 An automated rapid sediment analyser ( ARSA) 869 Fig. 1. View of the total ARSA system showing fall tube and associated electronics. Approximately 150-200 samples can be run before accumulated sediment must be removed through the bottom drain valve and the tube refilled with deionized or distilled water. Figure 2 shows the sample introduction device. A controlled electric motor mounted above the tube depresses a sediment coated screen onto the surface of the water column. The inverted sub-62 micron screen holds the sample in place (as seen in Fig. 2b) by water surface tension. Parallel contact of the sediment-laden screen and the water surface releases the particles and permits a gentle and simultaneous discharge of the grains. Sample sizes between 5-7 g are used for all ARSA analyses. 316 870 Terry A. Nelsen Fig. 2. (a) Showing variable speed sample introduction device, top of fall tube, and upper transducer pressure port, (b) Samole introduction device with sediment on screen ready for sample run. DATA ACQUISITION AND COMPUTATION PROGRAM A Hewlett-Packard Model #9810A calculator provides the heart of this ARSA data acquisition system. The calculator program includes subroutines for data acquisition, storage, computation, and hard copy output. Before sample introduction the transducer's output voltage is adjusted to an arbitrary small positive value which is simultaneously displayed on the digital volt- meter. Data acquisition starts when sample introduction causes a predetermined threshold millivoltage to be exceeded. The program then pauses for 5 s while surface oscillations caused by sample introduction damp. Following this, three voltage values are read in the next second and placed in memory. Later these values will be averaged and this average used in the computational subroutine as the 100% reference value. Based on fall-time values in the memory bank, the calculator then runs time com- parison do-loops against the system's crystal clock. When the time value for each 0-25-<> interval of the sand range ( — 1 -00-4-00 §) is satisfied, the calculator commands the digital voltmeter to read the transducer voltage and place this value in memory for future use in the data reduction subroutine. Successive voltage values decline in magnitude as grain fallout past the lower pressure port causes the transducer's dia- phragm to return to the null (baseline) position. After gathering digital voltage values 317 An automated rapid sediment analyser (ARSA) 871 for all the 0-25-<> intervals in the memory, the computational subroutine takes over. Since all samples analysed in the ARSA have been prescreened (wet and dry) at 4-0 frequency histogram on the x-y plotter. An example of this graphic display is shown in Fig. 3b. The entire process from sample introduction to printed statistics and plotted histo- gram takes 8 min. -2-0 -1-0 0-0 0 2-0 3-0 4-0 5-0 Fig. 3. Examples of the system's x-y plotter output (sample 10-0-A) for (a) manually imputted sieve data, cp mean = 1-39, s.d. 091, and (b) an ARSA run data, 9 mean 1-40, s.d. 0-93. SYSTEM PERFORMANCE Although the ARSA was developed as an instrument which gave phi means similar to sieve phi means (final correlation coefficient of 0-99), the final system also showed a remarkable similarity between ARSA frequency distributions and sieve frequency distribution data (Fig. 3) with an overall correlation coefficient for 0-25-4 intervals of 0-86. ACKNOWLEDGMENTS Throughout the development of this system, Donald J. P. Swift, Patrick G. Hatcher, and Charles Lauter offered constructive criticism and sound advice which was greatly appreciated. 318 872 Terry A. Nelsen REFERENCES Bascomb, C.L. (1968) A new apparatus for recording particle s;ze distribution. J. sedim. Petrol. 38, 878-884. Emery, K.O. (1938) Rapid method of mechanical analysis of sands. J. sedim. Petrol. 8, 105-111. Felix, D.W. ( 1969) An inexpensive recording settling tube for analysis of sands. J. sedim. Petrol. 39, 777-780. Gibbs, R.J. (1972) The accuracy of particle-size analysis utilizing settling tubes. J. sedim. Petrol. 42, 141-145. Krumbein, W.C. & Pettijohn, F.J. (1938) Manual of Sedimentary Petrography. Appleton-Century- Crofts, New York, U.S.A. Poole, D.M. (1957) Size analysis of sand by a sedimentation technique. J. sedim. Petrol. 27, 460-468. Sanford, R.B. & Swift, D.J. P. (1971) Comparison of sieving and settling techniques for size analysis, using a Benthos rapid sediment analyser. Sedimentology, 17, 257-264. Schlee, J. (1966) A modified Woods Hole rapid sediment analyser. J. sedim Petrol. 36, 403-413. Zeigler, J.M., Whitney, G.G. & Hayes, C.R. (1960) Woods Hole rapid sediment analyser. J. sedim. Petrol. 30, 490-495. Zeigler, J.M., Hayes, C.R. & Webb, D.C. (1964) Direct readout of sediment analysis by settling tube for computer processing. Science. 145, 51. (Manuscript received 21 January 1976; revision received 23 March 1976) 319 34 Reprinted from: Bulletin, Vol American Association of Petroleum Geologists 60, No. 7, 1078-1106. Tectonics of Southwestern North Atlantic and Barbados Ridge Complex1 Abstract More than 40,000 km of bathymetric, mag- netic, and gravity data and 2,000 km of seismic-reflec- tion data were obtained in 1971 and 1972 aboard the NOAA ships Researcher and Discoverer over the Bar- bados Ridge complex and the ad|acent southwestern North Atlantic. Most of the tracklines were oriented east-west and spaced closely (20 km) to attempt corre- lation between adjacent lines. About a dozen long, north-south-trending tracklines provided c >ntrol on the structural variations in that direction From bathymetric and magnetic dala t .vat. • Jtab- lished that from the Late Cre'-jceous '■:■■ v.;/ '.ie the development of the Mid-Atla;.i.c Ric^w i ENia aiea is essentially the same as in the rest of the North Atlantic. Indications for relatively recent tectonic activity were found on some seismic records along several east- west faults, some of which were in alignment with off- set zones of the magnetic-anomaly lineations. The im- plicit suggestion is that intraplate tectonic activity is common, and that the western extension or "dead traces" of transform faults may provide avenues where the accumulated tectonic energy within the oceanic plate is released. The influence of many of the major east-west faults extends westward from the Atlantic basin across the Barbados Ridge complex to the platform of the Lesser Antilles volcanic arc. Major topographic changes, as well as changes in the character of the geophysical anomalies and in the chemistry of the volcanic rocks across the fault lines suggest that the faults have played a significant role in the evolution of this area. As these faults apparently have affected the structure west of the shallow earthquake belt and the axis of the gravity minima, this area appears ideal to study possi- ble anomalies in the subduction process or, perhaps, the applicability of the concept itself. Introduction The area from the Romanche fracture zone, near the equator, to the Barracuda Ridge, about 16°N, is one of the more complex geologic areas of the Atlantic Ocean floor. Whereas the general evolution of the North and South Atlantic Oceans, on the basis of magnetic-anomaly linea- tions, had been understood by 1970, this region remained a problem area because of the many fracture zones, the close proximity of the magnet- ic equator, and the lack of adequate survey cover- age. Yet this area is in a key position for critical tests or refinements of the plate-tectonics hypoth- esis (Isacks et al, 1968; Le Pichon, 1968; Morgan, 1968). In addition to the still unresolved problem of the overlap of Central and South American Paleozoic rocks in the Bullard reconstruction of Pangea (Bullard et al, 1965), and the controversy GEORGE PETER2 and GRAHAM K. WESTBROOK^ Miami, Florida 33149, and Keele, Staffordshire, England about the age and origin of the Barbados Ridge (Meyerhoff and Meyerhoff, 1972), the various tectonic concepts contained in the papers that discuss the evolution of this part of the Atlantic include north-south extension (Funnell and Smith, 1968), north-south extension and left-lat- eral shear (Ball and Harrison, 1969, 1970), and sea-floor spreading along east-west-trending mid- oceanic-ridge segments (Dietz and Holden, 1970; Freeland and Dietz, 1972). To test some of these hypotheses, in 1971 and 1972 a systematic geophysical study of the sea floor was undertaken between the Lesser Antilles island arc and the Mid- Atlantic Ridge (Fig. 1). The northern and southern boundaries of the study area were approximately 18°N and 10°N, respectively. The specific scientific objectives were to: (1) investigate the possible presence of magnetic-anomaly lineations east of the Lesser Antilles island arc, establish their trend, and iden- tify them; (2) define topographic and structural trends, and establish the development of the Mid- Atlantic Ridge and the associated fault zones in the area; (3) determine the east-west extent of the Barracuda fault zone and its role as a major ©Copyright 1976. The American Association of Petroleum Geologists. All rights reserved. 'Manuscript received, July 23, 1975; accepted. January 20, 1976. 2NOAA, AOML. MG&GL. 3The University. The success of this work was due largely to the cooperation and dedication of the captains, officers, and crews of the NOAA ships Researcher and Discoverer. We are grateful to Omar E. DeWald, George Merrill, and Sam A. Bush of the Atlantic Oceanographic and Meteorological Laboratories (AOML) of the NOAA for their significant contributions to the data-collection and processing phases of this work, and to the preparation of some of the bathymetric and magnetic maps. We acknowledge George H. Keller and Bonnie C. McGregor for their critical reviews of the manuscript, and Claire Ulanoff for her cheerful editorial and typing work. Data presented in this paper over the Barbados Ridge complex south of 14°N were provided before their publication by Graham K. Westbrook. University of Keele. England. This work was supported by the Marine Geology and Geophysics Laboratory of AOML, NOAA, with contributions from NSF-IDOE Grant NO. AG-253 and AG-489. Although the writers generally agree on the interpretation presented, the senior writer takes full responsibility for challenging some of the concepts of plate tectonics. 1078 320 Southwestern North Atlantic and Barbados Ridge 1079 Fig. 1 — Trackline coverage east of Lesser Antilles island arc (NOAA 1971, 1972). transform fault or plate boundary; (4) describe in detail the southeast extension of the Puerto Rico Trench and the area of transition between it and the Barbados Ridge (of special interest was the determination of the role of the Barracuda and other fault zones as barriers to sediment deposi- tion); and (5) determine the structure of the Bar- bados Ridge, and investigate subduction and un- derthrusting as possible mechanisms for its formation. Data-collection techniques, instrumentation, and comments on data processing and accuracy were given by Peter et al (1973a, b) and Dorman et al (1973). In this paper the bathymetric, mag- netic, gravimetric, and seismic-reflection results will be discussed in light of the basic objectives. Mid-Atlantic Ridge Bathymetry Previous investigations have established that the overall trend of the Mid-Atlantic Ridge east of the Lesser Antilles island arc is north-south (Heezen and Tharp, 1961; Collette et al, 1969; van Andel et al, 1971; Collette and Rutten, 1972). Most of the NOAA tracklines (Fig. 2) were ori- ented perpendicular to this trend, and were ex- pected to reveal the development of the Mid-At- lantic Ridge with little interference from *Z 35 36- -r- 55* -15* -T- 50* T- 45* 20- -T 40- Fig. 2 — Identification of selected east-west tracklines. 321 1080 George Peter and Graham K. Westbrook 60°W 55°W 50°W 45°W n — t — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — r CO v^\ -i — i — i — i — i — i — i i i i_i i i i i i i i i ■ ' ' 500 2000 1000 1500 DISTANCE (KILOMETERS) Fig. 4 — Bathymetric profiles along southern half of east-west tracklines. 323 2500 1082 George Peter and Graham K. Westbrook in cc £ 6 s "o 7 1 — I I I I I I I I I I i i i i 1000 500 0 Km 2 — s \ 3- ^i J AWy 111 lUi i '\i/j '\ rwf lit - ' fh Akjl 4 — i ii iuU fiilfyw iwv \ 5- $ . ,.li.*Uiil \MMrf • K 6- PiWf"^"«l J.i _f_ }_ 30 25 20 13 6 5 ! 5 t_ t_ J I t / ( \ l3^_l2L _J!° J_ ; I i [30 }25 " |20 fa |6 I5 f' i l i i i i i i i i i i_ VEMA F Z 16° 14° 12° 10° 60° 58" 56° 54° 52° 50° 48° 46° 44° 42° 40° Fig. 1 1 — Identification and offset pattern of magnetic-anomaly lineations east of Lesser Antilles arc. Anomaly numbers are after Pitman et al (1968). Anomaly no. 1 is over axis of Mid-Atlantic Ridge. Tertiary, Quaternary turbidites (Damuth and Fairbridge, 1970; Embley et al, 1970). The block south of 12°54'N appears to be an exception, be- cause here the transparent zone is overlain by a sedimentary unit approximately 1,000 m thick (intermediate turbidite unit) characterized by strong incoherent reflections and by three or four weak reflecting horizons, which locally fade into the incoherent zone. From 12°20'N southward this unit is overlain by strongly stratified se- quence of late Tertiary -Quaternary turbidites. Whereas normal, reverse, and thrust faults commonly can be recognized on seismic records run normal to the trends of these faults, strike-slip motion cannot be established on the basis of these records alone. The studies of Currey and Nason (1967) of the seaward extension of the San Andreas fault revealed that a zone of chaotic re- flections, abruptly terminating coherent reflecting horizons, can be expected in a strike-slip fault zone. This zone also may involve complementary normal faulting. From these observations and the fact that some of the faults on Figure 17 lie at the ~T~V t -i500 400 300 200 100 0 300 Km Fig. 12 — Magnetic-anomaly profiles 41 and 70 showing correlation across Vema fracture zone. westward extension of magnetic-offset zones, the apparent normal faults centered on 15°15'N and 14°38'N, and the fault zones centered on 12°53'N and 11°05'N also may have had strike-slip dis- placements. The importance of near-bottom currents in car- rying and eroding the sediments is illustrated at 15°15'N (Fig. 17) where the trough left by the fault already is filled by transparent sediments, and at 11°40'N where there is a small erosion channel at the point where the dip of the sea floor changes from a southerly to a northerly direction. Profile A (Figs. 13, 15) illustrates currently ac- tive faults south of the Barracuda Ridge. Their relative youth is indicated by the step-like dis- placements of the sea floor, and the increased dis- placement of the deeper reflectors also indicates activity along these faults in the geologic past. Most of the faults are associated with scarps or steep slopes of the basement, suggesting that the origin of the tectonic activity is within the oceanic crust. The northernmost fault at 15°39'N appears to be related to the relative uplift of the Barracu- da Ridge, the troughs at 15°N and at 15°15'N are related to the east-west trending fault systems of this area. Barbados Ridge Complex Bathymetry The Barbados Ridge complex is an outer sedi- mentary-island arc that consists of a system of the north-south and east-west-trending ridges, troughs, scarps, and topographic lineaments, whose respective development varies along the 331 1090 George Peter and Graham K. Westbrook c o u o o C o E .a 332 Southwestern North Atlantic and Barbados Ridge 1091 O - M > S-i ^t/fm^y^%^^-.»,<-}-.y-V^'Mii If/a, ~-^;MM 0 ,.^ 8- - s i \ :.jft 4 . ; o- % t : . .,■ || ; | ; .,...,. ■• :,.. ,0 B~ » m « 1 ■£ * > 00 Q. B o 4 o e •» « « S0NO33S Nl 3WU 13AVU AVM'OMl S0NO33S Nl 3WI1 13AVH1 AVM-OM1 333 1092 George Peter and Graham K. Westbrook I6°00' N Fig. 15 — Retouched photograph of north-south seismic-reflection line along 56°30'W, directly south of Barracuda Ridge. 60° W I8°N 16°- 14°- 18° N -16* _l4< Fig. 16 — Magnetic-anomaly map over northern half of Barbados Ridge and adjacent Atlantic sea floor. 334 Southwestern North Atlantic and Barbados Ridge 1093 k?- > S0NO33S Nl 3WI1 13AV81 AVMOM1 mi. • >iM o Z g u. SQNOD3S Nl 3WI1 13AV81 AVMOAAi 335 1094 George Peter and Graham K. Westbrook Fig. 18 — Key physiographic elements of Lesser Antilles arc-Barbados Ridge complex. Contour interval 1 km. arc. The north-south-trending topographic ele- ments are part of the overall island-arc trends, the east-west elements are related to the fault systems of the adjacent Atlantic basin. These extend west- ward underneath the Barbados Ridge complex and have cut or modified the major north-south elements. The interaction of the north-south and east-west tectonic trends often produced locally complex structural and topographic patterns: compared to the trackline spacing, the wave- length of these features is too short and, there- fore, these were not incorporated into the bathy- metric presentation (Fig. 13). Four major north-south-trending topographic elements may be distinguished within the Barba- dos Ridge complex. These are: (1) the Tobago trough and Lesser Antilles Trench; (2) the Barba- dos Central Ridge province; (3) the Barbados Trough province; and (4) the Barbados Frontal Hills zone. The major east-west-trending topo- graphic elements are: (1) the Dominica Trans- verse Valley system; (2) the Martinique Trans- verse Valley system; and (3) the Sta. Lucia- Barbados Transverse Ridge system (Fig. 18). The north-south topographic elements are prominent from the continental shelf of South America northward to about 13°N; there is a transition zone between 13°N and 14°N; and the east-west elements are more abundant north of 14°N. The 14°N parallel also cuts the Barbados Central 336 Southwestern North Atlantic and Barbados Ridge 1095 w KM Fig. 19 — Photograph of seismic section L (Fig. 13). Ridge, and the average depth of the Barbados Ridge complex is more than 1,000 m greater north of here. South of 14°N, the backbone of the Barbados Ridge complex is the Barbados Central Ridge, on which the island of Barbados is located. In addi- tion to the change of strike at 13°N, the single Central Ridge also may have broken into a broader "ridge province," if one assumes that the isolated peaks north of Barbados are part of the Central Ridge. As an alternate interpretation, these peaks also could be related to the Sta. Lu- cia-Barbados Transverse Ridge system. The Bar- bados Central Ridge and the other elements of the Barbados Ridge complex are well developed south of 13°N. North of 14°N, the topographic high between the Dominica and the Martinique Transverse Valley systems may represent the northernmost element of the Barbados Central Ridge province. The Lesser Antilles Trench separates it from the volcanic arc on the west, and a prominent, nar- row depression at 58°45'W forms its eastern limit. The Lesser Antilles Trench has the same over- all width as the Tobago trough, but its northern half is "V" shaped with a graben 10 to 15 km wide in the center (Fig. 19). It is difficult to trace the Barbados Trough province between 13°N and 14°N because, as stated before, there are several minor valleys and ridges there, occupying an area more than 100 km wide. North of 14°N, the valley at 58°45'W is a 337 feature that may be a structural equivalent or a northern extension of the Barbados Trough prov- ince. Its northern terminus is at 15°N, where it merges with the Dominica Transverse Valley sys- tem. Among the major east-west-trending topo- graphic elements the Dominica Transverse Valley system is the northernmost (15°N). Although there is a limited amount of data available, it ap- pears that whereas the overall strike of the system is almost due east-west, it consists of four en-ech- elon, northwest-southeast-trending valleys, sepa- rated by narrow, sharp peaks. Its eastern terminus seems to be at 58°45W. On the west it merges with the Lesser Antilles Trench, although the bench on the slope of the volcanic-arc platform also may be genetically related to it. The Martinique Transverse Valley system is lo- cated between 13°55'N and 14°25'N. It is more than 50 km wide and has a definite east-west strike. In detail this valley system also is highly fragmented, but here by a combination of narrow troughs and ridges, striking both east-west and north-south. The broad saddle between the Tobago trough and the Lesser Antilles Trench is the western third of the Sta. Lucia-Barbados Transverse Ridge system (Weeks et al, 1971; Bassinger and Keller, 1972). The east-west-trending axis of the system shifts southward to 13°35'N east of the Central Ridge, and the system becomes much narrower. It may extend from the Central Ridge 1096 George Peter and Graham K. Westbrook o Z o u LU in 60° 00' W > < i— >- < i o KM Fig. 20 — Retouched photograph of western edge of the Barbados Central Ridge (sec. B, Fig. 13). to the Frontal Hills zone in the form of irregular topographic highs. Data are inconclusive to de- cide whether the topographic highs between 59°W and 59°30'W are part of this Transverse Ridge system or part of a broader Central Ridge province. North of the Barbados Frontal Hills zone, be- tween about 59°W and the island platform of the volcanic arc, lies a generally hummocky topog- raphy that resembles in geologic and geophysical character the Frontal Hills zone. The area has a roughly triangular shape with the volcanic-arc platform, the Frontal Hills zone, and the topo- graphic extension of the Puerto Rico Trench forming the sides of the triangle. The effect of both east-west and north-south faulting is indi- cated in the general bathymetric trends and in the basement configuration below the sediments (Schubert and Peter, 1973; Schubert, 1974). Seismic-Reflection Data Marine seismic-reflection data over several ele- ments of the Barbados Ridge complex already have been discussed by Chase and Bunce (1969); Collette et al (1969); Bunce et al (1970); West- brook (1973); and Peter et al (1974). Here only a few comments will be made about the NOAA 1971-1972 seismic-reflection data, as these illus- trate the overall structural division of the Barba- dos Ridge complex and the relative age of some of the tectonic events. Line J (Fig. 14) is a continuous line from the Atlantic basin, across the Barbados Ridge com- plex, the volcanic arc, into the eastern margin of the Grenada trough. Because the east-west topo- graphic elements are more significant north of 14°N, one can see from the topography (Fig. 13) that a 10-km north-south shift of the trackline would have highlighted different topographic ele- ments. The line extends across the southern mar- gin of the Lesser Antilles Trench and across the northern tip of the Barbados Central Ridge. On the west flank of the Central Ridge, subbottom reflectors of the Lesser Antilles Trench arch up and pinch out, or terminate at the sea floor over the Central Ridge. Only a thin layer of sediments covers the acoustic basement on the Central Ridge. This is shown well on line B (Fig. 20). Line J runs along the southern part of the Martinique Transverse Valley and, although good reflecting horizons are lacking, approximately Vi to 1 sec of penetration is indicated on the record. East of 59°W, the prominent valley on the record lies in the same structural province as the Barbados trough farther south; on the east lies the Frontal Hills zone. This zone is characterized by lack of reflectors and penetration along this line. The At- lantic basin on the east of this area (Fig. 13) is 338 Southwestern North Atlantic and Barbados Ridge 1097 O — Nov»n,0'^eo0- 8? iU :> * '"« h t»' , o~ "V m O •* np>^>n>ON.coo» S0NCO3S Nl 3WI1 13AVS1 AVM"OMl SGNOD3S Nl 3Wli 13AVS1 AVMOAAi o E o U 0 u. 339 1098 George Peter and Graham K. Westbrook Fig. 22 — Photograph of seismic section M (Fig. 13). gently arched, and the sediments are dipping away from 57°W, both toward the Atlantic and toward the island arc. A good onlap sequence of sediments is present east of the arch at 57°W and there is an abrupt termination of a strong, shallow reflector under the toe of the Frontal Hills. It of- ten is said that the lack of continuous reflectors and poor penetration under the toe of island arcs in general, and in this area in particular, is the result of very complex faulting (Chase and Bunce, 1969). Records will be presented here that show that when good reflectors are present, they can be detected even in cases of very intense faulting. The strong reflector that dips toward the Frontal Hills from the Atlantic basin does not show up on the Frontal Hills because it has not been deposit- ed there. The termination point of that reflector W marks a former edge of the Atlantic basin and the toe of the Frontal Hills zone, which subsequently became covered by transparent sediments, ex- tending the toe eastward. Most of the reflectors of the Lesser Antilles Trench on line K (Fig. 21) are uplifted over that part of the Central Ridge that lies between the two transverse valleys. Despite intense faulting the reflectors are recognizable. Along line M (Fig. 22) this entire upper sequence is shown clearly, suggesting that the valleys and ridges are the product of tectonics rather than erosion. The At- lantic basin dips gently toward the west along line K, and it appears that the reflectors become less coherent toward the west. Some hints of weak re- flectors suggest that this entire incoherent se- quence of sediments is uplifted in the toe of the KM Fig. 23 — Photograph of seismic section D (Fig. 13). 340 Southwestern North Atlantic and Barbados Ridge 1099 59°00'W > 10 i 10 20 KM Fig. 24 — Photograph of seismic section P (Fig. 13). Frontal Hills. Farther south, line D (Fig. 23) shows the uplift of the Atlantic sediments quite clearly; the well-stratified beds are progressively thinner as they are uplifted successively along at least three faults to form the toe of the Frontal Hills. These faults are either high-angle reverse, or normal faults. Farther up on the slope the re- cord shows an approximately 1-km thick zone of incoherent reflectors, which are similar to those shown in line K and on some other NOAA pro- files (unpub.) that were run directly east of the Frontal Hills zone. The uplifted sediments of the Atlantic basin on the Frontal Hills zone are in direct contrast with the gently westward-dipping (dip approx. 2°) sed- iments of the Atlantic, north of the Barbados Ridge complex. On line P (Fig. 24) these sedi- ments are overlain by a thick accumulation of an acoustically incoherent sediment pile and, as far as the instrumentation allowed them to be seen (3 sec penetration, 45 km from the edge of the over- lying sediment pile), reflecting horizons within these sediments are undisturbed. Line L (Fig. 19) illustrates the central graben and youthful tectonism of the Lesser Antilles Trench. Gravity Several aspects of the gravity anomalies of the Lesser Antilles island-arc system have been dis- cussed by Talwani (1966), Bush and Bush (1969), and Bunce et al (1970). The NOAA and Universi- ty of Durham investigations (Kearey, 1973; West- brook, 1973, 1975; Peter, 1974; Peter and West- brook, 1974a, b; Westbrook, 1974a, b; Kearey et al, 1975; Westbrook, 1975) allowed the mapping of the gravity field of this area in much greater detail than previously, and established the exten- sion of certain structural trends from the Atlantic basin into the Barbados Ridge complex. The most noticeable feature of the gravity field is the continuation of the negative free-air anom- aly band of the Puerto Rico Trench, which turns away from the topographic axis as 18°N. As the anomaly band extends farther south, it reflects the east-west structural discontinuities similar to the topography. These effects manifest them- selves as: (1) reduced amplitude of the gravity low (-192 mgal) between 19°N and 18°N (from Schubert, 1974); (2) sharp increase of the ampli- tude at 17°20'N (from -220 mgal to -276 mgal, Schubert, 1974); (3) a 45-km sinistral offset of the axis of the low at 16°30'N, and another similar offset at 15°10'N; (4) the interruption of the neg- ative, north-south anomaly band by a positive, east-west-trending free-air anomaly band at 13° 50'N; and (5) the development of two approxi- mately parallel negative free-air anomaly bands south of Barbados (13°N) with amplitudes about 100 mgal less than the amplitude of the single band on the north (Fig. 25). The axis of the free-air anomaly minimum does not follow the Lesser Antilles Trench south of the Sta. Lucia-Barbados Transverse Ridge; the sinis- tral offset at 15°10'N places it east of the Barba- dos Central Ridge. South of Barbados, the axis of 341 1100 George Peter and Graham K. Westbrook 342 Southwestern North Atlantic and Barbados Ridge 1101 E o c CL. a 0 343 1102 George Peter and Graham K. Westbrook one of the negative anomaly bands is over the eastern edge of the Tobago trough, the axis of the other is over the Barbados trough. These two bands are parallel near Barbados, but separate at the latitude of Tobago ( 1 1 ° 1 5'N ), where the west- ern band veers west-southwest onto the Paria shelf, and the eastern band continues south- southwest until about 10°30'N, where it sharply swings toward Trinidad. The free-air anomaly contour lines trend essen- tially east-west over the Atlantic basin, east of the Barbados Ridge complex. There are two promi- nent east-west-trending anomaly bands; one is a positive band at about 13°50'N, the other is a negative band north of it. These two seem to ex- tend from the volcanic platform to about 56° 30'W in the Atlantic basin. The Bouguer anomaly map of the area (Fig. 26) also is dominated by essentially north-south and east-west trends: the north-south trend follows the Barbados Ridge complex, and the east-west trend characterizes the Atlantic basin. When the regional gradient is removed from these data, some of the east-west-trending Bouguer anom- alies also cross the complex, and extend to the island platform of the Lesser Antilles arc. The north-south-trending Bouguer anomaly features include: (l) a gradient change under the Frontal Hills zone due to the dipping mantle; (2) a band of gravity lows over the greatest sediment accumulation, which on the south is over the Bar- bados Central Ridge (presumably this band marks the location of a former trench); and (3) a series of highs (170-240 mgal) that extend from the island platform east-southeast of Guadeloupe to the Sta. Lucia-Barbados Transverse Ridge over the lower slope of the island platform. Many of these highs are even larger than those reported over the volcanic islands themselves, which only reach 120-190 mgal (Kearey et al, 1975). Bouguer anomalies also were used to study fur- ther the east-west structures of the Atlantic basin. Two-dimensional modeling of the crustal struc- ture was performed using the Bouguer anomalies, NOAA seismic-reflection data (Peter and West- brook, 1974b), and University of Durham and earlier seismic-refraction data (Ewing et al, 1957; Westbrook et al, 1973). An interesting result that emerged was the fact that when Bouguer anom- alies were computed down to the acoustic base- ment with a realistic sediment density (2.0 g cm-3), then many of the large anomalies were eliminated (Fig. 27). These results suggest that in this area the buried topography of the basement is responsible largely for the Bouguer anomalies, with only very minor contribution from changes of mantle elevation. Discussion One of the main objectives of this paper is to present a large body of new data over a previous- ly little studied region of the southwestern North Atlantic, and the Barbados Ridge complex. Pa- pers by Westbrook (1973, 1974a, 1975) discussed in detail how some of these data may be fitted into an overall plate-tectonics scheme. We intend only to highlight here the relation of these new data to (1) the scientific objectives outlined earli- er; (2) some of the hypotheses advanced for the evolution of this area; and (3) some of the corol- lary assumptions of the plate-tectonics hypothe- sis. At the time this project was initiated in 1971, the sea-floor-spreading history was not known east of the Lesser Antilles arc. Although the mag- netic-anomaly lineation pattern presented (Fig. 1 1) is admittedly debatable in detail, the magnetic lineations and the east-west topographic profiles clearly establish that there is a well-developed Mid-Atlantic Ridge east of the Lesser Antilles, and that this segment of the ridge has evolved since the Late Cretaceous in the same way as in the rest of the North Atlantic. Two tracklines in- dicate similar development of the Mid-Atlantic Ridge even south of the Vema fracture zone. From geometric considerations of the original fit of the continents, we propose that the possible southern limit of this type of ridge development is the Doldrums fracture zone (approx. 8°N). We found neither topographic nor magnetic evidence for major breaks in the continuity of the Mid- Atlantic Ridge, which would have supported a major north-south extension of this area during the Cenozoic (Funnell and Smith, 1968; Ball and Harrison, 1969, 1970). As the Late Cretaceous magnetic lineations also are trending north-south, there is no indication for the existence of a ridge- ridge triple junction at that time, which might have provided some indirect support for the inter- pretation and identification of the east-west- trending anomalies in the Colombia basin as being Late Cretaceous (Christofferson, 1973), if this part of the Caribbean were formed in an At- lantic spreading regime. As part of the topographic and structural stud- ies of the Mid-Atlantic Ridge, several major (transform) and minor faults were located. Data from the northern half of the study area indicate that the northwest-southeast trend of the faults near the ridge axis changes to east-west between magnetic anomalies 6 and 13. The eastern half of the Barracuda Ridge follows an east-west fault pattern, and there is no indication in the bathy- metric data that it extends farther east and con- 344 Southwestern North Atlantic and Barbados Ridge 1103 425 250 400 225 mgal |g| ^ong 54 5C w TO ACOUSTIC TO SEABED 8&SEMENT - 425 250 400 225 375 200 350 I75 325 150 J50 175 325 150 0 5 km 10 0 "i 1 04 " 2 e 5 3 10 15 " 15 20 ! 1 1 1 1 1 1 1 1 _i 20 Fig. 27 — Bouguer anomaly profiles and crustal structure along 54°30'W. Upper half: Bouguer anomaly profile to seabed, lower numbers on scale; Bouguer anomaly profile to acoustic basement, higher numbers on scale. Lower half: crustal section from sea surface to mantle. Density of 1.04 g/cc = sea water, 2.0 g/cc = sediment, 2.8 g/cc = average crust, and 3.3 g/cc — mantle. Top of 2.8-g/cc layer was taken from our seismic-reflection record, top of 3.3-g/cc layer was computed to fit "acous- tic basement Bouguer profile," with nearby seismic-refraction data extrapolated as guide. Horizontal and vertical scales are in kilometers; for map location, latitude crossings also are shown. nects with the Fifteen-Twenty or Researcher frac- ture zones. Clearly additional bathymetric and magnetic data are needed to define better the to- pography and structure of the flank of the Mid- Atlantic Ridge where changes in the trend of these faults take place, and where major ridges, like the Barracuda and Researcher, seemingly ter- minate. One of the interesting results of our study was the discovery of relatively recent tectonic activity along many major and minor faults west of the exposed flank of the Mid-Atlantic Ridge. This ac- tivity was especially obvious in the bathymetric and seismic-reflection records from the Demarara abyssal plain directly south of the Barracuda Ridge (Fig. 15), and along the east-west faults on the 300-km-wide, uplifted crustal block adjacent to the Barbados Ridge complex. Our detailed sur- veys and the seismic records suggest that faulting and tectonic activity within an oceanic crustal plate may be common. The lack of any clear pat- tern of recorded earthquakes from these areas may indicate that (1) there are no very recent movements along these faults; (2) the displace- ments occur through creep; or (3) the magnitude of the earthquakes is so small that they cannot be recorded at far removed measuring stations. Because of the east-west-trending offset zones of the magnetic lineations and the fact that the faults on the seismic records often coincide, one also may conclude that the offsets of magnetic- anomaly bands may not be due entirely to trans- form faulting, or that the so-called "dead-traces" of transform faults remain zones of weakness where subsequent tectonic adjustments occur. Relatively recent motions along these faults also may support one of the contentions of Ball and Harrison (1969, 1970) that most of the faults cutting the Mid-Atlantic Ridge in the equatorial region are very slow-moving transcurrent faults. They argued that if the transcurrent motion is slow compared to the combined spreading mo- tion between the offset parts of the ridge crest, then the earthquake first-motion studies will show the "transform" movement but the transcurrent part may not be detected. Bathymetric and seismic-reflection data show that the Barracuda Ridge becomes a subbottom feature west of 58°50'W, and extends westward along the same trend to the island platform of the Lesser Antilles arc (Schubert and Peter, 1973; Schubert, 1974). However, it is only one of the many east-west structural elements in this area, and does not appear to have had a controlling influence by itself on the transition, of the south- eastward extension of the Puerto Rico Trench. As suggested by Peter et al (1974), the Puerto Rico Trench proper terminates at about 18°N, 60°W, where the zones of the gravity minima and the shallow earthquakes separate from the topo- graphic trough. If a definition of a trench as a structural feature accompanied by the zones of the gravity minima and shallow earthquakes is accepted, then the topographic trough extending 345 1104 George Peter and Graham K. Westbrook southeastward from the Puerto Rico Trench is not part of this trench. We believe that in this transition zone there is a buried trench, marked by the axis of the gravity minima and the co-lo- cated shallow earthquake zone, which connects with the Puerto Rico Trench on the north, and the Lesser Antilles Trench on the south. The high acoustic reflectivity of the sea floor and the manv incoherent subbottom reflectors make the structural analysis of this area very dif- ficult. However, the roughly east-west bathyme- tric trends, and the abrupt changes in the ampli- tude of the gravity minima most likely reflect the westward extension of the fault systems that char- acterize the Atlantic basin east of the Lesser An- tilles arc and the Barbados Ridge complex (Schu- bert, 1974). Some of these faults appear to have caused the offset of the axis of the buried trench in a left-lateral sense, and others probably are partly blocking it. Basically, however, it is not these faults, but the large amounts of sediments that have collected in the trenches and fault de- pressions east of the Lesser Antilles island arc that are responsible for the termination of the to- pographic expression of the Puerto Rico Trench. The hummocky topography in the transition zone (between the Puerto Rico Trench and the Barbados Ridge complex) is most likely slump and current-derived sediment that has been trap- ped between the east-west faults. The sediments overlie the gently westward-dipping Atlantic sea floor (line P, Fig. 24), causing the topographic low that extends southeast from the Puerto Rico Trench. As an interpretation alternative to that proposed by plate tectonics, it is possible that the upper sedimentary horizon (strong upper reflec- tor on line P) of the Atlantic sea floor is an un- conformity, rather than a thrust surface. Accord- ing to the subduction hypothesis the overlying sediment pile represents scrapings of sediments from the Atlantic sea floor. If such a process is possible, then in case of underthrusting not only the overlying sediment pile should be disturbed, but at least the upper, unconsolidated parts of the Atlantic sea-floor sediments as well. It is unlikely that the approximately l-km-thick, mostly unli- thified sediments covering the basement rocks of the Atlantic would underthrust a 2 to 3-km thick sediment pile to about 40 km, without undergoing any noticeable internal deformation. The increasing width and height of the Barba- dos Ridge complex toward the south generally are attributed to the availability of more sedi- ments closer to the South American source, and to the subsequent subduction of more sediments on the south (Chase and Bunce, 1969; West- brook, 1973). Our north-south seismic-reflection profiles east of the Barbados Ridge complex show that the thickness of sediments over the basement is essentially uniform between 13°N and 14°30' N; at about 15°30'N it increases rapidly toward the Barracuda Ridge, and at 12°20'N it increases substantially southward. West of the thickest sed- iments on the north, the Barbados Ridge complex terminates, and there are no changes in the Bar- bados Ridge complex south of 12°20'N either. These observations seem to indicate that there is no simple relation between the present sediment thickness of the Atlantic basin and the topog- raphy of the Barbados Ridge complex, and that other controlling factors should be considered. For this alternative we have suggested that the east-west faults of the Atlantic basin have extend- ed across the trough now occupied by the Barba- dos Ridge complex, and that these have formed dams against the northward-advancing sediments within the trough (Peter, 1974; Peter and West- brook, 1974a, b; Westbrook, 1974a, b, 1975). The influence of these fault zones also is manifested by the abrupt changes in the elevation of the Bar- bados Ridge complex, as they provided a suffi- cient discontinuity within the crust that uplifts of the sediment pile have occurred at different times and in different degrees between the various seg- ments of these faults. The seismic records are es- pecially clear in indicating a relatively recent (Pleistocene-Holocene) uplift of the Barbados Ridge complex north of 14°N (Fig. 19), but the uplift on the south has occurred earlier (Pliocene- Pleistocene?; Peter et al, 1974; Westbrook, 1975). In this context of decreasing age of the elements of the Barbados Ridge complex toward the north, the triangular-shaped hummocky region on the north — from its geophysical characteristics — is really the youngest member of this complex, that has not been subjected to significant uplift. If the major faults do define the boundaries of uplift, then it is reasonable to assume that the area be- tween the Barracuda Ridge and the present edge of the complex (approximately 16°N) will be uplifted next. The data presented here for the area of the Bar- bados Ridge complex show vertical uplift at the toe of the Frontal Hills zone (Fig. 2). Further, east-west fault zones, revealed by bathymetric, seismic, and gravity data, extend across the Bar- bados Ridge complex, the axis of the gravity mi- nima, and the shallow-earthquake zone to the vol- canic arc platform. At the largest of these fault zones, near 14°N, major changes occur both within the structure of the Barbados Ridge com- plex and the volcanic arc. These changes include: (1) the Barbados Ridge complex becomes 1,000 m deeper on the north; (2) the free-air gravity anomalies change character; (3) the positive Bouguer anomalies extending southward from Desirade over the lower slope of the island plat- form terminate; (4) the chemistry of the volcanic 346 Southwestern North Atlantic and Barbados Ridge 1105 rocks of the Lesser Antilles arc shows marked dif- ferences on the two sides of this fault (Stipp and Nagle, 1974; Wills, 1975); and (5) the trend of the volcanic arc changes. These observations and changes may be ex- plained by some anomaly in the subduction pro- cess (Westbrook, 1973, 1974a, b, 1975), and some of them may be only coincidences. In all certain- ty, however, these data are sufficient to question some of the current concepts of the subduction process as they are applied to this area, and sug- gest the need for serious, further investigations. Conclusions 1. The Mid-Atlantic Ridge in the area east of the Lesser Antilles arc developed from about the Late Cretaceous to the Holocene much as in the rest of the North Atlantic. Thus, the Barracuda Ridge and fracture zone is not a major disconti- nuity between oceanic crusts of different spread- ing history. 2. Relatively recent tectonic activity along the western extension of some transform faults sug- gests that these "dead traces" actually may pro- vide avenues for the release of tectonic energy in the oceanic plate. 3. Several east-west faults extend from the At- lantic basin to the island platform of the volcanic arc. These have cut the former trench east of the arc, have dammed the northward advance of sed- iments in the trough, and probably have caused the segmented differential uplift of the Barbados Ridge complex. Bathymetric and seismic-reflec- tion records indicate that the area south of 14°N was uplifted before the area on the north; the area north of 16°N has geophysical characteris- tics similar to the Barbados Ridge complex and may be thought of as the youngest member of the complex that has not received substantial uplift. 4. The Puerto Rico Trench terminates at ap- proximately 18°N, 60° W, where, because of sedi- ment fill and the influence of east-west faults, the topographic trough veers sharply eastward of the shallow earthquake zone and the axis of the gravi- ty minima. These, however, most likely mark the now buried part of the trench that connects southward with the Lesser Antilles Trench. 5. Because the east-west faults of the Atlantic basin cross the axis of the gravity minima and the shallow earthquake zone, and seemingly even in- fluence the structure of the volcanic arc and the chemical composition of its rocks, a simple sub- duction model probably is not applicable for this area. References Cited Ball, M. M„ and C. G. A. Harrison, 1969, Origin of the Gulf and Caribbean and implications regarding ocean ridge extension, migration, and shear: Gulf Coast Assoc. Geol. Socs. Trans., v. 19, p. 287-294. 1970, Crustal plates in the central Atlan- tic: Science, v. 167, p. 1128-1129. Bassinger, B. G., and G. H. Keller, 1972, Marine geo- physical observations across the Barbados Ridge-St. Lucia cross warp (abs.): 6th Caribbean Geol. Conf., Margarita, Venezuela, Trans., p. 379-380. Birch, F. S., 1970, The Barracuda fault zone in the west- ern North Atlantic: geological and geophysical stud- ies: Deep-Sea Research, v. 17, p. 847-859. Bullard, E. C, J. E. Everett, and A. G. Smith, 1965, The fit of the continents around the Atlantic: Royal Soc. London Philos. Trans., ser. A, v. 258, p. 41-51. Bunce, E. T., J. D. Phillips, R. L. Chase, and C. O. Bowin, 1970, The Lesser Antilles arc and the eastern margin of the Caribbean Sea, in The sea, v. 4: New York, Wiley-Intersci., p. 359-385. Bush, S. A., and P. A. Bush, 1969, Isostatic gravity map of the eastern Caribbean region: Gulf Coast Assoc. Geol. Socs. Trans., v. 19, p. 281-285. Chase, R. L., and E. T. Bunce, 1969, Underthrusting of the eastern margin of the Antilles by the floor of the western North Atlantic Ocean, and origin of the Bar- bados Ridge: Jour. Geophvs. Research, v. 74, p. 1413-1420. Chnstofferson, E., 1973, Linear magnetic anomalies in the Colombia basin, central Caribbean Sea: Geol. Soc. America Bull., v. 84, p. 3217-3230. Collette, B. J., and K. W. Rutten, 1972, Crest and frac- ture zone geometry of the Mid-Atlantic Ridge be- tween 10° and 16°N: Nature, Phys. Sci„ v. 237, p. 131-134. H. Schouten, and K. Rutten, 1973, Fracture zones and flow trajectories in the southern North At- lantic (abs.): Eos (Am. Geophys. Union Trans.), v. 54, p. 325. J. I. Ewing, R. A. Lagaay, and M. Truchan, 1969, Sediment distribution in the oceans: the Atlan- tic between 10° and 19°N: Marine Geology, v. 7, p. 279-345. H. Schouten, K. Rutten, and A. P. Slootweg, 1974, Structure of the Mid-Atlantic Ridge province between 12° and 18°N: Marine Geophys. Research- es, v. 2, p. 143-179. Curray, J. R., and R. D. Nason, 1967, San Andreas fault north of Point Arena, California: Geol. Soc. America Bull., v. 78, p. 413-418. Damuth, J. E., and R. W. Fairbridge, 1970, Equatorial Atlantic deep-sea sands and ice-age aridity in tropical South America: Geol. Soc. America Bull., v. 81, p. 189-206. Dietz, R. S., and J. C. Holden, 1970, Reconstruction of Pangea: Breakup and dispersion of continents. Perm- ian to present: Jour. Geophys. Research, v. 75, p. 4939-4956. Dorman, L. M., B. G. Bassinger, E. Bernard, et al, 1973, Caribbean Atlantic geotraverse, NOAA-IDOE 1971, Rept. 3, Gravity: NOAA Tech. Rept. ERL 277- AOML 11. Embley, R. W., J. I. Ewing, and M. Ewing, 1970, The Vidal deep-sea channel and its relationship to the Demerara and Barracuda abyssal plains: Deep-Sea Research, v. 17, p. 539-552. 347 1106 George Peter and Graham K. Westbrook Ewing, J. I., C. B. Officer, H. R. Johnson, and R. D. Edwards, 1957, Geophysical investigations in the eastern Caribbean: Trinidad shelf, Tobago trough, Barbados Ridge, Atlantic Ocean: Geol. Soc. America Bull., v. 68, p. 897-912. Ewing, M., G. Carpenter, C. Windisch, and J. Ewing, 1973, Sediment distribution in the oceans: the Atlan- tic: Geol. Soc. America Bull., v. 84, p. 71-88. Freeland, G. L., and R. S. Dietz, 1972, Plate tectonic evolution of the Caribbean-Gulf of Mexico region: 6th Caribbean Geol. Conf., Margarita, Venezuela, Trans., p. 259-264. Funnell, B. M., and A. G Smith, 1968, Opening of the Atlantic Ocean: Nature, v. 219, p. 1328-1333. Heezen, B. C, and M. Tharp, 1961, The physiographic diagram of the South Atlantic, the Caribbean, and the Scotia Sea, and the eastern margin of the South Pacific Ocean: Geol. Soc. America. Isacks, B., J. Oliver, and L. R. Sykes, 1968, Seismology and the new global tectonics: Jour. Geophys. Re- search, v. 73, p. 5855-5900. Kearey, P., 1973, Crustal structure of the eastern Carib- bean in the region of the Lesser Antilles and Aves Ridge: PhD thesis, Univ. Durham, 182 p. G. Peter, and G. K. Westbrook, 1975, Geophysi- cal maps of the eastern Caribbean: Geol. Soc. Lon- don Jour., v. 131, p. 311-321. Lattimore, R. K., P. A. Rona, and O. E. DeWald, 1974, Magnetic anomaly sequence in the central North At- lantic: Jour. Geophys. Research, v. 79, p. 1207-1209. Le Pichon, X., 1968, Sea floor spreading and continen- tal drift: Jour. Geophys. Research, v. 73, p. 3661- 3697. Merrill, G, G. Peter, and W. Sproll, 1973, Seismic re- connaissance east of the Lesser Antilles arc (abs.): Eos (Am. Geophys. Union Trans.), v. 54, p. 326. Meyerhoff, A. A., and H. A. Meyerhoff, 1972, Conti- nental drift, IV: the Caribbean "plate": Jour. 'Geol- ogy, v. 80, p. 34-60. Morgan, J. W., 1968, Rises, trenches, great faults, and crustal blocks: Jour. Geophys. Research, v. 73, p. 1959-1982. Paitson, L., C. H. Savit, D. M. Blue, and W. A. Knox, 1964, Reflection survey at Barracuda fault: Geophy- sics, v. 29, p. 941-950. Peter, G, 1974, Tectonic evolution of the eastern mar- gin of the Caribbean region (abs.): Geol. Soc. Ameri- ca Abs. with Programs, v. 6, p. 910. and G. K. Westbrook, 1974a, Tectonics of the Barbados Ridge and adjacent Atlantic basin (abs.): Eos (Am. Geophys. Union Trans.), v. 55, p. 284. 1974b, Interconnection between the tec- ERL 293-AOML 13, 29 p. C. Schubert, and G. K. Westbrook, 1974, tonic framework of Barbados Ridge and the adjacent Guiana basin (abs.): 7th Caribbean Geol. Conf. Abs., p. 50-51. O. E. DeWald, and B. G Bassinger, 1973b, Car- ibbean Atlantic geotraverse, NOAA-IDOE 1971, Rept. 2, Magnetic data: NOAA Tech. Rept. ERL 288-AOML 12. G. Merrill, and S. Bush, 1973a, Caribbean At- NOAA-IDOE Caribbean Atlantic geotraverse: Geo- times, v. 19, p. 12-15. R. K. Lattimore, O. E. DeWald, and G. Merrill, 1973c, Development of the Mid- Atlantic Ridge east of the Lesser Antilles island arc: Nature, Phys. Sci., v. 245, p. 129-131. Pitman, W. C, III. A. M. Herron, and J. R. Heirtzler, 1968, Magnetic anomalies in the Pacific and sea-floor spreading: Jour. Geophys. Research, v. 67, p. 2069- 2985. and M. Talwani, 1972, Sea floor spreading in the North Atlantic: Geol. Soc. America Bull., v. 83, p. 619-646. Schouten, H., 1974, Magnetic anomalies over fracture zones in the lower magnetic latitudes of the central North Atlantic (abs.): Eos (Am. Geophys. Union Trans.), v. 55, p. 232. Schubert, C, 1974, Seafloor tectonics east of the north- ern Lesser Antilles (abs.): 7th Caribbean Geol. Conf. Abs., p. 62-63. and G. Peter, 1973, Sea floor tectonics west of the Barracuda Ridge (abs.): Eos (Am. Geophys. Union Trans.), v. 54, p. 326. Sclater, J. G., R. N. Anderson, and M. L. Bell, 1971, Evolution of ridges and evolution of the central East- ern Pacific: Jour. Geophys. Research, v. 76, p. 7898- 7915. Stipp, J. J., and F. Nagle, 1974, A geochemical study of the Lesser Antilles island arc: regional distribution of Sr 87/86 initial ratios (abs.): 7th Caribbean Geol. Conf. Abs., p. 65. Talwani, M., 1966, Gravity anomaly belts in the Carib- bean in continental margins and island arcs, in Conti- nental margins and island arcs: Canada Geol. Survey Paper 66-15, p. 177. van Andel, Tj. H., R. P. Von Herzen, and J. D. Phillips, 1971, The Vema fracture zone and the tectonics of transverse shear zones in oceanic crustal plates: Ma- rine Geophys. Researches, v. 1, p. 261-283. Weeks, L. A., R. K. Lattimore, R. N. Harbison, et al, 1971, Structural relations among Lesser Antilles, Venezuela, and Trinidad-Tobago: AAPG Bull., v. 55, p. 1741-1752. Westbrook, G. K„ 1973, Crust and upper mantle struc- ture in the region of Barbados and the Lesser An- tilles: PhD thesis, Univ. Durham, 228 p. 1974a, The structure of the Barbados Ridge and buried trench of the Lesser Antilles subduction zone (abs.): 7th Caribbean Geol. Conf. Abs., p. 71. 1974b, The structure of the Barbados Ridge: 7th Caribbean Geol. Conf. Trans, (in press). 1975, The structure of the crust and upper man- tle in the region of Barbados and the Lesser Antilles: Royal Astron. Soc. Geophys. Jour., v. 43, p. 1-42. M. H. P. Bott, and J. H. Peacock, 1973, The lantic geotraverse, NOAA-IDOE 1971, Rept. 1, Pro- ject introduction — bathymetry: NOAA Tech. Rept. Lesser Antilles subduction zone in the region of Bar- bados: Nature, Phys. Sci., v. 244, p. 18-20. Wills, K. J. A., 1975, The geologic history of southern Dominica and plutonic nodules from the Lesser An- tilles: PhD thesis, Univ. Durham, 244 p. 348 35 Reprinted from: Earth and Planetary Science Letters, Vol. 30, No. 1, 135-142 Earth and Planetary Science Letters, 30(1976) 135-142 © Elsevier Scientific Publishing Company, Amsterdam - Printed in The Netherlands 135 OPENING OF THE RED SEA WITH TWO POLES OF ROTATION * EVAN S. RICHARDSON '<2 and C.G.A. HARRISON ] University of Miami. Rosenstie! School of Marine and Atmospheric Science, Miami, Fla. (USA) NO A A - Atlantic Oceanographic and Meteorological Laboratories, Miami, Fla. (USA) Received August 12, 1975 Revised version received January 5, 1976 Previous studies have shown that the Red Sea was formed by two stages of sea-floor spreading, with a quiescent period in between. We suggest that these two phases have occurred in different directions. The shape of the central trough indicates that the present-day motion is almost E-W, whereas the total opening, deduced from the shape of the coastlines, is NE-SW. If the axial trough has opened in an E-W direction, the earlier stage of opening was in a direction which made the Dead Sea Rift fall along a small circle to the pole of early opening, and hence suggests that the Dead Sea Rift was a transform fault during this early stage. The later movement gives almost pure extension along the Dead Sea Rift, and this should be seen by normal faulting. Available first-motion studies are not precise enough to confirm or deny this hypothesis. 1. Introduction 2. Coast to coast opening It has been suggested that there have been two stages of sea-floor spreading in the Red Sea [1,2]. The latter stage was responsible for the narrow axial trough in the center of the Red Sea, which is approximately 50 miles wide and is believed to have been caused by spreading over the last 3.5 m.y. The earlier stage was responsible for the formation of the bulk of the Red Sea. Between the two stages of spreading there was a period of quiescence, which allowed thick salt deposits to be accumulated. These thick salt deposits indicate that the oceanic crust formed during the earlier stage of spreading is at a greater depth than the crust form- ed during the later stage. The axial trough is caused by the absence of salt accumulations over the younger crust. Bathymetric and magnetic evidence has been published to support this two-stage concept [2]. * Contribution from the University of Miami's Rosenstiel School of Marine and Atmospheric Science, 4600 Rickenbacker Causeway, Miami, Florida 33149. Several poles of rotation for the opening of the Red Sea have been published in the past [1 ,3— 6]. These TABLE 1 Poles of rotation for the Red Sea Pole of rotation Method of calculation lat. long. Refer- ence 32.5° N 22.5° E shear along Dead Sea Rift and [3] Gulf of Aden data 29.0°N 32.0°E fit of coastlines and rates of [4] opening in Red Sea and Dead Sea Rift 36.5°N 18.0°E fit of coastlines and Suez Rift [5] 32.0°N 22.0°E fit of lines 20-30 km seaward [6] of coasts and fit of Danakil Horst 31.5°N 23.0°E fit of lines 52 km seaward of [1] coasts, Gulf of Suez 349 136 poles and the methods by which they were calculated are summarized in Table 1 . An important factor in obtaining a pole of rotation for the Red Sea is whether or not the sea opened from coast to coast. Davies and Tramontini [7| believe that much o\ the Red Sea is underlain by oceanic crust. The\ felt that the most important result to emerge from then work was the clear indication that the Red Sea is underlain by rocks whose seismic velocity is higher than the majority of velocities reported from the continents. They reported a layer with an approxi- mate seismic velocity of 6.63 km/sec which is in good agreement with the oceanic layer 3 average of 6.69 km/ sec. From where their refraction lines end to the shore- line, they were not prepared to generalize except to note that there is. apart from a superficial change in sediment reflection characteristics, no reason to suppose that the structure changes. Magnetic anomalies in the axial trough are lineated parallel to the strike of the topography. The pattern tits the anomaly sequence expected from sea-floor spreading, and it seems absolutely certain that the axial trough has been formed by sea-floor spreading over the past few million years. Recently, Girdler and Styles [2] have shown that at one place outside the axial trough there are also lineated magnetic anomalies. They believe that these have also been caused by sea- floor spreading during an earlier episode, thus confirm- ing the seismic evidence that almost all of the Red Sea is oceanic in nature. Girdler and Styles proposed that there was a period of quiescence between the formation of the main por- tion of the Red Sea, which they thought was formed between 41 and 34 m.y. ago, based on the magnetic time scale of Heirtzler et al. [8], and the axial trough, which was formed during the past 3.5 m.y. Although the timing of the earlier stage of opening seems to be in doubt, the difference in structure and the change in character of the older magnetic anomalies, compared with the younger ones, suggesting a deeper source for the older ones, certainly is strongly indicative of two stages of spreading with quiescence in between. 3. Establishment of the later pole of rotation The establishment of the pole of rotation for the total opening of the Red Sea by fitting together the Fig. 1. Oblique Mercator projection of the Red Sea with projec- tion pole at 36.5° N, 18.0°E. This was the position of the pole of rotation used by McKenzie et al. [5] to describe the coast- line fit on either side of the Red Sea. Lines connecting correspond- ing points on opposite coastlines are horizontal and represent small circles about the pole. The shorter lines connecting corre- sponding points across the axial valley are not horizontal. coastlines gives a pole at 36.5°N, 18.0°E [5]. Fig. 1 is an oblique Mercator projection of the Red Sea region using this point as the projection pole. It can be seen that lines connecting congruent points on each coast- line are horizontal lines, representing small circles about the pole. But we noticed on a detailed chart of the Red Sea prepared by Laughton [9] that the axial trough appeared to have been formed by a more or less E-W separation. On this chart, the western bound- 350 137 ary of the axial trough, which is marked by the 500- fathom contour, can be superimposed on the same contour on the eastern boundary, by an eastward translation. We digitized the 500-fathom margins of the axial valley at 15' intervals of latitude. These mar- gins are plotted on the transverse Mercator projection of Fig. 1 . Congruent points on either side of the axial trough are connected by the short lines, and it can be seen that these lines are not horizontal. This suggests that the axial valley was not formed by relative motion about the pole used to plot Fig. 1 , but about a differ- ent pole. In order to establish a possible pole of rotation for the formation of the axial trough, we traced the western 500-fathom boundary of the axial trough from Laugh- ton's map and slid it eastwards so that it matched the eastern boundary. Since the movement was almost exactly latitudinal, it is possible to do this because the Fig. 2. Oblique Mercator projection of the Red Sea with projec- tion pole at 15.2°S, 32.8° E. This is the pole of rotation calcu- lated for the later stage of opening. The lines connecting con- gruent points on either side of the axial trough (500-fathom isobath) are now horizontal, whereas the lines connecting the coastlines are not. map of Laughton is plotted on a Mercator projection. We were able to measure a vector of opening for the northern end of the axial trough and one for the south- em end. These two vectors enabled us to establish a pole of opening which was at 1 5.2°S, 32.8°E, with a total opening angle of 1 .08° about this pole. This is the pole for the clockwise rotation of Arabia from Nubia (or that part of Africa to the northwest of the East African rift valley). Using this rotation as the pro- jection pole of an oblique Mercator projection (Fig. 2), we can see that now congruent points on either side of the axial valley are horizontal, and therefore lie on small circles about this pole. One problem in establishing a pole for the later spreading episode is that there may have been flow- age of salt deposits into the axial trough, which may therefore be narrower than the amount of sea floor created during the more recent episode of sea-floor spreading. Girdler and Darracott [1 | have also suggest- ed that the later opening may be about a pole different from that describing the early opening, but without going into details concerning what the difference might be. However, in a more recent paper. Girdler and Styles [2] make no mention of different poles of rotation. Girdler and Whitmarsh [10] have found evidence at two DSDP sites in the southern and central Red Sea (sites 227 and 228 in Fig. 3) that there has been lateral flow of salt deposits into the axial trough. At both sites Pliocene sediments and Miocene evaporites were found overlying oceanic crust which was predicted by magnetic anomaly evidence to be younger than 2.5 m.y. Coleman [11] has also mentioned the possi- bility of lateral flow of Red Sea evaporites. He suggest- ed that the irregular bathymetry of the axial trough at 16.67°N may have resulted from salt flowage. At present there is no evidence that there has been similar flowage in the northern Red Sea. Clearly the bathymetric shape of the axial valley is less smooth in the north. If there has been flowage of evaporites in the south but not in the north, this will cause an error in our calculation of the pole of opening of the axial trough. The error will be mainly in the determi- nation of the latitude of the pole and not in the longi- tude. The true pole would lie further to the south than the position given above. If the axial valley in the south has been narrowed by 12 km due to flowage of salt, then the pole of opening should be shifted from a latitude of about 1 5°S to a latitude of about 70°S. 351 138 THE RED SEA EGYPT SUDAN ETHIOPIA Fig. 3. Map of the Red Sea showing the location of three earth- quakes (a, b, and c) [5,13,14) for which fault plane solutions have been obtained, and the locations of two DSDP sites (227 and 228) where it has been suggested that lateral flow of evap- orites has occurred tending to smooth the boundaries of the axial trough [10]. However, since the amount of flowage is difficult to assess quantitatively, we shall assume that the pole at 15.2°S, 32.8°E is correct, but will mention what the effect of moving it further south would be. A better way of establishing the distance to a pole of rotation is to rotate magnetic anomalies on either side of the spreading center into coincidence, as done, for instance, by Pitman and Talwani [12]. Magnetic anomalies represent oceanic crust which is more precise- ly dated than the locus of the thick salt deposits out- lined by the central valley described above. However, the development of magnetic anomalies is generally poor within the Red Sea. Estimates of spreading rate can only be obtained from a rather narrow latitudinal band from about 17°N to 21°N [13-15]. The uncer- tainty in observed spreading rates is so large that no trend from south to north can be seen. All that these data tell us is that the pole of rotation lies sufficiently far away from the Red Sea such that a variation of spreading rate of 20% or more is not produced within this 4° band. This puts the pole of rotation further away than about 20°. Beyond that the magnetic data cannot go. An alternative method of determining the direction of movement is to measure the strike of transform faults offsetting the axis of spreading. However, in the Red Sea we do not believe that there are any features which have been definitely identified as transform faults. Therefore this method of analysis is not available for this portion of the earth. 4. Establishment of early pole We accept the pole of McKenzie et al.(36.5°N, 18.0°E) [5] as a good approximation of the mean pole of opening for the complete history of the Red Sea. However, if a two-stage model is accepted, this pole becomes a resultant of the first and second episodes of spreading. Therefore it is possible to obtain a pole of rotation for the early stage of spreading by vectori- ally subtracting the axial trough pole from the resul- tant or total pole of McKenzie et al. [5]. Employing this method, we have calculated the early pole at 29.6°N, 20.6°E, some 6° south of the total pole open- ing. If there has been flowage of evaporites in the south- ern Red Sea, the later pole will be further south and this will tend to move the early pole to the north. 5. Dead Sea Rift One supposition of plate tectonics is that a trans- form fault always falls on a small circle to the relative pole of rotation of the two plates involved. The Dead Sea Rift has generally been accepted as such a bound- ary between the Arabian and Nubian plates. Fig. 4 shows the pole of McKenzie et al. [5] plotted on a map of the Mediterranean and Red Sea region. It is evident here that the Dead Sea Rift does not fall on a small circle to this pole. However, the Dead Sea Rift does fall quite close to a small circle about the early pole presented in this paper (Fig. 5). We feel that this is more than a fortuitous occurrence. In fact, depending 352 139 /L, sT L-- — l^__-L— -T \ \ \ \^^ \ i \ \ \ \^\ \ 1 Ni -—\-~~\ \ \ 1 \ r \ \ ^-V""^\ ' X h N ^ \ / \L_V-"T^ \ \ ■^ | Fig. 4. Azimuthal equidistant projection of the Mediterranean and Red Sea region with projection pole at 36.5° N, 18.0°E marked with a cross. This is the position of the McKenzie et al. [5] pole. The Dead Sea Rift is marked with plus signs which are located from south to north as follows: south end of the Gulf of Aqaba, north end of the Gulf of Aqaba, south end of the Dead Sea, north end of the Dead Sea, Sea of Galilee. These points do not lie on a small circle with respecj to the rotation pole. ^1 ixTVl ---L \ \ \J\^\ \ —\—L-\- ^ \ vl \ \ ^ ^ \ \ \^-^\ \ jK N^L» \ A-- \ \ \t y — JL 1 r — i \ / \jLr— -\ — \ jN v ^ '-■-- Fig. 5. Same as Fig. 4, except that the projection pole is at 29.6° N, 20.6° E, which is the rotation pole calculated for the early stage of opening of the Red Sea. The points along the Dead Sea Rift now lie close to a small circle about the rotation pole. on the amount of flowage of the evaporites from the main trough, the precise location of the early pole may be slightly north of its calculated position, therefore causing the Dead Sea Rift to fall even closer to a small circle about the pole. By assuming a two-stage model for the development of the Red Sea we may therefore find a key to under- standing the history of the Dead Sea Rift as a plate boundary. The strike of the fault was determined during the first stage of spreading in the Red Sea when most of the motion along the fault consisted of a left- lateral strike-slip component. Then when the second stage of spreading began, the fault took on a compo- nent of rifting. So, today the major source of earth- quakes along the Dead Sea Rift should be caused by normal rather than strike-slip faulting. We suggest that this hypothesis be tested by first motion studies. It should be noted that several authors have suggest- ed that a significant amount of left -lateral shear has occurred along the Dead-Sea Rift during the Quaternary. Quennell [16] inferred a Pleistocene movement of 45 km from geomorphic features, the most prominent of which is the shape of the deep depression of the Dead Sea. Zak and Freund [17] have recorded horizontal dis- placements (which are younger than the Lisan Marl - 23,000 years) of 1 50 m along the fault in the Dead Sea area. However, Freund et al. [18] do not hesitate to admit that a general agreement has not yet been reached concerning the Dead Sea Rift's lateral displace- ment. They refer to Neev and Emery [19] in discussing the geology of the Dead Sea as accepting the shear hypothesis, and Picard [20] as not accepting it. We also refer to Bender [22] who lists several reasons why he does not accept the shear hypothesis. Along the entire east side of the rift, he notes that there is over- whelming evidence of dip-slip movement along hundreds of faults and fault zones with vertical throws of up to 1000 m. He reports that evidence of lateral displace- ment (horizontal slickensides, etc.) is very rare (observ- ed at three places) and in the order of centimeters up to a few meters. Bender suggests that these minor later- al movements and some minor folding due to tangental compression can be explained as secondary structural phenomena. Perhaps the confusion throughout the literature con- cerning movement on the Dead Sea Rift is because both strike-slip and normal faulting have occurred, but at 353 140 different times. We feel that our model of strike-slip and then normal faulting is not inconsistent with the observed data. The Dead Sea Rift follows approximately a small circle about the early pole from its intersection with the Red Sea to as far north as the Huleh Depression in Lebanon. North of this, the Yamuneh Fault trends more or less northeasterly, clearly departing from the small circle. For this reason we postulate that the struc- tural continuation of the Dead Sea transform fault follows the Roum Fault which trends north toward the Mediterranean near Beirut. Dubertret [23] also suggested a similar continuation, but for different reasons. He postulated that motion (strike-slip) be taken up along the Roum Fault rather than along the Yamuneh because the structures in Lebanon are far too gentle to accommodate the amount of shortening necessary to explain the displacement along the Dead Sea Rift to the south. Of course today the Roum Fault is not active, but was probably a continuation of the Dead Sea Rift as a transform fault during the first stage of spreading in the Red Sea. One difficulty with our later pole is that, being south of the Red Sea, a much greater amount of exten- sion would be expected across the Dead Sea Rift than is actually observed. The largest amount of extension appears to be less than 20 km in the Gulf of Aqaba. One possible solution to this problem is that again, depending on the amount of flowage of evaporites in the southern Red Sea, the later pole may be pushed far enough to the south so that its anti-pole may be north of, and not far from the Dead Sea Rift. In this case, a smaller amount of extension along the Dead Sea Rift than in the Red Sea's axial trough would be expected. However, it is also possible that some extension has been taken up in crustal thinning rather than faulting and could not be ascertained by field investigations alone. 6. Red Sea spreading rate Girdler and Styles [2] published a spreading rate for the axial trough (recent episode) of 0.9 cm/yr. This rate was computed from a magnetic profile across the southern Red Sea trending N58°E. However, because we postulate a second-stage direction that is almost E-W, we calculate a new spreading rate. We use the same profile used by Girdler and Styles but have ob- tained a spreading rate of 1 .0 cm/yr. 7. Fault plane solutions Fault plane solutions have been previously publish- ed for three earthquakes that have occurred in the Red Sea [5,24,25]. The locations of these earthquakes are shown in Fig. 3 and given in Table 2, and the solutions are shown in Fig. 6. The earthquake occurring in the northern Red Sea (a in Fig. 3, Fig. 6A and Table 2) occurred at the extreme northern boundary of the axial trough and was generated by a normal fault. Be- cause of the relative paucity of data for this earth- quake, it would be possible to draw the fault plane and the accessory plane so that they were striking approximately N-S, in agreement with what we would expect for a normal fault generated within an axial valley where the motion was E-W. We feel that first motion data from this earthquake are not inconsistent with the motion which we propose is happening in the Red Sea today. Two of the earthquakes, one in the central and one in the southern Red Sea (b and c in Fig. 3, Fig. 6B and C, and Table 2), appear to have occurred on transform faults, although offsets of the axial trough are not evi- dent. However, some disagreement exists between various solutions of these two earthquakes which have appeared in the literature. For the earthquake in the central Red Sea, Fairhead and Girdler [24] have obtain- ed a focal plane with a strike of N53°E and dip 82°SE, whereas McKenzie et al. [5] published a solution for the same earthquake and obtained a focal plane with TABLE 2 First motion studies of earthquakes in the Red Sea Origin time Lat. Long. Ref. Northern Red Sea Central Red Sea Southern Red Sea 31 Mar. 1969 27.5°N 33.8°E [5,24] 13 Mar. 1967 19.7°N 38.9°E [5,24] 11 Nov. 1962 17.1°N 40.6°E [5,24,25] 354 141 Fig. 6. Focal mechanism for three earthquakes in the Red Sea (see Table 2 and Fig. 3). The dashed focal planes are those that have been previously published. A. Solution for earth- quake on March 31, 1969. B. March 13, 1967. C. November 11, 1962. magnitude for the East Africa Rift. We use the same pole of rotation for the Gulf of Aden as presented by McKenzie et al. [5]. We find that this new pole for the East Africa Rift lies at 51 .7°S, 3.5°E and that the magnitude of rotation is 2.65 X 10~7 degree/year. This pole lies farther to the southwest than that presented by Girdler and Darracott [1] but lies in the same gener- al direction with respect to the rift. They refer to seis- mological studies and gravity anomalies indicating a tensional stress field of approximately S30°E across the rift. These geophysical data are as consistent with our pole as they are with theirs. strike N68°E and dip 80° SE, a difference of 15° in the strike. We have no means of knowing which solu- tion is the more accurate, but point out that a slight modification to the solution of McKenzie et al. would give a motion close to the one which we predict. The earthquake in the southern Red Sea has even less data (see Fig. 6), and a slight modification to the fault planes would give good agreement to our predict- ed motion. In this case, the fault would be a right- lateral fault along the approximately E— W plane, rather than the left-lateral motion along the approximately N— S plane, as proposed by McKenzie et al. [5]. We conclude that the first motion studies of these three earthquakes are sufficiently inaccurate that it is impossible to use them to decide whether our proposed E— W motion is more correct than the NE— SW motion derived from the total opening of the Red Sea. The first motion studies fit either hypothesis equally well. 9. Summary We have shown that the shape of the axial valley of the Red Sea suggests that recent spreading has been in an E-W direction, rather than the NE-SW direc- tion of the earlier phase of spreading. If this is the case, then the Dead Sea Rift was a plate boundary with almost pure transform fault motion along it during the earlier phase of spreading. According to our model, present-day motion of the Dead Sea Rift is extensional, and should be marked by predominantly normal fault- ing. The available earthquake evidence is not capable of distinguishing between the E— W motion suggested in this paper and the motions suggested previously. The rotation pole position calculated for the later phase of spreading is probably not accurate, because of salt flowage. However, it can be taken as the northernmost limit of the true rotation pole. 8. Implications for the East Africa Rift McKenzie et al. [5] and Girdler and Daracott [1] have both made use of three-dimensional vector addi- tion to arrive at a pole and magnitude of rotation for the East Africa Rift system. This is possible because the rift is part of a three-plate spreading system; and by vectorially adding the poles and magnitudes of rota- tion for the Red Sea (Arabia— Nubia) and the Gulf of Aden (Arabia— Somalia), a pole and magnitude of rotation for the East Africa Rift (Nubia-Somalia) may be obtained. Because we have obtained a new pole of rotation for the Red Sea, we can also calculate a new pole and Acknowledgments We thank M.M. Ball, R.S. Dietz and F. Nagle for con- structive criticisms. The maps in Figs. 1,2,4 and 5 were drawn using FORTRAN program HYPERMAP, written by R.L. Parker, whom we thank. Research supported by NSF grant GA-42979 from the oceanog- raphy section and from NOAA. References 1 R. Girdler and B. Darracott, African poles of rotation, Comments Earth Sci.: Geophys. 2 (1972) 131-138. 355 142 2 R. Girdler and P. Styles, Two-stage Red Sea floor spread- ing, Nature 247 (1974) 7-11. 3 A. Laughton, The birth of an ocean, New Sci. 27 (1966) 218-220. 4 D. Roberts, Structural evolution of the rift zones in the Middle East, Nature 223 (1969) 55-57. 5 D. McKenzie, D. Davies and P. Molnar, Plate tectonics of the Red Sea and East Africa, Nature 226 (1970) 1-6. 6 R. Freund, Plate tectonics of the Red Sea and East Africa, Nature 228 (1970) 453. 7 D. Davies and C. Tramontini, The deep structure of the Red Sea, Philos. Trans. R. Soc. Lond.,Ser., A 267 (1970) 181-189. 8 J.R. Heirtzler, CO. Dickson, E.M. Herron, W.C. Pitman III and X. Le Pichon, Marine magnetic anomalies, geo- magnetic field reversals, and motions of the ocean floor and continents, J. Geophys. Res. 73 (1968) 2119-2136. 9 A.S. Laughton, A new bathymetric chart of the Red Sea, Philos. Trans. R. Soc. Lond., Ser. A 267 (1970) 21-22. 10 R. Girdler and R. Whitmarsh, Miocene evaporites in Red Sea cores and their relevance to the problem of the width and age of oceanic crust beneath the Red Sea, in: Initial Reports of the Deep Sea Drilling Project 23 (1974)913- 921. 11 R. Coleman, Geologic background of the Red Sea, in: Initial Reports of the Deep Sea Drilling Project 23 (1974) 813-819. 12 W.C. Pitman III and M. Talwani, Sea-floor spreading in the North Atlantic, Geol. Soc. Am. Bull. 83 ( 1972) 619-646. 13 T.D. Allan, Magnetic and gravity fields over the Red Sea, Philos. Trans. R. Soc. Lond., Ser. A 267 (1970) 153-180. 14 J.D. Phillips, Magnetic anomalies in The Red Sea, Philos. Trans. R. Soc. Lond., Ser. A 267 (1970) 205-217. 15 F.J. Vine, Spreading of the ocean floor: new evidence, Science 154 (1966) 1405-1415. 16 A.M. Quennell, The structural and geomorphic evolution of the Dead Sea Rift, Q.J. Geol. Soc. Lond. 114 (1958). 17 1. Zak and R. Freund, Recent strike slip movements along the Dead Sea Rift, Israel J. Earth Sci. 15 (1966) 33-37. 18 R. Freund, Z. Garfunkel, I. Zak, M. Goldberg, T. Weissbrod and B. Denn, The shear along the Dead Sea Rift, Philos, Trans. R. Soc. Lond., Ser. A 267 (1970) 107-130. 19 D. Neev and K.O. Emery, The Dead Sea, depositional processes and environments of evaporites, Bull. Geol. Surv. Israel 41 (1967) 147. 20 L. Picard, Thoughts on the graben system in the Levant, Geol. Surv. Can. Paper 66-14 (1966) 22-32. 21 L. Picard, On the structure of the Rhinegraben with com- parative notes on Levantgraben features, Israel Acad. Sci. Hum. 9 (1968) 34. 22 F. Bender, The shear along the Dead Sea Rift: discussion, Philos. Trans. R. Soc. Lond., Ser. A 167 (1970) 127-129. 23 L. Dubertret, Remarques sur le fosse de la Mer Morte et ses prolongements au nord jusqu'au Taurus, Rev. Geogr. Phys. Geol. Dyn. 9 (1967) 3-16. 24 D. Fairhead and R. Girdler, The seismicity of the Red Sea, Gulf of Aden and Afar Triangle, Philos. Trans. R. Soc. Lond., Ser. A 267 (1970) 49-74. 25 L. Sykes, local mechanism solutions for earthquakes along the world rift system, Bull. Seis. Soc. Am. 60 (1970) 1749- 1752. 356 36 Reprinted from: Earth and Planetary Science Letters, Vol. 30, No. 2, 173-175, Earth and Planetary Science Letters, 33(1976)173 175 © Klsevicr Scientific Publishing Company, Amsterdam Printed in The Netherlands 173 [4] OPENING OF THE RED SEA WITH TWO POLES OF ROTATION REPLY E.S. RICHARDSON K2 and C.G.A. HARRISON ' 1 Rosensticl School oj Marine and Atmospheric Science, University of Miami, Miami. Fla. (USA) 2NUAA - Atlantic Oceanographic and Meteorological Laboratories, Miami, Fla. (USA) Received August 2. 1976 We welcome the comments by Girdler and Styles 1 1 | concerning our recent paper [2]. However, the reason for their first argument evades us. Girdler and Styles state that the bathymetry of the Red Sea's axial trough is wider in the south than in the north, and that the extension in the Dead Sea Rift is even less than that in the Red Sea. However, it is obvious that the axial trough from Laughton's bathymetric chart of the Red Sea [3] (de- fined by the 500-fm contour) is narrower in the south- ern Red Sea than in the north. (We have examined some of the seismic reflection profiles published in the lit- erature [4,5] and have found that in the majority of profiles, the slope break on the walls of the axial trough occurs quite close to the 500-fm depth). With this in mind, we were not surprised to calculate a pole of rotation for the later stage of opening to the south of the Red Sea. Two criteria, however, caused us to doubt the lati- tudinal accuracy of this pole: ( 1 ) The Dead Sea rift has undergone a maximum amount of extension of less than 20 km (Gulf of Aqaba). (2) Girdler and Whitmarsh [6] and Coleman [7] have reason to believe that lateral flow of evaporites into the axial trough has occurred in the southern Red Sea, causing the axial trough to be narrower than the amount of sea floor created during the later stage of spreading. These two lines of evidence point to a pole of ro- tation to the north of the Dead Sea Rift. And we made note of this in our orginal paper, on page 1 40 (which Girdler and Styles obviously did not see). We stated that depending on the amount of evapo- rite flowage into the axial trough, the later pole may be pushed far enough to the south so that its anti-pole may be north of, and not far from, the Dead Sea Rift. Our example was that, if the axial valley in the south has been narrowed by 12 km due to salt flow- age, the rotation pole should be shifted from a lati- tude of about 1 5°S to a latitude of about 70°S. This is only an example, as Girdler and Whitmarsh [6] show evidence that the narrowing of the axial trough. in the south is much more than 12 km. They note the presence of Miocene evaporites overlying oceanic crust with a magnetic age of 2.5 m.y. Assuming an average sea-floor spreading rate of 1.0 cm/yr. we esti- mate that the flowage of evaporites is close to 25 km (12 km on either side). This would push the pole of rotation so far south that its antipole would approach 55°N. Girdler and Styles [1] object to our description of the motion of Arabia away from Nubia as clockwise. We wish simply to point out that an anticlockwise rotation about a pole requires a clockwise rotation about its anti-pole. Since the pole we describe is lo- cated south of the Red Sea, it would naturally require a clockwise rotation to moce Arabia away from Nubia. The situation is illustrated in Fig. 1. The other major point to which Girdler and Styles address themselves is that of northeast-southwest fea- tures within the axial trough. They go as far as to mea- sure 67 azimuths from magnetic, gravity, bathymetry, and interpretation maps to prove their point of recent northeast-southwest motion. However, they note that due to the small width of the axial trough their mea- surements "are not very accurate" and "give rise to large errors". We agree with this conclusion. In addi- 357 174 Or RED SEfl Fig. 1 . Movement of pole of rotation to account for 25 km of salt flowage. Note that anti-pole moves to a position north of the Red Sea. Rotation of Arabia from Nubia is clockwise about the pole, when the motion is viewed from outside the earth, but anti-clockwise about the anti-pole. tion, we note that none of the data referred to gives conclusive evidence of transform motion within the axial trough except the fault-plane solutions from Fairhead and Girdler [8]. Here, strike-slip motion is indicated. But because of scanty data, a precise azi- muth of the nodal planes cannot be determined, as already discussed in our paper. We have checked care- fully all the references used by Girdler and Styles [1] to make their measurements of azimuths, except that by Backer et al. [9]. To our surprise we found that in most cases the original authors made no suggestion that these features were transform faults. Searle and Ross [5] did suggest that the magnetic anomalies stud- ied by them could be best explained by northeast- southwest motion, but other interpretations are also possible. Phillips [10] suggested three possible models for the magnetic anomalies he studied in the Red Sea. He was unable to choose between these models ex- cept on the basis of other evidence for directions of motion. This other evidence was the direction estab- lished from earthquake first-motion studies, which we have already discussed in our paper, and which we have suggested do not make great constraints on the actual motion because of the rather poor recording of earthquakes in this region. One of the models suggest- ed by Phillips [10] was one in which there was east- west motion between Arabia and Nubia, in agreement with our model. Several authors have made note of magnetic linea- tions striking N60°E and N70°E [5,10-12]. In fact, Allan [13] makes special mention of five earthquake epicenters which are aligned in an east-west direction and show remarkable coincidence with his postulated offsets in the axial trough. He states that this is con- vincing proof of a tranform fault in this region. Girdler and Styles [14] suggest that there was a ces- sation in spreading in the Red Sea of about 30 m.y. Even though the later stage of spreading took place at the same geographic location as the earlier stage (in the center of the Red Sea) it would be surprising in- deed if the new direction of spreading were along ex- actly the same azimuth as the old direction, as the direction of spreading is controlled by processes un- derlying the lithosphere. If the processes ceased for 30 m.y., there is no reason to believe that the move- ments woul regenerate in the same direction as be- fore. Our conclusions are that the shape of the axial trough suggests an east-west movement of Arabia away from Africa (as shown by super-position of the 500-fm contours). Problems of salt flowage, however, preclude the calculation of an accurate latitude for the pole of rotation, whereas the meridian of the rota- tion pole is much more accurately known. Research supported by the National Science Foun- dation and by NOAA. References 1 R. Girdler and P. Styles, Opening of the Red Sea with two poles of rotation - some comments, Earth Planet. Sci. Lett. 33 (1976) 169-172. 2 E.S. Richardson and C.G.A. Harrison, Opening of the Red Sea with two poles of rotation, Earth Planet. Sci. Lett. 30 (1976) 135-142. 3 A.S. Laughton, A new bathymetric chart of the Red Sea, Philos. Trans. R. Soc. Lond., Ser. A, 267 (1970) 21-22. 4 J.D. Phillips and D.A. Ross, Continuous seismic reflexion profiles in the Red Sea, Philos. Trans. R. Soc. Lond., Ser. A, 267 (1970) 143-152. 5 R.C. Searle and D.A. Ross, A geophysical study of the Red Sea axial trough between 20° S and 22° N, Geophys. J.R. Astron. Soc. 43 (1975) 555-572. 6 R. Girdler and R. Whitmarsh, Miocene evaporites in Red Sea cores and their relevance to the problem of the width and age of oceanic crust beneath the Red Sea, in: Initial Reports of the Deep Sea Drilling Project 23 (U.S. Govern- ment Printing Office, Washington, D.C., 1974) 913-921. 358 175 R. Coleman, Geologic background of the Red Sea, in: Ini- tial Reports of the Deep Sea Drilling Project 23 (U.S. Government Printing Office, Washington, D.C., 1974) 813-819. D. Fairhead and R. Girdler, The seismicity of the Red Sea, Gulf of Aden, and Afar Triangle, Philos. Trans. R. Soc. Lond., Ser. A, 267 (1970) 49-74. H. Backer, K. Lange and H. Richter, Morphology of the Red Sea Central Graben (Valdivia Enzschlamme A & B, Preussag). J.D. Phillips, Magnetic anomalies in the Red Sea, Philos. Trans. R. Soc. Lond., Ser. A, 267 (1970) 205-217. 1 1 J.D. Phillips, J. Woodside and CO. Bowin, Magnetic and gravity anomalies in the Central Red Sea, in: Hot Brines and Recent Heavy Metal Deposits in the Red Sea, E.T. Degens and D.A. Ross, eds. (Springer-Verlag, New York, N.Y., 1969)98-113. F.K. Kabbani, Geophysical and structural aspects of the central Red Sea valley, Philos. Trans. R. Soc. Lond., Ser. A, 267 (1970) 89-97. T.D. Allan, Magnetic and gravity fields over the Red Sea, Philos. Trans. R. Soc. Lond., Ser. A, 267 (1970) 153-181 R.W. Girdler and P. Styles, Two-stage Red Sea floor spreading, Nature 24 7 ( 1 9 74 ) 7 1 1 . 12 13 14 359 37 Reprinted from: Earth and Planetary Saienae Letters, Vol. 30, No. 1, 74-75. Earth and Planetary Science Letters, 30 ( 1 976) 109-116 © Elsevier Scientific Publishing Company, Amsterdam - Printed in The Netherlands 109 [6] ASYMMETRIC FRACTURE ZONES AND SEA-FLOOR SPREADING PETER A. RON A NOAA, Atlantic Oceanographic and Meteorological Laboratories, Miami, Fla. (USA) Received August 12. 1975 Revised version received January 29, 1976 An asymmetric pattern is observed in the orientation of minor fracture zones about the axis of the Mid-Atlantic Ridge at five sites where relatively detailed studies have been made between latitudes 22°N and 51°N. The minor fracture zones intersect the axis of the Mid-Atlantic Ridge in an asymmetric V-shaped configuration. The V's point south north of the Azores triple junction (38°N latitude) and point north south of that junction. The rates and directions of sea-floor spreading are related to the asymmetric pattern of minor fracture zones at the sites studied. Half-rates of sea-floor spreading averaged between about 0 and 10 m.y. are unequal measured per- pendicular to the ridge axis. The unequal half-rates of spreading are faster to the west north of the Azores triple junction and faster to the east south of that junction. The half-rates of sea-floor spreading calculated in the directions of the asymmetric minor fracture zones are equal about the ridge axis within the uncertainty of the direction deter- minations. A discrepancy exists between minor fracture zones that form an asymmetric V about the axis of the Mid-Atlantic Ridge, and major fracture zones that follow small circles symmetric about the ridge axis. To reconcile this discrepancy it is proposed that minor fracture zones are preferentially reoriented under the influence of a stress field related to interplate and intraplate motions. Major fracture zones remain symmetric about the Mid-Atlantic Ridge under the same stress field due to differential stability between minor and major structures in oceanic lithosphere. This inter- pretation is supported by the systematic variation in the orientation of minor fracture zones and the equality of sea- floor spreading half-rates observed about lithospheric plate boundaries. 1. Introduction A discrepancy is becoming apparent between the overall symmetry of the Atlantic ocean basin and asym- metry of both topography and sea-floor spreading about the Mid-Atlantic Ridge. The overall symmetry of the Atlantic (Fig. 1) was first recognized from bathymetric profiles along widely spaced tracklines that revealed the nearly median position of the Mid-Atlantic Ridge [1], the nearly mirror image distribution of physiograph- ic provinces about the ridge axis [1], the trajectories of major fracture zones which follow small circles symmetric about the axis of the Mid-Atlantic Ridge [2,3], and the sequences of remanent magnetic anom- alies attributed to polarity reversals that indicate a grossly similar history of sea-floor spreading in the eastern and western basins [4,5]. Recent work summarized in Table 1 reveals asym- metry of both topographic features (Fig. 1) and half- rates of sea-floor spreading about the axis of the Mid- Atlantic Ridge at sites where relatively detailed in- vestigations have been made between major fracture zones. A problem exists in reconciling the overall symmetry of the Atlantic ocean basin with the asym- metry revealed by the recent work. 2. Asymmetric topography of the oceanic ridge crest Valleys with intervening ridges oriented transverse to the rift valley have been delineated where relatively detailed bathymetric surveys have been made at sites along the crest of the Mid-Atlantic Ridge (Fig. 1). The spacing between the transverse valleys ranges between 360 110 TABLE 1 Relation between topography and sea-floor spreading on the Mid-Atlantic Ridge Reference Topography Azimuth of transverse valleys Average Sense of off- and ridges spacing be- set of axis of bounding latitude on azimuth of tween trans- MAR at trans- lithospheric plates MAR axis of MAR Side of MAR verse valleys (km) verse valleys W E [33] America and 61-62°N 033° 280° _ 30 left lateral Eurasia 61-62°N 033° 280° - 30 left lateral 61-62°N 033° 280° - 30 left lateral [30] America and Eurasia 47-51°N 335 and 360° 280-295° 080- 090° 30 left lateral 47-51°N 335 and 360° 280-295° 080- 090° 30 left lateral [10,23,30] America and 45_46°N 019° 285 ± 10° 087 ± 10° 30 left lateral Eurasia 45-46°N 019° 285 ± 10° 087 ± 10° 30 left lateral left lateral Azores triple function [8,9,11, America and 36-37°N 018° 270 ± 10° 108° 50 right lateral 24,25] Africa 36-37°N 018° 270 ± 10° 108° 50 right lateral 36-37°N 018° 270 ± 10° 108° 50 right lateral [6,7,26,27] America and 25-27°N 025° 265° 115° 55 right lateral Africa 25-27°N 025° 265° 115° 55 right lateral 25-27°N 025° 265° 115° 55 right lateral [12] America and 22-23°N 020° 260 ± 10° 110 ± 10° 50 left lateral Africa 22-23°N 020° 260 ± 10° 110 ± 10° 50 left lateral 22-23°N 020° 260 ± 10° 110 ± 10° 50 left lateral [28,34] America and 6-8°S 350° 231 ± 10° 080c — left lateral Africa 6-8°S 350° 231 ± 10° 080° - left lateral 6-8°S 350° 215 ± 10° 080° - left lateral 6-8°S 350° 215 ± 10° 080° - left lateral MAR = Mid-Atlantic Ridge; - indicates no data. Italicized values are computed values (this paper). 