/ / o O CO Digitized by tine Internet Arciiive in 2012 witii funding from LYRASIS IVIembers and Sloan Foundation http://archive.org/details/easternberingsOOhood The Eastern Bering Sea Shelf The EaBtePi Berini Sea Bhelf: ani Mesoiipcef Edited by Donald W. Hood and John A. Calder oMme IlIM e * ^^^"^ "'Co. \ \ / ^^ATES O* ^ UNITED STATES DEPARTMENT OF COMMERCE Philip M. Klutznick, Secretary NATIONAL OCEANIC AND ATMOSPHERIC ADMINISTRATION Richard A. Frank, Administrator OFFICE OF MARINE POLLUTION ASSESSMENT R. L. Swanson, Director Published in 1981 by the Office of Marine Pollution Assessment of the National Oceanic and Atmospheric Administration with financial support from the United States Department of the Interior, Bureau of Land Management Library of Congress Catalog Card Number 81-600035 1 Distributed by the University of Washington Press Seattle, Washington 98105 Affix to page iv of Volume One Contents Foreword ix Preface xi Introduction xiii I: Physical Oceanography 3 A Perspective of Physical Oceanography in the Bering Sea, 1979 Thomas H. Kinder 5 Marine CUmatology of the Bering Sea James E. Overland 15 Recent Short-period Wintertime Climatic Fluctuations and Their Effect on Sea-surface Temperatures in the Eastern Bering Sea H. J. Niebauer 23 Hydrographic Structure Over the Continental Shelf of the Southeastern Bering Sea Thomas H. Kinder and James D. Schumacher 31 Circulation Over the Continental Shelf of the Southeastern Bering Sea Thomas H. Kinder and James D. Schumacher 53 Circulation and Hydrography of Norton Sound R. D. Muench, R. B. Tripp, and J. D. Cline 77 Reevaluation of Water Transports in the Vicinity of Bering Strait L. K. Coachman and K. Aagaard 95 Tides of the Eastern Bering Sea Shelf Carl A. Pearson, Harold O. Mofjeld, and Richard B. Tripp 111 II: Ice Distribution and Dynamics 131 Recent Fluctuations in Sea Ice Distribution in the Eastern Bering Sea H. J. Niebauer 133 Remote Sensing Analysis of Ice Growth and Distribution in the Eastern Bering Sea S. Lyn McNutt 141 Nearshore Ice Characteristics in the Eastern Bering Sea William J. Stringer 167 Bering Sea Ice-edge Phenomena Seelye Martin and Jane Bauer 189 Eastern Bering Sea Ice Dynamics and Thermodynamics Carol H. Pease 213 Anticipated Oil-Ice Interactions in the Bering Sea Seelye Martin 223 III: Geology and Geophysics 245 Sedimentary Processes and Potential Geologic Hazards on the Sea Floor of the Northern Bering Sea Matthew C. Larsen, C. Hans Nelson, and Deuin R. Thor 247 The Ice-dominated Regimen of Norton Sound and Adjacent Areas of the Bering Sea Verna M. Ray and William R. Dupre 263 Ice Gouging on the Subarctic Bering Shelf Deuin R. Thor and C. Hans Nelson 279 The Role of the Kaltag and Kobuk Faults in the Tectonic Evolution of the Bering Strait Region Mark L. Holmes and Joseph S. Creager 293 IV: Chemical Oceanography 303 Some Geochemical Characteristics of Bering Sea Sediments D. C. Burrell, K. Tommos, A. S. Naidu, and C. M. Hoskin 305 The Distribution and Elemental Composition of Suspended Particulate Matter in Norton Sound and the Northeastern Bering Sea Shelf: Implications for Mn and Zn Recycling in Coastal Waters Richard A. Feely, Gary J. Massoth, and Anthony J. Paulson 321 Some Heavy Metal Contents of Bering Sea Seals David C. Burrell 339 Preliminary Observations of the Carbon Budget of the Eastern Bering Sea Shelf Donald W. Hood 347 Organic Matter in the Bering Sea and Adjacent Areas N. Handa and E. Tanoue 359 Hydrocarbons of Animals of the Bering Sea D. G. Shaw and E. R. Smith 383 Organic Geochemistry of Surficial Sediments from the Eastern Bering Sea M. I. Venkatesan, M. Sandstrom, S. Brenner, E. Ruth, J. Bonilla, I. R. Kaplan, and W. E. Reed 389 Hydrocarbon Gases in Near -surface Sediment of the Northern Bering Sea Keith A. Kvenvolden, George D. Redden, Deuin R. Thor, and C. Hans Nelson 411 vi Distribution of Dissolved LMW Hydrocarbons in Bristol Bay, Alaska: Implications for Future Gas and Oil Development JoelD. Cline 425 V: Fisheries Oceanography 445 Overview of Fisheries Oceanography Felix Favorite 447 Shelf Environment W. James Ingraham, Jr. 455 Ichthyoplankton Kenneth D. Waldron 471 Halibut Ecology E. A. Best 495 Distribution, Migration, and Status of Pacific Herring Vidar G. Wespestad 509 The Biology of Walleye Pollock Gary B. Smith 527 Population Characteristics and Ecology of Yellowfin Sole Richard G. Bakkala 553 Trans-shelf Movements of Pacific Salmon Richard R. Straty 575 Finfish and the Environment Felix Favorite and Taivo Laevastu 597 Ecosystem Dynamics in the Eastern Bering Sea Taivo Laevastu and Felix Favorite 611 Vll Foreword President Nixon, in his energy message of 23 January 1974, set up Project Independence, the goal of which was to make the nation self-sufficient in oil production by the end of the 80's. Oil and gas deposits on the outer continental shelf of the nation were estimated to be the largest single source of petroleum available. Consequently, the expeditious development of the petroleum resources of the outer continental shelf became an essential and central element in the plan to attain the goal of Project Independence. Because of the extensive outer continental shelf of Alaska and the high estimated potential petroleum resources contained therein, Alaska was quickly placed at the forefront of the nation's OCS oil and gas development program. Potential lease areas or basins were identified encompassing significant parts of Alaska's vast OCS area extending from the northeast Gulf of Alaska to the demarkation line in the Arctic. Several of the basins of most promise are in the Bering Sea, an area of exceptionally high biological productivity which historically has provided bountiful harvests of fish and other living resources to the United States and other nations. The Bering Sea supports a wide variety of marine mammals and birds, many legally designated as endangered species, and others harvested commercially by foreign nations. Alaskan natives rely heavily on many of the Bering Sea living resources as the cornerstone of their subsistence culture. The requirement to develop sufficient energy sources for the present and future decades has been identified as certainly one of the paramount problems of expanding urgency which our nation faces. Development of the petroleum resources of the Bering Sea Outer Conti- nental Shelf offers a promise of substantial benefits to the nation. Associated with these developments is an element of risk to the marine environment and living resources of the area. Public concern for the conservation and protection of the marine environment and its living resources, especially in the pristine environments of the Alaska OCS area, is well recognized. Decisionmakers in the private sector as well as in government must be respon- sive to the expressed concern about possible environmental and resource damage caused by OCS oil and gas development. Decisionmakers are required by law to develop an under- standing of the possible risks to the environment and resources, to quantify, where possible, the probability and extent of likely damage, and to utilize this information and under- standing in their decisions about OCS oil and gas development. As manager of the Outer Continental Shelf Leasing Program, the Bureau of Land Manage- ment (BLM) of the Department of the Interior (DOI) initiated in 1975 the Alaska Outer Continental Shelf Environmental Assessment Program (OCSEAP). OCSEAP is a compre- hensive, multidisciplinary environmental studies program designed to provide BLM in particular, other decisionmakers, and the interested public with a source of adequate infor- mation to help them formulate their decisions and to generate management strategies that provide acceptable protective and mitigating measures against the possible ranges of undesirable or unacceptable impacts on the marine environment and living resources. OCSEAP is managed for BLM by the National Oceanic and Atmospheric Administration (NOAA) through an interagency agreement. ix An important goal of the environmental studies program in Alaska is to enable the BLM to organize and develop policy guidance for major decision points in their OCS leasing process. The implied premise is that sound policy governing OCS leasing and petroleum development will result from adequate knowledge and understanding of the environment, the ecological processes controlling the distribution and relative abundance of important populations of marine organisms, and their vulnerability and sensitivity to OCS petroleum development. In the management of OCS oil and gas development it is recognized that the viability of policies will be enhanced if both what is known and what is not known are included in the decisionmaking process. It also is apparent that the most effective and economical use of OCS study funds can be achieved when there is a clear understanding of the status of scientific knowledge in the region of concern. This also includes both what is known and what is not known. If the status of knowledge is sufficiently documented, it enables the studies managers to identify information needs that are attainable within limits of resources and time available, and which have the greatest relevancy to the issues of concern associated with OCS oil and gas development. The rationale above was the primary consideration which caused OCSEAP/BLM to produce the treatise. The Eastern Bering Sea Shelf: Oceanography and Resources. As recently as 1970, the Bering Sea was largely a frontier area for marine science. At that time most of the research in the Bering Sea had been conducted by the Russians, with lesser contributions by Japan and the United States. Emphasis was primarily on the assessment of commercial species of fish and shellfish with lesser efforts directed toward other marine organisms, such as marine birds and mammals, biological productivity, and the physical environment. In 1974, Hood and Kelley published a review of knowledge existing before 1970, Oceanography of the Bering Sea with emphasis on renewable resources. Within the last decade our knowledge and understanding of the Bering Sea have been significantly expanded. The BLM/NOAA Outer Continental Shelf Environmental Assess- ment Program contributed substantially to our increased understanding of the outer conti- nental shelf in the Bering Sea and the living marine resources of the area. The present publication consists of a selection of 73 chapters written by acknowledged experts in their fields, giving the most accurate and comprehensive description to date of the physical environment and resources of the outer continental shelf of the Bering Sea from the Aleutian Islands to the Bering Strait. The book should be invaluable as a source of scientific data and information readily available for use in decisions regarding the manage- ment of OCS petroleum development. It will be equally valuable to OCSEAP in planning and conducting continuing environmental studies in support of BLM's needs for environ- mental and resource data to guide policy decisions. Finally, the book will be of great interest and use to the scientific community in conducting a wide range of research projects in future years. Esther C. Wunnicke Manager, Bureau of Land Management, Alaska Outer Continental Shelf Office Herbert E. Bruce Director, Alaska Office, Office of Marine Pollution Assessment, National Oceanographic and Atmospheric Administration Preface The production of these volumes, entitled The Eastern Bering Sea Shelf: Oceanography and Resources, was conceived by the Juneau Project Office of the Outer Continental Shelf Environmental Assessment Program (OCSEAP) with the aim of providing a compilation of basic information on the Bering Sea Shelf to be used by the Bureau of Land Management in its primary responsibility for implementing most of the pre-sale oil and gas lease objectives. Of even greater long-term value for the protection of the Bering Sea environment, this treatise will help to build an understanding of how the Bering Sea functions as a system. Only by understanding how this ocean functions will we be able to predict the impact of human activities on specific parts or the whole of this unusually productive environment. It is the primary purpose of this publication to present in a single document what is now known about the natural science of the Eastern Bering Sea Shelf. Some criticism has arisen that this effort is premature— that so much work is presently in progress that a later treat- ment would be more definitive. It might be desirable to wait until present work is finished, reflected upon, and analyzed, but it is not the nature of interdisciplinary scientific effort for all disciplines to reach a comfortable plateau in status of knowledge simultaneously, allow- ing for integration, substantive analysis, and thorough documentation. Instead, one disci- pline feeds on the information of another as it becomes available and understandable, thus building toward an understanding of the way the system functions as a whole. Apart from the dictates of the workings of interscience disciplines, the expediency of the times demands that the best information possible be available for use in formulating impact statements for oil and gas development in the Bering Sea during 1980 and beyond. We have, therefore, done the best we could to set down all we know about the natural science of the Bering Sea Shelf. It is apparent that our scientific knowledge of this region is well advanced and that the analysis of certain of its elements is at a level equal to or exceed- ing that of any like region in the world. Our efforts here have not produced the ultimate document of the Bering Sea Shelf— one may never be written— but this is intended to be an important benchmcirk in Bering Sea literature upon which future science and the use of science for pragmatic purposes can be based. A secondary, but important, purpose of these volumes is to provide a credible scientific document from which estimates of the effects of oil and gas development in the outer continental shelf region of the eastern Bering Sea can be made. The creation of this work has provided a medium for presenting basic scientific evidence about how the natural sys- tem works, uncomplicated by the requirement of applying the findings to specific problems of oil and gas development. The application of this information is being accomplished through other workshop and synthesis efforts. Considerable time and expense are required to produce a high-quality publication con- taining the essential information, and such a project must usually be supported by a mission- oriented department of the federal government. Despite what may be administrative diffi- culty, it is important that provisions be made to allow for appropriate scientific contribu- tion to decisionmaking processes involving the uses of the ocean, whether it be for oil and gas development or for other human needs. xi The mechanism for preparing this document consisted of several steps. First, the extent of the pertinent information available on the Bering Sea Shelf had to be determined. Knowledgeable Bering Sea scientists were assembled; they explored the amount of material that could be brought together for publication at this time, and agreed to serve as associate editors of thirteen disciplinary and interdisciplinary sections. This gave us a tentative outline. Next these same scientists met again to propose authors and extent of material in their respective areas. A symposium was held in Anchorage, Alaska, in November 1979, on which occasion much of the material was presented orally, and some drafts of papers were collected which would eventually constitute some of the chapters presented here. By 15 January 1980 a draft of the total volume, consisting of fourteen sections containing sixty of the chapters in these volumes, was furnished to the Bureau of Land Management as an interim working document. Then began the long process of editorial, typesetting, composi- tion, and printing effort to produce the final form of the book. I hope that through this exercise we have all gained insight into making scientific infor- mation serve the environmental decision-making process so that all men, present and future, may properly benefit from the bounty of the ocean's resources. Many people contributed to the organization and preparation of these volumes, but we list only those who have taken a leadership role in specific aspects of its production. Printing preparation was done by Science Applications, Inc., in Boulder, Colorado, with Joseph G. Strauch, Jr., as principal investigator, Patricia Martin Gibby as technical editor, and Robert E. Peterson as graphics editor. Much of the material included in this book was presented at a symposium held in An- chorage, Alaska, in November, 1979, organized by Leland Hepworth and John A. Calder, Jr., OCSEAP program managers for the Bering Sea Studies. Working groups at this sympos- ium were led by E. Carmack for physical oceanography, P. Becker for benthic biology, D. Hood for chemical oceanography, D. Menzel for plankton ecology, B. Farentinos for birds and mammals, P. Fischer for geologic and ice hazards, D. Nyquist for ice-edge ecosystems, and O. Mathisen for fisheries biology. Each of the fourteen sections of these volumes had an associate editor, named at the beginning of each section, responsible for selecting authors, comprehensively covering the subject, and organizing the section internally in such a way as to be coherent with the whole treatise. Funding came from the Outer Continental Shelf Environmental Assessment Program, which is administered by the National Oceanic and Atmospheric Administration for the Bureau of Land Management. The encouragement and administrative support of H. E. Bruce, Juneau Project Manager, and R. L. Swanson, Director of the Office of Marine Pollu- tion Assessment of NO A A for this effort are greatly appreciated. D. W. Hood J. A. Calder xn Introduction Donald W. Hood In the eighteenth century, Kamchatka and Alaska were better known and much more the object of international attention than was California. Should wealth be the criterion? William R. Hunt (1975) Sandwiched between the northernmost land masses of the North American and Asian continents lies the Bering Sea. It has a mean depth of 1,636 m, a surface area of 2.3 X 10^ km' and a volume of 3.7 X 10^ km^ . These dimensions make it the third largest semi-enclosed sea of the world ocean, surpassed in size only by the Mediterranean and the South China seas. It derives its name from Vitus Bering, who in 1725-43 commanded an extensive Russian expedition which explored the coasts of the Kam- chatka Peninsula to the south of the Aleutian Islands and the southern mainland of Alaska as far as Prince William Sound (Hood and Kelley, 1974). From north to south the Bering Sea can be viewed as a sector with a radius of about 1,500 km with the Bering Strait approximately at the vortex. The southern arc represented by the Aleutian Islands and the Alaska Peninsula extends between 157° W on the shores of Bristol Bay on the east to 163° E at the coast of the Kamchatka Peninsula on the west, a total distance of almost 3,000 km. Included within the Bering Sea boundaries are three major bays: Bristol Bay in the southeast and Norton Sound in the northwest, both bordering Alaska; and the Gulf of Anadyr in the northwest, bordered by Siberia. Three major rivers— the Kusko- kwim and Yukon draining central Alaska and the Anadyr draining western Siberia— empty into the Ber- ing Sea. All these rivers contain glacial melt water and runoff from taiga forests and tundra characteris- tic of the far northern environment. The Bering Sea is often considered the northern extension of the Pacific Ocean. It is true that ex- change of surface water through the Aleutian Island passes occurs relatively freely and water character- istic of the north Pacific Ocean is evident within the Bering Sea. However, since water passage is limited by the sill depths of the passes except in the west and the link with the Arctic Ocean is the narrow (85 km), shallow (45 m) Bering Strait, this sea possesses the essential features of a well-defined ocean region and may be best looked upon as a con- fined sea with characteristics of its own sufficiently different from adjoining oceans to make it a unique ocean region. The eastern Bering Sea shelf (P^ig. 1), about one- half (1.2 X 10^ km' ) the total area of the Bering Sea, is exceeded in size only by the shelf common to the Chukchi, East Siberian, and Laptev seas, within the geographic boundary of the Arctic Ocean. Be- sides its size, it has several other features which may contribute significantly to the way this shelf func- tions oceanographically to make it one of the most productive of the world's ecosystems. Three features clearly differentiate the eastern Bering Sea shelf from most other shelves for which there is a sufficient data base for comparison. First, there is a recurring, although variable in extent, annual ice cover that in cold years reaches as far south as the Alaska Peninsula. Second, it is vented at the northern end by the Bering Strait, through which passes about 2 X 10^ m^ of water per second (2 Sv). This water has its origin in the North Pacific, with a small contribution (10 percent) from Alaskan rivers, and yet little of it penetrates the southern shelf area. Third, there are three persistent fronts, most domi- nant in the summer season, that lie roughly over the shelf break (150 m) and the 100-m and 50-m con- tours in the southeast shelf region. These fronts have great influence on the southeast Bering Sea and possibly the entire shelf ecosystem. Other features such as topography, wind patterns, and general climate also contribute heavily to the function of this shelf. The Bering Sea ice cover and ice dynamics aire the subject of chapters 9-14 of this volume; these repre- sent a substantial contribution to knowledge about physical and meteorological features that influence the dynamics of the ice system. When ice forms in the fall, it acts as a physical barrier to direct sea-air interaction and also as an effective insulator against energy transfer between the near-freezing shallow waters of the shelf and the overlying Arctic air mass. The formation of ice and the resulting increase in density of the surface water combine with wind mixing on the surface and tidal mixing at depth to cause complete vertical homogenization of the water column over the whole of the shelf during the early winter. The advances and retreats of the ice are variable, depending on broad-scale meteorological events, sea and air temperature, and surface winds. The period 1973-78 was a time of extreme fluctua- tions in ice conditions: the years 1973-76 were characterized by below-normal temperatures which brought about abnormally high ice coverage, which reached as far as Unimak Pass in 1976 and persisted well into May before sufficient warming occurred to Xlll 170° 175" 180° 175° 170° 165° 160° 155 150° I ^, [^ I — ' Ch%kchi Sea 65' 62' 59° 56° 53° 100 200 krr 50 0 50 100 miles ^ IGulf, of Anady iC phukotsk Peninsula ^( « ^ ,^ J |( Seward ^<^ J^medB Peninsula "-^'^m Islands Norton Sound n^::Xape St. Lawrence Island ' ^ukon River Navarin ^--* /Kuskoqulm River 65 62' 59'^ 56' 53"' 180 175" 170° 165' 160 Figure 1. The eastern Bering Sea shelf. bring on the spring melt. On the other hand, 1977-79 was a period of high air and sea temperatures and limited ice coverage brought on by southerly shifting winds (See Niebauer, Chapter 9, this volume). These widely varying physical conditions greatly influenced the primary productivity, and thereby all other biotic components, of the shelf region, as described in Sections VII and X of Volume 2. The oceanographic significance of the shallow northern opening of the Bering Sea through the Bering Strait into the Chukchi Sea is yet to be fully ascertained, particularly as it contributes to the whole of the biological system. The pathway to the Arctic from the Pacific through the Bering Sea has long been known to be of major importance to some migrating whales (See Frost and Lowry, Chapter 50, Volume 2). Furthermore, the significance of the Bering Strait passage for walrus, seals, and beluga and bowhead whales between the Arctic and Bering Sea is discussed by Burns, Chapter 46 and Lowry and Frost, Chapter 49 of Volume 2. Although some of the highest short-term primary production rates XIV ever measured in the world's oceans were found in the Bering Strait (McRoy et al. 1972), the processes which caused this production and its quantitative extent have not been elucidated. The influx of water into the Bering Sea through the Aleutian passes has been summarized by Favorite (1974). Several conflicting opinions persist as to the amount and location of flow through the passes. Arsenev (1967) believed that the Komandorski/Near islands Strait is the most significant source of inflow (14.4 Sv) with a considerably reduced flow in the central Aleutian passes (4.4 Sv); no net exchange was considered to occur in the eastern Aleutian passes. It is clear that much of this water must return to the Pacific through deep passes (Hughes et al. 1974), since the outflow to the Arctic Ocean is only 1-2 Sv and the Bering Sea is generally considered to have a positive net influx of fresh water. Of equal or perhaps greater importance to the pro- ductivity potential of the Bering Sea is the vertical transport (upwelling) caused by water flow through the passes as described by Hood and Kelley (1976) and Swift and Aagaard (1976). This transport was shown to be responsible for providing about 5 X 10'' g of nitrate nitrogen per km^ /day to the surface waters of a region of several thousand square kilo- meters observed north of Samolga Pass. Other passes are expected to contribute similarly. The fate of nitrate and other accompanying nutrients brought to the surface by upwelling in the Aleutian passes has not been determined, but this upwelling could be another important reason for the high biopro- ductivity observed in the Bering Sea. The third unique feature of the shelf, that of three ocean fronts, is a recent discovery made possible by detailed hydrographic surveys of the region through the Outer Continental Shelf Environmental Assess- ment Program (OCSEAP), sponsored by the Depart- ment of the Interior, and Processes and Resources of the Bering Sea (PROBES) studies sponsored by the National Science Foundation. The details of these oceanic structures are described in Kinder and Schumacher (Chapter 4, this volume) and the bio- logical significance to plankton ecology by Goering and Iverson (Chapter 56, Volume 2) and Cooney (Chapter 57, Volume 2). Historical setting The north Pacific Ocean south of 49° N lati- tude between the coasts of America and Japan was fairly reliably described on maps dating from the late sixteenth century; however, almost total ignorance prevailed about the region north of that transect. It was not until 1643 that a Dutch East India com- pany expedition commanded by Maerten Gerritszoon Vries discovered a previously much talked about mystery land, Yezo, to the north of Japan— Yezo is now known as the Kurile Islands. In 1639 a bargeload of Cossacks made their way down a Siber- ian river to the Sea of Okhotsk for a first view of the "Icy Sea" to the north; the Amur, Ud, Anadyr, and Okhota rivers flowed south to the "Eastern Sea." Still there was no knowledge of a passage be- tween the two bodies of water. The question of whether America and Asia were united attracted the attention of Peter I, Czar of Russia from 1682 to 1725, who initiated a program of westernization by utilizing Swedish prisoners of war in Siberian ports as teachers of shipbuilding and navigational arts. After the death of Peter I, the program was continued by his successor, Empress Catherine, when Vitus Bering, a Danish explorer, and Lt. Alexei Chirikov were commissioned for their first expedition. In 1728, Vitus Bering, with the two ships St. Peter and St. Paul, sailed through the Bering Strait to settle once and for all that Asia and America were separated. On the second voyage, fatal to Vitus Bering, who died of scurvy in 1841, much of the Arctic coast of Asia was charted, as well as the southwest Alaskan coast and the Aleutian Island arc. Even though Vitus Bering had remarkable success, his work was doubted and in fact discarded to some extent. The pressing question of a northern passage from the Pacific Ocean led the British Ad- miralty to issue instructions to their most talented explorer. Captain James Cook, to extend his third voyage in 1778-80 in further search of a northeast or northwest passage, from the Pacific into the Atlantic Ocean or the North Sea of Europe by way of Russia. At the helm of the Resolution and accom- panied by Lt. Charles Clarke on the Discovery, Cook passed through the Bering Strait into the Arctic Ocean in the summer of 1778. The expedition penetrated to 70° N, where Cook viewed the "Chuckchee" Sea and the extremities of both conti- nents, and the possibility of a northwest or northeast passage to India so long sought from the Atlantic Ocean side had to be renounced. The remarkable accuracy of Bering's earlier cartography was demon- strated. The expedition, on its way south to wait out the Arctic winter, visited Norton Sound and many inlets in south-central Alaska. The following sum- mer's expedition again sailed through the Bering Strait, but without Captain Cook, who had been murdered during the winter of 1779 by Sandwich Island natives. XV Physical dimensions and ice cover The Bering Sea is divided into a neritic area (0-200 m) and an abyssal region (over 1,000 m) of about equal extent, the sum totaling 87 percent of the entire area of 2.3 X 10^ km^ . The continental slope, representing only 13 percent, generally has a slope of 4-5° and essentially divides the Bering Sea in half in a northwest-southeast direction. The shelf feature of the eastern half continues north through the narrow Bering Strait into the Chukchi Sea. The shelf region contains several islands, important to marine mammal and bird ecology, which influence the circulation of water and the formation and movement of sea ice. St. Lawrence, the largest of the five major shelf islands, is nearly 200 km long and lies between Norton Sound and the Gulf of Anadyr, south of the Bering Strait. Nunivak Island, the sec- ond largest, lies near the coast of Alaska between the mouths of the Kuskokwim and Yukon rivers. St. Matthew, St. Paul, and St. George islands, the other three large islands, lie well offshore. St. Mat- thew is about midshelf, approximately 275 km south- west of St. Lawrence, whereas St. Paul and St. George, the fur seal islands, are near the shelf break some 300 km northeast of Unimak Pass. The deep basin of the Bering Sea, beginning west of Unimak Pass, is separated from the Pacific Ocean by the Aleutian Island arc, which rises from 7,600 m on the Aleutian Trench side and 4,000 m on the Bering Sea side. Passes through the arc deepen westward, the deepest being 4,420 m between Kamchatka and the westernmost Komandorski Island. North of this passage a sill at 3,589 m provides the deepest passage to the basin. Except for a narrow passage at 180° E, the sill depth east of 171° E is much less than 1,000 m. The total cross section permitting exchange of water through the southern boundary is only 701 km^ (Favorite 1974). The Bering Sea basin is a vast plain lying at a depth of 3,800-3,900 m with occasional gradual sloping hollows to depths of as much as 4,151 m. Two sub- marine ridges penetrate the basin. The Shirshov Ridge extends south from Kamchatka along 170° E longitude to near the Aleutian arc and separates the eastern and western Aleutian Basin. The North Rat Island Ridge (Bowers Bank) extends 300 km north in a counterclockwise direction. The four major rivers discharging into the Bering Sea are the Yukon, Kuskokwim, Kamchatka, and Anadyr. The flow of these rivers is much greater in the summer months, since the major runoff in all these drainage basins is from melt water. The Yukon, the largest, has a peak flow in August about equal to that of the Mississippi, and a mean flow for the year of about two-thirds that of the Columbia (for details see Ingraham, Chapter 29, this volume). Nonrenewable resources The nonrenewable resources most likely to be found in the Bering Sea are hydrocarbons and heavy metals. About 75 percent of the shelf airea holds some likelihood of hydrocarbon accumulation, and some areas of the shelf are extremely promising. The most promising areas are the St. George basin on the outer margin of the continental shelf, the North Aleutian basin, the Navarin basin, and the Norton basin. These sites have a long geological history of high organic productivity. The outer basins have 1-2 km of Oligocene to Recent deposits, mainly consisting of diatomite and diatomaceous sand, with some conglomerate. The North Aleutian Basin contains several kilometers of terrigenous and volcanogenic sediments. The Norton Basin was probably inundated during the late Miocene, and since the basin was probably receiving continental sediments from the late Oligocene, reservoir rocks should be common. The Yukon River has been a major source of sediment since the middle Miocene, and probably large sources of natural gas are present. The occurrence of gold, platinum, and tin placers on the shores of the Bering Sea is well known, but exploration for offshore deposits has thus far been disappointing. Almost all placers found have been in the nearshore zone, generally less than 5 km from shore. The only significant discoveries are rich, but small, submerged deposits of placer gold off Bluff and much larger, low-grade gold deposits on the sea floor off Nome. Commercial production from the latter deposits began in 1975. The search for tin in and near the Bering Strait and for platinum in and near Goodnews Bay has yet to yield more than minor results. Renewable resources The borders and islands of the Bering Sea have been inhabited by Eskimos in the northern parts and by Aleuts along the Aleutian arc as long as man has a historic record. These natives subsisted on a hunting and fishing economy for centuries in com- patibility with an abundance of fish and animals in the area. The ancient cultures are still viable in some areas (Fig. 2). Not until western man began exploiting the region in order to supply large popula- tions of people with furs, fish, and marine mammals was the natural living wealth of the area affected. As a result the most valuable asset, the sea otter, was exploited to the point of near extinction (Schneider, Chapter 51, Volume 2), and the Steller sea cow lasted XVI Figure 2. Eskimos walrus hunting in the Bering Sea (photograph courtesy of John J. Burns). only 27 years between discovery and extinction (Hunt 1975). At least part of the herring's ecological niche appears to have been replaced by the Alaskan pollock (Wespestad, Chapter 32, this volume); the sockeye salmon (Straty, Chapter 35, this volume), once endangered and still under heavy pressure, is being restored, but international cooperation will be necessary in order to avoid further, even acciden- tal, depletion of the stocks. Throughout recent history the wide diversity of the living resources of this region has been known, but only recently has it been fully appreciated (Favorite, Chapter 28, this volume). Large harvests of pollock, cod, ocean perch, black cod, halibut, rattails, tanner and king crab, and, more recently, shrimp are now being taken by many of the major fishery nations of the world. Until recently Japan harvested by far the most, followed by the U.S.S.R.; but since the passage of the 1977 Fisheries Conservation Act the United States is catching up and now clearly dominates the crab fishery. The commercial catch of finfish is discussed by Bakkala (Chapter 61, Volume 2) and of crabs by Otto (Chapter 62, Volume 2). The potential for other fisheries is treated by McDonald et al. (bivalve mollusks. Chapter 67); Hughes (surf clams. Chapter 68) and Macintosh and Sowerton (gastropods, Chapter 69, Volume 2). The large annual tonnage constituting the human harvest of fish represents less than 4 percent of the nekton and benthos required to support the extremely large population of mammals and birds dependent upon this resource (Laevastu and Favorite (Chapter 37, this volume). xvii Conclusion Due to its latitude, the Bering Sea lies in a region with great annual variations in properties. Incident radiation in the northern region varies annually from almost total daily darkness to total light; the vi^ind torque over the sea is an order of magnitude greater in winter than in summer; and the extensive ice cover in the winter is totally absent in the summer. Al- though warm seas contain more diverse popula- tions, the colder seas support much larger individual populations. The Bering Sea has one of the largest marine mammal populations, possibly the largest clam population, one of the largest salmon runs, one of the highest densities of birds, and the largest eelgrass (Zostera) beds in the world; the yields of its pelagic and benthic commercial fisheries are extremely high. Today the Bering Sea is the focus of more scien- tific research than at any time in its history. Three major efforts— the baseline studies sponsored by the Outer Continental Shelf Environmental Assessment Program of the Bureau of Land Management, U.S. Department of the Interior; intensified National Marine Fisheries efforts because of passage recently by the U.S. of the Fisheries Conservation and Man- agement Act of 1977; and the International PROBES study sponsored by the Office of Polar Programs of the National Science Foundation— are now in prog- ress. This treatise, therefore, might best be consid- ered an interim progress report on the Bering Sea. In the immediate years ahead much more information will doubtless be published than is now available. It is hoped that this document will provide an historical baseline and point the way to further studies of this significant region of the world's oceans. REFERENCES Arsenev, V. S. 1967 Currents and water masses of the Bering Sea. (In Russian, English summary.) Izd. Nauka, Moscow. (Transl., 1968 Nat. Mar. Fish. Serv., Northwest Fish. Center, Seattle, Wash.) Favorite, F. 1974 Flow into the Bering Sea through Aleutian Island Passes. In: Oceanog- raphy of the Bering Sea, D. W. Hood and E. J. Kelley, eds., 3-37. Inst. Mar. Sci., Occ. Pub. No. 2, Univ. of Alaska, Fairbanks. Hood, D. W., and E. J. Kelley 1974 Introduction. In: Oceanography of the Bering Sea, D. W. Hood and E. J. Kelley, eds., xv-xxi. Inst. Mar. Sci., Occ. Pub. No. 2, Univ. of Alaska, Fairbanks. Hood, D. W., and J. J. Kelley 1976 Evaluation of mean vertical transports in an upwelling system by CO2 measurements. Mar. Sci. Comm. 2:387-411. Hughes, F. W., L. K. Coachman, and K. Aagaard 1974 Circulation, transport and water ex- change in the western Bering Sea. In: Oceanography of the Bering Sea., D. W. Hood and E. J. Kelley, eds., 59-98. Inst. Mar. Sci., Occ. Pub. No. 2, Univ. of Alaska, Fair- banks. Hunt, W. R. 1975 Arctic passage. Sons, New York. Charles Scribner's McRoy, C. P., J. J. Goering, and W. E. Shiels 1972 Studies of primary productivity of the Bering Sea. In: Biological oceanog- raphy of the northern North Pacific Ocean, 199-216. Idemitsu Shoten, Tokyo. Swift, J. H., and K. Aagaard 1976 Upwelling near Samalga Pass. Limnol. Oceanogr. 21:399-408. xvm The Eagtem Bepimi Sea Shelf: ani Mesoiireet Section I Physical Oceanography Thomas H. Kinder, editor A Perspective of Physical Oceanography in the Bering Sea, 1979 1 I Thomas H. Kinder Naval Ocean Research aind Development Activity Bay St. Louis, Mississippi ABSTRACT Until recently, physical oceanographic research in the Bering Sea concentrated on broad spatial and long temporal scales, and much of the field work occurred off the shelf in water overlying the deep basins. Research concentrated on basin-wide phenomena of long duration, and this work deter- mined the oceanic climate or physical geography of the Bering Sea. Since about 1975, the focus of research has shifted toward shorter spatial and temporal scales, and also from the deep basins onto the shelf. Deviations from the large-scale mean state, such as interannual variability, fronts, eddies, tides, and vertical finestructure, are important biologically as well as physically, and this trend in research will probably continue through the next decade. INTRODUCTION About one-half of the Bering Sea overlies abyssal plain (depths >3,500 m), and about one-half overHes continental shelf (depths <200 m, Fig. 1-1). Until 1975 most physical oceanographic research concen- trated on either the waters above the deep basins or those above the continental shelf near Bering Strait. This bias was illustrated in the symposium volume edited by Hood and Kelley (1974): physical oceano- graphy papers concentrated on currents and water masses of the deep basins, and the monograph by Coachman, Aagaard, and Tripp (1975) focused on Bering Strait. With the exception of the Bering Strait region, little research was done on the continental shelf. Much of the work before 1975 emphasized deter- mination of mean conditions. Data were so few and so far apart, both in time and space, that variability could not be addressed rigorously. Investigators realized that smaller spatial and temporal scales were important, but they did not have adequate measure- ments to define these scales. Hydrographic station spacing was often 100 km and stations were reoc- cupied only annually, if at all. Except for Bering Strait, direct current measurements were essentially nonexistent. Most work before 1975 aimed at establishing the mean state of the Bering Sea on large spatial scales, emphasizing the deep basin. Since 1975 the focus has shifted from the longer temporal and broader spatial scales to shorter ones, and from the deep basins to the continental shelf. Investigations of interannual variations on broad spatial scales (climatic variability) and research in the deep basins continue, but the emphasis now lies elsewhere. Developments in instrumentation, trends in physical oceanography, and changes in funding have brought about this new emphasis. The large spatial and temporal scales inherent in earlier work were primarily determined by the instru- mentation and resources that were available. Hydro- graphic profilers (STD (salinity, temperature, depth) or CTD (conductivity, temperature, depth)) have increased vertical resolution from 10 m (or greater) to 1 m. Horizontal resolution was increased by generous amounts of shiptime dedicated to physical projects so that scales of 5 km or less were sampled. Satellite imagery permitted synoptic realizations of the surface thermal patterns and ice fields to scales of less than 1 km. Drogued drifting buoys tracked by satellite provided long tracks over large areas. Moored instru- ments measured currents to scales shorter than one hour. Sufficient shiptime was provided to repeat hydrographic surveys during a season, and to main- tain current meter moorings for more than one year. Horizontal and vertical spatial resolution and temporal resolution were improved by one order of magnitude. Repeated measurements made it possible 6 Physical oceanography to estimate persistence or change at these finer scales over the seasons and during different years. Interest in these smaller scales was partly stimulated by growing theoretical and observational evidence in the oceanographic community that under- standing broad-scale phenomena depended on under- standing smaller-scale features. For instance eddies (MODE Group 1978), fronts (J. Geophys. Res. 83 (C9) 1978), finestructure (J. Geophys. Res. 83 (C6) 1978), and shelf dynamics {Memoires de la Soci^te Royale des Sciences de Liege, X 1976) were becoming increasingly important in the scientific literature. Renewed interest in shelf dynamics, overcoming a physical oceanographic bias toward deep water work, was partly the result of a new perception. Before the mid-1979's physical oceano- graphers apparently believed that the continental shelves were rather chaotic, not amenable to study with standard oceanographic techniques, and in any case not as important as the deep ocean. Over the last decade we have found the shelves to be well organized (both hydrographically and dynamically), predisposed to fruitful study, and certainly important in their own right. In the Bering Sea, research on such topics as eddies, fronts, finestructure, and shelf dynamics naturally proceeded, based on the know- ledge of broad-scale mean conditions established by earlier work. The geographical shift was abetted by two large research programs: OCSEAP (Outer Continental 160'E I64'E I68'E 172' E I56"W 152* W ALASKA J t I I J I L_ AMCHITKA 4 UJ Q Q tr < o z < (/) X in UJ cr I- < < UJ 5 OCT NOV DEC JAN FEB MAR APR MAY JUN JUL AUG SEP ST. PAUL Figure 2-2. Annual temperature march and monthly standard deviations for Northeast Cape and St. Paul. There is greater annual range at the northern station. CLIMATOLOGICAL SUMMARY Data for this section were obtained from Climatic Atlas of the Outer Continental Shelf Waters and Coastal Regions of Alaska: Volume II— Bering Sea (Brower et al. 1977) and A Climatological Guide to Alaskan Weather (Grubbs and McCollum 1968). The data periods in Grubbs and McCollum "contain at least 10 years" while Brower et al. specify the number of observations. Fig. 2-1 is a map locating stations and marine areas used in the figures and tables. Marine areas A to E designate regions where ship reports form the climatological base. Temperature Since this region is oriented north-south, there is some latitudinal variation in temperature. This is more noticeable in winter when the total amount of sunshine varies from several hours in the south to almost none in the north. In addition the southern portion in winter may come under the influence of southeasterly flow of fairly warm, moist air from the north Pacific, while the northern portion is under the influence of cold, dry air generally flowing out of the interior of northern Alaska. While the ice serves as a barrier between air and sea, some heat is diffused through the ice from the much warmer ocean so that, although the arctic air is very cold, temperature minimums do not reach the low extremes of the Alaskan interior. Fig. 2-2 plots the mean monthly temperatures and Marine climatology 17 standard deviations for Northeast Cape (approx- imately 3,000 observations per monthly group) and for St. Paul (over 4,000 observations per monthly group) after Browser et al. (1977). Fig. 2-2 shows much larger annual range in the north. The August- minus-February temperature difference is 25 C, versus a 14 C difference at St. Paul. The standard deviation of monthly temperatures at Northeast Cape is greater in winter than in summer and greater than the standard deviations for all seasons at southern stations. Table 2-1 presents the mean maximum and minimum temperatures (°C) by month for Northeast Cape and Adak after Grubbs and McCollum (1968). When the mean daily ranges of temperatures (monthly mean maximum minus monthly mean minimum) for summer and winter are compared. Northeast Cape shows an 8 C range in late winter and a 4 C range in summer, while the range at Adak is 4-5 C, independent of season. The greater temperature ranges at Northeast Cape point to a continentality in the north during winter which is replaced by a maritime climate in summer. TABLE 2-1 Maximum and minimum mean monthly temperatures and monthly precipitation for Northeast Cape and Adak Temperatures in ° C Precipitation in mm NORTHEAST CAPE JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC Mean maximum temperature -12 -13 -9 -6 2 Mean temperature -16 -17 -13 -9 -31 Mean minimum temperature -19 -21 -17 -13 -3 Mean total precipitation 12.9 10.2 21.1 9.4 13 7 11 10 4 8 8 16 6 7 2-3 -13 5 0-4 -15 -2 -7 -17 21.1 9.4 13.7 12.9 28.2 98.8 113.0 56.4 39.9 14.5 ADAK JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC Mean maximum temperature 3 3 Mean temperature 1 1 Mean minimum temperature -1 -2 Mean total precipitation 160.0 134.6 4 6 7 2 3 5 -1 165.1 106.7 116. 