C- 75/ v. ^TOFco, c <"o o v^ ^7-ES O* * U.S. DEPARTMENT OF COMMERCE National Oceanic and Atmospheric Administration COLLECTED REPRINTS-1975 Volume II ATLANTIC OCEANOGRAPHIC AND METEOROLOGICAL LABORATORIES Digitized by the Internet Archive in 2012 with funding from LYRASIS Members and Sloan Foundation http://archive.org/details/collectedreprint1975atla ATMOSAl U.S. DEPARTMENT OF COMMERCE Juanita M. Kreps, Secretary NATIONAL OCEANIC AND ATMOSPHERIC ADMINISTRATION Robert M. White, Administrator ENVIRONMENTAL RESEARCH LABORATORIES Wilmot N. Hess, Director Collected Reprints 1975 Volume II ATLANTIC OCEANOGRAPHIC AND METEOROLOGICAL LABORATORIES ISSUED FEBRUARY 1977 Boulder, Colorado Atlantic Oceanographic and Meteorological Laboratories Miami, Florida 33149 For sale by the Superintendent of Documents, U. S. Government Printing Office, Washington, D. C. 20402 FOREWORD This is the tenth annual publication of the collected reprints of NOAA's Atlantic Oceanographic and Meteorolog- ical Laboratories. It brings together our research results published during 1975 in a wide range of scientific and technical journals as well as in some internal NOAA publications. Although the generation of new knowledge is of itself satisfying to the researcher who does it, the usefulness to others is largely a function of how well the knowledge is disseminated. For this reason, our collected reprints receive a wide distribution to the libraries of universities, research institutions, and government agencies in this country and abroad. The Atlantic Oceanographic and Meteorological Laboratories conduct research on the physical, chemical, and geological characteristics and processes of the ocean waters, the sea- floor, and the overlying atmosphere. During 1975, these research efforts were carried out by four major groups: Physical Oceanography Laboratory Marine Geology and Geophysics Laboratory Sea-Air Interaction Laboratory Ocean Remote Sensing Laboratory The reprints in this volume are arranged alphabetically by the last name of the first author within each of these groups Harris B. Stewart, Jr. Director, AOML Atlantic Oceanographic and Meteorological Laboratories 15 Rickenbacker Causeway Virginia Key Miami, Florida 33149, U.S. A in CONTENTS VOLUME I Page No, GENERAL 1. Stewart, H. B. , Jr Bologna Workshop on Marine Science-Concluding Observations. Report of the Marine Science Workshop held by the Johns Hopkins University, Bologna, Italy, October 15-19, 1973, Annex D, 73-78. 2. Stewart, H. B. , Jr. Book Review: Handbook of Marine Science, Volume I and II. Sea Frontiers 21, No. 1, 58. 3. Stewart, H. B. Jr. The National Oceanic and Atmospheric Administration and the Outer Continental Shelf. Proc. of the Estuarine Research Federation, OCS Conference and Workshop, Marine Environmental Implications of Offshore Oil and Gas Development in the Baltimore Canyon Region of the Mid-Atlantic Coast, VIMS, Wachapreague, Virginia, 27-38. PHYSICAL OCEANOGRAPHY LABORATORY 4. Beardsley, R. C, W. C. Boicourt, and D. V. Hansen. Physical Oceanography of the Middle Atlantic Bight. Proc. of Special Symposium, "The Middle Atlantic Continental Shelf and the New York Bight," American Museum of Natural History, New York City, 3-4-5 November 1975, 1 p. 20 5. Brown, W., W. Munk, F. Snodgrass, H. Mofjeld, and B. Zetler. MODE Bottom Experiment. Journal of Physical Oceanography 5, No. 1, 75-85. 21 6. Charnell, R. L. (Editor), D. A. Mayer, D. V. Hansen, D. Swift, A. Cok, D. Drake, G. Freeland, W. Lavelle, T. McKinney, T. Nelson, R. Permenter, W. Stubblefield, D. Segar, P. Hatcher, G. Berberian, L. Kiester, M. Weiselberg. Assessment of Offshore Dumping in the New York Bight, Technical Background: Physical Oceanography, Geological Oceanography, Chemical Oceanography. NOAA Technical Report ERL 332-MESA 3, 83 p. 32 7. Charnell, R. L., and D. A. Mayer. Water Movement Within the Apex of the New York Bight During Summer and Fall of 1973. NOAA Technical Memorandum ERL MESA-3, 32 p. 122 8. Chew, F. The Interaction Between Curvature and Lateral Shear Vorticities in a Mean and an Instantaneous Florida Current, A Comparison. Tellus XXVII, No. 6, 606-618. 154 9. Gordon, H. R. and W. R. McCluney. Estimation of the Depth of Sunlight Penetration in the Sea for Remote Sensing. Applied Optics 14, 413-416. 167 10. Hansen, D. V. Review of: Progress in Oceanography, Volume 6. Edited by B. A. Warren, Pergamon Press, New York. Bulletin of the American Meteorological Society 56, No. 9, 997-998. 171 11. Hazelworth, J. B., B. L. Kolitz, R. B. Starr, R. L. Charnell, G. A. Berberian, and M. A. Weiselberg. New York Bight Project, Water Column Sampling Cruises #6-8 of the NOAA Ship FERREL, April-June 1974. NOAA Data Report MESA-1, 177 pages. * 172 12. Hazelworth, J. B., and R. B. Starr. Oceanographic Conditions in the Caribbean Sea During the Summer of 1971. NOAA Technical Report ERL 344-AOML 20, 144 p. 200 13. Hazelworth, J. B., B. L. Kolitz, R. B. Starr, R. L. Charnell, G. A. Berberian, and M. A. Weiselberg. New York Bight Project, Water Column Sampling Cruises 9-12, of the NOAA Ship FERREL, July-November 1974. NOAA Data Report ERL MESA-3, 234 p.** 347 14. Kirwan, A. D., Jr., G. McNally, M. S. Chang, and R. Molinari. The Effect of Wind and Surface Currents on Drifters. Journal of Physical Oceanography 5, No. 2, 361-368. 348 "Text only; pp. 26-177 are station data. "-Abstract only; complete text on microfiche. VI 15. Maul, G. A. Circulation of the Eastern Gulf of Mexico and Its Possible Relation to Red Tides. Proc. of the Florida Red Tide Confer- ence, Florida Marine Research Publication, Florida Department of Natural Resources, Marine Research Laboratory, Sarasota, Florida, October 10-12, No. 8, 9-10. 356 16. Maul, G. A. An Evaluation of the Use of the Earth Resources Technology Satellite for Observing Ocean Current Boundaries in the Gulf Stream System. NOAA Technical Report ERL 335-A0ML 18, 125 p. 357 17. Maul , G. A. and S. R. Baig. A New Technique for Observing Mid-Latitude Ocean Currents from Space. Proc. of the American Society of Photogrammetry, 41st Annual Meeting, Washington, D. C, March 9-14 1975, 713-716. 485 18. Maul, G. A. and H. R. Gordon. On the Use of the Earth Resources Technology Satellite (LANDSAT-1) in Optical Oceanography. Remote Sensing of Environment 4, No. 2, 95-128. 489 19. Mayer, D. A. Examination of Water Movement in Massachusetts Bay. NOAA Technical Report ERL 328-AOML 17, 46 p. 523 20. Mayer, D. A. and D. V. Hansen Observations of Currents and Temperatures in the Southeast Florida Coastal Zone During 1971-1972. NOAA Technical Report ERL 346-AOML 21, 36 p. 572 21. Mofjeld, H. 0. and M. Rattray, Jr. Barotropic Rossby Waves in a Zonal Current: Effects of Lateral Viscosity. Journal of Physical Oceanography 5, No. 3, 421-429. 611 22. Mofjeld, H. 0. Empirical Model for Tides in the Western North Atlantic Ocean. NOAA Technical Report ERL 340-AOML 19, 24 p. 620 23. Molinari, R. L. A Comparison of Observed and Numerically Simulated Circulation in the Cayman Sea. Journal of Physical Oceanography 5, No. 1, 51-62. 647 vii 24. Molinari, R. L. and A. D. Kirwan, Jr. Calculations of Differential Kinematic Properties from Lagrangian Observations in the Western Caribbean Sea. Jouimal of Physical Oceanography 5, No. 3, 483-491. 659 25. Segar, D. A. and A. Y. Cantillo. Direct Determination of Trace Metals in Seawater by Flameless Atomic Absorption Spectrophotometry. Advances in Chemistry Series, Number 147 3 Analytical Methods in Oceanography, 56-81. 668 26. Segar, D. A. and G. A. Berberian. Trace Metal Contamination by Oceanographic Samplers. A Com- parison of Various Niskin Samplers and a Pumping System. Advances in Chemistry Series, Number 147 3 Analytical Methods in Oceanography 3 9-15. 694 27. Zetler, B., W. Munks H. Mofjeld, W. Brown, and F. Dormer. MODE Tides. Journal Physical Oceanography 5, No. 3, 430-441. 701 VOLUME II MARINE GEOLOGY AND GEOPHYSICS LABORATORY 28. Ballard, R. D., W. B. Bryan, J. R. Heirtzler, G. Keller, J. G. Moore, and T. van Andel . Manned Submersible Observations in the FAMOUS Area: Mid- Atlantic Ridge, Science 190, 103-108. 713 29. Dietz, R. S., J. F. McHone, and N. M. Short. Oman Ring: Suspected Astrobleme. Meteoritics 10s No. 4, 393. 719 30. Freeland, G. L. and D. J. P. Swift. New York Alternative Dumpsite Assessment-Reconnaissance Study of Surficial Sediments. IX International Congress of Sedimentology; Nice, France, Theme 10, 13-18. Also appeared in Offshore Technology Conference, Paper #0TC 2385, 505-511. 720 31. Freeland, G. L., D. J. P. Swift, W. L. Stubblefield, and A. E. Cok. Surficial Sediments of N0AA-MESA Study Areas in the New York Bight. Mid-Atlantic Shelf/New York Bight ASL0 Symposium, 24-25. 727 viii 32. Gadd, P. E., J. W. Lavelle, and D. J. P. Swift. Calculations of Sand Transport on the New York Shelf Using Near-Bottom Current Meter Observations. American Geophysical Union3 Fall Annual Meeting 56, No. 12, 1003. 729 33. Hatcher, P. G. and L. E. Keister. Carbohydrates and Organic Carbon in the New York Bight Sedi- ments as Possible Indicators of Sewage Contamination. Mid- Atlantic Shelf/New York Bight ASLO Symposium, 31-33. 730 34. Hatcher, P. G., B. R. Simoneit, and S. M. Gerchakov. Mangrove Lake, Bermuda; Its Sapropelic Sedimentary Environment. Seventh International Meeting on Organic Geochemistry, Madrid, Spain, 61. 733 35. Hulbert, M. H. and R. H. Bennett. Electrostatic Cleaning Technique for Fabric SEM Samples. Clays and Clay Minerals 23, 331-335. 734 36. Kearey, P., G. Peter, and G. K. Westbrook. Geophysical Maps of the Eastern Caribbean. Journal of the Geological Society 131 ■> 311-321. 739 37. Keller, G. H. Sedimentary Processes in Submarine Canyons off Northeastern United States. IX International Congress of Sedimentology, Nice, France, Theme 6, 77-80. 750 38. Keller, G. H. Book Review: The Ocean Basins and Margins - Volume 2: The North Atlantic. Edited by A.E.M. Nairn and F.G. Stehli, Plenum Publ., New York. Marine Geotechnjology 1, No. 1, 159-161. 757 39. Keller, G. H. and L. L. Minter. Cariaco Trench - Sediment Geotechnical Properties. Journal of Sedimentary Petrology 45, No. 1, 292-294. 760 40. Keller, G. H., S. H. Anderson, and J. W. Lavelle. Near-Bottom Currents in the Mid-Atlantic Ridge Rift Valley. Canadian Journal of Earth Science 12, No. 4, 703-710. 763 IX 41. Lavelle, J. W. , G. H. Keller, and T. L. Clarke. Possible Bottom Current Response to Surface Winds in the Hudson Shelf Channel. Journal of Geophysical Research 80, No. 15, 1953-1956. 771 42. Lavelle, J. W., D. J. P. Swift, H. R. Brashear, and F. N. Case. Tracer Observations of Sand Transport on the Long Island Inner Shelf. American Geophysical Union, Fall Annual Meeting 56, No. 12, 1003. 775 43. McGregor, B. A. and P. A. Rona. Crest of the Mid-Atlantic Ridge at 26° N. Journal of Geophysical Research 80, No. 23, 3307-3314. 776 44. McGregor, B. A., G. H. Keller, and R. H. Bennett. Seismic Profiles Along the U.S. Northeast Coast Continental Margin. EOS 56, No. 6, 382. 784 45. McHone, J. F. , Jr. and R. S. Dietz. Impact Structures on LANDSAT Imagery (Abstract). Geological Society of America Annual Meetings, Abstracts with Programs 7, No. 7, 1196. 785 46. Maupome, L., R. Alvarez, S. W. Kieffer, and R. S. Dietz. On the Terrestrial Origin of the Tepexitl Crater, Mexico. Meteoritics 10, No. 3, 209-214. 786 47. Maupome, L., R. Alvarez, S. W. Kieffer, and R. S. Dietz. Tepexitl Crater, Mexico: Not Meteoritic (Abstract). Meteoritics 10, No. 4. 454-455 792 48. Minter, L. L., G. H. Keller, and T. E. Pyle. Morphology and Sedimentary Processes in and Around Tortugas and Agassiz Sea Valleys, Southern Straits of Florida. Marine Geology 18, 47-69. 793 49. Permenter, R. W. , W. L. Stubblefield, and D. J. P. Swift. Substrate Mapping by Sidescan Sonar. Florida Scientist 38, Supplement 1, 13-14. 816 50. Richards, A. F., K. Oien, G. H. Keller, and J. V. Lai. Differential Piezometer Probe for an In Situ Measurement of Sea-Floor Pore-Pressure. Geotechnique 25, No. 2, 229-238. 817 51. Rona, P. A. Book Review: Sonographs of the Sea Floor. Journal of Geology 83, No. 4, 536. 827 52. Rona, P. A. Letters to the Editors: Minerals and Plate Tectonics. Science 190, No. 4213, 422. 828 53. Rona, P. A. Relation of Offshore and Onshore Mineral Resources to Plate Tectonics. Proc. Offshore Technology Conference, Paper No. OTC 2317, 713-715. 829 54. Rona, P. A. Salt Deposits of the Eastern and Western Atlantic (Abstract). Proc. International Symposium on Continental Margins of Atlantic Type, Brazil, 1-15. 833 55. Rona, P. A. Viewpoint. Encounter with the Earth, edited by L. F. Laporte, Canfield Press, San Francisco, CA, 83-86. 834 56. Rona, P. A., B. A. McGregor, P. R. Betzer, G. W. Bolger, and D. C. Krause. Anomalous Water Temperatures Over Mid-Atlantic Ridge Crest at 26° North Latitude. Veep-Sea Research 22, 611-618. 838 57. Scott, R. B., J. Malpas, G. Udintsev, and P. A. Rona. Submarine Hydrothermal Activity and Seafloor Spreading at 26° N, MAR. Geological Society of America, Abstracts with Programs 7, No. 7, 1263. 846 58. Stubblefield, W. L. and R. W. Permenter. Temporal and Spatial Substrate Variation in the New York Bight Apex. Geological Society of America, Abstracts with Programs 7, No. 7, 1285-1286. 847 XI 59. Stubblefield, W. L., J. W. Lavelle, D. J. P. Swift, and T. F. McKinney. Sediment Response to the Present Hydraulic Regime on the Central New Jersey Shelf. Journal of Sedimentary Petrology 45, No. 1, 337-358. 848 60. Swift, D. J. P. Barrier-Island Genesis: Evidence From the Central Atlantic Shelf, Eastern U.S.A. Sedimentary Geology 143 1-43 870 61. Swift, D. J. P. Response of the Shelf Floor to Geostrophic Storm Currents, Middle Atlantic Bight of North America. Sedimentary Mechanics, IX International Congress of Sedimentology, Nice, France, Theme VI, 193-198. 913 62. Swift, D. J. P. Tidal Sand Ridges and Shoal-Retreat Massifs. Marine Geology 18, 105-134. 921 63. Swift, D. J. P., G. L. Freeland, P. E. Gadd, J. W. Lavelle, and W. L. Stubblefield. Morphologic Evolution and Sand Transport in the New York Bight. Mid-Atlantic Shelf/New York Bight ASLO Symposium, 65-66. 950 SEA-AIR INTERACTION LABORATORY 64. McLeish, W. L. and G. E. Putland. The Initial Water Circulation and Waves Induced by an Airflow. NOAA Technical Report ERL 316-AOML 16, 45 p. 952 65. McLeish, W. and G. E. Putland. Measurements of Wind-Driven Flow Profiles in the Top Millimeter of Water. Journal of Physical Oceanography 53 No. 3, 516-518. 999 66. Ostapoff, F., J. Proni and R. Sellers. Preliminary Analysis of Ocean Internal Wave Observations by Acoustic Soundings. Preliminary Scientific Results of the GARP Atlantic Tropical Experiment, prepared by the International Scientific and Management Group (ISMG) of the World Meteoro- logical Organization, Volume II, GATE Report No. 14, 392-397. 1002 xii 67. Ross, D. A Comparison of SKYLAB S-193 and Aircraft Views of Surface Roughness and a Look Toward SEASAT. Proceedings of the NASA Earth Resources Survey Symposium, Houston, Texas, June 1975, Vol. 1-C, 1911-1936. 1008 68. Stegen, G. R., K. Bryan, J. L. Held, and F. Ostapoff. Dropped Horizontal Coherence Based on Temperature Profiles in the Upper Thermocline. Journal of Geophysical Research 80 , No. 27, 3841-3847. 1034 OCEAN REMOTE SENSING LABORATORY 69. Apel, J. R. Seasat: A Spacecraft Views the Marine Environment with Microwave Sensors. Remote Sensing: Energy -Related Problems , edited by Dr. T. Veziroglu, J. Wiley & Sons, 47-60. 1041 70. 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. Journal of Geophysical Research 803 No. 6, 865-881. 1055 71. Apel, J. R., J. R. Proni, H. M. Byrne, and R. L. Sellers. Near-Simultaneous Observations of Intermittent Internal Waves on the Continental Shelf from Ship and Spacecraft. Geophysical Research Letters 2, No. 4, 128-131. 1072 72. Proni, J. R. , and J. R. Apel. On the Use of High-frequency Acoustics for the Study of Internal Waves and Microstructure. Journal of Geophysical Research 80 3 No. 9, 1147-1151. 1076 73. Proni, J. R. , D. C. Rona, C. A. Lauter, Jr., and R. L. Sellers. Acoustic Observations of Suspended Particulate Matter in the Ocean. Nature 254, No. 5499, 413-415. 1081 xi n Reprinted from: Science 190, 103-108, 28 Manned Submersible Observations in the FAMOUS Area: Mid-Atlantic Ridge grams of rock samples, 50 water samples, and 42 sediment samples, all precisely lo- cated relative to a 5-fathom contour map of the inner rift valley floor. Some 15.000 color and black-and-white photographs were taken, and about 140 scientific man- hours of visual observation and notes documented the 27 km of sea floor tra- versed. Volcanic and tectonic processes associated with an active spreading center have been directly observed. R D. Ballard, W. B. Bryan, J. R. Heirtzler, G. Keller, J. G. Moore, Tj. van Andel Project KAMOUS (French- American Mid-Ocean Undersea Study) was con- ceived three years ago, its objectives being lo define the tectonic and volcanic process- es associated with genesis of new oceanic crust. A small area on the Mid-Atlantic Ridge centered at about 36°50'N was se- lected lor detailed study on the basis of sci- entific and logistic criteria (Fig 1) More than 25 cruises were made lo the area by surface ships from the United Stales. France, Canada, and Fngland, culminating in the first manned submersible studies of a mid-ocean ridge by the French submersible irchimede in 1973, and by Archim'ede. Cvana, and the American submersible Al- vin during the summer of 1974. The re- gional setting of the dive site was estab- lished by narrow-beam echo sounding, dredging, side-scan sonar and deep-tow surveys, photography, aeromagnetic sur- veys, and magnetic and gravitv surveys from shipboard The Glomar Challenger, on leg 37 of the Deep Sea Drilling Project (DSDP). drilled four holes starting about IX nautical miles west of the dive site. The final phase of the surface ship surveys, car- ried out by R V Knorr concurrently with the submersible program, consisted of dredging, coring, detailed photographic and thermal surveys, and deplovment of ocean bottom seismographs. In this article we present a preliminary summary and in- terpretation of some of the unique observa- tions made b\ the manned submersible Al- vin during the summer of 1974. Data from the surface ship surveys, most of which are yet unpublished, have contributed signifi- cantly to the success of the project and to some of the interpretations presented here The accompanying article describes the re- lated French program. The submersible Alvin carries two scien- tists and a pilot: normal bottom time was 4 to 5 hours, during which time the sub- mersible traversed from 1 lo nearly 4 ki- lometers across the bottom. The sub- mersible was continuously navigated in real lime relative to three acoustic bea- cons. Precision varied with bottom condi- tions but was generally belter than ± 5 me- ters. Headings were obtained by gyro- compass, with a backup magnetic system. Precise measurements of depth, height off bottom, and heading were automatically logged as a function of time: these com- bined with the navigation data were plotted as X. Y . Z coordinates of the track at 20- or 40-second intervals. Rock, water, and sediment samples were collected rou- tinely and stored in a compartmented "lazy susan" basket. Nearly continuous vi- sual records by the scientists were aug- mented by photographs recorded by semi- automatic external still cameras, internal still and movie cameras, and a television camera with video tape recording. The Al- vin completed 15 dives in the median valley and two dives in Fracture Zone B (Fig. I). Overall. Alvin collected about 400 kilo- Geologic Setting The area selected for the prime dive site (Fig. I) consists of a median valley seg- ment about 20 km long, bounded on the north and south by transform faults, which are called Fracture Zone A and Fracture Zone B, respectively The median valley floor is about I km wide at its center, with depths averaging about 1400 fathoms: at either end it widens to more than 8 km with depths reaching 1700 fathoms. The west wall rises abruptly within a distance of I to 2 km lo a broad, undulating terrace with an average depth of about 1000 fath- oms To the east, the slopes rise more grad- ually in a series of steps to the base of the main east wall 2 to 3 km from the center of the valley. This bathymetne asymmetry is matched by a corresponding asymmetry in magnetic anomalies, with an inferred spreading rate of about OX cm/year to the west and about 1 .3 cm/year to the east ( / ). Within 10 to 12 km from the center of the valley, rift mountain peaks rise to depths typically about 800 fathoms. Each of the transform faults is marked by a narrow trough and by a linear /one of micro- earthquakes (2). The center of the median valley floor is marked by a series of rugged linear hills which have been considered to be recent volcanic extrusions (3) Much of the sub- marine operation was centered around two of these hills, the French working on the flanks and to the north of Ml. Venus, and the American group working south of Ml. Venus and around Mt. Pluto (Fig. I ). R I) Ballard. W B Brvan and J R Heirt/lcr arc ai ihc Department ofGeolog) and Geophysics, Woods Hole Oceanographit Institution Woods ftole, Massa- chusetts 02543, G Keller is ai the Atlantic Occan- ographic and Meteorological Laboratories of the Na- tional Aeronautics and Space Xdministration. Miami. Honda 1.tl4s) J G. Moore is ai ihc US Geological Survev. Menlo Park. California 1600li Fig. 1 (A) Location of dive area and tracks of Alvin in the median valley. (B) Regional reference map; small rectangle in area covered by (A) 714 cordant at about 1300 fathoms, suggesting that they have grown in response to a com- mon hydraulic head of lava in a shallow magma chamber. Large tumuli, 4 m or more in diameter, were occasionally observed on flow sur- faces. Some resemble small-scale driblet cones, while others are better described as giant pillows. As on terrestrial lava flows, they appear to represent short-lived sec- ondary centers of tumescence and ex- trusion. Growth of the central volcanic hills and advancement of the lava flows appears to take place almost entirely by tumescence, fracturing of crust, budding, tumescence of the new extrusions, fracturing, budding, and so on. In fact, the large volcanic hills probably could be regarded as very large compound lumuli which have repeatedly fractured and budded lo give rise to the major lava flows The flows themselves fracture and bud down to tne scale of the individual lava fingers and toes, at which point the available lava presumably be- comes exhausted Detached, subsphencal pillows were rarely seen; the few examples appeared to have pulled loose from the ends of lava tubes on steep slopes, but this is not common Most of the recent extrusive activity ap- pears to be restricted to vents in the medi- an valley associated with Mt. Pluto or Mt. Venus. However, limited flank volcanism was observed on the lower walls of the val- ley immediately to the east and west of the valley floor. This /one of flank volcanism seems wider and better developed to the east than to the west, in harmony with the taster spreading rate to the east. However, in all traverses from the center of the valley outward to the flanks we were impressed by the rapid increase in sediment cover and bottom life and by the intense tectonic deg- radation to which the extrusive lava forms were subjected. Generally, within 300 m of the valley center to the west and within s00 m to the east, most of the delicate extrusive lorms had been destroyed, the flows were sliced and offset by numerous faults, and the surfaces were reduced to broken, jum- bled lava blocks and extensive talus fans at the base ol fault scarps. Preliminary petrographic study and analysis ol rock glasses by electron micro- probe have shown considerable variation in composition of the lavas, which can be related to their apparent age and position within the median valley. Table I shows examples of the more extreme composi- tions. The analysis shown in column A is typical of recent lavas from Mt. Pluto on the central volcanic axis and closely resem- bles that of the unusual low-TiO„ high- MgO glass reported from DSDP sites 3-14 and 3-18 (4). The analysis in column B is typical of lavas on the east and west flanks of the valley, as well as of older flows near the central axis that have not been cov- ered by the younger extrusions. This com- position is more typical of basalts reported from other parts of the ocean ridge system. The regional variation in percentage by weight of SiO, is shown in Fig. 4; a similar variation can be demonstrated for most of the other oxides of major elements, as im- plied by the data in Table I. Especially no- table are the increases in FeO/MgO, TiO,, and K.O from the center to the margins of the median valley. The slight asymmetry in the composi- tional distribution is consistent with the bathymetric asymmetry and differences in spreading rates and suggests that all of these features reflect operation of common processes at depth. The chemical varia- tions, although subtle, are matched by dis- tinct differences in liquidus mineral phases (microphenocrysts). As SiO: increases, oli- vine becomes less abundant, chrome spinel disappears, plagioclase becomes more abundant, and pyroxene appears. The chemical and petrographic variations are similar to those observed previously only over distances of hundreds of kilometers along the length of the Mid-Atlanlic Ridge (5). It seems unlikely that these variations, which we observe over a distance of a few kilometers, can be related to separate mantle sources or to deep mantle process- es, although transition element geoehemis- Fig 2 Flongaled bulbous and hun -shaped lava pillows. Note multiple fracturing ol ltusi due lo ex- pansion combined wiih surface quenching. Fig. 3. Layered interior of lava tube exposed at the top of ihe wall of a lensional fracture. 715 try implies separate magma types in the R.V. Charcot dredge samples from the FAMOUS area (6). Preliminary dating of lavas based on measurements of manganese and palago- nite crusts, along with our impressions of outcrop freshness based on direct observa- tion of sediment cover and faunal abun- dances, indicates that at least some of the apparently differentiated flank lavas are of about the same age as lava from the cen- tral valley, although the flank eruptions are much less voluminous This differentiation between central magma and contempo- raneous flank magma is well known at Kilauea volcano (7) and implies the exis- tence of a shallow, compositionally zoned magma chamber beneath the median val- ley (8). Calculations indicate that simple fractional crystallization of olivine and plagioclase, as suggested by the petro- graphic data, can account for most of the major element variation, although serious discrepancies remain for TiO: and alkalies. Frey el al. (4) showed that TiO, and FeO/ MgO are apparently the most critical ma- jor element parameters in sea floor lavas, and any process proposed to account for compositional differences in lavas across the median valley will have to account for the simultaneous variation in these param- eters. Tectonics In contrast to recent volcanic activity, which appears to be concentrated in a nar- row central zone, recent tectonic move- ment is evident throughout the entire width of the inner rift valley floor. Faults and fis- sures are numerous, striking 020° parallel to the rift avis (Fig. 5). Locally, faults trend perpendicular to the axis, but these are few in number and are truncated by the more dominant 020° structures. The dominance of tectonic activity is re- flected in the fine-scale geomorphology of the central valley From the west wall across the central depression between Mt. Venus and Ml. Pluto to the first major step fault of the east wall is a predominantly block-faulted terrain Normal faulting is common, and when combined with ten- sional extension it yields open fissures with difTerenlial movement between the two walls. Dips range from 45° to nearly verti- cal, with 70° to 80° common on smaller faults. Displacements vary from less than 1 to more than 100 m. 10 to 20 m being the most common. The faulted blocks are back-lilted 5° to 7° and in some instances have rotated away from one another, pro- ducing a rubble-filled V-shaped notch. Observations in the region of Mt. Pluto agree with those of the French on Mt. Fig. 4. Regional variation in SiO: content of rock samples recovered by Alvin. Dots are sta- tion locations for which data are available These stations are located along the dive tracks shown in Fig I Numbers are percentages of SiO;by weighl Venus to the north, that the morphology of the central topographic highs results from constructional volcanic processes, but the remainder of the inner floor to the east and west is in part shaped by tectonic forces. Even in the central region of Mt. Pluto, small postvolcanic tensional fissures occur. South of Mt. Pluto, one scissors fault was traced with displacement down to the west at its northern end and down to the east at its southern end. The nature of fissure development was varied. The most typical feature in the cen- tral volcanic zone was a small crack 2 to 4 cm wide wtth no vertical displacement, striking 020° across otherwise undisturbed volcanic terrain. Small cracks traced over the young sediment-free central region may extend 50 to 100 m before being ofTset in an en echelon pattern To the east and west the fissure widths vary, show vertical displacement, and may show considerable variation along strike. Beginning as l-m Table I. Electron microprobe analyses of basalt glasses. The analysis in column A is from ihe west flank of Mt. Pluto, dive 525. station 5. sample 2. Thai in column B is from ihe east flank of ihe median valley, dive 527. station 6. sample 3. Con- Percenlage stituent A B SiO- 48.9 51.1 TiO: 0X4 1.42 AIX>, Ib.l 14V FeO X.74 III 1 MnO 0 15 0 IX MgO 10.5 7X9 CaO 11. X II X Na.O 241 2.46 K.O 0.09 021 Cr.O, otw o.ox Totals 99.62 100.14 openings 10 m deep, they may then be completely filled with rubble or divide to form two fissures isolating a down- dropped block. Farther to the east and west vertical fault scarps predominate over open fissures. Proceeding from the central valley to- ward the walls, the submarine typically passed over a series of outward-dipping faults with displacement down toward the valley walls. This outward and downward displacement is maintained even though the general bottom gradient is upward. The structure of the central valley is thus a horst with a central sag. Beyond the mar- ginal graben, the normal faults increase in throw and become inward-dipping, as ex- emplified by the scarps forming the main east and west walls. One of these inward-dipping faults was investigated during four dives on the west wall. Here, the displacement observed on two of the dives was at least 300 m, from the top of the scarp to the top of the talus. The upper 1 30 m of the scarp consisted of truncated lava tubes and pillows lacking any intercalated sedimentary material. These grade through an interval of 30 to 50 m into a more massive outcrop character- ized by a blocky, angular surface without the elliptical outlines and radial joint pat- terns typical of truncated pillows. Lenticu- lar intrusions 20 to 30 cm wide and more than 2 m long strike obliquely across the exposed surface; they are most common I 5 to 50 m below the top of the massive out- crop. The total number of dike-like or sill- like intrusions observed on three dives, however, was surprisingly small on the order of a dozen or so. The nature of the fault zone on the west wall was strikingly different from thai of the faults in the inner valley. The latter were generally clean, fresh, and free of fault breccia and fault gouge On the lower west wall, on the other hand, the fault zone was 100 to 150 m wide, and exposed alter- nating, subvertical shells of fault breccia and massive basalt, dipping parallel to the face inward toward the inner rift vallev floor at about 80° (Fig. 6). Elsewhere the west wall appears to have been uplifted along a series of parallel normal faults, with individual displacements of less than 100 m showing exposures only of truncated lava. The upper surfaces of these fault slices form structural terraces capped by primary volcanic material which are back- lilted 5° to 7°. Major talus fans develop at the base of these scarps and. in many cases, on the back-tilted surfaces of the fault blocks. The faulting appears to be a continuing process with repeated episodes of dis- placement. Narrow fissures cut across indi- vidual pillow forms with no disruption oth- 716 Fig. 5. Varieties of minor and major lensional fractures, increasing in size from (a) to (d) In (d), the submarine is about 8 m into the fracture. er than a single plane of failure. Broken pillows on opposite sides of fissures up to I to 2 m wide can usually be visually matched, and indicate mainly dilation with minor vertical displacement. Coherent pil- low forms preserved in the fault scarps sometimes made it difficult to determine whether the slope was a fault scarp or a primary flow front. Renewed tectonic ac- tivity along older fault planes is indicated by faulted lithified calcareous sediments, a few centimeters to a meter in thickness, overlying the older pillow surfaces. During the diving operations, more than 50 fault scarps were inspected. In no in- stance were sediments found interbedded with the igneous rocks, ft was common, however, to see sediments accumulated in open fissures. Since tectonic activity domi- nates in the older flanking region, where sediments have had a greater time to accu- mulate, open fissures may represent the primary avenue by which sediments enter the geologic section. Two dives on the north and south sides of the active transform section in Fracture Zone B encountered heavy sediment cover in the topographic trough and on its walls. This cover consisted of a semiconsolidated blanket of calcareous sediment draped over the terrain, with only local exposures of weathered basalt talus or greenstone. On the north side of the transform, there was minor shearing and lilting of sediment and igneous rocks. No mafic or ultramafic rocks were recovered, although dredging on the topographic saddle to the south has recovered serpentmite. pyroxenite, and veined greenstone. No thermal anomalies were observed from the submarine, al- though many of the dredged greenstones are strongly fractured and brecciated, and are traversed by vuggy veins containing both calcite and aragonite which may be due to hydrothermal circulation. Fig. 6. Brecciated and vertically striated basalt on the west wall of the median valley 717 Summary Lava forms resemble those observed on terrestrial pahoehoe lava flows; the fea- tures that appear in truncated fault scarps as circular or elliptical pillows are elon- gated, tubular forms in three dimensions. Detached, subspherical pillows are very rare. The lavas show systematic chemical and mineralogical variation, with the oli- vine basalts associated with the central vol- canic highs and plagioclase-pyroxene ba- salts being typical of the west and east walls. Active volcanism is mainly restrict- ed to a narrow (0.5 to I km wide) central zone in the median valley. The central valley has a horst-like struc- ture which is bounded by graben at the base of the east and west walls. Intrusive sills and dikes are exposed only at the base of one 300-m scarp on the west wall Most fault displacements are less than 100 m and expose only breccia, truncated lava pillows, and tubes. In general, faulting appears to be a con- tinuing process, while volcanic activity is episodic. Structural deformation rapidly degrades the primary volcanic morphology typical of the. central highs, although vol- canic features are locally preserved on the wider structural terraces on the west and east flanks of the median valley. Dives in Fracture Zone B revealed minor deforma- tion of recent sediment cover, but there was no evidence of recent volcanic or hy- drothermal activity. References and Notes 1 H D Needham and J Francheteau, Earth Planet Set. Lett 22,29(1974). 2. R C Spindel el al. Nature (Lond I 246. 88 (1974). 3 J. G Moore, H. S. Fleming, J D Phillips, Geology 2.437(1974). 4. F. A. Frey. W B. Bryan, G Thompson, J Geophys Res 79, 5507 (1974). 5. J. G Schilling, Nature (Lond) 242. 565 (1973). 6 H Bougaull and R Hekinian, Earth Planet Set Lett 14, 249 (1974). 7. T. L. Wright and R S Ftske./ Petrol 2, I (1971). 8 J R Cann, Geophvs J R Astron Soc 39, 169 (1974); D Greenbaum, Nat Phys Sci 238, 18 (1972) 9 Supported by the Submarine Geology Branch, Na- tional Science Foundation, through grants GA- 35976 and GA^»1694; by the Seabed Assessment Program, International Decade of Ocean Explora- tion, through grant GX-36024; and by the Manned Undersea Science and Technology Office, Nation- al Oceanic and Atmospheric Administration, through grant 04-3-158-17. We wish to acknowl- edge the development of submersible in- strumentation and operation procedures under contract N0O0I4-71-C-O284, NR293-0O8 of the Advanced Research Projects Agency, and the sub- stantial contnbution of the U.S. Navy and the Naval Research Laboratories in providing narrow- beam bathymetry and detailed sea floor photogra- phy We are especially grateful to pilots J. Don- nelly, D Foster, L Shumaker, and V Wilson for their skillful handling of the Alvtn, and to Captain R Flegenheimer and the personnel of the R.V Lulu, and Captain E. Hiller and the crew of the R. V. Knorr for outstanding support. This article is Woods Hole Oceanographic Institution Contn- bution No. 3508 718 29 Reprinted from: Meteoritios 10, No. 4, 393. OMAN RING: SUSPECTED ASTROBLEME RobcrtS. Dietz*, John F. McHone**, Nicolas M. Short***, *NOAA, Atlantic Ocean, and Met. Lab., Miami, FL 33149 **Dcpt. Geology, University Illinois, Uibana, 1L ***NASA, Goddard Space Flight Center, Grecnbclt, MD A LANDSAT image (ERTS-1, 1217-06090, 25 Feb., 73) of Oman reveals a circular double ring structure (19°55'N, 56 58'E) in the desert region of Oman, on the Arabian Peninsula, which is a suspected astrobleme, ancient impact scar. Little is known about the surface geology except that the Oman Ring deforms a terrane of flat-lying Cretaceous rocks of marine shallow-water shelf facies. Although the exact geologic structure remains uncertain, it seems likely that the inner ring, 2.0 km across, mailcs a central domal uplift, while the outer ring, 6.0 km across, marks the outer ring fault, so that this suspected impact site is of moderate dimensions. The 1:3 ratio between the dome and the outer ring is typical of impact scars. The central dome suggests that the impacting body was a comet head (Milton and Roddy, 1972). I.M. LIBoushi of Oman has offered to make a Held investigation of this structure. Geologic maps show a NE-SW trending zone of Hornmz (Cambrian) piercement salt domes lying about 100 km to the N\V of the Oman Ring. Hut we have been unable to detect any of these on LANDSAT imagery, so it is doubtful that the Oman Ring is an outlying member of this field. 719 30 Reprinted from: IX International Congress of Sedimentology; Nice, France, Theme 10, 13-18. Also appeared in Offshore " Technology Conference, Pater #0TC 2385, 505-511. THIS PRESENTATION IS SUBJECT TO CORRECTION New York Alternative Dumpsite Assessment - Reconnaissance Study of Surficial Sediments By George L. Freeland and Donald J. P. Swift, NOAA ©Copyright 1975 Offshore Technology Conference on behalf of the American Institute of Mining, Metallurgical, and Petroleum Engineers, Inc. (Society of Mining Engineers, The Metallurgical Society and Society of Petroleum Engineers), American Association of Petroleum Geologists, American Institute of Chemi- cal Engineers, American Society of Civil Engineers, American Society of Mechanical Engineers, Institute of Electrical and Electronics Engineers, Marine Technology Society, Society of Explor- ation Geophysicists, and Society of Naval Architects and Marine Engineers. This paper was prepared for presentation at the Seventh Annual Offshore Technology Conference to be held in Houston, Tex., May 5-8, 1975. Permission to copy is restricted to an abstract cf not more thar 300 words. Illustrations may net be copied. Such use of an abstract should contain conspicuous acknowledgment of where and by v>hom the paper is presented . ABSTRACT Evaluation of offshore areas as potential dumpsites requires assessment of bottom sedi- ment character and transport. Preliminary evaluation of the potential for deposition of dumped materials, primarily sewage sludge, were made at two sites 60 nm from New York Harbor in 20 to 30 fm water depths. Since sludge particle density is barely over 1. 0, geological data were analysed for potential deposition and transport of fines in particular in addition to the sand-sized fraction. Results suggest a net southwestward bottom sediment transport, intensifying during winter storms. I. INTRODUCTION Purpose of the Study Waste disposal at the present New York Bight apex sewage sludge dumpsite is currently 5. 9 million yxr/yr. The dumpsite and its surroundings are presently undergoing intensiv^ study by NOAA's Marine Ecosystems Analysis Program (MESA) in order to determine the environmental impact of this dumping. MESA oceanographers are also examining two interim alternative dumpsites in the central part of the Bight shelf (Fig. 1). Preliminary results from these studies are presented in this paper. Areas of Investigation Location criteria of the northern site re- References and illustrations at end of paper. quires that it be a minimum of 25 nm from the Long Island shoreline, 10 nm from the axis of the Hudson Shelf Valley, and no more than 65 nn- from the New York Harbor entrance. The south- ern dumpsite is proposed to be located seaward of lines connecting the most shoreward 20 fm isobath, 10 nm from the Hudson Shelf Valley axis, and 65 nm from the New York Harbor entrance. Geological Setting of the Dumpsite Area The New York Bight is a pentagonal sector of the middle Atlantic shelf, approximately 100 nm wide at its maximum extent from the New York harbor mouth to the shelf edge (Fig. 1). Depths range between 20 and 30 fm over much of this area. The most prominent topographic feature of the shelf surface is the Hudson Shelf Valley, a broad, shallow channel extending from the Bight apex seaward as far as the outer shelf in the vicinity of Hudson Canyon. The shelf sediment surface has been pro- foundly modified by events associated with the Pleistocene ice age. During the last major ice advance, the Laurentide ice sheet extended to south-central Long Island and westward across New Jersey. Ancient shorelines near the head of Hudson Canyon indicate a sea-level lowering of about 450 ft. During this and preceeding periods of lowered sea level, stream erosion, amplified by meltwater discharge, dissected the uppermost shelf strata. The ancestral Northern 720 506 NEW YORK ALTERNATIVE DUMPSITE ASSESSMENT OTC 2385 New Jersey, Long Island, and Hudson Rivers scoured out their shelf valleys and incised the margins of the low divides that separated these drainage systems. As post-glacial sea level rose, shoreline ad- vance across the shelf resulted in erosion and beveling of the divides by the action of surf and coastal currents, and winnowing of fines from much of the shelf surficial sediment. Material released was transported by littoral drift south- westward along the slowly advancing shoreline into the shelf valley floors. The area of the Long Island Shelf Valley was a broad bay during much of this time, shielded from storm waves by a peninsula formed by the submerging Block Divi (Fig. l)(Swift, et al. , 1972). Thus, the pattern of trunk and tributary streams of the ancestral Long Island River are still clearly seen. The Hudson Shelf Valley, however, is so deeply in- cised that sedimentation by the transgressing shoreline and by subsequent storm-driven flows across the shelf have not sufficed to fill it in. the south, the North New Jersey Shelf Valley has been partly buried beneath sand ridges which stream southwestward from the crest of the Hudson Divide. Scour in the trough floors betwee|n the ridges has resulted in erosion of up to 20 ft. into early Holocene/late Pleistocene lagoonal clays and deposition of 20 ft. of sand on older strata to form ridges. Field Work Each of the two sites studied covered a 10 x 10 nm square and contained 36 sample stations, plus geophysical tracklines spaced 2 nm apart. In the northern area the grid was placed slightly seaward of the center of the proposed dumpsite to investigate, in part, a tributary valley systen which could be an area of deposition of dumped fines. The grid was oriented north-south so that sample lines would cross the valley. In the ern area, the broad, flat high area of the Hudso Divide and the strong ridge and swale topogra to the west were of interest, therefore track- lines were oriented northwestward. At sample stations a Smith-Maclntyre bottom grab sample and bottom photos were taken; at every fourth station bottom water samples 2 m from the bottom were collected. In the southern area, additional bottom grab samples (Shipek) were collected at 1/4 nm intervals along a centrally located line. On geophysical tracklines 3.5 kHz shallow-penetration seismic reflection profiling and side-scan sonar records were taken. In addition, the submersible NEKTON was used for two geological and three biological dives in the southern area. Field aliquots and two 15 cm Phleger cores were taken from the Smith-Maclntyre samples Size analysis consisted of splitting and dry screening freeze-dried samples on a 2-mm screen; passed material was washed of fines on a 230 mesh (62. 5 micron) screen, and run de through an automated rapid sediment analyzer for grain size distribution. Geophysical records were examined for areas of interest which were photographed. II. PHYSICAL NATURE OF THE PROPOSED DUMPSITES TolNorthern Site south ional physand for geological processing, and two vials of sample from the top centimeter were taken for chemical and fine-sediment analysis. From the 10-liter bottom water samples, two 2-liter samples were filtered onboard through a Wattman glass fiber GF/C filter for organic analysis, and a one 1-liter sample was filtered through a 0. 45 micron Nucleopore filter for suspended sediment analysis. Filters were immediately frozen. Data Processing Bathymetry : The area contains flat to gently sloping topography with depths from 25 fm in the northwest corner to nearly 32 fm in the east (Fig. 2). The most prominent feature is the rem- nant of an early Holocene/late Pleistocene north- west-trending stream tributary in the northern halfof the area, while to the southwest a flat shoulder gently slopes eastward. There probably has been only a slight smoothing of the basic topography formed before sea level rise. Stratigraphy : 3. 5 kHz seismic data reveal the presence of a thin (5-6 ft. ) surface layer, presumed to be Holocene, overlying a reflector similar to the early Holocene/late Pleistocene "basement" reflector which outcrops in the south- ern area. This upper layer is the sand sheet which was deposited over the Pleistocene eros- surface as sea level rose. The lack of njadditional, more modern sediment over the basal sheet is due to a limited sediment supply since transgression. In many places the "base- ment", a lagoonal clay, is exposed or eroded. Surficial Sediments: Sand, with some areas of over 5% gravel, comprise the surficial sedi- ment (Figs. 2 & 3). Fine sands lie in the north- eastern part of the area near the axis of the ancient stream valley, while coarser medium sands lie in the western and southern parts of the area. Mud deposits are restricted to only two stations with over 5% mud (5. 7% and 8. 2%). A pronounced gravel deposit (39% gravel) assoc- iated with coarser medium sand, is mapped in the southeastern part of the area. Assessment of sediment transport from grain-size distribution data is not very obvious as medium sand occurs in both topographic lows 721 OTC 2385 GEORGE L. FREE LAND and highs. The coarser sediment in the southern and western parts of the area could result from these areas being on northeast-facing slopes. Since the strongest storm-generated currents are throught to come from the northeast, these "upcurrent-facing" slopes may have been win- nowed of finer sand, leaving a coarser lag dep- osit. The lack of large areas of mud deposition indicates a) a nearly complete lack of mud as a source material, and/or b) what little, if any, mud being transported into the area is not re- maining. Bedforms : Bottom photographs indicate that the northern area is characterized by smooth, slightly undulatory, mounded or rippled bottom. These patterns reflect the competing activities of bottom wave surge, which tends to ripple the bottom, and the plowing activities of benthic fauna (mainly sand dollars) which smooth the bottom. Ripples form during storms and are slowly erased with intervening fair weather. Their spacing and height increase with increas- ing grain size: as ripples become larger, coarser sand appears on the crests, and finer sand in the troughs. If there is a wide range of grain sizes, or if sand or shell fragments are too coarse to be moved at all, the coarse par- ticles accumulate in the troughs along with med- ium sand, finer sand then collecting on crests. Under such conditions, sand dollars preferen- tially collect on ripple crests and flanks where & DONALD J. P. SWIFT 507 the areas in between the ridges. Stratigraphy: Seismic reflection profiles reveal the early Holocene/late Pleistocene re- flector outcropping in the area of ridge and swale topography, and a lower event probably within the Pleistocene. Surficial Sediments : Grain- size patterns in the southern area are more clearly related to bottom topography than in the northern area and thus are more easily related to sediment trans- port. The crest and east flank of the Hudson Divide are floored by coarse sand, the coarser fractions of medium sand (Fig. 5), and by up to over 20% gravel (Fig. 6). In the ridge and swale topography to the west of the Divide crest lie finer medium sand and fine sand. Submersible observations in the bottom of Veach and Smith Trough revealed the presence of a veneer of shelly, pebbly sand with large, angular clay pebbles in the troughs of ripples which were er- oded from the underlying early Hollcene/late Pleistocene substrate. The coarser sand and gravel deposits on the crest and east flank of the Hudson Divide point to southwest currents winnowing finer material from this "upcurrent" side of the Divide and de- positing it as longitudinal ridges in deeper water to the west. Currents must have scoured the intervening troughs at the same time. Bottom photos taken on the Hudson Divide show flat topo- graphy which has been rippled by bottom current^, they preserve ripple structure through armoring^^ reWorked by benthic organisms (surf clams, rather than degrading them sand dollars and worms). In the troughs of the Mesoscale bedforms appear on the sonograph ylige and swale topography, considerable coarse as highly reflective, low relief, east-west trend -shell debris lies in the ripple troughs along with ing zones 30 to 45 ft apart which are linear erosional windows exposing the basal Holocene pebbly sand. These zones appear to be aligned parallel to the main current flow direction, while degraded sand waves, which appear as poorly-defined patterns of repeating diamond shapes, are transverse to current flow. Southern Site Bathymetry: The southern area is consider- ably more complex than the northern area (Fig. 4). The crest of the Hudson Divide is aligned northwest-southeast through the center of the area separating irregular bottom sloping north- eastward into the Hudson Shelf Valley and Tiger Scarp from a strong-developed area of north- east-trending ridge and swale topography on the west side of the area. The ridge and swale topo- graphy probably formed from the pre-existing New Jersey Shelf Valley through the action of strong southwestward currents which have apparently transported sand from the crest and east flank of the Hudson Divide westward to buil<| ridges obliquely across the Valley and to scour coarser sand grains. The many small mounds of gray subsurface sediment brought up by the worms contrasts with the tan to reddish-brown color and more even appearance of undistrubed oxidized surface sediment. On ridge crests and upper flanks vast concentrations of small sand dollars occur, sometimes covering 100% of the bottom, amoothing ripples to the point where they are no longer distinguishable. Bedforms : Micro- relief in the southern area is quite similar to that in the northern area on the bottom away from the ridge and swale topo- graphy. Moderately rippled bottom, bioturbated ripples, and bioturbated mounded topography are present on the Hudson Divide and its flanks. In the troughs between ridges west of the Divide, strongly developed ripples and sand ribbon patterns appear in the thin veneer of coarse pebbly sand overlying the clay substrate. Most of the strong rippling probably forms during winter storms and then is modified during the summer. 722 508 NEW YORK ALTERNATIVE DUMPSITE ASSESSMENT OTC 2385 in. PROBABLE NATURE OF SEDIMENT TRANSPORT General The nature and pattern of sediment transport on the continental shelf is still poorly under- stood (Swift, et al. , 1972). Measurements of sediment transport rates have been largely con- fined to shallow laboratory flumes, while wave- driven transport and unidirectional transport have rarely been studied together at full scale. Empirical equations for sediment transport re- sulting from these laboratory studies must be applied with great caution to the shelf floor. Nevertheless, these laboratory studies, and field studies from other areas (usually nearer to the coast) permit the drawing of the following inferences concerning sediment transport in the study area. Sand Transport Cohesionless sediment (mainly sand) is en- trained by bottom flow only during major storms primarily in the winter. If the trajectory of a mid-latitude low pressure cell across the New York Bight is such that a day or more of intense northeast winds occurs, then a southward quasi geostrophic flow may develop across the entire shelf with velocities in excess of 40 cm/sec, sufficient to entrain bottom sand (Beardsley & Butman, 1974). Sand flow may be more intense than indicated by laboratory studies of uni- directional currents since the storm flow field has a marked wave surge component which "lubricates" the entrainment of sand. Storm trajectories with winds from other directions do not result in such close coupling between wind and water flow (ibid. ). Between storms the shelf flow field is rarely sufficiently intense to entrain sand. The response of the shelf floor to intermitten southwest-trending shelf flows may be seen in the gross pattern of distribution of surficial sediment and in the mesoscale morphologic pattern of ridges and swales. Shelf highs such as the Block, Long Island, and Hudson Divides are floored by coarser sand than are the lows, suggesting that the storm flow field tends to accelerate slightly over the highs due to the de- creased water cross-sectional area, and expand and decelerate over lows, with the result that finer sand is swept off the highs into the lows. Apparently, local interaction between substrate and flow has resulted in ridge and swale topo- graphy with relief of up to 40 ft, and ridge crests 2. 5 nm apart and several miles long. Side slopes are usually less than a degree. These drifts may possibly be the result of helical flow cells in the storm flow field coupling with the substrate. If so, zones of downwelling and flow acceleration occur over troughs causing scour, while upwelling and flow deceleration occurs over crests causing deposition. These velocity perturbations probably need be only a few percent of the ambient value to induce the observed topography. Smaller scale flow cells with spacing of tens to hundreds of feet are pro bably responsible for the streaky patterns (linear erosional windows) seen by side- scan sonar. In areas where regional flow fields chara- cteristically decelerate, sand willtend to settle from the flow more rapidly than it is entrained, creating constructional ridge topography. Sub- dued contours trending nearly east-west in the northern area (Fig. 2) may be a consequence of the overprinting of such a pattern on the old Long Island drainage system. The pronounced ridge topography in the western part of the southern area (Fig. 4) seems ,to be the result of the acceleration of storm flow over the Hudson Divide. The deeply incised troughs are a clear response to this flow. Co- hesive clays are locally exposed in trough axes and appear on seismic records as subsurface reflections that are truncated along ridge flanks. If these troughs are not being actively flushed, they should have been largely filled in by fine sand from adjacent ridges. In fact, the opposite may be occurring. Along the western side of the southern area, ridges tend to peak near the deep est portions of the adjacent troughs, suggesting build-up of crests at the expense of troughs. The ridge and swale topography, where developed, modifies the regional pattern of grain size distribution (Stubblefield, et al. , 1975). Trough axes contain discontinuous zones of coarse, rippled, pebbly sand, cohesive clay ;locally exposed in ripple troughs, and clay pebbles. Ridge crests tend to consist of medium- grained sand, while ridge flanks contain finer sand which may bridge across trough axes. The down-flank fining of grain size may reflect transport towards the end of storm periods, or milder events when flow is weaker and no longer couples with topography, when the finer sand fraction is mobilized on ridge crests and dis- perses down flank. Mud Transport General Finer particulate material consists of agglomerates of silt and clay- sized quartz and clay mineral particles bound together with organic matter. This material is cohesive, and settles out as mud patches on the sea floor after periods of peak flow when such muds are re- entrained into the water column. Unlike sands, muds may continue to remain suspended for days or weeks after the storm event, particularly ] 723 OTC 2385 GEORGE L. FREELAND & DONALD J. P. SWIFT ?09 in the near-bottom turbid layer. Wave surge tends to resuspend particles that settle out, creating a continuous exchange between the turbid layer and the bottom. Strong waves favor suspension in the turbid layer; wave decay favors deposition on the bottom. The behavior of sus- pended fine sediment is particularly relevant to the issue of ocean dumping of sewage sludge, as sludge mixes with natural suspended materials during dispersal. Suspended Sediment Concentrations : Over 90% of the fluvial sediment discharge from the north- east United States is effectively trapped near the coast in estuaries and coastal wetlands. Consequently, the terrigenous fraction of the suspended matter decreases rapidly seaward (Manheim, et al. , 1970; Meade, et al. , 1975; Schubel & Okubo, 1972). Solids in suspension at levels of the water column over the alternative dumpsites are predominately combustible plank- ton and their non- combustible remains (Drake, D., pers. comm.). Total suspended matter con- centration in surface water is from 100 to 500 ug/liter, comprised of 5% or less terrigenous matter, 80% combustible planktonic matter, and 15% siliceous and calcareous non-combusti- ble planktonic remains. Subsurface water con- centration is similar or somewhat less, except in the nepheloid layer 15 to 30 ft above the bot- tom. Concentration in this layer is 500 to 2000 ug/liter, consisting of 10 to 20% terrigenous matter, 30 to 60% combustible matter, and 50 to 80% non-combustible matter (ibid. ). Text- ural properties of bottom sediments in the alt- ernative dumpsites show that very little sedi- ment finer than 62 ug is deposited. Fairweather currents in these areas are generally to the southwest at net velocities of a few cm/sec. Because of greater average depths in the northern dumpsite, fines dumped there are more likely to be transported further prior to deposition than at the southern dumpsite. However, after deposition, fine sediments would be less likely to be resuspended, except during major storms. CONCLUSIONS Both dumpsites are floored by sand, pre- dominately medium-grained (0. 25 to 0. 5 mm). At the northern dumpsite, moderate sediment transport is indicated to the south and west over a gently sloping bottom incised by broad, low- gradient, pre-existing valleys. Evidence of more active sediment transport to the south- west is indicated at the southern dumpsite. This work is funded by NOAA's Marine Eco- systems Analysis (MESA) Program in the New York Bight. References Cited Beardsley, R.C., and Butman, B. , 1974. Cir- culation on the New England Continental Shelf: Response to Strong Winter Storms, Geophysical Research Letters, v. 1, pp. 181-184. Manheim, F. T. , Meade, R. H. , and Bond, G. C. 1970. Suspended Matter in Surface Waters of the Atlantic Continental Margin from Cape Cod to the Florida Keys. Science , v. 167, pp. 371-376. Meade, R.H. , Sachs, P. L. , Manheim, F. T. , Hathaway, J.C., and Spencer, D. W. , 1975. Sources of Suspended Matter in Waters of the Middle Atlantic Bight, Journal of Sedi- mentary Petrology , in press! Schubel, J.R. , and Okubo, A. , 1972. Comments on the Dispersal of Suspended Sediment Across the Continental Shelves, pp. 333- 346 in_, Swift, D. J. P. , Duane, D. B. , and Pilkey, O.H. (eds.), Shelf Sediment Trans- port: Process and Pattern j Dowden, Hutch- inson and Ross, Inc. , Stroudsburg, Pa. Stubblefield, W. L. , Lavelle, W.J. , McKinney, T. F., and Swift, D.J. P., 1975. Sediment Response to the Hydraulic Regime on the Central New Jersey Shelf. Journal of Sedi- mentary Petrology , in pre"ss! Swift, D.J. P., Kofoed, J. , Sears, P. , and Saulsbury, F. , 1972. Holocene Evolution of the Shelf Surface, Central and Southern Shelf of North America, pp. 499-574, in Swift, D.J. P., Duane, D. B. , and Pilkey, O.H. (eds.), Shelf Sediment Transport: Process and Pattern , Dowden, Hutchinson and Ross, Stroudsburg, Pa. 724 73°00' 72°00' Fig. 1 - Index map of the New York Bight. Contour interval U meters. Blocked out areas 2D1 and 2D2 are the proposed interim alternative dumpsites, MEDIUM SAND :¥*:: 1 75 - 1 99 ♦ / INI > 2 0O« FINE SAKD Fig. 2 - Northern alternative dump- site, area 2D1. Contour interval 1 fathom. Grain size distribution of the sand-sized fraction (l/l6 mm {hfi) to 2 mm [-10J ) . Dots are Smith-Ma clntyre bottom grab sample and photo stations. Fig. 3 - Northern alternative dump- site, area 2D1. Percent gravel over 5 percent. Bathymetric contour inter- val 1 fathom. Gravel contour interval 5 percent. 725 c C Fig. k - Southern alternative dump site, area 2D2. Bathymetric contour interval 1 fathom. + Smith Maclntyre bottom grab sample and photo station. - Geophysical trackline. —Submersible dives. Fig. 5 - Southern alternative dump- site, area 2D2. Grain size distribu- tion of the sand-sized fractions. Dots are Smith-Maclntyre grab and photo sta- tions. Contours are the 20 fathoms isobath. Fig. 6 - Southern alternative dump- site, area 2D2. Percent gravel (heavy contours) over %, contour interval 5%. Dot pattern is area of 1.00 to 1.2U0 sized medium sand from Fig. 5. Large dots are sample stations. Light contours are the 20 fathom isobath. 726 Reprinted from: Mid-Atlantic Shelf/New York Bight ASLO Symposium, 24-25. George L. Freeland, Donald J. P. Swift and William L. Stubblef ield , NOAA/AOML, 15 Rickenbacker Cswy., Miani, Florida 33149; Anthony E. Cok, Adelphi University, Garden City, New York 11530. SURFICIAL SEDIMENTS OF NOAA-MESA STUDY AREAS IN THE NEW YORK BIGHT The nature of bottom sediments and sediment particles suspended in the water column becomes of interest to environmental managers when man's activities In the ocean causi perturbations on the bot- tom and in the near-bottom water column. In addition to the immediate results of anthropogenic (man's) activities, one must also consider the effect of long term natural phenomena. Sediments on the ocean floor in the New York Bight are important as there is extensive ocean usage by man. Users of sedimentologicai data are fisheries biologists, sanitary and ocean engineers, public health experts, vessel captains, and a myriad of government planners. The more important categories of human usage of the shelf surface are for food resources, waste disposal, foundations, mineral resources and recreation. The distribution of surficial sediments across the surface of the Bight may be explained in terms of sea level fluctuation caused by continental glaciation during the past several million years. During the last such episode, when ihe North American ice sheet extended from Canada down as far as Long Island and northern Kew Jersev, sea level was lowered to about 155 meters below present sea level in the vicinity of Hudson Canyon. From the time of the maxinum ice advance about 15,000 years at,o (Hilliman a'ld Emery, 1968), to about 5,000 years B.P. the ice has been melting and the shoreline has advanced over the shelf to its present position. The surficial sediments and features seen on the shelf today are the result of this fall and rise of sea level. As transgression progressed, fluvial and older sediments of the land surf. ice now occupied by the shelf were first covered by estuni ine and lagoonal sediments behind barrier islands or directly reworked by the advancing shoreline. Sand eroded at the shoreface was swept partly back onto the barrier islands by storms and buried there, only to be re-exposed as the shoreface advanced. Most of the material, however, has been washed down- coast and seaward to accumulate as a discontinuous sand blanket 0 to 10 n thick (Stabl el a)., 1974). Thus t lie dominant material present on the shell floor is sand-sized sediment, unconsolidated fine- grained sediment haying been resuspended and transported back into the estuaries or off the shelf edge. Locally, the underlying stratum of transgressed lagoonal and estuarine semi-eunsolidated nud deposits are exposed on the sea floor (Swift et al., 1972: Stahl et al., 1974; Sheridan or al., 1974), recognizable by angular clay fragments and oyster shells. Studies off the New Jersey coast have been made of nearshore and central shelf areas of t idge and swale topography (McKinney et al., 1474; Stubb! ef ield et al., 1974; Stubblef in Id et al,, 1975; Stubblef iei .1 and Swift, 1975). Results suggest that fine sands are scoured i rom depressions during winter northeast storms, and are deposited on crests which appear to build up concurrently. The underlying early Holocene lagoonal clay is exposed in particularly deep troughs. Trough axes art locally floored with coarse sand, shell debris, occasional to common lagoonal clay fragments , or mnv have a mud covering in deep holes. Sandy bottoms are usually strongly rippled. Flanks are coveicd with fine to very fine sand; the crests consist of medium to fine sand. In the Bight apex, extensive sedimentologicai studies and a 1973 bathymetric survey reveal that the only significant change in bottom topography since 1936. is at the dredge spoil dumpslte, where t|he dumping of 88 million cubic meters of dredged material has caused up to 10 meters of shoaling. The center of the Christ iaensen Basic, a natural collecting area for fine-grained sediment, is no doubt contaminated with sludge, but shows no apparent sediment build-up during the intervening 37 years. Preliminary sediment distribution maps show the apex outside of the Christ iaensen r.isin tu be floored primarily by sand ranging from silty fine sand to coarse sand, with small areas of sandy grav- el, artifact (anthropogenic) gravel, and r.iud . The nearshore mud patches which have caused environ- mental concern appear to be periodic, at times being covered with sand, and occasionally scoured out. Side-scan sonar records show linear bedforms indicative of sand movement over most of the apex are... To the east of the apex along the south shore of I.rng Island, sediment samples and geophysical data have been collected a:,d are currently being processed. Closely-spaced d;.ta were collected in an area of ridge and swale topography near the terminus of the Suffolk County sewer outfall. Two mid-shelf areas, located 120 km from the entrance to Lower Bay and 1.8 km north and south, respectively, of the axis of the Hudson Shelf Valley have been designated by the MESA Project Office as proposed interim alternative dumping areas for sewage sludge and possibly dredge spoil. Dach of the areas studied is an 18 by 18 km square within which bottom samples, photographs, and side-scan sonar and seismic reflection profiles were taken. Tiie northern area is located in a tributary valley of the ancestral Long Island River System where the surficial sediments consist of fine sands to the northeast, coarser medium sands to the west and south, and 39X gravel at one station in the south. Only two stations contained over 5% mud (5.7% and 8.2%). Bottom photographs indicate that the area is characterized by smooth, slightly undulatory , mounded or rippled bottom. In the southern area coarser sand and gravel deposits lie on the crest and east flank of the Hudson Divide, while medium and fine sand occur in the ridge and swale topography to th. west. These distributions point to the winnowing of fine sediment from the crest and east flank of the Divide and deposition to the west. Submersible observations in Veatch and Smi tli 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. Based on these studies, if sewage sludse vere dumped, widespread dispersion, mostly Co the southwest, could be exported, with winter-time resuspensioa and transport out of the a^eas of fine material on the bottom. Higher vohrss of dumped particles would find their way into the rludsor Shelf Valley if the northern ^re.^ were -:seJ. Considerably more rapid build-up and possibly permanent build-up of dumped particles on the bottom could be expected if dredged material were cunpeo. 31 727 REFERENCES McKinney, T.F., W.L. Stubblef ield and D.J. P. Swift. 1974. Large-scale current llneations on the central New Jersey shelf: investigations by side-scan sonar. Mar. Gcol. 17:79-102. Milliman, J.D. and K.O. Emery. 1968. Sea levels during the past 35,000 years. Science 162:1121-1123. Sheridan, R.E., C.E. Dill, Jr. and J.C. Kraft. 1974. Holocenc sedimentary environment of the Atlantic inner slu'lf off Delaware. Bull. Ceol. Soc. Amer. 85:1319-1328. Stahl, I,., J. Koczan and D. Swift. 1974. Anatomy of a shorcf ace-connected sand ridge on the New Jersey shelf: implications for the genesis of the surficial s^nd sheet. Geology 2:117-120. Stubblef ield, W.I,., M. Dtcken and D.J. P. Swift. 197m. Reconnaissance of bottom sediments on the inner and central New Jersey shelf. N0AA-MESA Report No. 1, 39 pp. Stubblef ield, W.L., J.W. Lavclle, T.F. McKinney and D.J. P. Swift. 1975. Sediment response to the present hydraulic regime on the central New Jersey shelf. Jour. Sed. Petrol. 45:337-358. Swift, D.J. P., J.W. Kofoed, F.P. S.iulsbury and P. Jears. 1972. Holccene evolution of the shelf surface, central and southern Atlantic shelf of North America, p. 499-574. In D.J. P. Swift, D.B. Duane and O.H. Pilkey (eds.), Shelf Sediment Transport: Process and Pattern. Dowden, Hutchinson and Ross. Stubblef ield, W.L. and D.J.F. Swift. 1975. Ridge development as revealed by sub-bottom profiles on the central New Jtrsey shelf. In press _in_ Mar. Ccol. 728 32 Reprinted from: American Geophysical Union} Fall Annual Meeting 56, No. 12, 1003. CALCULATION'S OF SAND TRANSPORT ON THE NEW YORK S1ELF USING NEAR- BOTTOM CURRENT METER OBSERVATIONS P- E. Gadd J. W. Lavelle D. J. P. Swift fall at: ATLANTIC OdANOGRAPHIC 5 KETEOROLOGICAL LABORATORIES, 15 Rickenbacker Causeway, Miami, Florida 351-19) Using near-bottom current meter and suTficial sediment size observations in conjunction with empirical transport formulae, calculations of cohesionlcss bedload sediment movement within the New York Bight have been made. The study employed data collected from 23 long-term current meter records which span a 2 1/2 year period. Measured bottom current speeds suggest that shelf sand transport occurs in intense pulses, associated with storms, separated by relatively long periods of quiescence. The percentage of time that the threshold velocity is exceeded during an entire current meter record serves as a measure of the transport activity existing in that particular area. It has been determined that these percentages vary from less than 1% in offshore areas during fair weather flow regimes to 901 in the shoal regions adjacent to the New York Harbor Entrance which are regularly subjected to intense tidal" currents . On the gentle slope of the continen- tal shelf south of Long Beach, Long Island, transport quantities tend to increase logarith- mically as the water depth decreases. The Hudson Shelf Valley exhibits anomalously high sediment transporting capabilities coincide^' with periods of strong westerly winds. 729 33 Reprinted from: Mid-Atlantic Shelf/New York Bight ASLO Symposium, 31-33. Patrick G. Hatcher and Larry £. Keister NOAA/ACML, 15 Rickenbacker Causeway, Miami, Florida 33149 CARBOHYDRATES AND ORGANIC CARBON IN NEW YORK BIGHT SEDIMENTS AS POSSIBLE INDICATORS OF SEWAGE CONTAMINATION As part of NOAA's Marine Ecosystem Analysis Program in the Mew York Bight, sediment samples have been collected over a wide area of the Bight, including the Hudson Shelf Valley, and analyzed for total organic carbon (TOO and total carbohydrates (TCH). These parameters were primarily examined to provide a gross qualitative analysis of the bulk organic matter in the sediments. The TOC distribution indicates that organic matter is primarily associated with silty sands which accumulate in topographic lows such as the Christiaensen Basin and the Hudson Shelf Valley. The high concentrations of TOC (usually 3-5%) have previously been attributed to sewage sludge which is being released in large quantities to the area by ocean dumping (Gross, 1972; Pearce, -1972). However, the TOC distribution may largely be attributed to the fact that its occurrence in sediments is usually inversely related to particle size distribution (Hunt, 1961) and that fine particles tend to accumu- late in topographic lows. The TOC distribution would thereby be closely related to the isobaths. TCH distribution in sediments of the New York Bight is very similar to that of TOC. In fact, a correlation coefficient of .99 between TOC and TCH at the head of the Shelf Valley typifies their relationship in the Apex. Hcwever, progressing seaward this relationship deteriorates and the corre- lation coefficient drops to .8?. From this observation we can deduce that the organic matter in the Apex is fairly uniform in carbohydrate content (as if from a singular source) whereas seaward the organic matter is less uniform in carbohydrate, suggesting that it may be derived from more than one source. This could possibly be resulting from the fact that carbohydrates, introduced in large amounts in the Apex, are transported seaward and diluted with oceanic sedimentary organic matter which is generally depleted of carbohydrates (Degens, 1967). From the previous discussion we can see that both TOC and TCH are dependent on the particle size or the total amount of organic matter present. By reporting TCH as a percentage of TCC the new para- meter (TCH/TOC) becomes independent of particle size or total organic content. TCH/TOC is, thereby, a qualitative parameter for organic matter. Figure 1 shows the distribution of TO1/T0C in sediments ■of the New York Bight. Values generally range from 20 to 60 with the high values located near the Long Island shore, the head of the Hudson Shelf Valley, and in the Shelf Valley itself. The high values ar? at least a factor of 2 greater than expected for shelf sediments (Shabarova, 1955; Degens. 