30 and 55 km at the various sites (Table 1). The trans- verse valleys and intervening ridges are distinctly de- lineated by those surveys that include tracklines at spacings closer than 20 km oriented parallel to the axis of the Mid- Atlantic Ridge, such as the surveys at 26°N (Fig. 2) [6,7] , and at 36°N [8] . The transverse features are less distinctly delineated by surveys based only on tracklines oriented perpendicular to the axis of the Mid-Atlantic Ridge, as at the other sites (Table 1). Where distinctly delineated at 26°N (Fig. 2) [6,7] and at 36°N [8,9], the transverse valleys exhibit the characteristics of minor fracture zones associated with small ridge-ridge transform faults. These characteristics include ridge-ridge offset of the rift valley up to about 20 km, the association of earthquake epicenters with the zone of offset, the presence of a basin several hundred meters deep at the intersection of the zone of offset with the rift valley, and relief of hundreds of meters between the floors of the transverse valleys and the crests of the intervening ridges. It is inferred by analogy with their characteristics at 26°N and 36°N that the transverse valleys and intervening ridges at the 361 Ill Amount of offset of ,i\is o\ MAR .it transverse valleys (km) Ace o\ crust (m.y. B.P.) Sea-floor spreading average half- averaging side of rate of spread- interval MAR ing (cm/yr) (m.y.) Azimuth of spreading direction (relation to axis of MAR) <10 <10 <10 <10 <10 3-7 3-7 0-10 0-10 1.0 /./ 1.10 0-10 0-10 0-10 0-10 0-10 vv w E E w 303° (normal) 280° (oblique) 123° (normal) 070° (normal) 250° (normal) <10 <10 <10 0-10 0-10 1.10 1.28 1.19 ± 0.10 0-10 0-10 0-10 E w E 109° (normal) 289° (normal) 087 ± 10° (oblique) 20 20 20 <10 <10 <10 <10 <10 <10 0-10 0-10 0-10 0-10 0-10 0-10 0-10 0-10 0-10 0-4.5 0-4.5 4.5-10 4.5-10 1.3 1.0 ± 0.1 /./ ± 0.2 1.3 1.1 1.3 7.5 1.4 1.5 0.1 0.1 2.16 ± 0.24 1.89 ± 0.04 1.59 ± 0.24 1.12 ± 0.08 0-10 0-10 0-10 0-10 0-10 0-10 0-10 0-10 0-10 0-4.5 0-4.5 4.5-10 4.5-10 E W w E W W E W w E W E W 108° (normal) 288° (normal) 270 ± 10° (oblique) 115° (normal) 295° (normal) 265° (oblique) 110° (normal) 280° (nearly normal) 260 ± 10° (oblique) 080° (normal) 260° (normal) 080° (normal) 260° (normal) other sites described (Table 1) are also minor fracture zones associated with small ridge-ridge transform faults. The transverse valleys and intervening ridges at each site intersect the east and west sides of the rift valley in an asymmetric V-shaped configuration (Fig. 1; Table 1). The two sides of the rift valley are parallel. North of the Azores triple junction between 45°N and 46°N the angle of intersection of the transverse valleys and intervening ridges with the rift valley is oblique on the east side and nearly normal on the west side [10]. Between 47°N and 51°N the transverse valleys and intervening ridges appear to retain the same orienta- tion as between 45°N and 46°N, but the azimuth of the axis of the Mid-Atlantic Ridge changes from north- east to northwest resulting in oblique intersections on both sides of the rift valley. At both sites north of the Azores triple junction the V formed by the intersection of the transverse valleys and intervening ridges with the two sides of the rift valley points southward (Fig. 1). South of the Azores triple junction at 36°N (Table 1) [8,9,1 1 ], at 26°N (Fig. 1 ; Table 1 ) [6,7] , and at 22°N [12], the angle of intersection of the transverse valleys 362 112 80° W 70° 60° 50° 40° 30° 20° WW 0° 60" N - - 60° N - 10° N 10°S 80° W 70 30° 20° 10° W 0° Fig. 1. Map of the Atlantic ocean basin showing principal litho- spheric plates, axis of the Mid-Atlantic Ridge, major fracture zones that form small circles symmetric about the ridge axis, minor fracture zones that form V-shaped configurations asym- metric about the ridge axis delineated in areas of relatively de- tailed investigations (boxes). The configuration of inferred minor fracture zones in the area at 6°S is predicted rather than observed, as discussed in the text. and intervening ridges with the rift valley reverses be- coming nearly normal on the east side and oblique on the west side. The V formed by the intersection of the transverse ridges and intervening valleys at the sites south of the Azores triple junction points northward (Fig. 1). The characteristics of the minor fracture zones de- scribed are distinct in several respects from major frac- ture zones. Major fracture zones of the Atlantic like the Gibbs (latitude 52°N) [13], the Oceanographer (latitude 35°N) [14], the Atlantis (latitude 30°i\ J [15], the Kane (latitude 24°N) [16], and the Vema (latitude 10°N) [17], follow families of small circles symmetric about the axis of the Mid-Atlantic Ridge, generally ex- hibit ridge-ridge offsets of at least 100 km, and are spaced hundreds of kilometers apart along the ridge axis. 3. Apparent and true relative rates of sea-floor spread- ing Determination of rate and direction ol sea-floor spreading are related in that the apparent and true relative rates of spreading are a function ol direction. The principle method to determine spreading rate is based on the Vine and Matthews hypothesis [18]. Strips of crustal material that are alternately magnetized dur- ing spreading about an oceanic ridge are identified m the magnetic polarity reversal time scale. Relative hall-rates of sea-floor spreading may be derived from the distance between the axis of the oceanic ridge and the identified magnetic anomaly. The distance measured perpendicular to the axis of the oceanic ridge yields an apparent rela- tive half-rate of spreading. The distance measured par- allel to transform faults and their continuation as frac- ture zones that may be oblique to the axis of the ocean- ic ridge, yields a true relative half-rate ol spreading, because these features indicate the true direction of relative motion between diverging lithospheric plates [3,19-21 ]. In the case that the direction of a fracture zone is perpendicular to the axis of an oceanic ridge the apparent and true relative half-rates of spreading are equal. The apparent relative half-rates of sea-floor spread- ing determined perpendicular to the axis of the Mid- Atlantic Ridge at the sites studied are unequal (Table 1 ). To facilitate comparison between sites and to suppress shorter period variations [22] the spreading rates are averaged over the period 0-10 m.y. B.P. The average apparent relative half-rates of spreading are faster to the west of the Mid-Atlantic Ridge axis at latitude 45°N north of the Azores triple junction [23] , and are faster to the east at latitudes 36°N, 26°N, and 6°S south of that junction [24—28], The average true relative half-rates of sea-floor spreading in the directions of the minor fracture zones were calculated from the average apparent relative half-rates perpendicular to the ridge axis using simple trigonometric relations (Fig. 3). At latitude 26°N where the azimuths of the minor fracture zones are accurately known (Fig. 2; Table 1 ), the average true relative half- spreading rates are equal about the ridge axis. At the other sites at latitudes 45°N, 36°N, and 22°N, where the azimuths of the minor fracture zones are less ac- curately known (Table 1 ). the average true relative half-spreading rates are equal about the ridge axis 363 113 27° al46°00 W N 45°00 W 44°00 W 27c 46°00 W 45°00 W 44°00 W Fig. 2. Bathymetric map [7] contoured in hundreds of meters of a 180-km square on the Mid-Atlantic Ridge crest at 26°N latitude (Fig. 1; Table 1). Sounding tracks are dashed. Depths exceeding 3400 m are shaded to delineate the rift valley and transverse valleys that trend normal to the east side and oblique to the west side of the rift valley, as sketched in the inset. within the uncertainty of the azimuth determinations. The five sites described are the only sites known in sufficient detail to reveal the systematic variation in orientation of minor fracture zones and the equality of average true relative half-rates of sea-floor spreading about the Mid-Atlantic Ridge. If the relations observed between the orientation of minor fractures zones and half-rates of spreading are consistent, then the orienta- tions of minor fracture zones may be computed from half-rates of spreading. For example, at the Mid-Atlantic crest between latitudes 6° and 8°S the apparent relative half-rates of spreading are known [28] , and orienta- tions of minor fracture zones are unknown. The pre- dicted orientations of the minor fracture zones are computed from the apparent relative half-rates of spread- ing (Table 1). Detailed studies (line spacing closer than 364 114 Fig. 3. Geometry of spreading normal and oblique to the axis of an oceanic ridge, to = zero isochronal = isochron at unit time; /] and l^ = lengths of crust generated in t\ , normal to the axis of an oceanic ridge; rx and r^ = half-rates of spreading corresponding to l\ and /j ; '3 = length of crust generated in t\ along a direction defined by angle a oblique to the axis of the ridge; r$ = half-rate of spreading corresponding to 1$. An angle a exists such that: I3 = /2/cos a = /) and r$ = rj/cos a = rx . 20 km both perpendicular and parallel to the ridge axis) are needed at more sites along the Mid-Atlantic Ridge to test these relations. 4. Discussion Hypotheses to explain the observations presented of fracture zones and sea-floor spreading must consider the various characteristics described. In particular, the asymmetric V-shaped intersection of the inferred minor fracture zones with the axis of the Mid-Atlantic Ridge, the inversion of the V north and south of the Azores triple junction, the inequality of apparent relative spreading half-rates determined perpendicular to the ridge axis, the equality of true relative spreading half- rates determined parallel to minor fracture zones nor- mal and oblique to the ridge axis, and the existence of asymmetric features within the symmetric frame- work of the Atlantic ocean basin. Two alternative hypotheses are considered to ac- count for the observed relations between topography and sea-floor spreading, as follows: (1) Original orientation. The asymmetric orienta- tion of minor fracture zones about the axis of an oce- anic ridge is produced by asymmetric processes of development of the oceanic lithosphere. According to this hypothesis the relative motions of the litho- spheric plates follow the directions of the asymmetric minor fracture zones. This hypothesis poses problems in reconciling asymmetric with symmetric features of the ocean basin because asymmetric plate motions at minor fracture zones would be incompatible with symmetric plate motions at major fracture zones. (2) Reorientation. The processes of development of oceanic lithosphere are essentially symmetric and pro- duce both symmetric minor and major fracture zones associated with symmetric sea-floor spreading. The minor fracture zones are continuously reoriented while the major fracture zones maintain their original orienta- tions. As a consequence of this reorientation apparent relative half-rates of spreading determined perpendicu- lar to the axis of an oceanic ridge are unequal. True relative half-rates of spreading determined in the direc- tions of the reoriented minor fracture zones normal and oblique to the axis of an oceanic ridge are equal. This hypothesis is supported by the relations between spreading directions and rates determined at sites along the Mid-Atlantic Ridge (Table 1), and offers promise of reconciling the discrepancy between asym- metric and symmetric features of the Atlantic ocean basin. The continuous reorientation of minor fracture zones according to hypothesis 2 may be caused by the applica- tion of an external stress field deriving from different sources, as follows: ( 1 ) Forces related to magma tic processes. These forces are related to vertical and horizontal magmatic movements associated with the axial region of an oce- anic ridge. A type of regional magmatic movement proposed by Vogt [29] and applied by Johnson and Vogt [30] to account for V-shaped topography about the axis of an oceanic ridge depends on the principle of a geopotential gradient to drive asthenospheric flow from topographic highs over inferred mantle plumes such as at Iceland and the Azores. According to their hypothesis, the V should point in the direction of flow away from the high as the result of astheno- spheric flow along and sea-floor spreading about an oceanic ridge. The Vogt-Johnson hypothesis does not account for the orientation of the V-shaped topography 365 15 described to the north and south of the Azores because the V points toward rather than away from the Azores (Fig. 1 ). Forces related to magmatic processes un- doubtedly contribute to the stress field, but are con- sidered secondary rather than primary components. (2) Forces rehired re tectonic processes. These forces are related to interplate and intraplate motions and may be primary components of the stress field in- ferred to be reorienting the direction of minor fracture zones and sea-floor spreading along the Mid-Atlantic Ridge. The role oi' interplate and intraplate forces as primary components o[' the stress field is supported by the observation that the orientation of the minor frac- ture zones and of sea-floor spreading systematically changes about lithospheric plate boundaries. The orientation of minor fracture zones and of sea-floor spreading is different on the two sides of the rift valley of the Mid-Atlantic Ridge, a divergent plate boundary, and differs between the America and Eurasia plates north of the Azores triple junction and the America and Africa plates south of that junction (Fig. 1 ; Table 1). The Azores triple junction has been a major in- fluence in the development of the Atlantic at least since the early Mesozoic opening of the central North Atlantic [31], 5. Differential stability of symmetric and asymmetric structures in oceanic lithosphere The reorientation hypothesis allows the simultaneous development of small asymmetric structures and large symmetric structures in oceanic lithosphere. Minor fracture zones associated with small transform faults (offset <30 km) behave in an unstable manner at the relatively slow average half-rates of spreading (<2 cm/yr) prevalent at the Mid-Atlantic Ridge. The minor fracture zones are continuously reoriented under the influence of an external stress field as they are gen- erated by sea-floor spreading about the small transform faults. Major fracture zones associated with large trans- form faults (offset >50 km) behave in a stable manner at relatively slow average half-rates of spreading (<2 cm/yr). The major fracture zones maintain their orien- tation under the influence of the same external stress field as they are generated by sea-floor spreading about the large transform faults. Thickness of lithosphere re- lated to distribution of isotherms at a transform fault may be a determinant of the stability of fracture zones [32]. Asymmetric small structures may then develop within the overall symmetry of the Atlantic ocean basin as a consequence of the differential stability between minor and major fracture zones of the oceanic lithosphere. Acknowledgement 1 thank Walter C. Pitman, 111, for a helpful review. References 1 B.C. Heezen, M. Tliarp and M. Ewing, The floors of the oceans, I. The North Atlantic, Geol. Soc. Am. Spec. Paper 65 (1959) 122 pp. 2 B.C. Heezen and M. Tharp, Physiographic diagram of the North Atlantic Ocean, Geol. Soc. Am. Spec. Paper 65 (1968) revised. 3 W.J. Morgan, Rises, trenches, great faults and crustal blocks, J. Geophys. Res. 73 (1968) 1959. 4 W.C. Pitman, III and M. Talwani, Sea-floor spreading in the North Atlantic, Geol. Soc. Am. Bull. 83 (1972) 619. 5 G.O. Dickson, W.C. Pitman, III and JR. Heirtzler, Mag- netic anomalies in the South Atlantic and ocean-floor spreading, J. Geophys. Res. 73 (1968) 2087. 6 P. A. Rona, R.N. Harbison, B.G. Bassinger and R.B. Scott, Asymmetrical bathymetry of the Mid-Atlantic Ridge at 26°N latitude (abstract), EOS Trans. Am. Geophys. Union 54 (1973) 243. 7 P. A. Rona, R.H. Harbison, B.G. Bassinger, R.B. Scott and A.J. Nalwalk, Tectonic fabric and hydrothermal activity of Mid-Atlantic Ridge crest (26°N), Geol. Soc. Am. Bull. (1976) in press. 8 R.S. Detrick, J.D. Mudie, B.P. Luyendyk and K.C. Mac- donald, Near-bottom observations of an active transform fault (Mid-Atlantic Ridge at 37°N), Nature 246 (1973) 59. 9 I. Reid and K. Macdonald, Microearthquake study of the Mid-Atlantic Ridge near 37°N, using sonobuoys, Nature 246(1973) 88. 10 P.J. Bhattacharyya and D.I. Ross, Mid-Atlantic Ridge near 45°N, computer interpolation and contouring of bathy- metry and magnetics, Marine Sciences Directorate, Dep. of Environment, Ottawa, Mar. Sci. Paper 11 (1972) 9 pp. 1 1 H. Fleming, Naval Research Laboratory, personal com- munication. 12 T.H. van Andel and CO. Bowin, Mid-Atlantic Ridge be- tween 22° and 23° north latitude and the tectonics of mid- ocean rises, J. Geophys. Res. 73 (1968) 1279. 13 H.S. Fleming, N.Z. Cherkis and J.R. Heirtzler, The Gibbs fracture zone: a double fracture zone at 52°30'N in the Atlantic Ocean, Mar. Geophys. Res. 1 (1970) 37. 366 116 14 P.J. Fox, A. Lowrie, Jr. and B.C. Heezen, Oceanographer Fracture Zone, Deep-Sea Res. 16 (1969) 59. 15 J.D. Phillips, B.P. Luyendyk and D.W. Forsyth, Central North Atlantic plate motions. Science 1 74 (1971) 846. 16 P.J. Fox, W.C. Pitman, III and I . Shephard, CniMal plates in the Central Atlantic: evidence for at least two poles of rotation. Science 165 (1969)487. 17 T.H. van Andel, J.B. Corliss and V.T. Bowen, The inter- section between the Vema Fracture Zone and the Mid- Atlantic Ridge in the North Atlantic, J. Mar. Res. 25 (1967) 343. 18 F.J. Vine and D.H. Matthews, Magnetic anomalies over oceanic ridges, Nature 199 (1963) 947. 19 B.C. Heezen and M. Tharp, Tectonic fabric of the Atlantic and Indian oceans and continental drift. Philos. Trans. R. Soc. London, Ser. A 258 (1965) 90. 20 D.P. Mckenzie and R.L. Parker, The North Pacific: an example of tectonics on a sphere. Nature 216 (1 967) 1 276. 21 X. LePichon, Sea-floor spreading and continental drift, J. Geophys. Res. 73 (1968) 3661. 22 D.L. Anderson, Accelerated plate tectonics, Science 187 (1975) 1077. 23 B.D. Loncarevic and R.L. Parker, The Mid-Atlantic Ridge near 45°N, XVII. Magnetic anomalies and ocean-floor spreading. Can. J. Earth Sci. 8(1971) 883. 24 H.D. Needham and J. Francheteau, Some characteristics of the Rift Valley in the Atlantic Ocean near 36 48' North, Earth Planetary Sci. Lett. 22 (1974) 29. 25 D. Greenewalt and P. I . Taylor. Deep-tow magnetic mea- surements across the axial valley ot the Mid-Atlantic Ridge, .1. Gcoph>s. Res. 79 (1974) 4401. 26 R.K. Lattimore, P. A. Rona and O.I . DeWakL Magnetic anomaly sequence in the Central Niutli Atlantic. .1. Geo- phys. Res. 79 (1974) 1207. 27 B.A. McGregor, C.G.A. Harrison, J.W. Lavellc and PA. Rona, Magnetic anomaly pattern on Mid-Atlantic Ridge crest at 26''N. J. Geophys. Res. (1976) in press. 28 T.H. van Andel and T.C. Moore, Magnetic anomalies and sea-floor spreading rates in the northern South Atlantic. Nature 226 (1970) 328. 29 P.R. Vogt, Asthenosphere motion recorded by the ocean floor south of Iceland, Larth Planetary Sci. Lett. 13 (1971) 155. 30 G.L. Johnson and P.R. Vogt, Mid-Atlantic Ridge from 4 7" to 51° north, Geo!. Soc. Am. Bull. 84 (1973) 3443. 31 P. A. Rona and U.S. Fleming, Mesozoic plate motions in the eastern central North Atlantic. Mar. Geol. 14 (1973) 239. 32 P.R. Vogt, O.F. Avery, E.D. Schneider, C.N. Anderson and D.R. Bracey, Discontinuities in sea-floor spreading. Tectono- physics 8 (1969) 285. 33 P.R. Vogt and G.L. Johnson. Seismic reflection survey of an oblique aseismic basement trend on the Rcykjanes Ridge, Earth Planet. Sci. Lett. 15 (1972) 248. 34 T.H. van Andel and G.R! Heath, Tectonics of the Mid- Atlantic Ridge, 6-8° south latitude, Mar. Geophys. Res. 1 (1970) 5. 367 38 Reprinted from: Earth Science Reviews, Vol. 12, No. 1, 74-75. PLATE TECTONICS AND OIL Alfred G. Fischer and Sheldon Judson (Edi- tors), 1975. Petroleum and Global Tecton- ics. Princeton University Press, Princeton, N.J., 322 pp., US. $16.50. Petroleum and Global Tectonics is a col- lection of nine papers discussing geological processes relevant to the occurrence of oil from the point of view of plate tectonics. The papers, by scientists from universities and the petroleum industry, were presented at a symposium held at Princeton University in 1972 to honor Hollis D. Hedberg. The papers are meaningfully arranged and introduced by the editors In an over- view of plate tectonics, Sir Edward Bullard points out that while petroleum exploration is largely concerned with vertical crustal movements which allow the accumulation of sediments, plate tectonics is primarily concerned with horizontal movements. Five successive papers demonstrate how both ver- tical and horizontal movements determine the evolution of sedimentary basins as the sites of petroleum generation, accumulation and storage. W. Jason Morgan contributes theoretical background on the relation between heat flow and vertical movements of the oceanic lithosphere, one of the most thoroughly un- derstood of the phenomena producing verti- cal crustal displacements. AG. Fischer dem- onstrates that the vertical movements of oceanic lithosphere, combined with horizon- tal movements of plates, provide a plausible mechanism for basins that develop on conti- nental margins; basins developed on conti- nental interiors remain problematic. Those basins that originated by rifting contain the largest volume of prospective sediment. D.J.J. Kinsman explains their development by initial uplift and post-rifting subsidence related to subcrustal temperature and densi- ty distributions. J.D. Lowell, G.J. Genik, T.H. Nelson, and P.M. Tucker consider the evolution of the southern Red Sea as an example of how structural arching, rifting, subsidence, and breaching of continental li- thosphere act to control the occurrence of petroleum. J.R. Curray synthesizes different assemblages of marine sedimentary facies that constitute basins and shows that oroge- nic histories of these sedimentary assem- blages follow almost infinite variations in plate tectonic settings, rather than an invari- ant geotectonic cycle. The generation of petroleum is treated by J.G. Erdman who reviews the processes by which an organic fraction of sediment may be transformed into hydrocarbons by inorganic processes related to temperature, degree of oxidation, and the mineral matrix. H.D. Klemme assembles substantial data to examine relations between hydrocarbon occurrence and both the tectonics and ther- mal regime of productive basins. His evi- dence indicates that basins associated with significantly higher heat flow located along continental margins and rift zones provide optional conditions for the generation, mi- gration, and accumulation of petroleum. The papers of this symposium demon- strate that plate tectonics provides insight to problems relevant to the occurrence of pe- troleum including the origin of basins, sources of sediment, open basins, restricted marine circulation, basin geometry, basin re- lations on opposite continental margins, and thermal regimes within basins. Most of the 198 known giant oil fields discussed by J.D. Moody in the concluding paper were found prior to the advent of plate tectonics, but plate tectonics will play a significant role in finding the estimated 200 to 300 remaining giant fields. This book exemplifies the kind of creative interplay between academe and industry that results in intellectual and ma- terial advances. It is worthwhile reading both for scientists and informed laymen. Peter A. Rona, Miami, Fla. 368 39 Reprinted from: Geological Society of America, Microform Publication, Vol. 5, 490 p. Mid-Atlantic Ridge: Selected Reprints and Bibliography Edited by Peter A. Rona INTRODUCTION The following articles from publications of the Geological Society of America are assembled in chronological order and provide perspective of the development of geological knowledge of the Mid-Atlantic Ridge spanning a quarter century of research from early studies to the present frontier. Early bathymetric reconnaissance gradually revealed the regional morphology of the Mid-Atlantic Ridge (Tolstoy and Ewing, 1949; Tolstoy, 1951). Cross-sections of the deep crustal structure underlying the Mid-Atlantic Ridge determined by the two-ship seismic refraction method (Ewing and Ewing, 1959) are only now being refined by new methods. Groundwork on the regional distribution of sediment type by coring (Ericson and others, 1961) and of sediment thickness by seismic reflection profiling (Ewing and others, 1964) preceeded studies of sedimentary processes at representative sites on the Mid-Atlantic Ridge (van Andel and Komar, 1969; Ruddiman, 1972). Sampling of rocks from emerged (Le Maitre, 1962) and from submerged (Quon and Ehlers, 1963; Engel and others, 1965; Switzer and others, 1970; Melson and Thompson, 1973) portions of the Mid-Atlantic Ridge has contributed to recognition of the distinctive petrology of oceanic rocks, and has stimulated their comparison with ophiolites (Thayer, 1969; Green, 1970). 369 The designation of the Mid-Atlantic Ridge as a divergent plate boundary in the theory of plate tectonics has focused research on processes at the axial region of the ridge. Advances in magnetic interpretation made it possible to determine the history of generation about the ridge crest of Atlantic oceanic lithosphere (Pitman and Talwani, 1972). Studies of the thermal regime of the Mid-Atlantic Ridge are related both to its characteristic profile (Sclater and Detrick, 1973) and to the petrologic effects of hydrothermal activity (Anderson, 1972). Increasing realization of the complexity of axial processes has led to the concentration of studies at representative areas of the Mid-Atlantic Ridge crest (Ward, 1971; Johnson and Vogt, 1973; van Andel and others, 1973; Phillips and others, 1975). Interdisciplinary, cooperative investigations have been adopted as an effective research approach. These investigations of the crestal region of the Mid-Atlantic Ridge include work near lat 36°N by project FAMOUS (French-American Mid-Ocean Undersea Study; Heirtzler and Le Pichon, 1974), near lat 45°N by Canadian scientist (Loncarevic, this publication), and near lat 26°N by the Trans-Atlantic Geotraverse (TAG) project of the National Oceanic and Atmospheric Administration (Rona and others, 1976). No longer an enigmatic geographic feature, the Mid-Atlantic Ridge is being studied as the locus of processes that affect the entire Earth. 370 Reprinted from: Marine Geology, Vol. 21, No. 4, M59-M66, Marine Geology, 21 (1976) M59— M66 M59 © Elsevier Scientific Publishing Company, Amsterdam — Printed in The Netherlands Letter Section PATTERN OF HYDROTHERMAL MINERAL DEPOSITION: MID- ATLANTIC RIDGE CREST AT LATITUDE 26° N PETER A. RONA National Oceanic and Atmospheric Administration, Atlantic Oceanographic and Meteor- ological Laboratories, Miami, Fla. 33149 (U.S.A.) (Received February 25, 1976; revised and accepted May 20, 1976) ABSTRACT Rona, P.A., 1976. Pattern of hydrothermal mineral deposition: Mid-Atlantic Ridge crest at latitude 26° N. Mar. Geol., 21:M59-M66. Interdisciplinary studies of the TAG Hydrothermal Field on the Mid- Atlantic Ridge crest at latitude 26° N reveal two principal depositional patterns of hydrothermal minerals: (1) A pattern of deposition controlled by physical and chemical processes within the hydrothermal field. A major process in determining depositional pattern within the hydrothermal field is inferred to be sealing of talus by deposition of hydrothermal minerals from solutions discharged through underlying faults at and adjacent to the wall of the rift valley. The sealing of a given volume of talus is inferred to occur during a period of the order of 1 • 10 yr, causing successive migrations of the zone of discharge. The resulting pattern of hydrothermal mineral deposition within the hydrothermal field would be expec- ted to be a mosaic of deposits overlapping in time and space with a predominantly fault- controlled trend parallel to the axis of the rift valley. (2) A pattern of deposition controlled by sea-floor spreading encompassing the entire hydrothermal field. A linear zone of hydrothermal deposits will extend from an active depositional locality at a rift valley along the direction of sea-floor spreading depending both on the continuity of sea-floor spreading and the persistence in time of the special structural and thermal conditions that concentrate the hydrothermal activity. The special structural and thermal conditions that have concentrated hydrothermal activity at the TAG Hydrothermal Field have persisted during sea-floor spreading for at least 1.4 • 10 yr. INTRODUCTION Concentrated hydrothermal mineral deposits are known from several local- ities along divergent plate boundaries, including the Red Sea (Degens and Ross, 1969), the Galapagos spreading axis (Moore and Vogt, 1976), and the Mid- Atlantic Ridge at latitudes 36° N (ARCYANA, 1975), 26° N (M.R. Scott et al., 1974), and 23°N (Thompson et al., 1975). The minerals are deposited by sub-sea floor hydrothermal convection systems inferred to involve the circulation of seawater through oceanic crust driven by intrusive heat sources at sea-floor spreading centers (Spooner and Fyfe, 1973). Knowledge of the 371 M60 depositional pattern of hydrothermal minerals in time and space at localities along divergent plate boundaries would help to elucidate the nature of sub- sea floor hydrothermal convection systems and metallogenesis in oceanic crust (Rona, 1973; Bonatti, 1975). Interdisciplinary studies by the NOAA Trans- Atlantic Geotraverse (TAG) project of concentrated hydrothermal mineral deposits on the Mid- Atlantic Ridge at latitude 26° N provide the basis for a preliminary interpretation of their pattern of deposition (Rona et al., 1976). Evidence for past and present concentration of hydrothermal activity including at least a 10-km2 area at the southeast side of the rift valley has led to designation of this locality as the TAG Hydrothermal Field (Fig.l; R.B. Scott et al., 1974). OBSERVATIONS The bathymetric setting of the TAG Hydrothermal Field is a ridge that 44°55'W 44°30*W 00*N Fig.l. Bathymetric map (McGregor and Rona, 1975) contoured in hundreds of meters showing axis (solid line) of rift valley (shaded) of the Mid-Atlantic Ridge at latitude 26 N, two profiles (X, Y) along which temperature measurements and bottom photographs were made concurrently, and dredge stations at the southeast wall of the rift vallev (TAG 1972- 13 [72-13], TAG 1973-2A [73-2A] TAG 1973-3A [73-3A] and along a ridge (unshaded) trending orthogonal to the axis of the rift valley (TO-75AK61-1A [ 75-1 A ], TO-75AK59-1B [75-1B], TAG 1973-6A [73-6A]). The approximate known area of the TAG Hydrothermal Field is indicated (dashed line). 372 M61 trends orthogonal to the rift valley (Fig.l). The west end of the ridge forms the southeast wall of the rift valley. Abundant manganese oxide crusts were recovered from three dredge stations where the orthogonal ridge forms the southeast wall of the rift valley (Fig.l; dredge stations TAG 1972-13, TAG 1973-2A, TAG 1973-3A). The hydrothermal origin of the manganese oxide crusts is evidenced by their rapid rates of accumulation (200 mm per 106 yr) determined radiogenically, and by their extreme purity of composition (40% Mn), with only trace quantities of metals other than manganese (M.R. Scott et al., 1974), occupying the Mn-rich end member of Bonatti's hydrothermal classification (Bonatti, 1975). Manganese oxide crusts were also present at one station situated along the crest of the orthogonal ridge (Fig.l; dredge station TO-75AK61-1A). Manganese oxide crusts were absent at two other dredge stations along the crest of the orthogonal ridge (Fig.l; dredge stations TAG 1973-6A, TO-75AK59-1B), where ropy-textured, apparently fresh basalt with high K2 O content (0.3%) characteristic of off-axial extrusion was recov- ered (R.B. Scott et al., 1976). The hydrothermal manganese oxide occurs as a crust on talus of basalt frag- ments (Fig.2), as veins in the basalt fragments, and as a crust on and matrix in breccia of altered basalt fragments (Fig.3). The talus and breccia occur along the inner margins of steps on the southeast wall of the rift valley revealed by narrow-beam bathymetry and bottom photographs (McGregor and Rona, 1975). The steps are interpreted as the topographic expression of faults that may act as conduits for hydrothermal solutions. Two profiles combining water temperature measurements and bottom photographs were made over steps on the southeast wall of the rift valley between depths of about 2500 and 3500 m (Rona et al., 1975). A temperature anomaly about 300 m wide consisting of an increase in ambient temperature (+0.1°C) and an inversion of potential temperature gradient (0.015°C per m warming downwards), was measured within 20 m of the sea floor over a talus- covered step at a depth of about 3000 m on one of the two profiles (Fig.l; profile X; Rona et al., 1975). The character and geologic setting of this tem- perature anomaly favor its interpretation as due to convective transfer of heat by discharge of hydrothermal solutions focussed by faults in the wall of the rift valley and diffused by a porous and permeable layer of talus overlying the faults. A temperature anomaly was absent at the second profile (Fig.l; profile Y) situated 5 km along the step on the wall of the rift valley where the temperature anomaly was measured on profile A. Bottom photographs reveal- ed that breccia and pillow lavas are the predominant rock types present along profile B (McGregor and Rona, 1975, their fig.6). DISCUSSION The duration and position of deposition of the manganese oxide crusts observed at the TAG Hydrothermal Field may be deduced from their rates of accumulation and the local half-rate of sea- floor spreading (1.3 cm per yr; 373 M62 Fig.2. Bottom photographs (field of view approximately 4 X 6 m) showing talus of basalt fragments at 3000-m depth along profile X (Fig. 1) on the southeast wall of the rift valley where a water temperature anomaly was measured (Rona et al., 1975). The camera water current compass (length 34 cm) is visible suspended 5 m below the camera. Lattimore et al., 1974). The manganese oxide crust sampled attains a thick- ness of 42 mm at the southeast wall of the rift valley, 5 km from the axis of the rift valley (Fig.l; dredge station TAG 1972-13). U-Th dating of the man- ganese crust shows it to have accumulated at a rate of about 200 mm per 106 yr, with cessation of accumulation about 15 • 103 yr ago (M.R. Scott et al., 1974, their table lb, fig.3). Assuming a constant rate, the manganese oxide crust at dredge station TAG 1972-13 began to accumulate about 2 • 10s yr ago at a position 2 km from the axis of the rift valley and continued to ac- cumulate during sea-floor spreading nearly up to its present position at the wall of the rift valley. The manganese oxide crust at dredge station TO-75AK61-1A, situated 17 km from the axis of the rift valley (Fig.l), consists of two layers. An un- derlying layer of hydrothermal manganese up to 10 mm thick accumulated at a rate of about 35 mm per 106 yr (R.B. Scott et al., 1976), during a period of 3 • 10s yr, at distances between 8 and 12 km from the axis of the rift val- ley. The deposition of hydrothermal manganese oxide ceased 12 km from the axis of the rift valley when the underlying basalt was about 1 • 106 yr old. Then a layer of hydrogeneous ferromanganese oxide up to 3 mm thick accu- mulated at a rate of about 8 mm per 106 yr on the hydrothermal manganese (R.B. Scott et al., 1976) during a period of about 4 • 105 yr, at distances between 12 and 17 km from the axis of the rift valley. The reconstructed 374 M63 Fig. 3. Bottom photograph (field of view approximately 4 X 6 m) showing hydrothermal manganese oxide crust (upper left and central portion of photograph) on breccia of basalt fragments at 2600-m depth along profile Y (Fig. 1 ). The camera water current compass (length 34 cm) is visible suspended 5 m below the camera. sequence of events indicates that hydrothermal deposits began to accumulate on the floor of the rift valley and continued to accumulate through uplift of the floor to form the walls and adjacent orthogonal ridge during a period of about 8 • 105 yr. The duration of deposition of the hydrothermal manganese crusts is of the same order of magnitude at dredge stations TAG 1972-13 (2 • 10s yr) and TO-75AK61-1A (3 • 105 yr), in spite of the different distances from the axis of the rift valley at which the accumulation occurred. The similar duration of accumulation implies the operation of a process that may limit hydrothermal discharge at a given site adjacent to the rift valley to a period of the order of 1 '10s yr. It has been suggested that development of an impermeable sedi- ment cover may suppress hydrothermal discharge on an oceanic ridge (Lister, 1972). However, our studies indicate that sediment cover is negligible in the area of the TAG Hydrothermal Field (Rona et al., 1976). The observations presented of near-bottom water temperature, bottom photographs, and petrology, support the hypothesis that hydrothermal dis- 375 M64 charge becomes suppressed at sites within the hydrothermal field when deposition of hydrothermal minerals seals off a portion of the discharge zone. Self-sealing by mineral precipitation is a mechanism that has been recognized to operate in certain geothermal systems on continents (Elder, 1966; Facca and Tonani, 1967; Helgeson, 1968; Elders and Bird, 1974; Batzle and Simmons, 1976). The temperature anomaly attributed to convective transfer of heat occurs where the wall of the rift valley is covered by talus, a porous and permeable material through which hydrothermal discharge from under- lying faults can flow (Figs. 1,2; profile X). Hydrothermal manganese oxide recovered near profile X occurs as a crust on basalt talus (Fig.l; dredge stations TAG 1972-13 and TAG 1973-3A; Rona et al., 1976, their table 3). A tem- perature anomaly was absent along profile Y (Fig.l), where a high proportion of breccia and pillow lava was photographed (Fig.2), materials which are impermeable to hydrothermal flow. Hydrothermal manganese oxide occurs as a crust on and matrix in breccia recovered near profile Y (Fig.l; dredge station TAG 1973-2A; Rona et al., 1976, their table 3). According to this interpretation, the breccia may fcrm by alteration and cementation of the talus by concentrated hydrothermal activity. CONCLUSIONS A preliminary pattern of deposition of hydrothermal minerals at a locality along a divergent plate boundary is emerging from interdisciplinary studies of the TAG Hydrothermal Field. Two principal patterns may be discerned, that will require testing by more detailed studies at this and other localities: (1) A pattern of deposition controlled by physical and chemical processes within a hydrothermal field. A major process is inferred to be sealing of inter- stices in talus by deposition of hydrothermal minerals from solutions dis- charged through underlying faults at and adjacent to the wall of the rift valley. The duration of accumulation of hydrothermal manganese oxide crusts determined at two sites (Fig.l; dredge stations TAG 1972-13, TO-75AK61-1A), indicates that the sealing process, inferred to involve the conversion of talus to breccia, occurs during a period of the order of 1 • 10s yr. As the talus overly- ing fracture-focussed hydrothermal discharge becomes sealed, the zone of dis- charge gradually migrates to areas of unsealed talus. The migration of the hydrothermal discharge zone is probably controlled by the characteristics of the fracture system that focusses the flow. Consequently, the migration will follow the direction of faults at and near to the wall of the rift valley, which are primarily aligned parallel to the axis of the rift valley. Once sealed, the breccia may be fractured by tectonic forces opening the possibility of another generation of hydrothermal deposition; however, the fractured breccia probably would not regain the original porosity and permeability of the talus. An additional process that may suppress hydrothermal activity within the area of a hydrothermal field is off-axis intrusive and extrusive volcanism (Rona et al., 1976). The resulting pattern of hydrothermal mineral deposition within 376 M65 the hydrothermal field would be expected to be a mosaic of hydrothermal deposits overlapping in time and space, partially covered by extrusive volcanic rocks, with a predominant fault-controlled trend parallel to the axis of the rift valley. (2) A pattern of deposition of hydrothermal minerals controlled by sea- floor spreading encompassing an entire hydrothermal field. It was previously pro- posed (Rona, 1973; Rona et al., 1976) that a linear zone of relict hydrothermal deposits will extend along the direction of sea-floor spreading from an active depositional locality at a rift valley. The length of the linear zone would de- pend both on the continuity of sea- floor spreading and the persistence in time of the special structural and thermal conditions that concentrate the hydro- thermal activity. The width of the linear zone of relict hydrothermal deposits would equal the width of the associated hydrothermal field, which is 10 km in the case of the TAG Hydrothermal Field. The relict hydrothermal manganese crust recovered 17 km in the direction of sea- floor spreading from the axis of the rift valley (Fig.l; dredge station TO-75AK61-1A), indicates that the special structural and thermal conditions that have concentrated hydrothermal activity at the TAG Hydrothermal Field have persisted during sea- floor spread- ing for at least 1.4 • 106 yr. Off-axis extrusive volcanism, such as that eviden- ced at dredge stations TAG 1973-6A and TO-75AK59-1B (Fig.l), may cover linear zones of hydrothermal deposits that may extend along flow lines of sea- floor spreading. REFERENCES ARCYANA, 1975. Transform fault and rift valley from bathyscaph and diving saucer. Science, 190: 108—116. Batzle, M.L. and Simmons, G., 1976. Microfractures in rocks from two geothermal areas. Earth Planet. Sci. Lett., 30: 71—93. Bonatti, E., 1975. Metallogenesis at oceanic spreading centers. Annu. Rev. Earth Planet. Sci., 3: 401—431. Degens, E.T. and Ross, D.A. (Editors), 1969. Hot Brines and Recent Heavy Metal Deposits in the Red Sea— A Geochemical and Geophysical Account. Springer, New York, N.Y., 571 pp. Elder, J.W., 1966. Heat and mass transfer in the Earth: hydrothermal systems. N.Z.D.S.I.R. Bull., 169: 115 pp. Elders, W.A. and Bird, D.K., 1974. Investigations of the Dunes geothermal anomaly, Imperial Valley, California, II. Petrological studies, presented at the International Symposium on Water— Rock Interaction of the International Union of Geochemistry and Cosmochemistry, Prague, 14 pp. Facca, G. and Tonani, F., 1967. The self-sealing geothermal field. Bull. Volcanol., 30: 271. Helgeson, H.C., 1968. Geologic and thermodynamic characteristics of the Salton Sea geo- thermal system. Am. J. Sci., 266: 129. Lattimore, R.K., Rona, P.A. and DeWald, O.E., 1974. Magnetic anomaly sequence in the central North Atlantic. J. Geophys. Res., 79: 1207—1209. Lister, C.R.B., 1972. On the thermal balance of a mid-ocean ridge. Geophys. J. R. Astron. Soc, 26: 515—535. McGregor, B.A. and Rona, P.A., 1975. Crest of Mid- Atlantic Ridge at 26°N. J. Geophys. Res., 80: 3307-3314. 377 M66 Moore, W.S. and Vogt, P.G., 1976. Hydrothermal manganese crusts from two sites near the Galapagos spreading axis. Earth Planet. Sci. Lett., 29: 349—356. Rona, P.A., 1973. Plate tectonics and mineral resources. Sci. Am., 229 (1): 86—95. Rona, P.A., McGregor, B.A., Betzer, P.R. and Krause, D.C., 1975. Anomalous water tem- peratures over Mid-Atlantic Ridge crest at 26 north latitude. Deep-Sea Res., 22: 611— 618. Rona, P.A., Harbison, R.N., Bassinger, B.G., Scott, R.B. and Nalwalk, A.J., 1976. Tectonic fabric and hydrothermal activity of Mid- Atlantic Ridge crest (lat. 26 N). Geol. Soc. Am. Bull., 87: 661-674. Scott, M.R., Scott, R.B., Rona, P.A., Butler, L.W. and Nalwalk, A.J., 1974. Rapidly accu- mulating manganese deposit from the median valley of the Mid- Atlantic Ridge. Geophys. Res. Lett., 1: 355— 358. Scott, R.B., Rona, P. A., McGregor, B.A. and Scott, M.R., 1974. The TAG hydrothermal field. Nature, 251: 301-302. Scott, R.B., Malpas, J., Rona, P.A. and Udintsev, G., 1976. Duration of hydrothermal ac- tivity at an oceanic spreading center, Mid-Atlantic Ridge (lat. 26 N). Geology, 4: 233— 236. Spooner, E.T.C. and Fyfe, W.S., 1973. Sub-sea floor metamorphism, heat and mass trans- fer. Contrib. Mineral. Petrol., 42: 287—304. Thompson, G., Woo, C.C. and Sung, W., 1975. Metalliferous deposits on the Mid- Atlantic Ridge. Geol. Soc. Am. Abstr. Progr., 7: 1297—1298. 378 41 Reprinted from: Proc. of NOAA Marine Minerals Workshop, March 1976, 111-119, Resource Research and Assessment of Marine Phosphorite and Hard Rock Minerals Peter A. Rona National Oceanic and Atmospheric Administration Atlantic Oceanographic and Meteorological Laboratories INTRODUCTION The National Oceanic and Atmospheric Administration (NOAA) is involved in six projects related to assessment of marine phosphorite and hard rock minerals (Table 1) . NOAA involvement constitutes support through the Sea Grant Program of four of the projects (S-3, S-9, S-28, S-32) , and actual implementation of two of the projects (M-4 and S-37, NOAA Metallogenesis) . Brief summaries and a list of publications are presented for each of the six projects. PROJECT SUMMARIES Evaluation and Economic Analysis of Southern California Phosphorites and Sand-Gravel Deposits (S-3). The Principal Investigators of this project are Peter J. Fischer of California State University, Northridge, and Walter Mead of the University of California, Santa Barbara (Table 1). The project objective is to make a geological evaluation, integrated with economic and socio- economic assessment, of offshore and onshore sand and gravel and phosphorite deposits. The assessment of the sand and gravel resource potential of the southern California shelf is nearing completion. The study extends from the Mexican border north to Point Conception, a distance of 460 km. Based upon preliminary estimates, the volume of unconsolidated shelf sediments is 26.5 km^. Economic studies are in progress to determine which, if any, of these deposits are viable resources. With regard to phosphorite, a set of maps of the southern California continental borderland has been completed showing all available phosphorite resource data. Undersea Mineral Survey of the Georgia Continental Shelf (S-9) . The Principal Investigator of this project is John Noakes of the University of Georgia. The project was completed in 1975, accomplishing the following : 1. The technique of neutron activation analysis using a Californium 2 52 neutron source has been applied to both 379 shipboard and In situ identification of elements in seafloor minerals. 2. Field tests have demonstrated the potential of using a mobile sled equipped with radiation detection equipment to locate and differentiate between thorium associated with heavy mineral deposits and uranium associated with phosphorites, 3. Over 300 miles of Georgia coastal area have been covered by reconnaissance surveys. Lake Superior Copper Survey (S-20) . R. P. Meyer of the University of Wisconsin is the Principal Investigator of this project which was completed in 197 5. Accomplish- ments of the project include the following: 1. Five areas adjacent to the copper producing area of the Keweenaw Peninsula were investigated and were identified as possible target areas for future development. 2. Bottom-towed and surface-towed resistivity arrays were success- fully applied to the location of known copper-bearing veins and sand deposits with high heavy mineral content. 3. An active-source audiomagnetotelluric system with towed receivers successfully detected conductivity anomalies associated with known copper-bearing veins. 4. A first-order analytical method was developed to distinguish resistivity anomalies related to bottom topography from those due to changes in conductivity. Marine Lode Minerals Exploration (S-32) The Principal Investigator of this project is J. R. Moore of the Marine Research Laboratory of the University of Wisconsin. The project objective is to provide basic chemical, mineral, and textural explora- tion clues that will indicate the presence of sub-seafloor lode bodies, particularly ores of copper, lead, zinc, nickel, and barite. The project has already received cooperative assistance from Chromalloy Corp. and ASV Corp. for surveys at industry mining sites at Castle Island (barite) and Kllamar (copper), Alaska. Coronado Bank Phosphorite Deposit (M-4) The Principal Investigator of this project is B. B. Barnes of the former Marine Minerals Technology Center. The project was completed 380 in 1971, accomplishing the following: 1. A typical marine phosphorite deposit on Coronado Bank offshore southern California, was investigated to test equipment and techniques for phosphorite deposit delineation. The investigation included bathymetry, seismic reflection profiling, bottom photography, and dredging. 2. Areas of Coronado Bank that yielded the nighest percentage of P20- (nodules) were related to zones of deep weathering, fractures in the sea floor, and organic activity. Metallogenesis at Dynamic Plate Boundaries (A-]) In 1972 the NOAA Trans-Atlantic Geotraverse (TAG) project (P. 'A. Rona, Chief Scientist-) of the Atlantic Oceanographic and Meteorological Laboratories (AOML) , dredged hydrothermal manganese oxide crusts frcm the wall of the rift valley of the Mid-Atlantic Ridge at latitude 26° N. Subsequent multidisciplinary investigations in- cluding narrow-beam bathymetry, gravity, magnetics, bottom photography, near-bottom water temperature and chemistry measurements , dredging and coring revealed both active and relict hydrothermal manganese oxide de- posits covering at least a 15 km square area, in and adjacent to the rift valley, that has been designated the TAG Hydrothermal Field. The TAG Hydrothermal Field is hypothesized to be the discharge zone of a voluminous sub-seafloor hydrothermal convection system involving the circulation of seawater through oceanic crust driven by intrusive heat sources beneath the rift valley. From geochemical considerations and analogy with ophiolites, such as the Troodos Massif of Cyprus, massive copper - iron stratabound sulfide bodies, are inferred to underlie the hydrothermal manganese oxide crusts, although only dissem- inated sulfides have been sampled to date. A new NOAA project, Metallogenesis at Dynamic Plate Boundaries (see A-l) is being proposed to increase understanding of the hydrothermal process of metal concentration in oceanic crust, to develop exploration criteria for both active and relict hydrothermal deposits in oceanic .rust in situ and in ophiolites, and to determine the distribution of hydrothermal deposits in oceanic crust. The Principal Investigator of this project is P. A. Rona (AOML, Miami). Ophiolices, slices of oceanic i rust formed about an oceanic ridge and incorporated into certain islands and continents are presently accessible to exploitation, and are being mined for base and precious metals at certain localities such as Cyprus. 1 1 381 TABLE 1. NOAA Activities in Assessment of Marine Phosphorite and Hard Rock Minerals Project Principal Identification* Investigator (s) Title Term S-3 P- J- Fischer and Evaluation and economic 1975-76 W. Mead analysis of southern California's phosphorite and sand-gravel deposits S-9 J. Noakes Undersea mineral survey 1970-75 of the Georgia continental shelf S-28 R. P. Meyer Lake Superior copper 1971-75 survey S-32 J. R. Moore Marine lode minerals 1975-78 exploration M-4 B. B. Barnes Coronado Bank phosphorite 1968-71 deposit A-l P- A. Rona Metallogenesis at Dynamic 1976 (pursuant Plate Boundaries to work initio- in 1972) - 196 * S - Sea Grant Program * M - Marine Minerals Technology Center * A - Atlantic Oceanographic and Meteorological Labs, NOAA 114 382 REFERENCES BY PROJECT (TABLE 1) Evaluation and Economic ^malysis of Southern California's Phosphorite and Sand-Gravel Deposits (S-3) Ashley, R. , Berry, R. , and Fischer, P.J., 1975, Geology of the northern continental shelf of the Santa Barbara Channel from Gaviota to El Capitan : in, Studies on the Geology of Camp Pendleton and Western San Diego County, California, p. 77-79. Ashley, R. , Berry, P.., and Fischer, P.J., 1976, Geology of the northern continental shelf of the Santa Barbara Channel from Gaviota to El Capitan: Journ. cf Sedimentary Petrology, in press. Byrd, R. , Berry, R. , and Fischer, P.J., 1975, Quarternary geology of the San Diego - La Jolla Underwater Park: in, Studies on the geology of Camp Pendleton and Western San Diego County, California, p. 77-79 and p. 300. Drake, D. , Kolpack, R. , and Fischer, P.J., 1972, Sediment transport on the Santa Barbara - Oxnard shelf, Santa Barbara Channel, California: in, Swift, D.J. P., and others, editors/ Shelf sediment transport: Dowden , Hutchinson and Ross, Inc., p. 307-331. Mead, W. J., 1969, and Sorensen, P.E., 1969, A new economic appraisal of marine phosphorite deposits: Marine Technology Society, The Decade Ahead. Mead, W. J., and Sorensen, P. E., 1970, The principal external costs and benefits of marine mineral recovery: Offshore Technology Conference, Proceedings, V. 1. Wilcox, S. , Mead, W., and Sorensen, P.E., 1972, A preliminary estimate of the economic potential of marine placer mining: Marine Technology Society, Proceedings. 383 Undersea Mineral Survey of the Georgia Continental Shelf (S-9) Noakes, J. E. and Harding, J. L.,1971, New techniques on seafloor mineral exploration: Marine Technology Society, V. 5, No. 6, p. 41. Noakes, J. E. , Harding, J.L., and Spaulding, J.D. , 1974, Locating offshore mineral deposits by natural radioactive measurements: Marine Technology Society, V. 8, No. 5, p. 36-39. Noakes, J. E. , Harding, J. L. , Spaulding, J. D. and Fridge, D. S. , Surveillance system for sub-sea survey and mineral exploration Offshore Technology Conference, Paper 2239, p. 909-914. Noakes, J. E., Harding, J. L. , Spaulding, J. D. , and Hill, J., Radioactive monitoring of offshore nuclear power stations: Offshore Technology Conference, Paper OTC 1988, p. 501-506. Noakes, J. E., Smithwick, G. , Harding, J. and Kirst, A., 1971, Undersea mineral analysis with Californium-252 : Proceedings Am Nuclear Society Meeting, April. Lake Superior Copper Survey (S-28) Brzozowy, C. P., 1973, Magnetic and seismic reflections surveys of Lake Superior: University of Wisconsin, Sea Grant College Technical Report WIS-SG-74-220, 40 pp. Goodden, J. J. P., 1973, Surveying the lake floor in search of underwater copper reserves to revive an ancient mining district, Keweenaw Peninsula, Northern Michigan: University of Wisconsin - Madison Marine Research Laboratory, Sea Grant Underwater Minerals Program, 7 pp. Goodden, J. J. P., 1974, Sedimentological aspects of underwater copper exploration in Lake Superior : University of Wisconsin - Madison, Master's Thesis. Meyer, R. P., Moore, J. R. and Nebrya, E., 1975, Underwater copper explorations in Lake Superior II: Specific targets charted in 1974: Offshore Technology Conference, Paper OTC 2291, 16 pp. 116 384 Moore, J. R. , Meyer, R. P., and Wold, R. J., 1972, Underwater copper exploration in Lake Superior - prospects mapped in 1971: Offshore Technology Conference, Paper OTC 1648, p. II - . 307-322. Nebrija, E., Young, C. , Meyer, R. , and Moore, J. R. , 1976, Electrical prospecting for copper veins in shallow water: Offshore Technology Conference, in press. Smith, P. A., and Moore, J. R. , 1972, The distribution of trace metals in the surficial sediments surrounding Keweenaw Point, Upper Michigan: International Assoc. Great Lakes Res., Proc. 15th Conf. Great Lakes Res., p. 383-393; The University of Wisconsin Sea Grant College Reprint WIS-SG-73-341. Thornton, S. E., A shipboard geochemical prospecting technique for determining copper in Lake Superior sediments: University of Wisconsin, Sea Grant Underwater Minerals Program, 7 pp. Tuerkheimer, F. M. , 1974, Copper mining from under Lake Superior: The legal aspects: Natural Resources Lawyer, Winter issue, p. 137-155, University of Wisconsin, Sea Grant College Reprint WIS-SG-74-354. Marine Lode Minerals Exploration (S-32) Moore, J. R. , and Welkie, C. W. , 1975, Metal-bearing sediments of economic interest, coastal Bering Sea: Anchorage, Proc. Conference of the Alaska Geological Society, April. Moore, J. R. , and Van Tassel, J., 1976, Exploration research for marine gold placers: Grantley Harbor - Tuksuk Channel region, Seward Peninsula, Alaska: Sea Grant Technical Report, in preparation. Panel on Operational Safety in Marine Mining, Moore, T. R. , Chairman, 1975, Mining in the outer continental shelf and in the deep ocean: Washington, D.C., National Academy of Sciences, 119 pp. 117 385 Otjen, R. P., 1975, Texture and composition of surficial sediments between Cape Home and Rocky Point, Alaska: University of Wisconsin - Madison,, M.S. Report, 89 pp. Owen, R. M. , 1975, Sources and depositions of sediments in Chagvan Bay, Alaska: University of Wisconsin - Madison, Ph. D. Thesis, 201 pp. Welkie, C J., 1976, Noble metals placer formation: An offshore processing conduit: University of Wisconsin - Madison, M.S. Thesis, in preparation. Coronado Bank Phosphorite Deposit (M-4) Barnes, B. B. , 1970, Marine phosphorite deposit delineation techniques tested on the Coronado Bank, Southern California: Offshore Technology Conference, Paper OTC 1259, p. II - 315-347. Metallogenesis at Dynamic Plate Boundaries (A-l) Betzer, P. R. , Bolger, G. W. , McGregor, B. A., and Rona, P. A., 1974, The Mid-Atlantic Ridge and its effect on the composition of particulate matter in the deep ocean: EOS (Am. Geophys. Union Trans.), V. 55, No. 4, p. 293. McGregor, B. A. and Rona, P. A., 1975, Crest of Mid-Atlantic Ridge at 26° N: Jour. Geophys. Research, V. 80, p. 3307-3314. Rona, P. A. ,1973, Plate tectonics and mineral resources: Scientific American, V. 229, #1, pp. 86-95. Rona, P. A., Harbison, R. H., Bassinger, B. G. , Scott, R. B. , and Nalwalk, A. J., 1976, Tectonic fabric and hydrothermal activity of Mid-Atlantic Ridge Crest (lat. 26° N) : Geol. Soc. Am. Bull., V. 87, 661-674. 118 386 Rona, P. A., McGregor, B. A., Betzer, P. R. , and Krause , D. C. , 1975, Anomalous water temperatures over Mid-Atlantic Ridge Crest at 26° North latitude: Deep-Sea Research, V. 22, p. 611-618. Scott, M. R. , Scott, R. B. , Rona, P. A., Butler, L.W. , and Nalwalk, A. J., 1974, Rapidly accumulating manganese deposit from the median valley of the Mid-Atlantic Ridge: Geophysical Research Letters, V. 1, p. 355-358. Scott, R. B., Rona, P. A., McGregor, B. A., Scott, M. R. , 1974, the TAG hydrothermal field: Nature, V. 251, p. 301-302. J 19 337 4 2 Reprinted from: Special Volume of 'Annals of the Brazilian Academy of Sciences. ' Anais Acad. Brasil Ciencies (Suplemento) , Vol. 48, 256-274. SALT DEPOSITS OF THE ATLANTIC PETER A. BONA National Oceanic and Atmospheric Administration Atlantic Oceanographic and Meteorological Laboratories 15 Rickenbacker Causeway, Miami, Florida 33149 U.S.A. ABSTRACT The distribution in space and time of salt deposits beneath Atlantic continental margins and the adjacent ocean basin is presented in a map and synthesized in a table. Criteria for detecting the salt deposits are defined. The major features of the distribution of the salt deposits are summarized. The distribution of the salt deposits corresponds to the independently determined history of opening of the North Atlantic and South Atlantic. INTRODUCTION The present paper reviews the occurrence of salt deposits beneath continental margins and the adjacent ocean basin around the Atlantic (Fig. 1). Major features of the distribution in space and time of Atlantic salt deposits are deduced from this review. DISTRIBUTION OF ATLANTIC SALT DEPOSITS The distribution of salt deposits of the Atlantic is illustrated in Figure 1 and synthesized in Table 1. The location of each salt deposit and the association of salts present are listed in Table 1. Only those salt deposits that include halite (rock salt) are listed. The mode of occurrence of the salt is speci- fied in Table 1 as either diapirs or strata. Strata refers to beds of salt that may be undeformed or partially deformed. The thickness of salt present is given where known from physical evidence (drilling, outcrop, seismic reflection and /or refraction). The thickness given is that of stratified deposits and not of diapirs. Theore- tical computations are not used as evidence for thickness of a salt deposit. However, it is useful to recall that theoretical studies indicate that salt thicknesses of the . order of hundreds of meters beneath a sedimentary overburden of at least 600 m are generally necessary to produce diapirism (Nettleton, 1934; Parker and McDo- well, 1955). Evidence for the occurrence of the salt deposits described in Table 1 derives from the distinctive physical and chemical properties of salt given in Table 2. Drilling and outcrops furnish direct evidence of the presence of salt deposits. Indirect evidence of the presence of salt deposits is furnished by the following methods: 1 . Seismic reflection and refraction measur- ements based on density-velocity contrasts between salt and surrounding sediment. 2. Magnetic measurements based on the amag- netic properties of salt relative to surround- ing sediment. 3 . Gravity measurements based on the density differential between the salt and surround- ing sediment. In practice, the density of salt deposits varies widely depending on the mass of associated caprock and other factors. An. Acad. bras. Cienc. (1976). 48. (Suplemento) 388 266 PETER A. RONA 60"N RONA Fig-. 1 — Distribution of Atlantic salt deposits. T: Tertiary period; K: Cretaceous period; J: Jurassic period; TR: Triassic period: M: Mississipean (Carboniferous) period; S: Silurian period; DSDP 139, 140, 144; Deep Sea Drilling Project sites 4. Thermal gradient measurements based on the high conductivity of salt relative to surrounding sediment. 5. Salinity and chlorinity gradient measure- ments of interstitial water in unconsolidated sediment over salt deposits based on the high solubility of certain salts, especially halite, in water. The measured salinity gradients are given in Table 1 in parts per thousand (ppt). Seismic, magnetic, and gravity measurements alone may be inadequate to unambiguously dis- tinguish diapirs of salt from diapirs of mud or igneous rock. Thermal gradient and salinity measurements used in conjunction with the other geophysical methods can distinguish salt diapirs from those of other materials. Salinity gradients in interstitial water of unconsolidated sediment have been effectively used by the Deep Sea Drilling Project to indicate the presence of both salt diapirs and strata beneath the seabed (Manheim et al., 1973). As a consequence of the solubility of halite and rapid rates of ionic diffusion, vertical salinity 389 SALT DEPOSITS Of THE ATLANTIC 267 n 03 w "O &q 73 c O in § ^ S E-i § •* < ^ ft 6q aj 2 W" ^ bo .. dj pq 10 .3, •- 2 B c_ £ 2 s» o - O w "1 § -3 ? O O a H 2 2 t» . CO CO c bo So o 0 ttf) o CO CO co pcj il) 33 co &3 O ft; 81 S ft o 8.8 Q 5 § £ ^ ii 02 2 £ t5 « +j *; t- 0> •a >> o "3 c 5 o CO >> 3 £ a K fi CO ^ S3 £ CO CO c a "2 .2 PL, O ~ 1i W Ul ui H P W W H 3 w T3 1-5 O M *d 1-5 Ul a) CUD "w a) ui s J S CO < 01 S CO •a B CO £ .5 "O B cO CO ;- CO s sd CO - u CO T-, 3 x 3 -B a CO 02 fi a o p v-< CO CO B 8 0 CO CO "co CO r? ^ - Of 13 >> .- J= X 8 CO ■B £ 2 "« 2 ? 3 2 "C « riJo b" rco .. B ft W ft w 3 £ « Eg cS a £ cO 02 03 3 0 CO ft -5 S "d p B «< ° w ° p s 5 .S < s g co oT ,_," — t> ¥ '■* 2 S ra • pq -g .. -a 3 CX) c ^ CO CO 5" 05 CO t-i a >; •- tj a! co S S «n — CO ^-t E c CO 3 C3 i co ^ •" 2 g W g h , g 5 3 ~ B; s hoS< T2 B A ft >> be 1 CO CO 03* o CO 03 o o a 5 CO B O CO o fi '~ "co t^-i CO CO B O CJ w f-l '*'■< J i Hi bo B V bs o >i B CO bo c >. B CO CJ 5 CO s CO | -5 CO 3 fi ■5 cO Q CO C bo g •g co" 3 03 >0 "3 K a X bo B CO m --^ 2 « r"C E CO ^ p ■a * » S bo Dh o H w 1-5 > H 3 CO o 1-5 CO CO CO ^z: CO CO C3 rt o O CO B iJ )J hi CO •a B CO 22 ft ^2 "S co p 5° S 'O CO •sis -5 ■3 2fi^ o CO bo — w il co CO ^ O ft m CO O g CO JC J3 B S u CO co CO .- M CQ Q « 390 268 PETER A. RONA M > « st e. K) t> fc he Bq a H Q S S •~ 03 Bq Q S Q as Kl g s s ^ £ 15 0 s ^ fa OS tt) 0 K) El 0 1O Bq O £ T3 K) e CO « 1$ 01 £ .i: cs O CO " O .2 t- C Q "5 Cow c o o .2 ■= 1 .2 2-3 CH K ^ bC 03 SI — 03 a 3 >e ■- 03 >> ■3 ft a • S S J5 a 01 o" 03 « g . .0 9 01" o 01 cu — v ai Ti a w < W H u w 3 (Li S w Ch 3 *d 0! >. a .a '■? s O «3 .co a S * CD 05 > £ rH •c TS * b CO a 3 c 0 a « a '£ CJO ,Q 0) 0 a s n £ A :p M a M CO 03 - E Ml O CO co £ s~ s o .2 t. W 0) H-> 0 "3 E a > ■a 3 0} a h- 1 3 >» c >, H 0 03 a to 0 ■a S J c CO '53 CO „ £> 0 CO 0 C N 0 E 0 O 3 tJj 0 0} ;- 3 0 a a O 0 0 O O 0. rH 3 C--. t- X ■> en 1; en 0> -J - tU r-J 13 .. B < £ ** ^ OS 41 is n s U <- rH O to H rs S -S a M > t> W T3 C OJ — .v .-0 03 O .2 c CO B O >> 0 CO 8 3 CO bo 0 0 bo O O *5 ri 03 « Uj 7, 1> J- ft - -S 'K w & CO — . co >* Eco CU co t« M CM F< - Q cc 10 a o CO — xs B co .2 co cc b*? 03 8 S 0 fl ~ J 5 S . " £ a> CO » _ - NOB 0 c bo 0 co B £ 03 3 O CO Vh OOO 0 as uj Ph o o « 2 — cm w CO 03 03 w CO CO ^ -S Z 3 « « CO J (-1 ^ (D w ■3 r= & PS 3 O1 J= a CO 03 be 0 d cd GQ c « ■a r Cu ** r « « I •5 £ c > a « c ro U a, Q co 0 0 •r en or -a B H3 CO c 91 CO 03 5 g s .i: co 11 0 % "cO « ft i § « 10 C3 0) 01 si* 03 tT -p ■^. o d O r? , 0) KB W ■CP .- o g 3 CO 2 «• f ■«Ka 3 -'3 "} cO w b e bo o cS cS u TA "S ^ | S £ciiSa § 3 * Q a o S" 2 * < oj 42 "5 "8 cO S » in i-S ,- CO a 3 - ce to •m u . a) s ;>> *3 » •£> ■C -3 a .- 'm CO 03 to H s o 2 S I "o 4i q d C M h c 3 53 § ho 13 CO J2 o £ ££ f g m 5 «££ g .5 o « * S 43 ,£] a 0. cO oi CO 03 s 3 M § co a 2 «s « cO >, ♦J q; +J cy u in -~ OT - ® ? *J CD J 5 M o CO "S OJ cfl - CJ T3 ° c 'm to CO to 3 03 CO « U 3 13 EH ft 1-1 t. V 3 >> tO 9 M js 63 °i on "5 6J Eh c o o a S In w ~ 0 ft o g 3 £•§ 3 o • •S 2 ft o ho -^ C Ch O a ft a w 1 «9 0 03 '3 - to i 3 > s 0> £ s| Q 3 o CD s '0 S 3 CO >. 03 25 4= ft CO >. C >. i-t ^^ A 42 CO W) C o — a n s- ft 03 03 a £ ._ 03 co to 73 T3 C C CO CO 03 03 C pi .& d .2 t- b w .is ft m .2 a to a ■a a CO 03 ra a co & t; '5. .2 2 CO ft to 5 a co a « V T3 a !>. o • £ "C S 3 "O a w >. &l' CO M W CO m c ° to 3 £ 3 .2 co OJ CO to 43 "" CO - R, W BP| rt & 5 §, ci H J N < H <5 o o o o o t> OJ C4 t> 393 SALT DEPOSITS OF THE ATLANTIC 271 gradients may develop over halite deposits through several kilometers of water-saturated unconsolidated sediment overburden (Manheim, 1970). The horizontal salinity gradients that develop are approximately equal to the vertical salinity gradients, so that the distribution of vertical salinity gradients delineates the hori- zontal extent of an underlying salt deposit (Ma- nheim and Bischoff, 1969). The ages of the salt deposits listed in Ta- ble 1 are based on various criteria, in order of increasing reliability: 1. Stratigraphic relations: The stratigraphic position of salt strata or the base of salt diapirs in a known stratigraphic sequence. 2. Associated strata: Paleontologic or radio- genic dating of strata associated with stra- tified salt or incorporated in salt diapirs. 3. Caprock: Palynologic dating of the caprock of a salt diapir. 4. Salt: Palynologic dating of the salt of diapirs or strata. MAJOR FEATURES OF THE DISTRIBUTION OF ATLANTIC SALT DEPOSITS The information presented in Figure 1 and Table 1 on the distribution of Atlantic salt deposits is probably incomplete. Salt deposits are generally detected in their most spectacular manifestation as diapirs. Extensive areas of relatively thick salt strata may remain unde- tected beneath Atlantic continental margins and the adjacent ocean basin. For example, layers of competent materials such as carbonates and basalt flows and sills may suppress diapirism and mask underlying salt beds. However, major features of the distribution of Atlantic salt de- posits may be deduced from Figure 1 and Table 1, as follows: I. Salt deposits are present along those rifted portions of the continental margins of North America, South America, Africa, and Eurasia that trend nearly perpen- dicular to fracture zones of the Atlantic ocean basin. II. Salt deposits are absent along those sheared portions of the equatorial conti- nental margins of South America and Africa that trend nearly parallel to frac- ture zones of the Atlantic ocean basin. III. Salt deposits are absent in the South Atlantic south of the Rio Grande Rise and the Walvis Ridge. IV. The salt deposits of the rifted continental margins appear to extend continuously in basins opening seaward from the conti- nents to the deep ocean basin. V. The farthest seaward known extent of salt deposits in the Atlantic is beneath the lower continental rise off northwest Africa at least 450 km from the coast, as predicted from geophysical measurements (Rona, 1969, 1970) and confirmed by measurement of salinity gradients ( DSDP sites 139, 140; Waterman et al., 1972). Salt deposits beneath the Sao Paulo Pla- teau extend 700 km seaward from the coast. According to continental drift re- constructions of the Mesozoic opening of the Atlantic (Dietz and Holden, 1970), the extent of salt deposits off northwest Africa represents a half- width of opening. The extent of salt deposits of the Sao Paulo Plateau represents a full-width of opening. VI. Atlantic salt deposits exhibit a systematic distribution in time and space, as follows: 1. Late Silurian period: Eastern North America including the Michigan basin. 2. Mississippian period: Northwestern Atlantic including the Maritimes basin, Scotian shelf, and Grand Banks. 3. Late Permian period: Northeastern Atlantic including the North European basin and the North Sea. 4. Late Triassic and Jurassic periods: a. North Atlantic including the Grand Banks, Scotian shelf, Atlantic con- tinental margin of North America, Cuba, Bahama Banks, Senegal basin. Aaiun basin, offshore northwestern Africa. Essaouira basin, Portugal basin, Aquitaine basin, North Euro- pean basin, North Sea, and British Isles. b. Gulf of Mexico. c. Mediterranean region including the Atlas Mountains and the Algerian Sahara. 5. Aptian stage of the Cretaceous period: South Atlantic including the south- eastern continental margin of South America (Sergipe Alagoas, Reconcavo, Espirito Santo, Campos, and Santos basins), and the southwestern conti- nental margin of Africa (Mocamedes, Cuanza, Lower Congo, and Gabon basins). 394 272 PETER A EONA 6. Miocene epoch of the Tertiary period: Western Mediterranean Sea. 7. The age and geographic relations of inferred salt deposits beneath the De- merara Rise off the northeastern con- tinental margin of South America remain problematic. VII. Salt deposits of two different ages sepa- rated by intervals of nonsaliferous sedi- ments are superposed in at least two regions of the North Atlantic: 1 . Mississippean and Late Triassic through Jurassic salt deposits are superposed in the Scotian shelf-Grand Banks region. 2. Late Permian and Late Triassic salts are superposed in the North European basin — North Sea region. VIII. The distribution of Atlantic salt deposits in space and time (Fig. 1; Table 1) is genetically related to the independently determined history of opening of the North Atlantic in the Late Triassic and Jurassic (Rona, 1969, 1970; Schneider and Johnson, 1970; Pautot et al., 1970; Olson and Leyden, 1973), and the opening of the South Atlantic in the Early Cretaceous (Belmonte et al., 1965; Campos et al., 1974). ACKNOWLEDGEMENTS I thank Professor F. F. M. de Almeida, other members of the Organizing Committee, Professor H. Martin, and the Brazilian Academy of Sciences for the opportunity to participate in the International Symposium on Continental Margins of Atlantic Type. The research was supported by the National Oceanic and Atmos- pheric Administration (NOAA) as part of the Trans-Atlantic Geotraverse (TAG) project. REFERENCES ANTOINE, J. W., MARTIN, R., PYLE. T.. and BRYANT. W. R. — 1974 — Continental margins of the Gulf of Mexico, in Eurk. C. A., and Drake, C. L.. Geology of continental margins. New York. Springer Vet-lag, 1009 pp.. pp. 6S3-G94. AUDLEY-CHARLES. M. G. — 1970 — Triassic palaeo- geography of the British Isles. Geol. Soc. London Quart. Jour., v. 126, pp. 49-90. AUZENDE. J. M.. EONNIN, J., OLIVET, J. O., PAUTOT. G., and MAUFFRET, A. — 1971 — Upper Miocene salt layer in the Western Medi- terranean basin. Nature Phys. Sci., v. 230, pp. &2-S4. AYMe. J. M. — 1965 — The Senegal salt basin, in Salt Basins Around Africa. London, Institute of Petroleum, 112. pp., pp. S3-90. BALL, M. Jtf., LOCK, W.. HARRISON, C. G. A., NAGLE, F., JR., and WILLIAMS, G. J. — 1974 — Diapirs of the Old Bahama Channel. EOS, Amer. Geophys. Union Trans., v. 55, pp. 284. BALL, M, M., DASH. E. P., HARRISON, C. G. A., and AHMED, K. O. — 1971 — Refraction seismic measurements in the northeastern Bahamas (abs.). EOS, Trans. Amer. Geophys. Union, v. 52, n. 4, pp. 252. BARTLETT. G. A., and SMITH. L. — 1971 — Me- sozoie and Cenozoic history of the Grand Banks of Newfoundland. Canadian Jour. Earth Sci., v. 8, n. 1. pp. 65-S4. BAUMGARTNER, T. R., and VAN ANDEL, T. II, — 1971 — Diapirs of the continental margin of Angola, Africa. Geol. Soc. America Bull., v. 82, pp. 793-802. BELL, W. A. — 1958 — Possibilities for occurrences of petroleum reservoirs in Nova Scotia Dept. Mines, 277 pp. BELMONTE, Y., HIRTZ, P., and WENGER, R. — 1965 — The salt basins of the Gabon and the Congo (Brazzaville). A tentative paleogeographic interpretation, in Salt Basins Around Africa. Lon- don, Institute of Peti-oleum, 112 pp, pp. 55-74. BELT, E. S. — 1965 — Stratigraphy and paleogeo- graphy of Mabou Group' and related Middle Car- boniferous facies, Nova Scotia, Canada. Geol. Soc. America Bull., v. 76, pp. 777-802. EIDGOOD, D. E. T. — 1969 — The distribution and and diapiric nature of some Nova Scotia evaporites — a geophysical evaluation. 3rd Symp. on Salt, Cleveland, Ohio. BENZ.^A. — 1958 — Relations between oil fields and sedimentary troughs in Northwest German Basin. in Weeks, L. D., editor, Habitat of Oil. Tulsa, Amer. Assoc. Petroleum Geologists, 13S4 pp., pp. 1054-1066. BIJU-DUVAL, B., LETOUZEY, J., MONTADERT, L., COURRIER, P., MUGNIOT, J. F., and SANCHO, J. — 1974 — Geology of the Mediterranean Sea basins, in Burk, C. A., and Drake, C. L. editors, The geology of continental margins. New York, Springer- Verlag, 1009 pp., pp. 695-721. BONNARD, E., DEBOURLE, A., HLAUSCHEK, H., MICHEL, P., PEREBASKINE, V.. SCHOEFFLER, J., SERONIE-VIVIEN, R., and VIGNEAUX, M. — 1958 — The Aquitanian basin, southwest France, in Weeks, L. G., editor, Habitat of Oil. Tulsa, Amer. Assoc. Petroleum Geologists, 1384 p., pp. 1091-1122. BRINKMANN, R., and LoGTERS, H. — 1968 — Diapirs in Western Pyrenees and foreland Spain, in Brauns- teiw, J., and O'Brien, G. D., Diapirism and diapirs., a symposium. Tulsa, Amer. Assoc. Petroleum Geologists, Memoir 8, 444p., pp. 275-292. BROGNON, G. P., and VERRIER, G. R. — 1966 — Oil and geology in Cuanza basin of Angola, Amer. Assoc. Petroleum Geologists Bull., v. 50, pp. 108-158. BURK, C. A., EWING, M., WORZEL, J. L., BEALL, A. O., JR., BERGGREN, W. A.. BUKRY. D.. FISCHER, A. G., and PESSAGNO, E. A., JR. — 395 SALT DEPOSITS OF THE ATLANTIC 273 1969 — Deep-sea drilling' into the Challenger Knoll, central Gulf of Mexico. Amer. Assoc. Petroleum Geologists Bull., v. 53, pp. 1338-1347. BUROLLET, P., and DUMESTRE, A. — 1952 — Le diapir du Djebel Rhoeuis. Algeria, 19th Interna- tional Geological Congress, S.2, p. 6. BUTLER, L. W. — 1970 — Shallow structure of the continental margin, southern Brazil and Uruguay. Geol. Soc. America Bull., v. 81, pp. 1079-1096. CAMPOS, C. W. M., PONTE, F. O, MIURA, K. — 1974 — Geology of the Brazilian continental margin, in Burk, C. A., and Drake, C. L., editors, The geology of continental margins. New York, Springer- Verlag, 1009 pp., pp. 447-461. CLARK, S. O., JR., editor — 1966 — Handbook of physical constants: Geol. Soc. America, Memoir 97, 587 pp. DEMAISON, G. J. — 1965 — The Triassic Salt in the Algerian Sahara, in Salt basins around Africa. London, Institute of Petroleum, 122 pp., pp. 91-100. DIETZ, R. S. and HOLDEN, J. C. — 1970 — Recons- truction of Pangaea. Breakup and dispersion of continents, Permian to present. Jour. Geophys. Res., v. 75, pp. 4939-4956. DUPOUY-CAMET, J. — 1953 — Triassic diapiric salt structures, southwestern Aquitaine basin, France. Amer. Assoc. Petroleum Geologists Bull., v. 37, pp. 2348-2388. EMERY, K. O., UCHUPI, S., PHILLIPS, J. D., BOWEN, C. O., BUNCE, E. T., and KNOTT, S. T. — 1970 — Continental rise off eastern North America. Amer. Assoc. Petroleum Geologists, v. 54, pp. 44-108. EVANS, R. — 1970 — Sedimentation of the Mississipian evaporites of the Maritimes: an alternative model. Canadian Jour. Earth Sci., v. 7, pp. 1349-1352. FONSECA, J. I. — 1966 — Geological outline of the Lower Cretaceous Bahia supergroup, in Van Hinte, J. E., ed., Proc. 2nd West African Micropaleon- tological Colloquium, Ibadan, pp. 49-71. GUSSOW, W. C. — Carboniferous stratigraphy and structural geology of New Brunswick, Canada. Amer. Assoc. Petroleum Geologists Bull., v. 37, pp. 1713-1816. HODGMAN, C. D., WEAST, R. C, SHANKLAND, R. S., and SELBY, S. M., editors — 1963 — Handbook of chemistry and physics. Cleveland. Ohio, The Chemical Rubber Pub. Co., 3604 pp. HOWIE, R. O., and CUMMING, L. M. — 1963 — Basement features of the Canadian Appalachians. Geol. Surv. Canada, Bull. n. 89, 18 pp. JUX, V., 1961 — The palynologic age of diapiric and bedded salt in the Gulf coastal province. Louisiana Geol. Survey, Geol. Bull. 38, 46 pp. KEEN, M. J. — 1970 — A possible diapir in the Lau- rentian Channel. Canadian Jour. Earth Sci., v. 7, pp. 1561-1564. KENT, P. E. — 1969 — The geological framework of petroleum exploration in Europe and North Africa and the implications of continental drift hypotheses, in Hepple, P., editor. The exploration for petroleum in Europe and North Africa. London, Institute of Petroleum, pp. 3-17. KING, L. H., and MAC LEAN, B. — 1970 — A diapiric structure near Sable Island-Scotian shelf. Maritime Sediments, v. 6, pp. 1-4. KIRKLAND, D. C, and GERHARD, J. E. — 1971 — Jurassic salt, central Gulf of Mexico, and its temporal relation to circum-Gulf evaporites. Amer. Assoc. Petroleum Geologists Bull., v. 55, pp. 680-686. KR6MMELBEIN, K, and WENGER, R. — 1966 — Sur quelques analogies remarquables dans les miero- faunes cretacees du Gabon et du Bresil Oriental (Bahia et Sergipe), in Reyre, D., editor. Sedi- mentary basins of the African coasts, Part 1, Atlantic Coast. Paris, Association of African Geo- logical Surveys, 304 pp., pp. 193-196. LACOSTE, J. — 1934 — Etudes gcologiques dans le Rif Meridional. Protectorat Repub. Francaise au Maroc, Serv. Mines, Notes et Mem. 31 and 32, v vols., 660 pp. LANDES, K. K., EHLERS, G. M., and STANLEY, G. M. — 1945 — Geology of the MacKinac Straits Region. Michigan Geol. Survey, Pub. 44, Geol. Ser. 37, pp. 192. LANDES, K. K. — 1945 — Preliminary map no 40, The Salina and Bass Island rocks in the Michigan basin. U. S. Geol. Survey, Oil and Gas Inv. LEYDON, R., BRYAN, G., and EWING, M. — 1972 — Geophysical reconnaissance on African shelf: 2. Margin sediments from Gulf of Guinea to Walvis Ridge. Amer. Assoc. Petroleum Geologists Bull., v. 56, pp. 682-693. LEYDEN, R., DAMUTH, J. E., ONGLEY, L. K, KOSTECKI, J., and VAN STEVENICK, W. — 1975 — Salt diapirs on the Sao Paulo Plateau, Brazilian continental margin. Geol. Soc. America, Abstracts with programs, v. 7. n. 7, 1350 pp., pp. 1168-1169. MAHER, J. C, and APPLIN, E. R. — 1971 — Geologic framework and petroleum potential of the Atlantic coastal plain and continental shelf. U. S. Geol. Survey Prof. Paper 659, 98 pp. MANHEIM, F. T.. and BISHOFF, J. L. — 1969 — Geochemistry of pore waters from Shell Oil Com- pany drill holes in the continental slope of the northern Gulf of Mexico. Chemical Geol., v. 4, pp. 63. MANHEIM, F. T., SAYLES, F. L., and WATERMAN. L. S. — 197.'! — Interstitial water studies on small core samples, Deep-Sea Drilling Project: Leg 10, in Worzel, J. L., Bryant, W., ct al., Initial Reports of the Deep Sea Drilling Project, v. 10. Washington, D. C, U. S. Government Printing Office, 748, pp. 615-623. MANHEIM, F. T., and HORN, M. K. — 1968 — Com- position of deeper subsurface waters along the Atlantic continental margin. Southeastern Geology, v. 9, pp. 215-236. MATTICK, R. E., FOOTE, R. Q., WEAVER, N. L.. and GRIM, M. S. — 1974 — Structural framework of United States Atlantic outer continental shelf north of Cape Hatteras. Amer. Assoc. Petroleum Geologists Bull., v. 58, Part II of II, pp. 1179-1190. MAUFFRET, A., FAIL, J. P., MONTADERT, L., SANCHO. J., and WINNOCK, E. — 1973 — North- western Mediterranean sedimentary basin from seismic reflection profile. Amer. Assoc. Petroleum Geologists Bull., v. 57, pp. 2245-2262. MC IVER, N. L. — 1972 — Cenozoic and Mesozoic stratigraphy of the Nova Scotia shelf. Canadian Jour. Earth Sci., v. 9, pp. 54-70. 396 274 PETER A. RONA MEYERHOFF, A. A., and HATTEN, C. W. — 1968 — Diapiric structures in central Cuba, in Braunstein, J., and O'Brien, G. D., editors, Diapirism and diapirs, a symposium. Tulsa, Amer. Assoc. Pe- troleum Geologists, Memoir 8, 444 pp., pp. 315-357. MONTADERT, L., WINNOCK, E., DELTEIL, J. R., and GRAU, G. — 1974 — Continental margins of Galicia-Portugal and Bay of Biscay, in Burk, C. A., and Drake, C. L., editors, The geology of conti- nental margins. New York, Springer- Verlag, 1009 pp., pp. 323-342. MURRAY, G. E. — 1968 — Salt structures of Gulf of Mexico basin — a review, in Braunstein, J., and O'Brien, G. D., editors, Diapirism and diapirs, a symposium. Tulsa, Amer. Assoc. Petroleum Geologists, Memoir 8, 444 pp., pp. 99-121. NETTLETON, L. L. — 1934 — Fluid mechanics of salt domes. Amer. Assoc. Petroleum Geologists Bull., v. 18, pp. 1175-1204. OFFSHORE EXPLORATION STAFFS, AMOCO CA- NADA PETROLEUM COMPANY LTD., AND IMPERIAL OIL LTD. — 1974 — Regional geology of Grand Banks. Amer. Assoc. Petroleum Geolo- gists, v. 58, Pt. II of II, pp. 1109-1123. OLSON, W. S., and LEYDEN, R. J. — 1973 — North Atlantic rifting in relation to Permian-Triassic salt deposition; in Logan, A., and Mills, D. V. editors, The Permian and Triassic systems and their mutual boundary. Canadian Soc. Petroleum Geologists, Memoir 2, 766, pp., 720-732. OLSON, W. S. — 1974 — Structural history and oil potential of offshore area from Cape Hatteras to Bahamas. Amer. Assoc. Petroleum Geologists Bull., v. 58, Part II of II, pp. 1191-1200. PARKER, T. J., and MC DOWELL, A. N. — 1955 — Model studies of salt-dome tectonics. Amer Assoc. Petroleum Geologists Bull., v. 39, pp. 2384-2470. PAMENTER, B. — 1971 — Sands aplenty in Onondaga well. Oilweek, 22 November, pp. 22-24. PAUTOT, G., AUZENDE, J. M., and LE PICHON, X. — 1970 — Continuous deep salt layer along North Atlantic margins related to early phase of rifting, Nature, v. 227, pp. 351-354. QUEROL, R. — 1966 — Regional geology of the Spanish Sahara, in Reyre. D., editor, Sedimentary basins of the African coasts. Part 1, Atlantic coast. Paris, Association of African Geological Surveys, 304 pp., pp. 27-38. REYRE, D. — 1966 a — Particularites geologiques des Bassines de l'Ouest African (essai de recapitulation), in Reyre, D., editor, Sedimentary basins of the African coasts, Part I, Atlantic coast. Paris, Association of African Geological Surveys, 304 pp., pp. 253-304. REYRE, D. — 1966 b — Recapitulation generale : Les formations permiennes, Triassiques, Jurassiques et Cretacees dans les bassins cotiers Ouest-Africains, in Reyre, D., editor, Sedimentary basins of the African coasts, Part I, Atlantic coast. Paris. Asso- ciation of African Geological Surveys, 304 pp.. 3 tables. RONA, P. A. — 1969 — Possible salt domes in the deep Atlantic off North-west Africa. Nature, v. 224, pp. 141-143. RONA, P. A. — 1970 — Comparison of continental margins of eastern North America at Cape Hatteras and northwestern Africa at Cap Blanc. Amer. Assoc. Petroleum Geologists Bull., v. 54, pp. 129-157. RYAN, W. B. F., HSU, K. J., ct al. — 1973 — Initial reports of the Deep Sea Drilling Project, v. XIII, Parts 1 and 2. Washington, D. C, U. S. Gov't: Printing Office, 1447 pp. SANNEMANN, D. — 1968 — Salt-stock families in northwestern Germany, in Braunstein, J., and O'Brien, G. D., editors. Diapirism and diapirs, a symposium. Tulsa, Amer. Assoc. Petroleum Geologists, Memoir 8, 444 pp., pp. 261-270. SCHNEIDER, E. D., and JOHNSON, G. L. — 1970 — Deep-ocean diapir occurrences. Amer. Assoc. Pe- troleum Geologists, v. 54, pp. 2151-2169. SCIENTIFIC PARTY — 1975 — Basins and margins of the eastern South Atlantic, Leg 40 of the Deep Sea Drilling Project. Geotimes, v. 20, n. 6, pp. 22-24. SHERIDAN, R. E. — 1974 — Atlantic continental margin of North America in Burk, C. A., and Drake, C. L., editors, The geology of continental margins. New York, Springer-Verlag, 1009 pp.. pp. 391-407. SOCIeTe CHERIFIENNE DES PeTROLES — 1966 — Le bassin du sud-ouest Marocain, in Reyre, D., editor, Sedimentary basins of the African coasts, Part I, Atlantic coast. Paris, Association of African Geological Surveys, 304 pp., pp. 5-12. SWAIN, F. M. — 1952 — Ostracoda from wells in North Carolina. 2. Mesozoic ostracoda. U. S. Geol. Survey, Prof. Paper 234-B, pp. 59-192. TEMPLETON. R. S. M. — 1971 — The geology of the continental margin between Dakar and Cape Palmas, in Delany, F. M., editor, ICSU/SCOR Working Party 31 Symposium. Cambridge 1970. The geology of the East Atlantic continental margin. 2. Africa. Rep. N» 70/16, Inst. Geol. Sci., 209 pp., pp. 44-60. VIANA, C. F. — 1971 — Revisao estratigrafica da Bacia do Reconcavo-Tucano. Bol. Tec. Petrobras, Rio de Janeiro, v. 14, n. 3/4, pp. 157-192. VIANA, C. F. — 1966 — Stratigraphic distribution of ostracoda in the Bahia supergroup (Brazil), in Van Hinte, J. E., ed., Proc. 2nd West African Micropaleontogical Colloquium, Ibadan. pp. 240-256. VON HERZEN, R. P.. HOSKINS, H, and VAN ANDEL, T. H. — 1972 — Geophysical studies in the Angola diapirs field. Geol. Soc. America Bull, v. 83, pp. 1901-1910. WATERMAN, L. S., SAYLES, F. L., and MANHEIM. F. T. — 1972 — Interstitial water studies on small core samples, Leg 13, in Hayes, D. E., Pimm, A. C, et al.t Initial reports of the Deep Sea Drilling Project, v. XIV. Washington, U. S. Go- vernment Printing Office, 975 pp., pp. 753-762. WATSON, J. A., and JOHSON, G. L. — 1970 — Seismic studies in the region adjacent to the Grand Banks of Newfoundland. Canadian Jour. Earth Sci., v. 7, pp. 306-316. WEBB, G. W. — 1973 — Salt structures east of Nova Scotia. Earth Science Symposium on Offshore Eastern Canada, Geol. Survey Canada, Paper 71-23, pp. 197-218. WINNOCK, E. — 1973 — Expose sussinct de revo- lution paleogeologique de l'Aquitaine\ Bull. Soc. G6ol. France, 7eSerie, v. XV, pp. 5-12. ZIEGLER, P. A. — 1975 — Geologic evolution of the North Sea and its tectonic framework. Amer. Petro- leum Geologists, v. 59, pp. 1073-1097. 397 43 Reprinted from: Journal of Oaean Management, Vol. 3, 57-78, Ocean Management, 3 (1976) 57—78 © Elsevier Scientific Publishing Company, Amsterdam — Printed in The Netherlands Energy and Mineral Resources of the Pacific Region in Light of Plate Tectonics Peter A. Rona * and Lawrence D. Neuman ** ABSTRACT The Pacific is a closing ocean basin that is diminishing in size as it is consumed at convergent plate boundaries around three-fourths of its perimeter. Geothermal energy sites, areas of offshore petroleum potential, deposits of precious, base, iron and ferro- alloy metals are distributed along the convergent plate boundaries of the Pacific including the surrounding continents. The energy and mineral resources of the Pacific region are concentrated by geologic processes at the convergent plate boundaries. INTRODUCTION The Pacific is a region of geologic diversity. The Pacific region encom- passes the largest ocean basin on earth, extensive chains of volcanic islands that follow arcuate trends around the western margin of the Pacific Ocean, and seas occupying marginal basins between the island arcs and eastern Asia (Fig. 1). The theory of plate tectonics has gained wide scientific acceptance during the past five years, and offers a conceptual framework to unify the diverse geological phenomena of the Pacific region. The conceptual frame- work of plate tectonics is leading to a new understanding of the relation be- tween the geology and the distribution of energy and mineral resources of the Pacific region. An earlier paper treated principles of the relation between plate tectonics and mineral resources (Rona, 1973). This paper aims to apply these princi- ples to develop a basic understanding of the distribution of energy and min- * National Oceanic and Atmospheric Administration (NOAA), Atlantic Oceanographic and Meteorological Laboratories, Miami, Fla. 33149, U.S.A. ** Office for Ocean Economics and Technology, United Nations, New York, N.Y. 10017, U.S.A. 57 398 eral resources of the Pacific region. The approach taken is first to present salient features of the geology of the Pacific region from the point of view of plate tectonics (Figs. 1—3). Then to view the distribution of selected energy and mineral resources of the Pacific region with respect to the geology in a series of maps (Figs. 4—9) compiled from various sources (Van Roy an and Bowles, 1952; ECAFE, 1962, 1963, 1970; McKelvey and Wang, 1969; Dr. Peter A. Rona is presently Senior Research Geophysicist with the NO A A, Atlantic Oceanograph- ic and Meteorological Laboratories in Miami, Florida and Adjunct Professor of Marine Geology and Geo- physics at the University of Miami. He is Chief Scien- tist of the Trans- Atlantic Geotraverse (TAG), an international cooperative project to investigate the Earth's crust along a corridor across the Atlantic be- tween southeastern North America and northwestern Africa to gain an understanding of continental drift, sea-floor spreading, and the occurrence of seabed minerals. He generally spends several months of the year at sea leading explorations of the seabed. Prior to joining NOAA in 1969 he was Exploration Geologist with Standard Oil Company (N.J.) between 1957 and 1959. From 1960 to 1969 he was with Columbia University, Hudson Laboratories where he developed and became Head of Marine Geophysics. He received the degree of Ph.D. in 1967 from the Department of Geology and Geophysics at Yale Uni- versity. He has published about 50 scientific papers and is a member of 12 professional societies. Lawrence D. Neuman joined the Ocean Economics and Technology Office as Scientific Affairs Officer in 1973, specializing in the economic potential of the sea and the economic development of coastal areas. A native New Yorker, he received his bachelor's degree in physics at Columbia College and his doctorate in geology and marine geophysics at Columbia Univer- sity's Lamont-Doherty Geological Observatory. 58 399 Anon., 1972; DEMR, 1972; Jones, 1972; Eimon, 1974). Finally, to attempt to understand the distribution of the energy and mineral resources in terms of geologic processes (Fig. 10). Attention is focused on those energy and mineral resources associated with present plate boundaries of the Pacific region. Other deposits may then be interpreted in terms of past plate bound- aries following the uniformitarian principle of geology that the present is the key to the past. GEOLOGY OF THE PACIFIC Lithospheric plates The conceptual framework of plate tectonics, developed by many work- ers, views the earth as comprised of a rigid outer shell about 100 km (60 miles) thick, the lithosphere, that behaves as if it were floating on an under- lying plastic layer, the asthenosphere (Fig. 2). The upper, more brittle part of the lithosphere is termed crust, of which there are two major types, the granitic continental crust (about 30 km thick) and the basaltic oceanic crust LITHOSPHERIC PLATES 120*e 150* 180* 150* 120" 90* 60* DIVERGENT PLATE BOUNDARY ... CONVERGENT PLATE BOUNDARY _ TRANSFORM PLATE BOUNDARY __ UNCERTAIN PLATE BOUNDARY HALF RATE OF * * SEA FLOOR SPREADING (CM/YR) „ RELATIVE PLATE MOTION (CM/YR) „. DIP OF BENIOFF ZONE (UPPER 100 KM) Fig. 1. Lithospheric plates of the Pacific region showing directions (arrows) and half-rates of sea-floor spreading about divergent plate boundaries in the eastern Pacific, directions (arrows) and rates of convergence at convergent plate boundaries bordering the Pacific ocean basin, and angles of inclination (in degrees from horizontal) of Benioff zones beneath the convergent plate boundaries (Fig. 2). (Le Pichon et al., 1973) 59 400 (about 10 km thick). The lithosphere is segmented into a number of major plates, each of which may encompass a continent and part of an ocean basin, and numerous minor plates. The Pacific region includes portions of the Pa- cific, China, America, and Antarctic major plates, and several minor plates (Fig. 2). PLATE BOUNDARIES The boundaries of lithospheric plates are delineated by narrow earthquake zones where the plates are moving with respect to each other. Two types of boundaries are considered (Fig. 2). At the first type, a divergent plate bound- ary, two adjacent plates move apart as new lithosphere is added to each plate by the process of sea-floor spreading. Divergent plate boundaries extend around the globe through all the major ocean basins as part of a 65,000 km- (40,000 mile-) long undersea mountain chain. Divergent plate boundaries of the Pacific region including the East Pacific Rise are located in the eastern Pacific ocean basin off South America, Central America, and North America (Fig. 1). The sea floor is spreading about different segments of the East Pa- cific Rise at rates ranging between about 1 and 10 cm per year (Fig. 1). Fig. 2. Diagram showing plate motions at divergent and convergent plate boundaries. Lithospheric plates move like conveyor belts from a divergent plate boundary (oceanic ridge) to a convergent plate boundary where they either descend along a Benioff zone at an oceanic trench (subduction) or they override the adjacent plate (obduction). 60 401 At the second type of boundary, a convergent plate boundary, two adja- cent plates come together. In the general case, one plate descends under another plate along an inclined plane (Benioff zone) and is resorbed into the asthenosphere (subduction; Fig. 2). In the special case, one plate may temporarily override the other plate (obduction) until the situation reverts to subduction. The Pacific is bounded on three sides by convergent plate boundaries marked by oceanic trenches where lithosphere descends along Benioff zones at rates comparable to the rates of sea floor spreading. As a consequence of the crustal consumption at the convergent plate boundaries bounding the Pacific, the Pacific is a closing ocean basin that is diminishing in size, in contrast to the Atlantic which is an opening ocean basin that is growing larger. The coexistence of divergent plate boundaries where litho- sphere is created, and convergent plate boundaries where lithosphere is de- stroyed, implies that the diameter of the earth is not radically changing. The inclination of Benioff zones at convergent plate boundaries plays an important role in the development of basins marginal to continents and the generation of volcanism. The inclination of a Benioff zone is inversely pro- portional to the rate of convergence of adjacent plates at a convergent plate boundary (Luyendyk, 1970). Marginal basins are present in the western Pacific where the rates of plate convergence are relatively slow and the incli- nation of the Benioff zone exceeds about 35° (Fig. 1; Karig, 1971; Oxburgh and Turcott, 1971; Sleep and Toksoz, 1971; Bracey and Ogden, 1972). Mar- ginal basins are absent in the eastern Pacific where the rates of plate conver- gence are relatively fast and the inclination of the Benioff zone is less than about 35°. As will be shown, the presence or absence of marginal basins af- fects offshore petroleum potential. The inclination of Benioff zones also affects the composition of igneous rocks (rocks solidified from molten mate- rial) and associated metal deposits. AGE OF SEA FLOOR AND OF CONTINENTS The range and distribution of ages differs markedly between the Pacific Ocean basin and the surrounding continents (Fig. 3). The age of the Pacific sea floor, determined by dating of rock samples recovered by the Deep Sea Drilling Project (Fischer et al., 1971) and by the magnetic polarity reversal time scale (Pitman et al., 1974), ranges between about 150,000,000 years (Late Jurassic period) and the present. The distribution of the ages about the divergent plate boundaries in the eastern Pacific is regular, the sea floor being youngest adjacent to the boundaries and becoming progressively older away from the boundaries as a consequence of sea-floor spreading. The distribu- tion of ages at the convergent plate boundaries around the Pacific is irregu- 61 402 120'E 150* 1B0* 150* 120* 90* 60* 30*^ AGE OF SEA FLOOR % CENOZOIC ::: MESOZOIC CONTINENTAL STRUCTURAL PROVINCES % CENOZOIC • ::: MESOZOIC v. HERCYNIAN lllll CALEDONIAN PRECAMBRIAN 'CONVERGENT PLATE BOUNDARY = DIVERGENT PLATE BOUNDARY Fig. 3. Age of the Pacific sea floor and of continental structural provinces. Lithospheric plate boundaries are shown. lar, the age of the sea floor varying along the boundaries as a consequence of subduction. The circum-Pacific continents are divided into structural provinces accord- ing to the time of their most recent deformation during mountain building episodes (Fig. 3). Structural provinces are predominantly Cenozoic (0— 70,000,000 years) in western South America, and Mesozoic (70,000,000— 200,000,000 years) in western North America. The island arcs of the western Pacific are Cenozoic (0—70,000,000 years). Structural provinces of eastern Asia and Australia exhibit a more complex distribution spanning nearly the entire age range of the earth. The ages of most circum-Pacific metal deposits range from Mesozoic through Cenozoic (0—200,000,000 years ago), corre- sponding to the known history of lithospheric plate motions in the Pacific Ocean basin. DISTRIBUTION OF SELECTED ENERGY RESOURCES Geothermal Energy Geothermal phenomena including active volcanos, thermal springs, fuma- roles, geysers, and high values of heat flow are distributed along convergent 62 403 GEOTHERMAL ENERGY 120'E ISO" 180" 150" 120" 90" 60" 30* V '.'■>//////! 1H\\\ ■ GEOTHERMAL ENERGY SITES * ACTIVE VOLCANOES • THERMAL SPRINGS AND FUMAROLES :SS NUMEROUS THERMAL SPRINGS a GEYSERS HEAT FLOW •.