9 12 13 11 7 9 11 9 8 5 3 6 3 1 83.8 76.2 104.1 137.2 185.4 193.0 198.1 18 Physical oceanography Precipitation and cloud cover There is a general decrease in the amount of precipitation from south to north because the northern points are more distant from the moisture source, especially in winter. Table 2-1 shows that precipitation at Northeast Cape is low from December through June, when the northern region is dominated by the arctic air mass. There is a sharp spike during August and September, when storm tracks penetrate the northern Bering Sea. Adak has precipitation the year round with an increase in October through December which corresponds to a season of cyclogenesis in the southern Bering Sea. One of the most important controls effected by cloudiness is of the type of air which is synoptically in possession of the area. In winter the majority of the Bering Sea region is most frequently under north- easterly or northerly flow of cold, dry arctic air. In summer the entire region is under the influence of moist air from a north Pacific air mass. This leads to a larger number of clear days in the northern region in winter. However, even though the arctic air contains only a small amount of moisture, the cold air mass exhibits high relative humidities near the surface. Only a slight amount of lift, for example from a weak storm system, is required for formation of cloud cover. Fig. 2-3 shows the percent of obser- vations by month in which the observed cloud cover was five-eighths or more for Northeast Cape, St. Paul, and Adak (Brower et al. 1977). NORTHEAST CAPE FEBRUARY SPEED CLASSES 1-6 KM 7-16 KN = 17+ KN 0 10 20 30 40 50 SCALE (IN PERCENT OF TIME) ST PAUL MARINE AREA A MARINE AREA B MARINE AREA C MARINE AREA D MARINE AREA E Figure 2-4. February wind roses. (Direction from which the wind is blowing.) OCT NOV DEC JAN FEB MAR APR MAY JUN JUL AUG SEP 100% 80% 60%- 40%- - 20%- --- NORTHEAST CAPE - SAINT PAUL — ADAK Figure 2-3. Percent of observations reporting five-eighths cloud cover or greater. Surface Winds Fig. 2-4 shows wind roses for selected locations in the Bering Sea during February and Fig. 2-5 shows roses for the same stations during August. Wind roses show the percentage of observations from each of eight possible directions. The number in the center is the percent of light winds in the record. Each arm is divided into the percent of observations of 1-6 kn, 7- 16 kn, and 17 kn or greater from each direction. In winter the northern stations show a high per- centage of winds greater than 17 kn from the north and northeast, while in the south, the winds over marine area C are uniformly distributed over direc- tion with moderately large speeds. This marine rose is indicative of a fairly continuous progression of storms through the area. Wind speeds over the Bering Sea in summer are generally lower than in winter, Marine climatology 19 NORTHEAST CAPE AUGUST ST PAUL SPEED CLASSES I-6KN 7-16 KM =^ 0 10 20 30 40 50 SCALE (IN PERCENT OF TIME) MARINE AREA A MARINE AREA B MARINE AREA C MARINE AREA D MARINE AREA E Figure 2-5. August wind roses. (Direction from which the wind is blowing.) although conditions are seldom calm. Marine area A to the north shows little preferred direction, but the other stations show predominance of south and southwest winds. Runoff Freshwater inflow to the eastern Bering Sea is primarily from the Yukon, with only minor contri- butions from other sources, most notably the Kuskokwim. Fig. 2-6 plots the mean monthly discharge rates for the combined Alaska runoff into the Bering Sea and for the Yukon for 10 water -years 1967-77 (U.S.G.S. 1977). The 10-year monthly mean discharge rates show very little flow and varia- bility of flow during the months of December, January, February, March, and April. The months of greatest flow are also the months of greatest varia- bility: May and June. CIRCULATION There are two general approaches to classifying climatological types: a synoptic climatology which regards circulation patterns as an implicit function of the static sea level pressure (SLP) distribution (Barry and Perry 1973), and a kinematic approach in which synoptic weather maps are classified in terms of principal storm tracks (Klein 1957). Two synoptic climatologies which refer to the Bering Sea are those of Putnins (1966) and Barry (1978). Putnins establishes 22 patterns "in such a way that for every date of this period (1 January 1945 to 31 March 1963) a specific baric weather pattern could be assigned." Unfortunately, Putnins' emphasis centers on continental Alaska and applies only in a general manner to the Bering Sea. Barry developed a synoptic climatology for the Chukchi Sea which has 22 types and includes the maritime region of northern Bering Sea. Barry states that in winter his Type 1, arctic high pressure with subpolar easterlies at Kotzebue (Fig. 2-7) is dominant and is associated with a low- level atmospheric temperature inversion. Interrup- tions by anticyclonic systems are the most common. They are associated with cold continental air masses which reinforce the shallow arctic bound- ary layer. Cyclonic interruptions are less common DISCHARGE RATE IN CUBIC FEET PER SECOND OCT NOV DEC JAN FEB MAR APR MAY JUN JUL AUG SEP 600,000 400,000 200,000 / 1 / / \ \ 1 / 1 / / / / / 1 1 t 1 1 1 i \ ) \ \ \ \ s s \ ^^\ ttr—i ^^^ ^ / / / / / / / / / / / / // 1 'i 1 1967-1977 MONTHLY MEAN FOR THE KUSKOKWIM, YUKON AND KOYUKUK RIVERS 1967-1977 MONTHLY MEAN FOR THE YUKON RIVER AT RUBY Figure 2-6. Runoff rate as a function of month for the Yukon River and for total runoff into the eastern Bering Sea. 20 Physical oceanography 64° N 164° 160° I56°W Figure 2-7. Barry's (1978) dominant winter Type 1 sea-level pressure weather pattern. and bring warm air with larger amounts of cloudiness. Barry states that in summer there is a great variety of types, with both cyclonic and anticyclonic types apparent. Type 3 (Fig. 2-8) is the most common in July and is closest to the mean monthly SLP charts for summer given by Brower et al. (1977); one interpretation is that atmospheric circulation in the southern Bering Sea can be more readily charac- terized by mean storm tracks or the presence of low-pressure centers in certain sectors of the region than by static weather types. Fig. 2-9 plots the average number of low-pressure centers observed in a 10° X 10° latitude-longitude area during the nine-year period of record 1966-74. These areas are NW (60°-70°N, 170°-180°W), NE (60°-70°N, 160°- 170°W), SW (50°-60°N, 170°-180°W), and SE (50°-60°N, 160°-170°W). It is apparent that southern sectors have two to three times more storms than the northern sectors. However, the monthly variability is high, suggesting that the period of record is too short to make comparisons between months. OCT NOV DEC JAN FEB MAR APR MAY JUN JUL AUG SEP 6 I - - I 1 . ..-4- (>. ''%<■'■-.. i c' ■ .*^v ^^[^ ^'x . Ni 1 \ -^■* .-'■■• 4 ^^ ""■y L N ^~~ — 4 >. ^ 1 >-^^^ 1966-1974 MEAN NUMBER OF LOW CENTERS OBSERVED WITHIN 60°-70°N and 170°-|80°W 1966-1974 MEAN NUMBER OF LOW CENTERS OBSERVED WITHIN 60°-70°N and I60°-|70°W 1966-1974 MEAN NUMBER OF LOW CENTERS OBSERVED WITHIN 50°-60°N and I70''-I80''W 1966-1974 MEAN NUMBER OF LOW CENTERS OBSERVED WITHIN 50°-60°N and I60''-I70°W 64° N ~? Chukolsk \ Bering ^ ^->-Peninsulaj^5/Aa///>^g„g^j Peninsula 160° I56°W Figure 2-8. A frequently occurring summer weather type, from Barry (1978). Figure 2-9. Frequency of low-pressure systems by month for the northern and southern Bering Sea. CLOUD STREETS The advection of cold air southward from the north and northeast Bering Sea in winter produces ideal conditions for convective cloud development over the relatively warm waters to the south. The air is virtually unimpeded as it flows south across the ice, and the ice edge forms a sharp line of demarcation where sea temperatures can be as much as 15 C warmer than the air. As the air continues to flow south it is progressively destabilized by the upward transfer of heat and moisture from the ocean. The most frequently observed patterns are of the type displayed in Fig. 2-10, which shows uniform cloud streets at intervals averaging 5-6 km forming 20-70 km to the south of the ice edge and aligned in the direction of the surface wind (Streten 1975). They extend some 200-300 km downstream, display- ing only a smaU increase in cloud element dimensions. 0) < o o -^^ 3 O in QO C -a 3 O C o 00 CO CM o a CM 3 2i 22 Physical oceanography Beyond this distance there is a sharp transition from parallel streets to an open cell convection pattern with cloud elements 10 km in width separated by 20-25 km. With increasing distance from the source, the open areas grow to 50 km with cells of 20 km. ACKNOWLEDGMENTS This research is PMEL contribution No. 446. It was supported by the Bureau of Land Management through interagency agreement with the National Oceanic and Atmospheric Administration, under a multiyear program responding to needs of petroleum development of the Alaskan continental shelf, man- aged by the Outer Continental Shelf Environmental Assessment Program (OCSEAP) Office. Assistance from S. Schoenberg and the PMEL graphics and word-processing center is gratefully acknowledged. The anonymous reviews of the manuscript contri- buted to strengthening the final draft. Grubbs, B. E., and R. D. McCollum Jr. 1968 A Climatological guide to Alaskan weather. Scientific Services Section, 11th Weather Squadron, Elmendorf AFB, Alaska. Klein, W. H. 1957 Principal tracks and mean frequencies of cyclones and anticyclones in the Northern Hemisphere. U.S. Weather Bureau, Research Paper 40. Koppen and Geiger 1930 Handbuch der Klimatologie. Berlin. 5 Vols. Putnins, P. 1966 The sequence of basic pressure patterns over Alaska. Studies on the Meteor- ology of Alaska 1st Interim Report, Environmental Data Service, ESSA, Washington, D.C. Streten, H. A. 1975 Cloud cell size and pattern evolution in Arctic air advection over the north Pacific. Arch. Met. Geophys. Biokl., Ser. A, 24: 213-28. REFERENCES Barry, R. G. 1978 Study of climatic effects on fast ice extent and its seasonal decay along the Beaufort-Chukchi coasts. In: Environmental Assessment of the Alaskan Continental Shelf IX, Trans- port, NOAA, Environmental Research Laboratories, Boulder, Colo., 704-719. U.S.G.S. (United States Geological Survey) 1967-77 Water resources data for Alaska for water year 1977, Part I; Surface Water Records. U.S. Dep. of the Int. Geol. Surv.— Water Res. Div. Barry, R. G., and A. H. Perry 1973 Synoptic Climatology, Methods and Application. Methuen & Co., London. Brower, W. A. Jr., H. F. Diaz, A. S. Prechtel, H. W. Searby, and J. L. Wise 1977 Climatic atlas of the outer continental shelf waters and coastal regions of Alaska. AEIDC and NOAA, Anchor- age, Alaska. Estienne, P., and A. Godard 1970 Climatologie. Armand Colin, Paris. Recent Short-period Wintertime Climatic Fluctuations and Their Effect on Sea-surface Temperatures in the Eastern Bering Sea H. J. Niebauer Institute of Marine Science University of Alaska Fairbanks, Alaska ABSTRACT Upper air (700 mb) winter pressure patterns have shown sharp fluctuations over the period 1963-78. Mean annual sea- surface temperature (SST) fluctuations appear to be an effect of these short-term climatic fluctuations. The mid-1960's were a time of southerly flow of air leading to above-normal SST. A rather sharp reversal in atmospheric conditions led to a sharp drop in SST in the early to middle 1970 's. Since 1977, the upper air flow has become southerly, leading to a sharp rise in SST. Autocorrelation analysis of the SST suggests that these trends persist for at least two years. INTRODUCTION Since the early 1960's there have been several unusually strong short-term climatic fluctuations (time scale of one to several years) over the Northern Hemisphere. The most recent of these fluctuations is best illustrated by the severe North American winters of 1976-79. However, this latest fluctuation has also caused abnormally warm winters in Alaska and the adjacent areas. This chapter examines these short- term climatic variations and their relationship to fluctuations in sea-surface temperature (SST) in the eastern Bering Sea. Favorite and McLain (1973) did a computer analysis of two million observations of SST for the North Pacific Ocean for 1953-60. Evidence is given for an orderly transpacific movement of extensive warm and cold surface-water anomalies. They suggest that these features move about the North Pacific gyre with periods of five to six years and profoundly affect environmental conditions off the West Coast of the United States and Canada. These fluctuating environmental conditions are probably related to fluctuations in SST in the Bering Sea discussed in this chapter. The eastern Bering Sea SSTs in the mid-1960's were generally above normal (Niebauer 1978). Johnson and Seckel (1976) pointed to a climatic shift which occurred in the early 1970's with drastic effect on some of the Bering Sea fisheries. They cited low salmon catches in 1973 and 1974 as attributable to the unusuailly cold winters of 1971 and 1972. McLain and Favorite (1976) related the cold SST to large-scale changes in the atmospheric circulation which caused increased northerly winds over the Bering Sea. The onset of the decline in SST coin- cided with anomalous southward extent of the ice pack (Kukla and Kukla 1974). Fluctuations in the southern ice extent in the Bering Sea in relation to the short-term climatic fluctuations are considered in Chapter 9 of this volume. More recently, Niebauer (1980) has related a subsequent sharp rise in SST in this region in 1975-78 again to changes in the atmospheric circulation which caused southerly flow over the Bering Sea. Namias (1978) has suggested that the latitudinal SST patterns across the North Pacific in November 1976 foretold the strong and persistent air flow from the south over the Bering Sea during the winter of 1976-77. This may be related to the transpacific movement of warm and cold pools of water noted by Favorite and McLain (1973). This trend has persisted and intensi- fied through the winter of 1978-79 (Niebauer 1980). The following sections of this chapter consider a description of the upper-air (700 mb or 3,000 m) 23 24 Physical oceanography flow for the winters of 1962-79. The patterns are related to a 16-year time series of SST for the eastern Bering Sea. DATA Northern Hemisphere monthly 700 mb pressure charts from the period 1963-79 from Monthly Weather Review were analyzed. Representative examples were chosen to illustrate the salient points in the following sections. Sea- surface temperatures (SST) were obtained from the Naval Fleet Numerical Weather Central, Monterey, California, through D. R. McLain of the National Marine Fisheries Service. The SSTs are 12-hourly analyses of observations from ships of opportunity in the eastern Bering Sea. These observations have been arithmetically averaged to give the mean annual time series in Fig. 3-1. There are, at times, strong horizontal temperature gradients near the Pribilof Islands and in Bristol Bay that may bias the SST data. In addition, in years of above-normal ice extent, ships are forced out of portions of Bristol Bay and to the south of the Mean Annual SST "PC > Mean Annual SST VC i Pribilof Islands Bristol Bay 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 YEARS Figure 3-1. Sea surface temperatures from the Pribilof Island and Bristol Bay regions of the eastern Bering Sea (Niebauer 1978). Pribilof Islands. This may bias the SST to warmer temperatures. An example of this is a comparison of the SSTs for 1975 and 1976. In general, 1976 was a colder year (Niebauer 1980) and yet the SST was -0.3-0.5 C higher (Fig. 3-1). However, shelf-wide bottom temperatures for June 1975 and 1976 (Fig. 9-3, Niebauer, Chapter 9, this volume) were nearly identical. The greater ice extent in 1976 relative to 1975 may have been responsible for the slightly higher temperature bias in 1976. Although there may be occasionally nonperiodic bias to the SST, the mean annual SST time series in Fig. 3-1 is considered precise enough for this study. RESULTS Fig. 3-1 is a time series of mean annual SST from the region of the Pribilof Islands and Bristol Bay. The annual SST was near the 16-year mean of 4.2 C in 1963 before rising to 5.4 C in 1967. SST then fell to 2.8 C, or 1.4 C below normal, in 1975. Since then, there has been a rapid rise to 5.7 C, or 1.5 C above normal, in 1978. Similar data from Bristol Bay display similar characteristics. Thus, over the five- year period 1967-71, SST decreased by 2.3 C; over the last five years (1974-78) there has been nearly a 3.0 C rise in mean annual SST of the southeastern Bering Sea shelf water. Inspection of the monthly mean SST from which Fig. 3-1 is derived shows that in spring (March and April) 1975 the SST was about 2.7 C below the 16-year mean, while by spring (April and May) 1978 the SST was about 2.0 C above normal, giving a rise of nearly 5.0 C over three years. It has been suggested that these fluctuations in SST are due to large-scale changes in atmospheric circulation, or more specifically the winter circulation (Niebauer 1980), over the North Pacific and Bering Sea. To outline observations in support of this theory, mean winter 700 mb (3 km) pressure patterns are contrasted with the general winter 700 mb flow patterns (Fig. 3-2) over the period of mean SST illustrated in Fig. 3-1. The mean 700 mb winter flow is generally toward the northeast, parallel to the Aleutian Chain, with relatively weak flow over the Bering Sea. Fig. 3-3 is an example of upper-air flow from the south bringing warmer air from the Pacific Ocean over the Bering Sea in winter 1966-67. This correlates with, and is probably a primary cause of, the relatively high SST in the mid-1960's (Fig. 3-1). Fig. 3-4 is an example of a monthly mean 700 mb pressure chart in the winter of 1974-75, showing essentially meridional flow from the arctic, south into the Bering Sea. Some of the flow then turns east and Recent short-period wintertime climatic fluctuations 25 flows into the southeast Bering Sea. This cold arctic air appears to cool the underlying Bering Sea (McLain and Favorite 1976). The onset of the decline of SST coincided with anomalous southward penetration of the ice pack (Kukla and Kukla 1974) and had adverse effects on Alaska fisheries (Johnson and Seckel 1976). A subsequent rise in SST in the middle to late 1970's in the Bering Sea appears to be related, again, to large-scale changes in the atmospheric circulation, which caused southerly flow over the Bering Sea (Niebauer 1978). Fig. 3-5 illustrates the mean 700 mb contours for January 1978, and shows the Aleutian Low over the western Aleutians and strong meridional flow into the southeast Bering Sea from the North Pacific. These patterns are similar to those of the mid-6 O's and have generally persisted through much of the late 1970's. Namias (1978) has suggested that SST patterns in the North Pacific in November 1976 foretold the strong and persistent air flow from the south over the Bering Sea during winter 1976-77. These air-flow patterns can explain, to a large extent, the high SST in the Bering Sea during this period. DISCUSSION That fluctuations in the mean atmospheric circulation are the driving force behind the observed large interannual fluctuations in SST is consistent with the shelf circulation patterns deduced by Coachman and Charnell (1979) and by Reed (1978). Coachman and Charnell point out that the shorter time-scale circulation (~25 cm/sec) of this shelf is dominated by tides and wind events, but that the longer time-scale mean flow is weak. Seaward of the 100 m isobath flow is ~2-5 cm/sec; between the 100 m and 50 m isobaths it is nearly zero. This weak flow on the shelf has the effect of decoupling the transport of mass characteristics, such as heat, from the mean flow (see, for example, Csanady 1976) and makes their horizontal transport a function of large-scale diffusion. Reed (1978) considered changes in heat content of two 1° X 1° areas inshore of the 100 m isobath. Because of the low net flow in the region, advection of heat was neglected. Reed suggested that gain of heat through horizontal diffusion has little effect on Figure 3-2. Mean winter (Dec- ember-February) 700 mb con- tours (labeled in tens of feet) based on the 26-year period, 1947-72. Flow is generally along the isobars with low pressure to the left (Namias 1978). INTER mb Hek^t NORMAL-! 26 Physical oceanography Figure 3-3. Mean 700 mb contours (tens of feet) for February 1967 (Posey 1967). the heat budget and that net radiation is typically the dominant heat flux in the southeast Bering Sea in summer. During the early fall, evaporation becomes important due to increased wind and rapid cooling of the overlying atmosphere. The conclusion dravi^n from both of these studies is that net heat gain or loss comes about primarily through air-sea interaction, and that due to the low net flow on the shelf, fluctu- ations in heat are retained longer than in an advective system. To test this idea, an autocorrelation analysis was done on the deviations from mean monthly SST from the Pribilof Islands for the years 1963-78 (Fig. 3-6). Significant correlation up to 23 months' lag suggests that these anomalies in SST on the eastern Bering Sea shelf do, in fact, persist for at least two years. Thus, if the atmosphere does drive the ocean (Davis (1976) suggests that it does although Namias (1978) points out that there are probably many and varied feedback loops between ocean and atmosphere and that the subject is highly complex), then Fig. 3-6 suggests that short-term climatic fluctuations may also persist for at least two years. However, Namias and Born (1970) showed that temporal coherence among monthly mean SST patterns in the North Pacific is far greater than any known meteorological coherence; they also found that the SST coherence persists for as long as two years. That fluctuation in the mean winter atmospheric circulation is the driving force behind the observed large year-to-year fluctuations in SST is also con- sistent with the observation of Coachman and Charnell (1979) that there is significant correlation between June bottom temperatures on the eastern Recent short-period wintertime climatic fluctuations 27 Figure 3-4. Mean 700 mb contours (tens of feet) for December 1974 (Taubensee 1975). Bering Sea shelf and the freezing degree days of the previous winter. Moreover, Niebauer (1980) has found significant correlation between mean annual SST (Fig. 3-1) and mean winter cloud cover and mean winter north-south component of the surface winds but not with summer cloud cover or winds. This is interpreted to mean that the large-scale deviations from mean air and sea temperature in the eastern Bering Sea depend on winter atmospheric advective processes. Insolation is important in the seasonal heating and cooling cycle, but apparently has little effect on interannual deviations from the mean. The strong summer insolation also "caps" the shelf water between the 50 and 100 m isobaths through the formation of a strong thermocline. This, along with the low net flow, insulates the heat content of this water which has been acquired during the winter. This mechanism probably also accounts for the high SST autocorrelation coefficients up to two years later which, in turn, suggest that short-term climatic fluctuations may also persist for at least two years. ACKNOWLEDGMENTS I wish to thank Terri Paluszkiewicz, who did the autocorrelation calculations, and the Institute of Figure 3-5. Mean 700 mb contours (tens of feet) for January 1978 (Wagner 1978). 28 Recent short-period wintertime climatic fluctuations 29 1.0 - .8 4-i \ C \ 0) .6 - ♦ o \ »^ \ »^ 0) V o T " .4 - V Av c \ r •- o \ f \ 4-1 \ / \ CO VV V \ 1% Level o ^ _ \ K. V y » u .0 V^ \ 5% Level 1 2 ** \3 / Lags \/ (years) -.2 5% LfivPl 1% Level Figure 3-6. Autocorrelation analysis of 1963-79 monthly mean SST from the Pribilof Islands region. The .01 and .05 confidence intervals are indicated. Marine Science, University of Alaska Publication/ Drafting personnel, who helped with the manuscript. I also wish to thank the anonymous reviewers for their many helpful comments. This work. Contribution No. 402, Institute of Marine Science, University of Alaska, was supported by the National Science Foundation, Division of Polar Programs, Grant DPP 7G23340 A02 (PROBES), by the Alaska Sea Grant Program, Grant 04-8-MO 1-187, and by the University of Alaska, with funds appropriated by the State of Alaska. Coachman, L. K., and R. L. Charnell 1979 On later water mass interaction— A case study, Bristol Bay, Alaska. J. Phys. Oceanogr. 9: 278-97. Csanady, G. T. 1976 Mean circulation in shadlow J. Geophys. Res. 81: 5389-99. seas. Davis, R. E. 1976 Predictability of sea surface temperature and sea level pressure anomahes over the North Pacific Ocean. J. Phys. Oceanogr. 6: 249- 66. Favorite, F., and D. R. McLain 1973 Coherence in transpacific movements of positive and negative anomalies of sea surface temperature, 1953-60. Nature 244: 139-43. Johnson, J. H. 1976 and B. R. Seckel Use of marine meteorological observations in fishery research and management. Presented at the World Meteorological Organization's Techni- cal Conference on the Applications of Marine Meteorology to the High Seas and Coastal Zone Development. 22- 26 November 1976. Geneva, Switzerland. Kukla, G. J., and H. J. Kukla 1974 Increased surface albedo in the Northern Hemisphere. Science 183: 709-14. REFERENCES Brower, W. A., Diaz, and A. 1977 , Jr., H. W. Searby, J. L. Wise, H. F. S. Prechtel Climatic atlas of the outer continental shelf waters and coastal regions of Alaska. NOAA, Asheville, N.C. McLain, D. R., and F. Favorite 1976 Anomalously cold winters in the southeastern Bering Sea, 1971-1975. Mar. Sci. Comm., Basic and Applied 2: 299-334. Namias, J. 1978 Multiple causes of the North America abnormal winter 1976-1977. Monthly Weather Rev. 106: 279-95. 30 Physical oceanography Namias, J., and R. M. Born 1970 Temporal coherence in North Pacific sea surface temperature patterns. J. Geophys. Res. 75: 5952-5. Posey, J. W. 1967 Weather and circulation of February 1967. Cold in the East but con- tinued warm in the West. Monthly Weather Rev. 95: 311-18. Niebauer, H. J. 1978 On the influence of climatic fluctua- tions on the biological and physical oceanography of the southeast Bering Sea continental shelf. Presented at the Joint U.S. /Japan Bering Sea Ecosystem Seminar, Seward, Alaska, 7 August 1978. Reed, R. K. 1978 The heat budget of a region in the eastern Bering Sea, Summer 1976. J. Geophys. Res. 83: 3635-45. Taubensee, R. E. 1975 Weather and circulation of December 1974: Continued warm across much of the country. Monthly Weather Rev. 103: 266-71. 1980 Sea ice and temperature variability in the eastern Bering Sea and the rela- tionship to atmospheric fluctuations. Submitted to J. Geophys. Res. Wagner, A. J. 1978 Weather and circulation of January 1978: Cold with record snowfall in the Midwest and Northeast, mild and wet in the West. Monthly Weather Rev. 106: 549-555. Hydrographic Structure Over the Continental Shelf of the Southeastern Bering Sea Thomas H. Kinder' and James D. Schumacher^ ' Naval Ocean Research and Development Activity National Space Technology Laboratories Station, Mississippi ^ Pacific Marine Environmental Laboratory, Environmental Research Laboratory/ National Oceanic and Atmospheric Administration Seattle, Washington ABSTRACT We synthesize recent work conducted over this exceptionally broad (-500 km) shelf which generally has only slow mean flow (<2 cm/sec). Hydrographic structure is little influenced by this flow, but rather is formed primar- ily by boundary processes: tidal and wind stirring; buoyancy input from insolation, surface cooling, melting, freezing, and river runoff; and lateral exchange with the bordering oceanic water mass. Three distinct hydrographic domains can be defined using vertical structure to supplement temperature and salinity criteria. Inshore of the 50 m isobath, the coastal domain is vertically homogeneous and separated from the adjacent middle domain by a narrow (~10 km) front. Between the 50 m and 100 m isobaths, the middle domain tends toward a strongly stratified two-layered structure, and is separated from the adjacent outer domain by a weak front. Between the 100 m isobath and the shelf break (-170 m depth), the outer domain has surface and bottom mixed layers above and below a stratified interior. This interior has pro- nounced fmestructure, as oceanic water intrudes shoreward from the weak haline front over the slope, and shelf water (middle domain) intrudes seaward across the 100 m isobath. These domains and their bordering fronts tend to persist through winter, although the absence of positive buoyancy often makes the middle shelf vertically homogeneous. INTRODUCTION We selected the title hydrographic "structure" rather than simply "hydrography" because we wish to emphasize the structure, or organization, inherent in the hydrographic distributions. This approach focuses on the shapes of vertical profiles, or rather classes of shapes (e.g., two-layered), rather than on values of temperature and salinity or their corre- lation (TS diagrams). Thus, we find a large region of the shelf where the temperature and salinity are verticEilly homogeneous throughout the year, al- though the values of temperature and salinity fluc- tuate over a wide range. We concentrate on the persistent vertical homogeneity and label this region a hydrographic domain. Because vertical profiles control the hydrostatic stability of the water column, and because stability influences vertical mixing, this approach is physically meaningful and useful. We also concentrate on characteristics of small size, on what can be called the spatial variability. Thus the fronts that separate regions of uniform hydrographic structure (hydrographic domains) are discussed in some detail, as is the finestructure over the outer shelf. One front, for example, has a width of only 10 km and the finestructure has a typical vertical extent of 5 m. It is now possible to resolve such features as fronts and finestructure because samples are taken closer together than formerly. Our emphasis on hydrographic structure and small spatial scales is not opposed to examination of TS properties or broader spatial scales, but comple- mentEiry to it. Our description of the shelf hydro- graphic structure is more meaningful in considering shelf environment from a climatic point of view. We mostly ignore changes at intervals longer than annual, although interannual hydrographic variability is significant (e.g.. Overland, Niebauer, and Ingraham, this volume). The major features that we discuss here, however, were observed both in 1976 (the winter of 1975-76 was exceptionally cold, with 31 32 Physical oceanography extensive ice cover) and in 1977 (the w^inter of 1976-77 was exceptionailly mild, with reduced ice cover). Although altered by interannual changes, the features that we describe persist through these long-term vairiations. Because mean flow over the shelf is small, changes in hydrographic properties can be straightforwardly attributed to local processes rather than to advection. For instance, cold temperatures in the lower layer of the middle shelf persist throughout summer (Fig. 4-4). This was once believed to be evidence of mean flow from the northern shelf, but in fact the cold temperatures are caused by local processes: heat loss at the sea surface and complete vertical mixing during winter, followed by the establishment of strong stratification during spring and summer. This stratification insulates the lower layer from downward heat transfer. Especially over the inner two-thirds of the shelf, important characteristics of the hydrography can be explained by local phenomena and advective effects are unimportant. We complete the introduction by briefly discussing the oceanographic setting, reviewing previous work, and discussing the data. Then we define the hydro- graphic structure by discussing salient characteristics: domains, fronts, fine structure, winter structure, and river plumes. We then discuss some processes that affect the hydrographic structure: stirring and buoyancy addition, heat and salt transport, and upwelling. Finally we discuss and speculate about aspects of the hydrographic structure. Setting The southeastern continental shelf is bordered by the Alaska Peninsula, the Alaska mainland, and a line running southwest from Nunivak Island to the Pribilof Islands and thence following the shelf break southeastward to Unimak Pass. Waters above the shelf receive an annual excess of precipitation over evaporation, as well as freshwater runoff from num- erous rivers. Estimating precipitation either from Jacobs (1951) or from station data reported by Brower et al. (1977), and evaporation from Jacobs, a net of about 1 percent of the volume of water over the southeastern shelf is added annually by precipitation minus evaporation.' An additional 1 percent is added by river runoff, principally from the Kuskokwim and Kvichak (1,500 to 2,000 m^sec average discharge from all rivers: Rod en 1967, Favorite et al. 1976). In winter, ice covers over 50 ' Using recent precipitation estimates by Reed and Elliott (1979) would increaase this to nearly 2 percent. Reed and Elliott state, however, that their estimates may be inaccurate in the subarctic Pacific. percent of the shelf, initially appearing inshore in November, often expanding to cover more than 80 percent of the shelf by March, and rapidly disappear- ing between late April and early June (Favorite et al. 1976, Muench and Ahlnas, 1976). Ice appears to form near shore and is blown southward during the freezing season (see McNutt and Pease, this volume). Current meter records show that most of the hori- zontal kinetic energy of the shelf water is tidal: 60-95 percent of the variance in records 9-332 days long was tidal (see Kinder and Schumacher, Chapter 5). Vector mean speeds (<2 cm/sec) were one order of magnitude lower than tidal speeds (~20 cm/sec). Historical review There has been considerable Japanese, Soviet, and American work done on this shelf. Results of this work have been effectively summarized (Ohtani 1973, Takenouti and Ohtani 1974, Arsenev 1967, Dodimead et al. 1963, and Favorite et al. 1976), and this brief review places more recent results in perspective. Takenouti and Ohtani (1974) discussed waters above the shelf, which they realized were separated from ocean waters by a "discontinuous zone" (cf. Kinder and Coachman 1978). They further reported that the cold (<1.0 C) water near the bottom in the middle shelf (cf. Fig. 4-4) was not advected from the Gulf of Anadyr as Kitano (1970) believed, but was formed in situ in winter and insulated by strong stratification in summer. Their proposed classifica- tion of water masses over the southeastern shelf has been modified by recent findings. Takenouti and Ohtani defined a CW (coastal water) region by its low salinity, but we have found that at the end of winter the salinity may be higher there than in the adjacent convective area (C A— roughly corresponding to our middle domain). Furthermore, the Alaskan Stream (AS) region near the shelf break is misnamed— direct connection with the westward-flowing Alaskan Stream, which exists south of the Alaska Pennisula and Aleutian Islands, is not proved. At the same time, our map of hydrographic domains (Fig. 4-1) is congruent with theirs, and builds upon their insights. Ohtani (1973) discussed the southeastern shelf in more detail. He mentioned the thermal front that forms between the middle and coastal domains (cf. Fig. 4-6 and Schumacher et al. 1979), and correctly suggested the importance of tidal stirring in forming this front. Ohtani also emphasized vertical stratifica- tion in defining shelf water masses, and dwelt less on arbitrary temperature and salinity limits. Again, net inflow of Alaskan Stream water is more tenuous than Ohtani suggested; properties are certainly exchanged Hydrographic structure 33 166 162 160' 158" 59' 58 57' 56' A 164 160 Figure 4-1. Approximate boundaries separating the three shelf (coastal, middle, outer) and the oceanic hydrographic domains. The boundaries are three fronts: inner, middle, and shelf break. These fronts roughly coincide with the 50 m isobath, the 100 m isobath, and the 200 m isobath (shelf break). Profiles from the numbered stations appear in Fig. 4-2. through the eastern Aleutian passes by vigorous tidal currents, but the net flux of water is not known, and is probably small in any case because of small cross- sectional area (Favorite 1967).^ Arsenev (1967) wrote about water masses and currents of the entire Bering Sea, using many sources, but highlighting Soviet work. He discussed the importance of water mass transformation by fresh- water runoff, insolation, cooling, melting, and freez- ing. He also recognized the separation of oceanic and shelf waters, but virtually ignored the southeastern shelf in favor of the western shelf, especially the Gulf of Anadyr. Dodimead et al. (1963) and Favorite et al. (1976) summarized the regional oceanography of the North Pacific, including the Bering Sea. Dodimead et al. (1963) included an appendix on Bristol Bay, and ^ Favorite (personal communication 1979) has pointed out that the distribution of a temperature maximum along the eastern Bering Sea continental slope suggests net inflow to the Bering Sea through passes west of Unimak Pass between 170°Wandl72°W. noted several features that have been elaborated only recently. They reported the inner front that separ- ates the coastal and middle domains as a sharp boundary (cf. Schumacher et al. 1979), and also reported "marked changes" near the shelf break that correspond to the weak haline front there (cf. Kinder and Coachman 1978, Coachman and Charnell 1979). They also noted the patch of cold surface water within Bristol Bay, which they attributed to upwell- ing. Favorite et al. (1976) showed three domains across the shelf: shelf edge, mid-shelf, and West Alaska Coast (their Fig. 33). Their geographical boundaries nearly coincided with the three domains that we describe in the next section, but they were apparently based on TS relations (cf. Ingraham, this volume ). Favorite et al. (1976) also discussed the frontal zone over the slope. Thus, many of the features that we now recognize as important components of the hydrographic struc- ture of the shelf were reported previously. Among these features are the front over the slope and the inner front farther inshore, the division of shelf waters into distinct domains, and the possible upwell- 34 Physical oceanography ing in Bristol Bay. We now know more details and understand these features better, but it is clear that our progress has benefited from these earlier works. Data From August 1975 to February 1978, hydrographic casts were made with profiling instru- ments: STD (sahnity, temperature, depth), CTD (conductivity, temperature, depth), or XBT (expend- able bathythermograph). Covering all months from February to October, these 1,064 STD and CTD casts are biased towards summer (Table 4-1), but this bias is not a serious limitation because of adequate cover- age in February. The STD and CTD data were calibrated by a water sample, normally taken at the bottom during alter- nate casts. Calibration temperatures were determined by reversing thermometer, salinity by portable induction salinometer. We claim an accuracy of ±0.02 C and ±0. 