1967). Since TCH/TOC is qualitative, the contours are suggestive of the fact that a source of organic matter exists at the head of the Shelf Valley and it is being diluted seaward within the Valley. We propose that the major source of this high carbohydrate organic matter is sewage-derived, based on known concepts about the carbohydrate content of various organic matter sources entering the Bight. Organic matter derived from terrestrial soils and transported to the Bight by the Hu.lson River is expected to have a TCH/TOC of approximately 20. However, liLtlp of this material reaches the Bighr but, rather, is trapped in the estuary (Meade, 1969). Its contribution to the TCH/TOC in sediments of the Bight is, therefore, expected to be minimal. A major fraction of the particulate material in near-shore areas is composed of living planktonlc organisms. These contain substantial amounts of carbohydrates and the TCH/TOC is often around 30 to 80 (Strickland, 1965). Easily hydrolyzable sugars such as glucose and galactose compose the major fraction with structural carbohydrates such as cellulose and hemicellulose composing only a minute fraction (Parsons et a_l. , 1961). By the time these organisms die and settle to the sea floor as de- tritus, most of the carbohydrates (and also up to 90% of the organic matter) are decomposed with only the resistant ones surviving bacterial decomposition. The TCH/TOC thereby decreases to a value of less than 10 in the upper layers of sediment (Degens, 1967). In the New York Bight, phytoplankton production may well surpass all other inputs of organic matter in the water column with an average annual production of 800x10" kg carbon (Malone, 1975). However, if the sedimentary organic natter were to be exclusively derived from phytoplankton, we would observe a TCH/TOC of less than 10. The sediments in the Apex have a TCH/TOC ranging from 40 to 50, suggesting the presence of a high-carbohydrate source other than phytoplankton. Sewage sludge is presently being released to the New York Bight at a rate of 5x10 ra /yr (U.S." EPA, 1975) which represents roughly 160x10° kg C/yr. Sewage outfalls located along the Long Island and New Jersey shores discharge an undetermined amount of sewage to the Bight. Outfalls located in the Hudson River estuary, the Hudson River, the East River, and various other locations in Metropoli- tan New York also discharge an undetermined amount of sew?.ge. This material most likely settles out within the estuary but eventually may be dredged and dumped in the New York Bight. As we can see, the Hew York Bight receives a substantial amount of sewage-derived organic materials, possibly half that supplied by phytoplankton. Sewage contains a substantial amount of carbohydrates, mostly in the form of cellulose and hemi- cellulose (Hunter and Heukelekian, 1965). The TCH/TOC is roughly 30 for sewage sludge. Once sewage is released to the environment it undergoes relatively rapid decomposition and a certain amount of TOC is lost. However, more resistant components such as cellulose and hemicellulose may not undergo an equivalent amount of decay. The TCH/TOC may therefore increase to values of 40 to 60 or more as the TOC. is being pref errentially lost. Therefore, as sewage-derived components settle to the sediment the TCH/TOC increases and it may continue to increase as more decomposition takes place. For the New York Bight the value of TCH/TOC will be influenced by the relative contribution of organic matter from each of the sources and also the degree of decomposition which has occurred. The largest source of organic matter to the Bight is phytoplankton; however, up to 90% of this may be lost before being incorporated in the sediment. Sewage is the other major source and a much larger frac- tion is expected to reach the sediments due to the fact that most of its organic matter is relatively resistant to microbial decomposition. We therefore are confident in stating that the major source of organic matter to the sediments is derived from sewage and the high TCH/TOC ratios (30-50) observed in 730 -32- 73°40'W 7"3°00'W i> Fig. 1. The distribution of TCH/TOC in sediments of the New York Bight. The -fc denotes the sewage sludge dumpsites. the Bight are in turn du to the large percentage of sewage-derived organic matter. A TCH/TOC of 40 or 50 in the sediment indicates that the organic matter contains a substantial amount of sewage which has undergone some decomposition. A TCH/TOC of 20 or less suggests that sewage-derived organic matter is a small fraction of the sediment organic matter which is dominated by phytoplankton organic matter. REFERENCES Degens, E.T. 1967. Diagenesis of organic matter. In G. Larsen and G.V. Chilingar (ods.)i Diagenesis in Sediments. Elsevier. Cross, M.C. 1972. Geologic aspects' of waste solids and marine waste deposits, New York metropolitan region. Geol. Soc. Am. Bull. 83:3163-3176. Hunt, J.M. 1961. Distribution of hydrocarbons in sedimentary rocks. Geochim. Cosmochim. Acta 22:37-49. Hunter, J.V. , and H. Heukelekian. 1965. The composition of domestic sewage fractions. J. Water Pollut. Control Fed. 37:1142-1163. Malone, T. 1975. Phytoplankton productivity: nutrient recycling and energy flow in the inner New York Bight. Unpublished report to NOAA, Contract #03-4-043-310. 731 -33- Meade, R.H. 1969. Landward transport of bottom sediments in estuaries of the Atlantic Coastal Plain. J. Sed. Petrol. 39:222-234. Parsons, T.R. , K. Stephens, and J.D.H. Strickland. 1961. On the chemical composition of eleven species of marine phytoplankters. J. Fish. Res. Bd. Canada 18:1011-1016. Pearce, J.B. 1972. The effects of solid waste disposal on benthic communities in the New York Bight. In M. Ruivo (ed.), Marine pollution and sea life. Fishing News, Surrey, England. Shabarova, N.T. 1955. The biochemical composition of deep-water marine mud deposits (ocean bottoms). Biokhimiya 20:146-151. Strickland, J.D.H. 1965. Production of organic matter in primary stages of the marine food chain. In J. P. Riley and G. Skirrow (eds.), Chemical oceanography. Academic. U.S. Environmental Protection Agency. 1975. Ocean disposal in the New York Bight. Technical Brief Report 02. U.S. Govt. Printing Office. 732 34 Reprinted from: Seventh International Meeting on Organic Geochemistry, Madrid, Spain, 61. MANGROVE LAKE, BERMUDA; ITS SAPROPELIC SEDIMENTARY ENVIRONMENT PATRIC G. HATCHER (1), - BERND R. SIMONEIT (2), - SOL M. GERCHAKOV (3) Over 9000 years since its origin, Mangrove Lake, Bermuda has accumulated 18 m of sediment representing three different lithologies: a peat overlain by a fresh— water sapropel, and a brackish— wate sapropel. The organic geochemistry of this anoxic, organic —rich, and multi— banded gel was examined to identify various organic constituents (TOC, H, N, carbohydrates, proteins, alkanes, fatty acids, and other bitumens) and their early diagenetic changes. Major diagenetic changes are observed for carbohydrates and proteins which constitute a large fraction of the organic matter and may be undergoing condensation to humates. Total extracts of the sediment reveal the presence of hydrocarbons, fatty acids, polycyclic hydrocarbons, aromatic hydrocarbons, phenols, esters, and nitrogenous com- pounds. The n— alkanes and n— fatty acids are of algal, bacterial, and terrigenous source. No correlations between n— alkanes and n— fatty acids are observed to indicate any significant chemical maturation. In addition to diagenetic changes, distinct changes are observed in the depth distributions of all organic geochemical parameters relating to the abrupt changes in depositional environments within each of the lithologies. Mangrove Lake is presented as an interesting site for studies in early diagenesis in that it is an enclosed marine ecosystem where large amounts of organic matter are rapidly accumulating. (1) NOAA, Atlantic Oceanographic and Meteorological Laboratories. 15 Rickenbacker Causeway, Miami, Florida 33149. (2) Space Sciences Laboratory. University of California, Berkeley, California 94720. (3) Department of Microbiology, University of Miami, Miami, Florida 33152. 61 733 35 Reprinted from: Clays And Clay Minerals 23, 331-335, NOTES ELECTROSTATIC CLEANING TECHNIQUE FOR FABRIC SEM SAMPLES (Received 19 March 1975) Durmg clay fabric investigations of slightly consolidated submarine sediments, a technique was developed for elec- trostatically cleaning surfaces of scanning electron micro- scope (SEM) samples. One of the two surfaces resulting from a single fracture of an oven-dried sample was cleaned using the conventional peeling technique (100 applications of cellophane tape; Barden and Sides, 1971) (Fig. la). The opposite surface was cleaned using the electrostatic tech- nique now routinely employed in this laboratory (Fig. lb). Both surfaces were cleaned satisfactorily. In sharp contrast, uncleaned fracture surfaces of this sample (not shown) were noted to be debris-cluttered*. Both micrographs of Fig. I show predominately stepped face-to-face arrangement of the particles in regions which appear more dense whereas numerous oblique edge-to-face contacts occur in areas which appear less dense. The relatively greater proportion of the clay particles lying approximately perpendicular to the surface of the peeled sample may be an artifact pro- duced by the cleaning technique. The process of pressing tape against the sample and pulling it away not only removes debris but also may preferentially remove flat- lying particles of the sample or pull them up on edge. Although we have not shown definitively that peeling does produce such an artifact, numerous peeled surfaces have a similar 'lifted' appearance and these features are highly suspect. Slightly consolidated clay sediments which are prepared using air or oven-drying techniques have been shown to undergo severe stresses and deformation with large reduc- tions in pore space (Yong, 1972; Naymik, 1974). For this reason, in recent investigations of sediment fabric, samples have been prepared by freeze drying or by critical point drying. These less disruptive drying methods result in rela- tively fragile material when applied to uncemented, high- porosity clay sediments. Consequently, these dried samples may lack sufficient structural integrity for peeling tech- niques to be applied. Fracture surfaces of such fragile sam- ples can be cleaned using the electrostatic technique (Fig. 2). The micrograph in Fig. 2(a) shows bands running roughly diagonally which have relatively more edge-to- edge (EE) and edge-to-face (EF) particle contacts alternat- ing with bands which have the predominant face-to-face (FF) structure. The EE and EF contacts lead to a structure far more open in appearance than the FF contacts (Fig. 2b). * See also photomicrographs of uncleaned and peeled fractured surface in Tovey and Yan (1973). In electrostatic surface cleaning, loose particles are lifted from the sample in an electrostatic field without any physi- cal contact with the surface. A field of approximately 20 kV/cm at the sample surface is satisfactory. The field is produced in this laboratory by briskly rubbing a piece of cellulose acetate butyrate tubing with a piece of polyester cloth. The charged tubing is then moved slowly over the fracture surface at a distance of about 1 cm. In order to reduce charge build up. the specimen is fractured and attached to the microscope stub before cleaning, and the stub is grounded at least intermittently during cleaning. A sample can be cleaned in less than 1 min. No features have been observed that indicate any disturbance of the clay fabric due to electrostatic cleaning. The electrostatic cleaning technique is rapid and very economical and has performed satisfactorily in routine SEM sample prep- aration. Acknowledgements — We wish to thank Dr. D. Latham of the University of Miami for his assistance in measuring the electrical field. Dr. W. R. Bryant supplied the Missis- sippi Delta sample. Mr. W. B. Charm assisted with micro- scopy and Drs. G. H. Keller and W. D. Keller reviewed this manuscript. M. Hulbert acknowledges the support of a National Research Council Associateship. National Oceanic and Atmospheric Administration, Atlantic Oceanographic and Meteorological Laboratories, Marine Geology and Geophysics Laboratory, 15 Rickenbacker Causeway, Miami, Florida 33149, U.S.A. REFERENCES Matthew H. Hulbert Richard H. Bennett Barden. L. and Sides, G. (1971) Sample disturbance in the determination of clay structure: Geotech. 21, 211-222. Naymik, T. G. (1974) The effects of drying techniques on clay-rich soil texture In: Proc. 32nd Ann. Meeting Elec- tron Microscopy Soc. Am. (Edited by Arceneaux, C. J.) pp. 466-467. Saint Louis, MO. Tovey, N. K. and Yan, W. K. (1973) The preparation of soils and other materials for the SEM : Proc. Int. Symp. Soil Structure, Gothenberg, Sweden, pp. 59-67. Yong, R. N. (1972) Soil technology and stabilization. In: Proc. 4th Asian Regional Conf. Soil Mechanics, (Edited by Moh, Z. C.) Vol. 2, pp. 1 1 1-124. Bangkok, Thailand. 734 -r ' Fig. 1. Opposite surfaces from a fracture of oven-dried continental slope sediment (Wilmington Canyon; (a) peeled, (b) electrostatically cleaned. 735 736 Fig. 2. Critical-point dried, electrostatically cleaned submarine sediment (Mississippi Delta); (a) general view, (b) detail of EE, EF structure (rosette). 737 738 36 Reprinted from: Journal of the Geological Society 131, 311-321. Geophysical maps of the eastern Caribbean PHILIP KEAREY, GEORGE PETER & GRAHAM K. WESTBROOK SUMMARY The results of marine geophysical surveys by tive gravity anomaly belt east of the Lesser NOAA, Miami and the University of Durham Antilles, and reveal new features not previously with the Royal Navy in the eastern Caribbean mapped, such as easterly trending anomalies region are presented as maps of the bathym- east of the Lesser Antilles and the complicated etry, free-air gravity anomaly, Bouguer gravity anomaly pattern of the Aves Ridge. They anomaly and total field magnetic anomaly, provide a detailed coverage of the area bounded These maps show in greater detail features by latitudes io°n. and I7°n., and longitudes which were already known, such as the nega- 57°w. and 65°w. The eastern margin of the Caribbean Sea is formed by several arcuate structures (Fig. i). The westernmost is the Aves Ridge, which is a submarine feature with many of the characteristics of an old, submerged island arc. The most prominent is the Lesser Antilles Island Arc, which is a pronounced ridge with a series of volcanic islands situated along it, the oldest of which are of Eocene age. It is separated from the Aves Ridge by a sediment-filled depression, the Grenada Trough. The easternmost element is the Barbados Ridge Complex which is essentially a thick, complexly deformed sediment pile filling a crustal depression that once held the former trench of the Lesser Antilles Arc. In the late 1960s and early 1970s, Durham University with the Royal Navy, and the Atlantic Oceanographic and Meteorological Laboratories of NOAA, Miami, conducted an extensive marine geophysical survey programme over this area as part of the Cooperative Investigation of the Caribbean and Adjacent Regions (CICAR) . The results of these surveys, supplemented by earlier work where available, are presented here in the form of maps of the bathymetry, free- air and Bouguer gravity anomalies, and total field magnetic anomaly (larger scale dyeline copies of the maps are deposited in the Geological Society's Library). The maps are based on a system of closely-spaced track-lines, (Fig. 1). Naviga- tion in the Durham-Royal Navy survey areas was by Decca Lambda electronic fixing system, which provided an accuracy of approximately ±200 m, with positions fixed every ten minutes. Satellite and radar fixes near land were used for checks. Navigation for most of the NOAA survey lines was by satellites and Loran C, with positions fixed about every two hours. On some of the earlier NOAA lines and on those of other institutions, navigation was mainly by celestial fixes and Loran C. The method of positioning and data collected by the various ships concerned are given in Supplementary Publication No. 18008 (1 page), deposited at The British Library, Boston Spa, Yorkshire, U.K. and at The Geological Society Library. Bathymetry was monitored mostly by conventional sonic transducers (with the exception of the 1971-1972 NOAA lines, most of which were done with a narrow-beam transducer), the total magnetic field by proton precession Jl geol. Soc. Lond. vol. 131, 1975, pp. 311-321, 5 figs. Printed in Northern Ireland. 739 312 P. Kearey, G. Peter & G. K. Westbrook magnetometers, and the gravity by Graf-Askania or La Coste-Romberg sea gravity meters. Gravity was tied into several established reference stations on the islands (Masson Smith & Andrew 1965; Dorman et al. 1973). Key Durham University & Royal Navy 1971, 1972 N.O.AA 1971,1972 N.O.AA 1968,1969 Other institutions LESSER J^C;::5>- ANTILLES . Map Area Fig. 1. Chart showing the ships' tracks on which data vised in the map compilation were collected. 740 Geophysics of E. Caribbean 313 i. Bathymetry — Map 1 The bathymetry of the eastern margin of the Caribbean is presented at 200 m contour intervals. The depth values are given in corrected metres using Matthews tables (Matthews 1939), except in the area surveyed by H.M.S. HECLA, where depths were obtained for a seawater sound velocity of 1-5 km s_1. Uncorrected depths with this velocity assumption yield an error up to 40 m in depths over 4 km, but over most of the area the error is less than 10 m. Data reduction methods for the NOAA data are contained in Peter et al. (1973a). Bathymetric data were supplemented by existing charts to improve interpolation between survey lines south of latitude I3°N. and around the islands. The major bathymetric features from w. to E. are the Aves Ridge, the Grenada Trough, the Lesser Antilles Island Arc, the Tobago Trough-Lesser Antilles Trench, and the Barbados Ridge Complex. The Aves Ridge is a broad, linear feature characterized by a series of prominent, N.-s. trending ridges on its western flank and less well-developed ridges on its eastern flank. Its western edge is a fairly straight, N.-s. trending slope, but its eastern edge exhibits a radius of curvature (c. 400 km) which is similar to that of the Lesser Antilles Island Arc. The Grenada Trough exhibits subdued topog- raphy south of i5°N, while in the north it becomes increasingly rugged. The ridge upon which the Lesser Antilles Islands are situated is a major feature steeper on its western side than its eastern side, s. of Marie Galante. East of Guadeloupe, La Desirade lies on a prominent high which is the southernmost of a series of highs forming a shelf outside the northern islands of the arc. Between the Barbados ridge, on which the island of Barbados is located, and the Lesser Antilles Arc lies a smooth bottomed depression, the Tobago Trough. The so-called Lesser Antilles Trench is located east of Martinique. The two are separated by a saddle east of St. Lucia. East of these two depressions lies a broad region, approxi- mately 300 km across, of irregular bathymetry, referred to as the Barbados Ridge Complex. There are certainly many more minor peaks and troughs in this region than those that could be resolved with the trackline spacing of the surveys. The Barbados Ridge Complex is about 1000 m higher south of I4°n, than to the north. In the southern half there is a well-developed series of n.-s. trending ridges and troughs, a large trough being well developed on the eastern flank of the Barbados Ridge. These features are more subdued north of I4°N., and are cut by e.-w. trending troughs and ridges. To the n. the Barbados Ridge Complex terminates along a nw.-se. trend south of i6°n. (parallel with the trend of the Barracuda Ridge to the ne.), and to the south it merges with the S. American continental shelf. Its eastern margin is well defined by a relatively steep slope down to the Atlantic Ocean floor. 2. Free-air gravity anomaly — Map 2 In the regions surveyed in 1971 and 1972 the anomalies shown are accurate to within ±5 mgal, elsewhere they may be in error by up to ±10 mgal. The most conspicuous feature of the map is the large, linear negative anomaly, 741 £ E u pa b i « wis 6 «f h 2n 742 •a Pk >* a a n c (3 C4 ^ fr So > 6D 743 316 P. Kearey, G. Peter & G. K. Westbrook reaching —270 mgal, e. of the Lesser Antilles Island Arc. The presence of this anomaly was first noted by Hess (1933, 1938) and later workers have delineated it further (Andrew et al. 1970; Bunce et al, 1971 ; Bowin 1972). Our surveys show that the axis of this negative anomaly, instead of being a smooth arcuate curve as was once supposed is offset in a sinistral sense along the major e.-w. topographic trends that are present north of I4°n. South of i4°N, where the Barbados Ridge is best developed, the anomaly is split by the ridge into two distinct parts; one follows the eastern side of the Tobago Trough ( — 120 mgal), the other runs over the eastern flank of the Barbados Ridge. Over the eastern part of the Barbados Ridge Complex and the directly adjacent Atlantic Basin there are e.-w. trending anomalies. A linear belt of positive free-air anomalies (100-180 mgal) defines the Lesser Antilles Island Arc. The Aves Ridge is also characterized by positive free-air anomalies, which may reach maxima in excess of 100 mgal over topographic prominences. Over the Grenada Trough the free-air anomalies are negative in the south, reaching a minimum of at least —70 mgal; free-air anomalies increase northwards to small positive values. 3. Bouguer gravity anomaly — Map 3 Bouguer gravity anomalies were computed for an average crustal density of 2-67 g cm-3. Around the volcanic islands a value of 2-8 g cm-3 would have been more suitable, but in the region of sedimentary relief a density of 2-2 g cm-3 would have given a more realistic result. In view of these differences a compromise value of 2-67 g cm-3 was adopted, which is the commonly used average. It is also the value most suitable for the region of the continental shelf and slope. In computing the anomalies a correction for the two-dimensional effect of seabed relief was applied along most of the tracklines. The exception is in the ne. corner of the map where a simple Bouguer correction was applied over the region where the deep-sea topography is not of a two-dimensional nature. Bouguer anomalies along the 1968 and 1969 NOAA lines south of i3°N were adapted from Laving (197 1); those on the islands are from Masson Smith & Andrew (1965). The dominant feature of the Bouguer anomaly map is the large linear minimum that follows the crest of the Barbados Ridge in the south and the Lesser Antilles Trench in the north. The amplitude of the anomaly decreases northward from about —300 mgal at Barbados (with respect to the value over the Atlantic) to —220 mgal ne. of Guadeloupe. In the south the axis of the anomaly passes east of Tobago and Trinidad and apparently turns southwest towards the southern part of Trinidad. The Bouguer anomalies vary between 150 and 180 mgal over the islands of the Lesser Antilles. Between Guadelope and St. Lucia there are three higher maxima east of the islands. Over the Aves Ridge there is a relative Bouguer anomaly low with respect to both the Venezuela Basin and the Atlantic Ocean. Superimposed on this low are local maxima which are probably caused by struc- tures in the sediment/basement interface. Both the Aves Ridge minimum and the higher values over the Grenada Trough probably reflect the relative position of the Moho under these features. 744 745 746 Geophysics of E. Caribbean 319 4. Total field magnetic anomaly — Map 4 The total field magnetic anomalies of this map represent values with respect to the I.G.R.F. (LA.G.A. 1969). For this reason the values are generally negative with an average background of —230 gamma. Only the area north of i2°54'n. is shown because the line spacing south of this latitude is not sufficiently dense for contouring. In the area between latitudes i2°54' and I3°54'n. the data were corrected for ionospheric effects using magnetogram data from geomagnetic observatories at Paramaribo, Surinam and San Juan, Puerto Rico. The region of the Lesser Antilles Island Arc near the islands is characterized by numerous short-wavelength magnetic anomalies, some reaching an amplitude in excess of 600 gamma. These in most instances could not be contoured due to the line spacing of the surveys which much exceeds the wavelength of these anomalies. A series of somewhat smaller amplitude anomalies characterizes the platform e. of the islands, roughly following the same trend as the Bouguer anomalies. The western half of the Barbados Ridge Complex exhibits magnetic anomalies of low amplitude. Over the eastern half, however, and over the adjacent Atlantic Basin, there are prominent E.-w. trending anomalies at I5°n., and similar anomalies of somewhat reduced amplitude at I4°n. and i3°n. These are related to the e.-w. ridges and troughs in the igneous basement of the Atlantic (Peter et al. 1973a, 1973b; Peter & Westbrook 1974). Over the Grenada Trough the magnetic field is relatively smooth, but over the Aves Ridge there are numerous anomalies of a few hundred gamma amplitude. These anomalies, as well as those over the Lesser Antilles Island Arc, indicate the presence of mafic igneous rocks near the sea floor. 5. Discussion The most westerly unit of the map area, the Aves Ridge, is probably an ancient island arc which may have been active during Cretaceous time. The magnetic anomaly map of the Aves Ridge exhibits numerous short wavelength anomalies which may be expected over the lava flows and volcanic plugs characteristic of an island arc. Rock dredged from the ridge have calc-alkaline affinities and ages of 80 to 57 m.y. (Fox et al. 1971, Nagle 1971). Bouguer anomalies indicate the presence of an underlying mass deficiency probably caused by depression of the Moho beneath the Ridge. The calc-alkaline volcanism of the Lesser Antilles has existed since the upper Eocene. A general summary of the geology of the islands of the Lesser Antilles has been presented by Martin-Kaye (1969) and there has been much work performed on the petrology and source of magmas feeding the volcanoes (e.g. Lewis 1971; Robson & Tomblin 1966; Sigurdsson et al. 1973). The positive Bouguer anomalies and the short wavelength, high amplitude magnetic anomalies near the islands indicate the presence of near surface, high density igneous rocks. The Jurassic spilites and keratophyres on La Desirade (Fink et al. 1972) on the eastern edge of the island platform are in striking contrast to the 747 320 P. Kearey, G. Peter & G. K. Westbrook Eocene and younger calc-alkaline volcanic rocks of the rest of the Lesser Antilles. Bouguer anomalies suggest that buried structures containing rocks similar to those on La Desirade may continue south from La Desirade as far as St. Lucia. The negative Bouguer anomaly zone east of the Lesser Antilles over this area is caused by depression of the igneous basement which is filled with sediments reaching a maximum depth of 20 km below Barbados (Westbrook et al. 1973). The eastern limit of the seismicity associated with the Benioff zone which dips westward beneath the Lesser Antilles (Molnar & Sykes 1969, Tomblin 1972) coincides approximately with the axis of the negative Bouguer anomaly. Con- sequently one may suggest that subduction of the igneous crust of the Atlantic Plate beneath the Caribbean Plate, as required by plate tectonics theory, begins at the axis of the anomaly. Much of the rough topography of the eastern part of the Barbados Ridge Complex seems to be a direct result of deformation. Seismic reflection profiles provide ample evidence of the deformation of the underlying sediments (e.g. Chase & Bunce 1969, Peter et al. 1974). Both the topography and gravity anoma- lies indicate a major discontinuity across the Barbados Ridge Complex at around I4°n. The e.-w. ridges and trough in the sediment pile which cause local offsets of the negative free air anomaly seem to be directly related to e.-w. trending features in the Atlantic ocean floor that are probably ancient transform faults (Peter & Westbrook 1974). The purpose of this paper has been to present new data in a region which is geologically complex and the object of much interest. Detailed interpretations of some of the data have been published or are in preparation (Kearey 1973, Westbrook 1973, Westbrook et al. 1973, Peter & Westbrook 1974, Kearey 1974). acknowledgements. We gratefully acknowledge the assistance of the captains and crews of H.M.S. HECLA and the NOAA ships DISCOVERER and RESEARCHER in the data collection. We thank M. H. P. Bott for his advice and encouragement. This work was supported in part by N.E.R.C. research grant GR/3/937 awarded to M. H. P. Bott and NSF-IDOE grant AG-253 awarded to G. Peter. 6. References Andrew, E. M., Masson Smith, D. & Robson, G. R. 1970. Gravity anomalies in the Lesser Antilles. Inst. Geol. Sci. Geophys. Paper 5, 21 pp. Bowin, C. O. 1972. Puerto Rico Trench gravity anomaly belt. Mem. geol. Soc. Amer. 132, 339-50. Bunce, E. T., Phillips, J. D., Chase, R. C. & Bowin, C. O. 1971 . The Lesser Antilles Arc and the eastern margin of the Caribbean Sea. In A. E. Maxwell (ed.), The Sea 4 (2), 359-85. Chase, R. L. & Bunce, E. T. 1969. Underthrusting of the eastern margin of the Antilles by the floor of the western North Atlantic Ocean, and the origin of the Barbados Ridge. J. geophys. Res. 74, 1413-20. Dorman, C. M., Bassinger, B. G., Bernard, E., Bush, S. A., Dewald, O. E., Capine, L. A., Lattimore, R. K. & Peter, G. 1973. Caribbean Atlantic Geotraverse, IDOE 1971, Report No. 3, Gravity. NOAA Tech. Rept EP-L 277-AOML1 1, 35 pp. Fink, L. K. Jr., Harper, C. T., Stipp, J. J. & Nagle, F. 1972. The tectonic significance of La Desirade. Possible relict sea-floor crust. In C. Petzall (ed.) Trans. Vlth Caribbean geol. Conf. Margarita, Venezuela 197 1 , 302. Fox, P. J., Schreiber, E. & Heezen, B. C. 1971 . The geology of the Caribbean crust: Tertiary sediments, granites and basic rocks from the Aves Ridge. Tectonophysics 12, 89-109. 748 Geophysics of E. Caribbean 321 Hess, H. H. 1933. Interpretation of geological and geophysical observations in the Navy-Princeton Gravity Expedition to the West Indies in 1932. Bull. geol. Soc. Amer. 49, 1938. Gravity anomalies and island arc structures with particular reference to the West Indies. Proc. Amer. phil. Soc. 79, 71-96. Kearey, P. 1973. Crustal structure of the eastern Caribbean in the region of the Lesser Antilles and Aves Ridge. Ph.D. thesis, Univ. Durham (unpubl.). I974- Gravity and seismic reflection investigations into the crustal structure of the Aves Ridge, eastern Caribbean. Geophys. J. Roy. astr. Soc. 38, 435-48. Laving, G. J. 1971. Automatic methods for interpretation of gravity and magnetic field anomalies and their application to marine geophysical surveys. Ph.D. thesis, Univ. Durham (unpubl.). Lewis, J. F. 1 97 1 . Composition, origin and differentiation of basaltic magmas in the Lesser Antilles. Mem. geol. Soc. Amer. 130, 159-79. Martin-Kaye, P. H. A. 1969. A summary of the geology of the Lesser Antilles. Overseas Geology and Mineral Resources Inst. Geol. Sci. 10, 172-206. Masson Smith, D. J. & Andrew, E. M. 1965. Gravity and magnetic measurements in the Lesser Antilles. Overseas Geological Surveys (Geophysical Division) Preliminary Report and illustrations 16 pp. Matthews, D. J. 1939. Tables of the velocity of sound in pure water and sea water for use in echo sounding and sound ranging. Admiralty Office, London. 52 pp. Molnar, P. & Sykes, L. R. 1969. Tectonics of the Caribbean and the Middle America regions from focal mechanisms and seismicity. Bull. geol. Soc. Amer. 80, 1639-84. Peter, G., Merrill, G. & Bush, S. A. 1973. Caribbean Atlantic Geotraverse, NOAA-IDOE 1 97 1, Report No. i, Project Introduction — Bathymetry. NO A A Tech. Rept. ERL 293- AOML 13, 29 pp. , Dewald, O. E. & Bassinger, B. G. 1973. Caribbean Atlantic Geotraverse, NOAA-IDOE 1971, Report No. 2, Magnetic Data. NOAA Tech. Rept. ERL 288-AOML 12, 19 pp. , Schubert, C. & Westbrook, G. K. 1974. NOAA-IDOE Caribbean Atlantic Geotraverse. Geotimes 19, (8), 12-5. & Westbrook, G. K. 1974. Tectonics of the Barbados Ridge and the adjacent Atlantic Basin. Trans. Am. Geophys. Union 55, 284. Robson, G. R. & Tomblin, J. F. 1966. Catalogue of the active volcanoes of the world including solfatara fields. Part XX West Indies. International Association of Volcanology, Rome. Sigurdsson, H., Tomblin, J. F., Brown, G. M., Holland, J. G. & Arculus, R.J. 1973. Strongly undersaturated magmas in the Lesser Antilles. Earth planet. Sci. Letters 18, 285-95. Tomblin, J. F. 1972. Seismicity and plate tectonics of the eastern Caribbean. In C. Petzall (ed.) Trans. Vlth Caribbean Geol. Conf. Margarita, Venezuela 1971, 277-82. Westbrook, G. K. 1973. Crust and upper mantle structure in the region of Barbados and the Lesser Antilles. Ph.D. thesis, Univ. Durham (unpubl.). , Bott, M. H. P. & Peacock, J. H. 1973. The Lesser Antilles subduction zone in the region of Barbados. Nature Phys. Sci. 244, 18-20. Received 13 August 1974; revised typescript received 23 October 1974. Philip Kearey, Department of Geological Sciences, The University, South Road, Durham DHi 3LE. George Peter, N.O.A.A. Marine Geology and Geophysics Laboratory, Atlantic Oceanographic and Meteorological Laboratories, 15 Ricken- backer Causeway, Virginia Key, Miami, Florida 33149. Graham Karel Westbrook, Department of Geology, The University, Keele, Staffordshire ST5 5BG. 749 37 Reprinted from: IX International Congress of Sedi- mentology, Nice, France, Theme 6, 77-80. Sedimentary Processes in Submarine Canyons off Northeastern United States by George K. Keller Geological Oceanographer , National Oceanic and Atmospheric Administration, Atlantic Oceanographic and Meteorological Laboratories, Miami, Florida, USA. Submarine canyons off the northeast United States are characteristic- ally found to head at the shelf break, extending down to the upper rise and the abyssal plain in a few cases. Only in the case of the Hudson canyon is there visible surface expression of the drowned channel extending to the present Hudson river channel. The Hudson shelf valley and canyon system appears to tell the story for the origin of a number of the canyons bordering the northeast United States. Seismic reflection profiling has revealed in most cases buried channels extending from the vicinity of present canyon heads shoreward. Most of the canyons cannot be identified with earlier or present rivers as can the Hudson, but it is apparent from the reflection profiles that they were part of drainage systems extending across various parts of the continental shelf. During the Pleistocene and possibly earlier as well, the canyons and their drainage systems served as distinct channels for the transport of coarse-grained sediments from the shelf to the rise and abyssal plain. Horn and others (1971) , in their study of sediment cores from the western margin of the Atlantic basin, clearly showed the effectiveness of the canyons in channeling coarse sediments onto the abyssal plain (Fig. 1). Today, with the higher sea level, the canyons have lost their shelf valleys due to infilling and appear topographically only as canyons in the continental slope, heading on the order of 160 km off the coast. The rise in sea level has also served to eliminate the fluvial transport of coarse material across the shelf to the canyons. Horn, D. R. , Ewing, M. , Horn, B. M. , and Delach, M. N. , 1971. Turbidites of the Hatteras and Sohm abyssal plains, western North Atlantic. Marine Geology, 11: 2S7-3 23. 750 Recent studies in canyons off northeastern United States reveal that currents within the canyons are far from unidirectional, but possess a strong up- and down-canyon component as well as distinct across-canyon flow in some cases. This reversal in flow direction is a distinctive character- istic of both shelf and canyon dynamics. It is readily apparent that the tidal component is the most influencial factor on this flow regime (Fig. 2). Internal waves appear to play some role in canyon dynamics (Keller and 2 3 others, 1973 ; Shepard and others, 1974°), but an understanding of their influence has not yet been developed. For the most part, net transport within the canyons appears to be down- canyon based on both current measurements and the observation of such bottom features as ripples and leeside tailings. In Hudson and Hydrographer canyons bottom transport downcanyon is quite distinct (Fig. 3). It is also apparent from current observations that current shears are present at dif- ferent levels in canyons and flow within a canyon is far from uniform from its lip to the floor. In the Wilmington canyon area, it appears that the westerly set of shelf currents and , at greater depths , the Western Boundary Undercurrent strongly influence the current dynamics within the canyon. Current observations from depths of 700 and 900 m show the typical up- and down-canyon flow, but calculation of net transport over a 19-day period revealed a flow across (to the west) and up the canyon. Conflicting evi- dence , such as the orientation of ripple marks and leeside tailings , point to downcanyon transport of bedload material in the canyon head, but further downcanyon these same features indicate across-canyon flow (Fenner and others, 19714). 2Keller, G. H. , Lambert, D. , Rowe, G. , and Staresinic, N., 1973. Bottom currents in the Hudson canyon. Science, 180: 181-183. 3 Shepard, F. P., Marshall, N. F. , and McLoughlin, P. A., 1974. "Internal waves" advancing along submarine canyons. Science, 183: 195-198. 4Fenner, P., Kelling, G. , and Stanley, D. J., 1971. Bottom currents in Wilmington submarine canyon. Nature Fhys. Science, 229: 52-54. 751 In this case where bottom feature orientations contradict the measured flow regime, the long-term transport regime is undoubtedly that reflected in the bottom features and not that of the 19-day observations. North of the Hudson canyon there is a distinct difference in the sedi- ment being carried through the canyons. To the north, off Georges Bank the canyons in general appear to be quite active. For the most part, these canyons are areas of active coarse-grained sediment erosion and trans- port downslope. Relatively little fine-grained material is present in the upper parts of the canyons, with rock outcrops the rule rather than the exception. Hydrographer canyon appears to be extremely active with bottom currents of 55 cm/ sec having been measured and the migration of sand ripples down-canyon recorded photographically. In Hudson canyon and those to the south (Wilmington, Baltimore, and Norfolk) the overall sedimentarv regime is primarily one of fine-grained sediment (silt, clayey silt, and silty clay) with the sands and gravels being confined to the "head" portion of the canyons. These "southern" canyons take on the appearance of being relatively inactive, although bottom velocities of 25 to 35 cm/sec are common. Current velocities in the Georges Bank canyons do appear to be higher than those recorded in the "southern" canyons, but velocities in themselves do not seem to explain the differences in the sedimentary regimes found associated with these various canyons. Sediment source material undoubtedly plays a major role in regard to sediment transport through the canyons. In the case of Georges Bank canyons, the source material is primarily coarse-grained, reworked glacial deposits, whereas to the south, less and less of this coarse material is available. Although current veloci- ties of 35 cm/ sec have been recorded in Hudson canyon, there is no indication of coarse-grained material being transported downcanyon. Based on geochem- ical studies of the carbohydrate and total organic carbon content of the sedi- ments, however, there is evidence that the Hudson canyon does serve as a 752 conduit far the movement of fine silt and clay size material out onto the abyssal plain. Today the submarine canyons off the northeast United States appear to transport little, if any, coarse-grain sediment from the shelf to the abyssal plain. Exceptions to this may be same of the canyons off Georges Bank, where strong currents (55 cm/ sec) and sand migration downcanyon have been observed. Canyons are, however, zones of considerable current flow with common velocities of 15 to 25 cm/sec and frequently as high as 30 to 35 cm/ sec. The northeast canyons are certainly active, but appear now to be primarily the conduits for the transport of fine-grained, rather than coarse-grained sediment to the deep sea. 753 Fig. 1. Areas of coarse-grained turbidite concentrations-, attribuxad to transport through nearby canyons. After Horn and others (1971). 754 HUDSON CANYON (BM) o Q APRIL 19 (0830) Fig. 2. Characteristic up- and down-canyon flow with the tide superimposed on the current data. (BM) bottom meter. 755 DOWNCANYON (SOUTH) UPCANYON (NORTH) 5 ^ Fig. 3. Vector analysis plot showing the net transport over 390 hours through Hudson canyon (depth 230 m) (TM) top meter- 756 38 Reprinted from: Marine Geo technology 1, No. 1, 159-161. Book Review The Ocean Basins and Margins— Volume 2: The North Atlantic by A. E. M. Nairn and F. G. Stehli (Editors). Plenum Publishing, New York, 1974. 598 pp. $38.00. Reviewed by George H. Keller* This volume consists of fourteen chapters or papers by various authors who have compiled an up-to-date summary of the geological framework of the North Atlantic basin and its margin. It is organized so that the reader is led through a series of geological discussions starting with the Bahamas and moving north along the east coast of North America, east to the margins of Greenland and Norway, then south through the British Isles to Brittany. Although Spain and Portugal are not discussed, a substantial chapter is devoted to West Africa. The reader is thus given a fish-eye view of the geology ringing the North Atlantic. The geology and tectonism of the major oceanic islands are discussed in light of their relationship to the basin's development. Tectonic events influencing the margin along with their ages are adeptly summarized in one of the later chapters. A regional scan of the geophysical data available from the basin itself is presented in the final chapter in an effect to define the history of the North Atlantic. The editors, A. E. M. Nairn and F. G. Stehli, set the stage in the first chapter by briefly summarizing the model for tectonic evolution of the North Atlantic basin. Stehli in the following chapter draws heavily from the literature to discuss the geology of the continental shelf from North Carolina to Florida, the geology of Florida, the northern Antilles, and the Bahamas. Much of the discussion centers on the Bahamas and the hypotheses as to the origin of this platform. Mention is made of the age and distribution of surficial sedimentary sequences as well as outlining the geological section underlying the Blake Plateau. A good synthesis of the stratigraphy, sediments, physiography, and structure of the continental margin from Florida to Newfoundland, including a brief discussion of the magnetic quiet zone and the overall continent-to-ocean transition zone of ♦Atlantic Oceanographic & Meteorological Laboratories, Miami, Florida. Marine Geotechnology, Volume 1, Number 2 Copyright © 1975 Crane, Russak & Company, Inc. 159 757 160 GEORGE H. KELLER the western Atlantic is provided by M. J. Keen. The northeastern terminus of the Appalachian orogeny is discussed by H. Williams, M. J. Kennedy, and E. R. W. Neal, and is basically a synthesis and interpretation of Newfoundland geology and tectonism which pretty well defines the significant elements of the Appa- lachian system. T. Birkelund, D. Bridgewater, A. K. Higgins, and K. Perch-Nielsen present a summary of the geological and structural history of the Greenland margin. Particular attention is given to the hypothesis that Greenland represents a fragment of an earlier North Atlantic continental mass as well as pointing out the pronounced matching of stratigraphic units, both in type and age with those of Europe. A thorough summary of the geology of the Caledonides sequence of Scandinavia is given by R. Nicholson, with a discussion of the respective units and their tectonic setting. Some mention is made of the comparison of the Caledonides with those of the British Isles as well as commenting on the paleogeographic reconstruction of the Caledonides. Further, J. F. Dewey con- tinues the discussion of the Caledonides by summarizing the detailed geological data from the British Isles and then goes on to discuss the entire orogenic belt, its extent and geological variations in light of various global tectonic models. The geological and tectonic histories of the western approaches (Southern Ireland, Great Britain, and Brittany) are summarized by T. R. Owen. Not only is the terrestrial geology and structure discussed but the adjacent continental shelf geology as well. A good synthesis of the paleogeographic tectonism from the Precambrian to Quaternary is also presented. M. Vigneaux reviews the geological framework of the Bay of Biscay, discusses the available geophysical data, and addresses certain aspects of the geomor- phology and sedimentation pattern of littoral, continental shelf, and bathyal regions. It is a fine discussion of the sedimentary history as well as touching upon the origin and age for the opening of the Bay. Skipping to the south, an excellent survey of the geological history and stratigraphic sequences of West Africa from the Precambrian to the Holocene is presented by W. Dillon and J. M. A. Sougy. Discussion centers on specific geological aspects and their interre- lationships in light of the formation of the North Atlantic. The authors also briefly mention the geology and volcanic history of the Canary and Cape Verde Islands. A good summary of Cenozoic to Recent volcanism in and along the northern margins of the Atlantic has been compiled by A. Noe-Nygaard. The discussion is primarily concerned with the terrestrial geology rather than that of the seafloor. A brief summary of volcanism in relation to the theory of seafloor spreading is given. A fine discussion of the geological and tectonic framework of the seafloor in and around the Azores Islands is presented by W. L. Ridley, N. D. Watkins, and D. J. MacFarlane. They discuss the volcanism of the Islands as well 758 BOOK REVIEW 161 as giving considerable attention to the geophysical observations made in the vicinity of these mid-ocean islands. Speculation on the cause of changes in spreading direction and the overall tectonic model for this juncture of the seafloor are set forth. A well-organized overview for North America, Greenland, Eurasia, and Iberia of tectonic patterns and events that have influenced the margin of the North Atlantic basin as well as the later events involved in the breakup and formation of the basin is presented by F. J. Fitch and others. Also included is a discussion on the verification of events by means of radiometric-age dates. The final chapter by H. C. Noltimier is a summary of the geological and tectonic frame- work of the North Atlantic basin itself. Various geophysical data such as seismic refraction and reflection, magnetics, gravity, and heatflow are reviewed in light of plate tectonics and formation of the North Atlantic. Overall, this volume serves as an excellent reference to the student of plate tectonics and seafloor spreading. The authors have drawn heavily on the literature thus providing an excellent source book of previous studies. 759 39 Reprinted from: No. 1, 292-294. Journal of Sedimentary Petrology 45, CARIACO TRENCH-SEDIMENT GEOTECHNICAL PROPERTIES1 G. H. KELLER and L. L. MINTER' NOAA, Atlantic Oceanographic & Meteorological Labs 15 Rickenbacker Causeway, Miami, Florida 33149 Abstract: A study of the mass physical properties of a sediment core from the anaerobic Cariaco Trench shows there to be a noticeable difference in these properties relative to those of the eastern Caribbean and Atlantic basins, and an even greater difference from those of the anoxic Black Sea. The variations in physical properties appear to reflect the respective concentrations of organic carbon in the four areas compared. The Cariaco Trench, which lies parallel to and just off the northern coast of Venezuela, is one of the few deep-sea areas exhibiting' an- aerobic conditions. Circulation below 150 m (sill depth) is limited, and from about 375 m to the bottom at 1,400 m an anoxic environment exists (Richards and Vaccaro, 1956). Earlier investigators have discussed various oceanographic conditions (Richards, 1960; Richards and Vaccaro, 1956; Curl, 1960; and Fanning and Pilson, 1972) in the trench as well as its geological setting (Ball, et al., 1971; and Peter, 1972). Sediment cores have been col- lected from different parts of the trench and discussed to varying degrees by Heezen, et al., (1959), Athearn (1965), and Lidz, et al., (1969). These studies have dealt mainly with the discussion of sedimentary features, limited micropaleontological and chemical analyses, and geochronology as related to the occurrence of stagnant conditions in the trench. The trench, in addition to its physical con- figuration and anaerobic environment, is char- acterized by a very high rate of productivity (2.30 gC/m2/day) in the overlying waters (Curl, 1960) and a rather high sedimentation rate (50 cm/1000 yr) (Heezen, et al., 1959). During a cruise of the NOAA Ship Dis- coverer in 1972 a large diameter (8.9 cm) gravity core, 106 cm long, was taken from a depth of 1302 m in the eastern part of the trench (10°29.8'N, 64°29.2'W). This relatively undis- turbed sample was collected in order to examine in detail the geotechnical properties of the anaerobic trench deposits. •Manuscript received June 28, 1974; revised Au- gust 29, 1974. 'Present address: Dept. of Oceanography, Texas A&M Univ., College Station, Texas. The cored interval consists primarily of a silty clay with a considerable amount of fine layering. Some laminae are made up of dis- tinctly coarser material such as angular quartz grains, Foraminifera, ostracods, and pteropods. Others, however, are not texturally different but are distinguished by color changes in the silty clay. Radiographs of the silty-clay intervals do not distinguish the laminae indicating a uniform density. These coloration laminae may result from varying concentrations of organic carbon which in turn probably reflect the high produc- tivity levels of the overlying waters. Athearn (1965) suggested that these laminated sediments be attributed to the alternation of wet and dry seasons. He also speculated that some of the coarser layers containing Foraminifera and pteropods may have been due to "mass kills" on a local basis. In our examination of these layers we found both benthic and pelagic Foraminifera as well as angular quartz grains. The presence of shallow water benthic forms and a high concentration of quartz suggest transport of these coarser materials into the trench from the adjacent shelf. Turbidites ap- pear to be relatively rare (Athearn, 1965; Lidz, et al., 1969) and large turbidity currents in themselves do not seem to be a major means of transport into the trench. It would appear that runoff, transport of shelf sediments, and local slumping play the major role in the overall in- filling process. The one-meter interval sampled consists pri- marily of clay (54%) with lesser amounts of silt (24%) and sand (22%) (Fig. 1). Dis- regarding the three major sandy layers, the overall texture is more that of a silty clay (clay 63%, silt 24%, and sand 13%). Water content is given in percent dry weight as normally de- 760 CARIACO TRENCH-SEDIMENT GEOTECHNICAL PROPERTIES 293 WATER CONTENT UNIT WT SPECIFIC TOTAL ORGANIC SHEAR STRENGTH SAND, SILT, CLAY (%) (%) DRY WEIGHT (g/cm') GRAVITY POROSITY (%) CARBON (%) (g/cm*) ( 0- 20- ) 20 40 60 80 100 80 120 160 1.4 1.6 2.6 2.7 60 70 80 3.0 4.0 5.0 10 30 50 \ X — — —• PL LL V M t < I * 3/ I 1 • > i ♦ I I - 1 \ PL LL % : I i i V i~ r jT ** \ s / i i l *"7 r 1 > < / \ V \ 60- i ' / l / J ( \ \ - ; < I I > / k I i I I PL LL .' » > ; \ / 80- \ \ __ ^> i. < ^ i _ V \ •c V> / \ I \ <^ x V ,> 100- r r 1 1 • \ k - Fig. 1. — Physical and chemical properties for Cariaco Trench core. termined in engineering practice. Other proper- ties such as unit weight, porosity, shear strength, and the plastic and liquid limits to be discussed later, are adequately explained in any soil me- chanics text and will not be defined here. Water content varies from about 145-170% in the silty clay to 60-80% in the coarser sandy layers (Fig. 1). Liquid and plastic limits define the water content at the stages where a sediment flows when shocked under laboratory testing and when it loses its plasticity respectively. These limits for the silty clay intervals are remarkably similar within the one meter sampled, with a plastic limit of 47 and liquid limits of 114-117. Using Casagrande's (1948) classification, which is based on the liquid and plastic limits, these sediments are classed as "organic clays of medium to high plasticity." Few deep-sea sedi- ments fall into this class, but more commonly are classified as "inorganic clays of high plas- ticity," (Richards, 1962). Being classed as organic clays is, however, not surprising owing to the high concentrations of organic material common to the trench deposits. Unit weight or wet bulk density varies rela- tively little throughout the sampled interval ex- cept where layers of coarser material occur. The higher unit weights (1.51-1.65 g/cm3) of the sandy sequences are quite distinct from the 1.30-1.37 g/cm3 values found for the silty clay. Such low densities are attributed to the min- eralogy of the fines and the high organic content found associated with these finer sediments. Grain specific gravity varies considerably even within the silty clay sequences indicating that these intervals are not as uniform in com- position as some of the other data imply. Values range from 2.50-2.72, tending to average about 2.63. Porosity is strongly influenced by grain size as shown in Figure 1. Porosities of 60-70% characterize the coarser sediments whereas much higher values of 77-82% are identified with the silty clays. Shear strengths as determined by the labora- tory vane-shear test, using a 2.50 cm X 1.25 cm vane at a rotation speed of 70 degrees per min- ute, range from 39-50 g/cm2 for the silty clay intervals. In comparing natural versus remolded shear strength, sensitivities of 3^1.5 are found, indicating a 50-75% loss of strength due to remolding (Richards, 1962). Sensitivities such as these, in the case of the Cariaco Trench, are probably the result of the high organic content. A similar relationship has been reported for the organic-rich deposits of the Black Sea (Keller, 1974). Organic carbon, as might be expected, is ab- normally high in the sediments of the Cariaco Trench. Total organic carbon was determined by the combustion method (Gustin and Tefft, 1966) and showed values ranging from 4.17 to 5.12% except in a layer of coarse material where a value of 1.63% was recorded (Fig. 1). These values are slightly higher than those re- ported by Athearn (1965) for the same general area of the trench, but are well within the limits 761 294 G. H. KELLER AND L. L. MINTER of 1-6% reported by Lidz and others (1969) for the trench as a whole. Assuming grain size to be similar, it might be expected that sediments from anoxic and aerobic environments would exhibit physical properties quite distinct from one another. Such a distinction has been pointed out between open-ocean deposits and those of the Black Sea (Keller, 1974). Although the Cariaco Trench sediments are in an anoxic environment, they differ considerably from the Black Sea deposits yet contrast relatively less with the aerobic de- posits of the Caribbean and Atlantic basins. Water contents and porosities in the trench sediments are only slightly higher than those reported for the Atlantic (Keller and Bennett, 1970), quite similar to those in the Tobago Trough in the eastern Caribbean (Keller, et al., 1972), but much lower than the Black Sea de- posits. Unit weight and grain specific gravity values on the other hand are lower in the trench than is commonly found in the Atlantic or Caribbean sediments, but considerably higher than those reported from the Black Sea. In regard to mass physical properties, Cariaco Trench sediments, although from an anoxic environment, are more closely related to the aerobic Atlantic and Caribbean deposits than to those of the Black Sea. These relationships seem to best correlate with organic carbon content. Cariaco Trench deposits are much richer in organic carbon (4-5%) than Atlantic and Caribbean sediments (0.25-2%) (Trask, 1939; Froelich et al., 1971) ; yet are considerably below the very high con- centrations (5-20%) reported for the Black Sea (Ross et al., 1970). The difference in organic carbon content between the Cariaco Trench and the Atlantic and Caribbean deposits is much less than between the Trench and the Black Sea sediments. The Black Sea is basically an en- closed sea with unique circulation characteristics both in its upper and lower levels whereas, the Cariaco Trench has relatively good circulation in its upper waters with little or no influx of fresh water. The different circulation char- acteristics are undoubtedly a major cause for the contrasting concentrations of organic carbon in the two anoxic basins and in turn the differ- ing sediment physical properties. REFERENCES Athearn, W. D., 1965, Sediment cores from the Cariaco Trench, Venezuela : Woods Hole Ocean. Inst. Tech. Rep. 65-37, 31 p. Ball, M. M., G. C. A. Harrison, P. R. Supko, W. Bock, and N. J. Maloney, 1971, Marine geo- physical measurements on the southern boundary of the Caribbean Sea: in Donnelly, T. W. (ed.), Caribbean Geophysical, Tectonic, and Petrologic Studies, Geol. Soc. America Mem. 130, p. 1-33. Casagrande, A., 1948, Classification and identifica- tion of soils : Am. Soc. Civil Engineers Trans., v. 113, p. 901-931. Curl, H., Jr., 1960, Primary production measure- ments in the north coastal waters of South America: Deep-Sea Res., v. 7, p. 183-189. Fanning, K. A., and M. E. Q. Pilson, 1972, A model for the anoxic zone of the Cariaco Trench: Deep-Sea Res., v. 19, p. 847-863. Froelich, P., B. Golden, and O. H. Pilkey, 1971, Organic carbon in sediments of the North Caro- lina continental rise: Southeastern Geology, v. 13, p. 91-97. Gustin, G. M., and M. L. Tefft, 1966, Improved accuracy of rapid micro carbon and hydrogen method by modified combustion absorption tech- nique : Microchemical Jour., v. 10, p. 236-243. Heezen, B. C, R. J. Menzies, W. S. Broecher, and M. Ewing, 1959, Date of stagnation of the Cariaco Trench, southeast Caribbean (Abs.) : Geol. Soc. America Bull., v. 69, p. 1579. Keller, G. H., and R. H. Bennett, 1970, Varia- tions in the mass physical properties of selected submarine sediments : Marine Geology, v. 9, p.- 215-223. , D. N. Lambert, R. H. Bennett, and J. B. Rucker, 1972, Mass physical properties of Tobago Trough sediments : VI Conf . Geologica del Caribe, Margarita, Venezuela, Mem., p. 405- 408. -, 1974, Mass physical properties of some western Black Sea sediments : p. 332-337 ; in Degens, E. T., and D. Ross (eds.), Black Sea — Geology, Chemistry, Biology : Am. Assoc. Pe- troleum Geologists, Mem. 20, 633 p. Lidz, L., W. B. Charm, M. M. Ball, and S. Valdes, 1969, Marine basins of the coast of Venezuela : Bull. Marine Science, v. 19, no. 1, p. 1-17. Peter, G., 1972, Geology and geophysics of the Vene- zuelan continental margin between Blanquilla and Orchilla Islands: N.O.A.A. Tech. Rept. ERL 226-AOML 6, 82 p. Richards, A. F., 1962, Investigations of deep-sea sediment cores, II. Mass physical properties : U. S. Hydrographic Office Tech. Rept. 106, 146 p. Richards, F. A., and R. F. Vaccaro, 1956, The Cariaco Trench, an anaerobic basin in the Carib- bean Sea: Deep-Sea Res., v. 3, p. 214-228. , 1960, Some chemical and hydrographic ob- servations along the north coast of South Amer- ican I. Cabo Tres Puntas to Curacao, including the Cariaco Trench and the Gulf of Cariaco : Deep-Sea Res, v. 7, p. 163-182. Ross, D. A, E. T. Degens, and J. Macllvaine, 1970, Black Sea: recent sedimentary history: Science, v. 170, p. 163-165. Trask, P. D, 1939, Organic content of recent marine sediments: p. 428-453, in Trask, P. D. (ed.), Recent Marine Sediments, a symposium : Soc. Econ. Paleontologists and Mineralogists, Spec. Pub. 4, 736 p. 762 40 Reprinted from: Canadian Journal of Earth Science 12, No. 4, 703-710. Near-bottom Currents in the Mid- Atlantic Ridge Rift Valley1 George H. Keller Atlantic Oceanographic and Meteorological Laboratories, 15 Rickenbacker Causeway, Miami, Florida 33149 Susan H. Anderson Woods Hole Oceanographic Institution, Woods Hole, Massachusetts 02543 AND J. William Lavelle Atlantic Oceanographic and Meteorological Laboratories, 15 Rickenbacker Causeway, Miami, Florida 33149 Received September 18, 1974 Revision accepted for publication December 11, 1974 Near-bottom current measurements during a 46 day period (October to December 1973) in the rift valley of the Mid- Atlantic Ridge just south of the Azores reveal a mean flow at two stations of 2.6 and 8.2 cm/s, predominantly trending parallel to the northeast orientation of the valley. Maximum velocities recorded at the sites were 14.4 and 24.2 cm/s. Semi-diurnal tidal currents appear to strongly influence the flow within the valley causing current reversals, except in areas where local topography and constrictions in the valley apparently result in a unidirectional flow to the northeast. Les mesures de courant de fond durant une periode de 46 jours (octobre a decembre 1973) dans la vallee du rift de la crete Mid-Atlantique au sud des Azores revelent une coulee moyenne aux deux stations de 2.6 et 8.2 cm/s avec predominance d'une tendance parallele a l'orientation nord-est de la vallee. Les vitesses maximums enregistrees a ces endroits atteignaient 14.4 et 24.2 cm/s. Les courants de maree semi-diurne semblent influencer grandement la coulee dans la vallee causant ainsi des renversements de courant, excepte dans les regions oil la topographie locale et les ressercements dans la vallee resultent apparemment en une coulee unidirectionnelle vers le nord-est. [Traduit par le journal] Can. J. Earth Sci.. 12, 703-710 ( 1975) 763 704 CAN. J. EARTH SCI. VOL. 12, 1975 Introduction The Mid-Atlantic Ridge is a major topographic feature extending the length of the Atlantic Basin which, in many places, serves as a barrier to the circulation of bottom water from the eastern to the western Atlantic (Wiist 1933). Offsetting fracture zones running perpendicular to the ridge often provide controlling sills for bottom-water movement across the ridge system (Garner 1972). Independent of these large-scale water move- ments, considerable local variability of current dynamics may be expected due to extreme local variations in relief that characterize the ridge and axial rift valley. Through the French American Mid-Ocean Undersea Study (FAMOUS), a series of submer- sible dives were planned in the rift valley and fracture zones of the Mid-Atlantic Ridge between 36°30'N and 37°N. During preliminary dives made by the ARCHIMEDE in 1973, the French believed that they had encountered currents averaging 39 cm/s (0.75 knots) on four out of seven dives (J. Francheteau, personal communi- cation 1973). These relatively high velocities oc- curred in several places within the dive area but were not consistently encountered in areas where the submersible dived more than once. Further- more, the estimates were not consistent with five short-period measurements made in the same area by J. D. Phillips in 1972 (personal communi- cation). To obtain better data on the currents in the prime dive area prior to the major operations scheduled for July and August 1 974, three current meters were deployed in the rift valley from the USNS MIZAR and retrieved 46 days later (19 October to 5 December 1973) by the USNS LYNCH. Steep rift walls define the floor of the rift valley in the area chosen for the FAMOUS study. The valley floor is up to 5.8 km wide at its intersec- tions with the north and south bounding fracture zones, narrowing to 2.2 km midway between these fracture zones (Fig. 1). A series of ridges rising as much as 250 m above the surrounding floor and flanked by lateral troughs characterize the center of the valley floor (Needham and Francheteau 1974). The trend of the rift valley is to the north northeast, approximately 022°. To monitor the currents within this highly variable topography, two current meters were to be placed in lateral troughs and one on a central high in the narrower portion of the rift valley (Fig. 1). Although positioning was determined by satellite navigation and a bottom-mounted transponder beacon in the rift valley, it appears that the final positions of the moorings were slightly offset from the intended features by drift during descent to the sea floor, but were well within the rift valley. The moorings were at varying depths as desired, with the northern meter (A) being set at 36°49.2'N, 33°16.5'W at a depth of 2509 m (1394 fathoms), meter (B) at 36°48.43'N, 33°15.75'W at 2412 m (1340 fath- oms), and meter (C) at a depth of 235 1 m (1 306 fathoms) at 36°46.78'N, 33°16.46'W. Each current meter mooring had the same con- figuration: a model 102 Geodyne current meter (film recording) suspended 35 m above the sea floor on a dacron line and held in place by a 204 kg anchf r and a number of glass floats located a few metres above the current meter. The mooring was retrieved on demand by means of triggering an AMF model 324 acoustic trans- ponding release positioned on the line mid-way between the meter and the anchor. Current Observations The current meters recorded velocity and direc- tion every 10 min averaging over a period of about 50 s. An electronics failure resulted in unreliable velocity data from current meter (B), although good directional measurements were recorded. The records from current meters (A) and (C) along with the tidal height expected at Ponta Delgada, using the high and low water prediction tables (National Ocean Survey 1972), are plotted as a time series in Fig. 2. Current flow at all three stations was predominantly to the north and northeast, paralleling the general trend of the rift valley. At Station A, flow reversals of a semidiurnal nature occurred, presumably tidal, whereas at stations B and C a more or less con- stant northward and northeastward flow was recorded during the 46 day period. At station C velocity varied considerably during that time, but at both stations B and C the flow direction was always to the north of northeast. This would in- dicate that, at these particular sites, a north or northeasterly current is superimposed on the oscillatory flow regime so as to mask out the south and southwest flow recorded at station A. Flow direction can be more clearly visualized from progressive vector diagrams. The diagram for station A (Fig. 3) reveals the current fluctua- tions and reversals that are more typical of deep ocean currents. The more or less unidirectional 764 NOTES CONTOURS IN FATHOMS 705 Fig. 1 . Current meter stations and bathymetry of the rift valley of the Mid-Atlantic Ridge. Fracture zones A and B (F.Z. A; F.Z. B) delimit the northern and southern boundaries of the FAMOUS study area. Contours are in fathoms. flow observed at stations B and C (Fig. 4) is anomalous and would be suspect if recorded on the abyssal plain. Although an assumption, it seems reasonable that this unique flow pattern is the result of local topographic control on the flow regime. A similar unidirectional flow, but in a slightly different deep-ocean setting was re- ported from the Gibbs fracture zone by Garner (1972). In that case the unique topography of the fracture zone appeared to serve to funnel the flow from the east to the west side of the ridge. Current speeds at station A reached a maxi- mum of 14.4 cm/s (0.28 knots), but were com- monly much lower as shown by a mean speed of 2.6 cm/s (0.05 knots) for the 46 day period. At station C, current speeds were considerably greater, reaching a maximum of 24.2 cm/s (0.47 knots) and a mean for the recording period of 8.2 cm/s (0.16 knots). The stalling speeds of meters A and C were near 1 cm/s and 1.5 cm/s respectively, and the speeds reported above are biased low for that reason. However, only the average speed at station A where stalling occurred 35% of the time would be noticeably affected, 765 706 CAN. J. EARTH SCI. VOL. 12, 1975 r>z o -10 o V^^M^'MaM-"1 uyv _J 1 1_ ^pM^P^H^^^p •Vy- :< 5 U >w o -10 , *».,. -^ ,*i\AVr Aft ^.^^A^v^^v^vv^VirVVY^yVAiV^W^' '^waAtV^^Va,> v*k' ' -NT- _1 1 L_ zi - 20 ££§ '0 §§§ 0 o w -10 |^iMU^vWVlA/vW 5 _ 20 C< 5 0 ->LlJ o o - -10 w"^^W^MjWlAAlft*^^ Q 20 22 24 26 28 30 I OCT 3 5 7 9 II 13 15 17 19 21 23 25 27 29 I 3 NOV. DEC TIME (DAYS) Fig. 2. Current direction and velocity profiles for stations A and C along with the tidal height profile for the same period predicted at Ponta Delgada. and even then the bias amounts to at most 0.35 cm/s. The higher average speed at station C seems to reflect in part the narrowing of the rift floor at this location. The long period amplitude modulation evident in the tidal record over the 46 day period (tidal frequencies beating against each other) is reflected in the current measure- ments (Fig. 2). In November 1972, J. Phillips (personal com- munication) made a number of short duration (45-49 h) current measurements in the same general area of the rift valley. His observations indicated that current velocities were relatively low (ranging from 2.6 cm/s (0.05 knots) to 14.9 cm/s (0.29 knots)), just as we have reported. He too recorded basically a unidirectional flow to the north and northeast at a site adjacent (0.75 km southeast) to our station C. From this com- parison it might be cautiously stated that the current velocity and direction seems to be per- sistent, at least for this time of the year. Discussion In order to compare tidal components in records A and C, we have taken the raw data minus a 25 hour running average, and least squares fit that residual at five tidal frequencies. In Table 1, we have provided the tidal ellipse parameters for these components plus the velo- city means for the unfiltered data. From this can be noted: (1) the near linear nature of the semi- diurnal components and their orientation similar to the direction of the rift valley, (2) the relatively small (down by a factor of 10 from the principal semidiurnal contribution) and more circular diurnal tidal currents, (3) the high coherence of the semidiurnal tidal currents registered at the two meters, and (4) the extreme difference in the nonoscillatory velocity contribution which causes the vast dissimilarity in the progressive vector diagrams. In order to uncover additional periodic com- ponents in the current records, the data from 766 NOTES 707 S2 o o o CO — ! 2 cr o LJ o z: < i- co Q 0 5 DISTANCE EAST-WEST FROM ORIGIN (KM) Fig. 3. Progressive vector diagram for Station A. meters (A) and (C) were further subjected to spectral analysis. Figure 5 represents these results for meter (A). Because that meter is observed to have stalled in the first third of the record, the spectrum represents information contained in the last two-thirds of the record only. We have choosen a polarized spectral analysis as sug- gested by Gonella (1972) in order to sharpen a possible inertial frequency contribution. As is the convention, negative frequencies are clock- wise components while the positive frequencies represent an anticlockwise rotational flow. In order to remove scatter in the spectrum and at the same time resolve some of the low fre- quency contributions, we have 'decimated' (Blackman and Tukey 1958) the record by a factor of 16 and have used a Hanning spectral window on the resulting spectral averages. Be- cause we have forced the information into the lower end of the spectrum, we have run the risk of alaising higher frequencies. The smallest re- solvable period in this analysis is a fairly large 5.33 h. Mitigating information, we think, comes from a pilot analysis we performed using a Bartlett filter (Jenkins and Watts 1968) in which we found no peaks in the high frequency end of the spectrum and a rapid energy fall off. We have removed the average from the data before Fourier transforming. Turning to the results of the analysis, we find domination of the spectrum by the semi-diurnal tidal current which Fig. 2 also shows. The near equality of those spectral peaks is another demonstration of the near linearity of those tidal components, while the apparent disparity in the height of the small diurnal components again suggest circularity of the tidal ellipse. The inertial period at this latitude is 20.0 h, and there is indeed a clockwise peak at 20.0 h, albeit rela- tively small. Aside from the very low frequency peaks about which we make no comment, and the semidiurnal tides, the next most energetic contribution is the anticlockwise component near 6.2 h. This will be recognized as the product of nonlinear inter- actions of the semidiurnal components with themselves (first harmonics) or with each other. The small but recognizable peaks in both sides of the spectrum at 7.7 h, are believed to be the result of inertial/semidiurnal tidal interaction. It should be noted that it is the large amplitude of the semidiurnal tidal currents which cause many of these nonlinear interactions to rise above the background. Somewhat puzzling are the anticlockwise tidal side peaks at 11.1 and 14.6 h, which we do not believe are artifacts of the analysis. From the pilot spectral analysis for both meters where the high frequencies (> 1/5. 33h) were al- lowed, we found meter (A) observed a near — 5/3 power law behavior similar to that re- ported by Webster (1972) from the western Atlantic whereas meter (C) demonstrated a near — 1 frequency dependence. Neither (A) or (C) meters observed the preferred —2 dependence of Kolmogorov (1941). 767 708 CAN. J. EARTH SCI. VOL. 12, 1975 h- u E< E-- E< IS) o c 1.2 •— x 9. < E< I I + + + + t «") K, ^ (^ ^ 5" c c o a c = * o ■o *"* c u •?§ 3 8 E ii ■;- M C 5.2 3-5 - o 52 o E b c 5? (J = S.2 U ■ — "9.9 i- Q. .2,0 EH f c 4J E h c > >. 0 r o < > * u a 0 « * lH — 0 1- S+3 ■ago °-o c e u° o £ 0 h c' >. o£ ZtJ o 1- :\J a > 768 NOTES 709 250 -200 DISTANCE EAST-WEST FROM ORIGIN (KM) Fig. 4. Progressive vector diagram for Station C. For all intents and purposes the progressive vector plot for Station B is the same as that for Station C. Summary Current measurements made in the rift valley of the Mid- Atlantic Ridge indicate that the general flow regime is to the northeast, approxi- mating the topographic trend of the ridge itself. Reversals in flow direction occurred consistently along the western margin of the valley where the valley floor is 4-5 km wide. A more or less uni- directional flow to the northeast prevailed throughout the period of observation in the axial zone of the valley. This anomalous flow pattern is perhaps attributable to the influence of local topography in the vicinity of the current meters. Mean current velocities at station C in the nar- rower section of the rift-valley floor were also found to be higher by a factor of three than those at station A. The topography within the rift valley of the Mid-Atlantic Ridge does appear to 769 710 CAN. J. EARTH SCI. VOL. 12, 1975 iO! Q. 10 o 10* 10 l i i i r CONFIDENCE INTERVALS J(80% _ 95 % ♦ 180% 3 V o (E UJ z t-l <0 J o oi ■H o> (0 CO (E (E IE I I I K <0 •rt K CO CD I.I LJ .18 .16 .14 12 .10 .08 .06 .04 .02 .00 -.02-04 -.06 -.08 -.10 -.12 -.14 -.16 -.18 CPH Fig. 5. A rotary component spectrum for the currents observed at mooring A. The right hand side (negative frequencies) represents clockwise components, the left hand side (positive frequencies) represents the anticlockwise components. strongly influence the near-bottom current regime, although the currents seem to be of small amplitude. Acknowledgments We wish to thank the officers and crew of the USNS MIZAR and USNS LYNCH and par- ticularly George Lapiene, Atlantic Oceano- graphic and Meteorological Labs, Donald Koelsch, Woods Hole Oceanographic Institu- tion, James McGrath, Naval Research Lab, and Patrick Taylor, Naval Oceanographic Office for their assistance during the field phase of this study. We also acknowledge with many thanks the assistance of Thomas Clarke in processing the data and the critical review of our manuscript by N. P. Fofonoff. Blackman, R. B. and Tukey, J. W. 1958. The measure- ment of power spectra. Dover, New York, New York. Garner, D. M. 1972. Flow through the Charlie-Gibbs fracture zone, Mid-Atlantic Ridge. Can. J. Earth Sci. 9. pp. 116-121. Gonella.J. 1972. A rotary-component method for analyz- ing meteorological and oceanographic vector time series. Deep-Sea Res. 19, pp. 833-846. Kolmogorov, A. N. 1941. The local structure of turbu- lence in an incompressible fluid at very large reynolds numbers. Dokl. Akad. Nauk. SSSR 30. (4), pp. 299-303. Jenkins, G. M. and Watts, D. G. 1968. Spectral analysis and its applications. Holden-Day, San Francisco, California. National Ocean Survey. 