••V MEAN HEAT FLOW > 3.0 mi MEAN HEAT FLOW > 2.0 m MEAN HEAT FLOW < 2.0 III! MEAN HEAT FLOW < 1.0 CONVERGENT PLATE BOUNDARY DIVERGENT PLATE BOUNDARY Fig. 4. Map of geothermal energy of the Pacific region. Lithospheric plate boundaries are shown. plate boundaries around the Pacific (Fig. 4; Kennedy and Richey, 1947; Waring, 1965; Karig, 1971; Snead, 1972). These geothermal phenomena re- sult from heating due to mechanical factors (friction), chemical reactions (dehydration), and to the internal heat of the earth, as the lithospheric plates descend into the asthenosphere at convergent plate boundaries. Similar geo- thermal phenomena are inferred to occur along the East Pacific Rise and the divergent plate boundaries of the Pacific region where heat flow is high, as a consequence of the upwelling of magma during sea-floor spreading (Lang- seth, 1969;Sclater, 1972). Geothermal energy is being tapped at sites in the western United States, Japan, and New Zealand. In addition to its direct utilization, geothermal energy drives hydrothermal processes involving the circulation of hot solu- tions through the lithosphere which act to concentrate metals both at diver- gent and convergent plate boundaries, as will be discussed. Hydrothermal mineral deposits, that is, mineral deposits precipitated from hot aqeous solu- tions, constitute a major part of useful metallic ores on continents and may be important in ocean basins. 63 404 Organic energy: petroleum Areas of offshore petroleum potential conform with convergent plate boundaries around the Pacific (Fig. 5; McKelvey and Wang, 1969). Both the circum-Pacific trenches and the island arcs of the western Pacific create a habitat that is favorable for the accumulation of petroleum in several re- spects. The trenches and island arcs act as barriers that catch sediment and organic matter from the continent and ocean basin. Deep-sea sediment with variable content of organic matter is continuously transported into the trenches on a conveyor belt of spreading sea floor (Sorokhtin et al., 1974). The island arcs divide the ocean basin into marginal basins such as the South China Sea, the East China Sea, the Yellow Sea, the Sea of Japan, the Sea of Okhotsk, and the Bering Sea. The shape of the trenches and marginal basins acts to restrict the circulation of the ocean, so that oxygen is not replenished in the seawater and the organic matter is preserved. Geothermal heat in the trenches and marginal basins may facilitate the conversion of organic matter to petroleum (Fig. 4; Tarling, 1973; La Plante, 1974). Finally, geological structures that develop as a result of deformation of the sediment in the trenches and marginal basins by tectonic forces form traps that favor the accumulation of petroleum. In contrast to the areas of offshore petroleum ORGANIC ENERGY 150° 180* 150" 120* 90° 60" 30* \ PETROLEUM PRODUCING AREAS ONSHORE PETROLEUM POTENTIAL OFFSHORE PETROLEUM POTENTIAL SEDIMENTARY ROCKS %& CRYSTALLINE ROCKS CONVERGENT PLATE BOUNDARY = DIVERGENT PLATE BOUNDARY Fig. 5. Areas of petroleum potential and production of the Pacific region (adapted from Rona and Neuman, 1974, 1975; after McKelvey and Wang, 1969). 64 405 potential, the sedimentary basins from which petroleum is produced on con- tinents around the Pacific exhibit no apparent spatial relation to plate bound- aries (Fig. 5; Irving et al., 1974; Rona and Neuman, 1974). DISTRIBUTION OF SELECTED MINERAL RESOURCES Metal deposits at divergent plate boundaries Knowledge of the distribution of metal deposits with respect to divergent plate boundaries is limited because, as submerged oceanic ridges, these boundaries are less accessible to observation than convergent plate bounda- ries. This knowledge is necessary to evaluate the metallic mineral potential of oceanic lithosphere. All oceanic lithosphere is generated by sea-floor spread- ing about oceanic ridges and underlies all ocean basins which cover two- thirds of the earth. The Red Sea and the Atlantic Ocean provide evidence of the nature of processes that may be concentrating metals in oceanic litho- sphere of the Pacific. Evidence from the Red Sea and the Mid-Atlantic Ridge indicates that hydrothermal processes are concentrating metals in oceanic crust at diver- gent plate boundaries. The Red Sea represents the earliest stage in the growth of an ocean basin, the stage when a divergent plate boundary rifts a continent in two. About five years ago the richest known submarine metallic sulfide deposits were found in basins along the center of the Red Sea at a depth of about 2,000 m (6,600 ft.) below sea level (Degens and Ross, 1969). The sulfide minerals, in which various metals are combined with elemental sulfur, are disseminated in sediments that fill the basins to a thickness esti- mated between 20 m (66 ft.) and 100 m (330 ft.). The top 10 m (33 ft.) of sediment, which has been explored by coring the largest of the basins, has a total dry weight of about 80 million tons, with average metal contents of 29% iron, 3.4% zinc, 1.3% copper, 0.1% lead, 0.005% silver, and 0.00005% gold (Bischoff and Manheim, 1969). The deposits are saturated with (and overlain by) salty brines carrying the same metals in solution as those present in the sulfide deposits. The salty brines are considered to be the hydrother- mal solutions from which the sulfide minerals are precipitated. The most advanced growth stage of a divergent plate boundary is the oceanic ridge system including the Mid -Atlantic Ridge and the East Pacific Rise. An active area of submarine hydrothermal mineral deposits, the TAG Hydrothermal Field, was recently discovered at the crest of the Mid- Atlantic Ridge (26°N) by the Trans-Atlantic Geotraverse (TAG) project of the Na- tional Oceanic and Atmospheric Administration (NOAA). As the first of its kind discovered, the distribution of such hydrothermal fields along oceanic 65 406 ridges is unknown, but it is suspected that the TAG Hydrothermal Field may represent an important class of features. The TAG Hydrothermal Field includes both active and relict areas (Rona et al., 1976). The active area (15X15 km), including the east wall of the rift valley between depths of 2000 and 3500 m, is covered by a discontinuous layer of manganese oxide at least 5 cm (2 in.) thick (R. Scott et al., 1974; McGregor and Rona, 1975), that is being deposited by hydrothermal solu- tions enriched in various metals (Betzer et al., 1974). The hydrothermal so- lutions emanate as hot springs from fractures in the ocean bottom (Rona et al., 1975). The relict area comprises hydrothermal material that was deposit- ed in the active area adjacent to the rift valley and transported at least tens of kilometers away from the ridge crest on a conveyor belt of spreading sea floor (Rona, 1973). A hydrothermal origin for the metallic oxide present is indicated by its chemical purity (40% manganese with only trace quantities of iron and copper compared with manganese nodules which generally con- tain about 10% manganese and appreciable quantities of iron and copper), and rapid rate of accumulation (about 200 mm per 1,000,000 yr. which is about one hundred times faster than manganese nodules) (M. Scott et al., 1974). The TAG Hydrothermal Field not only confirms that metals are con- centrated in normal oceanic crust by hydrothermal processes, but indicates that such processes may occur at a divergent plate boundary more-or-less continuously from early (Red Sea) to advanced (Mid-Atlantic Ridge) stages of growth. Sediment samples directly overlaying the basalt that forms the foundation of the Pacific and other ocean basins recovered by the Deep Sea Drilling Pro- ject both at and away from oceanic ridges, reveal widespread enrichment by certain precious, base, iron and ferro-alloy metals (Bostrom and Peterson, 1969; Dymond et al., 1970; Von der Borch and Rea, 1970; Von der Borch et al., 1971; Cook, 1972; Piper, 1973; Sayles and Bischoff, 1973; Dasch, 1974). The observation that the enrichment is limited to sediment in the basal layer directly overlying basalt indicates that it occurred soon after the generation of the underlying basalt by sea-floor spreading about an oceanic ridge. The metal enrichment is ascribed to hydrothermal processes similar to those that produced the metalliferous sediments in the Red Sea and the metallic oxides at the TAG Hydrothermal Field. The concentration of metals in the widespread enriched sediments of the Pacific is only a fraction of that observed in the Red Sea, but higher concentrations may exist locally. Processes of metal concentration at divergent plate boundaries A model of metallogenesis at divergent plate boundaries, based on various lines of evidence (Spooner and Fyfe, 1973), considers that certain precious, 66 407 base, iron and ferro-alloy metals may be concentrated as deposits by sub-sea floor hydrothermal convection systems involving the circulation of seawater as a hydrothermal solution through rocks to a depth of about 5 km (Hart, 1973) beneath the ocean bottom. The development of such hydrothermal convection systems is favored by the supply of seawater, heat, and the in- tensely fractured basaltic rocks at divergent plate boundaries. According to the model, cold, dense seawater descends through fractures in the basalt of an oceanic ridge and is heated by contact with hot, intrusive bodies of mag- ma (molten rock material) and rock that upwell to form new lithosphere at the ridge crest. The warm, less dense seawater rises through the features and leaches metals disseminated in the basalt that are then transported in solu- tion as complexes with chlorides in the seawater. A fraction of the metals in solution combines with sulfur in the seawater and precipitates to form mas- sive statiform bodies of metallic sulfide including copper and iron, possibly associated with gold. It is suspected that such copper-iron sulfide bodies may underlie the TAG Hydrothermal Field, but it is technically infeasible at present to drill into the ocean bottom to test this idea (Rona, 1973; R. Scott et al., 1974; Rona et al., 1976). Metallic oxides, like the manganese oxide at the TAG Hydrothermal Field, precipitate under oxidizing conditions as the hydrothermal solutions discharge from the ocean bottom in hot springs. Amorphous particles of ferric hydroxide precipitate from the hydrothermal solutions in the overlying seawater. The ferric hydroxide scavenges the re- maining metals from solution and settles to deposit a layer of metalliferous sediment on basalt of the ocean bottom, like the metalliferous sediments ob- served in the Red Sea and the Pacific Ocean. Metal deposits at convergent plate boundaries Precious-metal deposits including gold, silver, and platinum are distributed along convergent plate boundaries around the Pacific Ocean (Fig. 6). In the eastern Pacific precious-metal deposits occur landward of convergent plate boundaries along the western margins of North America and South America. In the western Pacific, precious-metal deposits occur on island arcs situated along convergent plate boundaries including Japan, the Philippines, and In- donesia. Deposits are also present in eastern Asia and Australia, where they are separated by a gap from the active convergent plate boundaries. The distribution of light-metal deposits including aluminum, beryllium, lithium, and titanium appears unrelated to plate boundaries of the Pacific (Fig. 7). These metals are associated with granitic rocks of the continents that are compositionally different from the basaltic rocks of the ocean basin. The distribution of light-metal deposits on the circum-Pacific continents is 67 408 PRECIOUS METAL DEPOSITS A SILVER • COLD i PLATINUM CONVERGENT PLATE BOUNDARY =• DIVERGENT PLATE BOUNDARY Fig. 6. Map of precious-metal deposits of the Pacific region (adapted from Rona and Neuman, 1974, 1975). Lithospheric plate boundaries are shown. LIGHT METAL DEPOSITS 120°E 150" 180° 150* 120* 90* 60* 30** ■ ALUMINUM * BERYLIUM a LITHIUM • TITANIUM CONVERGENT PLATE BOUNDARY = DIVERGENT PLATE BOUNDARY Fig. 7. Map of light-metal deposits of the Pacific region. Lithospheric plate boundaries are shown. 68 409 BASE METAL DEPOSITS 120°E 150° 180" 150° 120" 90° 60° 30"W * ANTIMONY • COPPER ' LEAD d MERCURY " TIN o ZINC * * * 0 CONVERGENT PLATE BOUNDARY .DIVERGENT PLATE BOUNDARY Fig. 8. Map of base-metal deposits of the Pacific region (adapted from Rona and Neuman, 1974, 1975). Lithospheric plate boundaries are shown. related to the occurrence of particular minerals in granitic rocks and to the concentration of these minerals by processes of subaerial weathering. Base-metal deposits including antimony, copper, lead, mercury, tin, and zinc are distributed along convergent plate boundaries around the Pacific Ocean (Fig. 8), similar to the distribution of precious metal deposits (Fig. 6). The base-metal deposits occur landward of convergent plate boundaries along the western margins of the Americas in the eastern Pacific, and on is- land arcs along convergent plate boundaries of the western Pacific. In south- east Asia deposits of tin associated with tungsten, fluorite, bismuth, and molybdenum occur in belts of granites of predominantly Mesozoic age (70,000,000-200,000,000 years ago). Base-metal deposits also occur in east- ern Asia and Australia where they are separated by a gap from active conver- gent plate boundaries. The copper occurs associated with other metals along convergent plate boundaries of the Pacific region in two economically important classes of ore deposits — massive statiform sulfide bodies and porphyry ore bodies. 69 410 Massive statiform sulfide bodies, deposits confined to layers within bedded sequences of volcanic or sedimentary rocks, are present in western North America, Japan, and the Philippines (Eimon, 1974) where they range in age from Paleozoic through Cenozoic (0—500,000,000 years old). Porphyry ore bodies, disseminated deposits of copper-sulfide minerals associated with vol- canic rocks, constitute over one-half of the world's copper production. The majority of porphyry copper deposits lie in two belts of the Pacific region (Eimon, 1974): 1) the western Americas belt extending from Chile to Alaska where the deposits are Mesozoic and Cenozoic in age (0—200,000,000 years old), 2) the southwest Pacific belt including Taiwan, the Philippines, Borneo, West Siam, New Guinea (Papua), and the Solomon Islands, where the depos- its are Cenozoic in age (0—70,000,000 years old). Iron and ferro-alloy metal deposits including chromium, cobalt, manga- nese, molybdenum, nickel, tungsten, and vanadium are distributed along convergent plate boundaries around the Pacific (Fig. 9), similar to the distri- bution of precious (Fig. 6) and base (Fig. 8) metals. Iron and ferro-alloy metal deposits occur landward of convergent plate boundaries along the western margins of the Americas in the eastern Pacific, and on island arcs IRON AND FERROALLOY METAL DEPOSITS • IRON D CHROMIUM' g COBALT O MANGANESE * MOLYBDENUM a NICKEL ■ TUNGSTEN a VANADIUM MANGANESE NODULES UP TO 20% OF BOTTOM = 20-50% OF BOTTOM BOUNDARY BETWEEN RED CLAY & BIOGENIC OOZES CD COPPER CONTENT IN WEIGHT 1.5-2.0% © NICKEL CONTENT 1.5-2.0% 'CONVERGENT PLATE BOUNDARY = DIVERGENT PLATE BOUNDARY Fig. 9. Map of iron and ferro-alloy metal deposits including manganese nodules of the Pacific region (adapted from Rona and Neuman, 1974, 1975). Lithospheric plate bound- aries are shown. 70 411 along convergent plate boundaries of the western Pacific. These deposits also occur in eastern Asia and Australia where the deposits are separated by a gap from active convergent plate boundaries. Nodules containing variable percentages of manganese, copper, nickel, and cobalt are present on about two-thirds of the Pacific sea floor (Fig. 9; Strak- hov et al., 1968; Horn et al., 1972; Skornyakova and Andrushchenko, 1974). A zone of nodules of anomalously high copper and nickel content (1.5— 2.0%) and areal density (20—50% of sea floor) extends east- west across the central Pacific between latitutes 5° and 20° N coinciding with a region of high biological productivity and may be related to additional concentration of these metals from seawater by organisms (R.M. Garrels, pers. comm.). No apparent relation exists between the positions of the plate boundaries and either the overall distribution of the enriched zone of nodules in the Pacific Ocean basin. The nodules are formed by hydrogenous processes in which the metals are precipitated both from seawater and from interstitial water of un- derlying sediments. The metals are at least partially derived from hydro- thermal sources. In the special case of obduction, the upper layer of oceanic lithosphere is thrust up and overrides an adjacent plate at a convergent plate boundary (Fig. 2). The slice of oceanic lithosphere, up to several tens of kilometers thick, may contain the various types of precious, base, iron and ferro-alloy deposits that were described to form at oceanic ridges (divergent plate boundaries). Areas of obduction in the southwest Pacific are Papua, New Guinea (Davies and Smith, 1971), where gold and copper prospects exist (Eimon, 1974; Grainger and Grainger, 1974), and the island of New Caledonia (Arias, 1967), where nickel and chromium deposits are mined. The island arcs of the western Pacific, the Kamchatka Peninsula, and western North America from Alaska to Baja California are areas of former obduction which incorporate slices of oceanic lithosphere (Coleman, 1971, fig. 4). Processes of metal concentration at convergent plate boundaries The observation that precious, base, iron and ferro-alloy metal deposits are associated with convergent plate boundaries of the Pacific region (Figs. 6—9) has led to the interpretation of these deposits as genetically related to plate convergence. Models are being developed to gain an understanding of the sources of the various metals and the processes that concentrate the metal deposits. Prior to the theory of plate tectonics, the source for metals was generally considered to be anomalous metal concentrations in continental crust and mantle underlying the deposits (Krauskopf, 1967; Noble, 1970). Plate tec- 71 412 tonics has turned attention to the oceanic lithosphere as a source for a signif- icant fraction of the metals in deposits at convergent plate boundaries of the Pacific. Early models stressed metals concentrated by hydro thermal pro- cesses in particulate phases (metalliferous sediments) and in solid phases (oxides and sulfides) in oceanic crust as primary sources (Sawkins, 1972; Sil- litoe, 1972a). However, the amounts of those precious, base, iron and ferro- alloy metals disseminated in oceanic crust by magmatic processes more than suffice to quantitatively account for the majority of deposits of these metals observed along convergent plate boundaries (see NOTES, p. 75). Adequate sources and supplies of various metals exist to account for the metal of ore deposits (Krauskopf, 1967). The principal problems in metallo- genesis are extraction of metals from the sources, transport of the metals, their concentration and deposition. In simplest form, models of metallo- genesis at convergent plate boundaries envisage the extraction of metals from sea water-saturated oceanic crust as it undergoes partial melting under condi- tions of increasing temperature and pressure during descent of the oceanic lithosphere along a Benioff zone (Fig. 10; Sawkins, 1972; Sillitoe, 1972a). The metals ascend as components of magmas, are concentrated in fluids released from the magmas, and are deposited. The models are becoming increasingly complex to account for the actual characteristics of metal deposits at convergent plate boundaries of the Pacific region (Mitchell and Bell, 1973; Ridge, 1972). The distribution of metal de- posits parallel to convergent plate boundaries in metal provinces of the west- ern Americas (Fig. 10) may be related to progressive increase in temperature and pressure and change in chemical environment down the inclined plane of the Benioff zone which together act to separate different components of the oceanic lithosphere during partial melting (Sillitoe, 1972b). Different associations of metals and igneous rocks may be related to variation in com- position of magmas controlled by changes in the inclination of Benioff zones resulting from changes in rates of lithospheric plate convergence and sea floor spreading through time (Mitchell, 1973). The actual inclinations of Benioff zones are not constant as shown in models (Fig. 10), but vary with depth. Metals other than those present in oceanic crust such as tin, as well as ad- ditional quantities of metals and sulfur present in oceanic crust, may be derived from the asthenosphere and continental lithosphere overlying Benioff zones .(Fig. 10). The proportion of metals and sulfur derived from the various potential sources is unknown and is the subject of studies using sulfur, lead, and strontium isotopes as tracers (Corliss, 1974; Dasch, 1974). Volatile components such as hydrogen fluoride and carbon dioxide liberated from dry oceanic lithosphere at depths exceeding 200 km along a Benioff zone may lower melting points, assist in transporting metals, and liberate tin 72 413 and associated metals (tungsten, bismuth, fluoride, and molybdenum) from granite in overlying continental crust (Mitchell and Garson, 1972; Stern and Wylie, 1973; Oyarzun and Frutos, 1974). The tin and associated metals in eastern Asia and the various metal deposits in eastern Australia may have been deposited above former Benioff zones of shallow inclination adjacent to the continental margins related to relatively fast plate convergence. Sub- sequent increase in inclination of the Benioff zones related to relatively slow plate convergence has resulted in the seaward migration of the Benioff zones as a consequence of the growth of marginal basins (Fig. 10), leaving the ob- served gap between the deposits of the continental margins and active con- vergent plate boundaries of the western Pacific (Mitchell, 1973). Convergent plate boundaries are the loci of a multiplicity of interacting geologic processes that are difficult to differentiate. The models incorporate different processes to explain the factors that control the locations of ore deposits along the convergent plate boundaries of the Pacific region: (1) deep processes: variations in sources of metals, physico-chemical mecha- nisms, magmatic processes, seismic activity, rate and inclination of litho- spheric descent, and geologic structure associated with subduction along Benioff zones (Krauskopf, 1967; James, 1971; Sawkins, 1972; Sillitoe, 1972a, 1974; Mitchell, 1973); (2) shallow processes: regional and local vol- canism, magmatic processes, hydrothermal activity, geologic deformation and structure of circum-Pacific mountain belts and island arcs (Minato et al., 1965; Hollister, 1973; Solomon, 1974). The models of metallogenesis at con- vergent plate boundaries are becoming more complex as factors are added to successively approximate the actual deposits. The models are still inter- pretive in that they explain the distribution of known deposits. With further development these models may predict the locations of new deposits. 0 KM OKM 15,200 18,700 T 19,000 T T T 1 SOUTH AMERICA ANDES PETROLEUM j Co ^Cu.Au -i^-Eb'Zn'Cu'A'!» Fe ili^f . >-x :litho- SPHERE; :;::%v: tlASTHENpSPHERE ::%•>. Fig. 10. A diagrammatic east-west section across the central Pacific region shows the rela- tion of petroleum and metal deposits to divergent (East Pacific Rise) and convergent (Pacific margins) plate boundaries, as discussed in the text. Metals are disseminated in the rocks and concentrated as oxides and sulfides in oceanic crust represented by the black layer at the top of the oceanic lithosphere. 73 414 SUMMARY Despite the geologic diversity of the Pacific region, the distribution of selected energy and mineral resources follows a pattern with respect to lithospheric plate boundaries (Fig. 1), as follows: (1) Hydro thermal processes acting at divergent plate boundaries (oceanic ridges) concentrate metals in the upper layers of oceanic lithosphere as metalliferous sediments, metallic oxide deposits, and possibly as massive stratabound copper-iron sulfide deposits. Because all of the oceanic litho- sphere is generated by sea-floor spreading about oceanic ridges (Figs. 1, 2), the processes of metal concentration at divergent plate boundaries affect the oceanic lithosphere underlying two-thirds of the earth including the entire Pacific ocean basin (Fig. 10). (2) Oceanic trenches along convergent plate boundaries of the eastern, northern, and western Pacific, and marginal basins formed between island arcs and eastern Asia are areas of offshore petroleum potential (Figs. 5, 10). (3) Deposits of precious, base, iron and ferro-alloy metals occur along the landward side of convergent plate boundaries on continents and island arcs of the Pacific region (Figs. 6—9). (4) Models suggest that the observed distribution of petroleum and metal deposits of the Pacific region are genetically related to geologic processes acting at the circum-Pacific convergent plate boundaries (Fig. 10). The development of oceanic trenches and marginal basins create conditions that favor the accumulation of sediment and organic matter, the conversion of the organic matter to petroleum, and the trapping of the petroleum. Metals undergo multiple stages of concentration from various sources in two prin- cipal regions (Fig. 10): (a) Divergent plate boundaries: certain precious, base, iron and ferro-alloy metals are disseminated in oceanic lithosphere by magmatic processes and concentrated by hydrothermal processes. (b) Convergent plate boundaries: metals are concentrated from oceanic lithosphere descending along Benioff zones and from the overlying astheno- sphere and continental lithosphere by various physical and chemical pro- cesses. The deposits at convergent plate boundaries are products of complex histories that are only beginning to be understood. In conclusion, the conceptual framework of plate tectonics may be ap- plied to predict areas hundreds to thousands of kilometers in extent of the Pacific region where certain types of energy and mineral resources are likely to occur. Conventional geological, geochemical, and geophysical methods must then be employed to locate the deposits that may be only tens to thou- sands of meters in extent within the areas of potential occurrence identified from plate-tectonic considerations. The resolution of plate tectonics in pre- 74 415 dieting the locations of deposits will improve as models of geologic processes at plate boundaries are refined, but conventional exploration methods will continue to complement plate tectonics. ACKNOWLEDGMENTS We thank V. Baum, Chief of the Resources and Transport Division, and J.P. Levy, Chief of the Office for Ocean Economics and Technology of the United Nations for their important encouragement. The United Nations and the National Oceanic and Atmospheric Administration supported this work. NOTES The amount of many precious, base, iron and ferro-alloy metals dis- seminated in oceanic crust (basalt) is considerably greater than in con- tinental crust (granite). For example, the copper content of basalt (about 100 ppm) is approximately five times that of granite (Turekian and Wede- pohl, 1961; Vinogradov, 1962). An orebody containing 1,000,000 tons of copper is equivalent to only about 4% of the copper disseminated in a 100-km3 volume of oceanic basalt (1 ppm = 104 tons per mile3 or 25 • 102 tons per km3). Sulfur is also present both in oceanic crust (100—400 ppm; Vinogradov, 1962) and in seawater (1 g/1) in sufficient quantities to form the various sulfide ores. A continuous supply of metals and sulfur is provided by the conveyor belt of seawater-saturated oceanic crust that is generated at divergent plate boundaries, moves across the ocean basin, and is consumed at the convergent plate boundaries. An estimated volume of 100,000—250,000 km3 of oceanic crust has been overridden for every kilometer of leading edge along the western Americas (Gilluly, 1973). Four percent of the copper dis- seminated in this volume of oceanic crust is equivalent to between 1,000,000,000 and 2,500,000,000 tons, only a fraction of which is known to be concentrated in massive stratabound and porphyry copper deposits along the western margins of North and South America. REFERENCES Anon., 1972. The Times Atlas of the World. Bartholomew, Times Newspapers Ltd., Edinburgh, 4th rev. ed., 272 pp. Avias, J., 1967. Overthrust structure of the main ultrabasic New Caledonian massives. Tectonophysics, 4: 531—541. Betzer, P.R., Bolger, G.W., McGregor, B.A. and Rona, P.A., 1974, The Mid-Atlantic Ridge and its effect on the composition of particulate matter in the deep ocean. Am. Geo- phys. Union Trans., 55: 293. 75 416 Bischoff, J.L. and Manheim, F.T., 1969. Economic potential of the Red Sea heavy metal deposits. In: E.T. Degens and D.A. Ross (Editors), Hot Brines and Recent Heavy Metal Deposits in the Red Sea. Springer, New York, N.Y., 600 pp. Bostrom, K. and Peterson, M.N.A., 1969. The origin of aluminum-poor ferromanganoan sediments in areas of high heat flow on the East Pacific Rise. Mar. Geol., 7: 427—447. Bracey, D.R. and Ogden, T.A., 1972. Southern Mariana Arc: geophysical observations and hypothesis of evolution. Geol. Soc. Am. Bull., 83: 1509—1522. Coleman, R.G., 1971. Plate tectonic emplacement of upper mantle peridotites along continental edges. J. Geophys. Res., 76: 1212—1222. Cook, H.E., 1972. Stratigraphy and sedimentation. In: J.D. Hayes et al. (Editors), Initial Reports of the Deep Sea Drilling Project, IX. U.S. Govt. Printing Office, Washington, D.C., 1205 pp., pp. 933—943. Corliss, J.B., Dasch, E.J., Dymond, J.D., Field, C.W., Heath, G.R., Senechal, R.G. and Veeh, H.H., 1974. Metallogenesis in southeast Pacific Ocean: Nazca Plate Project. Am. Assoc. Pet. Geol. Bull., 58: 1434 (abstract). Dasch, E.J., 1974. Metallogenesis in the southeastern Pacific: a progress report on the IDOE Nazca Plate Project. Phys. Earth Planet. Inter., 9: 249—258. Davies, H.L. and Smith, I.E., 1971. Geology of Eastern Papua. Geol. Soc. Am. Bull., 82: 3299-3312. DEMR, 1972. Principal mineral areas of Canada, scale 1 : 7,600,200. Dept. Energy, Mines and Resour., Min. Resour. Branch and Geol. Surv. of Canada, Map 900A, 22nd ed. Dymond, J., Corliss, J.B., Heath, G.R., Field, C.W., Dasch, E.J. and Veeh, H.H., 1973. Origin of metalliferous sediments from the Pacific Ocean. Geol. Soc. Am. Bull., 84: 3355-3372. ECAFE, 1962. Oil and natural gas map of Asia and the Far East, scale 1 : 5,000,000. UN Econ. Comm. for Asia and the Far East (ECAFE), No. 16351, UN Publ. Sales No. 62-1-16, 4 sheets. ECAFE, 1963. Mineral distribution map of Asia and the Far East, scale 1 : 5,000,000. UN Econ. Comm. for Asia and the Far East (ECAFE). No. 63-15714, UN Publ. Sales No. 63-1-18, 4 sheets. ECAFE, 1970. Energy atlas of Asia and the Far East. UN Econ. Comm. for Asia and the Far East (ECAFE). UN Publ. Sales No. E.70.1.11, E/CN. 11/900, 25 pp. Eimon, P. I., 1974. World copper resources. Mineral Res. Company, Salt Lake City, Utah, map in 3 sheets. Fischer, A.G., Heezen, B.C., Boyce, R.E., Bukry, D., Douglas, R.G., Garrison, R.E., Kling, S.A., Krasheninnikov, V., Lisitsin, A.P. and Pimm, A.C., 1971. Initial Reports of the Deep Sea Drilling Project, VI. U.S. Govt. Printing Office, Washington, D.C., 1329 pp. Gilluly, J., 1973. Steady plate motion and episodic orogeny and magmatism. Geol. Soc. Am. Bull., 84:499-514. Grainger, D.J. and Grainger, R.L., 1974. The minerals deposits map of Papua New Guinea, 1: 2,500,00. Aust. Bur. Miner. Resour., Geol. Geophys., Bull., 148, PNG 8, 171 pp. Hart, R.A., 1973. A model for chemical exchange in the basalt-seawater system of ocean- ic layer II. Canadian J. of Earth Sci., 10: 799—816. Hollister, V.F., 1973. Regional characteristics of porphyry copper deposits of South America. Soc. of Mining Engineers Preprint No. 73-1-2. Horn, D.R., Horn, B.M. and Delach, M.N., 1972. Metal content of ferromanganese depos- its of the oceans. Lamont-Doherty IDOE/NSF GX33616 Tech. Rep. 3. Office for Int. Decade of Ocean Exploration, Natl. Sci. Found., Washington, D.C., 51 pp. Irving, E., North, F.K. and Couillard, R., 1974. Oil, climate, and tectonics. Can. J. Earth Sci., 11: 1-17. Isacks, B., Oliver, J. and Sykes, L.R., 1968. Seismology and the new global tectonics. J. Geophys. Res., 73: 5855—5899. James, D.E., 1971. Plate tectonic model for the evolution of the Central Andes. Geol. Soc. Am. Bull., 82: 3325—3346. 76 417 Jones, D.B. (Advisory Editor), 1972. Oxford Economic Atlas of the World. Oxford Uni- versity Press, 4th ed., 239 pp. Karig, D.E., 1971. Origin and development of marginal basins in the western Pacific. J. Geophys. Res., 76: 2542-2561. Kennedy, W.Q. and Richey, J.E., 1947. Catalog of the active volcanoes of the world. Bull. Volcanol. Suppl. Ser. II, VII: 1-11. Krauskopf, K.B., 1967. Source rocks for metal-bearing fluids. In: H.L. Barnes (Editor), Geochemistry of Hydrothermal Ore Deposits. Holt, Rinehart and Winston, New York, N.Y.,pp. 1-33. Langseth, M.G., Jr., 1969. The flow of heat from the Earth and the global distribution at the surface. Am. Inst. Aeronaut. Astronaut., Pap, 69-589: 1—10. La Plante, R.E., 1974. Hydrocarbon generation in Gulf Coast Tertiary sediments. Am. Assoc. Pet. Geol. Bull., 58: 1281—1289. Le Pichon, X., 1968. Sea floor spreading and continental drift. J. Geophys. Res., 73: 3661-3697. Le Pichon, X., Francheteau, J. and Bonnin, J., 1973. Plate Tectonics. Elsevier, Amster- dam, 300 pp. Luyendyk, B.P., 1970. Dips of downgoing lithospheric plates beneath island arcs. Geol. Soc. Am. Bull., 81 : 3411-3416. McGregor, B.A. and Rona, P.A., 1975. Crest of the Mid-Atlantic Ridge at 26°N. J. Geo- phys. Res., 80: 3307-3314. McKelvey, V.E. and Wang, F.F.H., 1969. World subsea mineral resources. U.S. Geol. Surv., Misc. Geol. Invest., Map 1-632. Minato, M., Gorai, M. and Hunabashi, M., 1965. The Geological Development of Japanese Islands. Tokyo, Tsukiji Shoka. Mitchell, A.H.G., 1973. Metallogenic belts and angle of dip of Benioff zones. Nature Phys. Sci., 245: 49-52. Mitchell, A.H. and Bell, J.D., 1973. Island-arc evolution and related mineral deposits. J. Geol., 81: 381-405. Mitchell, A.H.G. and Garson, M.S., 1972. Relationship of porphyry copper and circum- Pacific tin deposits to palaeo-Benioff zones. Inst. Min. Metall., Trans., Sec. B, 81: B10-B25. Oxburgh, E.R. and Turcotte, D.L., 1971. Origin of paired metamorphic belts and crustal dilation in island arc regions. J. Geophys. Res., 76: 1315—1327. Oyarzun, J.M. and Frutos, J.J., 1974. Porphyry copper and tin-bearing porphyries: a dis- cussion of genetic models. Phys. Earth Planet. Inter., 9: 259—263. Piper, D.Z., 1973. Origin of metalliferous sediments from the East Pacific Rise. Earth Planet. Sci. Lett., 19: 75-82. Pitman, W.C. Ill, Larson, R.L. and Herron, E.M., 1974. Age of the ocean basins deter- mined from magnetic anomaly lineations. Geol. Soc. Am., map. Ridge, J.D., 1972. Annotated bibliographies of mineral deposits in the western hemi- sphere. Geol. Soc. Am. Mem., 131: 681 pp. Rona, P. A., 1973. New evidence for seabed resources from global tectonics. Ocean Man- agement, 1 : 145—159. Rona, P. A. and Neuman, L.D., 1974. Plate tectonics and mineral resources of the Pacific (abstract). Am. Assoc. Pet. Geol. Bull., 58: 1456. Rona, P. A. and Neuman, L.D., 1975. Plate tectonics and mineral resources of the circum- Pacific. Circum-Pacific Energy and Mineral Resources — Am. Assoc. Pet. Geol., Mem. 25, in press. Rona, P. A., McGregor, B.A., Betzer, P.R. and Krause, D.C., 1975. Anomalous water tem- peratures over Mid-Atlantic Ridge crest at 26°N. Deep-Sea Res., 22: 611—618. Rona, P.A., Harbison, R.N., Bassinger, B.G., Scott, R.B. and Nalwalk, A.J., 1976. Tec- tonic fabric and hydrothermal activity of Mid-Atlantic Ridge crest (26°N). Geol. Soc. Am. Bull., in press. 77 418 Sawkins, F.J., 1972. Sulfide ore deposits in relation to plate tectonics. J. Geol., 80: 377-397. Sayles, F.L. and Bischoff, J.L., 1973. Ferromanganoan sediments in the equatorial East Pacific. Earth Planet. Sci. Lett., 19: 330—336. Sclater, J.G., 1972. Heat flow and elevation of the marginal basins of the western Pacific. J. Geophys. Res., 77 : 5705-5709. Scott, M.R., Scott, R.B., Rona, P. A., Butler, L.W. and Nalwalk, A. J., 1974. Rapidly ac- cumulating manganese deposit from the median valley of the Mid-Atlantic Ridge. Geophys. Res. Lett., 1: 355—358. Scott, R.B., Rona, P.A., McGregor, B.A. and Scott, M.R., 1974. The TAG hydrothermal field. Nature, 251 : 301-302. Sillitoe, R.H., 1972a. A plate tectonic model for the origin of porphyry copper deposits. Econ. Geol., 67: 184—197. Sillitoe, R.H., 1972b. Relation of metal provinces in western America to subduction of oceanic lithosphere. Geol. Soc. Am. Bull., 83: 813—818. Sillitoe, R.H., 1974. Tectonic segmentation of the Andes: implications for magmatism and metallogeny. Nature, 250: 542—545. Skornyakova, N.S. and Andrushchenko, P.F., 1974. Iron-manganese concretions in the Pacific Ocean. Int. Geol. Rev., 16: 863-919. Sleep, N.H. and Toksoz, M.N., 1971. Evolution of marginal basins. Nature, 233: 548— 550. Snead, R.E., 1972. Atlas of World Physical Features. Wiley, New York, N.Y., 158 pp. Solomon, M., 1974. Massive sulfides and plate tectonics. Nature, 249: 821—822. Sorokhtin, O.G., Ushakov, S.A. and Fedynskiy, V.V., 1974. Dynamics of crustal plates and origin of petroleum deposits. Akad. Nauk SSSR, Doklady, 214: 1407—1410. Spooner, E.T.C. and Fyfe, W.S., 1973. Sub-sea floor metamorphism, heat and mass trans- fer. Contrib. Miner. Pet., 42: 287—304. Stern, C.R. and Wylie, P. J., 1973. Water-saturated and undersaturated melting relations of a granite to 35 kilobars. Earth Planet. Sci. Lett., 18: 163—167. Strakhov, N.M., Shterenberg, L.E., Kalinenko, V.V. and Tikhomirova, E.S., 1968. Geo- chemistry of a sedimentary manganese ore-forming process. Akad. Nauk. SSSR Geol. Inst., Trans., 185: 495 pp. Tarling, D.H., 1973. Continental drift and reserves of oil and natural gas. Nature, 243: 277-279. Turekian, K.K. and Wedepohl, K.H., 1961. Distribution of the elements in some major units of the Earth's crust. Geol. Soc. Am. Bull., 72: 175—192. Van Royan, W. and Bowles, O., 1952. The mineral resources of the world. V. II, Atlas of the World's Resources. Prentice-Hall, New York, N.Y., 181 pp. Vinogradov, A.P., 1962. Average contents of chemical elements in the principal types of igneous rocks of the earth's crust. Geokhimiya, 7: 555—571. Von der Borch, C.C., Nesteroff, W.D. and Galehouse, J.S., 1971. Iron-rich sediments cored during Leg 8 of the Deep Sea Drilling Project. In: J.I. Tracey, Jr. et al. (Editors), Initial Reports of the Deep Sea Drilling Project, VIII. U.S. Govt. Printing Office, Washington, D.C., 1037 pp., pp. 725—819. Von der Borch, C.C. and Rex, R.W., 1970. Amorphous iron oxide precipitates in sedi- ments cored during Leg 5, Deep Sea Drilling Project. In: D.A. McManus et al. (Edi- tors), Initial Reports of the Deep Sea Drilling Project, V. U.S. Govt. Printing Office, Washington, D.C., pp. 541—544. Waring, G.A., 1965. Thermal springs of the United States and other countries of the world: a summary. U.S. Geol. Surv., Prof. Pap., 492. 78 419 44 Reprinted from: Papers from Circum-Pacific Energy and Mineral Resources Conference, Honolulu, Hawaii, August 26-30, 1974, publ. by Amer. Assoc, of Petroleum Geologists, Memoir 25, 48-57. Plate Tectonics and Mineral Resources of Circum-Pacific Region1 PETER A. RONA2 and LAWRENCE D. NEUMANP Abstract Distribution of selected energy (petroleum and geothermal) and mineral (precious, base, iron, and ferro- alloy metals) resources of the Pacific Ocean basin and Cir- cum-Pacific continents appears to be related to lithospheric plate boundaries. Divergent plate boundaries (oceanic ridges) are related in time to the development of strati- graphic traps in sedimentary basins of the continents, and in space to metalliferous deep-sea sediments and the possible occurrence of massive stratabound metallic sulfide deposits in oceanic crust. Convergent plate boundaries are related in space to areas of offshore petroleum potential and to on- shore deposits of precious, base, iron, and ferro-alloy met- als. Models suggest genetic relations between the observed distribution of deposits and geologic processes at plate boundaries, and may lead to the discovery of new re- sources. Introduction The Pacific Ocean basin and surrounding con- tinents provide a natural laboratory to develop and test ideas on the relation between plate tec- tonics and mineral resources. Our approach is as follows: 1. Outline the geologic framework of the Pacif- ic with particular attention to boundaries of the lithospheric plates. 2. Determine the spatial distribution of various mineral resources with respect to the plate bound- aries (Figs. 1—4). 3. Consider models of mineral-concentrating processes related to plate boundaries that may ex- plain the observed distribution of mineral re- sources (Fig. 5). The boundaries of lithospheric plates are delin- eated by narrow earthquake zones where the plates are moving with respect to each other. The theory of plate tectonics recognizes three types of plate boundaries (Isacks et al, 1968). One type, a convergent plate boundary, is where two adjacent plates move together and collide or where one plate descends under the other plate along a Be- nioff seismic zone and is subducted into the up- per mantle. The second type, a divergent plate boundary, is where two adjacent plates move apart because new lithosphere is added to each plate by the process of seafloor spreading. The third type is the transform plate boundary, where two adjacent plates slide past one another. A series of resource maps (Figs. 1-4), using the same base map (Van der Grinten projection) as that used by McKelvey and Wang (1969), was compiled from numerous sources (Van Royan and Bowles, 1952; Roberts and Irving, 1957; Anon., 1962, 1963, 1970, 1972a, b; Lafitte and Rouveyrol, 1965; McKelvey and Wang, 1969; Bonine et al, 1970; Dengo and Levy, 1970; Jones, 1972). Pacific Geologic Framework Lithospheric Plate Boundaries Divergent plate boundaries expressed as ocean- ic ridges divide the Pacific into several lithospher- ic plates (Fig. 1). The oceanic ridge system of the Pacific is not midoceanic but is located in the eastern Pacific off Central America and South America and northwestern North America. The average half-rates of seafloor spreading about the oceanic ridges of the Pacific, determined from the geomagnetic polarity-reversal time scale, range from 1.2 cm/yr at the Gorda Rise off the north- western United States to about 10 cm/yr off Chile and Peru (Table 1). The Pacific is a closing ocean basin surrounded on three sides by convergent plate boundaries ex- pressed as oceanic trenches (Fig. 1). In the global crustal budget, the Circum-Pacific convergent plate boundaries account for the major portion of crustal consumption. Linear rates of crustal con- sumption around the Pacific, assumed to be equal to rates of plate convergence, range between about 1 and 15 cm/yr (Table 2). The dip of the Benioff zone at convergent plate boundaries (Table 3) is inversely proportional to the relative rate of convergence of the adjacent plates (Luyendyk, 1970). For example, the dip of the Benioff zone in the upper 100 km is about 12° beneath Peru, where the relative rate of plate con- vergence is 8.8 cm/yr (Table 2), and the half-rate of seafloor spreading about the corresponding section of the East Pacific Rise is 9.5 cm/yr (Ta- ble 1). The dip of the Benioff zore increases to about 55° beneath the Kermadec Islands (Table 3), where the relative rate of convergence decreas- es to 5.8 cm/yr, and the half-rate of seafloor 1 Manuscript received. December 26, 1974. ^National Oceanic and Atmospheric Administration. Atlantic Oceanographic and Meteorological Laboratories, 15 Rickenbacker Causeway, Miami, Florida 33149. 3Office for Ocean Economics and Technology, United Nations. New York. New York 10017. 48 420 Plate Tectonics and Mineral Resources 49 180' 150* 120* 90* 60* 30*W • PETROLEUM PRODUCING AREAS ONSHORE PETROLEUM POTENTIAL m OFFSHORE PETROLEUM POTENTIAL CRYSTALLINE ROCKS SEDIMENTARY ROCKS DIVERGENT PLATE BOUNDARY CONVERGENT PLATE BOUNDARY TRANSFORM PLATE BOUNDARY UNCERTAIN PLATE BOUNDARY FIG. 1 -Areas of petroleum production and potential of Pacific region. Lithospheric plates and plate boundaries are shown. spreading about the East Pacific Rise decreases to about 5.5 cm/yr. Oceanic and continental crust are juxtaposed at convergent plate boundaries in the eastern Pa- cific (Fig. 1). Marginal basins generally underlain by oceanic crust intervene between convergent Table 1 . Seafloor-Spreading Half-Rates about Pacific Oceanic Ridges Latitude Longitude Half-spreading Location rate (+North, (+East, (cm/yr) -South) - West) 46.5 -129 1.5 Juan de Fuca Ridge 41.3 -127.5 1-2 Gorda Rise 12.5 -103.5 4.5 East Pacific Rise 3.2 -102 6.4 East Pacific Rise 3.0 - 83.5 3.2 Galapagos rift zone 2.0 - 97.5 3.2 Galapagos riff zoneJ 0.75 - 87.6 3.2 Galapagos rift zone - 5.6 -106.8 6.5 East Pacific Rise -9.2 -110.3 7-5 East Pacific Rise -19 -113 9-9.5 East Pacific RiseJ -28 -112 8-10 East Pacific Rise -35.6 -110.9 5.0 East Pacific Rise -43.2 - 82.5 3.0 Chile Ridge3 'Vine and Wilson, 1965. JAtwater and Mudie, 1973. *Herron, 1972. plate boundaries and continental crust in the western Pacific. The development of marginal ba- sins may be related to the dip of Benioff zones (Oxburgh and Turcotte, 1971; Karig, 1971; Sleep and Toksoz, 1971; Bracey and Ogden, 1972). Where the marginal basins are present in the western Pacific, the dip of the Benioff zone ex- ceeds about 35°; where marginal basins are ab- sent in the eastern Pacific, the dip is less than about 35° (Table 3). Ages of Seafloor and Continents The age of the Pacific seafloor, as determined by dating of remanent magnetic anomalies in the geomagnetic polarity-reversal time scale (Pitman et al, 1974) and dating of rock samples recovered by the Deep Sea Drilling Project (Fischer et al, 1971), ranges between Late Jurassic (about 150 m.y. ago) and recent. The distribution of ages about divergent plate boundaries is regular, and the seafloor becomes progressively older with dis- tance from these boundaries. The distribution of ages at convergent plate boundaries is irregular, and the seafloor is of various ages at these bound- aries. The distribution of ages on continents is delineated by structural provinces that reflect the youngest deformational event, as distinguished from the ages of the seafloor, which reflect the origin of the constituent rocks. Structural prov- inces of the eastern Pacific are predominantly Ce- nozoic along western South America and Meso- zoic along western North America. In the western Pacific, structural provinces of eastern Asia and Australia exhibit a complex distribution and 421 50 Peter A. Rona and Lawrence D. Neuman Table 2. Rate of Plate Convergence at Convergent Plate Boundaries of the Pacific1 Latitude Longitude Rate Azimuth Location (+ North, (+ East, (cm/yr) - South) - West) 51 160 7.2 114 Kurile Trench 43 148 7.5 107 Kurile Trench 35 142 8.6 101 Japan Trench 27 143 7.5 265 Japan Trench 19 148 4.9 282 Mariana Trench 11 142 2.3 243 Mariana Trench - 3 142 14.5 78 New Guinea -13 -172 9.9 97 N. Tonga Trench -34 - 178 5.8 95 S. Kermadec Trench -45 169 3.5 72 S. New Zealand -55 159 2.6 29 Macquarie Island -50 - 75 3.1 240 Cape Horn -35 - 74 8.7 74 S. Chile Trench - 4 - 82 8.8 77 N. Peru Trench 7 - 79 8.3 75 Panama Gulf 20 - 106 6.4 39 N. Middle America Trench 57 - 150 5.6 144 E. Aleutian Trench 50 - 178 6.9 126 W. Aleutian Trench 54 162 7.0 115 W. Aleutian Trench 'From Le Pichon et al, 1973, Table V. range in age from Precambrian through Ceno- zoic. Distribution of Geothermal Phenomena The distribution of geothermal phenomena, in- cluding heat flow, active volcanoes, thermal springs, fumaroles, and geysers, is spatially relat- ed to lithospheric plate boundaries. Values of heat flow in the Pacific Ocean basin (Langseth, 1969; Sclater, 1972) are relatively high (>2 HFU) at divergent plate boundaries and decrease basin- ward away from these boundaries. Values of heat flow at convergent plate boundaries and in mar- ginal basins (Karig, 1971) around the Pacific are variable. The distribution of heat-flow values on the Circum-Pacific continents is complex. Rela- tively high values (>2 HFU) are present in limit- ed areas of eastern Australia, eastern Asia, and western North America. Active volcanoes, ther- mal springs, and fumaroles are aligned along the landward side of convergent plate boundaries on continents and island arcs around the Pacific (Kennedy and Richey, 1947; Waring, 1965; Snead, 1972). These features are not evenly spaced but are grouped (Kelleher, 1972). It is apparent from the distribution of heat-flow values and thermal springs on islands and conti- nents around the Pacific that numerous potential sites exist for the development of geothermal en- ergy. Geothermal energy is being utilized at sites in New Zealand, Japan, and the western United States. Distribution of Petroleum Resources Areas of offshore petroleum potential conform with convergent plate boundaries around the Pa- cific (Fig 1 ; McKelvey and Wang, 1969). Areas of petroleum potential in the eastern Pacific com- prise thick sedimentary accumulations underlying the continental margins of western North Ameri- ca and South America and partially filling ocean- ic trenches seaward of the margins of Central America and South America along convergent plate boundaries. In the western Pacific, island arcs directly landward of convergent plate boundaries divide the ocean basin into marginal basins partially enclosed between the islands and eastern Asia — for example, the South China Sea, the East China Sea, the Yellow Sea, the Sea of Japan, the Sea of Okhotsk, and the Bering Sea. 422 Plate Tectonics and Mineral Resources 51 Table 3. Dip of Benioff Zone (upper 100 km)1 Dip (degrees) Location (counter-clockwise around Pacific) 45 New Zealand 55 Kermadec 50 Tonga (south) 30 Tonga (north) 65 New Hebrides (south) 55 New Hebrides (north) 65 Solomons 45 New Britain 40 Sunda: Flores Sea 60 Sunda: Java 50 Sumatra 50 Burma 55 Mindanao 45 Marianas 25 Izu-Bonin 40 Ryukus 35 Kurile 30 Honshu 40 Aleutians 35 Middle America 12 Peru 15 Chile (north) 'isacks and Molnar, 1971; Oliver et al, 1973. Areas of petroleum potential in the western Pacif- ic comprise sedimentary accumulations in the marginal basins as well as in oceanic trenches along convergent plate boundaries. The oceanic trenches along the eastern and western sides of the Pacific Ocean basin receive deep-sea sediments that presumably are trans- ported into the trenches on a "conveyor belt" of spreading seafloor during subduction of the oceanic lithosphere. The content of organic mat- ter in the deep-sea sediments that are transported into the trenches varies in space and in time; for example, a zone of high organic productivity ex- tending across the equatorial Pacific affects the composition of the sediments deposited in that region. Both the amount of organic matter deliv- ered to the oceanic trenches and the petroleum potential are expected to vary accordingly. Tem- perature and pressure conditions beneath the trenches and the marginal basins may facilitate the conversion of organic matter in the sediments to petroleum (Tarling, 1973; Sorokhtin et al, 1974). Areas of onshore petroleum production are sedimentary basins on continents with no appar- ent spatial relation either to divergent or conver- gent plate boundaries of the Pacific (Fig. 1). However, the development of the sedimentary se- quences in the basins may be related to divergent plate boundaries in time, if not in space, through the influence on global sea level of reversible vol- ume changes of oceanic ridges (Rona, 1973b; Rona and Wise, 1974). The volume of the world oceanic ridge system is not constant but appears to vary directly with rates of seafloor spreading. A volume increase in the world oceanic ridge sys- tem reduces the cubic capacity of ocean basins, resulting in transgression of the sea onto all the continents and deposition of a sedimentary se- quence that may contain organic source material and reservoir rocks. Conversely, a volume de- crease in the oceanic ridge system increases the cubic capacity of ocean basins, resulting in re- gression of the sea from all the continents and the development of widespread unconformities that may be associated with traps for the accumula- tion of petroleum. Reversible volume changes of the worldwide oceanic ridge" system have operat- ed on a time scale of tens of millions of years, as evidenced by the presence of six sedimentary se- quences separated by surfaces of unconformity in the Phanerozoic stratigraphy of North America (Sloss, 1963). The inferred relations of the volume of oceanic ridges to sedimentary sequences and unconformities may prove useful in exploration for stratigraphic traps. Distribution of Selected Mineral Resources Light Metal Deposits Deposits of light metals, including aluminum, beryllium, lithium, and titanium, exhibit no ap- parent relation to plate boundaries of the Pacific. These metals are associated with granitic rocks of the continents as opposed to basaltic rocks of the ocean basins. The distribution of aluminum, be- ryllium, lithium, and titanium on continents is re- lated to the occurrence of particular minerals in granitic rocks and to conditions of weathering. Metal Deposits at or Near Convergent Plate Boundaries Precious metal deposits — Deposits of precious metals, including gold, silver, and platinum, ex- hibit a spatial relation to convergent plate bound- aries (Fig. 2). In the eastern Pacific, precious met- al deposits are found along the western margins of North America and South America. Addition- al deposits are present in eastern North America 423 52 Peter A. Rona and Lawrence D. Neuman I2Q*E 150* 180* IW 120* 90* 60* 30*< * SILVER • GOLD O PLATINUM 120* E 190* ltO* 190* 120* FIG. 2-Map of precious metal deposits of Pacific region. Lithospheric plate boundaries are shown. in the area of the Canadian shield and in eastern South America in the area of the Guianian and Brazilian shields. In the western Pacific, precious metal deposits occur on island arcs (including Ja- pan, the Philippines, and Indonesia) situated along convergent plate boundaries. Deposits also occur in eastern Asia and Australia, where they are separated by a gap from active convergent plate boundaries. Precious metal deposits have not been found along those sections of conver- gent plate boundaries where oceanic lithosphere is juxtaposed and island arcs are absent (e.g., the eastern side of the Philippine Sea and the section between New Zealand and Samoa). Base metal deposits — Deposits of base metals exhibit a spatial relation to convergent plate boundaries (Fig. 3). Their distribution is grossly similar to that of precious metal deposits (Fig. 2). In the eastern Pacific, base metal deposits are present along the western margins of North America and South America. In the Peruvian Andes, porphyry copper and vein-type minerali- zation are associated with Neogene silicic volcan- ic rocks (Mitchell, 1973). Additional base metals occur in central and eastern North America in- cluding the area of the Canadian shield. In the western Pacific, base metal deposits are found on island arcs along convergent plate boundaries. Deposits of tin, tungsten, and fluorite with minor bismuth and molybdenum occur in belts of predominantly Mesozoic granites and acidic eruptive rocks along the southern margin of Alaska and the eastern margin of Asia (Mit- chell, 1973). Base metal deposits also occur in eastern Asia and Australia, which are separated by a gap from active convergent plate boundaries. As in the case of precious metals, base metal de- posits have not been found along those sections of convergent plate boundaries where oceanic lithosphere is juxtaposed and island arcs are ab- sent. Sediment samples recovered by coring near the crest of oceanic ridges and by deep-sea drilling away from the crest reveal widespread enrich- ment by base metals of those strata directly over- lying basalt of the Pacific Ocean basin (Bostrom and Petersen, 1969; von der Borch et al, 1971; Cook, 1972; Cronan et al, 1972; Dymond et al, 1973; Sayles and Bischoff, 1973; Piper, 1973). The base metals include antimony, copper, lead, mercury, and zinc, but no tin. The observation that the base metal enrichment is limited to the basal sedimentary layer overlying basalt implies that the enrichment occurred soon after the gen- eration of the underlying basalt by seaflooi spreading about an oceanic ridge at a divergent plate boundary. Iron and ferro-alloy metal deposits — Deposits of iron and ferro-alloy metals exhibit a spatial rela- tion to convergent plate boundaries around the Pacific (Fig. 4). Their distribution is grossly simi- lar to that of precious and base metals. In the eastern Pacific, iron and ferro-alloy metal depos- its occur along the western margins of North America and South America. Additional deposits occur in central and eastern North America, in- cluding the Canadian shield, and in the Guianian and Brazilian shields of South America. In the western Pacific, iron and ferro-alloy deposits oc- cur on island arcs along convergent plate bound- 424 Plate Tectonics and Mineral Resources 53 150* 120* W FIG. 3 -Map of base metal deposits of Pacific region. Lithospheric plate boundaries are shown. Sediments i enriched in base metals (shaded) overlie basalts of ocean basin. aries and in eastern Asia and Australia, where they are separated by a gap from active conver- gent plate boundaries. Iron and ferro-alloy metal deposits have not been found along those sections of convergent plate boundaries where oceanic lithosphere is juxtaposed and island arcs are ab- sent. Sedimentary strata directly overlying basalt of the Pacific Ocean basin are enriched in iron and ferro-alloy metals, in addition to base metals. Nodules containing variable percentages of manganese, copper, nickel, and cobalt are present on about two thirds of the Pacific seafloor (Fig. 4; Strakhov et al, 1968). A zone of nodules of anom- alously high copper and nickel content (1.5-2.0%) and areal density (20-50% of seafloor) extends east-west across the Pacific between latitudes 5° and 20° N. No apparent spatial relation exists be- tween plate boundaries and either the overall dis- tribution or the enriched zone of nodules in the Pacific Ocean basin. Metal Deposits at Divergent Plate Boundaries Knowledge of the distribution of metal depos- its with respect to divergent plate boundaries is * IRON Q CHROMIUM © COBALT O MANGANESE * MOLYBDENUM A NICKEL ■ TUNGSTEN * VANADIUM MANGANESE NODULES ' UP TO 20% OF BOTTOM = 20-50% OF BOTTOM BOUNDARY BETWEEN RED CLAY & BIOGENIC OOZES BE COPPER CONTENT IN WEIGHT 1.5-2.0% © NICKEL CONTENT 1.5-2.0% FIG. 4 -Map of iron and ferro-alloy metal deposits, including manganese nodules, of Pacific. Lithospheric plate boundaries are shown. 425 54 Peter A. Rona and Lawrence D. Neuman limited because, as submerged oceanic ridges, these boundaries are less accessible to observa- tion than are convergent plate boundaries. Sedi- ments deposited about divergent plate boundaries in the Pacific Ocean basin are enriched in certain base, iron, and ferro-alloy metals. The recently discovered TAG hydrothermal field (R. Scott et al, 1974; Rona et al, in press), named after the Trans-Atlantic Geotraverse (TAG) of the Nation- al Oceanic and Atmospheric Administration, rep- resents a type of metal deposit present in oceanic crust and may provide insights to processes of metal concentration at divergent plate bounda- ries, including those of the Pacific. The TAG hydrothermal field occupies an area at least 15 km by 15 km including the east wall of the rift valley of the Mid-Atlantic Ridge at 26°N. Manganese oxide was recovered consistently by dredging from deposits concentrated along steps in the wall that are interpreted as faults which have acted as conduits for hydrothermal fluids. Concentration of the manganese oxide by hy- drothermal processes capable of extreme chemi- cal fractionation is evidenced by an exceptionally rapid rate of accumulation, about 200 mm per million years, and extreme purity, about 40% manganese with only trace quantities of iron and copper (M. Scott et al, 1974). Present activity is indicated by a positive temperature anomaly with an inverted temperature gradient (Rona et al, 1975) and by metal enrichment of suspended par- ticulate matter (Betzer et al, 1974) in the seawater overlying the TAG hydrothermal field. Model of Metal Deposits at Divergent Plate Boundaries A model based on various lines of evidence (Spooner and Fyfe, 1973) considers that certain precious, base, iron, and ferro-alloy metals may be concentrated as deposits by sub-seafloor hy- drothermal convection systems that may develop at crests of oceanic ridges. According to the mod- el, cold dense seawater enters fractures in the ba- salt of an oceanic ridge. The seawater is heated as it encounters hot material intruded at the diver- gent plate boundary. The warm, less dense seawa- ter rises through fractures in the basalt and leach- es metals that are transported in solution as complexes with chloride in the seawater. A frac- tion of the metals then combines with sulfur in the seawater and precipitates as a massive strata- bound sulfide body under reducing conditions beneath the basalt-seawater interface. Manganese oxide precipitates under oxidizing conditions at the basalt-seawater interface as the hydrothermal solutions discharge from the seafloor. Colloidal ferric hydroxide precipitates from the hydrother- mal solutions in the overlying seawater. The ferric hydroxide scavenges the remaining metals from solution by colloidal absorption and settles to de- posit a layer of metalliferous sediments on basalt of the adjacent seafloor. An example of an economic mineral deposit interpreted as the product of a sub-seafloor hy- drothermal convection system occurs in the Troo- dos massif of Cyprus. The Troodos massif is in- terpreted as an obducted slice of oceanic crust generated at a divergent plate boundary (Gass and Masson-Smith, 1963; Moores and Vine, 1971). An "umber" of metallic oxides overlies massive stratabound copper sulfide ore bodies in basaltic rocks of the Troodos massif. The manga- nese oxide of the TAG hydrothermal field chemi- cally resembles the umber and may be underlain by massive copper sulfide bodies (Rona, 1973a; R. Scott et al, 1974; Rona et al, in press). Relict metallic oxide and sulfide deposits may be ex- pected to extend in belts along flow lines of sea- floor spreading transverse to the axis of an ocean- ic ridge; the extent of the belts would depend on the continuity of seafloor spreading and the per- sistence of the sub-seafloor hydrothermal convec- tion system which concentrates the deposits (Rona, 1973a). The metal deposits may be buried by off-axis volcanism. Model of Metal Deposits at Convergent Plate Boundaries A model to interpret the observed association of precious, base, iron, and ferro-alloy metal de- posits with the convergent plate boundaries of the Pacific is based on the following concepts: 1. The dip of Benioff zones is inversely propor- tional to the rate of plate convergence (Tables 1- 3; Luyendyk, 1970). 2. Marginal basins develop where the dip of Benioff zones exceeds about 35°. 3. Calc-alkalic andesitic volcanic rocks and to- nalitic plutons occur above steeply dipping Be- nioff zones (Mitchell, 1973). 4. Silicic volcanic rocks and granitic plutons occur above shallow-dipping Benioff zones (Mit- chell, 1973). 5. Metals in deposits along convergent plate boundaries are primarily derived from oceanic crust descending along the associated Benioff zone (Sawkins, 1972; Sillitoe, 1972) and from continental crust. The role of the mantle as a source of material remains to be evaluated. 6. The nature and volume of volatile matter ex- pelled from oceanic crust descending along Be- nioff zones influence the extraction, transport, 426 Plate Tectonics and Mineral Resources 55 concentration, and deposition of metals (Rub, 1972). The model presents two regimes to account for the past and present distribution of metals along convergent plate boundaries of the western and eastern Pacific, as follows: 1. Relatively fast seafloor spreading and plate convergence are associated with a shallow Be- nioff zone and the generation of silicic volcanic rocks and granitic plutons (Fig. 5a, b). High-level porphyry copper deposits occur in the silicic vol- canic rocks, and deep-level tin, tungsten, bismuth, fluorite, and molybdenum occur in the granites. The copper is primarily derived from metallifer- ous sediments and massive stratabound metallic sulfide deposits of the oceanic crust that descends along the underlying Benioff zone. The tin and associated metals, along with a portion of the granite (Stern and Wyllie, 1973), are derived from continental crust and their segregation is aided by the rise of volatile matter expelled from the de- scending oceanic crust. This regime is exemplified by western South America at present and by east- ern Asia in late Mesozoic time. 2. Relatively slow seafloor spreading and plate convergence are associated with a steep Benioff zone, the development of marginal basins, and the generation of calc-alkalic andesitic volcanic rocks and tonalitic plutons with associated por- phyry copper deposits (Fig. 5c, d). The calc-alkal- ic volcanic rocks and the copper are primarily de- rived from the oceanic crust (Jakes and White, 1972) that descends along the underlying Benioff zone; this crust includes metalliferous strata and massive strata-bound metallic sulfide deposits. The tonalitic plutons and some copper may be derived from the base of continental or island-arc crust (Jakes and White, 1971; Brown, 1973). Granitic plutons emplaced during the first regime (Fig. 5b) are unroofed by erosion to expose de- posits of tin and associated metals (Fig. 5d). The development of marginal basins (Fig. 5c, d) re- sults in the separation of island arcs from the con- tinent, producing a gap such as that observed be- tween the metal deposits of eastern Asia and Australia and the active convergent plate bound- aries of the western Pacific (Figs. 2-4). This re- gime is exemplified by eastern Asia at the end of the Mesozoic and at present. Summary Consideration of the distribution of selected energy and material resources with respect to li- thospheric plate boundaries of the Pacific (Figs. 1-4) leads to the following conclusions: 1. Divergent plate boundaries (oceanic ridges) are inferred to be related in time to the devel- opment of stratigraphic traps for petroleum in sedimentary basins on continents through the control of eustatic sea level by reversible volume changes of oceanic ridges. 2. Divergent plate boundaries are related in space to metalliferous deep-sea sediments and to the possible occurrence of massive strata-bound metallic sulfide deposits in oceanic crust. 3. Convergent plate boundaries are related in space to areas of offshore petroleum potential FIG. 5 -Model of metallogenesis at convergent plate boundaries of Pacific (modified from Mitchell, 1973). Diagrammatic cross sections through convergent plate boundaries show Benioff zone, oceanic crust (black) incorporating metalliferous sediments and massive strata-bound metallic sulfide bodies (white semicircles in oceanic crust), rising magma (solid vertical arrows), rising volatile matter (dashed vertical arrows), silicic volcanic rocks and granitic plutons (+), andesitic volcanic rocks and tonalitic plutons (X ), and known deposits of porphyry copper (PCu), tin (Sn), tungsten (W), fluorite (F), and antimony (Sb). For explanation, see text. 427 56 Peter A. Rona and Lawrence D. Neuman and to onshore deposits of precious, base, iron and ferro-alloy metals. 4. Certain precious, base, iron, and ferro-alloy metal deposits that are present in Precambrian shields and old orogenic belts distant from pre- sent plate boundaries may be related to former plate boundaries. 5. Models suggest genetic relations between the observed distribution of metal deposits and geo- logic processes at plate boundaries. The models are interpretative in that they attempt to explain the observed distribution of deposits. With fur- ther development these models may predict the locations of undiscovered metal deposits. References Cited Anonymous, 1962, Oil and natural gas map of Asia and the Far East, scale 1:5,000,000: U.N. ECAFE, no. 16351, U.N. Pub. Sales No. 62-1-16, 4 sheets. 1963, Mineral distribution map of Asia and the Far East, scale 1:5,000,000: U.N. ECAFE, no. 63-157 14, U.N. Pub. Sales No. 63-1-18, 4 sheets. 1970, Energy atlas of Asia and the Far East: U.N. ECAFE, U.N. Pub. Sales No. E-70-1-11, E/CN-1 1/900, 25 p. 1972a, Principal mineral areas of Canada, scale 1:7,600,200, 22nd ed.: Canada Dept. Energy, Mines and Resources, Mineral Resources Br., and Canada Geol. Survey, Map 900A. 1972b, The Times atlas of the world, 4th ed. (revised): Edinburgh, John Bartholomew and Son Ltd., Times Newspapers Ltd., 272 p. Atwater, T., and J. D. Mudie, 1973, Detailed near-bot- tom geophysical study of the Gorda Rise: Jour. Geo- phys. Research, v. 78, p. 8665-8686. Betzer, P. R., et al, 1974, The Mid- Atlantic Ridge and its effect on the composition of particulate matter in the deep ocean: Eos (Am. Geophys. Union Trans.), v. 55, no. 4, p. 293. Bonine, M. E., et al, 1970, Atlas of Mexico: Univ. Texas at Austin, Bur. Business Research, 138 p. Bostrom, K., and M. N. A. Petersen, 1969, The origin of aluminum-poor ferromanganoan sediments in areas of high heat flow on the East Pacific Rise: Marine Geology, v. 7, p. 427-447. Bracey, D. R., and T. A. Ogden, 1972, Southern Maria- na arc; geophysical observations and hypothesis of evolution: Geol. Soc. America Bull., v. 83, p. 1509- 1522. Brown, G. C, 1973, Evolution of granite magmas at destructive plate margins: Nature; Phys. Sci., v. 241, p. 26-28. Cook, H. E., 1972, Stratigraphy and sedimentation, p. 933-943 in J. D. Hays et al, eds., Initial reporjts of the Deep Sea Drilling Project, v. IX: Washington, U.S. Govt. Printing Office, 1205 p. Cronan, D. S., et al, 1972, Iron-rich basal sediments from the eastern equatorial Pacific; Leg 16, Deep Sea Drilling Project: Science, v. 175, no. 4017, p. 61-63. Dengo, G, and E. Levy, 1970, Estudios et mapa metalo- geneticos de America Central: Inst. Centroamericano Inv. y Tecnologia Indus., Pub. Geol. No. 3, 57 p., scale 1:200,000. Dymond, J., et al, 1973, Origin of metalliferous sedi- ments from the Pacific Ocean: Geol. Soc. America Bull., v. 84, p. 3355-3372. Fischer, A. G., et al, 1971, Initial Reports of the Deep Sea Drilling Project, v. VI: Washington, U.S. Govt. Printing Office, 1329 p. Gass, I. G., and D. Masson-Smith, 1963, The geology and gravity anomalies of the Troodos Massif, Cy- prus: Royal Soc. London Philos. Trans., Ser. A, v. 255, p. 417-467. Herron, E. M., 1972, Sea-floor spreading and the Ceno- zoic history of the east-central Pacific: Geol. Soc. America Bull., v. 83, p. 1671-1692. Isacks, B., J. Oliver, and L. R. Sykes, 1968, Seismology and the new global tectonics: Jour. Geophys. Re- search, v. 73, p. 5855-5899. and P. Molnar, 1971, Distribution of stresses in the descending lithosphere from a global survey of focal mechanism solutions of mantle earthquakes: Rev. Geophys. and Space Physics, no. 9, 103 p. Jakes, P., and A. J. R. White, 1971, Composition of island arcs and continental growth: Earth and Plane- tary Sci. Letters, v. 12, p. 224-230. and — 1972, Major and trace element abundances in volcanic rocks of orogenic areas: Geol. Soc. America Bull., v. 83, p. 29-39. Jones, D. B., advisory ed., 1972, Oxford economic atlas of the world, 4th ed.: London, Oxford Univ. Press, 239 p. Karig, D. E., 1971, Origin and development of marginal basins in the western Pacific: Jour. Geophys. Re- search, v. 76, p. 2542-2561. Kelleher, J. A., 1972, Rupture zones of large South American earthquakes and some predictions: Jour. Geophys. Research, v. 77, p. 2087-2103. Kennedy, W. Q., and J. E. Richey, 1947, Catalogue of the active volcanoes of the world: Bull. Volcanol., Supp. Ser. 2, v. 7, 11 p. Lafitte, P., and P. Rouveyrol, 1965, Carte miniere du globe sur fond tectonique au 1:20,000,000 (projection equiareale): Paris, Bur. Recherches Geol. et Minieres, 74, ed. provisoire, pt. 2. Langseth, M. G., Jr., 1969, The flow of heat from the Earth and the global distribution at the surface: Am. Inst. Aeronautics and Astronautics, Paper No. 69- 589, 10 p. Le Pichon, X., J. Francheteau, and J. Bonnin, 1973, Plate tectonics: New York, Elsevier, 300 p. Luyendyk, B. P., 1970, Dips of downgoing lithospheric plates beneath island arcs: Geol. Soc. America Bull., v. 81, p. 3411-3416. McKelvey, V. E., and F. F. H. Wang, 1969, World sub- sea mineral resources: U.S. Geol. Survey Misc. Geol. Inv. Map 1-632. Mitchell, A. H. G., 1973, Metallogenic belts and angle of dip of Benioff zones: Nature; Phys. Sci., v. 245, p. 49-52. Moores, E. M., and F. J. Vine, 1971, The Troodos Mas- sif, Cyprus, and other ophiolites as oceanic crust; evaluation and implications: Royal Soc. London Phi- los. Trans., Ser. A., v. 268, p. 443-466. 428 Plate Tectonics and Mineral Resources 57 Oliver, J., et al, 1973, Dynamics of the downgoing litho- sphere: Tectonophysics, v. 19, p. 133-147. Oxburgh, E. R., and D. L. Turcotte, 1971, Origin of paired metamorphic belts and crustal dilation in is- land arc regions: Jour. Geophys. Research, v. 76, p. 1315-1327. Piper, D. Z., 1973, Origin of metalliferous sediments from the East Pacific Rise: Earth and Planetary Sci. Letters, v. 19, p. 75-82. Pitman, W. C. Ill, R. L. Larson, and E. M. Herron, 1974, Age of the ocean basins determined from mag- netic anomaly lineations: Geol. Soc. America, map. Roberts, R. J., and E. M. Irving, 1957, Mineral deposits of Central America: U.S. Geol. Survey Bull. 1034, 205 p. Rona, P. A., 1973a, Plate tectonics and mineral re- sources: Sci. American, v. 229, no. 1, p. 86-95. 1973b, Relations between rates of sediment ac- cumulation on continental shelves, sea-floor spread- ing, and eustacy inferred from the central North At- lantic: Geol. Soc. America Bull., v. 84, p. 2851-2872. and D. U. Wise, 1974, Symposium: global sea level and plate tectonics through time: Geology, v. 2, no. 3, p. 133-134. et al, 1975, Anomalous water temperatures over Mid-Atlantic Ridge crest at 26° North latitude: Deep Sea Research, v. 22, p. 611-618. et al, in press, Tectonic fabric and hydrothermal activity of Mid-Atlantic Ridge crest (lat- 26°N): Geol. Soc. America Bull. Rub, M. G., 1972, The role of the gaseous phase during the formation of ore bearing magma tic complexes: Chem. Geology, v. 10, p. 89-98. Sawkins, F. J., 1972, Sulfide ore deposits in relation to plate tectonics: Jour. Geology, v. 80, p. 377-397. Sayles, F. L., and J. L. Bischoff, 1973, Ferromanganoan sediments in the equatorial East Pacific: Earth and Planetary Sci. Letters, v. 19, p. 330-336. Sclater, J. G., 1972, Heat flow and elevation of the mar- ginal basins of the western Pacific: Jour. Geophys. Research, v. 77, p. 5705-5709. Scott, M. R., et al, 1974, Rapidly accumulating manga- nese deposit from the median valley of the Mid-At- lantic Ridge: Geophys. Research Letters, v. 1, p. 355-358. Scott, R. B., et al, 1974, The TAG Hydrothermal Field: Nature, v. 251, p. 301-302. Sillitoe, R. H., 1972, A plate tectonic model for the ori- gin of porphyry copper deposits: Econ. Geology, v. 67, p. 184-197. Sleep, N. H., and M. N. Toksoz, 1971, Evolution of marginal basins: Nature, v. 233, p. 548-550. Sloss, L. L., 1963, Sequences in the cratonic interior of North America: Geol. Soc. America Bull., v. 74, p. 93-114. Snead, R. E., 1972, Atlas of world physical features: New York, Wiley, 158 p. Sorokhtin, O. G., S. A. Ushakov, and V. V. Fedynskiy, 1974, Dynamics of crustal plates and origin of petro- leum deposits: Akad. Nauk SSSR Doklady, v. 214, p. 1407-1410. Spooner, E. T. C., and W. S. Fyfe, 1973, Sub-sea floor metamorphism, heat and mass transfer: Contr. Min- eralogy and Petrology, Springer- Verlag, p. 287-304. Stern, C. R., and P. J. Wylie, 1973, Water-saturated and undersaturated melting relations of a granite to 35 kilobars: Earth and Planetary Sci. Letters, v. 18, p. 163-167. Strakhov, N. M., et al, 1968, Geochemistry of a sedi- mentary manganese ore-forming process: Akad. Nauk SSSR Geol. Inst. Trans., v. 185, 495 p. Tarling, D. H., 1973, Continental drift and reserves of oil and natural gas: Nature, v. 243, p. 277-279. Van Royan, W., and O. Bowles, 1952, The mineral re- sources of the world, in Atlas of the world's resources, v. II: New York, Prentice-Hall, 181 p. Vine, F. J., and J. T. Wilson, 1965, Magnetic anomalies over a young oceanic ridge off Vancouver Island: Sci- ence, v. 150, no. 3695, p. 485^189. von der Borch, C. C, W. D. Nesteroff, and J. S. Gale- house, 1971, Iron-rich sediments cored during Leg 8 of the Deep Sea Drilling Project, in J. I. Tracey, Jr., et al, eds., Initial reports of the Deep Sea Drilling Pro- ject, v. VIII: Washington, U.S. Govt. Printing Office, p. 725-819. Waring, G. A., 1965, Thermal springs of the United States and other countries of the world: a summary: U.S. Geol. Survey Prof. Paper 492, 383 p. 429 45 Reprinted from: Geological Society of America Bulletin. Vol. 87, 661-674. Tectonic fabric and hydrothermal activity of Mid-Atlantic Ridge crest (lat 26°N) PETER A. RONA REGINALD N. HARBISON* BOBBY G. BASSINGER* National Oceanic and Atmosphertc Administration, Atlantic Oceanographtc and Meteorological Laboratories, IS Rickenbacker Causeway, Miami, Florida 13149 ROBERT B. SCOTT Department of Geology; Texas A&M University, College Station, Texas 77843 ANDREW J. NALWALKf Marine Sciences Institute, University of Connecticut, Croton, Connecticut 06 ?40 ABSTRACT An asymmetric tectonic fabric was de- lineated by narrow-beam bathymetric profiles in a 180-km- area of the Mid- Atlantic Ridge crest at lat 26°N. Features of the tectonic fabric are a continuous rift val- ley offset by small (<10-km) transform faults and minor fracture zones expressed as valleys with intervening ridges that trend normal and oblique to the two sides of the rift valley. The discharge zone of a pos- tulated sub-sea-floor hydrothermal con- vection system is focused by faults on the southeast wall of the rift valley and driven by intrusive heat sources beneath the rift valley. The rift valley has a double structure consisting of linear segments, bounded by ridges, and basins at the intersections of the minor fracture zones. The double structure of the rift valley acts like a template that programs the reproduction of the tectonic fabric. The minor fracture zones form an asymmetric V about the rift valley at var- iance with the symmetric small circles formed by major fracture zones. To recon- cile the asymmetry of minor fracture zones with the symmetry of major fracture zones, it is proposed that the minor fracture zones have been preferentially reoriented by an external stress field attributed to interplate and intraplate motions. Major fracture zones remain symmetric under the same stress field owing to differential stability be- tween minor and major structures of oceanic lithosphere. Key words: oceanic ridges, Mid-Atlantic Ridge, tectonic fabric, fracture zones, transform faults, rift valley, hydrothermal activity, hydrothermal min- " Present address: U.S. Geological Survey, P.O. Box 7944, Metaine, Louisiana 70011. t Deceased. eral deposits, sub-sea-floor hydrothermal convection system. INTRODUCTION The crest of the Mid- Atlantic Ridge at lat 26°N in the corridor of the Trans-Atlantic Geotraverse (TAG) of the National Oceanic and Atmospheric Administration (NOAA) was selected for study because the oceanic crust in this region is believed to have un- dergone a long history of relatively normal development isolated from mantle plumes and multiple plate boundaries (Fig. 1; Rona, 1973a). The tectonic fabric of the ridge crest at lat 26°N should be useful as a standard of crust produced from a slow- spreading ridge (Menard, 1967). An interdisciplinary study of the Mid- Atlantic Ridge crest at lat 26°N performed as part of the TAG project in 1972 and 1973 included narrow-beam bathymetric, gravimetric, and magnetic profiling, dredg- ing, coring, measurements of the thermal structure and chemistry of the water col- umn, and ocean-bottom photography. The study resulted in the discovery of the TAG Hydrothermal Field, an area of at least 15 km2, including the southeast wall of the rift valley, where manganese oxide of hy- drothermal origin is present (Fig. 2; Scott, R. B., and others, 1974) and hydrothermal solutions may be discharging from the sea floor (Rona and others, 1975). Previous evidence of the concentration of metals by hydrothermal processes in ocean basins has come from widespread metal enrichment in sediment immediately overlying basalt gen- erated by sea-floor spreading about di- vergent plate boundaries (Degens and Ross, 1969; Bostrom and Peterson, 1966; Bostrom and others, 1972; von der Borch and Rex, 1971; von der Borch and others, 1971; Hollister and others, 1972; Dvmond and others, 1973; Sayles and Bischoff, 1973; Piper, 1973). This report describes the regional tec- tonic fabric of a 180-km square of the Mid-Atlantic Ridge crest at lat 26°N cen- tered on the TAG Hydrothermal Field and considers the relation of the tectonic fabric to hydrothermal activity. The tectonic fab- ric was delineated by a rectilinear grid of narrow-beam bathymetric, gravimetric, and magnetic profiles spaced about 10 km apart parallel to and 20 km apart perpen- dicular to the axis of the rift valley (Figs. 3 through 8). Directions of features are given in the azimuth system (0° to 360° clockwise 0 7 / / ^l^> - — -C_££cro,5 [ „., ' ' ,' 40'N"_ ' LWfT' — ^-—-LJ Figure 1. Inset shows location of study area on the crest of the Mid-Atlantic Ridge in the cor- ridor (striped) of the NOAA Trans-Atlantic Geotraverse (TAG) in the central North Atlantic Ocean between southeastern North America and northwestern Africa. Locations and trends of transverse valleys (minor fracture zones) de- lineated within the study area are shown by solid lines. The axis of the Mid-Atlantic Ridge is dashed. The locations of major fracture zones (Atlantis and Kane) are shown. Geological Society of America Bulletin, v. 87, p. 661-674, 1 1 figs.. May 1976, Doc. no. 60504. 661 430 662 RONA AND OTHERS Figure 2. Photograph of first specimen of hydrothermal manganese oxide, 15.0 cm in diameter and 4.2 cm thick, recovered from the site of the TAG Hydrothermal Field in 1972 (Table 2, Station TAG 1972-13, see footnote 1; photograph by Wayne Stevens). from north). Primary position control was by a satellite navigational system with esti- mated accuracy of ±1.0 km. BATHYMETRY Rift Valley Features of the tectonic fabric in the study area are the rift valley of the Mid- Atlantic Ridge, three valleys with interven- ing ridges transverse to the rift valley, and valleys that transect the transverse ridges (Figs. 3, 4). The rift valley trends northeast and consists of two structural elements that alternate along its axis. Linear segments of the rift valley, 20 to 40 km in length, alter- nate with irregularly shaped areas 15 to 25 km in length. The azimuth of the linear segments is 25°. The width of the linear segments at the 3,400-m isobath, which continuously delineates the floor of the rift valley, ranges between 5 and 15 km; the width of the irregularly shaped areas ranges between 15 and 25 km. The irregularly shaped areas between the linear segments are occupied by topographic depressions that form basins between 200 and 500 m below the 3,800-m isobath. The linear seg- ments of the rift valley are successively off- set in a right-lateral sense by as much as 10 km at the irregularly shaped areas. Transverse Valleys and Intervening Ridges The transverse valleys and intervening ridges on the southeast side of the rift valley trend about 115°, nearly perpendicular to the rift valley axis (Figs. 3, 4). Those on the northwest side trend about 265°, oblique (about 60°) to the axis of the rift valley (Rona and others, 1973). The trend of the transverse valleys and intervening ridges appears to change near the northwestern margin of the study area. Intersections of the transverse valleys and intervening ridges with the rift valley are topographically complex (Fig. 3). Certain of the transverse valleys are continuous with the rift valley (the southernmost trans- verse valley and the western half of the cen- tral transverse valley). Other transverse val- leys are blocked by topographic highs within 25 km of the rift valley (the north- ernmost transverse valley and the eastern half of the central transverse valley). The actual or projected juncture of each trans- verse valley with the rift valley occurs at one of the irregularly shaped areas where the linear segments of the rift valley are off- set. All transverse ridges continue up to the rift valley, where they form the walls along the linear segments. Narrow-beam bathymetric profiles (ef- fective total beam width about 20°) across the three transverse valleys and intervening ridges are shown in Figure 5. The normal and oblique orientations of the transverse valleys and ridges with respect to the axis of the rift valley are apparent from the align- ment of the transverse valleys. Width of the transverse valleys at their floors is about 10 km, and the average spacing between adja- cent valley floors is 55 km. Relief of the intervening ridges measured from the valley floors ranges between 1,000 and 2,000 m. The transverse valleys appear linear and continuous along their axis on the basis of the bathymetric profiles (Fig. 5) and isobaths (Fig. 3), with the exception of the topographic complexities noted at certain junctures with the rift valley. The floor of each transverse valley progressively deepens away from the rift valley, and the mean depth successively increases from northeast to southwest (Fig. 5). Narrow-beam bathymetric profiles that cross the study area transverse to the axis of the rift valley (Fig. 6; also see McGregor and Rona, 1975, Fig. 5) reveal four things: (1) Echo returns are predominantly diffrac- tions, as distinguished from specular reflections (Hoffman, 1957), indicating that the rock surface is irregular relative to the 12.5-cm wavelength of the projected acous- tic pulse. (2) The floor of the rift valley is irregular and has relief up to several hundred metres that is related to the pres- ence of discontinuous linear topographic prominences that project from either wall (Fig. 6, profiles C, D, E, H, J) or stand near the center of the rift valley (Fig. 6, profiles B through G), subparallel with the axis. (3) Steplike topographic levels with relief of hundreds of metres and widths of kilometres are present on both walls of the rift valley; the steplike levels may be corre- lated along either wall for distances corre- sponding to at least the width of each transverse ridge (about 30 km; Fig. 6, profiles B through G). (4) The mean inclina- tion of the two walls of the rift valley ranges between about 5° and 35°, with no systema- tic difference apparent between the walls. Branches df the Rift Valley The rift valley has branches that extend from either side (Figs. 3, 4). Branches ex- tending from the southeast side trend nearly parallel to the rift valley (25°). Some branches that extend from the northwest side trend nearly parallel (25°), and others trend oblique (355°) to the rift valley. Valleys parallel to the branches of the rift valley transect the transverse ridges at ir- regular intervals, with a minimum spacing of 5 km. These valleys generally extend only partly across the transverse ridges, al- though some valleys extend across the ridges. The valleys that transect the trans- verse ridges on the southeast side of the rift valley trend nearly parallel to the rift valley (25°) and perpendicular to the transverse ridges. The valleys that transect the trans- verse ridges on the northwest side of the rift valley have two trends corresponding to the two trends of the branches on the north- west side of the rift valley: (1) a trend (355°) oblique to the axis of the rift valley and nearly perpendicular to the transverse ridges, and (2) a trend (25°) parallel to the rift valley and oblique to the transverse ridges. The two trends appear to intersect. TAG Hydrothermal Field The TAG Hydrothermal Field occupies a salient of the southeast wall that projects into the rift valley between depths of about 3,500 and 2,000 m (Fig. 3). The salient is 431 45°30'W TECTONIC FABRIC AND HYDROTHERMAL ACTIVITY OF MID-ATLANTIC RIDGE CREST 45°00' 44°30' 44°00' 663 43°30'W 46°00'W 45°30 45°00' 44°30' Figure 3. Isobath map of the study area (Fig. 1) contoured in hundreds of metres corrected for ship's draft and vertical sounding velocity (Matthews, 1939). Depths exceeding 3,400 m are shaded. Sounding tracks are dashed. The known portion of the TAG Hydrothermal Field on the southeast wall of the rift valley is outlined (trapezoid). Dredge and core stations are marked with dots (Table 2). Dots are omitted for three dredge stations within the TAG Hydrothermal Field (Table 2, stations TAG 1972-13, TAG 1973-2A, TAG 1973-3A). Values of heat flow in heat-flow units are shown at five stations marked by crosses (Langseth and others, 1972). The northeast-trending sounding tracks of bathymetric profiles 1 through 4 shown in Figure 5 are labeled (solid lines). The northwest-trending sounding tracks of bathymetric profiles A through K shown in Figure 6 are labeled. Note the north arrow in the upper left corner. Contour interval 200 m. the end of a transverse ridge and forms the wall along a linear segment of the rift valley just north of the intersection with the cen- tral transverse valley. This transverse ridge is transected by more valleys perpendicular to its axis than any other ridge in the study area. The southeast wall of the rift valley at the TAG Hydrothermal Field has steplike topographic levels several kilometres wide at depths of about 3,200 and 2,500 m (Fig. 6, profile F). Two stereophotograph tran- sects of this area resolve steps on the wall that are 100 to 300 m wide with 75-m av- erage relief between depths of 3,400 and 3,100 m, 30 to 400 m wide with 40-m av- erage relief between depths of 2,800 and 2,400 m, and smaller steps between depths of 2,500 and 2,400 m, indicating that steps in the wall range in scale from metres to kilometres (McGregor and Rona, 1975). GRAVITY Measurements of gravity were made with a shipborne Graf-Askania gravimeter to an estimated accuracy of ±5 mgal. The free-air 432 664 RONA AND OTHERS gravity anomalies of the study area range between -40 and +80 mgal (Fig. 7). Nega- tive free-air gravity anomalies coincide with the rift valley and the transverse valleys. Positive free-air anomalies coincide with the transverse ridges, and largest values are on the highest parts of the transverse ridges adjacent to the rift valley. The TAG Hy- drothermal Field occurs adjacent to such an anomaly. The values of gravity agree in magnitude with a rough calculation of the terrain effect. These observations suggest that the free-air anomalies are primarily due to topography. MAGNETIC MEASUREMENTS Linear magnetic anomalies parallel the rift valley (Fig. 8). The linear anomalies are offset at transverse valleys in direction and amount corresponding to the offset of the linear segments of the rift valley. Low to negative values of residual magnetic inten- sity are associated with the transverse val- leys. The linear magnetic anomalies are in- terpreted to have been generated during the Brunhes (axial anomaly), Matuyama, Gauss, and Gilbert magnetic polarity epochs between 0 and 5.8 m.y. B.P. The positive axial anomaly does not coincide with the axis of the rift valley but is cen- tered over the southeast wall 5 km from the axis. About 2 km of the 5-km offset may be attributed to shift in the magnetic field owing to superposition of the sea floor and regional magnetic fields. The polarity of the residual magnetic field is indistinct between the end of the Gilbert epoch and anomaly 5 (10 m.y. B.P.) of the magnetic polarity re- versal time scale (Heirtzler and others, 1968) identified about 120 km to either side of the rift valley. Half rates of sea-floor spreading measured perpendicular to the axis of the rift valley and averaged over var- ious intervals to 10 m.y. B.P. are asymmet- ric, being faster to the southeast than to the northwest (Table 1). The TAG Hydrothermal Field is situated within the positive axial anomaly and is as- sociated with irregularities in the shape of that anomaly (Fig. 8). A more detailed magnetic survey reveals a pronounced low in the axial anomaly that coincides with the TAG Hydrothermal Field (Fig. 9; McGregor and Rona, 1975). Magnetic modeling indicates that the magnetic low is due to reduction in intensity of remanent magnetization (McGregor and others, 1976) that may be attributed to alteration of the magnetic mineral component in basalt by hydrothermal solutions (Luyen- dyk and Melson, 1967; Watkins and Pas- ter, 1971; Ade-Hall and others, 1971). Hydrothermal alteration of basalt is evi- denced by the presence of greenstone at 4S°30 — r- 1 \ ... ' vj* IRREGULARIY. ■SHAPta I \AREA TRANS RIDGE TRANSVERSE VAUEY AV VA>IlLl|l°>„I i • I N z. <•* .\0°v \lRREGUlARtY .™*NSV"SE VAllEY <-. Vf. \ - N SHAPED TRANSVERSE RIDGE S$&\ \ . IRREGUIARIY \* \ 0cA \ SHAPED 1* IV" A,EA I»ANSVE>SE .^ , r ^\ Figure 4. Diagram- matic representation of the tectonic fabric of the study area out- lined from Figure 3. Azimuths of the major trends are noted. The TAG Hydrothermal Field is outlined (trap- ezoid). Note the north arrow in the upper left corner. several sites along transverse ridges (Table 2"; stations TAG 1972-8, TAG 1972-15). HEAT FLOW Five measurements of conductive transfer of heat through the sea floor were made by Langseth and others (1972). The heat-tlow measurements show higher heat flow through the transverse ridges (3.26 to 8.60 HFU) than through the intervening trans- verse valleys (2.04 to 2.69 HFU; Fig. 3). From the limited number of measurements, it is ambiguous whether this distribution of values is related to topography or to dis- tance from the rift valley. A large variation in heat flow occurs over a horizontal dis- tance of 5 km on one of the transverse ridges (6.75 and 3.26 HFU). A water-temperature profile parallel to the ocean bottom over the southeastern wall of the rift valley at the TAG Hydro- thermal Field was made with a 4-m-long vertical array of three thermistors mounted on a towed deep-sea camera (Rona and others, 1975). The profile revealed an ab- rupt anomaly of +0. 1 1°C associated with a gradient of 0.008°C/m, warming down- ward within 20 m of the bottom along a horizontal distance of about 350 m be- tween depths of 3,030 and 2,950 at a step- like level on the southeast wall (Fig. 6, profile F), where hydrothermal material 'Table 3, "Rocks recovered from the Study Area of the NOAA Trans-Atlantic Geotraverse (TAG) on the Mid-Atlantic Ridge Crest at lat. 26°N," GSA sup- plementary material 76-4, may be ordered from Docu- ments Secretary, Geological Society of America, 3300 Penrose Place, Boulder, Colorado 80301. DISTANCE (KILOMETERS) 0 50 100 150 SW Figure 5. Digitzed reproductions of narrow- beam bathymetric profiles 1 through 4 recorded along sounding tracks parallel to the rift valley located in Figure 3. The three transverse valleys correlate (dashed lines) as continuous features on either side of the rift valley, consistent with the isobath map (Fig. 3). Vertical exaggeration is about x60. 