02^/00. The XBT profiles were calibrated against nearby CTD casts, and we claim ±0.1 C. The unusual data processing necessary to examine details of finestructure in vertical profiles of salinity is discussed elsewhere (Coachman and Charnell 1979). TABLE 4-1 Summary of STD and CTD data MONTH NUMBER OF CASTS February March April May June July August September October TOTAL 117 65 34 159 213 122 152 184 18 1064 These casts were taken during the period August 1975 to February 1978. HYDROGRAPHIC STRUCTURE Three hydrographic domains We have divided the shelf into three structural domains, called the coastal, middle, and outer domains (Fig. 4-1). These domains are nearly con- gruent with geographical boundaries previously defined by water masses (e.g., Favorite et al. 1976), and are approximately separated by the 50 m isobath, the 100 m isobath, and the shelf break (close to the 200 m isobath). Our structural domains broaden the criteria previously used for defining the shelf water masses, emphasizing the potential for stratification of the water column (Table 4-2; also cf. Fig 24 in Coachman and Charnell 1979). These domains are most prominent in summer, but are also discernible during the other seasons. Lying seaward of the outer domain, and separated from it by a weak haline front (shelf break front), is the oceanic domain. The oceanic domain completes our scheme, but it is outside the geographic focus of this chapter (Sayles et al. 1979 concentrate on the water overlying the deep basins). Defining water masses is most useful where temperature and salinity vary slowly at a location (e.g., no surface cooling), or where significant mean advection makes water masses useful in tracing flow. Thus tracing water masses has usefully revealed mean flows in the deep ocean. On this shelf, however, great changes in water mass properties occur annually (Coachman and Charnell 1979, Kinder et al. 1978), and there is little mean flow. Moreover, seemingly reasonable temperature-salinity parameters may prove deceptive. For example, a criterion previously used to describe coastal water has been its low salinity (Takenouti and Ohtani 1974, Favorite et al. 1976), but we now know that during early spring the coastal domain may be more saline than the adjacent middle shelf water (Kinder 1977). To overcome some of these ambiguities, we have added vertical structure to the criteria. Instead of water mass, we use the word domain, favored by Dodimead et al. (1963), to connote broader criteria than simple temperature-salinity correlations. These domains are geographic entities; energy balances forming vertical structure are closely tied to local geography, so that the domains are also nearly fixed geographically (see Stirring and buoyancy addition in the next section of this chapter). During the summer, vertical structural criteria permit easy separation of the shelf into three domains: homogeneous (coastal), two-layered (mid- dle), and stratified interior (outer; see Table 4-2). These categories are insensitive to particular values of temperature and salinity which vary from year to year (see Niebauer, Chapter 3, for variations of sea surface temperature, and Ingraham, this volume, for interannual hydrographic variations), but depend on the influence of buoyancy input, which Hydrographic structure 35 TABLE 4-2 Hydrographic domains in summer Coastal Middle Outer Vertical structure Stratification Depth Temperature Salinity homogeneous two-layer surface mixed layer stratified interior bottom mixed layer finestructure very low very high moderate <50m thickness of bottom (tidal) mixed layer 50 m < depth <100 m thickness of surface + bottom mixed layer >100m > surface + bottom mixed layers, thus an interior region exists very warm in late summer (efficient heat transfer throughout water column) (~ 8 to 12 C) very cold bottom temperature throughout summer (vertical heat transfer impeded by stratification) (~ -1 to 3 C) moderate (~ 3 to 6 C) generally low (<31.5^/oo), but may be relatively high following winter (> 320/00: brine drainage during freezing) moderately low (~ 31.5^/00) high (> 320/00) river runoff freezing melting adjacent water overlying deep Influences freezing basin; Bering Slope Current This table emphasizes summer conditions, when the domains are most clearly established and when our data are most extensive. These domains remain useful throughout the year: see the section in this chapter on winter structure. tends to stratify the water column, and tidal stirring, which tends to mix the water column. An example of each domain from early autumn 1976 (Fig. 4-2) illustrates the three structures. Station 126, in about 50 m of water, is nearly homo- geneous in both temperature and salinity, while station 101, in about 70 m of water, is strongly two-layered with vertical temperature and salinity differences of 4 C and 0.4*^/oo. In still deeper water, station 47 has a surface layer, a bottom layer (not completely mixed), and a stratified interior. Many stations in the outer domain display strong finestructure in temperature and salinity (Coachman and Charnell 1977, 1979; Kinder 1977; Kinder et al. 1978), but we smoothed the profiles in Fig. 4-2 to emphasize larger-scale features. A companion view of these domains is shown by plotting vertical temperature differences across the shelf in early autumn 1976 (Fig. 4-3). Shoreward of the 50 m isobath, this difference was generally <^1 C, while between the 50 and 100 m isobaths it commonly exceeded 7 C. Nearer the shelf break, A 20- 40- E 60- 4) 80- O 100- 200- 300-" 101 Temperature (°C) 4 5 6 7 Temperature (°C) SEP-OCT 1976 47 B 31 o. 4) a Salinity <%o) 32 33 _l 20- V 40- N A 60- 126 I . 80- 101 \ 00- 47 00- SEP-OCT 1976 6- U O — 5- (0 4- 4) Q. E 4) 3- 2- 1- 0 Ot--250 ot=25.4 0^ = 26 0 SEP-OCT 1976 32 —r- 33 Salinity (%o) —I 34 Figure 4-2. Typical stations from autumn 1976 illustrating the three domains. Coastal (homogeneous), station 126; middle (two-layered), station 101 ; outer (stratified interior), station 47 (see Table 4-2 for domain characteristics). (A) Temperature (°C). (B) Salinity {^loo). (C) Temperature-salinity (°C, *-*/oo). Station locations are shown in Fig. 4-1. 36 Hydrographic structure 37 170' 168 164 A Tmax (C Sep-Oct 19 170' 162" 156 Figure 4-3. Maximum vertical temperature difference, surface minus deep (°C). The largest differences are in tlie middle domain, and the smallest in the coastal domain (cf. Fig. 4-1). intermediate values near 4 C were found. A plot of bottom temperature in autumn 1976 (Fig. 4-4) illustrates the strong insulating effect on the stratifi- cation displayed in Fig. 4-3. Even in September, cold temperatures (<0 C) remained from the preceding winter, isolated by the very strong two-layered stratification in the middle domain. In contrast, the coastal domain was well mixed and bottom temperatures exceeded 9 C. In the deeper water of the outer domain bottom temperatures were inter- mediate, generally 3-6 C. Obviously, stratification is important to an understanding of the shelf. Using stratification as em adjunct to water mass analysis is valuable, but Coachman and Charnell (1979) also used the traditional method successfully. They defined a shelf water mass found in the middle domain, and an oceanic ("Alaska Stream/Bering Sea") water mass found above the continental slope (Fig. 4-5). They were then able to explain much of the structure of the outer domain in terms of the lateral mixing along isopycnal surfaces between the shelf and oceanic water masses. The shelf water mass was always less saline and, below 30 m, colder than the adjacent oceanic water. In spite of annual and interannual variations, there exist throughout the year two water masses, one cold and fresh and the other warm and saline, in juxtaposition along the outer shelf. One important evidence of the inter- action of these water masses, finestructure in vertical profiles, is discussed below in a separate section. A combination of categorizing by vertical struc- ture and by traditional water mass techniques is more useful than either used separately. For examining the shelf alone, structural categories are most distinct, but for understanding interaction with waters over- lying the deep basin, water mass analysis is useful. As Coachman and Charnell (1979) implied, these two views are interdependent. Fronts The four hydrographic domains (coastal, middle, outer, and oceanic) are separated by three fronts. The front separating the coastal and middle domain is much narrower than the domains themselves, and so is legitimately called a front. The other two transi- tions are much broader relative to their adjoining 38 Physical oceanography 166 54' TEMPERATURE (C) BOTTOM ( Sep-Oct W^S __ 168" 166' Figure 4-4. Bottom temperature (°C), late September and early October 1976. Even in autumn, low temperatures persist in the bottom layer of the middle domain. domains, but since they have been called fronts (e.g., by Iverson et al. 1980), we adopt this usage also. Proceeding seaward, we label these fronts as inner, middle, and shelf break (Fig. 4-1). The inner front separates the homogeneous coastal domain from the two-layered middle domain. It was hinted at by Dodimead et al. (1963), and noted by Muench (1976) farther north. Schumacher et al. (1979) have called it a structural front, to stress the separation of two vertical structures or marked change of stratification rather than the sepairation of two water masses. The front is about 10 km wide and generally follows the 50 m isobath (Fig. 4-1). Approaching shallower water from the middle do- main, isotherms, isohalines, and isopycnals all spread from the thin thermocline, halocline, and pycnocline over the middle shelf (Fig. 4-6). Within 10 km the vertical hydrographic structure changes from dis- tinctly two-layered to neairly homogeneous. Away from the front, within the strongly stratified middle domain, the thickness of the bottom mixed layer (as judged by nearly isothermal and isohaline profiles) is nearly 50 m, about the same as the total water depth where the front is found. In general, we find that over the middle shelf the bottom mixed layer is ~50 m thick, the surface mixed layer is 15-20 m thick, and the front occurs where the water depth approx- imately equals the thickness of the bottom mixed layer (i.e., 50 m); the strongest stratification occurs where the sum of the bottom and surface mixed layer thickness equals the water depth (i.e., 20 m + 50 m = 70 m). During winter, the middle and coastal domains are nearly vertically homogeneous following surface cooling, freezing, and vigorous wind stirring in fall and winter (but see our section on winter structure for an important exception). Frontogenesis occurs with the addition of meltwater during the ice breakup in spring. As this initial stratification forms, it is reinforced by insolation, so that later in summer thermal stratification is primarily responsible for vertical density differences. Schumacher et al. (1979), Kinder (1977), and Kinder et al. (1978) reported details of this front, and Simpson and Hydrographic structure 39 SHELF MW: 18,19 AC : 2 5,26.27, 43,54,55,71 ALASKA STR. /BERING MAY 77 ACONA ,2,36,37,42 Figure 4-5. Temperature-salinity correlations, middle domain (SHELF) and oceanic domain (ALASKA STR./BERING). Envelopes drawn from data gathered in August 1976 and May 1977 illustrate the warmer and saltier oceanic water at the same density as the cooler and fresher shelf water, and interleaving occurs across the outer domain. (From Coachman and Charnell 1979.) Pingree (1978) summarized features of similar fronts over the European continental shelf. The shelf break and middle fronts are less clearly describable. Overlying the continental slope, the shelf break front separates the oceanic domain from the outer domain, but the width of this front is similar to that of the outer domain. Similarly, the middle front which divides the middle and outer domains near the 100 m isobath (Fig. 4-1) is broad and ill-defined compared to the inner front. Never- theless, the shelf break and middle fronts aire both real and important components of the hydrographic structure. Kinder and Coachman (1978) described the shelf break front and recognized its essentially haline character. The front is revealed by a change in the horizontal salinity gradient (from nearly zero over the deep basin to about 4 X 10"^ g/kg/km over the outer COASTAL 140 MIDDLE SHELF A) >o 5 / " 20 km I 1 JUN 1976 TEMPERATURE (°C ) 40 80 SEP-OCT 1976 SALINITY (7oo) Figure 4-6. The structural (inner) front separating the coastal and middle domains. This line was between Nunivak Island and the Pribilof Islands. (A, B) Temperature and salinity cross sections, and (C) sequential temperature profiles from June 1976. The sections are based on stations separated by about 10 km. (D, E) The same section based on widely spaced CTD stations in autumn 1976. (From Schumacher et al. 1979.) 40 Hydrographic structure 41 shelf), by isopycnals extending from the shelf to intersect the sea surface above the slope, and by isolines downwarped beneath the front. Available winter data show that this front persists throughout the year. Coachman and Charnell (1979) and Coachman (1978) examined this region in more detail, and described this transition zone as two broad fronts, one over the slope and one farther inshore near the 100 m isobath. Between these two transitions, each of which has a large horizontal salinity gradient (~10 X 10""' g/kg/km), is a region of very small gradient (Fig. 4-7). The transition near the shelf break corresponds to the front described by Kinder and Coachman (1978), while the inner transition corresponds to the middle front separating the middle and outer domains (Fig. 4-1). At different times when examined by different distributions, these broad transitions do appear truly frontal. For instance Coachman and Charnell (1979) showed a mean cross-shelf temperature section for August 1976 that clearly showed a thermal front near the 100 m isobath, and Coachman (1978) showed strong evidence of a front delineated by particulate and chlorophyll a concentrations in April 1978. Over the slope Kinder and Coachman (1978) showed a shallow weak -density front in an August 1972 section, and we show a weak-salinity front from February 1978 (Fig. 4-9). Kinder and Coachman (1978) also showed large dissolved phosphate and nitrate gradients across the shelf break front in July 1974. Both the middle and shelf break fronts gen- erally appear broad and therefore weak, but oc- casionally they are manifest as sharp fronts in various properties. The shelf break front, however, can always be detected as a weak front in salinity. Finestructure and density inversions Finestructure, the layering of vertical profiles on scales from 1 to 25 m (Fig. 4-8), is a salient feature of the outer domain (Table 4-2). The distri- bution of the finestructure and the mixing physics associated with it are clues to understanding cross- shelf fluxes. Horizontal distributions of the occurrence of finestructure over the shelf showed that it was common between the shelf break and the 100 m isobath, and occurred elsewhere only rarely. During 33.0 r 32.5 SLOPE =9 3X10"^ %o/km o 32.0 - 31.5 SLOPE = 9.3x10-5 7oo/kr 50 km 100 m 150m ISHELF BREAK) Figure 4-7. Vertically (0-100 m) and horizontally (along-isobath) averaged sections across the shelf from May 1976, August 1976, and May 1977. Transitions in the salinity gradient mark the 100 m isobath (middle-outer domains; middle front) and the shelf break (outer-oceanic domains; shelf break front). (From Coachman and Charnell 1979.) 42 Physical oceanography 32 S^ /oo 32 5 I 33 T, °C Figure 4-8. A superb example of temperature and salinity finestructure in August 1976. Finestructure, although often less pronounced than this, was present at most outer domain stations. (From Coachman and Charnell 1979.) 1976, a year when the shelf was surveyed extensively, finestructure in the outer domain was reported in March (Coachman and Charnell 1977), in June (Kinder 1977), in August (Coachman and Charnell 1979), and in September-October (Kinder et al. 1978). Only a few stations with finestructure were reported outside the outer domain (e.g.. Kinder 1977, Fig. 22), and data from 1977 and 1978 also conform to these distributions. As Coachman and Charnell (1979) discussed, the finestructure occurs in the interior of the water column, below the surface mixed layer and above, the bottom mixed layer. Within this interior region, warmer and saltier oceanic water intrudes shoreward while cooler and fresher shelf water intrudes seaward. As interpreted by Coachman and Charnell (1977, 1979), the outer domain is a region of lateral (i.e., cross-flow, and here also cross-isobath) water mass interaction with interleaving of water masses occurring at finestructure scales. Such interleaving, when water masses of similar density but differing temperature and salinity values mutually intrude, has been observed in many other locations (e.g., see J. Geophys. Res. 83(C6) 1978). Occurrence of finestructure throughout the outer domain, best documented in 1976 (a year with extensive ice cover and late ice breakup) but also observed in 1977 and 1978, implies that finestructure is an inherent pairt of mixing across the outer domain. An essential stage in mixing large masses of water is the reduction of the spatial scale of identifiable water parcels, until a scale is eventually reached at which molecular diffusion is effective. As the spatial scales decrease, spatial gradients increase, as does the surface area of the boundary, and so mixing progresses. Interleaving on finestructure scales is the initial scale reduction. While finestructure features are only a few meters in vertical extent, they ap- parently extend horizontally for tens of kilometers. In both March and August 1976, Coachman and Charnell (1977, 1979) traced temperature-salinity correlations within layers for distances of about 100 km. One startling result of Coachman and Charnell's work was the discovery of a static instability in a layer about 10 m thick in March 1976 and many smaller-scale instabilities of a few meters' thickness in summer. The larger instability was clearly resolved by the instrumentation used (standard CTD vertical profiling system), and had an apparent lifetime of about one week. They speculated that it was formed by interaction between strong winds and the seasonal ice cover. The smaller instabilities were poorly resolved by the standard CTD profilers used, but Postmentier and Houghton (1978) measured nearly identical features over the oceanographically similar slope region south of New England using a higher- resolution profiler. Both Coachman and Charnell (1979) and Postmentier and Houghton (1978) invoked differential diffusion of temperature and salt to explain the smaller instabilities. Because heat diffuses more rapidly than salt at molecular scales (it is easier to transfer energy than mass), adjacent layers of water can become convectively unstable on small scales, either through salt fingers (warm and saline water overlying cool and fresh water) or through double diffusion (cool and fresh water above warm and saline water). In the outer domain, the condi- tions for salt fingers exist at the lower interface of shoreward-intruding basin water, while the condi- tions for double diffusion exist at the upper interface. Winter structure The discussion of hydrographic domains and fronts mostly reflects summer conditions, but winter conditions are more interesting than might be ex- pected. In winter, waters above most of the shelf usually are vertically homogeneous, with two excep- tions. In the outer domain warmer (but more saline I Hydrographic structure 43 and therefore denser) water from the oceanic domain intrudes beneath cooler and fresher shelf water, thus maintaining stratification. Elsewhere, low-salinity water from melting ice may stratify water that was well mixed during autumn and winter (by wind stirring and surface cooling). A cross section taken from southeast of the Pribilofs toward Cape Newenham in February 1978 (Fig. 4-9) illustrated intrusion of the basin water. Between the shelf break and the 100 m isobath (i.e., outer domain) water warmer than 3.5 C and saltier than 32.5*^/oo intruded beneath shelf water which was both colder and fresher. Inshore of the 100 m isobath the water column was well mixed, colder (<2.5 C) and fresher (<32.250/oo). Data from several stations with similar profiles, saltier and warmer near the bottom, were also taken near the Pribilof Islands during April and May 1978, and Coachman and Charnell (1977) showed data with this character taken in March 1976. There is sufficient coverage of the outer domain during late winter and early spring to suggest that cold and fresh shelf water overlies warmer and more saline basin water, and that this domain remains stratified during winter. station 85 0 Middle Shelf t Outer Shelf t Oceanic 22-23 Feb 1978 Figure 4-9. Temperature (°C) and salinity (°/oo) across the shelf in February 1978. This section is from southeast of the Pribilofs toward Cape Newenham. In the outer domain the deeper water is warmer, but more saline and therefore denser, than the shallower water. Melting ice in the middle shelf can also cause stratification during the winter, but inshore within the coastal domain mechanical stirring keeps the water column well mixed. In February 1978 we observed (by satellite imagery, see Fig. 5-11, Chapter 5) that ice near Nunivak Island moved about 100 km southeast, into an area previously free of ice. About ten days later we measured hydrographic properties near this ice, which was melting. Away from the ice (~20 km), sea surface temperatures were near 0 C, and temperature profiles were vertically homo- geneous (Fig. 4-10). Within the ice (where water depths exceeded 50 m), however, the water column was stratified in two layers. In the shallow layer temperatures were near freezing (~— 1.73 C) and the salinity was lower than in the homogeneous water. 27,0 28.0 29.0 30.0 31.0 32.0 33.0 Salinity (%») 2.0 -1.0 0.0 1.0 2.0 3.0 40 Temperature (°C) 0.00-1 — . ' " — . ' ' — r,— ' 1 12.00- 24.00- f 36.00- Q. O a 48.00- 60.00- 72.00 \H Sig Sig^S 21.00 22.00 23.00 24.00 25.00 26.00 27.00 Sigma-T Figure 4-10. Temperature (°C), salinity (°/oo), and density (kg/m^ ) profiles near the ice edge in February 1978. Dashed profiles were typical away (>20 km) from the ice or where water depth was less than about 50 m. Solid profiles were typical near the ice where water depth ex- ceeded 50 m. Below the weak pycnocline, salinity and temperature were similar to values away from the ice. The decrease of temperature and salinity probably result- ed from ice melting, about 30 cm of ice for Fig. 4-10. The transition between two-layered and homo- geneous conditions occurred near the 50 m isobath, as in summer. Inshore of the 50 m isobath, the water column was homogeneous with or without ice. We do not know how persistent this winter strati- fication is, but we suspect that the weak stratification found in water deeper than 50 m was fragile, depend- ent in part on the continued presence of ice. Once the upper layer cools to the freezing point, ice stops melting. As stirring erodes the pycnocline, however, heat remaining in the bottom layer is mixed upward, presumably melting ice and adding light meltwater. The continued presence of ice above a stratified water column in winter apparently favors continued strati- fication, suppressing wind stirring, limiting surface 44 Physical oceanography heat loss, and maintaining a reservoir of potential meltwater. The question of whether ice cover gener- ally affects the hydrographic structure over the middle shelf in v^^inter and to what extent this struc- ture in turn influences the resulting stratification during the ice-free season remains unanswered, but the ice may be important through the following summer. It is clear that the eventual melting of ice in spring is important in stratifying the middle shelf domain (Schumacher et al. 1979). River plumes The local effects of river discharge have received little attention because most oceanographic data have been collected away from the coast. Satellite images and sparse hydrographic data suggest that river plumes (defined, say, by salinity lower than 25*^/oo) remain near shore, flowing anticlockwise around Bristol Bay, and leaving the southeastern shelf to the north (much of this water may flow through Etolin Strait, inshore of Nunivak Island). Large discharges of fresh water can stratify the water column in the coastal domain, and may form fronts (see Garvine and Monk 1974 for a description of the frontal plume of the Connecticut River in Long Island Sound). Straty (1977) reported observations made in Bristol Bay during 1966. He traced the anticlockwise nearshore flow of river water around Bristol Bay using dye, drift cards, and salinity measurements. He reported no fronts associated with the Naknek, Kvichak, Egegik, and Ugashik rivers, probably be- cause of vigorous tidal stirring in the shallow (less than 20 m) bay. Conditions may be similar in Kuskokwim Bay farther west, but we have no data there. The direct effects of the river discharges appear to remain within a few tens of kilometers of the coast, providing an inshore boundary of the coastal domain. PROCESSES THAT AFFECT THE HYDROGRAPHIC STRUCTURE Many processes can form, alter, or erase features of the hydrographic structure. We have grouped such processes into three somewhat arbitrary categories. First we discuss the addition of heat and salt and their transport across the shelf, processes which directly transform water masses. Next we focus on the interplay of mechanical stirring and buoyancy addition. These processes determine the stratifica- tion, a key element of the structure and a strong influence on the transport. Finally, we speculate on the possibility of upwelling, which may affect the hydrographic distributions in Bristol Bay. Heat and salt transport Transformations of water masses on this shelf occur locally through the addition of heat and salt. Because of cooling and heating at the surface, evapor- ation and precipitation, and freezing and melting, relatively large fluxes of heat and salt occur at the sea surface. To a lesser extent horizontal mixing and river runoff at lateral boundaries influence temp- erature and salinity. Because of the low mean flow on the shelf, and because of the shelf's great width, water mass properties are more likely to result from local phenomena (e.g., insolation and melting) than from advection. Great changes occur annually in the flux of heat and salt at shelf boundaries, so that water masses vary annually also. In the middle and coastal domains the change in heat content of the water is primarily balanced by heat transfer through the sea surface; horizontal advection and horizontal turbulent diffusion are much smaller. Reed (1978) calculated a heat budget for an area (1° lat. X 2° long.) of the middle domain for summer 1976. The local rate of heat change was balanced over the summer by net surface exchange within 10 percent (excellent agreement for such budgets). During the summer, most of this surface exchange was radiative, but by early fall evaporation was important. Over the fall, winter, and early spring, the terms incorporating phase changes (evap- oration, freezing, and melting) share importance with radiation terms. The net surface exchange retains its importance, however: Coachman and Charnell (1979) found a high correlation (r =—0.96, n = 12) between mean lower-layer temperatures in June over the middle shelf and degree-days of frost for the preceding winters. Reed's (1978) results can probably be extrapolated into the coastal domain also. Although the vertical heat distribution differs there (Fig. 4-2), the horizontal terms and surface exchanges are probably similar. In the outer domain, however, horizontal terms apparently are more significant. Since mean flow is 2-10 cm/sec, advection cannot be ignored, and lateral exchange, as evidenced by finestructure, may be even more important. The strong annual cycle which Coachman and Charnell (1979) showed for this region— approximately the seaward half of the shelf waters and those over the slope— was caused by surface exchange. Below the surface mixed layer, however, they showed large-ampUtude finestructure (Coachman and Charnell 1977, 1979), with warmer shoreward intrusions originating in the oceanic domain and colder seaward intrusions originating over the shelf (Fig. 4-8). These lateral interleavings are strong evidence of lateral exchanges of heat and salt. Hydrographic structure 45 I with the shelf water (colder and fresher) gaining heat and salt. Various attempts have been made to estimate the horizontal heat flux in terms of a bulk heat conduc- tivity such that the turbulent horizontal heat flux is given by : ax p = density of water Cp = heat capacity (J m /sec) (kg/m^ ), (J/kg/°C), \ iLi = horizontal temperature (°C/m), 9X gradient and Kh = horizontal conductivity (m^ /sec = 10'' cm^ /sec). This is sometimes a poor approximation of the physical processes (which are hidden within Kh), and Kh is often not constant. Nevertheless, such estimates remain useful for modeling and estimating cross-shelf fluxes. Kinder et al. (1978) calculated a heat balance for the lower layer of the middle shelf over summer 1976 and estimated Kh~1.7 X 10^ cm^ /sec (=1.7 X 10^ m^/sec). They similarly estimated vertical conductivities in the middle shelf, and values ranged from 7 X 10"^ cm^ /sec to 5 X 10"' cm^ /sec. Because of the strong stratification, the lowest values approached molecular diffusion (~1.4 X 10"^ cm^/ sec). These estimates by Kinder et al. (1978) were probably maxima, since they assumed that all local change had been caused by one-dimen- sional diffusion, either vertical or horizontal. Coachman and Charnell (1979), applying a model proposed by Joyce (1977), estimated 2.8 X 10^ cm^ /sec and IX 10^ cm^ /sec for the middle and shelf break fronts and 20 X 10^ cm^ /sec for the outer domain between fronts. Kinder and Coachman (1978) calculated a salt budget for the entire shelf. Since fresh water is added annually at the coast by river runoff, and because precipitation exceeds evaporation over the Bering Sea, there must be a flux of salt shoreward to main- tain the long-term mean salinity distribution. For the shelf as a whole, the largest term (>99 percent of salt flux) is advection: relatively saline water from the oceanic domain flows onto the western shelf to supply the Bering Strait outflow (~1.0 X 10'' m^ /sec). Over the southeastern shelf, however, the salt balance is not advective (because of low mean flow). Kinder and Coachman (1978) estimated a diffusivity of 3 X 10^ cm^ /sec for the cross-shelf salt flux. Calculating diffusivity for the middle domain over summer 1976, Kinder et al. (1978) obtained 1.1 X 10^ cm^ /sec, using the same method as for thermal conductivity (the salt diffusion equation was analogous to that of heat). Kinder and Coachman (1978) suggested that the cross-shelf salt flux was driven by the tides, as a "tidal diffusion," because most (~90 percent) of the kinetic energy over the shelf is tidal (Chapter 5, this volume). The tidal current, if appropriately correlated with salinity variations over the tidal cycle, could cause a significant flux of salt across the shelf. Coachman and Charnell (1979) showed, however, that in the outer domain lateral interleaving on vertical fine- structure scales is ubiquitous and represents cross- shelf mixing. Tidal diffusion still remains tenable for the middle and coastal domains, and the tides do contribute most of the turbulent energy (via the bottom frictional layer and velocity shear) within the outer domain. Kinder et al. (1978) also reported another means of salt flux— ice transport. During the ice-covered part of the year, satellite imagery often shows open water south of east-west zonal coasts: south of St. Lawrence Island, south of Nunivak Island, and northern Kuskokwim Bay are typical examples (see Muench and Ahlnas 1976 and chapters by McNutt and Pease, this volume). During the spring of 1976 (Kinder 1977) and to a lesser extent in 1977, water with elevated salinities (>32.5°/oo in June 1976) was found in Kuskokwim Bay. Our explanation is that ice freezing in Kuskokwim Bay is blown seaward, leaving behind the brine that drains during freezing. We do not know accurately the amount of ice ex- ported from the coastal domain annually in this way, nor do we know the salinity of the exported ice. Kinder et al. (1978) estimated that this divergence of ice transport may account for a mean salt flux of 6 t/sec (1 t = 10^ g) shoreward from the middle to the coastal domains. This is about 10 percent of the mean salt flux (50 t/sec) estimated by Kinder and Coachman (1978) for the southeastern shelf. As Coachman (1979) pointed out, this mechanism may be generally important at high latitudes; hypersaline water relict from the previous winter has been found not only in Kuskokwim Bay, but recently in Norton Sound (Chapter 6, this volume), Kotzebue Sound (Kinder et al. 1977), and lagoons adjoining the Beaufort Sea (Wiseman 1979). Since this mechanism causes a net freshening of the middle domain, it may partly explain the correlation that Coachman and Charnell (1979) reported between yearly mean temperatures and yearly mean salinities over the middle shelf: both cooling of the middle shelf 46 Physical oceanography waters and the export of ice from the coastal to the middle shelf domains may be causally related to severe (cold and windy) winters, when south- ward outbreaks of cold and dry continental air cause more ice formation (see Overland, Chapter 2, this volume). Stirring and buoyancy addition A water column is stably stratified if the density increases towards the bottom. During spring and summer, this stable water column usually prevails because lighter, more buoyant fluid is added at the surface (ice melting, precipitation, or freshwater runoff) or because the surface waters become less dense as a result of warming (insolation). Alterna- tively, the addition of dense water at the bottom (intrusion of oceanic-regime water onto the shelf) makes the surface waters more buoyant than the bottom waters (Fig. 4-9). Processes that tend to stratify the water column stably by decreasing the density of the near-surface, we call positive buoyancy additions. If dense water is added at the surface (brine drainage during freezing) or if the surface waters are made denser (radiative cooling or evaporation), then the water column becomes less stratified or less stable. If water becomes denser than that below it, then the water column is unstable and vertical mixing (overturn) occurs. We call processes that destabilize the water column negative buoyancy additions. By buoyancy addition we mean any change in water properties that alters the mean density of the water column (as distinguished from mechanical stirring that redistributes the density). Mechanical stirring is an important process tending to mix the water column. Over this shelf, the main source of stirring is the tidal currents and a secondary source is the wind (Schumacher et al. 1979, Simpson and Pingree 1978). Most tidal stirring power (turbu- lence) is generated near the bottom, most wind stirring power (turbulence) at the surface. Thus, we attribute the surface mixed layer to wind stirring, and the bottom mixed layer to tidal stirring. Station 101 in Fig. 4-2 illustrates these two layers in which mechanical stirring is sufficient to keep temperature and salinity homogeneous over layers of 20 m or more thickness. At Station 126, stirring had over- come any stabilizing effects of positive buoyancy addition, and the entire (50 m) water column was well mixed (neutrally stable). Another way of viewing these two tendencies is to consider the potential energy of two water columns. Consider the first, like Station 101, to consist of two layers, while the second, like Station 126, is com- pletely mixed. If both columns have the same verticEilly averaged temperature, salinity, and thus density (assuming a linear equation of state), then all points from both stations fall on the same straight line on the TS plane; it is the vertical structure that differentiates between the two distributions. It requires mechanical work to mix the two-layered water column so that it looks like the homogeneous water column, because the center of mass of the well-mixed water column is higher than in the two- layer column. Over the Bering Sea shelf the primary source of this mixing energy is the tides. When the cross section across the inner front was made in 1976 (Fig. 4-6) we found two water columns like those just described. The homogeneous water column on the coastal domain side of the front could have been made by completely mixing the water column on the middle domain side of the front. On the shoreward side of the front, the tidal stirring was just competent to overcome the buoyancy addition from melting ice and insolation; thus fresh water and heat were mixed throughout the water column. On the seaward side of the front, however, stirring was inadequate. A wind-stirred surface layer and a tidally stirred bottom layer met at a sharp pyc- nocline. This interplay of buoyancy addition and stirring has some positive feedback. Stratification suppresses vertical mixing so that mixing is impeded after stratification forms, and as further buoyancy is added stratification increases. This added stratification further suppresses mixing, and so forth. This feed- back helps explain why the transition between the coastal and middle domains is so sharp: stratification enhances stratification, and well-mixed structure likewise tends to persist. Over the middle shelf the surface mixed layer and the bottom mixed layers meet, making the vertical structure distinctly two- layered. The middle front, separating the middle and outer domains, marks the limit of the ability of the two homogeneous layers to encompass the entire water column. Seaward of this front, in the outer domain, an interior region exists between the two mixed layers. Finestructure exists only in this interior; elsewhere it would be vertically mixed by stirring. Table 4-3 emphasizes the roles of stirring and buoyancy addition in forming the hydrographic domains. We can get a feeling for the reason why the domains are closely tied to the isobaths (50 m, 100 m, shelf break) by following the formalism of Simp- son and Hunter (1974), who examined a front like our inner front near the British Isles. They compared the rates of addition of potential energy by insolation and by stirring. Hydrographic structure 47 For a two-layered water column, potential energy (V) addition rate is approximately: dV^a^gh (Jm-^sec) (J = Joule) dt 2Cp a : a thermal expansion coefficient (kg°C-'m-3) Q: insolation (J m"^ /sec) g: acceleration of gravity (m/sec' ) h: water depth (m) C: specific heat (J/kg/°C) p : density (kg/m^ ) The major annual change is in the insolation term (Q); other buoyancy terms (e.g., melting ice) could be added easily. The turbulent energy available for stirring is simply: ^-kpU^ (Jm-^sec) dt where: k - drag coefficient, p = density (kg/m^ ), and U^ = mean of cubed speed (m^ /sec^ ). Most of this power (note that 1 J/sec = 1 watt) does not mix the water, but the relative amount (1 percent or so) that does go into mixing seems con- stant for a given flow regime (e.g., near the inner front). We can see that the buoyancy addition term has small changes across the shelf in all of its terms but h, while in the stirring term U^ changes across the shelf. The tidal current, U, is also a function of depth (h) and position on the shelf (Pearson et al.. Chapter 8, this volume), so that both buoyancy addition and stirring are dependent on location. Although neither buoyancy addition nor stirring changes very rapidly at a given location, the tidal currents vary signifi- cantly over two-week cycles (fortnightly tide), winds vary, and buoyancy input changes diurnally and annually; but an important variation across the shelf can be seen by taking the ratio of dE/dt and dV/dt. The result is a constant times U^ /h; across the shelf, from the outer to the coastal domain, this changing ratio reflects the changing balance between buoyancy and stirring. Nearshore, since U^ is large and h small, stirring prevails. Farther offshore, because U^ decreases and h increases, stratification (given sufficient Q or other buoyancy source) prevails. We even found that this held in February 1978 when we took measurements in melting ice: seaward of the 50 m isobath the water column was TABLE 4-3 Stirring and buoyancy addition in the iiydrograpiiic domains Coastal domain Tiiroughout tiie year tidal and wind stirring produce adequate mixing power to overcome normal sources of buoyancy: insolation, melting ice, and river runoff. Exceptions to this are probably short lived, except in river plumes within 10-20 km of the coast. (Water depth < thickness of tidal-mixed layer.) Middle domain Tidal and wind stirring are inadequate to mix the entire water column during the high buoyancy-addition season (spring and summer). The vertical structure during that season is two- layered: a wind-stirred surface layer and a tidal-stirred bottom layer separated by a sharp pycnocline. During fall and winter, when buoyancy addition is usually negative, the vertical structure is uniform, but the potential for stratification remains. Melting ice can establish two-layered stratification, even in winter. (Water depth = thickness of tidal-mixed layer + thickness of wind-mixed layer.) Outer domain The surface mixed layers and bottom mixed layers do not meet; the pycnocline is weaker than in the middle domain and there is an interior region separating the mixed layers. Fine- structure is ubiquitous within this stratified interior region. Even in winter, negative buoyancy and stirring are insufficient to mix the water column completely. More saline water from the oceanic regime makes the deep column denser than the surface waters, even if the surface waters are cooled to the freezing point. (Water depth > thickness of tidal-mixed layer + thickness of wind-mixed layer.) See also our Table 4-2 and Figure 24 in Coachman and Charnell (1979). two-layered, while shoreward of the 50 m isobath the column was well mixed (the 50 m isobath coincides with the inner front during summer). Thus the potential for stratification (expressed by U^ /h), is always present, requiring only sufficient buoyancy addition to be realized. Our data do not reveal variability in frontal loca- tion, either on short time scales such as diurnal 48 Physical oceanography or fortnightly, or longer scales such as interannual. There is undoubtedly some variation in the location of fronts on many scales, but the inner front is tied closely to its mean position by the variation of U^ /h, and similar considerations probably affect the middle front's position also. Stirring and buoyancy addition form the vertical hydrographic structure within the coastal and middle domains, and modify the structure of the outer domain. In summary, cool surface water appears often in Bristol Bay during spring and summer, and Myers (1976) presented hydrographic evidence that this results from upwelling. Whether this persistent feature actually results from upwelling, from another dynamic response to the flow regime, or from vertical mixing, however, is not known. DISCUSSION Upwelling The cold surface patch observed in Bristol Bay during summer has been ascribed to upwelling. Myers (1976) documented the frequent occurrence of cooler surface water in Bristol Bay during spring and summer, and our own data also showed this (Kinder et al. 1978). Myers attributed this to upwelling forced by an Ekman convergence in the bottom boundary layer. This convergence was caused by a mean cyclonic flow that approximately follows the 50 m isobath. Arguments based on hydrographic evidence from 1969-70 presented by Myers favored upwelling of water originating southwest of the cool surface patch rather than local vertical mixing, but the explanation of this upwelling was incomplete. The mean flow is only about 2 cm/sec, while tidal speeds are about 20 cm/sec (Chapter 5, this volume). Thus, the tidal kinetic energy is 100 times that of the mean flow, and tidal effects may be more important than the mean flow. For the Ekman convergence to work, water must be forced upward against stratification, rather than forced horizontally to the west or south- west (where there is no mean flow). Moreover, Myers hydrographically inferred that the source of upwelled water is southwest of Bristol Bay, but his proposed Ekman convergence requires a source to the east and north. A possible alternative, strong wind during summer, occurs too infrequently to account for this persistent feature. In the open ocean, with upper layers moving faster than lower layers, large (nearly geostrophic) cyclonic features are associated with isopycnals that dome upward; perhaps the mean flow does influence the observed distributions in Bristol Bay. A secondary circulation would then be neces- sary to maintain the density structure against tidal stirring and mixing. On the other hand, a combina- tion of vertical mixing driven by tidal currents and freshening of inshore waters by river runoff could produce the observed hydrographic distributions. This seems more in harmony with processes over the remainder of the shelf, but is no more proved than the upwelling hypothesis. Cross-shelf fluxes On the basis of conservation of heat and salt, we have discussed some estimates of cross-shelf fluxes in terms of diffusion coefficients. Knowledge of these fluxes, and of the mechanisms driving them, is important for both conservative and non-conservative material, e.g., larvae, nutrients, plankton, and petroleum. EspeciEilly in summer, dispersion charac- teristics differ in the different domains. Vertical exchanges are severely damped in the strongly strati- fied middle domain, while complete vertical mixing occurs rapidly in the tidaUy stirred coastal domain. Dispersion also differs horizontally in the three domains. In the outer domain interleaving on fine- structure scales is an important component of mixing processes, but no finestructure is found in the inner two domains. Mixing, although probably driven by the dominant tidal currents in both nearshore domains, differs between the middle and coastal domains. Over the two-layered middle shelf hori- zontal transport may differ markedly in each layer (e.g., a nearly estuarine two-layer flow), but this is unlikely in the vertically homogeneous coastal domain. There is also a question of steadiness of these fluxes: how much do they vary and over what time scales? Coachman and Charnell (1979) estimated that the horizontal salt flux in the outer domain was three to four times greater than the fluxes at the shelf break and middle fronts. This implies a divergence (depletion) of salt transport near the shelf break and a convergence (accumulation) near the 100 m iso- bath—an imbalance which cannot persist over long periods without altering the observed long-term salinity distribution. There is some annual variation in fluxes, as the hydrographic structure, wind stress, and ice cover all change significantly over the year. The variability of these fluxes, and particularly their timing (or phasing) with respect to critical biological events, may be more important than the mean fluxes. As yet, we can only roughly estimate mean values, and we do not understand the processes that drive these fluxes. Hydrographic structure 49 Fronts It is not clear whether the fronts separating the hydrographic domains are convergences or diver- gences—whether they enhance or inhibit mixing. These boundaries separate distinct hydrographic domains and probably dynamic ones, and they have large gradients in various properties. The methods of transport of passive properties, such as salt and nutrients, and of dynamic properties, such as momen- tum and vorticity, probably change across these fronts, and these changes are most clearly seen in the vertical hydrographic structure. Intuition suggests that fronts are convergences (James 1978), and that cross-frontal exchanges are impeded (e.g., methane distributions shown by Cline, this volume). A con- vergence throughout the depth and length of the inner front, for instance, seems unlikely; but neither observation nor modeling have answered the ques- tions of convergence and cross-frontal mixing. There is evidence of year-to-year (Coachman and Charnell 1979) and annual (Schumacher et al. 1979, Kinder and Coachman 1978, Coachman and Charnell 1979) variability of the fronts, and Schumacher et al. (1979) reported wavy features in satellite images of the inner front that imply more rapid variability. The longer-term fluctuations seem related to changes in atmospheric forcing (e.g., insolation, temperature, storms), and the wavy features may be frontal insta- bility (inherent). Further understanding of these changes will probably add knowledge of variations in cross-shelf fluxes. ward. Waters offshore are cooled directly not only by the atmosphere, but by melting ice that originally formed nearer shore. Because these processes are forced by weather, changes in the winter weather are manifest in ice cover and therefore in the shelf hydrography. Ice processes thus affect the shelf hydrographically in two ways: through melting and freezing, ice lo- cally redistributes salt and heat in the water column, changing the vertical stratification; and it directly influences shelf-wide heat and salt budgets by acting as an insulating cover while transporting salt and heat. SUMMARY The southeastern shelf has a distinct hydrographic structure. Proceeding seaward from the coast in summer one encounters the vertically homogeneous coastal domain, the inner (structural) front, the two-layered middle domain, the middle front, the outer domain, the shelf break front, and finally the bordering oceanic domain (Fig. 4-11). These features can best be understood by considering these simplifi- cations: 1. In the middle and coastal domains mean advection is negligible. 