1972. Tidal tables, 1973 (high and low water predictions); Europe and the West Coast of Africa. U.S. Dept. of Commerce, Rockville, Md. Needham, H. D. and Francheteau, J. 1974. Some characteristics of the rift valley in the Atlantic Ocean near 36°48' north. Earth Planet. Sci. Lett. 22, pp. 29^3. Webster, F. 1969. Turbulence spectra in the ocean. Deep- Sea Res., Supp. 16, pp. 357-368. Wust, G. 1933. Das Bodenwasser und die Gleiderun der Atlantischen Tiefsee, die Stratosphare des Atlan- tischen Ozeans. Dent. Atlantische Expedition "Meteor", 1925-1927, Wiss. Ergeb., 7(1). 770 Reprinted from: Journal of Geophysical Research 80 No. 15, 1953-1956. 41 Possible Bottom Current Response to Surface Winds in the Hudson Shelf Channel J. W. Lavelle, G. H. Keller, and T. L. Clarke NO A A Atlantic Oceanographic and Meteorological Laboratories, Miami, Florida 33149 Current measurements made in the Hudson Shelf Channel during the summer of 1973 show essentially channel axial bottom current even though the channel aspect ratio is small in the area of measurement Although the current record is of short duration, correlation of water movement with surface winds is suggested by the data. The sense of summertime nontidal bottom flow in the channel (up or down channel) would appear to be controlled by the surface wind direction (offshore or onshore). These results would suggest the likelihood of net down-channel flow during the summer months. Introduction The New York Bight, defined as that area of the continental shelf bounded by Long Island on the north and New Jersey on the east, is dominated topographically in the near-shore region by the Hudson Shelf Channel. The channel itself is the drowned ancestral Hudson River Channel and leads southward for some 24 km from the present bight apex and then southeastward to the vicinity of the Hudson Submarine Canyon. In order to gain some insight into the circulation system of the bight and more particularly the flow regime in the upper portion of the Hudson Shelf Channel, current measurements were made in the summer of 1973 with a single bottom-mounted current meter. The meter, recording over an 1 1-day period before internal failure, was located (40°9.7'N, 73°4I.2'W) approximately 45 km south of Long Island and 27 km east of New Jersey in the channel axis, where the water depth measured 62 m. The location of the axis was determined with a fathometer during a cross-channel transect. Surprising though it may seem, understanding of the cir- culation system in the bight is still relatively poor. Although there have been several surface and bottom drifter studies made, comparatively few reports of direct current measure- ments are yet available for this area [Charnell et al.. 1972]. Further offshore but physiographically related to the channel is the Hudson Submarine Canyon, where summertime bottom current measurements were recently recorded [Keller et al., 1973]. In that investigation, axial currents of sufficient magnitude to move coarse sediment were observed, but the length of the observations precluded correlations with possible forcing agents. Those measurements were made more than 100 km from the present site in much deeper water with greater relief. In another recent and important study, measurements of currents al a site southeast of Block Island showed a strong correlation with prevailing meteorological conditions [Beardsley and Butman, 1974]. Documentation of the role sur- face winds play in the movement of shelf water has been made in other recent literature [Huyer and Pattullo. 1972; Cannon et al.. 1972; Cannon. 1972]. These earlier studies caused us to look to the wind as a possible driving mechanism for the small nontidal currents observed at the channel site. With current measurements sampled every 10 min by a Geodyne 100-C internally recording Savonius rotor current meter positioned I m above the bottom and using weather records kept by the Coast Guard at Ambrose Tower and at Copyright © ll)7? by the American Geophysical Union. I Barnegat Light, New Jersey, we are able to show that the cur- rents are axial in the channel during the recording period and suggest that the sense of flow (up or down channel) may be controlled to a large extent in the summertime by the prevail- ing surface winds. We present current and wind observations in some detail in the following section. Comparison of Winds and Currents The data taken over the 1 1-day interval (August 3, 1973, at 0700 to August 14, 1973, at 0700) at 10-min intervals are given in component form in Figure 1, the solid line representing the raw data. In order to remove energy at diurnal tidal periods and below we have smoothed the data with a 25-hour boxcar averaging window, and that result is plotted as a dashed line. The highest recorded velocity (11.5 cm/s) occurred on August 9. The average speed over the 1 1-day period was 2.5 cm/s. On occasion the actual current speed fell below 1 cm/s, and the meter stalled. It is evident from the record that currents during this period were heavily tidally influenced, although the tidal component generally is no larger in magnitude than the nontidal compo- nent. In order to appreciate tidal constituents we have made a least squares fit of the raw data minus the 25-hour running average at five tidal periods (T = 25.82, 23.93, 12.66, 12.42, and 12.00 hours) and have reconstructed the tidal current con- tribution as the dot-dash line in Figure 1. The A/2 semidiurnal component dominates with a 2.6-cm/s contribution, although the ringing tidal record demonstrates the significant role of the other tidal constituents. In Figure 2 the progressive vector diagram demonstrates the near one-dimensional nature of the flow. The trend of the channel axis in this area is roughly 135°, as is the resultant flow direction in the diagram. The total I 1-day excursion amounts to 8.8 km down channel. The most prominent feature is the 4- day southeast pulse beginning midday on August 7, during which the average down-channel flow amounted to nearly 4 cm/s. Wind data taken at Ambrose Tower, some 34 km north- northwest of the site, and at Barnegat Light, New Jersey, some 56 km south-southwest of the site, are presented in Figures 3 and 4. At Ambrose Tower a log of meteorological and sea conditions is regularly kept, providing six wind velocity and pressure recordings per day over the study period. In the upper panels of Figure 3 the points represent the Ambrose Tower raw data in component form, and the solid line represents a 25-hour average. In the lower panels, irregularly kept wind es- timates from Barnegat Light (points) as well as a hand-drawn 953 771 1954 Lavelle et al.: Bottom Current 3° 5° z 2 Eu /\A^'Ai'^W^W^V'-/VvV t^t^t^f ^ TIME (DAYS) Fig. I. Bottom current measurements (solid line) taken in the Hudson Shelf Channel from 0700 on August 3 to 0700 on August 14, 1973. The dashed line represents a 25-hour running average of the same data. The dot-dash line represents a least squares tidal fit to the currents. average (solid line) are presented. At Ambrose Tower the wind record is clearly modulated by diurnal solar heating, although the predominant feature is the south-southwest wind begin- ning on August 7. The strongest winds recorded were 14.4 m/s on August 10, while the 5-day average amounted to 6.7 m/s. At Barnegat Light, south-southwest winds were evident from midday on August 7, with one interruption on August 10, until midday on August 1 1, when the wind began to rotate to the west. Although the wind data are not as accurate or as frequent as we would like for the second station, there seems to be some coherence in meteorological records. We think that it is reasonable to assume that in their coherent aspects the wind data reflect a regional pattern, inclusive of the current meter site. Besides the south wind evident over 4 days in the center of the record, a west wind was evident over roughly a 2-day period near August 5 and 2-day period beginning August 12. Turning now to the coherence between the wind and current meter data, we point out the correspondence between major features of the progressive vector diagrams (Figures 2 and 4). The residual bottom current began a trend down channel after midday August 7 and continued until the middle of August 1 1. The surface winds showed considerable strength from the south-southwest on August 7 until roughly midday on August HUDSON SHELF CHANNEL CURRENT (3 AUG/0700 - II AUG/0700,1973) Fig DISTANCE EAST- WEST FROM ORIGIN (KILOMETERS) Progressive vector diagram of the bottom currents in the up- per Hudson Shelf Channel. 1 1, when the wind came from the west, that is, during the same interval as the down-channel flow. The two periods of lighter west winds previously mentioned, beginning August 5 and 12, seem to be reflected in the current record as up-channel flows. During the first part of the obser- vation period (August 3 at 0700 to August 5 at 0700) there was also a down-channel flow with a southwest wind. If the correlation is as we propose, there would appear to be some hysteresis in the response of the system to the forcing. While the tidal currents are of the same magnitude as the residual currents and therefore the source of some masking, there appears to be considerably more lag in setting up a flow than in its decay. For example, the up-channel flow on August 12 may have required a half day of offshore winds to begin up- channel flow, while the down-channel flow of August 1 1 relaxed almost immediately to the shift in wind direction. The limited data that we have are clearly insufficient to support SURFACE WINDS AUG 3- AUG. 15, 1973 Fig. 3. Upper panels represent wind velocity components measured al Ambrose Tower with a 25-hour boxcar average (solid line) superimposed. The lower panels are velocity component es- timates from the log at Barnegat Light, New Jersey, with a hand- drawn average. Directions expressed in this plot are directions toward which the wind is blowing. 772 Lavelle et al.: Bottom Current 1955 anything more detailed than the broad correlation that we have proposed above. In the Block Island work, Beardsley and Butman [1974] sug- gested that a northeast wind blowing parallel to the Long Island barrier can cause a relatively intense geostrophic flow to the west. We think that the correlation that we suggest is the result of the same basic mechanisms, although they are now operating on a highly stratified water column rather than the well-mixed condition existing during the late winter and early spring months. We may be seeing a situation of onshore winds leading to setup along the coast and bottom seaward flow, while offshore winds may cause some setdown and a compensating flow up channel. The correlation in this case could be dependent on the stratification extant in the bight during the summer months. Conclusions From these observations we think there is reason to :gest a correlation between flow in the axis of the Hudsi "helf Channel and the prevailing surface winds. During the :mer period, winds which blow onshore (south-southwest 01 outh) seem to produce down-channel flow, while winds blowing offshore (west or northwest) appear to cause a current reversal and a net up-channel current. In both cases the strength of the wind-induced flow is comparable to the tidal currents. The evidence begins to appear for sustained (>1.5 days) and reasonably high winds (>5 m/s). For winds of highly variable direction, no significant signature in the channel flow should be observed. If the stratification of the bight is a req- uisite ingredient for the flow, it would be reasonable to expect a correlation of this kind beginning in June and continuing through September. It is informative to look at the mean meteorological condi- tions prevailing during the June-September months in the bight area. In a summary of meteorological conditions for American ports [Brower et al., 1972], monthly mean wind speeds and directions are compiled for the offshore Sandy Hook, New Jersey, area. There, mean winds are shown to come out of the south or southwest from May through September at an average velocity of 5 m/s. The remaining months of the year, the winds are from the northwest at slightly higher average speeds. With this wind direction over the June-September period of high stratification a net down- channel flow from June through September could be expected. This same general result, summertime down-axis flow, was suggested in unrelated observations by Keller et al. [1973] in the Hudson Submarine Canyon. Again we would like to mention the rectilinear nature of the tidal and residual currents during the measurements. While the Hudson Shelf Channel in the vicinity of the observation has a depth to width ratio of only 10 2,' the flow moves essentially up and down the channel axis, a result suggesting some topographic control. Finally, we would like to point to two recent studies made of the sediments in the Hudson Shelf Channel. In work on the organic material and carbohydrate concentrations in the chan- nel, P. G. Hatcher and L. E. Keister (personal communica- tion, 1974) were able to observe distribution of materials whose apparent source was the sewage sludge dumping site of New York City, which lies at the channel head. Organic dis- tributions at points sampled across the channel have maxima corresponding to the channel axis. Concentrations are highest near the source, as could be expected, but extend down the channel some 100 km following the bottom contours. 1 1 ,13 4000 WINDS AT AMBROSE TOWER ( 1 AUG-I5AUG.I973) 12 It II IS ■^^14 3000 /lO / 9 2000 /a 5 J 1-^v 6 v 1000 4 i N \ 2 S 0 \ 1 ORIGIN | 1 - Fig. 4. distance east -west from origin ( kilometers) Progressive vector diagram for the wind measured at Ambrose Tower. Anomalous organic carbon concentrations are also found in the Hudson Submarine Canyon [Keller, 1973]. In a slightly earlier study of the heavy metal contaminants in the surficial sediments of the bight, Carmody et al. [1973] found anom- alously high concentrations of trace metals following the Hudson Shelf Channel axis and extending to 30 km from the dump site. Both observations may be additional evidence for a consistent, if not considerable, bottom transport down the channel axis. While the suggestion of wind-forced axial flow in the Hudson Shelf Channel is based on just one 11-day suite of observations, we think it can serve as a useful working hypothesis for additional and lengthier observations. There may indeed be a general transport down the channel and out through the canyon at certain times of the year, but verifica- tion awaits further measurements. Acknowledgments. The authors would like to thank George Lapiene for his contributions to the engineering and field efforts in- volved in these measurements. We also acknowledge the helpful cri- tique of the manuscript by Donald Hansen and Robert Charnell. This work was supported in part by NOAA's New York Marine Eco- system Analysis Program (Mesa). References Beardsley, R., and B. Butman, Circulation on the New England con- tinental shelf: Response to strong winter storms, Geophys. Res Lett . I. 181-184, 1974. Brower, W. A., D. D. Sisk, and R. G. Quayle, Environmental guide for seven U.S. ports and harbor approaches, report, p. 55, NOAA Environ. Data Serv., Nat. Clim. Center, Asheville, N. C, 1972. Cannon, G. A., Wind effects on currents observed in Juan de Fuca submarine canyon, / Phys. Oceanogr., 2. 281-285, 1972. Cannon, G. A., N. P. Laird, and T. Ryan, Currents observed in Juan de Fuca submarine canyon and vicinity, 1971, Tech Rep. ERL252- POL14. Nat. Ocean, and Atmos. Admin., Boulder, Colo., 1972. 773 1956 Lavelle et al.: Bottom Current Carmody, D. J., J. B. Pearce, and W. E. Yasso, Trace metals in sedi- Keller, G. H., Sedimentary dynamics within the Hudson Submarine ments of New York Bight, Mar. Pollul. Bull., 4. 132-135, 1973. Canyon, Proces-Verbaux Relations Sedimentaires entre Estuaires Charnell, R., D. Hansen, and R. Wickland, Surface and bottom water et Plateaux Continentaux, p. 49, Inst, de Geol. du Bassin d'Aqui- movement in New York Bight, The Effects of Waste Disposal in the taine, Bordeaux, France, 1973. New York Bight, report, Coastal Eng. Res. Center, Washington, Keller, G„ D. Lambert, G. Rowe, and N. Staresinic, Bottom currents D. C, 1972. in the Hudson Canyon, Science. 180, 181-183, 1973. Huyer, A., and J. Pattullo, A comparison between wind and current observations over the continental shelf off Oregon, summer 1969, J. (Received August 6, 1974; Geophys. Res.. 77. 3215-3219, 1972. accepted February 13, 1975.) 774 42 Reprinted from: American Geophysical Union, Fall Annual Meeting 56, No. 12, 1003. TRACER OBSERVATIONS OF SAND TRANSPORT ON THE LO!JG ISLAND INNER S1EI.F J.W, lavelle 1). J.?. Swift (both at: Atlantic Oceanosraphic and Meteorological laboratories, NCAA, 15 Rickenbacker Causeway, Miami, fla. 331M9) K . P . Brashear F.N. Case (trth at: Oak Ridge National Laboratory, operated for ERDA by Union Carbide Corporation, Oak Ridge, Tennessee 37830) We liave observed both sprij:g-siimmer and fall- winter sand transport in two '-xperinents on tlie long Island Inner Shelf at water depths of 20 tc ?'t meters using a ran :o isotope sand tracer nystun. Dispersion patterns of the tagged material, milium to fine sand with a mean diameter of 0.125 nm, sampled biweekly over both 70-w^y experiments, support the hypothesis of wintertime storm activity ai> the principal a^ent of shelf sand transport. In tr.e late spring =*no early summer, movement is primarily diffusive in nature, at- tending ?C0 metorr- fran the injection point, while late fall-winter pattcins have strong idvectLve features, including an ellipsoidal smear Df material extending approximately 1200 meters longshore after the passage of several "northeasters." Near-bottom current observations made with Savenius rotor sensors show: a doubling of peak near-bottom velocities from approximately ?0 to 60 c^i/sec, from th-^ first to the second experiment; and the dominance of west- ward storr. flew along the Long Island Inner Shelf in transporting sand. 775 43 Reprinted from: Journal of Geophysical Research 80 No. 23, 3307-3314. Crest of the Mid-Atlantic Ridge at 26°N Bonnie A. McGregor1 and Peter A. Rona NO A A Atlantic Oceanographic and Meteorological Laboratories. Miami, Florida 33149 A relatively detailed investigation of the Mid-Atlantic Ridge crest at 26°N was conducted by using narrow-beam bathymetric data, total earth's magnetic field measurements, and underwater photographs. The Mid-Atlantic Ridge crest at 26°N appears to be hydrothermally active. The structural setting of this area is conducive to the occurrence of hydrothermal deposits. The walls of the rift valley are extensively faulted with blocks and steps ranging in size from kilometers to meters in width and relief. Underwater photographs show hydrothermal manganese associated with interpreted fault steps at depths between 3100 and 2500 m on the east wall, suggesting that the faults provide avenues for hydrothermal fluids. Small topographic highs in the floor of the rift valley are the sites of relatively recent volcanism and are believed to represent the top of an active dike emplacement zone. Bathymetric trend directions for this portion of the ridge crest are complex in comparison with plate rotation predicted trends for the Mid-Atlantic Ridge. The bathymetric grain is a function of processes active in the rift valley. Introduction During the 1972 trans-Atlantic geotraverse (TAG) program of the NOAA Atlantic Oceanographic and Meteorological Laboratories a geological and geophysical study of an area 1 80 km by 180 km of the crest of the Mid- Atlantic Ridge between 25°N and 27°N was carried out (Figure 1) [Rona el at.. 1973. 19756], As a continuation of this program in 1973, a relatively detailed study was made of an area 75 km by 55 km (area out- lined in Figure 1 ) centering on the TAG hydrothermal field [R Scott et ai, 1974], which was identified during the 1972 TAG program [M. Scon et ai.. 1974]. This paper reports results of the detailed study, the objectives of which were ( 1 ) to define the geomorphologic features present on a slow-spreading midoceanic ridge, (2) to establish the relationship of these features to axial rift processes, and (3) to determine the struc- tural framework associated with the hydrothermal mineral deposit. In the detailed study area, northwest-southeast track lines oriented perpendicular to the trend of the rift valley are 2 km apart, and the northeast-southwest lines parallel to the rift valley are 3.5 km apart (Figure 2). Narrow-beam echo- sounding data and total earth's magnetic field measurements were collected continuously along the track lines positioned by satellite navigation. Supplemental bathymetric data from the 10- and 20-km line-spaced survey (Figure 1) of the 1972 field season were also used. Small-scale features of the order of I m are revealed in a transect of stereophotographs made from the floor to the top of the east wall of the rift valley. The general setting for this area is a portion of ridge crest in the central North Atlantic between two major fracture zones, the Kane at 24.7°N and the Atlantis at 30°N. The oldest sea floor studied in the detailed area is slightly older than 3 m.y. based on a 1.2-cm/yr spreading rate [Laitimore el ai, 1974]. Bathymetry The general bathymetry of the Mid-Atlantic Ridge is shown in Figure I . The rift valley is composed of a series of closed basins trending N25°E. Three transverse valleys spaced 30 km apart appear as deep basins on both the east and the west sides of the rift valley. These basins intersect the rift valley with a different trend on either side. 90° on the east and 60° on the ' Also at Rosensliel School of Marine and Atmospheric Science. University of Miami. Miami. Florida 33149. Copyright © 1975 by the American Geophysical Union. west [Rona el ai. 1973, 19756]. The rift valley widens and deepens at the intersection of these transverse valleys. Two oblique-trending transverse valleys are present on the west side of the ridge, and in each case a side lobe of the rift valley with a trend due north occurs such that the transverse valley trends are perpendicular to these north-south rift trends. Figure 3 is a schematic drawing of the previously discussed trends. Shaded areas are visible bathymetric trends, and dashed lines are trends that are less obvious but are present. According to Johnson and Vogt [1973], the Mid-Atlantic Ridge between 47° and 5I°N can be divided into rift segments 30 km in length with either normal-trending (north-south) or oblique-trending axes. At 26°N the rift has a normal and an oblique trend in the same segment. The different trends may suggest minor adjust- ments in spreading direction for small blocks of sea floor. The east wall of the rift valley located at 26°06'N, 44°45'W is the site of the TAG hydrothermal field described by R. Scott el at. [1974]. The detailed bathymetry of the ridge crest in this area is show n in Figure 4, the heavy 3500-m contour outlining the rift valley. Three of the transverse valleys intersect the rift in this area, two on the west and one on the east. Where the transverse valley trends are orthogonal to the rift southeast of the axis, the flanking valleys in the rift mountains parallel the N25°b trend of the rift axis. Where the transverse valley trends are oblique to the main rift trend northwest of the axis, the flanking valleys parallel the north-trending lobe or segment in the rift valley (Figure 4). The flanking valleys appear to form by isolating portions of the rift valley as shown on the southeast side of the rift at 25°55'N, 44°45'W (Figure 4), where discontinuous highs are beginning to isolate the 3500-m basin to the east of the rift. A similar feature appears to be forming at 26°15'N on the west side of the rift. The axis of the rift valley within the detailed study area has two deep basins containing small linear topographic highs. One is located at the northern end of the southern basin, and the other is marked by the closed 4000-m contour in the northern basin (profiles 10-16 and 20-22 in Figures 4 and 5). Both features are 200-300 m high and are believed to be the site of recent volcanism. Fresh basaltic glass was recovered from the high in the northern basin (R. Scott, personal com- munication. 1974). Similar features were found in the Famous area to the north between 36° and 37°N [Macdonald et ai. 1974; Seedham and Francheleau, 1974]. The east wall of the rift at 26°06'N, 44°45'W is higher than the west wall, and the rift valley narrows in this area. Large 3307 776 3308 27 00 ,N - 26° 30' 26° 00' 25° 30' 25 00 ,N - McGregor and Rona: Crest of the Mid-Atlantic Ridge 46°00'W 45° 30' 45°00' 44° 30' 44°00' W - 1 1 1 1 1 hmLf o - ' ^1 4000 1 1 1 ^W CO" 40° WW 1 1 27°N 00' 26° 30' 26° 00' 25° 30' H25°N ^00' N 46°00' W 45°30' 45°00' 44°30' 44°00' W Fig. I . Bathymetry of a portion of the Mid-Atlantic Ridge (200-m contour interval, depths in corrected meters). Inset shows general location of area. Outlined box is location of detailed study area shown in Figure 3. Dots are earthquake epicenter positions. Heavy 3600-m contour outlines the rift valley. Bathymetric map is from Rona el al. [19756]. Fig. 2. Track control for Figu Figure 5. lines from re 4, and TAG phase 3, 1973. Track lines are data index numbers are track locations for blocks in the rift walls are apparent on the narrow-beam echo- sounding profiles and can be correlated for several kilometers along strike (arrows in profiles 4-8 of Figure 5). They are believed to have a fault origin, resembling fault blocks shown by Atwater and Mudie [1973] on the Gorda Rise. Suggested faults are indicated on Figure 5. The two peaks making up the east wall of the rift (profile 14) can be traced continuously from profile 4 to profile 20 (dashed line). Their elevation ap- pears to be a function of the fault displacement, these blocks also being associated with the TAG hydrothermal field. At least two scales of fault-derived features are present in the area: (1) those of the order of several kilometers with a minimum size of 0.4 km, visible on narrow-beam echo- sounding profiles, and (2) those of the order of meters, which can only be seen with bottom photography. Underwater Photography A transect of underwater photographs reveals the small- scale features and detailed structure of the east wall. Two 777 McGregor and Rona: Crest of the Mid-Atlantic Ridge 3309 camera runs, each being of approximately 5-hour duration, are shown in relation to the bathymetry (Figure 6). Two un- derwater survey cameras in stereographic configuration were used with 35-mm negative movie film in each camera. Pictures were taken automatically every 8 s, providing overlap of adja- cent photographs, while the cameras were maintained approx- imately 5 m above the bottom, at ship's speeds between 0.5 and I knot (0.25-0.5 m/s). During camera traverse 1, pictures of the bottom were taken continuously, whereas during camera traverse 2, a weak pinger return from the camera necessitated pulling clear of the bottom for an interval to iden- tify the pinger and bottom returns. Because of the large number of photographs obtained, one camera per run was selected, every third photograph scanned, and the percentage of bottom type recorded following the technique used by Moore and Fiske [1969], Five categories of bottom type were identified as being significant for interpreta- tion of the structure of the rift wall: ( 1 ) pillows, indicating lava flows, (2) talus, angular debris from gravel to boulder size, (3) breccia, cemented angular and subangular fragments, (4) sedi- ment, composed of silt and sand, and (5) hydrothermal manganese deposit (Figure 7). A total of 1800 photographs from the two stations were scanned. Because of the volume of data the percentage of bottom types was averaged over approx- imately 15-min intervals and plotted as bar graphs (Figure 6). Also shown in Figure 6 are locations of three dredge hauls with percentage of total material dredged (TAG leg 3, 1972, and leg 4, 1973, cruise report, R. Scott, personal communica- tion). The dredge hauls help identify some of the relationships and material present in the photographs. Sediment is present over most of the rift wall, although it may be only a few cen- timeters thick. Pillow basalts are found capping the top of this east wall and associated with a step on the wall at approx- imately 3000 m. This step was the site of a temperature in- crease in the bottom water [Rona el ai, 1975a]. Hydrothermal manganese having a layered appearance was photographed along both traverses between 2500 and 3100 m, blanketing the talus, and from the dredged material was found as a matrix fill- ing interstices and cementing the talus. Talus is abundant over most of the wall. The spatial relationship of bottom types gives an indication of the structure of the wall. A sequence of sedi- ment, talus, and breccia repeating itself suggests the presence of a terrace or step in the wall, sediment on the outer portion and talus plus breccia close to the wall implying a scarp as the source of the talus. The pinger and bottom trace, while they maintain the camera 5 m off the bottom, also confirm the presence of steps or terraces. On those portions of the camera runs that were made approximately perpendicular to the con- tours, the number of possible steps in the wall could be counted. Camera traverse I from 2800 to 2400 m had 10 such steps varying in width between 30 and 400 m with 40 m of average relief between each step to account for 400 m of relief of the wall. Smaller steps are present in the breccia between 2500 and 2400 m and in a small region of the pillow basalts at the top of the wall, as the camera periodically had to be raised between times of level towing while being maintained at ap- proximately 5 m off the bottom. The size of these steps was of the order of 10 m wide and 10 m high. In the region of pillow basalts these steps probably represent the edge of lava flows and are constructional features. During camera traverse 2, between depths of 3400 and 3100 m, four steps were identified by bottom type sequence. The steps varied in width between 100 and 300 m, each being separated by 75 m of average relief. If the correlation of bottom type with structure is correct, the Fig. 3. Schematic drawing of the ridge crest and topographic trends. A and B are north-south trending rift segments, and C repre- sents N25°E rift segments. Hatched bands are main transverse valley trends. All these trend directions make up the topographic fabric of the rift. wall consists of many steps. The association of the hydrother- mal manganese where it is present with talus and breccia at the juncture of the steps and wall suggests that the steps may be fault-controlled, the fault zone providing a conduit for hydrothermal solutions [R. Scon el ai, 1974; McGregor el ai, 1974a). Also present in the photographs was evidence of bottom current activity confined above 3000 m. Globigerina ooze in some places appears to have a lag deposit of debris, which in some cases is manganese-coated pteropods similar to those seen in the Famous area (G. Keller, personal communication, 1974). Oscillation ripple marks with varying wavelengths up to 0.6 m, amplitudes about 2 cm, and varying orientations in- dicate that current direction is probably controlled by topography. In some instances, slight drift of the sediment cloud stirred up by impact of the camera's compass on the bot- tom was to the north. A large number of siliceous sponges are present on camera traverse 1 at depths between 2350 and 2900 m, where a solid substrate facilitates their attachment. The sponges have a concave side (Figure 7), and in general, this side faces south. If these sponges orient to maximize feeding efficiency, the concave side may be expected to be directed toward the current. This would imply a current flow to the north, which is in agreement with observations from the rift valley to the north at 36°-37°N [Keller el al.. 1975). Although ripple marks were present, no strong bottom currents occurred during the camera transects, as is evidenced by the lack of dis- persion of the sediment cloud formed by the camera's compass in most cases. Magnetic Anomalies Total earth's magnetic field measurements were made along the ship's tracks shown in Figure 2, the numbered tracks refer- ring to the profiles in Figure 5. The international geomagnetic reference field 1965 coefficients [International Association of Geomagnetism and Aeronomy (IUGG) Commission 2, 1969] were used to reduce the data, with a 400-7 adjustment to the reference field based on the American world chart 1970 model [Hurwitz, 1970]. The large positive anomaly in the center of 778 3310 McGregor and Rona: Crest of the Mid-Atiantic Ridge 779 McGregor and Rona: Crest of the Mid-Atiantic Ridge 331 SHEADING IATI IICM/YI 4000 2000 4000 2000 4000 2000 4000 2000 PROFILE 24 22 20 18 16 14 12 10 _i i i_ _i i i_ -400 + 400 -400 ♦400 -400 ♦ 400 -400 ♦ 400 -400 ♦ 400 -400 ♦ 400 -400 ♦ 400 -400 ♦ 400 -400 + 400 -400 ♦ 400 -400 ♦ 400 20 40 60 DISTANCE (KIIOMETEISI BATHYMETRY IN METERS 20 40 60 10 DISTANCE (KIIOMETEISI MAGNETIC ANOMALY IN GAMMAS Fig. 5. Balhymetnc and magnetic profiles orthogonal (NW-SE) to the ridge. See Figure 2 for location Large arrows in- dicate axis of the rift valley. On profile 14, short lines indicate location of 10-m steps, and long lines indicate 100- to 300-m steps determined from photographic transects. Small arrows indicate fault blocks. Dashed line correlates large fault- controlled block in the east wall. Possible fault zones are suggested. each profile (Figure 5) is the axial anomaly. A magnetic model for the area assuming a constant spreading rate of 1.2 cm/yr [Laiiimore el ai, 1974] is shown in Figure 5. The oldest sea floor in the detailed study area is about 3 m.y. old. A significant feature of the magnetics is the different character of the axial anomaly in the area of the TAG hydrothermal field. A low is present in the axial anomaly decreasing in prominence from profile 14 north and south to profiles 20 and 8. The source of this magnetic anomaly is a portion of sea floor with reduced intensity of remanent magnetization [McGregor, 1974; McGregor el ai. 1974ft]. Since this is the site of the hydrothermal deposit and a bottom water temperature anomaly [Rona el ai, 1975a], hydrothermal alteration of the basalt and reduction of the intensity of rema- nent magnetization are expectable. Alteration of basalts by hydrothermal activity has been shown to reduce the magnetic susceptibility and intensity of remanent magnetization [Luyen- dyk and Melson, 1967; Watkins and Paster, 1971]. In addition to the hydrothermal manganese, altered basalts were dredged, including greenstones and zeolitized rocks (TAG leg 4, 1973. cruise report, R. Scott, personal communication). The Reyk- janes Peninsula of Iceland, a hydrothermal area, also has a prominent magnetic low caused by the hydrothermal altera- tion of basalt [Bjornsson el ai, 1972]. 780 3312 McGregor and Rona: Crest of the Mid-Atlantic Ridge 26°10'N 26°05'N 26°05'N 44o50*W *r4 ' > 44°45'W Fig. 6. Bathymetry, in hundreds of meters, of a portion of the east wall. Positions of two camera stations (heavy dashed lines) and three dredge stations (diamond pattern) are shown. Bar graphs are percentages of bottom type present along each camera run averaged over approximately 15 min of time. Bottom types are (I) pillow basalts, indicating flows, (2) talus, (3) breccia, (4) sediment, silt, and sand, and (5) hydrothermal manganese. Computer modeling studies were undertaken to interpret the magnetic anomaly pattern present on the Mid-Atlantic Ridge at 26°N [McGregor, 1974; McGregor el ai. 19746], Variations in the intensity of magnetization parallel to the ridge crest were found, a low in the magnetization being cor- related with an extensively faulted region in the east wall of the rift valley. Earthquake Activity The relationship of bathymetry and earthquake epicenter locations is shown in Figure 1 , with the epicenter data from the NOAA National Geophysical and Solar Terrestrial Data Center Hypocenter Data File (C. A. von Hake, personal com- munication, 1973). Eighteen events from 1953 to 1973 and one from 1937 occurred in the area with magnitudes between 4.0 and 5.0. The plotted epicenters have an uncertainty in location of 20 km. In spite of the uncertainty it is useful where detailed bathymetry is available to look for any associations. Within the study area the northern half of the rift is much more active than the southern half. Many epicenters appear to be associated with the wide, deep basins in the rift at the in- tersection of the transverse valleys. Some epicenters appear to be associated with the rift mountains, but this apparent as- sociation may reflect the uncertainty of location determina- tions. Coarse-grained gabbro dredged at 26°15'N, 44°20'W (R. Scott, personal communication, 1974) suggests exposure of a magma chamber. Such exposure is most probably related to faulting. A detailed microearthquake survey between 36° and 37°N [Spindel et ai. 1974] shows that activity is associated with rift walls and is related to faulting. In the area 25°-27°N, detailed seismic monitoring was not done, but from the large- magnitude earthquake epicenters (magnitude of about 5 on the Richter scale), favoring of the east wall for the zone of activity was suggested. Of the 18 events shown in Figure 1, only three would not lie along the east wall (these would lie in the rift val- ley), a maximum error in position of 20 km being assumed. Conversely, of those 18 events, even with a maximum error in position being assumed, four can reside only under the east wall. North of the Azores triple junction between 47° and 51°N, Johnson and Vogt [1973] found that the epicenters, in spite of the error in position, still favored the west side of the rift. The significance of earthquake swarms as indicators of pos- sible hydrothermal or magmatic activity was suggested by Sykes [1970], Earthquake swarms are present on the ridge crest at 28.4° and 31.4°N [Sykes, 1970], suggesting possible locations for additional detailed studies to be done. Conclusion Detailed studies are essential to defining the structure of ridges and to understanding rift processes. This portion of the 781 McGregor and Rona: Crest of the Mid-Atlantic Ridge 3313 WPp* . . .<"■>"" v: Tin •' V- J? ^*? ■§&■ K&3k ■ . ■ ■ RJ*' - ■ ob • jz j_ S a. 782 3314 McGregor and Rona: Crest of the Mid-Ati antic Ridge Mid-Atlantic Ridge (25°-26°N), with its associated slow spreading rate, is extensively faulted. Fault blocks vary in size from tens of meters to kilometers in width and relief. The TAG hydrothermal field is associated with fault steps 30-300 m in width on the east wall as well as with large faulted blocks, sug- gesting that the fault zones provide an avenue for hydrother- mal fluids. Hydrothermal manganese, identified from dredged specimens [M. Scott el al., 1974], was photographed blanketing talus at depths between 3100 and 2500 m. Small steps, about 10 m in width and relief, occur at the top of the east wall of the rift. Topographic features at 26°N bear certain similarities to those reported on other portions of the Mid-Atlantic Ridge between 36° and 37°N and between 47° and 51°N. The rift valley is composed of a series of discontinuous deep basins, with small linear topographic highs in the floor of the rift val- ley which are the site of recent volcanism and are believed to be the zone of active dike emplacement. The topographic fabric is complex in comparison with that predicted by plate tectonic theory. The rift valley itself is composed of two trend directions, topographic trends being a function of the portion of the rift valley from which they originated. Seismicity and magnetic anomaly patterns may be useful in locating areas where tectonic processes are conducive to the occurrence of hydrothermal activity. Acknowledgments. We gratefully acknowledge Dale C. Krause for his invaluable assistance in establishing the objectives and models tested by phase 3 of the NOAA-TAG 1973 field work. We thank Louis W. Butler, Mahlon M. Ball, and George H. Keller for enlightening discussions and review of the manuscript. We also thank J. William Lavelle for providing computer programing assistance, Sam A. Bush for assisting in the data reduction, and Martin R. Fisk for calculating magnetic held values. We express our appreciation to Charles A. Lauter, Jr., whose preparation and operation of the deep-sea camera made the photographic transects a success. L. L. Posey, W. S. Sim- mons, and the other officers and crew of the NOAA ship Researcher provided cooperation and diligence in executing the survey. Funds for the study were provided by the National Oceanic and Atmospheric Administration. References Atwater, T., and J. D. Mudie, Detailed near-bottom geophysical study of the Gorda Rise, J. Geophys. Res.. 78, 8665-8686, 1973. Bjornsson, S., S. Arnorsson, and J. Tomasson, Economic evaluation of Reykjanes thermal brine area, Iceland, Amer Ass. Petrol. Geol Bull.. 56. 2380-2391, 1972. Hurwilz, L., Mathematical model of the 1970 geomagnetic field (abstract), Eos Trans. AGU. 51. 269, 1970. International Association of Geomagnetism and Aeronomy (IUGG) Commission 2, Working Group 4, International geomagnetic reference field 1965.0, J. Geophys. Res.. 74, 4407-4408, 1969 Johnson, G. L., and P. R. Vogt, Mid-Atlantic Ridge from 47° to 5I°N, Geol. Soc. Amer. Bull.. 84. 3443-3462, 1973. Keller, G. H., S. H. Anderson, and J. W. Lavelle, Near-bottom cur- rents in the mid-Atlantic Ridge rift valley, Can. J. Earth Sci., 12. 703-710, 1975. Lattimore, R. K., P. A. Rona, and O. E. DeWald, Magnetic anomaly sequence in the central North Atlantic, J. Geophys. Res., 79. 1207-1209, 1974. Luyendyk, B. P., and W. G. Melson, Magnetic properties and petrology of rocks near the crest of the Mid-Atlantic Ridge, Nature. 215. 147-149, 1967. Macdonald, K. C, B. P. Luyendyk, F. N. Spiess, and J. D. Mudie, Near bottom geophysical studies of the Famous rift valley, Mid- Atlantic Ridge (abstract), Eos Trans. AGU. 55. 446, 1974. Matthews, D. J., Tables of the velocity of sound in pure water and sea water for use in echo-sounding and sound ranging, Publ. H.D. 282. 52 pp.. Admiralty Hydrogr. Dep., London, 1939. McGregor, B. A., Crest of Mid-Atlantic Ridge at 26°N: Topographic and magnetic patterns, Ph.D. thesis, Univ. of Miami, Coral Gables, Fla., 1974. McGregor, B. A., P. A. Rona, and D. C. Krause, Crest of Mid- Atlantic Ridge at 26°N (abstract), Eos Trans. AGU. 55. 293. 1974a. McGregor, B. A., C. G. A. Harrison, and P. A. Rona, Magnetic anomalies on the Mid-Atlantic Ridge crest at 26°N, Geol. Soc. Amer Abstr. Annu. Meet.. 6. 863, 1974ft. Moore, J. G., and R. S. Fiske, Volcanic substructure inferred from dredge samples and ocean-bottom photographs, Hawaii, Geol. Soc. Amer. Bull.. 80. 1191-1202, 1969. Needham, H. D., and J. Francheteau, Some characteristics of the nfi valley in the Atlantic Ocean near 36°48' north. Earth Planet. Sci. Lett.. 22, 29-43, 1974. Rona, P. A., R. N. Harbison, B. G. Bassinger, L. W. Butler, and R B Scott, Asymmetical bathymetry of the Mid-Atlantic Ridge at 26°N latitude (abstract), Eos Trans. AGU. 54. 243, 1973. Rona, P. A., B. A. McGregor, P. R. Betzer, G. W. Bolger. and D. C. Krause, Anomalous water temperature over Mid-Atlantic ridge crest at 26° north latitude. Deep Sea Res., in press, 1975a. Rona. P. A., R. N. Harbison. B. G. Bassinger, R. B. Scott, and A. J. Nalwalk, Tectonic fabric and hydrothermal activity of Mid-Atlanlic Ridge crest (26°N). Geol. Soc. Amer. Bull., in press, 19756. Scott, M. R , R. B. Scott, A. J. Nalwalk, P. A. Rona, and L. W. Butler, Hydrothermal manganese in the median valley of the Mid- Atlantic Ridge. Geophys Res. Lett.. I. 355-358, 1974. Scott, R. B., P. A. Rona, B. A. McGregor, and M. R. Scott, The TAG hydrothermal field. Nature. 251, 301-302, 1974. Spmdel, R C, S. B. Davis, K C. Macdonald, R. P. Porter, and J. D. Phillips, Microearthquake survey of median valley of the Mid- Atlantic Ridge at 36°30'N. Nature, 248. 577-579, 1974. Sykes, L. R., Earthquake swarms and sea floor spreading, J Geophys. Res.. 75. 6598-6611, 1970. Watkins, N. D., and T. P. Paster, The magnetic properties of igneous rocks from the ocean floor, Phil. Trans. Roy. Soc. London. Ser. A. 268. 507-550, 1971. (Received September 26, 1974; revised January 30, 1975; accepted February 20, 1975.) 783 44 Reprinted from: EOS 56, No. 6, 382, SEISMIC PR-TILES ALONG TiT U.S. NGKTKLAST COAST CONTINENTAL KAF3IN ?■>. f,. X-cTrepor (KOAA, Atlantic Cceanographic and l:ia tecrological laboratories, Miami, i'lorida 331U9) G.'rt. Keller (N'OAA, Atlantic Oceancgrephic and Meteorological Laboratories, Miami, Florida 33149) R.H. Bennett (NCAA, Atlantic Cceanographic and ^ateorological Laboratories, Miami, Florida ?;i4?) TVonty-seven hundred kilometers of seismic rcf. action prof:_e3 were collected between hy*.u'1ogr cipher and Wilmington Canyons paj^Lllel to the continental ir.-.:-g\r. o:. tr? outer shelf, "i ' Jle slcpe, lower slc-e, and contir.ental rise. On the oior*r and in the canyons extensive slumping has occurred. A stratified block separating twin canyons southwest of Slock Car.ycn appears to be tilted. A wedge cf sediment, probably Tertiary in age, with many reflectors thickens to the southwest between Alvin and Hudson Canyor.s. The slope in this area is cut by several large can- yons whereas in the vicinity of Hydrogr^pher Canyon end from Hudson to Wilmington Canyon the slope is extremely dissected with many small vaV.eys. reflecting horizons arc continuous, parallel to the margin. Pronounced horizons in many places appear to control the morphology. Unconformities are abundant on the margin indi- cating varying deposit iondl ar.d erosicnal en- virc:\TCnts. In the rise a buriod depression extends from Hudson to Llndenkohl Canyon. 784 45 Reprinted from: Geological Society of America Annual Meetings 3 Abstracts with Programs 7, No. 7, 1196. IMPACT STRUCTURES ON LANDSAT IMAGERY McHone, John F. Jr., Department of Geology, University of Illinois, Urbana, Illinois 61801; and Dietz, Robert S., National Oceanic and Atmospheric Administration, Miami, Florida 33140 Satellite generated imagery (I.ANDSAT 1:1,000,000) offers a "new look" at terrestrial meteorite craters and astrobleines (ancient impact scars). We have compiled an atlas of known impact sites from such images and, by comparison, selected other geologic features of possible extraterrestrial origin. Considering the plethora of endogenic circular geologic p. true t- ures, the identification of -an exogenic impact scar still require:- field evidence of shock metamorphism; nevertheless some structures can !>•■ al- most certainly recognised as astroblemes from LANDSAT imagery alone (e.g. Aragu'ainha Ring in Ma to -Grosso , Brazil). Established impact sites of the world are resolved by LANDSAT in1 ■■ five different styles: (I) large annular qrabens (Man i couagan) ; (?) hall's eye pattern.; composed of concentric disks (Araguninha) ; (3) gho.sl rings (Gosses Bluff); (4) bleached rings (Barringer and Lonar craters); and (5) deep, as indicated by delayed freezing, circular lakes, common]/ with central islands, in glaciated Precambrian shields (Clearwater Lakes, Mistastin. Lanpa jarvi ) . o A probable satellite-detected astrobleme is the Oman Ring (10 'ia'N, 56 58'E) in the Arabian Peninsula. This ghost ring structure, brought to our attention by Nicolas Short of NASA, is 6 km in diameter with a 2 km wide central dome. Another probable site is Or.ero Fl'gygytgyn (67°30'N, 1 V;'o0S ' F.) in north eastern Siberia. A crater lake 14 km across, the feature rescmbli New Quebec crater in Canada but is much larger and more maturely eroded. Imagery shows the rudely circular shoreline to be enclosed by highLy circular geomorphology bearing a definite rim along at least the north edge. Origin by impact. in the early Quaternary or late Pliocene seems indicated. Representative images of several other known and suspected impact sites have been selected for discussion. 785 46 Reprinted from: Meteovitics 10, No. 3, 209-214. ON THE TERRESTRIAL ORIGIN OF THE TEPEXITL CRATER, MEXICO Lucrecia Maupome lnstituto de A stronomia, UNA M Mexico 20, D. F. and Centro de Investigation, IPN Mexico 14, D. F. Roman Alvarez lnstituto de Geofi'sica, UN AM Mexico 20, D. F. Susan W. Kieffer Dept. of Geology, University of California Los Angeles, CA 90024 Robert S. Dietz National Oceanic and Atmospheric Administration Miami, FL 33149 The possibility of a meteoritic origin for Tepexitl crater in Mexico was proposed by Maupome (1974), owing to the evident explosive origin of the crater, its circularity and its strong topographical resemblance to Wolf Creek cr?ter, to the Pretoria Salt Pan, and to some lunar craters. Located in Zacatepec plain, the crater is a bowl-shaped, nearly circular structure with a complete rim, Fig la and lb; its present depth averages 75 m from the top of the rim to the flat floor. Its measured diameters vary from 1 180 m to 966 m. The rim varies in height above the crater floor from 57 m on the north to 92 m on the south. At the top the width is 2 to 5 m; the outer slopes are about 15° and the inner walls have slopes of about 27°. The crater has an inner ridge along the SE radius extending from the top of the rim to nearly the center of the crater. Tepexitl is located at 19°13'N, 97°26'W, in the NE part of the State of Puebla. The area is within the Mexican Volcanic Belt (Quaternary volcanism) and contains maars, calderas, volcanic cones, limestone outcrops and intrusive ryholitic domes. There are, however, few geological studies published on the area and those available cover it only partially (Ordonez, 1905 and 1906; Ohngemach 1973). Tepexitl apparently had not been formerly described in the literature. Geophysical studies were initiated in order to test the hypothesis of meteoritical origin (Alvarez eta I., 1975); aeromagnetometry and ground magnetometry surveys were carried out over volcanic cones, calderas, and Tepexitl, in order to study geophysically determinable differences. Such initial surveys yielded favorable results in support of an impact origin for Tepexitl since its magnetic response did not show the presence of a magnetic dipole detected at the majority of the other structures, thus suggesting a different origin for this crater. Meteoritics. Vol. 10, No. 3. September 30. 1975 209 786 "; •=:■•.' ■■.■ki>$r~ •.}*■• SAfe^I fo • Fig. la Aerial view of Tepexitl, the inner ridge is in the SE quadrant. North is to the top. (Fig. 4, Plate 6, Maupome 1974). 210 787 F F F FF F F F F T .1. XX X X X XX xxxxxxx Little hills of tefra (volcanic ash) Cultivated land Outer wall of the rim Highest part of the rim Inner wall of the rim Ridge Cultivated corn field Fig. lb Elements of the crater (Fig. 6, Maupome, 1974). Closer geological studies have indicated, however, that the origin of the structure is definitely volcanic. The crater is an ash ring composed primarily of volcanic glass (vitric) fragments. These were apparently ejected fairly cold as there is no evidence of welding of the ash into a welded tuff. Although the crater floor has a depression below the regional level, the volume of the rim exceeds the volume of the bowl, indicating that there has been addition of material to the neighboring surface. This is unlike a 211 788 meteorite crater where the volume of the ejecta is essentially equal to the volume of the hole since the volume of the impacting missile is negligible in comparison. The crater wall is made up mostly of dip-slope beds which retain their initial dips, dipping both outward and inward with respect to the rim crest. There is no evidence of inverted or overturned stratigraphy. Beds described earlier by Maupome' (1974) as "overturned" are in fact "draped" over the inside of the crater wall; these, however, are not stratigraphically overturned beds. In addition, at this crater there is no uplifted country rock showing a quaquaversal dip. In fact, no country rocks are seen at all in the rim. The rim is entirely constructional, made up of volcanic ash, which is contrary to the expectation if it were an impact crater. Further arguments arsily of California, Los Angeles, CA 90024 ****National Ocean, and Atmos. Adm., Miami, FL Although previously described as a possible meteorite crater (L. Maupome', Rcvista Mcxicana dc Aslronomia y Astrofisica, V. 1 , 1974), Tepexitl crater (19°13'N, 97°26'W) is definitely volcanic. It should not be included on any list of possible terrestrial impact sites. Tcpexitl crater, east of Puebla, Mexico, and in the Mexican Volcanic Belt, is 1.2 km across and averages 75 m deep below its rim. Our examination reveals that Tcpexitl is an ash ring built by successive magmatic phreatic eruptions of rhyolitic vitric ash. The remarkably symmetric ash ring suggest that the explosions were of the "gun-barrel" type formed under calm wind conditions. The associated magma chamber must have been shallow as only limestone and alluvial gravel xenolii'hs.wcic ejected without any deep crustal or sub-crustal rocks. Abundant ground water or possibly a shallow lake presumably fueled these steam explosions. Tepexitl is but one of a score of volcanic crater-form features in the region. Although it has some unique aspects, which initially suggested that it might be an impact site, it is definitely a member of this regional family of volcanic explosion craters, maars and collapse features. The following evidence opposes an impact origin: 1 , no meteorites were found; 2,no shatter cones were discovered; 3, we could identify no shock meta- morphosed rocks; 4, the rim is entirely composed of ash with no uplifted quaquavcrs.'diy dipping country rock; 5, there arc no overturned beds or overturned flap; 6, ejected blocks arc of small si/c, the largest observed being about 1 m3; 7, the ejected material is entirely volcanic; 8, the volume o( the rim far exceeds thai portion of the crater which is depressed below the regional level, so that much new material has been added to the earth's surface; 9, there is no evidence of a. central dome or ring synclinc; and 10, there is no sucvite, impact melt rock. 792 48 Reprinted from: Marine Geology 18, 47-69. MORPHOLOGY AND SEDIMENTARY PROCESSES IN AND AROUND TORTUGAS AND AGASSIZ SEA VALLEYS, SOUTHERN STRAITS OF FLORIDA LARRY L. MINTER' , GEORGE H. KELLER1 and THOMAS E. PYLE2 ' Atlantic Oceanographic and Meteorological Laboratories, National Oceanic and Atmo- spheric Administration, Miami, Fla. (U.S.A.) 1 Marine Science Institute, University of South Florida, St. Petersburg, Fla. (U.S.A.) (Received June 17, 1974; accepted for publication July 24, 1974) ABSTRACT Minter, L. L., Keller, G. H. and Pyle, T. E., 1975. Morphology and sedimentary processes in and around Tortugas and Agassiz Sea Valleys, southern Straits of Florida. Mar. Geol., 18: 47-69. Continuous seismic reflection profiling and new bathymetry data in the southern Straits of Florida over an area dominated by the Tortugas and Agassiz Valley systems have allowed a more detailed analysis of the morphology and sedimentary processes active in this region. Four dives in the submersible DSV " Alvin" supplement the seismic and bathymetric data. The continental slope in the study area can be divided into two physiographic pro- vinces: (I) an irregular topography controlled by the Florida Escarpment west of Tortugas Valley; and (II) the remainder of the continental slope which contains the majority of features under investigation. Seismic data indicate that the valleys are being filled shoreward of 290 fathoms (530 m) by a wedge of prograding sediments derived from the Florida shelf. The morphology of the two valley systems reflects probable differences of origin. Tortugas Valley appears to have originated coincident with the eastern terminus of the Florida Escarpment and province-I-type topography. The Agassiz valleys may have an origin associated with jointing patterns observed by divers aboard DSV "Alvin". Current meter readings and bottom photographs from "Alvin" indicate that currents are relatively sluggish and not very effective in the transport of sediment within the valleys. An area of undulations west of Pourtales Terrace was investigated and concluded to be erosional in origin. Slumping appears to have played a large part in shaping many features in the study area. The bottom morphology and sediment distribution on the continental slope and in the axis of the Straits of Florida suggest that bottom currents are active in shaping the entire area. INTRODUCTION The southern Straits of Florida separate the Florida-Bahama Platform, an extensive carbonate environment, from the crystalline and terrigenous sedi- 793 48 mentary rocks of Cuba (Hurley, 1964). The southern Straits extend from the Gulf of Mexico on the west to Cay Sal Bank on the east, where the easterly flowing Florida Current turns northward to enter the northern Straits of Florida. An intensive study of an area in the southern Straits of Florida (Fig.l) was undertaken for the purpose of investigating the unusual bottom morphology noted in earlier studies (Jordan and Stewart, 1961; Jordan et al. 1964; Malloy and Hurley, 1970). The submarine topography of this area is dominated by the Tortugas and Agassiz Valleys which were recognized, named, and classified as two separate systems by Jordan and Stewart (1961). For the purpose of identification, a meridian at 82°50'W is considered to separate the Tortugas Valley and its associated tributary system from the Agassiz Valley system lying farther to the east. All of these valleys are unusual in that their relief increases in the downslope direction, unlike most submarine valleys whose heads display the greatest relief. The anomalous character of the valleys, which do not cross the continental shelf, but head Fig.l. Location map showing regional features and area of study within dashed box. (From Uchupi, 1966a.) 794 49 in water depths of about 300 fathoms (549 m), has led to the hypothesis that the heads of these valleys are being infilled with recent sediments transported south and westward around the western end of the Marquesas Keys, located 35 km west of Key West, Florida (Fig.l), (Jordan and Stewart, 1961; Jordan, 1962). Another region of particular interest in the study area is a series of bottom undulations in the vicinity of latitude 24°16'N and longitude 82°22'W (Fig.2). The bathymetric map presented here (Fig.2) is based on soundings made by the USC & GSS "Hydrographer" from 1952 to 1960. The previous studies mentioned above were either lacking in detailed soundings or did not include the complete area of study; consequently, a new base map was constructed. A portion of Kofoed and Jordan's (1964) bathymetry of the Tortugas Terrace was used to complete a small area lacking in sounding density (Fig.3, block A). 83°30' 83°00 Fig.2. Bathymetric map of study area. Contours in fathoms. 795 50 83°50' 83°30' 82°30' 82°20' Fig. 3. Bathymetric map showing physiographic provinces, position of Florida Escarpment in the study area and location of DSV " AJvin" dive sites. (Block A after Kofoed and Jordan, 1964.) Seismic data consist of 10 cubic inch (163.9 cm3 ) air gun and 3.5 kHz profiles run by NOAA's Atlantic Oceanographic and Meteorological Laboratories (AOML) in October of 1971, sparker profiles (160,000 joules) produced by the U.S. Naval Oceanographic Office from the USNS "Kane" in 1969, and 3.5 kHz profiles from the University of South Florida aboard RV "Eastward" in September of 1971. A total of nearly 1500 km of seismic lines were run (Fig. 4). In conjunction with the seismic work we also made four dives with the deep-sea submersible DSV "Alvin". Two dives were into two of the Agassiz valleys and two were made in the area of bottom undulations west of Pourtales Terrace. During the dives numerous bottom photographs and current meter readings were obtained. Short sediment cores of about 30 cm were recovered from each dive site. 796 51 82 30 1 12 TRACKLINES AOML R/V EASTWARD USNS KANE Fig. 4. Chart showing position of tracklines used during study. REGIONAL GEOLOGY Morphology The southern Straits of Florida are asymmetrical in cross-section due to the steeper slopes of the Cuban side of the trough. The narrowest portion, between Key West and mainland Cuba, is 154 km wide and the length of the southern Straits from 84°00'W longitude to Cay Sal Bank is approx- imately 347 km. Morphological differences between the continental slopes of southern Florida and northern Cuba are readily apparent from previously published bathymetric maps (see Malloy and Hurley, 1970, fig. 3). The Cuban slopes exhibit steep escarpments indented by numerous embayments which result in a complex topography. West of 81°17'W there is essentially no continen- tal shelf along the northern coast of Cuba. The Florida margin reveals a wide continental shelf west of the peninsula and a moderately wide shelf, approximately 62 km, from southern Florida across Florida Bay to the shelf break. Just below the shelf break and paral- leling the lower Florida Keys is the Pourtales Terrace, a prominent feature of the southern Straits of Florida. Descriptions of the terrace and its geological setting have been presented by several investigators (Jordan, 1954; 797 52 Jordan and Stewart, 1961; Jordan etal., 1964; Malloy and Hurley, 1970; Gomberg, 1973), and will not be discussed here. The Pourtales Escarpment forms the seaward edge of the Pourtales Terrace, displaying a relief of as much as 150 fathoms (275 m) in places. At the base of the escarpment the bottom flattens and begins a gentle slope toward the axis of the Straits. This gently dipping bottom is interrupted by another scarp, the Mitchell Escarpment, which separates the continental slope from the axis of the southern Straits of Florida. The Florida Escarpment, which marks the outer limits of the west Florida continental slope, can be traced around the Florida Platform and eastward into the study area (Fig.l). The escarpment decreases in relief and depth to its base as it changes trend from south to southeast on entering the southern Straits. From the published bathymetry it can be seen that the continental slope south of 27°N between about 440 fathoms (805 m) and the base of the Florida Escarpment is of an irregular nature (Jordan and Stewart, 1959; Uchupi, 1966b). Embayments, valleys, isolated hills, and other rugged features are noted by irregular isobaths on the lower continental slope. Similar topography extends part way into the study area and can be seen between 440 fathoms (805 m) and the base of the Florida Escarpment eastward to 83°15'W (Fig. 2). The escarpment also terminates at this point. The axial trough of the southern Straits is a westerly dipping channel. Hurley (1964) has suggested that this gradient is due to sediment moving south and west through the Straits, just opposite the flow of the Florida Current. The existence of ripple marks indicative of southward flow along the bottom has been observed in the northern Straits of Florida (Hurley and Fink, 1963). Neumann and Ball (1970) and Duing and Johnson (1971) confirm these observations in portions of the northern Straits. As will be discussed later, bottom currents were observed at 235 fathoms (430 m) flowing in a westerly direction in the southern Straits during this study. A sediment "fan" entering the Gulf of Mexico from the southern Straits is considered by Hurley (1964) as possibly representing bottom transport by westerly moving currents. Seismic data of Pyle and Antoine (1973) show this "fan" to be a topographic feature with outcrops of reflectors at the edge of the sediment body. They interpret the "fan" as being primarily the result of erosional and not depositional processes. Structure Origin of the southern Straits of Florida has been a question for debate among geologists for many years. Agassiz (1888) suggested that the entire Straits of Florida were formed by erosion due to the flow of the Florida Current. Other investigators have ascribed them to graben faulting, subsidence and marginal upbuilding, and non-deposition or erosion coupled with reef growth of upbuilding (Pressler, 1947; Uchupi, 1966a; Antoine and Pyle, 1970). Due to the configuration of the southern Straits, Malloy and Hurley (1970) and Pyle and Antoine (1973) postulated a half-graben theory, with 798 53 downfaulting on the Cuban side and a generally smooth rise on the Florida side. Faulting has been suggested as the probable cause for the Pourtales and Mitchell escarpments (Jordan and Stewart, 1961; Jordan et al., 1964; Uchupi, 1966a). Malloy and Hurley (1970) and Pyle and Antoine (1973) see evidence for faulting along the Mitchell Escarpment but feel that the Pourtales Escarpment may be the result of sediment deposition against the steeper face of an upbuilding reef front. Little work has been completed along the nothern coast of Cuba, however, Khudoley (1967) has shown the probable association of the steep Cuban slopes with some fault zones on the island. The Florida Escarpment has been considered of fault origin by numerous workers (Jordan, 1951; Greenman and Le Blanc, 1956; Jordan and Stewart, 1959). However, magnetic data of Heirtzler et al. (1966), and Gough (1967) indicate that faulting has not occurred in the basement rocks. Antoine et al. (1967), Uchupi (1967), Uchupi and Emery (1968), and Bryant et al. (1969) believe the escarpment is due to upbuilding of carbonate banks during Cretaceous time along the outer edge of a subsiding continental block. The topographic irregularities noted above the Florida Escarpment have been considered a result of faulting, lateral folding, and landslide features by Jordan and Stewart (1959) and large-scale, post-Lower Cretaceous erosion by Antoine and Pyle (1970) along the outer edges of upbuilding reefs along the scarp edge. The smoother upper continental slope appears to have formed by sediment upbuilding and outbuilding once the reefs died and subsided (Uchupi, 1967). LOCAL GEOLOGY The study area is an area of transition between the Gulf of Mexico and the Florida Straits (Fig.l). Here, too, the Florida continental slope of the southern Straits continually changes its strike from east— west to northwest- southeast as the southwestern edge of the Florida Platform is approached. For descriptive purposes the bathymetry may be divided into two con- trasting sections, the axis and the continental slope of the southern Straits. The axis of the Straits is defined as that area below the continental slope break. East of 83°15'W the boundary is determined by onlapping sediments of the axis of the Straits over the steeper dipping continental-slope sediments (Fig.5D). West of this point the continental slope-axis boundary is delimited by the base of the Florida Escarpment. Axis of Straits For the most part the axis of the southern Straits is contoured in 50- fathom (91-m) intervals on Fig. 2. East of 83°15'W, or the terminus of the Florida Escarpment, a 10-fathom (18-m) interval was used to a depth of 1000 fathoms (1829 m) so as to better delineate several features of interest, most notably the lower Tortugas and Agassiz valleys. The lower sections of 799 54 Fig. 5. A. Entire line I-J. B. Prograding continental shelf above Tortugas Valley showing foreset bedding. The upper escarpment and probable reverse fault structure can be seen at the base of the prograding strata. C. Enlargement of "step-like" features. D. Axis- continental slope boundary east of Tortugas Valley. Distance between horizontal lines is equal to 40 fathoms (73 m). these valleys are here defined as those portions which exist in the axis of the Straits while the upper valleys are on the continental slope. The sediments of the axis are younger than those of the outer continental slope as can be seen by Fig.5D. Here the axial sediments butt against the steeper dipping continental slope beds and overlay the continuing sequence of continental-slope deposits. The appearance of the zone of contact here between the slope and axis deposits does not indicate that the axis is receiving sediments from the continental slope. The sharp contact and the flat-lying appearance of the axis beds are more indicative of deposition per- pendicular to the continental slope, or in an east— west direction. Bottom current activity, although poorly known in the Straits of Florida, is probably of great importance in shaping the axial as well as the continental-slope deposits. 800 55 Continental slope From bathymetry alone the continental slope in the study area can be divided into two physiographic provinces (Fig.3): (I) the irregular topography from 440 fathoms (805 m) to the base of the Florida Escarpment west of Tortugas Valley; and (II) the remainder of the continental slope in Fig. 2. Province I. The upper Tortugas Valley defines the easternmost extent of this province. Along the valley axis at 880 fathoms (1609 m) the transition from the upper to lower Tortugas Valley coincides with a westward shift of 20 km of the province boundary away from the valley (Fig.3). The boundary, at the 800-fathom (1463-m) isobath at this distance from the lower valley, then turns south and crosses the maximum slope of the area to 1000 fathoms (1829 m) depth. This marks the position considered the termination of the Florida Escarpment. The continental slope of province I is not as well covered by seismic pro- filing as is province II. However, the seismic data available in province I reveal a differing configuration of the subsurface continental slope sediments. Line AA-BB is most interesting since it provides the best penetration of the sub- surface. As can be seen in Fig.6, the surface of the continental slope east of Tortugas Valley (province II) is conformable, to the depth of penetration, with the deeper beds. West of the valley in province I structural deformities can be seen in the subsurface. Irregular bottom features, faulting, and warping of the beds are apparent. A major unnamed valley can be seen near the west- ern edge of province I (Fig. 2). The subsurface structure appears to be the controlling factor of the present continental slope morphology in province I. Province II. The continental slope in province II includes the majority of features under investigation in this study. The Tortugas and Agassiz valley systems and the area of undulations west of Pourtales Terrace are the most pronounced and will be discussed in later sections. Other features of interest in province II are described below. Fig.6. Line AA-BB relating province-I-type topography west of Tortugas Valley (T) on the left to province-II-type topography east of the valley. 2.0 seconds travel time equals 800 fathoms (1463 m) depth. 801 56 The continental-shelf break in the study area occurs at about 40 fathoms (73 m) (Fig. 2). The average gradient of the sea floor below the break is 25 m/km or 1.5°. At about 100 fathoms (183 m) the bottom slope increases and at 140 fathoms (256 m) it reaches 3°. The reason for this increase is that progradation of material from the shelf has resulted in an outbuilding of the shelf edge which can be clearly seen in seismic profiles (Figs.5B, 7). Deposition of this nature has been noted in the Florida Straits and around the Florida Platform (Uchupi, 1966a; Rona and Clay, 1966; Uchupi and Emery, 1968; Bryant et al., 1969; Malloy and Hurley, 1970). Uchupi and Emery (1967) have also noted this pattern on the Atlantic continental margin. Seismic profiles reveal a prograding sequence exhibiting foreset and bottom- set beds. Jordan and Stewart (1961) have postulated that great amounts of sediment are being transported from Florida Bay and the west coast of Florida onto the continental slope west of the Marquesas Keys, enough to account for the smooth slopes from 25 (46 m) to 300 fathoms (549 m). Milligan (1962) has suggested that some sediment is also being introduced from the Gulf of Mexico by the Florida Current. The smoothness of the upper continental slope may be the result of the Florida Current touching bottom at this depth as described by Jordan et al. (1964). D. N. Gomberg (personal communication, 1974) has concluded that the Florida Current is active in an easterly flowing direction, at times, to a depth of at least 120 fathoms (200 m). Below the prograding sequence of the upper continental slope, the sea floor east of Tortugas Valley is remarkably flat with slopes of less than 1° to a depth of about 540 fathoms (988 m). No discontinuities of the con- tinental slope can be detected other than the valleys. At 540 fathoms (988 m) the gradient of the continental slope progressively increases until the slope break is reached. The maximum gradient noted is 4° near the base of the continental slope on both sides of the Agassiz valley system (Fig. 2). Between the Tortugas and Agassiz valleys the slope break occurs at 850 fathoms (1554 m). East of the Agassiz valleys the break has decreased in depth to 780 fathoms (1426 m). o.o Ul < o 2.0 5.5kr Fig. 7. Prograding shelf sequence above Agassiz valleys along line H-G. Note absence of upper escarpment at this position. Depth between horizontal lines equals 40 fathoms (73 m). 802 57 Coincident with the increase in slope at 540 fathoms (988 m) between the two valley systems, a series of step-like features are observed on sub-bottom profiles I-J and D-E (Figs.5C, 8). A series of at least five and possibly six steps can be detected on the cont; ental slope. The relief of the steps is relatively constant, on the order of 10—12 fathoms (18—22 m), and four of them can be tentatively correlated across the area by their depth from differ- ing profiles. The origin of these step-like features is probably related to the increased slope just above the shallowest step. The beds beneath the steps appear to be continuous on the seismic records, therefore faulting is ruled out. It is possible that, as the Florida slope subsided, the outer continental slope subsided more rapidly, creating zones of weakness in the upper strata. These weaker zones may then have slumped along their lineations, thus producing the "steps". PROMINENT FEATURES Three prominent areas in the study region constitute the principal objec- tives of the study: (1) The Tortugas valley system; (2) the Agassiz valley system; and (3) the area of bottom undulations near the western end of Pourtales Terrace. The primary object of the study was to investigate the morphology of these features and to attempt to understand their formation and the dynamic processes presently associated with these features. SW NE 0.0 iu y> ui I— >- < :5.5km; 2.0 , ^.m^. »i.-J-ii(.»A m. *~- «'.iy-:»?>te^ Fig. 8. The two westernmost valleys along line D-E. Note apparent slumping in valley on right and possible secondary channel in the valley at left. Farther to southwest are additional "step-like" features. Vertical distance between horizontal lines is 40 fathoms (73 m). 803 58 Tortugas valley system The Tortugas valley system is composed of the major Tortugas Valley and at least nine tributary valleys (Fig. 2). Two local escarpments, one at 280 fathoms (512 m) and the other at about 440 fathoms (805 m), are major features within the system. The visible length of the entire Tortugas system is about 93 km, slightly more than half of that distance being on the continental slope. Tributary Valleys. The Tortugas tributary valleys appear to be governed by the two escarpments. The narrow, steep-walled tributary valleys originate at the base of the upper escarpment at a depth of 310 fathoms (567 m) (Fig. 2). From this locality the valleys run courses which are perpendicular to the regional slope. With the exception of the easternmost valley, which appar- ently does not enter the Tortugas head, the tributary courses are generally difficult to define once they pass seaward of the lower escarpment. Jordan and Stewart (1961) consider the easternmost valley to be younger than the Tortugas Valley because of its alignment with the maximum slope of the surface beds, which they believe postdate the formation of the Tortugas Valley. Bathymetry reveals that the tributary valleys change their morphological characteristics as they transgress downslope (Fig. 2). Near the base of the upper escarpment the tributaries are relatively shallow topographical expressions, but as they progress downslope toward the Tortugas Valley head, the width/depth ratio decreases rapidly. Near the lower scarp the valleys show near vertical walls with narrow widths. These deep channels dissipate as the lower scarp is crossed. The Tortugas tributary valleys show evidence that they have maintained their position on the continental slope while the walls have been built up (Fig. 9). The situation may possibly be much like that described off Cape Hatteras where Rona (1970) suggests that the intercanyon areas and valley walls are being deposited while currents act to keep the channels open. Some of the tributary valleys can be seen to possess fill while others are relatively clean. This implies that sweeping of the valleys is intermittent, or perhaps that some of the valleys are inactive now and are being filled. Escarpments. The two escarpments in the Tortugas valley system, as men- tioned above, define the limits of the Tortugas tributary valleys. Along lines I-J and U-V (Figs.5B, 10) the upper escarpment is visible at a depth of about 280 fathoms (512 m). This escarpment can be seen in Fig. 2 and has a max- imum relief of 30 fathoms (55 m). Line I-J (Fig.5B) provides the only evidence as to the possible origin of the scarp. This profile reveals that reverse fault- ing may have occurred with the bottomset beds of the prograding sequence overriding the seaward sediments at depth. The lower escarpment, which begins at about 440 fathoms (805 m), is considered here to define the northern limits of the Tortugas Valley since the 804 5<) Fig. 9. A. Complete line A-B. B. Section showing buried Agassiz valleys and upper escarp- ment at extreme right. C. Tortugas tributary valleys below upper escarpment. Note the building up of the channel walls and intercanyon areas while valleys have maintained their position. The same process can be observed in the subsurface along section B. D. Section across upper head of Tortugas Valley. The large valley at right is not delineated well by bathymetry. A buried valley can be seen to the right of this valley. 