433 TEC IONIC FABRIC AND HYDROTHKRMA1 ACTIVITY OF MID-ATLANTIC RIDGE CREST 665 was dredged .Tabic 2. dredge station I AC. 1973-3A). The source of the water- temperature anomah may be either dis- charge of hvdrothermal solutions or con- ductive transfer ot heat from the ocean bot- tom. 1 he abrupt, narrow character of the anomalv, inverted gradient, and associated geological and geochemical evidence favor a hvdrothermal source. PETROLOGY Rocks recovered by dredging and coring from the different features of the tectonic fabric in the study area are described in Table 2 i see footnote 1); sampling sites are marked in Figure 3. Fresh pillow basalt was recovered from the topographic promi- nences t)ii the floor of the ritt valley (Table 2, stations TAC, 1972-17, TAG 1973-4C, TAG 19~3-4G). Fresh basalt was also re- covered from the walls of the rift valley. A varied suite of altered and metamorphosed rocks and limestone in addition to basalt was recovered from the transverse ridges. A coarse-grained cumulate gabbro exposed in the wall of a valley that transects one of the transverse ridges may be a magma chamber formed in laver 3 of oceanic crust (Table 2, stations TAG 1973-7A, TAG 1973-7B, TO 75AK60-2A). Manganese oxide was consistently dredged from the southeast wall of the rift valley within the area of the TAG Hydro- thermal Field (Fig. 3; Table 2, stations TAG 1972-13, TAG 1973-2A, TAG 1973-3 A). The manganese oxide occurs as a crust up to 42 mm thick on basalt talus, as veins in the talus fragments, and as a crust on and matrix in a breccia of basalt fragments. Stereophotograph transects of the rift valley wall show that the manganese oxide is as- sociated with sediment-free talus and breccia-covered inner portions of the steps observed on the southeast wall between depths of 3,100 and 2,500 m (McGregor and Rona, 1975). The accumulation rate of the manganese oxide measured from the growth rate of Th-:i"and Pa-" toward secu- lar equilibrium with their uranium parents is about 200 mm/10" yr, two orders of magnitude faster than deep-sea hydrogen- ous ferromanganese nodules and crusts (Scott, M. R., and others, 1974). Atomic absorption spectrophotometry of the man- ganese indicates that the Mn content is 40 percent, Fe less than 0.1 percent, Al 0.1 percent, Zn 100 ppm, Cr 15 ppm, Ni 300 ppm, Co 20 ppm, and Cu 30 ppm; deep-sea hydrogenous ferromanganese nodules and crusts contain less Mn (10 to 20 percent) and considerably more of the other ele- ments (Scott, M. R., and others, 1974). The relatively rapid rate of accumulation and the pure composition of the manganese oxide indicate concentration by a hy- drothermal mechanism capable of extreme chemical fractionation. A crust consisting of an upper layer of hydrogenous man- ganese oxide up to 2 mm thick underlain by a lower layer of hydrothermal manganese oxide up to 10 mm thick was recovered from a site 12 km southeast of the sites at the southeast wall of the rift valley on the same transverse ridge (Fig. 3; Table 2, sta- TABLE HALF RATES OF SEA-FLOOR SPREADING ABOUT THE MID-ATLANTIC RIDGE Reference Latitude on Mid- Atlantic Ridge Orientation Averaging interval van Andel and Moore (1970) Lattimore and others (1974) McGregor and others (1976) McGregor and others (1976) Present paper Needham and Francheteau (1974) Needham and Francheteau (1974) Greenewalt and Taylor (1974) Macdonald and others (1975b) Macdonald and others (1975 b) Loncarevic and Parker (1971) 6° to 8°S Perpendicular to axis of rift vallev 26CN Perpendicular to axis of riff valley 26°N Perpendicular to axis of rift vallev 26°N Perpendicular to axis of rift valley 26°N Normal (10"°) and oblique (265°) to axis of rift valley (25°) 36°N Perpendicular to axis of rift valley 36°N Perpendicular to axis of rift valley 36°N Perpendicular to axis of rift valley 36°N Perpendicular to axis of rift valley 36°N Perpendicular to axis of rift valley 45°N Perpendicular to axis of rift valley Axis of rift valley (0 m. v. B.P.) to anomaly 3 (4.5 m.y. B.P.) Anomaly 3 (4.5 m.y. B.P.) to anomaly 5 (10 m.y. B.P.) Axial anomaly to anomaly 5 (10 m.y. B.P.) Axis of rift valley (0 m.y. B.P.) to Brunhes- Matuyama boundary (0.69 m.y. B.P.) Matuyama epoch (0.69 to 2.43 m.y. B.P.) Axis of rift valley (0 m.y. B.P.) to anomaly 5 (10 m.y. B.P.) Axis of rift valley (0 m.y. B.P.) to Brunhes- Matuvama boundary (0.69 m.y. B.P.) Matuyama epoch (0.69 to 2.43 m.y. B.P.) Axis of rift valley (0 m.y. B.P.) to Brunhes- Matuyama boundary (0.69 m.y. B.P.) Axis of rift valley (0 m.y. B.P.) to anomaly 2 (1.8 m.y. B.P.) Anomaly 2 (1.8 m.y. B.P.) approximately to anomaly 4 (8 m.y. B.P.) Center of Brunhes (0 m.y. B.P.) to anomaly 5 (10 m.y. B.P.) Average h; alf rate Direction (cm/yr) 2.16 ± 0.24 E 1.89 ± 0.04 W 1.59 ± 0.24 E 1.12 ± 0.08 W 1.3 SE 1.1 NW 1.7 SE 0.7 NW 1.3 SE 1.1 NW 1.3 SE 1.3 NW 1.5 E 0.7 W 1.3 E 0.9 W 1.3 E 1.0 W 1.33 E 0.70 W 0.95 E 1.35 W 1.10 E 1.28 W 434 666 RONA AND OTHERS nun TO 75AK61-1A; Scott and others, 1975). Seven attempts to dredgt the north- west wall of the n ft valley opposite the TAG Hydrothermal Field to determine whether hydrothermal deposits were sym- metrically disposed about the rift valley re- covered only a few fragments of basalt (Table 2, station 1 AC, I97.3-5C). The hy- drothermal material is sufficiently triable that samples would probably have been re- covered if present on the northwest wall. DISCUSSION Comparison of Tectonic Fabric The continuous rift valley at lat 26°N consisting of linear segments and basins (Figs. 3, 4) is similar to that observed along the Mid-Atlantic Ridge between lat 22CN and 23°N (van Andel and Bowm, 1968), at lat 36CN (Needham and Francheteau, 1974), at lat 45°N (Aumento and others, 1971), and between lat 47°N and 51CN (Johnson and Vogt, 1973). The topo- graphic prominences on the floor of the rift valley at lat 26°N, from which fresh basalt was recovered, appear similar to the discon- tinuous medial ridge as much as 250 m high by 1 km wide and as much as 4 km long described from the rift valle\ at lat 36CN (Bellaiche and others, 1974; Needham and Francheteau, 1 9^4; Moore and others, 1974; Macdonald and others, 1975a), in- terpreted as a locus of crustal emplacement and basalt eruption. Steplike levels and steps in the walls of the rift valley on scales ranging from metres to kilometres similar to those at lat 26°N have been observed in walls of the rift val- ley at lat 36GN (Needham and Francheteau, 1974) and at the Gorda Rise, where the steps are interpreted as fault blocks (Atwa- ter and Mudie, 1973). Block faulting involv- ing uplift is evidenced in the study area by the exposure of the cumulate gabbro from a deeper crustal level in an inferred fault scarp 1.2 km high (Table 2, stations TAG 1973-7A, TAG 19~3-7B, TO T"5AK6()-2A) and by the exposure of greenstone in the walls of transverse ridges. Transverse ridges with intervening valleys that intersect and offset the rift valley at lat 26°N at a spacing of 55 km occur at an average spacing of 65 km along the Mid-Atlantic Ridge between lat 10°N and 40°N (Perry and Feden, 1974) and are present at lat 36°N (Derrick and others, 1973). The portions of the trans- verse valley that transect the rift valley are the loci of earthquake epicenters at lat 26°N (McGregor and Rona, 1975) and at lat 37°N (Reid and Macdonald, 1973). The asymmetry in tectonic fabric, rates of sea-floor spreading, and position of the axial magnetic anomaly at lat 26CN is ob- served at other intensively studied sites along the Mid-Atlantic Ridge. Transverse valleys and intervening ridges that trend normal and oblique to the southeast and northwest sides of the rift valley, respec- tively, at lat 26°N are also present at lat 36°N (Detnck and others, 1973; H. Flem- ing, 1975, personal commun.), and at lat 36°N, where the normal and oblique sides reverse (Bhattacharyya and Ross, 1972). Between lat 47°N and 51°N, the azimuths of linear segments of the rift valley alternate north and northwest; transverse ridges occur adjacent to the former and interven- ing transverse valleys occur adjacent to the latter, forming a V-shaped pattern that is asymmetric about the rift valley (Johnson and Vogt, 1973, Fig. 7). Half rates of sea-floor spreading mea- sured perpendicular to the axis of the rift valley of the Mid-Atlantic Ridge and aver- aged over corresponding time intervals are faster to the east than to the west at lat 36°N, similar to the half rates at lat 26°N AXIAL VALLEY D H K 3000 h 4000_ OKM 100 200 Figure 6. Digitized reproductions of narrow-beam bathymetric profiles recorded along sounding tracks A through K transverse to the rift valley located in Figure 3. The rift valley and the TAG Hydrothermal Field (arrow on profile F) are noted. Steplike topographic levels on the rift valley walls are tentatively correlated by dashed lines between the profiles. Faults inferred between fault blocks are indicated by dashed lines on profiles. Vertical exaggeration is about X 15. 435 TECTONIC FABRIC AND HYDROTHERMAL ACTIVITY OF MID-ATLANTIC RIDGE CREST 667 (Table 1). Conversely, average half rates of spreading at lat 45°N are faster to the west than to the east (Table 1). The axial anom- aly is centered-over the southeast wall offset about 5 km from the axis of the rift valley at lat 36°N (Needham and Francheteau, 1974) and at lat 22°N (van Andel and Bowin. 1968, Fig. 13), similar to the posi- tion of the axial anomaly at lat 26CN (Fig. 8). Earthquake epicenters along the linear segments of the rift valley may favor the southeastern side at lat 26°N (McGregor and Rona, 1975) and lat 36°N (Spindel and others, 1974) and the west side between lat 47°N and 51°N (Johnson and Vogt, 1973). Tectonic Fabric and Hydrothermal Activity The concentration of a hydrothermal mineral deposit in the Earth's crust requires special physical and chemical conditions. The emplacement of igneous rocks and the propagation of fractures and faults at di- vergent plate boundaries are conducive to the development of sub— sea-floor hydro- thermal convection systems (Spooner and Fyfe, 1973; Hutchinson, 1973; Sillitoe, 1973; Rona, 1973b; Lister, 1974b; Lowell, 1975; Bonatti, 1975). Dense, cold sea water may penetrate down fractures and acquire heat at depth, and less dense hydrothermal water may ascend and discharge as sub- marine springs through fracture systems. The circulating sea water may remove ele- ments including heavy metals from the oceanic crust through which it circulates by high-temperature leaching and mass trans- fer (Krauskopf, 1956; Holland, 1967; Helgeson, 1964; Corliss, 1971; Hart, 1973). Metals including Fe, Mn, Cu, and Ni have been experimentally leached by sea water from basalt at 200° to 500°C and 0.5 to 1 kb (Mottl and others, 1974; Bischoff and Dickson, 1975). The convection of sea water as a hydrothermal solution through rocks emplaced at divergent plate bound- aries is indicated by several lines of evi- dence. (1) Low values of heat flow from oceanic ridge crests imply that heat must be removed by water circulation (Palmason, 1967; Deffeyes, 1970; Talwani and others, 1971; Lister, 1972; Anderson, 1972; Wil- liams and others, 1974; Sclater and others, 1974). (2) Hydrous metamorphosed oceanic crust and oceanic serpentine re- quire a voluminous source of water (Miyashiro and others, 1971; Christensen, 1972). (3) Isotopic compositions of the hydrated rocks require a low 501S isotopic source such as sea water (Muelenbachs and Clayton, 1972; Spooner and others, 1974). v The TAG Hydrothermal Field is hypoth- Figure 7. Map of free-air gravity contoured in milligals. Gravity highs (H) and lows (L) are indicated. The ship's tracklines are dashed. The TAG Hydrothermal Field is outlined (trapezoid). The axis of the rift valley along the linear segments is shown. Note the north arrow in the upper left corner. esized to be the zone where a voluminous sub — sea-floor hydrothermal convection system discharges through faults between steps of the southeast wall of the rift valley that act as conduits for the hydrothermal solutions (Figs. 9, 10). The system may be charged with sea water through fractures that underlie both the adjacent transverse valleys and the many valleys that transect the adjacent transverse ridge. Experimental (Elder, 1965) and theoretical (Donaldson, 1962; Elder, 1967) modeling shows that hydrothermal discharge is localized, and re- charge is delocalized. Discharge is confined to vertically rising narrow jets (Wooding, 1963) and fracture-focused streams (Elder, 1965). Recharge occurs over large areas and involves downward water flow at rates that are fast enough to reduce the upward conductive heat flux (Wooding, 1963). Heat-flow measurements from the study area and from other areas on oceanic ridges are consistent with downwelling of sea water at valleys and upwelling at ridges (Lister, 1972), a circulation pattern favored by a geometric forcing effect of the topog- raphy (Lister, 1974a, 1974b). Such hy- drothermal circulation can account for the low-intensity hydration and metamorphism under the influence of geothermal gradients that are higher than background value in certain rocks recovered from the transverse ridges (Table 2). The southeast wall of the rift valley at the TAG Hydrothermal Field projects over the locus of crustal emplacement beneath the rift valley, the source of heat that vigor- ously drives the ascending limb of the hypothetical hydrothermal convection sys- tem. As hydrothermal solutions enriched in heavy metals complexed with sea-water chlorides ascend through the rocks, metals may combine with sulfur in the sea water and precipitate sulfides under reducing conditions within the basalt (Meyer and Hemley, 1967). Chalcopyrite, pyrrhotite, and marcasite have been experimentally grown from sea water during reaction with oceanic tholeiite at 500°C and 0.8 kb (Ha- jash, 1974). Metallic oxides may form under oxidizing conditions at the basalt- sea-water interface. Amorphous ferric hy- droxide may precipitate as a colloid in the overlying sea water (Zelenov, 1964) and scavenge the metals remaining in solution or in admixed sea water by absorption (Krauskopf, 1956). Manganese oxide of hydrothermal origin has been sampled from the basalt -sea-water interface (Scott, M. R., and others, 1974), a positive tempera- ture anomaly (0.1 1°C) and inversion of the gradient have been measured in near- bottom water (Rona and others, 1975), and amorphous particulate matter enriched in iron and manganese has been sampled from the water column overlying the TAG Hy- drothermal Field (Betzer and others, 1974). 436 668 RONA AND OTHERS Metallic sulfides have not been sampled from the TAG Hydrothermal Field, but their presence is suspected from geochemi- cal considerations and analogy with ophio- lites. Disseminated pyrite occurs in green- stone dredged from a transverse ridge in the study area (Table 2, station TAG 1972-15), and metallic sulfides are common in altered oceanic rocks (Dmitriev and others, 1970; Bonatti, 1975). Certain ophiolites, includ- ing those of Newfoundland (Upadhyay and Strong, 1973) and the Troodos Massif of Cyprus (Constantinou and Govett, 1973; Hutchinson, 1973; Robertson and Hudson, 1973), exhibit an association of cupreous pyrite bodies, altered pillow lava, and over- lying manganiferous oxides attributed to hydrothermal processes similar to those in- ferred to be operating on the Mid-Atlantic Ridge at lat 26°N. Troodos-type massive stratiform cupreous sulfide bodies may be forming in pillow lava under the manganese oxide deposits at the TAG Hydrothermal Field (Fig. 10; Rona, 1973b; Scott, R. B., and others, 1974). It is infeasible to test this hypothesis by deep-sea drilling at this time because techniques constrain drilling to areas of sediment accumulation that, as topographic lows, are the inferred inflow areas of sub-sea-floor hydrothermal con- vection systems, and hydrothermal deposits would be absent. The massive stratiform sulfides would be expected to underlie hy- drothermal discharge areas at topographic highs expressed as transverse ridges. Development of Tectonic Fabric Why is a crustal layer of uniform thick- ness not generated about the rift valley in- stead of a layer of varying thickness that forms topographic highs (transverse ridges) and lows (intervening transverse valleys)? Why is the tectonic fabric not symmetric about the rift valley? How can the asym- metry of the tectonic fabric be explained? How is hydrothermal activity related to tec- tonic fabric? These questions pinpoint problems addressed in a model that de- scribes the development of the tectonic fab- ric of the study area (Fig. 11). In the initial configuration of the model, the sea floor spreads symmetrically at equal half rates (r ,) perpendicular to either side of the axis of a rift valley (Fig. 11a). The rift valley consists of alternating linear seg- ments and irregularly shaped areas. The ir- regularly shaped areas are transected by minor fracture zones that successively offset the linear segments in a right-lateral sense. These minor fracture zones are considered to originate as small ridge-ridge transform faults on the basis of their characteristics in the study area, including offset of the linear segments of the rift valley and linear rema- nent magnetic anomalies, presence of topographic depressions, and seismicity. 46'00'W 45-30'w J500W _s). a. Profiles 12 km north of TAG Hydrothermal Field; this magnetic profile has large axial lBrunhes) positive anomaly over rift valley typical of profiles outside TAG H\ drotherm.il Field, b. Profiles across ~[ AG Hydrothermal Field. Arrow on magnetic profile points to magnetic low in positive axial Rrunhes) anomaly. Reduction in magnetic intensity may provide a useful criterion in exploration for active or relict submarine hydrothermal mineral deposits. which is a zone' of extension [Hebzen and others. 1959; Matthews and Bath, \^b~; Harrison, |96S; C ami. |9_0). The dikes and the sides of the rift valle) are parallel. The crustal material remains near its level of emplacement in the form of relative topographic lows at the irregularly shaped areas because the cold walls at the intersec- tion of the rift vallev. with the transverse tractnres ma\ cause upwelling material to solidify below its isostatic equilibrium level (Sleep and Biehler, 1970). The crust is at least 1 . s km thinner at the transverse tractnres than at the intervening transverse ridges interred from the differ- ence of topographic relief. Crustal material also solidifies below its eventual isostatic equilibrium level beneath linear segments of the rift valley (Sleep, 1969), where the ma- terial is isostaticallv uplifted in fault blocks as a result of processes of crustal thickening and extension (Deffeyes, 1970; Osmaston, 1971; Lachenhruch, J973; Parker and Old- enburg, 19-s; S'eedham and Francheteau, 1974). Sea-floor spreading at equal rates per- pendicular to either side of the rift valley results in the generation of transverse val- leys about the topographic lows at the ir- regularly shaped areas of the rift valley and intervening transverse ridges about the topographic highs at the linear segments. The transverse ridges are constructed of fault blocks that are uplifted from the floor and accrete at the walls of the rift valley; each fault block is several kilometres wide, and its long axis lies parallel to the ntt valley. The transverse ridges are transected per- pendicular to their axes by valleys. I he \ al- ley s originate as branches nearly parallel to the rift \ allev that remain near the level of the rift valley floor during differential uplift of the adjacent fault blocks. The branches are inactive compared with the rift valley, which is active because it is coincident with the locus of crustal emplacement. The rift zones of Iceland (Kjartansson, 1960, 1962, 1965, 1968, 1969) and the Afar region (Lowell and Genik, 1972) also exhibit branches subparallel to the active rift. As the sea floor spreads, the branches split off from the rift valley and form the valleys that transect the transverse ridges. Lengths of crust / 1 and /_. are generated per unit time perpendicular to the axis of the rift valley, such that /, = /j. The second configuration of the model (Fig. lib) introduces apparent asymmetric directions and half rates of sea-floor spread- ing. The half rate of spreading r perpendicular to the rift valley decreases 15 percent to the left (r.,) relative to the right side (r,), corresponding to half rates of 1.1 and 1.3 cm/yr averaged over an interval of 10 m.y. in the study area (Table 1); the cor- responding lengths of crust generated per unit time are / ., and / ,, such that I , < I ,. The direction of spreading remains perpendicu- lar to the axis of the rift valley to the right and reorients 30° to the left, corresponding to the trends of the transverse valleys on either side of the rift valley in the study area (Figs. 3, 4). Solving for the half rate of spreading in the reoriented direction to the left (r:1) using values from the study area, r:l = r.,/eos 30° = 1.3 cm/yr. The crustal length generated at spreading half rate r , per unit time is/-,. LIsing values from the study area, r | = r ., and /, = /.,. The asymmetric directions and equal half rates (r,, r:!) of spreading introduced in the second configuration of the model (Fig. I lb) continue for 6.5 m.y. to produce the third configuration (Fig. I 1c). Transverse- valleys and intervening ridges are generated about topographic highs and lows, as previ- ously described (Fig. I la). However, these features are oriented normal (right side) and oblique (left side) to the rift valley, that is, aligned with the true relative directions of spreading. In spite of the reorientation of spreading direction, the long axes of the fault blocks uplifted from the rift valley floor remain parallel to the axis of the rift valley. This parallelism accounts for the parallelism of the walls observed along linear segments of the rift valley. Do the long axes of the fault blocks remain parallel to the axis of the rift valley during oblique spreading, or do the blocks rotate until the long axes become perpendicular to the re- oriented spreading direction? If the fault blocks rotated during spreading, then a delta-shaped gap would be expected to form between the unrotated blocks ad- jacent to the rift valley and the rotated blocks away from the rift valley on the side of oblique spreading. The apparent absence of such gaps in the study area (Fig. 3) makes such rotation unlikely. Rather, the long axes of the fault blocks probably remain nearly parallel to the axis of the rift valley during both symmetric and asymmetric spreading, as shown in the model (Fig. 1 1). Two sets of branches of the rift valley form, corresponding to the two sets ob- served in the study area (Figs. 3, 4). One set is parallel to the rift valley and the long axes of the fault blocks. This set controls the structural development of the valleys that transect the transverse ridges perpendicular to their axes on the normal spreading (right) side of the model and controls de- velopment of a secondary trend that tran- sects the transverse ridges on the oblique spreading (left) side of the model. The sec- ond set of branches is oblique to the rift valley and to the long axes of the fault blocks. This second set appears only on the oblique spreading (left) side, where it con- trols the structural development of the val- leys that transect the transverse ridges per- pendicular to their axes. The actual bathymetric features transecting the trans- verse ridges on the northwest side of the rift valley in the study area appear to be a com- posite of at least these two trends (Fig. 3). The central ridge on the right side of the rift valley is transected by several closely spaced valleys that originated as branches of the rift valley (Fig. lie). These valleys facilitate the inflow of sea water to charge a sub — sea-floor hydrothermal convection system. Hydrothermal deposits shown in black in Figure 1 lc form at and adjacent to the wall of the rift valley. Relict deposits extend along the transverse ridge as a con- sequence of sea-floor spreading (Rona, 1973b). The relict hydrothermal man- ganese recovered 12 km along flow lines of 438 670 RONA AND OTHERS 0 (KM) Figure 10. Diagrammatic sketch of a sub-sea-floor hydrothermal con- vection system like that hypothesized to exist at the TAG Hydrothermal Field (drawn from profile F, Fig. 6; vertical exaggeration is about x2). Arrows indicate directions of hydrothermal flow. Slant lines indicate direc- tions of maximum permeability controlled by structural grain, including fractures, faults, and dikes. Hydrothermal deposits are forming adjacent to the rift valley, and relict deposits are present away from the rift valley as a consequence of sea-floor spreading. Actually, relict deposits may be covered by off-axis volcanism. Symbols: +, zone of recharge; -, zone of discharge; x, zone of igneous intrusion. sea-floor spreading southeast of the active site at the wall of the rift valley (Table 2; station TO 75AK61-1A) indicates the per- sistence of the special structural and ther- mal conditions that maintain sub— sea-floor hydrothermal convection for at least 1 m.y. Off-axis extrusive volcanism may act both to suppress hydrothermal activity and to cover hydrothermal deposits. Consequent- ly, the actual extent of hydrothermal de- posits along flow lines of sea-floor spread- ing may be difficult to determine. Relative to magnetic measurements, the model (Fig. 11) is consistent with the ob- served general parallelism between rema- nent magnetic lineations and the rift valley (Fig. 8). Because remanent magnetization resides in the rocks of the fault blocks, the inferred parallelism between the long axes of these blocks and the rift valley ensures that the gross pattern of linear residual anomalies remains parallel to the axis of the rift valley in spite of the asymmetric tec- tonic fabric. The 3-km southeastward offset of the center of the axial (Brunhes) mag- netic anomaly from the rift valley that re- mains after removal of the magnetic field ef- fect may be alternatively interpreted as fol- lows: (1) if it is possible for the locus of crustal emplacement to shift from the rift valley to one of its branches, then a north- westward shift to the present rift valley could account for the off-center position of the axial anomaly, or (2) asymmetric half rates of sea-floor spreading perpendicular to the rift valley would produce a wider magnetic anomaly on the faster spreading side, resulting in a displacement of the geometric center of the anomaly in the di- rection of faster spreading while the locus of spreading remained at the rift valley. The second interpretation is supported by the : ' ^ ■ m • 1 ; ' \ \ * *■ ■ : W: ■■>. i 1 1 1 i m »• .-—.x' > r ; i :4>: . .♦:•:• Ul ;J-* :: ::; :::: i mir,; j crj i B ;•;• \ lp.r?.; < ml ^ ■:•:«•:•:♦ m ;i!i - — .' -■■■ mm i l:;i m miii \ m t- ij < :•:•*•:•: ¥ Si:?" * * ..*:§ •:■:•• ■> :■ :♦::». *:»W» •: "t'J':: I:*?: » t X :■:♦:•:• » « t | ?:+:•:• >:*¥: '5 * I **:::: m ■; ,„U„ :i ____<» r '* i * n '.».'.» If.'*'*.'* { ♦ * * ♦ :*. ¥,:♦. .*:•:♦. :*::*::±::*:;:* ■ t -y- •*• -T- ■ t urn 1 1 I:::*::*:*:::: J=>:*:::*:::*W f.-.*-'*:*:.-:: » * » I Hi i t « ■ ■■* J » * * t t < i ■{ » ± * ( ■.•:{ 1 I* f I * lit I tt t t t 1 ' > t ' ; ! ii i t i l 1 rf » t i in it ..! i 0 50 ■ i i 100 -I— I KM Figure 1 1 . Diagrammatic model for development of tectonic fabric and hydrothermal activity observed on Mid-Atlantic Ridge crest at lat 26°N (Figs. 3, 4). Axis and two sides of rift valley (heavy solid vertical lines) are parallel. Arrows indicate apparent directions of sea-floor spreading at apparent half rates r„ r2, and r , about rift valley. Transverse ridges (shaded) and intervening transverse valleys (open) are shown. Transverse ridges are transected by valleys perpendicular to their axes (open). Most of valleys only partly transect transverse ridges, and valleys to left of rift valley exhibit composite trends (Fig. 3). Transverse ridges are constructed of fault blocks with long axes (light vertical dashed lines) parallel to rift valley. Lengths of crust generated per unit time are bracketed between sides of the rift valley (heavy solid vertical lines) and solid vertical lines on transverse ridges to either side of the rift valley (b, c). Sites of active and relict hydrothermal mineralization (black) are shown along one of transverse ridges (c). Successive configurations of model (a, b, c) are explained in text. Distinct tectonic fabric of model is obscured in ocean basin by off-axis volcanism. 439 TECTONIC FABRIC AND HYDROTHERMAL ACTIVITY OF MID-ATLANTIC RIDGE CREST 671 asymmetric spreading rates in the study area (Table 1) and poses fewer mechanical problems than the first interpretation. Symmetric and Asymmetric Processes of Oceanic Ridges The asymmetric tectonic fabric of the study area occurs within the overall sym- metric framework of the central North At- lantic, as indicated by the nearly median position of the Mid-Atlantic Ridge; the nearly mirror-image distribution of physiographic provinces about the ridge axis (Heezen and others, 1959); the trajec- tories of major fracture zones such as the Atlantis and Kane (Fig. 1), which follow small circles symmetric about the axis of the Mid-Atlantic Ridge (Morgan, 1968); and the sequences of remanent magnetic anomalies that indicate a grossly similar history of sea-floor spreading in the eastern and western Atlantic (Pitman and Talwani, 1972) since the time of the magnetic quiet zone boundary (Rona and others, 1970). Two alternative hypotheses are considered to reconcile the development of the ob- served asymmetric tectonic fabric within a symmetric framework: 1. Original orientation. The asymmetric orientation of minor fracture zones about the axis of an oceanic ridge, such as the normal and oblique orientation of the minor fracture zones of the study area, is produced by asymmetric processes of de- velopment of oceanic lithosphere. This hypothesis poses problems in reconciling asymmetric with symmetric fractures of the ocean basin, because asymmetric plate mo- tions at minor fracture zones would be in- compatible with symmetric plate motions at major fracture zones. 2. Reorientation. The processes of de- velopment of oceanic lithosphere are essen- tially symmetric and produce both symmet- ric minor and major fracture zones as- sociated with symmetric sea-floor spread- ing. The minor fracture zones are continu- ously reoriented, whereas the major frac- ture zones maintain their original orienta- tions. As a consequence of reorientation, apparent- half rates of spreading deter- mined perpendicular to the axis of an ocean- ic ridge are asymmetric. True- half rates of spreading determined in the directions of the reoriented minor fracture zones normal and oblique to the axis of an oceanic ridge are equal. This hypothesis is supported by the relations between spreading directions and rates determined in the study area and may reconcile the discrepancy between asymmetric and symmetric features of the ocean basin. The continuous reorientation of minor fracture zones according to hypothesis 2 ' rcl.Hivt*. may be caused by the application of an ex- ternal stress field deriving from different sources, as follows: 1. Forces related to magmatic processes. These forces are related to magmatic movements associated with the axial region of an oceanic ridge. A type of regional magmatic movement proposed by Vogt (1971) and applied by Johnson and Vogt (1973) and Vogt and Johnson (1975) to ac- count for V-shaped topography about the axis of an oceanic ridge depends on the principle of a geopotential gradient to drive asthenospheric flow along the axis of the ridge from topographic highs over inferred mantle plumes, for example, the Azores about 1,000 km north of the sudy area. Ac- cording to their hypothesis, the V should point in the direction of flow away from the high as the vector resulting from astheno- spheric flow along and sea-floor spreading about an oceanic ridge. The Vogt-Johnson hypothesis does not account for the V-shaped configuration of the minor frac- ture zones about the Mid-Atlantic Ridge in the study area because the V points toward rather than away from the Azores (Fig. 1). Forces related to magmatic processes un- doubtedly contribute to the stress field, but they are considered secondary rather than primary components. 2. Forces related to tectonic processes. These forces are related to interplate and intraplate motions and may be primary components of the stress field that we infer to be reorienting the direction of minor fracture zones and sea-floor spreading along the Mid-Atlantic Ridge. The role of interplate and intraplate forces as primary components of the stress field is supported by the observation that the orientation of minor fracture zones and of sea-floor spreading differs about lithospheric plate boundaries. The orientation of minor frac- ture zones and of sea-floor spreading is dif- ferent on the two sides of the rift valley of the Mid-Atlantic Ridge and differs between the American and Eurasian plates north of the Azores triple junction and the American and African plates south of that junction (Table 1). The reorientation hypothesis allows the simultaneous development of small-scale asymmetric structures and large-scale symmetric structures in oceanic lithosphere. Minor fracture zones associated with small transform faults (offset ^ 30 km) like those in the study area (Fig. 1) behave in an un- stable manner at the relatively slow average half rates of spreading (=£2 cm/yr) prevalent at the Mid-Atlantic Ridge. The minor frac- ture zones are continuously reoriented under the influence of an external stress field as they are generated by sea-floor spreading about the small transform faults. Major fracture zones like the Atlantis and Kane associated with large transform faults behave in a stable manner at relatively slow average half rates of spreading. The major fracture zones maintain their orienta- tion under the influence of the same exter- nal stress field as they are generated by sea- floor spreading about the large transform faults. Thickness of lithosphere related to distribution of isotherms at a transform fault may be a determinant of the stability of fracture zones (Vogt and others, 1969). Asymmetric small-scale structures may then develop within the large-scale symmetry of the Atlantic Ocean basin as a consequence of the differential stability between minor and major fracture zones (Rona, 1976). CONCLUSIONS The tectonic fabric of oceanic crust that is slowly spreading about an oceanic ridge develops according to a definite geometry, which has been deduced from analysis of the asymmetric tectonic fabric of the Mid- Atlantic Ridge crest at lat 26°N within the overall symmetric framework of the central North Atlantic Ocean basin (Figs. 3, 4, 11), as follows: (1) The double structre of the rift valley consisting of linear segments be- tween transform faults, alternating with ba- sins at transform faults, acts as a template that programs the reproduction of tectonic fabric through control of the formation of topographic highs and lows. (2) The trans- verse ridges are constructed of fault blocks that are uplifted from the floor and accrete at the walls along the linear segments of the rift valley. (3) The transverse valleys are minor fracture zones aligned with the direc- tion of sea-floor spreading about the topo- graphic lows. (4) Branches of the rift valley extend to either side oriented parallel to the rift valley and perpendicular to the axis of the transverse ridges; the branches split off from the rift valley as a consequence of sea-floor spreading and form valleys that transect the. transverse ridges. (5) Minor fracture zones expressed as transverse val- leys between ridges may be asymmetric about the axis of a rift valley, tending to remain normal to one side and to reorient oblique to the other side of the rift valley. (6) Where tectonic fabric is asymmetric about the rift va|ley, apparent half rates of spreading measured perpendicular to the rift valley are also asymmetric, with faster half rates on the normal side and slower half rates on the oblique side. (7) The half rates of spreading measured in the direc- tions of the minor fracture zones normal and oblique to the rift valley tend toward equality over averaging intervals of millions of years. (8) The observed tectonic fabric may be explained by preferential asymmet- ric reorientation of minor fracture zones relative to symmetric major fracture zones, resulting from differential structural stabil- 440 6 2 RONA AND OTHERS irv under the influence of an external stress Held. Structural .uul thermal conditions at di- vergent plate boundaries .ire conducive to hydnithenn.il activity. The concentration of stllv- se.l-tloor h\ drotherm.il systems is favored by special conditions in the tectonic fabric ol an oceanic ridge crest, such as close spacing oi valleys and proximity to intrusive heat sources that promote vigor- ous circulation (Fig. 10). The distribution of hydrotherm.il convection systems along divergent plate boundaries, like the inferred system at the TAG Hydrothermal Field, can only be conjectured from the known dis- tribution of 17 active hydrothermal systems over a distance of 250 km in the neovol- canic /one of Iceland on the Mid-Atlantic Ridge (Bodvansson, 1961) and at least 14 systems over a distance of 900 km in the Red Sea (Degens and Ross, 1969; Backer and Schoell, 19~2). Because all of the oceanic crust that covers two-thirds of the Farth has been generated about divergent plate boundaries, the relation between tec- tonic fabric and hydrothermal activity on the Mid-Atlantic Ridge crest at lat 26°N is relevant to the metallic mineral potential of ocean basins and regions where oceanic crust has been incorporated into islands and continents. ACKNOWLEDGMENTS We lament the early death of our col- league and friend, Andrew J. Nalwalk, who generously contributed his prowess at sea to this work. We thank Louis W. Butler of the Na- tional Oceanic and Atmospheric Adminis- tration (NOAA) for his help in all phases of the work and Bonnie A. McGregor of NOAA for recontouring the bathymetnc map. We are grateful to Bruce C. Heezen for encouraging us to determine the charac- teristics of normal oceanic crust. We thank Gleb B. Udintsev and others onboard the R'V Akademik Kurchatov for obtaining the AK series of dredge samples (Table 2, see footnote 1) on a cooperative TAG cruise in 1975 as part of the U.S. -USSR Agreement on Cooperation in Studies of the World Ocean. We acknowledge the excellent coopera- tion of Captain Floyd J. Tucker, Jr., Cap- tain Lavon L. Posey, Cdr. Richard H. Allbntton, Cdr. Walter S. Simmons, Lt. Paul M. Duernberger, and the other officers and crews of NOAA Ship Discoverer and NOAA Ship Researcher during the 1972 and 1973 TAG cruises. REFERENCES CITED Ade-Hall, J. M„ Palmer, H. C, and Hubbard, T. P., 1971, The magnetic and opaque pet- rological response of basalts to regional hydrothermal alteration: Royal Astron. Soe. Geophys. Jour., v. 24, p. 137-174. Anderson, R. N., 1972, Petrologic significance of low heat flow on the flanks of slow- spreading mid-ocean ridges: Geol. Soe. America Bull., v. 83, p. 2947-2956. Atwater, T., and Mudie, J. D., 1973, Detailed near-bottom geophysical study of the Corda Rise: Jour. Geophys. Research, v. 78, p. 8665-8686. Aumento, F., Loncarevic, B. D., and Ross, D. I., 1971, IV. Regional studies; Hudson geo- traverse: Geology of the Mid-Atlantic Ridge at 45°N: Royal Soe. London Philos. Trans., ser. A, v. 268, p. 623-650. Backer, H, and Schoell, M., 1972, New deeps with brines and metalliferous sediments in the Red Sea: Nature Phys. Sci., v. 240, p. 153-158. Bellaiche, G., Cheminee, J. L., Francheteau, J., Hekinian, R., Le Pichon, X., Needham, H. D., and Ballard, R. D., 1974, Inner floor of the rift valley: First submersible study: Na- ture, v. 250, p. 558-560. Betzer, P. R., Bolger, G. W., McGregor, B. A., and Rona, P. A., 1974, The Mid-Atlantic Ridge and its effect on the composition of particulate matter in the deep ocean: EOS (Am. Geophvs. Union Trans.), v. 55, p. 293. Bhattacharyya, P. J., and Ross, D. 1., 1972, Mid-Atlantic Ridge near 45°N, computer interpolation and contouring of bathymetry and magnetics: Canada Dept. Environment, Marine Sci. Directorate, Marine Sci. Paper 11, 9 p. Bischoff, J. L., and Dickson, F. W., 1975, Seawater-basalt interaction at 200°C and 500 bars: Implications for origin of sea-floor heavy mineral deposits and regula- tion of seawater chemistry: Earth and Planetary Sci. Letters, v. 25, p. 385-397. Bodvarsson, G., 1961, Physical characteristics of natural heat resources in Iceland: Jokull, v. 11, p. 29-38. Bonatti, E., 1975, Metallogenesis at oceanic spreading centers, in Annual Reviews of Earth and Planetary Sciences, Vol. 3: Palo Alto, Calif., Annua] Reviews, Inc. p. 401-431. Bostrom, K., and Peterson, M.N. A., 1966, Pre- cipitates from hydrothermal exhalations on the East Pacific Rise: Econ. Geology, v. 61, p. 1258-1265. Bostrom, K., Farquarson, B., and Eyl, W., 1972, Submarine hot springs as a source of active ridge sediments: Chem. Geology, v. 10, p. 189-203. Cann, J. R., 1970, New model for the structure of the ocean crust: Nature, v. 226, p. 928-930. Christensen, N. I., 1972, The abundance of ser- pentimtes in the oceanic crust: Jour. Geol- ogy, v. 80, p. 709-719. Collette, B. J., 1974, Thermal contraction |oints in a spreading seafloor as origin of fracture zones: Nature, v. 251, p. 299-300. Constantinou, G., and Govett, G.J.S., 1973, Geology, geochemistry, and genesis of Cyprus sulfide deposits: Econ. Geologv, v. 68, p. 843-858. Corliss, J. B., 1971, The origin of metal-bearing submarine hydrothermal solutions: Jour. Geophys. Research, v. 76, p. 8128-8138. Deffeyes, K. S., 1970, The axial valley: A steady-state feature of the terrain, in |ohnson, FL, and Smith, B. L., ed>, The megatectonics of continents and oceans: Camden, N.J., Rutgers Univ. Press, p. 194-222. Degens, E. T., and Ross, D. A., eds., 1969, Hot brines and recent heavy metal deposits in the Red Sea: New York, Springer, 600 p. Detnck, R. S„ Mudie, J. D., Luyendyk, B. P., and Macdonald, K. C, 1973, Near-bottom ob- servations of an active transform fault (Mid-Atlantic Ridge at 37°N): Nature Phvs. Sci., v. 246, p. 59-61. Dmitnev, L. V., Barsukov, V. L., and Udintsev, G. B., 1970, Rift zone of the oceans and the problem of ore formation: Geokhimiva, v. 4, p. 937. Donaldson, I. D., 1962, Temperature gradients in the upper layers of the Earth's crust due to convective water flows: Jour. Geophvs. Research, v. 67, p. 3449-3459. Dymond, J., Corliss, J. B., Heath, J. R., Field, C. W., Dash, E. J., and Veeh, H. H, 1973, Origin of metalliferous sediments from the Pacific Ocean: Geol. Soe. America Bull., v. 84, p. 3355-3372. Elder, J. W., 1965, Physical processes in geo- thermal areas, in Terrestrial heat flow: Am. Geophvs. Union Geophvs. Mon. 8, p. 211-239. 1967, Steady free convection in a porous medium heated from below: Jour. Fluid Mechanics, v. 27, p. 29-48. Greenewalt, D., and Taylor, P. T., 1974, Deep- tow magnetic measurements across the axial valley of the Mid-Atlantic Ridge: Jour. Geophvs. Research, v. 79, p. 4401-4405. H.i|.ish, A., 1974, An experimental investigation of high temperature seawater-basalt in- teractions: Geol. Soe. America Abs. with Programs, v. 6, p. 771. Harrison, C.G.A., 1968, Formation of magnetic anomaly pattern by dyke injection: Jour. Geophys. Research,' v. 73, p. 2137-2142. Han, R. A., 1973, A model for chemical ex- change in the basalt-seawater system of oceanic layer II: Canadian Jour. Earth Sci., v. 10, p. 799-816. Heezen, B. C, Tharp, M., and Ewing, M., 1959, The floors of the oceans. I. The North At- lantic: Geol. Soe. America Spec. Paper 65, 122 p. Heirtzler, J. R., Dickson, G. O., Herron, E. M., Pitman, W. C. Ill, and Le Pichon. X., 1968, Marine magnetic anomalies, geomagnetic field reversals, and motions of the sea floor and continents: Jour. Geophvs. Research, v. 73, p. 2119-2136. Helgeson, H. C, 1964, Complexing and hy- drothermal ore deposition: New York, Per- gamon Press. Hoffman, )., 1957, Hyperbolic curves applied to echo sounding: Internat. H\drog. Rev., v. ?4, p. 45-55. Holland, H. D., 19f>-, Gangue minerals in hy- drothermal systems, in Barnes, H. 1... ed., Geochemistry of hydrothermal ore de- posits: New York, Holt, Rinehart &: Win- ston, p. 382-436. 441 IFCTONK FABRIC AND HYDROTHFRMAI ACTIVITY OF MID- ATLANTIC RIDGE CREST 673 Hollister, C . D.. huiit, |, I., Habih, IV, H.nh.i- wav, | (. .. Lancelot, V., l.uterhacker. H., Paulus, 1 |.. Poag, C . W.. Wilcoxon. |. C .. and \X oi mcII. P.. I9~2, Initial reports of the Deep Sea Drilling Project, Vol. II: Washing- ton, D.C., U.S. Clovt. Printing Office, 10- p. Hutchinson, R. W.. |9 5, \'olcanogenic sulfide deposits and then metallogenic significance: Econ. Geology, \. 68, p. 1223-1246. IAGA Commission 2. Working Croup 4. |9<,9, International geomagnetic reference field 1965.0: |our. Geophys. Research, \. ~4, p. 440~-440X. Johnson. Ci. 1.. and Vogt, P. R. 19~3, Mid- Atlantic Ridge from 4~ to si North: C.eol. Soc. America Bull., v. S4. p. 5443-3462. Kjartansson, C. I960, 1^2. I^hs, 19hK, 1969, Geological map of Iceland, sheets 1,2, \ 5, 6: Reykjavik, Mus. Nat. History. Krauskopf, K. B., 1956, Factors controlling the concentrations of thirteen rare metals in seawater: Geochim. et C osmochim. Acta, \. 9, p. 1-12B. Lachenbruch, A. H., 19~3, A simple mechanical model for oceanic spreading centers: Jour. Geophys. Research, \. -S. p. 3395-3417. Langseth, M. G-, |r.. Malone, I. E., and Book- man, C. A., I9~2, Sea floor geothermal measurements tram Win.i Cruise 25: Palisades, N.Y., Lamont-Dohertv Geol. Observatory Tech. Rept. 4, CL'-4-_2; Teeh. Rept. 2, CU-2-72, 159 P. Latrimore, R. K.. Rona, P. A., DeWald, O. E., 1974, Magnetic anomaly sequence in the central North Atlantic: |our. Geophys. Re- search, \. _9, p. 120"- 1209. Lister, C.R.B., 19~2, On the thermal balance of a mid-ocean ridge: Royal Astron. Soc. Geophys. Jour., v. 26, p. 515-535. 1974a, Water percolation in the oceanic crust: EOS (Am. Geophvs. Union Trans.), v. 55, p. -40-"42. 19-4h, On the penetration of water into hot rock: Roval Astron. Soc. Geophvs. |our., v. 39, p. 465-509. Loncarevic, B. D., and Parker, R. L., 1971, The Mid-Atlantic Ridge near 45°N. XVII. Magnetic anomalies and sea floor spread- ing: Canadian Jour. Earth Sci., v. X, p. 883-898. Lowell, J. D., and Cenik, J. D„ 1972, Sea-tloor spreading and structural evolution of southern Red Sea: Am. Assoc. Petroleum Geologists Bull., v. 56, p. 247-259. Lowell, R. P., 1975, Circulation in fractures, hot springs, and convective heat transport on mid-ocean ridge crests: Royal Astron. Soc. Geophys. |our., v. 40, p. 351-365. Luyendyk, B. P., and Melson, W. G., 196", Magnetic properties and petrology of rocks near the crest of the Mid-Atlantic Ridge: Nature, v. 215, p. 147-149. Macdonald, K., Luyendyk, B. P., Mudie, ). D., and Spiess, F. N., 1975a, Near-bottom geophysical study of the Mid-Atlantic Ridge median valley near lat 37°N: Pre- liminary observations: Geology, v. ?, p. 211-215. 1975h, Near-bottom magnetic measure- ments, asymmetric spreading, oblique spreading, and the zone of intrusion in the FAMOUS area of the Mid-Atlantic Ridge I i~°N): EOS (Am. Geophys. Union Trans.), v. 56, p. s-V Matthews, D. H., and Bath, J., 196"7, Formation ot magnetic anomaly pattern of Mid- Atlantic Ridge: Roval Astron. Soc. Geophys. |our., v. 13, p. 349-357. Matthews, D. |., 19 19, Tables of the velocity of sound in pure water and sea water for use in echo-sounding and sound ranging (2nd ed.): London. Admiralty, Hvdrographic Dept. Pub. 2S2, S2 p. McGregor, B. A., and Rona, P. A., 1975, Crest of Mid-Atlantic Ridge at 26°N: Jour. Geophvs. Research, v. 80, p. 3307-3314. McGregor, B. A., Harrison, C.G.A.. Lavelle, J. W., and Rona, P. A., 1976, Magnetic anomaly pattern on Mid-Atlantic Ridge crest at 26°N: Jour. Geophys. Research, v. SI (in press). McKen/ie, 1). P., Davis, D., and Molnar, P., 1970, Plate tectonics of the Red Sea and Fast Africa: Nature, v. 226, p. 243-248. Menard, H. W., 1967, Sea floor spreading, to- pography, and the second laver: Science, v. 157, p. 92 i- 924. Meyer, C, and Hemley, J. J., 1967, Wall rock alteration, in Barnes, H. L., ed., Geochemis- try of hydrothermal ore deposits: New York, Holt, Rmehart & Winston, p. 166-235. Miyashiro, A., Shido, F., and Fwing, M-, 1971, Metamorphism in the Mid-Atlantic Ridge near 24°N and 3()°N: Royal Soc. London Philos. Trans., ser. A, v. 268, p. 589-603. Moore, J. G., Fleming, H. S., and Phillips, J. D., 1974, Preliminary model for extrusion and rifting at the axis of the Mid-Atlantic Ridge, 36°48'N: Geology, v. 2, p. 437-440. Morgan, W. J., 1968, Rises, trenches, great faults and crustal blocks: Jour. Geophvs. Re- search, v. 73, p. 1959-1982. Mottl, M. J., Corr, R. F., and Holland, H. D., 1974, Chemical exchange between sea water and mid-ocean ridge basalt during hydrothermal alteration: An experimental study: Geol. Soc. America Abs. with Pro- grams, v. 6, p. 879-880. Muehlenbachs, K„ and Clayton, R. N., 1972, Oxygen isotope geochemistry of submarine greenstones: Canadian Jour. Earth Sci., y. 9, p. 471-478. Needham, H. D., and Francheteau, J., 1974, Some characteristics of the rift valley in the Atlantic Ocean near 36°48' North: Earth and Planetary Sci. Letters, v. 22, p. 29-43. Osmaston, M. F., 1971, Genesis of ocean ridge median valleys and continental rift valleys: Tectonophysics, v. 11, p. 387-405. Palmason, C, 1967, On heat flow in Iceland in relation to the Mid-Atlantic Ridge, in Bjornsson, S., ed., Iceland and mid-oceanic ridges: Soc. Sci. Islandica, v. 38, p. 111-127. Parker, R. L., and Oldenburg, D. W., 1973, Thermal model of ocean ridges: Nature Phvs. Sci., v. 242, p. 137-139. Perry, R. K., and Feden, R. H., 1974, Bathymetry of the Mid-Atlantic Ridge from 10°N to 40°N latitudes: Geol. Soc. America Abs. with Programs, v. 6, p. 909. Piper, D. Z., 1973, Origin of metalliferous sedi- ments from the East Pacific Rise: Earth and Planetary Sci. Letters, v. 19, p. 75-82. Pitman, W. C. Ill, and Talwani, M., 1972, Sea-floor spreading in the North Atlantic: Geol. Soc. America Bull., v. 83, p. 619-649. Reid, 1., and Macdonald'K., 1973, Microearth- quake study of the Mid-Atlantic Ridge near 37°N, using sonobuoys: Nature, v. 246, p. 88-90. Robertson, A.H.F., and Hudson, J. D., 1973, Chemical precipitates on a Tethyan ocean ridge: Earth and Planetary Sci. Letters, v. 18, p. 93-106. Rona, P. A., 1973a, Marine geology, in McGraw-Hill yearbook of science and technology: New York, McGraw-Hill Book Co., p. 252-256. 1973b, Plate tectonics and mineral re- sources: Sci. American, v. 229, no. 1, p. 86-95. 1976, Asymmetric fracture zones and sea-floor spreading: Earth and Planetary Sci. Letters (in press). Rona, P. A., Brakl, J., and Heirtzler, J. R., 1970, Magnetic anomalies in the northeast Atlan- tic between the Canary and Cape Verde Is- lands: Jour. Geophys. Research, v. 75, p. 7412-7420. Rona, P. A., Harbison, R. N., Bassinger, B. G., Butler, L. W., and Scott, R. B., 1973, Asymmetrical bathymetry of the Mid- Atlantic Ridge at 26°N latitude: EOS (Am. Geophys. Union Trans.), v. 54, p. 243. Rona, P. A., McGregor, B. A., Betzer, P. R., Bolger, G. W., and Krause, D. C, 1975, Anomalous water temperatures over Mid- Atlantic Ridge crest at 26° North latitude: Deep-Sea Research, v. 22, p. 611-618. Sayles, F. L., and Bischoff, J. L., 1973, Fer- romanganoan sediments in the equatorial East Pacific: Earth and Planetary Sci. Let- ters, v. 19, p. 330-336. Sclater, J. G., Von Herzen, R. P., Williams, D. L, Anderson, R. N., and Klitgord, K., 1974, The Galapagos spreading center: Heat flow low on the north flank: Royal Astron. Soc. Geophys. Jour., v. 38, p. 609-626. Scott, M. R., Scott, R. B., Rona, P. A., Butler, L. W., and Nalwalk, A. J., 1974, Rapidly accumulating manganese deposit from the median valley of the Mid-Atlantic Ridge: Geophys. Research Letters, v. 1, p. 355-358. Scott, R. B., Rona, P. A., McGregor, B. A., and Scott, M. R„ 1974, The TAG Hydrother- mal Field: Nature, v. 251, p. 301-302. Scott, R. B., Malpas, J., Udintsev, C, and Rona, P. A., 1975, Submarine hydrothermal activ- ity and sea-floor spreading at 26°N, MAR: Geol. Soc. America Abs. with Programs, v. 7, p. 1263. Silhtoe, R. H., 1973, Environments of formation of volcanogenic massive sulfide deposits: Econ. Geology, v. 68, p. 1321-1336. Sleep, N. H., 1969, Sensitivity of heat flow and gravity to the mechanism of sea-floor spreading: Jour. Geophys. Research, v. 74, p. 542-549. Sleep, N. H., and Biehler, S., 1970, Topography and tectonics at the intersections of fracture zones with central rifts: Jour. Geophys. Re- search, v. 75, p. 2748-2752. Spindel, R. C, Davis, S. B., Macdonald, K. C, Porter, R. P., and Phillips, J. D., 1974, Mi- 442 h~4 RON A WD OTHKRS croearthquake survev of median vallev of ocean rises: |our. Ceophvs. Research, v. "3, Drilling Project, Vol. 8: Washington, D.C., the Mid-Atlantic Ridge .it 56°30'N: Na- p. 12~V-129X. U.S. Govt. Printing Office, p. 829-833. turc, v. 248, p. s~"-5~9. van Andel, T. H., and Moore, I. C. I9~(), Watkins, N. [)., and Paster, 1. P., 1971, The Spooncr, F..T.C .. and F\ fe, XV. S., 1973, Sub-sea Magnetic anomalies and sea floor spreading magnetic properties of igneous rocks from floor metamorphisrn, heat and mass trans- rates in the northern South Atlantic: Na- the ocean floor: Royal Sue. London Philos. ter: Contr. Mineralog\ and Petrolog\, ture, v, 226, p. 528-330. Trans., ser. A, v. 268, p. 507-550. v. 42. p. 28~-304. Vogt, P. R., 19~1, Asthenosphere motion re- Williams, D. L, Von Herzen, R. P., Sdater, J. G., Spooner. E.T.C .. Beckinsale, R. D., Fyfe, W. S., corded hy the ocean floor south of Iceland: and Anderson, R. N.. 1974, The Galapagos and Sniewing, |. D., 19~4, O'" enriched Earth and Planctar\ Sci. letters. \. |3, spreading center: Lithospheric cooling and ophiolitic metahasic rocks from E. Liguria p. 155 — 160. hydrothcrmal circulation: Renal Astron. ItaK . Pindos Greece), and Troodos Vogt, P. R.. and Johnson. G. 1 . I9~5, 1 tans- Soc. Geophys. Jour., v. IN, p. 587-608. i. \prus : C onrr. Mineralog) and Petrology, form faults and longitudinal flow beneath Wilson. |. 1., 1V6.5, A new class of faults and v. 4". p. 41— "4, the midoceanic ridge: |our. Geophys. Re- their bearing on continental drift: Nature, Talwani. M.. VX'indish, C. ( .. and Langseth, search, v. 80, p. 1399-1428. v. 20" p. 543-347. M.C.. lr. |9~1, Reykjanes Ridge crest: A Vogt. P. R. Avery, O. F... Schneider. E. I).. An- Wooding, R. A.. 1 963, Convection in a saturated detailed geophysical study : Jour. Geophvs. derson, C. N.. and Braces. 1). R.. 1969, porous medium at large Ra\ leigh numbers Research, v. ~6, p. 473-51". Discontinuities in sea-floor spreading: Tec- and Peclet number: Jour. Fluid Mechanics. Turcotte, D. 1., 19~4, Are transform faults tonophysics, v. 8, p. 285-3 1". \. 15, p. 527-544, thermal contraction cracks?: Jour. von der Borch. C . C .. and Rex, R. NX.. |9~1, Zelenov, K. K., 1964. Iron and manganese in Geophys. Research. \. -9, p. 25~5-25--. Amorphous iron oxide precipitates in sedi- exhalations from the submarine volcano L'padhyay, H. D., and Strong, D. F., 1973, ments cored during Leg 5, Deep Sea Drilling Bano XX'uhu (Indonesia): Akad. Nauk SSSR Geological setting of the Betts Cove copper Pro|ect. in Initial reports of the Deep Sea Dokladv. v. 155, p. 1515-1520. deposits. Newfoundland: An example of Drilling Project, Vol. v Washington. D.C., ophiolite sulfide mineralization: Econ. U.S. Govt. Printing Office, p. s41-s44 Geology, v. 68, p. 161-167. von der Borch. C. C .. Nesteroft, \X . D., and Manuscript Received by the SOCIETY August van Andel, T, H.. and Bowm. C. O., 1968, Galehouse, |. S„ 1 9" 1. Iron-rich sediments 21, 1V~4 Mid-Atlantic Ridge between 22c and 25c cored during Leg 8 of the Deep Sea Drilling Revised Manuscript Received July 24, 1975 North latitude and the tectonics of mid- Project, in Initial reports of the Deep Sea Manuscript Accepted August 20, 1975 443 46 Reprinted from: Geology, Vol. 4, No. 4, 233-236. Duration of hydrothermal activity at an oceanic spreading center, Mid-Atlantic Ridge (lat 26°N) Robert B. Scott Department of Geology, Texas A&M University, College Station, Texas 77843 John Malpas Department of Geology, Memorial University, St. Johns, Newfoundland, Canada A1C5S7 Peter A. Rona National Oceanic and Atmospheric Administration-Atlantic Oceanographic and Meteorological Laboratories, 15 Rickenbacker Causeway, Miami, Florida 33149 Gleb Udintsev Institute of Oceanology, USSR Academy of Sciences, Moscow, USSR ABSTRACT Hydrothermal manganese oxide coats talus on the Mid- Atlantic Ridge at lat 26° N until spreading moves the rock past the thermally and structurally active rift- valley wall. Hydrothermal activity is re- placed by hydrogenous ferromanganese oxide precipitation on ocean crust older than 0.7 m.y. on the ridge-crest highlands. INTRODUCTION Abundant hydrothermal manganese oxide crusts were found coating basalt talus on the rift-valley wall at the Mid- Atlantic Ridge at lat 26°N during the Trans- Atlantic Geotraverse project of the National Oceanic and Atmospheric Ad- ministration in 1972 and 1973 (Scott, R. B., and others, 1974; Scott, M. R., and others, 1974). These crusts are almost pure man- ganese oxide with only a trace of Fe, GEOLOGY Co, Cu, and Ni and compositionally fall into the Mn-rich end member of Bonatti's hydrothermal classification (Bonatti, 1975). Very rapid growth of the manganese deposits is required because they are as great as 50 mm thick only 5 km from the rift axis. U-Th dating of the outermost layers of these manganese crusts shows them to be accumulating at 200 mm/m.y., nearly two orders of magnitude faster than typical hydrogenous ferromanganese crusts or nodules (Scott, M. R., and others, 1974). Bottom photographs of the hydrothermally active portion of the rift wall (McGregor and Rona, 1975) show the presence of similar coatings over talus. Betzer and others (1974) found abnormally high con- centrations of weak-acid-soluble Fe- and Mn-bearing particulate matter suspended over the Mid-Atlantic Ridge at lat 26° N. A region of low magnetic intensity within the Brunhes normal in the hydrothermal area may be related to the destruction of magnetic domains during hydrothermal alteration of basalt (McGregor and Rona, 1975). From these data, R. B. Scott and others (1974) concluded that cold, dense sea water flows down fracture systems (Lister, 1974), reacts with hot rocks or magma under the ridge crest, produces less dense hydrothermal fluids enriched in Ca, Si, Fe, Mn, and H2S and depleted in Mg (Hajash, 1975; Mottl and others, 1974; Bischoff and Dickson, 1975), and then is emitted as submarine springs, where oxy- genated sea water causes precipitation of manganese oxides on talus overlying frac- tured fault scarps. However, the distribution of dredge sites did not define the limits of this hydro- thermal activity in time or space. Defini- tion of the limits of activity at lat 26°N became one of the objectives of participa- tion of the NOAA Trans-Atlantic Geo- traverse project in the U.S.-U.S.S.R. Agree- ment for Cooperative Studies of the World Ocean. The hydrothermal region at lat 26° N was dredged during the spring of 1975 aboard the R/V Akademik Kurchatov by Soviet and American scientists. 233 444 DREDGED ROCKS Manganese oxide crusts and associated veins in basalt talus were recovered from site 75-1 A, 18 km from the rift axis (Fig. 1). Unlike the hydrothermal crusts found at sites 72-13, 73-2A, and 73-3A, the crusts at site 75-1 A have two distinct layers. The basal layers on the altered basalt have the same physical appearance as the hydro- thermal crusts dredged in 1972 and 1973. The basal layers have a smooth, slightly botryoidal surface over an undulating laminated interior, a uniform brownish- black to submetallic gray color, and a thickness as great as 10 mm. The slightly undulating botryoids are about 1 mm in diameter. In contrast, the upper layers are most similar to hydrogenous crusts; these upper layers have highly irregular micro- botryoidal surfaces. Small columnlike bot- ryoids are less than 0.1 mm in diameter and vary from a shiny grayish-black to an earthy, dark yellowish orange color. Organism tests are trapped between botryoid columns. The upper crusts are less than 2 mm thick. The two crust types were carefully separated for chemical, mineralogical, and scanning electron microscope (SEM) studies. The basalts have been altered from nearly fresh basalt to extremely friable grayish yellow green material. Numerous veins of manganese oxides similar to the lower crusts fill fractures that cut the talus of basaltic breccia. One vein of free- growing zeolite crystals (75-1A28) as much as 5 mm in diameter was found coating altered basaltic glass; a thin (<0.1 mm thick) irregularly botryoidal manganiferous crust coated these zeolites. RESULTS OF CHEMICAL, MINERAL- OGICAL, AND SEM INVESTIGATIONS The lower crusts (sample 75-1A24) only have x-ray diffraction patterns of birnessite (7.1, 3.5, and 2.46 A d-spacings). The upper crust (75-1A24) is apparently amor- phous to x-rays. The zeolite (75-1A28) is obviously an analcite from powder camera patterns, and chemically it is a potassic Na analcite (10 percent Na20, 1 percent K2Ot and trace of CaO). Atomic absorp- tion spectrophotometry shows the lower crusts to contain 40 percent Mn, less than 0.1 percent Fe, and 0.15 percent Cu+Co+Ni; in contrast, the upper crusts contain more than 16 percent Fe and 0.4 percent Cu+Ni+Co (Fig. 2). The low Cu+Ni+Co contents and low Fe contents of the basal crust are similar to hydro- thermal crusts found in 1972 and 1973. Figure 1. Hydrothermal field at lat 26°N. The 4-km-deep rift valley is on left; year and dredge number are used to identify dredge sites; contour intervals are 0.1 km. Hydrothermal sites 72-13, 73- 2A, 73-3A, and 75-1 A lie along an irregular ridge that trends southeast perpendicular to rift val- ley. This ridge seems to be cut into blocks by northeast-trending depressions parallel to rift that may represent normal faults. Bathymetry from McGregor and Rona (1975). (Cu + Co+Ni)xlO k40 15-2 .2B 19-3 • »A •I0G »4?£e/? --'A?« Fe 50 25 Mn Figure 2. Composition of hydrothermal and hydrogenous Mn deposits plotted on Mn portion of the Mn-Fe-(Cu+Co+Ni)10 ternary diagram. Average composition of three analyses of upper and lower crust of sample 75-1A24 are attached by dashed line. At Mn apex, 1972 and 1973 analyses of hydrothermal crust are shown. Hydrogenous materials are represented by points on left: P = aver- age Pacific and A - average Atlantic nodule; 10G and 2B are Mn crusts from the Atlantis Fracture Zone; 19-3 and 15-2 are Mn crusts from pillow lavas close to hydrothermal field at lat 26 N (Scott, M. R., and others, 1974). By atomic absorption spectrophotometry; precision expressed as percent of value determined: Fe±2 percent, Mn±l percent, Cu±l percent, Ni±5 percent, Co±5 percent. 234 APRIL 1976 445 The coating on the zeolites (Fig. 3d) with a 0.3 percent Co+Cu+Ni content is tran- sitional between hydrothermal and hydro- genous crust compositions; the high Fe content in this coating (18 percent) may be indicative of a hydrogenous origin. Even though no Fe-rich hydrothermal man- ganese deposits have been identified at lat 26°N, Bonatti (1975) showed their existence elsewhere. SEM photographs show the lower crust to have the texture of well-crystallized birnessite (Fig. 3a); this boxwork of plates is very similar to the texture found in birnessite from widely differing environments (Swanson, 1975; Brown and others, 1971; Fewkes, 1973). In contrast, the upper crust at the same scale (Fig. 3b) shows a smooth featureless surface with no textural indication of crystallinity. The smaller scale view in Figure 3c shows the overall micro- botryoidal form of the lower crust that is similar to columnar zones described by Sorem and Foster (1972) in hydrogenous ferromanganese. DISCUSSION Clearly, from the chemical and physical description and SEM observations, the manganese oxide lower crust 75-1A24 appears to be most similar to other hydro- thermal manganese crusts; the ferro- manganese oxide upper crust 75-1 A24 has strong affinities to hydrogenous ferro- manganese crusts and nodules. A compari- son of the compositions shown in Figure 2 with those of Bonatti (1975, Fig. 4) support these conclusions. The dramatic differences in iron and Cu-t-Co+Ni contents between the upper and lower crust samples taken only a few millimetres from one another suggest drastically different mechanisms of formation or different sources of fluids. Several authors (Bonatti and others, 1972; M. R. Scott and others, 1974) have noted an inverse relation between the rate of manganese crust accumulation and the content of trace metals. If data given by M. R. Scott and others (1974, Table 2, Figs. 2, 3) are representative of this relationship, then the approximate growth rate in millimetres/106yr = e14.8S- 1.54 In Cu+Co+Ni ppm The Cu+Co+Ni for the lower crust equals 1,520 ppm and for the upper crust equals 3,910 ppm. Growth rates for the lower and upper crusts are calculated to be 35 and 8 mm/106yr, respectively. The maxi- mum thickness of the upper crust is about 3 mm, and the underlying hydrothermal layer is as great as 10 mm thick; this im- plies that the hydrothermal activity may have continued to affect the talus for 0.3 m.y. before hydrogenous activity began. The magnetic anomaly age of the ocean crust under site 75-1 A is approximately 1.4 m.y. (McGregor and Rona. 1975). With this spacial relationship (Fig. 1), the half- spreading rate of about 1.3 cm/yr, and estimated growth rates of crusts, • se- quence of events can be postulated. When site 75-1 A rocks spread from the rift axis to the rift wall position of modern hydro- thermal activity, they were 0.4 m.y. old. Hydrothermal activity continued until this site was 0.7 m.y. old and had moved be- yond the influence of hydrothermal activity. Hydrogenous growths of manganese then ensued for 0.4 m.y. The total age of the site to account for this sequence of events would be 1.1 m.y., close to the magnetic anomaly age of 1.4 m.y. Thus, it seems that the most active hydrothermal region is on the rift wall near both high geo- thermal gradients in the rift and active faults scarps on the rift wall. A ^trip of hydrothermally altered oceanic crust results. A ropy-textured, seemingly fresh, thin basalt flow was dredged at site 73-6A and at 75-1 B; chemical analysis of basalt 73-6A2 shows an abnormally high K20 content of 0.3 percent, whereas fresh typi- cal rift-valley tholeiites in this region have only 0.05 to 0.10 percent K20 (Scott and others, 1973). Samples 6A and IB may be Figure 3. SEM photographs of manganese crust, (a) Lower crust 75-1A24 showing crystal plates of birnessite. (b) and (c) Botryoidsof upper crust 75-1A24 showing the absence of crystallinity. (d) Smooth interior of a ferromanganese crust that caps analcite crystals of a vein 75-1 A28. GEOLOGY 235 446 slightly alkalic younger off-axis basalts (Strong, 1974) that may have covered older hydrothermal areas. The Lag.f./Sme.f. ratio (0.6) and the rare-earth element abundances fit intraplate or ridge-crest criteria (Schilling and Bonatti, 1975). However, the possibility of simple low- temperature addition of K during weather- ing cannot be discounted (Scott and Hajash, 1975). Thus, no definite limit to the size of the hydrothermal field can be established. Three other localities of hydrothermal manganese have recently been located; one is close to the Galapagos spreading axis, and it has nearly identical physical, chemical, isotopic, and growth rate characteristics to the lat 26°N deposits (Moore and Vogt, 1975). Another occurs at lat 23° N on the Mid- Atlantic Ridge (Thompson and others, 1975). French sci- entists have also found a hydrothermal manganese deposit in a transform fault in the FAMOUS area (ARCYANA, 1975). Recognition of such occurrences imme- diately following publication of findings at lat 26°N suggests that these deposits may be common and were overlooked in the past. The most common hydrothermal deposit in oceanic spreading centers and related structures besides manganese crusts are sulfides precipitated as veins within the crust (Dmitriev and others, 1970; Bonatti, 1975). It is probable that these two phe- nomena are both part of the same com- plex hydrothermal process operating in the ocean crust. The association is strength- ened by Hajash's (1975) observation of experimental chalcopyrite and pyrrhotite precipitation in Fe- and Mn-rich sea water resulting from reaction with basalt at 400° to 500°C. Some mechanisms may even exist to precipitate sulfides at the water-rock interface from chloride-rich brines (Sillitoe, 1972; Constantinou and Govett, 1972; Searle, 1972; Upadhyay and Strong, 1973; Sato, 1973). Thus far, no massive sulfides have been observed in rocks dredged from open-ocean centers either within or on the rock-water interface. Obviously, both hydrothermal manganese and sulfides will have to be studied to understand the total chemical effects of cycling sea water through cooling oceanic crust. REFERENCES CITED ARCYANA, 1975, Transform fault and rift valley from bathyscaph and diving saucer Science, v. 190, p. 108-1 16. Better, P. R., Bolger. G. W., McGregor, B. A., and Rona. P. A., 1974, The Mid-Atlantic Ridge and its effect on the composition of particulate matter in the deep ocean: EOS (Am. Geophys. Union Trans.), v. 55, p. 193. Bischoff, J. L., and Dickson, F. W., 1975, Seawater-basalt interaction at 200 C and 500 bars: Implications for origin of sea- floor heavy-metal deposits and regulation of seawater chemistry: Earth and Planetary Sci. Letters, v. 25, p. 385-397. Bonatti, E., 1975, Metallogenesis at oceanic spreading centers, in Annual review of Earth and planetary sciences: Palo Alto, Calif., Annual Reviews, Inc., p. 401-431. Bonatti, E., Kramer. T., and Rydell, H. S., 1972, Classification and genesis of sub- marine iron-manganese deposits, in Ferro- manganese deposits on the ocean floor: Palisades, N.Y., Lamont-Doherty Geol. Obs., Columbia Univ., p. 146-166. Brown, F, H., Pabst, A., and Sawyer, D. L., 1971, Birnessite on colemanite at Boron, California: Am. Mineralogist, v. 56, p. 1057-1064. Constantinou, G., and Govett, G.T.S., 1 972, Genesis of sulphide deposits, ochre and umber of Cyprus: Inst. Mining and Metal- lurgy Trans., v. 81, p. B33-B46. Dmitriev, L. V., Barsukov, V. L., and Udintsev, G. B., 1970, Rift zones of the ocean and the problem of ore formation: Geokhimiya, v. 4, p. 937-944. Fewkes, R. H., 1973, External and internal features of marine manganese nodules as seen with the SEM and their implications in nodule origin, in Morganstein, M., ed., The origin and distribution of manganese nodules in the Pacific and prospects for exploration: Honolulu, Hawaii Inst. Geophysics, Univ. Hawaii, p. 21-29. Hajash, A., 1975, Hydrothermal processes along mid-ocean ridges: An experimental investi- gation: Contr. Mineralogy and Petrology (in press). Lister, C.R.B., 1974, Water percolation in the ocean crust: EOS (Am. Geophys. Union Trans.), v. 55, p. 740-742. McGregor, B. A., and Rona, P. A., 1975, Crest of Mid-Atlantic Ridge at 26 N: Jour. Geo- phys. Research, v. 80, p. 3307-3314. Moore, W. S., and Vogt, P. R., 197S, Hydro- thermal manganese crusts from two sites near the Galapagos spreading axis: Earth and Planetary Si. Letters (in press). Mottl, M. J., Corr, R. E., and Holland, H. D., 1974, Chemical exchange between sea water and mid-ocean ridge basalt during hydro- thermal alteration: An experimental study: Geol. Soc. America Abs. with Programs, v. 6, p. 879-880. Sato, T., 1973, A chloride complex model for Kuroko mineralization: Geochem. Jour., v. 7, p. 245-270. Schilling, J. G., andoBonatti, E., 1975, East Pacific Ridge (2 S-10 S) versus Nazca intraplate volcanism: Rare-earth evidence: Earth and Planetary Sci. Letters, v. 25, p. 93-102. Scott, M. R., Scott, R. B., Rona, P. A., Butler, L W., and Nalwalk, A. J., 1974, Rapidly accumulating manganese deposit from the median valley of the Mid-Atlantic Ridge: Geophys. Research Letters, v. 1, p. 355-358. Scott, R. B., and Hajash, A., 1975, Initial sub- marine alteration of basaltic pillow lavas: A microprobe study: Am. Jour. Sci. (in press). Scott, R. B., Hajash, A., Kuykendall, W. E., Rona, P. A., Butler, L. W., and Nalwalk, A. J., 1973. Petrological and structural significance of the Mid-Atlantic Ridge be- tween 25°N and 30°N: EOS (Am. Geophys. Union Trans.), v. 54, p. 249. Scott, R. B., Rona, P. A., McGregor, B. A., and Scott, M. R., 1974, The TAG hydro- thermal field: Nature, v. 251, p. 301-302. Searle, D. L., 1972, Mode of occurrence of the cupriferous pyrite deposits of Cyprus: Inst. Mining and Metallurgy Trans., v. 81, p. B189-B197. Sillitoe, R. H., 1 972, Formation of certain massive sulphide deposits at sites of sea floor spreading: Inst. Mining and Metal- lurgy Trans., v. 81, p. B14I-B148. Sorem, R. K . and Foster, A. R., 1972, Marine manganese nodules: Importance of struc- tural analysis: Internal. Geol. Cong., 24th, Montreal 1972. sec. 8, p. 192-200. Strong, D. F., 1 974, An "off-axis" alkali vol- canic suite associated with the Bay of Islands ophiolites, Newfoundland: Earth and Planetary Sci. Letters, v. 21, p. 301-309. Swanson, S. B., 1975, Two examples of second- ary alteration associated with mid-ocean ridges (Master's thesis | : College Station, Texas A&M Univ. Thompson, G., Woo, C. C, and Sung, W., 1975, Metalliferous deposits on the Mid-Atlantic Ridge: Geol. Soc. America Abs. with Pro- grams, v. 7, p. 1297-1298. Upadhyay, H. D.. and Strong, D. F., 1973, Geological setting of the Betts Cove copper deposits, Newfoundland: An example of ophiolite sulfide mineralization: Econ. Geology, v. 68, p. 161-167. ACKNOWLEDGMENTS Reviewed by Enrico Bonatti and Ronald Sorem. Research supported by the Institute of Oceanology of the USSR Academy of Sciences, the National Oceanic and Atmospheric Admin- istration, and National Science Foundation Grant DES 74-18567. We are indebted to the scientific colleagues and crew aboard the R/V Akademik Kurchatov Mark DiStefano and Steve Swanson of Texas A&M University significantly aided our SEM and x-ray research. MANUSCRIPT RECEIVED DEC. 4, 1975 MANUSCRIPT ACCEPTED JAN. 13, 1976 236 APRIL 1976 447 47 Reprinted from: Marine Geology, Vol. 20, No 4, 315-334. Marine Geology, 20(1976) 315—334 © Elsevier Scientific Publishing Company, Amsterdam — Printed in The Netherlands RIDGE DEVELOPMENT AS REVEALED BY SUB-BOTTOM PROFILES ON THE CENTRAL NEW JERSEY SHELF W. L. STUBBLEFIELD and D. J. P. SWIFT National Oceanic and Atmospheric Administration, Atlantic Oceanographic and Meteorological Laboratories, Miami, Fla., (U.S.A.) (Received December 12, 1974; accepted August 4, 1975) ABSTRACT Stubblefield, W. L. and Swift, D. J. P., 1976. Ridge development as revealed by sub- bottom profiles on the central New Jersey shelf. Mar. Geol., 20: 315—334. Closely-spaced 3.5 kHz seismic profiles were collected over the north-easterly trending ridge and swale system 50 km east-southeast of Atlantic City, New Jersey. They yield infor- mation on the Late Quaternary depositional history of the area, and on the origin of the ridge system. Four of the sub-bottom reflectors identified were sufficiently persistent to warrant investigation and interpretation. These reflectors, which have been cored, litho- logically identified, and radiocarbon dated, are stratigraphically higher than the reflectors dealt with by the majority of previous studies. The upper three reflectors are definitely mid- and post-Wisconsin in age and present a record of the most recent glacial cycle. The upper three units associated with the observed reflectors appear to exert a pronounced influence on the bathymetry. The gently corrugated ridge system of Holocene sand is formed over the regionally flat-lying upper unit, an Early Holocene lagoonal silty clay. The characteristically flat, broad depressions of the area are floored by this lagoonal material. Locally, however, marine scour has cut through the silty clay into an underlying unit of unconsolidated fine Pleistocene sand. Several stages of trough development appear to be represented. After penetrating the lagoonal clay, troughs are initially narrow, but when incised through, the sand into a lower, Pleistocene, silty -clay unit, the troughs become notably wider. As downcutting is inhibited by the lower clay, the upper clay is undercut as the trough widens in a fashion similar to a desert blowout. The sub-bottom reflectors indicate that ridge development on the central shelf has involved aggradation as well as erosion. Some ridges seem to have grown by vertical and lateral accretion from small cores. The internal structure of other ridges suggests that they formed by the coalescence of several small ridges. Others appear to have undergone appreciable lateral migration. The ridges appear to be in a state of continuing adjustment to the hydraulic regime of the deepening post-Pleistocene water column. INTRODUCTION A prominent system of northeasterly trending ridges and depressions exists on the shelf floor 50 km east-southeast of Atlantic City, New Jersey. Examination of a 1:125,000 scale ESSA bathymetric map contoured by Stearns (1967) suggests that the ridges comprise two basic populations in 448 316 2 J^^'AP" 15' Fig.l. Index map of the study area with the New Jersey coastline inset for regional setting. The bathymetric contour lines are in fathoms. terms of spacing and height (Fig.l). A small-scale ridge system is super- imposed on a large system. The latter appears to be impressed onto a broad shoal-retreat massif, a constructional feature resulting from the retreat of a littoral drift depositional center associated with a retreating estuary mouth (Swift, 1973). The ridge spacing of the larger ridge system averages 3.1 km with a mean flank dip of 0.4°. In addition to the two basic populations a third, smaller system of contrasting sediment bands of negligible relief, has been observed from side-scan records, bottom photographs, and submersible dives (McKinney et al., 1974). Genesis of the ridge topography of the surficial sand sheet on the inner and central continental shelf has remained an enigma to workers since the pioneer work of Veatch and Smith (1939). A historical school of thought suggests that the pronounced undulations of the sandsheet are fluvial or littoral features formed during the lower sea-level stands of the Pleistocene (Veatch and Smith, 1939; Shepard, 1963; Kraft, 1971; McKinney and 449 317 Friedman, 1970). Others questioned the feasibility of these structures surviving a marine transgression and suggest instead a post-transgressive response to the Holocene hydraulic regime (Uehupi, 1968; Swift et al., 1972; Stubblefield et al., 1975). Uehupi (1970) subsequently abandoned the hypothesis of recent reworking of the Holocene sands and proposed a mechanism of terraces and barrier beaches overstepped by a transgressive Holocene sea. In order to resolve the controversy surrounding the origin of the ridges, a dense network of high-resolution, shallow-penetration seismic-reflection profiles was collected in a 400 km2 area (Fig.l). The investigation was directed toward the internal structure of the sand sheet and the role which the sub-bottom reflectors contribute to the existing bathymetry. In addition, some of the reflectors were correlated with data of previous workers in an effort to establish geological continuity with other sections of the New Jersey continental shelf. METHODS Field methods The seismic reflection data were collected from the NOAA ship "Peirce" during August 1973 using a 3.5 kHz transducer. The transducer, with a 0.2 m/sec pulse length was towed 5—6 m beneath the surface at ship's speeds varying from 3.5 to 4.0 knots. The seismic record was recorded at a 250 m/sec scan rate. The seismic lines were run normal to the ridged features in an area previously vibracored (Fig.l). This approach ensured maximum delineation of the sand sheet's structure and permitted correlation between the core record and seismic reflectors. Raydist navigation provided an accuracy of ±10 m. Laboratory methods The 115 km of seismic records were scanned for bottom and sub-bottom reflectors and "hand-smoothed" to compensate for sea surface waves. Each reflector was converted to X— Y values, placed on computer data cards by means of a X— Y digitizer unit, and subsequently plotted by a Univac 1108 computer. With this method, the horizontal scale was reduced by a 1:10 ratio and the vertical by 1:2 resulting in a vertical exaggeration of 5. This exaggeration, together with that resulting from the speed of the vessel, yields a composite vertical exaggeration of 12:1. Such a degree of vertical exaggeration enables delineation of subtle features in the original record. A travel time of 1.65 km/sec was used through both the water and uncon- solidated sediment. The negligible error induced by a slightly fast travel time through the water (1.65 km/sec as opposed to 1.50 km/sec) is not in conflict with the purpose of the study. 450 318 LATE-QUATERNARY STRATIGRAPHY The seismic profiles reveal changes in acoustic impedence (reflectors) which, in turn, can be correlated with the lithology sampled by vibracores. As many as eleven reflectors were observed in the seismic records but only four were of sufficient consistency throughout the area to warrant discussion. The four reflectors of interest have been lithologically identified, strati- graphically dated from vibracores (Stubblefield et al., 1975) and described (Fig.2). Radiocarbon dates were obtained from analysis of shell material in I V-l ss (CREST) -W SEA LEVEL 100- 200 CM I I I i I I i_ 10 14 18 2 2 500 BP» 100--- 3,760 BP*' .** 4> -|-^HS 200- » v-4 (UPPER FLANK) V-3 (LOWER FLANK) CM | | SAND fHf SILTY CLAY \t*\ SHELL MATERIAL <^ CROSS -BEDDING -?- DISCONFORMITY Hs HOLOCENE SHELF SAND H,£.£l.£ SEISMIC REFLECTORS I I I 1 I J L_ 10 14 18 2 2 10,050 B.P.' 10,950 BP1 I00i 22,035 BP: 200 25,300 B.P.; 300- 36,600 BP. 400- CM ' *i* V-2 (TROUGH) 29,700 B P«: IOO-f>:*j i i i i i i i i , 10 14 I 8 2 2 4> fc^ 32,150 BP iP1 Si 200-?-^ "--- G CM S as< I I I I I I I I 10 1.4 18 2 2< Fig.2. Lithic log of four vibracores and mean grain size in quarter phi (0) units within the sand horizons. (Modified from Stubblefield et al., 1975). 451 a±. a 3 £ o 0 JQ X! 3 C 0 £ 4J 03 s "a! 'w ^ t/3 m 0 0 Oh o 03 c -a c 2 * 2 S - ■- § Ma? 3 Oi c WD J) taJO 03 15 O ■c £ <*H tm 0 < e 0 t-H -3 W a J •e < u (A '2 H Q D 0 S2 03 01 e 3 Oh U-i o 2 c ■3 c 41 03 03 £ 44> Oi 4J '£ o 0 c 0) C H- H 5? £ £ 44> C 5 03 0 '3 J u Oh o o CO O 0) CO +* 1« si o ■s: --1 a) a s § « -2 g "C s. a « ° * in <2 ii c rb Oh < Ph 3 w 0 03 -u CD CIS 0 0> =3 c .S « -a c o 03 _S 03 s c 03 & 2 So 2 > 0 0 0 o -a b -C a ■& 03 a. hS a « I* a w I I C 03 si I £ & £ 3 rt a PH wfc o o o o CO CO to CD CM CO V A 3 0 00 Irt o CO o a o 03 % 0 61) OJ J2 I/J u 441 "5 04 a 0 1/3 T3 IH £ O O I o -a >> 03 >> 03 i % Li 41 £ tac _ C w ZT >> C 3 £ JS « -a £ 2, >S -g - a; 3 O a v 3 "3 g c u tab -5 >> 03 £ -o ^ >> 4) *_; £ '33 -2 o C 03 !/) oi *j £ '35 o LO Irt CO 05 ©_ o" (N H C<1 V 452 320 the vibracores. To avoid the often confusing situation of reflector labeling versus stratigraphic-horizon labeling, each unit in Fig. 2 is denoted by the same label as its upper boundary reflector; e.g., reflector H marks the upper boundary of unit H. H denotes Holocene deposits, P the Pleistocene material, and sediment of questionable age is marked as G. Various observed features of the four seismic reflectors are summarized in Table I. QUATERNARY STRATIGRAPHY Unit G Unit G is the lowest unit in the section. It has been penetrated by vibra- coring only in its uppermost 10 cm (core V2, Fig. 2), where it appears as a clean, fine-grained, shell-free sand. Its textural character and its minimum constraining age of >36,000 B.P., suggests that this unit was deposited in a near-shore environment, perhaps during a period of an advancing sea marking the commencement of the Plum Point Interstadial (mid-Wisconsin). This inference is supported by the apparent absence of an unconformity between this supposed basal sand and the overlying offshore silty-clay deposit. Strati- graphically, however, this unit appears to correlate with the unit underlying Garrison's (1970) Pleistocene unconformity, which he suggests as Late Tertiary in age. Garrison's work was over a broad area of the continental shelf south of New England and to the northeast of our work area. The small size of the study area relative to Garrison's work may have resulted in the lack of detection of an unconformity. Until a more detailed coring program is completed, the age of unit G and thus an interpretation of its depositional environment remains uncertain. Unit PI This medium gray, silty clay is perhaps the most widespread of the Late- Quaternary sequence, as indicated by the persistency of its reflective surface, reflector PI (pp.324— 325). This unit is of Pleistocene age, with dates ranging from 25,300 ± 1040 B.P. to >36,000 B.P. The younger section of this unit was probably an offshore deposit formed in advance of the prograding shoreline represented by unit P. However, the older part of unit PI, approxi- mately 36,000 B.P. in age, may reflect the maximum glacial retreat during the Plum Point Interstadial as described by Goldthwait et al. (1965) and Milliman and Emery (1968). The age of unit PI, mid-Wisconsin including the Farmdalian substage, is comparable to that proposed by McMaster and Ashraf (1973) for their reflector, P2. They made a tentative correlation of their reflector with Garrison's (1970) Pleistocene unconformity. McMaster and Ashraf's work was to the east of this study on the eastern fringe of Long Island extending south to the shelf break. They traced their P2 reflector across most of the shelf at sub-bottom depths of 17—34 m, but fail to mention the amount of 453 321 regional dip of their P2 other than that it parallels the present shelf surface. In the present study area of this study the regional dip of PI was calculated to be 0.04° to the S61°E. If 0.04° dip is assumed, an approxi- mation of 17m/97 km (17 m/l° latitude) depth compensation may be applied. By projecting McMaster and Ashraf's reflector for an additional 80—90 km in a plane normal to the strike of the eastern Long Island coast- line, a depth comparable to that of our reflector PI results. Unit P The uppermost Pleistocene sand, dated at 22,035 ± 665 B.P., possesses a slightly irregular reflective surface and ranges in thickness from 1 to 8 m. Throughout most of the sample area, however, the thickness varies from 2 to 4 m. The maximum thickness of unit P is in the southeast sector. The upper reflective boundary of this unit, reflector P, has a dip of 0.02° and a strike of S38°E. The dip angle and strike direction were calculated using the reflector's depth throughout the study area. The strike direction is within 5° of the present beach orientation in the vicinity of Little Egg Inlet, New Jersey (Fig.l). Unit P is a clean, medium-grained upward-coarsening sand (Fig. 2). This characteristic and its date of 22,035 B.P. suggest a deposition environment of a prograding shoreline. If this inference is valid, the advance of the Holocene seas was controlled by the regional gradient established during periods of lower sea level, since the coast-concordant strike indicates only a slight reorientation of the beach during the last 20 millenia. After the Plum Point Interstadial, the ice sheets readvanced, the marginal seas withdrew, and the Pleistocene sand of this unit was exposed to subaerial processes. Fig. 3 suggests that the Pleistocene sand, which is 40—50 m below THOUSANDS OF YEARS BEFORE PRESENT 4 8 12 16 ii i 20 24 i i i i i 28 32 i i i i 36 i i ^~""^N. \ \ \ \ / / \ \ \ \ 1 1 \ \ >• \ \ / 1 \ \ \ / / \ \ \ / / \ / / \ / \ \J MIUIMAN & EMERY (1968) CURRAY (1965) 20 40 60 S 5 uu 80 X at 100 -120 140 Fig. 3. Comparison of data from the vibracores with sea level curves of Milliman and Emery (1968) and Curray (1965). The sample ages are represented by dots and the range in age by error bars. (From Stubblefield et al., 1975) 454 322 present sea level, was a positive area from 12,000 to 20,000 B.P. Radio- carbon dates from the four vibracores are in general agreement with the sea- level curves of Milliman and Emery (1968) and Curray (1965). Visual evidence in Core V-3 (Fig.2) indicates a possible disconformity at the top of the clean sand (unit P). Sheridan et al. (1974) report an extensive erosion surface on their upper Pleistocene unit, which probably correlates with P. Core V-2 (Fig.2) demonstrates further evidence for erosion, in that the Holocene lagoonal deposit and the bulk of the Pleistocene sand are both absent. However, much of the missing section in Core V-2 is thought to have been removed by the modern marine erosion as explained in a later discussion, rather than by subaerial erosion during Early Holocene times. Unit H The upper unit in Core V-3 (Fig.2) is a medium gray silty clay with a locally irregular and discontinuous upper boundary (reflector H). The reflector appears as an undulating surface with the deepest part found under a topographic high in the eastern sector of the study area. The depth of reflector H, below present sea level, ranges from 36 to 52 m and the thick- ness of unit H varies from 0 to 6 m. Unit H thickens to the southeast in the direction of its regional slope. The absence of this upper silty clay at various places throughout the study area is probably the result of both erosion and non-deposition. Examination of the bathymetry in Fig.l fails to suggest recent downcutting in those places where the upper silty clay is missing within the substrate, suggesting that its absence may be due to a positive area during deposition, rather than subse- quent erosion. However, where unit H intersects the surface, particularly as an outcrop in the deep topographical troughs (McKinney et al., 1974), and in those places where an underlying unit is surficially exposed (profile A, Fig. 4a; profile K, Fig. 4b) erosion of unit H is obviously occurring. The depositional environment of the upper silty clay is inferred from radiometric ages, lithology, depth of unit, and fragmented shell material. The silty clay is underlain by medium to fine sand (unit P) which is dated at 22,035 ± 665 B.P., and is overlain by coarse sand with shell material dated at 10,950 ± 360 B.P. These limiting ages indicate that the unit was deposited during a period in which landward passage of the shoreline occurred. The lithology of cored sections of this unit is similar to that described by Sheridan et al. (1974) as a Holocene lagoonal deposit, near the Delaware coast. In addition, the average depth of reflector H is approximately 42 m below present sea level which places the unit in that portion of the Emery et al. (1967, fig. 4) diagram described as lagoonal. The geographic limits of the Emery et al. (1967) study is sufficiently close to this work to allow application of its interpretations to our data. Shell material, too small to radiocarbon date, has been identified by Don Moore, University of Miami, as organisms capable of living in shallow, brackish environments (Crassostrea virginica and Mercenaria mercenaria), a conclusion which supports our inference of lagoonal deposition. 455 323 Holocene lagoonal deposits tend to occur during glacial retreat and marine transgression. The bracketing dates for unit H (> 10,950 <22,035 B.P.) include the time of maximum glacial advance which occurred 18,000 to 22,000 B.P. during the Woodfordian glacial cycle (Goldthwait et al., 1965; Schafer and Hartshorn, 1965). If this unit does in fact reflect deposition during glacial retreat subsequent to maximum Late Wisconsin ice advance, and if a date of approximately 16,000 years B.P. is accepted for the Pleistocene— Holocene boundary (Emery and Uchupi, 1972) a date of post- Pleistocene may confidently be applied to this silty clay. These four units, their related seismic reflectors, and their time-stratigraphic framework provide a record of a complete glacial cycle on the central New Jersey shelf. The retreat of the ice sheet, accompanied by the advance of the ocean is indicated by the lower unit G. PI possibly represents maximum glacial retreat and marine transgression during late mid-Wisconsin time. Unit P is then representative of the subsequent ice advance and retreat of the ocean. Unit H was deposited by the advancing Holocene lagoonal belt. SURFICIAL SAND SHEET Topography and internal structure The surficial sand sheet above reflector H is complexly structured. Large- scale ridges (shaded pattern of inset, Fig.l) appear as half-cylinders of sand resting on a relatively level reflector H (Fig. 4a, b). In some cases, internal structure may be observed. This may take the shape of apparent ridge "cores", formed by internal strata which parallel the ridge flanks (station 86, profile C, Fig.4a; record A, Fig. 5). Elsewhere, multiple "cores" within a ridge suggest coalescence of several nuclei during growth (station 77 to 80, record C, Fig. 4a). Internal reflectors with consistent direction of dip occur in some ridges (station 251 to 255, record G, Fig.4b) suggesting lateral ridge migration. Internal patterns are locally very complex; in record B, Fig.5, truncated reflectors suggest that a former ridge on the northwest side of the record has been leveled and the adjacent trough filled in; a new ridge has appeared to the southeast. Large-scale troughs (stations 120—140, profile D, Fig. 4a) appear to bottom in reflector H which is thinly mantled with a few centimeters of coarse, shelly or pebbly sand locally grading upward into centimeters of finer sand. This fine sand thickens towards the ridge flanks (Stubblefield et al., 1975). Locally, reflector H is without this surficial coating. Small-scale ridges (linear pattern of inset, Fig.l) likewise appear to be half cylinders of sand resting on reflector H. In the few cases where internal structure have been resolved, it appears to be similar to that of the large-scale ridges. Small-scale troughs, unlike large-scale troughs, locally penetrate through reflectors H and P, into unit PI (Fig.6). Two variants of such apparently erosional troughs appear. Small-scale troughs which penetrate into the sand of unit P tend to be "V" shaped in cross-section (station 60 to 62, profile B, Fig.4a; record A, Fig.6). Other small-scale troughs extend completely through 456 324 b 2, o in . \Ti b J o vk b b c 3WI1 13AV«i AVM0M1 457 325 C*-H >> 0 n (U CU £ £ ^3 >, Si 0> > 03 0 -^> < T3 j C 0) 03 > * m y s-< CS c« Gfl >i +-> -a c 3 01 09 01 V a J3 & O 0 +j 0) a> ja &M s 0 u a £ 2. 3 Dh « -5 ■o 3 o cfl .fi y • a I 2 0 A y « _ y GO JZ a 53 !S o Jh C|_| °^ -3 .2? y y C — T3 y c 3 2 o ° £ y i/j « y CO £ ^ -* m so CO 458 326 100m ® 78 79 PROFILE C *'*V* i*?**** -"' *"?*' '"» ~-""*vV,,v •»■>,» 100m Fig.5. Hand-enhanced, high-resolution seismic-reflection profiles. A. Ridge with internal "core", suggesting upward ridge growth by the addition of conformable beds. B. Zone of discontinuous ridge growth. Ridge at right was formed subsequent to filling of trough at left by the progradation of incline strata. See Fig.l for location. the unconsolidated Pleistocene sand and are floored by the silty clay of unit PI and assume a more nearly parabolic cross-section, with a rounded bottom and more gently inclined flanks. Evolution of the ridge topography Large-scale ridges are locally broken into segments by small-scale troughs which cross them at a low angle, suggesting that small-scale troughs formed after large-scale troughs (McKinney et al., 1974). The varieties of ridges and their internal structure suggest the following model for ridge evolution ( Fig.7). Large-scale ridges, hereafter called primary ridges, were initiated immediately after the passage of the shoreline, at about 10,000 B.P. (Fig.7a). They formed in the leading edge of the shelf sand sheet (Duane et al., 1972), which 459 327 ~^&>Z!S&£^^F£!^!!Z!!!%!!c£ t PROFILE B 100m PROFILE K , .,-, ,*»?'* «**■.' r l^f! ' 100m Fig. 6. Hand-enhanced, high-resolution seismic-reflection profiles. A. Immature, small-scale trough with a V-shaped profile. B. Mature trough with a parabolic profile with a broad axis and gentle side slopes. Note that the flat-lying strata abut against sides of troughs. The upper datum is approximately 5 m beneath the sea surface. advanced as the shoreface underwent erosional retreat (Stahl et al., 1974). Primary troughs formed concomitantly with primary ridges, by non- deposition between ridges, or by the movement of sand off reflector H onto the ridges. As the Holocene transgression continued and the water column deepened, the ridge topography underwent progressive modification. Ridge spacing, a function of flow depth (Allen, 1968), increased. Internal ridge structure suggests that this was accomplished by lateral ridge migration, or by coales- cence of several smaller ridges. Ridge width appears to have increased as a result of more intense sedimentation on ridge flanks than on ridge crests, so that the ridges expanded laterally, rather than building upwards. As a result internal reflectors are generally steeper than present ridge flanks. Scour of the trough floors has locally breached reflector H. Where this has 460 328 PRIMARY RIDGE DEVELOPMENT OF RIDGE TOPOGRAPHY Fig. 7. Hypothetical model for ridge development. A. Large-scale ridges are initiated in the nearshore environment. The ridges grow by vertical and lateral aggradation, resulting in "concentric" stratification. Large-scale troughs are zones of non-deposition. B. Small-scale, secondary trough is incised into older substrate. Profile is initially V-shaped. C. Secondary trough widens; the Pleistocene sand is readily eroded and the Holocene clay is subject to undercutting. Aggradation of secondary ridges is fed in part by sand released during trough erosion. occurred, rapid downcutting and removal of the noncohesive sand of under- lying unit P appears to have resulted in a relatively steep- walled small-scale trough, hereafter called a secondary trough (Fig.7B). At the same time, secondary ridges began to appear in the primary troughs. Some ridges seem to bear a levee-like relation to the secondary troughs (Fig.7B), as though excavation of the former supplied material to the latter. Where secondary troughs have penetrated as far as the silty-clay surface of reflector PI (Fig. 7C) the troughs are broader, as though the clay inhibited further downcutting, and encouraged lateral erosion and trough widening, after the fashion of a desert blow-out. This model for ridge formation may be compared with the inshore portion of the study area, traversed by profiles A— G (Fig.l). Here a broad primary trough between two primary ridges develops a secondary topography as it is traced southwest (profiles A— G, Fig.4). In Fig.8, bathymetry of transects A, B and C are matched with a common datum, approximately 13 m below sea level (computed with a travel speed of 1.5 km/sec through water), and adjusted laterally so that successive crests of the landward primary ridge (position 1) coincide. Since the secondary topography (positions 2—5) becomes increasingly better developed through profiles C, B and A to the south, this series may be approximately equivalent to a time series, and as such may be compared with Fig.7. 