2. Water mass transformations occur locally, primarily through heat and salt transfer at the surface. Role of ice Ice, with strong annual and interannual variation, influences the hydrography of this shelf in several different ways. These effects are both local and shelf-wide. Locally, ice affects the energy balance and vertical distributions of heat and salt. Ice cover effectively insulates the water and slows heat transfer (both radiative and conductive-con vective), and ice covered with snow has high albedo, reflecting incoming short- wave radiation. Freezing and melting also alter the distribution of heat and salt. Ice acts as a buffer for temperature as changing heat balance alters the amount of ice present at nearly constant temperature. Local freezing and then melting causes a vertical redistribution of salt, so that a water column that has uniform salinity in fall may have haline stratification in spring. Freezing and melting also influence shelf-wide distributions of heat and salt. Freezing nearshore and melting offshore transport both salt and heat shore- 3. Vertical profiles are determined by the interplay of buoyancy addition and mech- anical stirring, and in the outer domain also by lateral interleaving between shelf and oceanic waters. 4. Rates of buoyancy addition change annually, while stirring (primarily tidal) remains nearly constant with time and increases shoreward. During winter, the separation into these domains is less clear, and the addition of negative buoyancy and stronger wind stirring move the boundary of vertical homogeneity seaward through the middle shelf. Even during this season, however, the potential for stratification like that in summer remains, and melting ice can provide sufficient buoyancy to stratify waters in the middle domain. The hydrographic structure influences mixing, and the system of domains and fronts affects many distributions: e.g., salt, heat, momentum, vorticity, sediment, benthos, plankton, nutrients, fish, and 50 Physical oceanography pollutants. With few exceptions, we do not under- stand these effects, and in many cases we do not even know what the effects are. As we have suggested, some effects of salinity and temperature distributions and their interactions with the hydrographic structure are straightforward, but many others are not. Future studies of the shelf will have to consider the influence of hydrographic structure on many phenomena. OCEANIC OUTER MIDDLE COASTAL DOMAIN DOMAIN DOMAIN DOMAIN Shelf brear Middle 1 nner Front '' Front Front 500km 1 50km 1 I20km |50km| 150km |I0| 200 km Figure 4-11. A schematic of the cross-shelf density structure illustrating the system of hydrographic domains separated by fronts. This picture represents summer conditions, when the structure is clearest. Vertical profiles are shown beneath each domain. See Tables 4-2 and 4-3 for a tabula- tion of domain properties. ACKNOWLEDGMENTS Primary funding came from the Outer Continental Shelf Environmental Assessment Program, which is administered by the National Oceanic and Atmos- pheric Administration for the Bureau of Land Man- agement. While writing this paper T. H. Kinder was supported by the Naval Ocean Research and Devel- opment Activity. This is PMEL contribution number 425. Bob Charnell was a coprincipal investigator on this project, and Pat Laird was frequently chief scientist on project cruises. Both were lost at sea off Hawaii in December 1978. REFERENCES Arsenev, V. S. 1967 Currents and water masses in the Bering Sea. (Transl. 1968, Nat. Mar. Fish. Serv., Northwest Fisheries Cen- ter, Seattle, Wash.) Brower, W. A., Jr., H. F. Diaz, A. S. Prechtel, H. W. Searby, and J. L. Wise 1977 Climatic atlas of the outer continental shelf waters and coastal regions of Alaska. AEIDC, Anchorage, Alaska. Many people contributed to the work reported here, and we list only those who directly helped us prepare reports or manuscripts: L. K. Coachman, R. L. Charnell, R. B. Tripp, D. J. Pashinski, J. C. Haslett, N. P. Laird, R. L. Sillcox, and K. Ahlnas. L. K. Coachman was principal investigator with us during this project. There was also a small army of engi- neers, technicians, computer specialists, and secre- taries at the University of Washington, Pacific Marine Environmental Laboratory, and Naval Ocean Re- search and Development Activity whose efforts made this chapter possible. The officers and crews of Acona, Moana Wave, Discoverer, Surveyor, and Miller Freeman spent many uncomfortable hours at sea supporting the field program. F. Favorite read an earlier draft and made many helpful comments. T. C. Royer and A. W. Green reviewed this paper and made insightful criticisms. Coachman, L. K. 1978 Water circulation and mixing in the Southeast. Progress report on Proc- esses and Resources of the Bering Sea shelf (PROBES), Proc. Rep. Inst. Mar. Sci., Univ. of Alaska, Fairbanks. 1979 On the oceanographic role of Arctic shelves. Unpub. MS, presented at Fall AGU meeting, 4 Dec. 1978. Coachman, L. K., and R. L. Charnell 1977 Finestructure in outer Bristol Bay, Alaska. Deep Sea Res. 24: 869-89. 1979 On lateral water mass interaction— a case study, Bristol Bay, Alaska. J. Phys. Oceanogr. 9: 278-97. Hydrographic structure 51 Dodimead, A. J., F. Favorite, and T. Hirano 1963 Salmon of the North Pacific Ocean, Part II. Review of oceanography of the subarctic Pacific region. Inter. N. Pac. Fish. Comm. Bull. 13. Favorite, F. 1967 The Alaskan stream. Inter. Fish. Comm. Bull. 21: 1-20. N. Pac. Favorite, F., A. J. Dodimead, and K. Nasu 1976 Oceanography of the subarctic Pacific region. Inter. N. Pac. Fish. Comm. Bull. 33. Garvine, R. W., and J. D. Monk 1974 Frontal structure of a river plume. J. Geophys. Res. 79: 2251-9. Iverson, R. L., L. K. Coachman, R. T. Cooney, T. S. English, J. J. Goering, G. L. Hunt, Jr., M. C. Macauley, C. P. McRoy, W. S. Reeburg, and T. H. Whitledge 1980 Ecological significance of fronts in the southeastern Bering Sea. In: Ecologi- cal processes in coastal and marine systems. Plenum Press, N. Y. Jacobs, W. C. 1951 James, I. D. 1978 Joyce, T. M. 1977 Kinder, T. H. 1977 The energy exchange between sea and atmosphere and some of its conse- quences. Bull. Scripps Inst, of Ocean- ography 6(2): 27-122. A note on the circulation induced by a shallow sea front. Estuarine and coastal marine science 7: 197-202. A note on the lateral mixing of water masses. J. Phys. Oceanogr. 7: 626-9. The hydrographic structure over the continental shelf near Bristol Bay Alaska, June 1976. Dep. of Ocean- ography, Univ. of Washington. Tech. Rep., Ref: M77-3. Kinder, T. H., J. D. Schumacher, R. B. Tripp, and J. C. Haslett 1978 The evolution of the hydrographic structure over the continental shelf near Bristol Bay, Alaska, during summer 1976. Univ. of Washington, Dep. of Oceanography, Tech. Rep. Ref: M78-16. Kinder, T. H., J. D. Schumacher, R. B. Tripp, and D. Pashinski 1977 The physical oceanography of Kotze- bue Sound, Alaska, during late sum- mer, 1976. Univ. of Washington, Dep. of Oceanography, Tech. Rep. Ref: M77-99. Kitano, K. 1970 A note on the thermal structure of the eastern Bering Sea. J. Geophys. Res. 75: 1110-15. Muench, R. D. 1976 A note on eastern Bering Sea hydro- graphic structure, August 1974. Deep Sea Res. 23: 245-7. Muench, R. D., and K. Ahlnas 1976 Ice movement and distribution in the Bering Sea from March to July, 1974. J. Geophys. Res. 81: 4467-76. Myers, R. L. 1976 Ohtani, K. 1973 On the summertime physical oceano- graphy of Bristol Bay, 1969-1970. M.S. Thesis, Univ. of Alaska. Oceanographic structure in the Bering Sea. Memoirs of the Faculty of Fisheries, Hokkaido Univ. 21: 65-106. Postmentier, E. S., and R. W. Houghton 1978 Finestructure instabilities induced by double diffusion in the shelf/slope water front. J. Geophys. Res. 83: 5135-8. Correction, 1979, J. Geo- phys. Res. 84: 1847. Kinder, T. H., and L. K. Coachman 1978 The front overlaying the continental slope of the eastern Bering Sea. J. Geophys. Res. 83: 4551-9. Reed, R. K. 1978 The heat budget of a region in the eastern Bering Sea, summer, 1976. J. Geophys. Res. 83: 3636-45. 52 Physical oceanography Reed, R. K., and W. P. Elliott 1979 New precipitation maps for the North Atlantic and North Pacific Oceans. J. Geophys. Res. 84: 7839-46. Roden, G. I. 1967 On river discharge into the north- eastern Pacific Ocean and the Bering Sea. J. Geophys. Res. 72: 5613-29. Sayles, M. A., K. Aagaard, and L. K. Coachman 1979 Oceanographic atlas of the Bering Sea. Univ. of Washington Press, Seattle. Schumacher, J. D., T. H. Kinder, D. J. Pashinski, and R. L. Charnell 1979 A structural front over the continental shelf of the eastern Bering Sea. J. Phys. Oceanogr. 9: 79-87. Simpson, J. H, and R. D. Pingree 1978 Shallow sea fronts produced by tidal stirring. In: Oceanic fronts in coastal processes, M. J. Bowman and W. E. Esaias, eds., 29-42. Springer-Verlag, N.Y. Straty, R. R. 1977 Current patterns and distribution of river waters in inner Bristol Bay, Alaska. NOAA Tech. Rep. NMFS SSRF-713. Taken outi, A. Y., and K. Ohtani 1974 Currents and water masses in the Bering Sea: A review of Japanese work. In: Oceanography of the Bering Sea, D. W. Hood and E. Kelley, eds., 39-57. Inst. Mar. Sci. Univ. of Alaska, Fairbanks, Occ. Pub. #2. Simpson, J. H., and J. R. Hunter 1974 Fronts in the Irish Sea. 404-6. Wiseman, W. J. Nature 250: 1979 Hypersaline bottom water: Peard Bay, Alaska Estuaries 2 : 189-93. ^ Circulation Over the Continental Shelf of the Southeastern Bering Sea Thomas H. Kinder' and James D. Schumacher^ National Space Technology Laboratories Station, ' Naval Ocean Research and Development Activity, Bay St. Louis, Mississippi ^ Pacific Marine Environmental Laboratory, Environmental Research Laboratories/National Oceanic and Atmospheric Administration Seattle, Washington I ABSTRACT Using extensive direct current measurements made during the period 1975-78, we describe flow over the southeastern Bering Sea shelf. Characteristics of the flow permit us to define three distinct regimes, nearly coincident with the hydrographic domains defined in the previous chapter. The coastal regime, inshore of the 50 m isobath, had a slow (1-5 cm/sec) counterclockwise mean current and occasional wind- driven pulses of a few days' duration. The middle regime, bounded by the 50 and 100 m isobaths, had insignificant (<1 cm/sec) mean flow but relatively stronger wind-driven pulses. The outer regime, between the 100 m isobath and shelf break (—170 m), had a 1-5 cm/sec westward mean and low- frequency events unrelated to local winds. Over the entire shelf most of the horizontal kinetic energy was tidal, varying from 60 percent in the outer regime to 90 percent in the coastal regime. About 80 percent of the tidal energy was semidiurnal. Mean flow over the shelf is well described qualitatively by dynamic topographies, and shallow current data from both coastal and outer regimes agree quantitatively as well. Two meteorological conditions that force the observed current pulses have been identified. In summer eastward -traveling low atmospheric pressure centers caused low-frequency pulses in the middle regime, and weaker pulses in the coastal regime. In winter, outbreaks of cold and dry continental air forced pulses within the coastal and middle regimes. INTRODUCTION Until recently, few direct current measurements were available over this shelf, so that ideas about flow were based partly on indirect methods and partly on intuition (see the historical review which follows). Since 1975, however, we and many colleagues have gathered numerous direct current measurements with concurrent hydrographic data. We are thus able to base our characterization of the flow over this shelf on plentiful information. Our analysis of this suite of data (Table 5-1) is still incomplete. We expect further analysis building on this preliminary report to improve understanding, but not to change fundamen- tally the conclusions that we present here. The most important discovery as a result of these new data is the existence of three distinguishable flow regimes over the shelf. These flow regimes corres- pond closely to the hydrographic domains outlined in the preceding chapter. The coastal regime is shore- ward of the 50 m isobath, the middle regime is between the 50 m and 100 m isobaths, and the outer regime is between the 100 m isobath and the shelf break. Seaward of the shelf break lies the oceanic regime. Although they nearly coincide, for clarity we refer to flow regimes and hydrographic domains in this and the previous chapter. We emphasize the characteristics of these regimes as we examine the frequency distribution of the hori- zontal kinetic energy, the mean circulation, and seasonal variations. Before discussing our findings, we review the physical setting, highlight earlier work, and discuss our measurements. Setting The southeastern shelf is bounded by the Alaska Peninsula, the Alaskan mainland, the shelf break 53 54 Physical oceanography running from Unimak Pass to the Pribilof Islands, and an arbitrary line running from the Pribilofs to Nuni- vak Island (Fig. 5-1). The shelf break occurs at an average depth of 170 m (Scholl et al. 1968), and the shelf shoals gradually over a featureless expanse of 500 km. Flow over the continental slope is highly variable (Kinder et al. 1975, Coachman and Charnell 1979, Kinder et al. 1980) with mean flow to the northwest at 5-10 cm/sec. The shelf regime is separa- ted from the adjacent oceanic regime by a weak haline front (Kinder and Coachman 1978, Coachman and Charnell 1979). We have found no convincing evidence for exchange of mass or momentum be- tween the shelf and oceanic regimes by eddies or rings. This characteristic is probably the effect of some combination of the width of the shelf, the front overlying the slope, and the weakness of the boun- dary current above the slope. At the same time, a considerable volume of water must flow across this long (~ 1,000 km) shelf break somewhere: about 1 X 10^ m'^ /sec flows northward through the Bering Strait into the Chukchi Sea (Coachman and Aagaard, 62°- 61°- 60°- 59°- 58° 57°- 56° 55°- 54' St Matthew 1 BERING SEA Figure 5-1. Locations of current-meter moorings. Most moorings had two current meters, one 10 m above bottom and one 20 m below the surface. BC signifies Bristol Bay project and FX signifies Frontal Experiment. Sequential deploy- ments at a mooring site are designated by letter suffixes, e.g., BC-9A. The three moorings southwest of Nunivak Island numbered 1-3 have FX prefixes; all others have BC prefixes. Circulation over the shelf of the southeastern Bering Sea 55 Chapter 8, this volume) and river runoff and precipi- tation account for only about 2 percent of this. There is strong evidence that little of this required flow occurs southeast of the Pribilofs, and hydrographic distributions suggest that it occurs near Cape Navarin on the Siberian coast. As we have explained more fully in the previous chapter, the shelf has three distinct hydrographic domains delimited by the 50 m and 100 m isobaths and by the shelf break (170 m isobath). Freshwater runoff from rivers at the coast and an excess of precipitation over evaporation maintain a surface salinity difference across the shelf of about 2*^/oo. First-yeair ice covers most of the shelf during winter, and the weather also changes with the season: these changes are discussed more fully in the chapters on ice (Pease, McNutt, this volume) and on weather (Overland, Niebauer, and Ingraham, this volume). In summary, five factors influence shelf circula- tion: (1) The width of the shelf and the weak haline front over the slope separate much of the shelf from the adjacent oceanic water masses. (2) There is no strong boundary current adjacent to the shelf. (3) The shelf has three distinct hydrographic domains. (4) River runoff maintains a salinity gradient across the shelf. (5) Weather and ice cover vary annually. HISTORICAL REVIEW Until recently, investigators faced the serious problem that there were few data from which infer- ences about circulation could be made. And yet the need existed for those conducting such non-physical studies as geological, biological, and fisheries studies to know the circulation. For this reason circulation schemes were proposed based on various mixtures of current measurements, hydrographic measurements, drifter measurements, and large doses of intuition. It is not surprising that many of these inferences do not compare favorably with our present knowledge, but some ideas resulting from these early attempts do agree with our present ideas. There are long reference lists of egirlier work in Arsenev (1967), Ohtani (1973), and Favorite et al. (1976). Before 1975 direct current measurements virtually did not exist. A single 24-hour current measurement taken close to Nunivak Island during August 1934 indicated northward flow at 17 cm/sec (Barnes and Thompson 1938). During the summers of 1955 and 1956 the Tokei Maru attempted current measure- ments north of the Alaska Peninsula, but mean currents were obscured by the tides (International North Pacific Fisheries Commission 1957). Hebard (1961) reported June 1957 current measurements at four sites. Each measurement lasted 39 hours, and he inferred a cyclonic (counterclockwise) mean flow around Bristol Bay. Kinder and Coachman (1975) placed three current-meter moorings in Pribilof Canyon for two weeks during July 1974. They measured vector mean speeds from 0.6 to 10.8 cm /sec, and directions mostly parallel to the local isobaths. Some efforts were also made to infer currents from drift bottle experiments (e.g., Thomp- son and Van Cleve 1936, Dodimead et al. 1963), but little was learned from the few recoveries. Attempts were also made to describe the circula- tion using dynamic topographies. On the face of it these attempts seem forlorn, for neglected forces (e.g., friction, sea surface slope, wind stress) may be more important than those retained (i.e., Coriolis and the pressure gradient caused by varying density). In spite of this, mean flows on shelves are often qualita- tively described by dynamic topographies, and some features in the Bering topographies agree with direct measurements. Natarov (1963) showed a surface topography (relative to 1,000 db' ) with broad northward flow across the shelf for summer 1959. Arsenev (1967) constructed a similar map of summer surface current based on 1,000 db. His map showed about 10 cm/sec cyclonic circulation in Bristol Bay, two eddies farther seaward, and northeastward (i.e., shoreward) flow between the Pribilofs and Nunivak Island. Adding wind-driven (Ekman) current weak- ened one eddy and changed the Pribilofs-Nunivak flow to northwestward. Neither Natarov nor Arsenev explained their method for extrapolating dynamic topography relative to 1,000 db from the deep basin across the shelf. Ohtani (1973) also calculated a surface dynamic topography, using a reference level of 50 db (his calculations did not expend inshore of the 50 m isobath). He showed a broad northwest- ward flow, but cautioned that such topography has limited application in shallow water. Attempts were also made to deduce the flow from examination of hydrographic properties. Kitano (1970) apparently thought that the cold bottom water over the middle shelf in summer was advected southeastward from the Gulf of Anadyr (north- western Bering Sea shelf). Conversely, Takenouti and Ohtani (1974) showed flow in the opposite direction ' One decibar (db) is approximately equivalent to one meter depth. 56 Physical oceanography through the middle shelf based on their perception of shelf water masses. Finally there were circulation schemes, opinions of various authors based on the meager useful data available and their own intuition. The U.S. Navy Hydrographic Office (1958, reproduced by Hughes et al. 1974 as their Fig. 3.4) reported mean currents of 0.1- 0.5 knots (5-25 cm/sec) across the southeastern shelf based on ship drift; a later publication (Naval Oceanographic Office 1977) revealed that this erron- eous result was not based on data (Brower et al. 1977 showed a similar circulation in their Figs. 3 and 4). The most ambitious attempt was by Favorite et al. (1976, their Figs. 41 and 42). They hypothesized mean flow paralleling the shelf break (transverse current, identical to Kinder, Coachman, and Gait's 1975 Bering Slope Current), flow onto the shelf paralleling the Alaska Peninsula (West Alaska Cur- rent), cyclonic flow around Bristol Bay, no flow over most of the middle shelf, and a southwestward current from near Kuskokwim Bay toward the Pribilofs (Pribilof Current). They did not assign speeds to these flows, but recent data support their general scheme, except for the Pribilof Current. In general, eairly attempts to describe flow over the shelf were unsuccessful and even deceptive. The few direct measurements used were overwhelmed by tidal currents and of too short duration to resolve the mean. Dynamic heights referred to 1,000 db were alleged to show flow over the shelf. Local transfor- mations of water masses were ignored and distribu- tions of temperature and salinity were attributed solely to advection. Circulation schemes were drawn showing more detail than the evidence justified. We present our own circulation scheme of the mean flow later in this chapter but we hope that it is used with two questions in mind: What is the statistical sig- nificance of this depiction of the mean flow? What is the physical significance of the mean flow? METHODS Measurements Our inferences are based on 60 current-meter time- series records made during 1975-78 (Table 5-1, Fig. 5- 1). Usable record lengths varied from 9 to 246 days, and a total of about 16 record-years of data were obtained. All data were acquired by RCM-4 Aanderaa current meters on taut-wire moorings. Typical instrument locations were 20 m below the surface and 10 m above bottom, as we avoided both the most active part of the surface layer and the bottom boundary layer. Since we also wanted the shallowest instrument on each mooring to be above the seasonal pycnocline, which averaged about 20 m deep, shallow instrument placement was a compromise. Much has been vnritten about the peculiarities of these instru- ments, and their possible errors due to "rotor pump- ing," mooring motion, and rotor stalling (Quadfasel and Schott 1979 give several references). In our measurements these errors were minimized by : (1) use of subsurface flotation (although the upper float was typically within 20 m of the surface); (2) short mooring length, all less than 200 m and typically less than 100 m; (3) use of taut moorings, about 1,000 lb (1 lb = 4.45 newton) buoyancy; (4) large and persistent (tidal) scalar speeds, pre- venting rotor stalling; (5) large tidal currents, which maintained rela- tively steady vane alignments. All the data that we use have been averaged in some manner, so that random errors are not a serious problem. Biases will not average out, and we estimate direction accurate to ±5° and speeds within ±1 cm/sec (exclusive of rotor pumping and mooring mo- tion). Mayer et al. (1979) discussed similar moorings, and found an erroneous increase in energy up to two- fold (i.e., 40 percent speed increase) in winter on the middle Atlantic shelf. In the Bering Sea strong sum- mer storms are infrequent, and ice cover severely damps surface waves during much of the winter, so that these estimates seem too high for our moorings. We estimated the speed error by examining the most energetic tidal constituent, the principal lunar semidiurnal (M2 ). From records BC-9B and BC-20B (see Table 5-1 and Fig. 5-1) we took one-month segments from the shallow (23 m and 22 m depth) instruments. We compared periods when the mooring was beneath ice cover to periods when the surface was free of ice, assuming that the ice damps surface waves (e.g., Wadhams 1978) so that there was no error induced by waves during the period of ice cover. We further assumed that the M2 constituent was unchanged by ice cover (it may be affected, but it is probably diminished by ice, making our calculations conservative; see Chapter 8, Pearson et al., this volume). We found a 19-26 percent speed increase during the ice-free season. Although the errors probably vary significantly with weather, we believe that our speeds are accurate to about 25 percent near 20 m depth, and the speeds from deeper instru- ments are probably more accurate than this. We also used drifters, floating buoys with "window shade" drogues at 10 m or 17 m depths. The tracks of these surface buoys closely correspond to the tra- TABLE 5-1 Current-meter records Vector Water Meter Scalar mean Record Mooring depth depth speed speed Direction length Period (m) (m) (cm/sec) (cm/sec) CT) Days COASTAL REGIME FX-IA 48 38 21.2 1.7 294 63 7/20-9/20/78 FX-2A 43 20 32.4 1.9 306 63 7/19-9/20/78 FX-3A 46 14 33.1 2.3 319 63 7/20-9/20/78 36 21.9 1.7 295 63 7/20-9/20/78 BC-9A 41 17 28.0 0.1 248 60 6/2-7/31/76 27 24.8 0.4 251 60 6/2-7/31/76 BC-9B 41 23 28.6 4.4 309 231 9/24/76-5/12/77 33 16.1 2.9 311 231 9/24/76-5/12/77 BC-9C 41 23.5 24.6 1.0 316 122 5/12-9/10/77 33.5 19.3 0.9 279 120 5/12-9/8/77 BC-14A 51 20 34.9 1.4 086 31 5/29-6/28/76 20 33.1 4.8 065 33 8/27-9/28/76 37 27.4 2.1 211 31 5/29-6/28/76 37 20.6 0.6 345 33 8/27-9/28/76 BC-15A 50 20 28.8 2.4 274 118 5/31-9/26/76 BC-15C 50 20 29.9 2.2 273 130 5/4-9/10/77 34 22.4 1.0 304 130 5/4-9/10/77 BC-16A 49 20 28.2 0.5 314 131 5/3-9/10/77 37 21.2 0.2 Oil 131 5/3-9/10/77 BC-18A 31 20.5 24.5 1.2 Oil 122 5/12-9/12/77 BC-19A 28 22.5 27.6 0.8 MIDDLE REGIME 156 12 5/12-5/23/77 BC-2A 65 20 17.2 0.5 306 58 9/8-11/5/75 50 10.0 0.9 301 59 9/8-11/5/75 BC-2B 65 50 17.6 1.2 089 192 11/5/75-5/14/76 BC-2C 65 20 14.8 0.8 324 119 5/31-9/26/76 BC-2D 65 21 17.1 1.9 080 194 9/27/76-4/8/77 BC-2E 65 20 16.9 0.7 214 130 5/4-9/10/77 50 15.3 0.3 251 130 5/4-9/10/77 BC-4A 55 30 29.3 2.7 299 58 9/8-11/4/75 47 21.9 0.6 311 58 9/8-11/4/75 BC-4B 55 30 25.4 2.0 301 209 11/5/75-5/31/76 BC-4C 55 25 27.8 2.7 312 61 6/1/76-7/31/76 52 19.5 1.3 299 61 6/1/76-7/31/76 BC-4D 55 20 31.9 3.6 310 119 9/25/76-1/21/77 48 18.2 3.0 313 174 9/25/76-3/17/77 BC-4E 55 20 27.7 1.1 300 64 5/13/77-7/15/77 48 18.3 0.4 274 45 5/13/77-6/28/77 BC-4G 55 18 30.1 3.9 294 64 7/19-9/20/78 46 16.9 2.1 302 64 7/19-9/20/78 BC-5A 70 20 19.6 1.2 253 31 5/29-6/28/76 20 15.5 0.2 350 33 8/27-9/28/76 50 27.8 0.9 298 31 5/29-6/28/76 50 31.6 1.8 170 33 8/27-9/28/76 BC-6A 76 20 11.3 0.6 012 123 5/29-9/28/76 50 23.3 0.7 023 31 5/29-6/28/76 50 18.2 1.4 059 33 8/27-9/28/76 BC-8A 73 26 21.8 0.9 089 60 6/1-7/31/76 54 21.9 0.6 348 60 6/1-7/31/76 57 58 Physical oceanography TABLE 5-1, cont. Vector Mooring Water Meter Scalar mean Record depth depth speed speed Direction length Period BC-lOA 66 49 21.9 2.0 171 68 6/1-8/8/76 BC-12A 95 39 9.2 0.6 OUTER REGIME 278 84 3/19-6/11/76 BC-3A 115 20 29.0 3.2 Oil 130 11/7/75-3/16/76 BC-3B 116 25 27.8 2.8 334 9 3/17-3/25/76 105 17.4 1.2 342 73 3/17-5/28/76 BC-3C 114 20 31.8 17.2 012 123 5/29-9/28/76 100 20.5 6.7 000 123 5/29-9/28/76 BC-13A 122 20 20.6 3.3 353 69 3/22-5/29/76 100 11.9 1.6 348 87 3/22-6/16/76 BC-13B 115 100 17.3 1.6 335 36 6/6/76-7/12/76 BC-13C 108 22 25.2 5.3 333 202 9/29/76-4/19/77 96 16.1 4.4 321 83 9/29-12/21/76 BC-17A 104 96 18.1 3.2 ST. MATTHEW 296 142 9/22-3/11/77 BC-20A 64 22 23.8 3.4 335 157 9/17/77-2/20/78 52 16.7 1.4 356 177 9/17/77-3/13/78 BC-20B 64 51 14.6 2.8 334 54 7/21-9/13/78 BC-21A 42 29 32.4 1.5 292 246 9/16/77-5/20/78 BC-21B 42 20 28.8 2.3 338 55 7/20-9/13/78 32 20.3 0.8 317 55 7/20-9/13/78 The prefix (i.e., BC, FX) indicates the project (Bristol Bay or Frontal Experiment), and indicates the number of the mooring location. The suffix refers to a particular deployment, and individual instruments are identified by their depths. Thus BC-9B 20 m refers to a record obtained at 20 m depth from the second deployment at mooring 9 by the Bristol Bay Project (Fig. 5-1 shows locations). Some records are separated because of midsummer biological fouling problems encountered near the Alaska Peninsula (e.g., BC-14A). jectories of water parcels at drogue depths. Using the Nimbus satellite, the position of these drifters was fixed about four times daily. This time series of posi- tions (error = ±4 km) was then processed to yield velocities. Kinder et al. (1980) discuss this technique in more detail. Processing Sample intervals of the current meters vairied from ten minutes to one hour, depending on the expected length of deployment. All data were filtered, with one or two low-pass filters (i.e., they pass low fre- quencies and block higher ones). All data were originally filtered with a three-hour filter to minimize high-frequency noise. To examine lower frequencies, a further 35-hour low-pass filter was used to remove the tides. Table 5-2 lists basic filter properties, and Charnell and Krancus (1976) discuss processing de- tails. The 3-hour low-pass filter removes unwanted (for our purposes) high-frequency data, while the 35- hour filter removes diurnal and semidiurnal signals while leaving periods longer than about two days intact. Note that 50 percent amplitude corresponds to 25 percent kinetic energy; the commonly used "half-power point" occurs at longer periods (lower frequencies) than the 50 percent amplitude point. TABLE 5-2 Filter properties Period (hours) with given % amplitude remaining 3-Hour Filter 35-Hour Filter 0.1% 2 25 50% 2.9 35 99% 5 hours 55 hours Circulation over the shelf of the southeastern Bering Sea 59 FREQUENCY DISTRIBUTION OF HORIZONTAL KINETIC ENERGY One of the most useful ways of examining time- series data of water velocity is to calculate the kinetic energy distribution as a function of frequency. Some of the ocean's processes occur at distinct frequencies, or in discernible frequency bands: often a chaotic jumble in plots of velocity versus time becomes clear when energy versus frequency is plotted. The most striking example is the tidal currents, whose frequencies are determined by astronomical constants. For the southeastern shelf, we define three fre- quency categories: mean, low-frequency (subtidal), and tidal. The mean flow is the vector average over a few months or longer. We have in mind the flow over a season or longer, but in practice we have usually defined the mean by the length of the current record. Two tidal periods stood out, diurnal (about 24-hour period) and semidiurnal (about 12-hour period). Low-frequency (long period) flow then fell between seasonal and daily periods. Although there were no definite periods associated with the low-frequency energy, one week was typical. In most records, these three frequency categories contained over 90 percent of the kinetic energy. Tidal frequencies contained most of the energy over the shelf, ranging from 60 percent of the fluctuating energy^ near the shelf break to more than 95 percent in some records obtained inshore of the 50 m isobath. Roughly 80 percent of this tidal energy was semidiurnal, and 20 percent diurnal (Pearson et al.. Chapter 8, this vol- ume, describe the tides more fully). Vector mean speeds were usually 10 percent or less of the mean scalar speeds (Table 5-1), so that the kinetic energy of the mean flow was about 1 percent of the totEil horizontal kinetic energy (KE^^MV^ /2, or per unit mass KE/M=V^ /2). Of the 60 records in Table 5-1, only three had vector mean speeds exceeding 5 cm/sec, while only six records had scalar mean speeds below 15 cm/sec. Low-frequency flow, that appearing at frequencies between those of tidail and mean flow, accounted for 3-20 percent of the energy. Over the outer shelf, energies in this frequency band were higher than farther inshore. Frequencies in these bands matched frequencies of weather phenom- ena, say 2-10 days (see weather chapters), and of the inherent variability of the mean flow over the outer shelf and slope (e.g., eddies and meanders). The ^ The fluctuating kinetic energy (per unit mass) is the total kinetic energy, less the kinetic energy of the mean (FE=V2 (U-U)^ ). It is often referred to simply as the energy or total energy and is equal to one-half the variance. inertial period (about 14 hours at 58° latitude), which is important in mainy open-ocean records of kinetic energy, accounted for only 1 percent or less of the total energy in our records. Table 5-3 illus- trates the frequency distribution of the horizontal kinetic energy in some typical records. Basic plots We can illustrate the character of flow over the shelf most clearly by showing plots of typical velocity records. Although there were significant differences between records and between regimes, we attempt to highlight common features of the records. Fig. 5-2 shows unfiltered data from BC-15A (Fig. 5-1) at a depth of 20 m during September 1976.^ Samples are plotted every 20 minutes (the current- meter sample interval) as two components. The upper plot is the east-west component (U), with eastward flow positive, and the lower plot is the north-south component (V), with northward flow positive. Two features are obvious from this depic- tion. First, a tidal signal overwhelmed any other variable signals present, and it was mostly semidiurnal (about 12-hour period). This record came from the coastal regime, where the tides contributed 90 percent of the fluctuating kinetic energy, and where the semidiurnal tide typically accounted for 80 percent of the tidal energy. Second, the upper plot is displaced below the zero axis, while the lower plot is centered about this line. Thus, for the week shown, o 50n A SEPT Figure 5-2. A portion of BC-15A 20 m in September 1976. This record is unfiltered, and the sampling interval was 20 minutes. A. East-west(U) velocity component and B. North-south(V)velocity component. Strong semidiurnal and less strong diurnal tides show clearly. The U compo- nent is displaced below the axis, indicating a mean flow to the west. ^ The draftsman has inadvertently low -pass filtered this record somewhat. Original plots made with fine pens show consider- ably more jiggles (high-frequency energy). 60 Physical oceanography TABLE 5-3 Frequency distribution of horizontal kinetic energy (cm^ /sec^ )' Record Mean Fluctuating^ Low Diurnal Semi- diurnal Inertial Coastal Regime BC-9B, 23 m, winter BC-9C, 23 m, summer Middle Regime BC-2B, 50 m, winter BC-2C, 20 m, summer Outer Regime BC-3C, 100 m, summer BC-3C, 20 m, summer BC-3A, 20 m, winter ' cm^ /sec^ is a unit of energy per unit mass, i.e., erg/g (1 erg/g = 10^ J/kg). ^ Fluctuating kinetic energy is the total kinetic energy in the records, less the kinetic energy of the mean. It is one-half the record variance. Note that the semidiurnal tidal energy does not change much between summer and winter. Percent of fluctuating energy is shown in parentheses. 2(0) 1(0) 434(100) 340(100) 59(14) 6(2) 74(17) 73(21) 271(62) 256(75) 1(0) 0(0) 1(0) 0(0) 203(100) 175(100) 11(5) 3(2) 68(33) 49(28) 120(59) 110(63) 0(0) 1(1) 20(8) 8(29) 5(1) 266(100) 510(100) 511(100) 30(11) 149(29) 112(22) 42(16) 60(12) 65(13) 165(62) 251(49) 285(56) 2(1) 7(1) 6(1) the mean current was westward (cf. record mean of 2.4 cm/sec at 274°T). The entire BC-15A record had one of the clearest and most consistent means of all those that we have examined, but usually attempting to measure average currents in flows like this with records of a few days would be futile. Fig. 5-3 shows a different presentation, progressive vector diagrams (often abbreviated PVD), all smoothed with the three-hour low-pass filter. Pro- gressive vector diagrams show what the trajectory of a water parcel would have been if it had had the velocity measured at the site of the current meter. This is much different from familiar depictions of flow such as trajectories (paths of particles) or streamlines (lines parallel to velocity vectors). Fig. 5-3 shows progressive vector diagrams from each of the three regions: outer (BC-13A 20 m), middle (BC-6A 50 m), and coastal (BC-15A 20 m). The record mean was largest at the outer mooring (note different distance scales), and least at the middle mooring. All records had vigorous clockwise tidal motion, which was nearly circular in the inner record, and elliptical in the middle record (major axis parallel to the local isobaths). During much of the outer record the tidal circles were superposed on strong low- frequency flow so that they appear as cusps. These subtidal frequency pulses typically lasted for a few days (time may be inferred in Fig. 5-3 by counting clockwise semidiurnal circles), and can be seen in all three records. In the outer record, both speed and direction of pulses were variable. The middle and coastal records showed pulses that ap- peared similar at each occurrence, northward at BC-6A and westward at BC-15A. Our division of the flow into three frequency bands (mean, low- frequency, tidal) can be seen in each of these exam- ples, and the examples also emphasize the unsteadi- ness of the flow. Low-frequency plots To examine the low-frequency (subtidal) flow, the 35-hour low-pass filter was applied to records before plotting. Plots of components versus time (Fig. 5-4 A, 4B) and stick plots (Fig. 5-4C) were then made. Stick plots are a time series of vectors, the base of the Circulation over the shelf of the southeastern Bering Sea 61 OUTER SHELF MIDDLE SHELF COASTAL o N KILOMETERS EAST Figure 5-3. Progressive Vector Diagrams (A) Outer regime, BC-13A 20 m. (B) Middle regime, BC-6A 50 m. (C) Coastal regime, BC-15A, 20 m. These figures are the displacement of a water parcel having the same velocity as measured by the current meter, and may be quite different from the actual trajectory of any parcel. The outer record showed tides (cusps), strong pulses, and a northward mean. The middle record showed strong tides (clockwise ellipses), weak northward pulses, and perhaps a weak northeastward mean. The coastal record showed strong tides, westward pulses (extending the tidal circles like a loose spring or slinky toy), and a strong westward mean. Note different scales. (S signifies the start and F the finish of each plot.) vector indicating the time, the direction away from the base indicating the direction toward flow, and the length of the vector indicating speed. The record from BC-15A at a depth of 20 m shows generally westward flow at 5 cm/sec, with higher speeds later in the record. We found that during the generally low wind speeds (<5 m/sec) in summer there was little correlation between wind and current. During strong winds (>10 m/sec), especially in autumn when higher winds are common and we have many records, wind and currents were much better correlated. The record segments with close correlation seemed to occur for two to four days during the passage of atmospheric low-pressure centers through the area. Spectra Except for truly dominant frequency components (e.g., tides), it is difficult to describe the frequency distribution of the records from the plots we have presented thus far. A useful technique for doing this (and the one used to make the estimates for Table 5-3) is spectral einalysis. A time series is mathemati- cally (Fourier) transformed, so that instead of veloc- ity as a function of time, we have kinetic energy as a 22 21 AUG 76 10 SEP 76 Figure 5-4. Low-frequency flow. Low-pass (35 -hour) filtered data from BC-15A 20 m. ■ (A) East-west compo- nent, (B) North-south component, and (C) Vector sticks (cm/sec). Much of the low-frequency flow occurred during strong (>10 m/sec) wind events. 