40 fathoms (73 m) between horizontal lines. main valley cannot be traced shoreward of the scarp's position. Relief along the lower escarpment varies from 90 fathoms (165 m) near the west end to 40 fathoms (73 m) at the eastern end (Fig. 2). The depth and shape of the lower escarpment suggest a relationship associated with both the Florida Escarpment and the Tortugas Valley. The depth at which the lower escarpment originates is the upper limit of province-I-type topography. Additional scarps can be seen to the west of this escarpment at the same depth (Fig. 2). The shape of the escarpment around the head of Tortugas Valley implies a genetic relation to the valley head. Two possible explanations of origin can be deduced from this infor- mational) the scarp represents the eastern upper end of province-I-type topography; or (2) the scarp has been formed around the Tortugas Valley head. Seismic profiles reveal no evidence of fault origin for this escarpment. It has already been shown that the Tortugas Valley marks the eastern end 805 60 0.0. w Ilk m= U->- £ 4.0 Fig. 10. Seismic record along line U-V-W. Note blocky material on the surface to left of point V. This material is considered to be a slump mass derived from the continental shelf toward point W. 2.0 sec. travel time corresponds to 800 fathoms (1463 m) depth. of province-I-type topography. That the escarpment transgresses into province-II-type topography argues against it being a result of upper province-I irregularities. Also, since the escarpment forms an approximate semi-circle around the Tortugas Valley head, it is felt that the second explana- tion is more plausible. It is probable that this scarp was initiated as a slump scar by slow creep of the continental slope sediments into the head of Tortugas Valley. The tributary valleys are younger than the scarp and continental slope, as evidenced by their perpendicularity to both the maximum slope of the continental slope above the escarpment and the escarpment itself. Most of the tributary valleys never cut deep enough into the continental slope to reach the base of the escarpment and, therefore, appear to terminate at the escarpment. Tortugas Valley. Tortugas Valley originates in what appears to be a pseudo bowl-like area bordered by the steep escarpment at 440 fathoms (805 m) to the north. Below the escarpment gently dipping beds converge to form a steep V-shaped trough with a minimum relief of 40 fathoms (73 m) from the top of the valley walls. The average gradient of the valley is 13 m/km (Jordan and Stewart, 1961). In contrast to the tributary valleys, the major valley broadens downslope. This broadening is accompanied by a gentle curve to the east. Along this curving section the walls of the valley are relatively steep. At a depth of 880 fathoms (1609 m) there is a rapid decrease in wall height which corresponds to the transgression of the valley from off the continental slope 806 6 1 into the axis of the Straits. The lower Tortugas Valley can be seen to be wider and shallower than the upper valley (Fig. 2), perhaps due to a difference in sediment characteristics between the continental slope and axis as suggested by the continental slope-axis boundary (Fig.5D). Profiles crossing the upper part of Tortugas Valley reveal several changes taking place in a downslope direction (Figs. 6, 11). The channel becomes wider and deeper with depth. Along lines AA-BB and 10-11 (Figs. 6, 11C) the bottom of the channel can initially be seen without sidewall reflection interference. From these profiles it appears that faulting and slumping are important factors in the present morphology of the valley. Profile AA-BB w 1.0 w -8.3km- 2.0 V 8.3km- (Al (Bl w 10 V* ^ 7.4km ICl s o 5 /"— 2 0 Fig.ll. 3.5 kHz profiles across Tortugas Valley showing changing characteristics of the valley downslope. A. Portion of line 8-9. Valley walls are very steep. An unidentified valley exhibiting natural levees can be seen west of the main channel. B. Crossing of Tortugas Valley along line 9-10. Valley is wider than in A. The easternmost tributary valley can be seen east of the main valley. C. Tortugas Valley along line 10-11. "Step- like" features can be seen on the eastern side. Notice apparent slumping along valley walls and orientation of bedding planes in the axis suggesting mass movement. 2.0 sec. travel time equals 800 fathoms (1463 m) depth. 807 62 (Fig. 6) appears to show faulting along both sides of the valley in the sub- surface. Line 10-11 (Fig.llC) shows evidence of fault blocks on the surface as well as slump features along the western side of the channel bottom and the eastern channel wall. The origin of the Tortugas valley system is considered to be related to its position in respect to the Florida Escarpment. The escarpment can be traced no farther eastward than 83°15'W. The typical irregular topography above the Florida Escarpment, physiographic province I (Fig.3), can be seen to extend eastward to the upper Tortugas Valley. As seen from Fig.6, the deformed sediments west of the valley are in marked contrast to the con- formable units to the east. It is probable that this transition line may then have been a zone of weakness, thus a likely place for the valley to form. Agassiz valleys system The Agassiz valleys, as a system, consist of at least ten north— south trending valleys on the Florida continental slope and a lower main valley which exists in the axis of the Straits (Fig.2). The lineal extent of the system, about 95 km, closely approximates that of the Tortugas valley system. In addition to the bathymetric and seismic profiling work in the area, two dives (numbers 354 and 355) in the submersible DSV "Alvin" were under- taken in the two westernmost valleys of the Agassiz system (Fig.3). Each dive lasted about 6 hours As Jordan and Stewart (1961) noted, the Agassiz valleys do not have the dendritic pattern of the Tortugas system but maintain nearly parallel trends throughout most of their length. However, at least three pairs of the valleys do appear to merge to form single channels. The two easternmost valleys merge at 490 fathoms (896 m), the westernmost two valleys merge at about 640 fathoms (1170 m), and the two valleys just east of these also unite at 490 fathoms (896 m) (Fig.2). The escarpment which coincides with the heads of the Tortugas tributary valleys does not extend eastward to the Agassiz valleys although both systems seem to originate at about the same isobath (290 fathoms; 530 m). Seismic profile A-B (Fig.9B) reveals that the heads of the Agassiz valleys are being filled in at this depth. A profile downslope, D-E, shows the changing patterns of the valley morphology (Fig. 8). All of the valleys tend to increase in relief downslope to the 520-fathom (951-m) isobath. Below this depth only three valleys remain visible (Fig.2). The three valleys continue their increasing relief to 650 fathoms (1189 m) where they terminate abruptly. South of the Agassiz valleys and below 640 fathoms (1170 m), isobaths are generally irregular and not consistent with the surrounding submarine topography (Fig.2). Profile V-W (Fig. 10) suggests a possible explanation. Between point V and 780 fathoms (1426 m), or 1.95 sec, the surface of the axis deposits is not smooth and continuous but displays a disorderly structure. At 780 fathoms (1426 m), or 1.95 sec. in Fig. 10, the smooth dipping conti- nental-slope beds are seen emerging from beneath the sediment fill. It 808 63 appears that this zone of irregular topography and substrate may be a localized slump mass, having initiated on the continental slope and transgressed out onto the axis of the Florida Straits. Initiation of such a slump would account for the sudden termination of the valleys at 650 fathoms (1189 m). Another valley heading at 600 fathoms (1097 m) between the two "slump" terminated valleys continues its path farther downslope than the 650-fathom (1189-m) contour. The fact that this valley has a wide, rounded head and a different appearance than the other valleys suggests that it is anomalous to the other valleys. It may be a younger feature formed after the slumping occurred. The lower Agassiz Valley begins at a depth of 850 fathoms (1554 m) and crosses the axis to a depth of 990 fathoms (1810 m) (Fig. 2). The width and axial gradient of this valley are very similar to those of the lower Tortugas Valley, and it seems reasonable to assume that the relationship of the lower valley to the upper valleys of each system is similar; they are the axial transgressions of the valleys which cross the continental slope. The relation- ship between the lower Tortugas Valley and the upper valley system is obvious from Fig. 2, but the lower Agassiz Valley has no such tie. A solution to this anomaly can be presented if we conclude that the two westernmost valleys of the Agassiz valley system were terminated by a local slump at 650 fathoms (1189 m). Upslope projection of the lower Agassiz Valley is found to coincide with the downslope projection of the two western valleys (Fig. 2). It would seem justifiable to presume that these two upper valleys merged and were connected to the lower Agassiz Valley prior to the slumping which obliterated all evidence of the one time valley connections between 650 (1189 m) and 850 fathoms (1554 m). Two dives with DSV "Alvin" reveal some of the differences in bottom morphology at various locations along the two westernmost Agassiz valleys. Dive 355 reached bottom at 477 fathoms (873 m) in the second valley from the west (Fig. 3). Observers described the area as a gently undulating bottom, not anything like a valley, and having the appearance of a field blanketed in snow. Fine gray carbonate mud, silty-clay, makes up the bulk of the sediment, the remainder being mostly pelagic foraminifers, pteropod fragments, ostracods, and some sponge spicules. The bottom is relatively flat with no indications of scour marks or striations, suggesting that currents are playing a very minor role in transporting sediments along the bottom at this location. A number of pits of burrowing organisms and tracks, probably those of crabs, also appear undisturbed in the fine bottom deposits. No rock out- crops were seen in the area. The sea floor adjacent to the valley had an ap- pearance similar to the axis. Dive 354 reached the bottom at 607 fathoms (1110 m) in the axis of the westernmost valley (Fig. 3). The bottom morphology differed from that of dive 355 in many respects. Slabs of chalky limestone, too friable to sample, were found in the canyon axis, as were boulders of hard rock 2—3 m in diameter. A small ridge could be seen extending across the valley axis. At the ridge crest a semi-consolidated brownish gray rock protruded through the mud cover of the valley floor. 809 64 The walls of the valley at this location were 55 fathoms (100 m) above the valley floor. The upper portions were vertical rock, too hard to be sampled, which appeared to be coralline. The lower halves of the walls had slopes of 20°— 30°. Distinct contacts could be noted where the walls began to appear Vshaped, indicating a possible strata change at this level. However, observers felt that the cause of this contact may have been due to sedimenta- tion and talus accumulation from above. The vertical walls trend in a north— south direction and have the appear- ance of being fault or joint controlled due to their straight alignment. Observers aboard "Alvin" believed that jointing could be seen in the limestone walls. Subsequent inspection of seismic line D-E (Fig. 8), which crosses the valley near the position of dive 354, and other profiles over the Agassiz valleys reveal no indication of faulting, although slumping is recogniz- able on the records. Dive 354 revealed some channel axis characteristics similar to dive 355. Bottom conditions appeared to be depositional rather than erosional. Again photographs supported the suggestion that the valleys are filling as no scour marks or striations could be detected. Observers aboard "Alvin" noted material being introduced into the valley axis over the edge of the walls. This move- ment of sediment outside the valleys, although of unknown velocity and direction, is indicative of bottom currents which are probably aiding in shaping the general sedimentation pattern of the entire study area. This is also a direct observation of one method of valley fill in addition to fallout of fine material in suspension and tests of organisms. Dives 355 and 354 provided short-term current measurements that aid in the interpretation of sediment transport within the valleys. Three bottom current readings were taken on dive 355 with values ranging from 0—9.2 cm/sec. (0.18 knots). The current flow was always to the south in the valleys. Dive 354 made one current meter reading of 2.5 cm/sec. (0.05 knots) downcanyon. These readings, together with the photographs, indicate little transport within the valleys. Sediments of dive 354 at 607 fathoms (1110 m) recovered by the "Alvin" corer were slightly different than those farther upslope on dive 355. The deeper sediments contained a greater amount of pelagic foraminifers than those farther upslope. This should be expected as the carbonate muds, some of which originate from shore, decrease in a seaward direction and the fallout of foraminifers becomes a more important source of sediments. Local slumping has probably played an important role in the introduction of sediments into the valleys at the position of dive 354. The large blocks of limestone and boulders described by divers aboard "Alvin" are probably part of the walls which have broken off along jointing planes. That these features are being covered by fine sediments suggests that the current meter readings, although of only short duration, are representative of current activity in the valleys today. However, short-term current readings must be regarded with skepticism as the valleys may be periodically scoured or filled. The Agassiz valleys, between 320 (585 m) and 450 fathoms (823 m), show 810 65 the same evidence of maintaining their position on the continental slope while the walls and intercanyon areas have built up as the Tortugas tributary valleys. The aforementioned suggestion by Rona (1970) is concluded to be the same for both systems. This pattern can be seen in the Agassiz valleys in the subsurface of line A-B (Fig.9B) although the channels are now being filled by flat-lying sediments derived from progradation of the shelf edge. Although firm conclusions as to the origin of Agassiz valleys cannot be made at this time, the best evidence is probably provided by observers aboard the submersible DSV "Alvin". They noted jointing patterns in the walls of the valley on dive 354 and it is possible that these jointing planes produced zones of weakness in the rock which were later eroded to form the valleys. Their straight alignment along the continental slope is suggestive of some type of structural control. Undulation area An area of slight undulations located at the western end of Pourtales Terrace, 24°16'N and 82°23'W (Fig.2) has been investigated by Jordan and Stewart (1961) and Jordan et al. (1964). Both works have described the morphology of the features, noting that they are asymmetrical in cross-section with the steep sides facing east. The crest to crest length varies from 900 m to 1150 m and the crest to trough depths range from 7 (13 m) to 15 (27 m) fathoms. Like the valley systems, the undulations trend parallel to the regional slope. Jordan and Stewart (1961) considered these features to be step faults associated with faulting on the southern edge of the Pourtales Terrace. Jordan et al. (1964) see no evidence in seismic profiles to support the fault hypothesis nor do we. They suggest that the undulations are giant sand waves created by an easterly flowing bottom current associated with the swift- moving surface flow of the Florida Current. Their evidence can be seen in their fig. 16. This figure shows a westerly thickening of unconsolidated bottom sediments at the western edge of the Pourtales Terrace indicating that an eastward setting current has been active in sediment transport over this end of the terrace. Seismic profiles and the results of two dives (numbers 356 and 357; Fig. 3) by DSV "Alvin" were used in an attempt to determine what proces- ses are active in the area today. Profile aa-bb was run in an approximate west to east direction (Fig. 4). A strong reflector below the surface strata is apparently the western extension of the Pourtales Terrace buried beneath a westerly thickening overburden (Fig. 12). Line bb-cc was run nearly perpen- dicular to the undulations (Fig.4). Again westerly sediment thickening can be seen above a strong basement reflector (Fig.12), supporting Jordan et al.'s (1964) contention that an easterly flowing current was responsible for covering the western end of the Pourtales Terrace. A total of seven current measurements were taken on dives 356 and 357 in this area. The range of values was 9.2 cm/sec. (0.18 knots) to 23.4 cm/sec. 811 .,<; 0.0 bb Fig.12. Lines aa-bb and bb-cc. A strong reflector 0.1 to 0.2 sec. below surface along aa-bb is probably the buried western end of Pourtales Terrace. Sediment thickening is to west. Undulations west of Pourtales Terrace seen in cross-section along bb-cc. Strong reflector can be seen 0.2 to 0.3 sec. below undulations. Features appear erosional. Vertical distance between horizontal lines equals 40 fathoms (73 m) depth. (0.43 knots) all in a westerly direction with one exception, an 11.4 cm/sec. (0.24 knots) flow to the south in the trough of one undulation. The depth of both dives was approximately 235 fathoms (430 m). The observations of D. N. Gomberg (personal communication, 1974) and the results of these current measurements indicate that bottom flow reverses somewhere between 120 (200 m) and 235 fathoms (430 m). From seismic data it would appear that the prominent direction of flow is, or has been in the past, toward the east over the entire depth range (Fig.12). This information suggests that bottom currents do reverse at times in the southern Straits. The depth range over which these currents reverse cannot be defined by present data, but is most likely greater than the aforementioned range. No attempt at relating southerly flow in the northern Straits and westerly flow in the southern Straits has been attempted from these measurements. The southerly flowing current measurement leaves open the possibility that these undulations are valley or erosional features. Although easterly flowing currents are responsible for covering the western end of Pourtales Terrace it does not appear that these features are sand waves as suggested by Jordan et al. (1964). From Fig.12 it can be seen that the axes of the undula- tions are considerably lower than the depth of the surrounding continental slope. The crests are a little, if any, higher than the local topography. Farther downslope seismic profiles show them to be lower. For these reasons it seems unacceptable to conclude that the features are depositional in origin. The fact that they incise the prograding sediment mass is evidence of their youth relative to the Tortugas and Agassiz valleys. Submersible pictures and observers' reports indicate that this area is one of present deposition. 812 67 Evidence is not conclusive as to their origin and more data, especially con- cerning currents, need to be gathered before further speculation is warranted. The sediments obtained from dives 356 and 357 differed from those of 354 and 355 in particle size and content. Sediments of the undulation area are coarser than those in the valleys, as would be expected if they were being derived mainly from shoreward sources. Sponge spicules are more abundant and diverse as well as mollusc fragments and bits of coral. Foraminifera and pteropod fragments are still abundant but are of less proportion to the sand- size fraction of the sediments than in the deeper water samples. CONCLUSIONS The area of investigation is one of transition from a relatively narrow strait to a deep ocean basin. The new bathy metric map (Fig. 2) and seismic profiles reveal the relationship between and the possible origin of several morpholo- gic features. Of primary interest are the Tortugas and Agassiz valley systems. The Tortugas tributary valleys and the Agassiz valleys are aligned with the maximum slope and may have formed contemporaneously. The southern Florida continental-shelf edge has been built out and upward resulting in the filling of the head of the Agassiz valleys shoreward of 290 fathoms (530 m). The Tortugas tributary valleys originate at the base of an apparent fault-controlled escarpment at the base of the prograding material from the shelf edge. Undulations incising this prograding sequence are believed to be erosional in origin. The Florida Escarpment has been shown to be associated with a unique morphology between approximately 440 fathoms (805 m) to its base from 27° N to its termination at 83°15'W longitude. Above this depth the conti- nental slope is generally smooth with few irregularities. Bottom currents are probably playing a large role in the shaping of the continental slope and axial sediments. Bottom currents between 120 (220 m) and 235 fathoms (430 m) apparently reverse at times and it is probable that this depth interval may be greater. Over-all, the study area is considered to be one of deposition, although some erosional processes, mainly slumping, appear to have been important in the formation of a number of features. Some valleys may also have intermittent scour. The full extent to which erosional processes are active today require additional study. ACKNOWLEDGEMENTS The writers acknowledge with many thanks the officers and crews of the RV "Lulu", RV "Gosnold", RV "Eastward" and DSV "Alvin" for their most valuable assistance. Support in the field operations by George Lapiene and Douglas Lambert (AOML); F. Kelly (RV "Eastward"); D. Wallace, V. Maynard, R. Clingan, J. McCarthy and P. Jones (USF); A. Ekdale, E. McHuron and J. McCrevey (Rice); and G. Griffin (University of Florida) are gratefully acknowledged. Funds for the AOML field operation were provided by the 813 68 NOAA, Manned Undersea Science and Technology Office. RV "Eastward" time was made available to the University of South Florida by NSF through the Oceanographic Program of the Duke University Marine Laboratory. REFERENCES Agassiz, A., 1888. Three Cruises of the Blake. Houghton, Mifflin, New York, N.Y., 314 pp. Antoine, J., Bryant, W. R. and Jones, B., 1967. Structural features of continental shelf, slope, and scarp, northeastern Gulf of Mexico. Bull. Am. Assoc. Petrol. Geologists, 51:257-262. Antoine, J. W. and Pyle, T. E., 1970. Crustal studies in the Gulf of Mexico. Tectono- physics, 10:477—494. Bryant, W. R., Meyerhoff, A. A., Brown, N. K., Furrer, M. A., Pyle, T. E. and Antoine, J. W., 1969. Escarpments, reef trends, and diapiric structures, eastern Gulf of Mexico. Bull. Am. Assoc. Petrol. Geologists, 53:2506—2542. Duing, W. and Johnson, D., 1971. Southward flow under the Florida current. Science, 173:428-430. Gomberg, D. N., 1973. Drowning of the Floridian Platform margin and formation of a condensed sedimentary sequence. Geol. Soc. Am. Bull., Abstr. Progr., p. 640. Gough, D. I., 1967. Magnetic anomalies and crustal structure in eastern Gulf of Mexico. Bull. Am. Assoc. Petrol. Geologists, 51:200—211. Greenman, N. N. and Le Blanc, R. J., 1956. Recent marine sediments and environments of northwest Gulf of Mexico. Bull. Am. Assoc. Petrol. Geologists, 40:813—847. Heirtzler, J. R., Burckle, L. H. and Peter, G., 1966. Magnetic anomalies in the Gulf of Mexico. J. Geophys. Res., 71:519—526. Hurley, R. J., 1964. Bathymetry of the Straits of Florida and the Bahama Islands, Southern Straits of Florida. Bull. Mar. Sci. Gulf Carib., 14:373-380. Hurley, R. J. and Fink, L. K., 1963. Ripple marks show that countercurrent exists in Florida Straits. Science, 139:603—605. Jordan, G. F., 1951. Continental slope off Apalachicola River, Florida. Bull. Am. Assoc. Petrol. Geologists, 35:1878—1933. Jordan, G. F., 1954. Large sinkholes in Straits of Florida. Bull. Am. Assoc. Petrol. Geologists, 38:1810-1817. Jordan, G. F., 1962. Submarine physiography of the U.S. continental margins. U.S. Dept. Comm. Tech. Bull., 18:28 pp. Jordan, G. F. and Stewart, H. B., 1959. Continental slope off southwest Florida. Bull. Am. Assoc. Petrol. Geologists, 43:974—991. Jordan, G. F. and Stewart, H. B., 1961. Submarine topography of the western Straits of Florida. Geol. Soc. Am. Bull., 72:1051-1058. Jordan, G. F., Malloy, R. J. and Kofoed, J. W., 1964. Bathymetry and geology of Pourtales Terrace, Florida. Mar. Geol., 1:259—287. Khudoley, K. M., 1967. Principal features of Cuban geology. Bull. Am. Assoc. Petrol. Geologists, 51:668—677. Kofoed, J. W. and Jordan, G. F., 1964. Isolated fault scarps on the continental slope off southwest Florida. Southeastern Geol., 5:69—77. Malloy, R. J. and Hurley, R. J., 1970. Geomorphology and geologic structure: Straits of Florida. Geol. Soc. Am. Bull., 81:1947-1972. Milligan, D. B., 1962. Marine Geology of the Florida Straits. Thesis, Florida State Univ., Tallahassee, Florida, 120 pp. Neumann, A. C. and Ball, M. M., 1970. Submersible observations in the Straits of Florida: geology and bottom currents. Geol. Soc. Am. Bull., 81:2861—2874. Pressler, E. D., 1947. Geology and occurrence of oil in Florida. Bull. Am. Assoc. Petrol. Geologists, 31:1851-1862. 814 69 Pyle, T. E. and Antoine, J. W., 1973. Structure of the west Florida Platform, Gulf of Mexico. Texas A and M Univ., Tech. Rept, 73-7-T: 168 pp. Rona, P. A., 1970. Submarine canyon origin on upper continental slope off Cape Hatteras. J. Geol., 78:141—152. Rona, P. A. and Clay, C. S., 1966. Continuous seismic profiles of the continental terrace off southeast Florida. Geol. Soc. Am. Bull., 77:31—44. Uchupi, E., 1966a. Shallow structure of the Straits of Florida. Science, 153:529—531. Uchupi, E., 1966b. Map showing relation of land and submarine topography DeSoto Canyon to Great Bahama Bank. U.S. Geol. Surv., Misc. Geol. Inv. Map 1-47 5, 1 sheet. Uchupi, E., 1967. Bathymetry of the Gulf of Mexico. Trans. Gulf Coast Assoc. Geol. Soc, 17:161-172. Uchupi, E. and Emery, K. O., 1967. Structure of continental margin off Atlantic coast of United States. Bull. Am. Assoc. Petrol. Geologists, 51:223—234. Uchupi, E. and Emery, K. O. , 1968. Structure of continental margin off Gulf coast of United States. Bull. Am. Assoc. Petrol. Geologists, 52:1162—1193. 815 +-) sz ^- S- .—I CL I O) CO C£ .-I 1 ; E oj a 40 iT> p 4T P co p P o 0) P aj a.o bed *^ ID >-i a) £ P tj P H U5 p o q o co £ x , bc+j q 0! ■H & o) ^ o u id a H ^p-ucej p: P 0 P rd XZ O CO CO p 1 1 i — ! o p oo c x: ,G o >-, -P-r-t P O, q ■a ■<3) 4P a P tj p St.- • 4- p xz pu^ co co b:p p:cq P p P 0 ■P (70 CD q QJ-H P P • * H \HM 2 P aCO OJ J & 4n f-i rd C co cr • • O O TJ'O 0 ;H £2 s CD r >-< P P r-j 4h 4h P ^ 3 rrj 0 r« *^ H gson^ q 0 a-H^ k c & DC co OO s CD •H OO ri^ a OP 4-1 CO LO CJ — I Ph rd a) co CJ.— ;-h X B o 'd ^XZ cj-h Ph oj ■ -P »H rH -H +J O >. ro 4-J U: rt p cu -h rd >h a, w P 0 ^ q co un ru e s CO-H > Q 0 CO CJ O -rJ Ph O 'C • '■. Hlfi£ (1) -!-> 1?; ,_) r rd Cc-irit! 0 ' r~i -t-J 4J O OJ • ,Q pi H ■fa C-H UJJTJ ro a; co - rd o cu en CJ •r-l C >, P 0! Cfi-H P •H " •m CO O 4->r-i - P Qij t-> c: p co .—i co T-i 0 *- CH 4' C CO (li o H >, O^r, r. P u rd w C/J P ro a o s q P Oj "j P Ti r3 Q O c c a-H 00 0 rd (;.: O p^ t; -P • QJ %"* u ;H L- t) >>fU rd CO o O-rt J-1 +J co aO gj •H lO P O cj •-: co co p ^ CJ CJ 4-> ^ > rd o ,y c: co .—: c >tj rd rj ?v o P 'P _ q a p c o. 40 t) 6 "'-r3 "^-!-J ^ hf 1 -j f-i O-'J C) OO co 0 q p (D CJ 4-J ^ C> C-H CO m Q C) W rd ' d -^ P' "e o d 4-! I> l>> C V-- £* " |— j :-! s ^ ^-h i'Vj '0 £4 rd 4-i " 0 Co 0£ CJ r. O Ei JJ P 4-J o LO C -^ ro ri o rH co goo • x' -_j > ■ r— ' rP ">■< rd 4-1 frj q •> rr ;■ oo o cj £ T-i & ■d — 4-J P P co J. J p O X >> C -' rd CJ Lo f O CO CD Xj rd CO 4-1 ■§ ?3 P rj 4_> 4_. jz C O O PrC o o i — i P 00 4_i P • H C)-") OU rd-H <; '>/ CO 4^ i-i Ci, rd O O CO >^~-. o^ CO rrj CJ 0) •H ro rrj CO TO U _ C C cCbO •P O >^rrj CO ' CJ CJ rCH -P O rd O rd P C4-!40 CO H -H CJ C; rH 00 4=1 rrj C P o o •H •P >^o P rrj H Ph H P OP ^ P. o c >d ^ o co p OJ 42 P £ p(h rd' PO rd CJ P CO §6 'rJ P 0 P n C p t3.',q fc p o a o co P 00 43 O CO OJ co P O bij ci) ■:-■ .c. x P CJ ^ p o p o q CO ,P I i 0 co Oh Ph rd X : lO - M JZ O P/H P P1 ' cr1 4-J ?4 ^ 00 CO • Ph *~2 O CO*^ 00 CJ rH HO C P CD vm O P P.-H P O P P 5 S CJ, o co P fc: O O P O H-r4 p C :><: p p co o P P 0 CO U H oo p-h q q Ph P ■ P CD E CO; OP O Ph P SJ 431 ro H P P > P CO ^P rrj O 40 P r— ; o CO 'm a c CO p CO p p H d a p p o o S P CJ 4T 2: O i bi O P CO OP oO f0 CJrH X o P G p p O H p p p 40 "a • no. -.-4 P 1 4h H o o J a d CP P ■ H p P o p CO X CO 1 — i a P ■H rH r* o p.p rrj t% PH 816 50 Reprinted from: Geotechnique 25, No. 2, 229-238. Differential piezometer probe for an in situ measurement of sea-floor pore-pressure A. F. RICHARDS*, K. 0IENt, G. H. KELLERJ and J. Y. LAI§ A telemetering differential piezometer was designed Un telemetre/piezometre differentiel fut projete' et and constructed at the Norwegian Geotechnical construit a l'institut Geotechnique Norvegien, afin Tntt;t„t» «« m»o<.„™ « minimnm ^;ff„»n.;,i „„oo„~ de mesurer une pression differentielle minimum Institute to measure a minimum dinerential pressure , . . ..,.„ . , , (pression interstitielle se rapportant a la pression (pore-pressure referenced to hydrostatic pressure) of hydrostatique) de 34 kPa et une pression maximum 34 kPa and a maximum pressure of 294 kPa in de 294 kPa dans des profondeurs d'eau atteignants water depths of up to 500 m. An emplacement sys- jusqu'a 500 m. Un systeme de mise en place fut tem was built at the University of Illinois. One construit a l'universite d'Ulinois Le succes d'un e , . ,A , . . ,,,.„ . _, . test in situ dans le bassm de Wilkinson, dans une successful in situ test in the Wilkinson Basin, ,n a profondeur d-eau de 274 m> produisit une pression water depth of 274 m, yielded a maximum excess interstitielle avec un exces maximum de 59 kPa, pore-pressure of 59 kPa after the probe was driven apres avoir enfonce la sonde d'environ 3-2 m dans about 3-2 m into the silty-clay bottom. An excess ,e fond de l'argile limoneuse. Un exces de pression «„,., ~.«— „-rnoin jci/m. interstitielle de 9-8 kPa fut mesure cinq a dix heures pore-pressure of 9-8 kPa was measured 5-10 h ... , , . , JJ .. ^ . , apres la mise en place de la sonde. On discute des after emplacement of the probe. Implications of consequences d'exces de pression interstitielle excess pore-pressures cyclically generated by storm engendr^s d'une facon cyclique par la tempete et par and internal wave loading of sea-floor soils is dis- )e chargement interne de sols provenant du fond de cussed. It is concluded that a better understanding la m,er' Produit Par les vaf uf- Ori conclut qu'une - , . , , ... . . . .. . meilleure comprehension de la stabihte sous-manne of submarine slope stability through the use of the d,une pente par rutilisation du principe de contrainte effective stress principle should now be possible by effective, devrait maintenant etre possible, en me- measuring pore-pressure in situ. surant la pression differentielle in situ. More than a decade ago it was proposed that the likely relationship between excess pore-water pressure in cohesive sediments and potential submarine slides, slumps, and possibly turbidity currents could be investigated by the in situ measurement of pore-pressure in sea-floor sedi- ments (Richards, 1962). Negotiations with the Norwegian Geotechnical Institute (NGI) in 1964-5 led to the acquisition by the University of Illinois, Urbana, of two differential piezo- meter probes and a counter-computer in 1966. This Paper describes the probe system that was built around these units and the results of its limited testing at sea in 1967. Previously, the system was mentioned by Richards (1968) and Richards and Keller (1968). One test was very briefly summarized by Lai et al. (1968). DIFFERENTIAL PIEZOMETER SYSTEM Principal of operation and specifications The magnitude of hydrostatic pressure increases linearly by about one decibar (10 kPa) per metre of sea-water depth. Thus, the measurement of very small pressures above or below hydrostatic in water depths of several hundred metres is difficult. The measurement of only • Marine Geotechnical Laboratory, Lehigh University, Bethlehem, Pennsylvania, t Norwegian Geotechnical Institute, Tasen, Norway. X NOAA, Atlantic Oceanographic and Meteorological Laborities, Miami, Florida. § Manila, Philippines. 817 230 A. F. RICHARDS, K. 0IEN, G. H. KELLER AND J. Y. LAI the small difference in pore-pressures greater or less than hydrostatic in cohesive sea-floor sedi- ments is relatively easy. This latter relationship is schematically shown in Fig. 1, which illustrates the 274 m water depth occurring at the principal test site that will be discussed later. The piezometer was designed to measure a minimum differential pressure of less than 0-35 kg/cm2 (34 kPa) and a maximum of 2 kg/cm2 (196 kPa) from hydrostatic; after construc- tion the maximum was found to reach 3 kg/cm2 (294 kPa). Accuracy of the system during calibration testing was about ±63-5g/cm2 (±6-2kPa). The design working depth was 500 m, equivalent to a hydrostatic pressure of 50 kg/cm2 (4-9 MPa). It was estimated that by utilizing the free-fall method the probe might penetrate up to 20 m below the bottom under its own weight. The measuring principle is shown in Fig. 2. A pressure differential is measured between the hydrostatic pressure, u2, and the pore-water pressure, ux. The pressures are transmitted in water-saturated tubes. One long plastic tube extends from the u2 bourdon tube, through the weight stand, to above the water-sediment interface. The pore-pressure in the sediments surrounding the porous bronze filter at the tip of the probe is transmitted through a second plastic tube to the ux bourdon tube. Any difference in pore-pressure relative to the reference hydrostatic pressure deflects the bourdon tubes, shown in Fig. 3, causing a change in tension and hence the change in frequency of the vibrating wires that are clamped to the end of each tube. The relationship between the vibrating wires and electro-magnets is shown in Fig. 4. A linear relationship exists between the magnitude of the differential pressure and the change in the difference of the squares of the frequencies of the vibrating wires. The vibration frequencies are telemetered over an electrical logging cable to the ship. Aboard the ship an electronic frequency counter and computer displays the squares of the frequencies of the vibrat- ing wires and the difference between the squares (Fig. 2). A block diagram of the principal components in the counter-computer is shown in Fig. 5. Weight stand In Urbana, a weight stand was constructed 3-66 m long from the base of the support cone to the bail (Fig. 6). A 0-95 m long stainless steel tube separated the base of the support cone and the top of the piezometer. The entire probe system was 4-9 m long, tip to bail. Lead weights were cast having diameters of 15 and 20 cm. A few of the larger diameter weights were placed at the bottom of the weight stand to make the hole in the sea floor suffi- ciently large to reduce wall friction upon probe retrieval. The entire probe system, when fully weighted for deployment, had a maximum weight of nearly 570 kg in air. SEA TESTS Operational procedures The entire probe system (Fig. 6) was prepared for deployment in the summer of 1966; how- ever, because the scheduled ship unexpectedly could not be made available, the tests were post- poned one year. The probe system was deployed a number of times during June 1967 in the Wilkinson Basin, Gulf of Maine, which is located about 125 km east of Boston. In this basin, from the bottom surface to a depth of about 3 m, the average properties of the normally consolidated silty clays arte: w= 163, h>l= 124, wP = 47 and ysat= 1*33 Mg/m3. At the bottom of the basin the water temperature was only 5-6°C (Schopf, 1967). Consequently, after assembly, the tip of the probe was packed in ice for 0-5-1 h each time before it was lowered to the sea floor. 818 DIFFERENTIAL PIEZOMETER PROBE FOR AN IN SITU MEASUREMENT OF SEA-FLOOR PORE-PRESSURE 231 i 4} U E e £ 5. £ ois- • Q. S E © "c 8 £ «, n s £ ** _ o> S g "3 M>> i. s ^-© <-* CO « -a e a, k — e 3 %. O ■* 3 .= =1" E "" > « ai _> S-e o ai^.2 ■fi «"=> C *" ^ O eg a> -H_g I -5- o« £ ^ ft* q. «5 DC *- s n ■ -a O J* 41 >> E S 5> C8 ** l" L_ _ ■g n^^ cn.y-3 S ^ CI — © u M efOfc ?5 819 232 A. F. RICHARDS, K. 0IFN, Ci. H. KILLER AND J. Y. LAI mM 9 Fig. 3. Detail of bourdon gauges Fig. 4. Detail of electro-magnet and vihrating-wire assembly (left) and entire probe tip (right), which is 37-5 cm long 820 DIFFERENTIAL PIEZOMETER PROBE FOR AN IN SITU MEASUREMENT OF SEA-FLOOR PORE-PRESSURE 233 PULSE INPUT VIBRATING WIRE GAUGE OSCILLATOR (V.W.G OSC) VIBRATING WIRE GAUGE OSCILLATOR nr VWG OSC CONTROL CIRCUITS VWG OSC. CONTROL CIRCUITS CONTROL LOGIC r|ff-ff DISPLAYED FUNCTION TIME REFERENCE GENERATOR FREQUENCY REGISTER ACCUMULATOR Fig. 5. Block diagram of the electronic counter-computer POWER SUPPLY BATTERIES/ 110 or 220 Vac 821 234 A. F. RICHARDS, K. 0IEN, G. H. KELLER AND J. Y. LAI Fig. 6. Telemetering NGI-Illinois differential-piezometer probe system hanging over side of NOAA Survey Ship Davidson, June 1967. The probe is 4-9 m long 822 DIFFERENTIAL PIEZOMETER PROBE FOR AN IN SITU MEASUREMENT OF SEA-FLOOR PORE-PRESSURE 235 The following operational procedure was followed after the probe was lowered. It was held a short distance above the sea floor until the zero reading remained steady, indicating that ambient temperature and pressure equilibrium had been attained. The probe was then raised several metres above the bottom to gain potential energy, and then dropped into the bottom by allowing the winch to freewheel. The free-fall method (see Richards, 1973) could not be utilized during the 1967 testing at sea. While the probe was in the sediment, changes in the vibrating wire frequencies were recorded. After the readings stabilized, the probe was pulled out of the bottom and the zero reading again recorded to determine if any changes had occurred. It was intended that the penetration depth would be measured in a comparable manner to that used to determine the penetration depth of the NGI Torpedo sampler (Richards, 1973). This method would consist of echo-ranging on a hollow acoustic reflector attached by a line to the bail of the probe. In operation, the length of the line was to be sufficiently long to ensure that the reflector was always located above the bottom. However, due to a misunder- standing on the last lowering, the required co-ordination of effort was not maintained and this method could not be used. There is no reason to believe that the method would hot be suc- cessful if it were properly used, in light of the comparable operational experience obtained in Norway. RESULTS Only one full test was completed at sea. Problems with connectors, cable failures, and other difficulties unrelated to the functioning of the probe system consumed the remaining available time for testing. Fig. 7 records the dissipation of excess pore-pressures generated when the probe was emplaced on 12 June, 1967, an estimated 3-6 m below the bottom in a water depth of 274 m at latitude 42° 35T' north and longitude 69° 36-8' west. The lack of measurements between 80 and 306 minutes occurred because one of the vibrating wires stopped vibrating, presumably because of an intermittent connexion in a pair of mating electrical plugs that had partly pulled apart. DISCUSSION In the interpretation of the data (Fig. 8), the following assumptions have been made: (a) the probe actually was at the estimated depth below the bottom; (b) all the excess pore-pressure generated by the probe emplacement was dissipated in the ten hour measurement interval; (c) only pore-water entered the porous bronze filter, i.e. there was no water leakage to the porous filter from an elevation higher than the filter; (d) free gas was not present in the soil at the time of the measurement. The first assumption is not critical; either a greater or a lesser depth would not significantly change the conclusions if the other assumptions were valid. Although Fig. 7 shows little change in the magnitude of excess pore-pressure sensed during the last 5 h the probe was in place, it may be debated whether all of the excess pore-pressure generated by driving the probe into the bottom was dissipated during the relatively short time interval monitored. An interval at least three to five times longer would have permitted a more valid assessment of whether excess pore-pressure existed at the approximate depth and location tested. The length of time is dependent on the depth of penetration, consolidation characteristics of the sediment, and other factors. This uncertainty cannot be resolved from the data presented. It is therefore concluded that while the measurement is suggestive, it is by no means conclusive. 823 236 A. F. RICHARDS, K. 0IEN, G. H. KELLER AND J. Y. LAI hours 0-60 0 50 — 040 EXCESS P0RE- PRESSURE: „2 0-30 kg /car 0-20 — O'lO — 10 0-5 1 5 10 - 1 1 1 1 - - 1 III 1 I 1 I 1 o "* 1 1 1 1 1 1 1 1 20 40 60 80 100 200 400 600 LOG TIME AFTER PENETRATION OF PROBE ;min 1000 Fig. 7. Measured values of excess pore-pressure against log time after penetration and eye-fitted curve for 12 June, 1967 test The third assumption is believed valid considering the design of the probe tip, the distance of the tip to the base of the weight stand, the method of emplacement, and the low permeability of the silty clay indicated by the mean grain size being less than 2 y.m. Whether or not gas was present in the soil is not known for certain. Free gas has not been observed in cores raised from the deeper parts of the Wilkinson Basin. The soil in the vicinity of the test site is exceptionally transparent to acoustic energy from echo sounders, which suggests that free gas probably was not present in the soil at the test site. SUBSEQUENT STUDIES Since this investigation, a number of developments have occurred that are relevant to the future use of differential piezometers. Morgenstern (1967) assessed submarine slumping and the formation of turbidity currents in terms of effective stress; he clearly recognized the im- portance of excess pore-pressures. Sangrey et ah (1969) interpreted soil behaviour under repeated loading in terms of effective stress; they concluded that the build-up of pore-pressure was critical in bringing the soil to the effective stress failure envelope. Henkel (1970) postu- lated that differential loading by waves could cause submarine soil failure seaward of the Mississippi River Delta. Wilson and Greenwood (1974) reported that cyclically loaded soil samples failed at stresses less than the compressive strength of the soils obtained from standard strength tests. The theme of slides induced by cyclical storm wave pressure loading was am- plified by Bea (1971), Wright and Dunham (1972), Mitchell et ah (1973), Bea and Arnold (1973), Bea and Bernard (1973) and others. Southard and Cacchione (1972), in discussing earlier laboratory and theoretical work by Cacchione, hypothesized that internal waves may break on the outer continental shelf and upper continental slope, which also could cause cyclical loading. Bea ( 1 974) wrote that side-looking sonar records of the shelf edge immediately after hurricane Camille showed mysterious 'ripples' that disappeared 4-8-6-4 km shoreward of the shelf-break. He hypothesized that these ripples may have resulted from internal waves break- ing at the edge of the shelf. 824 DIFFERENTIAL PIEZOMETER PROBE FOR AN IN SITU MEASUREMENT OF SEA-FLOOR PORE-PRESSURE PRESSURE: kg /cm2 237 2851 I I I I I 1 I I I I Fig. 8. Interpretation of results of the test shown in Fig. 7 The time appears appropriate to measure the pore-pressure in situ; to compute the total stress by conventional means or from in situ nuclear densitometer measurements (Hirst et al., 1975); and to derive the effective stress actually existing in the floor of the ocean, particularly in areas susceptible to storm or internal wave loading. A knowledge of the total, effective and pore-water stress in sea-floor soils will significantly contribute to an understanding of submarine slope stability. CONCLUSIONS The NGI-Illinois differential piezometer probe system was successfully operated at a depth of 278 m below the sea surface, which corresponds to a hydrostatic pressure of about 27 kg/cm2 (2-65 MPa). A maximum excess pore-pressure of 0-6 kg/cm2 (59 kPa), resulting from driving the probe approximately 3-2 m into the bottom, was measured immediately following emplacement. There was little change in the 0-1 kg/cm2 (9-8 kPa) magnitude of excess pore-pressure mea- sured between five and ten hours after emplacement. While the magnitude of apparent excess pore-pressure remaining after nearly ten hours indicates that a small excess pore-pressure existed in sediments believed to be normally con- solidated, the time interval was insufficiently long to permit a more definite assessment to be made. It is possible that this apparent excess pore-pressure might have been largely or entirely dissipated if the pore-pressure was monitored for a considerably longer time after the emplace- ment of the probe. The free-fall, probe-emplacing method was not tested. Consequently, the maximum depth in which this probe may be emplaced in soft cohesive sediments was not determined. The bourdon gauge pressure sensing system performed adequately. However, the use of bellows, differential pressure transducers, or other more sensitive transducers probably would 825 238 A. F. RICHARDS, K. 0IEN, G. H. KELLER AND J. Y. LAI result in a more accurate measuring system. A bellows differential piezometer was built at the NGI after the bourdon gauge units were delivered ; it has not been tested at sea. A differential conductance piezometer has been designed, but not built, at the Marine Geotechnical Labora- tory. With the technology existing to measure pore-pressure in situ on the ocean floor, a better understanding of submarine slope stability is now possible. A knowledge of pore-pressure also will significantly contribute to long-term engineering investigations utilizing the effective stress principle. ACKNOWLEDGEMENTS Tests at sea were made from the NOAA Survey Ship Davidson, Lt Cdr W. Jeffers command- ing. At the University of Illinois, Professor V. J. McDonald and Mr J. Sterner ably assisted in electrical-electronic maintenance and operation. Mr R. G. Bea, Shell Oil Company, is thanked for his constructive review of this Paper. This investigation was performed under Office of Naval Research contract NONR 3985(09), NR 081-260, to the University of Illinois. The Paper was written under the sponsorship of ONR contract N00014-67-A-0370-0005, NR 083-248, to Lehigh University. REFERENCES Bea, R. G. (1971). How sea-floor slides affect offshore structures. Oil and Gas Jul 48, 88-92. Bea, R. G. (1974). Private communication. Bea, R. G. & Arnold, P. (1973). Movements and forces developed by wave-induced slides in soft clays. Preprints Offshore Tech. Conf. 2, 731-742. Bea, R. G. & Bernard, H. A. (1973). Movements of bottom soils in the Mississippi delta offshore. Offshore Louisiana Oil and Gas Fields. Lafayette Geol. Soc. and New Orleans Geol. Soc, 13-28. Henkel, D. J. (1970). The role of waves in causing submarine landslides. Geotechnique 20, 75-80. Hirst, T. J., Burton, B. S., Perlow, M., Jr, Richards, A. F. & Van Sciver, W. J. (1975). Improved in situ gamma-ray transmission densitometer for marine sediments. Ocean Eng. 3, May. Lai, J. Y., Richards, A. F. & Keller, G. H. (1968). In place measurement of excess pore pressure in Gulf of Maine clays (abstract). Am. Geophy. Union Trans. 49, 221. Mitchell, R. J., Tsui, K. K. & Sangrey, D. A. (1973). Failure of submarine slopes under wave action. Proc. 13th Coastal Eng. Conf. 2, 1515-1541. New York: American Society of Civil Engineers. Morgenstern, N. R. (1967). Submarine slumping and the initiation of turbidity current. Marine geo- technique, 189-220. Urbana: University Illinois Press. Richards, A. F. (1962). Unpublished report to Royal Norwegian Council of Scientific and Industrial Research. Richards, A. F. (1968). Discussion to session 1, shear strength of soft clay. Proc. Geotech. Conf, Oslo 2, 131-133. Oslo: Norwegian Geotechnical Institute. Richards, A. F. (1973). Geotechnical properties of submarine soils, Oslofjorden and vicinity, Norway. Norwegian Geotech. Inst. Tech. Rept 13, 107 pp. Richards, A. F. & Keller, G. H. (1968). Measurement of shear strength, bulk density, and pore pressure in recent marine sediments by in situ probes: results of 1967 shallow water tests (abstract). Am. Assoc. Petroleum Geol. Bull. 52, 547. Sangrey, D. A., Henkel, D. J. & Esrig, M. I. (1969). The effective stress response of a saturated clay soil to repeated loading. Canad. Geotech. Jnl 6, 241-252. Schopf, T. J. M. (1967). Bottom-water temperatures on the continental shelf off New England US Geol. Survey Prof. Paper 575-D, D192-197. Southard, J. B. & Cacchione, D. A. (1972). Experiments on bottom sediment movement by breaking internal waves. Shelf sediment transport: process and pattern, 85-97. Stroudsburg, Pa: Dowden, Hutchinson and Ross. Wilson, N. E. & Greenwood, J. R. (1974). Pore pressures and strains after repeated loading of saturated clay. Canad. Geotech. Jnl 11, 269-277. Wright, S. G. & Dunham, R. S. (1972). Bottom stability under wave induced loading. Preprints Offshore Tech. Conf 1, 853-862. 826 Reprinted from: Journal of Geology 83, No. 4, 536, 51 Sonography of the Sea Floor. By R. H. Bbldersox, N. H. Kenyon, A. H. Stkide. and A. R. Stubbs. Amsterdam: Elsevier Publishing Company, 1972. 1S5 pages, 1C3 figures. §27.75. "Sonograph" is a term used by the authors to describe the records made by side-scan sonar. Side-scan sonar is a technique devel- oped over the past 15 years to determine the shape of the ocean bottom and is finding new applications in biology, oceanography, and engineering. The side-scan technique is a variation of echo sounding in which the sonar transducer is directed oblique instead of normal to the ocean bottom. Sonographs of the ocean bottom have been compared to oblique aerial photographs of the land with the important difference that the sonograph is an acoustic rather than an optical picture. The authors briefly consider the acoustics of side-scan sonar and concentrate their considerable collective experience on the interpretation of sonographs. They present a picture atlas of 163 sonographs from the collection of the Bntish National Institute of Oceanography. The atlas is divided into 109 geological sonographs (continental shelf, upper continental slope, deepsca floor) and C4 other sonographs (marine life, sea effects, man-made objects). The sonographs are clearly reproduced on 22 by 30.5 cm (81 by 12 inch) pages. Each is accompanied by a brief description and a lino drawing which differentiates data from acoustic effects and gives overall dimensions of the area en- sonified. A problem in the use of the sono- graphs presented is that their widths are slightly nonlinear because distance to scat- tering features is a slant -range; techniques have subsequently been developed to convert slant-range to true-range display. Most of the sonographs were made by towing the instrument package within tens of meters of the ocean bottom on a cable suspended from a ship proceeding at a speed of about 1 knot. These, sonographs have a slant -range of less than 1 km and show small-scale geological, biological, and engineering fea- tures. Ten of the sonographs were made with an instrument called GLORIA (from Geo- logical Long Range Inclined Asdic) with a slant-range up to 22 km which is towed near the ocean surface from a ship at speeds up to 7 knots. The resolution of features is roughly one thousandth of the range. The book is useful as a manual of interpretation for everyone using side-scan sonar and any- one interested in the capabilities of side-scan sonar for geological, biological, oceano- graphic, and engineering applications. A list of 120 references covering the physical principles, instrumentation, and application of side-scan sonar between 1954 and 1971 is included. Peter A. Rona National Oceanic and Atmospheric Administration 827 52 Reprinted from: Science 190, No. 4213, 422 Minerals and Plate Tectonics In liis article. "Minerals and plate tec- tonics (II): Scawalcr and ure formation" (Research News, 12 Sept . p. 868), Allen L. Hammond cites examples of active hydro- thermal systems known on the sea floor. An additional documented example is the TAG hydrothermal field (/). discovered on the crest of the Mid-Atlantic Ridge at 26°N by the Trans-Atlantic Geotraverse (TAG) project of the National Oceanic and Atmospheric Administration. Manga- nese oxide of hydrothermal origin (2) present on ore wall of the rift valley (i) is inferred to have been deposited by hot (■/), metal-enriched (5), aqueous solutions dis- charging from fractures in the sea floor (6). The existence of such sub-sea floor hydro- thermal systems at sites along oceanic ridges indicates that hydrothermal activity may concentrate metals in oceanic crust throughout the opening of an ocean basin, from the Red Sea stage to the Atlantic Ocean stage, by sea floor spreading about a divergent plate boundary. Peter A. Rona A tlantic Oceanographic and Meteorological Laboratories. National Oceanic and A tmospheric Administration, Miami, Florida 33149 References 1. R B Scon. P. \ Rona. B A. McGregor. M. P.. Scoit. Saiwe I Lond. / 251. 301 ( 1974V 2. M R. Scott. R. B. Scon, P A. Rona. L. \V. Buiier, A. J. Nai^alk. Ceophvs. Res Leu 1. 355(1974). 3. B. A. McGreuor and P A Rona, J. (icjphvs. Res. 80, 3307(1975). 4 P. A. Runa. B A McGregor. P. R. Bct/er, D. C. Kraube. Deep-Sea Res . in press. 5. P. R. Bcu.r. G. W. Boker. B. A McGregor. P. A. Rnna. LOS. Im Gcophvs. Union Trans 55. 293 (1974) 6. P. A. Rona R N. Harbison, B G Bassinger R. B. Scott, A J. Nalwalk. Oeol Soc. Am. Bull., in press. 828 Reprinted from: Proc. Offshore Technologv Conference, Paper No. OTC 2317, 713-715. 53 THIS PRESENTATION IS SUBJECT TO CORRECTION Relation of Offshore and Onshore Mineral Resources to P I ate Tecton i cs By Peter A. Rona , NOAA ©Copyright 1975 Offshore Technology Conference on behalf of the American Institute of Mining, Metallurgical, and Petroleum Engineers, Inc. (Society of Mining Engineers, The Metal lurgicaV Society and Society of Petroleum Engineers), American Association of Petroleum Geologists, American Institute of Chemi- cal Engineers, American Society of Civil Engineers, American Society of Mechanical Engineers, Institute of Electrical and Electronics Engineers, Marine Technology Society, Society of Explor- ation Geophysicists, and Society of Naval Architects and Marine Engineers. This paper was prepared for presentation at the Seventh Annual Offshore Technology Conference to be held in Houston, Tex., May 5-8, 1975. Permission to copy is restricted to an abstract cf not more thar 300 words. Illustrations may net be copied. Such use cf an abstract should contain conspicuous acknowledgment of where and by v*hom the paper is presented . ABSTRACT The Pacific and Atlantic are natural laboratories to study relations between mineral resources and plate tectonics. The distribu- tion of mineral deposits about convergent lithospheric plate boundaries may be best observed in the Pacific. The Pacific is sur- rounded by convergent plate boundaries where oceanic crust is consumed by subduction along Benioff zones. Areas of offshore petroleum potential are associated with the convergent plate boundaries which create oceanic trenches along the eastern Pacific and small ocean basins along the western Pacific. Precious, base, iron and ferro-alloy metal deposits occur onshore from the convergent plate boundaries on continents in the eastern Pacific and on island arcs and continents in the western Pacific. The distribution of mineral deposits about divergent plate bounda- ries is less known than about convergent plate boundaries because the former are submerged oceanic ridges. The Atlantic is bisected by the Mid-Atlantic Ridge where oceanic crust is created by sea floor spreading. Certain metals are being concentrated in oceanic crust at divergent plate boundaries by hydrothermal processes. The TAG Hydrothermal Field, an active hydrothermal area discovered by the NOAA Trans-Atlantic Geotraverse (TAG) on the Mid- Atlantic Ridge at 26°N, is providing information of how -metals are concentrated in oceanic crust at divergent plate boundaries. Present models References and illustrations at end of paper. which treat the relation between mineral de- posits and lithospheric plate boundaries are largely interpretive , explaining the observed distribution of deposits. The development of these models may lead to the discovery of new deposits both offshore and onshore. INTRODUCTION The Atlantic is an opening ocean basin that is inferred to have been growing wider at a rate of several centimeters per year for approximately the past 200,000,000 years. The Pacific is a closing ocean basin that is inferred to have been diminishing in size at a rate comparable to that of the opening of the Atlantic. The contrasting histories of the Atlantic and Pacific are explained according to the theory of plate tectonics in terms of motions of lithospheric plates (Fig. 1) about divergent and convergent plate boundaries (Fig. 2 ) . The occurrence of mineral deposits includ- ing petroleum may be related to plate motions and, in particular, to geological processes associated with the boundaries between the plates. Differences in plate motions have resulted in different distributions of minerals in the Atlantic and Pacific regions that may be characteristic of divergent and convergent plate boundaries. 829 714 RELATION OF OFFSHORE AND ONSHORE MINERAL RESOURCES TO PLATE TECTONICS DIVERGENT PLATE BOUNDARIES Relations between the opening of an ocean basin about a divergent plate boundary and the occurrence of mineral deposits may be inferred from the development of the Atlantic as follows: 1. At an early stage of opening following conti- nental rifting a sea is formed with its circu- lation restricted by tectonic conditions and by the positions of the surrounding continents1. Conditions in this sea favor both the accumula- tion of organic matter that may become petroleum and of rock salt that may subsequently form domes to trap the petroleum2 . Metalliferous sediments and possibly massive stratabound sulfide bodies are concentrated by hydrothermal processes at the divergent plate boundary. An example of this early stage of opening is the Red Sea3. 2 . Widespread metalliferous sediments immedi- ately overlying basalt of the ocean basins indi- cate that hydrothermal activity at a divergent plate boundary may continue as the ocean basin opens ** ~ 6 . 3. The recent discovery of the TAG Hydrothermal Field by the NOAA Trans-Atlantic Geotraverse (TAG) project, where metal oxides and possibly sulfides are being deposited on the Mid-Atlantic Ridge at 26°N, indicates that hydrothermal processes may concentrate metals in oceanic lithosphere at a divergent plate boundary from early (Red Sea) to advanced (Atlantic Ocean) stages of opening7'8. CONVERGENT PLATE BOUNDARIES The Pacific Ocean basin is inferred to be closing as a consequence of consumption of oceanic lithosphere at convergent plate bounda- ries around three-fourths of its perimeter. Relations between the consumption of an ocean basin at a convergent plate boundary and the occurrence of mineral deposits may be inferred from the Pacific9 , as follows : 1. The structural framework at a convergent plate boundary expressed as an oceanic trench or an island arc that delineates a marginal basin creates conditions that favor the development of offshore petroleum. 2 . The subduction of oceanic lithosphere along a Benioff zone at a convergent plate boundary is genetically related to the occurrence of iron and ferro-alloy, base, and precious metal depo- sits , such as those distributed along the conti- nental margins and island arcs of the Pacific10 1? SUMMARY A model depicts processes at divergent and convergent plate boundaries of the Atlantic and Pacific (Fig. 3). Petroleum is related to early stages of opening of an ocean basin about a divergent plate boundary and to basins formed at convergent plate boundaries. Certain iron and ferro-alloy , base , and precious metals are concentrated in oceanic lithosphere by hydro- thermal processes at a divergent plate boundary from early to advanced growth stages of an ocean basin. The metals undergo further concentration when the oceanic lithosphere is subducted at a convergent plate boundary. The metals are deposited along the convergent boundary where they are most accessible for exploitation. REFERENCES 1. Rona, P. A. : "Possible Salt Domes in the Deep Atlantic off Northwest Africa," Nature (1969) 224, 141-143. 2. Rona, P. A. : "Comparison of Continental Margins of Eastern North America at Cape Hatteras and Northwestern Africa at Cap Blanc," Bull. Am. Assoc. Pet. Geologists (1970) 54, 129-157. 3. Hutchinson, R.W. and Engels, G.G. : "Tectonic Evolution in the Southern Red Sea and its Possible Significance to Older Rifted Conti- nental Margins," Geol. Soc. Am. Bull. (1972) 83_, 2989-3002. 4. Bostrom, K. and Peterson, M.N. A. : "Origin of Aluminum- Poor Ferromanganoan Sediments in Areas of High Heat Flow on the East Pacific Rise," Mar. Geol. (1969) 7_, 427-477. 5. Dymond, J., Corliss, J.B., Heath, G.R., Field, C.W., Dash, E.J. , and Veeh, H.H. : "Origin of Metalliferous Sediments from the Pacific Ocean," Geol. Soc. Am. Bull. (1973) 84, 3355-3372. 6. von der Borch, C.C., Nesteroff, W.D. , and Galehouse, J.S.: "Iron- Rich Sediments Cored During Leg 8 of the Deep Sea Drilling Project," In, Tracey, J.I. Jr., et al. (eds. ) . Initial Reports of the Deep Sea Drilling Project (1971) 8, 725-819. 7. Scott, R.B., Rona, P. A. , McGregor, B.A. , and Scott, M.R.: "The TAG Hydrothermal Field," Nature (1974) 251, 301-302. 8. Scott, M.R., Scott, R.B., Rona, P. A. , Butler, L.W. , and Nalwalk, A.J. : " Rapidly Accumulating Manganese Deposit from the Median Valley of the Mid-Atlantic Ridge," Geophys. Res. Letters (1974) 1, 355-358. 9. Rona, P. A. , and Neuman, L.D. :~" Plate Tec- tonics and Mineral Resources of the Pacific," Bull. Am. Assoc. Pet. Geologists (1974) 58, 1456. 830 PETER A. RONA 715 10. Sawkins, F. : "Sulfide Ore Bodies in Relation to Plate Tectonics," J. Geol. (1972) 80, 377-397. 11. Sillitoe, R.H. : "Relation of Metal Provin- ces in Western America to Subduction of Oceanic Lithosphere , " Geol. Soc. Am. Bull. (1972) £3_, 813-818. 12. Mitchell, A.G.H. : "Metallogenic Belts and Angle of Dip of Benioff Zones," Nature Phys. Sci. (1973) 245, 49-52. 13. Le Pichon, X.: "Sea Floor Spreading and Continental Drift," J. Geophys. Res. (1968) 73_, 3661-3697. 14. Rona, P. A. : "New Evidence for Seabed Resources from Global Tectonics," Ocean Management (1973) 1, 145-159. 15. Isacks, B. , Oliver, J., and Sykes, L.R. : "Seismology and the New Global Tectonics," J. Geophys. Res. (1968) 73, 5855-5899. 16. Rona, P. A. : "Plate Tectonics and Mineral Resources," Scientific American (July 1973) 229, 86-95. 1 | VrV EURASIA PLATE BOUNDARIES ■*- DIVERGENT -c CONVERGENT Fig. 1 - Boundaries of six principal lithospheric plates1^ l^. Divergent plate boundaries where lithosphere is created and convergent plate boundaries where lithosphere is destroyed are indicated. 831 4= M i C o o rH •i- c 43 ■> c CU IS\ cc Jh CO H CL -p CU X CJ ■p !■ o CJ CO c •H T3 s c c O cc CO E co CD O 03 E o cj •H o o CJ ^ SH CO cc Cm a -r CO ~ r. 03 >> B -p 43 1-1 c H +3 CO 43 CU CO T3 42 CD a CU a 3 -P m OJ o cc O 4J Xj r-\ ;■ P 01 CU E > c h -P c o O C o N •H ■H -p C H a O CU cO cj QJ pq > *s TJ o CO >5 S3 E M co bD CO y— >» C X) -p co O C C 0) H 3 QJ rH CO o bO •H 43 m E TJ 0) OJ CU > O -p •P •rl \D CJ CO Td N_^ J3 H T3 P 3 3*! 43 o O 3 ^ -p o Sh CJ ^ 43 CO •H o T1 -P 43 OJ C CO CO -p 33 • £ co 0 w E CO H 43 C co 01 p. o h -p CU .H bO CO p P -P CD H a CO o •H P- CU H E o bO p cj SH 0) 1 rH CU p -p M > S3 CO CM CO • H CU rH 43 t3 bD p P- M bO co 0) rH •H O CU > CO Cn £ bO £3 -p ■P T3 o a H • H CJ O (TOO OTOO H3dd00 •rt\D CMrH •H CO CJ QJ cn •H Ph Sn CO. Tl Tl £3 £3 CO =1 0 cj 43 ■H +-> QJ c •P co CO rH r-\ -P P. < -P CU c 43 CD -P on SH Cm ai O > C a 0 o o • M +J T) o u CU ai CO 1 •p CO C CO QJ 0 bf) u M CJ CU > -p •H CO T) CU £ O 1 -P CO T) H c E co co M H w: ri CD o> •H rH Tl o u CO +J 01 bu P C •H Cm CO O D a rH o 01 ■H ■a -p o co E r-{ fl -p • s bD O ■rl 43 fr< CO o 832 Reprinted from: Proc. International Symposium on Conti- 54 nental Margins of Atlantic Type, Brazil, 1-15. SALT DEPOSITS OF THE EASTERN AND WESTERN ATLANTIC Peter A. Rona Atlantic Oceanographic and Meteorological Laboratories National Oceanic and Atmospheric Administration Miami, Florida, U.S.A. ABSTRACT Sufficient information is available from onshore and offshore geological surveys and drilling to relate the distribution of evaporites in time and space to the develop- n i e n t of the Atlantic. Evaporites had long been recognized in marginal basins of the Atlantic but their seaward extent was thought to be restricted by shelf-edge barriers. The discovery of salt d e p o s i t s in the Atlantic ocean basin be- neath the continental rise indicated that the real restrictions on evdporite deposition were the tectonic sot fin g produced by the rifting of the continents and the configuration of the Atlantic Sea during early stages of drift. Questions raised include distinctions between salt deposits of Atlantic marginal basins and ocean basin, shallow versus d e e p water evaporites, and the relation of the salt deposits to horizontal (sea floor spreading) and to vertical (rifting, subsidence, and eustacy) movements. The Atlantic exhibits a classic distribution pat- tern of evaporites of an opening ocoan basin. 833 55 Reprinted from: Encounter with the Earth, edited by L.F. Laporte, Canfield Press, San Francisco, Ca, 83-86. Viewpoint Peter A. Rona Peter A. Rona is a research scientist with the National Oceanic and Atmospheric Administration, Miami, Florida. Dr. Rona's recent work has been on the possible link between plate tectonics and economically valuable mineral deposits. In this Viewpoint Dr. Rona reviews the geologic factors controlling such deposits along plate boundaries. Learning Where to Look for Mineral Resources The revolutionary advance in our understanding of internal processes of the earth made during the past ten years is leading to economic returns through the discover/ of new minerai resources. The problem of exploring the earth's crust for mineral deposits is the proverbial one of searching 83 834 Internal Processes for a needle in a haystack. Just as the needle is tiny relative to the size of the haystack, so petroleum and metal deposits are small relative to the volume of the earth's crust. Many valuable mineral deposits occupy areas less than that of a city block. The problem of discovering a deposit, like that of finding the needle in the haystack, is facilitated by knowing where to look. In the past we have found only the most accessible deposits on the continents, largely by trial and error, without a real understanding of why and where the deposits occurred. As a result of our increased understanding of internal processes of the earth we are gaining insights that tell us where to probe for new sources. £>-' Seem like a good place to look- How do internal processes control mineral deposits? An important clue is the observation that many minerals lie along the boundaries between the plates that divide the earth's crust (Figure 3-1). Petroleum and various base and noble metals occur along the margins of the Pacific Ocean, which are plate boundaries. These deposits include the metal provinces of western North and South America, and the petroleum of Indonesia and the west coast of the United States. We are exploring the plate boundaries submerged beneath the ocean along midoceanic ridges and have found that overlying sediments are enriched in various metals and that solid metal deposits appear at certain sites along the Mid-Atlantic Ridge. For this reason we suspect that metal deposits including copper, manganese, iron, nickel, lead, zinc, chromium, cobalt, uranium, and gold may be present at sites on midoceanic ridges. We do not think that midoceanic ridges contain any petroleum. Sediments on the floor of the Red Sea, where a plate boundary lies between Africa and Eurasia, are enriched in iron, zinc, copper, lead, silver, and gold. Sediments full of organic matter that may eventually become petroleum are preserved at the margins of the Red Sea. Why should certain mineral deposits be concentrated along plate bound- aries? Plate boundaries are where the geologic action takes place. As crustal material is being created at divergent plate boundaries and is being de- stroyed at convergent plate boundaries, processes are working to concen- 84 835 Viewpoint trate minerals in deposits along the boundaries. The metals that appear at divergent plate boundaries on midoceanic ridges and in the Red Sea are being deposited from hot solutions that concentrate metals from the rocks lying at these boundaries (Figure 3-2a). The metals at convergent plate boundaries around the Pacific Ocean are also deposited from hot solutions rich in metals possibly derived from the melting of the Pacific plate as it plunges beneath the adjacent continents (Figure 3-2b). Many mineral deposits lie in areas far from present plate boundaries. To understand the reason for the location of these deposits, we must consider how the sizes, shapes, and positions of the continents and ocean basins have changed through time. For example, the Atlantic originated as a sea at an early stage in the development of the divergent plate boundary and widened into an ocean over a period of 200 million years. Various metals and organic matter may have accumulated in the Atlantic Sea, as they have in the present Red Sea, so that metal and petroleum deposits may be present at sites under the miles-thick sediments along the eastern and western margins of the Atlantic Ocean. The separation of continents by continental drift about divergent plate boundaries may divide preexisting mineral provinces. Fitting the continents together in their positions prior to continental drift in ?. global jigsaw puzzle may reveal the continuity of certain mineral provinces between the continents. Diamonds found in northwest South America (Guyana and Venezuela) appear to have been derived from source rocks in west Africa (Liberia, Ivory Coast, and Ghana) when South America and Africa were joined prior to the opening of the Atlantic Ocean. Gold-bearing formations in these two regions of South America and Africa can be matched across the Atlantic. Metal provinces of southeast Africa, Southern India, and western Australia apparently match as well, agreeing with the positions of these continents when joined prior to continental drift. Mineral deposits may also be present along former plate boundaries that are no longer active. Metal deposits of the Appalachian Mountains of southeastern North America and the Ural Mountains between Europe and Asia may have originated when these mountain ranges developed at former convergent plate boundaries. Still other mineral deposits within continents do not appear to conform either to present or former plate boundaries, and their origins remain problematic. The patterns of mineral distribution that are emerging from our in- creased understanding of internal processes of the earth are guiding man's search for new mineral deposits. However, many factors should be kept in mind concerning mineral resources for the future. Nature is not manu- facturing new mineral resources at plate boundaries as fast as man is depleting known mineral deposits. The rate of accumulation of metal deposits from hot solutions at midoceanic ridges is about 8 millionths of an inch per year (2 hundred-thousandths of a centimeter per year). The minimum time required to form petroleum from organic matter and 85 836 Internal Processes concentrate it into pools is of the order of tens cf thousands of years Many existing mineral deposits are inaccessible to us because they are too deeply buried in the crust to be detected and recovered by present technology; and when deposits are accessible, often the cost or production would be higher than the market value. Even with our increased under- standing of where to look for new mineral deposits, evaluation of a new area of land or sea mav require vears of exploration and expenditure with uncertain outcome. Following the discovery of a valuable petroleum or metal deposit, five to ten vears of developmental work are generally re- quired before the prospect is ready for production. In conclusion, our greater understanding of the earth's internal processes can be expected to accelerate the discovery of mineral resources on the continents and in the ocean basins, but still offers us no promise of Utopia. Summary The earth's present-day surface results from large-scale processes operating throughout its crust and upper mantle. The all-encompassing theory of plate tectonics explains the broad features and geologic activity that we observe on the face of the earth. Relatively thin, rigid plates move either by gravity or convection over the thicker, plastic upper mantle. Plate boundaries are marked by long, linear geologic features like mountain ranges, submarine basaltic ridges, volcano and earthquake zones, and deep sea trenches— depending on whether plates are moving together, moving away from each other, or slipping past one another. Continents are thick slabs of lighter rocks floating more or less in isostatic balance on the denser, underlying mantle. Ocean basins are underlaid with thinner slabs of heavier rock resting on the mantle. Continents are regions of higher elevation continually eroding,- ocean basins are areas of lower elevation that receive sediments deposited from the continents. As oceanic plates collide with continental plates, the oceanic sediments may be scraped off and piled against the continents or metamorphosed and granitized as they descend below the continent. Plate collisions also deform rocks by faulting and folding them. Basaltic lavas continuously well up from the mantle and form long, linear ridges within the central portions of ocean basins along diverging plate boundaries. Plate movement away from the midoceanic ndges brings these mantle materials from the oceans and eventually adds them to the continents. Plates thus grow at their diverging boundaries and are con- sumed at converging ones. Consequently, throughout geologic history- continents have grown with the continuous addition of mantle material and have shifted spatially as plate movement has carried them along. 