461 329 Fig.8. Overlay of bathymetry from transects A, B, and C. See Fig.l for relative geographic positions. The five labeled positions are explained in the text. Possible relation of ridge topography to hydraulic regime The progressive evolution of the ridge topography appears to reflect an attempt on the part of the sea floor to maintain an equilibrium response to the slowly changing hydraulic regime during the period of water column deepening and shoreline retreat associated with the Holocene transgression. While a morphologic progression may be inferred on the basis of seismic reflection and related data, the character of the hydraulic forcing mechanism must remain speculative until the nature of the flow field on the New Jersey shelf is adequately documented. Our present, rather unsatisfactory state of knowledge may be summarized as follows. Beardsley and Butman (1974) have noted that sustained, high-velocity currents in the Middle Atlantic Bight occur primarily during those winter storms whose trajectories, relative to the shelf, permit a prolonged period of northerly winds. Such winds cause an Ekman transport of surface water to the coast, and result in coastal setup on the order of 40—60 cm (Beardsley and Butman, 1974). When this occurs, a shelf-wide southward geostrophic flow ensues. Studies conducted by Csanady and Scott (1974) under somewhat different circumstances suggest that the coastal margin of such geostrophic flow may assume a jet-like character, with velocities greater than those of the main flow. Limited data from the North Carolina coast (Swift, 1975) suggests that during periods of peak flow such accelerated coastal flows may experience downwelling and that inner shelf ridges may in fact be initiated by such flow, with the downwelling jet local- ized between the ridge and the shoreface. Ridges left behind on the shelf floor by the retreating shoreface continue to be maintained by the flow field. Pre-recent substrate continues to be exposed in the troughs, and as noted above, secondary patterns of ridges may appear. Continued ridge maintenance requires that the storm flow field be structured; a homogeneous flow field would serve to degrade ridge crests and fill in troughs. Such structure has not been observed, but it is predicted by theoretical and experimental work of Ekman (1905), Faller (1963, 1971), Faller and Kaylor (1966), Lilly (1966), Hanna (1969) and Brown (1971). Ekman (1905) has shown that wind-driven shelf flows would tend to have a three-layered structure. An upper boundary layer consists of wind-driven water, whose speed is in excess of the geostrophic value induced by regional set-up. The upper boundary layer is characterized by an Ekman spiral in 462 330 which wind-driven surface water moves at 45° to the right of the wind (in the northern hemisphere). With increasing depth, velocity vectors shift progressively to the right, and speed decreases to the geostrophic value. Depth-averaged flow in the upper boundary layer is 90° to the right of the wind direction. Beneath the upper boundary layer, the core flow is geo- strophic in nature, moving approximately parallel to the coast in response to the pressure field induced by wind set-up. Below the core flow, water is sheared against the stationary sea floor, causing a lower boundary layer that is characterized by a reverse Ekman spiral. As the bottom is approached, flow is diverted progressively toward the left (in the northern hemisphere), and the speed is progressively reduced. The thickness of the boundary layers is a function of velocity at the top of each layer, and the characteristic eddy viscosity of the layer. Very little is known about the values that these parameters attain on the Atlantic shelf during storms. However, it seems probable that during storm flows, the boundary layers would expand at the expense of the core flow and would partially or completely overlap (Leetmaa, 1975). Above a critical Reynolds number the internal character of the boundary layers, as well as their thickness, must change markedly. The boundary layers become unstable. However, since the flows are still subject to the Coriolis effect, the instability is not randomed but ordered (Faller and Kaylor, 1966; Faller, 1971). The flow divides into relatively sharply defined zones of down- welling, high-velocity surface water, alternating with more diffuse zones of upwelling lower-velocity bottom water (Faller, 1971). The result is a series of helical vortices, with alternate cells rotating with the opposite sense. The extent to which this scheme applies to storm flows on the Atlantic shelf is not known. However, if such a cellular flow structure should couple with the cohesiveless substrate of the Atlantic shelf during periods of peak flow, sand ridges might be expected to localize, and be localized at, zones of bottom-current convergence; and troughs at zones of bottom-current divergence. According to this model the ridges would be classic bedforms, in the sense of morphologic responses to secondary flow patterns within a sheared flow (Wilson, 1973). The nature of such coupling is problematic. The flow cells, if they indeed occur, are very flat. In the study area the ratio of ridge height to crest-to-crest distances averages 1:12 for the secondary ridges and 1:120 for the larger primary ridges. The ratio for the secondary ridges is similar to flow-cell spacing described by Faller and Kaylor (1966) for rotating tank experiments. They noted a spacing of 11 D for "small scale" cells where D is equal to the thickness of the Ekman layer. The spacing is "much greater" for large-scale cells. The sharply defined nature of the secondary troughs corresponds with the sharply defined nature of downwelling zones. Large-scale, primary troughs, however, are broad and flat. If they are responses to zones of down- welling, then the focus of downwelling must shift across the trough during the storm event in order to produce substrate mobility over large areas. Faller and Kaylor note that small-scale cells are aligned to the left of the 463 331 mean flow, while the large-scale cells are aligned to the right of the smaller cells. Large-scale, primary ridges, in fact, tend to be aligned to the right of small-scale, secondary ridges. By this criterion, primary and secondary ridges may be of synchronous origin. However, morphologic relationships suggest that primary ridges formed prior to secondary ridges. The sequence of primary ridges can be traced landward on the northern flank of the Great Egg Shelf Valley (Swift, et al., 1972) where they appear to be presently form forming on the shoreface. The offshore primary ridges tend to be J-shaped, hooking landward into the Great Egg Shelf Valley, as though affected by the tidal flow of the shelf valley when it was an active estuary mouth (McKinney et al., 1974). Secondary ridges only occur in this offshore zone. Their associated troughs pass through primary ridges, as though they were a later overprinting. If this is the case, then the secondary ridges may be a response to a change in the hydraulic regime initiated by a critical alteration in the shoreline configuration and bathymetry during the course of the Holocene transgression. Source of the Holocene sand sheet The seismic reflection observations presented above, including published data, place some constraints on possible sources for the Holocene sand sheet of the central New Jersey shelf. As noted by Meade (1969), Atlantic coastal estuaries are not sources of fluvial sand, but instead serve as sinks for both fluvial and littoral sands. Thus sands overlying the Holocene lagoonal carpet must be of other than fluvial origin. Possible sources are: (1) erosional retreat of the barrier face; (2) in-situ origin by breaching of the lagoonal carpet and erosion of the underlying sand; and (3) southerly transport on the shelf surface during storms. The first possibility is difficult to evaluate. Since the Late Holocene reduction in the rate of sea-level rise, New Jersey coastal barriers appear to be at least locally growing upwards in place, being nourished by the inner shelf sand sheet, rather than vice versa (McM aster, 1954). The reverse may have been true during the earlier period of rapid sea-level rise, with the sand sheet forming as a debris blanket resulting from erosional shoreface retreat (Stahl et al., 1974). Since the barriers themselves rest on the lagoonal carpet deposited landward of them, their source of sand during this period must have been from updrift, from eroding headlands, or from zones where the shoreface was incised through the lagoonal carpet (unit H) into the under- lying Pleistocene sand (unit P). This hypothesis is in accord with the regional ridge pattern (Fig.l) in which the sequence of ridges extending from the study area back to the New Jersey coast appears to form a shoal-retreat massif, marking the retreat path of the littoral drift depositional center on the north side of the ancestral Great Egg Estuary (Swift et al., 1972). The second potential source, from the excavation of secondary troughs into the pre-recent substrate, is by itself inadequate to account for the Holocene sand sheet. Examination of Fig.4 indicates that of the 115 km of 464 332 seismic record, less than 9 km reveal erosion through the silty clay of unit H into the loose sand of unit P beneath. Furthermore, the surficial sand sheet is 2—12 m thick, but the Pleistocene sand, where still capped by unit H, is 1— 8 m thick; its volume is inadequate to serve as a sole source. We note, however, that northeast of the study area, the ridge topography of the massif gives way to a nearly flat surface with broad shallow hollows (Uchupi, 1970, pl.l). Deflation of this surface by shelf flows may also have contributed to ridge growth in the study area. SUMMARY The ridges occur in a belt trending across the shelf normal to the shore (Fig.l). The axis of individual ridges extend across the ridged zone, parallel or sub-parallel to the shore. The ridge sequence is inferred to be a shoal- retreat massif, the retreat path of the littoral-drift convergence localized on the northeast side of the ancestral Great Egg Estuary. Nearshore members of this sequence appear to be forming as shoreface-connected ridges in response to coastal storm flows (Duane et al., 1972). A little further seaward, similar ridges may have been detached from and abandoned by the shoreface during Holocene sea-level rise. Yet further seaward, in the study area described by this paper, larger ridges are spaced further apart. This may be an innate characteristic, due to the more intense nature of tidal flows associated with the Great Egg Estuary when it still received the drainage of the ancestral Schuylkill River (Swift et al., 1972), or it may reflect an adjustment of the ridge topography to the increasing depth of the flow. The character of internal reflectors suggests that this response took the form of a dominance of flank over crestal aggradation, so that narrow, steep-sided ridges became broader with more gently inclined flanks, and that lateral migration of ridges also occurred. The large-scale, offshore ridges appear to have undergone a second stage of evolution, in that a pattern of small-scale, more southerly trending ridges have been imprinted over the first pattern. Secondary troughs have locally been incised into the Early Holocene silty clay that underlies the surficial sand sheet. These are relatively steep-walled features. However, where they have penetrated to the underlying Pleistocene sand, undercutting and lateral erosion have resulted in broader, more gently rounded features. Secondary ridges may have been nourished in part by material released during the formation of secondary troughs. ACKNOWLEDGEMENTS We are indebted to the officers and crew of NOAA ship "Peirce" for their professional abilities and cooperative attitude. We thank Sue O'Brien and Dave Senn for drafting, Thomas Clarke of the University of Virginia for computer programming, and Drs. G. H. Keller and H. B. Stewart, Jr. for critical review. Radiocarbon dates were provided by facilities at the Department of Geology, University of Miami, Florida. 465 333 This study is part of the National Oceanic and Atmospheric Admini- stration's Marine Ecosystem Analysis (MESA) program. REFERENCES Allen, J. R. L., 1968. The nature and origin of bed-form hierarchies. Sedimentology, 10: 161-182. Beardsley, R. C. and Butman, B., 1974. Circulation on the New England Continental shelf: Response to strong winter storms. Geophys. Res. Lett., 1: 181—184. Brown, R. A., 1971. A secondary flow model for the planetary boundary layer. J. Atmos. Sci., 27: 742-757. Csanady, G. T. and Scott, J. T., 1974. Baroclinic coastal jets in Lake Ontario during IFYGL. J. Phys. Oceanogr., 4: 524—541. Curray, J. R., 1965. Late Quaternary history continental shelves of the United States. In: H. E. Wright, Jr. and I. G. Frey, (Editors), The Quaternary of the United States. Princeton Univ. Press, Princeton, N. J., pp.723— 735. Duane, D. B., Field, M. E., Meisburger, E. P., Swift, D. J. P. and Williams, S. J., 1972. Linear shoals on the Atlantic inner continental shelf, Florida to Long Island. In: D. J. P. Swift, D. B. Duane and O. H. Pilkey, (Editors), Shelf Sediment Transport: Process and Pattern. Dowden, Hutchinson and Ross, Stroudsburg, Pa., pp.499— 575. Ekman, V. W., 1905. On the influence of the earth's rotation on ocean currents. Ark. Mat. Astron. Fys., 2: 1—53. Emery, K. O. and Uchupi, E., 1972. Western North Atlantic Ocean; Topography, rocks, structure; water, life, and sediments. Am. Assoc. Pet. Geol. Bull., Memoir 17: p. 532. Emery, K. O., Wigley, R. L., Bartlett, A. S., Rubin, M. and Barghoorn, E. S., 1967. Fresh water peat on the continental shelf. Science, 158: 1301—1307. Faller, A. J., 1963. An experimental study of the instability of the laminar Ekman boundary layer. J. Fluid Mech., 15: 560—576. Faller, A. J., 1971. Oceanic turbulence and the Langmuir Circulations. Ann. Review Ecology and Systematics, 2: 201—233. Faller, A. J. and Kaylor, R. E., 1966. A numerical study of the instability of the laminar Ekman boundary layer. J. Atmos. Sci., 23: 466—480. Garrison, L. E., 1970. Development of continental shelf south of New England. Am. Assoc. Pet. Geol. Bull., 54: 109-124. Goldthwait, R. P., Dreimanis, A., Forsyth, J., Karrow, P. F. and White, G. W., 1965. Pleistocene deposits of the Erie Lake. In: H. E. Wright, Jr. and J. G. Frey (Editors), The Quaternary of the United States. Princeton Univ. Press, Princeton, N. J., pp.85 — 97. Hanna, S., 1969. The formation of longitudinal sand dunes by large helical eddies in the atmosphere. J. Appl. Meteorol., 8: 874—883. Kraft, J. C, 1971. Sedimentary facies patterns and geologic history of a Holocene trans- gression. Geol. Soc. Am. Bull., 82: 2131—2158. Leetma, A., 1975. Some simple mechanisms for steady shelf circulation. In: D. J. Stanley and D. J. P. Swift, (Editors), Marine Sediment Transport and Environmental Manage- ment. Wiley Interscience, New York, N. Y., in press. Lilly, D. K., 1966. On the instability of Ekman boundary flow. J. Atmos. Sci., 23: 481-494. McKinney, T. F. and Friedman, G. M., 1970. Continental shelf sediments off Long Island, New York. J. Sediment Petrol., 40: 213-248. McKinney, T. F., Stubblefield, W. L. and Swift, D. J. P., 1974. Large-scale current lineations on the central New Jersey shelf: investigation by side-scan sonar. Mar. Geol., 17: 79-102. McMaster, R. L., 1954. Petrography and genesis of New Jersey beach sands. State of New Jersey Dept, Conservation and Econ. Development, Geol. Surv. Bull., 63: 239 pp. 466 334 McMaster, R. L and Ashraf, A., 1973. Drowned and buried valleys on the southern New England continental shelf. Mar. Geol., 17: 79—103. Meade, R. H., 1969. Landward transport of bottom sediments in the estuaries of the Atlantic Coastal Plain. J. Sediment. Petrol., 39: 222—234. Milliman, J. I. and Emery, K. O., 1968. Sea levels during the past 35,000 years. Science, 162: 1121-1123. Schafer, J. P. and Hartshorn, J. H, 1965. The Quaternary of New England. In: H. E. Wright, Jr. and J. G. Frey (Editors), The Quaternary of the United States, Princeton Univ. Press, Princeton, N. J., pp. 113—127. Shepard, F. P., 1963. Submarine Geology. Harper and Row, New York, N. Y., 557 pp. Sheridan, R. E., Dill, C. E. and Kraft, J. C, 1974. Holocene sedimentary environment of the Atlantic inner shelf off Delaware. Geol. Soc. Am. Bull., 85: 1319—1328. Stahl, L., Koczan, J. and Swift, D. J. P., 1974. Anatomy of a shoreface connected sand ridge on a New Jersey shelf: Implications for the genesis of the shelf surficial sand sheet. Geology, 2\ 117—120. Stearns, F., 1967. Bathymetric Maps of the New York Bight, Atlantic Continental Shelf of the United States, Scale 1:125,000. National Ocean Survey, National Oceanic and Atmospheric Administration, Rockville, Md. Stubblefield, W. L., Lavelle, J. W., McKinney, T. F. and Swift, D. J. P., 1975. Sediment response to the present hydraulic regime on the Central New Jersey shelf. J. Sediment. Petrol., 45: 337-358. Swift, D. J. P., 1973. Delaware Shelf Valley: Estuary retreat path, not drowned river valley. Geol. Soc. Am. Bull., 84: 2743-2748. Swift, D. J. P., 1975. Coastal sedimentation. In: D. J. Stanley and D. J. P. Swift (Editors), Marine Sediment Transport and Environmental Management. Wiley Interscience, New York, N. Y., in press. Swift, D. J. P., Kofoed, J. W., Saulsbury, F. P. and Sears, P., 1972. Holocene evolution of the shelf surface, central and southern Atlantic shelf of North America. In: D. J. P. Swift, D. B. Duane and O. N. Pilkey (Editors), Shelf Sediment Transport, Process and Pattern. Dowden, Hutchinson and Ross, Stroudsburg, Pa., pp.499— 575. Uchupi, E., 1968. Atlantic continental shelf and slope of the United States — Physiography. U. S. Geol. Surv. Prof. Pap., 529-C: 30 pp. Uchupi, E., 1970. Atlantic continental shelf and slope of the United States — shallow structure. U. S. Geol. Surv. Prof. Pap., 529-1, 44 pp. Veatch, A. C. and Smith, P. A., 1939. Atlantic submarine valleys of the United States: The Congo submarine valley. Geol. Soc. Am. Spec. Pap., No. 7: 101 pp. Wilson, I. G., 1973. Equilibrium cross-section of braided and meandering rivers. Nature, 241: 393-394. 467 48 Reprinted from: Marine Sediment Transport and Environmental Management, D. J. Stanley and D. J. P. Swift, editors, John Wiley and Sons, Inc., Chapter 14, 255-310. CHAPTER 14 Coastal Sedimentation DONALD J. P. SWIFT Atlantic Oceanographic and Meteorological Laboratories, Miami, Florida The preceding chapters have discussed sedimentation in the intracoastal zone of lagoons and estuaries which lie seaward of the main shoreline, and on the open beach and associated surf zone. This chapter looks at sedimen- tation in the coastal zone as a whole, from the shoreline out to an indeterminate distance on the order of 5 km, where shelf flows are no longer affected by proximity to shore. From this perspective, the system of longshore sand transport beneath the zone of shoaling and break- ing waves can be examined together with a deeper sys- tem of longshore sediment transport driven by inter- mittent wind or tidal flows. Time and space patterns of sediment input into this double system, the character of sediment transport, zones of temporary storage or per- manent deposition, and the bypassing of sediment onto the shelf surface are analyzed. More complex patterns of sediment transport are also described, which result when coastal flows associated with straight coastal com- partments interact with circulation in the erosional re- entrants of rocky coasts or constructional inlets of la- goons and river mouths. ONSHORE-OFFSHORE SEDIMENT TRANSPORT In considering coastal sediment transport, it is convenient to divide the movement of sediment into an onshore- offshore component and a coast-parallel component, and to consider these separately before examining the coastal sediment budget as a whole. Coast-parallel transport is many times more intensive than onshore-offshore trans- port, but it is the latter that determines morphologic changes at given coastal transects. Hence this chapter begins by examining the coast in profile. Hydraulic Zones and Morphologic Provinces When examined in cross section, the inner shelf is seen to consist of a regular succession of morphologic prov- inces, each associated with a distinctive zone of hydraulic activity (Fig. 1). Subaerial environments of open coasts are most highly developed on barrier islands, where a zone of storm washover and eolian activity results in washover fiats and dune belts, respectively. The intertidal swash zone builds the beach foreshore. The foreshore progrades seaward dur- ing fair weather by the addition of successive inclined sand strata to form the beach prism, a body of stored sand. The upper surface of the beach prism is the beach back- shore. The zone of breaking waves may be divided into the breaker line, which tends to maintain a breakpoint bar, and a surf zone, in which a wave-driven littoral current flowing parallel to the beach is overriden by the bores of breaking waves. The littoral current tends to scour a longshore trough. On unconsolidated coasts capable of relatively short- term response to the hydraulic regime, the inner shelf seaward of the breakpoint bar tends to exhibit two mor- phologic elements. A more steeply dipping shore/ace ex- tends to depths of 12 to 20 m. Its upper slope may be as steep as 1:10; its seaward extremity, at 2 to 20 km from shore, may slope as gently as 1 : 200. Beyond it lies 468 255 256 COASTAL SEDIMENTATION AEOLIAN ZONE STORM WASHOVER SWASH SURF ZONE BREAK ZONE OF ZONE LITTORAL POINT SHOALING CURRENT WAVES OLDER S DEPOSITS BERM SWASH BAR DUNE FORESHORE BREAKPOINT BACKSHORE LONGSHORE BAR SHOREFACE TROUGH SHOREFACE FIGURE 1. Morphologic elements oj the open coast and corresponding hydraulic provinces. the flatter inner shelf floor proper; the transition may be abrupt or very gentle. The upper shoreface, to a depth of perhaps 10 m, corresponds to the hydraulic zone of shoaling waves. The lower shoreface and inner shelf flow also experience the surge of shoaling waves, but their slopes, textures, and bed forms are equally a response to unidirectional shelf flows. The Beach Profile Circulation in the surf zone and the morphologic response of the substrate are described in Chapter 13. This sec- tion deals with the net effect of such hydraulic process and substrate response on the onshore-offshore sediment budget. As a consequence of the enormous and nearly continu- ous expenditure of energy in the beach and surf zones, the topographic features of cohesionless sand found there may only exist as equilibrium or near-equilibrium re- sponses to the circulation patterns described in the pre- ceding chapter. The equilibrium is not a static one, however, as the characteristics of the wave regime that force the response are constantly changing, often more rapidly than the morphologic response can accommo- date. As a consequence, the nearshore beach and surf zone topography is endlessly destroyed and rebuilt ac- cording to a complex cycle, as the nearshore wave regime and circulation pattern alternate between fair- weather and storm configurations, and on a larger scale between the summer season of infrequent storms and the winter season of frequent storms (Davis and Fox, 1972); see Fig. 2. FAIRWEATHER PHASE! BEACH AND BAR BUILDING. The cycle is controlled by two mechanisms: the wave regime and the net circulation pattern driven by it. During fair weather, waves tend to be far-traveled swells, of low amplitude and long period. The asymmetry of associated bottom wave surge is marked, with the landward stroke beneath the wave crest being significantly more pro- longed and more intense than the seaward stroke be- neath the trough (Chapter 8, Fig. 8). Peak orbital velocities may be separated by periods of 8 seconds or on windward coasts, markedly longer. These same fair- weather swells tend to result in a relatively weak near- shore circulation pattern. Momentum flux, which is a function of wave height, is relatively low during fair weather both seaward and landward of the breaker, hence discharge through the littoral circulation cells is relatively low. During fair weather, these two mechanisms, bottom wave surge and the littoral circulation pattern, cooperate to store sand in the beach prism. The wave regime ap- pears to serve as a fractionating mill, dividing the avail- able sand into a fraction undergoing mainly bed load transport, and a fraction undergoing mainly suspensive 469 (a) WAVE DRIFT RETURN FLOW NET FLOW LITTORAL CURRENT RIP CURRENT FIGURE 2. Comparison of (a) fair-weather and (b) storm hydraulic regimes. Based on Longuet-Higgins (1953), Schiffman (1965), and Ingle (1966). 470 257 258 COASTAL SEDIMENTATION transport. Sand coarser than a critical size threshold will be driven landward as bed load, by the landward asymmetry of bottom wave surge, toward the breakpoint. Longuet-Higgins (1953) and Russell and Osorio (1958) have undertaken calculations and experiments to determine the shore-normal components of flow averaged over a wave cycle, in the nearshore zone of shoaling waves. Their results indicate an increase in net land- ward flow with increasing height off the bottom, as a result of the failure of wave orbitals to close. Super- imposed on this is a mid-depth return flow resulting in a three-layer flow system (Fig. 2). It is not entirely clear, however, if the latter component of flow would exist in nature as a response to wave setup, or if it was merely a wave tank artifact, induced by the continuity require- ments of a closed system. At the breakpoint, much of the energy heretofore available to drive sand landward over the bottom is lost to turbulence, and sand tends to accumulate as a break- point bar. Waves of oscillation turn into waves of trans- lation (bores), in which water moves forward as a mass, and there is some evidence to indicate that landward of the breakpoint the velocity cross section averaged over the wave cycle changes to a two-layer system (Schiffman, 1965); see Fig. 2. An upper layer moves landward as a series of bores, and tends to be compensated by a basal return flow. This two-layer system is of course super- imposed on the generally much stronger coast-parallel flow characteristic of the longshore trough. To the ex- tent that the two-layer flow prevails, the bar crest is a zone of flow convergence, and its sand storage capability is readily understood. The bar builds upward until the rate of deposition of sand at the conclusion of wave breaking is equaled by resuspension during wave break- ing. Depth of water over the bar at equilibrium is gen- erally a third of the water depth prior to formation of the bar (Shepard, 1950). Breakpoint bars tend to orient themselves normal to wave orthogonals. When deep-water orthogonals make a high angle to the shore, wave refraction does not fully eliminate this angle near the beach. Under these condi- tions, the bar tends to consist of series of en echelon seg- ments, each aligned obliquely with respect to the beach, and alternating with rip current channels (Sonu et al., 1967). Bar position is very sensitive to wave height, as this determines breakpoint position (Keulegan, 1948). If the tide range is appreciable, bar position will shift detect- ably through the tidal cycle. New bars tend to form during the peak or waning phases of a storm and to be slowly driven onshore as waves diminish during the en- suing fair-weather period, although an abrupt decrease in wave height may cause a second bar to form landward of the first. During the period of landward migration of the bar, coarser bed load sand may bypass the bar and move onto the beach, if the waves are sufficiently long in period to re-form after breaking (King and Williams, 1949). Such bypassed sand will tend to accumulate as a swash bar (intertidal bar), or the plunge point bar itself will tend to migrate landward to the point where it is captured by intertidal processes, and becomes a swash bar (Fig. 3). As noted by King (1972), a swash bar may only form when the beach slope is lower than the maxi- mum potential slope permitted by the grain size of the available sand; swash bars thus comprise attempts by the regime of wave swash and backwash to build to this ideal beach profile. Unlike plunge-point bars which are formed at a bottom current convergence, swash bars are formed by an abrupt bottom current deceleration. Their seaward slopes are swash current graded, but the land- ward slopes are lower than the angle of repose, and have the same net landward sense of sand transport. Swash bars are the dominant bar on fine, flat beaches such as those of the central Gulf of Mexico, where the wave climate is mild and the supply of fine sand is abundant. They also tend to form on beaches with a high tidal range, where the bar is exposed to swash and backwash throughout much of the tidal cycle (ridge and runnel systems). The landward movement of coarser fine sand during fair weather on open beaches may thus proceed as a sheet flow bypassing the bar, or migration of the bar up the beach, or more commonly as both. The result of this landward flux of sand is the formation of the beach prism of gently inclined sand strata, differentiated into the backshore beach (constructional upper surface sub- ject to eolian action) and foreshore beach (swash-graded forward surface) separated by the berm (Fig. 1). If swash bar migration is the dominant mode of beach aggrada- tion, then the berm will prograde seaward mainly by the welding to it of successive swash bars, and the inter- nal structure of the beach prism will consist of seaward- dipping cross-strata sets, whose internal structures dip more steeply landward (Davis et al., 1972). The ease with which breakpoint and swash bars can be constructed in wave tanks strongly suggests that these are indeed basic genetic types of bars. These two rela- tively simple types belong to a broad class of bed forms that arise in response to the mutual interaction of flow with the substrate. However, it has recently become ap- parent that much more elaborate patterns of bars may form more or less passively, in response to an innate pattern within the velocity field. Crescentic bars that form in response to standing edge wave patterns have been described in Chapter 13. On gently inclined shore- faces, shore-parallel bars may form in arrays of up to 30 471 ONSHORE-OFFSHORE SEDIMENT TRANSPORT 259 FIGURE 3. Sequence of maps showing bar migration and erosion at South Beach, Oregon. Bars form below mean sea level and advance up beach at rate of 1 to 5 ml day. Under the influence of strong, southward-flowing currents they migrate southward at 10 to I5m/day. When they reach midtide level, they become stationary, or welded to the beach. From Fox and Davis (1974)- units. Bowen (personal communication) has suggested that such multiple bar systems may form in response to standing waves generated by the partial reflection of low- amplitude, long-period (1-2 minutes) incident waves. Such complex bar patterns clearly amplify the fair- weather storage capacity of the surf zone. As noted above, the fair-weather littoral hydraulic regime is a fractionating mill, which splits the available sand into bed load and suspended fractions. The be- havior of the bed load fraction has been traced above. Sand thrown into suspension at the breakpoint and fine enough to stay in suspension in the turbulent surf zone will tend to be fluxed alongshore by the longshore flow in the surf zone, and out through a rip channel to rain out on the shoreface (Cook, 196*_,.. storm phase: beach and bar destruction. During a storm, the wave regime and the littoral circulation patterns cooperate to withdraw littoral sand stored dur- ing the preceding fair-weather period. Wave steepness (ratio of wave height to wavelength) increases beyond a critical value (Johnson, 1949), at which point bottom wave surge asymmetry is no longer efficient in driving coarser sand landward as bed load. Waves during storms are locally generated, and they tend to be shorter in period and higher (more energetic) with higher maxi- mum orbital velocities. More sand is thrown into sus- pension and the critical grain-size threshold between suspensive and tractive sand fractions is shifted to favor suspension. Suspension is more nearly continuous. At the same time, discharge through the littoral circulation cells is increased manyfold. During the advent of a severe storm the sudden sea- ward shift in breaker position, plus the great intensifi- cation of seaward sand transport, may be sufficient to destroy the bar and beach prism altogether. Some sand is driven across the back beach and over the dunes in the form of a washover fan (if this area is low enough to be so flooded), but most is transported seaward through rip channels and in rip current plumes. Toward the end of the storm, fallout from rip currents accumu- lates as a series of coalescing aprons of sand on the shoreface. Lagoons that are flooded during the period of rising storm surge may cut new inlets and break out 472 260 COASTAL SEDIMENTATION through their barrier islands. The associated sand-laden jets may greatly add to this shoreface fallout (Hayes, 1967). As the storm wanes, the bar re-forms, and the cycle begins anew. The cyclic nature of sand storage on beaches has been quantitatively assessed by Sonu and Van Beek (1971) in a study of northern North Carolina beaches (Fig. 4). They observed a sequence wherein a storm-degraded concave beach profile, representing minimum storage, passed by means of swash bar accretion to a. convex profile of maximum storage, during a four-month pe- riod. They noted that the sense of sedimentation (ero- sion or accretion) was more strongly correlated with the direction of wave approach (and hence with wind direc- (a) ^ — -<■ >v >• c \ c I t I y ->■ Accretion >- Erosion @ 5/15-5/22 Sea level 0 10 20 30 40 50 Horizontal scale (meters) FIGURE 4. (a) Characteristic sequences of beach profile (d) Time history of sand storage. From Sonu and Van Beek change, (b) Observed sequences, (c) Observed sequences as a (1971). function of sediment storage (Q) and beach width (S). 473 THE SHOREFACE PROFILE 261 (c) 50 40 30 n Ay, -*- Accretion -^. Erosion 30 40 Beach width, S (m) 50 A * Wave height K A A \ */ A A \w WVi ^"\ rV l\ V/v L, . _L i i i i i i w 1 1 1 1 10 20 10 20 10 20 10 20 10 20 10 20 Dec, 1963 Jan., 1964 Feb. Mar. Apr. May FIGURE 4. (continued) tion) than with the wave steepness. Waves arriving from the northeast tended to cause erosion. These were asso- ciated with strong onshore winds and probably a wind- driven bottom return flow. It appears that during pe- riods of strong alongshore or onshore winds, the system of littoral sand transport is no longer a closed system but discharges its sand into the wind-driven flow of the adjacent shelf floor. This deeper, intermittent system of transport is described in the following sections. THE SHOREFACE PROFILE Hydraulic Climate of the Shoreface Far less is known about the circulation patterns of the shoreface and inner shelf than is known about the cir- culation patterns of the surf zone. Classical coastal workers, long preoccupied by the surf, have been indif- ferent to this topic, as have been physical oceanographers, whose habit has long been to hurry in their ships across the inner shelf, to the intellectual challenges of the large- scale planetary flows of the deep ocean basins. This situation is being rapidly reversed in view of rising public concern over the coastal environment (see Chapter 2), but old mental sets still linger. The shoreface and inner shelf are a zone of transition, where the wave climate is still a major factor in shaping the seafloor, but where the shelf flow field is becoming of increasing significance in a seaward direction. There is some justice in the indifference of classical coastal workers to this hydraulic province. During periods of fair weather, the shelf flow on most coasts may be many 474 262 COASTAL SEDIMENTATION times less intense than littoral drift (Fig. 2A). Its veloci- ties, on the order of 1 to 10 cm/sec, are capable of mov- ing whatever fines happen to be in suspension, but are not significant transporters of sand, although sand is re- peatedly suspended by bottom wave surge at the crests of wave-generated ripples. Fair-weather flows, however, may be relatively complex in pattern, with nearshore reversals of the open shelf flow, induced by coastal prom- ontories and by interaction with the tidal streams of inlets and estuary mouths. Two kinds of inner shelf flows are quite significant in transporting sand and in molding coastal topography. On coasts with high tidal ranges, midtide current veloci- ties associated with the passage of the coastal tidal wave may exceed 2 knots and locally attain 4 knots a few hundred meters seaward of the surf. Enormous volumes of sand are shifted on each tidal cycle, with significant net transport in the direction of the residual tidal cur- rent. Coastal tidal flows are poorly understood and tend to be rather complex because of strong interactions be- tween tide-built topography and the tidal flow. Some examples are discussed in later sections (see pp. 294-295). Intense coastal flows may also develop during storms (see Chapter 4). Such flows are far more infrequent than semidiurnal tidal currents, but unlike the latter, they occur on every coast, whether or not strong tidal cur- © I I r^ HIGH PRESSURE >\. | •* GRADIENT <^j FORCE CORIOLIS \^ FORCE ^*% \ \ PRESSURE GRADIENT LOW FIGURE 5. Geostrophic flow on the continental shelf. (A) Parcel of water at a reference depth moves seaward in response to pressure gradient force. As it accelerates, it experiences a Coriolis force impelling it to the right of its trajectory. Eventually trajectory parallels isobars, and pressure force and Coriolis force balance. (B) Cross section of hypothetical shelf experiencing geostrophic flow: illustrating REFERENCE DEPTH RESULTANT FORCE .-', PRESSURE ^ FORCE > HIGH relationship of sea surface slope, isobaric surfaces, and reference depth. (C) Relationship between geostrophic flow and flow in bottom boundary layer. In latter case, a friction term enters the equation of motion, and the balance of forces occurs among a friction term, a pressure term, and a Coriolis term. Modified from Strahler (1963). 475 THE SHOREFACE PROFILE 263 rents occur. They, too, are significant transporters of sand. Without these storm-driven flows, the coasts of our planet would have a markedly different appearance. Storms, whether of tropical or extratropicalorigin, are rapidly moving counterclockwise wind systems that may be a thousand or more kilometers in lateral extent. Winds intensify rapidly toward the storm center, and in hurri- canes, by definition, exceed 74 mph. The extent to which the shelf water column will couple with storm winds depends on the trajectory of the storm with respect to the geometry of the shelf. Sustained regional coupling of water flow with wind flow appears to occur when the winds blow equatorward along the length of eastward- facing coasts (Beardsley and Butman, 1974) or blow poleward along the length of westward-facing coasts (Smith and Hopkins, 1972). Under such conditions, water in the surface layer will be transported landward as a consequence of the Coriolis effect. Coastal sea level will rise until the coastal pressure head balances bottom friction, and bottom water can flow seaward as rapidly as surface water flows landward. Beardsley and Butman report up to 100 cm of coastal setup under such condi- tions. Since the sea surface is inclined against the coast, there is a gradient of seaward-decreasing pressure at any reference depth. A parcel of water, accelerated by the pressure force, has its trajectory steadily deflected to the right by the Coriolis "force," until finally, it is flowing along the isobars and the pressure and Coriolis terms balance (Fig. 5). inner shelf velocity field. The complex velocity structure of the coastal zone is best approached in terms of the interaction of three major flow strata (Ekman, 1905; see Neumann and Pierson, 1966, p. 202). These are an upper boundary layer, a core flow, and a lower boundary layer (Fig. 6). The reader is advised to review Chapters 3 and 4 for a better understanding of this section. The upper velocity boundary layer experiences strong wave orbital motion and, much of the time, a vertical velocity gradient imposed on it by wind stress. When the surface boundary layer is fully developed, surface water tends to move at 45° to the right of surface wind as a consequence of the Coriolis effect. Each successive lower layer moves at slower speed than the one above it, and is deviated successively further to the right (Ekman spiral). Net flow averaged over the depth of the layer trends 90° to the right of the surface wind. Above a critical Reynolds number this Ekman velocity structure becomes unstable, and is overprinted by a more com- plex structure, in which zones of upwelling and down- welling alternate, forming a pattern of horizontal helical vortices aligned parallel to or at a small angle to the zor WAVE DRIVEN ZQNE QF OW FRICTION-DOMINATED FLOW ZONE OF GEOSTROPHIC FLOW UPPER BOUNDARY LAYER LOWER BOUNDARY LAYER FIGURE 6. Velocity structure of the shore/ace and inner shelf. (A) General form of velocity profiles through the upper boundary layer, core flow, and lower boundary layer, and relative values of eddy viscosity. (B) V elocity structure during a period of relatively mild flow. (C) Velocity structure during peak flow. mean flow direction (Langmuir circulation: Langmuir, 1925). The transition tends to occur at surface wind speeds of 10 km (Assaf et al., 1971). The coefficient of eddy diffusion A, is relatively large in the surface layer as a consequence of wave-generated turbulence (Fig. 6); it must undergo an abrupt increase at the onset of Langmuir circulation. Below the base of the layer, core flow extends, un- modified, down to the bottom boundary layer. In the core, water flows in slablike fashion, with little vertical shear. Core flows are generally geostrophic in the sense that in the equation of motion, the pressure term is pri- marily balanced by the Coriolis term (Fig. 5). However, a steady state geostrophic balance is rarely maintained for any length of time. The shelf pressure field is in a state of continual change, in response to the passage of the diurnal tidal wave and to the passage of weather systems. As the pressure field builds up and then decays, the flow must accelerate and decelerate in sympathy, 476 264 :OASTAL SEDIMENTATION constantly changing direction so that the pressure and Coriolis terms may balance. Such time-dependent flows are referred to as rotary tidal currents if mainly tide- forced, or inertial currents if mainly wind-forced. The character of the bottom velocity boundary layer differs fundamentally from its surface analog. The sur- face boundary layer is externally forced, by the wind. Its velocity gradient, wave surge, and secondary flow patterns are overprinted on the core flow, and are car- ried along with it. The bottom velocity boundary layer is caused by frictional retardation of the core flow as it shears over the motionless substrate. Its lowermost meter exhibits a logarithmic velocity profile (Chapter 7), but the lower boundary layer as a whole is a thicker stratum, characterized by a velocity profile that is a reverse Ek- man spiral. Frictional retardation of flow results in a deviation of boundary flow direction to the left of core flow, so that Coriolis and frictional terms may together balance the pressure term (Fig. 5C). The lowest layers, experiencing the greatest retardation, are deviated the furthest. Theoretical studies (Faller, 1963; Faller and Kaylor, 1 966) suggest that this layer is also subject to helical flow structure above a critical Reynolds number. However, no field studies of this phenomenon have been undertaken. Such innate flow stabilities, and also turbu- lence induced by bottom roughness elements, would lead to an eddy coefficient larger than that of the core flow (Fig. 6A). Three hydraulic provinces may be defined on the inner shelf on the basis of flow structure (Figs. 65, C). Near the beach, the two boundary layers of the shelf flow field must completely overlap. In this zone the effects of the regional pressure gradient on water be- havior are largely damped out as a consequence of fric- tional retardation. Oscillatory wave surge is the domi- nant water motion, giving rise to the complex nearshore circulation pattern described in the preceding chapter. A little farther seaward, the two boundary layers are more or less separate, but still occupy most of the water column. Flow is frictionally dominated; in the equation of motion the wind stress is largely balanced by friction. The effect of the Coriolis term is negligible in shallow water and there is little or no deviation of boundary flow with respect to core flow. The flow is Couette-like, in that there is a more or less linear velocity gradient from top to bottom. Still further seaward, the two boundary layers diverge significantly. The geostrophic core flow dominates the water column. This pattern of coastal flow zonation must vary with the intensity of the regional and local wind fields. An intensified regional wind will accelerate core flow and increase the thickness of the bottom boundary layer. Intensification of the local wind field will cause the upper boundary layer to thicken, though not necessarily at the same rate. The intensification of local wind may either lead or lag the intensification of wind on the adjacent shelf, depending on the trajectory of the weather system. The net effect of a storm is to expand the width of the coastal flow zones and to displace the outer two zones seaward. There are few data available for such situa- tions (see Chapter 4). From theoretical considerations, it appears that the upper and lower boundary layers may overlap far out on the shelf. Zonation becomes pri- marily a function of depth (Fig. 5C). In the zone of friction-dominated flow, the water accelerates in response to direct wind stress until the stress is balanced entirely by friction; the Coriolis term is not significant, and flow in this zone may take on the dimensions of a coastal jet (Csanady and Scott, 1974). The zone of friction-domi- nated flow will be a downwelling zone if local winds have an onshore component, or if regional coast-parallel winds result in onshore surface transport. It will be an upwelling zone if the reverse situation prevails (Cook and Gorsline, 1972). The deeper, offshore flow may retain a primarily geo- strophic balance of forces during a storm, although the friction term is necessarily more prominent. If overlap of the boundary layers extends through this zone, it is theoretically possible (Faller, 1971) that there be top to bottom overturn as a consequence of Ekman instability, with high-velocity, wind-driven surface water delivered to the seafloor in zones of downwelling. The velocity structure of the shelf water mass follows a seasonal cycle that is coupled to the cycle of density stratification. During the summer, this upper velocity boundary layer is the same as the upper mixed layer. Wave turbulence and Langmuir circulation maintain the layer's mixed character, while the pycnocline tends to decouple upper boundary flow from core flow. During the fall, the thermal contrast is weakened by surface cooling. The increasing frequency and severity of storms cause steady erosion of the lower, stratified portion of the water column by Langmuir circulation (Faller, 1971) and the upper mixed layer thickens at the expense of the stratified water below. Meanwhile, a lower mixed layer may be induced by intensified turbulence in the bottom boundary layer, and may thicken until the den- sity structure has simplified to a two-layer system (Char- nell and Hansen, 1974). Further vigorous storm action will drive the weakening pycnocline down to the sea- floor, so that there is no further impediment to top-to- bottom overturn by secondary flow components. Sedimentation on the Upper Shoreface The shoreface slope, with its gradient of seaward-de- creasing grain size, occurs primarily in the zone of wave- 477 THE SHOREFACE PROFILE 265 HSEDIMENT INPUT WAVE CLIMATE -Hgrain SIZE ^T DEPTH AS A FUNCTION OF DISTANCE FROM — SHORE FIGURE 7. Relationships of variables controlling slope of the shoreface. driven flow (Fig. 6) although its lower portion tends to extend into and be modified by the zone of friction- dominated flow. The slope and grain-size gradient of the shoreface have been generally considered to com- prise a response to the regime of shoaling waves seaward of the breakpoint, in which depth as a function of dis- tance from shore is itself a function of littoral wave power, sediment discharge, and grain size (Fenneman, 1902; Johnson, 1919, p. 211; Johnson and Eagleson, 1966; Price, 1954; Wright and Coleman, 1972; see Fig. 7. Johnson (1919, p. 211) has described this equilibrium relationship as follows: The subaqueous profile is steepest near land where the debris is coarsest and most abundant; and progressively more gentle further seaward where the debris has been ground finer and reduced in volume by the removal of the part in suspension. At every point, the slope is precisely of the steepness required to enable the amount of wave energy there developed to dis- pose of the volume and size of debris there in transit. The main line of inquiry into the forces maintaining the shoreface profile has led to the null-line hypothesis, evaluated in Chapter 8. The hypothesis has been ex- pressed in its most complete form by Johnson and Eagle- son (1966). It envisages shoreface dynamics in terms of a Newtonian balance of forces experienced by a sand particle on the shoreface, in which the downslope com- ponent of gravitation is opposed by the net fluid force averaged over a wave cycle. Since in shallow water, bottom orbital velocities are asymmetrical, with stronger landward surge (Chapter 8, Fig. 8), fluid forces are directed upslope. The gravitational force becomes more intense as the shoreline is approached and the slope increases. However, the fluid force increases yet more rapidly. As a consequence, for a given grain size there should be a null isobath, seaward of which particles of the critical size tend to move downslope, and landward of which they tend to move upslope. The equilibrium grain size should decrease with increasing depth. Hence, the shoreface sand sheet should tend to become finer down- slope, as indeed it does. The shoreface slope at each point should be uniquely determined by the grain size of sub- strate and the intensity of bottom wave surge. However, attempts to utilize null theory in the field have met with ambiguous or negative results (Miller and Zeigler, 1958, 1964; Harrison and Alamo, 1964). Objections include: (1) slopes are not sufficiently steep over much of the shoreface (Zenkovitch, 1967, p. 120), and (2) slope sorting by waves tends to be overwhelmed by other processes, which as the authors of the theory admit, are not accounted for in null theory. No account, for instance, has been taken of the process of ripple sorting as described in Chapter 8 (p. 117). Wells (1967) has shown that divergence of onshore-offshore transport of a given grain size from its null isobath should occur as an innate response to higher order wave interactions, without regard to the gravitational force acting on the grains. A perhaps more telling criticism of null-line theory is that a significant portion of shoreface sand travels not as bed load, but in suspension. Murray (1967) has per- formed tracer studies that indicate that on the upper shoreface, the dispersal of sand corresponds to the pre- diction of diffusion theory. Field observations by Cook and Gorsline (1972) have led them to conclude that the seaward-fining grain-size gradients of the shoreface are more likely to be caused by rip current fallout rather than by the null-line mechanism. It may be more fruitful to approach the problem of shoreface maintenance from the point of view of ener- getics, rather than from the point of view of a balance of forces. Such an approach would view the depth at each point of a shoreface profile as a function of wave power at that point. The ideal wave-graded profile would be one that experiences at each point a maximum bot- tom orbital velocity equivalent to the threshold velocity of the size class of available sand. It should be possible to construct an algorithm for calculating water depth as a function of wave characteristics and bottom sedi- ment grain size, based on the equations for bottom or- bital velocity, for friction energy loss to the bottom, and for the shoaling transformations of waveform that have been presented in Chapter 6. Lower Shoreface Sedimentation : Onshore-Offshore Sand Budget It seems doubtful that such a model for maintenance of the shoreface profile by the wave regime would be suf- ficient to fully account for the distribution of slopes and 478 TRANSECTS DEPTH 6? 20 20 20 Li ULlLL UULV 12 14 40 4 0 BERM BREAKER UPPER SHORE FACE LOWER SHORE FACE DIAMETER 75°30 75°00' DEPTH (M) 40 20 0 60 40 20 0 40 _ 20 55 o i- 20 = 203'05 EC l90 5 o 2ogns 0 112.5 20 0 1160 20 0 40320 ° | 2sl-A moo 2S~ DEPTH (M) i3.0 BEACH BREAKER PPER RE FACE q6.o I ■ UPPE 1 iMm )SHO nS.O J. -,10 0 320.0 OWER SHORE FACE DIAMETER (H FIGURE 8. Distribution of grain sizes on retrograding coasts. (A) TAe storm-dominated coast of Virginia-northern North Ijmuiden ^ Hoek van Holland Carolina. Data from Swift et al. (1971). (B) Dutch coast. Data from Van Straaten (1965). 266 479 THE SHOREFACE PROFILE 267 grain sizes associated with observed shoreface profiles. For instance, modern coasts whose historical records in- dicate that they are undergoing erosional retreat tend to consist of two distinctive grain provinces. From the breaker to a depth of about 10 m, the upper shoreface consists of fine, seaward-fining sand (Fig. 8). Seaward of 10 m, grain size on the lower shoreface and adjacent shelf floor is far more variable and generally markedly coarser. We may account for the fine, upper shoreface sand province as a mantle of rip current fallout, whose slope is adjusted by the regime of shoaling waves (Cook, 1969). However, the lower shoreface province of coarse variable sand does not fit the model for wave mainte- nance of the shoreface. We may consider the hypothesis that it is instead a response to the deeper, intermittent high-intensity flows of the zone of friction-dominated flow (Figs. 2B and 6B, C). Observations by Moody (1964, pp. 142-154) on the erosional retreat of the Delaware coast lend some sup- port to this hypothesis (Fig. 9). In this area, the shore- face steepens over a period of years toward the ideal wave-graded profile, during which time the shoreline remains relatively stable. The steepening is both a de- positional and erosional process. Moody notes that steep- ening was accelerated after 1934 because a groin system initiated then "presumably trapped sand, causing the upper part of the barrier between mean low water and — 3 m to build seaward" (Moody, 1964, p. 142). How- ever, erosion continued offshore at depths of 6 or 7 m below mean low water. The steepening process is not continuous, but varies with the frequency of storms and duration of intervening fair-weather periods. The slope of the barrier steepened from 1:40 to 1:25 between 1929 and 1954, but erosion on the upper barrier face between 1954 and 1961 regraded the slope to 1 :40. The steepening process is terminated by a major storm, during which time the gradient is reduced and a signifi- cant landward translation of the shoreline occurs. Moody ( 1 964, p. 1 99) describes the Great Ash Wednesday Storm of 1962, bracketed within his time series, as having stalled for 72 hours off the central Atlantic coast. Its storm surge raised the surf into the dunes for six suc- cessive high tides. The shoreline receded 18 to 75 m during the storm. While much of the sand was trans- ported over the barrier to build washover fans over 1 m thick, much more was swept back onto the seafloor by large rip currents and by the storm-driven seaward- trending bottom flow of the shoreface (Moody, 1964, p. 114); see Fig. 2B. Moody's observations allow us to present a general model of shoreface maintenance, in terms of the on- shore-offshore sediment budget. There seems to be little MEAN LOW WATER V.N « v. 1645 KILOMETERS FIGURE 9. Retreat of the Delaware coast, based on U.S. Coast and Geodetic Survey records and a survey by Moody. From Moody (1964). reason to doubt the applicability of the conventional wave-grading model to the upper shoreface, even if we cannot yet present this model in a quantitative manner. Upper shoreface textures and slopes are time-averaged responses to two opposing mechanisms, the seaward flux of suspended sand in rip currents on one hand, and the landward creep of bottom sand in response to the net landward sense of bottom wave surge on the other hand. The upper shoreface profile varies in cyclic fashion, with storage of sand mainly on the beach during the fair- weather summer season, and storage of sand mainly on the upper shoreface during the winter. For long periods of time, the upper shoreface profile may oscillate about the ideal wave-graded configuration. During major storms, however, the upper shoreface system of wave-driven longshore sand flux interacts with coastal boundary of the storm flow field. Sand eroded from the beach and bar by storm waves passes seaward 480 268 COASTAL SEDIMENTATION in intensified rip currents to the zone of friction-domi- nated flow (Fig. 2B), which during storms may take the form of a downwelling coastal jet. When this occurs, bottom flow on the lower shoreface will have a seaward component of flow, and the coastal sand transport system is transformed from a closed system of net sand storage to an open system of net sand loss. Sand raining out of rip currents will not come to rest, but will be transported obliquely seaward. If the storm is severe enough, the mantle of rip current fallout that accumulated during the preceding fair-weather period will be stripped off, and the underlying strata will be exposed to erosion. This hypothetical scheme has not yet been adequately tested by field observations. However, as a hypothesis, it has a number of advantages. It provides a rationale for the Bruun model of erosional shoreface retreat (Fig. 10). Bruun (1962; see also Schwartz, 1965, 1967, 1968) noted the characteristic exponential curve of the inner shelf profile, and accepted the hypothesis that it consti- tuted an equilibrium response to the hydraulic climate. With this premise adopted, it follows that a rise in sea level must result in a landward and upward translation of the profile, as long as coastwise imports of sand into the coastal sector under study are equaled by coastwise exports. The translation necessitates shoreface erosion and provides a sink for the debris thus generated be- neath the rising seaward limb of the profile. Moody's time series shows that over a 32 year period, shoreface erosion on the Delmarva coast was in fact nearly compensated by aggradation on the seafloor in accordance with the Bruun principle (Table 1). The small deficit is probably attributable to loss to washover fans, and through littoral drift to nearby Cape Henlopen spit. Moody's studies provide us with insight into the proc- esses governing the Bruun model. His observations indi- cate that the process of erosional retreat of the shoreface is not continuous. It is cyclic in a manner analogous to the annual cycle of the upper shoreface profile, but the period is related to the frequency of exceptional storms, and is on the order of years. The model also provides a more detailed and satis- factory explanation for the origin of the surficial sands of shelves undergoing transgression than does the relict- Recent sediment model of Emery (1968). The surficial sand sheet of the shelf is a lag deposit created during the process of erosional shoreface retreat by the seaward transfer of sand during storms and its deposition on the adjacent shelf floor (Fig. \0A). The nearshore modern sands of the upper shoreface are a transient veneer of rip current fallout. Both textural provinces are "modern" in the sense of being adjusted to the prevailing hydraulic regime; both are "relict" in the sense of being derived TABLE 1. Sediment Budget from the Delmarva Coast Sediment Source Period Average Volumetric Change* (m3/year) Barrier (mean low water to toe of sand barrier) Sand dunes (mean low water to top of sand dunes) Offshore erosion (principally on north- west side of ridges) Erosion from bay inside Indian River Inlet 1929-1961 148,000 1954-1961 -100,000 (estimated) 1919-1961 100,000 -69,000 Total erosion -417,000 1939-1961 +120,000 1939-1961 +5,700 Site of Deposition Tidal delta Barrier south of Indian River Inlet Offshore accretion 1919-1961 +256,000 Total accretion +381,700 Total erosion -417,000 Total accretion +318,700 Net erosion — 98, 300 m3/year Source. From Moody (1964). * "+" indicates accretion; "— " indicates erosion. from the underlying substrate. The role of shoreface re- treat in generating shelf sediments is explored further in Chapter 15. Deposits of the Coastal Profile: Textures and Bed Forms textures of the shoreface. The patterns of onshore and offshore sediment transport described in the pre- ceding sections give rise to systematic distributions of sediment types and bed forms over the beach and shore- face. The beach and surf zones consist of alternating belts of finer and coarser sand, the absolute grain-size values depending on grain sizes available to the coast and on the hydraulic climate of the coast (Bascom, 1 95 1 ) . The coarsest grain sizes are found on the crest of the berm, in the axis of the longshore trough during the erosional phase of the beach cycle, and on the crest of the plunge point bar. The distribution of grain sizes on retrograding shore- faces has already been described (Fig. 8). Upper shore- face sands tend to be fine grained to very fine grained, and become finer in a seaward direction. The grain-size 481 ONSHORE-OFFSHORE SEDIMENT TRANSPORT 269 ® ® VECTO« RESOLUTION OF PROFILE TRANSLATION AGGRADING OAR Will CAPTURE SHORELINE ZONE OF AGGRAOATION U^ WASHOVER CYCLE OF BARRIER SANDS HIGH HITO«Al 0«IFI DISCHARGE FIGURE 10. Dynamic and stratigraphic models for (A) a retrograding and (B) a prograding coast during a rise in sea level. IJ coastwise sand imports are balanced by or are less than coastwise sand exports, the hydraulically maintained coastal profile must translate upward and landward by a process of gradient is perhaps originally the result of progressive sorting (see p. 162) operating on the suspended sand load of rip current plumes; the underlying deposit be- comes finer down the transport direction in the manner of loess or volcanic ash deposits. It is perhaps secondarily the result of adjustment to the landward increase in bot- tom orbital velocity, and to the mechanism of ripple sorting (p. 117). On retrograding coasts such as the North Carolina and Dutch coasts (Fig. 8), the lower shoreface consists of variable but generally coarser sand. It is texturally adjusted to the coastal boundary currents associated with peak flow events. This material is of negligible thickness and constitutes a residuum mantling the erod- ing surface of the underlying older deposits. Its final resting place appears to be the adjacent seafloor, where it forms a discontinuous layer up to 10 m thick (Stahl et al., 1974). Bimodal sands tend to occur at the contact between the two provinces, where the rip current fallout blanket thins to a feather edge. This contact advances down- slope during fair-weather periods of upper shoreface ag- gradation, and retreats upslope during periods of storm erosion of the entire shoreface. shoreface erosion and concomitant aggradation of the adjacent seafloor (Brunn, 1962). If coastwise sand imports exceed exports, as is the case for deltaic coasts, then the profile must translate seaward and upward. Based on Curray et al. (1969). On prograding coasts, such as the western Gulf of Mexico (Bernard and Le Blanc, 1965) or the Costa de Nayarit (Curray et al., 1969), more sand is delivered by littoral drift during fair weather than can be removed by storms. The fine, seaward-fining sand of the upper shoreface extends down to the break in slope where it may become as fine as mud, and continues across the shelf floor (Fig. 105). Visher (1969) has observed size frequency distribu- tions in the surf zone and on the shoreface (see Fig. 11). Moss' theory may be applied to his observations (see p. 162), but caution must be used, as Moss' theory re- lates to quasi-steady flows, while in the coastal marine environment a high-frequency flow oscillation due to wave surge tends to be superimposed on a steady flow component. Peak wave surge regularly induces the up- per flow regime (Moss' rheologic regime), while the in- tervening flow may consist of one of the less intense stages. In Fig. 12, the supercritical flows of swash and back- wash in the intertidal zone have resulted in complex subpopulation assemblage (lower foreshore, 1.5 ft sam- ples). The contact (C) population, consisting primarily of shell debris, comprises up to 10% of the total distri- 432 270 COASTAL SEDIMENTATION SHORE - PROFILE TIDE GOING OUT 5 FEET " 11 FEET ■ LOWER . FORESHORE 0 12 3 PHI SCALE 4 - 15FEET ■ - 99 I PREDOM : SHELL i 90 50 10 1 0 1 LOW TIDE FOREST BEACH S. C.= FIGURE 1 1. Representative size frequency distributions of the shoreface. From Visher (i969). bution. A large C population is characteristic of Moss' rheologic regime; the B population, however, is less than 1%. While the hydraulic microclimate of rheologic flow is conducive to the incorporation of a large B popu- lation into the bed, such a response is presumably in- hibited by the gross hydraulic structure of the surf zone; suspended fines are steadily flushed seaward through rip channels. Two framework (A) populations are locally present, reflecting perhaps discrete responses to swash and the slightly higher backwash velocities. Subtidal surf zone populations (5 and 6.5 ft) are similar, but the framework population is better sorted (corresponding segment of the cumulative curve is steeper). This sorting is further improved in the upper foreshore sample ( 1 1 ft sample). This sample has a reduced contact (C) popu- lation and an enriched interstitial (B) population. B pop- ulation enrichment reflects the heavy rip current fallout of fine suspended sand experienced at this depth, and perhaps also the presence of Moss's fine ripple regime. BED FORMS OF THE SHOREFACE. CliftOn et al. (1971) have noted that the high-energy shoreface of southern Oregon is characterized by zones of primary structures that reflect the hydrodynamic subenvironment (Fig. 12). An "inner planar facies" occurs beneath the reversing supercritical flows of the swash zone; the associated structure within the deposit consists of thin beds and laminae of gently inclined sand. The rhomboid ripple marks and antidunes that form in each backwash are rarely preserved. Beneath the surf zone of gently sloping beaches lies an "inner rough facies" of shore-parallel ridges and troughs 1 to 2 m across and 10 to 50 cm deep. The flat-topped ridges tend to be steepest on the seaward side, and the ridges migrate seaward. During periods of strong littoral currents, troughs are more nearly perpendicular to land, and migrate downcurrent and offshore. The internal structure of this facies consists of medium-scale (units 4-100 cm thick) seaward-dipping trough cross-bedding. Beneath the breakpoint lies an "outer planar facies." No bar existed here during the period of Clifton's study. Small ripples may form during the initiation of trough or crest surge, but the flow becomes supercritical during maximum surge, and the ripples are destroyed. The in- ternal structure is horizontal lamination. Clifton et al. describe the upper shoreface of the Ore- gon coast as the "outer rough facies." The characteristic bed forms are lunate megaripples, 30 to 100 cm high, and with spans (terminology of Allen, 1968a, pp. 60-62) between 1 and 4 m. Concave slopes face landward, and the ripples migrate landward at rates of 30 cm/hr. 483 LONGSHORE SEDIMENT TRANSPORT 271 WAVE ACTIVITY WAVE TYPE ORBITAL VELOCITY: OFFSHORE SWELL NEARSHORE c sea surface Fe" iCO 0 2C 4C Meters SEA ^LOOR- STRUCTURAL FACIES FIGURE 12. Relationship of depositional structures to wave type and activity. From Cl ij ton etal. (1971). ASYMMETRIC RIPPLE INNER PLANAR Crestal sands are notably coarser than trough sands. The resulting internal structure is a medium-scale cross-strati- fication with foresets dipping steeply landward. The lower portion of the upper shoreface of the Oregon coast has an "asymmetrical ripple facies" of short-crested wave ripples 3 to 5 cm high with chord lengths (Allen, 1968a. pp. 60-62) of 10 to 20 cm. They may reverse asymmetry with each passing crest and trough, or reveal a persistent landward asymmetry. Crests become lower, longer, and straighter as the pattern is traced seaward. Interfering sets are common, weaker sets tending to occur as ladderlike rungs in the troughs of the stronger set. The angle between the two sets is bisected by the wave surge direction. The internal structure of this facies tends to consist of small-scale (less than 4 cm) shoreward-inclined ripple cross-lamination, inter- fingering with gently dipping, medium-scale scour and fill units. A similar sequence of bed forms and textural prov- inces has been reported from the Georgia coast by Howard and Reineck (1972). The Georgia coast has a milder wave climate than those described above, al- though it is also characterized by strong tidal flows. The equilibrium configuration of the shoreface is rather different here (Fig. 13): the slope is much gentler, and the break between the fine sand of the upper shoreface and the coarse sand of the lower shoreface occurs as far seaward as 14 km from the beach. An inner planar zone of laminated sand is equivalent to that of Clifton et al.'s i 1971 ) but is markedly wider, extending from 0 to — 1 m. 200 m from the beach. An inner rough facies (1-2 m depth) is equivalent to that of Clifton et al.'s, but is expressed as rippled, laminated sand, rather than megaripples. An outer planar facies (5-10 m depth) consists of laminae and thin beds with sharp, erosional lower contacts, grading upward into bioturbate texture. Howard and Reineck (1972) suggest deposition during storm intervals, alternating with pe- riods of fair weather and bioturbation. An upper shore- face facies of fully bioturbated, muddy fine sand has no parallel in Clifton's study of the high-energy Oregon coast, and is a consequence of the high input of fine sand and reduced wave energy. Seaward of 10 m, the muddy, gently sloping shoreface becomes markedly coarser, then gives way to a flatter seafloor of medium to coarse sand, characterized by heart urchin bioturbation and trough cross-stratification. Clifton and co-workers did not extend their study suffi- ciently far seaward to detect such a coarse lower shore- face and seafloor facies However, an equivalent facies does appear on retreating coasts of both North Carolina and Holland (Fig. 8). Reineck and Singh (1971) have described shoreface and inner shelf deposits from the low wave energy, high mud input, prograding coast of the Gulf of Gaeta, Italy. The inner facies are rather similar to those of the Georgia coast. Ripple bedding is the main sedimentary structure out to 2 m. Below 2 m, laminated bedding becomes the main structure, and bioturbation becomes prominent, increasing seaward. Laminae are inferred to be deposited from graded suspensions after storms. At 6 m, sand gives way to silty mud, heavily bioturbated by Echnocardium cordatum. There is no equivalent of the coarse offshore facies of retreating coasts. LONGSHORE SEDIMENT TRANSPORT The seasonal cycle of onshore and offshore sand migra- tion in the surf is superimposed on a much more intensive flux of sand parallel to the beach, under the impetus of the wave-driven littoral current. The mechanisms driving 484 BB-N BB-S CEC-2 CAB CEC-7 STRUCTURES Laminated Sand Small Scale Ripples Meganpples Bioturboied Sand Sand|lt„„culo, B.dd.ng and \ f'n« Interbedd.ng Mud Icoone Interbeddii Mud Shells I- 27 28 29 30 £ FIGURE 13. Primary structures of the Georgia shorejace, as their traversing the north flank of an estuary mouth shoal, revealed by box cores. Complexity oj lines N and S are due to From Howard and Reineck (1972). 272 485 LONGSHORE SEDIMENT TRANSPORT 273 this littoral drift and equations for determining its trans- port rate are described in Chapter 13. The propensity of this coastwise sand flux to aggrade or erode the shoreline can be understood by reference to a convenient graphical model presented by May and Tanner ( 1973). As a consequence of refraction of waves about a coastal headland, such a headland will tend to concentrate the wave rays on it and hence wave energy (see p. 7b). As a consequence, the wave energy density (proportional to the spacing ol wave rays) decreases steadily from point a on the headland to point e in the bay. These relationships are shown in highly schematic fashion in Fig. 14. The longshore component of wave power Pi. is a function of both the wave energy density and the breaker angle (see Equation 2, Chapter 13). It must therefore pass through a maximum between the point of greatest wave energy density a, and the point of greatest breaker ® © FIGURE 14. Model for littoral sediment transport. (A) Wave refraction pattern, with wave approach normal to coast. (B) Re- sulting curves for energy density at the breaker E (dimensions AIT-2)/ longshore component of littoral wave power PL {dimen- sions MLT~3); and the littoral discharge gradient dq/dx (dimen- sions L2T~l). (C) Advanced state of coastal evolution. After May and Tanner (1973). angle at c. The maximum is attained, however, closer to point c, as the gradient of wave energy density on this subdued model coast is relatively flat. Both the sand transport rate //. (see p. 247) and the discharge of sand (q) are proportional to Pi and vary with it. Therefore, the longshore discharge gradient dq/ d.v varies as the derivative of Pi (Fig. 14). The sediment continuity equation (p. 190) states that the time rate of change of seafloor elevation along a streamline in the littoral current is proportional to the littoral discharge gradient under conditions of steady flow. In other words, if more sand is moving into a given section of shoreface than is moving out (negative dq/dx), then the seafloor of that section must aggrade. If, on the other hand, more sand is being exported than imported (positive dq/dx), the seafloor of that sector must erode (see the discussion on p. 190). In general, erosion occurs along a positive discharge gradient, and deposition occurs along a nega- tive discharge gradient. In the model of Fig. 14, this relationship means that the shoulder of the headland, from a to c, should erode, with the material being transported into the bay, to fill sections c through e. The same sort of process should occur on the other side of the bay (not shown) and the other side of the headland (not shown). If the direction of wave approach were held constant and normal to the regional trend of the coast, then a very peculiar coast- line should eventually result. It would straighten out to a nearly east-west line running through c, but with a needlelike projection at a, the point of littoral drift divergence, and a similarly narrow indentation at e, the point of littoral drift convergence. On real coasts, how- ever, the direction of wave approach is not constant, but fluctuates about the mean value, with changes occurring on a scale of hours to days. As the direction of wave ap- proach fluctuates, so do the positions of points a and e, and the development of coastal reentrants and projections is suppressed. If waves tend to approach at an angle instead of ap- proaching normal to shore, and if coastal relief is more deeply embayed (Fig. 15), then a rather different dis- tribution of longshore wave power will result. The locus of maximum deposition will be shifted from the bay head toward the tip of the adjacent headland where the gradi- ents of wave energy density and breaker angle are the steepest. During a storm when the intensity of littoral drift discharge is the greatest, deposition at this point may be so intense that a discontinuity in the shoreface may occur, in the form of a spit that builds out across the bay as an extension of the headland shoreface. As the shoreline matures, headland retreat, spit extension, and bay head beach progradation occur simultaneously and in this model also, the final coastline is again straight. 486 274 COASTAL SEDIMENTATION FIGURE 15. Variant of the littoral transport model with a more deeply embayed coast and an oblique direction of wave approach. Conventions as in Fig. 14. Thus, as a consequence of the submarine refraction o waves about the shoals off headlands, the shoreface tends toward an equilibrium plan view as well as an equilib- rium profile. Headlands tend to be suppressed and bays filled, because their existence leads to longshore wave power gradients that transfer sand from headland to bay head. A similar smoothing process operates at deeper levels on the shoreface, where sand transport occurs in response to tide- and wind-driven currents. Such flows accelerate past the projecting headlands which impede them, and expand and decelerate off the bays. The result is a posi- tive discharge gradient on the upcurrent sides of the headlands, and a negative discharge gradient on the downcurrent side. This pattern reverses when the cur- rents themselves reverse, so that the lower shoreface off headlands experiences net erosion, while the lower shore- face off bays experiences aggradation. Thus the equilib- rium plan view of a coast tends to be straight, to the extent that variations in the homogeneity of the sub- strate and the rate of sand supply to the surf zone will permit. A second characteristic of equilibrium coastal con- figuration is the adjustment of the trend of the coast to the angle of wave approach as mediated by the rates and locations that sand is put into and taken out of a littoral drift cell. A perfectly straight and infinitely long coast could ideally maintain any angle to wave ap- proach, if the only source of sand were its own shoreface erosion. In fact, however, such an ideal straight coast is rarely attained. Sea level is rising or has been until very recently, and the straightening process must operate con- tinuously as successive portions of the irregular subaerial surface are inundated. Depending on the degree of in- duration of the coast, an effective equilibrium is attained with less than the "climax" degree of coastal straight- ness; some irregularity usually persists, with at least sub- dued headlands serving as sand sources, and embayments serving as sand sinks. Locally, river mouths may serve as point sources of sand with high sand input rates. These result in deltas and an effective coastal equilib- rium at less than climax straightness. Komar (Chapter 13, Figs. 12, 13) has provided two ex- amples of the adjustment of coastal trend to the angle of wave approach, and to the location of sources and sinks. His river mouth (Fig. 12) injects sand at a point in the littoral drift system at a rate greater than the system can initially accommodate, and the coastline progrades. As it does so, orientation of the coast at each point adjusts so that the river mouth protrudes as a delta. Eventually, an equilibrium configuration is attained so that each point along the shoreline maintains an incident wave angle sufficient to bypass the same amount of sand at every other point. Komar's beach (Fig. 13) has no point source of sand. He starts with a straight beach and a landward-convex waveform, a form that might arise from offshore refrac- tion over the rocky headlands enclosing a pocket beach. The center of the beach becomes a source and the ends become sinks; the shoreface adjusts to the wave refrac- tion profile. Komar's two examples correspond to two basic cate- gories of coastlines. In the swash alignment (Davies, 1973, p. 123), each point on the coast tends to be oriented normal to the direction of wave approach, either as an initial condition or because the configuration of the shoreface and the wave refraction pattern have inter- acted until this is the case. Such an adjustment is only possible if there is a projecting headland in the down- drift direction as in the case of Komar's beach, or another reason to cause a sand sink and allow coastal progradation. Drift alignments (Davies, 1973, p. 123) are more nearly like Komar's delta model. These coasts are 487 TRANSLATION OF THE SHOREFACE 275 □ " ecent sediments | Olde r bedrock FIGURE 1 6. Zetajorm bays on the offset coast of eastern Malaya. Major wave approach direction is from the northeast. From Davies (2973). stabilized by competition between two opposing trends for control of the littoral transport system. As a coastal compartment becomes more nearly normal to wave or- thogonals, littoral energy density increases; however, the longshore component of wave energy decreases. Maxi- mum discharge tends to occur when the orthogonals of prevailing wave trains make an angle of 40 to 50° with the coast. A coast captured in this alignment will tend to be stable, as it is the alignment of maximum transport. Coasts with closely spaced barriers to littoral drift, in the form of river mouths or projecting headlands, may form offset coasts consisting of successive zetaform bays (Halligan, 1906). These fishhook beaches have a curved swash-aligned beach in the shadow zone immediately downdrift from the barrier, and a straight drift-aligned beach extending downdrift from the swash segment to the next headland (Fig. 16). If the variance of the direction of wave approach is high, then the shadow zone behind the barrier will be exposed to the direct approach of waves for a significant part of the time and will intermittently operate as a drift-aligned beach of reverse drift. The apex of the barrier will become a zone of net drift convergence, and hence a self-sustaining constructional feature. The dynamics of zetaform bays have been discussed in detail, with summaries of earlier observations, by Silvester (1974, pp. 71-90). TRANSLATION OF THE SHOREFACE The preceding evaluation of the alongshore and on- shore-offshore components of sediment transport per- mits us to take a more general look at the coastal sedi- ment budget and its effect on the shoreface stability. It is clear from the preceding discussion that the shoreface profile will translate either landward or sea- ward (the coast will retreat or prograde) depending on whether the effects of the fair-weather regime, which tends to aggrade the shoreface, or the storm (or tidal) regime, which tends to erode it, are dominant. The sense of coastal profile translation further depends on coastwise gradient of sand discharge (whether sand im- ports and exports for the sector under consideration sum to a surplus or deficit). As a consequence of the onshore- offshore cycle of sand exchange, the nature of coastal translation will finally depend on which shoreface prov- ince, if any, is actually subjected to a sand surplus. The coastal transport system may be visualized as two coast- parallel pipes, corresponding to the wave-driven littoral drift near the beach and the intermittent storm- or tide- driven sand flux that occurs en the shoreface and inner shelf seaward of the breaker. These two pipes are con- nected by valves, corresponding to the onshore-offshore cycle of sand exc iange. The factors listed above deter- mine what vah s are open, for how long, and the net sense of flow through the valves. We do not have the measurements of onshore-offshore sand transport that would allow us to document the manner in which this system actually works. Until we do, we must be satisfied with an exploration of possible limiting cases, by means of deductive reasoning (Fig. 17). Four basic cases may be distinguished. On modern coasts, undergoing relatively rapid sea-level rise, the gradient of littoral drift discharge (dq/dx) is either posi- tive or is so slightly negative that the resulting sand surplus is not sufficient to balance offshore transport 488 276 COASTAL SEDIMENTATION A TRANSGRESSION ( RISING SEALEVEL ) A UPPER vj ' ' 1 SHOREFACE ^^.V ^~"SJ5^7'-> LOWER I — I SHOREFACE ' ' INNER SHEEF ^^i DEPOSITIONAL REGRESSION RISING SEALEVEL ) \\ EROSIONAL REGRESSION ( FALLING SEALEVEL SOUR TRANSPORT SYSTEM MODERN DEPOSITS FIGURE 17. Schematic models for the shore face sand budget. (A) Retrograding coastal sector with rising sea level and balance or deficit in coastwise sand flux. (B) Near- stillstand coastal sector with effect of rising sea level compensated by sand surplus associated with coastwise sand flux. (C) Prograding coastal sector with effect of rising sea level and reversed by sand surplus associated with coastwise sand flux. (D) Prograding coastal sector with falling sea level and balance in or sand deficit resulting from coastwise sand throughput. during storms. Under these conditions Bruun coastal retreat must prevail (Fig. 17.4). Storm erosion must pre- dominate over fair-weather aggradation on the shore- face as it translates landward and upward in response to sea-level rise. During fair weather, sand may be tem- porarily stored on the upper shoreface and beach (see Fig. 4A). The barrier superstructure becomes a long- term reservoir, receiving sand from the eroding shore- face by storm washover, storing it, and finally releasing it to the eroding shoreface. The inner shelf floor tends to become a sand sink, retaining the coarser sand trans- mitted to it by seaward bottom flow during storms, and releasing the finer fraction to the coastwise shelf flows. The coast of northern New Jersey appears to be under- going such a retreat (Stahl et al., 1974). Locally, however, the sand discharge gradient associ- ated with the deeper storm-driven shelf flows may be steeply negative (Fig. \7B), resulting in a sand surplus. The surplus tends to be absorbed by inner shelf and lower shoreface aggradation and also by storm washover on the barrier. The upper shoreface, however, is rela- tively unaffected. The trajectory of the shoreface profile appears to be nearly parallel to the upper shoreface slope, resulting in stillstand of the shoreline. The barrier system of coastal North Carolina immediately north of Cape Hatteras appears to be undergoing such a depositional stillstand, resulting in the opening of an anomalously wide lagoon behind it as sea level con- tinues to rise (Swift, 1975). If the littoral drift transport system has a strongly negative discharge gradient, as is the case downdrift from river mouths with a high sand discharge, the satu- ration of the upper shoreface with sand causes the flooding of all other inner shelf provinces as well, and the shore- face progrades by the successive capture of upper shore- face bars as beach ridges (Figs. 105 and 1 1C). The analy- sis of the Holocene history of the Costa de Nayarit by Curray et al. (1969) provides an excellent case history of such a coast. Finally, we must consider the case of falling sea level. If, under such conditions, the net littoral drift input is negligible, not all portions of the shoreface will prograde. The shoreface must translate seaward down the gradient of the shelf in a reversal of the Bruun process. Under these conditions, successive beach-upper shoreface sand prisms undergo subaerial capture, and the lower shore- face and inner shelf undergo erosion as sea level drops (Fig. \1D). However, should the sand surplus due to 489 TRANSLATION OF THE SHOREFACE 277 littoral throughput increase, the sand budget must ap- proach that of the subsiding prograding coast and a more rapid shoreline regression must result. Modern examples of such falling sea-level budgets are confined to regions of glacial rebound or tectonic uplift, but the surfaces of the world's coastal plains were molded by it during the withdrawal of the Sangamon (Riss-Wiirm) Sea (Oaks and Coch, 1963; Colquhoun, 1969). A word on the factors controlling the steepness and curvature of the shoreface profile is in order at this point. Grain size is the most obvious control (Bascom, 1951); the coarser the sediment supplied to the coast, the steeper are the shoreface profiles. Shorefaces built of shingle may attain 30° slopes near the beach; shorefaces of sand are rarely more than 10° at their steepest, while shorefaces on muddy coasts are so flat as to be virtually indistinguishable from the inner shelf. Sediment input and the wave climate also affect the shape of the profile. In general, inner shelves experiencing a higher influx of sediment and a lower wave energy flux per unit area of the bottom are flatter, whereas inner shelves with a lower influx of sediment and a higher wave energy flux per unit area of the bottom are steeper (Wright and Coleman, 1972). Because of the complex interdependence of the process variables, cause and effect are difhcult to ascertain; on a steeper shelf, for instance, grain size is coarser because the steeper slope results in more energy being released per unit area of the bottom; more energy is released because the coarser grain size results in a higher effective angle of repose. Or a reduced input of sand will allow the profile to attain the maximum steep- ness permissible under the prevailing wave climate, with a resultant higher rate of energy expenditure on the shoreface, and a consequent coarsening of its surface (Langford-Smith and Thorn, 1969). The relationship between the rate of sea-level dis- placement and the shape of the profile requires some thought. A number of workers have assumed that rap- idly translating coasts are in a state of disequilibrium, and that equilibrium can only be realized on very slowly translating or stillstand coasts. This view results from an inadequate appreciation of the equilibrium con- cept and is tantamount to stating that only chemical reactions that have gone to completion are equilibrium reactions. It is important to clearly distinguish between the con- cept of coastal maturity on one hand, and the concept of coastal equilibrium on the other. Davis (in Johnson, 1919) has assembled a spectrum of coastal types that suggest that the coastal profile passes through stages of '"youth, maturity, and old age" in which the profile be- comes increasingly flatter, until a final profile of static equilibrium is reached — ultimate wave base, in which the continental platform has been shaved off to a level below which further marine erosion occurs so slowly as to be negligible. The scheme is unrealistic in that it fails to recognize the continuous nature and mutual depend- ence of the process variables of an equilibrium system. Some of these stages will occur as transient states after the sudden rejuvenation of a tectonic coast. But as the profile becomes increasingly mature, its rate of change decreases, until it attains the equilibrium configuration required by existing rates of such other process variables as sediment input and eustatic sea-level change. At this point the profile must continue to translate according to the Bruun (1962) model of parallel shoreface retreat, until the rate of one or another variable changes again. This equilibrium, of course, is only apparent if the coastal profile is examined over a sufficiently long period of time — on the order of decades. Shorter periods of obser- vation will resolve the apparent "equilibrium" into a series of partial adjustments to periods of fair weather and periods of storms. Only in cases of relatively rapid tectonism may hys- teresis, or lagged response, occur, and strictly speaking, the term "disequilibrium" should be applied only to such cases. Slower changes in a process variable will allow continuous and compensating adjustment of pro- file, and while its shape changes, the profile is at all times in equilibrium. Coastal disequilibrium tends to be more apparent on rocky coasts, because of the greater response time of the indurated substrate, and because such coasts are more likely to be subject to tectonism. Consequently, the effect of the rate of sea-level dis- placement in the equilibrium profile must depend on the initial slope of the substrate. On low coasts, where the initial slope is flatter than the maximum potential slope of the equilibrium profile, then the more rapid the sea- level displacement, the flatter is the resulting equilibrium profile (e.g., see Van Straaten, 1965). This relationship may be viewed as a function of work done on a substrate to build the optimum shoreface. As a coast advances more rapidly, successive shorelines experience the ero- sive effect of shoaling waves for shorter periods of time and the resulting profile is flat (immature). If, however, a coast undergoes stillstand, the climax or fully mature configuration can develop, which is the steepest profile possible for the available grain size of sand, rate of sediment influx, and hydraulic climate. On high, rocky coasts, however, the initial slope of the substrate may be steeper than the mean, or even the maximum slope of the steepest profile permitted by these variables. Under such circumstances, the more rapidly transiting shorelines, since these have the least work done on them, have the least modified and hence steepest (most immature) profiles, while the most slowly moving 490 278 COASTAL SEDIMENTATION shorelines are the most modified and hence flattest profiles. As noted above, existing measurements of the coastal hydraulic climate and resulting sand transport are gen- erally inadequate to define the coastal sand budget. It is possible, however, to extend the inferential models presented in Fig. 1 7 so as to take into account the effect of these variables on shoreface slope and curvature (Fig. 18). COASTAL ENVIRONMENTS The preceding sections of this chapter have described the onshore-offshore component of sediment transport, and also the flow of sediment parallel to the coast. Modes of shoreface displacement in response to rising and fall- ing sea level have been considered. These insights are prerequisites to an examination of specific patterns of coastal sedimentation. But before we proceed to such an EROSIONAL TRANSGRESSION (PROFILE INVARIANT) B DEPOSITIONAL TRANSGRESSION (PROFILE INVARIANT) STILLSTAND (PROFILE CURVATURE INCREASING) DEPOSITIONAL REGRESSION RISING SEA LEVEL (PROFILE INVARIANT) 3 EROSIONAL REGRESSION (PROFILE INVARIANT) DEPOSITIONAL REGRESSION FALLING SEA LEVEL (PROFILE CURVATURE DECREASING) ;.;.;.:•; ZONE OF DEPOSITION ZONE OF EROSION FIGURE 18. Modes of shore/ace translation as a function of curvature. Envelopes of .erosion and aggradation are shown. (1) direction of profile translation and (2) change in profile Terms from Curray (1964). 