62 Physical oceanography function of frequency. When plotted the spectrum gives a concise picture of the frequency distribution of the kinetic energy (variance). Of course, the characteristics that showed clearly on other plots should also be evident in the spectral plots. Calculating the spectrum for one series (autospec- trum) can be useful, but this can also be done for two records (cross spectrum). Normalizing such a spec- trum by the variances produces a "coherence" as a function of frequency. The coherence is the frequen- cy-dependent counterpart of the correlation function between two series as a function of time lag; instead of showing the time lags at which records are corre- lated, it shows the frequencies at which records are phase locked (the calculation also yields the corres- ponding phase spectrum, a frequency-dependent time difference). We plot the square of coherence because, like the squared correlation coefficient, the squared coherence corresponds to the percent of the variance at the frequency that is "explained" by a linear relationship. Coherence is always positive, but phase information (not shown) conveys sign. For instance, two sine waves of identical frequency but opposite sign have a coherence of 1.0 and a phase of 180°. We illustrate with examples that summarize the frequency distribution of kinetic energy over the shelf. Autospectra for BC-15C, 20 m and 34 m depth, from summer 1977 are plotted in Fig. 5-5. The vertical axis is kinetic energy density, i.e., per unit frequency, (cm/sec)^ /CPD (CPD is cycles per day). In this presentation the vector series have been decomposed into clockwise ( an ticy clonic) and counterclockwise (cyclonic) rotating components, instead of into north-south and east-west compo- nents. When there is no strong local orientation (e.g., a shoreline or strong bathymetric gradient), the clockwise/counterclockwise decomposition is more useful than the east/north decomposition. The most striking feature of both records is the high diurnal (1 CPD) and semidiurnal (2 CPD) peaks, with the semidiurnal higher. This conforms to our knowledge that motion over the shelf is tidally dominated (cf. Figs. 5-2 and 5-3). For the shallow record 89 percent of the energy was tidal (70 percent semidiurnal) and for the deeper record 94 percent was tidal (67 per- cent semidiurnal). At low frequencies there were small irregulair peaks. This also corresponds to expectation, as the low-frequency filtered records did not reveal any dominant frequencies (cf. Fig. 5-4). Near the inertial frequency (1.7 CPD), in the clock- wise components only, there is a small, statistically insignificant peak. One characteristic of inertial 10000^ 10000 :ii 0. 1 95Z A:: 1.50 l'. 00 l'. 50 2'. 00 2.50 CYCLES/DRY 10000a 10000 .. 1000 .. 100 10 0. 1 1.00 0.50 1 .00 1 .5C CYCLES/DRY 2.00 2.50 Figure 5-5. Autospectra, BC-15C 20 m and 34 m. The energy density (cm^ sec"^ CPD'' ) versus cycles per day shows strong diurnal and semidiurnal tidal peaks, and irregular low-frequency peaks. Clockwise (solid) and counterclockwise (dashed) current components are plotted. Circulation over the shelf of the southeastern Bering Sea 63 oscillations (in the Northern Hemisphere) is clockwise rotation, and the presence of similar peaks near the inertial frequency in most records has led us to label these peaks inertial. An example of vertical coher- ence from mooring 5A at 20 m and 50 m depth during summer 1976 is shown in Fig. 5-6. Again, we used clockwise and counterclockwise components. Not surprisingly, the tidal peaks show clearly. Tidal motion at 20 m depth was coherent with motion at 50 m depth. In our records vertical coherence > 0.9 at tidal frequencies was common, and since most of the energy was at tidal frequencies, this implies that shallow and deep motions were generally coherent. As with most of the records there are also some low-frequency peaks above the 95 percent confidence limit (~0.26). Examining individual records (plotted versus time), we often found that shallow and deep records were highly correlated during strong winds, but over the entire record low-frequency flow was often weak and uncorrelated. This is consistent with coherence estimates like that in Fig. 5-6. In general then, we find that motion throughout the sampled water column was highly coherent at tidal frequencies, and either weakly coherent or not 0.99j_ 0.90.. 0.20.., I 0.10..^ 0.00. (VERTICAL COHERENCE)' 1.0 K5~ CYCLES/DRY . 95Z 2.5 Figure 5-6. Vertical coiierence squared. The squared co- iierence of BC-5A 20 m with BC-5A 50 m is shown, in clockwise (solid) and counterclockwise (dashed) compo- nents. Highly significant coherences in tidal bands are common to all records. Some also show strong low- frequency coherence and inertial (1.7 CPD, clockwise) coherence. coherent at other frequencies. When the coherence was high, the phase differences were usually small, so that the flow at these frequencies was in the same direction at both depths. During strong winds, the low-frequency flow was vertically correlated (see below. Seasonal Variations and Meteorological Forc- ing^ and this probably caused the occasional peaks in coherence at low frequencies. Coherence calculations do not reveal relative speeds, but an examination of Table 5-1 shows that shallow instruments usually recorded faster speeds than deeper ones. This may be instrumental (see Introduction) or it may reflect an actual decrease with depth. Just as it is possible to calculate the vertical coher- ence to help determine whether motions are similar throughout the water column, horizontal coherences calculated between records from different moorings help determine if motions are similEir over broad areas. The low-frequency sticks in data from July 1976 suggested to us that the water motion at low frequencies might be coherent over part of the shelf (see Fig. 5-10); we calculated coherence between BC-5A (50 m), BC-2C (20 m), and BC-6A (50 m). The deep instrument on mooring BC-2C failed, but the current meter at a nominal depth of 20 m was below the pycnocline. Calculations between these instruments showed coherence at separations of 30, 42, and 72 km. Again, the calculated tidal coher- ences squared exceeded 0.9 (Fig. 5-7). At low frequencies there was also high coherence squared in a broad peak which decreased with separation: at 30 km it was ~0.7, and at 72 km it was ~0.55. These peaks suggested that the middle shelf was similarly affected by low-frequency motions, and probably most of these motions were generated by wind events similar to those discussed under Seasonal Variations and Meteorological Forcing. Values of coherence squared between the middle shelf and coastal records at comparable distances (~60 km) were lower (~0.4), but still significant. MEAN CIRCULATION Although the kinetic energy distributions were dominated by tidal currents, many of the records had significant means (e.g., Fig. 5-3A). We examine the statistical significance of the record means, and construct a circulation scheme using them. The mean flow is energetically two orders of magnitude smaller than the tidal flow over much of the shelf, however, and the mean circulation may not be the most important flow component in many situations. A vivid example is the flow through the Aleutian 64 Physical oceanography BC 5fl BC 2C i.e 1.0 1.5 CYCLES/DRY HORIZONTAL COHERENCE SQUARED BC 5R BC 6fi BC 6fl BC 2C 0.0 0.5 1.0 1.5 CYCLES/DRY 2.0 0.5 1.0 1.5 CYCLES/DRY 2.0 Figure 5-7. Horizontal coherence squared. Even at a separation of 72 km, low-frequency (-0.3 CPD) coherence squared is nearly 0.6, and at 30 km it exceeds 0.7. The horizontal coherence was higher over the middle shelf (shown here) than in the coastal or outer regimes. Clockwise components have solid lines. Passes (such as Unimak Pass). Flows in many of these passes have high mean scalar speeds, exceeding 100 cm/sec (see U.S.D.C. 1964), but these are mostly caused by tidal currents and may average to a vector mean of zero. Properties (e.g., salt and drift cards) can be exchanged between the North Pacific and Bering Sea through such a pass; it is unlikely that the same volume of water flows back and forth on ebb and flood, and the vigorous stirring near coast and bottom prevents such a volume of water from remain- ing intact. Practically, it is difficult to measure the mean flow in such passes with present techniques. For many oceanographic questions, the mean flow is not as important as the exchange of properties between two oceanic regimes and the mixing by vigorous tidal stirring.'* Just as obviously, the mean circulation over the Bering Sea shelf, even if much smaller than the tidal currents, can importantly influence many processes. For instance, plankton advected by a 2 cm/sec vector mean flow would travel about 150 km over the summer. Such a small mean flow is difficult to measure in a highly variable and active (noisy) back- ground. The 2 cm/sec mean flow may be important, but separating it from the tidal and other variable currents with statistical significance requires measure- ments of long duration. For example, we had six "Mean flow is often assumed through Unimak Pass into the Bering Sea; we know no evidence for this, only for an ex- change between the Bering and the North Pacific. different moorings at location BC-4 (Fig. 5-1), and for each mooring we computed a mean (Table 5-1) from records 45-209 days long, and instruments 18-52 m deep. These means accounted for less than 1 percent of the horizontal kinetic energy in this tidally dominated flow. All means fell in the northwestern quadrant, varying from 274 to 311°T with speeds from 0.4 to 3.6 cm/sec. The unweighted means for these records (± one standard deviation) were: 2.1 ± 1.2 cm/sec and 301 ± 11°T. Although we cannot compare this to the actual mean, the repeatability gives us confidence in our data. Some of the variabili- ty was not due to measurement error: seasonal differences (wind stress, ice cover, and stratification), vertical velocity shear (the means of shallow records averaged 1.2 cm/sec faster than the deeper ones), and interannual changes accounted for some variation. A reasonable qualitative estimate of the signifi- cance of the mean flow can be inferred from progres- sive vector diagrams (Fig. 5-3). To quantitatively estimate the statistical significance of the means, we used low-pass (35-hour) filtered records. We define a time scale^(T) of the low-frequency flow and then estimate the number of independent samples of the mean by dividing into total record length (T). Thus an estimate of the error (at 67 percent confidence) is: ^ The time scale is estimated by computing the area under the autocorrelation function; it is thus twice the "integral time scale" (Kundu and Allen 1976). Circulation over the shelf of the southeastern Bering Sea 65 E = iT (T/rr E: root mean square error estimate (cm/sec) a: standard deviation of record (after low-pass filtering; cm /sec) T: record length (sec) T: time scale (sec), the length of an independent sample. An example of these calculations is given in Table 5-4, illustrating the significance of means for summer 1976. For these records, the mean was usually significant in the coastal regime, where it was about three times the estimated error. Over the middle TABLE 5-4 Means and standard errors Record U±Eu (cm/sec) V±Ev (cm/sec) S, E' Coastal BC-15A (20 m) -2.3 ± 0.3 0.1 ± 0.5 2.3, 0.6 BC-14A(20m) 3.3 ± 0.4 1.2 ± 1.0 3.5, 1.1 (40 m) -0.4 ±0.4 -0.6+0.4 0.7, 0.6 Middle BC-2C (20 m) 0.4 ± 0.4 0.6 ± 0.3 0.7, 0.5 BC-5A(20m) -0.1+0.5 -0.5 ± 0.4 0.5, 0.7 (50 m) -0.1 ±0.3 0.1 ±0.7 0.1, 0.8 BC-6A (20 m) 0.6 ± 0.6 0.2 + 0.4 0.6, 0.7 (50 m) 0.6 +0.4 0.8 ±0.4 1.0, 0.6 Outer BC-17A (96 m) -2.9 ± 0.1 1.4 ± 0.4 3.2, 0.4 BC-13C(22m) -2.4 ± 1.9 4.7 ± 1.5 5.3, 2.4 U and V are the record means, and Eu and Ev are the estimated standard errors (see text). S and E are the vector mean speed and the vector sum of the errors. 'S= (U^ + V") '/^ E = (Eu^ + Ev^ )'/=> shelf, however, the mean was generally equal to or less than the standard error. In the outer regime, the mean was much greater than the error. These exam- ples conform to a general pattern: in the outer regime, means were greater than errors, in the coastal regime means were somewhat greater than errors, and in the middle regime means were insignificant. This generalization must be modified for measurements taken just seaward of the inner front separating the coastal and middle hydrographic domains (e.g., BC-4, Fig. 5-1). Such measurements show significant means because of their proximity to the front. These measurements also illustrate that the boundaries of the flow regimes nearly coincide with those of the hydrographic domains, but not exactly. Computed means for each mooring location (Fig. 5-8) show the general circulation pattern. Over the entire shelf, significant mean flow usually pEiralleled local isobaths. In the outer regime mean flow was parallel to the shelf break, flowing northwestward at 1-5 cm/sec. In the middle regime mean flow was in- significant, usually <1 cm/sec. In the coastal regime between Cape Newenham and Nunivak Island flow was westward at 1-3 cm/sec, while along the Alaska Peninsula flow paralleled the nearby shoreline at 2-5 cm/sec. Drifters, tracked by satellite, provided confirma- tion of flow character in the outer and middle re- gimes. Three drifters launched in June 1976 (near BC-2, -5, and -6) confirmed low vector mean speeds there, <1 cm/sec in the absence of strong winds. In 1977, six drifters released over the continental slope in the outer and oceanic regimes all drifted north- westward with vector mean speeds ~5 cm/sec (Coach- man and Chamell 1979, Kinder et al. 1980). Two of these drifters wandered into the middle regime near the Pribilofs, where their vector mean speed was -^1 cm/sec. These drifters gave no evidence of exchange across the shelf break (i.e., crossing the front over- lying the continental slope. Kinder and Coachman 1978), except in the area just north of the Alaska Peninsula. Kinder et al. (1978) found water lying along the Alaska Peninsula below 20 m that was both warmer and more saline than surrounding waters. Muench and Ahlnas (1976) noted that the region just north of the Alaska Peninsula was often free of ice when nearby water was ice-covered. These distributions could result from the advection of warmer oceanic water onto the shelf north of the peninsula. More- over, one of the drifters deployed over the slope in 1977 crossed the shelf break north of the Unimak Pass, and traveled about 80 km eastward before grounding on the peninsula (Kinder et al. 1980). 66 Physical oceanography 62°- 61 60°- 59°- 58° 57°- 56°- 55°- 54°- BC CURRENT METER MOORINGS ~IOm ABOVE BOTTOM ~20m BELOW SURFACE 180 DAYS PER BARB KILOMETERS 50 meters I 00 meters 200 meters 2000 meters Figure 5-8. Mean flow. The mean for all records at each mooring site is shown. Coastal and outer regime moorings generally had statistically significant means, while middle shelf sites did not. The domains refer to hydrographic structure (preceding chapter), but the domains and flow regimes are nearly coincident. These measurements are strong evidence for flow across the shelf break near the Alaska Peninsula (note BC-14, Fig. 5-8), but hydrographic and current measurements also demonstrated that this flow is probably intermittent. Mean flow over both the outer and coastal regimes appeared to be driven by pressure gradients caused by density differences. The method of inferring flow from dynamic calculations, referred to the deepest common depth of nearby hydrographic stations, gave approximate agreement with direct measurements (Coachman and Charnell 1979, Schumacher et al. 1979). Dynamic topographies in the outer regime (Kinder 1977, Coachman and Charnell 1979, Kinder et al. 1978) indicated westward flow of ~5 cm/sec. Calculations across the inner front (near the 50 m isobath) yielded 1-2 cm/sec flow counterclockwise along the front and westward between Cape Newen- ham and Nunivak Island (Schumacher et al. 1979). Thus mean flow in both the outer and coastal regimes can be inferred from the density field, but this is not true of the middle shelf, where the mean is nearly zero. Dynamic topographies do suggest a very weak southeastward flow across the middle regime (e.g., Kinder 1977, Reed 1978, and Kinder et al. 1978), but the current records (Fig. 5-8) did not confirm this flow. Synthesizing our knowledge of the shelf circu- Circulation over the shelf of the southeastern Bering Sea 67 Figure 5-9. Estimated mean circulation. No strong distinction is made as to season or deptii, altiiough it is biased toward summer and the surface. The dashed arrows in the northern coastal regime express uncertainty, while the arrow along the Alaska Peninsula expresses intermittency. Flow over the shelf is mostly tidal, so that the instantaneous flow is quite different from this depiction. For instance, over the middle shelf we expect the speed at any time to exceed 20 cm/sec even though the long-term vector mean is less than 1 cm/sec, and instantaneous directions seldom agree with the arrows at any location. The scheme is also incomplete: the source of water for the westward transport in the coastal regime is not shown. We speculate about this source in the text. lation, we have constructed a circulation scheme (Fig. 5-9). We emphasize again that the mean circulation is only a few percent of the total kinetic energy, and that even our lairge data sets do not define the mean flow adequately at some locations. That is, the arrows in Fig. 5-9 will often be incorrect instantane- ously and scalar speeds will be much higher than the long-term vector means. Directions apply throughout the water column and speeds at the surface; speeds decrease significantly with depth seaward of the 100 m isobath. Our data are most extensive in summer (Table 5-1), but most data from other seasons agreed with this depiction. In the coastal regime we show counterclockwise flow extending northward to St. Matthew Island with speeds of a few cm/sec. Within Kvichak and Bristol bays, measurements show that the counterclockwise flow extends to the coast (Straty 1977). The dashed arrow paralleling the Alaska Peninsula signifies intermittent flow. In the middle regime we show 68 Physical oceanography weak mean flow (<1 cm/sec), direction indetermi- nate. In the outer regime and over the slope we show strong (1-10 cm/sec) flow paralleling the shelf break and flowing westward. Wherever the flow was well defined, the entire water column moved in the same direction, but speed decreased significantly with depth in the deeper water seaward of the shelf break. This flow is a continuation of eastward flow north of the Aleutian Islands which then curves cyclonically toward the west in the southeastern corner of the deep Bering Sea (Kinder et al. 1980). We also note that eddies seem to be common over the deeper water, so that "mean" currents there may radically change direction for weeks or months (Kinder and Coachman 1977, Kinder et al. 1980). We have not observed comparable eddies in the shelf waters. We found no evidence of large flow onto the shelf; the flow leaving the shelf to the west in the coastal regime should also have small volumetric transport. We can estimate this outflow by assuming a 2 cm/sec flow over an average depth of 25 m along a section 150 km long, approximating conditions southwest of Nunivak Island to the inner front (i.e., spanning the coastal regime). This transport is 2 cm/sec X 25 m X 150 km = 0.08 X 10'' m^ /sec, or less than 10 percent of the Bering Strait outflow (~1 X 10*' m^/sec), but still larger than the river runoff (0.0015 X 10*" mVsec: Roden 1967). Water to supply this west- ward flow may come from the middle shelf; across the middle regime, between the front and the Pribi- lofs, a southeastward transport of 0.08 X lO*" m^/sec across 200 km averaging 70 m deep would be only 0.5 cm/sec, detectable in neither our current nor our hydrographic data. Inflow along the coastal regime north of the Alaska Peninsula (30 km wide by 25 m deep) would need to be a steady 10 cm/sec, which is not supported by data, although some lesser inflow does occur there. Thus the circulation scheme shown in Fig. 5-9 is incomplete because the source of water necessary to supply the westward transport in the coastal regime is not shown. We conjecture that this supply is mostly from the shelf northwest of our study region and that it flows southeastward into the middle regime. It may then enter the coastal regime in Bristol Bay, where the inner front appears weakest (see Chapter 4). Since the evidence for this circulation is weak, we exclude it from Fig. 5-9. SEASONAL VARIATIONS AND METEOROLOGICAL FORCING Some characteristics of the shelf environment change annually: insolation, river runoff, ice cover. wind stress, and atmospheric pressure. Even though the tides, which have little seasonal variation (Pearson et al.. Chapter 8, this volume), force most of the current, seasonal variation occurs in the non-tidal currents. These seasonal changes are small but still significant. We showed some of the hydrographic changes in the preceding chapter. An example of changes in atmospheric forcing is shown in Table 5-5 (see also Overland and Niebauer, Chapters 2 and 3, this vol- ume). During the winter, mean wind speeds are higher and storms are more frequent. Stratification is much weaker over most of the shelf, and we might thus expect more direct response to the wind in the absence of ice cover, which may alter the effect of the wind stress. Records from BC-2 and BC-9 indi- cate some seasonal changes, even when allowance is made for instrumental problems from more energetic surface waves (Table 5-6). Seasonal changes that we describe exceed the 25-30 percent error estimates (see Introduction).'' At BC-9 the shallow summer mean was 1.0 cm/sec toward 316° T. Decreasing the winter mean by one- third (a high estimate of the false inflation of speed by surface waves), the meein was 3.2 cm/sec toward 315°T. Likewise, BC-4 had higher vector mean speeds during winter, with the same direction as in summer. Another site with sufficient data to suggest TABLE 5-5 St. Paul weather 1975-76 Mean Mean pressure speed (mbar) (m/sec) Vector Variance mean (m^ /sec^ ) (m/sec, ° ) Summer 1976 Winter 1975-76 1011 1006 6.2 8.5 43 85 1.3 at 004 2.4 at 202 The weather station on St. Paul Island (northernmost of the Pribilofs) probably represents weather over the southeastern shelf better than the other stations, which are strongly influ- enced by topography. ' The seasonal differences are slightly different from what might be inferred from Tables 5-1 and 5-3 because: Tables 5-1 and 5-3 use entire record lengths instead of smaller seasonal segments and we examine the meteorological frequency band (1.5-10 day) for seasonal effects, a subset of the low-frequency category in Table 5-3. A sequential 29-day tidal harmonic analysis was used to estimate errors in the Introduction, whereas Table 5-6 lists wide-band tidal estimates (identi- cal with Table 5-3). Circulation over the shelf of the southeastern Bering Sea 69 TABLE 5-6 Seasonal kinetic energy (cm^ /sec^ ) Record Season Fluctuating Low' Semidiurnal Meteorological' Coastal BC-9C Summer 23 m 5/12-9/10/77 BC-9B Winter 23 m 9/24/76-5/12/77 Middle BC-2C Summer 20 m 5/31-9/26/76 BC-2B Winter 50 m 11/5/75-5/14/76 Outer BC-3C Summer 20 m 5/29-9/28/76 BC-3C Summer 100 m 5/29-9/28/76 BC-3A Winter 20 m 11/7/75-3/16/76 340 434 175 203 510 266 511 59 11 149 30 112 256 271 110 120 251 165 285 47 74 21 69 k ' Low-frequency energy includes all subtidal (less than diurnal frequency) energy in the record. Meteorological frequency energy is only that within the 1.5-10 day (0.67-0.10 CPD) band. Note the relative constancy of the semidiurnal tidal energy at each mooring site. seasonal changes is BC-2. Like BC-9 and BC-4, cur- rents at BC-2 were faster in winter, with a vector mean speed (~1.5 cm/sec) about double the summer value. Unlike the other two sites, however, at BC-2 the direction reversed from westward (~300°T) in summer to eastward in winter. The kinetic energy spectra showed that the greatest relative change occurred in the meteorological band, between 1.5 and 10 days (Table 5-6). This band corresponds to the increased variance seen in St. Paul wind data (Table 5-5). In the middle regime a record for 20 m in the summer of 1976 (BC-2) had meteoro- logical kinetic energy of 2 cm^ /sec^ , while the deep 50 m record was threefold greater in the winter of 1976-77—7 cm^ /sec^ . In the coastal regime 23 m records (BC-9) showed greater differences between seasons. In summer 1977 meteorological band variance was 4 cm^ /sec^ , but during the winter of 1976-77 it was 47 cm^ /sec^ . These changes for the middle and coastal regimes exceeded our error estimates, but changes in the outer regime (BC-3) did not. This reinforces the idea that subtidal energy in the middle and coastal regimes is mostly meteorologi- cally forced, while low-frequency energy in the outer domain is not forced by local weather. In the outer regime, the low-frequency energy is probably associ- ated with variations in the Bering Slope Current. Two different meteorological conditions can force low-frequency current pulses. One dominates the summer pulses and the second occurs only in winter. The summer condition is an eastward-moving low atmospheric pressure center, and the winter condition 70 Physical oceanography is a southward outbreak of cold continental air. Strong storms associated with atmospheric low- pressure cells also occur in other seasons, but the continental air outbreak is exclusively a wintertime phenomenon. Such a summer event occurred in late July 1976 when several current-meter moorings and drifters were placed in a line running northwest from the Alaska Peninsula (Fig. 5-10). (We suspect biological fouling of the upper instruments, especially BC-5 (20 m) and BC-6 (20 m), so that speeds in some records are too low.' ) As the wind increased to 10 m/sec and rotated towards the east, currents at moorings 5, 6, and 2 responded similarly, with about 30 5H 0 ^i:^^^;^\\\\\\ -^^ BC 5 20nn BC 5 50m BC 6 20nn BC 6 50m \\\\ ;~z^" \\l/^^... BC 2 20m BC 15 20m DRIFTER 535 — DRIFTER 544 — BC 14 20m — BC 14 50m speed ^ 1.0 cm/sec GEOSTROPHIC MODEL WIND FOR BC 2 Figure 5-10. Current response to a summer wind event, middle and coastal regimes. Stick diagrams for current meters, drogued drifters tracked by satellite, and estimated wind. As the wind increased beginning on 27 July 1976, currents also increased and followed the wind in a clock- wise rotation. Biological fouling probably decreased speeds at BC-5 20 m and BC-6 20 m. 'In Chapter 28, F. Favorite cites an 1894 report of jellyfish fouling fishing gear near the Alaska Peninsula. Our fouling problem in the same area was also caused by jellyfish. a one-day lag. That is, speeds increased, and direction rotated slightly clockwise toward the east after peak speeds were reached. The shallow current meter at mooring 14 showed a similar speed increase, but direction remained essentially parallel to the local bathymetric contours (Figs. 5-1 and 5-8). The two drifters still operating at that time had shown aimless wandering at mean speeds <1 cm/sec before the strong winds began. They were still near the moored instruments in late July (they were launched in early June), and responded strongly during this event. Both drifters accelerated to speeds near 20 cm/sec, and changed directions clockwise toward the east. The current pulse illustrated by Fig. 5-10 was typical of many similar pulses in the current records, and is the best documented such event. These pulses also seem characteristic of the low-frequency currents in the two nearshore regimes. In our records, 50 percent of the low-frequency (subtidal) variance lies in the band 0.7-0.1 cycles per day (1.5-10-day per- iod), which includes the repetition rate for storms. The continental air outbreak, which occurs fre- quently in winter, apparently affects both the cur- rents and ice distribution (see preceding chapter). High pressure over the Alaska mainland results in an offshore wind of cold, dry air blowing southward off the ice. This type of weather permits excellent satellite imagery (Fig. 5-11; see also Muench and Charnell 1977). Crossing the southern ice boundary, the cold, dry air receives heat and moisture from the underlying water. Clouds form downstream from the ice edge as long (hundreds of km) streamers parallel- ing the wind. A record from BC-2B (Fig. 5-12) concurrent with a southward outbreak of cold continental air showed a strong (20 cm/sec) current pulse to the south during the period 19-22 January 1976. As the satellite image shows, winds were blowing SSE near BC-2 by 20 January in excess of 20 m/sec, and rotated slightly from south to south-southeast (cf. Fig. 5-10). Similar pulses were observed during winter 1977 at BC-2. In January, a five-day pulse averaged 15 cm/sec to the southeast and in February flow exceeded 30 cm/sec to the northeast for 30 hours. Within the two inner regimes, winter brought ener- getic response to meteorological forcing (Table 5-6). Mean winds were stronger then (in the absence of storms), and strong winds occurred more frequently. Although the tides still dominated energy spectra, energy at meteorological frequencies played an in- creasing role, leading to the reversal of mean flow at BC-2. Ice cover probably modifies this forcing, but even under the ice (BC-9) wintertime winds cause current responses. Circulation over the shelf of the southeastern Bering Sea 71 '^■» •*">w NUNIVAK ISLAND iL GIL 020:21:10:23 5399 lOFE'SOO 20JAN76 H4 13S OlE Figure 5-11. An infrared satellite image of the Bering Sea on 20 January 1976. The Alaska Peninsula, Pribilof Islands, Nunivak Island, and other land features show clearly. The ice pack extended south of Nunivak Island, with some typical lineations, mostly tending east-west and parallel to the ice front, near the edge of the ice. Thin streamer clouds, normal to the ice front and parallel to the wind, extended southward. Fig. 5-12 shows measured currents and estimated winds. FLOW REGIMES AND HYDROGRAPHIC DOMAINS We have used flow regimes as a framework for discussion in the preceding sections, and now we summarize their characteristics (Table 5-7). As with the nearly coincident hydrographic domains (preced- ing chapter), our description is biased toward sum- mer. The discussion of seasonal variation showed, however, that many of the differences between regimes persist throughout the year. Energy of the fluctuating flow was 90 percent tidal over the middle shelf, about 80 percent in the coastal regime, but only 60-70 percent in the outer regime, where up to 30 percent of the energy was at periods greater than two days (Table 5-3). Inertial energy was about 1 percent of the energy in the outer regime, but barely detectable in the two shoreward regimes. The increase in tidal energy per unit mass in these two regimes is a consequence of shoaling. Larger fluctuating kinetic energy values at low frequencies in the outer regime probably reflect variability of the persistent westward-flowing current there. Oceanic currents usually have unsteady components with periods of a few days or longer. Spatial differences also appeared in coherence 12 Physical oceanography -.aexAl//^ CURRENT CM/SEC NIND M/SEC IjWV^'^^^- I I I I I I I I I I I I I I I I I I I I I I I 11 16 21 26 31 JANUARY 1976 Figure 5-12. Low-pass (35-hour) filtered current (upper, cm/sec) and wind (lower, m/sec), from BC-2B 50 m. Tiie wind data are from Fleet Numerical Weather Central estimates. Note the strong wind-driven pulse to the south during 19-23 January. Fig. 5-11 shows concurrent wind and ice conditions. calculations. Vertical coherence was significant at tidal frequencies everywhere, but at low frequencies only in the coastal regime, perhaps because the water column there is homogeneous, unlike the other two regimes, which have strong stratification (during summer). This stratification may confine responses (e.g., to wind events) to only part of the water column. Horizontally, tidal coherences were large everywhere. Significant low-frequency squared co- herences (~0.6 at 70 km separation) existed over the middle shelf during the summer of 1976, but were lower (~0.4) at similar distances between the middle and coastal regimes (i.e., across the inner front). Coherence squared was low in the outer regime and in the coastal regime, except at tidal frequencies (the number of records and their separations make this a tentative conclusion). The high coherence in the middle regime may be caused by the low background —the amount of low-frequency energy was small, and was mostly forced by wind events which have large spatial scales. Other processes may cause low- frequency motions in the outer and coastal regimes, and these competing processes can lower coherences calculated over long records. At low frequencies the clear meteorologically driven pulses, as illustrated in Figs. 5-10 and 5-12, were observed only over the middle shelf. Some vestige of these was seen in coastal records, but not in the outer regime. It may be that the response to these wind events is so weak that it is swamped in the regimes with more low-frequency energy (coastal and outer, see Table 5-3), but not in the middle regime. These spatial changes in flow thus occur at many frequencies: tidal, subtidal, and mean (see preceding section). They are nearly congruent with hydro- graphic domains, so that the fronts delimiting these domains also bound flow regimes. Some of the relationship between the hydrographic domains and the flow regimes is clear. In the outer regime the mean flow agrees with the dynamic topography, as is true in the coastal regime and near the inner front. Variations in response to wind forcing are less clear, but are probably related to the changes in stratifica- tion and water depth. SUMMARY With a preliminary analysis, we have characterized the flow over the southeastern Bering Sea shelf. The spatial and temporal coverage of the measurements were perhaps two orders of magnitude better than previously available. We have calculated the frequen- cy distribution of the flow, defined the separation of the shelf into three regimes, suggested annual varia- tions, and identified meteorological forcing. Flow above this shelf was mostly tidal: the scalar mean tidal speeds were about 20-25 cm /sec, increas- ing shoreward. About 80 percent of the tidal energy was semidiurnal (see Chapter 8, this volume). For individual records, up to 95 percent of the fluc- tuating energy was tidal, but both subtidal fluctu- ations and (record) means were significant. Differ- ences in these low-frequency and mean flows defined distinct flow regimes nearly coincident with previ- ously defined hydrographic domains (see preceding chapter). In the outer regime, between the shelf break and the 100 m isobath, mean flow was westward at 1-5 cm/sec, agreeing with the dynamic topography. Energy at subtidal frequencies appeared unrelated to local meteorology, and may have resulted from oscillations of the Bering Slope Current and the associated front overlying the continental slope. Farther inshore, between the 100 m and 50 m iso- baths, the middle regime had insignificant mean flow (although vigorous tidal currents exceeded 20 cm/sec). Near the inner front mean flows existed, however, at speeds of 1-5 cm/sec parallel to the 50 m isobath in a counterclockwise sense similar to calcula- tions of geostrophic flow across the inner front. North of the Alaska Peninsula there appeared to be a net eastward flow during summer, but it seemed to be intermittent. Ice cover and winds vary annually, and changes in flow were discernible over the seasons. In summer. I Circulation over the shelf of the southeastern Bering Sea 73 TABLE 5-7 Flow regimes Outer Middle Coastal Fluctuating horizontal kinetic energy Vertical coherence' Horizontal coherence' Wind events Mean^ 60-70% tidal 1% inertial only tidal only tidal obscure 1-5 cm/sec westward geostrophic balance 90% tidal 80% tidal inertial (?) inertial (?) only tidal tidal and low frequency tidal and only tidal low frequency (C^ ~0.7@60km) clearly present present 0.5 cm/sec 1-5 cm/sec random counter- Note^ clockwise geostrophic balance ' The characteristics of coherence are particularly tentative. Missing records, improperly positioned instruments (e.g., both current meters on the same mooring in lower layer), and poor spatial coverage prevent strong conclusions. (C^ is coherence squared.) ' In the vicinity of the inner front (near the 50 m isobath) speeds are 2-5 cm/sec and parallel the isobath (along-frontal) in a counterclockwise sense. ^ Also see Figs. 5-9 and 5-10. storms are less frequent and winds (in the absence of storms) are generally weaker (Chapter 3, this volume) than in winter. Currents directly driven by the wind were stronger and more common in winter than in summer within the coastal and middle regimes, but were not detectable in the outer regime in either season. Further analysis of data already collected will make the description of shelf circulation more com- plete. Moreover, progress will be made in understand- ing the dynamics of the shelf circulation and its interactions with the variations in hydrographic structure across the shelf with the help of clues from more detailed examination of annual changes. Annu- ally the mean wind stress, the wind stress variance, the stratification and its horizontal variation, the horizontal density (salinity) gradient, the freshwater discharge from rivers, and the extent of ice cover all undergo great changes. Understanding the effects of these on the circulation of this shelf will increase our understanding of shelf dynamics in general, and help us to understand the ecosystem of the Bering shelf. ACKNOWLEDGMENTS Many people contributed to the work reported here, and we list only those who directly helped us prepare reports or manuscripts: L. K. Coachman, D. V. Hansen, R. L. Charnell, R. B. Tripp, D. J. Pashinski, J. C. Haslett, N. P. Laird, R. L. Sillcox, and K. Ahlnas. L. K. Coachman was principal investigator with us during this project. There was also a small army of engineers, technicians, computer specialists, and secretaries at the University of Washington, Pacific Marine Environmental Laboratory, and Naval Ocean Research and Development Activity whose efforts made this report possible. The officers and crews of Acona, Moana Wave, Discoverer, Surveyor, and Miller Freeman spent many uncomfortable hours at sea supporting the field program. L.K. Coachman, 74 Physical oceanography J.R. Holbrook, and three reviewers made many help- ful criticisms of this chapter. Funding came from the Outer Continental Shelf Environmental Assessment Program, which is admin- istered by the National Oceanic and Atmospheric Administration for the Bureau of Land Management. This is PMEL contribution number 426. While writing this paper, T. H. Kinder was supported by the Naval Ocean Research and Development Activity. Bob Charnell was a co-principal investigator on this project, and Pat Laird was frequently chief scientist on project cruises. Both were lost at sea off Hawaii in December 1978. Dodimead, A. J., F. Favorite, and T. Hirano 1963 Salmon of the North Pacific Ocean, II, Review of oceanography of the sub- arctic Pacific region. Inter. N. Pac. Fish. Comm. Bull. 13. Favorite, F., A. J. Dodimead, and K. Nasu 1976 Oceanography of the subarctic Pacific region. Inter. N. Pac. Fish. Comm. Bull. 33. Hebard, J. F. 1961 Currents in the southeastern Bering Sea. Inter. N. Pac. Fish. Comm. Bull. 5: 9-16. Hughes, F. W. 1974 REFERENCES Arsenev, V. S. 1967 Currents and water masses in the Bering Sea. Transl. 1968, Nat. Mar. Fish. Serv., Northwest Fisheries Cen- ter, Seattle. Barnes, C. A., and T. G. Thompson 1938 Physical and chemical investigations in the Bering Sea and portions of the North Pacific Ocean. Univ. Wash. Pub. in Oceanogr. 3 (2): 33-79. Brower, W.A. Jr., H.W. Searby, J.L. Wise, H.F. Diaz, and A.S. Prechtel 1977 Climatic atlas of the outer continental shelf waters and coastal regions of Alaska, II, Bering Sea. Arctic Envi- ronmental Information and Data Center, Anchorage, Alaska. Charnell, R. L., and G. A. Krancus 1976 A processing system for Aanderaa current meter data. NOAA Tech. Mem. ERL PMEL-6. Coachman, L. K., and R. L. Charnell 1979 On lateral water mass interaction— a case study, Bristol Bay, Alaska. J. Phys. Oceanogr. 9: 278-97. L. K. Coachman, and K. Aagaard Circulation, transport, and water exchange in the western Bering Sea. In: Oceanography of the Bering Sea, D. W. Hood and E. J. KeUey, eds., 59-98. Inst. Mar. Sci., Univ. of Alaska, Occ. Pub. No. 2. International North Pacific Fisheries Commission 1957 Annual report for the year 1956. INPFC, Vancouver, Canada. Kinder, T. H. 1977 Kinder, T. H. 1975 The hydrographic structure over the continental shelf near Bristol Bay, Alaska. Univ. of Washington, Dep. of Oceanogr. Tech. Rep., Ref.: M77-3. and L. K. Coachman An analysis of current meter measure- ments from Pribilof Canyon. Univ. of Washington Dep. of Oceanogr. Tech. Rep.,Ref.:M75-121. 1977 Observation of a bathymetrically trapped current ring. J. Phys. Ocea- nogr. 7: 946-52. 1978 The front overlaying the continental slope of the eastern Bering Sea. J. Geophys. Res. 83: 4551-9. Kinder, T. H., L. K. Coachman, and J. A. Gait 1975 The Bering slope current system. J. Phys. Oceanogr. 5: 231-44. Circulation over the shelf of the southeastern Bering Sea 75 Kinder, T. H., J. D. Schumacher, and D. V. Hansen 1980 Observation of a mesoscale eddy: An example of mesoscale vairiability in the Bering Sea. J. Phys. Oceanogr. In press. Kinder, T. H., J. D. Schumacher, R. B. Tripp, and J. C. Haslett 1978 The evolution of the hydrographic structure over the continental shelf near Bristol Bay, Alaska, during summer 1976. Univ. of Washington, Dep. of Oceanogr. Tech. Rep., Ref.: M78-16. Kitano, K. 1970 A note on the thermal structure of the eastern Bering Sea. J. Geophys. Res. 75: 1110-15. Kundu, P. J., and J. S. Allen 1976 Some three dimensional characteris- tics of low frequency current fluctua- tions near the Oregon coast. J. Phys. Oceanogr. 