86 837 56 Reprinted from: Deep-Sea Research 22, 611-618, Anomalous water temperatures over Mid-Atlantic Ridge crest at 26° North latitude Peter A. Rona,* Bonnie A. McGregor,* Peter R. Betzer,-j- George W. Bolger-|- and Dale C. Krause^: (Received 7 August 1974; in revised form 30 January 1975; accepted 24 February 1975) Abstract — Two positive temperature anomalies with different characteristics were measured in the water over the walls of the rift valley of the Mid- Atlantic Ridge at 26°N latitude. A water temperature profile parallel to the ocean bottom was made with three thermistors mounted in a 4 m long vertical array on a towed deep-sea camera over a hydrothermal mineral deposit, the TAG Hydrothermal Field, situated on the southeast wall of the rift valley. It revealed an abrupt anomaly of + 01 1°C associated with an inverse gradient of 8 x 10~3 °C m_1 within 20 m of the bottom along a horizontal distance of about 350 m between water depths of 3030 and 2950 m. Over the northwest wall of the rift valley, a vertical water temperature profile measured with reversing thermometers revealed a relative warming of 0-20°C between 2000 m and the ocean bottom at 2508 m. Either conductive transfer of heat or discharge of hydrothermal solutions from the ocean bottom could be the heat source for the observed temperature anomalies. The different geologic setting in the two areas suggests that the anomaly over the southeast wall is due to the discharge of hydrothermal solutions; that over the northwest wall to in situ conductive transfer of heat or possibly to the advec- tion of a bolus of warm water into the area. INTRODUCTION Mid-oceanic ridge systems are recognized as regions of considerable geologic activity. Large volumes of basic rock are intruded or extruded, fault systems displace sections of the crust, and the epicenters of shallow earthquakes are located at ridge crests. Although most of these processes occur over long time periods, they may at times measurably affect the physical-chemical environ- ment of adjacent deep ocean waters (Knauss, 1962; Lubimova, Von Herzen and Udintsev, 1965). An interdisciplinary investigation of the environment of a hydrothermal mineral deposit on the Mid-Atlantic Ridge at 26°N was performed in the fall of 1973 (R. Scott, Rona, McGregor and M. Scott, 1974; M. Scott, R. Scott, Rona, Butler and Nalwalk, 1974). The investigation, part of the Trans-Atlantic Geo- traverse (TAG) project of the National Oceanic and Atmospheric Administration, included the measurement of near-bottom water temperatures. This report summarizes data obtained from a deep-towed camera equipped with thermistors, from an STD (salinity, temperature, depth) system, and from standard hydrographic casts over the TAG Hydrothermal Field on the south- east wall of the rift valley and adjacent portions of the Mid-Atlantic Ridge crest (Rona, McGregor, Betzer and Krause, 1974). INSTRUMENTATION Water temperature measurements parallel to the ocean bottom were made with a system similar to that used to make heat flow measurements in the sea floor. It consisted of thermistor tempera- ture probes (Fenwall part number K2365) and a thermoprobe recorder (modified Alpine model 323) mounted on the frame of a deep-sea camera. Three thermistor temperature probes were moun- ted at heights of 0, 3, and 4 m above the base of the camera frame. The thermistors were calibrated at the factory prior to the experiment and recalibrated at 0°C in an ice bath following the experiment; drift correction of the order of 0-0 1°C (2 Q) was applied to the data. The accuracy of the thermistors is ± 0-01 °C and their precision is ± 0-004°C. The output of each thermistor was recorded sequentially at an interval of 45 s. A tiltmeter in the thermoprobe recorder used mercury switches to measure inclination from the vertical in 15° increments. The instrument package included thermistors, recorder, two deep-sea cameras (EG & G model 207 A), two light sources (EG & G model 208), *National Oceanic and Atmospheric Administration, Atlantic Oceanographic and Meteorological Laboratories, 15 Rickenbacker Causeway, Miami, Florida 33149, U.S.A. tDepartment of Marine Science, University of South Florida, St. Petersburg, Florida 33701, U.S.A. ^Division of Oceanography, UNESCO, Paris, France. 838 612 Peter A. Rona, Bonnie A. McGregor, Peter R. Betzer, George W. Bolger and Dale C. Krause and an acoustic pinger (EG & G model 220). It was towed at ship's speeds between 15 and 45 m min-1 and maintained at a distance of about 5 m above the ocean bottom by means of a hydraulically-controlled deep-sea winch. This distance was determined to an accuracy of 2 m by the acoustic pinger. The visible wire angle was nearly zero so that the water depth measured by the ship's narrow-beam echo sounder (6° total beamwidth) was used as the depth of the ocean bottom at the instrument package. The estimated accuracy of depth below sea level of the instrument package is ± 100 m with corresponding uncer- tainty in the computation of potential temperature. Vertical profiles of water temperature and salinity were made with two techniques, an STD system (Plessey model 9040) and hydrocasts. The accuracy of the STD system is ± 002°C for temperature, ± 002%o for salinity, and ± 10 m for depth, with a precision of ± 0005°C, ± 0-001%o, and ± 1 m, respectively. The hydro- casts involved simultaneous temperature measure- ments on two protected reversing thermometers and a single unprotected reversing thermometer to estimated accuracies of ± 002°C, and ± 0-1 °C, respectively; the temperatures presented are the average of the measurements on the two protected thermometers. Quadruplicate measure- ments of salinity were made on each of two water samples from each Niskin bottle. All of the analyses were carried out at sea in an air condi- tioned laboratory with a Beckman (Model RS7B) induction salinometer using standard Copenhagen water as a reference. The average between duplicate analyses of the same water sample was 0005%o. An acoustic pinger was used to determine height above the bottom of the STD and the hydrocasts, both of which were lowered to about 15 m from the ocean bottom. In situ temperature measurements were converted to potential temperature following Bryden (1973). Primary navigational control was by a satellite navigational system. RESULTS Two profiles parallel to the ocean bottom were made with the thermistor temperature probes mounted on the deep-sea camera. Thermistor 45°00'W 44°40'W 26°N 0SN Fig. 1 . Bathymetric map in hundreds of corrected meters of the area of investigation including the southeast and northwest walls of the rift valley of the Mid-Atlantic Ridge at 26°N based on a 2 x 4-km grid of narrow beam bathy- metric profiles (McGregor and Rona, in press). The TAG Hydrothermal Field is outlined by dashed lines (R. Scott, Rona, McGregor and M. Scott, 1974). Locations are shown of thermistor profiles 1 and 2, hydrocast Stas. 6, 9 and 10, and STD profiles 2 and 3. Satellite fixes along thermistor profiles 1 and 2 are indicated by squares and times bracket temperature measurements listed in Table 1. Lines A-A' and B-B' are the locations of the bathymetric profiles shown in Figs. 5 and 6. 839 Anomalous water temperatures over Mid-Atlantic Ridge crest at 26° North latitude 613 1000 HORIZONTAL DISTANCE (M) 2000 Fig. 2. Thermistor profile 1 of in situ (1ST) and potential (PT) temperature between 2 and 20 m above the ocean bottom over a portion of the TAG Hydrothermal Field on the southeast wall (Fig. 1). profile 1 (Figs. 1 and 2) is about 2 km long and crosses a portion of the TAG Hydrothermal Field on the southeast wall of the rift valley between depths of 3420 and 2910 m (Table 1). An abrupt increase of 0-1 6°C occurred while the instrument package was between 2 and 20 m above the ocean bottom along a horizontal distance of about 350 m between water depths of 3030 and 2950 m (Table 1). The water depths determined during thermistor profile 1 are corroborated by a relatively detailed isobath map (Fig. 1). The temperature increase attributable to adiabatic heating for the 80-m change in depth is about 005 C, as deter- mined from an STD profile in an adjacent area, leaving a residual positive temperature anomaly of 0T1°C. Although the value of the temperature increase was taken from the lowermost thermis- tor, similar temperature increases were recorded on all three thermistors (Table 1). A temperature gradient of 8 x 10~3 °C m_1 warming toward the ocean bottom was measured over the 4-m long vertical thermistor array at the maximum of the temperature anomaly. This 0T1°C step-like positive temperature anomaly is equivalent to a Table 1. Temperatures measured along thermistor profiles 1 and 2. Thermistor profile L (Fig. 2) Date Time (GMT) Depth of Bottom (n) Distance of lowermost thermistor above bottom (m) In situ temperature C°C) Thermistor position in vertical array (m above lowermost ther- mistor) 4 3 0 •10 Oct. 1973 1530 3420 6 2.703 2.691 2.665 1540 3410 8 2.706 2.693 2.665 1550 3405 4 2.706 2.693 2.670 1600 3400 4 2.706 2.693 2.667 1610 3300 10 2.703 2.639 2.660 1620 3120 11 2.700 2.638 2.660 1630 3060 4 2.698 2.6e4 2.660 1640 3030 2 2.726 2.715 2. 695 1644 3000 G 2.777 2.762 2.811 1647 2990 20 2.789 2.769 2.819 1650 2975 7 2.843 2.833 2.799 1653 29C0 6 2.860 2.853 2.845 1655 2950 8 2.848 2.929 2.681 1653 2920 10 2.709 2.702 2.631 1700 2910 6 2.741 2.722 2.697 840 614 Peter A. Rona, Bonnie A. McGregor, Peter R. Betzer, George W. Bolger and Dale C. Krause Table 1. continued Thermistor profile 2 Date Tine (GMT) Depth of Bottom (m) Distance of lowermost thermistor above bottom Cm) In situ temperature C°C) Thermistor position in vertical array (m above lowermost thermistor) 4 3 0 3 Oct. 1973 2100 2800 2 2.936 2.929 2.895 2110 2750 4 2.946 2.929 2.895 2120 2700 8 2.951 2.929 2.914 2130 2680 5 2.946 2.929 2.905 2140 2620 10 2.946 2.939 2.914 2150 2590 7 2.955 2.935 2.901 2200 2520 8 2.942 2.929 2.905 -2210 2500 i* 2.951 2.939 2.914 2220 2150 2 2.984 2.967 2.952 2230 2400 9 3.012 2.977 2.943 2210 2390 H 2.984 2.967 2.943 2250 2380 11 3.003 2.986 2.990 2300 2370 2 3.060 3.044 3.015 vertical rise and fall of the thermistor instrument package of about 300 m (estimated from Fig. 4) in horizontal distances of 100 and 50 m (Fig. 2). A vertical fluctuation of this magnitude is ruled out because no abrupt changes occurred in the ship's speed, the winch meter wheel readings, the bottom topography, or in the position of the camera relative to the bottom. Thermistor profile 2 (Fig. 1) is 2-5 km long and traverses a portion of the TAG Hydro- thermal Field. A gradual increase of 0-1 2°C was measured on the lowermost thermistor between 2 and 1 1 m above the bottom along a horizontal distance of 1 -67 km between depths of 2800 and 2370 m (Table 1). The temperature increase was recorded on all three thermistors. Unlike the temperature increase registered along thermistor profile 1, this increase is wholly attributable to adiabatic heating for the 850-m change in depth as determined from an STD profile in the area. Hydrocasts were made over the southeast wall (Sta. 6; 26°09'W: 44°47'W; Fig. 1), the median valley (Sta. 9; 26°10'N, 44°53'W; Fig. 1), and the northwest wall (Sta. 10; 26°15'N, 44°57'W; Fig. 1) of the Mid- Atlantic Ridge. The values of potential temperature calculated from reversing thermometers are similar at the three stations down to a depth of about 2000 m (Fig. 3). Below 2000 m the temperature profile over the southeast 1000 2000 - 3000 1 1 1 **^--STA 6 EAST WALL STA 9 MEDIAN VALLEY r -yrr - STA 10 WEST WALL _l ' ' 2.0 3.0 4.0 5.0 6.0 POTENTIAL TEMPERATURE (°C) Fig. 3. Potential temperature profiles determined by reversing thermometers over the southeast (Sta. 6) and northwest walls (Sta. 10) of the rift valley (Sta. 9). Large dots are the depths of the reversing thermometers; the lowermost thermometers were about 1 5 m above the ocean bottom, shown beneath the profile at Sta. 10. 841 Anomalous water temperatures over Mid- Atlantic Ridge crest at 26° North latitude 615 wall remains similar to the temperature profile in the rift valley while the temperature profile over the northwest wall attains a temperature O20°C warmer than the profile in the rift valley. The salinity distribution over the northwest wall was normal from 2000 to 2493 m, 15 m above the bottom (Sta. 10; Fig. 3). STD profiles were made from the surface to within about 15 m of the ocean bottom on the southeast wall (STD 2; 26°08'N, 44°45'W; Figs. 1 and 4) and northwest wall (STD 3; 26°16'N, 44°58'W; Figs. 1 and 4) of the rift valley. Both STD profiles exhibit subadiabatic temperature gradients to a depth of about 2500 m where the gradient in STD 2 over the southeast wall becomes nearly adiabatic between 2500 m and the ocean bottom at about 2650 m. The adiabatic temperature gradient, using our measured values of temperature and pressure in the area investigated, was 1 x 10-5 °C m-1. DISCUSSION Previous observations indicate the occurrence of relative warming and superadiabatic gradients of water at certain sites above or near mid- oceanic ridges (Knauss, 1962; Lister, 1963; Lubimova, von Herzen and Udintsev, 1965; Bodvarsson, Berg and Mesecar, 1967; Sclater and Klitgord, 1973; Williams, von Herzen, Sclater and Anderson, 1973). Two of the principal processes that may account for positive temperature anomalies are conductive transfer of heat through the ocean bottom and discharge of hydrothermal solutions from the ocean bottom into the overlying water column. Evidence exists both for high conductive transfer of heat and for hydrothermal activity on the Mid-Atlantic Ridge at 26°N. High values of heat flow ranging between 2-0 and 86 HFU typical of oceanic ridges have been measured in sediments here (Langseth, Malone and Bookman, 1972). Suspended particulate matter re- covered in hydrocasts beginning 15 m above the ocean bottom in this crestal region of the Mid- Atlantic Ridge contains anomalous concentrations of iron and manganese relative to suspended matter at corresponding depths in the adjacent Atlantic away from the ridge crest (Betzer, Q 2500L_1__1_J L Fig. 4. STD profile 2 (Fig. 1) over the southeast wall and STD profile 3 (Fig. 1) over the northwest wall of the rift valley. The curve on the extreme right is the computed adiabatic gradient. 842 616 Peter A. Rona, Bonnie A. McGregor, Peter R. Betzer, George W. Bolger and Dale C. Krause Bolger, McGregor and Rona, 1974). The aluminum content and Fe/Al ratios in near- bottom suspended materials over the ridge (2-6% and 1-5, respectively) differed markedly from ridge sediments (5-6% and 1-0, respectively). The aluminum content of near-bottom suspended matter did not increase over the ridge. It is apparent from these observations that the suspended material collected and analysed was not simply resuspended ridge sediments. Colloidal metal hydroxides have, of course, been observed to precipitate from hydrothermal solutions upon contact with the oxidizing oceanic environment (Zelenov, 1964). The metal enrichment observed in the suspended particulate matter over the Mid-Atlantic Ridge is consistent with the hypo- thesis that hydrothermal exhalations account for the increase in sedimentary trace metal concentra- tions on oceanic ridges (Bostrom and Peterson, 1966; Corliss, 1971). The processes responsible for the observed enrichment of particulate matter in metals would be expected to add silica and fluoride to near- bottom water. Contrary to expectation, no significant increases were found in either silica/ chlorinity or fluoride/chlorinity ratios measured in the same near-bottom water samples from the Mid-Atlantic Ridge crest at 26°N (Fanning, Betzer, Bolger, Miller, McGregor and Rona, 1974). These data indicate that, if hydrothermal solutions had been discharged from this region of the ridge crest, then the silica and fluoride had been diluted to near-background levels and only a small residual of metal-rich particulate matter remained in suspension. Positive temperature anomalies were measured at sites over the southeast and northwest walls of the rift valley at 26°N. The regional evidence cited allows either conductive transfer of heat or hydrothermal discharge as a heat source for these anomalies; however, the geologic setting provides some insight into the nature of the individual anomalies. Over the TAG Hydrothermal Field on the southeast wall of the rift valley an abrupt and narrow temperature increase of 0-1 1°C accompanied by a gradient inversion was meas- ured parallel to the ocean bottom in thermistor 4000-^^^^^^^^^^^^""^^^"^""^- 4000 . 0 10 20 30 , DISTANCE (km) Fig. 5. Reproduction of narrow beam bathymetric profile' A-A' (Fig. 1) across the area investigated. Ther- mistor profiles I and 2, hydrocast Stas. 6 and 9, and STD profile 2, are projected on to the bathymetric profile. profile 1 (Figs. 1 and 2). This temperature anomaly occurred in the vicinity of a step-like topographic level interpreted as a portion of fault block in the wall of the rift valley (Fig. 5) Bottom photographs made concurrently with the temperature increase reveal pillow lavas and basalt breccia with what appears to be hydrothermal material (McGregor and Rona, in press). Identification of the material dredged from the site of the temperature anomaly as hydrothermal is based on the rapid rates of accumulation and purity of manganese oxides present (M. Scott, R. Scott, Rona, Butler and Nalwalk, 1974). Sediment clouds induced by occasional impact of the camera compass in thin sediment patches in the vicinity of the temperature anomaly indicate that near-bottom oceanic currents were negligible during the period of thermistor profile 1. The narrow, abrupt character of this anomaly, as well as its occurrence at a site where hydrothermal material is abundant and where a fault zone may act as a conduit for hydrothermal solutions, favor interpretation that the anomaly resulted from hydrothermal discharge. Over the northwest wall of the rift valley a hydrocast (Sta. 10) measured a relative tempera- ture increase of 0-20°C in the lower 425 m of the water column (Figs. 1 and 3). No increase in either silica/chlorinity or fluoride/chlorinity ratios was found in water samples recovered concurrently with the temperature recording. Downward mixing of overlying water is ruled out as a source 843 Anomalous water temperatures over Mid-Atlantic Ridge crest at 26° North latitude 617 for the temperature increase because the over- lying water column is subadiabatic and therefore stable as determined by STD profile 3 (Figs. 1 and 4). Advection of a warmer water mass from the east or west is considered unlikely because the water was cooler at corresponding depths to the east in the rift valley (Figs. 1 and 3) and to the west in STD 3 (Fig. 6); this observation does not exclude the possibility of advection of warmer water as a bolus rather than as a laterally con- tinuous water mass. Only a few fragments of basalt were recovered in seven attempts to dredge the northwest wall at the site of the temperature anomaly. Hydrothermal material dredged from the southeast wall at the TAG Hydrothermal Field is sufficiently friable so that, if present, it is likely that samples would have been recovered from the site dredged on the northwest wall. The absence of hydrothermal material from the site of the anomaly suggests either that the anomaly is due to in situ conductive transfer of heat, or to advection of a bolus of water which had already been warmed by conductive transfer or hydro- thermal discharge. The question remains why temperature anomalies were not observed at thermistor profile 2 and at Sta. 6 over the TAG Hydro- thermal Field on the southeast wall of the rift valley (Figs. 1 and 3). The intermittent activity of oceanic currents inferred from the presence of ripple marks in thin sediment revealed by bottom photography and the lack of topographic closure create conditions which favor the rapid dissipation of heat. These conditions on the Mid-Atlantic Ridge at 26°N are in marked contrast to those at hydrothermal sites in the Red Sea (Degens and Ross, 1969). In the Red Sea, temperature increases of the order of 40°C were measured within tens of meters of the sea floor where hypersaline solutions discharge into enclosed basins (Ross, 1972). The most probable answer to the question posed is that spatial and temporal variations exist in the generation of temperature anomalies in near-bottom water and that these variations are related to the nature of the heat sources and to the movements of the water. Long-term tempera- ture, salinity, and near-bottom current measure- ments over active oceanic ridges should help to determine the frequency and significance of heat fluxes to the deep ocean (Williams and von Herzen, 1974). Acknowledgements — Charles A. Lauter, Jr., of NOAA, assembled and operated the heat flow system which Dr. George Peter of NOAA loaned to us. Dr. Ants Leetmaa of NOAA advised us on the reduction and interpretation of temperature data and read the manuscript. Drs. Kendall Carder and Susan Betzer of the University of South Florida made helpful suggestions for improving the manuscript. Captain Lavon L. Posey, Cdr. Walter S. Simmons, Lt. Paul M. Duernberger, and other officers and crew of the NOAA ship Researcher provided strong support. This work was supported by NOAA and the Office of Naval Research under contract N 0001 4-72 A-0363-0001 to the University of South Florida. 2000 x 3000 4000 B 0 2000 3000 4000 30 B' 10 20 DISTANCE (km) Fig. 6. Reproduction of narrow beam bathymetric profile B-B' (Fig. 1) across the area investigated. Hydro- cast Sta. 10 and STD profile 3 are projected on to the bathymetric profile. REFERENCES Betzer P. R., G. W. Bolger, B. A. McGregor and P. R. Rona (1974) The Mid-Atlantic Ridge and its effect on the composition of particulate matter in the deep ocean. American Geophysical Union Transactions, EOS, 55, 293. Bodvarsson G., J. W. Berg, Jr. and R. S. Mesecar (1967) Vertical temperature gradient and eddy diffusivity above the ocean floor in an area west of the coast of Oregon. Journal of Geophysical Research, 72, 2693-2694. Bostrom K. and M. N. A. Peterson (1966) Precipi- tates from hydrothermal exhalations on the East Pacific Rise. Economic Geology, 61, 1258-1265. Bryden H. L. (1973) New polynomials for thermal expansion, adiabatic temperature gradient and potential temperature of sea water. Deep-Sea Research, 20, 401-408. 844 618 Peter A. Rona, Bonnie A. McGregor, Peter R. Betzer, George W. Bolger and Dale C. Krause Corliss J. B. (1971) The origin of metal-bearing sub- marine hydrothermal solutions. Journal of Geo- physical Research, 76, 8128-8138. Degens E. T. and D. A. Ross, editors (1969) Hot brines and recent heavy mineral deposits in the Red Sea, Springer- Verlag, 600 pp. Fanning K. A., P. R. Betzer, G. W. Bolger, G. R. Miller, B. A. McGregor and P. R. Rona (1974) Silica and fluoride over the TAG Hydrothermal Field. American Geophysical Union Transactions, EOS, 55, 293. Knauss J. A. (1962) On some aspects of the deep circulation of the Pacific. Journal of Geophysical Research, 2493-3954. Langseth M. G., Jr., L. E. Malone and C. A. Bookman (1972) Sea floor geothermal measure- ments from Vema cruise 25: New York. Lamont- Doherty Geological Observatory, Technical Report 4, CU-4-72, Office of Naval Research Contract N00014-67-A-0 108-0004, Technical Report 2, CU-2-72, National Science Foundation Grant GP5356, Technical Report 2, CU-2-72, National Science Foundation Grant GA1412, 159 pp. Lister C. R. B. (1963) Geothermal gradient measure- ment using a deep-sea corer. Geophysical Journal, 7, 571-583. Lubimova E. A., R. von Herzen and G. B. Udintsev (1965) On heat transfer through the ocean floor. In: Terrestrial heat flow, W. H. K. Lee, editor, American Geophysical Union Geophysical Mono- graph 8, pp. 78-86. McGregor B. A. and P. A. Rona (in press) Crest of Mid-Atlantic Ridge at 26°N. Journal of Geophysical Research. Rona P. A., B. A. McGregor, P. R. Betzer and D. C. Krause (1974) Anomalous water tempera- tures over Mid-Atlantic Ridge crest at 26°N. American Geophysical Union Transactions, EOS, 55, 293. Ross D. A. (1972) Red Sea hot brine area: revisited. Science, 175, 1455-1457. Sclater J. G. and K. D. Klitgord (1973) A detailed heat flow, topographic, and magnetic survey across the Galapagos spreading center at 86°W. Journal of Geophysical Research, 78, 6951-6975. Scott M. R., R. B. Scott, P. A. Rona, L. W. Butler and A. J. Nalwalk (1974) Rapidly accumulating manganese deposit from the median valley of the Mid-Atlantic Ridge. Geophysical Research Letters, 1 (8), 355-358. Scott R. B., P. A. Rona, B. A. McGregor and M. R. Scott (1974) The TAG hydrothermal field. Nature, 251, 301-302. Williams D. L. and R. P. von Herzen (1974) Heat loss from the earth: new estimate. Geology, 2, 327-328. Williams D. L., R. P. von Herzen, J. G. Sclater and R. N. Anderson (1973) Lithospheric cooling on the Galapagos spreading center, East Pacific. American Geophysical Union Transactions, EOS, 55, 244. Zelenov K. K. (1964) Iron and manganese in exhala- tions from the submarine volcano of Banu Wuhu (Indonesia). Dokladv Akademii nauk SSSR, 155, 1317-1320. 845 57 Reprinted from: Geological Society of America, Abstracts with Programs 7, No. 7, 1263. ANiiVAL 'MEETINGS, SALT LAKE CITY, UTAH 1263 SUBMARINE HYDROTHERMAL ACTIVITY AND SEAKLOOR SPREADING AT 2G°N, MAR Scott, Robert B. , Department of Ceology, Texas A&M University, College Station, Texas 77843; Malpas, John, Department of Ceology, Memorial University, St. Johns, Newfoundland, Canada; Udintscv, Gleb, Institute of Oceanology, USSR Academy of Sciences, Moscow, USSR; Rona, Peter A., NOAA-AOML, Miami, Fla., 331^9 On a joint USSR-US cruise of the R/V K.URC1IAT0V in 1975, hydrothermal deposits were found 18 km southeast of the rift valley in the direction of crustal spreading from the TAG hydrothermal field located on the eastern rift valley wall at 26°N, MAR. Veins of hydrothermal manganese oxides and analcite arc within altered basalt talus. The talus is also encrusted with two layer:; of manganese oxides: a basal layer of hydro- thermal manganese (40% Mn, 0.1% Ec, 550ppm Cu, 920ppm Ni , 50ppm Co, 2S00 ppm Ba) that varies between 1 and 10 mm 'thick and an upper layer of hydrogenous forromangancse (11.1% Mn, 16.4% Fe, 670ppra Cn, 1400ppm Ni, 184Qppin Co, 2000ppm Eh) that: varies between 0.5 and 2 mm thick. The hydrotherr.al crust consists of birncssitc that forms subn.etallic , gray- ish-brown laminations; from SKM studies the crust shows the typical box- work of birncssitc plates. In contrast, the hydrogenous layer is extremely botryoidal, amorphous to X-rays and shiny black; S£M studies show eacli botryoid to be massive, without any crystalline textures. The initiation of hydrogenous f erromangancse growths mark the end of local hydrolhcrin.il activity. Using a growth rate of 25 irnn/my based upon the Cu+Ni+Co content, the hydrothermal activity lasted up to 0.25 my after the formation of the rift valley wall. Hydrothermal activity must persist along segments of the rift valley wall for periods of at least one my, forming strips of hydrothennally altered ocean crust. The hydrothermally affected segment of the rift valley wall form-', the young end of a highland ridge that is transverse to the rift axis and parallel to the crustal spreading direction; an abnormal concentration of faults strike parallel to the rift axis and segment this ridge. Cold seawater enters these fault rones, reacts with hot basaltic lock and exits along sin, liar fault zones, forming hydrothermal deposits. 846 58 Reprinted from: Geological Society of America, Abstracts with Programs 7, No. 7, 1285-1286. TEMPORAL AMD SPATIAL SUBSTRATE VARIATION Ifl THE MEW YORK EIGHT APEX Stubblef iel d, W. L., Atlantic Ocecinographic and Meteorological Laboratories, Miami, Horida 331l,9; Permenter, R. U. , Atlantic Oceanographi c and Meteoi olog i ca i Laboratories, Miami, Florida 331'j9, NOAA Quarterly monitoring of selected areas in the New York Bight Apex, over a one and one-half year span, indicates a pronounced temporal and spatial variation within the upper few millimeters of the substrate. By means of sidescan sonar, bottom grab sampling, and bottom photo- graphy, bottoms ranging from coarse, clean sand to muddy, very-fine sand were observed. A temporal variation became apparent when the sampling stations were reoccupied with the aid of precision navigation, The monitoring, which included samples taken shortly before and after a decadal storm, suggests that the substrate mobility is most preva- lent in the vicinity of Long Island and New Jersey shorelines and in Chr I s t iaensen Basin which marks the head of the Hudson Shelf Valley. Sidescan records from nearshore Long Island indicate development of sand wave-like forms during the winter and subsequent degradation during the summer months. The sand wave development is probably associated with the peak-flow events of the winter storms. The fea- tures are of negligible relief, and may be degraded by the action of bottom wave surge and benthic organisms. Chr i s t i aensen Basin, which is characterized by muddy, very-fine sand in its center, is the settling site for much of the sewage sludge material presently being dumped by New York City. Sidescan records show that the Basin is characterized by a patchy bottom pattern. The patches are irregular In shrpe, frequently elongated to the northwest and vary in shape and position between observations which support the suggestion of temporal mobility cf the bottom sediments. 847 59 Reprinted from: No. 1, 337-358. Journal of Sedimentary Petrology 45, SEDIMENT RESPONSE TO THE PRESENT HYDRAULIC REGIME ON THE CENTRAL NEW JERSEY SHELF1 WILLIAM L. STUBBLEFIELD, J. WILLIAM LAVELLE and DONALD J. P. SWIFT National Oceanic and Atmospheric Administration, Atlantic Oceanographic and Meterological Laboratories, 15 Rickenbacker Causeway, Miami, Florida 33149 THOMAS F. McKINNEY Dames and Moore Inc., 14 Commerce Drive, Cranford, New Jersey 07016 Ahstkact: Petrographic data, from vibracores and grab samples collected on the Central New Jersey Shelf, suggest a substrate still actively responding to the hydraulic regime. Radiocarbon dates of shell material from the ridge and swale topography indicates aggrada- tion of the ridge's crest during the last 500 years and exposure of earlier Holocene material in the deeper troughs of the area. The samples from both the cores and the surficial samples were investigated for heavy mineral percentages and grain size analysis in addition to radio- carbon dating. The concentration of heavy minerals into disseminated bands, as observed in the vibracores, is compatible with sediment transport by sand ripples on the ridge's flanks. The grain size variation was subjectively analyzed by applying a 0~mode factor analysis which produced three distinct groupings of the grain size distribution. Each grouping is found to characterize a particular part of the ridge topography. Fine sand and moderate sorting occurs on the flanks, medium to fine sand and moderate sorting occurs on the crests whereas two populations are found in the troughs ; coarse, poorly sorted sands and very line, well sorted sands. This textural variation supports a hypothesis of up- flank rheologic and suspensive transport of medium and fine sand during intense storms and subsequent down- flank winnowing of fine sand during less intense meterological events. The radio- carbon dates indicate that size fractionation and heavy mineral concentrations are subsequent to isolation from a beach environment. INTRODUCTION This paper assesses petrographic and strati- graphic aspects of the central New Jersey Shelf as a guide to the origin of surficial sediments, and of the ridge and swale topography into which they are molded. Origins proposed for this topography include: (a) mainland and harrier beaches overstepped by a transgressive sea (Veatch and Smith, 1939; Shepard, 1963; Emery, 1966; Garrison and McMaster, 1966; Uchupi, 1970) ; (b) subaerial beach ridges dat- ing from earlier Pleistocene stillstands (Sanders, 1962; Dietz, 1963; Kraft, 1971); (c) modern submarine ridges formed subsequent to the lloloccne transgression (Schlee and Pratt, 1972; I Inane, ct al., 1(>72) ; (d) drowned stream inter- fluves ( McKinney and Friedman, 1970); and (c) formation at the foot of the shoreface with subsequent modification by the shelf hydraulic regime ( Swift, ct al., 1972a). 1 Manuscript received January 10, 1974; revised September 23, 1974. This study is one of several detailed exami- nations of the continental shelf substrate in the New York Bight being pursued by NOAA's Miami-based Atlantic Oceanographic and Me- teorological Laboratories in support of NOAA's Marine Ecosystem Analysis (MESA) program. The area was selected as a prominent offshore example of the ridge and swale topography of the Middle Atlantic Bight. This class of shelf floor topography is extensively fished for both surf clams {Spisula) and bottom fish, and is under increasing pressure for use as an ocean dumpsite. For both purposes, it is important to determine the stability of the substrate, and the pattern and rate of sediment transport. This particular study is concerned with present re- sponse of ridge and swale topography on the central New Jersey shelf to the hydraulic regime as indicated by analysis of grab samples and vibracores gathered by R/V Venture during July, 1972. The reader is referred to McKinney, ct al. (1974) for an analysis based on side- scan records and submersible observations. 848 338 W. L. STUBBLEFIELD ET AL. FIRST ORDER HIGHS " CRESTLINES, SECOND ORDER HIGHS Fig. 1. — Index map of the New Jersey shelf with the study area outlined. The first order highs are shown as stippled areas and the crest line of the second order highs as solid lines (From McKinney, et al, 1974). Study Area Fifty kilometers east-southeast of Atlantic City, New Jersey, on the central continental shelf, there exists a pronounced system of large scale ridges and swales at two characteristic wavelengths (Fig. 1). A third, small scale sys- tem consists of bands of contrasting sediment types with negligible relief (McKinney, et al., 1974). These three features appear to be superimposed on 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 et al. (1972b) have related this massif with the an- cestral Schuylkill River. McKinney, et al., (1974) have inferred that the large first order ridges were initiated in a nearshore or estuary mouth environment, possibly tide dominated, while the second order forms were formed at a later period on the open shelf. Examination of a 1:125,000 scales ESSA bathymetric map contoured at one-fathom in- tervals (Stearns, 1967), suggests that the first and second order ridge systems are oriented in a northeasterly direction averaging 54° for the first order system and 40° for the second (Fig. 1). Ridge spacing averages 3.1 km (1.9 nm) with the range varying from 1.2 km (0.75 nm) to 4.85 km (3 nm). Average side slopes are 0.4° with the minimum being 0.1° and a maxi- mum of 0.9°. METHODS Field Methods Grab samples were collected in an area bounded by 39°00'N and 39°10'N latitude; 73°45'W and 74°00'W longitude with naviga- tion provided by dual, automatic tracking Loran A/C receivers. The grab samples were taken by a Shipek bottom sampler approximately every 0.8 km (0.5 nm) along LORAN lines 1.62 km (1 nm) apart (Fig. 2). A more dense sample net, transecting local topography and normal to the initial LORAN lines, was com- pleted in the western sector of the sampling area. It is in this western sector that the maxi- mum ridge slope is found. Each sample was categorized as to crest, flank, and trough from Stearns (1967) bathy- metric map. The classification of the samples was achieved by drawing cross sections normal to the regional bathymetry and dividing the ridge system into respective environments on the basis of change in slope. These environ- ments, crest, flank, and trough, were identified throughout the sectors of the study area by particular depths which correspond to the change in slope. Each sample was associated with an environment by evaluation of that sample relative to its geographic position and depth. Position of sample relative to flank, crest, or trough was checked by comparison with the record of the fathometer which was left running and was annotated during the period of grab sampling. This approach af- fords an objective labelling of each sample relative to the ridge system. In addition to the 191 grab samples collected, four vibracores were obtained. Core V-l was taken on a crest, core V-2 from a trough, core V-3 from the lower reaches of a flank, and core V-4 represents the upper flank (Fig. 2). These four cores, from different parts of the ridge system, afford a total cumulative pene- tration relative to sea level of 15.6 m (Fig. 3). Laboratory Techniques The Shipek grab samples were split from a kilogram to approximately 60 gm using the random scoop technique suggested by Shepard and Young (1961), and were dry sieved using 3-inch U. S. Standard Sieve to separate the fraction coarser than 2 mm. The sand and silt 849 SEDIMENT RESPONSE TO HYDRAULIC REGIME 339 fraction was weighed on a top-loader Metier balance and placed in an ultrasonic water bath for a period of 1 to 3 hours depending- on the amount of silt and clay present in the sand. The ultrasonic bath tends to disaggregate the finer particles (Gipson, 1963) and disperse both the silt/clay and sand size particles (Kravitz, 1966) allowing for a more complete separation of the silt/clay from the sand. The samples were wet sieved using a size 230 mesh to re- move the silt/clay < 0.625 mm and dried. After drying, the samples were weighed to determine the weight percentage of silt/clay. The vast majority of the 191 samples contained less than 1% fines with only 3 samples exceeding 3% fines and 19 samples surpassing 2% concentra- tion of fines. The sand fraction was further reduced by splitting with a microsplitter to 10-gm samples and analyzed in regard to size distribution by a settling tube [Rapid Sediment Analyzer (USA)], calibrated against sieves (Sanford and Swift, 1971). The RSA, when electron- ically coupled with a Hewlett-Packard 9810 calculator, prints out a grain size distribution in quarter phi units and the mean, standard deviation, skewness, and kurtosis of each sam- ple. The values are derived by moment calcula- tions based on a sand fraction recalculated to 100%. The silt-clay fraction was not included in the moment calculations in order to maintain a liydrodynamically similar sample, since the silt/clay material possesses values for incipient motion and entrainment velocity that differ widely from those of sand-size material. Weight percent of heavy minerals was de- termined on selected sand fractions from both the Shipek grab samples and the vibracores. Since the heavier minerals are generally finer than 0.42 mm, the sand was sieved to this size so as to increase the percent of heavy minerals per sample. We were concerned with relative concentration among samples rather than an absolute measurement of concentration. The samples were dried, weighed, and placed in separator)' funnels using tetrabromoetbane ( CI Ilir.j-CI 1 1'.r._, ) to effect separation. The heavy minerals were drained from the funnel, dried, weighed, and a weight percentage calcu- lated (Table 1 ). The four vibracores, after removal from the core barrel, were maintained in an upright position at approximately 5°C. The cores were transversely cut into 70 cm lengths and radio- graphed. The cores were then longitudinally split, immediately described for color, structure, lithology, fauna] content, and photographed. Table 1. — Percent of heavy minerals in the sand V-l V-4 Surficial Sample % Sample % Sample % 75 cms 0.85 2 cms 2.27 19 (T) 15.16 85 1.14 10 4.29 22 (F) 3.92 95 6.64 20 1.63 24 (F) 4.34 100 3.74 40 1.86 26 (C) 7.99 133 0.04 45 1.31 28 (C) 4.21 140 1.35 50 4.98 30 (F) 5.34 150 1.54 60 5.13 38 (F) 4.50 160 1.19 70 2.43 40 (F) 3.90 170 0.24 80 3.30 42 (F) 4.99 175 0.40 90 4.35 48 (C) 22.09 185 0.17 100 2.12 58 (T) 39.4 195 1.18 110 3.97 7.1 (C) 6.18 129 1.03 76 (T) 5.59 130 4.33 80 (C) 4.54 140 3.52 102 (C) 6.09 160 4.78 103 (C) 6.46 170 2.61 104 (C) 6.10 ISO 1.87 112 (T) 5.39 l')0 2.52 200 1.75 210 3.61 220 1.56 240 0.64 250 1.60 260 2.15 270 2.51 The samples from a core are prefixed by a "V." The surficial sands are marked to reflect the part of the ridge system the sample was taken, e.g. Crest (C), Flank (V), or Trough (T). One half of the split core was sampled every 10 cms within the sand units or at any observ- able grain size discontinuity and processed sim- ilarly to the grab samples in order to determine grain size distribution and related statistical parameters. The other half of the core was partially impregnated with C1BA epoxy #6005 and 6010 and CI HA hardeners #830 and 850 to a recipe of 4:4:1:3. The partial impregna- tion, in addition to providing archival material, facilitates the investigation of in situ grain size distribution as well as other sedimentary struc- tures. Shell material from cores V-2, V-3, and V-4 were radiocarbon dated by facilities at the De- partment of Geology, University of Miami, Florida. Samples were selected in an attempt to establish the Plcistoccnc-I foloeene boundary and to determine if ridge aggradation has oc- curred subsequent to passage of the beach en- vironment. See Table 2 for the samples, sample location, depth in the core, and radiocarbon dates. Factor Analysis To detect patterns in the regional distribution of grain size characteristics in the l('l grab 850 340 W. L. STUBBLEFIELD ET AL. Fig. 2. — Bathymetry of the study area. The Shipek grab samples are shown as solid circles and the vibracores as squares. The core number appears beside the position. samples, a Q-mode factor analysis was applied. Factor analysis, besides its inherent ability to process copious quantities of bulky data quickly, provides an objective approach to the interpreta- tion. Factor analysis is rapidly becoming a com- mon tool among sedimentologists and thus does not warrant a detailed discussion of techniques. Interested readers, however, are referred to the papers of Imbrie and Van Andel (1964) and Klovan (1966). The data input consists of the grain size distribution of individual samples. For this study three eigenvalues were found to represent 94% of all sample interrelations. The remaining 6% are scattered among several eigenvalues and may be discarded as ambient noise. With the exception of three, all samples have a com- munality above 0.80 and the majority exceed 0.90, highly suggestive that a satisfactory group- ing description is obtained for the bulk of the samples. EVOLUTION OF THE STUDY AREA Quaternary History Based on Radiocarbon Dating and Stratigraphy The four vibracores from the study area (Fig. 2) provide an insight to its late Quaternary history. Radiocarbon dates of fauna found in the cores (Fig. 3) give maximum ages varying from less than 500 years B.P. for the ridge sand units to greater than 36,000 years B.P. for silty-clay beneath the surficial sands in the troughs. Using a date of approximately 16,000 years B.P. for the Pleistocene-Holocene bound- ary in this area (Emery and Uchupi, 1972), the cored part of the ridges are late Holocene, the surficial trough sands are middle Holocene, and the underlying silty clay (cores V-3 and V-2) is late Pleistocene. These dates and the lithol- ogies of their matrices allow us to infer the late Quaternary history. Radiocarbon dates, site elevation, and depth of core penetration suggest that cores V-2 and V-3, which represent late Pleistocene marine Table 2. — Radiocarbon dates of selected fauna from the study area Core Position Fauna dated Depth below sea level Apparent age (years B.P.) V-2 39°05.1'N 73°55.6'W V-2 V-3 Same 39°05.7'N 73°50.5'W V-3 V-3 Same Same V-3 Same V-4 39°06.9'N 73°31.3'W V-4 Same Trough 39°05.6'N Sample #156 73°55.1'W Mcrccnaria mercenaria Crassostrea virginica Mercenaria mercenaria Ensis directus Mercenaria mercenaria Mercenaria mercenaria Placopecten magcllanicus Placopecten magcllanicus Crassostrea virginica Mercenaria mercenaria 45.63 m (83) 46.95 m (125) 41.88 m (8) 43.30 m (150) 44.30 m (250) 45.50 m (370) 36.90 m (60) 37.55 m (125) 44.00 m 29,700 ± 650 32,150 10,950 600 360 22,035 ± 665 25,300 ± 1040 1200 > 36,000 <500 3,760 ± 70 10,050 ± 170 The depth from top of the core to the dated fauna, in centimeters, Below Sea Level. appears in parentheses beside Depth 851 SEDIMENT RESPONSE TO HYDRAULIC REGIME 341 SEA LEVEL- _^ CROSS -BEDDING [ ] SAND 1 SILTY CLAY SHELL MATERIAL HEAVY MINERAL CONCENTRATION (ANALYSIS) HEAVY MINERAL CONCENTRATION (RADIOGRAPH) -■>-?- DISCONFORMITY PHI SIZE () SORTING M SKEWNESS (Sk) H/P HOLOCENE PLEISTOCENE CONTACT HOLOCENE LAGOONAL DEPOSIT HOLOCENE SHELF SAND Fie. 3. — Lithic lop of the 4 vibracores and mean grain size in quarter plii (. ( ) = Standard deviation of the sample populations. process which tended to remove the coarser par- ticles from the quartz rich parent population, with the efficiency of coarse grain removal in- creasing with decreasing specific gravity. His conclusions were based on a comparison of grain diameters of the sample with settling velocity of equivalent diameters for each mineral species. Although Stapor does not specify, such selective removal of coarser grains would presumably result in a finer sand substrate exhibiting posi- tive skewness described by Everts (1972) and ourselves. Stapor also describes heavy mineral popula- tions on "sheltered beaches," which apparently result from a concentrating process which tends to remove finer particles from the heavy mineral population. Samples of such a deposit would presumably be coarser and negatively skewed as in the case of the lower heavy mineral zones of core V-4. Stapor's samples are from beaches and there- fore, as clearly indicated by Everts (1972), can- not be directly compared with heavy mineral deposits found beneath mean low water. How- ever, Stapor believes that the heavy mineral deposits of his "sheltered" Gulf beaches were concentrated on the shelf floor and were then delivered as slugs of heavy mineral-rich sand to the beaches. He suggests that the heavy mineral deposits of sheltered beaches were fine- truncated by asymmetrical wave surge on the shelf whereas the deposits of the open beaches were coarse-truncated in a higher energy en- vironment where coarser grains on the deposi- tional surface experienced greater dispersive pressure (see Bagnold, 1954), and hence a greater probability of entrainment. The geographic connotations of Stapor's analysis are not directly applicable to the off- shore cores of this study, and his evaluation of the responsible hydraulic mechanisms should be applied to our heavy mineral deposits with caution. However, it seems that two different concentrating mechanisms were operating dur- ing the formation of the heavy mineral bands in cores V-l and V-4. Differential response to dispersive pressures experienced by grains on a current-swept depositional surface may indeed have been responsible for the coarse-truncated crestal deposits. In the case of the fine-trun- cated sands in the lower portion of core V-4, asymmetrical wave surge is a feasible mecha- nism, especially in that the seaward-sloping outer flank is presumably a dynamic analog of the regime of shoaling waves on the sloping shore face. This is particularly true during storm events when wave surge can penetrate that deep. PRESENT HYDRAULIC PROCESSES AND SEDIMENT RESPONSE Petrography of Surficial Sediments Distribution of mean diameters. — The mean grain size of the sand fraction within the study area varies from 0.63 mm to 0.13 mm, corre- sponding to a range of coarse to fine sand on Wentworth's (1922) scale of particle size. The textural variation is considered relative to a particular part of the complete ridge system with each sample grouped as to crest, flank, or trough (see methods). The statistical moments of each separate grouping appears as part of Table 3. When the samples are plotted on Stearns (1967) bathymetric map, the textural variation appears to be a complex function of bathymetry (Fig. 6). This is especially apparent in the western 855 SEDIMENT RESPONSE TO HYDRAULIC REGIME 345 39°10'N 39°00'N 73"55'W 73°45'W Kic. 6. — Isopleth map of the mean grain size of the study area in relation to the 20-fathom contour. Where control is poor, textural boundaries have been defined by depth, since depth has been independently shown to be a major control of grain size (Fig. 9). sector of the study area where the grab sample net is sufficiently dense to deduce subtle trends in grain size facies. The crestal sands are of medium grain size, well sorted (low standard deviation), and con- tains only a trace of a population larger than sand size (Table 3). The flank sands, when compared to the crestal sands, are finer, of comparable sorting, but possess a larger amount of the very coarse material. The trough sands, however, tend to be either fine or coarse. The fine trough sands are much finer than the flank sands and with a higher degree of sorting. The amount of granular size material (> 2.0 mm) in the fine trough sands is much greater than that on the crest but is somewhat less than that found on the flanks. The coarse sand is on the opposite end of the spectrum, as reflected in Table 3, and is atypical of any other part of the ridge system. The coarse sands have mean values on the coarse end of Wentworth's (1922) medium grain classification, sorting which is very poor (large standard deviation) and con- tains large amounts of material coarser than sand-size. In a few of the coarse trough sands the amount of material larger than 2.0 mm exceeds 20%. This granule size material in the troughs ap- pears as lag deposits of detritus, clay pebbles, and large fragmented pelecypods, primarily Cras- 856 346 W. L. STUBBLEFIELD ET AL. 39°10'N 73°55'W 39o0CCN 73°45'W Fig. 7. — Isopleth map of material finer than sand size particles (< 4.0 io SAMPLE 85 X =2.57 ct = 0.34 Sk = 0.01 ■1.0 10 2.0 3.0 4.0 PHI SIZE SAMPLE 76 _ X =1.69 £ 20r Sk=0.38 SAMPLE 4 _ X = 0.94 S. 0--O.72 > 20r Sk-0.29 10 u. 0 ■1.0 1.0 2.0 PHI SIZE 3.0 4.0 Fig. 8. — Normalized factor components of the surficial samples and associated histograms of representa- tive samples. Factor I denotes the eigenvalue of greatest influence on all the samples, whereas factor III represents that of least influence. The samples are symbolized to indicate the part of the ridge system from which they were taken. The histograms display the mean grain size (x), standard deviation ( 2.5 * 3.5 4, 13.19 25.46 25.46 25.46 207.98 305.21 267.59 180.38 11 X 108 13 X 10" 92 X 108 50 X 108 The % F.xceedence column reflects the artificiality introduced when considering current data only in 5 cm/sec intervals and not as a continuous function. The suggested quantity of sediment transport is hased on Shield's (1936) function of critical velocity and Laursen's (1958) equation for total load concentration. to suspensive transport events. This suggestion is supported by the CM diagram (Fig. 10) with the fine grained factor TT sands plotting on the portion of Passega's (1964) diagram where transport is postulated to be via suspension and some rolling. Model for Hydraulic Process and Substrate Response Hydraulic Regime. — The stratigraphic and petrographic observations of ridge and swale sands do not provide a unique solution for the hydraulic regime that has distributed the sands. They do, however, provide con- straints for deducing the nature of this regime. Baldly stated, our data lead us to infer that since the passage of the beach, the troughs of the ridge and swale system are subjected alter- nately to coarse, high intensity, bedload de- position and fine suspensive deposition ; crests are uniquely subjected to a regime characterized by high intensity flow and suspensive deposition with an occasional interval of dune formation ; and flanks are subject to all of these regimes at different times, with suspensive transport dom- inating. These inferences must be considered relative to our scanty knowledge of the hydraulic cli- mate of the central New Jersey shelf. During the summer, the water column is stratified and undergoes a southerly drift, partly as a baro- clinic response to terrestrial runoff ( Bumpus, 1973). Surface tidal current velocities are generally less than 2D cm/sec (Redfield, 1956), and the tides are therefore generally inadequate to move bottom sand. During the fall, decreasing temperatures and the increased frequency of storms homogenizes the water column. During the months of De- cember, January, and February, storms cross the New Jersey shelf approximately every ten days. Repeated episodes of intense northeast winds might be expected to drive shelf water land- ward, and where constrained by the shoreline, southward. Harrison and Pore (1967), have inferred that in deeper water a geostrophic component of flow, also southward, may occur in response to such a wind setup. Whereas there are no published measure- ments of the winters' flow field, a partial docu- mentation of the summer flow field velocity is available from limited current observations made by McClennen (1973). McClennen's "C" meter, 100 cms above the substrate, was de- ployed during a ten-day period of the quiet summer season, but seems to have caught a mild summer storm in which sustained velocities of 35 cm/sec occurred. The velocity record was used to calculate the percent of time that the threshold velocity was exceeded for four repre- sentative size classes of sand and the resulting transport, by mass and number of grains (Table 4). The figures should not be taken too literally as many assumptions are involved and the equations used (Shields, 1936; Laursen, 1958) are based on shallow, steady flow in flumes, but they suffice to show that significant trans- port occurs. Ten transport events of this mag- nitude per year would be able within several decades to transport, over the surface area of a ridge, an amount of sediment equal to the volume of the ridge, yet the ridges rest on a surface ten millennia old. We tentatively con- clude that either the ridges are quite young, or else the flow field in the millennia since trans- gression has been structured in such a way as to maintain the ridges, otherwise they would have been degraded and flattened. A mechanism capable of maintaining ridges may exist in the storm flow field, but very little is known about this subject. Current- parallel bedforms are generally attributed to 864 354 W. L. STUBBLEFIELD ET AL. -SEA LEVEL' -SEA LEVEL*- FAIR WEATHER TRANSITIONAL 111,11 INTENSE STORM 111(467.), 11(25%) SCHEMATIC DISTRIBUTION OF FACTORS VERTICAL EXAGGERATION = 230 „ SAND TRANSPORT UNDER BIOGENIC AND GRAVITATIONAL FORCES SUSPENDED SAND TRANSPORT COMPONENT TRACTIVE SAND TRANSPORT COMPONENT BOTTOM CURRENT COMPONENT NORMAL TO PLANE OF PAPER. SIZE DENOTES RELATIVE STRENGTH OF COMPONENT. ,11 SAND TYPE ACTIVE FOR A PARTICULAR STAGE Fig. 11. — Model for sediment transport with no interpretation of internal structure intended. The numerals denote the percent of the three factor sands found on the crest, flank, or in the trough. current-parallel perturbations in flow (Allen, 1968), and these are usually described as arrays of horizontal helical vortices, with alternately right-handed and left-handed rotation, so that a cell shares a rising limb with a cell on one side, and a descending limb with a cell on the other (Wilson, 1972). The secondary flow com- ponent is perhaps a few percent of the main flow ; both components are most intense in the downwelling limbs. The case is perhaps best documented on the sand seas of the world's deserts, where wind parallel sand dunes as large as the New Jersey shelf ridges have required up to 10 millennia to form (Wilson, 1972). Cloud streets are some- times aligned with their crests (Hanna, 1969), and cloud streets are generally attributed to helical flow structure in the planetary boundary layer of the atmosphere (Brown, 1971), with clouds forming along the rising limbs of the half-cells. The tidal sand ridges of the southern bight of the North Sea have also been attributed to helical flow (Houbolt, 1968), although the situation is rendered complicated by residual ebb and flood currents (Huthnance, 1973). Theo- retical studies (Faller, 1963; Faller and Kaylor, 1966) suggest that all large-scale flows on a rotating planet that are shallow relative to their depth (Ekman boundary layer flow) are un- stable above critical Reynolds values, and that the instability tends to take the form of helical flow structure. Coupling between such large scale flow cells and the ridge topography if it does occur, may be a sporadic phenomenon. Sidescan records from the study area (McKinney et al., 1974) show a pervasive pattern of sand ribbon- like features occurring mainly in troughs. These are themselves probably the consequence of small-scale helical flow structure. The ribbons trend uniformly along the topography, making a 30° oblique angle with the trough axes. If this is a record of the last major storm, then this storm, or at least its last stage, does not appear to have had a flow direction appropriate for coupling, since the sand ribbons do not form herring bone patterns about trough axes. Storm flow directions on continental shelves are variable (Sternberg and McManus, 1972, Fig. 78) and may rotate within a single storm. 865 SEDIMENT RESPONSE TO HYDRAULIC REGIME 355 If flow parallels topography during peak flow intensity, then flow cells could lock onto topog- raphy during this period, to scour troughs and aggrade crests. Tf the trend of second order ridges ( Fig. 1 ) is determined by mean storm flow, then such coupling may be sufficiently frequent to maintain ridges. Whereas no direct evidence exists for the perhaps chimerical helical flow, the petrography of the J Tolocene sands and their morphology and stratigraphy do indicate sediment transport normal to the ridges, as well as parallel to them. The pattern of transport, furthermore, appears to be time dependent, varying with the rigor of the hydraulic event (Fig. 11). Substrate Response to High Intensity Flow. — The axial zones of coarse factor 111 sands on trough floors is attributed to a high intensity flow field, possibly of the nature described above, in which high velocity, wind-driven, sur- face flow converges over troughs, descends and diverges on the bottom. The regime is an ero- sive one; submersible dives (McKinney, ct al., \i>74) show that zones of the thin factor III sands locally have been swept away, exposing ribbon-like windows of coarser shelly, pebbly sand. This lag deposit includes shells of Cras- sostrca virginica and Mcrccnaria mcrccnaria dated at 10,050 years H.I'. These brackish water forms appear to have been scoured out of the underlying clayey substrate. The texture of factor I sands of the crests appears to also form during high intensity flow. Core \ -4 (Fig. 3) indicates that the high in- tensity regime of the crests is at least locally depositional. Radiocarbon dates from pelecypod shells ( I'lacopecten inagellanieus) show that this core has received over a meter of sand in the last 3,760 years, and 60 cm in the last 500 years. I luring flow sufficiently intense to activate coarse factor III sands of the troughs and medium factor 1 sands of the crest, the fine factor I I sands must also be activated. Their texture implies suspensive transport. The up- slope coarsening of these sands, noted also by McKinney, ct al. (1974), could result in part from high intensity flow events if trough di- vergence and crestal convergence of bottom flow are in fact associated with these events. Progressive sorting would result as bottom flow decelerates from the intense flow of the down- welling zone, to the more diffuse flow upwelling over crests. McKinney, ct al. (1974), have noted that the sand ribbon-like patterns of trough floors locally extend into the finer sands of flanks. The bed forms develop up to 2 m of relief in this fine, deep substrate, and become widely spaced. Substrate Response to Moderate Intensity Flow. — The presence of fine factor II trough sands and of medium, well sorted crestal sands which contain a negligible component of factor II re- quires a yet more complicated scheme. The ridge crests at 31 m depth must be activated much more frequently than the troughs at 45 m. Acti- vation restricted to crests should result from long period swells (storm forerunners), waning of an intense storm, or from wind stress to short-lived for downward momentum transfer to reach the bottom during periods of partial stratification. Such activity might reasonably be expected to winnow the crests and deposit fine sand, silt, and clay on flanks and in troughs as a graded suspension (Fig. 11 ). The finest sediments, when deposited, present hydraulically smooth surfaces whose grains do not protrude above the laminar sublayer, and in addition, tend to have some cohesiveness. Sub- mersible dives have revealed widespread "rusty bottoms" during fair weather; sectors in which the sand-water interface has developed an algal film. As a consequence of these several mecha- nisms, a higher velocity is required to entrain these materials than to deposit them. This be- havior probably explains why factor II sands extend across the margins of troughs whose axes appear to be subject to periodic high energy events. Substrate Response to Fair Weather Regime. — During fair weather circulation, which must prevail the majority of the time, the ridge crests may undergo some surficial ripple sorting, but the deeper surfaces appear to be dominated by biologic activity (Fig. 11). Submersible ob- servations indicate that sand dollars and stone crabs are abundant on the crests and upper flanks, and crabs are active on the lower flanks and in the troughs as well. Roth rippling and benthic activity would presumably induce minor downslope transport. CONCLUSION Substantial evidence exists to suggest that the ridge and swale system in the central New Jersey shelf is actively responding to the modern hydraulic regime. Despite considerable evi- dence for mobility of the bottom sediment, the ridges and swales maintain their form. Troughs expose windows of coarse lag sands over pre- recent substrate, indicating an erosional regime. The troughs show little evidence of sediment 866 356 IV. L. STUDBLEFIELD ET AL. accumulation, and many samples in fact present textural evidence indicative of high intensity flow. Ridge crests likewise present evidence of high intensity flow, but are at least locally ag- gradational, as a core has been dated to reveal a meter of aggradation in the last 3,760 years, and 60 cm in the last 500 years. The petrographic and stratigraphic data pre- sented do not provide a unique solution for the flow field that activates the surficial sediments of the central New Jersey shelf. They do, how- ever, provide inferential constraints. The sim- plest model compatible with the data recognizes three stages of activity (Fig. 11). During fair weather circulation, the activity on the ridge system is mainly biogenic with sand dollars being prevalent on the crest and upper flank and crabs most active in the troughs. A second stage of activity results from storms which fail to entrain the complete water column on the shelf, because wind velocities are not sufficiently intense or sustained or because the water column is too stratified. During such periods crests may undergo winnowing by wave surge and by the unidirectional component of the storm flow field. Factor I sand texture may be formed on crests, while factor II sand tex- tures are formed on the flanks and troughs, when fine sands move off crests as bedload or graded suspension or both. A third stage may be inferred to result from major storms, when the entire shelf water column is set into motion. Such movements may involve large scale secondary circulation with descent and divergence of high velocity surface water in the troughs. The troughs undergo scour, during which time the fine factor II sands develop their characteristic texture of abundant "B" and "C" subpopulations. Since pre-Recent substrate underlies trough axes at the depths of less than a meter where sampled, the trough bottom currents may diverge, thus sweeping out the troughs and returning sand to ridge flanks and crests. Some of this ma- terial may aggrade flanks and crests where factor II and I textures are respectively im- printed but the finer sand cannot undergo permanent deposition on the crests as wave surge is most intense here, and returns to the flanks and troughs as the storm wave surge and the mean flow return to the weaker intermediate stage of activity. The resultant facies include medium grained sand on the crests, an admixture of grain-sizes on the flanks but with a preponderance of fine- grained sand, and both coarse and fine deposits in the troughs with little medium-grained sand present (Fig. 11). The ridges appear to be maintained by differential sediment transport with net upflank transport of medium sand, and net down flank transport of fine sand. Modern sediment movement in the ridge and swale topography is confined to the outermost skin of sand; the cores may have been formed throughout much or all of the preceding period of Holocene transgression. However, it is inadvisable to call these features relict for this reason; all bedforms consist of active skins and inert cores. Large scale bedforms migrate much more slowly than small scale bedforms (Allen, 1968), and current parallel bedforms move yet more slowly since their migration is a response mainly to secondary components of flow. The criterion for bedform activity should not be age of the core, but instead whether or not the bedform's skin processes serve to maintain it. ACKNOWLEDGMENTS The authors are grateful to Dr. H. B. Stewart, Jr. and Dr. G. H. Keller for constructive crit- icism offered to this study. Special apprecia- tion is extended to Sue O'Brien and David Senn for the illustrations. Dr. Don Moore, of the Rosenstiel School of Marine and Atmospheric Science, University of Miami, Florida, provided biological identification. REFERENCES Allen, J. R. L., 1068, The nature and origin of bedload hierarchies: Sedimentologv, v. 10, p. 161-182. Bagnold, R. A., 1954, Experiments on a gravity-free dispension of large solid in a Newtonian fluid under shear : Proc. Royal Soc. London, A 225 p. 49-63. Brown, R. A., 1971, A secondary flow model for the planetary boundary layer : Jour. Atmo- spheric Sciences, v. 27, p. 742-757. Bumi'US, D. F., 1973, A description of the circula- tion on the continental shelf of the east coast of the United States : Progress in Oceanography, v. 6, chp. 4, p. 117-157. Curhay, Joseph R., 1965, Late Quaternary history continental shelves of the United States : in Wright, H. E., Jr., and D. G. Frey (eds.), the Quaternary of the United States; Princeton, N. J., Princeton Univ. Press, p. 723-735. Dietz, R. S., 1963, Wave-base, marine profile of equilibrium and wave built terraces ; a critical appraisal : Geol. Soc. America Bull, v. 74, p. 971-990. 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 Swift, D. J. P., D. B. Duane, and O. H. Pilkey (eds.), Shelf Sediment Trans- port Process and Pattern ; Dowen, Hutchinson, and Ross, Inc., Stroudsburg, Pa., p. 447-449. Emery, K. O, 1966, Atlantic continental shelf and 867 SEDIMENT RESPONSE TO HYDRAULIC REGIME 357 slope of the United States geologic background. U. S. Geol. Survey Prof. Paper 529-A, p. 1-23. — , and E. Uchupi, 1972, Western North At- lantic Ocean ; Topography, Rocks, Structure ; Water, Life, and Sediments: Am. Assoc. Pe- troleum Geologists Bull., Memoir 17, p. 532. Everts, Craig H., 1972, Exploration for high energy marine placer sites: Univ. of Wise. Sea Grant Technical Report 72-210, p. 179. Fai.LER, A. J., 1963, An experimental study of the instability of the laminar Ekman boundary layer: Jour. Fluid Mech., v. 15, p. 560-576. , and R. E. Kaylor, 1966, A numerical study of the instability of the laminar Ekman bound- ary layer : Jour. Atmospheric Sci., v. 23, p. 466-480. Frank, W. M., and G. M. Friedman, 1973, Con- tinental Shelf Sediments off New Jersey: Jour. Sed. Petrology, v. 43, p. 224-237. Garrison, L. E, and R. C. McMaster, 1966, Sedi- ments and geomorphology of the continental shelf off southern New England: Marine Ge- ology, v. 14, p. 273-289. Gipson, Mack, Jr., 1963, Ultrasonic disaggregation of shale : Jour. Sed. Petrology, v. 33, p. 955- 958. Hai.d, A., 1952, Statistical Theory with Engineering Application, New York, John Wiley and son, 783 p. Hanna, S., 1969, The formation of longitudinal sand dunes by large helical eddies in the atmosphere : Jour. Applied Meterology, v. 8, p. 874-883. Harrison, W., and N. A. Pore, 1967, An approach to correlation and prediction in the drift-run- off-wind system: (In) Circulation of continen- tal shelf waters off the Chesapeake Bight, ESSA Adm., Prof. Paper 3, p. 1-82. Houboi.t, J. J. H. C, (1968) Recent sediments in the southern bight of the North Sea : Geol. en Mijnbouw, v. 47, p. 245-273. Huthnance, J. M., 1973, Tidal current asymmetries over the Norfolk sandbanks : Estuarine and Coastal Marine Science, v. 1, p. 89-99. Imbrie, J., and T. H. Van Andel, 1964, Vector Analysis of Heavy Mineral Data: Geol. Soc. America Bull., v. 7~5, p. 1131-1156. Kennedy, J. F., and F. A. Locher, 1972, Sediment suspension by waves: in Meyer, R. E. (ed.). Waves on Beaches and Resulting Sediment Transport : Academic Press, London. Ki.ovan, J. E, 1966, The use of factor analysis in determining depositional environments from grain size distribution : Jour. Sed. Petrology, v. 36, p. 115-125. Kraft, J. C, (1971) Sedimentary facies patterns and geologic history of a Holocene transgression: Geol. Soc. America. Bull., v. 82, p. 2131-2158. Kravitz, H. H., 1966, Using an ultrasonic disruptor as an aid in wet sieving : Jour. Sed. Petrology, v. 36, p. 811-812. Laursen, E. M., 1958, The total sediment load of streams: Proc. Amer. Soc. Civil Engineers, v. 84 (HYI), p. 1530-31 to 1530-33. McCi.ennon, C. F.., 1973, New Jersey continental shelf near bottom current meters records and recent sediment activity: Jour. Sed. Petrol., v. 43, No. 2, p. 371-380. McKinnev, T. F., W. L. Stuhbii iiii.d, and D. J. P. Swift, 19/4, Large scale current lineations on the Great Egg Shoal Retreat Massif, New Jersey Shelf: Investigations by sidescan sonar: Marine Geology, v. 17, p. 79-102. and G. M. Friedman, 1970, Continental shelf sediments of Long Island, New York: Jour. Sed. Petrology, v. 40, p. 213-248. Mii.liman, J. D., and K. O. Emery, 1968, Sea levels during the past 35,000 years: Science, v. 162, p. 1121-1123. Moss, A. J., 1962, The physical nature of common sandy and pebbly deposits (Part I) : Am. Jour. Sci., v. 260, p. 337-373. , 1963, The physical nature of common sandy and pebbly deposits. (Part II) : Am. Jour. Sci., v. 261, p. 297-343. , 1972, Bedload sediments: Sedimentologv, v. 18, p. 159-219. Passega, R., 1964, Grain size representation by CM patterns as a geological tool : Tour. Sed. Pe- trology, v. 34, p. 830-847. , A. Rizzini, and G. Borghetti, 1967, Trans- port of sediments by waves, Adriatic coastal shelf, Italy: Am. Assoc. Petroleum Geologists Bull, v. 51, p. 1304-1319. Redfiei.d, A. C, 1956, The influence of the continen- tal shelf on the tides of the Atlantic coast of the United States: Jour. Marine Res, v. 17, p. 432-448. Sanders, J. F, 1962, North-south trending submarine ridge composed of coarse sand off False Cape, Virginia: Abstr. Am. Assoc. Petrology Ge- ologists Bull, v. 46, p. 278. Sanford, R. B, and D. J. P. Swift, 1971, Compari- son of sieving and settling techniques for size analysis, using a Benthos Rapid Sediment Ana- lyzer: Sedimentology, v. 17, p. 257-267. Schi.ee, J, and R. Pratt, 1972, Atlantic continental shelf and slope of the United States: U. S. Geol. Survey Prof. Paper 529-H. Shepard, F. P, 1963, Submarine Geology, New York, Harper and Row, 557 p. , and R. Young, 1961, Distinguishing between beach and dune sands : lour. Sed. Petrology, v. 31, p. 196-214. Shields, A, 1936, Anwendung der ahnlichkeits mechanik und der turbulenzforsechung auf die Geschiebe Beweguna: Preussische Versuchan- stalt fur Wasserbau und Schiffbau, Berlin. Soi.ohub, J. T, and J. E. Kolovan, 1970, Evalua- tion of grain-size parameters in lacustrine en- vironments : Jour. Sed. Petrology, v. 40, p. 81- 101. Southard, J. B, and L. A. Boguchwai, 1973, Flume experiments on the transition from ripples to lower flat bed with increasing grain size : Jour. Sed. Petrology, v. 43, p. 1114-1121. Siapor, F. W, Jr., 1973, Heavy mineral concentrat- ing processes and density/size equilibrium in the marine and coastal dune sands of the Apolachiola, Florida, region : Jour. Sed. Petrology, v. 43, p. 396-407. Stahl, L, J. Koczan, and D. J. P. Swift, 1974, Anatomy of a shoreface-connected sand ridge on a New Jersey shelf: Implications for the genesis of the shelf surficial sand sheet: Marine Geology, v. 2, p. 117-120. Sternberg, R. W, and D. A. McManus, 1972, Implications of sediment dispersal from long- term bottom current measurements on the con- 868 358 W. L. STUBBLEFIELD ET AL. tinental shelf of Washington : in Swift, D. J. P., D. B. Duane, and O. H. Pilkey (eds.). Shelf Sediment Transport : Process and Pattern : Dowden, Hutchinson and Ross, Stroudsburg, Pa., p. 181-195. Sterns, 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 Ad- min. Rockville, Md. Swift, D. J. P., B. Holliday, N. Avicnone, and G. Shideler, 1972a, Anatomy of a shorefaced ridge system, False Cape, Virginia: Marine Geology, v. 12, p. 59-84. , J. W. Kofoed, F. P. Saulsburg, and P. Sears, 1972b, Holocene Evolution of the Shelf Surface, Central and Southern Atlantic Shelf of North America: in Swift, D. J. P., D. B. Duane, and O. H. Pilkey (eds.), Shelf Sediment Transport Process and Pattern; Dowden, Hutchinson and Ross, Inc., Stroudsburg, Pa., p. 499-575. Uchupi, E., 1970, Atlantic continental shelf and slope of the United States shallow structure : U. S. Geol. Survey Prof. Paper 529-1, 44 p. Veatch, A. C, and P. A. Smith, 1939, Atlantic submarine valleys of the United States and the Congo submarine valleys : Geol. Soc. America Spec. Papers no. 7, 101 p. Visher, G. S., 1969, Grain size distribution and depositional processes : Jour. Sed. Petrology, v. 39, p. 1074-1106. Wentworth, C. K., 1922, A scale of scale and class terms for clastic sediments : Jour. Geology, v. 30, p. 377-392. Wilson, I. G, 1972, Aeolian bedforms — their de- velopments and origins : Sedimentology, v. 19, n. 3/4, p. 173-210. 