491 COASTAL ENVIRONMENTS 279 BEACH DOMINATED COASTS DRIFT COASTS SWASH COASTS U_/uJ R5SP INLET & RIVERINE COASTS ROCKY COASTS INE If ESTUARINE COASTS *«* ** INLET & RIVERINE COASTSX DELTAIC COASTS LOBATE DELTAS ft A DELTAIC /COASTSX & ^ RECESSED DELTAS ESTUARINE DELTAS FIGURE 19- Descriptive taxonomy of coasts analysis, we should consider a scheme for classifying the coastal settings in which the transport patterns occur. A study of maps of the world's coastlines suggests that the apparently unlimited variety of coastal configurations falls into a relatively small number of repeating patterns. Considerable thought has gone into coastal classifica- tion, and the reader is referred to the excellent summary of existing classifications presented by C. A. M. King (1972; also Chapter 1 5) .The operational classification that is used in this text is presented in Table 2 and Fig. 19. The most basic practicable division appears to be into coasts with substrates of crystalline or lithified sedimen- tary rock versus coasts bordering coastal plains, with substrates of unlithified sediment. Both lithified and un- lithified coasts may adjust their configurations in re- sponse to the coastal wave climates but they do so at different rates, and in response to somewhat different mechanisms. Patterns of sedimentation may be relatively simple on straight rocky coasts, but coasts of structured rock may be so deeply embayed as to greatly complicate the pattern (crenulate rocky coasts). Unconsolidated coasts are floored by easily eroded and flat-lying strata, and the surficial sediment tends to be both abundant in quantity and continuous in extent. On such coasts, wave-driven currents in the littoral zone and wind- or tide-driven currents farther offshore tend to build straight coastal segments of the sediment avail- able to them. A basic second-order division in %he mor- phology of unconsolidated coasts depends on the relative importance of straight coastal segments versus inlets that alternate with them. TABLE 2. A Coastal Taxonomy Criterion Coastal Type Substrate indurated Coast-parallel anisotropy Coast-transverse anisotropy Rocky coasts Straight rocky coasts Crenulate rocky coasts Substrate unconsolidated Littoral drift dominant Low drift angle High drift angle Transverse drainage dominant Fluvial drainage dominant Mild wave climate Moderate wave climate Moderate-strong wave climate Strong wave climate Tide modified Tidal drainage dominant Wave-modified tidal drainage Unconsolidated coasts Beach and barrier coasts Drift coasts (straight) Swash coasts (cuspate) Inlet and riverine coasts Deltaic coasts Lobate deltas Arcuate deltas Cuspate deltas Recessed deltas Estuarine deltas Estuarine coasts Inlet coasts 492 280 COASTAL SEDIMENTATION Inlets occur at river mouths, where they may be main- tained by river and tidal flow or by purely tidal flow, where there is a tidal exchange between lagoons and the sea. A dense subaerial drainage net or a high tide range may cause inlets with their coast-normal flow to occupy over 509c OI tne shoreline, resulting in deltaic, estuarine, or tidal inlet coasts. Open coasts with rigorous wave climates and frequent strong wind-driven currents tend to have fewer inlets than coasts not so affected, resulting in mainland beach-barrier beach coasts. This chapter has so far dealt mainly with the sedi- mentary regime of such simple, two-dimensional coasts. The succeeding sections examine in greater detail the modes of sand storage on beach-barrier island coasts, and also the modes of sediment storage on more complex coasts. The following chapter on shelf sedimentation stresses the role of varying coastal configurations in by- passing sediment to the continental shelf, and thus modu- lating the shelf sedimentary regime. SAND STORAGE IN THE SHOREFACE Storage in Low Retreating Shorefaces: Barrier Spits and Islands barrier formation. On most retreating coasts, the most important form of sand storage is within the shore- face itself, in the form of barrier spits and barrier islands. It would seem that along many coastal sectors, the coastal sedimentary regime rejects the primary shoreline formed by the intersection of the subaerial continental surface with the sea surface, and instead builds a secondary "barrier" shoreline seaward of the primary one. A char- acteristic of the equilibrium shoreface surface that as much as any mechanism is the basic "cause" of barrier islands and spits is its innate tendency toward two- dimensionality, its tendency to be defined by a series of nearly identical profiles in the downdrift direction. The equilibrium shoreface does not "want" a lateral bound- ary, since the wave and current field to which it responds does not generally have one. The initial conditions dur- ing a period of coastal sedimentation may, however, in- clude such discontinuities, as in the case of a coast of appreciable relief (bay-headland coast) beginning trans- gression. On such a coast shoreface surfaces will tend to be in- cised into the seaward margins of headlands exposed to oceanic waves, and will propagate by constructional means in the downdrift direction as long as material is available with which to build, and a foundation is avail- able to build on. The basic mechanism is that described by May and Tanner (1973); see Fig. 14. Where the shoreline curves landward into a bay, the longshore component of littoral wave power decreases, and the alongshore gradient of sediment discharge (dq/dx) is negative. The shoreface at that point must aggrade until the gradient approaches zero at that point, and the zone of negative gradient has moved downdrift. We give the lateral propagation of the shoreface into coastal voids the descriptive term "spit building by coastwise progradation" (Gilbert, 1890; Fisher, 1968). However, the tendency of the shoreface to maintain lateral continuity also acts to prevent discontinuities as well as to seal them off after they have formed. In order to illustrate this, we may consider another set of initial conditions — a low coastal plain with wide, shallow val- leys after a prolonged stillstand during which processes of coastal straightening by headland truncation and spit building have gone to completion. As this coastline sub- merges, the water, seeking its own level, will invade valleys more rapidly than headlands can be cut back. The oceanic shoreline, however, cannot follow, for if it should start to bulge into the flooding stream valleys, the bulge would become a zone of negative discharge gradient; hence the rate of sedimentation would increase to compensate for any incipient bulge. The shoreface would translate more nearly vertically than landward at this sector, until continuity along the coast was restored (Fig. ME). Thus a straight or nearly straight oceanic shoreline must detach from an irregular inner shoreline, and be separated from it by a lagoon of varying width. This process of mainland beach detachment was first proposed by McGee (1890), and later described in de- tail by Hoyt (1967); see Fig. 20. COASTWISE SPIT PROGRADATION VERSUS MAINLAND beach detachment. Much of the debate concerning origin of barriers deals with the relative importance of spit building versus mainland beach detachment (Fisher, 1968; Hoyt, 1967, 1970; Otvos, 1970a,b); see Chapter 12 (p. 223). The problem can be fully answered only by careful study of the field evidence, and as noted by sev- eral authors (Otvos, 1970a,b; Pierce and Colquhoun, at 15 £ Ui I3oS 12 3 4 KILOMETERS FIGURE 20. Barrier island formation by mainland beach detachment. Modified from Hoyt (1967). 493 SAND STORAGE IN THE SHOREFACE 281 1970) the evidence has frequently been destroyed by landward migration of the barriers. However, it is pos- sible, in the time-honored deductive fashion of coastal morphologists, to consider the conditions most favorable to these two modes of barrier formation. Spits are cer- tainly characteristic of coasts ofhigh relief undergoing rapid transgression as described above [see the papers in Schwartz (1973)]. It seems probable that under such conditions mainly beach detachment would be severely inhibited. Even allowing for ideal initial conditions with a classic coast of old age (Fig. 21), where alluvial fans are flush with truncated headlands, detached mainland beaches would have a limited capability for survival. With significant relief, the submarine valley floors ad- jacent to retreating headlands must lie in increasingly deeper water after the onset of transgression. As the barrier grows into the bay, its submarine surface area must increase, and the capacity of littoral drift to nourish it may eventually be exceeded. As this point is approached, the combination of storm washover and shoreface erosion will cause the barrier to retreat until equilibrium is restored, a position which may be well inland from the tips of headlands. Both littoral wave power and sediment supply may be deficient in these inland positions, further jeopardizing the survival of A. STILLSTAND B. BEACH DETACHMENT C. CYCLIC SPIT PROGRADATION FIGURE 21. Barrier formation with spit-building dominant. As a rugged coast passes from stillstand to transgression, a mature configuration is replaced by a transient state of mainland beach detachment, then by a quasi-steady state regime of cyclic spit building. This diagram also illustrates the relationship between the concepts of coastal equilibrium and coastal climax, since it consists of Johnson's (1919) stages of coastal maturity — portrayed in reverse sequence! the barrier. As the loop of the barrier into the bay becomes extreme, sediment supply from headlands is liable to capture by secondary spits formed during storms. These may prograde out toward the drowned valley thalweg until capacity is again exceeded and their tips are stabilized, further movement being limited to retreat coupled with that of the headland to which they are attached. Finally the survival of primary barriers on such a coast would be limited by the tendency of submerging headlands to form islands. A spit tied to a promontory that becomes an island can retreat no further if a drowned tributary valley lies landward of it, but must instead be overstepped. The few unequivocal examples of trans- gressed barriers on the shelf floor appear to be over- stepped, rock-tied spits (Nevesskii, 1969; McMaster and Garrison, 1967). On the other hand, transgression of a coast of very subdued relief, such as is the case for most coastal plains, would tend to promote mainland beach detachment at the expense of spit formation, given initial conditions of a straight coast (Fig. 22). The depth of water in which detached bay mouth barriers would be built would be less, because the relief would be less. The upper, ero- sional zone of the shoreface (Fig. 10/1) would be more likely to extend down into the pre-Recent substrate (Fig. 23^4); hence erosion of the inner shelf floor would become as important a source of sand for the barrier as the erosion of adjacent headlands. With a rise in sea level, valley-front dune lines would grow upward. River mouths, initially deltaic, would flood as estuaries, while lagoons would creep behind the beaches toward the headlands on either side. Barriers would retreat in cyclic, tank-trend fashion by means of storm washover, burial, and reemergence of the buried sand at the shoreface (Fig. 10.4). Coastal discontinuities sufficient to induce coastwise spit progradation would occur only locally. Thus, on a low, initially straight coast, barrier spits and barrier islands would preferentially form by mainland beach detachment rather than by coastwise progra- dation. Storage in Prograding Shorefaces The preceding discussion has identified barrier islands and spits as forms of sand storage on retreating coast- lines. On prograding coastlines, sand storage occurs in beach ridges and cheniers; the two forms differ in that beach ridges are separated by sand flats, whereas che- niers are separated by, and rest on, mud deposits. Sequences of beach ridges 15 to 200 m apart may form subaerial strand plains tens of kilometers wide. These are smaller scale features than the barriers, which 494 282 COASTAL SEDIMENTATION A STIUSTAND B. BEACH DETACHMENT C. BARRIER RETREAT FIGURE 22. Barrier formation with mainland beach detachment as the dominant process. A mature low coast passes via main land beach detach- ment into a steady state regime oj barrier retreat. characterize retreating and stillstand coasts; and bar- riers may, in fact, be locally comprised of beach ridge fields, as a consequence of minor frontal progradation or more extensive distal, coast-parallel migration (Hoyt and Henry, 1967). Curray et al. (1969) have presented a detailed study of what has been recognized as a classic strand plain coast, the Costa de Nayarit (Fig. 24). They postulate that each ridge forms as a plunge point bar, which in the presence of an oversupply of littoral sand, builds up close to mean low water. During a period of constant low swells, the bar may grow above this level as tides rise to the spring tide value (0.98-1.25 m); the bar becomes a subaerial feature during the subse- quent neap phase, and continues to grow by eolian activity (Fig. 105). Chenier plains form on coasts with a high suspended sediment input. In the classic chenier plain of the Louisiana coast west of the Mississippi Delta, the sand ridges support stands of live oaks (French, chene), hence the name (Price, 1955). The formation of chenier plains has been ascribed to rapid progradation of mud flats during periods of high suspended sediment discharge from nearby rivers or delta distributaries. When distribu- taries crevasse and the subdeltas are abandoned, the 495 SAND STORAGE OFF CAPES 283 rrr^TTT777T7 LAGOONAL DEPOSITS FIGURE 2 3. Contrasting sand budgets of a barrier built (A) on a gentle submarine gradient as in Fig. 22, and (B) on a steep submarine gradient as in Fig. 21. In (A), zone of shoreface erosion penetrates to Pre-Recent substrate, which becomes "income" for barrier nourishment. In (B) the barrier may only "borrow" from its own "capital" through shoreface erosion, and the heavy ex- penditure involved in paving the shelf with sand during barrier retreat may "bankrupt" the retreating barrier, which must either accelerate its retreat or be overstepped. In either case shoreface continuity is liable to be broken, resulting in cyclic spit building. downdrift coast becomes sediment starved. The mud flat erodes back and a beach ridge formed of the coarse sediment is thrown up by storm high tide. Todd (1968) has stressed the role of estuary mouths and inlets in the localization of chenier plains. He notes that littoral cur- rents updrift of the inlet tend to decelerate during ebb tide because of a reduction of the coastwise pressure gradient by the ebb jet, resulting in sediment deposition. Littoral currents in the same locality are accelerated by proximity of the inlet during flood tide, but fine sediment deposited requires a greater velocity for erosion than for deposition, and in any case has already compacted. Hence chenier plains tend to be localized on the muddy, updrift sides of tidal inlets. Downdrift of the inlet, the coast may instead be starved for fines as a result of sea- ward transport or "dynamic diversion" of the littoral current by the ebb jet, and the littoral sand deposits have more nearly the character of a beach ridge sequence. Otvos (1969) recognizes the role of inlets in localizing deposition, but notes that chenier deposition goes on for long distances beyond inlets. He cites new chronologic evidence from the Mississippi chenier plains to indicate that chenier ridge formation cannot be closely correlated with the abandonment of a subdelta mouth, and sug- gests that the intermittent shielding effect of nearby sub- delta growth on the wave climate plays a greater role in cyclic chenier plain growth. Beall (1968) has examined in detail sediment distribu- tion and stratigraphy in the present shoreface of the western Louisiana shoreline (Fig. 25). He distinguishes between three main stratigraphic patterns. Mudflats are defined on their seaward margins by a break point bar zone of very fine sand. Midtidal and upper tidal flats are distinguished by progressively finer sand and in- creasing percentages of silt and clay. A thin sand storm beach may rest on eroded marsh sediments. The strati- graphy is complex. Apparently a period of increasing littoral sediment discharge results in progressive flatten- ing of a shoreface, until the bar zone is triggered and becomes the maximum locus of sedimentation, prograd- ing both landward and shoreward. A (submarine) mud flat zone is thereby initiated in the sheltered longshore trough, and progrades toward the bar and landward. Transitional beaches have largely erosional profiles, with thin bar, beach, and washover sands overlying the ero- sional surface near the high-water line. The sequence is typical of that of erosional transgression, where the thin sand cap is a transient fair-weather veneer. However, Beall interprets these transitional beaches as prograda- tional, with rates of progradation intermediate between those of mud flats and those of "normal beaches." Normal beaches consist of up to 1 .7 m of seaward-fining fine sand, prograding seaward over an outer shoreface facies. Washover fans of normal beaches are thicker than those of transitional beaches. The three types of beaches described by Beall would appear to illustrate a temporal as well as a spatial sequence. Periods of rapid mud flat progradation are presumably followed by erosion, then the formation of transitional beaches, which prograde to become cheniers, then prograde more rapidly as mud flats. SAND STORAGE OFF CAPES South of the Middle Atlantic Bight of North America, the generally southwest-trending coastline has been molded into a series of large-scale cuspate forelands (Fig. 26). They are the response of the shoreface regime to a mod- erate to intense wave climate and a high variance in the direction of wave approach (Swift and Sears, 1974). Storm waves approach from the northeast, as is the case in the Middle Atlantic Bight, but the coast is also exposed to waves from more distant storms in the southeastern Atlantic. As a result, the cuspate forelands have been self-maintaining features throughout the postglacial pe- riod of sea-level rise and erosional shoreface retreat. Each foreland apex is a zone of littoral drift convergence. 496 284 COASTAL SEDIMENTATION 106° I05°30' 23°N FIGURE 24. The strand plain of the Costa de Nayarit, showing beach ridge sequences. Rio de la Cahas meanders through interlocking spits, indicating reversal of drift directions. From Curray et al. (2969). The resulting surplus of sand at the apex creates a coastal shoal. The shoal in turn maintains a pattern of wave re- fraction that drives littoral drift convergence (Swift and Sears, 1974); see Fig. 27. The question arises as to how such a closely coupled feedback system begins. The answer is that in a sense, it does not matter. It is a truism that as process variables approach the instability threshold, any singularity in a water-substrate system will excite the feedback of the process and response that lends to the formation of bed forms. In the case of the cuspate Carolina coast, the initial conditions were probably the sequence of shelf- edge deltas during the Late Wisconsinan low stand, cor- responding to the Peedee, Cape Fear, Neuse, and Pam- 497 SAND STORACE IN INNER SHELF RIDGE FIELDS 285 TIDAL -MUDFLAT JTTTTTrUL. LOCALITY A HIGH TIDE- PROTECTED MUDFLAT FACIES OUTER SHOREFACE SEDIMENTS OFFSHORE GULF BOTTOM MUD ^JHnNER- SHOREFACE ^ BREAKER BARS WASHOyER FAN EROSIONA SURFACE 'TRANSITIONAL BEACH" — HIGH TIDE' irar LOCALITY B MUDFLAT SEDIMENTS ERODED SHOREFACE/ OUTER SHOREFACE OFFSHORE GULF BOTTOM MUD "NORMAL BEACH" LOCALITY "C" vWASHOVER FANS HIGH TIDE* OFFSHORE GULF BOTTOM MUD 100 FIGURE 2 5. (A) The chenier plan of southeastern Louisiana. (B) Characteristic beach configurations. See text for explanation. From Beall (1968). lico rivers (note river system of Fig. 26 and compare Fig. 27). The Appalachicola cuspate foreland of the Florida Panhandle is particularly suspect as having been formed by this mechanism (Swift, 1973). Other cape- associated shoals may occur as a consequence of the re- duction in intensity of littoral drift around a rock-de- fended cape with consequent reduction in the compe- tence of littoral drift (Tanner, 1961; Tanner et al., 1963). On offset coasts (forelands separated by zetaform bays; p. 275), the forelands may be triggered by cape extension shoals, either river mouths or rocky promon- tories (Davies, 1958). Forelands and cape-associated shoals also occur on swash-aligned coasts. These coasts tend to be inherently unstable, breaking into short "arcs 498 286 COASTAL SEDIMENTATION H SHOAL RETREAT MASSIF FIGURE 26. Cuspate coast of the Carolinas. Values for littoral drift are in yd/year X 10~3. From Langjelder et al. (1968). of equilibrium," terminating in cuspate forelands with neither rivers nor outcrops required for cusp formation. Tidal inlets may evolve into cuspate forelands with asso- ciated shoals. Chincoteague Shoals, on the Delmarva coast, is an example of a barrier-overlap inlet that has become a cuspate spit with associated shoal (Fig. 36). Shoals developing over cuspate forelands may extend seaward the width of the shelf. Such cape extension shoals do not result from the seaward transport of sand, but rather from the landward translation of the cuspate foreland in response to rising postglacial sea level, to- gether with the retreat of the associated littoral drift convergence. The seaward-trending shoal marks the re- treat path of this convergence. Its response to the shelf regime is discussed in the next chapter. SAND STORAGE IN INNER SHELF RIDGE FIELDS Storm-Induced, Shoreface-Connected Ridges A major category of inner shelf sand storage found on low retreating coasts is storage in shoreface-connected ridges (Fig. 28) and in associated inner shelf ridge fields. These features are up to 10 m high, 2 to 5 km apart, and their crestlines may extend for tens of kilometers. Side slopes are rarely more than a degree. They typically converge with the shoreface at angles of 25 to 35° (Duane et al., 1972) and may merge with it at depths as shoal as 3 m. The best known development is on the coast of the Middle Atlantic Bight of North America, but they may be found on coastal charts as far south on the Atlantic coast as Florida, and around the Gulf coast littoral as far as Alabama. They also appear locally on the Texas coast. Allersma (1972) has reported them on the muddy coast of Venezuela, where they are dom- inantly composed of mud. They have been detected by ERTS satellite imagery on the Mozambique coast (John McHone, personal communication), and also appear on the southern littoral of the North Sea. With the excep- tion of the Venezuelan coast, most settings are that of a low, unconsolidated coast undergoing Bruun erosional retreat (Fig. \0A) in response to a moderate to strong wave climate and periodic intense storm or tidal flows. Where best studied, on the Virginia-northern North Carolina coast (McHone, 1972), the ridges appear to have some of the response characteristics of wave-built bars at their inner ends where they merge with the shoreface. Like wave-built bars, their landward ends are asymmetrical, with steep landward flanks, although the 499 (a) FIGURE 27. Model for the transformation of a stillstand delta into a retreating cuspate foreland. From Swift and Sears (1974). (b) B FIGURE 28. (a) Shoref ace-connected ridges of the Delmarva from (a). Dots represent hypothetical fixed points during a inner shelf, contoured at 2 fathom intervals. From ESS A bathy- period of shoref ace retreat and downcoast ridge migration. See metric map 0807 N-57. Ridges are in varying stages of detachment. text for explanation. From Swift et al. (1974) ■ (b) Schematic diagram of detachment sequence as inferred 500 287 288 COASTAL SEDIMENTATION BASE LINE 200 DISTANCE FROM BASE LINE IN METERS 400 600 800 1000 1200 1400 SMOOTHED SEA FLOOR BASIC PROFILE) FIGURE 29. Superimposed profiles of the inner shorejace-connected ridge at False Cape, Virginia. From McHone (1972). reverse asymmetry tends to prevail further seaward. En- velopes of profiles indicate that, as in the case of their small-scale break-point counterparts, ridges built to a height of approximately one-third water depth, at which point wave agitation is sufficiently intense to preclude further growth. The troughs between the ridges and the shoreface are similarly excavated to one-third of water depth below the smoothed profile (McHone, 1972); see Fig. 29. At the False Cape Ridge Field, Virginia (McHone, 1972; Swift et al., in press), analysis of the wave climate suggests that waves are capable of breaking on some part of the inner ridge crest about 10% of the time. As a consequence of their oblique orientation and varying crestal depth, such ridges may utilize energy from a relatively broad spectrum of wavelengths. As wave-built bars, however, the low-angle ridges are anomalous. They are much larger than surf zone bars and their oblique orientation is more nearly parallel to the direction of wave approach than normal to it. The ridges may be primarily a response to a downwelling coastal jet that comprises the coastal margin of the storm flow field (see p. 275), although storm wave action is clearly a complementary mechanism. At False Cape, Virginia, a 28 hour current-meter station revealed a steady southward and offshore flow on the order of 15 cm/sec at a distance of 8 cm off the bottom, subsequent to the passage of a cold front with winds in excess of 25 knots (Fig. 30). During this period, however, the anchored observation vessel maintained a wake trending southward and shoreward. The inferred structure of the coastal flow field during the observation period is pre- FIGURE 30. Progressive vector diagram of storm bottom flow at the innermost ridge at False Cape, Virginia. Vectors represent velocities taken for 3 minute intervals every 30 minutes by two orthogonal Bendix Q-18 meters mounted in a plane parallel to the seafloor, 16 cm off the bottom. After passage of a cold front, bottom flow trended southeast obliquely seaward over ridge crest at velocities up to 18 cm/sec, while wake of anchored observation vessel streamed southeast, toward shore. Based on Holliday (1971). 501 SAND STORAGE IN INNER SHELF RIDGE FIELDS 289 SURFACE WIND DRIVEN CURRENT BOTTOM WIND DRIVEN CURRENT SURFACE WAVE DRIVEN FIGURE 31. Hypothetical structure of the coastal boundary of the storm flow field, based on Figs. 2 and 30. sented in Fig. 31. The observed pattern is interpreted as downwelling coastal flow intensified by constriction of the trough toward its southern end, and also by the setup of waves breaking on the inshore end of the trough. Mapping of the junction of this ridge with the shoreface on four successive occasions has revealed the presence of a shifting saddle, where storm flows presum- ably break out over the ridge base (McHone, 1972). A grain-size profile over the ridge is extremely asym- metric (Fig. 32). Sands are coarsest in the landward trough and become steadily finer up the landward flank, are of relatively constant grain size across the crest, and become finer again down the seaward flank. Sorting is variable on the landward flank and crest but increases steadily down the seaward flank. The profile is characteristic of a flow-transverse sand wave, and suggests that the ridges are responding as would a sand wave to the cross-shoal component of flow. As described in Chapter 10 (p. 166), bed shear stress increases up the upcurrent flank of a sand wave, attain- ing a maximum at the crest or just forward of it, then decreases down the downcurrent flank. Grain size would tend to decrease monotonically across such a shear stress maximum as a consequence of the progressive sorting mechanism; as sand is eroded out of the trough, the coarser grains are more likely to be trapped out in the initial portion of the transport path (Chapter 10, p. 162). On this particular ridge, however, size character- istics do undergo a reversal on the landward side of the crest, where maximum shear stress is to be expected. 502 290 COASTAL SEDIMENTATION o - i l3S 06 - 2 - DEV o z tat O = 05 - 4 - DIA ^__ ^Ak \ z o 6- BOTTOM PROFILE^^^ STANDARD o Co — 1 Uj > 8 - u 02 - 0 1 100 1 METERS 200 300 1 1 400 1 • 500 5 2 TIDAL TRANSPORT ► INTERTIDAL SUSPENSIVE TIDAL TRANSPORT ► V4 METERS FIGURE 40. Sedimentation patterns at the mouths of Georgia estuaries as inferred from Oertel (1972) and Oertel and Howard (1972). 510 297 Region I Region III Channel Processes Region IV Buoyant expansion, wave, wind, and tide induced mixing Velocity scale 0 1 2 J I L^J m/sec 6 8 10 Distance seaward. x/b„ Region I Region IV Weak buoyant expansion, wave wind and tide induced mixing 6 8 10 Distance seaward, x/bn FIGURE 41. (A) Cross section of density and flow during flood stage {Aprils, 1973). Both sections taken structure of the South Pass of the Mississippi River during flood tide. From Wright and Coleman (1974). during low river stage (October 25, 1969) and (B) 298 511 SAND STORAGE AT COASTAL INLETS 299 Despite the relatively rapid postglacial rise in sea level, some river mouths have been able to maintain equilib- rium channels in which cross-sectional area is adjusted to discharge, as deltas (prograding river mouths) or as equilibrium estuaries (slowly retrograding river mouths; see Figs. 42A,B). Most, however, have not. Disequilib- rium estuaries have resulted whenever aggradation of the estuary floor in millimeters per year has been less than the rate of sea-level rise, so that before any given segment of channel could close down to the required cross- sectional area, the main shoreline had passed it by. Such "drowned" or disequilibrium estuaries are gen- erally nearly funnel-shaped, rather than trumpet-shaped, as are the equilibrium forms. As a consequence of their higher ratio of saltwater to fluvial discharge, their river mouth shoals are retracted into the throat of the estuary and the interpenetration of ebb and flood channels be- comes marked (Fig. 42C). With a yet further increase of tidal over fluvial dis- charge such a coastal indentation may no longer be ap- propriately called an estuary, but simply a bay. Large bays experiencing high tidal ranges may build a tide mmm m A. CONSTRUCTIONAL CHANNEL RIVER DOMINATED FLOW B. CONSTRUCTIONAL CHANNEL TIDE DOMINATED FLOW C. PARTLY EROSIONAL CHANNEL TIDE DOMINATED FLOW INTERTIDAL SHOAL gig: SUBTIDAL SHOAL 1 FLOW DOMINANCE OF CHANNEL FIGURE 42. Varieties of river mouths. (A) Prograding delta distributary entering tideless sea (based on Mississippi River Delta). (B) Equilibrium {discbarge adjusted) tidal estuary mouth (based on Georgia coast estuaries). (C) Disequilibrium estuary mouth (based on Thames estuary). From Swift (1975b). 512 300 COASTAL SEDIMENTATION flat-tidal channel complex at their heads as a consequence of net landward sediment transport by the shoaling tidal wave. These deposits are the functional equivalent of the tide-molded deposits of a disequilibrium estuary. The patterns of sand storage in estuary mouths may be extremely elaborate. These dynamic topographies are of major concern to port authorities concerned with the maintenance of deep-water approaches. As in the case of the systems of open coasts, estuary mouth sand storage systems are in a state of continuous reorganization in re- sponse to the postglacial rise in sea level. Kraft et al. (1974) have attempted to trace the trans- gressive history of the mouth of Delaware Bay by equat- ing a series of transects across the modern bay with the time series of profiles that would be expected at a single point during transgression (Fig. 43). Here ridges first appear as subaqueous tidal levees on the edge of tidal flats marginal to tidal channels. Unlike the tidal sand ridges of open shelf seas, these ridges migrate away from their steep sides (Weil et al., 1974). As transgression proceeds, the channels service a larger and larger tidal prism and tend to widen. The effect on the levees is erosion on the steep, channel-facing side, and aggrada- tion on the gently sloping side facing away from the channel. Weil et al. (1974) have attributed the submarine levees of Delaware Bay tidal channels to density-driven secondary flow associated with the tidal cycle (Chapter 10, Fig. 26). Inner estuary channels tend to be ebb-dominated perhaps because the upper estuary water mass tends to flood as a sheet, but tends to preferentially ebb through the channel system under the impetus of gravity dis- charge. Further down the estuary, as levees begin to build, the interfluves tend to become flood-dominated channels in their own right, although the dominance of channel and interfluves may locally be reversed. As previously noted, retardation of the tidal wave in the estuary results in a phase lag across the estuary mouth shoal, causing an interdigitation of ebb and flood channels, separated by partition ridges, across the crest of the shoal. Thus ridges initiated in the upper estuary may undergo a complex evolution as successive estuary environments and associated flow regimes pass over them. Individual ridges may maintain their integrity through this process or be replaced by related forms maintained by somewhat different mechanisms. Modification of ridge morphology intensifies as the regional shoreline passes, and the lower estuarine regime is replaced by an open shelf regime. If the wave climate is intense, then the outer surface of the estuary mouth shoal is pushed back by erosional shoreface retreat in a fashion similar to that transpiring on the adjacent main- 1500 YR B P. LIKE PRESENT KITTS HUMMOCK LIKE PRESENT PORT MAHON 3000 YR. B,P. LIKE PRESENT BOMBAY HOOK 5000 YR B. P. LIKE PRESENT NEW CASTLE 7000 YR BP FIGURE 43. Late Holocene evolution of the mouth of Delaware Bay, as inferred from cross sections across the modern bay. Apron of sand extending into bay mouth is assumed to have prograded up the bay concurrently with the landward movements of the shoreline on either side. From Kraft et al. (1974). 513 SAND STORAGE ON ROCKY COASTS 301 FIGURE 44. Representative examples of inlet morphology. (a) Fire Island Inlet, Long Island, a barrier-overlap inlet on a drift-aligned coast. Littoral drift dominates the ebb tidal jet. (b) Ocracoke Inlet. North Carolina. Nearly symmetrical inlet land coast. Frequently, however, the retreat path of the estuary is visible in the form of a cross-shelf channel and a ridge on the updrift side of the channel. On the Dela- ware inner shelf, such a ridge can be seen to mark the retreat path of the shoal on the north side of the estuary mouth, while the associated channel has formed by the retreat of the main flood channel of the estuary mouth (see Fig. 12, Chapter 15). Coastal Inlets and Littoral Drift The morphology of narrow estuary mouths and their analogs, coastal inlets, depends on the relative strengths of the river mouth or inlet jet, the wave-driven littoral current, and the tidal- and wind-driven components of the shelf flow field. Distributary mouths, subject to pe- riodic flooding and entering relatively tideless, wave- sheltered seas, consist of subaerial levees capped by a lunate distributary mouth shoal (Fig. 42^4). As a con- sequence of the Coriolis effect, flow is more intense on the right-hand side of the channel looking downstream, and as a consequence, the right-hand levee tends to ex- tend itself farther seaward as in the case of Mississippi distributary mouths. If the inlet faces an open or tidal sea, then the wave- and tide-driven coastal flow is diverted seaward around the ebb tidal jet (Todd, 1968) and the shoal assumes a half-teardrop shape (Fig. 425). On barrier coasts, the pattern of sand storage at tidal inlets tends toward one of three basic patterns: overlap, symmetrical, or offset inlets (Fig. 44). While these pat- terns have long been recognized, the responsible trans- port systems and sand budgets are imperfectly under- stood (Hayes et al., 1970; Byrne et al., in press; Gold- smith et al., in press). As noted by Byrne (personal com- munication), the patterns appear to reflect the relative intensities of gross littoral drift (both up and down the coast) and net drift (the difference between mean annual upcoast discharge and mean annual downcoast dis- charge). If both the gross rates of drift and the net rate are high, a disproportionately high volume of sand stor- age may occur in the updrift barrier segment, and an overlap barrier may result (Fig. 44^4). Where moderate gross rates of drift are associated with a strong net rate of drift, the situation favors a barrier offset inlet, in which the storage of sand on the downdrift side of the exterior shoal is favored (Fig. 44C). In one of the best studied barrier offset inlets, Wachapreague Inlet on the flow on a swash-aligned coast has resulted in sand storage in the wave-protected interior (flood delta) shoal. Ebb tidal jet domi- nates over littoral drift, (c) Absecon Inlet, New Jersey. Ebb- dominated flow has resulted in sand storage on the downdrift side of the inlet and an offset of the flanking barrier islands. 514 302 COASTAL SEDIMENTATION Delmarva coast, the role of the lagoonal reservoir in modifying the hydraulic characteristics of the inlet is of paramount importance (Byrne et al., in press a). In lagoon-inlet systems where the ratio of the in tertidal water prism to the subtidal volume is very large, a strong time- velocity asymmetry develops (see Postma, 1967). The strongest currents occur just before high tide, when the tidal channels have filled and the vast marsh surface is beginning to flood, and just after high tide, when the marshes are draining. Flows around low tide are weaker, as they are associated with the much slower discharge and recharge of the tidal creek system. In addition, flood and ebb durations are dissimilar, with a greater ebb duration. This phenomenon is a con- sequence of the lagoonal basin's morphology and fric- tional characteristics (Byrne et al., in press a), and has been predicted by shallow water tidal theory for storage systems with sloping banks (Mota-Oliveira, 1970; King, 1974). In physical terms, the hydraulic head generated across the inlet by the flood tide is imposed on the deepest part of the lagoon relatively early in the tidal cycle. Here frictional retardation of flow is least efficient, and the resulting sea surface slope propagates rapidly across the lagoon, resulting in rapid water influx. The greatest potential drop across the lagoon surface during the ebb half-cycle occurs when the marsh surface is still un- covering. Frictional retardation of flow is more effective in the thin landward portions of the lagoonal water column, and the ebb is prolonged. As a result of these modifications of the tidal cycle, the inlet operates in a bypassing mode. Sand is swept into the inlet from the updrift side, but does not pene- trate very far before it is swept out again, and the pro- longed ebb carries it into the storage area on the down- drift side of the external shoal. Here sand storage is en- hanced by the refraction pattern of shoaling waves (Goldsmith et al., in press). Symmetrical inlets are favored by swash-aligned coasts, where the ratio of the littoral component of wave power to tidal power is relatively low (Fig. 445). Symmetrical inlets, particularly those backed by lagoons with rela- tively small intertidal prisms and relatively large sub- tidal volumes, tend to store sand primarily in the interior shoal within the lagoon. SAND STORAGE ON ROCKY COASTS Rocky coasts display the greatest complexity in three dimensions. Rocky hinterlands in a mature state of dis- section result in embayed coasts with deep reentrants between rocky salients. If the substrate consists of folded metamorphic rocks, then it may have a well-defined anisotropy of its own and truly baroque patterns may result (Fig. 45). The fields of wave refraction developed over the seaward extensions of headlands result in fre- quent reversals of the sense of littoral drift cells and closely spaced alteration of zones of littoral drift diver- gence and convergence. Because of the relative steepness of the regional seaward slope and the resistant nature of the substrate, wave energy is concentrated along a very narrow intertidal zone. Waves breaking against vertical surfaces can generate enormous instantaneous forces of tens of metric tons per square meter (Zenkovitch, 1967, p. 139). Rocky shores yield along planes of weakness to become mantled with boulders under this assault (Fig. 46) and the intertidal and subtidal talus slopes become grinding mills where attrition produces finer debris and continues to grind it finer until, at about the grade of medium sand, the immersed weight of grains is no longer adequate to result in significant chipping or cracking — as long as the particles are able to escape the proximity and nutcracker behavior of coarser particles. The inter- action of intertidal and shallow subtidal wave forces with the three-dimensional complexity of rocky coasts results in such erosional forms as stacks, arches, and sea caves, and the constructional forms of looped, fringing, recurved, and cuspate spits, and tombolos that have long been the delight of coastal morphologists (Fig. 47). The constructional forms constitute localized depositional re- gressions and are usually comprised of sets and subsets of beach ridges reflecting stages in the feature's growth. If the net rate of sedimentation is sufficiently high rela- tive to the rate of sea-level rise, these forms tend to grow and coalesce, and will ultimately form a continu- ous shoreface. Rocky coasts are more nearly likely to be tectonically active than low, unindurated coasts, other things being equal, and the resistant character of their substrate may result in delays in the adjustment of the incised equilib- rium profile to the crustal movement, if this adjustment is indeed attained. Comparison of rocky coasts from dif- ferent parts of the world has revealed a continuum of adjustment from coasts as irregular as the margins of newly dammed reservoirs, to coasts whose adjustment has been complete, so that, by a combination of head- land truncation and the filling in of bays, the coastline has been straightened in plan view and the shoreface has received the characteristic exponential curvature. This continuum led Davis (1909) and Johnson (1919) to the concept of a cycle of coastal evolution in which, after an initial relative movement of sea level, the shore- line is straightened and the equilibrium profile passes through a cycle of youth, maturity, and old age. Zenkovitch (1967) has objected to the simplified assump- tions of the model and suggests that three types of em- 515 70°00' 43°30 69°50'W 69°45' FIGURE 45. Lpper: A portion of the coast of southern Maine. Bedrock is iso- clinally folded schist and gneiss. Lower: Beginning of formation of constructional shoreface and estuary mouth shoal at mouth of Kennebec River; see the upper diagram for location. 303 516 304 COASTAL SEDIMENTATION la MASS MOVfMENT FIGURE 46. Diagrammatic representation of major processes of cliff retreat and evolution. (7a) Undercutting and rapid removal of collapsed material, (lb) Undercutting and slow removal of collapsed material. (2) Mass movement and removal at various rates. From Davies (1973). bayed coasts may be distinguished on the basis of the relationship between the submarine slope and the equi- librium profile generated on it, as follows: (1) deep- water coasts where the submarine bottom passes immedi- ately below the equilibrium profiles; (2) coasts with deep-water capes, where this is true only off capes, and (3) shallow-water coasts, where the submarine slope is everywhere above the equilibrium profile. The term "effective wave base" is probably best substituted for equilibrium profile here, for Zenkovitch concludes that sectors of coasts "above the profile of equilibrium" are those sectors that develop forms of accumulation (sandy beaches, barriers, spits, and tombolos; Fig. 47), and that shallow water coasts develop the most complex ar- ray of these features. Zenkovitch further traces subcycles of coastal evolution caused by feedback between evolving accumulation forms and the rocky substrate, or between two forms, whereby the growth of some spits into wave shadows behind headlands may distort their sub- sequent pattern, and the growth of other spits may shield and starve younger spits, or induce yet others where none existed. These subcycles are probably more common than the Davis-Johnson cycle, which requires an isostatic crustal movement or eustatic sea-level jump for rejuvenation. They may be observed on all stable rocky coasts under- going transgression by postglacial sea-level rise. Such coasts probably do not evolve at all in the Davis-Johnson sense, but undergo steady state subsidence in a state of perpetual youth, maturity, or old age, depending on the degree of induration of the substrate and the amplitude of the inherited relief. The relationship between the rate of sea-level rise and the relief and induration of the substrate also deter- mines the geometry of sediment storage (Fig. 48). Cores off transgressed crystalline coasts of high relief might be expected to reveal a residual rubble overlain by fine- grained bay deposits. Overlying sand deposits of com- plex shape would reflect the passage of the outer shore- line with its array of accumulation forms. The upper surface of the sand horizon will have been beveled at least locally by shoreward profile translation, and off- shore sands or muds may locally have accumulated over the surface of marine erosion. Off high crystalline coasts, the full sequence will rarely develop and will be com- pletely missing off capes, where surf-rounded boulders may litter bare rock surfaces for kilometers offshore. Pocket beaches and spits may locally survive the trans- gressive process relatively intact; a rock-tied spit cannot retrograde with the ease of a low coast barrier. On steep coasts transgressive deposits may be minimal. On steep coasts with very narrow shelves, submarine canyons may penetrate almost to the shoreface, to tap the littoral drift, through such gravity processes as sand creep. On steep deep-water coasts, prisms of beach shingle intermittently cascade to bathyal depths down steep rock slopes that may be erosion-modified fault scarps; sediment passes through the coastal zone by gravity bypassing (Fig. 49). As the coast is lower and softer, so will the sequence more nearly resemble the uniform sequence typical of the low coast transgression. Bay muds will more nearly resemble lagoonal muds, capped perhaps by nearly uniform sheets of backbarrier and shelf sands instead of lenticular remnants of spits and tombolos. Regressive deposits occur on some rocky coasts, as a consequence of the Late Holocene reduction in the rate of sea-level rise, where sediment input is sufficient to reverse the sense of shoreline migration. In extreme cases, alluvial gravel cones may build out across the transgres- sive deposits. Bouldery topset beds may pass into foreset sands and then into bottomset muds within a few hun- 517 SUMMARY 305 FIGURE 47. Types of coastal accumulation forms, according to Zenkovitch (1967). Fringing: a. beach nourished from offshore; h, beach nour- ished from alongshore; c, beach filling an indenta- tion; d, cuspate beach with bilateral nourishment; e, asymmetrical, cuspate beach with bilateral nourishment, attached at one end; (, spit with unilateral nourishment; g, arrow (spit with bilateral nourishment); h, spit on smooth coast; i, bay mouth barrier; \, midbay barrier; k, tombolo; \, interisland tombolo, doubly attached; m, looped spit with bilateral nourishment; n, looped spit with unilateral nourishment; o, cuspate spit, detached; p, barrier island; q, barrier island resulting from cutting of inlet; r, estuary mouth swash bar; s, barrier sequence. Symbols: (1) mainland and active cliff; (2) dead cliff and coast with beach; (3, 4) major and minor transport directions. dred meters. On steep, unstable coasts, such masses may periodically slump down the submarine slopes to the basin floor. SUMMARY In considering coastal sediment transport, it is conven- ient to divide the movement of sand into an onshore- offshore component and a coast-parallel component. Onshore-offshore transport occurs in two provinces. In the nearshore province of beach, longshore trough, plunge point bar, and upper shoreface, onshore-offshore transport is controlled by the regime of shoaling and breaking waves. Breakpoint bars are initiated during the waning phases of storms. During the ensuing fair- weather period they tend to migrate onshore, and weld 518 306 COASTAL SEDIMENTATION LAND py-^j OUTCROP [p%3 GRAVEL □ SAND I I MUD 1 2 3 KILOMETERS FIGURE 48. Hypothetical stratigraphy of a rocky coast undergoing transgression. to the berm. The high, steep waves of storms tend to strip sand from the beach and transport it out to the surf zone, and the cycle begins anew. The cycle tends to be linked to the cycle of seasons in that offshore transport dominates during the period of winter storms, while onshore transport tends to dominate during the summer season of fair weather. The lower shoreface is a second province subject to onshore-offshore transport. The corresponding hydraulic regime is the zone of fric- tion-dominated unidirectional flow that constitutes the coastal boundary of the shelf flow field. During storms (or peak tidal flows) velocity in this zone may be more intense than in the zone of quasi-geostrophic flow further offshore. Downwelling and a seaward component of bot- tom flow may occur in this zone during some storm flows, at the same time that sand is moving seaward in the surf zone, so that sand is transported off the shore- face altogether. The interrelated behavior patterns of the zone of shoaling and breaking waves and zone of friction-domi- nated flows give rise on many coasts to a long-term cyclic pattern of advance or retreat of the coastal profile. The upper shoreface undergoes net aggradation and pro- gradation over a period of years tending toward the ideal wave-graded profile. A major storm or period of severe storms will result in large-scale seaward transport of sand, causing flattening and significant landward trans- lation of the profile. On coasts experiencing a net littoral drift surplus, fair-weather progradation is more effective than storm erosion, and the profile translates seaward and (in compensation for postglacial sea-level rise) up- ward. On coasts experiencing a net littoral drift deficit, the storm regime controls the offshore-onshore sand budget, and the coastal profile undergoes landward and upward translation through a process of erosional shore- face retreat. The debris resulting from this process nour- ishes the leading edge of the surficial sand sheet that mantles the shelf. The cycle of onshore-offshore transport is superim- posed on a much more intensive flux of sand parallel to the beach, under the impetus of the wave-driven littoral flows, and wind- and tide-driven coastal currents. As a result, there is an innate tendency toward two-dimen- sionality of the shoreface, in that successive downcoast profiles tend to be very similar. Headlands experience a greater littoral wave energy density, greater breaker angles, and decreasing littoral sand discharge along the beach toward the adjoining bay. A pattern of transport away from headlands toward bays is superimposed on a regional direction of littoral sand transport determined by the prevailing direction of deep-water wave approach. The resulting alternation of littoral drift divergences and convergences may impose a three-dimensionality on an unconsolidated coast in the form of alternate cuspate forelands and zetaform bays. Three-dimensionality may also be inherited from the relief of a rocky surface under- going transgression, or may be induced on an unconsoli- dated coast in the form of constructional river mouths and tidal inlets. The beach and shoreface comprise major reservoirs of sand in the coastal sediment transport system. During a transgression, the superimposition of a straight, wave- maintained upper shoreface on an irregular surface re- sults in the formation of two shorelines. An inner, la- 519 SUMMARY 307 VAR AND PAILLON CANYONS Recen: Mud L__JSand Q .* ^^J cH L ^> ■1 - ■-»- vsMH SHU 'dJft*^ "*■'** tH Hi. •- v v. *,- *!■ D FIGURE 49. Gravity bypassing on a recently formed continental margin, Provencal coast of France. (A) Axial fades of the Var and Paillon canyons. (B) Paillon River mouth (left) and pebble beach, Baie des Anges, Nice. (C, D) Boulder (up to 50 cm) mud admixtures at diving locality shown in (B),- depth 25 m. (E) Large blocks of Jurassic limestone overgrown with Poseidonia near Cap Ferrat. From Stanley (1969). goonal shoreline approximates the intersection of still water level with the dissected subaerial surface under- going erosion. An outer, oceanic shoreline of barrier spits and islands results from ( 1 ) the detachment of drift-nourished, wave-maintained beaches from the main- land as the rising sea floods the swales behind them, and (2) the lateral propagation of the shoreface from head- lands across the mouths of adjoining bays. Sand is also stored in shoals that form at littoral drift convergences, and in oblique-tending, shoreface-connected sand ridges that form at the foot of the shoreface in response to the storm wave regime and storm coastal currents. Where the littoral drift system intersects with the river- and tide- driven jets of river mouths and inlets, sand is stored in arcuate, seaward-convex shoals whose crests bear com- plex patterns of interdigitating ebb and flood channels, separated by sand ridges that build into the intertidal zone as swash platforms. 520 308 COASTAL SEDIMENTATION REFERENCES Allen, J. R. L. (1968a). Current Ripples. Amsterdam: North-Hol- land, 433 pp. Allen, J. R. L. (1968b). The nature and origin of bedform hier- archies. Sedimentology, 10: 161-182. Allersma, E. (1972). Mud on the oceanic shelf off Guiana. In Symposium on Investigations and Resources oj the Caribbean Sea and Adjacent Regions. UNESCO, Paris (Unipub, N.Y.), pp. 193- 203. Assaf, G., R. Gerard, and A. L. Gordon (1971). Some mecha- nisms of oceanic mixing revealed in aerial photographs. J. Geophys. Res., 76(27): 6550-6572. Bascom, W. N. (1951). The relationship between sand size and beach slope. Trans. Am. Geophys. Union, 32: 866-874. Beall, A. O. (1968). Sedimentary processes operative along the western Louisiana shoreline. J. Sediment. Petrol., 38: 869-877. Beardsley, R. C. and B. Butman (1974). Circulation on the New England continental shelf: Response to strong winter storms. Geol. Res. Lett., 1: 181-184. Bernard, H. A. and R.J. Leblanc (1965). Resume of the Quater- nary geology of the Northwest Gulf of Mexico Province. In H. E. Wright, Jr. and D. J. Frey, eds., The Quaternary oj the United States. Princton: Princeton Univ. Press, pp. 137-185. Bruun, P. (1954). Migrating sand waves and sand humps, with special reference to investigations carried on in the Danish North Sea coast. Con/. Coastal Eng., 5th Proc, 460-468. Bruun, P. (1962). Sea level rise as a cause of shore erosion. Proc. ASCE J. Waterw. Harbors Div., 88: 117-130. Byrne, R. J., P. Bullock, and D. G. Tyler (in press a). Response characteristics of a tidal inlet: A case study. In M. O. Hayes, ed., Second International Estuarine Conference Proc, Myrtle Beach, S.C., Oct. 15-18, 1973: New York, Academic Press. Byrne, R. J., J. T. De Alteris, and P. A. Bullock (in press b). Chan- nel stability in tidal inlets. 14th Conf. Coastal Eng. Proc, Copenhagen, 1974. Caston, V. N. D. (1972). Linear sand banks in the southern North Sea. Sedimentology, 18: 63-78. Caston, V. N. D. and A. H. Stride (1970). Tidal sand movement between some linear sand banks in the North Sea off north- east Norfolk. Mar. Geol., 9: M38-M42. Charnell, R. L. and D. V. Hansen (1974). Summary and Analysis of Physical Oceanography Data from the New York Bight Apex Col- lected During 1969-70. MESA Rept. 74-3, National Oceanic and Atmospheric Administration. Washington, D.C.: U.S. Govt. Printing Office, 44 pp. Clifton, H. E., R. E. Hunter, and R. L. Phillips (1971). Deposi- tional structures and processes in the non-barred, high energy nearshore. J. Sediment. Petrol., 41: 651-670. Colquhoun, D. J. (1969). Geomorphology oj the Lower Coastal Plain oj South Carolina. MS-15, South Carolina Development Board, Div. Geol., 36 pp. Cook, D. O. (1969). Occurrence and geologic work of rip-currents in southern California. J. Sediment. Petrol., 39: 781-786. Cook, D. O. and D. S. Gorsline (1972). Field observations of sand transport by shoaling waves. Mar. Geol., 13: 31-55. Csanady, G. T. and J. T. Scott (1974). Baroclinic coastal jets in Lake Ontario during IFYGL. J. Phys. Oceanogr., 4: 524-541. Curray, J. R. (1964). Transgressions and regressions. In R. L. Miller, ed., Papers in Marine Geology — Shepard Commemorative Volume. New York: Macmillan, pp. 175-203. Curray, J. R., F. J. Emmel, and P. J. S. Crampton (1969). Lagunas costeras, un simposio. In Mem. Simp. Int. Lagunas Costeras. UNAM-UNESCO, Nov. 28-30, 1967, Mexico, D.F., pp. 63-100. Davies, J. L. (1958). Wave refraction and the evolution of curved shorelines. Geogr. Stud., 5: 1-14. Davies, J. L. (1973). Geographical Variation in Coastal Development. New York: Hafner, 204 pp. Davis, R. A. and W. T. Fox (1972). Coastal process and near- shore sand bars. J. Sediment. Petrol., 42: 401^412. Davis, R. A., W. T. Fox, M. O. Hayes, and J. C. Boothroyd (1972). Comparison of ridge and runnel systems in tidal and non-tidal environments. J. Sediment. Petrol., 32: 413—421. Davis, W. M. (1909). Geographical Essays, 1954 Dover edition. New York: Dover Publications, 777 pp. Dolan, R. (1971). Coastal landforms: Crescentic and rhythmic. Geol. Soc. Am. Bull. ,82: 177-180. Duane, D. B., M. E. Field, E. P. Meisburger, D. J. P. Swift, and S. J. Williams (1972). Linear shoals on the Atlantic inner continental shelf, Florida to Long Island. In D. J. P. Swift, D. B. Duane, and O. H. Pilkey, eds., Shelf Sediment Transport: Process and Pattern. Stroudsburg, Pa.: Dowden, Hutchinson & Ross, pp. 447^-99. Ekman, V. W. (1905). On the influence of the earth's rotation on ocean currents. Ark. J. Mat., Astron. Phys., 2: 1-53. Emery, K. O. (1968). Relict sediments on continental shelves of world. Am. Assoc. Pet. Geol. Bull., 52: 445^64. Faller, A. J. (1963). An experimental study of the instability of the laminar Ekman boundary layer. J. Fluid Mech., 15: 560- 576. Faller, A. J. (1971). Oceanic turbulence and the Langmuir cir- culations. Ann. Rev. Ecol. System., 2: 201-233. Faller, A. J. and R. B. Kaylor (1966). A numerical study of the instability of the laminar Ekman boundary layer. J. Atmos. Sci., 23: 466^80. Farmer, D. G. (1971). A computer simulation model of sedimen- tation in a salt wedge estuary. Mar. Geol., 10: 133-143. Fenneman, N. M. (1902). Development of the profile of equi- librium of the subaqueous shore terrace. J. Geol., 10: 1-32. Fisher, J. J. (1968). Barrier island formation: Discussion. Geol. Soc. Am. Bull., 79: 1421-1426. Fox, W. T. and R. A. Davis (1974). Beach Processes on the Oregon Coast, July 1973. Tech. Rep. 12, Office of Na"al Research, Washington, D.C., 85 pp. Gilbert, G. K. (1890). Lake Bonneville. U.S. Geol. Surv. Mono. 1, 438 pp. Goldsmith, U., R. J. Byrne, A. H. Sallenger, and R. Driecker (in press). The influence of waves on the origin and develop- ment of the offset coastal inlets of the southern Delmarva Peninsula, Virginia. In M. O. Hayes, ed., Second International Estuarine Conference Proceedings, Myrtle Beach, South Carolina, Oct. 15-18, 1973. New York: Academic Press. Halligan, G. H. (1906). Sand movement on the New South Wales coast. Proc Limnol. Soc. NSW, 31: 619-640. Harrison, W. and R. M. Alamo (1964). Dynamic Properties oj Im- mersed Sand at Virginia Beach, Virginia, Coastal Eng. Res. Cent., Tech. Memo. 9, pp. 1-52. Hayes, M. O. (1967). Hurricanes as Geological Agents: Case Studies oj Hurricanes Carla, 1961, and Cindy, 1963. Texas Bureau of Eco- nomic Geology, Report of Investigation 61, 54 pp. 521 REFERENCES 309 Hayes, M. O., U. Goldsmith, and C. H. Hobbs (1970). Offset coastal inlets. In 12th Conference Coastal Engineering Proceedings, pp. 1187-1200. Holliday, B. VV. (1971). Observations on the hydraulic regime of the ridge and swale topography of the inner Virginia shelf. Masters Thesis, Old Dominion University, Inst, of Oceanog- raphy, Norfolk, Va. Howard, J. D. and H. E. Reineck (1972). Physical and biogenic sedimentary structures of the nearshore shelf. Senckenbergiana Marit.,4: 81-124. Hoyt, J. H. (1967). Barrier island formation. Ceol. Soc. Am. Bull., 78: 1125-1136. Hoyt, J. H. (1970). Development and migration of barrier islands, northern Gulf of Mexico, discussion. Geol. Soc. Am. Bull., 81: 3779-3782. Hoyt, J. H. and V.J. Henry (1967). Influence of island migration on barrier island sedimentation. Geol. Soc. Am. Bull., 78: 77- 86. Ingle, J. C. (1966). The Movement oj Beach Sand. New York: American Elsevier, 221 pp. Johnson, D. VV. (1919). Shore Processes and Shoreline Development. New York: Hafner (1965 facsimile), 584 pp. Johnson, J. VV. (1949). Scale effects in hydraulic models involving wave motion. Trans. Am. Geophys. Union, 30: 517-527. Johnson, J. VV. and P. S. Eagleson (1966). In A. J. Ippen, ed., Estuary and Coastline Hydrodynamics. New York: McGraw-Hill, pp. 404-492. Keulegan, G. H. (1948). An Experimental Study oj Submarine Sand Bars. Beach Erosion Board Tech. Rep. 3, 42 pp. King, C. A. M. (1972). Beaches and Coasts. New York: St. Martins Press, 570 pp. King, C. A. M. and VV. VV. Williams (1949). The formation and movement of sand bars by wave action. Geogr. J., 113: 70-85. King, D. B., Jr. (1974). The Dynamics oj Inlets and Bays. Tech. Rep. 22, College of Engineering, University of Florida, Gainesville, 82 pp. Kraft, J. C, R. E. Sheridan, R. D. Moose, R. N. Strom, and C. B. Weil (1974). Middle-Late Holocene evolution of the mor- phology of a drowned estuary system — the Delaware Bay. In International Symposium on Interrelationships oj Estuarine and Con- tinental Shelj Sedimentation. Inst. Geol. du Bassin d'Aquitaine, July 9-14, 1973, pp. 297-306. Langfelder, J., D. Stafford, and M. Amein (1968). A Reconnaissance oj Coastal Erosion in North Carolina. Dept. Civil Eng., North Carolina State University, Raleigh, N.C., 127 pp. Langford-Smith, T. and B. G. Thorn (1969). New South Wales coastal morphology. J. Geol. Soc. Aust., 16: 572-580. Langmuir, I. (1925). Surface motion of water induced by the wind. Science, 87: 119-123. Leopold, L. B., M. G. Wolman, and J. P. Miller (1964). Fluvial Processes in Geomorphology. San Francisco: Freeman, 522 pp. Longuet-Higgins, M. S. (1953). Mass transport in water waves. Phil. Trans. Roy. Soc. London, 245: 535-581. Ludwick, J. C. (in press). Tidal currents, sediment transport, and sandbanks in Chesapeake Bay entrance, Virginia. In M. O. Hayes, ed., Second International Estuarine Conj. Proc. Myrtle Beach, S.C., Oct. 15-18, 1973: New York, Academic Press. McGee, W. D. (1890). Encroachments of the sea. The Forum, 9: 437-449. McHone, J. F., Jr. (1972). Morphologic time series from a sub- marine sand ridge on the south Virginia coast. Masters The- sis, Old Dominion University, Norfolk, Va., 59 pp. McKinney, T. F., W. L. Stubblefield, and D. J. P. Swift (1974). Large scale current lineations on the Great Egg shoal retreat massif: Investigations by sidescan sonar. Mar. Geol. , 17 : 79-102. McMaster, R. L. and L. E. Garrison (1967). A submerged Holo- cene shoreline near Block Island, Rhode Island. Mar. Geol., 75: 335-340. May, J. P. and W. F. Tanner (1973). The littoral power gradient and shoreline changes. In D. R. Coates, ed., Publications in Geomorphology. Binghamton: State University of New York, 404 pp. Miller, R. L. and J. M. Zeigler (1958). A model relating dynamics and sediment pattern in equilibrium in the region of shoaling waves, breaker zone, and foreshore. J. Geol., 66: 417-441. Miller, R. L. and J. M. Zeigler (1964). A study of sediment dis- tribution in the zone of shoaling waves. In R. L. Miller, ed., Papers in Marine Geology — Shepard Commemorative Volume. New York: MacMillan, pp. 133-153. Moody, D. W. (1964). Coastal morphology and processes in re- lation to the development of submarine sand ridges off Beth- any Beach, Delaware. Ph.D. Thesis, Johns Hopkins Univer- sity, Baltimore, 167 pp. Moore, G. T. (1970). Role of salt wedge in bar finger sand and delta development. Am. Assoc. Pet. Geol. Bull., 54: 326-333. Mota-Oliveira, I. B. (1970). Natural flushing ability in tidal in- lets. Proc. 12th Coastal Eng. Conj., Washington, D.C., pp. 1827- 1845. Murray, S. P. (1967). Control of grain dispersion by particle size and wave state. J. Geol., 75: 612-634. Murray, S. P. (1970). Bottom currents near the coast during Hur- ricane Camille. J. Geophys. Res., 75: 4579-4582. Neumann, G. and W. J. Pierson, Jr. (1966). Principles oj Physical Oceanography. Englewood Cliffs, N.J.: Prentice Hall, 545 pp.. Nevesskii, E. N. (1969). Some data on the postglacial evolution of Karkinit Bay and the accumulation of bottom sediments within it. In V. V. Longinov, ed., Dynamics and Morphology oj Seacoasts. Akademiya Nauk SSSR Trudy Instituta Okeano- logii, Vol. 48, 371 pp., translated from Russian by Israel Program for Scientific Translations; reproduced by the Clear- inghouse for Federal Scientific and Technical Information, Springfield, Va., pp. 92-110. Niedoroda, A. W. and W. F. Tanner (1970). Preliminary study of transverse bars. Mar. Geol., 9: 41-62. Oaks, R. Q., Jr. and N. K. Coch (1963). Pleistocene sealevels, southeastern Virginia. Science, 14X): 979-983. Oertel, G. F. (1972). Sediment transport of estuary entrance shoals and the formation of swash platforms. J. Sediment. Petrol., 42: 858-863. Oertel, G. F. and J. D. Howard (1972). Water circulation and sedimentation at estuary entrances on the Georgia coast. In D. J. P. Swift, D. B. Duane, and O. H. Pilkey, eds., Shelj Sediment Transport: Process and Pattern. Stroudsburg, Pa.: Dow- den, Hutchinson & Ross, pp. 411-427. Otvos, E. G. (1969). A sub-recent beach ridge complex in southern Louisiana. Geol. Soc. Am. Bull., 80: 2353-2358. Otvos, E. G. (1970a). Development and migration of barrier islands, northern Gulf of Mexico. Geol. Soc. Am. Bull., 81: 241-246. Otvos, E. G. (1970b). Development and migration of barrier islands, northern Gulf of Mexico: Reply. Geol. Soc. Am. Bull., 81: 3783-3788. Pierce, J. W. and D. J. Colquhoun (1970). Holocene evolution of 522 310 COASTAL SEDIMENTATION a portion of the North Carolina coast. Geol. Soc. Am. Bull., 81: 3697-3714. Postma, H. (1967). Sediment transport and sedimentation in the marine environment. In G. H. Lauff, ed., Estuaries. Washing- ton, D.C.: Am. Assoc. Adv. Sci., pp. 158-180. Price, W. A. (1954). Dynamic environments: Reconnaissance mapping, geologic and geomorphic, of continental shelf of Gulf of Mexico. Trans. Gulf Coast Assoc. Geol. Soc, 4: 75-107. Price, W. A. (1955). Correlation of Shoreline Type with Offshore Bot- tom Conditions. Dept. of Oceanography, Texas A&M Univer- sity, Project 65: 75-107. Reineck, H. E. and I. B. Singh (1971). Der Golf von Gaeta (Tyrrhenisches Meer). III. Die gefuge von vorstrand — und schelfsedimenten. Senckenbergiana Marit., 3: 135-183. Robinson, A. H. W. (1966). Residual currents in relation to shoreline evolution of the East Anglian coast. Mar. Geol., 4: 57-84. Russell, R. C. H. and J. D. C. Osorio (1958). An experimental investigation of drift profiles in a closed channel. Proc. 6th Conf. Coastal Eng., pp. 171-183. Schiffman, A. (1965). Energy measurements in the swash-surf zone. Limnol. Oceanogr., 10: 255-260. Schwartz, M. L. (1965). Laboratory study of sea level rise as a cause of shore erosion. J. Geol., 73: 528-534. Schwartz, M. L. (1967). The Bruun theory of sea level rise as a cause of shore erosion. J. Geol., 75: 76-92. Schwartz, M. L. (1968). The scale of shore erosion. J. Geol., 76: 508-517. Schwartz, M. L. (1973). Barrier Islands. Stroudsburg, Pa.: Dow- den, Hutchinson & Ross, 451 pp. Shepard, F. P. (1950). Longshore-Bars and Longshore-Troughs. Beach Erosion Board Tech. Memo. 15, 32 pp. Silvester, R. (1974). Coastal Engineering II. New York: American Elsevier, 338 pp. Smith, J. D. and T. S. Hopkins (1972). Sediment transport on the continental shelf off of Washington and Oregon in the light of recent current meter measurements. In D. J. P. Swift, D. B. Duane, and O. H. Pilkey, eds., Shelf Sediment Transport: Process and Pattern. Stroudsburg, Pa.: Dowden, Hutchinson & Ross, pp. 143-180. Sonu, C.J. and J. L. van Beek (1971). Systematic beach changes in the outer banks, North Carolina. J. Geol., 74: 416-425. Sonu, C. J., J. M. McCloy, and D. S. McArthur (1967). Long- shore currents and nearshore topographies. Proc. 10th Conf. Coastal Eng., pp. 55-549. Stahl, L., J. Koczan, and D. Swift (1974). Anatomy of a shore- face-connected ridge system on the New Jersey shelf: Impli- cations for genesis of the shelf surficial sand sheet. Geology, 2: 117-120. Stanley, D. J. (1969). Submarine channel deposits and their fossil analogs (fluxoturbidites). In D. J. Stanley, ed., The New Con- cepts of Continental Margin Sedimentation. Washington, D.C.: Am. Geol. Inst., pp. DJS-9-1 to DJS-9-17. Strahler, A. (1963). The Earth Sciences. New York: Harper and Row, 681 pp. Stubblefield, W. L., J. W. Lavelle, T. F. McKinney, and D. J. P. Swift (1975). Sediment response to the hydraulic regime on the central New Jersey shelf. J. Sediment Petrol., 45: 337-358. Swift, D. J. P. (1973). Delaware Shelf Valley: Estuary retreat path, not drowned river valley. Geol. Soc. Am. Bull., 84: 2743- 2748. Swift, D. J. P. (1975a). Barrier island genesis: Evidence from the Middle Atlantic Shelf of North America. Sediment. Geol., 14: 1-43. Swift, D. J. P. (1975b). Tidal sand ridges and shoal retreat massifs. Mar. Geol., 18: 105-134. Swift, D. J. P., D. B. Duane, and T. McKinney (1974). Ridge and swale topography of the Middle Atlantic Bight: Secular response to Holocene hydraulic regime. Mar. Geol. ,15:227-247. Swift, D. J. P., B. W. Holliday, N. F. Avignone, and G. Shideler (1972a). Anatomy of a shoreface ridge system, False Cape, Virginia. Mar. Geol., 12: 59-84. Swift, D.J. P., J. W. Kofoed, F. P. Saulsbury, and P. Sears (1972b). Holocene evolution of the shelf surface, central and southern Atlantic coast of North America. In D. J. P. Swift, D. B. Duane, and O. H. Pilkey, eds., Shelf Sediment Transport: Process and Pattern. Stroudsburg, Pa.: Dowden, Hutchinson & Ross, pp. 499-574. Swift, D.J. P., R. B. Sanford, C. E. Dill, Jr., and N. F. Avignone (1971). Textural differentiation on the shoreface during ero- sional retreat of an unconsolidated coast, Cape Henry to Cape Hatteras, North Carolina. Sedimentology, 16: 221-250. Swift, D. J. P. and P. Sears (1974). Estuarine and littoral deposi- tional patterns in the surficial sand sheet, central and southern Atlantic shelf of North America. In International Symposium on Interrelationships of Estuarine and Continental Shelf Sedimentation. Inst. Geol. du Bassin d'Aquitaine, Bordeaux, Mem. 7, pp. 171-189. Tanner, W. F. (1961). Offshore shoals in an area of energy deficit. J. Sediment. Petrol., 31: 87-95. Tanner, W. F., R. G. Evans, and C. W. Holmes (1963). Low energy coast near Cape Romano, Florida. J. Petrol., 33: 713- 722. Todd, T. W. (1968). Dynamic diversion: Influence of longshore current-tidal flow interaction on chenier and barrier island plains. J. Sediment. Petrol., 38: 734-746. Valentin, H. (1954). Der landverlust in Holderness, Ostengland von 1852 bis 1952. Die Erde, 6: 296-315. See also, Land loss at Holderness, 1852-1952. In J. D. Steers, ed., Applied Coastal Morphology. Cambridge, Mass.: MIT Press, 1971, pp. 1 16-137. Van Straaten, L. M. J. U. (1965). Coastal barrier deposits in south and north Holland — in particular in the area around Scheveningen and Ijmuden. Meded. Geol. Sticht., NS 17: 41-75. Visher, G. S. (1969). Grain size distribution and depositional processes. J. Sediment. Petrol., 34: 1074-1 106. Weil, C. B., R. D. Moose, and R. E. Sheridan (1974). A model for the evolution of linear tidal built sand ridges in Delaware Bay, U.S.A. In G. Allen, ed., Estuary and Shelf Sedimentation: A Symposium. University of Bordeaux, July 7973. Wells, D. R. (1967). Beach equilibrium and second order wave theory. J. Geophys. Res., 72: 497-509. Wright, L. D. and J. M. Coleman (1972). River delta morphology: Wave climate and the role of the subaqueous profile. Science, 176: 282-284. Wright, L. D. and J. M. Coleman (1974). Mississippi River mouth processes: Effluent dynamics and morphologic devel- opment. J. Geol., 82: 751-778. Zenkovitch, V. P. (1967). Processes of Coastal Development. New York: Wiley, 738 pp. 523 49 Reprinted from: Marine Sediment Transport and Environmental Management, D. J. Stanley and D. J. P. Swift, editors, John Wiley and Sons, Inc., ley Chapter 15, 311-350. Offprints from: Marine Sediment Transport and Environmental Management Edited by D. J. Stanley and D. J. P. Swift Copyright 1976 by John Wiley & Sons, Inc CHAPTER 15 Continental Shelf Sedimentation DONALD J. P. SWIFT Atlantic Oceanographic and Meteorological Laboratories, Miami, Florida The preceding chapter considered in detail the nature of hydraulic process and substrate response along the coast. This chapter examines patterns of sedimentation on the shelf as a whole. It reexamines the coastal boundary of the shelf as a source of sediment for the rest of the shelf, and as a zone which thus regulates the rate and character of sedimentation on the shelf surface. Chapter 14 described a "littoral energy fence" imposed upon coastal sedimentation by the landward-directed asymmetry of wave surge in shoaling water, which causes sediment to be retained on the shoreface. This chapter concerns itself with the mechanisms by which this dynamic barrier is penetrated, along the shoreface or at river mouths, and by which sediment is injected into the shelf dispersal system. The relative efheiencies of shoreface and river mouth bypassing during periods of transgression on one hand, and during periods of regression on the other are described. These varying efheiencies lead to two distinct shelf regimes: a passive regime in which the shelf sand sheet is generated by erosional shoreface retreat (autochthonous sedimenta- tion) and a more active regime in which river mouth bypassing causes deposition across the shelf surface (allochthonous sedimentation). The chapter analyzes the transport patterns associated with these two regimes, and the resulting patterns of morphology, stratigraphy, and grain-size distribution. Portions of the material in this chapter have been presented elsewhere (Swift, 1974). 524 MODELS OF SHELF SEDIMENTATION One of the first comprehensive models for the genesis of clastic sediments on continental shelves was a by- product of Douglas Johnson's (1919, p. 211) attempt to apply Davis' geomorphic cycle of youth, maturity, and old age to the continental shelf (see p. 277). Johnson saw the shelf water column and the shelf floor as a sys- tem in dynamic equilibrium, in which the slope and grain size of the sedimentary substrate at each point control, and are controlled by, the flux of wave energy into the bottom. He described the resulting surface as an exponential curve in profile, concave up, with the steeper segment being the shoreface. Grain size was considered to decrease as a function of increasing depth with distance from shore, as ,a consequence of the diminishing input of wave energy into the seafloor. The model derived its sediment from coastal erosion rather than from river input, a more broadly applicable interpretation than many subsequent textbooks have realized. Despite its qualitative expression and limited appli- cability, the model was in advance of its time in its dynamical systems approach. However, this model could not withstand in its initial form the subsequent flood of data on the characteristic of shelf sediments. Shepard (1932) was the first to challenge it, noting that nautical charts of the world's shelves bore notations indicating that most shelves were veneered with a 311 312 C O N T I N K N T A I SHE! F SEDIMENTATION complex mosaic of sediment types, rather than a simple seaward-fining sheet. He suggested that these patches were deposited during Pleistocene low stands of the sea, rather than during Recent time. Emery (1952, 1968) raised this concept to the status of a new conceptual model. He classified shelf sediments on a genetic basis, as autlugenic (glauconite or phosphorite), organic (fcramini- fera, shells), residual (weathered from underlying rock), relict (remnant from a different earlier environment such as a now submerged beach or dune), and detrual, which includes material now being supplied by rivers, coastal erosion, and eolian or glacial activity. On most shelves, a thin nearshore band of modern detrital sedi- ment' is supposed to give way seaward to a relict sand sheet veneering the shelf surface. A third, more generalized model for shelf sedimenta- tion has been primarily concerned with the resulting stratigraphy. It incorporates elements of both the Johnson and Emery models. Like the Johnson model, it views the shelf surface as a dynamic system in a state of equilibrium with a set of process variables. The rate of sea-level change, however, is one of these variables; hence the effects of post-Pleistocene sea-level rise, as noted by Shepard and Emery, may be accounted for. The model may be referred to as the transgression- regression model, since it is generally expressed in these terms, or the coastal model, since it focuses on the behavior of this dynamic zone. It was first explicitly formulated by Grabau (1913), and more recently by Curray (1964) and Swift et al. (1972). In this model, the rate of sediment input to the continental shelf S, the character of the sediment G (grain size and mineralogy), the rate of energy input E, the sense and rate of relative sea-level change R, and slope L are seen as variables that govern the sense of shoreline movement (trans- gression or regression) and ultimately the character of shelf deposits. The relationship may be expressed in quasi-quantita- tive form as SG E R L oc T The processes controlling shelf sedimentation are much too complex to be adequately described by this equation and there is no way to evaluate its variables adequately. The expression is useful, however, in helping to sort out relationships. The first term, SG E, might be called the effective rate of coastal deposition. It increases with increasing 5\ the rate at which sedi- ment is delivered to the shore. It increases with increas- ing grain size G. since coarser sediments are less easily bypassed across the shelf. It decreases with increasing E, the rate of wind and tidal energy input, since a more rigorous hydraulic climate causes more sediment to be bypassed across the shelf. The second term, R/L, might be called the effective rate of sea-level movement. It increases with increasing /t, the absolute rate of sea-level movement (eustatic or tectonic), but decreases with increasing slope, L, of the coast. The steeper the slope, the greater the fall of sea level must be in order for the coast to advance a given distance. Also, with a greater slope, a greater volume of sediment must be delivered to the shoreline in order for the shoreline to prograde a given distance shoreward. The equation tells us that the rate and sense of shore- line movement, T, whether landward (negative) or seaward (positive), depends on the relationship between these two terms. Basic elements of the relationship are presented graphically in Fig. 1, according to a scheme of Curray (1964). In Fig. 2, the history of the Nayarit coast of Mexico has been plotted. The Coastal Boundary as a Filter: Shelf Sedimentary Regimes The fundamental determinants of shelf sedimentation are the areal extent of the adjacent continent undergoing denudation, and its relief, climate, and drainage pattern. These factors control the quantity of sediment delivered to the shoreline, and its textural and mineralogical composition. However, the rate and sense of shoreline movement, as determined by the parameters described above, have a modulating effect on the shelf sedi- mentary regime. It is helpful to think of the coastline as a '"littoral energy fence" (Allen, 1970b, p. 169) in which the net landward flow associated with bottom wave surge tends RELATIVE SEA LEVEL FALLING SEA LEVEL OR EMERGENCE RISING SEA LEVEL OR SUBSIDENCE RAPID SLOW STABLE SLOW RAPID FIGURE 1. Diagram of the effects of sea-level movement and the rate of coastal deposition on lateral migration of the shoreline. See text for explanation. From Curray (1964). 525 0 25 50 £ 75 o k 100 < STRAND PLAIN > ^> J[[ I C=3 MODERN ALLUVIUM _L SHELF hACIES MUDS X MODELS OF SHELF SEDIMENTATION 313 BASAL SANDS _L 20 15 10 0 5 NAUTICAL 10 MILES 15 20 25 30 RELATIVE SEA LEVEL FALLING SEA LEVEL OR EMERGENCE RISING SEA LEVEL OR SUBMERGENCE RAPIO SLOW STABLE SLOW RAPID o S i ' 1 \ \ I \ 1 \l 4 Z g O -J ( l\ 1 \ irs i r h- 2 i id O I (9 X \ T i i ■ • \* FIGURE 2. Above: Diagrammatic section off the Costa de Naya- rit, Mexico. See Fig. 10 A, Chapter 14 for details of coastal stra- tigraphy. Below: Schematic representation of shoreline' migration. From Curray (1964). to push sediment toward the shore. There are two basic categories of "valve" which regulate the passage of sediment through this dynamic coastal barrier into the transport system of the shelf surface. The shoreface may serve as a zone of sediment bypassing. The erosional retreat of the shoreface during a marine transgression bevels the subaerial surface being transgressed (Fig. 10A, Chapter 14), and spreads the resultant debris as a thin sheet over the shelf floor. The process by which the sedi- ment is so transferred is described in the accompany- ing text. The process is a passive and indirect one; the sediment that is released has undergone long-term storage as flood plain, lagoonal, or estuarine deposits, or has been derived from an earlier cycle of sedi- mentation. A second, more active route by which sediment may pass through the littoral energy fence is via the ebb tidal jet or flood stage jet of a river mouth. Patterns of river mouth bypassing are illustrated in Chapter 14 (Fig. 42). River mouth bypassing is more direct than shoreface by- passing, but sediment must still undergo storage. Sand is stored in the throat of the river mouth, and fines are stored in marginal marshes and mud flats until the period of maximum river discharge, when the salt wedge moves to the shoal crest, and stored sediment is bypassed to the shoreface of the shoal front (Wright and Coleman, 1974); see Chapter 14, Fig. 41. It may undergo a second period of storage on the shoreface and inner shelf until the period of maximum storm energy (Wright and Coleman, 1973). The mode of operation of these valves is dependent on basic parameters of coastal sedimentation. The spacing of river mouths is the fundamental determinant of the relative roles of shoreface versus river mouth bypassing. The character of the hydraulic climate is also im- portant; an intense tidal regime increases the efficiency of river mouth bypassing, whereas an intense wave climate increases the efficiency of shoreface bypassing. The rate and sense of coastal translation as described in the preceding section strongly affect the relative 526 314 CONTINENTAL SHELF SEDIMENTATION © k 'X RIVER ESTUARY SHELF '^L"'^L"^CJ\. \^l1""^1"'^L shoreface /////// t..»»»,»t RAPID TRANSGRESSION / / / / / / © SLOW TRANSGRESSION ^TTT^^ a-""^ / / / J / / / REGRESSION FIGURE 3. Sense of net sediment transport for (a) rapid trans- gression, (b) slow transgression, and (c) regression. Offshore com- ponent of transport is exaggerated for continuity. See text for ex- planation. roles of river mouth and shoreface bypassing (Fig. 3). Rapid transgression results in disequilibrium estuaries which become sediment sinks (see Chapter 14, Fig. 42 and associated text), and shoreface bypassing must dominate (Fig. 3a). The resulting deposits consist of a transient veneer of surf fallout on the upper shoreface, and the residual sand sheet of the lower shoreface and adjacent shelf (see Chapter 14, Fig. 8 and discussion, p. 265. These two deposits correspond to the nearshorc modern sand and shelf relict sand, respectively, of Emery (1968). Both deposits are relict in the sense that they have been eroded from a local, pre-Recent substrate, and both are modern in the sense that they have been redeposited under the present hydraulic regime. They are, in fact, palimpsest sediments (Swift et al., 1971) since they have petrographic attributes resulting from both the present and the earlier depositional environment. The term relict is best reserved for those specific textural attributes reflecting the earlier regime. Perhaps the most effective term for describing the relationship of these materials to the present depositional cycle is autochthonous (of local origin: Xaumann, 1858), and a shelf sedimentary regime characterized by rapid transgression and by- passing via shoreface erosion is described in this chapter as a regime of autochthonous shelf sedimentation. With a slower rate of translation (Fig. 3b), estuaries can equilibrate to their tidal prisms (see Chapter 14, Fig. 42 and associated text). River mouth as well as shoreface bypassing becomes a significant source of sediment. More subtle, but equally important, is the effect of a slow transgression on the grain size of bypassed sediment. With a slower rate of shoreline translation the intra- coastal zone of estuaries and lagoons can aggrade nearly to mean sea level. The resulting surface of salt marshes (or in low latitudes, mangrove swamps), threaded by high-energy channels, tends to serve as a 'low-pass, or bandpass filter, in the sense that the finer fraction of the sediment load is preferentially bypassed, while the coarsest fraction (and in the bandpass case, the finest fraction as well) is preferentially trapped out. In this process, migrating channels tend to select coarse ma- terials for permanent burial in their axes. The surfaces of the tidal interfluves receive the finest material for prolonged storage or permanent burial. However, fine sands and silts are deposited as overbank levees and tend to be reentrained by the migrating channels; hence they have the highest probability of being by- passed to the shelf surface. This material is sufficiently fine to travel in suspension for long distances. The estuaries of the Georgia coast have built a gently sloping shoreface of fine to very fine sand up to 20 km wide (Pilkey and Frankenberg, 1964; Henry and Hoyt, 1968); see Chapter 14, Fig. 13. This unusually wide and broad shoreface may be built by the combined contributions of shoreface and river mouth bypassing. Recent studies (Visher and Howard, 1974) suggest that the reversing tidal flows within the estuary consti- tute an efficient mechanism for the sorting of sands into size fractions, the spatial segregation of these fractions, and the bypassing of the finest sand out onto the shoreface. There is clearly a contribution of sand from shoreface erosion; however, shoreface sands, like the adjacent shelf sands, contain trace amounts of phosphorite (Pilkey and Field, 1972), indicating erosion of the Miocene strata which underlie the shoreface between the closely spaced estuaries, and which floor the deep scour channels of the estuary mouths (Barby and Hoyt, 1964). As the sense of coastal translation passes through stillstand to progradation (Fig. 3c), the shoreface becomes a sink rather than a mechanism for bypassing. Distributary mouths must further partition their pre- filtered load between sand sulliciently coarse to be captured by the littoral drift and buried on the shoreface, and sand fine enough to escape in suspension in the ebb 527 MODELS OF SHELF SEDIMENTATION 315 Channtl (C-F) Cloyey silt and silly cloy (C) - coortt gr Sond | (M) ~ madium V (F) - fin* gr 1VF)- vtry fin* gr. FIGURE 4. Schematic illustration of the depositional environ- ments and sedimentary Jacies of the Niger Delta and Niger shelf. Progressive size sorting of sediment results in a decrease in grain size through successive depositional environments in a seaward direction. From Allen (1970a). tidal jet, and be entrained into the shelf dispersal sys- tem. The shoreface behaves more nearly as a sediment trap, and bypassing occurs primarily through river mouths. The Niger-Benue delta system is one of the best studied examples of differential sediment bypassing through a prograding, deltaic environment (Allen, 1964, 1970a). The Niger-Benue river system delivers about 0.9 X 106 m3 of bed load sediment and about 16 X 106 m3 of suspended sediment (Allen, 1964) to its delta each year. During peak discharge from September to May, average flow velocities range from 50 to 135 cm sec, and gravel as well as sand is in violent trans- port. During low stages, flow velocities decrease to 37 to 82 cm sec, enough to transport sand and silt. In the higher part of the flood plain, the Niger is braided; in the rest the Niger shows large meanders (Fig. 4). During high stages, levees are overtopped, crevasse develops, and bottom lands are flooded. Gravel and coarse sand are deposited as a substratum of braid bars and meander point bars, respectively, and are veneered with a top stratum of overbank clays. Silt undergoes temporary deposition in levees in the lower flood plain but these tend to be undermined, so that their deposits reenter the transport system. Thus the flood plain environment serves as a skewed bandpass filter, with preferential bypassing of the medium and finer grades, preferential entrapment of the finest material over bank, and much coarse material being deposited in channel axes. This process continues through the tidal swamp environment, where the entrapment of fines dominates. Reversing tidal flows 528 316 CONTINENTAL SHELF SEDIMENTATION generate velocities of 40 to 180 cm, sec in tidal creeks, enough to move sand and gravel. Entrapment of fines overbank in the mangrove swamps is enhanced by the phenomena of slack high water and the prolonged period of reduced velocity associated with it. Fines then deposited begin to compact, and require greater veloci- ties to erode them than served to permit their deposition. Major channels, which pass through the intertidal environment to the sea, must store their coarser sedi- ment during low water stages at the foot of the salt wedge, where the landward-inclined surface of zero net motion intersects the channel floor. During high water stages, stored bottom sediment must be rhythmically flushed out of the estuary mouth by the tidal cycle. Sand coarser than the effective suspension threshold of 230 /u (Bagnold, 1966) will be deposited on the arcuate estuary mouth shoals, where, after a prolonged period of residence in the sand circulation cells of the shoal (see Chapter 10, p. 177), it leaks into the downcoast lit- toral drift system. Finer sand is entrained into suspension by large-scale top-to-bottom turbulence in the high- velocity estuary throat (Wright and Coleman, 1974) and will be swept seaward with the ebb tidal jet, to rain out on the inner shelf (Todd, 1968) where it is accessible to distribution by the shelf hydraulic regime. Shelves undergoing slow transgression or regression (Figs. 3b,c) thus experience a contrasting regime of allochthonous shelf sedimentation (Naumann, 1858) characterized by significant river mouth bypassing. In this regime there is a massive influx of river sediment whose grain size has been modified by passage through the coastal zone. Sheets of mobile fine sand and mud stretch from the coast toward the shelf edge. Shorefaces are broad and gentle and merge imperceptibly with a shallow inner shelf. Sedimentation on tectonic continental margins is a special case of allochthonous shelf sedimentation so distinctive as to warrant designation as a third and equal category. Shelves subject to such a regime are narrow and steep, if developed at all. River mouth bypassing and fractionation of the sediment load occur here also. Rubble subaerial fans may pass over short distances into sandy marine deltas with bottomset mud beds. Gravity dispersal becomes a significant coastal bypassing mechanism. Submarine canyons may cut completely across narrow shelves to tap the littoral drift (Shepard, 1973, p. 140) and divert sand seaward by slow or rapid mass movements. Where shelves are altogether lacking, coarse littoral prisms cascade inter- mittently down slopes that are nearly tectonic surfaces, to bathyal depths (Stanley, 1969). Tectonic regimes on incipient shelves are beyond the scope of this chapter, partly because they are more appropriately discussed in the chapter on slope sedimentation, and partly because of our ignorance, as this category is one of the last to be better known in the rock record (Stanley, 1969) than in modern environments. AUTOCHTHONOUS PATTERNS OF SEDIMENTATION Morphologic-Stratigraphic Patterns Shelves undergoing autochthonous sedimentation char- acteristically have a varied and systematic pattern of relief. The pattern tends to be correlated with both the distribution of surficial sediment and the internal structure of the surficial sediment mantle, and hence is a morphologic-stratigraphic pattern. On shelves of high relief, the pre-Holocene surface is exposed at the surface over wide areas, and constitutes an additional control of the pattern. Survival of Subaerial Patterns On high-latitude shelves, relief may exceed 200 m. Much of this relief may be the consequence of pre-Holocene fluvial and glacial erosion of a crystalline substrate (Holtedahl, 1940, 1958), and of the dissection of flat- lying or gently inclined Cenozoic strata into cuestas and plateaulike remnants. On the North American Atlantic shelf, the Fall Line, where turbulent piedmont streams pass onto the coastal plain strata, intersects the shoreline at New Jersey (Fig. 5). To the north, the Fall Line cuesta, of gently inclined coastal plain strata, forms first islands (Long Island, Nantucket), then offshore banks (Georges Banks, the Nova Scotian Banks). Basins landward of the drowned Fall Line (Long Island Sound, the Gulf of Maine, the Nova Scotian basins) have inner margins of crystalline rock thinly veneered with coarse detritus. The basin centers have a lower stratum of glacial lake deposits overlain by Holocene marine mud. Shelves of lower relief tend to be divided into broad, flat, plateaulike compartments by shelf valleys excavated during Quaternary low stands of the sea (Figs. 6 and 7). The outer margins of such shelves tend to consist of low- stand deltas, whose fronts are seaward-bulging shelf-edge scarps and whose landward margins may be marked by V-shaped, seaward-facing scarps that rise to the level of the inner shelf. Subaerial morphologic elements smaller in scale than cuestas, basins, and shelf valleys seem in general to have been destroyed by erosional retreat of the shore- face, and the larger scale elements have often been subtly but pervasively modified by this process. This point can usually be demonstrated by a comparison of 529 AUTOCHTHONOUS PATTERNS OF SEDIMENTATION 317 47° 67° 45°64c 38° 71° 70° 37° FIGURE 5. Bathymetry of the Gulf of Maine and Nova Scotian shelf. Dashed line is submerged extension of the Fall Line, separating gently dipping coastal strata from the crystalline substrate of the Appalachian orogenic belt. From Uchupi (1968). shelf morphology with the morphology of the associated subaerial surface. The coastal plains of the world bear a delicate fabric of high-stand scarps, separated by terraces overprinted with beach ridge fields, commonly dating from the last interglacial, or a high stand during the Wurm-Wisconsin glacial epoch (see, e.g., Colquhoun, 1969; Oaks and Coch, 1963; Bernard and Le Blanc, 1965). However, most submarine shelves are relatively featureless (the Aquitaine shelf: Caralp et al., 1972) or bear complex patterns of sand ridges that are the conse- quence of marine systems of sediment transport initiated after the passage of the shoreline (ridge and swale topography of Fig. 6). Major exceptions to this rule are the littoral bed forms of carbonate coasts; fringing reefs, beach rock, and calcarenite dunes cement as they form, and are far more resistant to the destruction during the passage of the shoreline. Carbonate littoral and sublittoral features have been reported from many shelves (Kaye, 1959; Ginsburg and James, 1974; Van Andel and Veevers, 530 <3 S T . Z ~ «■£ •a st ** s « * a. 3 318 531 3 1 vS 4 14 ■5 5 1 5 8 8 «• a. »4 S (*{{.'*, , , 818 DRILL HOLE DISCONFORMITY MEDIUM-FINE SAND PEBBLY SAND FINE, SILTY SAND SILTY CLAY A A' PLEISTOCENE-HOLOCENE CONTACT B B' TERTIARY-QUATERNARY CONTACT MOTTLED, DESICCATED, L^~ SILTY CLAY © RADIOCARBON DATE ON PROFILE FIGURE 8. Surficial sand of the inner New Jersey shelf. Stratum gressiou. The barrier superstructure is represented in this se- Hi is a shelf sand. Stratum H2 is a backbarrier sand. Stratum Hi quence by an unconformity; its forward face underwent continuous is a lagoonal mud. Thick zone in Hi is inferred to be a filled tidal erosional retreat (Chapter 14, Fig. 10A) and the resulting debris scour channel whose axis is normal to the plane of the diagram. The accumulated seaward of the shoreface as the leading edge of Hi. sequence was produced by coastal retreat during the Holocene trans- From Stahl et al. (1974). 533 NNW GULLY TROUGH SABLE ISLAND BANK AUTOCHTHONOUS PATTERNS OF SEDIMENTATION SSE 321 Maximum late-Wisconsin Low stand of seo level outwash plain 66 fbthom terrace 20,000 to 18.000 years BP ^SUSPENDED SEDIMENT ^ JL— /Pv-r ? m$ Cfer FIGL'RE 9. Evolution oj the surficial sand shelf on a glaciated shelf of appreciable relief {Nova Scotia shelf; compare with Fig. 5). From Stanley (1969)- tidal flow has served to reduce the relief of the buried subaerial surface. Tidal inlets scour trenches into the lagoonal stratum, then backfill these trenches as they migrate downdrift (Hoyt and Henry, 1967; Kumar and Sanders, 1970). The lagoonal carpet is itself discon- tinuous; Pleistocene beach ridges and other topographic highs protrude through the lagoonal deposits as penin- sulas or islands during their formation and are sheared off by shoreface retreat during passage of the main shore- line (Sheridan et al., 1974). On rocky coasts, lagoonal and estuarine deposits are confined to shelf valleys. Passage of the main shoreline results in destruction of the barrier and the upper part of the lagoonal sequence, and in the deposition of a second major stratum, a sheet of residual sand. This sand sheet overlies a surface of marine erosion whose areal geology is a patchwork of remnant lagoonal deposits and older substrate. On shelf sectors where the lagoonal carpet is well developed, this sand must travel from eroding headlands along the shoreface of retreating barriers, before being spread over the lagoonal carpet; or it is released as the retreating shoreface cuts into tidal inlet fills, or into estuarine channel sands scoured out of the pre-Recent substrate (Andrews et al., 1973). On shelves with poorly developed lagoonal strata, the retreating shoreface may be incised all the way through the lagoonal deposits and into 534 322 CONTINENTAL SHELF SEDIMENTATION (HO)SUBAERIAL GRAVEL Qjjj MAINLAND MARSH (H2) TIDAL CREEK LAG GRAVEL (H3)LAGOONAL SANDY MUD (H4) MARSH, PEAT, WASHOVER SAND (H5)lNLET SAND (H6) DUNE, WASHOVER SAND (H8) BASAL SHELF GRAVEL (H9) SHELF SAND (ni^) SHELF MUD (Pj) PLEISTOCENE LAGOONAL SANDY MUD (P2) PLEISTOCENE BARRIER SAND (P3) PLEISTOCENE INNER SHELF MUD (H7) BEACH, SHOREFACE SAND FIGURE 10. Stratigrapbic model for a low coast undergoing erosional shoreface retreat, Pleistocene sands, whose erosion provides material for the surficial sand sheet. The basal layer of the surficial sand sheet is a thin. discontinuous gravel (Powers and Kinsman, 1953; Belderson and Stride, 1966; Yeenstra, 1969; Xorris, 1972) or shell hash rich in backbarrier and beach species (Fischer, 1961; Merrill et al., 1965; Milliman and Emery, 1968; Field, 1974). More exotic clasts are clay pebbles eroded from Early Holocene lagoonal deposits, elephant teeth (Whitmore et al., 1967), and concretions from Tertiary strata (Stanley et al., 1967). The basal gravel is rarely more than a meter thick. It is SUCCESSIVE POSITIONS OF SHORE EACE DURING INTERMITTENT TRANSGRESSION TRANSGRESSION (7-11) AND BARRIER RETREAT STIUSTAND 16,7) WITH BARRIER NOURISHMENT AND UPWARD GROWTH TRANSGRESSION (1-6] AND BARRIER RETREAT LAGOON TIDAL CHANNEL BARRIER RIDGED SAND SHEET (DESTRUCTIONAll TRUNCATED BARRIER SAND LAGOONAL DEPOSITS PRE-HOLOCENE SUBSTRATE FIGURE 1 1. Above: Schematic illustration of intermittent shoreface retreat. As shoreface profile translates primarily landward in response to rising sea level, material eroded from shoreface accumulates on adjacent shelf as ridged sand sheet. Periods of primarily vertical translation of profile followed by periods of resumed landuard trans- lation result in truncated scarp. Below: Resulting stratigraphy. From Swift et al. (197 i). 