6: 181-99. Mayer, D. A., D. V. Hansen, and D. A. Ortman 1979 Long-term current and temperature observations on the middle Atlantic Shelf. J. Geophys. Res. 84: 1776-92. h Muench, R. D., and K. Ahlnas 1976 Ice movement and distribution in the Bering Sea from March to July, 1974. J. Geophys. Res. 81: 4467-76. Muench, R. D., and R. L. Charnell 1977 Observations of medium scale features along the seasonal ice edge in the Bering Sea. J. Phys. Oceanogr. 7: 602-6. Natarov, V. V. 1963 On the water masses and currents of the Bering Sea, Tr. VIRNO, 48; Izv. TINRO. 40 (Transl. 1968: Soviet fisheries investigations in the north- east Pacific, 110-130, NTIS, Spring- field, Virginia.) Naval Oceanographic Office 1977 Surface currents Bering Sea including the Aleutian Islands. Spec. Pub. 1402 NP5. Ohtani, K. 1973 Oceanographic structure in the Bering Sea. Memoirs of the Faculty of Fisheries, Hokkaido Univ. 21: 65-106. Quadfasel, D., and F. Schott 1979 Comparison of different methods of current measurements. Dt. hydrogr. Z. 32: 27-38. Reed, R. K. 1978 The heat budget of a region in the eastern Bering Sea, summer, 1976. J. Geophys. Res. 83: 3635-45. Roden, G. I. 1967 On river discharge into the north- eastern Pacific Ocean and the Bering Sea. J. Geophys. Res. 72: 5613-29. Schumacher, J. D., T. H. Kinder, D. J. Pashinski, and R. L. Charnell 1979 A structural front over the continental shelf of the eastern Bering Sea. J. Phys. Oceanogr. 9: 79-87. Scholl, D. W., E. C. Buffington, and D. M. Hopkins 1968 Geologic history of the continental margin of North America in the Bering Sea. Marine Geol. 6: 297-330. Straty, R. R. 1977 Current patterns and distributions of river waters in inner Bristol Bay, Alaska. NOAA Tech. Rep. NMFS SSRF-713. Takenouti, A. Y., and K. Ohtani 1974 Currents and water masses in the Bering Sea: A review of Japanese work. In: Oceanography of the Bering Sea, D. W. Hood and E. Kelley, eds., 39-57. Inst. Mar. Sci., Univ. of Alaska, Fairbanks, Occ. Pub. No. 2. Thompson, W. F., and R. V. Van Cleve 1936 Life history of the Pacific halibut. (2) Distribution and early life history. Rep. Inter. Fish. Comm. 9. U.S.D.C. 1964 United States coast pilot. Vol. 9, Pacific and Arctic coasts. U.S. Dep. of Commerce, 217. Wadhams, P. 1978 Wave decay in the marginal ice zone measured from a submarine. Deep Sea Res. 25: 23-40. Circulation and Hydrography of Norton Sound R. D. Muench,''^ R. B. Tripp,^ and J. D. Cline' ' Pacific Marine Environmental Laboratory/ National Oceanic and Atmospheric Administration Seattle, Washington ^ Department of Oceanography University of Washington Seattle, Washington ^Presently at Science Applications, Inc. /Northwest Bellevue, Washington ABSTRACT Norton Sound was two-layered vertically in temperature and salinity during summer, the eastern portion being more strongly layered than the western. Weak cyclonic mean flow in the eastern upper layer was not reflected in the lower layer, which was nearly stagnant and contained remnant cold, saline water formed locally during winter. Maintenance of this extreme layering in the eastern sound despite shallow depths (~15 m) was due to buoyancy input as solar heating and fresh water, sufficient to offset vertical mixing generated by tidal currents. In the western sound, coupled northerly upper- and lower-layer flows reflected a northward net flow over the shelf west of the sound. At times a strong baroclinic coastal flow occurred off Nome, and the central western sound was a locus for vertical mixing due to impingement of currents upon a shoal area. During winter, the sound approached vertical homogeneity due to vertical convection consequent to cooling and ice formation. Although it was possible to define mean flows, the currents were dominated by events which reflected regional wind and atmospheric pressure patterns in an as yet uncertain fashion. Such flow events may exert a primary control over such features as the Yukon River plume, which was never observed to enter the eastern sound even though the upper-layer salinity there suggests that Yukon water may have entered the sound prior to our observations. INTRODUCTION Observations of temperature, salinity, and currents have been obtained from Norton Sound, Alaska, as part of an investigation of transport processes on the Alaskan continental shelves. The general physical oceanographic conditions within Norton Sound, as deduced from observations made between 1976 and 1978, are addressed in this paper. Since analyses of the observations are still in progress, this should be considered as an interim, working document rather than a set of firm, final conclusions. Physical setting Norton Sound is a shallow, high-latitude embayment extending eastward from the northern Bering Sea and forming an indentation in the central west coast of Alaska (Fig. 6-1). Its east-west length is about 220 km, and its width about 150 km. Depths vary from less than 10 m in the southern portion to more than 30 m in a trough-like feature which trends east-west in the nearshore region just south of Nome; average depth in the sound is about 20 m. Two promontories extend into the sound about two-thirds of the way toward its eastern end. Cape Darby from the north and Stuart Island from the south (Fig. 6-1). Norton Sound is located in a region of extreme seasonal variability. During the approximately June- September summer the sound is ice free and air temperatures are well above freezing. The waters are exposed for 24 hours to daylight, though not neces- sarily direct sunlight, during part of this period, and generally to light and variable winds. By November, air temperatures drop well below freezing and ice formation has normally begun along the northern shore, with first ice typically forming on the surface in Norton Bay at the northeast corner of the sound. 77 78 Physical oceanography 170* 166' 162' NORTON BAY Figure 6-1. Geographical location and bathymetry of the Norton Sound region. Ice growth continues southward until, by mid- December, the entire sound is more or less covered. This ice cover, which usually persists until April or May, consists primarily of loose pack 0.5-1.0 m thick except for shorefast ice near shore and for some distance offshore in the region of the Yukon River Delta. Direct observations of the ice cover are limited, and much information on its distribution and extent has been obtained through the use of satellite data (Muench and Ahlnas 1976, Ahlnas and Wendler 1979). During winter the ice cover would be expected to markedly reduce air-sea exchange of heat, moisture, and momentum. This would mitigate effects on the water of low air temperatures and winter storms. Oceanographic background Our existing knowledge of northern Bering Sea circulation before the present series of studies was summarized by Coachman et al. (1975). The regional circulation is dominated by a northward net water transport of about 1.5 X 10^ m^ /sec over the shelf between Norton Sound and Siberia. [Editorial note: this value may be high. See next chapter. D.H.] More recently Muench et al. (1978), on the basis of near-bottom recorded current measurements, have reported northward net flow southeast of St. Law- rence Island and in Bering Strait from October 1975 through April 1977. They also noted that the currents were characterized by north-south flow events having Circulation and hydrography of Norton Sound 79 speeds of 50-100 cm/sec, fast relative to mean flow speed of about 15 cm/sec. These flow events had time scales of several days, or the same duration as meteorological events. Spatial variability of the northward flow and its interaction with the waters of Norton Sound remain uncertain. Before this study, little oceanographic infor- mation was available from Norton Sound itself. Bottom sediment distributions within the sound suggest that mean circulation is cyclonic and pene- trates to the easternmost sound (Drake et al. 1980), but flow rates have not been estimated. Cyclonic flow in western Norton Sound was depicted quali- tatively in the compilation by Hughes et al. (1974) and mentioned as a probability by Coachman et al. (1975). The Yukon River enters the Bering Sea in the southwest comer of Norton Sound (Fig. 6-1). A minor portion of its discharge flows directly into the southern sound, while the remainder enters the Bering Sea along the coast farther west. Both the volume and the pathway of Yukon water which enters the sound remain uncertain. River flow can be as high as about 3 X lO'' m-'/sec, as gauged at Pilot Station, some 60 km above the river mouth. Most of the annual river influx occurs between about April and November, and flow is an order of magnitude greater in midsummer than in midwinter. FIELD PROGRAM Oceanographic data were obtained from Norton Sound during 1976-78. Temperature and salinity data were acquired from vessels in summer 1976, 1977, and 1978 and through the ice from a helicopter in winter 1978. The periods of these cruises are summarized in Table 6-1. Twenty-four-hour time- series current observations were obtained from anchor stations in summer 1976, and instantaneous current profiles were obtained from anchor stations in summer 1977. Ancillary data obtained in summer 1977 included vertical profiles of transmissivity. In addition to shipboard data, current data were obtained using taut-wire moorings during both winter and summer periods. Statistics for these moorings are summarized in Table 6-2. Summer temperature and salinity data were obtained using a Plessey conductivity/tempera- ture/depth (CTD) profiling unit with calibration samples on each cast. Winter data were obtained with a Plessey Model 9400 profiling unit operating through the ice from a helicopter. Temperature and salinity data are accurate to within 0.02 C and 0.02^/00, respectively. TABLE 6-1 Summary of cruises to Norton Sound Vessel Cruise ID Dates No. of Stations Discoverer RP4-D1-76B Leg V Sea Sounder None Surveyor RP4-SU-77B Leg IV Discoverer RP4-D1-78B Leg I Discoverer RP4-D1-78B Leg III NOAAUH-IH W-29 26 Sept - 9 Oct 1976 33 8-12 July 1977 26 11 Aug- 2 Sept 1977 10 July - 3 Aug 1978 10-29 Sept 1978 17 Feb - 5 March 1978 55 56 46 37 Currents were observed on the taut-wire moorings with Aanderaa Model RCM-4 current meters. Our discussion addresses the non-tidal, low-frequency flow components, as tides are treated in Chapter 8 of this volume. In order to remove tidal and higher- frequency components, the records were processed according to Charnell and Krancus (1976) and run through a 35-hour low-pass filter. Anchored time-series stations were obtained using an Aanderaa current meter modified to operate through a deck readout unit. Speed and direction were measured hourly at 5-m intervals through the water column. Hourly CTD casts were also made at the time-series stations. Instantaneous current profiles were obtained using a Hydro Products current meter. Observations were not taken when the vessel exhibited yawing motions, easily detectable by rapid variations in heading. In addition to the water column observations, wind speed and direction and air temperature were ob- tained from shipboard and were also recorded at the nearby weather station at Nome. OBSERVATIONS Distribution of density, temperature, and salinity The waters of Norton Sound are characterized during the summer by two layers separated by a strong pycnocline. The upper layer is warmer and of 80 Physical oceanography TABLE 6-2 Norton Sound taut-wire mooring summary Usable Bottom D Meter D Mooring record L Latitude Longitude Vector mean Mooring (m) (m) time period (days) CN) CW) speed dir NC-14 32 22 8/21/76-6/25/77 309 64° 21.6' 165° 21.6' 2.2 cm/sec 073°T NC-15 19 15 8/21/76-1/30/77 131 64° 06.5' 165° 17.7' 5.9 cm/sec 018°T NC-20 19 6 7/8/77-8/25/77 48 63° 59.7' 165° 29.4' 7.8 cm/sec 340° T NC-20 19 14 7/8/77-8/25/77 None 63° 59.7' 165° 29.4' malfunctioned NC-21 25 6 7/9/77-7/22/77 -13 64° 08.2' 163° 15.2' malfunctioned* NC-21 25 20 7/9/77-8/26/77 48 64° 08.2' 163° 15.2' 0.1 cm/sec 198° T NC-22 16 8 7/9/77-lost None 63° 41.0' 163° 00.1' not recovered LD-5 27 20 7/25/78-9/4/78 42 64° 08.3' 163° 00.2' 0.8 cm/sec 309° T *Record became increasingly error-ridden with time, so no mean is presented here, but early part of record was useful for comparison purposes. lower salinity, and hence is less dense than the lower layer. This layering was more pronounced and consistent in its characteristics in 1976 and 1977 (Figs. 6-2 and 6-3) than in 1978 (University of Washington data not shown). In 1976 and 1977 temperature and density differences between the layers were greater in the eastern than in the western sound. A near-surface lens of warm (>12 C), low- density (<13 a I units) water in the southwestern sound was especially pronounced in July 1977 (station 23), and indicated the presence there of Yukon River water which is characterized by low salinity and corresponding low density. Concurrent temperature increases and density decreases with time during July-August 1977 were evident in both layers. The data in summer 1978 were insufficient to define spatial variations, but were adequate to detect higher bottom-layer temperatures in the eastern sound (~7 C, as opposed to 1-2 C during 1976 and 1977) than during the previous two seasons. Bottom-layer salinities were also lower in the eastern sound in 1978 than during 1976 or 1977 (about 320/oo, as com- pared to 34°/oo). The tendency for increased stratification in the eastern as compared to the western sound is exempli- fied by vertically averaged density gradients for July 1977 (Fig. 6-4). The single isolated high value off the Yukon River was due to elevated near-surface tem- peratures. A pronounced region of minimum strati- fication was present in the central western sound. Closely spaced stations obtained through this region in August 1977 defined vertical density structure through the area of minimum stratification (Fig. 6-5). This feature coincided with a shallow rise (depths <20 m) in the bottom southeast of Nome (Fig. 6-1). Though there was considerable year-to-year var- iation in detail, several features of the horizontal temperature and salinity distributions persisted. The surface and near-bottom temperature and salinity dis- tributions adequately represent the upper and lower layers, respectively, and are used to depict these features (Figs. 6-6 and 6-7). The upper layer was dominated by an eastward-trending high-salinity, low- temperature tonguelike feature which originated in the northwestern portion of the sound. This feature was best defined in July 1977, and by August it extended east-southeastward the full length of the sound but was patchy in nature. In September 1976 the tongue was restricted to the region just off Nome, and in 1978 the data were inadequate to detect its presence or absence. Generally, lower-salinity, warmer water lay north and south of the tongue. The lowest salinities which were observed occurred off the Yukon River in July 1977. This may have been a 9-12 JULY 1977 C -10 - 20 26-29 AUGUST 1977 Figure 6-2. Vertical distributions of temperature along selected north-south sections across Norton Sound at three different times. Section locations are shown in Fig. 6-6. 9-12 JULY 1977 -10 20 26-29 AUGUST 1977 Figure 6-3. Vertical distributions of density (aj along selected north-south sections across Norton Sound at three different times. Section locations are shown in Fig. 6-6. 81 82 Physical oceanography 166° 162° MEAN VERTICAL DENSITY GRADIENTS (sigma-t units/m) Figure 6-4. Horizontal distribution of the mean vertical density gradient in sigma-t units/m for the upper 15 m of the water column during July 1977, obtained by vertically averaging, from the surface to 15 m, gradients computed for 1-m increments. Dotted line defining shoal area is schematic only. sampling artifact, as measurements were obtained closer to shore at that time than during other cruises and probably penetrated farther into the Yukon River plume. The maximum observed temperatures (>16 C) occurred in the northeastern sound during August 1977. The horizontal range of temperatures within the upper layer during September 1976 was only about 1 C, in sharp contrast to stronger gradients observed in the 1977 summer data. The lower layer in the eastern and northeastern sound was characterized during summer 1976 and 1977 by particularly cold, saline water. The coldest (<0 C), most saline (>34°/oo) water observed was present in July 1977, whereas by August tempera- tures had increased to 2-3 C, salinities had decreased to 32-33°/oo and the locus of maximum salinity had shifted somewhat westward. As for the surface layers, bottom-layer salinities were lowest off the Yukon River mouth. A band of relatively low-salinity (22-23°/oo) water paralleled the northern coast of the sound in September 1976 and August 1977 and appeared to be a westerly extension of warm, low-salinity water then occupying the upper layer in the northeastern sound. 10 20 10 g X 20^ 30 to - 20 28-29 AUGUST 1977 Figure 6-5. Vertical distributions of density (a J along selected sections normal to the coastline of northern Norton Sound at two different times. Section G-G' was occupied twice in rapid succession on the same day. Locations are shown in Fig. 6-7. SEPT 1976 SURFACE JULY 1977 SURFACE AUGUST 1977 SURFACE /^ SEPT 1976 BOTTOM JULY 1977 BOTTOM />r AUGUST 1977 BOTTOM /^ Figure 6-6. Horizontal distribution of upper- and lower-layer temperatures in Norton Sound at three different times. Location of sections in Figs. 6-2 and 6-3 are indicated. SEPT 1976 SURFACE JULY 1977 SURFACE .ns^ AUGUST 1977 SURFACE r^ r^ SEPT 1976 BOTTOM JULY 1977 BOTTOM AUGUST 1977 BOTTOM .^n^ /"^ Figure 6-7. Horizontal distribution of upper- and lower-layer salinity in Norton Sound at three different times. Locations of sections in Fig. 6-5 are indicated. 83 84 Physical oceanography Station coverage during July 1977 and in 1978 was inadequate to determine whether or not the feature was present. Horizontal salinity gradients were stronger in both layers during 1977 than in 1976. Other temperature-salinity features were evident only part of the time, but nevertheless can contribute to our understanding. In September 1976, temper- ature on the Ot = 21 surface showed two tongues (Fig. 6-8): a warm (8.5-9.0 C) easterly-directed tongue in the southern sound and a colder (7.0-8.5 C) tongue extending westward in the northern part, neither of which was evident in summer 1977. Although distributions on isopycnal surfaces in shallow water like this, particularly near the top of the pycnocline, must be interpreted with caution, the distribution shown here agrees generally with the concept of a cyclonic circulation as discussed below. In September 1976 the near-bottom salinity distribu- tion revealed two tongues of relatively high-salinity water (>31°/oo) penetrating the sound from the west, roughly coincident with the two troughs in bottom topography south of Nome (Fig. 6-7). These features were not evident in summer 1977. A west- ward baroclinic coastal flow was present off Nome during August 1976 (Fig. 6-5), and was also reflected in the time-series current observations at station 22 (Fig. 6-9). The same region during 1977 appeared to be the site of vertical mixing 10-15 km offshore rather than of such a baroclinic current. Winter observations were obtained in February 1978, and the temperature and salinity distributions 65 166° 164° 162° 63' TEMPERATURE (°C) Figure 6-8. Distribution of temperature (°C) on a^ = 21 in September 1976. at that time are summarized on vertical sections in Fig. 6-10. The entire water column was at the freezing point, or about —1.6 C. Density differences were due to salinity differences, and were small with variations less than one a^ unit. Since observations did not extend into the eastern portion of the sound because that region was ice free during the field work, conditions there remain uncertain, but they were pro- bably similar to those observed in the western sound. The western sound was characterized by weak hori- zontal and vertical density gradients. Density was about 0.2-0.3 lower in the eastern than in the western portion, and lower by about the same amount near the bottom than near the surface. The two-layered structure which was present during summer had disappeared, its place having been taken by more or less uniform and weak vertical density stratification. Convective cooling and ice formation had created vertically uniform temperatures at the freezing point, but some vertical salinity, hence density, stratifi- cation remained. CURRENT OBSERVATIONS Current measurements were obtained from taut- wire moorings, as time series from anchored vessels and as instantaneous vertical profiles from anchored vessels. Statistics of the taut-wire mooring current measurements are given in Table 6-2. Five time-series current profiles were obtained at anchor stations along a transect south of Nome during summer 1976 (Fig. 6-14). While these were of too short duration to use in estimating mean flow, four of the five (22, 24, 25, and 26) were long enough to allow averaging out of the tidal signal. Moreover, they provide the only vertical distributions of current obtained in the region south of Nome. They are summarized as vector-averaged currents at each depth and presented along with vector-averaged winds observed from the anchored vessel (Fig. 6-9). The highest speeds observed were about 50 cm/sec and occurred near the surface at Station 22. Speeds decreased with increasing depth at that location to about 10 cm/sec near the bottom, and the direction rotated from northwesterly near the surface to west-southwesterly near the bottom. Mean speeds at stations 24, 25, and 26 were highest— 12-18 cm/sec near the surface— and decreased to about 5 cm/sec near the bottom. Flow directions at these locations were in the northwest quadrant throughout the water column. Short-term fluctuations, having time scales of a few hours, were superposed upon the mean flow at each location (Fig. 6-11). The surface wind was considerably steadier with time than the currents. Circulation and hydrography of Norton Sound 85 STATION 22 24 (26 HOURS (25 HOURS: 30 SEPT. -I OCT.) 1-2 OCT) N3 25 (10 HOURS: (25 HOURS: 4 OCT.) 4-5 OCT) 26 (25 HOURS: 5-6 OCT) \ \ \ ^ ¥ E X I- Q_ UJ Q 7 22 27 N i WINDS CURRENTS 5 M/SEC 10 50 CM/SEC 100 MEAN \ WIND Figure 6-9. Vector-averaged currents and surface winds at the five time-series stations in western Norton Sound during September and October 1976. Mean current vectors are depth averaged. Station locations are shown in Fig. 6-14. Two 44-day and one 14-day record were obtained from taut-wire moorings in summer 1977 (Table 6-2 and Fig. 6-12). An additional 45-day record, LD-5, was obtained during summer 1978 from nearly the same location as NC-21 off Cape Darby. Near- bottom records from this location during both summers indicated flows of order 1 cm/sec or less. The summer 1977 near-surface record at the same location was cut short due to instrument malfunc- tion, but indicated northwesterly flows of 10-15 86 Physical oceanography FEB 24-25. 1978 FEB 26- MAR I, 1978 63*40 50 I I SIGMA- 1 168° 167° 166° 165° 164° 163° 162 °W I ""^^^^^ii^ 1 ... NOME I c%.. .# 30' — i«f'--<^^ii:> .ii^ ^ \/f :/ " 30' 64* N - ,r ii • • 13 14 2« 3« 4» H • ,•5 16 '^ 6 7» •|8 19 C/CAPE - DARBY STUARL 64° N 30' - ^^^-1^ ^%^ €aL 30' 63° 1 1 1 .*^L '^I'P "r VTT' ^ '"1 cx« 168* 167° 166° 165* 164° 163° I62°W Figure 6-10. Vertical distribution of salinity and density (at) along two sections during winter 1978. cm/sec. The summer 1977 near-surface record from NC-20 in the central western sound indicated north-northwesterly flow at 10-15 cm/sec. Overwinter records were obtained from moorings NC-14 and NC-15 in the western sound. These were both near bottom and bracketed the period from late summer to mid or late winter 1977. They indicate a mean current which was northerly in the central sound and northwesterly along the northern shore, with speeds of about 5 cm/sec. All of the moored current records revealed non- tidal speed fluctuations which were large relative to mean flow speeds and had time scales of several days (Fig. 6-12), including occasional flow reversals. DISCUSSION Winter hydrographic observations made through the ice have established that Norton Sound becomes vertically well mixed during winter. This is common in high-latitude regions, as a consequence of vertical thermohaline convection due to surface cooling and ice formation. Observations during three summers reveal that well-mixed structure gives way to a regime which is strongly two-layered in temperature and salinity and hence also in density. Such a layered structure is generally typical of shallow oceanic regimes subject to tidal and wind mixing and buoyan- cy input. The upper layer is a consequence of the combined effects of wind mixing, freshwater influx, and solar warming, while the lower layer acquires vertical homogeneity through turbulence generated by currents at the bottom. The persistence of this structure in Norton Sound for several months each summer suggests that the time rate of change is small and that a small horizontal advection is balanced by other processes. The strong pycnocline between upper and lower layers inhibits vertical exchange of heat and salt. Our discussion examines these premises. Circulation and hydrography of Norton Sound 87 WIND \\^ w CO UJ en en T. (r \- 3 a- o iJJ Q I 7 TIME DAY (2) TIDES AT NOME KEY V ^^ v^ v^ //tl\. i\ >v>^^N -T^-^ J> ^ ^ 1^ ^^ — TT^ \ \ \\ d \ \ V- tT ^^7^ . ^\\ ^/ / K>^ o o ro O O in o o O O O O o o m CO I I I I I I I I I I 4 OCT NT o o o o o o o o o o _ ro in (s. (7) O O O O O I I I I I I I I I I I I 1 1 1 o o o o ro 5 OCT L WIND — I 10 0 M/SEC H L CURRENT 0 50 100 CM/SEC H N i Figure 6-11. Hourly surface winds and currents at time-series station 25 in western Norton Sound during October 1976, illustrating short-term fluctuations. Circulation Horizontal property distributions tend to reflect circulation processes. During winter there was little horizontal variability; in summer the lowest tem- peratures and highest salinities in the upper layer occurred in the central western part of the sound. Elsewhere in the upper layer, temperature and salinity distributions, particularly in September- October 1976 and August 1977, indicate water mass continuity between the easternmost sound and a near-coastal band farther west along the northern coast. This structure suggests that a baroclinic coastal flow was transporting water from the eastern sound westward along the northern coast. The existence of this flow was substantiated both by subsurface density observations (Fig. 6-5) and current observa- tions from station 22 (Fig. 6-9). The barochnic feature was not present off Nome in August 1977, and hydrographic data obtained during July 1977 were inadequate to detect its presence or absence. Upper-layer current observations farther east off Cape Darby during July 1977 indicate that a westerly mean flow paralleled the coastline but was superposed upon a highly variable flow which included reversals. We conclude that westerly flow along the northern coast is a usual feature but may vary in intensity and extent. It was more pronounced and uniform during October 1976 than during August 1977. Lower-layer temperature and salinity distributions differed from those in the upper layer, which suggests that advective processes were different in the two layers (Fig. 6-13). Temperature was lower and salinity higher in the eastern lower layer than in its 88 Physical oceanography NC-20 Figure 6-12. Low-pass filtered currents at 5 m depth at station NC-20 in Norton Sound, summer 1977; u and v are east and nortli components, respectively. Mooring loca- tions are indicated in Fig. 6-14. western part. This was most apparent in July 1977, less so in August 1977, and least obvious in Septem- ber-October 1976. Deep temperatures in the eastern sound increased during summer, while salinity de- creased; during July -August 1977 the record from a current meter (NC-21) moored in the eastern lower layer indicated that these changes were nearly linear with time. The cold (<0 C), saline (>340/oo) deep water present in the eastern sound during summer 1976 and 1977 cannot have originated on the shelf to the west because water there was too warm and of too low salinity except in September 1976. This deep water can only have been a remnant of the preceding winter's convective layer, with elevated salinities resulting from salt exclusion as ice was formed. Observations during summer 1978 were sufficient to establish that this was not the case at that time as temperatures and salinities were similar in the eastern and western lower layers. Sluggish circulation in the lower layer of the eastern sound was substantiated by observations. An advection rate along the lower boundary of the upper-lower layer interface can be estimated for September 1976 using temperature and salinity observations if we assume that the tonguelike low- temperature protrusion from the eastern sound (Fig. 6-8) reflected a steady -state balance between hori- zontal advection and lateral diffusion. A horizontal length scale of 100 km, applied to the empirical findings of Okubo and Ozmidov (1970), yields a lateral eddy conductivity of order 10^ cm^ /sec. 16 14 o ^ 10 LU % 8 4 2 0 -2 28-30 SEPT, 1976 / 7 . /_ * t 1 (i^ / / ^ ^^ / WEST y NEAR- f-^^BOTTOM / - ^\/ / - EAST ^ / NEAR- ' BOTTOM / - 1 1 1 1 1 8-12 JULV , 1977 ^ 1 / 7 -\ 6^ / i ' 1 \ i\i^ L/ / / - m ■ — >\yv - - WEST// NEAR- BOTTOM 1 1 m 1 1 EAST NEAR- / BOTTOM- 1 26-29 AUG, 1977 1 1 1/ 1 / / C\ \ ^ ^7' \> / \\ \ \ />!' '\j ^ \\ \\ "** /' ' 'I % - \ \\ \ ^ ! 1) ^^ \\\ \ ^ \\\ \ '^ 1 (^ Ml!/ 1 / \ f / nI / EAST rV k / NEAR- (S^ P \ / BOTTOM - WEST ^ NEAR - ■^ / j " - BOTTOM 1 1 1 I 1 1 4 26 28 30 32 34 3^ SALINITY (7oo) EAST NORTON SOUND WEST NORTON SOUND Figure 6-13. Temperature-salinity curves from selected stations in eastern and western Norton Sound during the summers of 1976 and 1977. Temperature and salinities were in fact continuous from east to west across the sound, but stations from the central sound have been omitted in the interest of clarity. Circulation and hydrography of Norton Sound 89 Application of the graphical method described by Proudman (1953) to the geometry of the tongue leads then to estimates of a westward advection rate of less than 1 cm/sec along its axis. During July- August 1977 the observed lower-layer current speed at NC-21 off Cape Darby was 0.13 cm/sec, a compar- able value. During summer 1978 a near-bottom mooring at the same location (LD-5) showed a net flow of less than 1 cm/sec to the northwest— again in agreement. The cause of this sluggish circulation in the eastern lower layer is uncertain, but it is probably the extreme layering which could decouple the lower layer from upper layer motions. It does not appear that basin configuration plays a major role, although the promontory extending from the southern shore formed by Stuart Island may be sufficient to deflect to the north whatever easterly flow exists along that shore, preventing it from entering the eastern sound. There exists, however, no sill between these prom- ontories which might prevent interchange of deeper waters, and there are no observations to support this hypothesis. The case for decoupling of motions at the horizontal interface between upper and lower layers will be examined in more detail below. While upper- and lower-layer circulation appeared decoupled in the eastern sound, time-series observa- tions obtained from the vessel during September and October 1976 did not reveal such decoupling in the western sound. Rather, there was a monotonic decrease in current speed, along with a slight rotation of flow direction, with increasing depth (Fig. 6-9). The four stations with long enough records to allow at least a crude attempt at averaging out the tidal signal (stations 22, 24, 25, and 26) all indicated northwesterly surface flow which became westerly near the bottom. This was in rough agreement with the moored current record from July-August 1977 (NC-20) which indicated a northwesterly near-surface flow in the same region, with an overwinter record at NC-15 which showed northerly flow, and with a near-bottom record off Nome which indicated a weaker westerly flow during September 1976-March 1977. These current observations suggest that western sound circulation was dominated by flow which penetrated only slightly eastward into the sound and was driven primarily by the northward flow over the shelf between Norton Sound and St. Lawrence Island (Fig. 6-14). Low-frequency (of order two days and longer) current fluctuations observed at all of the moorings NC-14, NC-15, NC-20, NC-21, and LD-5 were a major characteristic of the flow; they suggest that the cyclonic circulation driven by northward regional flow may penetrate eastward into the sound to varying extents and may at times Figure 6-14. Schematic representation of circulation and mixing in Norton Sound based on the summer 1976 and 1977 hydrographic and current data. Locations of time- series current and CTD stations occupied in September and October 1976 and taut-wire current moorings are shown. form a cyclonic loop within the western sound. The occurrence of northwesterly flow fluctuations at NC- 21 off Cape Darby suggests that at least the upper layer of the eastern sound responds to fluctuations in the western portion, although a coherency analysis of near-surface current records from stations NC-20 and NC-21 did not reveal a significant coherency between currents at the two locations. Penetration of flow eastward to Stuart Island would require, by con- tinuity, increased westward flow off Cape Darby. That remnant water was not present in the lower layer in summer 1978 suggests also that advection had occurred in the eastern lower layer prior to our observations. Therefore, it appears that at times eastward flow may penetrate into the eastern sound. The role of northward flow on the Bering Sea shelf in inducing a circulation within Norton Sound is un- certain. A probable mechanism invokes the high-lati- tude tendency for conservation of potential vorticity to constrain streamlines to parallel isobaths. The northward regional flow would tend to follow bot- tom contours into the sound and contribute to a cyclonic circulation there. The potential vorticity argument requires, however, that friction be neg- lected, an assumption of debatable validity in view of the shallow depths in the sound. Regardless of cause, easterly flow entering the sound along its southern 90 Physical oceanography shore, curving cyclonically to the north, would then be deflected to the west at the northern coast to satisfy volume continuity constraints. Subtidal flow pulses observed in the time-series current records are a dominant feature of regional flow. Their cause is uncertain, but the shallow depths and broad horizontal extent of the regional shelf suggest that winds and atmospheric pressure events may play major roles. Coachman et al. (1975) have suggested that flow through Bering Strait is character- ized by pulsations that depend upon north-south atmospheric pressure differences across the strait. A detailed analysis of this problem is beyond the scope of this chapter, and is treated elsewhere in this volume (Chapter 7). Maintenance of layering The extremely sluggish circulation and consequent isolation of the lower layer in the eastern sound in the summers of 1976 and 1977 are of particular interest because of their basic scientific implications and because similar conditions may also exist in other high-latitude bodies of water subject to similar con- ditions--for example, Kotzebue Sound (Kinder et al. 1977). Retention of temperature and salinity chairac- teristics by this water for four to five months in a total water depth of about 20 m, despite wind and tidal mixing and insolation, suggests: (1) horizontal advection of bottom water into the eastern sound was negligible; (2) insolation, particularly significant during early summer because of long daylight hours, had been unable to penetrate the bottom layer sufficiently to cause appreciable warming; and (3) vertical mixing through the pycnocline was extremely slight through the summer. Of these possibilities, the first has been shown by temperature, salinity, and current observations to be the case. The second may be examined summarily in the light of observations of transmissivity made by using a beam transmissometer in the eastern sound during summer 1977. In the northeastern sound off Norton Bay, transmissivity was of order 10 percent. Transmissivities were some- what higher farther south, but were still too low to allow appreciable penetration of solar radiation below the upper layer. Transmissivities in the lower layer were uniformly of order 10 percent throughout the eastern sound. The net result was to effectively prevent solar radiation from penetrating to the lower layer. The extreme layering in the eastern sound can be examined in terms of maintenance of the horizontal interface between layers. This interface is character- ized by high vertical gradients; as high as 4 C/m in temperature and l.S^/oo/m in salinity, with a cor- responding Vaisala frequency of about 0.1 /sec. Using the upper- and lower-layer current observations obtained from mooring NC-21 off Cape Darby, in conjunction with estimated mean vertical density distributions obtained at the beginning and end of the mooring period, a bulk interfacial Froude Number of about 10"' can be estimated. This value suggests that little exchange due to turbulent mixing occurs across the interface. We may estimate a vertical mixing coefficient through the interface using a simple conservation of heat argument. Assume (1) that bottom water temperature at the end of each winter was near freezing (~1.7 C), as borne out by the winter 1978 observations, (2) that upper-layer temperature had reached 8 C by early August follow- ing a linear increase with time from early May when the ice first melted, and (3) that advective sources and sinks of heat can be neglected so that heat entering the lower layer all passes through the inter- face. The computed vertical mixing coefficient through the pycnocline was of order 10"^ cm^ /sec for both summer 1976 and 1977. These values are small compared with those computed for other oceanic regions, but are reasonable for the extremely strong stratification observed. Buoyancy input into the eastern sound appears to be critical to maintenance of the two-layered struc- ture there. Comparison of our data with those presented by Coachman et al. (1975) from the Bering Strait suggests that the vertical density structure in the western sound is continuous with that on the Bering Sea shelf to the west. The enhanced summer 1976 and 1977 layering in the eastern sound can be accounted for by additional dilution of the upper layer by local freshwater influx and solar insolation. Assuming an initial salinity of 30*^/oo (see Fig. 6-10), about 10* m^ of fresh water are required for the observed dilution. This can be roughly accounted for by using a mean annual runoff of 30 cm extrapolated from Nome over a watershed area of about 2.5 X 10^ km^ . The Yukon River is a second source of fresh- water influx and during midsummer runoff of order 10^ m^ /sec could supply water sufficient to effect the observed dilution in about 10 days if the water were all to flow into the eastern sound. This amount of freshwater influx is equivalent to a buoyancy input of about 10"^ W m"^ . Assuming that solar insolation of 0.2 ly/min yields a buoyancy input to the upper layer of the same order, both insolation and freshwater influx are important to the buoyancy input. This compares with a bottom tidal dissipation of about 200 X 10^ W m"^ computed using the same method as Schumacher et al. (1979). In other shallow shelf regions, 1-2 percent of the tidal dissipa- Circulation and hydrography of Norton Sound 91 tion is sufficient to vertically mix the water column. Our estimates suggest that buoyancy input to eastern Norton Sound is sufficiently high that tidal dissipa- tion is inadequate to vertically mix the water column. We conclude that the persistence of this extreme layering in the eastern sound is due to the roles of freshwater influx and solar insolation in generating sufficient buoyancy to overcome tidal mixing. Low net advection in the lower layer is then due to de- coupling from the upper layer across the sharp interface maintained by buoyant forces. Hydrocarbon observations September 1976 observations of dissolved hydro- carbons provide additional evidence in support of our proposed general horizontal circulation scheme for Norton Sound. Although the observations included other hydrocarbons, we consider here distributions only of methane, ethane, and propane. Most meth- ane found in shallow shelf waters arises from bio- chemical reactions occurring in anoxic near-surface sediments. A prominent local source of methane may be used as a short-term water mass tracer, since it appears to be quasi-conservative over time and space scales such as we are dealing with in Norton Sound. The near-bottom methane distribution in Norton Sound suggests a dominant source (2,242 nl/1) at the eastern end of the basin, with a sluggish, ill-defined northwesterly drift (Fig. 6-15). Methane-deficient water (250 nl/1) was apparently entering from the south near the Yukon River Delta with relatively little transport eastward along the southern coast. The high methane concentration in the southeast corner of the sound suggests either an increased source at that location or an extremely sluggish circulation, or a combination of these. Coincident with the methane observations, a near- bottom plume of ethane- and propane-rich water was detected about 40 km south of Nome near station 24 (Cline and Holmes 1977). The hydrocarbons ap- peared to originate from a point source and drift northward toward Nome, then northwest along the coast, in agreement with the circulation scheme deduced above. The Yukon River plume The fate of Yukon River water remains problemat- ic, though it seems likely that it is dominated by flow events rather than mean flow. The temperature- salinity and hydrocarbon observations obtained dur- ing the summers of 1976 and 1977 yielded no evi- dence that Yukon River water was advected eastward into the eastern sound. Rather, it appeared that Stuart Island might have contributed to blocking the easterly flow along the southern coast (Figs. 6-5 and Figure 6-15. Distribution of dissolved methane 5 m above the bottom, September 1976, in nl/1. 