869 60 Reprinted from: Sedimentary Geology 14, 1-43, BARRIER-ISLAND GENESIS: EVIDENCE FROM THE CENTRAL ATLANTIC SHELF, EASTERN U.S.A. DONALD J. P. SWIFT Atlantic Oceanographic and Meteorological Laboratories, Miami, Fla. (U.S.A.) (Submitted July 9, 1974; revised and accepted January 31, 1975) ABSTRACT Swift, D.J. P., 1975. Barrier-island genesis: evidence from the central Atlantic shelf, east- ern U.S.A. Sediment. Geol., 14: 1-43. Since most barrier systems appear to have retreated into their present positions from further out on the continental shelf, the continental shelf is a logical place in which to investigate barrier genesis. The Middle Atlantic Bight of North America, one of the best known shelf sectors, does not appear to contain any drowned barriers. Instead, a series of terraces bear on their surfaces a discontinuous carpet of lagoonal sediments beneath a discontinuous sand sheet formed by erosional barrier retreat. Scarps separating terraces are the lower shorefaces of stillstand barriers whose superstructures were destroyed when shoreface retreat resumed. Thus the "origin" of most barriers is that they have retreated in from the position of their immediate predecessors. Barrier genesis, in the classic sense of large-scale, coastwise spit progradation or mainland-beach detachment, could only have occurred at Late Wisconsin lowstand, when the sense of sea-level displacement was reversed. The relative roles of coastwise spit progradation and mainland-beach detach- ment depend on coastal relief and slope, with steep, rugged coasts favoring spit prograda- tion at the expense of mainland-beach detachment. Since most major barrier systems form on flat coastal plains, it would appear that mainland-beach detachment is the more important mode of barrier formation. During stillstands or periods of reduction in the rate of sea-level rise, coasts can more nearly approach their climax configuration, in which the shoreline is relatively straight, and the shoreface is well developed and of maximum possible slope. Coastal adjustments during such periods may require localized mainland-beach detachment and coastwise spit progradation, in order to attain such a configuration. INTRODUCTION Contemplation of the origin of barrier islands has long been one of the favorite pastimes of coastal geomorphologists. Schwartz's collection of sig- nificant papers on this topic (Schwartz, 1973) will no doubt stimulate a further flurry of interest of which this paper is perhaps an example. How can the present generation claim to add anything new? Perhaps we do have a 870 significant perspective inaccessible to our predecessors; those of us interested in shelf geology at this point in history have, thanks to technological ad- vances, a far better collection of data from the shelf surface and subsurface — critical information if the uniformitarian principle is to be applied to the problem of barrier genesis during a transgression. I would like to present a paradigm for a barrier island that I think is a helpful simplification; a barrier island is a littoral sand body consisting of (1) a shore face maintained by the prevailing hydraulic regime, and (2) attached wash over fans whose surfaces are modified by aeolian and by biological (including human) activity. Barriers are in some respects analogous to sand waves, which consist of a current-graded slope attached to an avalanche slope. If this viewpoint is accepted, then the beach and shoreface response surfaces are clearly the critical zones; the existence, geometry, and behavior of other parts of the barrier are dependent on the behavior of the shoreface. The delicate balance of this system, and the rapidity with which the system adjusts to human interference, has recently been documented in an intri- guing case history by Dolan et al. (1973). Some workers have described the origin of modern barriers in terms of their response to the Late Holocene reduction in the rate of sea-level rise (Curray, 1969, p. JC-II-16). Others have stressed that most barriers have reached their present positions by migration from more seaward positions (Otvos, 1970a,b; Dillon, 1970; Pierce and Colquhoun, 1970; Stapor, 1973). However, few seemed to have taken the concept of barrier migration to its logical uniformitarian conclusion. To do so may require not only a change in the solution, but a change in the nature of the question being asked. If the conclusions of Stapor, Otvos, and Pierce and Colquhoun can be applied to the general case, then the question of how a barrier is initiated is in a sense, trivial. Firstly, if barriers have migrated into their present positions as a consequence of post Pleisto- cene sea-level rise, then the problem of genesis is transferred "out there" to some Late Holocene stillstand position on the shelf or perhaps to the shelf edge. Secondly, as Otvos and Stapor make very clear, migrating barriers are in a constant state of morphologic flux, with individual barrier islands and barrier spits continually undergoing expansion, contraction, fragmentation and integration. Thus a barrier system might have migrated to its present position from the shelf edge, but no two barrier islands or -spits will neces- sarily have existed during the same fraction of this time interval. Those aspects of barrier behavior are also characteristic of smaller, simpler bed forms; the nature of the maintaining process of, say, current ripples, is more significant than the nature of the initiating process, since the sand-water interface is metastable when the critical-velocity threshold is achieved, and any small perturbation will begin the constructive feedback between flow and substrate. Likewise, in such systems the individual forms, ripples, are ephemeral with respect to the lifetime of the ripple train. It therefore stands to reason that much of the evidence concerning the 871 genesis of the barriers of eastern North America must be sought on the Atlantic continental shelf surface (see Field and Duane, in preparation, for a related viewpoint). However, before examining this area, let us review ele- ments of shoreface dynamics relevant to barrier formation. SHOREFACE DYNAMICS The shoreface as an equilibrium surface The concept of the shoreface as a responsive interface between a substrate and a water column in a state of dynamic equilibrium with each other was first expressed in the geological literature in a highly deductive, qualitative way be Fenneman (1902) and Douglas Johnson (1919, p. 211). Johnson states: "The essential nature of the shore profile . . . [is] that nice balance between the amount of work required to remove the debris and the ability of the waves to do removal work which we call "equilibrium". 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 part in suspension. At every point the slope is precisely of the steepness required to enable the amount of wave energy there developed to dispose of the volume to sediment there in transit." Earlier, however, an Italian engineer, Cornaglia (1887) had observed a critical mechanism for the grading of the shoreface; depth and grain size on a point on the shoreface profile is determined by the Newtonian balance of forces on a sand grain; the value of the downslope component of gravitation- al force versus that of the net fluid force averaged over a wave cycle. Cornag- lia's ideas were not available to workers in the United States until the Coastal Engineering Research Center (Anonymous, 1950) published an edited trans- lation of an address by the Swedish coastal scientist, Munch-Peterson. U.S. workers developed the null-line concept, as it came to be called, into a complex numerical model for the shoreface equilibrium (summarized in Johnson and Eagleson, 1966). Further field work has shown that the null- line mechanism is only one of the processes controlling the shoreface equilib- rium (Murray, 1967; Cook and Gorsline, 1972). The downslope gravitational component is probably not sufficiently intense over most of the shoreface for the null-line mechanism to be an adequate model of sedimentation. However, the structure of the wave-velocity field itself may induce a diver- gence of shore-normal sediment transport about a null line (Wells, 1967). An adequate theory of shoreface sedimentation does not exist. However, no one who has watched a major storm strip away 50 m of beach, only to return the material in much the same configuration as before, can doubt that at least the upper shoreface exists in a state of equilibrium with the hydrau- lic regime. 872 Since the concept of equilibrium will be referred to again in this paper, it is important to define it carefully, and also stress the importance of the time scale of observation to the equilibrium concept. LeChatelier's principle is the most basic definition; an equilibrium system is one which when stressed, reacts in such a way as to relieve the stress. When "stressed" by a storm, a shoreface responds by widening and flattening through upper-shoreface ero- sion and offshore deposition. When fair weather returns, the lost sand (or its equivalent) also returns, to upper shoreface and beach storage. Systems whose characteristics oscillate about fixed values in this manner are systems in a state of homeostatic equilibrium. Whether or not a system can be described as an equilibrium system depends in part on its innate characteris- tics, but also on the time and space scales of observations. A coastal engineer is less likely than a geologist to find the term "equilibrium" applicable to the shoreface. The shoreface as he sees it is in a state of continual change and usually exhibits a slope intermediate between the extreme fair weather and storm pro-files. The geologist is concerned with the long-term near configura- tion of the shoreface as observed over periods of years or decades. From his point of view, the response time of the shoreface to hydraulic change is instantaneous when compared with such long-term processes as the migra- tion of the shoreline in response to sea-level rise. In these terms, the equilib- rium concept is a useful one. The shoreface and the sediment continuity equation Our thought about this dynamic control zone on the forward surface of barrier islands has been clarified by the insistence of C.A.M. King in succes- sive editions of her definitive text (most recently 1972) in considering the relative roles of onshore— offshore coastal transport, versus shore-parallel transport. This is implicitly the continuity concept of classical physics: to assess the coastal sediment budget, one considers a unit volume of a coastal water column (Fig. 1), and considers its imports and exports of water and sediment, through the faces of the box. These transport components are not merely convenient, but reflect the basic forcing mechanisms; the wave-driven longshore current and wind-driven coastal flow parallel to the beach on one hand, and upwelling and downwelling with associated coast-normal bottom transport, due to wave surge and wind and wave setup on the other hand. The continuity relationship allows us to assess the relationship of the longshore sand flux to coastal progradation and retrogration. The sediment continuity equation is: |£ + v- vc at d Xdt) Where V is a vector velocity, C is average concentration of sediment, p is sediment density, d is water depth, h is height from baseline, and V is the 873 OUTPUT TO DUNES, OVERWASH LITTORAL DRIFT INPUT LITTORAL DRIFT OUTPUT INPUT FROM SUBSTRATE EROSION COASTAL CURRENT OUTPUT OUTPUT TO SHELF FLOOR Fig. 1. The unit volume of coastal water column and its substrate. Arrows indicate components of sediment and water transport. Coast-parallel water movements are more intense than coast-normal movements, hence resultant transport vectors tend to make low angle with coast. divergence, an operator indicating net difference. The equation states that time rate of change of sediment concentration, plus the net difference be- tween sediment advected into the box and sediment advected out of the box, is proportional to the time rate of change of height of the water— sedi- ment interface (aggradation or erosion). In other words, the shoreline tends to aggrade when the sediment-discharge gradient across the box is negative, or if the concentration of sediment is increased with time; erosion requires the reverse conditions. May and Tanner (1973) have used a simplified version of the sediment continuity relationship in a model for the straightening of the coastline. With waves approaching normal to a coast composed of a series of headlands and bays, wave refraction will cause a greater wave-energy density (E) on the headland beaches (Fig. 2 A) with a corresponding reduction of this character- istic on the bay beaches (Fig. 2B). However, the angle between wave crests and the beach will be greatest on the side of the headland. These relation- ships are indicated in a very schematic fashion in Fig. 2. May and Tanner describe the longshore sediment transport rate as propor- tional to littoral power: PL = 0.5 EC sin 2 $ where PL is the longshore component of wave power, E is the wave-energy density (in Fig. 2A, proportional to the spacing of wave rays), C is wave- 874 VTION ® * * ■"**. ^ *^s« -Q^^ ■b- — ^c d2 dx © Fig. 2. Model for littoral sediment transport. A. Wave-refraction pattern with wave approach normal to coast. B. Resulting curves for energy density at the breaker. E (dimensions MT~2); longshore component of littoral wave power PL (dimensions MLT- 3); and the littoral-discharge gradient dq/dx (dimensions L2!^1). C. Advanced state of coastal evolution. After May and Tanner, 1973. phase velocity, and (3 is breaker angle. Thus the sand-transport rate is at a maximum between the point of maximum wave-energy density, a, and the point of maximum breaker angle, but much closer to the latter since on this gently embayed coastal model, the wave-energy density gradient is relatively flat (Fig. 2B). In this two-dimensional situation, where only the longshore component of sand transport is considered, the sediment continuity equation reduces to: dh_ _ _ dq dt dx where e is a dimensional constant related to sediment porosity. The term dq/dx is the rate of change of littoral drift discharge with distance along the 875 beach and varies as the derivative of PL (Fig. 2B). The equation states that where the sand-discharge rate decreases along an isobath of the shoreface (negative dq/dx) the shoreface at a point along the isobath must build up at a rate proportional to the drop in discharge across that point. In Fig. 2, dq/dx is positive from a to c, hence the shoreface off the headland must erode, and the headland shoreline retreat. Maximum erosion will occur at b, where the gradient is steepest. Between c and e it is negative, hence the bay shoreface must aggrade and the bay shoreline advance. Maximum deposition will occur at d. If waves advance normal to this shoreline over a sufficiently long period of time, a very peculiar shoreline must result. It will be essentially a straight line normal to wave approach running through point c. However, a sliver-like peninsula will protrude at a, where dq/dx = 0, and a narrow re-entrant will occur at e where dq/dx ~ 0 also. In nature, however, the direction of wave approach varies, day by day around the mean orientation, and positions a © F\n. 3. Variant of littoral-transport model, with a more deeply embayed coast and an oblique direction of wave advance. 876 through e shift accordingly, hence such anomalies do not occur at the mean positions of a and e, and in our ideal model, a smooth, straight shoreline must result. The coast is shown in an intermediate stage of evolution in Fig. 2C. It is interesting to consider a variant of this model, in which the degree of embayment of the coast is more extreme, and the mean direction of wave approach is at an angle to the coast (Fig. 3). Under these conditions, the decrease in wave-energy density from headland to bay head may no longer be nearly linear, but may be more nearly a step function. The littoral-dis- charge gradient is no longer primarily controlled by the breaker angle, but is strongly modified by the discontinuity in the wave-energy density. These relationships are very sensitive to the nature of the nearshore bathymetry, and become quite difficult to predict. They are presented in qualitative fashion in Fig. 3B. The main effect of the new configuration is the capture of the point of maximum littoral deposition, d, by the step in the wave-energy density distribution. Shoaling at this point may be sufficient to result in a discontinuity in the shoreface, in the form of a recurved spit which pro- grades into the bay, as the headland retreats and the bay-head beach ad- vances (Fig. 3C). The shoreface in profile The unit-volume model can be simplified to a coast-normal profile to assess the consequences of landward and upward translation of the equilib- rium profile in response to rising sea level. Bruun (1962; see also Schwartz, 1965, 1967, 1968) has shown that the geometry of such a translation re- quires shoreface erosion and a corresponding aggradation of the adjacent sea floor, assuming dq/dx = 0 parallel to the coast through the unit volume (Fig. 4A). Moody (1964, pp. 142—154) has described landward profile translation on the Delaware Coast, as induced by storm sedimentation. The barrier steepens over a period of years towards the ideal wave-graded profile, by both upper-shoreface aggradation and lower-shoreface erosion. During pe- riods of fair weather, the landward asymmetry of fair-weather wave surge traps sand on the upper shoreface, resulting in a trend towards increasing sand storage in the shoreface and beach prism. Mild storms erode the shore- face; but upper-shoreface deficits tend to be made up. This trend is abruptly terminated by a major storm which results in massive shoreface and beach erosion, a markedly flattened gradient, and landward profile translation on the order of tens of meters. The cycle then begins anew. Thus the longterm behavior of the shoreface during a marine transgres- sion, whether retrogradation or progradation, depends on the balance be- tween fair-weather accumulation and storm erosion over the observational interval. If erosion predominates, the mean trajectory of the shoreface must be landwards and upwards (Fig. 4 A). The barrier superstructure may retreat in cyclic, tank-tread fashion, by a process of storm washover of sand, its burial, and re-emergence at the upper shoreface. Debris from the eroding 877 B AGGtADING BAI Will CAPIUIE SHQIEUNE nEISTOCENE SAND .„. Fig. 4. A. Erosional shoreface retreat, after Bruun, 1962. On an unconsolidated coast, a rise of sea level results in shoreface erosion, and concomitant aggradation of the inner- shelf floor as long as coastwise sand exports equal or exceed coastwise sand imports. B. Depositional advance of the shoreface, in response to an excess of coastwise sand imports over exports. After Curray et al., 1969. lower shoreface will accumulate beneath the rising lower limb of the profile on the adjacent sea floor. The transfer process is poorly understood. Transfer from the shoreface to the adjacent shelf floor probably occurs during storms when the coastal boundary of the storm flow field takes the form of a down-welling jet flowing along the contours of the shoreface, in response to coastal-wind setup (Swift et al., in press). However, if sedimentation pre- dominates over erosion on the shoreface, the profile will translate seaward and upward, usually by means of the cyclic outward growth of beach ridges (Curray et al., 1969; Fig. 4B this paper). A word on the factors controlling the steepness and curvature of the 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 the shoreface profiles. Shorefaces built on shingle may attain 30° slopes near the beach; shorefaces of sa.nd are rarely more than 10° at their steepest, while shorefaces on muddy coasts are so flat as to be virtually indistinguish- able 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, while inner shelves with a lower influx of sediment and a higher wave-energy flux per unit area of the bottom are steeper (Wright and Cole- man, 1972). Due to the complex interdependence of the process variables, cause ^and effect are difficult 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 878 10 coarser grain size results in a higher effective angle of repose. Or starvation of the profile will result in steepening, with a resultant higher energy flux and coarsening of its surface (Langford-Smith and Thorn, 1969). The relationship between the rate of sea-level displacement and the shape of the profile requires some thought. A number of workers have assumed that rapidly 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 concept and is tantamount to stating that only chemical reactions that have gone to completion are equilibrium reactions. It is important to distinguish clearly between the concept of coastal matu- rity on one hand, and the concept of coastal equilibrium on the other. Davis (in Johnson, 1919) has assembled a spectrum of coastal types which suggests that the coastal profile passes through stages of "youth, maturity, and old age" in which the profile becomes 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 dependence of the process variables of an equilibrium system. Some of these stages will occur as tran- sient 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 (Fig. 4A), until the rate of one or another variable changes again. Only in such cases of relatively rapid tectonism may hysteresis, or lagged response occur. Strictly speaking, the term "disequilibrium" should be ap- plied only to such cases. Slower changes in a process variable will allow continuous and compensating adjustment of profile, 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 displacement in the equi- librium 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 the resulting equilibrium profile (see for instance 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 energy flux of the regime 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 configu- 879 11 ration 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 mean, or even the maximum slope of the steepest profile per- mitted by these variables. Under such circumstances, the more rapidly tran- siting shorelines, since these do the least work, have the least-modified and hence steepest (most immature) profiles, while the most slowly moving shorelines are the most modified and hence flattest profiles. Variation in curvature of the shoreface, plus variation in the angle of translation of the shoreface lead to a variety of possible modes of transla- tion. Fig. 5 illustrates some possible scenarios. In Fig. 5A, the coast is under- going Bruun erosional retreat, as in Fig. 4A. The trajectory of shoreface- EROSIONAL TRANSGRESSION (PROFILE INVARIANT) B DEPOSITIONAL TRANSGRESSION (PROFILE INVARIANT) STILLSTANO (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 Fig. 5. Modes of shoreface translation as a function of ( 1 ) direction of profile translation and (2) change in profile curvature. Envelopes of erosion and aggradation are shown. Terms from Curray, 1964. 880 12 profile translation is parallel to the surface being transgressed. River sand is trapped in the throats of estuaries. Littoral-discharge gradients are positive along most parts of the coast. Most modern retreating coasts fall into this category. In Fig. 5B, the rate of sea-level rise has decreased. Estuary mouths have adjusted to their tidal discharge, and are capable of bypassing sand. The littoral-drift gradient is still positive but is more gentle, hence shoreface erosion is less severe. The trajectory of shoreface-profile translation diverges from the surface being transgressed, and the back-barrier zone becomes a widening expanse of marshes or lagoons. The transgressing shoreface no longer destroys all back-barrier deposits as in Fig. 4A, as these deposits are now much thicker and the shoreface merely translates through their upper portion. Basal back-barrier deposits may be found beneath a thick layer of shelf- floor sand, that has accumulated seaward of the shoreface. As the widening belt of marshes and lagoons taps more of the river sediment, less is fed to the outer beaches, the trajectory of profile translation ceases to di- verge from the surface being transgressed, and the oceanic shoreline and inner lagoonal shoreline transgress at more or less the same rate. The Lumbee Group of the Cretaceous coastal plain of North and South Carolina was probably formed in this manner (Swift and Heron, 1969). In Fig. 5C, sea-level rise has decreased. There is sufficient input of sand into the littoral-drift system to cause the upper shoreface to prograde, steep- ening the profile. The Late Holocene reduction in the rate of sea-level rise of 4,000—7,000 years ago may have had this effect on portions of the world's coasts. In Fig. 5D, a markedly greater injection of river sand into the coastal drift nourishes the entire shoreface. The profile translates upwards and outwards. The Costa de Nayarit (Curray et al., 1969) is a classic example of such a coast. In Fig. 5E, sea level is falling. The Bruun process is reserved and the shoreface profile translates seaward down the sloping shelf surface. The upper shoreface progrades. The lower shoreface and inner-shelf floor subject to intensified wave action by a shoaling water column, erode. In Fig. 5F, sea level is falling, but there is a sufficient input of mobile fine sediment that both the shoreface and inner-shelf sectors aggrade. The shoreface in plan view Coastal straightening. A characteristic of the equilibrium shoreface surface that as much as any mechanism is the basic "cause" of barrier islands 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 boundary, since the wave and current field to which it responds does not generally have one. The initial conditions during a period of coastal sedimentation may however include such discon- 881 13 tinuities; as in the case of a coast of appreciable relief (bay— headland coast) beginning transgression. On such a coast shoreface surfaces will tend to be incised into the seaward margins of promontories exposed to oceanic waves, and will propagate by constructional means in the downdrift direction as long as material is avail- able with which to build, and a foundation is available to build on. The basic mechanism is presented in schematic fashion in Fig. 3. Where the shoreline curves landward into a bay, the longshore component of littoral wave power decreases, and dq/dx is negative. The shoreface at that point must aggrade until dq/dx approaches zero at that point, and the zone of negative dq/dx has moved downdrift. We give the lateral propagation of the shoreface into coastal voids the descriptive term "spit building by coast- wise 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 valleys after a prolonged stillstand during which processes of coastal straightening by headland trunca- tion and spit-building have gone to completion. Bay-head beaches have pro- graded to the position of adjacent headlands, or have been sealed off by spits and filled in by marshes. Fig. 2 is a better model from which to deduce the consequences for this set of initial conditions. As this coastline submerges, the water, seeking its own level, would invade valleys more rapidly than headlands could be cut back. The oceanic shoreline, however, would not follow, for if it should start to bulge into the flooding stream valleys, the bulge would become a zone of negative dq/dx; hence the rate of sedimenta- tion 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 would be restored. The beach and dune, nour- ished by littoral drift, would be able to grow upwards at the same rate as sea-level rise, but the swale behind the dune would not. Shallow water bodies would extend into these swales from the sides of estuaries. Thus a straight or nearly straight oceanic shoreline must detach from an irregular inner shore- line, 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 detail by Hoyt (1967) and Hails and Hoyt (1968). Coastal rotation. In addition to straightening the coast by infilling or inhibit- ing bays, the trend of the coast towards equilibrium in plan view may involve rotation of coastal segments into preferred orientations with respect to pre- vailing waves. In order for such rotation to transpire the coastal segment must be a more or less closed system with respect to littoral drift. The simplest case is a groin field, whose groins extend seaward through the zone of maximum wave action, so that drift between groins tends to accumulate against the downdrift groin while the upper end of the compartment, starved 882 14 by the updrift groin, tends to retreat. Pocket beaches, between rocky head- lands, behave in a similar manner, tending to rotate into a normal orientation with respect to prevailing wave orthogonals. If the initial angle between the prevailing wave direction and the coastline is less than 50°, then an equilibrium alignment at less than normal orien- tation may prevail. Davies (1973; p. 173) has noted that many beaches tend to build parallel to the line of maximum drift, which lies in the region of about 40°— 50° to the direction of wave approach. A homeostatic equilib- rium tends to prevail for such beaches. Any increase in beach angle causes a reduction in the beach-parallel component of littoral wave power while a decrease in beach angle causes a reduction in wave-energy density and there- fore a reduction in total littoral wave power. Davies refers to such an oblique equilibrium alignment as a drift alignment. Large-scale coastal segments gen- erally cannot rotate very far. As the downdrift sector finds itself in progres- sively deeper water, there is more surface area of shore face to nourish, and more and more sand is required for each angular unit of rotation. The extent to which rotation transpires probably also depends on the rate of sea-level rise or fall; with a slowly rising sea-level, for instance, equilibrium might be reached at a beach-wave angle closer to the climax configuration, than in the case of a more rapid rise of sea level. Coastal compartments not captured by the drift alignment will proceed to the swash alignment (Davies, 1958) or alignment normal to wave ortho- gonals. The adjustment is a feedback process, since as orientation of the shoreface shifts, the pattern of refraction of surface wave trains which in- duces the shift must also change. Straight swash coasts are inherently unstable, so that slight initial varia- tions of substrate configuration tend to initiate a constructive feedback, until the coast stabilizes in a series of wave-like forms, such as the cuspate forelands of the Carolina coast, or as an alternation of swash and drift segments, such as the offset coasts of successive zetaform beaches (Davies, 1973, p. 136). Shoreface maintenance and barrier formation Barrier island formation by upward growth. We have arrived at two time- honored mechanisms for barrier formation by a consideration of shoreface dynamics. Before considering the relative roles of coastwise spit progradation and mainland-beach detachment, we should consider a third basic mecha- nism, namely De Beaumont's (1845; in Schwartz, 1973) initiation and up- ward growth of offshore bars. In a sense, this is a special case of the main- land-beach detachment hypothesis in that it stresses onshore— offshore sedi- ment transport, versus coastwise sediment transport. There are two basic problems associated with this mode of barrier formation that have been faced with varying degrees of success by its proponents. The first concerns initial conditions. Johnson (1919) solved this by requir- 883 15 ing a previous withdrawal of the sea, such that at time zero, a metastable condition prevails, in that the sea-floor slope is gentler than that required by the equilibrium profile. However, as previously noted, the response time of the shore face is instantaneous with respect to even glacioeustatic sea-level fluctuations. Prograding shorelines continue to maintain shorefaces as they prograde, generally by the cyclic formation of closely spaced beach ridges (Fig. 4B). The result of stillstand passing into regression may be the expan- sion of the stillstand barrier into a beach-ridge plain tens of kilometers wide (Curray et al., 1969). As many authors have noted, a beach ridge is not the same thing as a barrier. Beach ridges are generally much narrower than barriers (on the order of 50 m versus perhaps 500 m for a barrier; Curray, 1969), although the largest beach ridges are as large as narrow barriers. Beach ridge "lagoons" are usually marshy swales no wider than the ridges them- selves (Curray, 1969). Barrier surfaces may consist of beach-ridge complexes, but these sequences have been tacked on to the forward face of a much larger feature — the barrier foundation. The subaerial surface of the Atlantic coastal plain is webbed with such beach-ridge plains between successive high stand barrier systems (Oaks and Coch, 1963; Colquhoun, 1969), demonstrat- ing the general inadequacy of a prograding shoreline to build an offshore barrier and wide lagoon according to Johnson's method. In genetic terms, beach ridges (or if the influx of fine sediment is sufficient, cheniers) appear to be the response of prograding shorelines to shoreface processes, while barriers are the response of stillstand or retrograding shorelines. The second and closely related problem associated with the upbuilding-bar hypothesis is the inadequacy of known mechanisms of swash-bar formation for building barriers of appropriate scale and distance from shore. Break- point bars are small-scale features which form beneath the breaker, on main- land beaches and on barriers. On retrograding coasts, they tend to form as storms wane, then migrate onshore as swash bars (intertidal bars) during the ensuing fair-weather period, to weld to the beach (King, 1972). On prograd- ing coasts, break-point bars are stabilized by accretion on their seaward faces, become swash bars, then during periods of neap tide, may build high enough to become the new shoreline and initiate dunes (Curray, 1969). The resulting feature, however, is a beach ridge, not a barrier island. The only other available mechanism appears to be intertidal swash plat- forms associated with the ebb-tidal deltas of estuary mouth or tidal-inlet shoals (Oertel, 1972). These are commonly shore-normal, levee-like struc- tures, but the "new barrier island" figured by Shepard (1973, fig. 7b) from the Gulf Coast of Florida may be of this origin since it closes off a tidal inlet. This small feature appears to represent a new superstructure element formed on an existing barrier platform. Otvos (1970a) advocates the emergence of offshore bars as a mechanism for generating barrier islands. His "subsurface evidence", however, is ambigu- ous, in that bore-hole spacing is not shown, and criteria for distinguishing barrier subenvironments are not discussed. In any case, his cross-sections 884 16 cannot distinguish between the offshore initiation and emergence of barrier islands, and their lateral migration in the direction of drift, which he so clearly describes and figures (Otvos, 1970a, fig. 3). Hoyt's (1967) conclu- sions appear to be still valid: ". . . minor barriers may form from bars, but the observed cases have been small, short-lived features formed close to the shoreline, and can hardly be compared with the major features under dis- cussion." Coastwise spit progradation vs. 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, 1968, 1970; Otvos, 1970a, b). The problem can be fully answered only by careful study of the field evidence, and as noted by several authors (Otvos, 1970a,b; Pierce and Colquhoun, 1970) the evidence has frequently been destroyed by landward migration of the barriers. However, it is possible, in the time-honored deductive fashion of coastal morphologists, to consider the conditions most favorable to these two modes of barrier formation. Spits are certainly characteristic of coasts of high relief undergoing rapid transgression as described above (see papers in Schwartz, 1973). It seems probable that under such conditions mainland-beach detachment would be severely inhib- ited. Even allowing for ideal initial conditions with a classic coast of old age (Fig. 6), where alluvial fans are flushed with truncated headlands, detached mainland beaches would have a limited capability for survival. With signifi- A. STILLSTAND B. BEACH DETACHMENT C. CYCLIC SPIT PROGRADATION Fig. 6. Evolution of the shore face 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. Johnson's (1919) stages of coastal maturity — portrayed in reverse sequence! 885 17 cant relief, the submarine valley floors adjacent to retreating headlands must lie in increasingly deeper water after the onset of transgression. As the sub- marine surface area of the barrier requiring nourishment increases, the ca- pacity of littoral drift to nourish it may eventually be exceeded. As this point is approached, storm washover 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 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. A STIUSTAND B BEACH DETACHMENT C BARRIER RETREAT Fig. 7. Evolution of the shoreface as a low coast passes from stillstand to transgression. A mature coastal configuration passes via mainland-beach detachment into a steady-state regime of barrier retreat. 886 18 Finally the survival or 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 transgressed barriers on the shelf floor appear to be overstepped, rock-tied spits (McMaster and Garrison, 1967; Nevesskii, 1969). 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. 7). The depth of water in which detached bay-mouth barriers would be built would be less, because the relief would be less. Littoral drift would be generally adequate to nourish the lesser submarine surface area of the barrier face. The upper, erosional zone of the shoreface (Fig. 4 A) would be more likely to extend down into the pre-recent substrate (Fig. 8); 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. Rivermouths, initially deltaic, would flood as estuaries, while lagoons would creep behind the beaches toward the headlands on either side. Coastal discontinuities ~rr7T7 ■ LAGOONAL p™l BA""1^ i 1 INNER nFPn- OlDfl SUSSIIAK HOIOCENE SAND SHUT Fig. 5. A model for cellular storm flow. Note extreme vertical exaggera- tion; cells would have height to width ratios of at least 1:15. 920 ^Reprinted from: Marine Geology 18, 105-134. 62 TIDAL SAND RIDGES AND SHOAL-RETREAT MASSIFS DONALD J. P. SWIFT Atlantic Oceanographic and Meteorological Laboratories, Miami, Fla. (U.S.A.) (Received June 10, 1974; accepted November 1, 1974) ABSTRACT Swift, D. J. P., 1974. Tidal sand ridges and shoal-retreat massifs. Mar. Geol., 18:105—134. In 1963, Off defined a bedform type which he described as rhythmic linear sand bodies caused by tidal currents. He figured twelve examples from around the world. Since then, the morphology and dynamics of sand transport in one of these areas, the tidal shelf seas around Great Britain, have undergone intensive study. The tidal sand ridges emerge as anomalies, in that they do not fit into the sequence of morphologic provinces which characterize the major sediment transport paths. It is suggested here that the ridge fields are analogous to the shoal-retreat massifs of the Middle Atlantic Bight in that they have been inherited from a nearshore regime during the course of the Holocene transgression. Shoal-retreat massifs are low, broad, shelf-transverse sand bodies which mark the retreat paths of coastal depocenters associated with littoral drift convergences. Two main types of shoal-retreat massifs in the Middle Atlantic Bight are: (1) estuarine shoal-retreat massifs; and (2) cape shoal-retreat massifs. Two similar classes of shoal-retreat massifs may develop in tidal shelf seas, but the mechanism is somewhat different. Class-1 tidal massifs are tidal ridge fields whose ridges were hydraulically packaged in an estuarine environment. If, upon transgression, they find themselves in a broad tidal bight which continues to funnel tidal flow, the ridges may survive for long distances out into the bight. The ridge fields of the Southern Bight of the North Sea may have undergone such an evolution. Class-2 tidal massifs occur off promontories in tidal seas that are swept by the edge waves generated by amphidromic tidal systems. Here the debris of shoreface erosion tends to be stored as shoreface-connected, tide-maintained ridges. Such ridges are also pre-adapted to survive with modification for long distances out on the associated shelf, as the water column deepens during a marine transgression. INTRODUCTION Sediment genesis and dispersal on a storm-dominated shelf The extremely detailed bathymetric maps available for the Atlantic shelf of North America, together with supplementary information on surficial sediments and shallow structure, have resulted in a clear picture of the mechanisms by which the Holocene surficial sand sheet was formed (Swift et al., 1972, 1974; Swift and Sears, 1974). The sand sheet is the blanket of debris produced by the erosional translation of the more steeply inclined 921 106 shoreface scarp into the flat coastal plain terrace during the Holocene trans- gression. The surficial sand is thus relict in the sense that it is derived from the underlying substrate. However, size frequency distributions of its samples generally reflect Holocene depositional mechanisms rather than earlier ones (Stubblefield et al., 1974). The term autochthonous is therefore preferred as a descriptor for the mode of sedimentation, since the term relict implies inherited rather than equilibrium textures (Curray, 1965; Swift et al., 1971). Autochthonous sedimentation, wherein the shelf cannibalizes its own substrate, may be contrasted with allochthonous sedimentation where the bulk of the surficial sediment is of recent fluvial origin, as in the case of the Niger shelf (Allen, 1965). Holocene fluvial and lagoonal sediments and Pleistocene sediments pass through an "energy fence" (Allen, 1965) of landward asymmetrical wave surge on the retrograding shoreface by one of two modes. In shoreface bypassing, sediments eroded from the retreating shoreface are shifted down- current (generally southward on the Atlantic shelf) by littoral drift, and by inner shelf storm currents during periods of strong onshore winds, and tends to be deposited just seaward of the shoreface, to form the leading edge of the shelf surficial sand sheet (Fig.l). Material retained for longer periods of time in the coastal transport system is ultimately deposited at littoral drift convergences, a mode termed downdrift bypassing. Sand shoals form at such convergences and translate landward with the retrograding shoreline. The low, shelf-transverse sand ridges which mark their retreat paths have been termed shoal-retreat massifs (massif in the sense of a topographic high composed of smaller scale highs; Swift et al., 1972; Fig.2 this paper). The surficial sand sheet of the Atlantic continental shelf exhibits a variety of morphologic— stratigraphic patterns, reflecting differing coastal regimes, and varying patterns of shoreface and downdrift bypassing (Swift and Sears, 1974). In the Middle Atlantic Bight with its deeply incised, widely spaced estuaries and moderate wave regime, shoal-retreat massifs that are the retreat paths of estuary-mouth shoals alternate with the thinner shoreface retreat blankets of the coastal compartments. Both are overprinted with a ridge and swale topography (Swift et al., 1974), a response of the surficial sand sheet to the storm flow field (Fig. 3). Ridges may form at the foot of the retreating shoreface and be detached as the transgression continues, or may be impressed directly on shoal-retreat massifs. Off North and South Carolina, the more intense wave regime has created a series of cuspate forelands. As a consequence of the littoral drift conver- gences associated with the foreland apices, cape shoal-retreat massifs alternate with shoreface retreat blankets (Swift et al., 1972). To the south as the scale and spacing of forelands decrease, the massifs tend to overlap, and the resulting sand sheet is more accurately described as a cape shoal-retreat blanket. This sector has likewise been overprinted with a ridge and swale, topography (Fig.4). The Atlantic shelf of North America is a storm-dominated shelf. It does 922 107 VECTOR RESOLUTION OF PROFILE TRANSLATION WASHOVER CYCLE OF BARRIER SANDS Fig.l. Shoreface bypassing. Landward and upward translation of the shoreface profile results in shoreface erosion and concomitant deposition of debris as the leading edge of the transgressive shelf sand sheet (Swift et al., 1974.) not experience the sustained wave intensity of coasts in the lee of the prevailing westerlies such as the California or Irish coasts. Except in the Gulf of Main, in the Georgia Bight, and in the mouths of estuaries, tidal currents are not sufficiently amplified to significantly modify bottom topography. Sediment genesis and dispersal on a tidal shelf As noted by Stride (1963), Belderson and Stride (1966), Belderson et al. (1971) and Kenyon and Stride (1970) the fabric of the surficial sand sheet of tide-dominated shelves is rather different. Sedimentation in their study area, the shelf around the British Isles, may still be inferred to be autochtho- nous at least with respect to sand, since estuary mouths are sinks rather than sources with respect to marine sand, even in the case of estuarine 923 108 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 Fig. 2. Downdrift bypassing. Littoral drift from the New Jersey coastal compartment to the north has been delivered to the mouth of Delaware Bay during its retreat across the shelf in response to postglacial sea-level rise. The retreat path of this littoral drift depositional center forms a low, shelf-transverse ridge, or shoal-retreat massif. (From Swift, 1973.) 109 03 O -2 « bs 03 C 0/ I/i is 0 > * a (A -C h BO Si -O 0 a U 0 w -j c Oi 09 03 < 'SI -a c 03 p, 0) an -o 111 -C h +J t+i cw 0 0 a u 0 03 E id -o c y u 0£ a 0 0) > 0 o J3 p. T| s-i c 0 03 s c<- CO cm 03 fa E 925 110 E a QJ >. O 2 a o J3 T3 C cs O u 2 t- as GO £ 2 a >> r3, a O es T 0 rr 0 926 Ill mouths of the Rhine distributaries (Oomkens and Terwindt, 1960). Shore- face bypassing has apparently been the dominant mechanism of supply (Belderson and Stride, 1966). However, the more intense hydraulic climate of the tide-swept shelf of Great Britain appears to be capable of erasing most traces of earlier coastal environments from the sea floor. Instead the British workers report a complex systematic pattern on sea floor sediment dispersal (Fig. 5) in which sand streams move through a succession of textural and morphologic facies down the gradient of mid-tide velocities (Fig. 6 ). Dominant bedforms include erosional furrows (Stride, 1973), sand ribbons (Kenyon, 1970b), sand waves (Stride, 1970), and patchy sand sheets (Kenyon, 1970b), in order of decreasing mid-tide velocities. TIDE-MAINTAINED SAND RIDGES Categories of tidal ridges A rather striking large-scale bedform that does not appear to be accounted for in their scheme is the current-parallel tidal sand ridge. Off (1963) has shown that these features have a world-wide distribution on shelf sectors in which the tide has undergone resonant amplification (Off, 1963; Fig. 7 this paper). Most of the tidal ridge fields figured by Off tend to fall into two cate- gories. A first category consists of ridges in embayments, or the mouths of embayments. This class of ridge is generally parallel to the axis of the embay - ment and normal, or oblique, to the regional trend of the shoreline. A second class of ridges occurs off capes and promontories. The ridges tend to be coast-parallel, but the ridge field as a whole tends to be elongated normal to shore. A third class, not figured, tends to occur on shelf edges where they intercept oceanic tidal streams. The first two classes occur together in the Southern Bight of the North Sea (Fig.7D; the Norfolk ridge field trends seaward from the Norfolk Promontory, while the Dutch, Flemish, and Thames ridge fields occupy the funnel-shaped Southern Bight of the North Sea. Sand storage and stability of tidal ridges These ridges are anomalous with respect to the scheme of tidal transport paths of Stride, Belderson, and Kenyon in two respects. Firstly, they appear to occupy no preferred location with respect to the facies sequence of the transport paths. Secondly, unlike the bedforms of the transport paths, they appear to be largely closed systems. Tidal transport paths, as described by British workers, start from "bedload partings," or zones of divergence of bottom sediment transport, and end in bedload convergences or over the shelf edge. The proximal sand ribbons and distal sand waves of these trans- port paths are in a sense hydraulically maintained sediment traps, in which grains undergo prolonged residence in their intermittent journey down the 927 112 Fig.5. Pattern of sediment transport around the British Isles. (From Kenyon and Stride, 1970.1 113 ZONE I BEDROCK & GRAVEL ZONE II SAND RIBBONS ZONE II SAND ZONE IV SMOOTH ZONE V SAND WAVES SAND PATCHES DECREASING MID-TIDE SURFACE VELOCITY; BOTTOM GRAIN SIZE Fig. 6. Succession of morphologic provinces along a tidal transport path, based on Belderson et al., 1971. transport path. Sand ribbons are hypothesized to be zones of temporary deposition beneath bottom convergences due to secondary flow cells in the tidal stream (Allen, 1970). Sand waves are known to advance tank-tread fashion, due to deposition on the slip face, burial, erosion on the upcurrent side and redeposition on the slipface (Allen, 1968a). Tidal sand ridges appear to be hydraulically maintained sand traps of a higher order of efficiency. As noted by Houbolt (1968; Fig. 8 this paper), and by Caston (1970), Caston and Stride (1970), the tidal sand ridges of the North Sea tend to be partitions separating tidal currents with a dominantly flood discharge on one side from tidal currents with a dominantly ebb discharge on the other; the ridges are as a consequence sand circulation cells which comprise closed or nearly closed loops in the sediment transport system. This characteristic is clearly indicated by bedform patterns. Sandwaves tend to climb obliquely up both sides of the ridge, with their orthogonals meeting head on along the crest; or if the ridge is asymmetrical in cross-section (Fig. 8 above) they move obliquely up the gentle side only; however, grain orientation on the steeper side suggests that sand here moves with the opposite sense. Caston (1970) notes that the stability of tidal sand ridges extends to larger-scale ridge patterns. Ridges may migrate laterally due to stronger residual tidal flows on one side than on the other. However, different segments tend to migrate at different rates, so that a side-slipping ridge will become offset, then sigmoid as two internal channels develop, and finally split into three parallel ridges (Fig.8A— D). Ridge evolution is thus cyclic rather than linear. Rates of side-slipping and ridge evolution are slow, however, and ridge stability is sufficient that the course of 17th-century 929 114 -a c o -a 3 O £ 0) 2 to -4J » ' «W • O £ c/i O a — 03 w 930 115 INFERRED SEDIMENT TRANSPORT <3> SAND WAVE 0 GRAIN ORIENTATION ABC D — BANK CRESTLINE ff DOMINANT CURRENT; SAND STREAM // MAJOR, MINOR BANK MOVEMENTS Fig. 8. Morphologic patterns in tidal ridge fields. Above: structure of Well Bank, Norfolk ridge field, from Houbolt, 1968. Below: cyclic evolution of Norfolk ridges, from Caston, 1970. Dutch naval battles may be traced on modern maps (Brouwer, 1964). The galleons and caravels had to avoid the same ridges that atomic submarines manoeuver about today. Such stability is probably characteristic of all large- scale current-parallel bedforms. Wilson (1973) has calculated that the great wind-parallel dunes of the sand-sea deserts may have taken ten thousand years to attain their present size. Their rates of migration are so slow that oases between ridges can migrate with the ridges. Ridges that have required much of the duration of the Holocene to develop must have been affected by the secular changes in Holocene climate (for desert ridges) and Holocene sealevel (for tidal shelf ridges). It seems reasonable, therefore, to examine tidal ridge fields to see if this is a class of tidal shelf topography whose formation is contingent on shoreline migration. The most landward occurrence of tidal sand ridges is within tidal estuaries. A logical course of study is therefore a consideration of estuarine ridges and their associated flow fields, and comparison of these features with the ridges and flow fields of associated shelf sectors, to see if initially estuarine ridges may have survived transgression to become shelf ridges during the Holocene transgression. 931 116 SHELF RIDGES FROM ESTUARINE RIDGES Estuarine flow The cross-sectional area of river channels is a power function of river discharge (Leopold et al., 1964). Where rivers enter the sea, a salt wedge intrudes beneath the fluvial jet, whose discharge is amplified by entrainment of the underlying, more salty water, resulting in two-layer (estuarine) circulation. Most rivers enter tidal seas. River mouth discharge is increased by the volume of the reversing tidal flow which propagates for some distance upstream. The estuarine (two-layer) circulation that is residual to the tidal cycle is itself increased by the efficient tidal mixing. Thus a river mouth whose channel is in equilibrium with total discharge must expand rapidly through the tide-influenced zone toward the sea. At the river mouth proper, a variety of processes conspire to construct an arcuate, seaward-convex sand shoal (Fig. 9). The most fundamental factor is the hydrodynamic continuity relationship; expansion of the fluvial jet results in rapid deceleration and loss of capacity, and sediment is deposited in the form of a shoal. Estuarine circulation also plays a role; river mouth morphology and the circulation interact, so that the crest of the shoal becomes the leading edge of the salt wedge during flood stage; or if the tidal CZ=Z> V fe\ mm WAVE TRANSPORT ~4 TIDAL TRANSPORT ::: INTEtTIDAL < 4 METERS -+ SUSPENSIVE TIDAl TRANSPORT Fig. 9. Sediment transport patterns on an estuary-mouth shoal. (After Oertel, 1972.) 932 117 component of river mouth discharge is very large, the spring ebb-tide terminus of the salt wedge; or both (Moore, 1970). The crest of the shoal thus becomes a bottom current convergence during periods of maximum sediment trans- port, and hence a reservoir for sand storage. A third major process maintain- ing the river mouth shoal is littoral drift, which is diverted seaward along the shoal crest and also serves to nourish it. Sediment storage in river mouth shoals is mediated by the behavior of the tidal wave as it passes over the shoal crest. The tide within the estuary is retarded by friction and is out of phase with the shelf tide; it continues to ebb after the shelf tide has already begun to flood. The two water masses tend to interpenetrate, with the main ebb jet passing out over the center of the shoal, and the oceanic tide flooding on either side of it. The response of the shoal surface to this repeated flow pattern is an interdigitation of ebb and flood dominated channels, separated by a discontinuous, zig-zag system of sand banks (Ludwick, 1974b). Since opposite sides of these banks expe- rience flow residual to the tidal cycle in opposite directions, each bank becomes a sand circulation cell or closed loop in the sand transport system (Fig.9). As such it is a hydrodynamic trap for sand supplied by littoral drift and tends to build up into the intertidal zone as a wave-dominated swash platform (Oertel, 1972). Despite the relatively rapid postglacial rise in sea level, some river mouths have been able to maintain equilibrium channels, as deltas (prograding river mouths) or as equilibrium estuaries (slowly retrograding river mouths; see Fig.lOA, B). Most, however, have not. Disequilibrium estuaries have resulted whenever aggradation of the estuary floor (in mm yr"1 ) has been less than the rate of sea-level rise, so that before any given segment of channel could close down to the lower limiting cross-sectional area required for fluvial and tidal discharge the main shoreline had passed it by. Such "drowned" or disequilibrium estuaries are generally more nearly funnel-shaped, rather than trumpet-shaped, as are the equilibrium forms. As a consequence of their higher ratio of salt water to fluvial discharge, their river-mouth shoals are retracted into the throat of the estuary and the interpenetration of ebb and flood channels becomes marked (Fig.lOC). With a yet further increase of tidal over fluvial discharge such a coastal indentation may no longer be appropriately called an estuary, but simply a bay. Large bays experiencing high tidal ranges may build a tide flat — tidal channel complex at their heads as a consequence of net landward sediment transport by the shoaling tidal wave (for instance, the German bight of the North Sea — Reineck, 1970). These deposits are the functional equivalent of the tide-molded deposits of a disequilibrium estuary. Transgressiue evolution of an estuary Ridge initiation in the upper estuary. Many of the tidal ridge systems figured by Off appear to be active wide-mouth estuary or tidal-gulf morphologies. Other examples are probably explicable in terms of landward translation 933 118 I CONSTRUCTIONAL CHANNEL RIVER DOMINATED FLOW m*? i Itp B. CONSTRUCTIONAL CHANNEL TIDE DOMINATED FLOW \ ;:>>s C. PARTLY EROSIONAL CHANNEL ^•v: TIDE DOMINATED FLOW INTERTIDAL SHOAL SUBTIDAL SHOAL FLOW DOMINANCE OF CHANNEL Fig. 10. Varieties of river-mouth morphology. of nearshore marine environments in response to the postglacial rise in sea level, with concomitant restructuring of estuarine ridge fields as shelf ridge fields. In order to understand such a transition, it is necessary to trace the evolution of a tidal ridge in a wide-mouth (disequilibrium) estuary during the course of a transgression, from its first emergence at the head of the estuary, through the estuary mouth, on to the adjacent shelf. Kraft et al. (1974) have attempted to trace the transgressive history of the mouth of Delaware Bay by equating 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. 11). Here ridges first appear as sub- aqueous 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. The origin of these tidal levees (as well as the origin of river levees) requires some thought. It seems probable that channeled flow, both in bankful stage (as in a river) and with overbank flow (as in a flooded river or tidal estuary), experien- ces a secondary circulation in the form of a double helical flow cell, with a common descending limb in mid-channel (Jeffreys, 1929; Leopold et al., 1964, p. 283; Wilson, 1973); at bends the pattern is deformed into a single helix. The mechanism is poorly understood. In flumes with rectangular cross- sections a somewhat more complex secondary flow pattern has been attri- buted to the unequal distribution of Reynolds (turbulent) stress components 119 o -A o £ CO C o -a G i' 03 15 > '3 a* .2 .5 ■g « -3d .i C J3 03 0 w c 0 03 c n 4J £ oj u 0 3 O > fe a> U cm a> r^ '^-, 75'I55 Fig.14. Ebb-directed (above) and flood-directed (below) sediment transport at the bed, north side of Chesapeake Bay mouth. Stream-lines are for the sediment transport vector, rouioo! see Ludwick, 1974 for computations. Depths in meters. Vertically ruled areas (shoaler than 5.5 m) are the shoals marginal to an interdigitating ebb— flood channel system. (From Ludwick, 1974.) 939 124 nels (see Outer Thames estuary in Houbolt, 1968, end paper map). On such openwork shelf ridge systems, ebb and flood sinuses still occur, but as minor zig-zags or bifurcations of ridges. Channels between main ridges are themselves divided into ebb dominant and flood dominant sides (Caston and Stride, 1970; Fig. 15 this paper). All channels have the same sense of shear. Thus the ridges are sand circulation cells, but the sense of circulation is the same from ridge to ridge, instead of alternating between clockwise and counter clockwise as on the estuary-mouth shoal. Huthnance (1973) attributes this open-shelf flow pattern 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 up-current flank as the water column shoals, and influence of friction becomes 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 down-current flank, the restoration of the ridge-parallel flow to ambient velocity is similarly lagged. When the tide changes, up-current and down-current flanks reverse roles. When flow is averaged over the tidal cycle, a clockwise pattern of residual around the ridge results (or counter 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 requirement that there be a significant angle between the axis of the tidal stream and that of the \ dj AIAA % '. >\MEAN BOTTOM FLOW "^ BOUNDARIES FOR MEAN FLOW SHEARS mM SHOAL Fig. 15. Residual current patterns around estuarine and open-shelf ridges. 940 125 ridge is problematic. 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 equilibrium 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 throughout most of the high velocity part of the tidal cycle. According to slender body theory the cross- shoal component of flow during this period can be considered to be two- dimensional, and driven by the cross-shoal pressure 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; Fig. 16 this paper) and Stride (1973). Due to 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. In the large scale, unbounded flows of the open shelf, helical flow struc- ture may continue by mechanisms other than, or in addition to, those oper- ating in river channels and estuaries. Since shelf flow fields are shallow relative to their extent, and occur on a rotating planet, they comprise Ekman boundary flows, whose velocity profiles, above the logarithmic boundary layer, should be upward-expanding Ekman spirals, as a consequence of frictional retardation of the flow by the sea floor. Above critical Reynolds numbers, such flow becomes unstable (Faller, 1963; Faller and Kaylor, 1966; Faller, 1971). However, the instabilities occur within the Ekman field; hence are ordered, not random, and in fact take the form of arrays of horizontal helical flow cells, with adjacent cells rotating in the opposite sense. Such helically structured flow, driven by the semi-diurnal tide, may Fig. 16. Lag effects in a rotating tide. Sand entrainment starts at velocity VI , and continues through velocity V2. (From Postma, 1967.) 941 126 couple with shelf-ridge topographies such as those of the Southern Bight of the North Sea, as has been suggested by Houbolt (1968). To summarize the above considerations, it seems probable that tidal sand ridges may be packaged by a confined estuarine or bay flow regime of reversing tides in a form that will permit them to survive transgression and adapt to a tidal shelf environment. The Southern Bight of the North Sea may provide a case history of such a transition, with the Thames estuary still actively generating ridges, and the Flemish and Zeeland ridge fields and other ridge fields of the Southern Bight marking early zones of ridge generation in a somewhat more confined, estuary-like flow regime, prior to the cutting of the Straights of Dover (Fig.7D). TIDAL SHOAL-RETREAT MASSIFS OF OPEN COASTS Ridge-generating processes of open tidal coasts The estuary to tidal bight transition is less useful as an explanation of several other major ridge fields of tidal shelves, because they may be traced landward to open coasts, rather than into estuaries. One well-studied example, the coast of southeast England, is characterized by a remarkable capacity for subtidal storage of sand. Spits extend across stream mouths, but otherwise the coast is a bluffed mainland coast in which the attempt of the wave regime to build barriers has been inhibited, apparently by the competition of subtidal sand bodies for the debris of shoreface erosion. The shoreline of southeast England, like most unconsolidated shorelines, is separated from the gently dipping shelf floor by a more steeply dipping shoreface, a surface in equilibrium with the hydraulic climate (Fenneman, 1902; Swift et al., 1974). The upper shoreface appears to be a response to the regime of shoaling waves. King (1972, p. 156) has described swash bars which migrate up through the intertidal zone to weld to the beach on the coast of southeast England. The lower shoreface, however, experiences mid- tide surface velocities in excess of 100 cm/sec (2 kn) and is instead tide- dominated. As a consequence of the amphidromic tidal regime of the North Sea, the tidal wave passes south along the coast as a progressive edge wave. Shallow-water distortion of the tidal wave results in a strong flood residual, oriented south, parallel to the coast. The shoreface has responded to the tidal regime by deforming into a series of accretionary bulges, whose subaerial expressions are known as nesses (Robinson, 1966; Fig. 17 this paper). The nesses tend to be connected to submarine sand ridges which extend obliquely upcoast and seaward. As a consequence of their orientation with respect to the coastal tidal wave, the channels landward of them are flood-dominated, while their seaward sides are ebb-dominated. The coast of southeastern England is retreating at rates generally in excess of a meter a year (King, 1972, pp.470— 474) and each ridge crest must serve as a zone of bottom current convergence on which is 942 127 5: ' '\ X ' v \ V \ "south \ v. \ ^X-Vo'v'/ Fig. 17. Shoreface-connected, tide-maintained ridges of the Norfolk coast. (From Robinson, 1966.) 943 128 deposited material eroded from the upcoast shoreface, and also from the downcoast shoreface in the zone of reverse drift behind the ness. Historical records examined by Robinson indicate that the nesses tend to migrate downcoast. Their downcoast migration, coupled with landward translation of the shoreface, appears to have left a poorly organized sequence of detached tidal ridges on the adjacent inner shelf, in a manner similar to the detachment of storm-generated ridges described from the North American shelf (Swift et al., 1974). The Norfolk ridge field as a shoal-retreat massif It is important to remember that a bedform is only half of a closely coupled feedback system, in which a perturbation in the flow field maintains a perturbation in the substrate, and vice versa. The detached offshore ridges (Fig. 18) are "relict" in the sense that they were born in the somewhat different regime and geometry of the shoreface. They are, however, equilib- rium features in that they continue to induce the flow perturbation that originally induced them; otherwise they could not remain as ridges of unconsolidated sand in the face of such high velocities. As the ridges are traced seaward across Fig. 18 to the offshore ridge field a systematic change is, however, apparent; ridge spacing and amplitude increases. This is in accord with Allen's (1968b) contention that the wavelength of large-scale bedforms is proportional to flow depth. Like estuarine ridges, the Norfolk ridges pass through the transition of Fig. 15, from ebb— flood channel couplets to sand cells with the same sense of circulation (Caston and Stride, 1970; Huthnance, 1973). Ridges out to, but not including Hewitt Ridges tend to be sigmoidal in plan view and are subject to Caston's (1970) scheme of cyclic evolution (Fig.6). Seaward of this limit, ebb— flood channel systems are not apparent. Since ebb— flood channel systems are a wave-phase lag phenomenon, the transition may occur at a characteristic Froude number. Nantucket shoals ridge field as a shoal-retreat massif The apparent evolutionary sequence of shoreface-connected ridges, sigmoidal inner-shelf ridges, and arcuate outer-shelf ridges of the Norfolk ridge field suggest that it is a shoal-retreat massif, marking the retreat path of the coastal tidal regime of the Norfolk salient across 90 km of the North Sea floor. As such, it is probably not a unique phenomenon. The great ridge field of Nantucket shoals (Fig. 19) on the U.S. Atlantic shelf may be an analogous case. Here mid-tide flood velocities regularly attain 100—200 cm/sec (2—4 kn; Atlantic Coastal Pilot). There is a continuous sequence of shoal-transverse tidal ridges up the axis of the shoal towards Nantucket Island. The sequence is completed by a series of shoreface-connected ridges on the seaward margin of Nantucket Island (shading denoting shoreface-connected ridges in Fig. 19 is based on ESSA map 0708N-52 with 1 fathom resolution). As in the case of the Norfolk ridge field, ridge spacing 944 129 and amplitude increases steadily in a seaward direction through this sequence. The ridges exhibit seaward asymmetry, as do the Norfolk ridges. The Nantucket map is sufficiently detailed to resolve the sandwave pattern as well as the coarser pattern of ridges. Sandwaves are normal to trough axes, but climb obliquely up the landward flanks of ridges to swing into parallelism with the crest. Paired ebb and flood sinuses occur locally (arrows, Fig. 19) as in the inner Norfolk ridges. However, the asymptotic relationship of sand- woo' 24 24 53' 30 53°00 01*30' 02*00' 02*30' Fig. 18. The Norfolk ridge field. (From Houbolt, 1968.) 945 130 70 '00 69 145 Fig. 19. Nantucket shoals ridge field. Arrows indicate sense of residual flow in major ebb— flood sinus couplets, as inferred from morphology. From C & GS map 0708N-53. Dashed lines are the positions of sand-wave crests inferred from the 1 fathom contours of the original map. waves to the crests of asymmetric ridges suggests that most ridges are of the open-shelf type, with each ridge having the same sense of residual current circulation around it (Fig. 15). As indicated by its sea-truncated drainage pattern, Nantucket Island is the remnant of a much larger subaerial surface, the ancestral Nantucket 946 131 peninsula. The spatial series of ridges apparent in Fig. 19 may be viewed on uniformitarian grounds, as equivalent to a time series. This equation suggests that as the Nantucket shoreface has undergone erosional retreat, the resulting debris has been packaged as shorei'ace-connected ridges, which in turn have evolved in response to a deepening flow field. REFERENCES Allen, J. R. L., 1965. Late Quaternary Niger Delta and adjacent areas: sedimentary environment and lithofacies. Am. Assoc. Pet. Geol. Bull., 49: 547—600. Allen, J. R. L., 1968a. Current Ripples. North-Holland, Amsterdam, 433 pp. Allen, J. R. L., 1968b. The nature and origin of bedform hierarchies. Sedimentology, 10: 161-182. Bagnold, R. A., 1966. An approach to the sediment transport problem from general physics. U.S. Geol. Surv. Prof. Pap., 422—1, 37 pp. Belderson, R. H. and Stride, A. H., 1966. Tidal current fashioning of a basal bed. Mar. Geol., 4: 237-257. Belderson, R. H., Kenyon, N. H. and Stride, A. H., 1971. Holocene sediments on the continental shelf west of the British Isles. In: F. M. Delany, ICSU/SCOR Working Party 31 Symposium, Cambridge, 1970, pp.157— 170. Brouwer, J. E. J., 1964. The North Sea. In: L. P. Stamp (Editor), The North Sea and the law of the Continental Shelf. World Land Use Survey Occ. Pap., 5, 41 pp. Caston, V. N. D., 1970. Linear sand banks in the southern North Sea. Sedimentology, 18: 63-78. Caston, V. N. D. and Stride, A. H., 1970. Tidal sand movement between some linear sand banks in the North Sea off northeast Norfolk. Mar. Geol., 9: M38— M42. Curray, J. R., 1965. The Quaternary of the United States. In: H. E. Wright Jr. and D. G. Frey (Editors), The Quaternary of the United States. Princeton University Press, Princeton, N. J., pp.723— 735. 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. Rev. Ecol. Syst., 2: 201-235. 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. Fenneman, N. M., 1902. Development of the profile of equilibrium of the subaqueous shore terrace. J. Geol., 10: 1—32. Gessner, F. B., 1973. The origin of turbulent secondary flow along a corner. J. Fluid Mech., 58: 1-25. Houbolt, J. J. H. C, 1968. Recent sediments in the southern bight of the North Sea. Geol. Mijnbouw, 47: 245—273. Huthnance, J. M., 1973. Tidal current asymmetries over the Norfolk sandbanks. Estuarine Coastal Mar. Sci., 1: 89—99. Jeffreys, H., 1929. On the transverse circulation in streams. Proc. Camb. Phil. Soc, 25: 20-25. Johnson, D. W., 1919. Shore Processes and Shoreline Development. Hafner, New York, N.Y. 1965 facsimile, 584 pp. Kennedy, J. and Fulton Jr., J. F., 1961. The effect of secondary currents upon the capacity of a straight open channel. Trans. EIC, 5: 12 — 18. Kenyon, N. H., 1970a. The origin of some transverse sand patches in the Celtic Sea. Geol. Mag., 107: 389—394. Kenyon, N. H., 1970b. Sand ribbons of European tidal seas. Mar. Gel., 9: 25—39. Kenyon, N. H. and Stride, A. H., 1970. The tide-swept continental shelf sediments between the Shetland Isles and France. Sedimentology, 14: 159—175. 947 132 King, C. A. M., 197 2. Beaches and Coasts, St. Martin's Press, New York, N.Y., 570 pp. Kraft, J. C, Sheridan, R. E., Moose, R. D., Strom, R. N. and Wiel, C. B., 1974. Middle- Late Holocene evolution of the morphology of a drowned estuary system. In: G. Allen (Editor), Estuary and Shelf Sedimentation: a Symposium. University of Bordeaux. (in press). Leopold, L. B., Wolman, M. Gordon, and Miller, J. P., 1964. Fluvial Processes in Geomorphology. W. H. Freeman, San Francisco, Calif., 522 pp. Ludwick, J. C, 1970. Sand waves and tidal channels in the entrance to Chesapeake Bay. Inst. Oceanography, Old Dominion Univ., Tech. Rep. 1, 7 pp. Ludwick, J. C, 1974a. Tidal currents, sediment transport, and sandbanks in Chesapeake Bay entrance, Virginia. In: M. O. Hayes (Editor), Int. Estuarine Conf. Proc, 2nd, Myrtle Beach, S. C, Oct. 15—18, 1973, in press. Ludwick, J. C, 1974b. Tidal currents and zig-zag shoals in a wide estuary entrance. Geol. Soc. Am. Bull., 85: 717-726. Moore, G. T., 1970. Role of salt wedge in bar finger sand and delta development. Am. Assoc. Pet. Geol. Bull., 54: 326—333. O'Brien, M.P., 1931. Estuary tidal prisms related to entrance areas. J. Civ. Eng., 1: 738-743. Oertel, G. F., 1972. Sediment transport of estuary entrance shoals and the formation of swash platforms. J. Sed. Petrol., 42: 858—863. Off, T., 1963. Rhythmic linear sand bodies caused by tidal currents. Am. Assoc. Pet. Geol. Bull., 47: 324-341. Oomkens, F. and Terwindt, J. H. J., 1960. Inshore estuarine sediments in the Haringvliet (Netherlands). Geol. Mijnbouw, 39: 701-710. Postma, H., 1967. Sediment transport and sedimentation in the estuarine environment. In: G. H. Lauff (Editor), Estuaries. Am. Assoc. Adv. Sci., Washington, D.C., 757 pp. Reineck, H. E., 1970. Marine Sandkorper, rezent und fossil. Geol. Rundsch., 60: 302~ 321. Robinson, A. H. W., 1956. The submarine morphology of certain port approach channel systems. Inst. Navig. J., 9: 20—46. Robinson, A. H. W., 1966. Residual currents in relation to shoreline evolution of the East Anglian coast. Mar. Geol., 4: 57—84. Smith, J. D., 1969. Geomorphology of a sand ridge. J. Geol., 77: 39—55. Stride, A. H., 1963. Current swept sea floors near the southern half of Great Britain. Q. J. Geol. Soc. Lond., 119: 175-199. Stride, A. H., 1970. Shape and size trends for sand waves in a depositional zone of the North Sea. Geol. Mag., 107: 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, July, 1972, p. 97. Stride, A. H., Belderson, R. H. and Kenyon, N.H., 1972. Longitudinal furrows and depositional sand bodies of the English Channel. Mem. Bur. Rech. Geol. Minieres, 79: 233-244. Stubblefield, W. L., Lavelle, J. W., McKinney, T. F. and Swift, D. J. P., 1974. Sediment response to the hydraulic regime on the central New Jersey shelf. J. Sed. Petrol., in press. 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. and McMullen, R. M., 1968. Preliminary studies of intertidal sand bodies in the Minas Basin, Bay of Fundy, Nova Scotia. Can. J. Earth Sci., 5: 175—183. Swift, D. J. P. and Sears, P., 1974. Estuarine and littoral patterns in the surficial sand sheet, central and southern Atlantic shelf of North America. In: G. P. Allen (Editor) Estuary and Shelf Sedimentation: a Symposium. University of Bordeaux, Talence, 1973, in press. Swift, D. J. P., Stanley, D. J. and Curray, J. R., 1971. Relict sediments, a reconsideration. J. Geol., 79: 322-346. 948 133 Swift, D. J. P., Kofoed, J. W., Saulsbury, F. P. and Sears, P., 1972. Holocene evolution of the shelf surface, central and southern Atlantic coast of North America. In: D. J. P. Swift, D. B. Duane, and 0. H. Pilkey (Editors), Shelf Sediment Transport: Process and Pattern. Dowden, Hutchinson and Ross, Stroudsburg, Pa., pp.499— 574. Swift, D. J. P., Duane, D. B. and McKinney, T., 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., Lavelle, J. W. and McHone, J., 1975. Maintenance of the coastal equilibrium profile: Some data from the Middle Atlantic Bight, U.S.A. Sedimentology, in press. Weil, C. B., Moose, R. D. and Sheridan, R. E., 1974. A model for the evolution of linear tidal built sand ridges in Delaware Bay, U.S.A. In: G. Allen (Editor), Estuary and Shelf Sedimentation: a Symposium. University of Bordeaux, Talence, 1973, in press. Wilson, I. G., 1972. Aeolian bedforms — their development and origins. Sedimentology, 19: 173-210. Wilson, I. G., 1973. Equilibrium cross-section of meandering and braided rivers. Nature, 241: 393-394. 