535 AUTOCHTHONOUS PATTERNS OF SEDIMENTATION 323 commonly overlain by 1 to 10 m of sand, with a sub- littoral molluscan fauna (Shideler et al., 1972). terraces and scarps. Most continental shelves are terraced, with terraces separated by scarps 10 m or more in height (Fig. 6). Their counterparts on the adjacent subaerial coastal plains mark Quaternary (or earlier) high stands of the sea. Shelf scarps appear to have resulted primarily from stillstands of the returning Holocene sea, although they may in some cases be reoccupied Pleistocene shorelines. On the Georgia coast, for example, the modern barrier island system is perched on the forward face of a Pleistocene shoreline, and the modern tidal inlets are reoccupied Pleistocene inlets (Hoyt and Hails, 1967). Shelf scarps are drowned shorelines only in the broadest sense; more specifically, they are relict lower shorefaces. To form such a scarp would require a period of near stillstand during a general transgression. The shoreface profile will translate more nearly upward than landward (Fig. 11) during this period, by means of upper shoreface and barrier surface aggradation. At the resumption of rapid transgression, the super- structure of the stillstand barrier will resume its land- ward migration through the process of shoreface erosion and storm washover, leaving behind a truncated lower shoreface. If both the lower and upper shoreface undergo aggradation during the stillstand period, so that the ideal profile is realized at all times, then there will be no surface expression of the stillstand shoreline, although seismic profiles may reveal a buried scarp (Stanley et al., 1968). SHELF VALLEY COMPLEXES AND SHOAL RETREAT MASSIFS. A second nearshore marine pattern of relief and sedi- ment distribution that may survive from the nearshore environment during a marine transgression is a shelf valley complex. This term refers to the groups of morphologic elements that occur along the paths of retreat of estuary mouths on autochthonous shelves. Shelf valley complexes are composed of deltas, shelf valleys, and shoal retreat massifs (Figs. 6 and 12). A shoal retreat massif is a broad, shelf-transverse sand ridge of subdued relief that marks the retreat path of a zone of littoral drift convergence (Swift et al., 1972). It may be dissected by subsequent storm or tidal flows into a cross-shelf sequence of smaller coast-parallel ridges, and the term massif is used in the sense of a composite topographic high, itself consisting of smaller highs. Such complexes are locally well developed on low- relief shelves such as the Middle Atlantic Bight of North America. Here they are largely constructional features molded into the Holocene sand sheet. The sand MODERN ESTUARY MOUTH SHOAL, TIDAL CHANNELS PAIRED FLOOD CHANNEL RETREAT TRACK, ESTUARINE SHOAL-RETREAT MASSIF 40M SCARP TRANSGRESSED CUSPATE DELTA; (CAPE SHOAL- RETREAT MASSIF) 60M SCARP FIGURE 12. The Delaware shelf valley complex, Delaware shelf of North America. Southward littoral drift of New Jersey coastal compartment is injected into reversing tidal stream of mouth of Chesapeake Bay. The resulting shoal is stabilized as a system of interdigitating ebb and flood channels, north of the main couplet of a mutually evasive ebb and flood channel. The shelf valley complex seaward of the bay mouth is the retreat path of the bay mouth sedimentary regime through Holocene time. Retreat of the main flood channel has excavated the Delaware shelf valley; retreat of the bay mouth shoal has left a seaward-trending shoal retreat massif on the shelf valley's north flank. From Swift (1973). sheet tends to completely fill the former subaerial valley cut by the river into the Pleistocene strata, and the shelf valley complex and the buried river channel may not everywhere coincide (Fig. 13). Shelf valley complexes are built in serial fashion by the retreating shoreline. It is important to remember that the last high-energy depositional environment ex- perienced by any given segment was the nearshore zone. As a consequence of remolding of preexisting deposits in this zone, elements of shelf valley complexes are not always what they seem. For instance, in Fig. 12, the topographic characteristics of the Delaware midshelf 536 324 CONTIN KNTAI. SIIKI.h SEDIMENTATION REHOBOTH BEACH SECTION HEN ♦ CHICKENS SHOAL a. ED LAGOONAL MUDS, CLAYS (H) ESTUARINE - SHALLOW MARINE SILTS E3 NEARSHORE MARINE SANDS 55 GRAVELS Q SHELLS, ENSIS, MULINIA 60" 200" DISTANCE (NAUTICAL MILES) 6 7 B 9 DISTANCE (KILOMETERS) FIGURE 13. Section across the head of the Delaware shelf valley complex based on vibracores and a 5.5 kHz seismic profile. Marine sand sheet with constructional tidal topography rests on Holocene lagoonal and older Pleistocene deposits. Delaware shelf valley occurs entirely within Pleistocene sands. Note offset between shelf valley and buried river channel. Prom Sheridan et al. (1974). delta suggest that the surface of this stillstand feature was successively remodeled at the resumption of transgression, first as a retreating cuspate foreland, then as cape shoal retreat massif, as illustrated in Chapter 14, Fig. 22. After a further period of stillstand indicated by a 60 m scarp, the coastal regime again changed, and the Delaware River mouth resumed retreat, this time as an estuary. The retreat path of this estuary mouth consists of a sharply de- fined submarine channel (shelf valley) flanked by a shoal retreat massif. The origins of these two features are easily deduced from uniformitarian reasoning. The shoal retreat massif may be traced into the modern north side shoal of the Delaware estuary mouth. This shoal is a sink for the littoral drift of the New Jersey coastal compartment, and is stabilized by a system of interdigitating ebb and flood channels. The shelf valley may be traced into the flood channel of a large ebb channel-flood channel couplet on the south side of the estuary mouth that accommodates most of the tidal discharge. On the central and southern Atlantic shelf of North America, four basic morphologic provinces may be described on the basis of constructional morphologic elements inherited from the retreating shoreline (Fig. 14). In the Middle Atlantic Bight (Fig. 6), widely spaced master streams have resulted in widely spaced shelf valley complexes. The plateaulike intcrfluves between the shelf valley complexes bear ridge fields that were also generated by shoreface retreat (see Chapter 14, Fig. 28). The more intense wave climate experienced by the Carolina salient has elicited a different response from the retreating river mouths. Capes Romain, Fear, Lookout, and Hattcras may have originally been cuspate deltas, associated with the Peedee, Cape Fear, Neuse, and Pamlico rivers (Chapter 14, Fig. 26). Retreat of these forelands has left large widely spaced shoal retreat massifs. South of Cape Romain the retreat of small, closely spaced cuspate forelands has generated a blanket of coalescing shoal retreat massifs on the adjacent shelf 537 A tTOCHTHONOUS PATTERNS OF SEDIMENTATION 325 CAPE COD SUSPENDED- SEDIMENT DISCHARGE DEPOSITIONAL PROVINCES SHELF VALLEY COMPLEX AND SHOREFACE RETREAT BLANKET CAPE RETREAT MASSIF AND SHOREFACE RETREAT BLANKET CAPE RETREAT BLANKET ESTUARY RETREAT BLANKET CAROLINA SALIENT WAVE CLIMATES = >40% >5 FT. — >30% >5 FT. ••• >20% >5 FT SOUTHERN ATLANTIC BIGHT \ 100 KM FIGURE 14. Coastal sediment discharge (Meade, 1969) and wave climate of the Middle Atlantic Bight (Do/an et a/., Wl) and resulting depositional provinces. From Sui/t and Sears (W4)- i Fie. 13). Vet further south, the Georgia Bight ex- periences a high tide range, a milder wave climate, and the closely spaced river mouths are estuarine in con- figuration. Their retreat has generated a blanket of coalescing shelf valley complexes (Fig. 16). Initiation of Modern Patterns TEXTL'RAL AND MORPHOLOGIC PATTERNS ON A STORM- DOMiNATF.D shllf. On two of the best studied autoch- thonous shelves, the Middle Atlantic Bight of North America and the shelf around the British Isles, the hydraulic climate is sufficiently intense to overprint older subaerial and nearshore marine patterns of the surficial sand sheet with a modern textural and morpho- logic pattern. In the Middle Atlantic Bight fair-weather flows are driven by the geostrophic response of the stratified shelf water column to freshwater runoff and to winds (MeClenncn. 1973; Bumpus. 1973); see Fig. 17. How- 538 FIGURE 1 5. Cuspate forelanch ami cape shoal-retreat massifs (stippled) of the South Carolina shelf. Sote overprinting by ridge and swale topograph). Contours in fathoms. From Swift et al. (19~2). 326 FIGURE 16. Morphologic pattern of estuarine shoal retreat blanket, overprinted by ridge and swale topograph}. South Carolina coast. Highs are stippled. Contours in fathoms. Irom Swift and Sears (IT 4). 539 AUTOCHTHONOUS PATTERNS OF SEDIMENTATION 327 39°30 - 39-00 38-30' 74-30' 74-00 73-30' 73-00' 72-30' FIGURE 17. Fair-weather hydraulic regime of the Sew Jersey shelf as indicated by Savoniits rotor current meters mounted 1.5 to 2.0 m above the seafloor at jour stations on the New Jersey shelf, for periods of 9 to 11 days in late spring. Progressive vector diagrams indicate a general southerly water drift, partly correlatable with wind directions (McC/en- nen, 7973). Loops, spikes, and bulges on progressive vector diagrams are modulation by the semidiurnal tide. In percent exceedence diagrams, current velocities are compared with bottom wave surge calculated from wave climate data. All data from McC/ennen (1973). ever, neither these unidirectional flow components nor the superimposed wave oscillations (McClennen, 1973) and tidal oscillations (Redfield, 1956) are strong enough to result in significant bed load transport over broad areas. During the winter period of frequent storms, the water column is not stratified, and air-water coupling is more efficient (see discussion, pp. 263-264). The geometry of the Middle Atlantic Bight is especially conducive to strong flows during this period. When low-pressure systems pass over the bight, so that the isobars of atmospheric pressure parallel the isobaths of the shelf surface, the resulting winds blow southward down the length of the Middle Atlantic Bio;ht, parallel- ing the curve of the shoreline, and induce a uniform setup of the shelf water mass against the coast of 40 to 60 cm. High-velocity "slablike" flows of remarkable longshore coherence result (Beardslcy and Butman, 1974; Boicourt, personal communication). The coastal boundary of these storm flows appears to initiate ridge topography at the foot of the shorefacc (I)uane et al., 1972); see Chapter 14, Figs. 28 and 31. However, it is clearly an oversimplification to describe the ridge topography of the Middle Atlantic Bight as a purely inherited topographic pattern. The ridges maintain their characteristic 10 m relief and textural patterns across the central shelf (Swift et al., 1974); see Fig. 18. Troughs retain erosional windows in which Early Holocene lagoonal clays are thinly veneered with a lag deposit of pebbly sand. Calculations based on current-meter records suggest that the unidirectional components of storm flows arc sufficient to mobilize the sandy bottom (Fig. 19), and to slowly level the ridge topographs', if the topography were not in fact a continuing response of the seafloor to the modern hydraulic climate. In several areas on the Middle Atlantic shelf, there is evidence to suggest that ridge topography may be initiated on the central shelf, if not already present as a survival from the nearshore environment. Off South Carolina, shoal retreat massifs are overprinted by a ridge topograpfiy even though the modern nearshore zone is not apparently forming ridges (Fig. 15). Else- where, the ride;e pattern appears to have changed as the water column deepened and the shoreline receded during the course of the Holocene transgression. The estuary mouth shoal that is the landward end of the Delaware Massif (Fig. 12) has impressed into it a tide-maintained ridge pattern that trends normal to the shoreline and parallel to the sides of the estuary mouth. As the crest of the massif is traced seaward, the trend of the ridges and troughs superimposed on it shifts toward a shore-parallel orientation. The bay 540 328 CONTINENTAL SHELF SEDIMENTATION 39°I0'N 39°05' 74°00' FIGURE 18. Distribution of grain sizes on the central New Jersey shelf. Medium to fine sands occur on ridge crests. Fine to very fine sands occur on ridge flanks and in troughs. Locally, 73°45' W erosional contours in troughs expose a thin lag of coarse, shelly, pebbly sand over lagoonal clay. From Stubblefield et al. (in press). mouth ridges are oriented parallel to the reversing tidal flows of the bay mouth; the offshore ridges appear to par- allel instead the geostrophic storm flows of the open shelf. The Great Egg Massif, associated with the former course of the Schuylkill River across the shelf, has been heavily dissected into a transverse ridge pattern. Seaward of a scarp whose toe lies at 90 m, a second, small-scale ridge pattern with a somewhat different trend has been superimposed on the first (Fig. 20). Stubblefield and Swift (1975) have presented a model for the evaluation of the compound ridge pattern based on vibracores. and 3.5 kHz seismic profiles collected in the area (Fig. 21). Radiocarbon dates indicate that the large-scale ridges appear to have formed immediately subsequent to the passage of the shoreline at approximately 11,000 BP (Fig. 2\A). Internal stratification indicates that large-scale ridges grow by the accretion of conformable beds. Wide, large-scale troughs appear as zones of bare Pleistocene substrate, where the surficial sand sheet was never formed, or where its material was swept away to nourish the growth of adjacent ridges. With continuing transgression and deepening of the water column, the ridges appear to have increased their spacing by means of lateral migration or the coalescence of adjacent ridges. Internal strata tend to dip more steeply than present ridge flanks, suggesting that toward the latter part of their history, ridge growth was mainly the consequence of lateral rather than vertical accretion Small-scale troughs transect large-scale ridges, and tend to break large-scale ridges up into en echelon seg- ments (Fig. 20). Where small-scale troughs cross large- scale troughs they are incised into the flat-lying Early Holocene and Pleistocene strata that floor the large- scale troughs. Small-scale troughs are commonly narrow features that do not penetrate through the Early Holocene lagoonal clay (Fig. 2\B). Where this clay is in fact breached, so that the small-scale troughs penetrate the underlying sand, the troughs are notice- ably wider, as though they had expanded by under- cutting of the clay in a fashion analogous to the growth of a blowout on a grass-covered eolian flat (Fig. 21C). The ridge topography of the Middle Atlantic Bight is accompanied by mesoscale bed form patterns, whose relationship to the ridge pattern is not clearly under- stood. The most ubiquitous mesoscale bed forms are the current lineations, which occur as sand ribbons or more 541 AUTOCHTHONOUS PATTERNS OF SEDIMENTATION 329 75°48' /S'M.V • / I • • • • 36°32' SIZE CLASS COARSE SAND MEDIUM SAND CLASS % VOLUME BOUNDARIES EXCEEDENCE TRANSPORT jFINE SAND jVERY FINE J SAND (PHI) 0-1 1-2 2-3 3-4 6.0 13.0 15.8 17.0 (m3/m/T) 3.4 8.0 75 5.2 30 u u 20 S 10 CURRENT METER STATION • STATIONS J NET TRANSPORT DIRECTION 12 16 DAYS 20 24 28 FIGURE 19- Sediment transport in response to the unidirectional component of flow during the month oj November 1972, in an inner shelf ridge field, False Cape, Virginia. Estimates based on Shield's threshold criterion, a drag coeffi- cient of i X 10~3, and Laursen's (2958) total load equation. Values expressed as cubic meters of quartz per meter transverse to transport direction for time elapsed. Solid line is the 10 m isobath. commonly as linear erosional furrows. They may trend parallel to the trend of the ridge topography, or may cut across it, so as to make a larger acute angle with the shoreline (Fig. 22). Toward the southern end of the Middle Atlantic Bight, the shelf surface shoals, narrows and curves to the east. Sand wave fields appear, perhaps indicative of the acceleration of storm flows in response to the decreasing cross-sectional area of the shelf water cokimn. Ridges molded into the Albermarle shoal retreat massif bear sand waves on their crests (Fig. 23). Sand waves locally attain 2 m heights and angle of repose slopes. Sand wave crestlines are not quite normal to shore, suggesting that the ridge crests on which they are found experience a seaward component of flow during storms. At Diamond Shoals, the southern extremity of the Middle Atlantic Bight, sand waves up to 7 m high occur between sand ridges, forming a reticulate pattern (see Fig. 27, Chapter 16). Grain-size patterns in the Middle Atlantic Bight suggest that the storm flows that interact with the ridge topography and the mesoscale bed forms are capable of transporting at least the finer grades of sand for appreci- able distances. The inner shelf sectors before the seaward- convex coastal compartments of the Middle Atlantic Bight exhibit a repeating pattern of grain-size distribu- tion (Fig. 24). The northern half of each of these inner shelf sectors, where south-trending storm flows must 542 330 CONTINENTAL SHELF SEDIMENTATION 39° 00 NX 38° 45'NN 7^T 74°00 W FIRST ORDER HIGHS CRESTLINES. SECOND ORDER HIGHS FIGURE 20. Great Egg shelf valley and shoal retreat massif. Large-scale ridges in inset may date from a period when the ancestral Great Egg estuary was active. Nearshore large-scale ridges were probably formed by shoreface de- tachment during erosional retreat of the shoreface, after capture of the ancestral Schuylkill River by the Delaware River, and consequent reduction in discharge of the Great Egg estuary. See Fig. 6 for relationships of Schuylkill, Delaware, and Great Egg rivers. From Stubblefield and Swift (in press). presumably converge with the shoreline, tend to be floored with primarily medium- and coarse-grained sands, molded into a well-defined ridge topography. On the southern halves of the coastal compartment, where the shoreline tends to curve to the west, storm flows might be expected to expand and decelerate. Here the fine sand blanket of the shoreface extends across the inner shelf floor, as though nourished by material swept out of the ridge topography to the north. The schematic flow pattern in the lowest panel of Fig. 24 is not basic on detailed observations. It is intended to indicate that current flowing generally southwest parallel to the long dimension of the shelf will tend to converge with the northeastern portion of the shoreline and diverge from the southwestern portion of the shoreline. A somewhat closer relationship appears to exist between flow geometry and sediment distribution in the vicinity of the shoal retreat massifs (Fig. 25). The ridge topography attains its maximum relief where it has been molded onto the crests of the massifs. The massifs do not exhibit bilateral symmetry; troughs are deepest and widest on the northern sides. As a trough axis is traced across the massif, erosional windows exposing the basal pebbly sand or the underlying clayey sub- strate become less frequent. The fine sands of the trough flanks tend to bridge across the trough floor. 543 AUTOCHTHONOUS PATTERNS OF SEDIMENTATION 331 ® PRIMARY RIDGE PRIMARY TROUGH PRIMARY RIDGE ■HOLOCENE SILTY CLAY -^-PLEISTOCENE SAND :':>:o:::o: ::>■ -'^ ' ■ ' T- ■"' ^m* -«- PLEISTOCENE SILTY CLAY PRIMARY RIDGE DEVELOPMENT OF RIDGE TOPOGRAPHY FIGURE 21. Ridge evolution on the central New Jersey shelf. (A) Ridge nuclei are formed during the process of ridge detachment and shoreface retreat, or by other means in the nearshore zone. Sand continues to be swept out of troughs onto ridges as water column deepens during the course of the Holocene transgression. (B) Seafloor scour during storms locally penetrates the Early Holocene lagoonal clay carpet, and a secondary trough forms, initially by downcutting. (C) Downcutting in the secondary trough decreases and lateral erosion increases as second silly clay layer is exposed. Secondary trough widens by undercutting of upper clay in "blowout" fashion. Sand from similar excavations upcurrent forms secondary ridges. From Stubblefield and Swift (in press). The trough axis tends to climb toward a low sill on the southern side of the massif; beyond this the seafloor drops off rapidly to the adjacent shelf valley. The valley floor commonly consists of fine to very fine featureless sand. The topography and grain-size pattern suggest that south-trending flows converge with the rising sea- floor and accelerate up the northern flanks of the massifs. Fine sand swept out of the troughs is deposited in the zone of flow expansion and deceleration over the shelf valley south of the massif. TEXTURAL AND MORPHOLOGIC PATTERNS ON A TIDE- DOMINATED shelf. The tide-swept shelf around the British Isles (Stride, 1963) provides an interesting con- trast with the storm-induced sedimentation of the Middle Atlantic Bight. Surges are at least as frequent here as in the Middle Atlantic Bight (Steers, 1971). However, much more work is done on the seafloor by the semidiurnal tidal currents associated with the amphidromic edge waves that sweep the margins of mar- ginal shelf seas of western Europe (see Chapter 5, Fig. 4 and discussion, p. 60). The rotary tidal currents associated with these tidal waves are in fact analogous in some respects to the inertial wind-driven currents generated by storms. Midtide surface velocities in excess of 50 cm/sec (1 knot) are sustained over vast areas, and locally exceed 200 cm/sec. Ebb-flood dis- charge differentials result in currents residual to the tidal cycle, whose velocities may be as great as a tenth of the midtide value. As a consequence of the higher rate of expenditure of energy on the seafloor, morphologic and textural patterns inherited from the retreating nearshore zone have been largely erased. Erosional shoreface retreat has resulted in a surficial sediment sheet that is com- parable in many respects to that of the Middle Atlantic Bight (see Belderson and Stride, 1966). However, the poorly resolved sand transport patterns of the Middle Atlantic Bight are replaced by well-defined transport paths, with sand streams that diverge from beneath 544 332 CONTINENTAL SHELF SEDIMENTATION 39°05'N 73°58W 73°57' 73°56' TREND OF BOTTOM UNEATION TRACKLINE FIGURE 22. Current I in eat ion patterns on the central New Jersey shelf. Bars indicating lineations are over 10 times as long as features that they rep- resent. They locally represent sets of lineations. High areas are stippled. Contours in fathoms. From McKinney et al. (2974)- tide-induced "bed load partings" and flow down the gradient of maximum tidal current velocities until either the shelf edge or a zone of "bed load conver- gence" and sand accumulation is reached (Stride, 1963; Kenyon and Stride, 1970; Belderson et al., 1970); see Fig. 26. Each stream tends to consist of a sequence of more or less well-defined zones of characteristic bottom mor- phology and sediment texture (Fig. 27). Streams may begin in high-velocity zones [midtide surface velocities in excess of 3 knots (150 cm/sec)]. Here rocky floors are locally veneered with thin (centimeters thick) lag deposits of gravel and shell. Where slightly thicker, the gravel may display "longitudinal furrows" parallel to the tidal current (Stride et al., 1972), a bed form related to sand ribbons (see Chapter 10, p. 170). Between approximately 2.5 and 3.0 knots (125-150 cm/sec) sand ribbons are the dominant bed form (Kenyon, 1970). These features are up to 15 km long and 200 m wide, and usually less than a meter deep. Their materials are in transit over a lag deposit of shell and gravel. Kenyon has distinguished four basic pat- terns that seem to correlate with maximum tidal current velocity and with the availability of sand (Chapter 10, Fig. 15). Further down the velocity gradient, where midtide surface velocities range from 1 to 2 knots (50-100 cm/sec), sand waves are the dominant bed form. Where the gradient of decreasing tidal velocity is steep or transport convergence occurs, this may be the sector of maximum deposition on the transport path. Over 20 m of sediment has accumulated at the shelf-edge convergence of the Celtic Sea, although it is not certain that this sediment pile is entirely a response to modern conditions. The Hook of Holland sand wave field off the Dutch coast is one of the largest (15,000 km2) and the best known (McCave, 1971). The sand body is anomalous in that it sits astride a bed load parting; the sand patch as a whole may be a Pleistocene delta or other relict feature. Sand waves with megaripples on their backs grow to equilibrium heights of 7 m with wavelengths of 200 to 500 m in water deeper than 18 m; in shoaler water, wave surge inhibits or suppresses them. Elongate tidal ellipses favor transverse sand wave formation, and the sand waves tend to be destroyed by midtide cross 545 AUTOCHTHONOUS PATTERNS OF SEDIMENTATION 333 FIGURE 2 3. Sand ridges with superimposed sand waves on the northern North Carolina inner shelf. Topographic highs are stippled. Contours in feet. From Swift et ah (1973). flow when the ellipse is less symmetrical. Under the latter condition, linear sand ridges may be the pre- ferred bed form, as midtide cross flow would tend to nourish rather than degrade them (Smith, 1969). The triangular sand wave field is limited by a lack of sand on the northwest, by shoaling of the bottom and in- creasing wave surge on the coast to the south, and by fining of sand to the point that suspensive transport is dominant to the north (McCave, 1971). Further down the velocity gradient, beyond the zones of obvious sand transport, there are sheets of fine sand and muddy fine sand and in local basins, mud. They lack bed forms other than ripples, and appear to be the product of primarily suspensive transport (McCave, 1971) of material that has outrun the bed load stream (see discussion, Chapter 10, p. 160). These deposits may be as thick as 10 m (Belderson and Stride, 1966), but where they do not continue into mud, they break up into irregular, current-parallel or current-transverse patches of fine sand less than 2 m thick, resting on the gravelly substrate. The complex pattern and mobile character of the shelf floor around the British Isles have led British workers to reject the relict model for the shelf sediments (Belderson et al., 1970). They note that it correctly draws attention to the autochthonous origin of the sediment, but that it fails to allow for its subsequent dynamic evolution. They propose instead a dynamic classification: 1 . Lower sea-level and transgressive deposits, patchy in exposure, but probably more or less continuous beneath later material; largely the equivalent of a blanket (basal) conglomerate. 546 76° 37' 76° VERY COARSE TO MEDIUM SAND FINE SAND .VERY FINE SAND, I SILTY CLAY WOODS HOLE DATA VA. INST. MAR. SCI. DATA LITTORAL DRIFT STORM DRIVEN CURRENTS ♦ ■" TIDAL CURRENTS FIGURE 24. Above: Bathymetry of the Delmarva inner direction of currents responsible for bed load sediment shelf. From L'chupi (1970). Center: Distribution of sediment. transport. Reproduced from Suift (1975). From Hathaway (1971) andSichols (1972). Belou: Inferred 334 547 ALLOCHTHONOUS PATTERNS O F SE DI M E N T A TI O N 335 76° OO'W % ; GRAVEL :■:■:■:■ coarse sand : : (INSET: 1.0-1.5 + ) □ MEDIUM SAND (INSET: 2.0-2.5*1 FINE SAND (INSET: 2.0-2.5*) 75° 29 W xxvxxv VERY FINE tt SAND (INSET: 2.5-3.0*) FIGURE 2 5. Grain-size distribution on a portion of the Virginia Beach massif, and adjacent shelf valley. 2. Material moving as bed load (over the coarser basal deposits) mainly well sorted sand and in places first-cycle calcareous sand. 3. Present sea-level deposits (category 2 sediment having come to permanent rest) consisting of large sheets to small patches, which range from gravel and shell gravel to sand and calcareous sands, muddy sands, and mud. The implication is that of a shelf surface moving toward a state of equilibrium with its tidal regime. The degree of adjustment appears to be greater than in the case of the North American Atlantic Shelf, in that there is less preservation of nearshore depositional patterns. As a consequence of the intensity of the hydraulic climate, there is less on shelf storage (category 3) and more material in transit. Locally, sand ridges similar to those of the Middle Atlantic Bight do occur. Like those of the Middle Atlantic Bight, they tend to be grouped in discrete fields. In some cases, it is possible to infer that these ridge fields are in fact shoal retreat massifs, generated by the retreat of a near shore depositional center during the course of the Holocene transgression (Swift, 1975). The clearest case may be made for the Norfolk Banks (Houbolt, 1968; Caston and Stride, 1970; Caston, 1972); see Fig. 28. Here a series of offshore sand ridges may be traced into a modern nearshore generating zone (Robinson, 1966; see Chapter 14, Fig. 39) where sand is packaged by the specialized tidal regime of the shoreface into shapes hydrodynamically suited for survival on the open shelf (see discussion, p. 180). The Nantucket Shoals sector of the North American Atlantic shelf appears to constitute a similar evolu- tionary sequence of ridges (see Chapter 10, Fig. 30). The Norfolk Banks are analogous to the cape shoal retreat massifs of the Carolina coast of North America, in that the generating zone is a coastal salient that serves as a sink for the nearshore sand flux. Other, more poorly defined ridge fields in the southern bight of the North Sea (Fig. 28) may be analogous to the estuarine shoal retreat massifs of the Middle Atlantic Bight in that they may have been generated by the retreat of the ancestral Rhine and Thames estuaries. ALLOCHTHONOUS PATTERNS OF SEDIMENTATION Shelves undergoing allochthonous sedimentation differ from autochthonous shelves in a variety of character- 548 336 CONTINENTAL SHELF SEDIMENTATION 1 FIGURE 26. Generalized sand transport paths around the British Isles and France, based on the velocity asymmetry of the tidal ellipse and the orientation and asymmetry of bed forms. From Kenyon and Stride (1970). istics. The most obvious is that allochthonous shelves tend to be floored by fine sands, fine muddy sands, or muds that have escaped from adjacent river mouths: autochthonous shelves in contrast are generally covered by coarser grained sand of local origin. Although sur- faces of allochthonous shelves are constructional in nature, they tend to be smooth and featureless; their fine materials have traveled primarily in suspension, and the effective underwater angles of repose of the sediment may be too low to result in such large-scale bed forms as sand waves or sand ridges. However, such features are not totally unknown. Allersma (1972) has reported "mud waves" from the Venezuelan shelf that appear to be very similar to the shoreface-connected ridges of the Middle Atlantic Bight. Transport on Allochthonous Shelves Mechanisms of sediment transport on allochthonous shelves have been generally described by Drake in Chapter 9. Since this chapter stresses regional transport patterns, it seems worthwhile to summarize Drake et al.'s (1972) study of river-dominated sedimentation on the southern California shelf. This carefully docu- mented, real time study of the dispersal of flood sedi- ment is probably the most detailed report on the nature of allochthonous sediment dispersal available at the time of writing. In January and February of 1969, southern California experienced two intense rainstorms which resulted in a record flood discharge. The freshly eroded sediment was a distinctive red-brown in contrast to the drab hue of ZONE I BEDROCK & GRAVEL ZONE II SAND RIBBONS ZONE III SAND WAVES ZONE IV SMOOTH SAND ZONE V SAND PATCHES T^P^ DECREASING MID-TIDE SURFACE VELOCITY; BOTTOM GRAIN SIZE FIGURE 27. Succession of morphologic provinces along a tidal transport path. Based on Belderson et al. (1970). 549 AUTOCHTHONOUS PATTERNS OF SEDIMENTATION 337 FIGURE 2 8. Tidal ridge fields of the southern bight of the North Sea. Northernmost ridge field appears to constitute a shoal retreat massif, marking the retreat path of the nearshore tidal regime of the Norfolk coast. Ridge fields in the approach to the English Channel may have been initiated in an earlier, more nearly estuarine environment, from Houbo/t (1968). the reduced shelf sediments. The flood deposit on the shelf could therefore be repeatedly cored and isopached, and its shifting center of mass traced seaward through time. USGS stream records show that 33 to 45 X 106 metric tons of suspended silt and clay and 12 to 20 X 106 metric tons of suspended sand were introduced by the Santa Clara and Ventura rivers. By the end of April 1969, more than 70% of this material was still on the shelf in the form of a submarine sand shoal extending 7 km seaward, and a westward-thinning and -fining blanket of fine sand, silt, and clay existed seaward of that (Fig. 29.4). By the end of the summer of 1969, the layer extended further seaward, had thinned by 20%, and had de- veloped a secondary lobe beneath the Anacapa current to the south (Fig. 29B). Eighteen months after the floods, the surface layer was still readily detectable. Considerable bioturbation, scour, and redistribution had occurred south of Ventura, but the deposit was more stable to the north (Fig. 29C). A concurrent study of suspended sediment distribution in the water column revealed the pattern of sediment transport (Fig. 29D). Vertical transparency profiles, after four days of flooding, showed that most of the suspended matter was contained in the brackish surface layer, 10 to 20 m thick. Profiles in April and May revealed a layer 15 m thick, with concentrations in excess of 2 mg 1, and a total load of 10 to 20 X 104 metric tons. Since this load was equal to river discharge for the entire month of April, it must have represented lateral transport of sediment resuspended in the near- shore zone. Vertical profiles over the middle and outer shelf for the rest of the year were characterized by sharply bounded turbidity maxima, each marking a thermal discontinuity. These also were nourished by lateral transport from the nearshore sector where the discontinuities impinged on the sloping bottom. The near-bottom nepheloid layer was the most turbid zone in the inner shelf. This nepheloid layer was invariably the coolest, and was invariably isothermal, indicating that its turbidity was the result of turbulence generated 550 34°20' - 34°10' - 34°20'- 34°10 34°20' 34°10' _L_2 S ■ • ' 119°50' 119°30' 119°10' (d) T3S<1 T3343 T3SS0 13588 13586 13372 sot- -"70 -1 - - . 70 - -., e° e-i FIGURE 29. Thickness of flood sediment (centimeters) on the Santa Barbara-Oxnard shelf in (a) March-April 1969; (b) ,M«>-v4»£».r/ i969/ and (c) February-June 1970, based on cores, (d) East-west cross section showing vertical distribution of light-attenuating substances over Santa Barbara-Oxnard shelf. For clarity, the bottom 20 m of the water column is not contoured, but the percent transmission value at the bottom is noted. From Drake et al. (1972). 338 551 J^^y 6' ■ LLOCHTHONOUS PATTERNS OF SEDIMENTATION 339 YOUNGER SUITE COARSEST GRADE FINE SAND AND COARSER VERY FINE SAND CLEAN VC. SILT CLAYEY SILT SILTY CLAY 5° OLDER SANDS V. COARSE SAND COARSE SAND MEDIUM SAND FINE SAND VERY FINE SAND FIGURE 30. Distribution of sediments on the Niger shelf. Young suite is of allochthonous origin; older suite of autochthonous sand is exposed in nondepositional windows. From Allen (1964). by bottom wave surge. Bottom turbidities ranged from 50 mg 1 during the flood to 4 to 6 mg 1 during the next winter, but were at no time dense enough to drive density currents. Drake et al.'s study suggests that the transport of sus- pended sediment across shelves undergoing allochthon- ous sediment action starts with introduction by a river jet, and continues with deposition, resuspension, and intervals of diffusion and advection by coastal currents in a near-bottom nepheloid layer. Depositional Patterns on Allochthonous Shelves Fine sediments deposited on allochthonous shelves may occur as a seaward-thinning sheet (Fig. 30), or as a series of strips of fine sand or mud oriented generally parallel to the shoreline, see Figs. 31 and 32 (see also Venkatarathnam, 1968; McMaster and Lachance, 1969; and Niino and Emery, 1966). On shelves of equant or irregular dimensions, shelf sectors surfaced by far- traveled, fine-grained sediment may be more irregular in shape (Niino and Emery, 1966; McManus et al., 1969; Knebel and Creager, 1973). Such allochthonous deposits tend to be separated by, or to enclose, nondepositional "windows" in which relatively coarse autochthonous sands are exposed. The disposition of these strips and sheets of allochthonous sediment is generally meaningful in terms of what is known of regional circulation patterns. Locally, the strips may underlie turbid, brackish water plumes that extend from river mouths under the impetus of buoyant ex- pansion and inertial flow (Chapter 14, Fig. 41). Where such flows of high-turbidity water extend for long dis- tances parallel to the coast or seaward across the shelf at promontories, they have been described by McCave (1972) as "advective mud streams" (Fig. 33). He cites Jerlov (1958) as describing such a mud stream running south from the Po Delta, over the mud bed shown in Fig. 34. However, the presence of windows of older sand does not necessarily mean that the sediment pattern is a transient one, which must be eventually followed by a total masking of the old surface of transgression by fine sediment. Instead, the pattern may be a steady state one, 552 29*N . *> neorshore drift river outflow »^- residual currents FIGURE 31- Distribution of mud and generalized transport pattern on the western Gulf shelf oj North America. I'rom McCave (l()72), after Van Andel and Curray (/960). - zyoo' - 22°30 22°00 21°30' - 2I°00 05°00' FIGURE 3 2. Distribution of sediment and generalized transport pattern off the Nayarit coast. Pacific side of Mexico. Autochthonous sands occur in non- depositional windows. Vrom Curray (1969). 340 553 AUTOCHTHONOUS PATTERNS OF SEDIMENTATION 341 Extinction (against distilled water) I I (Ml I I 0.4-0.6 0.2-0.4 0.6-0.8 0.8-1.5 (> 0.8 along Dutch coast) ^>1.5 ^> Residual currents FIGURE 33. Turbidity (light extinction) given by Joseph (1955) with the residual cur- rent pattern in the southern North Sea. Two advective mud streams are illustrated, one crossing from the English to the Dutch side of the area, the other running up the Dutch coast from the mouths of the Plime River. Actual sediment ccncentrations are higher in the latter. From McCave (1972). determined by the local relationship between the hy- draulic activity (primarily wave surge) on the bottom and the near-bottom concentration of suspended sedi- ment (Fig. 35), as well as by the regional transport pattern. On autochthonous shelves, sand transport is primarily advective in nature, occurring during short, intense episodes of wind-driven or tidal flow, and textural gradients tend to reflect the direction of sand transport, with transport becoming finer down the transport path (see Figs. 25 and 27). The transport of fine sand and mud on shelves undergoing allochthonous sedimentation is also primarily advective in nature, in that the turbid water tends to flow as a mass in response to the regional circulation pattern. However, because of the greater role of reversing tidal currents and wave surge in dis- tributing fine sediment it is convenient to think of fine sediment transport on allochthonous shelves as consisting of a dominant advective component, driven by the regional circulation pattern, and an important but subordinate diffusive component, driven by reversing tidal flows and wave surge. The diffusive component of transport not only in- fluences the regional pattern of fine sediment deposition as noted in Fig. 35 and Chapter 9, Fig. 15, but may also result in textural gradients within allochthonous sediment sheets that trend at an angle across the advective trans- port direction. On the Niger shelf, for instance, the domi- nant, advective transport direction is from east to west, under the impetus of the Guinea current (Allen, 1964). However, bottom sediments tend to become finer in a sea- ward (north to south) direction (Fig. 36). Sharma et al. 554 FIGURE 34. Distribution of sediment in the northern Adriatic Sea. From Van Straaten (7965). 342 muddy coast nearshore mud belt ■/////A mid-shelf mud-belt \y/////////j%, mz OAST SHELF EDGE outer-shelf mud-belt * or under advective mud stream FIGURE 3 5. Schematic representation of five cases of sites of shelf mud accumulation. Compare with Fig. 7 5 in Chapter b). Physical Processes of Sedimentation. New York: American Elsevier, 248 pp. Allersma. E. (1972). Mud on the oceanic shelf off Guiana. In Symposium on Investigation', and Resources of the Caribbean Sea and Adjacent Regions. UNESCO, Pans (Unipub, N.Y.). pp. 193- 203. Andrews, E., D. G. Stephens, and D. J. Colquhoun (1973). Scouring of buried Pleistocene barrier complexes as a source of channel sand in tidal creeks. North Island Quadr, ingle. South Carolina. Geol. Soc. Am. Hull., 81: 3659-3(5(52. Bagnold, R. A. (1966), An approach to the sediment transport problem from general physics, I'.S. Geol. Sun: Prof. Pap. ■122 I, 37 pp. Barby, D. G. and J. II. Hoyt (19<)4). An upper Miocene dredged fauna from tidal channels of coastal Georgia. ./. Paleontol., 38: (57 73. Beardsley, R. C. and B. Butman (1974). Circulation on the New- England conlinent.il shelf: Response to strong winter storms. Geol. Re. Lett., Is 181-184. Belderson, R. 11. and A. II. Stride (1966). Tidal current fashion- ing of a basal bed. Mar. Geol., I: 237-257. Belderson, R. II., N. 11. Kenyon, and A. II. Stride (1970). Holo- cene Sediments on the Continental Shelf II est of the Bulnh Isles. Inst. Geol. Sci. Rep. No. 70, pp. 157 170. Bernard, II. A. and R.J. Le Blanc (19(>5). Resume of the Quater- nary geology of the northwest Gulf of Mexico province. In II. E. Wright, Jr., and D.J. Irey, eels., The Quaternary of the I'nited Slates. Princeton, N.J.: Princeton University Press, pp. 137 185. Brouwer, J. E. (I9ti4). The North Sea. In L. D. Stamp, cd.. The Xorlh Sea and the Law of the Continental Shelf. Bude, Cornwall, England, Geographical Publications Ltd., World Land Use Survey Occ. P.ipers 5, pp. 3 14. Rumpus, D. F. (1973). A description of circulation on the conti- nental shelf of the east coast of the United States. Prog. Oceanogr., 6:117 1 57. Caralp, M., A. Klingebiel, C. Latouche, J. Moves, and R. Prud- honinie (1972). liilan Cartografihique un le Plateau Continental dWquitame. Rull. Inst Geol. Bassin d'Aquitaine, Numero Special. 351 Cours cle la Liberation, \.'> Talence, France, 25 pp. Carter, L. (1973). Surficial sediments of Rarklcy Sound and the adjacent continental shelf, West Vancouver Island. Can. ,/. Earth Sci., 10: 441 459. Caston, V. N. 1). (1972). Linear sand banks in the southern North Sea. Sedimentology, lit: l>3 78. (aston, V. N. I), and A. II. Stride (1970). Tidal sand movement between some linear sand banks in the North Sea off north- ( astern Norfolk. Mar. Geol., 9: M38- M42. Colquhoun, I). J. (19(>9). Gecnwrphology of the Lower Coastal Plain of South Carolina. Geological Notes, Division of Geologv, Co- lumbia, >outh Carolina State Development Board, 3(5 pp. (array, J. R. (1904). Transgressions and regressions. In R. L. Mills, ed.. Papers in Marine Geology: Shepard Commemorative I'olume. New York: MaeMillan, pp. 175-203. (urray, J. R. (19(59). History of continental shelves. In The Xew Concepts of Ci t.linental Margin Sedimentation, Washington, D.C.: Am. Geol. Inst., pp. JC-6-1 toJC-6-18. Dolan, R., B. Hayden, G. I loenenberg, J. Zeeman, and M. Vincent (1973). Classification of the coastal environments of the world. Part 1. 'The Americas. Tech. Rep. 1, Office of Naval Research, Washington, D.C., 1(53 pp. Drake, D. E., R. L. Kolpack, and P. J. Fischer (1972). Sediment transport on the Santa Barbara Oxnard shelf, Santa Barbara channel, California. In D.J. P. Swift, D. B. Duane, and O. H. Pilkcy, eds., Shelf Sediment Transport: Process and Pattern. Stroudsburg, Pa.: Dowden, Hutchinson & Ross, pp. 301-332. Duane, D. B., M. E. Field, E. P. Meisburger, D. J. P. Swilt. and S. J. Williams (1972). Linear shoals on the Atlantic inner continental shelf, Florida to Long Island. In D. J. P. Swift, D. B. Duane, and O. H. Pilkey, cds.. Shelf Sediment 'Transport: Process and Pattern. Stroudsburg, Pa.: Dowden, Hutchinson & Ross, pp. 447-499. Emery, K. O. (1952). Continental shelf sediments of southern California. Geol. Soc. Am. Bull., 6.3: 1105-1108. Emery, K. O. (1968). Relict sediments on continental shelves of the world. Am. Assoc. Pet. Geol. Bull., 52: 445-4t.4. Field, M. E. (1974). Buried strandline deposits on the central Florida inner shelf. Geol. Soc. Am. Bull., 85: 57-t>0. Fischer, A. G. (19(>1). Stratigraphic record of transgressing seas in the light of sedimentation Atlantic coast of New Jersey. Am. Assoc. Pel. Geol. Bull.. 15: 1(55(5-1(5(5(5. Ginsburg, R. N. and N. P. James (1974). Holocene carbonate sediments of continental shelves. In C. A. Burk and C. C. Drake, eds.. The Geology oj Continental Margins. Berlin and New York: Springer-Verlag, pp. 137-15(5. Grabau, A. \V. (1913). Principles of Stratigraphy, 19u0 facsimile edition of 1924 revision. New York: Dover, 1185 pp. Hathaway, J. C (1971). Data File, Continental Margin Program, Atlantic Coast of the United States, Vol. 2, Samples Collection and Analytical Data. Woods I lole Oceanographic Institution Ref. 71-15, U.S. Geol. Surv., Woods Hole, Mass., 44b pp. Henry, V. J., Jr. and J. II. Hoyt (19(58). Quaternary paralic shelf sediments of Georgia. Southeast. Geol., 9: 195-214. Iloltedahl, II. (1958). Some remarks on the geomorphology of the continental shelves off Norway, Labrador, and southeast Alaska. ./. Geol., 66: 4(51 471. Iloltedahl, O. (1940). The Submarine Relief off the Norwegian Coast. Oslo: Norske Videnskaps-Akade, 43 pp. Houbolt, J. J. II. C. (19(58). Recent sediments in the southern bight of the North Sea. Geol. Mijnhouu; 17: 245-273. Hoyt, J. II, and J. R. Hails (19(57). Pleistocene shoreline sediments in coastal Georgia: Deposition and modification. Science, 155: 1541-1543. Hoyt, J. II. and V.J, Henry, Jr. (19157). Influence of island mi- gration on barrier island sedimentation. Geol. Soc. Am. Bull., 78: 77 86. 561 REFERENCES 349 Jerlov, N. G. (1958). Distribution of suspended material in the Adriatic Sea. Arch. Oceanogr. Limnol., 11: 227-250. Johnson, D. (1919). Shore Processes and Shoreline Development. New York: Wiley, 585 pp. (2nd ed., 1938). Joseph, J. (1953). Die Trubangsverbaltnisse in der Sudwestlichon Nordsee VVahrend der "Gauss Fahrt" im Februar, Marz, 1952. Ber. Dtsch. U'iss. Komm.f. Meeresforsch. 13: 93-103. Kaye, C. A. (1959). Shoreline features and quaternary shoreline changes, Puerto Rico. U.S. Geol. Surv. Pro/. Pap., 317-B, 140 pp. Kenyon, N. H. (1970). Sand ribbons of European tidal seas. Afar. Geol.. 9: 25-39. Kenyon, N. H. and A. H. Stride (1970). The tide-swept conti- nental shelf sediments between the Shetland Isles and France. Sedimenlology, 14: 159-175. King, L. H., B. Maclean, and G. Drapeau (1972). The Scotian shelf submarine end moraine complex. In 24th Int. Geol. Congr., Montreal, pp. 237-239. Knebel, H.J. and J. S. Creager (1973). Sedimentary environments of the east-central Bering Sea continental shelf. Mar. Geol., 15: 25-47. Kumar, N. and J. E. Sanders (1970). Are basal transgressive sands chiefly inlet filling sands? Marit. Sediments, 6: 12-14. Laursen, E. M. (1958). The total sediment load of streams. Proc. Am. Soc. Civil Eng., 84(HV1): 1530. McCave, I. N. (1971). Sand waves in the North Sea oflf the coast of Holland. Mar. Geol., 10: 149-227. McCave, I. N. (1972). Transport and escape of fine-grained sedi- ment from shelf areas. In D. J. P. Swift, D. B. Duane, and O. H. Pilkey, eds., Shelf Sediment Transport: Process and Pattern. Stroudsburg, Pa.: Dowden, Hutchinson & Ross, pp. 225-248. McClennen, C. F. (1973). New Jersey continental shelf near bot- tom current meter records and recent sediment activity. ./. Sediment. Petrol., 43: 371-380. McClennen, C. E. and R. L. McMaster (1971). Probable Holo- cene transgressive effects on geomorphic features of the conti- nental shelf off New Jersey, United States. Marit. Sediments, 7: 69-72. McKinney, T. F. (1974). Large-scale current lineations on the Great Egg shoal retreat massif, New Jersey shelf. Investiga- tion by side-scan sonar. J. Sediment. Petrol., It: 79-102. McKinney, T. F., VV. L. Stubblefield, and D. J. P. Swift (1974). Large-scale current lineations on the central New Jersey shelf: Investigations by side scan sonar. Mar. Geol., 17: 79-102. McManus, D. A., J. C. Kelley, and J. S. Creager (1969). Conti- nental shelf sedimentation in an arctic environment. Geol. Soc. Am. Bull., 80: 1961-1984. McMaster, R. L. and T. P. Lachance (1969). Northwestern African continental shelf sediments. Alar. Geol., 7: 57-67. McMaster, R. L., T. P. Lachance, and A. Ashraf (1970). Conti- nental shelf geomorphic features off Portuguese Guinea, Gui- nea, and Sierra Leone (West Africa). Mar. Geol., 9: 203- 213. Meade, R. H. (1969). Landward transport of bottom sediments in the estuaries of the Atlantic coastal plain. J. Sediment. Petrol., 39: 229-234. Merrill, A. S., K. O. Emery, and M. Rubin (1965). Oyster shells on the American continental shelf. Science, 147: 395-400. Milliman, J. D. and K. O. Emery (1968). Sea levels during the past 35,000 years. Science, 162: 1121-1123. Milliman, J. D., C. P. Summerhayes, and H. T. Barretto (in press). Quaternary sedimentation on the Amazon continental margin: A model. Geol. Soc. Am. Bull. Naumann, C. T. (1858). Lehrbuch der Geognosie, Bd. 1, 2nd ed. Leipzig: Wilhelm Engelmann, 1000 pp. Nichols, M. M. (1972). Inner Shelf Sediments off Chesapeake Bay. I. General Lithology and Composition. Spec. Sci. Rep. 64, Virginia Institute of Marine Science, Gloucester Point, Va., 20 pp. Niino, H. and K. O. Emery (1966). Continental shelf off north- eastern Asia. J. Sediment. Petrol., 36: 152-161. Norris, R. M. (1972). Shell and gravel layers, western continental shelf, New Zealand. N.Z. J. Geol. Geophys., 15: 572-589. Oaks, R. Q., Jr. and N. K. Coch (1963). Pleistocene sea levels, southeastern Virginia. Science, 140: 979-983. Pilkey, O. H. and M. E. Field (1972). Onshore transportation of continental shelf sediment: Atlantic southeastern United States. In Shelf Sediment Transport: Process and Pattern. Strouds- burg, Pa.: Dowden, Hutchinson & Ross, pp. 424-446. Pilkey, O. H. and J. Frankenberg (1964). The relict-Recent sedi- ment boundary on the Georgia continental shelf. Georgia Acad. Set. Bull., 22: 1-4. Powers, M. C. and B. Kinsman (1953). Shell accumulations in underwater sediments and the thickness of the traction zone. J. Sediment. Petrol., 23: 229-234. Redfield, A. C. (1956). The influence of the continental shelf on the tides of the Atlantic coast of the United States. J. Mar. Res., 17: 432-448. Robinson, A. H. W. (1966). Residual currents in relation to shoreline evolution of the East Anglian coast. Mar. Geol., 4: 57-84. Sarnthein, M. (1972). Sediments and history of the postglacial transgression in the Persian Gulf and Northwest Gulf of Oman. Mar. Geol., 12: 245-266. Sharma, G. D., A. S. Naidu, and D. W. Hood (1972). Bristol Bay: Model contemporary graded shelf. Am. Assoc. Pet. Geol. Bull., 56: 2000-2012. Shepard, F. P. (1932). Sediments on continental shelves. Geol. Soc. Am. Bull., 43: 1017-1034. Shepard, F. P. (1973). Submarine Geology. 3rd Edition. New York: Harper & Row, 517 pp. Sheridan, R. E., C. E. Dill, Jr., and J. C. Kraft (1974). Holocene sedimentary environment of the Atlantic inner shelf off Dela- ware. Geol. Soc. Am. Bull., 85: 1319-1328. Shideler, G. L., D. J. P. Swift, G. H. Johnson, and B. W. Holliday (1972). Late Quaternary stratigraphy of the inner Virginia continental shelf: A proposed standard section. Geol. Soc. Am. Bull., 83: 1787-1804. Smith, J. D. (1969). Geomorphology of a sand ridge. J. Geol., 77: 39-55. Stahl, L., J. Koczan, and D. Swift (1974). Anatomy of a shore- face-connected ridge system on the New Jersey shelf: Impli- cations for genesis of the shelf surficial sand sheet. Geology, 2: 117-120. Stanley, D. J. (1969). Submarine channel deposits and their fossil analogs (fluxoturbidites). In D. J. Stanley, ed., The New Con- cepts of Continental Margin Sedimentation. Washington, D.C.: Am. Geol. Inst., pp. DJS-9-1 to DJS-9-17. Stanley, D. J., G. Drapeau, and A. E. Cok (1968). Submerged terraces on the Nova Scotian shelf. Z. Geomorphol., Suppl., 7: 85-94. 562 350 CONTINENTAL SHELF SEDIMENTATION Stanley, D. J., D. J. P. Swift, and H. G. Richards (1967). Fossilif- erous concretions on Georges Bank. J. Sediment. Petrol., 37: 1070-1083. Stanley, D. J., D. J. P. Swift, N. Silverberg, N. P. James, and R. G. Sutton (1972). Late Quaternary progradation and sand spill-over on the outer continental margin cff Nova Scotia, Southeast Canada. Smithsonian Contr. Earth Sciences, 8: 88 pp. Steers, J. A. (1971). The East Coast floods, 31 January-1 February, 1953. In J. A. Steers, ed., Applied Coastal Geomorphology. Cam- bridge, Mass.: MIT Press, pp. 198-224, 227 pp. (reprinted). Stride, A. H. (1963). Current-swept sea floors near the southern half of Great Britain. Q. J. Geol. Soc. London, 119: 175-199. Stride, A. H., R. H. Belderson, and N. H. Kenyon (1972). Longi- tudinal Furrows and Depositional Sand Bodies of the English Chan- nel. Mem. Bur. Rech. Geol. Min. No. 79, pp. 233-244. Stubblefield, VV. L., J. W. Lavelle, T. F. McKinney, and D. J. P. Swift (1975). Sediment response to the present hydraulic regime on the Central New Jersey shelf. J. sediment. Petrol., 45: 337-358. Stubblefield, W. L. and D. J. P. Swift (in press). Ridge develop- ment as revealed by sub-bottom profiles on the central New Jersey shelf. Submitted to Mar. Geol. Swift, D. J. P. (1973). Delaware shelf valley: Estuary retreat path, not drowned river valley. Geol. Soc. Am. Bull., 84: 2743-2748. Swift, D.J. P. (1974). Continental shell sedimentation. In C. A. Burke and C. L. Drake, eds., The Geology oj Continental Mar- gins. Berlin and New York: Springer-Verlag, pp. 117-133. Swift, D.J. P. (1975). Tidal sand ridges and shoal retreat massifs. Mar. Geol., 18: 105-134. Swift, D. J. P. (in press). Continental shelf sedimentation. In R. Fairbridge, ed., Encyclopedia of Sedimenlology. Princeton, N.J.: Van-Nostrand-Reinhold. Swift, D. J. P., D. B. Duane, and T. F. McKinney (1973). Ridge and swale topography of the Middle Atlantic Bight, North America: Secular response to the Holocene hydraulic regime. Mar. Geol., 15: 227-247. Swift, D. J. P. and S. D. Heron, Jr. (1969). Stratigraphy of the Carolina Cretaceous. Southeast. Geol., 10: 201-245. Swift, D.J. P., J. W. Kofoed, F. P. Saulsbury, and P. Sears (1972). Holocene evolution of the shelf surface, central and southern Atlantic coast ot North America. In D. J. P. Swift, D. B. Duane, and O. H. Pilkey, eds., Shelf Sediment Transport: Process and Pattern. Slroudsburg, Dowden, Hutchinson & Ross, pp. 499-574. Swift, D. J. P. and P. Sears (1974). Estuarine and littoral deposi- tional patterns in the surficial sand sheet, central and southern Atlantic shelf of North America. In International Symposium on Interrelationship of Estuarine and Continental Shelf Sedimentation. Inst. Geol. du Bassin d'Aquitaine, Mem. 7, pp. 171-189. Swift, D. J. P., D. J. Stanley, and J. R. Curray (1971). Relict sediments, a reconsideration. J. Geol., 79: 322-346. Todd, T. W. (1968). Dynamic diversion: influence of longshore current-tidal floor interaction on Chenier and barrier island plains. J. Sediment. Petrol., 38: 734-746. Uchupi, E. (1968). Atlantic continental shelf and slope of the United States— Physiography. U.S. Geol. Surv. Prof. Pap. 529-C, 30 pp. Uchupi, E. (1970). Atlantic continental shelf and slope of the United States: Shallow structure. U.S. Geol. Surv. Prof. Pap. 524-1, 44 pp. Van Andel, Tj. H. and J. R. Curray (1960). Regional aspects of modern sedimentation in the northern Gulf of Mexico and similar basins, and Paleogeographic significance. In F. P. Shepard, F. B. Phegler and T. H. Van Andel, eds., Recent sediments, northwest Gulf of Mexico. Tulsa: Am. Assoc. Pet. Geol., pp. 345-364. Van Andel, Tj. H. and J. J. Veevers (1967). Morphology and Sediments of the Timor Sea. Commonwealth of Australia, Bureau of Mineral Resources, Geology and Geophysics, Bull. 83, 172 pp. Van Straaten, L. M. J. V. (1965). Sedimentation in the north- western part of the Adriatic Sea. In VV. F. Whitland and R. Bradshaw, eds., Submarine Geology and Geophysics. London: Butterworths, pp. 143-162. Veenstra, H. J. (1969). Gravels of the southern North Sea. Mar. Geol., 7: 449-464. Venkatarathnam, K. (1968). Studies on the sediments of the con- tinental shelf off the regions Visakhapatnam-Pudimadaka and Pulicat Lake-Penner River confluence along the east coast of India. Natl. Inst. Set. India Bull., 38: 472-482. Visher, G. S. and J. D. Howard (1974). Dynamic relationship between hydraulic and sedimentation in the Altamaha estu- ary. J. Sediment. Petrol., 44: 502-521. Whitmore, F. C, Jr., K. O. Emery, H. B. S. Cooke, and D. J. P. Swift (1967). Elephant teeth from the Atlantic continental shelf. Science, 156: 1477-1481. Wright, L. D. and J. M. Coleman (1973). Variations in morphol- ogy of major river deltas as functions of ocean wave and river discharge regimes. Am. Assoc. Pet. Geol. Bull., 57: 320-348. Wright, L. D. and J. M. Coleman (1974). Mississippi River mouth processes: Effluent dynamics and morphological development. J. Geol., 82: 751-778. 563 50 Reprinted from: Marine Sediment Transport and Environmental Management, D. J. Stanley and D. J. P. Swift, editors, John Wiley and Sons, Inc., Chapter 10, 159-196. CHAPTER 10 Substrate Response to Hydraulic Process: Gram-Size Frequency Distributions and Bed Forms DONALD J. P. SWIFT Atlantic Oceanographic and Meteorological Laboratories, Miami, Florida JOHN C. LUDWICK Institute of Oceanography, Old Dominion University, Xorfolk, Virginia Chapters 8 and 9 dealt with the entrainment of sand and mud, respectively, on the continental shelf. In addi- tion, Chapter 8 discussed the most ubiquitous shelf bed form, the sand ripple formed by bottom wave surge, since it plays a critical role in the entrainment and trans- port of sand on the continental shelf. This chapter explores further the response of the shelf floor to the hydraulic climate. Two key responses that are used to infer regional patterns of sediment transport are grain-size frequency distributions and substrate bed forms. The chapter also describes a numerical model for estimating sediment transport and areas and rates of erosion and deposition. GRAIN-SIZE FREQUENCY DISTRIBUTIONS Krumbein (1934) was the first to bring to popular atten- tion the concept that the size frequency distribution of sand samples tends to be log-normally distributed. It has become a tenet of conventional wisdom that this dis- tribution, as defined by its mean and standard distribu- tion, is the signature of the depositional event, and that deviations from log normality, as measured in terms of standard deviation, skewness, and kurtosis, reflect both the provenance and subsequent hydraulic history of the sediment (see Inman, 1949; Friedman, 1961; Visher, 1969). Genesis of the Normal Curve Recent theoretical studies (Middleton, 1968; Swift et al., 1972b) have attempted to present this hypothesis in a more rigorous manner by consideration of probability theory. The reader is referred to these papers for the mathematical foundation of the following discussion. The probability model for the genesis of a log-nor- mally distributed grain population considers a flow over a sand substrate in which the total load is adjusted to flow conditions. If deposition is to occur, there must be a decrease in bottom shear stress ( — dro/dx) and dis- charge ( — bq/dx) down the transport path. The distri- bution of grain sizes in the load undergoing transport down this shear stress gradient and the absolute value of the gradient is such that for each grain-size class, an upstream portion of the path is experiencing supercritical stress, and a downstream portion is experiencing sub- critical stress. We are concerned with the central portion 564 159 160 SfBSTRATF. RESPONSE TO HYDRAULIC PROCESS of the transport path, where a series of transition points for critical shear stress occur, with each successive down- stream transition point being associated with a succes- sively finer grain-size class. The grains are assumed to travel down the transport path in a series of discrete hops as a consequence of the turbulent structure of the flow, and as a consequence of a larger scale cycle of flow events separated by periods of quiescence. The model is thus a stochastic model, with an inherently random aspect to its behavior, and the problem may be dealt with in terms of probability theory. Under these conditions, it is conceptually possible to define the grain-size frequency distribution at each point as the product of two probability vectors, an admittance vector and a retention vector (Fig. 1). The admittance vector is the sequence of probabilities of entrance of the size classes present, ordered in sequence of decreasing grain size. The retention vector is similarly the sequence of probabilities of retention of successively finer grain sizes. PHI UNITS (b) FIGURE 1. Grain-size frequency distributions as a product of a retention vector and an admittance vector. See text for explanation. From Swift et al. (1972b). If P]n is an element in an admittance vector, where j denotes thej'th station in the transport path and n de- notes one of n grain-size classes, and if P'}n is a corre- sponding element in a retention vector for the same station, then the product of the two probabilities, Pjn( 1 — P'jn) gives the probability that the particle in the local input enters but does not leave the station. The product of the input vector with all corresponding ele- ments in the admittance and retention vectors for a sta- tion gives the frequency distribution for that station (Fig. 1). This is a restatement, in probabilistic terms, of the intuitively apparent fact that the modal diameter of a deposit is that grain size most likely to arrive and least likely to be carried away from the place of deposition under prevailing flow conditions; progressively coarser sizes are progressively less frequent because they are less likely to arrive, and progressively finer sizes are progres- sively less frequent because they are more likely to be carried away. In Fig. la, the two linear numerical filters (admit- tance and retention vectors) are applied to a local input frequency distribution that is uniform in nature and a symmetrical retained frequency distribution results. If, however, the local input has a skewed distribution (Fig. lb), then the retained distribution is still skewed, al- though it has been modified by the station probabilities. If the filters are not linear, then further modification of the input vector occurs. In Fig. 2, various hypothetical input distributions are subjected to sorting down the stations of a hypothetical transport path according to the probabilistic algorithm described above. In column A, an initially rectangular distribution is seen to evolve into a distribution with a distinct mode, and the mode is seen to shift toward the finer end of the distribution at successive stations. The coarse flank of the mode becomes visibly sigmoid (S-shaped) as is characteristic of the side of the normal distribution frequency curve. The increasingly sigmoid shape is the consequence of the multiplication of succes- sive admittance vectors in order to obtain the coarse admixture of the local input frequency distribution. For instance, if the admittance vector has the form 0.100, 0.200, 0.300, 0.400, 0.500, 0.600, 0.700, 0.800, 0.900, 1.00 and retains this form from station to station, then at the third station the size frequency distribution of the coarse admixture will be determined primarily by the third power of the admittance vector, 0.001, 0.008, 0.027, 0.064, 0.125, 0.216, 0.343, 0.512, 0.729, 1.00 565 RECTANGULAR INPUT 20 INPUT 0 20 STN 1 •1 20n STN 2 0J 20 STN 3 0 20 STN 4 iilllll i 20] STN 5 > 0 o 1 z => 30 o J STN 6 STN 7 STN 8 40 -i 40- r -rr{]\ 40- STN 9 0 Ldrif. 50 STN 10 B NORMAL INPUT 30 30 0]i k. 30 ,iiL •1 ' 4 30- ,:.. 30 30 -rl l 1 11 h. 30 m 3" Jlllk 4 i 30 1 it 30- [}l cr62 X-34 _=£fl 30-i ■1 4 PHI UNITS •1 4 PHI UNITS EXPONENTIAL EXPONENTIAL SIZE VALUES DISTANCE VALUES 30n - JTL 20H m oimllL oVfllL 0A 30 - ro 30 rfh 30 rliii 30n \ i-jliilK. i 20^ 20-| 0 m i 30 -| L o 30-i ■J A •l 30n HI 30- Jm oLJml l PHI UNITS 40- JL. nil 11 Ik 40 40n 40-, 40 40- 40 ] m 40 40 60^ J L ■1 4 PHI UNITS FIGURE 2. Grain-size frequency distributions along sediment transport paths under different conditions. See text for explanation. From Swift et al. (1972b). 566 161 162 Sl'BSTRATE RESPONSE TO HYDRAULIC PROCESS The initially small probabilities have decreased more than the initially large ones, and the resulting curve of frequency against size class will be exponential in form. The modal shift is the phenomenon of progressive sorting (Russell, 1939) whereby the deposit becomes finer down the transport path, as a consequence of the steady de- pletion of the transported material in coarser particles. In physical terms, this means that the coarsest particles tend to get left behind whenever the bottom is eroded by a flow event that is weaker than the one that preceded it. If it is assumed that the input distribution is normal to begin with (Fig. 2, column B), and if it is assumed that the probabilities of admittance and retention vary linearly with grain size, then the mode shifts toward the fine end of the distribution as the sediment is traced down the transport path with no change in the shape of the normal curve. However, in a more realistic case, the probabilities of admittance and retention are assumed to vary exponentially with grain size; in other words, the transport rate varies exponentially with grain size. As a consequence, vector multiplication acts on the two sides of the frequency curve in a dissimilar fashion (col- umn C). The greater range of transport probabilities as- signed to the coarser sands results in greater efficiency of sorting on that side of the curve, and progressive steepening of that side, as the sediment is traced down the transport path. The sediment becomes increasingly enriched in the fine admixture at the expense of the coarse admixture (becomes fine-skewed), as well as be- coming finer down the transport path. This effect is particularly marked where the intensity of the flow field is made to decrease down the transport path (column D). Size Frequency Subpopulations and Flow Regimes Thus there are at least theoretical reasons supporting the concept that the size frequency distribution of fluid-de- posited sands constitutes hydraulic signatures of the flow process. Attempts to interpret these signatures have in general generated more heat than light (Emery and Uchupi, 1972, p. 375). However, the analysis of the subpopulations constituting sand samples has proved more fruitful. The basic work has been undertaken by Moss (1962, 1963, 1972). He notes that most grain-size frequency distributions of sand deposited from fluid flow do not plot as a straight line on probability paper as they should if they arc normally distributed. Instead the curves are Z-shaped (Fig. 3). He has demonstrated that these Z-shaped curves are composite distributions and are the consequence of the presence of three or more log-normally distributed subpopulations, and that these subpopulations are an outcome of the manner in which PHI SCALE 067 MM SCALE FIGURE 3- Cumulative curve of a swash zone sample. In Moss' terminology, A is framework population, B is interstitial population, and C is contact population. From Visher (J969). the bed is built (Fig. 4). A. framework population (A popu- lation) constitutes the bulk of the sample. Its modal di- ameter is a function of the average dimensions of the relatively large spaces between grains on the aggrading surface. There is a strong feedback in this system between deposit grain size and bed load grain size; the dimen- sions of grains selected from bed load for deposition in such holes depend on the dimensions of grains already deposited, which in turn depend on the dimensions of available grains in the bed load, and ultimately on the dimensions of the hydraulic parameters of the flow. A fine interstitial subpopulation (B population: fine tail of the size frequency distribution curve) consists of grains that are small enough to filter into the interstices of the grain framework of the deposit. Their average diameter is not that of the bowl-shaped openings on the bed surface but the smaller average diameter of the interstices within the deposit. A coarse contact subpopulation (C population: coarse tail of the frequency curve) consists of grains that are too coarse to fit into or through the surface openings as do the grains of the A and B populations. Instead they accumulate as slowly moving to stationary clogs of mu- tually interfering coarse grains on the bed surface. When a critical area of these rejected coarse particles has ac- 567 GRAIN-SIZE FREQUENCY DISTRIBUTIONS 163 10- FIGURE 4. Size-frequency curves of sands from various environments. Curves have been dissected to indicate subpopulations. From McKinney and Friedman (1970). cumulated, it will be buried beneath further layers of A population grains. It is important to assess the relationship between Moss* rather sophisticated theory of subpopulation genesis, and the prevailing equation of transport modes with sub- population characteristics. It has been generally assumed [see the review by Yisher ( 1969)] that the contact popu- lation represents particles moved by dragging; or rolling, the framework population represents particles moving by saltation, and the interstitial population represents par- ticles traveling in suspension. There is a correlation be- tween these differing modes of transport and the per- centage of respective subpopulations in the deposit, be- cause each of these modes is most likely to carry the appropriate size of material for the subpopulations with which they have been correlated. The relative percent- ages of subpopulations, however, are a direct consequence of mechanisms of bed construction, and only indirectly reflect modes of transport. Moss has shown, for instance, that both B and A subpopulations may be generated from saltative transport alone. The percentages of these three populations in a given deposit will vary, within limits set by grain geometry and grain interaction processes, according to the regional TABLE 1. Nomenclature and Grain-Size Characteristics of Sediment Flow Regimes Population Southard and Moss (1972) Boguchwal (197 3) A Fine ripple stage Ripples Dominant Coarse ripple stage Ripples Dominant Dune stage Dunes Dominant Rheologic stage Transition Upper flat Antidunes bed Dominant Mean Diameter (Moss. 1972) Abundant Scarce Scarce Scarce Scarce Scarce Abundant Abundant 0.07-0.25 mm (3.75-2.00*) 0.25-0.92 mm (2.00-0.25*) 0.25-2.2 mm (2.0 to -1.1*) 0.17-4.8 mm (2.6 to -2.3*) 568 164 SUBSTRATE RESPONSE TO HYDRAULIC PROCESS availability of the three populations, and also according to the hydraulic microclimate of the bed. Moss (1972), on the basis of flume studies and studies of river deposits, has described five bed regimes. These may be correlated w ith the flow regimes described by Southard and Boguch- wal (1973, Fig. 23). Each tends to form a characteristic admixture of subpopulations (Table 1). Moss (1972) notes that in the fine ripple stage, grains do not pro- trude through the laminar sublayer of the bottom bound- ary layer of the flow and microturbulence is absent from the bed surface. Fine particles can become concentrated near the bed. and can pass copiously into the interstices. Hence the fine ripple regime is characterized by an abundant B population. In the coarse ripple stage and dune stage, grains protrude through the lamina sublayer. Fluid dynamic lift and bed grain turbulence operate to keep fine particles from being concentrated near the bed, and the interstitial (B) population is normally a minor bed constituent. In the rheologic stage, flow is supercritical, and bed load particle behavior is dominated by the dispersive pressure associated with grain collisions (Bagnold, 1954,). These pressures force the particles against and into the bed. This effect is evidently dominant over the lift forces which act at the bed, and the interstitial B population again passes copiously into the bed. The rheologic stage is furthermore the only stage in which Moss observed an abundant contact (C) population. Moss' theory may thus be used to infer flow regime from the grain-size distribution. It must be applied with caution, however, as it was developed for quasi-steady flows, and the continental margin environment tends to be subjected to an additional oscillatory flow component because of wave surge. Grain-size distributions conse- quently tend to indicate more intense unidirectional flows than actually exist (Stubblefield et al., 1975). BED FORMS In this section.it is necessary to deal with more varied and larger bed forms than the wave ripples described in Chapter 7. Sand wave fields and sand ridge fields may generate bed form spacings of a kilometer or more, and bed form amplitudes of up to 30 m. Such large-scale bed form arrays become significant storage elements in continental margin sediment budgets, and such budgets cannot be understood without an awareness of bed form mechanics. Furthermore, large-scale bed forms impact directly on human usage of the continental margin. Large tankers navigate the Thames estuary channels (Langhorne. 1973) with scant meters of clearance over sand wave crestv Sewage outfalls and nuclear power plants are planned or are being constructed in the'inner shelf ridge fields of the Atlantic shelf. Seafloor well heads are subject to burial by migrating bed forms. concepts. A bed form is an irregularity in the par- ticulate substrate of a fluid flow. This definition includes the subaqueous sand wave and sand ridge fields of the earth's shelves, the subaerial dune fields of the earth's deserts and those photographed on Mars, and the bed forms of the base surge deposits surrounding the lunar craters, sedimented out of a transient fluid of gas, dust, and debris generated by the impact of meteors. Bed forms are not independent phenomena; they are equilibrium configurations of the interface between a mobile, usually cohesionless substrate, and an overlying flow field, and tend to occur in repetitious arrays rather than alone. They are the product of feedback between flow structure and substrate structure. The three-dimensional pattern of flow does not "cause" the bed form to arise, nor does the bed form "cause" the deformation of the boundary layer of the flow field; instead, strictly speaking, these two elements of a flow-substrate system interact to cause each other. Wilson (1972, p. 204) notes that when a fluid is sheared, either against another fluid, against itself, or against a rigid boundary, there are many situations in which secondary flows develop. Secondary flows are regularly repeated patterns of velocity variation super- imposed on the mean flow. The primary flows satisfy the three continuity laws (of mass, energy, and momen- tum), but in such a way that any small disturbance is initially self-aggravating; in other words, the flow is an unstable system. In sheared flows, this usually involves the development of any combination of such secondary flows as transverse internal waves, or transverse or flow- parallel vortices. Such secondary flows may occur simul- taneously at several scales. Wilson further notes that sheared fluids may become unstable in response to almost any sort of strong gradient in velocity, pressure, viscosity, temperature, or density in the direction normal to the shear force. These may arise over completely plane beds. Eventually, however, as the perturbed flow and the bed deform in response to each other, a new stable state is attained. The theory of fluid instability has been outlined by Lin (1955), Chandrasekhar (1961), Rosenhead (1963), and Yih ( 1965), and these authors have discussed many- cases to which it has been applied. Allen ( 1968a, p. 50) has summarized their computational approach. The al- gorithm requires that equations of motion be set up to describe the fluid motion of interest. These equations are solved to discover whether a small sinusoidal dis- turbance of one variable will be damped or amplified 569 BED FORMS 165 under the chosen limits for other variables. The motion is stable if the disturbance is damped, but unstable if it is amplified. In nature the unstable disturbance is ampli- fied until the other variables of the system set some lim- iting condition on the amplification and a new stale of quasi equilibrium is attained. Stability analysis has been successfully applied to the problem of ripple and sand wave formation (e.g., Smith, 1969) and it seems likely that all bed forms will ultimately prove susceptible to this mode of attack. BASIC MODES OF BED FORM BEHAVIOR. Most bed forms fall into two basic categories: those that are oriented across the flow direction, such as sand waves and ripples, and those that are oriented parallel to the flow direction, such as sand ribbons. These two basic patterns must correspond to two basic patterns within the flow field itself, a transverse pattern in which zones of scour and aggradation alternate down the flow path, and one in which zones of scour and deposition alternate across the flow path (Figs. 5 and 6). There is considerable evidence to indicate that this is the case, although the basic mech- anisms are far from clear. (a) (b) DEVELOPMENT gj^ 2223 ^r lA |K7 FIGURE 6. The development of a longitudinal bed form. (a) The pattern of secondary flow over longitudinal bed form elements: PP, flow attachment lines along ridge trough; QQ, flow separation lines along crests, (b) Development of longitudinal elements in vertical cross section perpendicular to mean flow direction. (i) Form and flow components; (ii) components in z direction; (Hi) components in x direction. Numbers as in Fig. 4- Note alternate notation of coordinate axes. From Wilson (1972). FIGURE 5. The development of a transverse bed form. (A) Initiation; (B) growth; (C) equilibrium. (I) Sand transport rate; (2) shear velocity at bed; (3) erosion rate; (4) streamlines. From Wilson (1972). Transverse Bed Forms mode of formation. As noted by Wilson (1972), most transverse bed forms are probably caused by trans- verse wave perturbations in the flow. The problem is a complex one, and the solutions offered to date have not been altogether satisfactory. Summaries are presented by Allen (1968a, pp. 130-149), Kennedy (1969, p. 151), and Smith (1970, p. 5928). Smith points out that many of these studies are un- necessarily restrictive; they assume an eddy viscous mean flow but neglect the inertial terms in the equation of motion (Exner, 1925, in Raudkivi, 1967) or assume in- viscid irrotational flow (Kennedy, 1969). These assump- tions require an a priori phase shift in the velocity field relative to the interface disturbance in order for insta- bility to occur. Smith (1970) has undertaken a stability analysis employing inertial terms in the equations of motion. His results indicate that the interface is unstable with respect to infinitesimal perturbations of wavelength greater than the wavelength for which the inertia of the grains is important (wavelengths less than 10 times mean 570 166 SUBSTRATE RESPONSE TO HYDRAULIC PROCESS grain diameter). Smith utilizes the sediment continuity equation, which may be presented in its simplest two- dimensional form as drj dq dt dx (1) where 77 is height of the interface above a datum, / is time, k is a constant, q is sediment discharge at a level near the bed, and x is horizontal distance. In physical terms, the time rate of change of the height of the inter- face at a point above the datum is proportional to the horizontal discharge gradient at that point, assuming sat- uration of the boundary flow with sediment; a decrease in discharge across the point ( — bq/dx) must result in aggradation, while an increase (dq/dx) must result in erosion. Smith has rewritten the equation in terms of boundary shear stress and discharge: (2) dt \to dro/ dx where c0 is the boundary concentration of sediment, q is the mean volume flux of sediment per unit width, and To is the local mean shear stress on the bed. Smith's analysis divides the nonuniform horizontal ve- locity along the waveform interface into an in-phase component and an out-of-phase component. The in- phase component consists of accelerating flow over crests with maximum shear stress at those points, as required by flow continuity. Since boundary erosion varies di- rectly with dro/dx, this in-phase component simply causes upstream erosion of the interface perturbation and downstream deposition; the perturbation moves downstream with neither growth nor decay. However, the inertia of the high-velocity water of the upper part of the water column causes it to converge with the rising bed on the upstream side of an interface perturbation, and there is as a consequence an out-of-phase component of velocity and bottom shear stress which attains its maximum value at this zone. This maximum persists when the components are added; hence, since deposition is proportional to — dro ''dx, some sand must always be deposited on the crest of the perturbation. Smith (1970) cites Exner's (1925) earlier stability analysis as qualitatively correct, despite neglect of the inertial terms in the momentum equations. Exner had shown that when downcurrent spacing is wide, the crests of perturbations move faster than the troughs be- tween them, resulting in oversteeping of the downcur- rent slope to the angle of repose (30° underwater), and consequently, in the formation of a horizontal roller eddy (wake, flow separation bubble) downstream of the crest. The perturbation is now a mature ripple or sand wave. The generation of a wake behind a growing bed form results in propagation of interface instability in the down- stream direction. Smith (1970) cites Schlichting (1962, p. 200) who has studied the development of a turbulent wake. Behind a negative step such as the avalanche slope of a growing transverse bed form, flow accelerates downstream of the attachment line (Fig. 7) and at the same time a boundary layer is initiated that grows in height downstream. Shear increases downstream because of flow acceleration, then decreases as the effects of boundary layer growth dominate over the effects of flow acceleration. Here, in a zone where dr0/i>x < 0, sand is deposited and a new ripple grows, which in turn de- forms, develops a wake, and triggers a third. Smith's stability analysis does not specify wavelength for growing bed form perturbations, and it is apparent that this parameter must be defined by spatial adjustments in the turbulent velocity field. As downstream ripples grow in height and their separation bubbles in width, they must grow in length, which is accomplished by the smaller ripples moving faster, and stretching out the ripple field. Saporotlon line Separation line Surface of separation Surface of separation Attachment line FIGURE 7. Three-dimensional separated flows, (a) Roller; (b) vortex. Note alternative notation of coordinate axes. From Allen (1970). 571 BED FORMS 167 Smith's scheme of transverse bed form formation by the spontaneous deformation of the interface into a moundlike perturbation, its increasing asymmetry, the formation of a separation bubble, and the downstream propagation of the instability, has been strikingly con- firmed in experimental work by Southard and Dingier (1971). Their work suggests that if the critical flow threshold is approached slowly, preexisting bed irregu- larities may trigger downstream ripples in the interval of metastability before the threshold is attained. How- ever, if the threshold is passed rapidly, or if marked preexisting irregularities do not occur, mounds will spon- taneously appear and transform themselves into regular ripples. Other schemes for the formation of transverse bed forms have been proposed, in which the wavelike per- turbation of flow precedes bed deformation, rather than arising from interaction with the bed. Cartwright (1959) has proposed that the shelf-edge sandwave field of La Chapelle Bank in the Celtic Sea are responses to stationary internal waves (tidal lee waves) in the strati- fied water column. Furnes (1974) has analyzed the for- mation of sand waves in response to internal waves of a fluid whose density stratification is a consequence of its suspended load. While compatible with the field evi- dence, these models for sand wave formation remain un- confirmed. They are important contributions, however, if only in that they reduce the bias toward the results of experimental laboratory work. The space and time scales and the internal structure of shelf flows are qualitatively different than those of laboratory flumes and there is no reason to assume that such further modes of sand wave formation do not exist in nature. types of transverse bed forms. Field and labora- tory observations show that there tend to be two over- lapping populations of transverse bed forms: ripples, with wavelengths up to 0.6 m, and sand waves, with wave- lengths in excess of 0.6 m (Fig. 8). Sand waves com- monly bear ripples on their backs. The two populations appear to be responses to two distinct genetic mech- anisms. As small forms grow up through the velocity gradient of the boundary layer, the zone of maximum stress on the upcurrent flank shifts to the crest, at which point the entire upcurrent slope is erosional and the lee slope depositional; upward growth is stopped, and the ripple migrates at constant speed (Wilson, 1972, p. 200). Transverse bed forms of larger wavelengths are insensi- tive to the boundary layer velocity gradient and their upflank zone of maximum shear stress shifts to the crest only when the whole flow is significantly deformed by their upward growth. As a consequence, the equilibrium height of sand waves in shallow flows is proportional to flow depth, while the equilibrium height of ripples is 10" 8 6 A, I03 8 6 2 - 10' 1 1 — — 1 1 1 • i -i T - 1 — • o „0. o - o 8o° 0 •' 8/1 • 0 •/ • o • 0 8 * • t / . O j • •/ 0 o/ • *s • o • 0 o y .•» • ." K =10000 . _jC_ - / 0 » •:. - 0 * • • • • • oa •• - • *» •• ■ • Ripples] 0.19,0.27,0.28,0.45 l • Dunes j & 093 mm sonde ■ ■ i i 1 J — i — 10' 6 8 6 8.0« FIGURE 8. Wavelength parallel to flow of experimental ripples and sand waves in relation to flow depth. Data of Guy etal. (1966). From Allen (1970). depth independent for all flow depths (Allen, 1967); see Fig. 8. Expressions for the equilibrium heights of ripples and sand waves have been considered by Kennedy and used by McCave (1971). Stride (1970) has plotted measurements of height versus depth for sand wave fields of the North Sea at depths of 90 to 60 m, and found no correlation. Large bed forms grow slowly, and equilibrium heights may be rarely obtained in such shallow tidal seas subject to strong periodic storm surges. Deep-sea sand waves (Lons- dale and Malfait, 1974) can obviously never equilibrate with total water depth, although the significant flow depth may be only a small fraction of total depth, because of density stratification. The distinction between small- and large-scale bed forms may be due to more than interaction of wavelength with the velocity gradient. Kennedy ( 1 964) has suggested that small transverse bed forms represent perturbations of the traction and saltation loads that move very near to or on the bed, and hence must react quickly to changes of flow speed. Larger transverse bed forms, on the other hand, could reflect a perturbation of the sus- pended load, which will tend to respond slowly, and therefore over a large distance to a change in flow speed. 572 168 SUBSTRATE RESPONSE TO HYDRAULIC PROCESS cm 10 S - Separotion point or line A - Attachment point or lins FIGURE 9. Skin friction lines and streamlines associated with a portion of a bed of experimental ripples in fine-grained quartz sand. Mean flow velocity 22 cm /sec from left to right. Mean flow depth 4.5 cm. Note alternative notation of coordinate axes. From Allen (1970). In continental margin sand wave fields, there are often three orders of transverse bed forms: current ripples, sand waves, and larger sand waves. McCave (1971) sug- gests that the two classes of sand waves may be the con- sequence of Kennedy's two categories of substrate re- sponse. Because of the turbulent diffusion of sand normal to the flow direction, an initially equant interface pertur- bation will tend to extend itself normal to flow, hence the quasi-two-dimensional nature of ripples and sand waves (ripple profile does not change down the length of the ripple crest). However, at increasing values of mean velocity, and therefore of turbulent instantaneous velocity component, transverse bed forms tend to become three-dimensional (Znamenskaya, 1965). As crests be- come locally inclined to the mean flow direction, the horizontal "roller eddy" of the separation bubble be- comes a horizontal helical vortex (Fig. 7) and irregular patterns of skin flow result (Fig. 9). Under yet more in- tense flows, the irregularities may take on ordered pat- terns (Fig. 10). Bagnold (1956) attributes one particu- larly common pattern, that of the lingoid ripple, to "... the partial diversion of grain flow . . . and its fun- neling into channels between existing ripples; deposition (of a new lingoid ripple) would take place immediately downstream of such a funnel." A diagonal or diamond- like pattern of lingoid ripples results. TRANSVERSE BED FORMS AND FLOW REGIMES. It has long been known that as a shallow flow over a nonco- hesive substrate intensifies, a sequence of bed configura- tions transpires (Simons et al., 1961; Simons and Rich- ardson, 1963; Guy et al., 1966). The flow variables gov- erning this sequence are h, depth of flow; u, mean velocity of flow; p, density of fluid; ps, density of sediment; /i, viscosity of fluid; and D, mean diameter of sediment. The critical parameters are fluid power (proportional to «3; see Chapter 8) and grain size (Fig. 11). Grain density is variable to the extent that heavy minerals may be present; and fluid density and viscosity vary somewhat with temperature and salinity. Flow depth determines whether or not the flow is subcritical or supercritical as expressed by the dimensionless Froude number F = u/(gk)112, where (gh)112 is the celerity of a shallow water wave. In supercritical flows (F > 1), surface waves couple with substrate perturbations (anti- dunes) that tend to migrate upcurrent. On the continen- tal margin supercritical flows are confined to the swash and breakpoint zones of the surf, and to tidal flats; and the antidunes and rhomboid ripples that form in these zones are ephemeral. Southard (1971) and Southard and Boguchwal (1973) have argued that bed configuration diagrams such as Fig. 1 1 should be presented in terms of dimensionless depth, velocity, and grain size to eliminate the overlap- 573 BED FORMS 169 FIGURE 10. (a) Lingoid ripple pattern on shelf floor off Cape Hatteras, Sorth Carolina, (b) Lingoid ripples on back oj sand ivave, straight ripples in trough, same area. ping of fields that occurs in diagrams utilizing fluid power or bed shear stress. Figure 1 1 shows that dunes (sand waves) occur at higher values of fluid power than do ripples by them- selves. This fact is consonant with Kennedy's suggestion that sand wave formation involves suspended load trans- port, which requires higher values of fluid power than does bed load transport. TRANSVERSE BED FORMS AND TIDAL FLOWS. Tidal flows, which reverse every 6 hours, generate transverse bed forms in a cohesionless substrate. Tidal current ripples are no different than ripples generated by unidirectional currents, except that their sense of asymmetry is reversed as the tide changes. Small sand waves (height of 1 m or less) may have their asymmetries partly or wholly re- versed by strong reversing tidal currents (Klein, 1970). Larger sand waves tend to display a time-integrated response to reversing tidal flows, maintaining an ebb or flood asymmetry in accord with the dominant flow com- ponent residual to the semidiurnal cycle. "Cat-backed" sand waves are large sand waves that have a sloping upcurrent side, a flat top, and (in profile) an •■ear" perched on the edge of the downcurrent slope (Van Veen, 1936). The ear is a response to the subordinate portion of the tidal cycle. Tide-formed sand waves in areas of equal ebb and flood flow are commonly symmetrical. As distance from shore increases, the tidal current is no longer reversing but rotary (Chapter 5). The advent of midtide cross flow tends to inhibit the formation of sand waves large enough to survive through the tidal cycle (McCave, 1971). Under such circumstances longi- tudinal bed forms are favored (Smith, 1969). Longitudinal Bed Forms Wilson (1972) comments that practically all longitudinal bed form elements, whether formed in wind or water, are initiated by regular helical vortices with axes parallel to flow. His reasons for his admittedly sweeping assertion are as follows: 1 . Longitudinal helical flow cells occur in many dif- ferent kinds of situations. They are the only kind of flow perturbations known to fluid mechanics whose wave- length is measured normal to the mean flow direction. 2. With the exception of alternating parallel lanes of fast and slow flow, the double helical pattern is the only one that meets the theoretical requirements, namely bi- lateral symmetry parallel to flow, regular repetition nor- mal to flow, and conformity with the law of continuity. 3. Many investigations of flow over longitudinal bed forms resulted in some evidence for the occurrence of helical flow over the longitudinal elements, for instance, model ripples and dunes (Allen, 1968a); in riv«r chan- nels (Gibson, 1909); over tidal sand ridges (Houbolt, 1968); and over desert dunes (Hanna, 1969). The theory of longitudinal flow perturbation is less well developed than the theory of transverse flow per- turbation. Such perturbations are not as obvious in lab- oratory flumes as transverse (streamwise) flow pertur- bations, and many occur at scales far beyond those of laboratory flumes. As in the case of transverse bed forms, longitudinal bed forms appear to be able to form in re- sponse to perturbations of boundary flow, or in response to perturbations of the whole flow field. As in the case of transverse bed forms, they appear to form during the course of flow-substrate interaction, and also in response to the preexisting internal structure of a sheared flow. Preexisting flow structures appear to be more impor- tant than in the case of transverse perturbations. Perhaps the most general statement that can be made is that in a sheared flow that is wide relative to its depth, a sig- nificant portion of flow energy must be diverted to an ordered secondary flow component, in order to maintain lateral flow continuity. At least three basic varieties of such secondary flow structure exist. 