6-6). It is of course feasible that sporadic flow of Yukon water had occurred earlier during either or both seasons, and that the observed upper-layer dilution was due in part to remnants of past Yukon water inflows. This may have been true before July 1977, when particularly low (<19°/oo) near-surface salini- ties were observed in the eastern sound (Fig. 6-7). The limited observations which indicated that the eastern basin had flushed out in summer 1978, resulting in deep-layer temperatures and salinities similar to those in the outer part of the sound, strengthen the suppo- sition that appreciable eastward flow events may occur. The summer 1978 data were not however adequate to determine whether or not Yukon-diluted waters were in fact continuous between the eastern and western sound. A Yukon River origin for bottom sediments in the eastern sound suggests that Yukon water can enter there (Drake et al. 1980). The pulse-like events observed in the current records suggest, moreover, that water motion in the sound, particularly in the form of easterly migrations in the cyclonic flow, may be adequate to cause such addi- tions. We have, however, no direct supporting evidence of such flow. R. Feely (PMEL unpublished data) has observed in data from summer 1979 a narrow coastal band of low-salinity surface water along the southern coast, which may be eastv/ard- flowing Yukon water. However, no flow rates are available. SUMMARY Observations obtained from Norton Sound indicate 92 Physical oceanography that in summer the regime was strongly two-layered in both temperature and salinity. The eastern and western sound were effectively separated into two distinct flow regimes at a line roughly coincident with the constriction formed by Cape Darby on the north and Stuart Island on the south. The eastern sound was more strongly layered than the western, and upper- and lower-layer flows were decoupled. The surface layer exhibited a weak tendency toward cyclonic flow, which was reflected both in the hydrographic observations and in moored current records. The eastern lower layer exhibited a near- zero mean flow allowing remnant cold, saline water from the previous winters' convective regimes to retain its identity in both 1976 and 1977 despite the shallow depth of the eastern basin. Sluggish lower-layer circulation was borne out by near-bottom moored measurements south of Cape Darby. Absence of the remnant water in summer 1978 suggested either that a flow event had flushed out the eastern basin or that less cold, saline water was produced in 1977-78. Short-term time-series measurements in the western sound suggest that the upper and lower layers were not decoupled as in the eastern sound. These observations, in conjunction with moored current records and dissolved hydrocarbon distributions, support the concept of a northerly flow in the westernmost sound. The extent of this flow eastward into the sound may vary; it was apparently more intense in 1976, as evidenced by a strong westerly coastal current off Nome, than during 1977, when such a coastal feature was not observed. Large fluctu- ations relative to the mean flows, with time scales of several days, were observed at all moorings and suggest that instantaneous flow patterns in the sound may vary considerably. The central western sound was a locus for vertical mixing as currents, primarily tidal, impinged upon a relatively shallow area. This was the only area where a breakdown in vertical layering was observed. We hy- pothesize that the relatively deep channels bracketing the shoal allowed occasional eastward flow of deeper, saline water that was then mixed upward leading to the observed higher local surface salinities. At no time did we see flow of the Yukon River plume into eastern Norton Sound. It remains uncer- tain where the Yukon water goes: it appears likely that pulse-like flow events such as those observed at the current mooring sites may intermittently trans- port Yukon water into the eastern sound, or con- versely, westward completely out of the sound. It may be entrained into the general northward flow west of the sound. Disposition of the Yukon River plume must be considered an appreciable regional problem. The near-bottom water in the eastern sound presents an interesting case. It is highly unusual for water at these depths (about 15 m) to retain its temperature-salinity identity for several months with- out undergoing vertical mixing due to winds or tides or sufficient lateral advection or diffusion to alter its heat or salt content appreciably. Maintenance of this regime is due to a large buoyancy input, in the form of insolation and fresh water, to the upper layer. ACKNOWLEDGMENTS This study was supported in part by the Bureau of Land Management through interagency agreement with the National Oceanic and Atmospheric Ad- ministration, under which a multiyear program responding to needs of petroleum development of the Alaskan continental shelf is managed by the Outer Continental Shelf Environmental Assessment Program (OCSEAP) Office. We are indebted to Dr. L. K. Coachman for suggestions in preparation of the manuscript, and to the anonymous reviewers who provided suggestions for improving it. The work could not have been carried out without the cooper- ation of officers and crew of R/Vs Discoverer, Sur- veyor, and Sea Sounder. REFERENCES Ahlnas, K., and G. Wendler 1979 Sea-ice observations by satellite in the Bering, Chukchi, and Beaufort seas. Proc. Conf. on Port and Ocean Engi- neering Under Arctic Conditions, Trondheim. In press. Charnell, R. L., and G. A. Krancus 1976 A processing system for Aanderaa current meter data. NOAA Tech. Rep. ERL-PMEL-6. Cline, J. D., and M. L. Holmes 1977 Submarine seepage of natural gas in Norton Sound, Alaska. Science 198: 1149-53. Circulation and hydrography of Norton Sound 93 Coachman, L. K., K. Aagaard, and R. B. Tripp 1975 Bering Strait: The regional physical oceanography. Univ. of Washington Press, Seattle, Wash. Drake, D. E., D. A. Cacchione, R. D. Muench, and C. H. Nelson 1980 Sediment transport in Norton Sound, Alaska. Mar. Geol. In press. Hughes, F. W., L. K. Coachman, and K. Aagaard 1974 Circulation, transport and water ex- change in the western Bering Sea. In: Oceanography of the Bering Sea, D. W. Hood and E. J. Kelley, eds., Inst. Mar. Sci., Univ. of Alaska. Occ. Pub. #2. Kinder, T. H., J. D. Schumacher, R. B. Tripp, and D. J. Pashinski 1977 The physical oceanography of Kotze- bue Sound, Alaska, during late sum- mer, 1976. Univ. of Washington, Dep. of Oceanogr. Tech. Rep. M77-99. Muench, R. D., and K. Ahlnas 1976 Ice movement and distribution in the Bering Sea from March to June 1974. J. Geophys. Res. 81: 4467-76. Muench, R. D., C. A. Pearson, and R. B. Tripp 1978 Winter currents in the northern Bering Sea and Bering Strait. EOS Trans. Amer. Geophys. Union 59: 304. Okubo, A., and R. V. Ozmidov 1970 Empirical dependence of the coef- ficient of horizontal turbulent diffu- sion in the ocean on the scale of the phenomena in question. Is. AN/SSSR Fizika Atmos. i Okeana. 6: 534-6. Proud man, J. 1953 Dynamical oceanography. Press, N.Y. Dover Schumacher, J. D., T. H. Kinder, D. J. Pashinski, and R. L. Charnell 1979 A structural front over the continental shelf of the eastern Bering Sea. J. Phys. Oceanogr. 9: 79-87. Reevaluation of Water Transports in the Vicinity of Bering Strait L. K. Coachman and K. Aagaard Department of Oceanography University of Washington Seattle, Washington INTRODUCTION The general northward flow through Bering Strait transports water and associated properties from the Pacific Ocean into the Arctic Ocean. Although the strait's cross section is small (85 km X 50 m) and the average transport is not large by oceanic standards (~1 Sv), the consequences of the flow reach well to the south into the Bering Sea and to the north into the Arctic Ocean. To the south, the effect of the Bering Strait flow is that of a continuous "leak" drawing off water from the Bering Sea across the large eastern continental shelf, where it significantly influences the flow field. To the north, Bering Strait water has been specifically traced to the North Pole (Coachman and Barnes 1961), and a recent analysis (Coachman 1978) suggests that the trans-ocean transport of the Arctic Ocean subsurface layer, to which the Bering Strait flow is a major contributor, may be much greater than previously supposed. Therefore Bering Strait water may have influence on the whole of the Arctic Ocean. Even though the existence of northward flow has been known for over 300 years, and oceanographic studies have been made in the region using reliable methods since the 1920's (Sverdrup 1929), systemat- ic studies of the flow employing direct current measurements on sections traversing the system between the Alaskan and Siberian boundaries began only in 1964 (Coachman and Aagaard 1966). Such measurements are necessary for determination of transport because of the spatial variability in the flow field. Results of these systematic studies through 1973 were summarized in Coachman et al. (1975). The available data came from 11 anchored current station sections across Bering Strait and another 10 sections north and south of the strait and extending entirely across the system between Alaska and Siberia. The latter can be used in assessment of transport because runoff into the intermediate region is at least an order of magnitude less than the oceanic transport. Although a number of years are repre- sented in the 21 sections, 19 were made during July and August, one in late September, and one in early October; hence, the results represent only summer conditions. The main findings were: (1) The north transport averaged close to 1.5 Sv.' (2) There were temporal variations in transport equal to the mean flow on a time scale of days. The data included variations as large as 0.9 Sv in one day and 2.0 Sv in three days. There was one case of southerly transport (—0.2 Sv). (3) The force driving the transport was a sea- level slope generally down to the north, which remained unexplained. Variability in transport was due to a combination of wind and regional atmos- pheric pressure distribution. Local wind could modify transport by about 0.5 Sv for each dyne/cm^ of sectional mean wind stress but could not account for the observed variations. The overall variability could be reasonably forecast by a linear correlation with surface atmospheric pressure at Nome with a phase lag of one day. (4) There was no evidence of an annual variation in transport. Among the many scattered current measurements available from the region, only two sets were made at times other than summer from which possible seasonal variations could be inter- preted : one set from the Strait of Anadyr in February and another from the eastern channel of Bering Strait in April. These data provided no support for the seasonal variation in transport suggested by Maximov ' Throughout we use the convention that positive values are northerly and negative southerly. 95 96 Physical oceanography (1945), Fedorova and Yankina (1964), and Antonov (1968). One earlier report on transport through Bering Strait not discussed by Coachman et al. (1975) was that of Bloom (1964). He presented data on flow through the eastern channel for various times during the 1950's based mainly on electromagnetic meas- urements. Coachman and Aagaard (1966), in a critical review, concluded that the results could not be interpreted as transports because of uncertainties in the techniques and calibrations. RECENT OBSERVATIONS Since 1973 we have accumulated sufficient new data to allow a reassessment of the transport in the vicinity of Bering Strait, particularly of flow during seasons other than summer and of its variability. The new data include: (1) Current records from meters moored at 11 locations from St. Lawrence Island to Cape Lisburne from September 1976 to June 1977 (Fig. 7-1), with the exception of mooring NC16, which was not deployed until the end of September. Each mooring, with subsurface flotation located well beneath the influence of ice, contained one Aanderaa current meter positioned 10 m above the bottom. The meters were set to record speed and direction at 40-minute intervals, and therefore nominally could record for one year; however, all the records ended for one reason or another between April and July. Concurrent data from all meters (except NC16) were obtained for the seven months September to March inclusive. (2) Water level records obtained for the same period by Aanderaa pressure gauges mounted on the anchors of moorings NC7, NCIO, and NC17, which recorded sea level at one-hour invervals. (3) Sections of current meter stations between Alaska and Siberia from R/V Moana Wave during September 1976 (Fig. 7-1). These sections were the same as those reported in Coachman et al. (1975), and the measurements were taken with a deck read- out Aanderaa meter with the vessel at anchor. (4) Five moorings with current meters in the area from outer Norton Sound to eastern St. Lawrence Island (in Fig. 7-1, the area from NC14 to NC16-17) for one and one-half months during summer 1978. The main goals of these studies have been to clarify the temporal variations (periods of one day and longer) in regional transport, which have been shown to be very large on a short time scale, to secure the relationships between the variations and likely causal mechanisms (atmospheric phenomena), and to 62° 160° Current meter moorings Location of anchored current stations Weather stations I I L Figure 7-1. Bering Strait and environs, showing location of long-term current meter moorings, sections of anchored current measurement stations, and weather stations. assess the consequences of the variations for the ocean conditions of the northern Bering and southern Chukchi seas. Thus, among the new data the seven- month concurrent current records receive primary attention. A reevaluation of the transport in the vicinity of Bering Strait using the new data requires an orderly series of analyses, to wit: (1) Calibration of the single record from Bering Strait in terms of transport. This can be accom- plished in two ways: by computing transports from the measurements of the six meters located along the Cape Lisburne section, which was a nearly closed section from Alaska to Siberia, and correlating the flow measured at NCIO with these; and correlating the transports from the 11 previous cross-sections with the flow at the position of meter NCIO in the section. (2) Next, cahbration of the flow measured in the St. Lawrence Island section (meters NC16, NC17) as transport. This is a more tenuous process, because although the flow passes through channels on both sides of St. Lawrence, the meters were placed only in the east channel. However, transport estimates are possible using thie two available measured sections transecting both channels (in 1968 and 1976). (3) Next, testing of the transport variations as functions of various atmospheric parameters, in- cluding wind, pressure, and pressure gradients, which are available from daily surface pressure charts and weather stations located around the region. (4) Finally, analysis of the response of the southern Chukchi and northern Bering seas to the ob- served variations in transport. The present chapter, which should be considered part one of these analyses, concentrates on the Cape Lisburne section and the calibration of the flow measured in Bering Strait as transport, and concludes with a qualitative description of the relationship of transport to atmospheric pressure. CAPE LISBURNE SECTION The particulars of the Cape Lisburne moored section are given in Table 7-1, and the mooring locations are shown in Fig. 7-1. At each mooring the current meter, an Aanderaa RCM-4 cycling at 40- minute intervals, was suspended 10 m above the bottom. Except for mooring NC5, which was not recovered, complete monthly records were obtained from all the sites from September through March, and the present analysis concentrates on these seven months during which records from the entire section are available. Table 7-2 gives the monthly and seven-month mean Water transport in the vicinity of Bering Strait 97 velocities at the six instrumented sites. Four fea- tures seem particularly important: (1) The long-term mean speeds are quite low, typically near 5 cm/sec, whereas we had a priori ex- pected them to be 15-25 cm/sec even in the central Chukchi Sea (Coachman et al. 1975). These low mean speeds reflect not so much a quiescent regime as a regime highly Vciriable in flow direction, thus resulting in a small vectorial mean. For example, every current meter recorded 40-minute mecin speeds in excess of 40 cm/sec; at NC7 the maximum 40- minute mean speed was 68 cm/sec. However, as we shall see in detail in subsequent representations of the flow field, the directional variability is large. This is apparent even in the monthly means (Table 7-2), where from September to March every mooring except NC2 shows resultant monthly vectors in three separate quadrants. (2) The monthly mean flow was most frequently toward the northwest or north at NCl-4, but during the seven-month comparison period had a southerly component nearly one-half the time at NC6 and 7. In fact, the strongest mean flows at the latter two moorings were southerly. (3) There is some indication of a seasonal trend, with southerly flow most common in fall and winter. Thus, only two of 22 instrument-months showed a mean southerly flow component during March- September, while 10 of 30 did so during October- February. (4) Table 7-2 also suggests some degree of hori- TABLE 7-1. Cape Lisburne section, moorings NC1-NC7 Mooring Latitude Longitude Depth of current meter Series start Series end NCI NC2 NC3 NC4 NC6 NC7 68°15.4'N 68° 29.7' 68° 44.2' 69° 00.7' 68° 57.2' 68° 55.2' 172°40.6'W 171° 55.3' 171° 06.2' 169° 59.2' 168° 18.6' 167° 21.3' 39 m 41 45 43 41 36 2200 GMT 0520 GMT 25 August 1976 17 June 1977 0000 1520 26 August 1976 4 April 1977 0220 0020 26 August 1976 7 August 1977 0600 1600 25 August 1976 6 May 1977 0220 1140 25 August 1976 20 April 1977 2140 2340 24 August 1976 27 June 1977 98 Physical oceanography TABLE 7-2. Mean velocities 7 -mo. Mooring Sept Oct Nov Dec Jan Feb Mar mean April May June July NCI cm /sec 3.5 1.2 2.4 6.8 6.4 3.8 3.1 2.6 2.8 4.4 @°T 149° 057° 309° 327° 324° 353° 346° 335° 329° 339° NC2 cm/sec 4.0 2.0 1.3 9.4 8.2 5.4 3.7 4.1 — — — — @°T 292° 143° 280° 325° 316° 344° 326° 320° NC3 cm /sec 6.7 2.5 0.8 6.1 9.7 5.6 5.7 4.5 3.8 5.0 12.4 8.1 @°T 302° 175° 247° 333° 320° 322° 330° 317° 325° 340° 321° 307° NC4 cm/sec 7.4 1.1 2.0 2.8 6.8 2.4 4.8 3.2 4.4 — — — @°T 334° 043° 132° 021° 344° 341° 008° 354° 335° NC6 cm/sec 8.0 3.4 1.7 12.9 5.5 5.1 3.0 1.1 — — — — @°T 359° 022° 061° 186° 230° 207° 012° 217° NC7 cm /sec 6.7 6.4 5.0 17.8 3.5 4.9 2.1 3.1 5.3 4.3 5.7 - @°T 043° 076° 085° 182° 203° 223° 093° 145° 029° 035° 021° (27 days) zontal coherence between adjacent instruments so that we might a priori have some expectation of meaningful transport calculations. Figure 7-2 shows the daily mean vectors for NCI and NC7 for September through March. The values have been smoothed by a 25-hour running mean. Mooring NCI has been taken as representative of the western and central parts of the Chukchi (NC1-NC4) and NC7 as representative of the easternmost Chuk- chi (NC6-NC7), where the strongest flows have previously been encountered (cf. Coachman et al. 1975, pp. 141-2). At NCI the flow during the first ^wM^^^^^\vW^\//^ik\VV,xN\^^^^^^ Figure 7-2. Daily mean cunrent vectors at NCI and NC7 for the period September through March, 1976-77, smoothed with a 25-hour running mean. The first day of each month is indicated by the appropriate capital letter. Water transport in the vicinity of Bering Strait 99 three months of deployment was highly variable in direction, generally alternating between periods with northerly and southerly flow components. Peak speeds were typically 15-20 cm/sec and the vari- ability time scale in the range of 3-10 days. In early December the speeds decreased to a slower level, typically 10 cm/sec or less, but much more nearly uniform in direction, generally NNW. The record from NC7 is quite different. The flow there was considerably faster, commonly exceeding 20 cm/sec and not infrequently twice that or more. The first one-fourth of the record shows relatively few southerly flow events, but from late October on such events dominate much of the record. The time scale of these events is rather uneven. On the one hand, from early November to early December there was a remarkably regular pattern to the flow rever- sals, with a northerly set for two to three days followed by a southerly one for a like time. On the other hand, there were times of prolonged southerly flow, the most conspicuous beginning in mid- December and persisting for 23 days. Other flow reversal sequences are intermediate to these extremes. During the last month of the record (March) the speeds were decidedly lower, but the directional reversals continued. Figures 7-3 to 7-9 show the temporal development of the current cross section month by month. The vertical axis is the time in days, and the six moorings are placed in their relative positions along the hori- zontal axis. The isotachs represent the vector daily mean flow component normal to the current-meter section (cf. Table 7-3 below). The hatched areas represent incidents of reversed flow, i.e., nominally southward. The generally higher speeds on the eastern side, TABLE 7-3. Transport section geometry Line segment Mean deptii, Normal direction Mooring length, 1; h, of 1; km m °T NC-2 NC-3 NCI 58.8 50 318 NC2 41.9 53 322 NC3 48.5 54 325 NC4 57.3 53 346 NC6 53.0 50 005 NC7 57.5 42 006 NC-6 NC-7 Figure 7-3. Temporal development of flow in the Cape Lisburne section for September 1976. Ordinate is days, and current meters are placed in their relative geographic locations along the abscissa. Isotachs denote flow compo- nent normal to the section; hatched areas represent areas and times of reversed (nominally southward) flow. NC-1 NC-2 NC-6 NC-7 rim>iNi!in/.',TTm7Tm?//////m////:,{fM ' '^'.iM'i 1^ - Figure 7-4. Same as Fig. 7-3, for October 1976. NC-1 NC-2 NC-6 NC-7 Figure 7-5. Same as Fig. 7-3, for November 1976. the frequent flow reversals, particularly on the eastern side, and the coherence between adjacent mooring sites are all readily apparent. An interesting feature of these representations is the tonguelike character of the isotachs. That is, as a rule major 100 Physical oceanography NC-1 NC-2 NC-6 NC-7 NC-1 NC-2 NC-3 NC-6 NC-7 Figure 7-6. Same as Fig. 7-3, for December 1976. Figure 7-9. Same as Fig. 7-3, for March 1977. Figure 7-7. Same as Fig. 7-3, for January 1977. ES^ZZZ55^^^- Figure 7-8. Same as Fig. 7-3, for February 1977. flow events axe most frequently connected with the ends of the section and show up at points increas- ingly further inside the section with reduced magni- tudes. This is true of the strongest flow events, both northerly and southerly. Prime examples are 15-18 September (Fig. 7-3), 26-28 October (Fig. 7-4), 13-17 November (Fig. 7-5), and a whole series of southerly flow events along the eastern end of the section during December to February (Figs. 7-6 to 7-8). On occasion, a flow change at one end of the section appears appreciably later in the interior of the sec- tion. For example, southerly flow through the eastern end during 22 to 25 September appeared in the interior 28 to 30 September (Fig. 7-3). Further- more, a short time lag of a day or so for flow events at the more interior stations is extremely common in these figures. The analysis and modeling of these flow events is outside the scope of the present de- scription, but many of them are certainly suggestive of long waves propagating along the coasts (in this case, coasts with rather complicated geometry). Figures 7-10 and 7-11 show the spectral distribu- tions of energy for moorings NCI and 7, again taken as representative of two different flow regimes. At NCI the energy is relatively low, except for a strong NC-1 ROTATED 9/01/76 0000 V-COMP 1 OELTfl-T =40 niN 500. T cx >- LJ :z. UJ ID a LlJ 10 10"' 10 FREQUENCY IN CYCLES/DRY Figure 7-10. Spectral energy distribution for meter NCI, September-March 1976-77. Ninety-five percent confi- dence limits are indicated. "Rotated" indicates the spec- trum is for the component of measured velocity normal to the section line, i.e., toward or away from Bering Strait (cf. Fig. 7-12). 10^ Water transport in the vicinity of Bering Strait 101 NC-7 ROTATED 9/01/76 0000 V-COMP DELTfl-T z40 MIN 400, Q_ cr >- CJ z: LU ZD o LlJ 0. 10 1-1 10"' 10' FREQUENCY IN CYCLES/DRY Figure 7-11. Same as Fig. 7-10, for meter NC7. 10' signal at the M2 tidal frequency. In particular, there is essentially no variance at time scales less than about three days. At lower frequencies the energy level increases gradually, with the bulk contained at fre- quencies less than 0.1 cycle per day. At NC7 the flow is of an entirely different tem- poral nature. The M2 signal is about one-third as large, but the low-frequency bands are far stronger everywhere below the diurnal. By far the largest energy concentration is the one with a period close to five days. We note also that the energy does not drop off at the lowest frequencies, as is the case for NCI. In summary, then, the current records show a flow field which in the central and western Chukchi (NCl-4) is less energetic (except in the semidiurnal tidal band) and far less variable than in the eastern Chukchi (NC6 and 7). At the same time, the flow appears sufficiently coherent laterally to encourage transport calculations, with the expectation that there will be large variations at time scales longer than a few days. Cape Lisburne transports Figure 7-12 shows the transport section defined by the six moorings NC1-NC7, and Table 7-3 lists the geometric parameters. The length of Ij , i.e., the effective horizontal extent over which the current measured at NCI is ap- plied for transport purposes, was determined in the following way. Examination of Brown Bear (1960), Northwind (1963), and Oshoro Maru (1972) data (cf. Coachman et al. 1975 for data references) indicate that in the vicinity of 1, , the influence of the south- 178° 176° 174° 172' 170°W 168° 162° 160° Figure 7-12. Geography of the Cape Lisburne transport section. eastward-flowing Siberian Coastal Current extends seaward somewhat more than 90 km. The south- westward extension of line 1^ was therefore termi- nated at the indicated transition point to the cold, low-salinity Siberian coastal water. There is no guarantee that the Siberian Coastal Current will not from time to time cross line 1, , and indeed the current and temperature record from NCI indicates that on at least one occasion the direct influence of the coastal current may have extended as far seaward as the mooring position. On the other hand, the cross-correlations generally show significant length scales of 100 km or more, which is considerably greater than the distance from NCI to the southwest end of li , about 38 km. We can therefore expect the flow to be reasonably coherent over the distance 1, , and the current record from NCI to be indeed representative of the line 1, . We return to the calcu- lation of flow correlations later. Line I7 was extended eastward to the 30 m isobath, where the Oshoro Maru measurements still showed northeasterly flow; this is 6-7 km from shore. Since the water is quite shallow over the remaining distance, it is doubtful that much of the transport is missed. Lines I2-I6 join points midway between each adjacent pair of mooring positions. A more difficult problem is vertical extrapolation of the velocity from the instrument position 10 m above the bottom. Two shipboard current-meter sections have been made along the mooring line, one from the Oshoro Maru in July 1972 and the other from the Moana Wave in September 1976. For each of these sections we calculated the ratio between the mean cross-section velocity component in the layer deeper than 30 m and that in the water column taken as a whole. Omitting two apparently anomalous 1 02 Physical oceanography stations of the total of 26, the mean ratio is 1.1: i.e., the deep flow was slightly higher than the mean over the water column. During the seven or eight winter months, when the water is vertically homogeneous, any baroclinic shear would of course be absent, and only the frictional shear (including that against the ice) would remain. In the absence of more extensive information, we therefore assume that the current 10 m above the bottom is representative of the water column as a whole. We further conjecture that the transport error introduced by this assumption is within 10 percent. The temporal and spatial correlations of the flow field are critical for transport calculations. The spectral calculations (Figs. 7-10, 7-11) show relatively little energy at frequencies higher than the diurnal, except for the M2 tide. This is in fact the case at all the mooring sites, as shown in Table 7-4, where the autocorrelations indicate a time scale of inde- pendence that exceeds two to three days. We should therefore expect that transport calculations done on a daily basis will cover all the important time scales. In Table 7-4 the autocorrelations have been calculated for four equal portions of the seven-month compari- son period, and we note that the statistics vary not only from mooring to mooring, but also in time. For the transport calculations to be meaningful, it is essential that the spacing of the current meters be less than the correlation length (cf. Fandry and Pillsbury 1979). The cross-correlations are shown in Table 7-5, again over 53-day intervals, as in Table 7-4. Table 7-5 provides a measure both of the mooring-to- mooring coherence and also of the correlation func- tion from west to east (first column) and from east to west (last row). Except for the first portion of the record, the flow field is well correlated from NCI to NC4, a distance of 138 km, and from NC7 to NC4, a distance of 160 km. Throughout the record, adjacent moorings (diagonal) are correlated. The resulting daily transports (arithmetic estimates only, contrast Fandry and Pillsbury 1979) are presented in Fig. 7-13. Bering Strait transports The daily mean transports normal to the Cape Lisburne section are well correlated with the daily mean north velocities measured at NCIO (r^ = 0.81). Transports through Bering Strait were calculated from the linear regression and are shown together with the Cape Lisburne transports in Fig. 7-13. As a check, the eleven detailed Bering Strait sections were examined (10 of these are shown in Coachman et al. 1975, Fig. 48). From the isotachs of north-south flow, the velocities at the location of the NCIO current meter were estimated and related to the measured transport. The results, together with the linear regression of Cape Lisburne transport on Bering Strait north velocity, are depicted in Fig. 7-14; the good agreement suggests that the mean daily velocities measured at NCIO provide reasonable estimates of the transport through Bering Strait. An obvious conclusion is that the whole southern Chukchi Sea, from Bering Strait to north of Cape Lisburne, is behaving as a coherent unit in response to the driving forces. There are two reasons why the coherence between Cape Lisburne transports and NCIO velocities is not even better: (1) The shear in Bering Strait is not invariant. Thus, velocities measured at one point in the cross- section can only represent transport within some TABLE 7-4. Number of days required for autocorrelation of component of velocity normal to sections to be statistically indistinguishable from zero Period 1 Period 2 Period 3 Period 4 Mooring 1 Sept. - 23 Oct. 24 Oct. - 14 Dec. 15 Dec. — 5 Feb. 6 Feb. - 30 Mar. NCI 5 4 6 3 NC2 6 4 4 4 NC3 5 4 3 3 NC4 5 4 10 4 NC6 3 3 8 3 NC7 3 3 4 3 Water transport in the vicinity of Bering Strait 103 ARCH BERING STRAIT CAPE LISBURNE BERING STRAIT TRANSPORT SECTIONS / Figure 7-13. Daily mean transports through the Cape Lisburne section and Bering Strait. Bering Strait transports were calculated from the correlation depicted in Fig. 7-14. limits; Fig. 7-14 suggests the limits to be less than ±lSv. (2) The surface area of the southern Chukchi Sea, including Kotzebue Sound, between the measurement locations (~1.4 X 10^ km^ ) rises and falls, partly in response to differential transports between the sections. (In this regard, inflow into Kotzebue Sound of the Noatak and Kobuk rivers is insignificant, the TABLE 7-5. Cross-correlation of normal component of velocity, zero lag. For duration of numbered periods, see Table 74 NCI NC2 NC3 NC4 NC6 NC2 NC3 NC4 NC6 NC7 Per. 1 0.58 Per. 2 0.85 Per. 3 0.72 Per. 4 0.75 Per. 1 * 0.81 Per. 2 0.75 0.91 Per. 3 0.50 0.62 Per. 4 0.63 0.72 Per. 1 - -0.37 0.45 Per. 2 0.53 0.87 Per. 3 0.56 0.53 Per. 4 0.63 0.62 Per. 1 * 0.27 Per. 2 * 0.80 Per. 3 0.30 0.69 Per. 4 0.48 0.45 Per. 1 * * * * 0.68 Per. 2 * * 0.43 0.73 0.95 Per. 3 * 0.29 0.57 0.52 0.88 Per. 4 0.45 0.40 0.36 0.32 0.78 Asterisks denote velocity components uncorrelated at the 95 percent confidence level. NORTH VELOCITY , cm/sec Figure 7-14. Correlation between mean daily Cape Lisburne transports and north velocity measured at NCIO in Bering Strait. Ninety-five percent confidence limits are indicated. Also shown are the transports from the previous 11 Bering Strait cross-sections of anchored sta- tions related to the velocities at the location of NCIO in the section. (N.B. The correlation line is for Cape Lisburne transports vs. daily mean north component of velocity measured at NCIO, n = 211, data not shown.) 104 Physical oceanography maximum being <10'* m-'/sec.) Examination of the data in Fig. 7-13 shows that although sometimes the transports at the two locations were in phase, more frequently they were out of phase by a day in either direction, and this buffering effect of the southern Chukchi Sea reduces the coherence in transports between the two locations. DISCUSSION There are several surprises in the results, rela- tive to the transport estimates of Coachman et al. (1975). These include the low mean values for longer-term flows (>1 month) and the extreme magnitudes of both the daily southerly transports and the accelerations. The first surprise is the low value of long-term mean transport— over the seven months of record T = 0.3 Sv, only about 20 percent of the annual mean value estimated by Coachman et al. (1975). This suggests that the previous estimate of mean annual transport of 1.5 Sv must be revised downward sub- stantially, despite the fact that the mean daily transport values are of the same order as measured earlier and include even greater extreme values of both northerly and southerly transport. The cause of a reduced annual mean is the frequent incidence of southerly transport events during fall-winter, the season not covered previously by measurements. All previous data were from summer (July to early October), and of the 22 summer sections only two showed southerly transport (—0.1, —0.2 Sv.). In contrast. Table 7-6 summarizes the incidence and duration of southerly flow events from the present mooring records. TABLE 7-6. Incidence and duration of major southerly transport events, September 1976 through March 1977 Month No. of events Total duration, days September 3 October 2 November 5 December 4 January 3V2 February iy2 March 2 7 11 9 12 8 8 6 A relatively high incidence of southerly transport events in fall agrees with Bloom's (1964) measure- ments, made with electrodes spaced over a distance of 7.5 km west from Cape Prince of Wales. Although the potential measurements cannot be converted to transports, they suggest possible southerly flow events and their duration. Fig. 7-15 compares such suggestions with measured southerly flow events during 1976 and 1977. In both years the period September through December showed significant southerly flow, but January through March 1976-77 showed definitely more southerly flow events than the similar period in 1957. SOUTH FLOW EVENTS A FROM POTENTIAL MEASUREMENTS BLOOM 1964 SEP ' OCT ' NOV I DEC ' JAN ' FEB ' MAR ' APR ' MAY ' JUN ' JUL ' AUG 1976- 7 1 JIL A 1 FEB ' MAR Figure 7-15. Incidence and duration of southerly flow events in Bering Strait, from the 1976-77 measurements and from the electrical potential measurements of Bloom (1964) for 1956-57. We can construct a revised estimate of mean annual transport. The Cape Lisburne transports were extrap- olated to the period April-June by linear regression of the September-March transports on the mean speeds (normal to the station line) measured at NC3 and NC7, which had records extending through June. The correlation was r- = 0.8; the regression equation gives for April and May, T = 0.50 Sv, and for June, T = 0.84 Sv. If we now assume that 1976-77 was a representative year, and that the previously deter- mined average of 1.5 Sv applies to July and August, the mean annual transport would be 0.6 Sv. Soviet oceanographers have published many estimates of the long-term transports, but nowhere can we find a description of the data or observational methods on which their results are based. Table 7-7 contrasts our results with the values reported by Fedorova and Yankina (1964), who included results from eight other studies. There appears to be an annual cycle of transport magnitude, from low mean monthly values in winter (0-0.5 Sv) to high values in August (1.5-2.0 Sv). But we do not beheve the cycle should be taken too literally, because in any one month the mean transport can vary significantly, depending on the incidence and duration of southerly flow events that affect the longer-term mean trans- port values. These monthly values, in turn, will Water transport in the vicinity of Bering Strait 105 TABLE 7-7. Monthly and annual Bering Strait transports in Sv Previously measured Soviet results'* cross sections 1941- 1976- No. of 1961 Range 1977 Averages sections Jan .60 .60-.86 .54 Feb .50 .50-.94 .24 Mar .46 .46-.70 .33 Apr .59 .58-.80 .50^ May .86 .76-1.00 .50^ June 1.24 .94-1.24 .84'' July 1.56 1.01-1.77 1.0 (14) Aug 1.62 1.13-2.25 1.7 ( 5) Sept 1.35 0.38-2.37 .46 1.3 ( 2) Oct .99 0.19-2.13 .03 1.3 ( 1) Nov .80 0.38-1.66 .32 Dec .68 0.56-1.20 .13 Mean Annual .95 0.31-1.42 0.6 I '^from Fedorova and Yankina, 1964 ''extrapolated probably depend primarily on the pattern of atmospheric conditions which happens to prevail during the period. In any event, our monthly mean values appear to be, if anything, a little lower than the compiled Soviet results. Fedorova and Yankina also reported annual means for 1952 through 1961 which ranged between 0.85 Sv and 1.08 Sv to the north. Comparison of the Soviet results with those from 1976-77 and the earlier measurements suggests that the 1976-77 annual estimate of 0.6 Sv is prob- ably a minimum, and also that there undoubtedly are interannual variations. Under these considerations, our present best estimate is 0.8 ± 0.2 Sv for the annual mean. A second major surprise in the 1976-77 records is the high values of daily mean transport. The range of values, 3.1 Sv northerly to —5.1 Sv southerly, extends the earlier range estimates (Coachman et al. 1975) considerably, particularly on the southward end. In the earlier results, the highest value was 2.2 Sv northerly; in the new results only three values from Bering Strait and four from Cape Lisburne exceed this. However, in southerly flow the earlier extreme value was only —0.2 Sv. In 1976-77, 26 daily values exceeded —1 Sv, of which eight were associated with the major flow reversal of Oct. 24 through Nov. 1 (Fig. 7-13). In the first attempt to relate mean daily trans- ports to atmospheric conditions. Coachman et al. (1975) noted that low northward transport seemed to be correlated with low atmospheric pressure at Nome with a phase lag of ~ 1 day, and vice versa. They also noted that the best correlation for the 21 cases (r^ - 0.74) included both atmospheric pressure and a northwest-southeast pressure difference across the system (pressure at Nome minus pressure at Cape Schmidt). As the first step in interpreting the relationship between transport variations and atmospheric condi- tions for the seven months of record, daily mean surface pressures were obtained from Nome (P„ ) and Kotzebue (P^ ) on the eastern side of the system and Provideniya Bukhta (Pp ) and Cape Serdtse-Kamen (Pgk) on the western side, the latter two stations being located south and north of Bering Strait. (N.B. Apparently Cape Schmidt is no longer a reporting station; however, Cape Serdtse-Kamen, to the north- west of Bering Strait, lies due west of Kotzebue with the same east-west separation from it as Nome from Provideniya Bukhta, see Fig. 7-1.) All possible correlations were investigated, with the following results: (1) The surface atmospheric pressure pairs on the eastern (P^ and P^ ) and western (Pp and P^^ ) sides of the system were each highly correlated, but there were significant day-to-day E-W pressure differences and variations. The east-west pressure differences south (P^-Pp) and north (Pk-Psk) of the strait were well correlated (r' =0.80). (2) The Bering Strait transports were well cor- related with the east-west pressure differences. The best correlation (r' = 0.70) was a two-component correlation using both (P^-Pp) and (Pk-Psk)> but the correlation was essentially the same (r^ =0.68) using only (Pn-Pp). The relationship was nearly in phase. The correlation between transport and (P^-Pp) the same day was r^ = 0.62, but only 0.53 for (P^-Pp) one day earlier; the best result was obtained by averaging (Pn-Pp) for the same day with one day earlier. The resulting equation is T = 1.10 + 0.20 (P„-Pp), where T is in Sv and (Pn-Pp ) in mb. The daily mean Bering Strait transports and east-west pressure differ- ences for the seven-month records are shown in Fig. 7-16. (3) The transports were not correlated with any of the individual station pressures or other combina- tions of pressures, regardless of lag. This result is 106 Physical oceanography contrary to the earlier suggestion (Coachman et al. 1975) that the surface pressure at Nome one day earUer could be used as an index of the transport. The mean atmospheric pressure at Nome for Septem- ber-March was about 10 mb lower than that for the earlier transport calculations, and strong south transport did not correlate with low pressure at Nome; about one-half the incidences of south trans- port occurred when ?