949 Reprinted from: Mid-Atlantic Shelf /New York Bight ASLO £3 Symposium, 65-66. Donald J.P. Swift. George L. Freeland, Peter E. Gadd, J. William Lavelle and William L. Stubblefield NOAA/AOML, 15 Rickenbacker Causeway, Miami, Florida 33149 MORPHOLOGIC EVOLUTION AND SAND TRANSPORT IN THE NEW YORK BIGHT The New York Bight is a roughly pentagonal sector of North America's Middle Atlantic Bight (Fig. 1). It is divided into a series of compartments by transverse shelf valleys, including the Block, Long Island, Hudson, North New Jersey, Great Egg, and Delaware Shelf Valleys (1). These f eatures'were incised repeatedly into the shelf surface by rivers during Pleistocene low stands of the sea, although not always in the same position. At the onset of the Holocene transgression the valleys served as retreat paths for estuaries, and they are consequently partly or largely filled with estuarine depo- sits. As the shoreline passed over a given segment of the buried channel, the channel was in many cases partially re-excavated by estuary mouth scour, while its northern margin tended to be aggraded by littoral drift from the adjacent coastal compartment. As a result of this diverse activity, shelf valley morphology may be quite complex. The shelf valleys themselves are in many cases the retreat paths of estuary mouth scour trenches and do not everywhere directly overlie the buried river-cut channel. They may be flanked by shoal retreat mas- sifs; the retreat paths of estuary mouth shoals. Shelf valleys commonly extend seaward into mid- shelf or shelf-edge deltas. The Holocene transgression appears to have been intermittent in nature, and during stillstands, scarps were locally cut into the shelf surface. The plateau-like interfluves between shelf valleys also underwent considerable modification dur- ing the Holocene transgression. At any given time, the nearshore hydraulic climate tended to maintain an inner shelf profile, of more steeply dipping shoreface, and more gently sloping shelf floor. As sealevel rose, this profile was translated landward by means of storm erosion of the shoreface. The eroded sand moved along the coast under the impetus of wind- and wave-driven littoral currents, to be deposited at the tips of spits or in estuary mouth shoals. However, during this process, much of it leaked out onto the adjacent shelf floor. As a result, the shelf floor has been veneered with 0-10 m of sand of primarily local origin. Modification of the shelf floor did not cease with passage of the shoreline. As a consequence of its geometry, the Middle Atlantic Bight between Cape Cod and Cape Hatteras is subjected to an unusu- ally rigorous hydraulic climate (2). The storm tracks of mid-latitude lows tend to pass northward along the length of the Bight. Many storms nest for an interval in the Bight, so that the isobars of atmospheric pressure parallel the isobaths of the shelf surface, and intense northeast winds blow down the entire arc of the Bight. The result of this "scale-matching" phenomenon is a strong geostrophic coupling between wind and water flow, so that the shelf water column may translate uniformly southward, in slab-like fashion at mid-depth speeds in excess of 30 cm sec-1, for periods up to several days. Within a few km of the Deach a coastal jet may form in response to direct wind stress. Flows up to 60 cm sec have been observed in this nearshore zone (as measured by Savorius rotor meters 1 m off the bottom; [3]). Comparison of current meter records with the dispersal patterns of sand labeled with a radioiso- tope (Fig. 2) indicates that the transport of sand on the shelf occurs in intense bursts, when storms accelerate near bottom (1-2 m) flow above a threshold of 20 cm sec . Transport events may last for hours or days, and are commonly separated by days to weeks of quiescence. Transport is generally parallel to isobaths, and' may trend either southwesterly or northeasterly, depending on the prevailing direction of the flow event. During the period November 11 to January 31, a single intense flow asso- ciated with a 3-day "northeaster" drove labeled fine sand for 1200 m to the west of the injection point along the Long Island inner shelf. Calculations of fluid power expended on the sea floor sug- gest that this northeaster induced a greater volume of sand transport on the sea floor than did the sum of the eight other flow events of the observation period (3). The effect of this regime on the shelf floor is various. Wave oscillation ripples and trains of sandwaves up to 1 m high may be seen after storm flows. Sand and gravel ribbons up to 50 in wide by a km long may also appear. However, the most striking response of the sea floor to storm flows is a ridge and swale topography, of sand ridges up tc 10 m high, 2-4 km apart, and with side slopes less than a degree. Ridge crests may be traced for tens of km, converging to the south with the shoreline at angles of 15 to 35°. Unlike smaller scale bedforms which appear to be responses to a single flow event or season of events, the ridges appear to be long term, time-averaged responses. Ridges on the Central New Jersey Shelf yield successively older dates with depth, that span the duration of the Holocene transgression. On the New Jersey and Long Island coasts, ridges appear to be presently forming at the foot of the shoreface. Ridge sequences may be traced as far as 20 km offshore in apparent genetic sequence. The ridges appear to be responses to the wind-sheared flow of the inner shelf, but the formative mechanism is not clear. Storm flow is predominantly southward and seaward, obliquely over the ridge crests, and there is local evidence that nearshore ridges migrate very slowly offshore and downcoast, extending their crestlines to keep contact with the retreating shoreline as they do so (4) . Shifts of a unit width in 50 or more years have been noted, although the actual movement may have been confined to a few major storxs. Seismic studies of ridges on the Central New Jersey shelf indicate that ridge modi- fication continues even after contact with the retreating shoreline is finally broken; troughs con- tinue to be scoured, crests aggraded, and ridge positions slowly shift. (l)Swift, D.J. P., Kofoed, J.W., Saulsbury, F.P. and Sears, P., 1972. Shelf Sediment Transport: "rocesa and Pattern. Dowden, Hutchinson and Ross, Stroudsburg, Pa., p. 499-574. (2)Beardsley, R.C. and Butman, B., 1974. Ceophys. Res. Letters, 1:181-184. (3)Lavelle, J.W. and others. Paper submitted to Geophys. Res. Letters. (4)Moody, D.W., 1964. Ph.D. Thesis, Johns Hopkins University, 167p. 950 •66- 76°39° 75° 40° 74° 41° 73° 42° ) 42° 1 72° _„,, — > SURFACE CHANNEL „_, .... SUBSURFACE CHANNEl ^ CU£ST» ,SCARp . 4 SHELF EDGE, MID-SHEIF DELTAS gg§ SHOAL RETREAT MASSIFS EH SAND RIDC-ES . 71° 41° 73° 38" 72° 39° 71° 40° Fig. 1. Morphologic framework of the New York shelf surface. 400 V) tc. w2O0i 25 NOV 74 1000 800 ' 600 4C0 200 ~0 ' 200~ 400 METERS '.■■-"- J - . i '",\\ ' ' . ^JO'. ' X ' . 400- n B £ ui200- 2 10 JAN 75 ■., . , ^ix&mi't;.^^', ; \ 0- — I 1 1 1 1 1 1 1 1 1 1 1 1 1000 800 600 400 200 200 400 METERS %, TRANSPORT O ,nn 0 20 0 20 w 100 1 data. 964 J. 00 2.00 4.00 6.00 .4 1 1 i U. CM/SEC 8.00 10.00 12.00 — « H-J 14.00 16.00 18.00 20.00 -10.00 -8-00 -6-00 -4.00 -2.00 0-00 V. CM/SEC 2-00 4.00 6.00 8.00 10-00 Figure 5. Mean water '-velocity profile. Downstream (+) and cross-stream (x) velocity components were derived from bubble lines. The solid line is an exponential curve fitting the deeper downstream data. 965 J,- 00 r>' 2.00 ♦ .00 5.00 U, CM/SEC 8.00 10.00 12.00 14.00 16.00 18.00 20-00 + + + + + + + + + + + + ♦ 9.60 7.20 4.80 U + 2.40 0.00 Figure 6. Downstream component of mean water-velocity profile. The straight line is fitted to the slope near the surface 3 and the curved departure represents the profile at a solid boundary in dimensionless units Y+ and U+. 966 Table 1. Water Velocity Measurements from Bubble Photographs Figure 3 4 5 6 Number of frames analyzed 46 11 17 37 Duration of experiment, sec 3.5 10 26 4.3 U , cm/sec a 900 550 600 550 Fetch, cm 24 25 69 382 Location early wave above wave within wave below wave rows rows rows rows Best exponential fit a, cm/sec 18.3 13.6 — — b, cm" 12.4 8.6 — — rms G. , cm/ sec 0.15 0.15 — — W, cm -0.017 -0.002 +0.005 -0.056 Stress, dynes/ cm2 2.0 1.0 1.2 0.7 967 curve in figure 5 approximates only a portion of the profile. The best exponential fit and the rms departure of the mean velocity values from the fit were obtained with a computer. The mean depth difference be- tween the bases of successive bubble lines is given in the table as W where a positive value represents a mean rise. The surface stress t0 was calculated from the exponential curves in figures 3 and 4 and from the smoothed surface velocity gradient in figures 5 and 6. The longer durations in figures 4 and 5 resulted from analyzing only 1 frame/sec. Another analysis of the run in figure 6, not presented, extended over 85 sec but gave a profile similar to that shown in the figure. The bubble-line photographs allow measurements of water velocity very near the surface, even among waves of some steepness, although the accuracies in figures 5 and 6 were decreased by wave shapes. The mean velocity profiles in figure 3 were obtained in the early laminar-turbu- lent zone where the flow was largely laminar in nature. The down- stream velocity component U fit a smooth curve well. The cross-stream component V also fit a smooth curve, although it was offset. Both pro- files showed departures near the surface, attributed to airflow inter- ference by equipment mounted asymmetrically above the bubble wire. The buoyancy- induced vertical motion between frames was small. The velocities in figure 3 fluctuated rapidly between frames. The rms fluctuation at a depth was about 15 percent of the mean velocity at that depth. The depth-averaged variance, reduced by the number of frames in the average, led to a calculated rms fluctuation of the aver- age of 0.16 cm/sec, comparable to the observed rms departure of the mean 968 profile from the fitted smooth curve. The correlation of fluctuations at depth with the surface fluctuations showed a rapid decrease with depth, reaching a value of 0.5 at a depth of 0.035 cm, then remaining near 0.3 to 0.4 at greater depths. Because the theoretical correlation caused by sinusoidal water waves is unity at all depths, the shape of the correlation curve indicates very shallow, rapid fluctuations pos- sibly caused by fluctuations in air stress. The kurtosis of the fre- quency distribution of surface fluctuations was -0.2, compared with 0 for a normal distribution and with -1.2 for a uniform distribution. This value is in accord with the possibility that the surface fluctua- tions were largely turbulent in origin. A mean surface stress of 2.0 dynes/ cm2 was indicated by the sur- face shear in figure 3. It appears that the stress was not constant with fetch, but was disturbed by an irregularity at the upwind edge of the water. This is supported by observations of changes in the fetch at which the waveforms underwent transition when the water level was changed slightly. The data in figure A were obtained at a lower airspeed well before transition so that the flow was solely a developing laminar boundary layer. The profile fit a smooth curve well, with a calculated stress of 1.0 dynes/cm2. Figure 5 was obtained within the laminar-turbulent transition shown by the wave patterns and is of an intermediate shape not fitting a simple analytic curve. The indicated stress is close to that in figure A. 969 Figure 6 was obtained well downstream where both the circulation and the waves represented the little-changing pattern of development be- yond the early transition. A straight line representing the surface shear is included, and a curve is drawn to show the departure from a linear profile observed in a turbulent transition layer at a solid boundary. This curve was calculated using the observed surface stress and average measurements from graphs of turbulent boundary-flow profiles given by Mellor and Herring (1969) and by Kline et al. (1967) in non- dimensional form. The viscous sublayer at the free surface is roughly one-half as thick as at a solid boundary. The different thickness may indicate that the value of von Kantian' s constant k is larger in turbu- lence at a free boundary. Also, the indicated surface stress supported by water shear is less than upstream, although other measurements in- dicate the total stress to be greater. The decrease might be attributed to the effects of waves on momentum transport. Figure 7 shows mean water velocity measured by a hot-film ane- mometer at different depths as a function of fetch with a midstream airspeed of 850 cm/sec. Mean water velocities obtained from a sub- merged hot-film anemometer extend to greater depths than do the bubble- line photographs, although such data are not available near a wavy sur- face. The "zero" depth readings here were obtained at the minimum con- tinuously submerged depth of the probe. The dip in the "zero" depth plot was found in two additional series of measurements. The water velocities in figure 7 show an initial rise in surface water velocity 970 Sis 10 - S 3 Figure 7. Mean water velocity versus fetch at different depths. 971 and deeper increases farther downstream, as expected for a developing laminar boundary layer. A dip in surface velocity and a rapid acceler- ation below mark the beginning of vertical mixing by turbulence and occur at the rapid growth of waves. Another set of these measurements is contoured in figure 8 as a vertical longitudinal section of water velocity. A laminar boundary layer developed to a depth of 0.5 cm in the first 20 cm of fetch. Then the near-surface velocity decreased between 25 and 35 cm of fetch while the flow below accelerated rapidly. Farther downstream, the mean ve- locities increased again as a turbulent boundary layer. The onset of vertical mixing is most clearly demonstrated in the velocity difference between the surface and a depth of 0.5 cm. The mean difference was 11 cm/sec at a fetch of 20 cm and 4 cm/sec at 30 cm of fetch. Figure 9 contains two sets of measurements of wave energy versus fetch with an airspeed of 850 cm/sec. A nearly fixed position identi- fied at uhe beginning of each wave row is referred to as a "source point." The fetches of the source points and the approximate ends of wave rows are shown. Two different high-pass filters nominally passing frequencies greater than 1 and 10 Hz were used in the rms voltmeter for these experiments. This figure shows drastic changes in wave energy at the locations of the water current changes. The two plots of wave energy versus fetch show very little growth before the source points at a fetch of about 15 cm, rapid growth to a fetch of 25 to 30 cm, and then slower growth as if approaching a saturation. 972 x (cm) Figure 8. Vertical longitudinal section of water velocity. The contour interval is 1 cm/ sec. 973 6j» o OS +i •< CO CO CO 00 3 I -a CO in CO 3 CO ?h 00 CO OS CO K 974 Wave energy at the same airspeed passed by a sharp high-pass filter with a cutoff at 8 Hz is shown in figure 10. There is a dis- tinct maximum in wave energy at a fetch of 35 cm. The energy at this fetch was nearly as great as the less filtered values in figure 9. After the initial wave development, the filtered high-frequency wave energy was nearly constant to a fetch of at least 400 cm, although the lower frequency waves were seen to be much greater. The capillary- gravity waves grew and approached saturation before lower frequency waves developed significantly. Figure 11 shows wave energy at a fetch of 400 cm as a function of time as the airflow was increased suddenly. Air leakage over the water surface before opening the blower, and possibly motor vibration, led to very small amplitude waves with a frequency of 2 crests/ sec. These apparently did not play a role in the further development of the wave field. Only slight wave growth occurred during the first 15 sec of fast airflow; the waves had a frequency of 15 crests/sec. The rate of wave energy growth next increased abruptly to a peak in energy. Approximately 95 percent of the wave energy at the peak arose during the period of rapid growth. The crest frequency of the rapidly growing waves was ?nly 7 crests/sec. The wave energy then decreased to about 10 percent of the peak. Repeated experiments give somewhat different results, but the pattern of wave growth described here was found repeat- edly. The measurements of wave energy at different fetches in figures 9 and 10 were replotted on a semilogarithmic graph in figure 12. The 975 ^ E 100 Figure 10. Filtered wave energy versus fetch. 976 500r20r 400 300 200 ••• •• •• • •••••••• • uc •• •••• •• ••••••. Figure 11. Wave energy versus time after the start of increased airflow. The wave energy was measured with a meter (E ) and calculated from the ■ m wvdth of the wave envelope (E ). Airspeed (U ) and crest frequency (f) are also plotted. 911 LU 0 -I * xx -2L x - «- 0 0 0 10 10 20 20 30 10 20 30 40 X(cm) 30 40 50 40 _J 50 50 Figure 12. Wave energy versus fetch during three experiments in figures 9 and 10. 978 measurements of wave energy versus time in figures 11 and 15 were re- plotted in figure 13. Line segments were drawn to show linear sections of the plots. The very different energy levels in figure 13 are attri- buted to the different airspeeds. The plot of wave energy versus fetch in figure 12 shows that approximately 90 to 95 percent of the wave growth occurred in an exponential manner between about 15 and 25 to 30 cm of fetch. In the less filtered data, the early wave growth was ob- scured by unrelated lower frequency waves, but the plot of filtered data shows an earlier exponential growth stage of small amplitude waves. Similarly, the replot of wave energy versus time shown in figure 13 shows two stages of exponential wave growth. The waves in the first, minor stage were found to have a frequency near 15 crests/sec, and those in the major stage had a frequency near 7 crests/ sec. Figure 14 contains hot-film anemometer measurements of water velocities at a depth of 0.2 cm and a fetch of 50 cm as the airflow was started slowly. The normalized square of the width of the veloc- ity fluctuation envelope and the airspeed were also plotted. This figure shows that when the airflow increased gradually, the water velocity increased rapidly; but it decreased abruptly as the steep waves developed. The velocity decrease may be associated with the dips in surface velocity in figures 7 and 8. A similar experiment with a fetch of 365 cm, a velocity probe depth of 0.3 cm, a wave gage, and a rapidly started airflow is shown in figure 15. The water velocity probe gave erratic readings when the 979 o Figure 15. Wave energy versus time. Data are replotted from figures 11 (+) and 15 (x) . 980 Ua Uw (Uw')2 cm cm cm2 sec sec secz 500 r l0 r l0 400 - 8 -0.8 300 - 6 -0.6 200 - 4 -0.4 100 - 2 -0.2 0 - 0 - 0' Figure 14. Near-surface water velocity versus time. The short-term aver- age velocity U (dashed line) 3 the square of the velocity fluctuations (U ') (continuous line) 3 and the airspeed U (dots) are shown, w a 981 Ua cm E erg sec cm2 800 700 600 500 400 300 200 100 200 Uw cm sec - 14 - 12 - 10 - 8 - 6 - 4 - 2 175 - 150 125 - 100 - 75 - 50 - 25 F c sec ^^--».^ *'"*' "***" "V»*— •.—•.-.!.— •■- -*Ua ^'~\./ - 20 - 15 - 10 - 5 A* —j 0L 0L 0L l- -2 J 1 2 4 6 8 10 12 14 T (sec) Figure 15. Near-surface water velocity and wave energy versus time. The mean velocity (U ) is interrupted as the wave energy (E) increases. Airspeed (U ) and wave crest frequency (F) are included. 982 waves appeared. Wave energy was calculated from the width of the wave envelope, and the frequency plot represents numbers of wave crests. The near-surface velocity rose rapidly when the fast airflow began, and only later did significant waves appear. In experiments in which the velocity probe was at a depth of 0.5 cm or greater, the increase in velocity was delayed until the steep waves appeared. In a different but similar wind tunnel installation, an infrared radiation thermometer viewed a 30-cm diameter circular area of the water surface at a fetch of 150 cm and a maximum airspeed of 600 cm/sec. Figure 16 shows the surface temperature changes when the blower was turned on. The water surface temperature changes are con- sistent with the changes in water circulation described above. The water surface temperature decreased as the airflow developed, rose suddenly as steep waves appeared in the measurement area, and then became steady at an intermediate value. The rise in surface tempera- ture is taken to represent the beginning of vertical mixing indicated in figures 7 and 8. In addition, the difference between surface and bulk water tem- peratures was recorded using the infrared radiation thermometer mounted above a water surface at a fetch of 190 cm at different airspeeds. As the airspeed increased, the line of source points moved upstream past the radiometer, and the water surface there became rough. Figure 17 shows a series of measurements over smooth and rough surfaces and a second more detailed series over a rough surface. The variation in surface temperature depression at a single fetch with airspeed further 983 25- 20- 15 - 10 - T (sec) 5 - 0 - <*r M J «r — f- i-li^J h 0.5°C ^ Figure 16. Water surface temperatures versus time after an air- flow began. Arrows mark the start of the airflow and also show the time that the water surface beneath the radiometer became rough. 984 0.5 r o o < 0.4 0.3 0.2 0.1 7 & L L X = SMOOTH + ■ ROUGH 200 400 U.($) 600 800 Figure 17. Two series of measurements of water surface temperature de- pression versus airspeed. In one seriesi the surface was characterized as either smooth or rough during each measurement. 985 demonstrates changes in water circulation at the transition. Although there is a gap in the data at low airspeeds, the data show the depres- sion to increase with airspeed where the surface was smooth, the appear- ance of an apparent discontinuity, and then a lessening of the depres- sion with airspeed above the rough surface. The waveform transition was within the measurement area at the discontinuity in the temperature depression curve. A second, higher set of readings shows the shape of the temperature depression curve downstream of the transition to be concave, opposite to the shape of the curve upstream. The photograph in figure 18 shows the initial wind waves on a water surface. The width of the scene was about 90 cm. The small- amplitude water waves were regular and long-crested. Rapidly growing waves developed only well downstream, had short-crested "dimple" shapes, and were mostly in rows alined along the direction of the airflow. Irregular larger waves formed downstream of the wave rows. Figure 19 is a nearly vertical photograph of the water surface where the steep waves first appeared. The regular precursor waves had a wavelength of about 1.8 cm. Dimple-shaped waves developed from the troughs of the precursor waves. The oblique photograph in figure 20 is a 0.5-sec time exposure of the surface reflection of a floodlamp on the same region of a water surface. The time exposure demonstrates that succeeding dimples formed in the same location, and each followed the previous one to give rows of dimples with bands of smooth water between. The source points of the 986 Figure 18. Upwind portion of a water surface disturbed by an airflow. 987 Figure 19. Initial wind waves. The length of the arrow represents 1.8 em. Bright spots in the dark wave troughs resulted from multiple reflections. 988 Figure 20. Time exposure of transitional waves in rows. The diameter of the area of reflection was about 50 am. The light bands in the upwind portion of the reflec- tion were given by steep, short-crested waves moving downwind; the dark bands between them were regions of smooth water. 989 steep waves remained in one position for several seconds before shifting rather abruptly to a new configuration. Figure 21 shows the variation in fetch of source points with air- speed as observed by eye. The mean fetch as well as the mean crosswind separation varied approximately as the reciprocal of the square of the airspeed. The fetch was less with more turbulent airflows and also varied with water level. As the level was raised and water overflowed into the entrance duct, the source points moved upstream and passed into the entrance duct when the water depth in the duct was on the order of 0.2 to 0.3 cm. When alternatively a false bottom was installed in the tank to give shallow water depths, the fetch to source points and the initial waves appeared to be affected only at depths of less than 0.3 cm. A drop of fluorescein dye solution falling a short distance onto the upwind surface of water beneath a slow airflow separated into a patch of surface film and a subsurface blob connected by a thin streak of dye. The surface film drifted downstream and separated into patches on the surface passing between wave rows. The dye streak became stretched downstream between two wave rows and appeared to undergo rotation about a longitudinal axis. When the dye encountered the edge of a compacted surface film, it moved as if the compacted film were a solid boundary. Floating dust particles also showed these surface circulations. 990 cj cvj o en e CVJ O X o 50 100 x (cm) Figure 21. Fetch of source points (a) and traverse spacing (b) versus airspeed. 991 V. DISCUSSION The present experiments furnish a description of the intermediate patterns of wave and current flow during the development of wind waves and wind-driven water turbulence. Different stages in the development of waves and current preceded the appearance of the gravity and capil- lary wave pattern characteristic of wind waves and water turbulence. In most respects, the upstream patterns at different fetches with a steady airflow were duplicated at a fixed fetch downstream when an airflow was started abruptly, and the two situations are combined in a description of the major features of wave and current development. The first developing waves were regular with long crests. In these experiments, the wave length was near 1.8 cm and the frequency was near 15 Hz, about that of capillary-gravity waves of minimum phase speed. The energy of these precursor waves increased with fetch some- times to a value near 0.2 to 0.3 ergs/cm, approximately 5 percent of the wave energy after the next growth phase. With a steady airflow, the form of the early wave energy growth was obscured by reflected and other unrelated waves on the surface, but the energy through a high- pass filter increased exponentially. These other waves were absent when the airflow was started abruptly, and the first growth rate was clearly seen to be exponential. A laminar boundary layer with a smooth velocity profile developed in the water. The transition between the first and second phases of wave de- velopment was marked by a transverse line of source points on the water 992 surface, each of which emitted a rapid series of steep, dimple-shaped waves. Each dimple developed from the trough of a small-amplitude pre- cursor wave. For some distance downstream of the source points, the dimples tended to travel in rows parallel to the direction of the air- flow. When the airflow was started abruptly and after several seconds of development of precursor waves, the row of source points appeared at the appropriate fetch in the upwind portion of the water tank. The transition from precursor waves to dimples in rows moved rapidly down the tank, possibly at the airspeed, so that rows of dimples momentarily covered nearly the entire water surface. The transition region also had an exponential increase in wave energy. The growth factor was more than 10 times greater than previous- ly measured when the airflow was started abruptly, and most of the wave growth occurred during this phase. The wave frequency decreased by about one-half before the waves became highly irregular. The wave amplitudes decreased after this phase and later developed again. With an abruptly increasing airflow, the steep transition waves appeared to be in packets. The pattern of rows of dimples continued until a wave energy of about 5 erg/cm2 was attained. The waves then developed, apparently through some intermediate patterns, to long-crested, separate gravity and capillary waves. With further development, the gravity and capil- lary waves continuously showed greater differences in frequency; the gravity waves increased in amplitude while the capillary waves appeared 993 to be saturated in energy. The final pattern was characteristic of wind waves on larger, open bodies of water. In all cases, the waveform transition was accompanied by a transi- tion from a laminar waterflow to turbulence. Particles floating on the surface and dye within the upper water indicated shallow longitudinal water vortexes with surface convergences between the rows of waves. The depth region of significant currents increased from less than 0.5 to more than 2 cm, and the surface speed decreased slightly over a short distance. The water surface temperature reflected this flow change. At successively greater airspeeds, the surface temperature at one position decreased (when the water was being cooled by the air) at a decreasing rate; but the temperature increased with airspeed after the transition, again at a decreasing rate. When the airflow was started abruptly, the water surface began to cool; but at the appearance of steep waves the surface warmed abruptly, and the temperature became constant at an intermediate value. The velocity profile in the turbulent flow region (fig. 6) shows a linear region near the surface. This is believed to be the first direct demonstration of a viscous sublayer in water at the free surface. The viscous sublayer is significantly thinner at a free surface than at a solid wall. The difference could result in part from the less restric- tive boundary condition at the air-water interface. The relation between the fetch to the wave source points and the airspeed is in accord with a general relation between air stress t and 994 transition fetch x of current and waves obtained through a dimensional c analysis. Several quantities appear in the analysis, but considerations of the mechanisms reduce the number of independent quantities. Stress and viscosity u are considered to act only upon water currents, with density p , as in the combination x/pu. Surface tension r and gravity g act with density only to propagate capillary-gravity waves, and a ratio having dimensions of velocity can be written (Tg/p) . One form of di- xx mensionless ratio of the pertinent quantities is tt » — . Some PuCrg/p)"* previous laboratory studies have found stress approximately proportion- al to the square of the airspeed (e.g., Shemdin, 1972) so that transi- tion fetch may vary as the reciprocal of the square of the airspeed. The relation xx ■ constant implies that the kinetic energy of the flow c at transition has a fixed value. The variation of transition fetch with the smoothness of the intersection at the bottom of the air entrance duct and the water surface demonstrated a variation in stress as did the fetch dependence on the turbulence in the air. However, observations that the fetch was not affected by bottom depth at depths greater than a few millimeters illustrate the shallowness of the circulation deter- mining the transition. Water velocity shear near the surface leads to an interaction be- tween the flow and waves at transition. Some shear persists downstream, and some interaction must continue throughout a rough water surface. The correlation between waves and current may have implications to prob- lems of air-sea interaction. The Miles and Phillips theories of wind- wave generation treat only the influence of air pressure (Kinsman, 1965) 995 and do not pertain to a waterflow-mediated generation of initial steep waves. Longuet-Higgins (1969) suggested that major waves grow by energy transfer from short waves, but gave no further suggestion of the means of short wave growth. Additional study of the energy transfer from wa- ter current to waves appears desirable. Also exchange rates between the sea and atmosphere frequently are limited by the rate of diffusion a- cross the viscous sublayer in the water. The thickness of the layer thus partly controls the exchange rate, and this thickness is less than in corresponding flows at a solid boundary. Further, current theories of laminar-turbulent transition at a solid boundary cannot apply near a free surface. VI. CONCLUSIONS The initial growth of steep waves beneath an airflow proceeds by a unique mechanism involving three-dimensional wave shapes and an inter- action with the shear flow of the water near the surface. Water turbu- lence near the surface is modified by the waves and the free motions of the surface. A viscous sublayer exists beneath the water surface, but is thinner than near a solid boundary. VII. ACKNOWLEDGMENTS William Everard designed and constructed the special electronic equipment. Figures 16 through 21 were obtained while at the Scripps Institution of Oceanography, La Jolla, Calif. 996 VIII. REFERENCES Hires, R. I. (1968), An experimental study of wind-wave interactions, Reference 68-5, Chesapeake Bay Institute, The Johns Hopkins Univ. , Baltimore, Md., 169 pp. Kinsman, B. (1965), Wind Waves; Their Generation and Propagation on the Ocean Surface, Prentice-Hall, Englewood Cliffs, N. J., 676 pp. Kline, S. J., W. C. Reynolds, F. A. Schraub, and P. W. Runstadler (1967), The structure of turbulent boundary layers, Fig. 9b, J. Fluid Me- chanics, 30 (Pt. 4):741-773. Longuet-Higgins, M. S. (1969), A nonlinear mechanism for the generation of sea waves, Proc. Roy. Soc. A 311:371-389. McLeish, W. , R. A. Berles, W. H. Everard, and G. E. Putland (1971), The SAIL 6-m wind-water tunnel facility, NOAA Tech. Memo. ERL AOML-12, Environmental Research Labs., Miami, Fla. , 19 pp. Mellor, G. L. , and H. J. Herring (1969), Two methods of calculating turbulent boundary layer behavior based on numerical solutions of the equations of motion, Fig. 3a, in Proceedings Computation of Turbulent Boundary Layers — 1968 AFOSR-IFP Stanford Conference, Vol. 1., eds. S. J. Kline, M. V. Morkovin, G. Sovran, and D. J. Cockrell, Thermosciences Div. , Dept. of Mechanical Engineering, Stanford Univ., Palo Alto, Calif., pp. 331-345. 997 Schraub, F. A., S. J. Kline, J. Henry, P. W. Runstadler, Jr., and A. Littell (1964), Use of hydrogen bubbles for quantitative de- termination of time dependent velocity fields in low speed water flows, Report MD-10, Div. of Engineering Mechanics, Stanford Univ., Palo Alto, Calif., 66 pp. Shemdin, 0. H. (1972), Wind generated current and phase speed of wind waves, J. Phys. Oceanogr. , 21:411-419. 998 Reprinted from: No. 3, 516-518. Journal of Physical Oceanography 5 65 Measurements of Wind-Driven Flow Profiles in the Top Millimeter of Water William McLeish and Gerald E. Putland Atlantic Oceanographic and Meteorological Laboratories, NOAA, Miami, Fla. 33149 (Manuscript received 13 January 1975, in revised form 4 February 1975) ABSTRACT Shapes of mean water velocity profiles measured with microscopic bubble tracers in developing laminar flows are recognizably different from those in a turbulent flow. A previously deduced viscous sublayer occurs at the surface, although it is thinner than an analogous sublayer computed for a solid boundary. The differing thickness leads in part to decreased surface temperatures at slicks. 1. Sublayers at a water surface A cold surface layer is often found at the ocean surface, almost always so at night. McAlister (1964) suggested that within this boundary layer there must be a conductive sublayer with a nearly linear temperature profile. His measurements with a specialized infrared radiometer at low wind speed indicated a conductive heat flow from the sea in approximate agreement with the total heat flow estimated from measurements in the air. Later, Timofeev (1966) asserted that the boundary layer exists only when the water surface is relatively smooth and disappears at higher wind speeds. However, further radiometric measurements by Mc- Alister and McLeish (1969) demonstrated the existence of a conductive sublayer in laboratory experiments at moderate air speeds, and McAlister et al. (1971) measured the total heat flow from the ocean by this technique. Since the turbulent water motions advecting heat also transport momentum, the existence of a conductive sublayer also implies a viscous sublayer. The viscous sublayer in water should be about twice as thick as the conductive sublayer (Wu, 1971). Informa- tion on the thickness of these sublayers at a water surface is necessary to measure ocean temperature and heat flow radiometrically and to predict exchanges of various materials between the atmosphere and the ocean. Measurements showing boundary layer thickness at a water surface have not been found among previous studies. The radiometric measurements in a conductive sublayer indicate only a minimum thickness, not its total thickness. Velocity profile measurements at 0.1 cm depth intervals by McAlister and McLeish (1969) had insufficient resolution to show a linear surface region. Laboratory velocity profiles by Wu (1968) appeared linear but fitted calculated logarithmic profiles satisfying the experimental conditions and so do not indicate viscous sublayers. Accepting the general assumption that boundary layers over a solid surface and the air-sea interface are similar, Wu (1971) calculated the thicknesses of conductive and viscous sublayers as a function of wind velocity. Katsaros and Businger (1973) made further calculations concerning the determination of heat flow from the ocean using radiometric measurements. In both papers, the authors emphasized the need for measurements to determine the flow structure near the surface. The present study furnished mean water velocity profiles very near the surface in both laminar and turbulent laboratory flows. The viscous sublayer at a free surface is considerably thinner than at a solid boundary. 2. Velocity profiles in the upper water Profiles of velocity in the top millimeter of water in a wind-water tunnel have been measured from cine photographs of clouds of microscopic hydrogen bubble tracers. Experimental methods were described by McLeish and Putland (1975). The buoyant rise of the bubbles was small, and the bubbles were photographed shortly after release. By using neutrally buoyant markers, the bias errors of floating tracers in an irregular flow described by McLeish (1968) were avoided. Laminar and turbulent regions of the water surface could be distinguished with the bubble patterns, by the mixing of dye streaks, and, it has been found, through the shapes of the waves. The present depth interval of 0.005 cm between measurements provides several velocity readings within the depth region in which a linear viscous sublayer might be expected. Fig. 1 shows downwind U and crosswind V mean water velocity profiles in the upstream region beneath 999 July 1975 WILLIAM McLEISH AND GERALD E. PUTLAND 517 Z.CM V.CM/SEC Fig. 1. Water velocity profile in an essentially laminar flow, + downwind and X crosswind. The fitted line is an exponential curve. an air flow with a midstream velocity of 900 cm s_1. The water flow there was essentially laminar but beginning a transition to turbulent flow. An exponential curve fitted the U readings well, although such a curve is not necessarily characteristic of a developing laminar boundary layer. In fact, Kunishi (1963) reported transient laminar velocity profiles with a different exponential shape. Instead, the fitted curve illustrates the precision of the individual measurements. In particular, an apparently erroneous mean velocity defect is seen in the upper 2-3 readings that might have been caused by air flow interference of equipment mounted asymmetrically in the air above the measure- ment area. The fitted curve indicates a surface shear from which a stress of 2.0 dyn cm-2 was calculated. Calculations based on estimated wave parameters indicated that the wave drift current was small. The crosswind velocity profile, although offset, shows generally little variation with depth. The velocity fluctuations in Fig. 1 were rapid, and the rms value at a depth was roughly 15% of the mean value there. The depth-averaged variance, reduced by the number of photographs, was similar to the variance of the differences between the mean values and the fitted curve. These departures, then, could be random fluctuations. The correlation between fluctuations at a depth and at the surface decreased to a value of 0.5 at a depth of only 0.035 cm and remained near 0.3-0.4 below. Since the vertical correlation of wave motions should remain unity, the rapid decrease in correlation with depth suggests that wave orbital motions were largely removed by the Y6 s averaging between photographs. Furthermore, the kurtosis of the frequency distribution of surface velocity fluctuations was —0.2, as compared with zero with a normal distribution (turbulence) and —1.2 for a uni- form distribution (approximately, waves). Rough estimates indicate that turbulent fluctuations in the air flow could cause the very shallow water velocity fluctuations. The velocity profiles in Fig. 2 represent fully laminar flow with an air speed of 550 cm s_1. The shape of the profiles were similar to those in Fig. 1, although the surface shear indicated a stress of only 1.0 dyn cm-2. Fig. 3 was obtained at the same air speed as Fig. 2 but downstream where the water flow was fully tur- bulent. The overall shear in the turbulent boundary layer was considerably less, and the shear zone was thinner than in the laminar flow case. Although the water surface was much rougher than upstream, the indicated stress was less, 0.7 dyn cm-2. The linear portion of the profile at the surface represents a viscous sublayer. This was thinner than the computed profile at a solid boundary with the same stress, as represented in Fig. 3 by a straight line and a departing curved line. The curved line was derived from nondimensional Z.CM (t/'(2)f/'(0)/{[c/'(Z)]2[c/'(0)]2}i V, CM/SEC Fig. 2. Water velocity profile in a fully laminar flow, + down- wind and X crosswind. The fitted line is an exponential curve. 1000 518 JOURNAL OF PHYSICAL OCEANOGRAPHY Volume 5 U, CM/SEC 1.0. Z.CM Y+ 18 8 96 7.2 4.8 2.4 0.0 U+ Fig. 3. Water velocity profile in a turbulent flow. The straight line fits the surface slope, and the curved departing line follows the mean profile at a solid boundary. values by Kline et al. (1967) and by Mellor and Herring (1969, see Fig. 3a). 3. Discussion The linear surface segment in Fig. 3 is believed to represent the first direct measurement of a viscous sublayer at a water surface. Although thinner than at a solid boundary, the viscous sublayer supports the previous evidence for a conductive sublayer sufficiently thick for radiometric heat flow measurements with moderate wind speeds. Ocean slicks commonly are cooler than the adjacent water surface and, at certain scales and times, represent the dominant surface temperature fluctuations at sea (McLeish, 1970). This effect can be attributed in part to the horizontal rigidity of a slick. A slick acts on the water turbulence below much as does a solid boundary. The conductive sublayer beneath a slick, then, is thicker than elsewhere and, in hindering the normal loss of heat from the sea, gives an increased surface temperature depression. Reduced air stress on the smooth surface at a slick also gives increased sublayer thickness. The thickness of the viscous sublayer at a free surface is determined in part by the horizontal water motions occurring there and can differ from that at a solid boundary. The present observations indicate that previous estimates of sublayer thickness based on an analogy to a layer at a solid boundary are of the correct order of magnitude but could be improved substantially through direct profile measurements. Acknowledgments. We are indebted to Mr. F. Ostapoff, Dr. C. Thacker and Ms. S. Worthem for comments on the manuscript. REFERENCES Katsaros, K. B., and J. A. Businger, 1973: Comments on the determination of the total heat flux from the sea with a two-wavelength radiometer system as developed by Mc- Alister. /. Geophys. Res., 78, 1964-1970. Kline, S. J., W. C. Reynolds, F. A. Schraub and W. C. Run- stadler, 1967: The structure of turbulent boundary layers. /. Fluid Mech., 30, 741-773. Kunishi, H., 1963: An experimental study on the generation and growth of wind waves. Disaster Prev. Res. Inst. Kyoto Univ. Bull., No. 61. McAlister, E. D., 1964: Infrared optical techniques applied to oceanography, 1, Measurement of total heat flow from the sea surface. Appl. Opt., 3, 609-612. , and W. McLeish, 1969: Heat transfer in the top millimeter of the ocean. /. Geophys. Res., 74, 3408-3414. , , and E. A. Corduan, 1971 : Airborne measurements of the total heat flux from the sea during BOMEX. /. Geophys. Res., 76, 4172-4180. McLeish, W., 1968: On the mechanisms of wind-slick generation. Deep Sea Res., 15, 461^69. , 1970: Spatial spectra of ocean surface temperature. /. Geophys. Res., 75, 6872-6877. , and G. E. Putland, 1975 : The initial water circulation and waves induced by an air flow. NOAA Tech. Rept. ERL 316- AOML 16. [Available from Superintendent of Documents, Washington, D. C] Mellor, G. L., and H. J. Herring, 1969 : Two methods of calculating turbulent boundary layer behavior based on numerical solu- tions of the equations of motion. Proc. Computation of Turbulent Boundary Layers — 1968 AFOSR-IFP Stanford Conference, S. J. Kline, M. V. Morkovin, G. Sovran and D. J. Cockrell, Eds., Thermosci. Div., Dept. Mech. Eng., Stanford University, 331-345. Timofeev, Yu. M., 1966: Thermal sounding of surface water layers by means of thermal radiation. Izv. Atmos. Oceanic Phys., 2, 772-774. Wu, J., 1968: Laboratory studies of wind-wave interactions. /. Fluid Meek., 34, 91-112. , 1971 : An estimation of oceanic thermal-sublayer thickness. /. Phys. Oceanogr., 1, 284-286. 1001 gg Reprinted from: Preliminary Scientific Results of the GARP Atlantic Tropical Experiment, prepared by the International Scientific and Management Group (ISMG) of the World Mete- orological Organization, Volume II, GATE Report No. 14, 392-397. - 392 - Preliminary Analysis of Ocean Internal Wave Observations by Acoustic Soundings Feodor Ostapoff, John Proni , and Ron Sellers Atlantic Oceanographic and Meteorlogical Laboratories NOAA, Miami, Florida INTRODUCTION One of the major objectives of the Oceanographic Sub-program (GATE Report No. 8, 1974) is to "concentrate on the physics of small scale processes' including studies of surface waves, internal waves, oqeanic fronts, and the development of the mixed layer." In order to achieve this goal, it is necessary to observe perturbations at the bottom of the mixed-layer and. in the upper thermocline (R. Pollard, 1974; Kraus and Ostapoff, 1974). One of the first attempts to accomplish this goal in the open ocean was made during Phase III of GATE in the C-Scale array on board the R. V. COLUMBUS ISELIN. In addition to direct measurements by moored current meter systems and towed CTD/ STD sensors, indirect acoustic sonsine- tschnioues "ays been applied. This preliminary note gives a brief description of the sensing technique and the first results of an intercomparison experiment during GATE between the acoustic record and a corresponding STD time series. An example will also be presented illustrating the application of the data to the analysis of the observations obtained during the joint roving operations between DISCOVERY and ISELIN, as well as QUADRA, DISCOVERY, and ISELIN. 1002 - 393 TECHNIQUE An acoustic transducer mounted in a towing fish was operated in such a mode that simultaneous STD profiles or time series as well as towed STD profiles were feasible. The transducer operates at 20 KHz with a pulse duration of 2.2 msec, and a repetition rate of .75 sec. (Proni and Apel, 1974). The beam width is approximately 120 by 18° (to the half power points) . Figure 1 shows- an example of the sonar record obtained on September 19 showing wave-like motion at the top of the thermo- cline. Heavy slanted lines are sonar reflections from the STD which was profiling down to 100 meters. In a sense, the acoustic technique as .applied here to ocean phenomena is quite similar to the acoustic echo sounder techniques used in studies of the atmospheric boundary layer (Little, 1969; Frisch and Clifford, 1974). It is important to realize that this technique does not require biological scatterers to be present such as in the deep-scattering layer (DSL) ; but that density fluctuations in a layered fluid are sufficient to produce a higher acoustic scattering cross-section (Proni and Apel, 1974; Little, 1969; Apel, Proni, Byrne, and Sellers, 1974) . PRELIMINARY RESULTS Sections of the acoustic record (as seen in Figure 1) have been analyzed for the time period during which the STD instrument was held at a constant depth of 45 meters. This level corresponds to the first solid scattering return on the acoustic record. The acoustic record was then scaled by eye and digitized. Figure 2 shows the undulations of the scattering layer (dashed curve) at about 45 meters for a 45-minute period. The solid line represents the temperature record from the STD. It should be noted that at this level the density field is more or less determined by temperature, because at this level the salinity goes through a broad maximum. It is evident from Figure 2 that the temperature fluctuations are approximately 180° out of phase with the fluctuation of the acoustic scattering layer as it should be if these fluctu- ations represent internal waves. As the layer is lifted, the temperature decreases and vice versa. Moreover, the double amplitude for the waves on the right hand side of the top 1003 _ 394 - curves can be determined from the vertical temperature dis- tribution, i.e. 6T/AT ~ 11.5 meters AZ " where 6T = 3.2°C from the time series and AT/AZ = 0.28°r from the STD cast. This compares to an estimated double ampli- tude from the acoustic record of - 12.5m. The lower curves in Figure 2 simply indicate the phase relation- ships, since the acoustic record at 45m depth was difficult to read and another scattering layer at 65 meters was chosen for this analysis. This is justified because for periods of several minutes the acoustic record shows that these layers move up and down in phase. However, a comparison of the amplitudes is not possible with this information. Spectral analysis shows that in the upper time series in Figure 2 most energy lies in a period band of 9-17 minutes, with a smaller peak at 17 minutes, while in the lower record around 7-8 minutes. The R. V. COLUMBUS ISELIN recorded some 160 hours of acoustic data in various modes: 1. While steaming jointly with DISCOVERY and QUADRA during the batfish operation. 1. While in free drift mode. 3. While hovering over a cyclesonde buoy. It is intended to digitize the acoustic data which also were recorded on magnetic tape and to process the data objectively by computers. This analysis will be carried out jointly with our British and Canadian colleagues in the study of the upper ocean dynamics. 1004 - 395 - REFERENCES Apel, J. R. , J. R. Proni, M. H. Byrne, and R. L. Sellers, 1974: Near Simultaneous Observations of Intermittent Internal Waves from Ship and Spacecraft. (Submitted for publication in Geoph. Res. Letters) Frisch, A. S. and S. F. Clifford, 1974: A Study of Convection Capped by a Stable Layer Using Doppler Radar and Acoustic Echo Sounders. JAS, Vol. 31, No. 6, pp. 1622-1628. Kraus, E. B. and F. Ostapoff, 1974: AS-II: Mesoscale Phenomena in the Boundary Layers of the Ocean and the Marine Atmospheres. EOS, Trans. AGU, Vol. 55, No. 8, pp. 763-764. Little, C. G. , 1964: Acoustic Methods for the Remote Probing of the Lower Atmosphere. Proc . IEEE, Vol. 57, No. 4, pp. 571-578. Pollard, R. , 1974: Wind-driven Deepening of the Oceanic Mixing Layer. Presented at the IAMAP-IAPSO combined First Special Assemblies, Melbourne, Australia, 14-25 January 1974. Proni, J. R. and J. R. Apel, 1974: On the Use of High-frequency Acoustics for the Study of Internal Waves and Micro- structures. (Accepted for publication in JGR) 1005 T3 O tfl O U ■H > a) 13 01 C •H C D O W o •H -P 03 o u (0 1 >1 > 4J 1) c h (0 • w w -* < Q U En O OJ £1 n -P a ^ w o (C X! C Cn-P c cj •H '-I 3 o T3 -H 01 5-1 O -P 0) I 2 ai c w CPrH O O 0) O . .h w (3 a) f0 M (3 tu > 3 H o ■P -P • -P C in 3 O U — C X! V U ^ t0 O Q) U A 0) -* u 0> a e • QJ to E-t M Q) P 0 e CQ in • 1X1 u x: d) -p -p it) a) (JilH c o •H M (U +J 4J to u -P -r cue r-H in U "* ■H 1 P co 3 O u ro U) .H U ft XI u 3 -P T3 (0 C (0 (/} O) •H M d) oo i. o 00 c 0) oo ro 2 o t. > 00 00 ro D. O .c oo 4J s- o O S- CD 1018 07i Figure 2. Growth of wave spectra for offshore wind conditions during JONSWAP-69. Fetch Three shape parameters a , o increases from Station 5 through Station 11 and the inset, y is simply the ratio of the energy at f to that which would be predicted by the Pierson-Moskowi tz (1964) form of the spectrum. b' and y> and the scale parameters f and a suggested by Hasselmann are shown in 1019 t — rn — i i ii it] i ■n i i- 1 i i m| iii r i mm T ~, fmU10 Tm" g "■ E 1 _- — .27 fm=1.9X — - 0 O oH o % X oo m ICH U4 E=l.2xlO"7xiJ 8° o nQ o "0 o • *r ° • - /•• - *LJ lO"3 o — - o o o o o • N.SEA x BaW UPWIND ■ B6W DOWNWIND o JAN 27 -» m.4 « . „ 1 1 1 ! I 1 1 1 1 1 1 , j_j i i i i i ii I i iii i i i t i JL I0; I03 10' I05 x = : xg U,2o ■ tire 3. Dimensionless parameters E and f versus dimensionless fetch X for high wind- r m speeds (15 to 25 m/sec) as determined by aircraft experiments in the North Sea and off Cape Fear, N. C, on 27 January (Ross and Cardone, 1974, and Barnett and Wilkerson, 1969). 1020 IO-2r 0.3 0.4 0.5 0.6 0.8 f m Figure 4. The behavior of E and f during growth. A sudden change of windspeed by a factor of 1.5 results in a departure from the mean followed by readjustment to mean condition along the cur/ed lines indicated. Nondimensional fetch relation- ships X = E = yf 2 f°r particular f are included. The upper and lower bound- U 1 o m aries (C = 10 and C = 10" ) represent the limits of momentum hain't tr.it fcrri'd to the w.ivo r.{«,f tru'n. 1021 a fmB"sfmU 0.03 1 0.005 T~i — ft- 0.4 0.5 0.60.70.80.9 f rr* ?^re 5. The Phillips parameter a versus f . Corresponding fetches E, are indicated by the tick marks. The behavior of a during readjustment to a change in local wind (U) conditions is along the curved lines. 1022 T? o o X h- Z> M < (Z O U. Q UJ f- O UJ a: a:* o o < a' 1- h- -> a: UJ LE UJ o a> a) m m H rn o O 2 h- UJ UJ QC O) t- \- IaJ • — ■ CO _j _J CD UJ _i _J o LJ ?: Ul UJ UJ 3 t- H O UJ -^ < < ^s CO CO < co ~ X X X Q X X X Z b° H° K-° £ O D ii ii < © I I p 4 k & # I I 6 6 < < cr or o O o o (M i o O o o o o CVJ (qp)HH°^> («HHDi LO (r/uj) °'n 0 in CVJ u 0 ■- St CM O o — . ro c CVJ _^. c 0 LU •r- cvJ Q CJ CVJ :;:* 3 l- l~3 h- r- < r i 4- 0 C CVJ Q. o ro o ■ • CVJ 00 0 0! an 5- 0 CO o o o 3 cr 1023 Tk 1 F 1 ro :1k I « n i/i ** i ' jQ *§ 7 flj I. w'- >5 • . \ ^ oo 1 4- (=5 O ^-^ r u .V W £T) IS) l> o Q. w-: E 1—1 o o 1 OJ CO 1 LJ Q-J u 3 CD 5- 1024 (>Ua± o o ° l)02n (qp) o - 92=8981 - 61=8981 -K)=898l 6i?=Z98l t>£=2.93l 81 = 2.981 2=8981 91=8981 00=8981 -|9i>=ZG0i 02=2.981 91=2.981 "3 > < CD C (C u 5- S- 3 ro -a i -a c re oo (>lo)9l 1025 CD i/> CO u • sr r*~ 00 "O o N- n3 t- CL (.) cr, "o r— CD +j 0> >, n3 1- O (T3 •!- ?->>- 3 -O m rv fO -r- o < O) *2! +-> < <+- o •r- C >> O r— x: (T< 1/5 13 CD S_ i— fO o >-, •r- -O CD 3 O 4-> i 1/1 o i- s- O O) CD "O *-> c CU 3 £E a> cri a r— 03 I -J3 U- UO o QJ O". o O to fO o 1026 >- a: UJ -J -J < en o CD II CD X X o < i i "D i i o> ? / / i «k % CO >0 tfSna i o 4 b in \ lO t? ce 0\ \ UJ t? c/> X | GQ o1 i O b ^ to M CD x X a i i C d \ * 1 1 J—L .J-L™~~«J_ o e o '~ CO C vH 4-J vl T3 ,74 C m 3 m «H U CO >o m «-j 2 Dl c o LO 1/1 01 in re >«-« UJ a !£ — — 4T)HHd1 ° o o o OfO cvj -« 0) o m ro LO lO (P/Ui)0ln 1027 CO -z. o "T o (£> ii 0 m o m ii r 1 1 — 91 iij CO CQ o LlI -J CD < LU co < 0 GD © 0 © — >- < UJ _l LU CJ> o 2 — ._J UJ UJ O LJ o Q < |— CD "3 < CO d O (X) 2 2 0 o □ — 0 LU O o 0 O - — J,. J | [ | 0 m in ro O o > to 3 C in ■"3 0\ o -D Ol) -o in C/l at > u «/> o JH s- 03 T3 if) i CO O I- tr Ld CO (jj CQ o ro 2 ~D "! lO roro LlJLiJ "D-D CD" CO D O < ( > O CO o < roro Lo^y o 33< ^^-D-O U- LO~" St" < — C\J CC a 4 © o o " o o o o o o H rf1 3 0! ra S- a> Cl r 11 CD + • c GJ .-. j rU (J -^ O 5- .13 T-l .- (qp) °-o • , HHn (Ho) 1 1029 320-i 280 240 ( ) Ixl CO i 200 h- Li_ 160 Zj 1 OC\ z 1 CSJ a co < 80 40- N0. ATLANTIC Hs= 15.22 Ft. AVA — Hs= 14.87 Ft. 0 T 0 .10 .20 .30 FREQUENCY (HZ) A u UJ 0') D Z o CO < NORTH SEA AVA Hs = 16.85 Ft. .10 .20 .30 FREQUENCY (HZ) B o UJ 00 I I- o CO < NORTH SEA 0 .10 .20 FREQUENCY (HZ) j-ire 13. Have spectra measured during the aircraft penetration of Hurricane Ava along with North Sea spectra for fetch-limited 25 m/sec winds. 1030 1 i / I lt- 1 o o o o o c/ o i \ \ o o / I o o ro M o/ o r i t uV r / Ik \ ■O > fO i* o o v. -H \ a 1 \ \ \ oV o orii j,m . . -J OO Xo • I _L_ • ;£Jro . *». - • ^ * CD ^ ro to Wx' .-^ '•— — _ .— — — "''' * CTlr— C T- O S *"- rC re '_> re CJ -t-> C ro <0 X3 O •i — ••-> s- 4- S- ro r \ S- X \ \ U S- i. •r- O V \ \ \ TO 4- o c S- O \ "V- \ — \ ■o > QJ < \ +-> UJ \ 3 ■.- -J V S. _J \ +-> — c c 2> < I < > » o ■- o < V \ \ \ \ I UJ 2: > C < < < o o c; *1 a: 7) ro iO "U u :r. X ■** 1 i. > 2 OJ1 a i .~ C •/- ■r- l-J 3 ro • CD O) / O i- s ra 3 T- oo 3 •V 1031 1U 1 \ 1 1 r 1 1 1 11 1 1 1 1 1 1 1 1 1 1 1 ] 1 III 1 ! 1 1 rr Jj p Eg2 h" u4 E-1.0xl0"4 R - • AVA - 10-3 - 0 WEST DELTA x SOUTH PASS j CAMILLE • • • X . yO cXoX0 °X 0 OO Q. - x . • _ X _ - X - 10 r- - E 4— 7 _fmU±o ■m" g X X rst ^—2 5 fm=i.6R- • X x 8x • x • v X 0 d • 0 a ox • 9 0 0 A 0 • - 10"1 1 111 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 111 1 1 1 1 _u_ 10 i 10 R = rg 10 10 u 10 igure 15. Behavior of nondimensional energy, E, and peak frequency, f , versus radial distance from eye. R « 102 constitutes the region of maximum winds. 1032 1 1 1 1 1 — n - _ - _ - - o o _ (3 o o° o □ D — O < — - 8 ~! - o - < « i i CO -z. - o — r- < - > " en UJ " GO _ o ro rO ro X X o b LU r- r- N- H IxJ LjJ LU • ~ _| ^ z 2 - _j 3 ID 3 LJ ~3 "0 ~D o \- m CD ■MM m < ii 01 a O < CD J- 1 1, 1 1 ! Q i OJ I O o lO I g i lO I o OJ I lO CM I o ro i LO ro i c 0) ir, i— Cl; +j T3 re IT OJ > cv fO I 2 _J 00 4- o a JZ a; ■*■' cr rO <*. fJ a 01 14- l/l 3< 4- 1/1 >- >^ o; (J > c ty o ■^ D GJ S- * 4- i- OJ ^ +-> rc +-> Cj ro Cl a i/i 1 — _^' rO u C rO c n •I— to s_ c ra Oi T3 — ro • r- a; ■o U3 CU 5- n HH, (qp) -o 1033 gg Reprinted from: Journal of Geophysical Research No. 27, 3841-3847. Dropped Horizontal Coherence Based on Temperature Profiles in the Upper Thermocline Gilbert R. Stegen' Department of Geological and Geophysical Science. Princeton University, Princeton. New Jersey 08540 Kirk Bryan and Joann L. Held Geophysical Fluid Dynamics Laboratory. NOAA. Princeton University. Princeton. New Jersey 08540 Feodor Ostapoff Atlantic Oceanographic and Meteorological Laboratories. NOAA. Virginia Key. Florida S3 14V A series of 66 temperature profiles taken in the open ocean 350 km north of Puerto Rico are analyzed to deter mine the displacement spectrum and the dropped horizontal coherence. The results are not inconsis- tent with the Garrett and Munk model- One interesting exception is a tendency for the horizontal coherence not to fall off monotonically with decreasing vertical scales. Introduction For years, oceanographers have been aware of small-scale high-frequency variability in the ocean thermocline This variability gives rise to a major source of sampling error in the determination of large-scale features of the thermocline from standard hydrographic data. Only recently have experimental techniques allowed direct measurements of the fine structure of temperature, salinity, and velocity in the main thermocline. Vertical profiles are now routinely obtained by detailed CTD (conductivity, temperature, and depth) measurements and free falling devices. Extensive measurements along horizontal tracks have also been obtained by towed thermistors. The most important source of data has been long time series in velocity and temperature obtained from moored instruments, sometimes located in small arrays to allow measurements of spatial coherence at different frequencies. The time series measurements of velocity in the mam ther- mocline carried out by the Woods Hole Oceanographic Institution [Fofonoff. 1969] showed that a great deal ol the fluctuating energy in the thermocline is contained in the fre- quency range of mertial gravity waves. N > w > f. where u> is the observed frequency and N and/are the Brunt-Vaisala and inertial frequencies, respectively. Using the W KB approxima- tion and the theory of internal inertial gravity waves. Fofonoff [1969] and Webster [1968] have been able to explain several important features of the current meter statistics. These ideas have been extended much further in an important paper by Garrett and Munk [1972] in which a model of the displacement and velocity spectra in the internal inertial gravity wave range is proposed. The WKB approximation is relaxed, and l he model depends only on the total depth of the thermocline, the local inertia! frequency, and the local range of Brunt-Vaisala frequency. The model makes use of relationships expected from small amplitude theory of internal inertial gravity waves but is calibrated against existing observations. Whether all details of this very interesting synthesis are verified by future measurements is perhaps less important than the tremendous ' Now at Flow Research. Inc.. Kent. Washington 98031. Copyright © 1975 by the American Geophysical Union. contribution of the model in providing a guide to other in- vestigators in analyzing and scaling diverse observations of variability in the ocean thermocline One prediction of the Garrett and Munk [1975] revised model that has not yet been tested against observations is the DHC. or dropped horizontal coherence. This quantity is a measure of the horizontal persistence of the velocity, or ver- tical displacement, of surfaces of constant density. The present study describes a set of measurements that provide a preliminary measure of the DHC in the tropical North Atlan- tic, approximately 350 km north of San Juan, Puerto Rico In this region the temperature structure is well-defined with a mixed layer down to approximately 60 m and a strong de- crease in temperature down to 500 m. At 500 m there is a layer of nearly uniform temperature Below this, there is another layer of rapidly decreasing temperature in the main ther- mocline In the present study we are only concerned with temperature profiles taken in the upper layers from the surface down to 450 m. The profiles were taken with expendable bathythermograph probes (XBT's) launched from a moving ship, the NOAA ship Discoverer The XBT's were launched every 2 mm as the ship maneuvered in a basic X pattern at about 6 km/hr. the minimum required for control. The track of the ship is shown in Figure I . All the XBT launches are in- dicated in this plot. However, only the solid circles represent positions of profiles actually used in this study. The ship navigated in relation to a Nomad (AN/SMT-I) buoy which was drogued at 30 m. Fixes of the buoy were taken from the bridge every 5 mm during the experiment. The fixes were com- bined with the drift pattern of the buoy to generate the ab- solute track of the ship shown in Figure I. In order to use the data from the XBT records to compute DHC it is necessary to subtract out the mean temperature profile. The next step is to divide the temperature deviation in each profile by the vertical gradient of the horizontally aver- aged temperature. This converts the temperature anomalies into displacement anomalies. In comparing the DHC computed from these anomalies with the Garrett and Munk [1975] predic- tions it must be recognized that the measurements are not truly synoptic, since there is a finite time interval between launches 3841 1034 3842 Stegen et al.: Temperatire Prohiis in Uppi r Tiikrmoci.ini: Fig. I Course ol the NOAA ship Discocerer during the experi- ment on October 12. 1971 Each open circle marks an XBT drop \ solid circle indicates that the temperature profile was used in this stud\ . Some loss of coherence due to this lime lag must be expected Ultimately, completely synoptic measurements of conductivi- ty, temperature, and depth must be carried out lor different separation distances, but it will require a major effort in- volving a joint operation of at least two research vessels. Experiment Site, The measurements were made on October 12. 1971. at 22C46'N and 67°03'W. The average temperature profile from all the XBT casts used in our analysis is shown in Figure la. A persistent 'bump' in the profile is indicated al about 160 m. Its origin is not clear, but it is probably the result ol an earlier water mass intrusion. Such features have been lound by several investigators [e.g.. Mazeika. 1974] in this part ol the tropical Atlantic and appear to be similar to the. steps lound by Tail and Howe [J%8] associated with the Mediterranean out- flow. The present study is concerned w ith those features ol the temperature structure which might be expected to be found al manv sites, so that the persistent intrusions are a source of contamination which we attempt lo remove Irom the profiles bv subtracting the mean shown in figure 2a Irom each in- dividual profile. Estimates of the Brunl-Vaisalii frequency arc- obtained from an STD cast made at the site a lew hours before the main experiment. The data from the STD cast arc first con- verted to /V2 and then smoothed in the vertical. The profile ol N shown in Figure lb is obtained from the smoothed curve ol ,V\ (It should be noted that the alternate procedure ol lirsl calculating a profile of N and then smoothing gives quite different results!) Mean values of A lor the upper layers can also be derived by using salinity data obtained Irom hvdrographic stations made during the IGY [Fuglister. I96()| in the same area and the temperature profile shown in Figure la. The results shown in Table I show lair agreement between the two sets of data. X BT data The XBT probes were the standard Sippican model T-4 The absolute temperature accuracy ol the XBT system is ±0.2°C. the errors being equally divided between probe and recorder. A more important source of error is in the accuracy of the probe depth, which is specified as ±2S. The depth errors are of two kinds. Since the XBTsystem measures BRUNI.VAISAIA FREQUENCY |cph) 20 40 60 80 "I I H^TZ r 1 I i 1 450 Fig. 2. (a) Average temperature profile Temperature profiles from 74 XBT casis were used. (/>) Brunl-Vaisula Irequency representative of the area at the time of the experiment It shows the lrequenc\ alter nine-point smoothing was applied to STD data taken at the site The vertical bars indicate the range ol frequencies included in each ol ihe three ^ases considered in this study: A. .V = 6.2189 cph lor depth range 75-203 m. B. \ = 3.4547 cph lor 150-406 m: C. /V = 4.6212 cph for 69-420 m. 1035 Stegen et al.: Temperature Profiles in Upper Thermocune 3843 TABLE 1. Comparison of the Average Values of N 203 m "STD 6.22 7.17 3.45 3.50 Values were obtained from a single STD cast and from a combination of the salinity data obtained during the IGY and the temperature data given in Figure 2a. Units are cph. depth relative to the point at which the probe enters the water, the absolute depth below an average surface will have a ran- dom error introduced by surface waves (wave height is 1 .5 m in the present study). Another source of error is due to deviations from the calibrated fall velocity of the probe. This second source gives a relative error which increases as the total depth increases. These two sources of error will be discussed further in connection with the method of data reduction. The standard output from the XBT system is an analog trace on pressure sensitive paper Conversion of these traces to digital data for computer analysis poses a formidable challenge In the present study the analog traces were photographed with 'litho' film to produce a high-contrast 4 X 5 in negative This entire negative was then digitized with a scanning microdensitometer which had a 1000 : I range in sen- sitivity to film density and gave a matrix of 1000 X 800 values over the negative The digitized density values were written directly onto a nine track 800 bits/in. IBM compatible magnetic tape. A computer search of the tape identified all points above a threshold intensity and delineated the trace. The data were digitized in this manner rather than by hand or by some other process in order to obtain better resolution, reduce the noise from the digitizing process, and save time Unfortunately, not all of these objectives were realized The XBT recording paper grid scale was almost the same intensity as the trace, so that it was often not possible to differentiate between the trace and the grid Thus some original records which were otherwise perfectly good could not be included in the final data set. Using fiducial marks photographed with the traces facilitated calibration of the temperature and depth scales and correction for the nonlineanties in the temperature and depth scales of the Sippican recorder The data were interpolated to equal depth increments of 0 5 m. Finally, the computer com- patible data set was plotted on an 8- x 10-in. scale and com- pared to the original data set Traces whose temperatures did not agree to within ±0.05°C everywhere were discarded The sorted records were averaged to form the average temperature-depth curve shown in Figure la. In view ol the uncertainties in absolute depth an additional quality control check was applied. For each individual profile the following integrals were calculated. S(Az) = ^ E lT(z.) - T(z + Az)] and '(Az) = Jj £ [fto T(z„ + Az)]2 where T(z) is the average profile in Figure la Figure 3 shows a typical plot of o-(Az) as a function of the vertical displacement. In most cases the displacement Az at which S(Az) and , is the dimensionless inertial frequency. With this notation the energy spectrum specified by GM75 [Garrett and Munk, 1975] is for the range «,/ « ar~'«,g«/M(P/fl*) E{a, p) = a*, 2 2 , T^T- 0*(n a + u, p ) 0 < a < 0(1 - u>,7*2]"2 (1) (2) where £ is a dimensionless constant and 0* is a scale vertical wave number. Both constants are chosen by Garrett and Munk [1975] to fit the data. Here ,4(0/0*) is a shape function. A = 1.5(1 + 0/0*) (3) Let Z2 be the fraction of the total energy which is potential energy, where (4) art -+- a>, p Isotropy in the horizontal plane is assumed. The displacement spectrum is given as Fr(0) = // Z2E(a, 0) do (5) For any two points separated by some distance X the expres- sion for the cross spectrum is (e.g., Phillips, 1966. pp. 74-78] C(X, p) Z7E(a, 0) • exp (iaX cos <£) d da (6) Since i0" exp {iaX cos )d = 2wJ0(aX), the coherence of ver- tical scales may be expressed as R(X, p) = [ Z*-E(a, p-)Jn(aX) da [ (7) Z'E(a, 0) da It is interesting to note that shape function ,4(0/0*) is indepen- dent of « and cancels out in the formula for the DHC. Since the difference between GM72 and GM75 is associated with ,4(0/0*), both models would predict the same curve lor the DHC. Making use of ( I) and (4) to eliminate E{a, 0) and Z', respectively, the expression for DHC is R(X, 0) = GifiutX/n) Let y = an/tlw, and G(0 w,X/n) = - [ y\l + y2y* Jn(0u,yX/n) dy IT Jn (X) (4) In the next section we will compare the predictions of F; given in (5) and DHC given in (8) and (9) with our data set. Analysis of the Data The displacement spectra obtained by combining all 66 profiles are shown in Figure 4. Three different depth ranges are shown. The abscissa is the log of the vertical wave number normalized by n. No band smoothing or tapering of the records is involved in the spectral estimates shown in Figure 4. Sixty-six degrees of freedom imply an 80% confi- dence limit of ±40% [Blackman and Tukey. 1958]. The spectra are shown with the predicted curve for the GM75 Fig. 4. Displacement spectra of the data in this study compared to CjM7.v F and B are normalized h\ n. the average Brunt-Vaisala frequency for the section under consideration, al (a) 75 m < — z < 203 m. (b) 150 m < z < 406 m, and (r) 69 m < - z < 420 m. A total of 66 XBT casts were used in these calculations. The 80"; confidence limits are wiihin a range of ±40%. 1037 Stegen bt al.: Temperature Profiles in Upper Thermocline 3845 E 10' = to' 1 r MIUARD (19721 Fig. 5. Displacement spectra from Garrett and Munk [ 1 975). Millard [1972], and this study. Comparison of the predicted curve (GM75) and experimental results. model. In all. cases the spectra lie below the predicted curve, but the trend is roughly the same. The difference is generally within the ±40% uncertainty of the spectral estimate. Millard [1972] has made a series of measurements with a CTD in the North Atlantic which may also be used to compute a displacement spectrum. The Millard data have been used by Garrett and Munk [1975] as a test of GM75. A rough com- parison of the Millard [1972] data and the present data set is provided by Figure 5. The Millard [1972] measurements were made at the Mode site at 28°N for various depth ranges below the seasonal thermocline. in all cases the stability is quite a bit less than that in the present study, in which we are considering the upper thermocline in a tropical area. Thus the present measurements provide information on a lower range of $n~'- with only a small overlap with the Millard [1972] spectra. From the theory of internal gravity waves we expect that disturbances near the inertial frequency will have a constant aspect ratio (vertical scale/horizontal scale) much less than f/N. For disturbances of this type, horizontal coherence should decrease for both increasing vertical wave number and increasing horizontal separation. This is the behavior of the DHC predicted by the GM75 model. Results based on XBT profiles are shown in Figure 6. The ordinate is $, and the abscissa is k, the separation distance. Log-log coordinates have been used so that the predicted coherences based on (9) plot as straight diagonal lines. Since the actual separation dis- tances at which the original observations were made have to be lumped together in broad categories resulting in an equivalent spatial smoothing, it is appropriate to smooth in wave number as well. In- calculating the coherence a 10% cosine bell is used [Bendat and Piersol. 1971] to taper the records, a procedure omitted in calculating the displacement spectra shown in Figures 4 and 5. Band averaging is carried out by smoothing the cross spectrum and spectral components separately before dividing to find the coherence. The number of points smoothed depends on the length of the record. For the up- I""»I0 (UK»|0 1038 3846 Stegen et al.: Temperature Profiles in Upper f hermocline 1 u ^V^ DEPTH RANGE 75203m vc* 08 OO78*0|cpm|< 032 >X*x 90s X|m| • 1200 \\ \ X ■ v 06 . sxX\ x »^W^ > ^ 0< N. • ^^^ •" **^.» S"> s ^^^. "*■*■*— 0 2 ^"» ^**^*""»«^_ 00 i i i i i i 02 0.6 0.8 /3|U>i/n| X Fig. 7. DHC for ihe upper depth range, 75 m < -z < 203 m. Solid line is the theoretical prediction of GM72. Dashed lines are 80% con- fidence limits based on the sample size. The open symbols represent the first wave number. per depth range, smoothing is done for three points, for the lower depth range six points, and for the combined profile nine points. The most consistent behavior is shown in Figure 6 for the upper depth range. The isopleths of 0.75 and 0.50 coherence have the expected trend, except for a conspicuous bulge centered at $ = 0.04 cpm. This bulge corresponds to a vertical wavelength of 25 m. Those features of the plot above $ = 0.07 cpm are probably not significant because of the limited vertical resolution of the original measurements. For the other depth ranges there is a strong tendency for the coherence to decrease at the lowest wave numbers, corresponding to vertical wavelengths greater than 50 m. To test the sensitivity of the coherence to the particular statistical treatment used, the coherence was computed without tapering and by using different band averaging. Although details are different and the patterns are quite broken up if no band averaging is per- formed, the loss of coherence at low vertical wave numbers is a persistent feature in ail cases. It was pointed out in the introduction that some loss of coherence is present due to the difference in time between XBT drops as the ship moves along its track. The measurements of lagged vertical coherence by Hayes [1975] are carried out in the thermocline with a background stability frequency N of about 2.5 cph. Hayes's measurements indicate a loss of coherence of unity per hour for vertical wavelengths between 25 and 37.5 m. Dividing this loss by the speed of the ship (6 km/hr), we obtain a loss of coherence due to the time lag alone of 0.14/km. From Figure 6 we can see that the expected loss of coherence is so much greater that the lime interval between drops does not seriously affect the interpretation of the data. In Figures 7 and 8 the same results are plotted in a different way. In Figure 7 the coherence results for the upper range, 75 m < -z < 203 m, are plotted as a function of the nondimen- sional product f)X. Based on an average of 30 pairs and three- point smoothing, the sampling error has been determined from tables by Amos and Koopmans [1963]. The envelope corre- sponding to an 80% confidence level is shown by dashed lines centered on the predicted curve based on the GM75 model. At low values of the coherence the confidence limits are clearly unsymmetrical indicating a large systematic bias of/? at small values. In this presentation of the data the agreement with the GM72 model is surprisingly good. The 0.5 level is a useful measure of the boundary between c llV 10 -2 10 io' 10 r 10 - 10" s 1 1 1 1 1 1 1 1 1 — I ■ "I t I 1 II It 1 1 1 1 1 L X v -\\ ^ x \ -\ xs x X V ^-GM 75 X Vv^-^* * I0.x)= Vz ^ v\ x \^ X X.N s X^ V XX N XX x X.x s \v x Xs ^ XX x X \ X N . \ X X. X - © X X x -> aji © x^x x — X XX Vv XX x rijax - - X^£ X x X x ~ 0 75- 203m x X x xX.x xxx - A 150- 406m \\\ - © 69-420m xxx i ' l II l I ll ' i i i i ■ i il L: — 1 — 1 1 11 lYu 10' -ioJ s U O IH 1-1 rH co CO i-4 X ■H O •H co > CO > CO to -^ CO c CO • 00 rH T3 01 01 tJ 01 l-i (J ^ h to U 0 I-I 4-1 3 01 4J CO 0) c/j u rH 11 rH QJ x: co 43 TJ x: J= H O 4-1 CJ 11 rH CO H tO >i O. 4-1 a. u •o 01 h u CO >. 