574 170 SUESTRATE RESPONSE TO HYDRAULIC PROCESS 20,000 10.000 8000 _ 6000 u S 4000 £ 2000 3t 1000 o. 800 § 600 w 400 200 40 PLANE BED (Even lominotion) RIPPLES (Cross-lamination) PLANE BED (Even lamination) NO SEDIMENT MOVEMENT 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.1 D (cm) FIGURE 1 1. Bed forms in relation to stream power and grain size. Data of G. P. Williams, H. P. Guy, D. B. Simons, and E. V. Richardson. From Allen (1970). MICROSCALE LONGITUDINAL BED FORMS (PARTING lineation). It has been repeatedly suggested that the logarithmic boundary layer tends to be so patterned, although an adequate analytic model has not yet been devised (Schlichting, 1962, pp. 500-509). Kline (1967) and Kline et al. (1967) have conducted dye experiments in flumes which suggest that the laminar sublayer and the lower part of the buffer sublayer of the turbulent boundary layer have a structure characterized by vig- orous transverse components of flow (see discussion, p. 94). Dye introduced into the boundary layer forms into bands that are more slowly moving than those in the intervening water zones. Although the streaks are randomly generated, they have a mean transverse spac- ing of Xr = \00 i>u* in which v is the kinematic viscosity and u> is the shear velocity (Kline, 1967). The response to helically structured boundary flow over a cohesionless particulate substrate is, however, well known; it is the ubiquitous parting lineation (Sorby, 1859), so named for the tendency of flagstones (silty sandstones with strong bedding fissility) to exhibit lineations on bedding planes. Closer examination reveals a waveform bedding surface whose undulations parallel flow direction; ridges are a few grain diameters high and are up to several centimeters apart (Allen, 1964; 1968a, pp. 31-32); see Fig. 12. There is clear evidence for the divergence and convergence of bottom flow in that the azimuths of long grains are bimodal, although this evidence does not re- solve the secondary flow pattern. A similar structure has been reported from mud beds (Allen, 1969). Here the notches are frequently narrower than the ridges. Coupling probably occurs between bed and flow struc- ture, in that the grain ridges localize flow cells. Also, the sand of the ridges is coarser (Allen, 1964) and the result- ing roughness would tend to slow crestal flow. This feature would cause downstream growth in the retarded wake of the grain ridges, and would perhaps induce upward ridge growth until ridge crests reach a level whose flow is rapid enough to counteract growth. MESOSCALE LONGITUDINAL BED FORMS (CURRENT lineations). "Current lineations" (McKinney et al. 1974) is a generic term for low-amplitude strips of sand resting on a coarser substrate (sand ribbons) and for strips of coarse sand or gravel flooring and elongate de- pression of slight depth (longitudinal furrows). Current lineations are a larger scale of longitudinal bed form, with spacings ranging from a few meters to many hun- 575 BED FORMS 171 15 20 25 30 35 40 M«an flow velocity (cm/sec) FIGURE 12. The mean transverse spacing of parting lineations as a junction of mean flow velocity and flow depth. From Allen (1970). dreds of meters (Allen, 1968a). They are best observed by means of sidescan sonar (Figs. 13 and 14). The large- scale patterns are characteristic of shelves with strong tidal flows (Kenyon, 1970); see Fig. 15. Sand ribbons and longitudinal furrows are probably the most common mesoscale bed form on the continental shelf, being widely distributed on both shelves dominated by tidal flows (Kenyon, 1970; Belderson et al., 1972) and storm-domi- nated shelves (Newton et al., 1973; McKinney et al., 1974); see Fig. 14. Unpublished data of the Atlantic Oceanographic and Meteorological Laboratories, Mi- ami, Florida, show them to be characteristic of large sectors of the Middle Atlantic Bight. Relief is negligible relative to width. Kenyon would restrict the term "sand ribbon" to features having length-to-width ratios of 1 : 40 and refers to shorter, broader features as elongate sand patches, but the distinction seems arbitrary. Unlike parting lineation ridges, sand ribbons tend to consist of streamers of finer sand in transit over a coarser substrate which may, in fact, be a gravel. A continuum may exist between a sand ribbon pattern of sand and gravel streets of equal width, to a "longitudinal furrow" pattern (Stride et al., 1972; Newton et al., 1973) in which widely spaced, elongate erosional windows in a thin sand sheet reveal a coarser substrate. Ribbon width relative to the width of the interribbon zone does not appear to be simply a function of the height of a sinus- oidal surface of sand layer over a coarser substrate, since the windows in profile are notchlike affairs separated by flat plateaulike zones (Fig. 16). Furthermore, the ribbons are commonly rather asymmetrical, as though the sand sheet occurs at minimum thickness on one side, and in- creases very slowly to maximum thickness on the other side. Such asymmetrical ribbons could be interpreted as degraded sand waves, but the sharpness of the contacts plus the lack of relief suggests instead asymmetrical helical flow cells (Fig. 17). Small sand ribbons may be large-scale analogs of the responses described in the preceding section that involve the entire logarithmic boundary layer. However, most shelf sand ribbon patterns have spacings of tens or hun- dreds of meters, and as noted by Allen (1970, p. 69), can only be responses to the entire depth of flow. Theoretical studies (Faller, 1971; Faller and Kaylor, 1966; Brown, 1971; Lilly, 1966) and experimental studies (Faller, 1963) show that there is a mechanism by which a helical flow pattern may be induced in the large-scale flows of the continental margin. When such flows are in geostrophic balance (pressure term balanced by Cori- olis term in the equation of motion; see p. 25). The lower portion of the flow is an Ekman boundary layer (see p. 97). The basal meter behaves as a logarithmic boundary layer in that flow speed decreases rapidly to a zero value or nearly so at the seafloor. Flow direction (in the northern hemisphere) is to the left of the free- stream direction, however, since the Coriolis term is re- duced along with mean velocity; the equation of motion more nearly constitutes a balance between friction and pressure terms. With increasing height off the bottom, flow is more nearly geostrophic and its direction is more nearly parallel to the isobars, until free stream condi- tions are reached. Thus velocity vectors at successively higher levels constitute a left-handed spiral. On the con- tinental shelf, this lower Ekman boundary layer may extend to the base of the mixed layer, if it exists, or to the surface, where it is overprinted with a right-handed Ekman spiral (upper Ekman boundary layer) because of direct wind stress (Ekman, 1905, Plate 1). Above a critical Reynolds number, this Ekman layer is unstable. However, because the instability transpires in an Ekman field subject to the Coriolis effect, the in- stability does not result in random turbulence, but in- stead in a regular pattern of secondary flow (Faller, 1971, pp. 223-225). In this pattern zones of surface convergence, downwelling, and bottom divergence alter- nate with zones of surface divergence, upwelling, and bottom convergence. The resulting flow structure con- sists of horizontal helical cells with alternating right- and left-hand senses of rotation (Fig. 6). Angles of con- vergence and divergence (pitch) are generally a few degrees; in other words, the secondary component of flow is weak, relative to the main geostrophic component. The flow cells may occur at several scales (Faller and Kaylor, 1966). In laboratory studies (Faller, 1963), smaller scale cells have a spacing of approximately 1 ID, where D is a characteristic depth of the Ekman layer, and tend to be oriented up to 14° to the left of the mean flow. They occur at Reynolds numbers above 125. Larger 576 172 SUBSTRATE RESPONSE TO HYDRAULIC PROCESS A\ 111 UMIN.'.TlON DIRECTION 'ERO RANGE °-^— -*■ PRO' IE 0. SEA FLOOR ^^R DE BENEATH The Ship pL \ I VID UNES OF ECHOES 2 FROM SIDE LOBES NEAR EDGE OF MA DISTANT EDGE OF MAIN BEAM FIGURE 13. (A) Sidescan sonar. (B) The resulting record. A, Bottom of seafloor; B, turbulence in water column due to ship's wake; C, zigzag pattern is due to refraction of sound in density- stratified water; D, main lobe (see above); E, side lobe. From Belderson et al. (2972). scale cells have wavelengths much greater than 1 \D and are oriented to the right of geostrophic flow. They occur at much lower Reynolds numbers. Helical flow structure may occur in the upper Ekman layer where its wind-driven stirring creates the mixed layer above the thermocline (Faller, 1971), or may occur in the lower layer (Faller, 1963). In surface helical flow the downwelling zones that collect the high-velocity wind-driven surface water are more sharply defined than the upwelling zones (Langmuir, 1925). In bottom helical flow, downwelling zones deliver higher velocity water to the seafloor, and may also be more sharply defined than the upwelling zones. During intense flows, when strati- fication breaks down and the layers partly or completely overlap, a compound top-to-bottom helical flow struc- ture might be expected. Observational and theoretical studies required to link this scheme to the observed shelf sand ribbon patterns have not been undertaken; however, there are obvious points of compatibility. The ribbons tend to be parallel, or oriented at a small angle to the regional trend of shelf contours, and presumably to the mean geostrophic flow direction. The greater intensity of downwelling zones would explain the dissimilar width of ribbons and inter- vening erosional windows. The Reynolds numbers re- quired are not excessive for either tide- or wind-driven shelf flows. 10 m and spacings measured in kilometers, are called sand ridges (Off, 1963; Swift et al., 1974); see Fig. 18. They are comparable in scale to the seif dunes, and the yet larger "draas" of the sand seas of the world's deserts (Wilson, 1972), except that as befits submarine sand bodies, their side slopes are much lower, usually being measured in fractions of a degree. Sand ridges appear to form in two basic types of situ- ations. They are characteristic of the reversing flows of tidal estuaries and bays, where they tend to form in complex arrays parallel to the estuary axis (Figs. \8A,B). They also appear on inner shelves of coasts undergoing erosional retreat (Figs. 18C,Z)), where they appear to be specific responses to the coastal boundary of the shelf flow field (Duane et al., 1972; Robinson, 1966; Swift et al., 1972a); the mechanism is discussed in detail in Chapter 14. On the inner shelf, either tidal or storm flows may be the forcing mechanism (Duane et al., 1972; Swift, in press). The ridges tend to extend obliquely seaward from the shoreface. Like sand ribbon patterns, the generally larger scale sand ridge fields tend to com- prise discontinuous sheets of finer sand over a coarser substrate. However, where sand ridges build up into the wave-agitated zone on open coasts, their crests tend to be coarser than their flanks although generally not as coarse as the substrate exposed in trough axes (Houbolt, 1968; Swift et al., 1972b; Stubblefield et al., 1975). longitudinal sand ridges. Large-scale longitudinal bed forms of the continental margin, with relief of up to SAND RIDGES AS RESPONSES TO WIND-DRIVEN FLOWS. Sand ridges are found on continental shelves seaward 577 BED FORMS 173 SEA FLOOR NORTH OF SHIP WATER COLUMN { WATER { COLUMN SEA FLOOR SOUTH OF SHIP FIGURE 13— Continued. of active inner shelf-generating zones, and occur as well on some shelves whose inner margins are not actively forming them (Swift et al., 1974). It appears that shelf flow fields can continue to maintain these ridges of coastal origin after the retreating shoreline has abandoned them, and can even- locally generate them afresh (see discussion in Chapter 14). Without a main- taining mechanism, shelf flows might be expected to de- grade sand ridges by leveling crests and filling in troughs. In fact, however, the ridges of the Atlantic continental shelf tend to expose compact clays or lag gravels in their troughs, indicating continuing trough scour (Swift et al., 1972a; McKinney et al., 1974). There are a variety of competing hydraulic mech- anisms that may serve to explain the formation and maintenance of large-scale sand ridges on the continental margin, none of which is clearly understood. On the open shelf, cellular flow structure in storm flows, as de- scribed in the preceding section, may couple with the shelf floor. Such cellular flow structure might generate ridges along the coastal boundary (see discussion in Chapter 14) where wind-driven flows are frequent and intense and there is an abundant supply of sand. As these ridges have been left behind by the retreating shoreline during the Holocene transgression, the same cellular flow structure may be continuing to maintain them on the outer shelf. If this analysis is correct, then sand ribbons and sand ridges may differ in that sand ribbons represent responses to one flow event or a flow season while sand ridges rep- resent time-averaged responses to repeated flow events, whose emerging relief tends to localize the position of large-scale flow cells. Events capable of forming such large-scale flow cells would presumably be peak storm 578 01*2 WW kW-f wmmv* ittr ridges of asymmetrical ribbons; 3: pinna (pele- cypod) bed. (b) 1: Symmetrical ribbon; 2; sand waves. From Newton et al. (1973). i\1± FIGURE 14. Sand ribbon patterns from the Spanish Sahara shelf. Light is sand; dark is coarser sand and gravel. Distortion ellipse with scales on first record, (a) 1, 2: Sharply defined 174 579 BED FORMS 175 TYPE A appron horizontal sraie FIGURE 1 5. Categories of sand ribbon from the shelf around the British Isles, and associated current velocities. From Kenyon (1970). or tidal flows, in which secondary circulation involves the entire water column. The most problematic aspect of shelf ridge fields is the depth-to-width ratios of the troughs, which range from 1 : 10 to 1 : 150. The smaller ratios are compatible with the "type I" flow cells of Faller's (1963) experi- mental work, whereas the large ratios may derive from Faller's "type II" cells which have "much greater" di- ui) FIGL'RE 16. (a) Wide sand ribbons alternating uith narrow streets of coarse sand (erosional uindous) due to intersection of sinusoidal surface of sand sheet uith horizontal substrate, (b) Same pattern due to notchlike incision of erosional uindous. The latter pattern is a common one. mensions. It is perhaps easier to conceive of such flat- tened cells if it is remembered that the central down- welling zone is the only sharply defined portion of a double helical flow cell; the marginal zones of diffuse upwelling may take up much of the "stretch," serving to complete flow continuity in a fashion analogous to the role of "ground" in electrical circuitry. The advent of appreciable relief in a growing system of sand ridges may bring other hydraulic mechanisms into play. Secondary flow cells appear to be an innate response to channeled flow. It has long been known (Gibson, 1909; Jeffreys, 1929; Einstein and Li, 1958; Leopold et al., 1964, pp. 251-284; Wilson, 1972) that driftwood or ice in a river tends to move toward the center, whose surface is elevated slightly above that of the margins, and that the thread of maximum velocity tends to be depressed below the surface. The result is a double helical flow cell, in which bottom water spreads, rises along the margins, converges over the center, and sinks there. Flume and theoretical work (Kennedy and Fulton, 1961; Gessner, 1973) indicates that in flumes of 580 176 SUBSTRATE RESPONSE TO HYDRAULIC PROCESS (A) AS SAND WAVE (PROFILE) PLAN) PLAN) FIGURE 17. Interpretation oj asymmetrical sand ribbons; B is more probable. square cross section the unequal distribution of turbulent (Reynolds) stresses will result in secondary flow from the center toward the corners. The resulting multiple flow cells do not form the double helical pattern postu- lated for natural channels, however, and their applica- bility to the natural situation is uncertain. Bagnold (1966, pp. 112-115) offers an independent explanation. He suggests that there is asymmetrical exchange of mo- mentum between the bottom boundary layer of a river and the overlying flow in that tongues of boundary water abruptly penetrate the overlying flow, to be compensated by a general sinking of the latter (see Chapter 7, p. 98). This results in elevation of the water surface over the channel axis where this exchange is most intense. The ensuing pressure head, he suggests, drives the secondary component of flow. The preceding discussion has dwelt on double helical flow cells as mechanism for generating a large-scale sand ridge topography. An attempt has been made to match (a) (c) (b) FIGURE 18. Patterns of sand ridges on tide-dominated shelves. From Off (1965). (d) 581 BED FORMS 177 theoretical and experimental studies with characteristics of shelf ridge fields. However, such large-scale coupling of flow with substrate has not yet been observed in the field. It is worth noting that there is an independent mechanism that is theoretically capable of maintaining a ridge topography, either by itself or in conjunction with other mechanisms. The mechanism described by Smith (1969) requires that ridges be aligned with mean flow direction and that the variance in flow direction be high, either because the flow is a rotary tidal flow; or because it is storm-driven, and the direction of flow varies during a storm and also among storms (see Chap- ter 4). As a consequence, most flows intense enough to entrain sand will be aligned at an oblique angle with the ridges during most of their duration. Flow across the ridge can be treated two-dimensionally according to slender body theory (Smith, 1969) and the stability analysis of Smith (1970) applies (see the preceding sec- tion). First one flank then the other flank of the ridge will be eroded, with sand transferred to the crest and far flank each time. SAND RIDGES IN RESPONSE TO TIDAL FLOWS. The reversing nature of nearshorc tidal flows adds another mechanism capable of maintaining a ridged topography. The velocity of the tidal wave is a function of water depth, and flow over a step or across a sill in a cohesion- less substrate will result in a phase discontinuity between the behavior of the tidal wave on either side of the sill (Fig. 19). Thus, when the tide is in the last stages of ebb on one side, it may be already beginning to flood on the other, so that there is an opposing sense of flow over the crest of the sill. If the flow is broad relative to its depth, and if the sill is a relatively large-scale feature, then this is an inherently unstable situation. Slight irregularities in the seafloor on either side of the sill will result in in- equalities in the rate of propagation of the tidal wave, and during the brief period of opposing flow the two water masses on either side of the sill will tend to inter- penetrate along a zigzag front. The tongues of flow on either side of the sill crest will tend to scour its channels until the crestline of the sill has also become zigzag (Fig. 19). A channel that is on the side of the sill facing the oncoming tidal wave and opens in that direction is called a. flood sinus (Ludwick, 1973). It experiences an excess of flood over ebb discharge (is flood-dominated). A channel on the other side of the sill is called an ebb sinus, and is ebb dominated. Scour in the interdigitating channels of such an ebb- flood channel complex is matched by aggradation of the interchannel shoals. This transfer is perhaps aided by the secondary circulation mechanisms described in the preceding section. As a consequence of the residual cur- rent pattern, net vectors of bottom flow integrated over the tidal cycle meet obliquely head-on over crests (Fig. 20), with the result that each ridge becomes a sand cir- culation cell, or closed loop in the sand transport pattern. Mean dro dx is negative along these vectors toward the m A m ® \ V/y 'J' '///A ^ ® PRE-EXISTING HIGH TIDAL CURRENT NEAR SLACK WATER SAND RIDGE FIGURE 19. Hypothetical scheme of development of an ebb-flood channel topography as a consequence of the phase lag, experienced by the tidal wave in its passage across a submarine sill. 582 178 SUBSTRATE RESPONSE TO HYDRAULIC PROCESS ■J VM v'/> ilAA Of • 1 "^\MEAN BOTTOM FLOW ~~^x BOUNDARIES FOR MEAN FLOW SHEARS mm SHOAL FIGURE 20. Nearshore and offshore patterns of tidal flow about sand ridges. Based on Luduick (1970b) and Caston and Stride (1970). zone of residual current shear at the ridge crest. The ridges therefore tend to aggrade toward the intertidal zone where they become "drying shoals" or swash plat- forms dominated by wave processes (Oertel, 1972). On open coasts, however, wave surge erosion may balance tidal current construction when the crests are still sub- tidal. Tidal sand ridges that partition ebb- and flood-domi- nated flows usually experience a stronger residual cur- rent on one side than on the other, and tend to migrate away from that side (Fig. 21). In such cases, where the cross-ridge component of flow is strong, a ridge may itself deform into a sigmoid pattern, and eventually into two or three separate ridges (Caston, 1972); see Fig. 21. Sills with interdigitating ebb and flood channel sys- tems occur at the mouths of most tidal estuaries (Fig. 22), as a consequence of frictional retardation of the tidal wave within the estuary, and the resultant phase lag. On the Bahama Banks, they occur on the inner sides of islands, where the two wings of the tidal wave meet as they refract around the island (Fig. 23). The evolution of such a system portrayed in Fig. 19 probably rarely occurs in nature; the channel systems form simultane- ously with such sills, not afterward. For instance, the ebb-flood channel system of the Chesapeake Bay mouth shoal appears to have formed during the Late Holocene reduction in the rate of sea-level rise (Ludwick, 1973). It can be inferred from the present morphology that the sill prograded south across the bay mouth, fed by the littoral drift discharge of the Delmarva coastal com- partment. The ridges would have developed in zig- zag fashion, alternately and progressively segregating the flow into ebb-dominated and flood-dominated channels (Fig. 24). Tidal flows often occur in the presence of salinity stratification so intense as to persist for part or all of the tidal cycle despite the powerful mixing effect of flow turbulence. Flow structure may be yet more complicated as a result. In Fig. 25, the residual circulation over the Hudson estuary mouth shoal (New York Harbor en- trance) is seen to be a resultant response to flow inter- digitation due to the phase lag effect (Fig. 25C) and to estuarine (two-layer) circulation (Fig. 255). £Z> SAND WAVE 0 GRAIN ORIENTATION — BANK CRESTLINE ff DOMINANT CURRENT; SAND STREAM // MAJOR, MINOR BANK MOVEMENTS FIGURE 21. Above: Anatomy of a tidal sand ridge. From Houbolt (1968). Below: Evolution oj a tidal sand ridge. From Caston (1972). 583 FIGURE 2 2. A hydraulic and geomorphic interpretation of the net nontidal {residual) flow pattern at the bottom in the entrance to Chesapeake Bay. Numbers are measured flood and ebb durations at the bottom in hours; small arrows show measured direction of near-bottom currents. Stippled areas are shoaler than 18 ft. Ruled areas show where there is an ebb or a flood flow predomi- nance. From Ludwick (1970a). 584 179 180 SUBSTRATE RESPONSE TO HYDRAULIC PROCESS FIGURE 23. Charles True. Ebb-flood channel pattern on the Great Bahama Bank. Altitude 3000 ft. Photo: Stratification may also play a role in the formation of ridge topography within the estuary. Weil et al. (in press) describe the formation of subtidal levees in Dela- ware Bay as the consequence of the penetration of sub- surface saline tongues up the channels during flood tide, resulting in an internal pressure head that can drive channel axis downwelling (Fig. 26), and as a conse- quence of the overriding of the tongues by fresher water during the ebb tide, with similar effects. One of us (Ludwick, in press) has mapped near-bottom con- vergences and divergences of flow in the Chesapeake Bay mouth during flood tide. These are absent during the more thoroughly stratified ebb. Here stratification appears to inhibit channel axis downwelling and bottom current divergence (see Fig. 31). Velocity profiles of Chesapeake Bay mouth tidal flows tend to be parabolic but with markedly sigmoidal perturbations (Ludwick, 1973), and may imply the presence of standing internal waves or wakes from shoals. Tidal flows in confined estuary mouths thus tend to develop an interdigitating pattern of ebb- and flood- dominated channels, whose sequence of partitioning ridges tends to alternate between clockwise and counter- clockwise current flows (Fig. 20). On the offshore shelf, however, the tide becomes rotary rather than reversing and a different pattern tends to appear (Caston and Stride, 1970). Ridges appear in free-standing sets rather than in continuous zigzag arrays. Residual current shears occur in channel axes as well as on ridge crests, and successive ridges experience residual flows with the same sense of rotation. Huthnance (1972) attributes this open shelf flow pat- tern to interaction of the ridges with the shelf tide. His model considers a rectilinear reversing tide whose flow directions make an oblique angle with the ridge axis. The cross-ridge component of flow must accelerate over the ridge crest for continuity reasons. The ridge-parallel component of flow must decrease up the upcurrent flank as the water column shoals, and influence of friction be- comes proportionately greater. However, because high- velocity fluid is being transported into the shoal region, the decrease in the ridge-parallel flow component lags behind the decrease in depth. On the downcurrent flank, the restoration of the ridge-parallel flow to am- bient velocity is similarly lagged. When the tide changes, upcurrent and downcurrent flanks reverse roles. When flow is averaged over the tidal cycle, a clockwise pattern of residual flow around the ridge results (or counter- 585 BED FORMS 181 FIGURE 24. Evolution oj "submarine zigzag spit" across Chesapeake Bay mouth. Based on Luduick (1972). clockwise, depending on whether the oblique, reversing tidal stream is sinistral or dextral with respect to the ridge). Huthnance proposes a second mechanism whereby in the northern hemisphere, Coriolis force also results in clockwise circulation. Huthnance's mechanisms are interesting, but the re- quirement that there be a significant angle between the axis of the tidal stream and that of the ridge presents a problem. The ridges are a response element within the flow field-substrate system, not an independent forcing element. It seems doubtful that ridges of cohesionless sand could maintain a significant angle with the tidal stream for any length of time, unless it were somehow an equi- librium response to flow. Smith (1969) notes that tidal sand ridges might be expected to orient themselves parallel to the long axis of the tidal ellipse, as the sand body would then be at a small angle of attack through- out most of the high-velocity part of the tidal cycle. According to slender body theory the cross-shoal com- ponent of flow during this period can be considered to be two-dimensional and driven by the cross-shoal pres- sure gradient. It would thus sweep sand first up one side and then up the other as the tide rotated. Possibly the dilemma is resolved by the lag effect cited by Postma (1967) and Stride (1974); see Fig. 27. Because of a lag in the entrainment of sand, the period of maximum sand transport is believed to lag behind maximum flood flow, and again behind maximum ebb flow. The result should be to align the response element (sand ridge) obliquely across the major axis of the tidal ellipse. It also seems likely that the large-scale, un- bounded tidal flow field of the open shelf might at least locally generate Ekman flow structure during midtide, and couple with inner shelf ridge fields in the fashion that has been suggested for wind-driven flows. Limiting Conditions of Bed Form Formation on the Continental Margin In attempting to apply the elements of bed form theory presented on the preceding pages to analyses of conti- 586 182 SUBSTRATE RESPONSE TO HYDRAULIC PROCESS SANDY HOOK 0 2 0 t . i ' ROCKAWAY 8 e o © FIGURE 2 5. (A) Profile across the Hudson estuary mouth (mouth oj New York Harbor) contoured for velocity residual to the tidal cycle. The flow pattern is a resultant response to component flow patterns shown in (B) and (C). (B)'Schematic diagram oj two-layered, estuarine flow pattern. (C) Schematic diagram oj com- ponent oj flow pattern resulting jrom phase lag oj tidal wave. From Duedall et al. (in press), after Kao (1975). nental margin sedimentation patterns, it is useful to keep in mind some generalizations presented by Allen (1966; 1968a, pp. 50-53; 1968b). Allen, following Bagnold (1956), notes that the grain-fluid system is a decidedly multivariate one, and that we should expect to find co- existing instabilities of several different modes and scales. Flows that experience both transverse and streamwise perturbations may develop bed form associations consisting of two different bed form types, for instance a reticulate pattern with sand waves overprinted on sand ridges. Likewise, flows tend to experience one or more instability modes at several different spatial scales, resulting in a bed form hierarchy as, for instance, in the case of the Diamond Shoals sand wave field (Hunt et al., in press), where photos show that current ripples are superimposed on sand waves (Fig. 10) and sidescan sonar records show in turn that sand waves are superimposed on giant sand waves (Fig. 28). Elaborate hierarchical associations of bed forms occur over vast areas of the earth's surface, in subaerial sand seas (Wilson, 1972), and also in widely disparate environments on the continental margin (com- pare Fig. 29 with Fig. 30). The physical scale at which bed forms occur affects their response characteristics, and in turn the flow fre- quency to which they are tuned. For instance, on the crests of the drying sand ridges of the Minas Basin, current ripples reflect radial drainage at the last stages of ebb, sand waves are oriented with slip faces seaward as responses to peak ebb flow, while larger dunes locally are landward facing, reflecting a stronger flood than ebb flow (Swift and McMullen, 1968; Klein, 1970; Dalrymple, 1973). The largest scale transverse and longitudinal bed forms have had to readjust to continuous environmental change associated with Holocene deglaciation and the accompanying transgression of the continental margins. In some cases, they appear to have taken nearly the duration of the Holocene to form. Sand ridges on the central New Jersey shelf have basal strata containing 11,000-year-old shells (Stubblefield et al., 1975). These features and many other shelf ridge fields appear to have been formed by shoreface ridge formation and detach- ment (Swift, in press) during the Holocene transgression; see Fig. 28, Chapter 14. Plan geometry and internal structure of Atlantic Shelf ridge fields suggest that ridge spacing has increased by ridge migration or coalescence as the water column deepened (Swift et al., 1974). 587 BED FORMS 183 A - FLOOD ISOVELS, cm/ttc DENSITY ISOPLETHS B-EBB C -TRANSPORT DOMINANCE FIGURE 26. Tidal sand ridge as a submarine levee, formed in response to stratified flow. From Weil et al. (in press). FIGURE 27. Lag effects in a rotating tide. Radial arrows are vectors of tidal current velocity at intervals through the tidal cycle. Sand entrainment starts at velocity Vi and continues to velocity V2. Net sand transport is to right and onshore. From Postma (1967). The response of larger bed forms tends to lag beyond the peak flow event or may comprise an average re- sponse to repeated events. Allen (1973) notes that maxi- mum sand wave height in the Fraser River and Gironde estuary is lagged behind peak tidal flow by as much as a quarter of the tidal period. Ludwick (1972) notes that tidal sand waves are symmetrical over portions of the Chesapeake Bay mouth where the tidal cycle is symmetrical, but are asymmetrical when there is flood or ebb asymmetry in the tidal cycle. Thus their response to reversing tidal flow is time-averaged in a manner entirely analogous to the response of oscillation ripples to wave surge (Chapter 8). Tidal sand waves in Chesapeake Bay mouth attain their greatest height and slopes during the summer months when wave activity is FIGURE 28. Sidescan sonar record of sand waves on the back of giant sand waves, Cape Hatteras, North Carolina. Sand waves are larger in coarse sand of trough than on finer sand of giant waves. Giant sand waves are 120 m apart, 7 m high. Unpublished data of Swift and Hunt. 588 184 SUBSTRATE RESPONSE TO HYDRAULIC PROCESS 28"20'N 28" 15' N - 74°25'W 74°20'W FIGURE 29. Erosional jurrows and large-scale silt ridges on the Blake-Bahamas outer ridge, 4700 m depth. From Hollister et al. (1974)- at a minimum; they are degraded by the more intense wave activity of winter months (Ludwick, 1970a,b, 1972). The Piatt Shoals sand wave field on the open Virginia shelf appears to be induced by storm flows, hence it may have the opposite behavior pattern; sand waves would be highest in the winter months and would tend to be degraded by fair-weather wave surge and burrowing organisms during the summer (Swift et al., 1974). ESTIMATES ON SEDIMENT TRANSPORT A Numerical Model port systems, and to determine the rates of erosion, trans- port, and sedimentation associated with these elements. Much of the material in the following chapters is devoted to available information of sediment sources, pathways, and sinks on the continental margin. However, there have been very few attempts to estimate rates of sedi- ment transport. It should be possible to measure the time history of a marine flow by means of a current- meter array, then employ the empirical relationships developed by hydraulic engineers to estimate the time history of sediment transport. The difficulties however, are formidable. In situ recording current meters are ex- pensive and difficult to maintain. Data processing is com- 589 1 I :. 70 15 70 00 69 45 FIGURE 30. Pattern of sand waves (dark lines) and sand ridges appear to have initially formed as sboref ace-connected ridges at Nantucket Shoals. Significant highs are stippled and ebb-flood similar to those attached to the shoreface of modern Nantucket channel couplets are indicated by arrows. Ebb and flood sinuses as Island, and to have been stranded on the shelf floor as the shoreface inferred from morphology, are indicated by arrows. The ridges underwent erosional retreat. From Swift (2975). 185 590 186 Sl'BSTRATE RESPONSE TO HYDRAULIC PROCESS wave surge. There is no general agreement on the most satisfactory transport equation, or on the applicability of equations developed under laboratory conditions to the complex deep-water flow fields of the continental margin. A simple numerical model for estimating rates and patterns of sediment transport in areas of tidal flow has been devised by one of us (Ludwick, in press). It is summarized below. structure of the model. The model requires deter- mination of the distribution across the study area of a sediment transport index ro«ioo over a tidal cycle. The index is derived from Bagnold's (1956) work (see p. 1 13), in which sediment discharge q is set proportional to fluid power to defined as so that 0> = Toll q = KtqU where t» is bottom shear stress and u is the depth-aver- aged flow velocity. For convenience of measurement, Ludwick substitutes Mioo, the velocity measured 100 cm off the bottom. With this information, it is possible to use the sediment continuity equation (p. 166) to deter- mine the distribution and relative rates of aggradation and erosion along streamlines of sediment transport. determination of Tn. In order to determine the distribution of to, current velocities were measured over 27 hour intervals at 24 stations in the mouth of Chesa- peake Bay. A Kelvin Hughes direct reading current meter was employed from an anchored ship. At each station the current meter was used successively through 1 1 different depth levels. Hourly profiles with 4 minute observation periods were obtained at each level. These speed values were then reduced to pseudo- synoptic data sets for standard times and depths at each station (see Ludwick, in press). Each data set was fitted to Hama's (1954) parabolic velocity defect law (see the discussion in Chapter 7, p. 96). This empirical func- tion pertains to outer boundary flow, at distances greater than 0. 1 5/?, where h is the thickness of a turbulent bound- ary layer, or water depth in the case of fully developed flow in a uniform channel. The equation is ^ -"(•-!)' where ux is the free stream velocity, u is velocity at dis- tance z above the bed, and u* is the friction or shear velocity. An estimate of u* on the bottom is then obtained by least squares curve fitting. The value can be converted to an estimate of ro, the boundary shear stress, through the relationship u* = (to p)1'2, where p is fluid density. This measurement, obtained by observation of the entire water column, provides a far more reliable estimate for ro«ioo, the fluid power, than does ro determined simply from the product Cioop(«ioo)2, due to uncertainties in determining Cioo (see Chapter 7, p. 99). maps of bed sediment transport. Values of the sediment transport index obtained for 24 stations must be converted to maps of near-bottom streamlines of sedi- ment transport. The values are adjusted to the mean tidal range, a process described by Ludwick (1973). They are further corrected by subtracting 150 dyne-cm,' sec cm'-', a threshold value for the initiation of sediment movement (Fig. 31). The value at each station is inte- grated separately over each flood and each ebb half- cycle, and the results are averaged for ebb and flood. After averaging and integrating, the units of the sedi- ment transport index are dyne -cm/cm2 per average ebb (or flood) half-cycle. The values obtained at points on the field grid of 24 irregularly placed stations must then be redistributed over a systematic grid. This is a problem in vector inter- polation. The first step is to prepare separate maps of the north-south and east-west components of ro«ioo for the flood cycle. Each map is contoured. The flood com- ponent maps are superimposed. Resultant vectors may now be calculated at any point, if the contour interval is sufficiently small. The density of resultant vectors may be increased in areas of complex flow. Finally, stream- line maps may be prepared by drawing lines that are everywhere tangent to the vectors (Fig. 32-4). The proc- ess is repeated for the ebb half-cycle and the vector sum of ebb and flood (Figs. 33.4 and 34.4). FIGURE 31. Tidal current speed and bottom shear stress at a flood channel station, Chesapeake Bay mouth. Speed values are for a distance of 18.5 ft off the bottom. Total depth, 56 ft. Observed speeds were corrected from mean tidal range and averaged over six cycles, zo is the roughness length estimated from vertical velocity profiles, k. is the height of bottom roughness elements, and rc is the critical shear stress, calculated from the Shields entrainment diagram. From Luduick (1970a). 591 76*1, W FIGURE 3 2. Ebb-directed sediment transport at the bed. (a) Streamlines of the sediment transport vector toUioo; depths are in meters; vertically ruled areas are shoaler than 5.5 m. (b) Erosion-deposition chart on which erosion is positive ( + ) and deposition is negative 75°|S5 (— ),• units are dyne-cm /cm2 per ebb half-tidal cycle per 463 m of transport X 10'*; cross-hatched areas indicate erosion intensity greater than —400 units; stippled areas indicate deposition intensity greater than + 400 units. From Ludwick {in press). 187 592 FIGURE 33. Flood-directed sediment transport at the bed. (a) Streamlines of the sediment transport vector T0Uioo." depths are in meters; vertically ruled areas are shoaler than 5.5 m. (b) Erosion-deposition chart on which erosion is positive ( + ) and deposition is negative ( — ); units are dyne-cm I cm1 per flood halj- tidal cycle per 46J m of transport X 10'*; cross-hatched areas indicate erosion intensity greater than — 4OO units; stippled areas indicate deposition intensity greater than +400 units. From Ludwick (in press). 593 FIGURE 34. Vector sum of ebb and flood sediment transport at the bed. (a) Streamlines of the resultant sediment transport vector roUioo," depths are in meters; vertically ruled areas are shoaler than 5.5 m. (b) Erosion-deposition chart on which erosion is positive (+) and deposition is negative ( — ); units are dyne-cm I cm1 per tidal cycle per 463 m oj resultant transport X 10*; cross-hatched areas indicate erosion intensity greater than — 400 units; stippled areas indicate deposition inten- sity greater than + 400 units. From Ludwick (in press) . 594 189 190 SUBSTRATE RESPONSE TO HYDRAULIC PROCESS It is important to realize the limitations of the stream- lines of bottom sediment transport so determined. The redistribution of information has not in any way in- creased the accuracy or resolution of the original data. Sediment input in a stream tube does not necessarily equal sediment output, since deposition or erosion may occur. There is no underlying stream function in the method, and the spacing of the streamlines is not a measure of transport rates. It is assumed, however, that transport is confined to a path of unit width that con- forms to the bathymetry of the seafloor, and that the streamline is the center line of this path. It is further assumed that conditions are steady and nonuniform for the entire pattern. net sedimentation maps. As a separate and ensuing procedure, it is possible to estimate the extent of areas of erosion and deposition, and also the rates at which these processes occur. The estimate utilizes the sediment continuity equation written in terms of discharge: dr) dq dt dx where r\ is bed elevation relative to a datum plane, / is time, 6 is a dimensional constant related to sediment porosity, q is the weight rate of bed sediment transport per unit width of streamline path, and x is distance along the streamline. Discharge (q) may be taken as proportional to to"ioo and the right-hand partial derivative may be approxi- mated by a finite difference: dq dx Aq (t0Uu — - — K - Ax (TqUioo)] *•> — Xi The term x-> — *i is held constant arbitrarily at a value of 465 m, hence dt oc — AroUioo Thus a decrease in transport rate along a transport path- way induces deposition; an increase causes erosion. The resultant vector map for a half-cycle is super- imposed on the equivalent streamline map. The mag- nitude of ro«ioo is determined at equispaced points along each streamline; AtoMioo is determined as a positive or negative value, and mapped over the area of study as an estimate of relative erosion and deposition intensity. In Figs. 325, 33B, and 345, net sedimentation maps have been prepared for the ebb and flood half-cycles and for the vector sum of ebb and flood. utility of the model. Such a manipulation of the data from 24 current-meter stations extracts a surprising amount of information from them. Streamlines of bed sediment transport associated with the ebb tidal jet are seen to pass over the bay mouth shoal in parallel fashion. Flood streamlines, however, form a pattern in which flow divergence and flow convergence alternate across the flow in sympathy with the topographic pattern of interdigitating ebb and flood channels. The vector sum map shows a complex pattern of flow dominance that is also correlated with bottom morphol- ogy. Patterns of net sedimentation do not correlate as closely with the topography, probably because they do not indicate the areas of maximum relief, but instead areas undergoing maximum change. In particular, the parabolic shoals that envelop each ebb or flood sinus are seen to be subject to a systematic pattern of sedimenta- tion. The sides of shoal segments that face the dominant flow, however obliquely, are eroding. The crest and downcurrent sides, however, are undergoing aggrada- tion. Thus, the processes that Smith (1970) has inferred to cause sand waves (see p. 165) appear also to be applicable to ebb-flood channel topography. The model can be generalized for portions of the shelf dominated by storm flows if each flow event is treated in the same fashion as Ludwick treated a tidal half-cycle, or sediment transport can be integrated over an arbi- trary period of observation. Transport Estimates from Tracer Dispersal Studies One of the main stumbling blocks in divising quantita- tive estimates of sediment transport has been the limited applicability of empirical relationships based on labora- tory observations to the complex flows of the marine environment. The model partially circumvents this prob- lem by recourse to the sediment transport index, based on Bagnold's generalized evaluation of fluid power (Chapter 8, p. 113). In doing so, it provides only a relative answer. Sediment transport is proportional to fluid power, and the proportionality constant remains unevaluated. Despite this sacrifice, the model has not resolved the problem of adequately treating the complex time and space scales of marine flows. In particular, it fails to deal with the vexing problem of the role of bot- tom wave surge in "lubricating" bottom sediment trans- port by reducing the effective transport threshold for a unidirectional flow component (see discussion, Chapter 8, p. 115). This wave surge factor becomes part of the pro- portionality constant. Wave-surge-amplified transport is not that critical a problem in the analysis of a primarily tide-built topography. It becomes critical, however, in open shelf transport, where wind-driven unidirectional flows attain their maximum intensities just as the wave regime does. 595 ESTIMATES ON SEDIMENT TRANSPORT 191 It is clear that the best resolution of a marine sediment transport system will be obtained when a model such as the one presented above is employed together with an independent method for evaluating the proportionality constant. The most promising method to date is the de- ployment of radioisotope tracers. Fluorescent tracers have been widely used (see Ingle and Gorsline, 1973; Inman and Chamberlain, 1959). However, since count- ing of labeled particles must be done in the laboratory, the analysis is tedious, and it is generally not possible to watch the development of dispersal patterns in real time. Furthermore, fluorescent tracers have a very limited ap- plicability seaward of the surf, as a consequence of the limited sensitivity of the method and the difficulties of hand sampling. Tracer dispersal can usually be observed in an area 50 m in diameter or less, under fair-weather conditions. After a storm, when a major displacement of sediment has occurred, the tracer grains are liable not to be there at all. Radioisotope tracers avoid much of this difhculty. The RIST (Radio Isotope Sand Tracer) system, devel- oped by Oak Ridge National Laboratories and the Coastal Engineering Research Center (Duane, 1970) de- tects radioisotope-labeled tracers by means of a towed scintillometer. The data logging system provides for real time readout, which greatly aids mapping of the dis- persal pattern. A relatively long-lived isotope such as ruthenium- 103, with a half-life of 40 days, permits effective tracing for three times that duration, or an entire storm season. A numerical estimate of sediment transport may be fine-tuned by quantitative analysis of radioisotope tracer dispersal patterns. The procedure requires not only the mapping of successive outlines of the tracer pattern but the ability to account for all of the labeled particles at each stage. In order to establish such a mass balance, it is necessary to know the depth of reworking, which is the depth to which labeled particles have penetrated during dispersal. This depth can be calculated from the known ability of the sediment to absorb radiation, if it is assumed that the tagged particles are mixed into the reworked layer in a homogeneous fashion. If tracer particles can be accounted for through successive map- pings of the dispersal pattern, then the rate of sediment transport as indicated by dispersal of tracers may be checked against the rate of transport as estimated from current-meter records in one of two ways. Transport rates may be determined directly from the dispersal pattern and compared with estimates based on current- meter records. Or current-meter records may be used to simulate tracer dispersal patterns, and these ideal patterns may be compared with observed dispersal pat- terns. Figure 35 shows a series of radioisotope dispersal pat- terns from an experiment conducted by J. W. Lavelle and his associates, Atlantic Oceanographic and Meteor- ological Laboratories, Miami, Florida. Water-soluble bags of labeled sand were released along a line in 20 m of water on the south shore of Long Island during April and May of 1974. Over a period of 69 days, a typical fair-weather dispersal pattern formed (panels B-F). The data in these panels have been corrected for decay, but in the last panel, the corrected values on the margins of the pattern are so much weaker than background, that they were lost when smoothed background values were subtracted. The mild summer storms during this period only briefly generated flows strong enough to transport sand, and much of the labeled sand remained in close proximity to the drop line, where it was not readily resolved by the towed scintillometer. It will be necessary to apply a statistical smoothing function to the data, in order to undertake a mass balance calcu- lation for the dispersed tracer sand. If continued experi- mentation leads to improved field techniques and data processing, then radioisotope tracers should prove a fruitful method for calibrating numerical models of sedi- ment transport. SUMMARY The size frequency distribution of marine sand samples tends to be log-normally distributed. This distribution, as defined by its mean and standard deviation, is the "signature" of the depositional event, and deviations from log normality, as measured in terms of skewness and kurtosis, may be taken to reflect both the provenance and the hydraulic history of the sediment. The modal diameter of a sand deposit is that grain size most likely to arrive and least likely to be carried away from the place of deposition; progressively coarser sizes are progressively less frequent because they are progressively less likely to arrive, and progressively finer grain sizes are progressively less frequent because they are more likely to be carried away. This intuitively ap- parent concept can be explained in terms of probability theory. As sand progresses down a transport path by inter- mittent hops, it tends to leave its coarser grains behind, and the deposits are progressively finer in the direction of transport (have undergone progressive sorting). They also will tend to be fine-skewed, particularly if the in- tensity of hydraulic activity also declines down the transport path. Moss has shown that the size distributions of marine sands tend to be made up of three log-normal popula- 596 192 SUBSTRATE RESPONSE TO HYDRAULIC PROCESS * GRAB SAMPLES 40°29'50" 40°29'45" 73°42'00' 73°41 30 B DAY 0 - PRELIMINARY i D f/A, DAY 22 - 4 PRELIMINARY DAY 48 % TRANSPORT W 20 % TRANSPORT 0 5 J">VJV>W. ° n DAY 3 RIST DROP i — i — i — r 10 DAY 69 % TRANSPORT 0 10 % TRANSPORT 0 20 THRESHOLD VELOCITY =20 OCM/SEC ~r~~i — i — "i — r — i — i — r^ — n — r — i — i — i — i — i — i — ^^n — i — i- "i — i — i — i — ' — i — i — 20 J DAY 22 i i 30 % TRANSPORT 0 20 40 "i — r 50 ~i 1 r 60 DAY 48 o 30 £ O 0 O FIGURE 3 5. Time sequence oj dispersal oj radioisotope- labeled sand, south shore of Long Island, April 22-July 2, 1974- (A) Background radioactivity in arbitrary units. Heavy black line is line oj emplacement oj sand labeled with ruthenium-105. (B-F) Maximum extent oj detectable signal ajter removal oj background on successive mapping days. Data have been corrected jor decay. Bottom panel: time-velocity record {jagged line). Height oj vertical bars indicates percentage oj total bed load transported, as determined from a normalized sediment transport index, Vioo — Vr- Width oj vertical bars indicates duration oj transport event. Rose diagrams indicate direction oj transport. Length oj radial bar indicates percentage oj transport during that event; width is proportional to direc- tion and intensity. Velocity was sampled every 10 minutes. Unpublished data oj Lavelle et al., Atlantic Oceanographic and Meteorologic Laboratories, Miami, Florida. tions as a consequence of the fashion in which the bed is built; the main subpopulation (A population) com- prises the framework of the deposit. A fine B population is interstitial; a coarse C population is the consequence of "traction clogs." The A:B:C ratio varies with the flow regime. Bed forms are irregularities in the particulate sub- strate of a fluid flow. Sheared flow is innately unstable, and tends to develop repeated patterns of velocity vari- ation, either parallel or normal to the flow direction. Such instabilities tend to interact with the bed so as to cause rhythmic variations in elevation. Flow and bed 597 REFERENCES 193 perturbation amplify each o'ther until equilibrium is attained. Bed forms occur in associations (more than one genetic type present), in hierarchies (successive scales of bed forms of similar genesis), and in hierarchical associations. Trans- verse bed forms interact with wavelike perturbations of flow transverse to the flow direction. Current ripples are small-scale transverse bed forms that appear to result from boundary layer instability; their wavelength is in- dependent of depth. Sand waves result from perturbations of the whole flow field, or a density-homogeneous por- tion of it. Several scales of sand waves may occur to- gether; the smaller scale may perhaps be a response to primarily tractive transport, whereas the larger scale may be a response to primarily suspensive transport. Longitudinal bed forms are caused by velocity perturba- tions parallel to flow. In some cases, the perturbation takes the form of horizontal, flow-parallel vortices whose sense of rotation is alternately right and left handed, and this may be true for all cases. Parting lineations are small-scale longitudinal bed forms. They are sand ridges a few grain diameters high and a few centimeters apart. Current lineations have wavelengths ranging from a few meters to hundreds of meters; their heights are negligible relative to width. In a characteristic pattern, sand ribbons occur on a gravel substrate. In longitudinal furrow patterns, the lows are more sharply defined than the intervening highs. Sand ridges may have wavelengths of hundreds of meters to several kilometers, and amptitudes of 10 m or more. They are induced by tidal flows at estuary mouths, by tidal or wind-driven flows on the shelf, and perhaps by boundary undercurrents on the continental rise. They appear to be time-averaged responses to intermittent flow, and in many cases have survived successive en- vironmental transitions associated with the Holocene transgression. A simple numerical model for estimating bed load transport on the continental margin requires as input current-meter measurements. Streamlines of bottom sedi- ment transport may be based on the sediment transport index. The index is derived from Bagnold's energetics, in which sediment discharge is set proportional to fluid power, equal to bottom shear stress times the depth- averaged velocity. The sediment continuity equation is used to predict areas and relative intensities of erosion and deposition. In this equation, the discharge gradient along a streamline is related to the time rate of change of bottom height above a datum by a dimensional constant. It may be possible to evaluate Bagnold's proportion- ality constant for sediment transport by means of mass balance assessments of radioisotope dispersal patterns. However, improvements in field tracer techniques and data processing are required before such an evaluation is possible. ACKNOWLEDGMENTS We thank J. R. L. Allen and R. L. Miller for critical review of the manuscript. SYMBOLS Cioo drag coefficient determined from measurements 100 cm above the bottom D grain diameter h water depth K a constant ks height of bottom roughness elements q sediment discharge t time u velocity u depth-averaged velocity u^ time-averaged free stream velocity Kioo velocity 100 cm off the bottom u* shear velocity; shear stress with velocity units x distance downstream y distance above the bed Z distance transverse to flow Zo roughness length « dimensional constant related to sediment porosity X wavelength p fluid density p., sediment density rj elevation of a surface above a datum t shear stress ro shear stress at the bed rc critical bed shear stress v kinematic viscosity a; fluid power REFERENCES Allen, J. R. L. (1964). Primary current lineation in the lower old red sandstone (Devonian) Anglo-Welsh Basin. Sedimentology, 3: 89-108. Allen, J. R. L. (196b). On bedforms and paleocurrents. Sedimen- tology, 6: 153-190. 598 194 SUBSTRATE RESPONSE TO HYDRAULIC PROCESS Allen, J. R. L. (1967). Depth indicators of clastic sequences. Mar. Geol., 5: 429-446. Allen, J. R. L. (1968a). Current Ripples. Amsterdam: North-Holland Publ., 433 pp. Allen, J. R. L. (1968b). The nature and origin of bedload hier- archies. Sedimentology, 10: 161-182. Allen, J. R. L. (1969). Erosional current markings of weakly co- hesive mud beds. J. Sediment. Petrol., 34: 607-623. Allen, J. R. L. (1970). Physical Processes of Sedimentation. New York: American Elsevier, 248 pp. Allen, J. R. L. (1973). Phase difference between bed and flow in natural environments and their geologic relevance. Sedimen- tology, 20: 323-329. Bagnold, R. A. (1954). Experiments on a gravity-free dispersion of large solid in a Newtonian fluid under shear. Proc. Roy. Soc. London, A225: 49-63. Bagnold, R. A. (1956). The flow of cohesionless grains in fluids. Proc. Roy. Soc. London, A249: 235-250. Bagnold, R. A. (1966). An approach to the sediment transport problem from general physics. U.S. Geol. Surv. Prof. Paper 4 22- 7, 37 pp. Belderson, K. H., N. A. Kenyon, A. H. Stride, and A. R. Stubbs (1972). Sonographs of the Sea Floor. Amsterdam: Elsevier, 185 pp. Brown, R. A. (1971). A secondary flow model for the planetary boundary layer. J. Atmos. Sci., 27: 742-757. Cartwright, D. E. (1959). On submarine sand waves and tidal lee waves. Proc. Roy. Soc. London, A253: 218-241. Caston, V. N. D. (1972). Linear sand banks in the southern North Sea. Sedimentology, 18: 63-78. Caston, V. N. D. and A. H. Stride (1970). Tidal sand movement between some linear sand banks in the North Sea off north- east Norfolk. Mar. Geol., 9: M38-M42. Chandrasekhar, B. (1961). Hydi adynamic and Hydromagnelic Stability. New York and London: Clarendon Press, 654 pp. Dalrymple, R. W. (1973). Preliminary investigations of an inter- tidal sand body, Cobequid Bay, Bay of Fundy. Marit. Sedi- ments, 4: 21-28. Duane, D. B. (1970). Synoptic observations of sand movement. Proc. 12th Coastal Eng. Con/., Washington, DC, 799-813. Duane, D. B., M. E. Field, E. P. Meisburger, D. J. P. Swift, and S. J. Williams (1972). Linear shoals on the Atlantic inner shelf, Florida to Long Island. In D. J. P. Swift, D. B. Duane, and O. H. Pilkey, eds., Shelf Sediment Transport: Process and Pattern. Stroudsburg, Pa.: Dowden, Hutchinson & Ross, pp. 447-498. Duedall, I., H. B. O'Connors, J. B. Parker, R. Wilson, W. Miloski, and G. Hulse (in press). The seasonal and tidal variation of nutrients and chlorophyl in the New York Bight Apex. Marine Sciences Research Center, State University of New- York. Einstein, H. A., and H. Li (1958). Secondary currents in straight channels. Trans. Am. Geophys. Union, 39: 1085-1088. Ekman, U. W. (1905). On the influence of the earth's rotation on ocean currents. Ark. Mat. Astron. Fysik, 2: 1-53. Emery, K. O. and E. Uchupi (1972). Western North Atlantic Ocean: topographs, rocks, structure, water, life, and sedi- ments. Am. Assoc. Pet. Geol. Mem., 17: 532. Exner, F. M. (1925). L'ber die wechselw irking zwischen wasser und geschiebe in flussen. Set-itngsbenchte der Academic der W'issen- schaften. Vienna Heft 3-4. pp. 165-180. Faller, A. J. (1963). An experimental study of the instability of the laminar Ekman boundary layer. J. Fluid Mech., 15: 560- 576. Faller, A. J. (1971). Oceanic turbulence and the Langmuir cir culations. Ann. Rev. Ecol. System., 2: 201-233. Faller, A. J. and R. B. Kaylor (1966). A numerical study of the instability of the laminar Ekman boundary layer. J. Atmos. Sci., 23: 466-480. Friedman, G. M. (1961). Distinction between dune, beach, and river sands from their textural characteristics. J. Sediment. Petrol., SI: 514-529. Furnes, G. K. (1974). Formation of sand waves on unconsolidated sediments. Mar. Geol., 16: 145-160. Gessner, F. B. (1973). The origin of turbulent secondary flow along a corner. J. Fluid Mech., 58: 1-25. Gibson, A. H. (1909). On the depression of the filament of maxi- mum velocity in a stream flowing through an open channel. Proc. Roy. Soc. London, A82: 149-159. Guy, H. P., D. B. Simons, and E. V. Richardson (1966). Summary of alluvial channel data from flume experiments 1956-1961. U.S. Geol. Sun: Prof. Pap. 462- J , 96 pp. Hama, F. R. (1954). Boundary characteristics for smooth and rough surfaces. Trans. Soc. Naval Architects and Marine Engi- neers, 62: 333-358. Hanna, S. R. (1969). The formation of longitudinal sand dunes by large helical eddies in the atmosphere. J. Appl. Meteorol., 88: 874-883. Hollister, C. D., R. D. Flood,. D. A. Johnson, P. Lonsdale, and J. B. Southard (1974). Abyssal furrows and hyperbolic echo traces on the Bahama outer ridge. Geology, 2: 395-400. Houbolt, J. J. H. C. (1968). Recent sediments in the southern bight of the North Sea. Geol. Mijnbouw, 47: 245-273. Huthnance, J. M. (1972). Tidal current asymmetries over the Norfolk sandbanks. Estuarine Coastal Mar. Sci., 1: 89-99. Ingle, J. C. and D. S. Gorsline (1973). Use of fluorescent tracers in the nearshore environment. In Tracer Techniques in Sediment Transport. Vienna: Int. At. Energy Agency, pp. 125-148. Inman, D. L. (1949). Sorting of sediments in the light of fluid dynamics. ./. Sediment. Petrol., 19: 51-70. Inman, D. L. and T. K. Chamberlain (1959). Tracing beach sand movement with irradiated quartz. J. Geophys. Res., 64: 41-47. Jeffreys, H. (1929). On the transverse circulation in streams. Proc. Cambridge Phil. Soc, 25: 20-25. Kao, A. (1975). A study of the current structure in the Sandy Hook-Rockaway Point transect. M.S. Research Paper, Ma- rine Sciences Research Center, State University of New York, Stony Brook. Kennedv.J. F. (1964). The formation ol sediment ripples in closed rectangular conduits and in the desert. J. Geophys. Res., 69: 1517-1524. Kennedy, J. F. (1969). The formation of sediment ripples, dunes and antidunes. Ann. Rev. Fluid Mech., 1: 147-168. Kennedv, R. J. and J. F. Fulton (1961). The effect of secondary currents upon the capacity of a straight open channel. Trans. EIC, 5: 12-18. Kenyon, N. H. (1970). Sand ribbons of European tidal seas. Mar. Geol., 9: 25-39. 599 REFERENCES 195 Kiein, G. D. (1970). Depositional and dispersal dynamics of inter- tidal sand bars. J. Sediment. Petrol., 40: 1095-1127. Kline, S.J. (1967). Observed structure features in turbulent and transitional boundary layers. In G. Sovran, ed., Fluid Me- chanics of Internal Flow. Amsterdam: Elsevier, pp. 27-68. Kline, S. J., VV. C. Reynolds, F. A. Schraub, and P. W. Runstler (1967). The structure of turbulent boundary layers. J. Fluid Mech., 30: 741-773. Krumbein, W. C. (1934). Size-frequency distributions of sedi- ments. J. Sediment. Petrol., 4: 65-77. Langhorne, D. N. (1973). A sandwave field in the outer Thames estuary. Mar. Geol., 14: 129-143. Langmuir, I. (1925). Surface motion of water induced by the wind. Science, 87: 119-123. Leopold, L. B., M. G. Wolman, and J. P. Miller (1964). Fluvial Processes in Geomorphology. San Francisco: Freeman, 522 pp. Lilly, D. K. (1966). On the instability of Ekman boundary flow. J. Atmos. Set., 23: 481^94. Lin, C. C. (1955). The Theory oj Hydrodynamu Stability. London and New York: Cambridge University Press, 155 pp. Lonsdale, P. and B. Malfait (1974). Abyssal dunes of foraminiferal sand on the Carnegie Ridge. Geol. Soc. Am. Bull., 85: 1697- 1712. Ludwick, J. C. (1970a). Sand waves in the tidal entrance to Chesapeake Bay: Preliminary observations. Chesapeake Sci., 11: 98-1 10. Ludwick, J. C. (1970b). Sandwaves and Tidal Channels in the Entrance to Chesapeake Bay. Technical Report 1, Inst, of Oceanography, Old Dominion University, Norfolk, Va., 7 pp. Ludwick, J. C. (1972). Migration of tidal sandwaves in Chesa- peake Bay entrance. In D. J. P. Swift, D. B. Duane, and O. H. Pilkey, eds., Shelf Sediment Transport: Process and Pattern. Stroudsburg, Pa.: Dowden, Hutchinson & Ross, pp. 377-410. Ludwick, J. C. (1973). Tidal Currents and Zig-Zag Sand Shoals in a Wide Estuary Entrance. Technical Report 7, Inst, of Oceanog- raphy, Old Dominion University, Norfolk, Va., 23 pp. Ludwick, J. C. (in press). Tidal currents, sediment transport, and sand banks in Chesapeake Bay entrance, Virginia. In M. O. Hayes, ed., Second International Estuarine Conf. Proc. Myrtle Beach, South Carolina, Oct. 15-18, 1973. McCave, I. N. (1971). Sandwaves in the North Sea off the coast of Holland. Mar. Geol., 16: 199-225. McKinney, T. F. and G. M. Friedman (1970). Continental shelf sediments of Long Island, New York. J. Sediment. Petrol., 40: 213-248. McKinney, T. F., VV. L. Stubblefield, and D. J. P. Swift (1974). Large scale current lineations on the central New Jersey shelf: Investigations by sidescan sonar. Mar. Geol., 17: 79-102. Middleton, G. V. (1968). The generation of log normal size fre- quency distributions in sediments. In Problems of Mathematical Geology. Leningrad: Science Press, pp. 37-46. Moss, A. J. (1962). The physical nature of common sandy and pebbly deposits (Part I). Am. J. Set., 260: 337-373. Moss, A. J. (1963). The physical nature of common sandy and pebbly deposits (Part II). Am. J. Sci., 261: 297-343. Moss, A. J. (1972). Bedload sediments. Sedimentology, 18: 159-219. Newton, R. S., E. Seibold, and F. Werner (1973). Facies distribu- tion patterns on the Spanish Sahara continental shelf mapped with sidescan sonar. Meteor. Forsch. Ergebnisse, 15: 55-77. Oertel, G. F. (1972). Sediment transport of estuary entrance shoals and the formation of swash platforms. J. Sediment. Petrol., 42: 857-863. Off, T. (1963). Rhythmic linear sand bodies caused by tidal cur- rents. Am. Assoc. Pet. Geol. Bull., 47: 324-341. Postma, H. (1967). Sediment transport and sedimentation in the marine environment. In G. H. LaufF, ed., Estuaries. Washing- ton, D.C.: Am. Assoc. Adv. Sci., pp. 158-180. Raudkivi, A. J. (1967). Loose Boundary Hydrodynamics. Oxford: Pergamon, 331 pp. Robinson, A. H. W. (1966). Residual currents in relation to shoreline evolution of the East Anglian coast. Mar. Geol., 4: 57-84. Rosenhead, L. (1963). Laminar Boundary Layers. London and New York: Oxford University Press (Clarendon), 687 pp. Russell, R. D. (1939). Effects of transportation on sedimentary particles. In P. P. Trask, ed., Recent Marine Sediments. London: Thomas Murphy, pp. 32-47. Schlichting, H. (1962). Boundary Layer Theory. New York: McGraw- Hill, 647 pp. Simons, D. B. and E. V. Richardson (1963). Forms of bed rough- ness in alluvial channels. ASCE Trans., 128: 284-302. Simons, D. B., E. V. Richardson, and M. L. Albertson (1961). Flume studies using medium sand (0.45 mm). U.S. Geol. Surv. Water Supply Pap. 1498A, 76 pp. Smith, J. D. (1969). Geomorphology of a sand ridge. J. Geol., 77: 39-55. Smith, J. D. (1970). Stability of a sand bed subjected to a shear flow of low Froude number. J. Geophys. Res., 30: 5928-5940. Sorby, H. C. (1859). On the structures produced by the current present during the deposition of stratified rocks. Geologist, 2: 137-147. Southard, J. B. (1971). Representation of bed configurations in depth-velocity-size diagram. J. Sediment. Petrol., 41: 903-915. Southard, J. B. and L. A. Boguchwal (1973). Flume experiments on the transition from ripples to lower flat bed with increasing grain size. J. Sediment. Petrol., 43: 1114-1121. Southard, J. B. and J. R. Dingier (1971). Flume study of ripple propagation behind mounds on flat sand beds. Sedimentology, 16: 257-263. Stride, A. H. (1970). Shape and size trends for sand waves in a depositional zone of the North Sea. Geol. Mag., 469-477. Stride, A. H. (1973). Interchange of sand between coast and shelf in European tidal seas (abstract). In Abstracts Symposium on Estuarine and Shelf Sedimentation, Bordeaux, France, July 1972, p. 97. Stride, A. H. (1974). Indication of long term tidal control of net sand loss or gain by European coasts. Estuarine Coastal Mar. Sci., 2: 27-36. Stride, A. H., R. H. Belderson, and N. H. Kenyon (1972). Longi- tudinal Furrows and Depositional Sand Bodies of the English Channel. Memoire Bureau Recherches Geologique et Minieres, No. 79, pp. 233-244. Stubblefield, W. L„ J. W. Lavelle, T. F. McKinney, and D. J. P. Swift (1975). Sediment response to the present hydraulic regime on the central New Jersey shelf. J. Sediment. Petrol., 15: 227-247. 600 196 SUBSTRATE RESPONSE TO HYDRAULIC PROCESS Swift, D.J. P. (1975). Tidal sand ridges and shoal retreat mas- sifs. Mar. Geol., 18: 105-133. Swift, D J. P., D. B. Duane, and T. F. McKinney (1974). Ridge and swale topography of the Middle Atlantic Bight, North America: Secular response to the Holocene hydraulic regime. Mar. Geol. Swift, D.J. P., B. Holliday, N. Avignone, and G. Shideler (1972a). Anatomy of a shoreface ridge system, False Cape, Virginia. Mar. Geol., 12: 59-84. Swift, D. J. P., J. C. Ludwick, and W. R. Boehmer (1972b). Shelf sediment transport: A probability model. In D. J. P. Swift, D. B. Duane, and O. H. Pilkey, eds., Shelf Sediment Transport: Process and Pattern. Stroudsburg, Pa.: Dowden, Hutchinson & Ross, pp. 195-223. Swift, D. J. P. and R. M. McMullen (1968). Preliminary studies of intertidal sand bodies in the Minas Basin, Bay of Fundy, Nova Scotia. Can. J. Earth Sci., 5: 175-183. Van Veen, J. (1936). Under zoekingen in der Hookden. The" Hague: Algemene Landsdrukkerij, 252 pp. Visher, G. S. (1969). Grain size distribution and depositional processes. J. Sediment. Petrol., 39: 1074-1106. Weil, C. B., R. D. Moose, and R. E. Sheridan (in press). A model for the evolution of linear tidal built sand ridges in Delaware Bay, U.S.A. In G. Allen and A. Klingbiel, eds., Symposium International: Relations Sedimentaires entre estuaires el plateaux con- tinentaux, University of Bordeaux, July 9-14, 1973. Wilson, I. G. (1972). Aeolian bedforms — their development and origins. Sedimentology, 19: 173-210. Wilson, I. G. (1973). Equilibrium cross-section of braided and meandering rivers. Nature, 241: 393-394. Yih, C. S. (1965). Dynamics of Non-homogeneous Fluids. New York: MacMillan, 235 pp. Znamenskaya, N. S. (1965). The use of the laws of sediment dune formation in computing channel .formation. Soviet Hydrology Selected Papers (Am. Geophys. Union 1966), 5: 415-432. 601 51 Reprinted from: Middle Atlantic Shelf and the New York Bight, ASLO Special Symposia, Volume 2, 69-89. Section 3 Geological processes Morphologic evolution and coastal sand transport, New York— New Jersey shelf1 Donald J. P. Swift, George L. Freeland, Peter E. Gadd, Gregory Han, J. William Lavelle, and William L. Stubblefield Atlantic Oceanographic and Meteorological Laboratories, NOAA, Virginia Key, Miami, Florida Abstract The surface of the New York-New Jersey shelf has been extensively modified by land- ward passage of nearshore sedimentary environments during the postglacial rise of sea level. The retreat of estuary mouths across the shelf surface has resulted in shelf valley complexes. Constituent elements include shelf valleys largely molded by estuary mouth scour, shoal retreat massifs left by the retreat of estuary mouth shoals, and midshelf or shelf-edge deltas. The erosional retreat of the straight coast between estuary mouths has left a discontinuous sheet of clean sand 0-10 m thick. During the retreat process, a sequence of oblique-trending, shoreface-connected sand ridges formed at the foot of the shoreface. As a consequence, the surficial sand sheet of the shelf floor bears a ridge and swale topography of sand ridges up to 10 m high and 2-4 m apart. The mechanics of sedimentation in these two nearshore environments ( estuary mouth and interestuarine coast ) are now being investigated for purposes of environmental man- agement as well as for further understanding of shelf history. In late fall and winter 1974, current meters were deployed on the Long Island coast and a radioisotope tracer dispersal pattern was traced over an 11-week period. Eastward or westward pulses of water were generated during this period of successive weather systems. Flows in excess of the computed threshold velocity of substrate materials were sustained for hours or days and were separated by days and weeks of subthreshold velocities, and the sand tracer pattern expanded accord- ingly. A single intense westward flow transported more sand than all the other events com- bined. The storm was anomalous with respect to the short term observation period, but it may in fact have been representative of the type of peak flow event that shapes the inner shelf surface. Systematic observations of sedimentation in New York Harbor mouth have not yet been initiated. However, reconnaissance data reveal a complex pattern of ebb- and flood-domi- nated zones that control the pattern of sand storage. We review in this paper our knowledge of the surface of the continental shelf off New York and New Jersey by considering two distinct topics: the geological history of this surface and the nature of sand transport across it. Our knowledge of the New York- 1 Contribution of the New York Bight Project of the NOAA Marine EcoSystems Analysis (MESA) Program. AM. SOC. LIMNOL. OCEANOGR. New Jersey shelf surface is primarily the result of a decade of work by K. O. Emery and his colleagues at the Woods Hole Oceanographic Institution. A summary of this information and much more has re- cently been provided by Emery and Uchupi (1972). As the work of the Woods Hole group drew to a close, we attempted to con- sider in greater detail some aspects of the morphologic evolution of the Middle At- gg SPEC. SYMP. 2 602 70 Geological processes lantic Bight surface (Swift et al. 1972, 1974; Swift 1973; Swift and Sears 1974; Stubble- field et al. 1974). A summary of this work constitutes the first section of this paper. As participants in NOAA's MESA (Ma- rine EcoSystems Analysis) program, we have been asked not only to evaluate the geological history of the New York Bight, but also to provide quantitative estimates of sediment transport that will be of direct use to environmental managers. It turns out that these two goals are closely related. Our sur- veys of the shelf surface have led us to in- fer that it has been shaped by the landward retreat of two basic sedimentary regimes during the Holocene transgression: tide- dominated sedimentation at estuary mouths, and the sand transport pattern of the ad- jacent shoreface and adjacent inner shelf. Environmental engineers and managers must deal with these same regimes. To satisfy their needs, we have initiated real-time studies of fluid motion and sub- strate response. State-of-the-art techniques for such studies are inadequate and progress has been slow. We report in the second por- tion of this paper fragments of our studies of sand transport to encourage colleagues engaged in similar studies. Our own initial experiments have raised more questions than they have answered. Evolution of the continental shelf surface Evolution of shelf valley complexes — The New York Bight is a pentagonal sector of the North American Atlantic shelf, extend- ing 800 km from Cape May, New Jersey, to Montauk Point, Long Island. Off New York, the shelf is 180 km wide ( Fig. 1 ) . The sandy shelf floor is divided into com- partments by shejf valley complexes extend- ing from the shoreline toward shelf edge canyons ( Fig. 1 ) . The most obvious ele- ments of these complexes are the shelf val- leys themselves which may appear as nar- row, well defined channels ( Delaware Shelf Valley; Hudson Shelf Valley) or as broad, shallow depressions which barely perturb the isobaths defining the shelf surface ( Block Shelf Valley, Long Island Shelf Val- ley, North New Jersey Shelf Valley, Great Egg Shelf Valley). Shelf valley complexes generally contain other morphologic ele- ments. The north rims of the shelf valleys 76°39° 75° 40° 74° 41° 73° 42° 74c --■SURFACE CHANNEL • ••• SUBSURFACE CHANNEL ' SCARP -200m- L\X| CUESTAS SHELF EDGE, MID-SHELF DELTAS ftS SHOAL RETREAT MASSIFS S-= SAND RIDGES -X. /\ ^L 73° 38° 72° 39° 71° 40° Fig. 1. Morphologic framework of the New York-New Jersey shelf. (Modified from Swift et al. 1972.) 603 New York-New Jersey shelf 71 tend to be elevated like levees above the adjacent shelf. Seaward ends of shelf valleys often terminate in delta terraces. Shelf val- ley complexes tend to be broken into seg- ments by coast-parallel scarps, which may have been formed by temporary stillstands of the returning Holocene sea. The origin of the shelf valley complexes is best inferred from the configuration of the Delaware Shelf Valley (Fig. 2), which can be traced without interruption into its modern estuary. The Delaware estuary mouth has a sill of sand nourished by littoral drift from the New Jersey coast ( Swift 1973). The sill is stabilized by an inter- digitating system of ebb- and flood-domi- nated channels, whose discharge inequali- ties are a consequence of the phase lag of the tidal wave in its passage across the sill ( Swift and Ludwick in press ) . The Dela- ware Shelf Valley may be traced directly into the flood channel of the main ebb channel-flood channel couplet. Its leveelike north rim may be traced directly into the complex of smaller ebb channels, flood channels, and sand banks on the north side of the estuary mouth. This shoal area serves as the depositional center for the littoral sand discharge of the New Jersey coast. The shelf valley complex, then, is not a drowned river valley, but is rather the track left by the retreat of the Delaware estuary mouth across the shelf during the Holocene sea-level rise. The shelf valley is the retreat path of a flood channel. The north flank levee is the retreat path of the estuary mouth shoal or is a shoal retreat massif — massif in the sense of a compound topo- graphic high consisting of smaller scale highs (Swift 1973). The surface channel does not directly overlie the buried river- cut channel but is offset to the south ( Sheri- dan et al. 1974). As the estuary retreated up the river valley, it not only tended to fill the river valley but in the final, estuary- mouth stage decoupled from it altogether by migrating to the south. The largely constructional nature of the Delaware Shelf Valley complex is also char- acteristic of the Great Egg Shelf Valley complex (Fig. 1), although the associated massif has been heavily dissected by the posttransgressional regime of southerly storm flows. To the north, however, the Hudson and Block Shelf Valleys occur on a terrain of innately greater relief. There are cuestalike highs, and the estuarine deposits only partly fill the shelf valleys. The deeply incised nature of the Hudson Shelf Valley may reflect the era when it received Great Lakes meltwater ( Veatch and Smith 1939 ) . Evolution of interfluves — Plateaulike in- terfluves between the shelf valleys have likewise been intensively modified by pas- sage of the shoreline. Interfluve surfaces range from exceedingly flat plains (slopes of 1:2,000) to irregularly undulating sand ridge topography (Fig. 3). Sand ridges ex- hibit up to 10 m of relief, are spaced 2 to 4 km apart, and their crestlines may be traced for tens of kilometers. Side slopes are gener- ally less than a degree. Crestlines are not quite parallel to the regional trend of the isobaths but tend to converge to the south- west with the shoreline (Fig. 1). Ridges at- tain their maximum development on the northeast sides of shoal retreat massifs. The ridges are molded into a surficial sheet of relatively homogeneous, well sorted sand, 0-10 m thick. In trough axes the sheet thins to a basal shelly, gravelly sand several decimeters or less thick, and a more hetero- geneous older substrate is locally exposed (Donahue et al. 1966; Stubblefield et al. 1974). This is commonly a muddy sand or mud deposited behind the retreating Holo- ce.ie barrier system (Stahl et al. 1974; Sheri- dan et al. 1974), but it is locally absent due to erosion or nondeposition, so that the Holocene sands rest directly on older Pleis- tocene sands. To understand the genesis of this post- glacial stratigraphy, it is necessary to con- sider the dynamics of a transgressing shore- line. We are indebted in this regard to Bruun (1962) and Fischer (1961) who ap- pear to have independently appreciated the role of the landward translation of the wave- and current-maintained coastal pro- file in generating transgressive stratigraphy. In the New York Bight, as along most low, unconsolidated coasts, the coastal profile consists of a steeply sloping nearshore sec- tor ( the shoref ace ) and a gently dipping in- 604 72 Geological processes MODERN ESTUARY MOUTH SHOAL, TIDAL CHANNELS PAIRED FLOOD CHANNEL RETREAT TRACK, ESTUARINE SHOAL-RETREAT MASSIF 40M SCARP TRANSGRESSED CUSPATE DELTA; (CAPE SHOAL- RETREAT MASSIF) 60M SCARP 605 New York-New Jersey shelf 73 39*I0'N 39"05' 74*00' 73*45' W Fig. 3. Simplified bathymetry and distribution of grain sizes on a portion of the central New Jersey shelf. Medium to fine sand occurs on ridge crests. Fine to very fine sand occurs on ridge flanks and in troughs. Locally, erosion in troughs has exposed a thin lag of coarse, shelly, pebbly sand over lagoonal clay. ( Reprinted from Stubblefield et al. 1974 by permission of the Journal of Sedimentary Petrology. ) ner shelf floor. The break in slope, which may be well defined or gently rounded, generally occurs at depths of 12 to 18 m, some few kilometers from the shoreline. Bruun (1962) pointed out that if this profile is in fact an equilibrium response of the seafloor to the coastal hydraulic cli- mate, then a rise in sea level must result in a landward and upward translation of the profile (Fig. 4A). Such a translation neces- sitates erosion of the shoreface. Much of the resulting debris will presumably be entrained in the littoral current and move downcoast, but during periods of onshore storm winds, the littoral drift may leak sea- ward, due to an offshore component of bot- tom flow, to be deposited beneath the rising seaward limb of the equilibrium profile on the adjacent inner shelf floor. Evidence for such seaward bottom trans- port is varied. Murray (1975) described periods of offshore bottom flow on the gulf coast, when winds are onshore and the wa- ter column is not stratified. Sonu and Van Beek (1971) noted that sand loss from North Carolina beaches correlates poorly with periods of high waves but correlates well with high waves generated by onshore northeast winds. On the Long Island inner shelf, we used sidescan sonar to map inner Fig. 2. Delaware Shelf Valley complex. Southward littoral drift along the New Jersey coast is injected into the reversing tidal stream of Delaware Bay mouth. The resulting sand shoal is stabilized as a system of interdigitating ebb- and flood-dominated channels. The shelf valley complex seaward of the bay mouth was formed by the retreat of the coastal sedimentary regime through Holocene time. Retreat of the main flood channel has excavated the Delaware Shelf Valley; retreat of the bay mouth shoal has left a levee- like high on the shelf valley's north flank. ( Reprinted from Swift 1973 by permission of the Geological Society of America Bulletin.) 606 Geological processes VECTOR RESOLUTION OF PROFILE TRANSLATION WASHOVER CYCLE OF BARRIER SANDS Fig. 4. Models tor a coast undergoing erosional shoreface retreat during a rise in sea level. A — Rise in sea level results in landward and upward translation of coastal profile ( Bruun 1962). B — Translation is accomplished. Wind and storm washover transport on the barrier surface and erosion of the shoreface and seaward transport of the resulting debris (Fischer 1961). shelf current lineations that form an east- ward-opening angle with the beach (Fig. 5). A poorly defined asymmetry is appar- ent: the western sides of the lineations are gradational, whereas the eastern sides are sharply defined. The origin of this pattern is not clear. The dark bands are strips of coarse or gravelly sand that may either be troughs between low amplitude, current- transverse sand waves or troughs between current-parallel sand ribbons. However, considering the angle that the lineations make with the beach, sand ribbons seem unlikely for reasons of flow continuity. If sand waves, the lineations are responses to strong bottom flows trending westerly and offshore. Fischer (1961), Stahl et al. (1974), and Sanders and Kumar (1975) described the stratigraphic consequences of erosional shoreface retreat, based on their observa- tions of the New Jersey and New York coasts. The barrier superstructure will re- treat over the lagoonal deposits by a cyclic process of storm washover, burial, and re- emergence at the shoreface ( Fig. 4B ) . Lower shoreface sands will tend to be trans- ported seaward to accumulate over the 607 New York-New Jersey shelf 75 40° 2 5' i 74°00' 73-55' 73°50' 73°45' 100 METERS Fig. 5. Sidescan sonar records of current lineations on the Long Island coast, collected at three dif- ferent periods. Positioning by Raydist. Current lineation pattern (bands A-F) expands to south during observation period. Apparent change in orientation in last panel is due to ship maneuvering. ( From Stubblefield et al. in prep. ) eroded surface of the lagoonal deposits as the leading edge of a marine sand sheet. Bruun's hypothesis is compatible with the stratigraphic evidence and with our limited knowledge of coastal hydraulics. A more rigorous test requires bathymetric time se- ries to document changes in the coastal pro- file. Limited data of this sort are becoming available. Harris (1954) undertook a study of the Long Branch, New Jersey, dredge spoil dumpsite to determine if dumping was nourishing the beach (Fig. 6). In fact, dur- ing a 4-year period, the shoreface under- went between 5 and 26 cm of erosion, while an irregular pattern of deposition prevailed on the inner shelf floor. A somewhat longer time series has been prepared by Kim and Gardner (Woodward-Clyde Assoc.) during study of proposed sewage outfall routes for the Ocean County, New Jersey, sewerage authority ( Fig. 7 ) . Two out of three profiles taken indicate 1.5-2.0 m of erosion over 20 years. The third profile is immediately south of a shoreface-connected sand ridge; here comparable aggradation has occurred as a consequence of southward ridge migration. Growth of ridges — Erosional shoreface retreat on the Atlantic cannot be adequately described by a two-dimensional model such as Fig. 4 because the shoreface appears to be the formative zone for sand ridge topog- raphy as well as for the sand sheet into which it is impressed. Clusters of shoreface- sand ridges occur on the New Jersey coast between Brigantine and Barnegat Inlets, on the north New Jersey coast between Mana- squan and Sea Bright, and on the Long Is- land coast from Long Beach to the shore- face of eastern Fire Island. The shoreface-connected ridges are named for their oblique, fingerlike exten- sions of the shoreface, causing seaward de- flections of isobaths as shoal as 5 m. The ridges tend to be asymmetric in cross-sec- tion, with steep seaward flanks, although this relationship may be reversed at the base of the ridge where it joins the shore- face. Seaward flanks tend to be notably 608 76 Geological processes Fig. 6. Erosion and deposition near Long Branch, New Jersey, dredge spoil dumpsite during a 4- yr period. Recorded changes are 0.4-1.4 ft. Shoreface lias undergone erosion; adjacent seafloor primarily has undergone aggradation. ( From Harris 1954. ) 6 4 2 en. 0 oc u -? h- uj -4 -6 -8 -10 200 400 600 800 METERS c , , J 1- NN^ 1973 1953 ^""~:u^zr-- i i . j i i i — i — i ^^r* — i 200 400 600 800 METERS 1000 200 400 600 800 1000 1200 75°00' 45' 30' METERS 15 74°00 Fig. 7. Profiles of proposed sewage outfall sites on the New Jersey coast. Sites A and B have eroded over a 20-yr period. Site C, immediately downcoast of a shoreface-connected sand ridge, has aggraded. ( Reprinted from Kim and Gardner 1974 with permission of Woodward-Clyde Assoc. ) 609 New York-New Jersey shelf 77 finer than landward flanks. Off Brigantine Inlet and off the New Jersey coast, shore- face-eonnected ridges are associated with free-standing inner shelf ridges that can be traced seaward for tens of kilometers in ap- parent genetic sequence. The ridges form on the shoreface in response to south-trend- ing coastal storm currents ( Duane et al. 1972 ) and become detached from the shore- face as it retreats. They tend to migrate downcoast (to the south or west) and off- shore, extending their crestlines so as to maintain contact with the shoreline (Fig. 8). Eventually, however, contact is broken, and they are stranded on the deepening shelf floor. Downcoast ridge migration is part of a general pattern of southwesterly sand transport on the Atlantic shelf. In the offshore ridge topography, this pattern is indicated by the tendency of both ridge crests and trough talwegs to rise toward the southwest. Locally, it is indicated by patterns of erosion and deposition near wrecks ( Fig. 9). Sand transport on the inner shelf The preceding description of the mor- phologic evolution of the New York shelf surface is based primarily on the interpre- tation of bathymetric maps, aided by local substrate inventories in which the bottom is 39°29'N 39°28N 74°I6'W 74°I5'W 74°I4' Fig. 8. Patterns of erosion and deposition on Beach Haven Ridge, New Jersey, between 1935 and 1954, superimposed on 1954 bathymetry. Pat- tern of north flank erosion and south flank deposi- tion indicates downcoast migration of ridge. (From DeAlteris et al. in press. ) examined by grab sampling, photography, Vibracoring, and seismic profiling. The con- clusions are qualitative but nonetheless valid. However, fuller understanding of the behavior of the shelf surface requires a different approach. We must directly measure fluid and sedi- ment transports involved in the two basic mechanisms that have shaped the shelf sur- face: tidal flow and sand storage at estuary mouths, and erosional retreat of the shore- face between estuary mouths. Environ- mental managers who must make decisions about dredged channels, sewage outfalls, sewage and dredge spoil dumpsites, deep- water tanker terminals, and offshore power plants need to understand these processes before they can evaluate the stability of the inner shelf surface. The nature of coastal sand transport dur- ing storms is the first major problem we will consider. Fluid motions in the surf zone have been studied for decades, and the role of longshore currents driven by shoal- ing and breaking waves has been described (e.g. Bowen 1969). In the New York area, ® ■1 WRECK r~l ACCRETION E3 SCOUR ® MN'0 CZ! 1-0-1.5 I |lJS-gJ0 V7A 20-2-5 iH<2.5 -GRAB SAMPLE "BOX CORE Fig. 11. Distribution of grain sizes over the Tobay Beach ridges, LINS area. Size classes in phi units. taken from Fig. 12, obscures the brief time- scale flow associated with the storm. The 3-h low-pass record, which is only slightly smoothed and still contains the tidal signal, shows a period of offshore flow more clearly. These results must be viewed cau- tiously. The Aandaraa current meters which were used have large direction and speed errors when used in shallow water with sur- face wave amplitudes as large as were pres- ent during the event described here. During the November-December exper- iment on the Long Island inner shelf, esti- mates of sand transport were made from calculations from current meter records (Lavelle et al. in press) and also from radio- isotope tracer dispersal patterns ( Lavelle et al. unpublished data). To generate the pat- terns, about 500 cm3 of indigenous fine to very fine sand was surface-coated with 10 Ci of ,,,:,Ru (half-life, 39.6 d). On 12 No- vember, equal portions of tagged sand were released in water soluble bags at three points at the east end of the main trough (Fig. 14). The injection points formed an equilateral triangle with sides roughly 100 m long. The developing dispersal pattern of labeled sand was surveyed at intervals by scintillation detectors mounted in a cylinder towed across the bottom. Navigation was by a Raydist system with 10-m resolution. Four postinjection surveys were made dur- ing the 11-week tracer experiment. Disper- sal patterns mapped 2 and 8 weeks after in- jection are shown in Fig. 14. After 2 weeks (25 November) roughly ellipsoidal smears trended east from each of the three injection points (Fig. 14A). Each smear could be traced for about 200 m before the signal was lost in the background radiation. After 8 weeks (10 January) the three eastward smears had been replaced by a single, more extensive pattern extending 700 m to the west (Fig. 14B). Partially processed data from an intermediate survey ( 17-19 Decem- ber) indicate that the reversal in fact had 612 80 Geological processes 12 NOV 74 13 14 15 16 17 18 N0V74 19 20 21 22 23 24 25 NOV 74 26 27 28 29 30 ® i5*otf- *~E W — DIR SPD B 64° 47 M 68° 69 S 81° 120 (b) long term bottom, middepth.and near-surface current means — eastward flow 5cm/s DIR SPD B 25 3° 36 M 241° 69 S 239° 96 LONG TERM BOTTOM, MIDDEPTH, AND NEAR-SURFACE CURRENT MEANS- WESTWARD FLOW ENT DIRECTION AND SPEED Fig. 12. Summary of flow data for the LINS experiment. A — Vector time series of representative near-bottom flow. Data have been subjected to a 40-h low-pass filter. B, C — Long term velocity averages of eastward and westward flow for meters grouped by depth in water column. Bottom, middepth, and near-surface groupings are labeled B, M, and S. (From Lavelle et al. in press.) 613 Netc York-New Jersey shelf 81 A (BOTTOM FLOW) ipnisiiininiiungii'iiii^ 0 20 cm/s B WIND) COASTLINE V />. cm/s X 29 30 DEC 1 Fig. 13. Vector time series for bottom current and wind velocities during the 1-4 December storm. A — 40-h low-pass filtered record ( Lanczos filter with response. —6 db at 36 h and —20 db at 40 h ) . B — Wind record from Ambrose Tower. C — 3-h low-pass filtered record (Lanczos filter with response. —6 db at 2.5 h and —20 db at 3 h). occurred before this and that it initially had been at least 1,200 m long. The temporal pattern or sediment trans- port over 60 days may be inferred from Fig. 14C. Current speed, measured 1.5 m from the bed, is plotted against time. The horizontal line at 18 cm/s is an estimated threshold for the fine to very fine sand (mean diameter, 3.0 (f>) found at the site. It is based on the work of Shields and sub- sequent workers (Graf 1971: p. 90) and on a choice of 3.0 X 10 ~3 for the drag coefficient ( Sternberg 1972 ) . This choice of threshold velocity was supported by empirical evi- dence obtained during the course of the experiment (Lavelle et al. in press). Esti- mates have been prepared for the relative role each transport event played in the overall transport record, based on the con- cept of factional energy expenditure pro- portional to the transport volume (Bagnold 1963). For each event where velocities ex- ceeding threshold were recorded, a trans- port volume was calculated: Qi = a \ (|«| - lutnD-'dt, [ where |u| is measured current speed, \utj,\ is threshold speed, a is a constant of pro- portionality, and f; is the duration of the transport event (Lavelle et al, in press). Expression of sand transport as a power of the difference of measured and threshold velocity is supported by Kennedy's ( 1969 ) analysis of stream transport data. Without assigning a value to a, we can calculate the rate of transport of one flow event relative to the next or in relation to the sand dis- charge that occurred over the entire dura- tion of the current meter record. The sec- ond of these options has been used in Fig. 14C, where relative sand transport as per- cent of total transport has been represented as solid bars superimposed on the current meter record. Bar height is a measure of volume percent of transport; bar width is a measure of duration of the transport event. Despite the exceedence of the sediment transport threshold at many points in the record, only the solid bars centered on 2 and 16-17 December are visible in the fig- ure. Thus sand transport during observation 614 82 Geological processes consisted of periods of quiescence separated by brief, intense transport events. Further- more, since discharge is calculated as a power function of excess velocity, intense storms are far more efficient transporters of sand than mild ones. Although the trans- port index calculated for the 1^4 December storm may be biased by the choice of threshold speed as well as by the functional dependence on velocity, it seems probable that any reasonable parameterization would lead to the same general conclusion: the storm event of 1-4 December moved more sand at 20-m water depth than the com- bination of all other transport events. Attempts have also been made to calcu- late sediment transport indices over longer periods of time in the New York Bight apex. The following computation is based on 30- 80-day Aandaraa current meter records ( Fig. 15 ) . Data in each current meter rec- ord consist of an average speed, u, and an instantaneous direction, 6, taken for each 10-min sampling interval. For each inter- val in which an assigned threshold speed, |u,J, is exceeded, a sediment transport in- dex, Q, has been computed, as follows: Q = u — \utn\y, (|w| - \uth\) >o. For each current meter, the set of vectors of flow direction, 6 (0°^ 0^359°), and of sediment transport index, Q, is sorted into 10-degree classes. The results are plotted as 400 V) K UJ W200 25 NOV 74 mm^M 400- B u200 2 1000 800 600 400 200 200 400 METERS 10 JAN 75 — i 1 1 1 1 r 1000 800 600 400 200 ,r *■■:;. :^> <-■■■■ ■ " 1 II II ■ II |l 1 1 1 0 200 400 METERS S ioo £ 80- 2 60- g 40 <