01 00 a 01 2 0) rH B C 9 c vO 01 vO 01 \£) *H X> wS 43 h CO rH is a a 00 43 C CO -H S -H CO 0) CO o. CO CO CO 4J CO 43 C CO § 3 O CO CO O. CO CO 0) 3 O T3 i-l to O to hI Hi » CO ■H H J3 - 1-1 ^H to •H 3 T3 0) >1 CO 4i •H cj 00 u CO 1 CO 1 58 CO 00 CO 0) CO O rH 11 00 CJ 01 U »H rH c to 43 o 43 O .O -H J3 r-4 43 .H a e C CO rH 1-1 •O CO o m o m o m O .H o x: O CJ O CJ a o O IH O. O to 43 rH iJ rH CM •H .-I rH •^ ■H ^ iH *— ' ■H ^ CO rH 4J 8 CO O O co 00 00 00 00 00 00 to CO lu rH to 43 00 00 u 4= j: n 43 43 c o A 43 43 u m 4J -H 4-1 :-■ 4J l-i •H 4J 4-1 4-1 > 43 co i-l J CO -o CO -H CO iH U CO 43 to CO o 4J a oi £ 01 3 -o 3 co 3 13 3 -o « 3 4J o 3 3 S co CO co 1 CO CO c to CO CO CO a. cr CO o CO to 3 h a O 43 C C CO 01 3 CN 3 co O CO CM 3 &u 14*1 4J -» CO J^ ^ lu i oo to I-l 1 JB O 1 X O- O CO 1 3 o o 3 o to 1 0 O J3 1 3 O O I 3 O O 6 1 O 1 o i o H CN O 43 5 O co O 43 O 43 00 rH 1 o o o CN co CO 4J ON m co CT> CO rH CO CN o rH rH ON Z o M > B O •H 4J fl 4J O O J o J ►J H o 3 f-H a o m CN 4-1 o O 4-> o o 1 > A*. J 4J J*. rH I-l J o o o r-~ a. 1 CO x o. m Ou iH O. r~ O o o o O rH u-j x a O 0) CO CO CO 1 (k o a a r~ a r^ 1 2 o CM m lu CO VI vt rH M O CO X CO X CO u-t 1 m ■H CN CN CN rH rH r>2 B>8 M I-l o o 01 01 z 1-1 •3 o 43 u 43 o fr«0 S-SO ■H o 4J •H 4J M a s-s 1 o o 1 o o rH CO CO IS CO O 1 rH rH I .-1 a> to to CJ B U 11 l-l ITI rH 1 +1 +1 I +1 +i 1 0 o 3 u 3 •+l o u fl o BO o CN CN a i ^ u-\ +1 +1 rH rH CO 0) rH rH rH co -< O CO ■< © CN iH Pu +t +1 +1 +1 +' CO O CO M a c a O to O o 0 jj CO CJ a •H ■H O a 5 oo § B s Sg o CJ a 4-1 4J •H ao m o O 3 3 4J u o o o o O O CO o o o CO iri o o rH rH 3 o CN CHvO J2 i co 3 mo in \o CN CO "? 5 + CO CN rH O to O CO rH O 1 g o i 3 m o C 1 o 3 vo oi I I S3 o o 1 1 01 01 to CO o 4-1 u u M 0) O fl a IH i-t O CM 1 o o o 43 00 •H 43 00 ■H 8 co r< © CO .< © 3© 3© r^ r^ 43 43 rH u IH rJ oi uS u n 01 01 co *j U HI 01 U I-I 4-1 44 H IH H | oi 4-) CO O 01 u 4J 41 01 01 4J 01 4J CO 01 V 1 rH < T3 ^w CO CO e2 a 4-1 01 01 a 01 a U 0) It 01 a 5! a •H ^ c2 4-1 3! 4J 01 ■U oi a o o a a 4J 4J a 1 3 U CO > o u 1-1 O o =3 4J 5! 4-1 O o pd C 00 l-i CO M 01 ■o ■H ■H c a oo •H ■H H 0) C 4J 3 4J u s T3 T) u 01 c •o •a CO Z a) 0) 01 •H OOQ cj M-l 0) u CO cS & 01 CO u u 01 CO M 01 ■H oo i & H rH 43 a cn I a. o 3 1-1 43 rH X n 3 0 CN CO CO 3. oi § 3 O 3 O Ou 3 A u 1-1 ^ M PL* CJ 0u CJ M M a CI •o •H £3 u n T3 4) . rH •H 43 CO ►» •o o O. 01 I iH •o 01 CO > - i X 01 •H . 00 o to H •H >s O >, a 3 o 2 >. u to Cw CO 43 a o . "0 0) * U X Q. O, O < <4-l X 01 01 •H CO U o a. ■a u 43 rH U <4-l m ° 11 BS CO i-i 3 rH 00 CO •H « | CO 0) 4-> U O CO • a n »-. ■H CO CO O. 01 rH 14 ^ lu iH > « Cm 1-1 4J 01 O l-i CO 01 to J! SO 5 * 0) 43 O H 3 4-1 4J w o 3 ,-T •o a, 00 a 4J CO 4-1 c S- o a § >. i-l 0) 01 IH CO CO 4) « 4J ■H - ■H eg M CJ CO <3£ Pu X co r< 3 H X a 0) CO O. eo a ^- CO ^ CO 4J 01 rH M s > M •H 3 CO 3 CO 3 O 43 01 • CO X 11 co CJ 0) C 0 s 3 Q CO O b U CJ u s~' V— > 1053 60 JOHN R. APEL CONCLUSION Seasat-A promises to be an exceptionally useful and productive program. It should have a large impact on earth science and on a community of users in the general populace. The United States space program has produced satellite systems that have looked at the thin envelope of the atmosphere, at the browns and greens of rocks and plants, and at the bright thermonuclear fires of the sun and stars. The time has now arrived to mount a space-oriented investigation of the sea, that last remaining member of the ancient elemental quartet of air, earth, fire, and water. 1054 Reprinted from: No. 6, 865-881. Journal of Geophysical Research 80 , 70 Observations of Oceanic Internal and Surface Waves From the Earth Resources Technology Satellite John R. Apel, H. Michael Byrne, John R. Proni, and Robert L. Charnell Atlantic Oceanographic and Meteorological Laboratories. Environmental Research Laboratories, NO A A Miami. Florida 33149 Periodic features observed on the ocean surface from the Earth Resources Technology Satellite I have been interpreted as surface slicks due to internal wave packets. They appear to be generated at the edge o( the continental shelf by semidiurnal and diurnal tidal actions and propagate shoreward. Nonlinear effects apparently distort the wave packets as they progress across the shelf. This observational technique con- stitutes a new tool for delineating two dimensions of the internal wave field under certain limited con- ditions. Surface Observations Visible manifestations of internal waves on the ocean sur- face may be due to at least two mechanisms, either of which modulates the short surface wave structure rather than causes any change in optical reflectivity or light absorption at depth. The first mechanism suggests that the high velocity of surface water arising from the large internal wave amplitude sweeps together surface oils and materials to form a slick in regions of surface water convergence, and thus the surface reflectivity is increased [Ewing, 1950; Shand. 1953]. The second mechanism predicts that capillary wave energy is focused in the con- vergence zones due to surface stress, and thus the small-scale roughness is enhanced and the surface reflectivity is decreased [Gargett and Hughes. 1972; Thompson and West. 1972]. In either case, quasi-periodic surface features will be produced under light wind conditions, and such effects are very often seen at sea. Figure \a is a photo taken from an aircraft in the Caribbean during Bomex in 1969 that shows a series of what almost cer- tainly must be internal waves propagating from the left overlaid at nearly right angles by much shorter wind waves (R. B. Grossman, private communication, 1974). A ship wake crosses from left to right. The length of the internal waves is to a very rough approximation near 200 m. Figure \b illustrates internal waves being generated by a tide rip and shows at least three distinct wave trains propagating at different angles [La- fond. 1962], Periodic surface slicks may generally be seen from the air under conditions favorable for viewing oil on the sea, i.e., at low sun angles or in the sun glitter, when light or moderate winds prevail. Generation Mechanisms Internal waves can be generated by a wide variety of mechanisms. Wind stress that excites a spectrum of surface waves is one source [Phillips, 1966]. In this case, wave-wave scattering between two surface oscillations of wave vectors and frequencies k,, k2 and oj,, ui2, respectively, generates an internal wave whose parameters are set by the conservation rules 3 = a;, - u>2. The amplitude of the internal wave is partially determined by the surface wave amplitudes; the in- terference pattern decays exponentially with depth with a scale given by 1 /|«3| and thus makes itself felt at levels where inter- nal waves may propagate. Copyright © 1975 by the American Geophysical Union. A second mechanism for generating internal waves is the scattering of barotropic tidal energy into internal baroclinic tides by bottom roughness or by the discontinuities presented by the edges of continental shelves or island arcs [Rattray, I960; Halpern, 1971]. The relative importance of different topographies in transforming tidal motions into internal wave motions is unknown. This phenomenon is difficult to observe in nature; extant evidence suggests that up to to of the 3 x 1012 W energy dissipation required to account for the lengthening of the day may be ascribed to such scattering [Hendershott. 1973]. Thus island arcs, continental shelf edges, and shallow submarine sills become important regions in which to make observations of internal waves. A third mechanism that can cause internal waves is shear flow instability [Rattray. 1960; Phillips, 1966]. In regions of the ocean having large baroclinic currents accompanied by horizontal stratification, small fluctuations in depth of a given stratum can grow to large amplitudes in an oscillatory way by mechanisms generally akin to the Kelvin-Helmholtz instabili- ty. An indication of the onset of this growth can be obtained by evaluating the gradient Richardson number Ri, where Ri = N*/(?ut/dz)*. Here N is the Brunt-Vaisala frequency, and du0/dz is the vertical shear in the horizontal baroclinic current ■speed Uo(z), which may be due to a mean motion or a tidal current. The onset of the growth generally occurs when Ri ap- proaches W (or less). The oscillation amplitude is usually a maximum at some intermediate depth, i.e., is an unstable in- ternal wave. This mechanism may be responsible for the gen- eration of short-period internal waves from baroclinic internal tides of semidiurnal periods or in regions of western boundary currents. Erts I Observations of Internal Waves Periodic or quasi-periodic features in or on the ocean have been detected on many Erts 1, Skylab, and U.S. Air Force Dapp meteorological satellite images taken over both con- tinental shelf and deep ocean areas; to date we have observed these features off the North American east and west coasts, the Gulf of California, the African east and west coasts, northern South America, the central Pacific, and the Celebes Sea. These features are almost certainly surface manifestations of internal waves that are made visible to the spacecraft by one of the mechanisms described above that serve to 'tag' the waves dur- ing winds which are well below whitecapping speeds The observations will first be presented, and then arguments will be made to support the internal wave interpretation. 865 1055 866 Apel et al.: Satellite Observations Fig la. Aircraft observations of internal waves during Bomex (1969) (latitude 9.5°N. longitude 59.5°W). Also shown are surface waves, a ship wake, and wind rows. ■■*":■■ .'•■■■ . "• v >»SS$ Fig. \b. Additional aircralt photo of internal waves [Shand. 195.1 Surface effects on the continental shelves. Figure 2 is a stan- dard trts image taken with the Erts I multispectral scanner southeast of New York; it is printed as a positive image and shows Long Island at the top and New Jersey at the left. The scene dimensions are 184 X 184 km'; north lies approximate!) 102° counterclockwise from the horizontal edges. The scene was taken in orange-red light between 600- and 700-nm wavelength, at approximately 1007 L.T on August 16, 1972. The large white objects in the center of the picture are high-altitude clouds; southeast of the mouth of the Hudson River two U-shaped acid-dumping events are visible; little else of oceanographic interest is apparent. 1056 Apel et al.: Satellite Observations 867 mdttti Fig. Standard NASA Erts image. Area shown is the New York bight taken on August 16, 1972. Figure 3 is a negative print of the same image, computer- enhanced through a process known as contrast stretching, with the land portions sacrificed in favor of ocean information. There is a wealth of data observable, including the Hudson River plume, acid and sewage sludge dumps, color discon- tinuities, and in the southeast corner, periodic surface slicks. (This figure is also geographically stretched by a factor of 1.4: I in the east-west direction due to computer/photo recorder peculiarities.) Figures 4, 5, and 6 are similar computer-enhanced images ol the New York bight area taken on May 31, July 6, and July 24, 1973; each shows periodic striations in the southeast corner of somewhat similar characteristics. No such features were observed on scenes taken over the same area between August 1972 and May 1973. Figure l(a, b, c, and d) is a geographically corrected line drawing of the periodic features observed in the four scenes above superimposed over a bathymetric chart of the area. The general orientation of the lines with the local bathymetry and their concentration in the region of the Hudson Valley argues for strong interaction with the bottom topography. We shall return to this figure later. The nature of these periodic features has recently been verified by direct measurement [Apel et al, 1974], and a simple highly plausible model has been constructed whose predictions are in good quantitative agreement with the observations. It is suggested that the features are surface slicks overlying internal waves that have been generated by scattering of the barotropic tide into baroclinic modes at the edge of the continental shelf, which then propagate up on the shelf, to be absorbed and reflected when the wave breaks on the sloping bottom. We first eliminate several other possible sources of the periodicities. (1) Atmospheric waves causing the depth of im- age modulation seen here do exist but would require much more atmospheric water content than is observed in the pic- tures; nor would atmospheric events reoccur in the same 1057 868 Apel et al.: Satellite Observations Fig. 3. Figure 2 after computer enhancement via contrast stretching. Many oceanographic features are now visible including packets of internal waves separated by 12 to 15 km (lower right-hand corner). geographical location over and over or be so well oriented with respect to the bottom topography. (2) Neither could the features be large surface gravity waves. The observed wavelengths are of the order of 400-1000 m; if these represented surface waves, their periods would be 16-25 s. Even if the slopes were as small as one in 30, the surface wave heights would be of the order of 7-15 m, and such a small rms slope could not account for the one-in-eight brightness change that the features represent on the image. On Figure 3. for ex- ample, the observed sea state was World Meteorological Organization code 1-2 with significant wave height, HU3 = 0.5 m. Therefore surface waves are eliminated as a possibility. (3) Finally, discussions with NASA Goddard Space Flight Center personnel (W. Hovis, private communication, 1972) have ruled out instrument artifacts as a source because of the somewhat irregular nature and orientation of the periodic variations. The remaining possibility is that the features are the result of oceanic internal gravity waves. In light seas, such oscillations often have surface slicks associated with them whose origins may be as described above. The result of either mechanism is a periodic variation in surface roughness and hence in optical reflectivity that defines the internal wave field beneath. Such slicks are common at sea [Shand, 1953; LaFond, 1962). To support further this hypothesis, refraction calculations have been performed by using published bottom topography and compared with the propagation patterns on Figure 3. A simple two-layer model [Lamb. 1932], together with topographic data and density profiles derived from bathy- thermographs for the area in question (obtained a day after the Erts I overflights), was used to compute the progression of a plane wave, originating in deep water, up on the continental shelf in the vicinity of the Hudson Canyon. The predicted phase speed (w/k) for conditions extant at the time of over- flight is about 0.25 m/s. The wave train was assumed incident normal to the edge of the continental shelf. This is probably not a necessary assumption, since for wavelengths of the order of these, phase orientation due to refraction is so strong that by the time they have progressed to the area of the image, the wave fronts would be rendered nearly parallel to contours of local topography. Figure la also shows the results of this calculation by way of a comparison between the observed wave packets and an ar- bitrarily spaced set of phase fronts (dashed lines) as computed 1058 apel et al.: Satellite Observations 869 Fig. 4. The New York bight, May 31, 1973. Note the weak development of an internal wave held in the lower right-hand corner and the large number of ships and ship wakes visible. from the two-layer dispersion relation. Both theory and obser- vation show the wave fronts advancing more rapidly in the deep water over the Hudson Valley than on the adjacent con- tinental shelf. The overall agreement is quite good. Thus all the evidence indicates that the periodic features in these Erts I photos are internal waves that are being detected by their associated surface slicks. These waves have most likely been generated by tidal action at the shelf edge and are being refracted as they move up onto the shelf. One pair of Erts 1 images has afforded a possible determina- 1059 870 Aj>el et al.: Satellite Observations The New York bighl, July 6, 1973. Note the more pronounced internal waves compared with those in Figure 4. tion of both the propagation speed, via time delay measurements, and the generation interval between the inter- nal wave packets. Figure 8 was taken on July 23, 1973, exactly 24 hours before and approximately 110 km to the east of Figure 6. These two images have a good deal of overlap in coverage, allowing observation of the same geographical area. Visible are two sets of packets in the lower center left of Figure 8, separated by 29 to 34 km. Figure 9 is a line drawing showing the packets on Figures 6 and 8. These packets could be standing lee waves created by the bottom topography but more likely are propagating groups of waves. When it is as- sumed that the packets are propagating, it is possible to arrive at an advance of about 30-33 km for a given packet during 24 hours (solid to dashed). Note as well the aforementioned spatial interval between packets observed on a given day, also about 30 km on both days (solid to solid or dashed to dashed). This near-equality of the between-packet intervals with the distance of advance argues for daily or near-daily genera- tion and implies a phase speed of about 0.35 m/s, which is what one would expect from a 50-m-deep mixed layer with a density contrast \p/p a< 10 3 in the depth of water in this region. Figure 10 shows how the wavelengths on Figure 9 vary with front-to-back position in the packets, as is taken at two separate sectors in the easternmost group on that image. There is clearly a strong tendency observed here and in most other satellite images for a given packet to have longer wavelengths at its front than at its rear. This is probably due to a combina- tion of a nonlinear finite amplitude effect with the dispersive character of the waves, as was recently discussed by Lee and Beardsley [1974]. A detailed study of these types of features should yield considerable insight into the physics controlling large amplitude internal wave propagation. Figure II, which was made on November 28, 1972, off southwest Africa just north of Cape Town, shows six packets of surface signatures apparently radiating from a small source at the center left. The area visible is about 180 by 250 km. On Figure 12 a line drawing of the packets superimposed on bot- tom topography is shown, with the depth given in meters. If the interpretation of these signatures as propagating tidally ex- cited internal wave groups is accepted, then the image shows evidence that internal waves may have lifetimes on the shelf of several days. (Density data for this time and place have not yet been obtained, and so no refraction calculations have been made.) Also shown in Figure 11 are the intervals between packets, ranging from 41 to 20 km as shore is approached, as well as the lengths of the leading waves, which increase from 1 .6 to 3.5 km as shoal water is reached. The dependence of the 1060 Apel et al.: Satellite Observations 871 Fig. 6. The New York bight, July 24, 1973. Note the well-developed internal wave fields. In this picture there a ship wake longer than 40 km. packet separations and length of leading wavelength as a func- tion of distance and water depth is shown in Figure 13. This curious behavior, wherein the overall packet suffers retarda- tion as the depth decreases while the leading edge portion is apparently accelerated, may be an anomaly, or it may be a highly significant dynamic effect. More study and additional data are obviously required. A summary of the important characteristics of the surface slick patterns in continental shelf regions follows. 1 . The waves occur in groups, or packets, separated by dis- tances that are of the order of either 15 or 30 km, which, together with the calculated phase velocities, suggest a semidiurnal or diurnal tidal origin. 2. They are nearly always oriented parallel to the local bot- tom contours, as is expected of shallow water waves. 3. None have yet been found in Erts images during the winter or spring in the areas investigated to date. 4. A repeated concentration of energy occurs in the Hud- son Valley in the vicinity of 60- to 80- m depth, probably ow- ing to generation and focusing of waves by the headlands near the Hudson Canyon at the edge of the shelf. 5. The wavelengths fall between 300 and 4000 m and have 1061 872 Apel et al.: Satellite Observations 16 AUG 1972 Fig. la. 31 MAY 1973 Fig. lb. 6 JULY 1973 Fig. 1c 24 JULY 1973 Fig. Id. Fig. 7. Geographically corrected line drawing of the internal wase fields observed in Figures 3, 4. 5. and 6 superimposed on the bottom topography. The dashed lines on (a) show isophase contours as calculated from a simple internal wave model. 1062 Apel et al.: Satellite Observations X71 Kig. 8. Eastern Long Island area, July 23, 1973 (photographically enhanced), 24 hours prior to and I 10 km to the east of the area shown in Figure 6. Note packets separated b\ about 30 km in lower left center. phase speeds near 0.25 m/s, with a longer distance between crests and a greater extent of the packet occurring on the in- shore side and shorter wavelengths and smaller packet extent on the offshore side; the reasons for some of these characteristics are not yet clear. 6. There is some evidence of a continued lengthening ol the distance between crests, especially in the front of the packets, as the group progresses up on the shelf; an accounting for this may be had by a combination of finite amplitude and dispersive effects. Surface effects observed in deep water. The data shown thus far have been taken on or near continental shelves, in water depths less than about 500 m. Very limited numbers of Erts 1 images have been taken over deep water in the open ocean, of which two scenes are shown in Figures 14 and 15. Figure 14 illustrates a photographically enhanced image of a portion of the deep Indian Ocean taken on December 4, 1972; equatorial east Africa is visible on the left. The continental shelf in this part of Africa extends only a few tens of kilometers off the coast. Quasi-periodic structures are apparent in the cloud-free center portion of the scene having a scale of 2-4 km and relatively little coherence. In the lower right there is a definite change in the orientation of the structure. These features lie approximately in the southward-flowing Somali current; the image was taken at a time of year when the current in the region usually undergoes sharp spatial and temporal 1063 874 Apel et al. Satellite Observations 1 ig 4 (ieographica'lb corrected line drawing showing wave packets from lieures d and X. variations due to the shifting monsoon It is thought thai the structure in question ma) he a surface signature ol deep water internal waves; however, no sea truth is available lor this area, and so such an interpretation must remain speculative. Another ocean scene is shown on Figures 15 and 16. These hrls 1 data were taken in the North Pacific at about 4UCN. 160°W in September 1973. Figure 15 has dimensions ol 92 ■ 1X4 km' and is a computer-enhanced negali\e print showing cumulus clouds in the south and a peculiar surface mottling in the north having a scale of a lew kilometers. In this latter region, a microscopic examination shows an expanse ol sur- face waves approximately 200 m in length propagating toward the southeast away from a low pressure cell several hundred kilometers away, having wind speeds of 10 to 15 m/s. A more detailed look at a 1 3 x 18 km sector of this region is shown on Figure 16a, which clearly illustrates two trains of surface gravity waves of nearly equal wavelength intersecting at approximately 15°. as well as the mottling mentioned above: the latter is seen to be very roughly periodic with an average 'wavelength' of about 2.5 km and approximately a north-south orientation. Figure \bb is the logarithm o! the two-dimensional Fourier transform of Figure 16a and shows the spatial frequency components of the mottling near the origin and those of the surface waves as two concentrations ol energy separated by about 15° at nearly equal wave numbers. Under the suspicion that the origin of the mottling might be found either in ( 1 ) surface wave interference effects or (2) in the kind of internal wave generation process arising from wavc- wave scattering that was discussed earlier, we have applied the wave vector conservation rule nil «o„ih,t-...-i m 1065 876 Apel et al.. Satellite Observations 18-00' E 17°00' E 30°00' S 31°00' S 32«00' S 30-00' S 31°00' S 32°00' S 17-00' E 16-00' E Fig. 12. Line drawing of packets of Figure 1 I together with bottom topograph). Wave Packet Characteristics Southwest Africa 11/28/72 T » • Position of front of packet , c(x)t A' —A Length of front wave in packet , X.d) ■— --- -• Depth of water, front of packet , h(n) 5 --4 J "-3 k" i --1 Fig. 13. Leading wavelength and interpacket separation as a lion of distance: from Figures 11 and 12 velocities above whitecapping speeds, whereas the above mechanism may. Surface effects in the Florida current. Figure 1 7 is presented as a final example of possible internal waves observed from Erts 1. This is a contrast-stretched negative image of the Florida current made on August 18, 1972, with Miami located near the upper portion of the photograph. At the bottom center near the mean axis of the current is a group of periodic striations oriented roughly northwest-southeast and extending approximately 30 km along the direction of mean flow. The distance between stripes is about 800 m. It is thought that these are surface signs of internal waves excited by and propagating along the Florida current. Periodic surface slicks in this general region are common at sea, and acoustic echo-sounding from ships has indicated that internal waves often underlie 1066 Apel et al.: Satellite Observations 877 such slicks [Proni and Apel. 1975]. The mechanism for their generation is quite possibly a shear flow instability. On the westward side of the current in the region of cyclonic horizon- tal velocity shear the vertical velocity and density gradients are sometimes such that the Richardson number for the flow ap- proaches W. Diiing's [1973] data give values for Ri that indicate incipient instabilities may occur. Once again, the theoretical conditions that indicate the existence of a mechanism lor generating internal waves have to be verified by simultaneous field and satellite measurements. Summary and Conclusions Convincing proof of the assertions that (1) the periodicities observed on Erts I images are surface slicks and (2) the slicks tag internal waves under conditions of light winds has been ob- tained by the authors in a recent cruise between New York and Bermuda. The data are only partially analyzed and will be published in the near future [Apel et al., 1974]. The limitations of the satellite technique appear to be as follows: light wind and clear sky conditions must prevail, high- frequency internal waves on the shelf are most visible, and coverage is limited by satellite dynamics and sensor characteristics. No amplitude information is currently de- rivable from the spacecraft images. However, as the ac- companying paper shows [Proni and Apel, 1975], this missing dimension to the internal wave field may be partially ob- tained via acoustic echo sounding, so that a more nearly complete picture of the three-dimensional field may be obtained. Fig. 14. Photographically enhanced Erts image of a portion of the deep Indian Ocean oil equatorial east Africa taken on December 4. 1972. 1067 X7K Apel et al. Satellite Observations Fig 15 Enhanced Erts image (negative) taken over the North Pacific al 40°N, I60°W in September 1973. Image dimensions are 92 x 184 km2. Note the peculiar surface mottling and surface wave held. 1068 6 6 = s --S 8 — "7 ^ 3 i E - ~ 5 s n 1! u u -•§ - E S c o = e a u — at) g * ll 1069 880 Apel et al.: Satellite Observations Fig. 17. Contrast stretched negative Ens image ol the Florida current made on August IX. 1972. Note the periodic striations present al the bottom center ol the image. 1070 Apel et al.: Satellite Observations 881 Acknowledgment. This project was partially supported by funds from the Advanced Research Projects Agency, which assumes no responsibility for the correctness of the results. References Apel. J. R . J. R, Proni, H. M. Byrne, and R. L. Sellers, Near- simultaneous observations of intermittent internal waves from ship and spacecraft, submitted to Geophys. Res. Leu.. 1974. Duing. W.. Observations and lirst results from Project Synop 71. Sci. Rep. UM-RSMAS 73010. Univ. or Miami, Miami. Ha.. 1471 Ewing. G.. Slicks, surface films and internal waves. J. Mar. Res. 9. 161. 1950. Gargett. A. h.. and B. A. Hughes. On the interaction ol surface and in- •:rnal waves. J. Fluid Mech., 52, 179-191, 1972. Halpern. D., Semidiurnal internal tides in Massachusetts Bay. J. Geophys. Res.. 76. 6573-6584. 1971. Hendershott, M., Ocean tides, Eos Trans. AGU. 54. 76-86, 1973. Lafond, E. C, Internal waves, in The Sea, vol. I. edited by M. N, Hill. pp. 731-763. Interscience, New York. 1962. Lamb. H . Hydrodynamics, p. 370, Cambridge Universilv Press. Lon- don. 1932. Lee, C. and R. C. Beardsley. The generation of long nonlinear inter- nal waves in a weakly stratified shear flow, J. Geophys. Res.. 79{i), 453-462, 1974. Phillips, O. M.. The Dynamics of the Upper Ocean, pp. 161-197, Cam- bridge University Press, London. 1966. Proni. J. R, and J. R. Apel. On the use of high-frequency acoustics for the study of internal waves and microstruclure. J. Geophys. Res.. 80, in press. 1975. Rattray, M.. On the coastal generation of internal tides, Tellus. 12. 54-62. 1960. Shand. J. A.. Internal waves in the Georgia Strait, Eos. Trans. AGU. 54. 849-856. 1953. Thomson. A. J., and B. j. West, Interaction of non-saturated surface gravity waves with internal waves. Tech. Rep. RADC-TR-72-280. Phys. Dvn. Inc.. Berkeley. Calif.. (ARPA order 1649). October 1972. (Received July 24. 1974; accepted November 6. 1974.) 1071 71 Reprinted from: No. 4, 128-131. Geophysical Research Letters 2, NEAR-SIMULTANEOUS OBSERVATIONS OF INTERMITTENT INTERNAL WAVES ON THE CONTINENTAL SHELF FROM SHIP AND SPACECRAFT John R. Apel, John R. Proni, H. Michael Byrne, and Ronald L. Sellers Ocean Remote Sensing Laboratory Atlantic Oceanographic and Meteorological Laboratories Environmental Research Laboratories National Oceanic and Atmospheric Administration Miami, Florida 33149 tract. Internal waves on the continental Abs shelf off New York have been observed from ship and the ERTS-1 spacecraft, and positive correla- tions made between surface and subsurface measure- ments of temperature, acoustic volume reflectivi- ty, and surface clicks. The spacecraft imagery senses the quasi-periodic variations in surface optical reflectivity induced by the internal waves. The waves appear to be tidally generated at the shelf edge and occur intermittently in packets, which propagate shoreward and disappear in water near 50-m depth. Introduction Surface manifestations of internal waves have frequently been identified during observations from ships, aircraft, and offshore towers, with "surface slicks" or "regions of enhanced capil- lary waves" being terms used to describe the state of the sea surface overlying the internal wave field. In a recent publication, we report on what appear to be surface signs of internal waves detected in visible and near- infrared imagery obtained from the multi-spectral scanner on the NASA Earth Resources Technology Satellite, ERTS-1 (Apel et al. , 1975). These signs usually take the form of periodic varia- tions in optical reflectivity of the ocean sur- face; they have wavelengths of order 500 to 5000 m and appear repetitively in several groups, or packets, which are separated by intervals ranging from about 15 to 40 km. The surface manifesta- tions usually occur during conditions of light winds and relatively clear skies. While the absolute identification of the features in the spacecraft imagery as being due to internal waves had not been attempted until now, it has nevertheless been possible to synthe- size a simple, consistent internal wave model that accounts for most of their major character- istics. They appear intermittently on the conti- nental shelf at intervals which suggest their generation by semidiurnal and diurnal tides at the edge of the shelf (Halpern, 1971) ; they seem to be refractively controlled in phase speed and propagation direction by the mixed layer and water depths and the Brunt-Va'isala profile; and they largely disappear where the mixed layer comes to occupy a substantial fraction of the water column. Nonlinear, dispersive behavior is indicated by certain of their characteristics, Copyright 1975 by the American Geophysical Union. including a sharp onset and the frequent appear- ance of the longest wavelengths at the front of the packet (Lee and Beardsley, 1974) . New York-to-Bermuda Remote Sensing Experiment (NYBERSEX) A cruise was scheduled during June and July 1974 aboard the R/V Westward, a 30-m auxiliary staysail schooner selected in part for her quiet acoustic characteristics. The experiment was primarily intended to observe internal waves in the New York Bight coincident with three conse- cutive ERTS-1 overpasses of that region (12 June, 30 June, 18 July), during the early summer, a season when previous ERTS observations of the periodic surface features had been most frequent (Apel et al. , 1975). The ship was instrumented with a salinity- temperature-deptl device (STD) , expendable bathy- thermographs (XBT) , a horizontally towed under- water hull containing temperature and depth sensors, an echo sounder (20-kHz, 2-ms pulse, 12° x 15° beam) that viewed vertically downward to delineate internal wave motion at depth (Proni and Apel , 1975), and color and infrared films for photographing surface features. Preliminary Results of the Experiment Measurements were made during approximately 20 days of ship operation and 16 separate space- craft overpasses. Because of weather, it was not possible to obtain one set of simultaneous observations involving all of the available instrumentation. Nevertheless, an ample quantity of cross-correlated data was gathered to enable us to assert unecuivocally that the spacecraft imagery does indeed contain surface manifesta- tions of internal waves. Figs. 1(a) & (b) illustrate the geographical area under investigation on 17 July 1974; a portion of the photographically enhanced ERTS-1 negative image (numbered 1724-14475-6) taken in the near-infrared between 700 and 800 nm; and a line drawing interpretation of the periodic surface features. The ship's track is located in the lower left-hand corner of Fig. 1(a), with the circle denoting its position at the time of the satellite overpass — 1447.5 GMT--and the triangle the location of a well-defined internal wave field encountered later on, at approximately 2115 GMT. Four packets are visible in the image. 128 1072 Apel et al . : Internal Waves 129 ^40° >Ae— ■+ 1— w — •- 1 i ' Figure 1(a) (Left). Line drawing showing internal wave packets south of Long Island (solid lines) overlain by isobaths in meters (dotted lines) . Ship track at bottom shows position at time of ERTS-1 overpass (dot) and at encounter with internal wave packet (triangle) . Vertical dashed line is western edge of space- craft image. Figure 1(b) (Right). Section of negative image from ERTS showing „ internal wave field from which Figure 1(a) was derived. Dimensions 154 x 108 km. Black areas are clouds. generally oriented along isobaths and separated by 24 to 33 km. They have wavelengths at the front of the packets ranging from approximately 400 to 1000 m and appear to be propagating shore- ward, a view that is substantiated in part by the absence of packets seaward of the shelf break and in part by the observed doppler shifts discussed below. No surface slicks were visible from the ship at the time of the overpass, nor are there any apparent at that point in the image. Nor would any waves be expected there, since the ship was beyond the edge of the shelf and hence outside of the boundary of the hypothesized generation region. However, at the position where the internal wave field was encountered (at the triangle), slicks were visible from the ship. Color photographs taken from the ship show that the slicks have the color of reflected skylight, in the low winds of 3 to 4 m/s; see Fig. 2. Figure 3 illustrates temperature and acoustic data gathered during a maneuver involving course reversals that was executed through the region indicated by the triangle on Fig. 1. The upper trace shows the temperature at a depth of 29 m Figure 2. Photo of internal wave slicks, taken from deck height (photo courtesy of R. L. Hughes) 1073 i 30 Apel et al.: Internal Waves WNW ESE DISTANCE ALONG TRACK X (km) Figure 3 (upper) . Temperature trace at 29 m depth taken near triangle on Figure 1; (lower) . acoustic reflections from two scattering layers at 20 and 33 m depth, and from bottom at 80 m. Ship course was reversed midway through. as obtained from the towed thermistor. The left half of the record was obtained during a 2.5-km segment of a west-northwest course, during which the internal wave packet of about 2 km total length was encountered; the right half shows the trace obtained during the reverse course. The left half shows down-shifted and the right half up-shifted oscillations. The maximum tempera- ture excursions were of order 6°C. The lower record was obtained from the 20-kHz echo sounder and clearly shows two acoustic reflecting layers oscillating vertically with peak-to-trough ampli- tudes of approximately 15 m. The temperature and acoustic records are in very good correlation, with a 180 phase shift and with the troughs of the wave corresponding to downward-moving warm water from the mixed layer and the crests to upward-flowing cold water. An STD cast made approximately one-half hour before the onset of this wave packet showed a temperature profile which, if oscillated vertically by 15 m, would yield approximately the temperature excursions shown in the upper portion of Fig. 3. This same graph can be used to derive the dominant wavelength, A, period, T, and the pro- jection of the internal wave phase velocity, ?, along the ship velocity, u, by assuming negative and positive doppler shifts due to the ship motion on the reciprocal courses. From the down- and up-shifted wavelengths on that figure, one obtains approximately X sj 500 m. 20 min c •£ 0.35 m/s, where a ship speed u = 1.5 m/s has been used. The phase and group speeds for a 500-m wave- length are about 0.5 m/s for X = 500 m, as cal- culated from the observed density profile, for which 2TT/N - 140 s. r, • raaX Returning to rig. 1, the wave packet that is closest to the one shown on Fig. 3 has a wave- length of approximately 600 m and a packet length of about 3 km, values quite near to those ob- served from the ship. In addition, the isophase lines of the packet appear to be nearly parallel to the isobaths, and it is quite reasonable to assume that the isophase fronts of the wave group marked by "0.6 km" can be extended toward the southwest parallel to the bottom topography, out of the ERTS picture and to the region of ship observation. During the seven-hour time delay between satellite overpass and shipboard measurements, the waves would have propagated only about ten km toward the northwest and could thus be expected at about the location shown. Assuming constant phase and group speeds of 0.35 m/s, the wave packet would progress about 33 km shoreward in 25 hr. This is close to the spatial intervals observed on Fig. 1, which suggests once-daily generation. Additional spacecraft and shipboard data show weakly developed internal wave striations on 12 June 1974, a calm, moderately clear day. However, on 13 June 1974 no evidence of slicks was visible either on the ERTS image or from the ship; this was a day of fresh, 7-8 m/s northeasterly winds, and slicks apparently do not form under such conditions. Another experiment (a) established that the echo-sounder was indeed quantitatively observing internal waves; and (b) showed the location of the reflecting layers relative to small-scale temperature gradients. Figure 4(a) (left) shows two XBT traces obtained on 1 July 1974; one was taken at the crest and the other at. the trough of an internal wave packet that was propagating past the ship while it was adrift. The packet was accompanied by a well-defined series of slicks. The temperature traces show the thermo- cline oscillating up and down with approximately a 20-min period by an amount r|(z) that depends on depth z; by differencing the two traces, one may obtain the displacement, which is shown on the right. The lowest order mode is clearly 1074 Apel et al . : Internal Waves 1 3 i CZES3 REFLECTING LAYERS J J_ 1 I 10 15 TEMPERATURE T(z), CO 20 0 2 4 6 8 iO DISPLACEMENT ij(ii.(fivt Figure 4 (Left) . Temperature profile at internal wave trough (solid line) and crest (dotted line). Cross-hatching shows regions of intense acoustic reflec- tions, as in Figure 3. (Right) . Displacement of isotherms shows dominant first- order mode. the dominant one, but higher order modes are also present. Coincident STD data demonstrate that the water column is statically stable, with the temperature inversion shown being offset by increased salinity in that region. The coincident towed thermistor and echo- sounder records also yielded data on this inter- nal wave field. These traces (not reproduced here) appear quite similar to those on Fig. 3. The acoustic record shows two reflecting laminae that undergo vertical excursions whose ampli- tudes and phases are in excellent agreement with those shown on the towed temperature and XBT records. The locations of these reflecting layers are indicated by the hatched regions of Fig. 4(a) and are clearly associated with depths where the macroscopic curvature of the temperature record, 32T/3z2, is the largest. This is in accord with models of acoustic scat- tering from fluctuations in the index of re- fraction. Thus the experiment yields some evidence that, in addition to scattering from biological material — the mechanism that is usually invoked to explain such reflections — these signals may be partially returned from microstructure and turbulence in the water column. Conclusions The combined measurements of horizontal and vertical temperature variations, density struc- ture, acoustic reflections, and spacecraft imagery have shown that the periodic features seen in ERTS images taken over the New York Bight are indeed surface slicks associated with oceanic internal waves; that such waves appear to be generated at the edge of this part of the continental shelf approximately once a day; that they propagate as intermittent high-frequency packets whose wavelengths, phase speeds, and periods are consistent with the vertical density profiles and depths for the area; and that they are detectable by high-frequency acoustic echo sounders. Thus the combination of spacecraft imagery, which yields synoptic scale data, and echo sounding, which provides amplitude infor- mation on a more restricted scale, appears to offer useful tools for investigations of higher- frequency internal waves on the continental shelves . Acknowledgements. The authors are grateful to NASA for gathering open-ocean ERTS-1 data and to the officers, crew, and apprentices of the Westward for their help. This research was partially supported by funding from the Advanced Research Projects Agency, which assumes no responsibility for the content cf this report. References 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. Geophys. Res. , 80, 865, 1975. Halpern, D. . Semidiurnal internal tides in Massa- chusetts Bay, J. Geophys. Res. . 76, 6573, 1971. Lee, C, and R. C. Beardsley, The generation of long nonlinear internal waves in a weakly stratified shear flow, J. Geophys. Res. , 79. 453, 1974. Proni, J. R. , and J. R. Apel, On the use of high- frequency acoustics for the study of internal waves and microstructure, J. Geophys. Res. (to be published), 1975. (Received November 11, 1974; accepted March 14, 1975.) 1075 72 Reprinted from: Journal of Geophysical Research 80, No. 9, 1147-1151. On the Use of High-Frequency Acoustics for the Study of Internal Waves and Microstructure John R. Proni and John R. Apel NOAA Atlantic Oceanographic and Meteorological Laboratories. Miami, Florida 33149 Experimental data and theoretical calculations on the scattering of high-frequency acoustic signals from oceanic internal waves are presented. Acoustic data on internal waves are compared with simultaneous temperature (towed thermistor) data. The comparisons have shown a high degree of cor- respondence between the temperature and the acoustic data. Theoretical calculations for the acoustic scattering cross section a are made by assuming that temperature lluctuations give rise to the acoustic scattering. An enhanced cross section for scattering from layered temperature lluctuations is to be ex- pected, in agreement with the 1973 calculations of W. H Munk and C. Garrett. For some lime it has been known that internal waves can be observed by using high-frequency (5- to 25-kHz) acoustic echo sounders [Hersey, 1962]. Chance observations of internal waves using high-frequency echo sounders have generally been attributed to scattering occurring from biological organisms riding upon isopycnal layers oscillating in the vertical due to the passage of internal waves. Recently, however, another mechanism has been proposed that may also contribute to making detectable scattering occur from internal waves. This mechanism is the scattering of high-frequency acoustic signals from temperature and density microstructure and microstruc- tural fluctuations. If it can be demonstrated that the vertical amplitudes and phase variations of upper-ocean internal waves can be measured by using echo sounders from ships and that these measurements do not necessarily rely on the presence of plankton and fish for their general utility, a new tool will be recognized that will enable rapid surveys of internal wave directional spectra to be made. These data, coupled with the synoptic view of wave patterns obtained from spacecraft [Apel el ai. 1975], would come close to specifying the three- dimensional vector wave field that is of ultimate interest. The possibility of detection of oceanic internal waves via acoustic scattering from temperature microstructure fluc- tuations has received support from several different areas, some of which are as follows: 1. Theoretical calculations carried out by the coauthors and others [Munk andGarreu. 1973] indicate that an enhanced acoustic scattering cross section is possible for the type of layered microstructural fluctuations thought to be associated with internal waves. 2. Analogous experiments on acoustic echo sounding in the atmosphere [Hall, 1972] have clearly revealed atmospheric internal waves where there is an obvious absence of biological scatterers. Some additional experiments [Little, 1969] have been carried out by comparing in situ measurements of at- mospheric microstructural fluctuations with those obtained acoustically, the result being that the microstructure seems theoretically and observationally capable of producing the backscatter. 3. Recent measurements on oceanic microstructure suggest a generally horizontally layered model of fluctuations with considerable fine structure present [Woods, 1968; Osborn Copyright © 1975 by the American Geophysical Union. and Cox, 1972]. This model is consistent with acoustic obser- vations and supportive of layered acoustic scattering theory. Again measurements of microstructure and acoustic backscatter in the ocean imply that the temperature fluc- tuations are capable of backscattering at detectable levels. The virtue of obtaining data on internal waves by using high-frequency acoustic echo sounding is that large volumes of data on both vertical and horizontal distributions of the waves can be gathered rapidly while the vessel is underway at speeds large in comparison with the phase speed of the waves. In other words, by this method, compared with thermistor chain towing, for example, a large-area synoptic body of data can be gathered relatively quickly. Directional spectra may be con- structed by sampling along five-point star patterns, for in- stance, and the phase relationships between surface slicks and the underlying internal waves may be studied conveniently. (Mel Briscoe has correctly pointed out that perhaps a factor of 2 is involved in increased data over traditional thermistor methods. He has also mentioned the article by Voorhis and Perkins [1966] as indicating some limitations on obtaining in- ternal wave spectra from moving platforms.) This kind of data is extremely valuable in its own right as well as essential for comparison with satellite data on internal waves. Woods advances the model of a vertical stratification that consists of alternate homogeneous layers of meter thickness separated by narrow (e.g., 10-cm) sheets in which there exist sharp temperature gradients. We assume that these layers con- tain microstructural temperature fluctuations that are capable of scattering high-frequency acoustic signals. The regions of sharp relatively stable temperature gradients are apparently also capable of scattering high-frequency sound, as is men- tioned earlier; the fact that a layering is present gives rise to an enhanced acoustic return. This model is considered in greater detail in the section on theoretical calculations. Finally, these layers and sheets are assumed to oscillate vertically with the passage of internal waves, their oscillation thus permitting high-frequency sound to be used to observe the internal waves. Example of Acoustic Data A typical example of the type of data obtained by the coauthors is shown in Figure I. A 20-kHz hull-mounted echo sounder was used in obtaining the record shown in Figure I . In this picture, time runs from left to right, time marks being placed every 5 min. The record was made in May 1970 ap- proximately 320-480 km off the Ecuadorian coast. The deep 47 1076 148 Proni and Apel: Acoustic Observations Fig. 1 . Internal wave observations made during a rise of the scatter- ing layer. Time marks are given every 5 min. scattering layer (DSL) is plainly visible, rising from about 225 m at the left of the figure to a depth of less than 100 m at the center of the figure. Notice a thin layer of high acoustic reflectivity at a depth of about 30 m at the left of Figure 1 . This layer undergoes erratic oscillations with a gradual increase of the acoustic return. As time goes on, returns from this reflecting layer become merged with returns from the DSL as it rises; 20-25 min later a reflec- ting layer becomes visible again at about a depth of 30 m. In addition, other reflecting layers have also become visible at depths of 70, 100, and 130 m. It appears from the data shown in Figure 1 and from a great deal of other similar data that the model of acoustic reflectors distributed in discrete vertical layers is tenable. Indeed these layers do appear to outline the vertical distribution of internal wave oscillations present. The fundamental questions are, of course. What are the sources of the acoustic return and how accurately does the acoustic record portray the vertical inter- nal wave field distribution? Theoretical Calculations First a calculation is carried out for acoustic scattering from a completely homogeneous volume of temperature fluc- tuations (no layering is assumed). A calculation is then carried out for acoustic scattering from a layered region, and the results are compared with results from the homogeneous calculation. It is assumed in the homogeneous calculation that the temperature fluctuations can be characterized as having a Kolmogorov spectrum. Acoustic scattering from homogeneous isotropic temperature fluctuations. The acoustic scattering cross section a for this case can be written as [Tatarski, 1961) a(6) = 0.03X"3 cos" 6 (sin " *m« + C0 cos - 2 with the speed of sound in seawater written as C(7) s 1449 + 4.627" - 0.05P m/s and where nifi) acoustic scattering cross section per unit volume; 8 scattering angle (for direct backscatter, 8 = 180°) X acoustic wavelength; C sound speed; CV structure constant for velocity fluctuations; T temperature, °K; CT structure constant for temperature fluctuations. (1) (2) Reliable measures of CV and CT are difficult to obtain. However, a survey of measured values for CT in the ocean has been given by two Russian authors, Gostev and Shvachko [1969]. A representative value for CT in the ocean is 10"' deg m""3. By way of comparison, CT for the atmosphere is typical- ly 5 X 10"2 deg m-"3. An estimate for 3(k) dx, dKz dnv (7) The layered geometry also puts restrictions directly on <£,(*) and indirectly on K. Our layered model assumes that all vertical (i.e., z) fluctua- tion wavelengths are less than a layer thickness; hence 03(k) ^ 0 for k, > 2w/AZ (8) The restriction on the horizontal wave number Kr(=(»Cx2 + «y2)"2) is not modeled so easily as that for kz (given by (8)). Munk and Garrett assume a relationship between k, and kt, namely Kr < (XK; (9) where a is a constant, and also assume that when the above condition is satisfied, one may write 4>3(k2, kx, Ky) = 4>3(Kz) (10) In the present development a different point of view is taken with regard to the above assumptions. By using an assumption different from that in (9) but in agreement with that in (9) a different scattering dependence on acoustic wavelength appears. A restriction that may be imposed on Kr for awide variety of possible fluctuation spectra is that *r < 2v/ARc (ID where ARC is the horizontal fluctuation correlation distance. This is a measure of the width of the spectrum in horizontal wave number space. As of yet no relation has been made between the allowable range for «2 and that for nr. To recover the two assumptions of Munk and Garrett (but in an amended form), we ourselves make the following assumptions: (1) ARC > AZ and (2) AR, < ARC. From assumption 1 it follows im- mediately that the allowable range of values for k, is greater than that for xr; i.e., allowable kz is greater than allowable Kr. From assumption 2, assumption (10) of Munk and Garrett follows, since assumption 2 implies that over any given area il- luminated by an acoustic beam, no appreciable fraction of a fluctuation wavelength is incorporated (i.e., near-specular con- ditions are present). To reiterate, by not making assumption (9) in the form of Munk and Garrett, a different dependence on acoustic wavelength of the acoustic scattering cross section will emerge. Also beam width information is contained in the factor / and does not come directly from the relation between k2 and Kr. Thus according to our model, only a certain volume in k space (a cylinder) contains fluctuation wavelengths. Only in this volume will impinging K (acoustic difference vectors) en- counter fluctuation wavelengths from which to scatter. This volume may be symmetric about the kz axis, or it may not be. Now <£,(«,) = I I 3(k2, kx, k„) dnz Jail «..«„ J dKy (12) /** / an cv ni* / an ct / 03(K., Kx, *v) d^ dKv (1 2»/4S,, •>-2t/ARiz 3) (Henceforth we take ARCZ = ARcy = ARC, where ARC is the horizontal correlation distance of the temperature fluc- tuations.) We assume, following Munk and Garrett and as- sumption 2, that 03(xj, kx, Ky) = i(Kz). Then 3(Kt) = (AiV/32ir2W>.(« .) (14) with 03(»O = 0 when either \KZ\ > 2-k/AR, (15a) |K„| > 2ir/AKc (15ft) |K,| < 2tt/AZ (15c) We may now write (7) as . ,,4 rKr.+ (2i/4J,l rA'»t(2>/4B, da 4k 1 I dQ ir Jk,-(2i/i8,) -"/f.-izi/a*,) Jk,- <2» ,/AZ, ^4 /AZ) 16 AZ2 A;V 4 32tt2 <£i(k;) rf«, dKz dKy (16) (Christopher Mooers has kindly pointed out that the integral over kz in (16) should be broken into two integrals with limits of integration from -« to K, - (2t/AZ) and from Kz + (2ir/AZ) to + °°, respectively; this is, of course, correct, so that one must add that the contributions at +» and -<*> are zero (for the spectrum given in (17)) and do not affect the resulting value of the integral.) A possible form for #,(/0 according to Gregg and Cox [1972) is 4>,(0 = Bkz'2 where B = 5 X 10"'0. Substituting (17) into (16) and integrating, one has (17) ^- = ^-BARr-AR^AZ dU 32 ill AZ2 I6ir" 2 AZ (18) where 8 is the angle between incoming and scattered (received) acoustic transmissions. We can now make a numerical estimate for da/dQ. We take for vertical incidence K, = 2 • k = 1.67 x !0+2 m"1 (19) (at 20 kHz, X = 7.5 cm). We take Az = 5 m, and then .1 da A:4 5 X 10"" AR,2AZ dQ -n*ARc (20) 32 (1.7 X io2r alAR?AZ (21) da'/dQ = 2.8 x 10"8 ■ ARC We take for ARC, ARC = 20 m. (Different researches report values of ARC ranging in size from the order of meters [Leiber- mann, 1951] to the order of tens of meters [Black, 1965]. A value 10 m is chosen as a reasonable numerical estimate for ARC.) Then da'/dQ = 2.8 X 10"' m7m3 (22) per unit solid angle, which is a significant increase over the value given for nonlayered scattering in (4). Note that for X « Az ■ sin (8/2), da_ dQ ■K2BARr2AR, AZ 8 sin" 0/2 (23) Thus da/dQ varies with X" 1078 150 Proni and Apel: Acoustic Observations In summary then, under conditions of layered acoustic temperature and density fluctuations, compared with three- dimensional isotropic fluctuations, two major differences appear: (1) An enhanced acoustic return of perhaps 20 dB or so over three-dimensional fluctuations is observed. (2) A X-2 acoustic wavelength dependence is observed, as compared with A~"3 for three-dimensional fluctuations and X~' for particulate matter [Tatarski, 1961]. The latter result suggests that a multifrequency system would be useful in providing information on the nature of the acoustic scatterers. This conclusion is well known and arises from other acoustic studies such as scattering from bubbles. Additional Example of Acoustic Data (The data in this section may be compared with those ob- tained by Curtinand Mooers [1975] along the west coast of the United States.) The following data were gathered on the U.S. continental shelf off New York (40°7.25'N, 71°34'W) on June 30, 1974. The data shown here were gathered during the New York to Bermuda remote sensing experiment. This is the area referred to in the paper by Apel el al. [1975]. Acoustic reflections from a packet of internal waves are shown in Figure 2. The packet is propagating shoreward (to the right in the figure) in water of 85-rn depth. Immediately below the acoustic record of the packet is a temperature record made by using a towed thermistor. The thermistor was towed in a vee fin device with minimal vertical movement. (The total vertical excursion for the duration of the record shown in Figure 2 was less than 1.2 m, the vee fin being centered at a depth of 3! m.) The records represent Doppler-shifted internal waves, the ship proceeding at about 2.5 m/s in the general direction of propagation of the wave packet, as is evidenced by surface slicks and bottom topography. It can be seen that cor- related oscillations occur in both records (to quantify the cor- respondence between the two records, the cross correlation is presently being calculated for a many-kilometer-long stretch, of which the data in Figure 2 form a part). Note that troughs in the acoustic signal correspond to troughs in the temperature trace; i.e., warm water is sweeping by the towed thermistor. A periodogram for acoustic records is shown in Figure 3; also shown in Figure 3 is an expendable bathythermograph record taken during the encounter with the internal wave packet. Note that the dominant peak in both records occurs at K = 2.2 cycles/km, or X = 450 m. This is roughly the domi- TEMPERATURE CO — 1 1 1 ' 1 ' ii., ' C _ £ 50 - H 2 - ■ X J; IOO Q 150 7 ft 1 i ■ r*- X<450m 3 - » iT. I 4 - 5 £3 - 2 2 1 \A*~" x 300m 2000 3000 DISTANCE (ml Fig. 2. Acoustic and temperature traces lor an internal wave packet. (Note the location of the XBT mark.) Fig. 3. (Top) XBT trace made during the passage of theipacket shown in Figure 2. (Bottom) Periodogram for the acoustic trace shown in Figure 2. (This is not a variance-preserving plot.) nant wavelength that one would eyeball-estimate from the data records of Figure 2. This periodogram was made from data that were demeaned, detrended (i.e., a linear trend), and cosine-tapered. The wave packet is a transient phenomenon of limited duration. The periodogram has 2 d.f. (There is an interesting problem here in spectral analysis. The high-frequency/short-wavelength internal waves present in New York to Bermuda remote sensing experiment NYBERSEX) data occur in the form of discreet packets. These packets appear to contain only a few cycles (e.g., 5-10) of the dominant internal wave wavelengths. Thus the number of degrees of freedom attainable is limited. The process of packet generation is intermittent (e.g., semidiurnal tidal excita- tion) and is nonstationary and nonhomogeneous. Further- more, the internal wave amplitude r\ regarded as a random variable is non-Gaussianly distributed (t; is distributed in fact somewhat as a sine distribution). The approach taken in this paper may be compared with that of Halpern [1971], in which a sequence of internal wave packets was analyzed as a unified time series. By using Halpern's approach a high number of degrees of freedom are available. In this paper, one realization of a wave packet is analyzed.) An amplitude estimate based on the temperature record can be constructed by using the XBT data and the towed ther- mistor data. Assuming the amplitude y can be written as [De- fant. 1961] •-" AIL (7" is temperature) one has, using a value of ATtaken from the 1079 Proni and Apel: Acoustic Observations 1151 peak-to-peak oscillation present in the 3000- to 3500-m range, r, = (17.4° - l0.5°)/0.38°C/m = 18.1 m Reading directly from the acoustic record, one has a double amplitude of approximately 18-20 m. This comparison has been made for only a small portion of the internal wave packet, but if one assumes that dT/dZ does not vary much over the length of the packet, several points for amplitude es- timate comparisons can be had. In general, the acoustic and temperature internal wave amplitude estimates have been found to lie within 10% of one another. The greatest source of uncertainty in reading the internal wave amplitudes from the acoustic records lies in the width of the acoustic trace. (Flow noise, ship motion, and cavitation noise all contribute to the widening of the acoustic trace.) The records can be enhanced by signal processing (the data are recorded on magnetic tapes), so that less ambiguous amplitude estimates can be made. Summary and Conclusions It appears that high frequency acoustics may be a more po- tent tool lor the study of internal waves than has been suspected in the past. This will be particularly true if, as is in- dicated in this paper, the acoustic technique does not have to rely upon the presence of biological or other material to out- line the presence of internal waves. The calculations presented in this paper indicate a higher expected acoustic scattering cross section from layered fluctuations than from nonlayered fluctuations provided the many assumptions made in the calculations prove to be correct. The fundamental question remaining, of course, concerns which type of scatterer is con- tributing which part of the acoustic return. It appears that ex- tensive research remains to be carried out in order to deter- mine the relative contribution of such ocean features as temperature microgradients, temperature fluctuations, tur- bulence, biological material, gas microbubbles, and suspended sediments to acoustic return in various oceanic conditions. Additional Comments The NYBERSEX cruise, recently (July-August 1974) carried out by the Ocean Remote Sensing Laboratory of NOAA, had the goal of obtaining concurrent satellite and acoustic data on internal waves. Goals of the cruise included obtaining acoustic records concurrently with the passage of surface slicks or signatures of internal waves (see the accom- panying paper by Apel et al. in this issue). Goals also included studying the correlation of isotherm oscillations, temperature step oscillations, and acoustic records of internal wave os- cillations Preliminary reduction of some of the data (some of which were presented) from this cruise indicates a significant degree of correlation among temperature, surface slick, and acoustic records. This paper has dealt with high-frequency (5- to 25-kH/) acoustic interactions with internal waves. Not discussed has been the field of low-frequency (e.g., 5-Hz to 5-kHz) acoustic interactions with internal waves. The prime effect of internal waves on low-frequency sound propagation is on the propaga- tion times of acoustic signals from source to receiver. Several researchers over the last 20 or so years have studied th.e effects of internal waves on low-frequency multipath acoustic propagation [Clark and Kronengold. 1974]. There have also been experiments in which concurrent high- and low- frequenc) acoustic measurements on internal waves have been made [Weston el al., 1970]. One of the most interesting current research problems in the propagation of long-range low- frequency acoustic signals is determining how to use a best model of an internal wave spectrum to predict the spectrum of acoustic phase fluctuations. An exciting possibility is that if satellite observations prove to be a means of obtaining reasonable estimates of internal wave fields in a given area, then it may prove to be possible to make predictions on what range of fluctuations may be imposed on acoustic signals by the internal wave field in that area. From illustrations con- tained in the paper by Apel et al. it is clear, for example, that there is a significant difference between the type of internal wave field modulation to be expected in shelf areas and that to be expected in the deep ocean. Acknowledgments. Melvin Briscoe and Christopher Mooers provided many useful comments and criticisms. Manuel Huerta, Henry Diaz, and Fred Newman assisted in clarifying the ideas in this paper. Thanks are extended to them all. References Apel. J R . H M Byrne. R L Charnell. and J R Prom. Obser- vations of oceanic internal and surface waves from the tarth Resources Technology satellite. / Geophvs. Res . 80. 865-871, 1975. Black. C F.. The Turbulent Distribution of Temperature in the Ocean. Bissett-Berman Corporation. Santa Monica. Calif. 1965. Clark. J G , and M. Kronengold, Long period fluctuations of CW signals in deep and shallow water, J. Acoust Soc. Amer.. 56(4), 1071-1083, 1974. Curtin, T. B., and C. N. K. Mooers, Observation and interprelalion- of a high-frequency internal wave packet and surface slick pat- tern, J Geophys. Res.. 80. 872-894, 1975. Defant, A . Physical Oceanography, vol. 2, p. 537. Pergamon, New York. 1961. Gregg. M.. and C. Cox, The vertical microstructure of temperature and salinity. Deep Sea Res.. 19. 355-376, 1972 Gostev. V. S, and R. F Shvachko. The microstructure of the temperature field in the ocean, Atmos Ocean Phys. 5(10). 1066-1074. 1969. Hall. F. F . Acoustic remote sensing of temperature and velocity struc- ture in the atmosphere. Collected Reprints: 1970-197 1 , pp. 168-181. Wave Propagation Lab., Boulder, Colo., 1972. Hersey. J B . Sound scattering by marine organisms, in The Sea. vol. I, pp. 498-566, U.S. Government Printing Office, Washington, DC. 1962. Halpern. D.. Semidiurnal internal tides in Massachusetts Bay, J Geophys Res . 76. 6573-6584, 1971. Liebermann, L., The effect of temperature inhomogeneities in the ocean on the propagation of sound, J Acoust Soc Amer., 23, 563-570, 1951. Little, C G., Acoustic methods for the remote probing of the lower at- mosphere, Proc IEEE. 57(4), 571-578, 1969. Munk. W H , and C. Garrett, Internal wave breaking and microstructure (the chicken and the egg). Boundary Layer Meteoroi. 4. 37-45, 1973 Tatarski, V. I., Wave Propagation in a Turbulent Medium, chap 4, McGraw-Hill, New York, 1961. Voorhis. A. D., and H. T Perkins, The spatial spectrum of short-wave temperature fluctuations in the near surface thermocline. Deep Sea Res . 13. 641-654. 1966. Weston, D E., J. Review, F. R. Jones, and J. W. Ranster. An echo- sounding record and a sound transmission record showing internal waves. Rep. ARL/L/R88. Admiralty Res. Lab., Teddington, Middlesex, England, 1970. Woods. J. W '., Wave-induced shear instability in the summer ther- mocline. J Fluid Mech . 32. 791-800. 1968. (Received April 25. 1974; accepted November 4, 1974.) 1080 Reprinted from: Nature 254, No. 5499, 413-415. 73 Acoustic observations of suspended particulate matter in the ocean Clouds of suspended particulate matter in the ocean have been observed by satellite1, sensed by nephelometers2, and sampled by water bottles. We present here evidence that it may be possible to observe oceanic suspended particulate matter acoustically. The particulate matter discussed is thought to have arisen from a dredging operation. It is also thought that the suspended matter is present at lower concentrations than any in any other case recorded acoustically3. Certain questions, as yet unresolved, regarding acoustic scattering strengths are also discussed. 80°I0'W 25°50'N 25°45'N- 25°50 N 25°45' N 80°I0'W 80°05'W Fig. 1 Typical trackline in area of investigation (May 22. 1974, RV Virginia Key). A-D indicates locations shown on Fig. 2. An experiment to determine the feasibility of acoustic surveys of suspended sediments was conceived during field testing of a 20-kHz LODAR echo-sounding system. The LODAR system has a downward-looking acoustic beam of 12° by 18° and a pulse duration of 2.2 ms. Initial acoustic records of a cloud-like feature were coincident with visual sightings of suspended sediment down-current of a working dredge. During this initial experiment, the dredge was operating for five days out of eight and data were taken on both incoming and outgoing tides. In conjunction with the LODAR records, measurements were made with expendable bathythermographs (XBTs) and a continuous recording light beam transmissometer (LBT). On the five days when the hydraulic suction dredge was working, a visible cloud of suspended sediment and the acoustic 'cloud' were observed down-current of the dredge. On incoming tides the sediment flowed into the bay where the water is too shallow to operate the LODAR system. On outgoing tides the suspended sediment flowed out of the channel; the surface manifestation disappeared between 200 and 600 m seaward of the channel, depending on the state of the sea. Figure 1 is a typical trackline and Fig. 2 is the acoustic record from part of that track. The same general trackline was traversed each day. The minimum volume of the cloud estimated from acoustic records was about 60,000 m3. During the three days when the dredge was not in operation, clouds were not observed either by the acoustic equipment or by visual observation of any surface manifestation. The continuous recording LBT records the percentage transmission of a beam of white light 2 cm in diameter over a l-m path; filters maximise the sensitivity at 475 nm. Although the light transmission characteristics involve more than one parameter34, the instrument is useful for obtaining approximate particulate concentrations and distributions. LBT station depths were limited by the length of telemetering cable aboard, but shallow casts (less than 100 m) indicated a reduction in transmittance of between IO°0 and 15°0 in the upper regions of the cloud. The developed cloud occupied a region between inflection points on the XBT temperature traces (Fig. 3). A volumetric concentration on the order of 0.01-1 %, obtained, from the acoustic data, enabled us to carry out a first-cut check of acoustic and transmissometer readings. For the LBT used5, a reduction in transmittance from 85 °0 to 60°„ implies a particulate concentration of roughly 0.03 %, so that the acoustic and transmissometer devices are consistent. The observed acoustic 'cloud' cannot be accounted for either by temperature anomalies or by a package of con- taminated bay water which would occur independently of dredging operations (see ref. 6). Temperature anomalies, however, seem to influence the location in depth of the acoustic cloud which occupies a region between inflection points on the _XBT records (Fig. 3). This observation is consistent with the finding that temperature-induced density gradients are an important controlling factor in the transport of fine-grained sediment7. Large biological reflectors cannot account for the cloud as there is no evidence of point source reflectors on the records. The possible role of small biota in producing such a cloud requires further investigation. Although unlikely, the stirring up of bottom nutrients by the dredging operation may have produced a short-lived bloom of microscopic biota. It is unlikely that the cloud results from microbubbles produced by organic processes in the sediment and released by Fig. 2 Acoustic LODAR record along (ypical trackline. locations as indicated on Fig. I. The shallow scattering layer is indi- cated. May 22, 1974. 1081 - 100 - 150 20C 20O Fig. 3 Sketch of the acoustic cloud with overlay of Sigma-7"(oT) curves (T. Lee and D. Mayer of National Oceanic and Atmospheric Administration) with XBTs on same scale. May 22, 1974. aT = (p-1) X 1,000 where p = water density). o> is obtainable from STD (salinity-temperature-depth) measurements. Gulf Stream core based on salinity of March 27, 1974. the dredging operation. An extraordinary amount of decay of organic matter would have to occur to produce clouds of the magnitude observed on every outgoing tide when the dredge was in operation. Such extraordinary quantities would have been observed as seeps emanating from the bottom when the dredge was inoperative. The observed behaviour and properties of the acoustic cloud support our interpretation that it is the product of acoustic scattering from suspended sediment particles in the water column. The nature of a suspended sediment cloud and its movements would be affected by the tides, temperature- induced density gradients, and the northward flow of the Gulf Stream current. The observed cloud varied in size, shape, and location corresponding to the changes in dredging opera- tions and tides. The tides affect both the direction of flow in Government Cut and the flow of warmer bay waters out to sea on outgoing tides, thus inducing the density gradients at the outflow of the cut. Also, the observed acoustic clouds moved to the north on leaving the cut, with the easternmost extent of the cloud approaching the Gulf Stream. How is it, though, that particles with diameters that may be of the order of 1-100 urn can reflect detectable sound signals at 20 kHz? The present hypothesis is that the sediment cloud may have regions of relatively high particulate concentration which are detectable but not resolvable using the acoustic system described here. Studies are under way to determine the validity of that hypothesis. Theoretical work and further field work are under way to develop what may prove to be a valuable survey tool, with immediate applications for the sedimentologist, geologist, ecologist and pollution monitoring agencies. Dr David Drake assisted with the transmissometer measure- ments; his comments and advice are appreciated. We thank the referee who pointed out the consistency check between the acoustic level received and the transmissivity decrease. J. R. Proni D. C. Rona C. A. Lauter Jr R. L. Sellers Atlantic Oceanographic and Meteorological Laboratories, Environmental Research Laboratories, National Oceanic and Atmospheric Administration, Miami, Florida 33149 Received December 11, 1974; revised February 25, 1975. 1 Klemas, V., Borchardt, J. F., and Treasure, W. M., Remote Sensing of Environment, 2 (4) (1973). 2 Ewing, M., and Thorndike, E. M., Science, 147. 1291-1294 (19651. > Gallenne. B., Estuar. coast, mar. Set., 2. 261-272 (1974). 4 Jerlov. N. G.. Optical Oceanographv (Elsevier, Amsterdam, 1968). s Beardsley, G. F.. Pak, H., Carder, K... and Lundgren, B., J. geophys. Res., 75, 2837-2845 (1970). -o Proni, J. R., and Apel, J. R.,J. geophys. Res. (in the press). 7 Drake, D. E . Shelf Sediment Transport (edil. by Swift, D. J. P.. Duane. D. B..and Pilkey. O. H.), 307-331 (Dowden, Hutchinson, and Ross, Stroudsburg, 1972). Printed in Great Britain by Henry Ling Ltd., at the Dorset Press. Dorchester. Dorset 1082 ■A U.S. Government Printing Office: 1977-778-938/160 Region 8 PENN STATE UNIVERSITY LIBRARIES ADDDD7Sfl327m