„ was greater than 1,000 mb. Apparently, there were too few data in the previous analysis, and furthermore those cases were from summer, when there is more sustained northerly transport and atmospheric pressure at Nome tends to be higher. To explore the relationship between atmospheric weather patterns and transport variations, we ex- amined the mean daily surface atmospheric pressure charts for the northern hemisphere, using the chart for 0000 Z to represent the previous day (the longi- tude of Bering Strait, 165°W, is time zone +11). The water level records from pressure gauges mounted on moorings NC7 and NC17 (Fig. 7-1) provide additional data useful for interpreting the transport accelera- tions. Because the meters were at different depths, the measurements can be used only to suggest the relative changes in north-south water level along the axis of the system. The mean daily values are shown in Fig. 7-16 as anomalies of water level from their mean values over the seven months. Also plotted is the difference between water level anomalies at the two meters, with positive values corresponding to the sea level at St. Lawrence being higher than normal, relative to sea level at Cape Lisburne; that is, sea level sloped more strongly down to the north. First, examination was made of pressure patterns surrounding all the major south flow events, e.g., 16 September, 10-11 October, 26-28 October, 22-23 November, et seq. In every case the large-scale atmospheric pressure patterns were the same. One day before a peak in southerly flow, a strong low- pressure system was centered some distance to the southeast of Bering Strait, in the area of Bristol Bay, Kodiak, Anchorage, and the northern Gulf of Alaska. At the same time the Siberian high was centered some distance west or west-northwest of the strait. The isobars signifying the strongest pressure gradient between pressure centers were located precisely over the Bering Strait region and, most significantly, they had a nearly north-south orientation which extended from over the Chukchi Sea south into the central Bering Sea— completely across the northern Bering Sea shelf. If the north-south orientation of the isobars did not extend totally across the northern shelf, or if the isobars were oriented northeast- southwest (the most typical configuration), strong southerly flow events did not occur. Fig. 7-17, surface pressure charts for 20-23 No- vember, illustrates a typical case. A low over the Gulf of Alaska on 20 November moved over Kodiak on 21 November and deepened. At the same time, the center of Siberian high pressure moved west. Thus, on 21 and 22 November the strongest pressure EAST-WEST PRESSURE DIFFERENCES ""[Pnome PrrovideniyaJ [Pkotzebue Pserdtse-kamenJ WATER LEVEL ANOMALIES — NC 17 (ST LAWRENCE) --NC 7 (CAPE LISBURNE) WATER LEVEL ANOMALY DIFFERENCES (PLUS DOWN TO north) Figure 7-16. Daily mean values of transport through Bering Strait, E-W surface atmospheric pressure differences across the region ((Pn-Pp) is south of Bering Strait and (Pk-Psk) north, cf. Fig. 7-1), water levels measured at St. Lawrence Is. (NC17) and Cape Lisburne (NC7, cf. Fig. 7-1) plotted as anomalies from seven-month mean values, and the differences in water-level anomalies (south minus north). Water transport in the vicinity of Bering Strait 107 20 NOV 22 NOV 21 NOV 23 NOV 140° 160° 180° 160° 140° 120° ^ \ x' ^\A///i^ ^ m ^^^:^*^ /< y ^'C^mi Ir ^ r ij \V^^f\ ^ 6^ \Aid;x^i^ 60° ^\ V>L3yy' j m; iDrVVr^ ^ 60° 55° 5! ^1 ^' / ^ 55' 50° t^ ■^^// \^ 1^ 4 ^V-12V^ V \ 50" L -r^ i ^25v / ; 1 t- 1 \v \ \ A -tT \ \ \ 170° 180° 170° 160° 150° Figure 7-17. Charts of surface atmospheric pressure for 0000 Z 21-24 Nov. 1976, taken to represent midday conditions the day before. The chart covers the area from the north Pacific to the Arctic Ocean, with Alaska on the right and Siberia on the left. gradients were directly over the Bering Strait region and the isobars had an extended north-south orienta- tion. The situation held over 22 November, although the pressure gradient was less as the low began filling, and by 23 November these conditions had dissipated. Southerly transports of —1.9 Sv and —1.8 Sv were recorded on 22 and 23 November (Fig. 7-16). The mechanism which drives major south flow events now seems clear. Strong north winds must develop over the whole northern Bering Sea, not just over the immediate region of Bering Strait. Large- scale, strong atmospheric pressure cells are required, a low well to the southeast and a high well to the west. The strong northerly winds generated thereby move water southward off the entire northern Bering Sea shelf. Removal of sufficient water off the northern shelf generates a sea-level slope down to the south (sea-level slope has been shown to be the major force driving transport through the strait (Coachman et al. 1975)). This, together with the strong north winds caused by the east-west atmospheric pressure gradient aligned over the system, drives enhanced southerly transport. It apparently requires about one day for development of these conditions, so that maximum south transport occurs the following day. Because the system behaves to a marked degree as a coherent 108 Physical oceanography unit, water levels at both St. Lawrence and Cape Lisburne fall together and are nearly in phase with the transport (cf. Fig. 7-16). This mechanism also accounts for the nearly in- phase behavior of transport and (Pn-Pp). Since it is the removal of water from the whole northern shelf area (largely located south of St. Lawrence Island) that is responsible for major south transport, the strong east-west pressure difference associated with the proper atmospheric condition does not have to lead the transport if the weather system develops from the south. Water removal from the northern shelf can begin before the pressure gradient indicating a strong north wind is registered at Nome and Provi- deniya Bukhta, located north of St. Lawrence and closer to Bering Strait. Northward transport stands in contrast to the southerly transport events. Periods of northerly flow tend to be more persistent and not so great in magni- tude, nor do they show the marked episodic character of the southerly flows (Fig. 7-16). The greater persistence of northerly flow must reflect the basic driving force, a higher sea level in the Bering Sea than in the Arctic Ocean (Coachman et al. 1975), which still remains unexplained. There were, however, a number of relatively rapid northward accelerations of transport during the seven months of record, which appear to have two basic causes: (1) After strong south transport events, rapid accelerations commonly occur which can be thought of as compensatory accelerations. When atmospheric conditions causing the southerly transport event dis- sipate, water is not being removed from the northern shelf, but there is still large southerly transport in the system. Water "piles up" in the region around St. Lawrence Island and Norton Sound, a condition reflected by a strong positive difference in water level anomalies, and following this by about one day a strong northward acceleration occurs. For example, on 23 November the atmospheric conditions driving the southerly flow of 22-23 November had dissipated (Fig. 7-17) and on 23-24 November the St. Lawrence water level anomaly rose significantly above that at Cape Lisburne (Fig. 7-16). Bering Strait transport changed from —1.8 Sv on the 23rd to +1.1 Sv on the 25th. (2) Occasionally, major northward accelerations appear to be, at least in part, directly driven by atmospheric conditions. Specifically, these are a strong pressure low centered in the western Bering Sea, southwest of Bering Strait, or a deep trough from the central Aleutians toward the northwest, so that the isobars in the strong pressure gradient are directed northward from the central Bering Sea along the axis of the system. This configuration creates strong southerly winds which can move water from the central Bering Sea onto the northern Bering Sea shelf, raising the water level in the vicinity of St. Lawrence Island and enhancing the sea-level slope down to the north. The best example during the seven months is depicted in Fig. 7-18. The proper isobar configuration was present on 5 January, while northerly transport accelerated to 2.5 and 2.6 Sv on 6 and 7 January, respectively (Fig. 7-16). In this case northerly transport was probably enhanced by northward extension of the east-west pressure gradi- ent past Cape Lisburne, indicating southerly winds over the southern Chukchi Sea. These could have abetted northerly flow by removing water from the downstream side of the strait. Water levels at both St. Lawrence Island and Cape Lisburne rose con- siderably as a consequence of water being advected into the system from the south, and the increases were in phase with the transport, reflecting the coherent behavior of the system. 5 JAN 1977 Figure 7-18. Same as Fig. 7-17, for 5 Jan. 1977. The prolonged southerly flow event of 24 October through 1 November (Fig. 7-16) illustrates the proposed set of mechanisms well. Development of the proper atmospheric conditions for southerly flow was first noticeable on the chart for 24 October. The low was centered over Anchorage, and the maximum pressure gradient west of the low extended from Pt. Barrow to the Pribilof Islands. It reached an extreme value on 16 October (low = 968 mb, high = 1041 mb), followed by the maximum measured southerly transport of —4.5 Sv on 27 October. The low re- mained stationary and had partially filled by the 29th, but was replaced by another intense low (956 Water transport in the vicinity of Bering Strait 109 mb) on the 30th. Thus, the northward acceleration of 28 and 29 October, caused by an enhanced sea surface slope down to the north following the partial filling of the first low (strong positive water level anomaly difference on the 27th: Fig. 7-16) was arrested through 1 November by the new intense low forming on 30 October. The pattern of closely spaced isobars extending north-south on 31 October had disappeared on 1 November, after which there was another northward acceleration. SUMMARY AND CONCLUSIONS New data on the flow regime in the vicinity of Bering Strait obtained since the comprehensive analysis of Coachman et al. (1975) include seven- month (September 1976-March 1977) current-meter records along a nearly closed section from Cape Lisburne to Siberia and from one instrument in Bering Strait. Daily mean currents normal to the Cape Lisburne section showed that the flow field of the southern Chukchi Sea divides into two regimes. In the western half currents were weak (<20 cm/sec). Except for a strong M2 tidal signal, the bulk of the energy was contained at frequencies less than 0.1 cycles per day. In the eastern half, the currents were much swifter (up to 68 cm/sec), but with less energy at the M2 tidal frequencies. At subtidal frequencies, the energy levels were far higher than to the west, with the largest energy concentration centered near the five-day time scale. Unlike what the records show for the western half, the energy did not drop off at the lowest definable frequencies. Throughout the section there were numerous current reversals (to the south) which were concentrated in the fall and early winter. These southerly flow events tended to begin at the ends of the section and propagate into the interior of the sea and are suggestive of long waves propagating along the coasts. Daily mean transports across the Cape Lisburne section were well correlated (r^ = 0.81) with mean daily north velocities measured in Bering Strait. Thus the whole southern Chukchi Sea from Bering Strait north past Cape Lisburne is responding as a coherent unit to the driving forces. These data, taken together with the previous 22 cross sections of measured flow, allow calibration of the Bering Strait measurements as transports. Daily mean values ranged from +3.1 Sv (north) to —5.0 Sv (south), extending the range of previously measured values considerably, particularly toward the south. There were very large daily accel- erations in transport, in six instances exceeding 2 Sv/day. Longer-term mean transport values (monthly, annually) are determined by the incidence and duration of southerly flow events. Such events were most frequent in fall and winter 1976-77, correspond- ing with the results from electric-potential measure- ments in Bering Strait over the year 1956-57 (Bloom 1964). More frequent and prolonged southerly flow events in fall-winter probably give rise to an annual cycle in Bering Strait transport, with monthly mean values varying from 0-0.5 Sv northerly in winter to 1.0-2.0 Sv northerly in summer. The mean annual transport appears to be ~0.8 ± 0.2 Sv. Analysis of the relationship between transport and the atmospheric pressure field suggests a good correlation (r^ = 0.7) between transport and the east- west pressure difference across the system. When pressure on the east side (Nome, Kotzebue) is greater than on the west side (Provideniya Bukhta, Cape Serdtse-Kamen), the regional air flow is towards the north, as are the water transports, and vice versa. Transports are not correlated with atmospheric pressure at Nome one day before, as suggested earlier. Southerly transport through the strait has a marked episodic character, and events are caused by particular large-scale atmospheric conditions. When a strong east-west pressure gradient lies over the strait and extends in a north-south direction from the Chukchi Sea to the central Bering Sea, entirely crossing the northern Bering Sea shelf, extensive northerly winds move water southward off the shelf. This produces a sea-level slope down to the south which, together with the northerly winds, drives southward transport. The atmospheric pressure pattern causing this condition is always a strong low located a considerable distance southeast of the strait (e.g., over Kodiak) together with the Siberian high centered some distance west of the strait. Peak southerly transport follows development of such a condition by about one day. Northward transport is more persistent, does not show the episodic character of southerly trans- port, and the transport values are lower than those of peak southerly flow. The basic driving force is an as yet unexplained higher mean sea level in the Bering Sea than in the Arctic Ocean. Strong northward accelerations have two causes. Most common is a compensatory acceleration following a southerly transport event; dissipation of the driving force for the southerly flow event, but continued southerly transport through Bering Strait, raises water level in the vicinity of St. Lawrence Island/Norton Sound, resulting in an enhanced sea-level slope down to the north driving northerly acceleration. The second mechanism, which rarely occurred during the seven months of record, was the development of a strong low-pressure center in the western Bering Sea, south 110 Physical oceanography and west of the strait, with isobars in the strong pressure gradient directed northward from the central Bering Sea to the strait. The extensive strong south- erly winds abetted northerly accelerations by moving water into the northern Bering Sea shelf, thereby enhancing the sea-level slope down to the north. REFERENCES Antonov, V. S. 1968 The nature of water and ice move- ment in the Arctic Ocean. AANII. Trudy 285:154-82. (transl.) Bloom, G. L. 1964 Water transport and temperature measurements in the eastern Bering Strait, 1953-1958. J. Geophys. Res. 69(16):3335-54. Coachman, L. K. 1978 On the oceanographic role of Arctic shelves. Invited Lecture, Fall AGU meeting, December 1978. (mimeo- graphed) Coachman, L. K., and K. Aagaard 1966 On the water exchange through Bering Strait. Limnol. Oceanogr. ll(l):44-59. Coachman, L. K., K. Aagaard, and R. B. Tripp 1975 Bering Strait: The regional physical oceanography. Seattle and London: Univ. of Washington Press. Coachman, L. K., and C. A. Barnes 1961 The contribution of Bering Sea water to the Arctic Ocean. Arctic 14 (3):146-61. Fandry, C, and R. D. Pillsbury 1979 On the estimation of absolute geo- strophic volume transport applied to the Antarctic Circumpolar Current. J. Phys. Oceanogr. 9:449-55. Fedorova, A. P., and A. S. Yankina 1964 The passage of Pacific Ocean water through the Bering Strait into the Chukchi Sea. Deep-Sea Res. 11: 427-34. (transl.) Maximov, L V. 1945 Determining the relative volume of the annual flow of Pacific water into the Arctic Ocean through Bering Strait. Probl. Arktiki, No. 2:51-8. (transl.) Sverdrup, H. U. 1929 The waters on the North Siberian Shelf. The Norweg. North Polar Exped. with the Maud 1918-1925. Sci. Res. 4(2). Tides of the Eastern Bering Sea Shelf Carl A. Pearson,' Harold O. Mofjeld,^ and Richard B. Tripp- ' National Ocean Survey, assigned to: Pacific Marine Environmental Laboratory/NOAA Seattle, Washington ^ Pacific Marine Environmental Laboratory/NOAA Seattle, Washington ^ Department of Oceanography University of Washington Seattle, Washington ABSTRACT INTRODUCTION The acquisition of a substantial amount of pressure-gauge and current-meter data on the Bering Sea shelf has permitted a much more accurate description of the tides than has previ- ously been possible. Cotidal charts are presented for the Mj and, for the first time, the N, , K, , and O, constituents, and tidal current ellipse charts for M^ and K, . S^ , normally the second largest semidiurnal constituent, has not been included because it is anomalously small in the Bering Sea. The tide enters the Bering Sea through the central and western Aleutian Island passes and progresses as a free wave to the shelf. Largest tidal amplitudes are found over the southeastern shelf region, especially along the Alaska Peninsula and interior Bristol Bay. Each semidiurnal tide propagates as a Kelvin wave along the Alaska Peninsula but appears to be converted on reflection in interior Bristol Bay to a Sverdrup wave. A standing Sverdrup (Poincarfe) wave resulting from cooscillation in Kuskokwim Bay is evident on the outer shelf. The semi- diurnal tides are small in Norton Sound where there is an amphidrome. The diurnal tides, which can have only Kelvin wave dynamics, cooscillate between the deep basin and the shelf. Amphidromes are found between Nunivak Island and the Pribilof Islands, and west of Norton Sound. Throughout most of the shelf the tide is of the mixed, predominantly semidiurnal type; however, the diurnal tide dominates in Norton Sound. Tidal models by Siinderman (1977) (a vertically integrated M2 model of the entire Bering Sea) and by Liu and Leendertse (1978, 1979) (a three-dimensional model of the southeastern shelf incorporating the diurnal and semidiurnal tides) are discussed. Good qualitative agreement is found between the models and observations. As with most continental shelves, the tides and tidal currents on the eastern Bering Sea shelf play important roles in such oceanographic processes as the maintenance of the density structure, sediment resuspension and transport, and the distributions of benthic and intertidad organisms. A knowledge of the tides and tidal currents is therefore necessary in order to understand the region's oceanography. The tides of the Bering Sea have been of interest to physical oceanographers and astronomers for a long time (e.g., Jeffreys 1921, Munk and MacDonald 1960, Cartwright 1979). This interest has been based on the premise that the vast continental shelves of the Bering Sea (Fig. 8-1), with their proximity to the Pacific Ocean, act as a major sink of the world's tidal energy. Yet many aspects of the tides and tidal currents in the Bering Sea have remained unknown because inadequate data made it impossible to draw definitive cotidal charts or to obtain reliable boundary condi- tions for numerical models. Fortunately, in recent years a large number of pressure-gage and current- meter observations have been made on the eastern Bering Sea continental shelf, and the new data make possible a more detailed description of the tides in the eastern Bering Sea. Ill 112 Physical oceanography Figure 8-1. Bathymetric chart of the Bering Sea. Concurrent with the recent field work has been the development of several numerical models for tides in part or all of the Bering Sea. One of these, the Liu and Leendertse (1978, 1979) model of Bristol Bay, will be described in detail. The other models are discussed briefly for completeness and to show the reader the scope of current theoretical work. The field work is also continuing. In a sense then, this chapter is a progress report as well as a review of past work and a description of new results. The tides that we shall be concerned with are the principal tidal constituents Nj and Mj in the semi- diurnal band (periods of about 0.5 days) and d and Ki in the diurnal band (periods of about 1.0 day). Ordinarily, the principal solar semidiurnal constituent S2 would be included in the discussion. However, S2 is anomalously small throughout the Bering Sea, possibly because it has small amplitudes in the adjacent North Pacific Ocean. Whatever the reason, no consistent distribution for S2 appeairs in the field data above the background noise level. The compli- cated distributions of semidiurnal and diurnal tides in the Bering Sea produce a rich variety of tidal types, ranging from fully semidiurnal in some regions to Tides 113 fully diurnal in others. Sample time series will be shown later in the chapter to illustrate the tidal types. Probably the most important figures are the cotidal charts for the four principal constituents, because from these charts it is possible to infer much about the dynamics of the tides and to obtain harmonic constants for tidal predictions. Empirical cotidal charts derived from recent data and theoretical cotidal charts from models will be presented. The discussion of tidal currents in the Bering Sea will be presented with less certainty than that of the tides, because the few early measurements of tidal currents were of short duration and mostly limited to harbors and nearshore regions, and the recent obser- vations from offshore current meters occasionally suffered from errors due to biological fouling and effects of wave motion. Still, with careful study and editing of the recent data reasonably accurate tidal current harmonic constants can be obtained for most areas. These results help define the dynamics control- ling the tidal motion. The setting for the tidal and tidal current distribu- tions can be obtained from previous work (Harris 1904, Leonov 1960, Office of Climatology and Oceanographic Analysis Division 1961, Defant 1961, Coachman et al. 1975) which for the most part describes the semidiurnal tide. The tide wave enters the Bering Sea as a progressive wave from the North Pacific Ocean, mainly through the central and west- ern passages of the Aleutian-Komandorski Islands. The Arctic Ocean is a minor secondary source of tides which propagate southward into the north Bering Sea where they comphcate the tidal distributions. The North Pacific and Arctic Oceans are also sinks of tidal energy for tides propagating out of the Bering Sea. Tides in the Bering Sea are considered to be the result of cooscillation with large oceans. Once inside the Bering Sea, each tidal constitent propagates as a free wave subject to the Coriolis effect and bottom friction. The tide wave propagates rapidly across the deep western basin. Part of it then propagates onto the southeast Bering shelf where large amplitudes are found along the Alaska Peninsula and in Kvichak and Kuskokwim Bays. Another part propagates north- eastward past St. La\snrence Island and into Norton Sound. Over most of the shelf region the tide is mainly semidiurnal, but in Norton Sound the diurnal tides predominate. A number of amphidromic systems are formed on the eastern Bering shelf and in Norton Sound as a result of the interference of tides propagating from various directions. Various authors have different opinions about the details of the tidal height and current distributions, including the exist- ence, shape, and location of amphidromic systems. Some of the controversy can be resolved with the recently acquired data; the rest will require more observations and/or complete models. NUMERICAL MODELS Several numerical models have been applied to the tides in the Bering Sea. They are of two basic types: vertically integrated models which simulate horizon- tal distributions and three-dimensional models which simulate vertical variations as well. Some of these models are still under development. The vertically integrated models by Hastings (1976) and Siinderman (1977) superimpose uniform- ly spaced grids over the Bering Sea Shelf and the entire Bering Sea, respectively. Finite difference approximations to the dynamic equations with quadratic bottom friction are solved over the grids as initial value problems in time with lateral friction included to stabilize the calculation. The tides enter the models as boundary conditions along the open boundaries; time series of sea level at the open boundaries drive the motion in the interior with zero flux and no-slip conditions imposed along coasts. The imposed time series are derived from harmonic constants interpolated from observed values at islands and on coasts. After integrating the models through an initial transient, the motion can be analyzed for tides and tidal currents. Since Hastings (1976) does not take this step, we show only results of the model by Siinderman (1977) , which is limited to the Mj semidiurnal constituent (Fig. 8-2) and its consequences in time-averaged properties and higher harmonics. A distinctly different, vertically integrated model is under development by Preisendorfer (1979), who uses Bristol Bay to illustrate a new technique. The technique involves computing the linear response of a region to a long wave of a given frequency, such as a tidal constituent, as a synthesis of responses to simple subregions. Since realistic boundary conditions were not used in the Bristol Bay example, the results of the calculation may be considered pre- liminary. Further work on the technique is planned (Preisendorfer, personal communication). A three-dimensional model has been developed by Leendertse and Liu (1977) and Liu and Leen- dertse (1978, 1979) to predict tides and wind-driven currents on the southeast Bering shelf for the predic- tion of oil spill trajectories and for risk analysis. The grid for the model (Fig. 8-3a) is uniform in the horizontal dimensions but packed in the vertical dimension to allow higher resolution of the pycno- 114 Physical oceanography :^2^ 50 cm/sec Figure 8-2. Charts of (a) coamplitude (cm), (b) cophase (Greenwich lag in degrees), and (c) tidal current ellipses (cm/sec; radial line within each ellipse represents the tidal velocity at Greenwich transit) from the vertically integrated model by Siinderman (1977) of the M, tide in the Bering Sea. For the finite-difference model: the grid size is 75 km; the time step is 223.5 sec; the bottom drag coefficient is 0.003; and the lateral viscosity is lO' cm^ /sec. The numbers appearing in boxes are observed values. (Repro- duced with permission from J. Siinderman, 1977, Deut- sche Hydrographische Zeitschrift, 91-101, Figs. 3, 4, 5). cline. The dynamic variables are the horizontal and vertical velocity components, temperature, salinity, density, pressure, and the energy density at sub- grid scales; a passive contaminant can also be in- cluded in the model. The large-scale (resolved by the grid) variables satisfy a relatively complete set of nonlinear dynamic equations averaged over each vertical layer and specialized to the extent that the hydrostatic equation is used in the vertical direction. The subgrid scale turbulence satisfies a one-equation closure model in the turbulent energy density. Within the dynamic equations are viscous and diffusive terms. The associated viscosities and diffu- sivities in the horizontal direction have contributions from the large-scale motions through the local vorti- city gradient. The contributions of the subgrid tur- bulence for each direction depend upon a turbulent Richardson number which includes the local den- sity gradient and the local subgrid energy density. The formulas for the turbulent contributions are modified from Mamayev (1958) and include empir- ical coefficients chosen to produce reasonable behavior of the turbulence in the presence of stratifi- cation. Liu (personal communication) indicates that newer versions of the model under development will use a different formulation for the turbulent viscos- ities and diffusivities which eliminates many of the empirical coefficients. The motion in the lowest layer is subject to bot- tom boundary conditions. The bottom is assumed to be impervious and insulating so that zero flux of heat, salt, and water occurs through the bottom. The bottom boundary condition for momentum is a quadratic friction law in which the Chezy coefficient is adjusted to match model currents with observed currents. At the free surface, a uniform wind stress can be imposed over the region to produce wind- driven currents in addition to the tidal motion. The model is integrated as an initial value problem with predicted tide-level time series specified along the open boundaries; these time series are obtained through interpolation of harmonic constants derived from bottom pressure data. To simulate tides and tidal currents on the south- east Bering shelf, Liu and Leendertse (1978, 1979) chose a period, 16-18 June 1976, for which extensive observations existed, some collected specifically in support of this modelling effort. The model was run for a total of 63 hours, which included an initial transient preceding the comparison period. Time series of sea level were then Fourier analyzed for composite semidaily and daily tides from which cotidal charts (Fig. 8-4a,b) were constructed. A more 170" 165' 160' 60* 55-i %. _^--^-' 'C^'._\. "^£^«'^i>p!^^ ■ ^^■>^c--^-^.- „M., Figure 8-3a, b. (a) The horizontal grid and (b) a section showing the vertical grid of the three-dimensional, turbulent energy model of the south- eastern Bering Shelf by Liu and Leendertse (1978, 1979). The hori- zontal grid spacing is 21.82 km; the vertical spacing in this 14-layer model is variable to allow high resolution in the pycnocline. The time step for the model is 180 sec, and the Chezy coefficient of bottom drag is 700 (cm'/2/sec). (Reproduced with per- mission from S. K. Liu and J. J. Leen- dertse 1979). 0- 10- 20- 30- 40- W £ 50- H UJ 5 60- z - 70- I £ 80- o 90- 100- 110- 120- 130- B 5 10 15 20 25 30 1 2 3 4 5 6 7 8 9 10 11 « £E 12 m > < -1 13 14 15 /^5S> • • ",•,*•'.'• • « • •.•-•,*,* •*• • •,•,'•*•'•*• • • •."•'•'•" /.' • •.••*•*•'•*•*• ■ •,*.*•*•'•*• , 1 , , — ^ /.*. •''•/X'M'-'-*'''','*'' * ~ ~ -—"* 1 ~ ~ M 5^':-:-x ■«••'•'•'•'•/*" ' /< '^y/y//////// ' : 55 /v:*:v::SvXv>>'* 0^ t N=19 K^ / J \ ^ t- / ^^ ( lO CM O O O O U rH CD 00 1-1 T-H O O O O O o<<<< <:<:o<: t- '^ a> o eg th cm C» C- Tji CO o CO -^ LO lO lO tH T-H C- i-H 00 C- CC CD 00 i-H ■^ o a> CT> CO CD T-i oo' lo" C-" CM CM 1-1 1-1 CM lO -^ CO LO CM 1-1 1-1 CO CM O lO LO O LO 00 C^ O CD 1-1 lO O CM 00 o ^ CM t-_ CM 00 ^_ o" oo' 1-i o o CM 1-1 CM CM CM r- lO 1— I CO lO o" CTJ t-" CD O CT3 CD lO CO LO CO o 1—1 1— I rH iH CM Tfi C- CD 1— I iH lo' Co' oo' rH 1-J CO iH CM CM CM CO 00 1— I 1-1 CM O 05 O 1-1 CM 1-1 CM 00 CM CO 1-1 CD 00 O 00 O^ lO ^ T-I t~ iH CO 1-H lo iH -^ a> ■^' !>•' CM lO ■^' O 05 LO lr-_ ^' O o" CD CD lO lO i-l lO LO 00 CO CO 1—1 CO ■rr CO CM -^ c^ o 00 ^ CO tH CM 00 lo CO o CD t--' CO 00 Pi O O o o o O o o O O <: a u O O o < < < < 0 < 0 < 00 CD CM q CO 00 1-1 00 00 q tH c- q lO t- q q q C— 1—1 ^ lO lO ^_ .s X 1-1 d cm' ^' t~' 00 lO ^' lO CM d CM oo' CM oo' i-i oo' 1— I 1—1 1— i d d d CM s Q CO o in o LO 1—1 CD 05 CD 05 o 1—1 o^ CD 1—1 00 CD 00 CD 00 CD I-l CD CD 00 CO o tH 00 CM 10 0 CO 1—1 0 CM 1—1 0 00 tH CD CM 1-1 0 0 05 o O CD CM 1—1 1-1 oo CM lO 00 CM o CD o 00 1-1 o> CO CD 00 o> 00 CD CD 00 00 1—1 o CM 1—1 00 CO CM 00 CO CM 1—1 05 00 00 in 00 CM C- IT- CD CO 1—1 1—1 s ^. 05 co_ ^. CD o> q 00_ 00 q CM lO q q ^. 1—1 q q q q c-_ q q q X CD ^' lo' CD oi '^t^' IT-' d d LO cm' 1—1 ^' oo' d d ^' lO ^ ^' I— 1 cm' cm' i-I 00 Di o O o o O o o < < < < o o o o O < < 0 0 0 0 0 < O iH 00 CO CM_ 00 00 1—1 LO q q q q 1—1 ^. q CM t— c— c- q t- q q c X 1—1 CM cm" '^ CD' ^' cm' d l-i 1—1 d Tj^' ^' '^ -*' oo' I-l i-i d i-i d LO' d -^' a Q tr- io 00 00 CD lO o 1—1 CO CO 1-1 1-1 00 1-1 oo 1-1 CD 1-1 I-l 00 o r-l 1-1 00 O 00 lO 00 o CO LO oo LO CO 00 oo CO 1-1 iH tH 00 Oi 0 00 0 tT- 05 CD o O CJ5 CM iH 00 CM 03 00 CM CD CM CO o lO OO o CT3 in CO CD in 00 00 Csl OO o CO CM 1-1 CO 1-1 CM 1—1 CM I— 1 CO CM CO CD CM IT- CM 0:1 CM 0 tH 00 05 CM CM CM CM CM CM CM r- LO CM CM s 00 '^ o 1—1 C-; o 1—1 o 1-1 00 CM 1-1 o 00 1-1 oo 1—1 1—1 CM o 1-1 c- 00 CD o CO CM lO 00 rH lO OO 00 lO CO in oo CO era 1-1 CT5 00 0 CM CO CD 01 CM 0 r- CO CD Oi CD %.4 o O lO CD CM CM CM CM 1-1 CO o CO 00 1—1 00 CO CM CO 00 oo 00 oo 00 o CO CM O^ CM CM CM CD 00 1-1 00 00 1-1 t- 00 1-1 CM 1-1 CM 05 00 1-1 1-1 CO CM 1-1 IT- CM CM CM CM 00 1-1 10 1—1 00 C- lO CM 00 00 00 « 05 00 CD Lrt ^. 00 LO C-; c- CD q q o q "*. Oi 00 q q Oi in q q E X C-' C-' lo' cd' t-' c~' 00 cm' 1—1 C-" 00 co' 1—1 ^' oo' C-' oo' CM d 1-1 00' ^' 1—1 00' 00' 1—1 ■^' d 1—1 o LO 1-1 o LO 1-1 00 lO I— 1 CD CD CM 1-1 lO CM 1-1 lO CM CO 00 1-1 CM OO 1—1 CO CM 1—1 O 1-1 CO CM lO 1—1 CM o CM CM o CO CM o CM 00 o CM CO o CM 0 CM LO CO iH 00 0 CM 0 CM in 0 CO en 00 1-1 ^ CD 00 0 CM CM CD CD CD CD lO lO c- t- c- c- c- lO c- CD c- 00 00 c- 00 c- 00 IT- 00 IT- 00 00 00 00 00 c- CD 00 o I-l o 1—1 OO 1—1 1-1 co o CD 1-1 CM CM o lO lO o o 1-1 o T-I CD CM CD CM c- 0 LO 0 LO 00 0 0 0 10 05 CM CM 0 CM 0 \3uoq lO CD CD in CD 1-1 CD 1-1 00 CD 1—1 00 CD iH CD 1-1 t- CD 1-1 lO CD iH 00 CD 1—1 1-1 CO 1—1 1-1 iH 05 CO 1—1 05 CO iH o c- tH O C- iH CD CD 1—1 IT- CO 1—1 00 CO 1—1 00 CO 1—1 CD CD 1—1 10 CD 1-1 LO CO CD CD 1—1 tH CM O CM O o CO CO OO IT- CO CM 1-1 o 05 lO O 00 o CD CM CM CM 00 00 0 00 00 10 00 tH 0 0 c- 0 0 CM 00 CM 0 N?Bi lO lO lO lO in CD LO 00 lO 00 LO o> LO in LO CD LO o CD o CD O CD 1-1 CD 1-1 CO CD CM CD 00 CD CD CD CD CD CD mdaa o CM o o 1—1 o o 1-1 CD OS o 00 00 00 o CM IT- 00 o lO O CM 1-1 lO o CM CO oo 1-1 05 oo 0 1—1 c- 1-n' CM 00 00 CD 1—1 m iH 0 CM CM q?daa I-l 1-1 iH lO 1-1 O 1-1 1-1 LO lO 1—1 1-1 00 o LO m CD 1-1 lO CD CM CM 00 00 1—1 CD CO 00 CM 00 0 CM 05 1— 1 CM IT- CO CM lion B^S 00 O CQ CO o CQ CQ CO 1-1 O m r-i O PQ O m 00 1—1 O PQ CD I-l o pq CM O CO iH O pq O CM O m 1—1 CM o PQ 1— 1 CM O PQ CM O '^ IT- CM 1-1 tH CM oo O Q O Q Q ^ J Z J J 0 CM 2 I-l 10 ^ 0 c m 0 0 CO < ^ 3 c U3 n S-1 0 en ^ -^ a> 0 U3 ^ 0) CJ ^ 0 M 0 0 M 0 c n () ■4^ CS n n) XI U3 a. 0> (an n) 1 S o o - CJ o — 05 — 'O >< "2 CO rt >-( a o 01 CO ,4_3 £ - 0.2 .ili O c -a .. o o'-g -3 (DO) X! O ^ 5^ ^ O ,=* 125 126 Physical oceanography region of Bristol Bay (stations BC16, BC2, BC14) and in the approaches to Norton Sound (stations LDl, LD2, NC17). Typical amplitudes over the southeast shelf axe 10-20 cm/sec. Largest Ki currents are found within Norton Sound where amplitudes range from 20 to over 30 cm/sec. West of Norton Sound amplitudes are very small. Ellipses for the smaller N2 and Oi constituents have not been plotted. Generally they are similar to the M2 and Ki respectively. Oi major axis amplitudes are 60-75 percent of Ki in the south- eastern Bering Sea and 40-60 percent in the Norton Sound region. N2 is about 25-40 percent of M2 throughout the Bering Sea. DISCUSSION Both models and observations show that the largest tides in the Bering Sea occur on the inner southeast Bering shelf. As shown in the figures, there is general agreement that the tides propagate as free waves onto the southeast shelf from the deep basin. In the vicinity of the Alaska Peninsula, the exponential decay of amplitude northward from shore and a com- parison of the tidal heights and currents at BC14 indicate that the semidiurnal and diurnal tides are Kelvin^ waves near the peninsula, the semidiurnal wave is progressive as at BC2, where the M2 phase difference is 12°, and at BC14, where the current is nearly in phase with the tide. Conversely, the Ki tide has a much larger standing wave component, with a phase difference of 64° at BC2 and about 40° at BC14. When the Kelvin waves propagate into the inner shelf, there is an interesting difference between the reflected waves for the two tidal species. The critical latitudes of the diurnal tides (~30°N) lie south of the Bering Sea; diurnal tides can there- fore obey only Kelvin-wave dynamics on the open shelf. The reflection of each diurnal tidal constituent produces a classic amphidromic system on the south- east shelf with the amphidromic point well offshore. The critical latitudes of the semidiurnal tides (75°N for M2 ) lie north of the Bering Sea. As a ^ Kelvin and Sverdrup waves are long vi^aves under the in- fluence of the earth's rotation. In the northern hemisphere Kelvin waves propagate with the coast on the right and am- plitude decreasing exponentially away from the coast. Current motion is rectilinear (in the absence of friction). Sverdrup waves (also called Poincar^ or inertial-gravity waves) have horizontal wave crests and the currents form a clockwise- rotating ellipse. Sverdrup waves can exist only where the wave period is less than the inertial period. In areas where both Kelvin and Sverdrup waves can exist, the actual wave is some combination of the two. For further discussion, the reader is referred to a text such as The Oceans (Sverdrup et al. 1942). result, semidiurnal tides can obey Sverdrup-wave as well as Kelvin -wave dynamics. The presence of a virtual amphidrome near Cape Newenham for each semidiurnal constituent suggests that a major portion (if not all) of the incident, semidiurnal Kelvin waves are dissipated or converted on reflection into Sverdrup waves. Perhaps the acute apex angle of Kvichak Bay, together with the presence of sharply protruding peninsulas, may produce an efficient conversion to semidiurnal Sverdrup waves. A large part of the semidiurnal tidal energy may also be dissipated over the extensive mud flats and shoals of the Kvichak and Nushagak Bays and Rivers. On the open shelf, the broad semidiurnal tidal ellipses have approximately 45-50° phases relative to the tidal heights in the direction of propagation. This fact and the fact that the offshore cophase lines are roughly parallel to the coast and widely separated indicate that the semidiurnal tides act as a standing Sverdrup (Poincare) wave in this region due to cooscillation in Kuskokwim Bay. Further cooscillation is evident at BCll, just southwest of Nunivak Island, where the M2 tide actually leads that at stations farther sea- ward. The Coast Pilot, published by the U.S. Department of Commerce (1964), notes that to the north of the southeast Bering shelf the currents in Etolin Strait between Nunivak Island and the mainland are suf- ficiently strong to prevent ice formation in winter. This may be due to tidal currents associated with the great differences in tidal phases between the north and south sides of Nunivak Island. There is much less consensus about the tides in the north Bering Sea. The tides around St. Lawrence Island are particularly confusing. Harris (1904) shows the tide progressing from west to east along the south shore and then east to west along the north shore, so that all phases of the tide are found around the island. Leonov (1960) has branches of the tide progressing through the passes on the west and east sides of the island and meeting on the north side together with the tide from the Arctic Ocean; he discusses areas of convergence and divergence and abrupt changes in velocity as a result of the meeting of the tides of the Pacific and Arctic Oceans. Conversely, the Office of Climatology and Oceano- graphic Analysis Division (1961) chart shows very little effect on the tide from St. Lawrence Island with a fairly smooth progression of the tide north- ward across the shelf and through the Bering Strait. Coachman et al. (1975) inferred the presence of an amphidrome south of St. Lawrence Island from tidal differences of stations listed in the tide tables. According to Siinderman's (1977) charts (Fig. Tides 127 8-2), cophase lines converge on Southeast Cape, and there is probably cooscillation in the bight on the south side of the island with a reversal of phase progression. This may be responsible for the large phase differences noted by Coachman et al. (1975). The tide appears to progress through the pass be- tween Northwest Cape and Siberia and thence west to east along the north shore of St. Lawrence Island to the vicinity of Northeast Cape, where it meets the wave progressing north along the Alaskan mainland. This results in nearly 180° phase difference between the Mj currents at LDl, near the Yukon Delta, and at LD2 near Northeast Cape. That is, current is northerly at LDl when it is southerly at LD2. Station NC17, between these two stations, is in a transition zone; most motion is cross-channel. Coachman et al. (1975) discussed measurements made at a current-meter mooring 90 km south of the Bering Strait. Tidal currents were mainly semi- diurnal with amplitudes of 1-12 cm/sec and the current ellipse orientation was generally northeast- southwest to north-south. In the Bering Strait, Ratmanoff (1937) and Fleming and Heggarty (1966) did not find noticeable tidal currents while Bloom (1964) and Coachman and Aagaard (1966) found that tidal currents modulated the net northward flow. Semidiurnal currents are especially small in Norton Sound south of Nome, for example at stations NCI 4 and NC20. These stations are in an antinode one- half wavelength from the head of the bay. Generally, semidiurnal tides and tidal currents are small through- out the Norton Sound and Bering Strait region, perhaps as a result of frictional dissipation across the broad Bering Sea shelf and Kelvin-wave dynamics (i.e., the Norton Sound amphidrome and small amplitudes to the west of a northward-progressing wave). The diurnal tides appear to be simpler than the semidiurnal tides because they are restricted to Kelvin-wave dynamics and have longer wavelengths. Tidal height/tidal current phase relationships indicate predominantly standing-wave characteristics in south- east Bering Sea and Norton Sound although between St. Lawrence Island and the Alaskan mainland the tide is more progressive. Thus the diurnal tide in the deep western Bering Sea basin cooscillates with Bristol Bay and Norton Sound. Currents are highest in Norton Sound because of the shallower depth, even though the diurnal tides are higher at the head of Bristol Bay. Between St. Lawrence Island and the Bering Strait, the diurnal tides virtually disappear, again due to Kelvin -wave dynamics and dissipation. While there is good agreement between the models and the observations on the major features of the tides of the eastern Bering Sea shelf, there are some differences. For instance, the M2 cotidal chart of Siinderman (1977) has systematically higher ampli- tudes and later phases than the observations on the southeast shelf, but lower amplitudes in Kvichak Bay. Also, the actual amphidrome in Norton Sound appears to be further ashore. These differences may be due to the coarse grid (75 km) and high horizontal eddy viscosity (10^ cm^ /sec) used in the model. It is more difficult to evaluate the Liu and Leendertse (1978, 1979) model (Figs. 8-3-5) since many of the observations were used for boundary conditions and for tuning empirical parameters. The model and observations Eire therefore not independent. The cotidal charts (Fig. 8-4) present composite tides based on a Fourier analysis of 50 hours of computed data; therefore the cotidal cheirts represent sums of constituents within each species, with an arbitrary composite phase rather than one referred to Greenwich. For these reasons, no attempt will be made here to assess the accuracy of this model. APPENDIX Moorings usually consisted of one or two Aanderaa RCM-4 current meters on a taut wire, with the upper current meter at a depth of about 20 m, just below the subsurface float; the deeper one was below the pycnocline, about 10 m above the bottom. In the Norton Sound area, where depths are generally less than 30 m, moorings contained only one meter. The Aanderaa TG2 or TG3 pressure gauge, if present, either was placed in a well within the anchor or was attached to acoustic release, just above the anchor. Pressure-gauge resolution is claimed by the man- ufacturer to be better than 0.5 cm. Individual moorings were in place for periods of up to one year, with observations of some locations spanning as much as three years. Data were processed by methods similar to those described in Chamell and Krancus (1976). No correction was made for atmospheric pressure in the pressure-gage data. In Aanderaa current meters, the recorded speed is an average over the data interval (15, 20, 30, 40, or 60 minutes) while direction is essentially instantaneous. To remove the phase difference between speed and direction, speeds at times Tn and T„+i were averaged to give the speed corresponding to the direction at Tn before converting to east and north components of velocity. The data were low-pass filtered to remove noise (i.e., energy at frequencies >0.5 cycle/hr) and resampled at hourly intervals. A second-order polynomial was then used to interpolate to even hours. 128 Physical oceanography The Munk-Cartwright response method (Munk and Cartwright 1966) was used for tidal height analysis using procedures based on those suggested by Cartwright et al. (1969). The pressure record from Station BC20 was selected as a reference for all other stations because of its length (300 days) and location (near the center of the study area), and because the tide there is thought to be representative of that entering the shelf from the deep basin of the Bering Sea. The tide potential was used as a reference for the BC20 pressure record, using weights of 0, ±2, and ±4 days for the diurnal and semidiurnal bands. For the other stations the reference series was a complex prediction based on the sixteen largest diurnal and semidiurnal harmonic constituents derived from the BC20 analysis using weights of 0, ±2 days. Results of high and low water analyses for many locations in the Bering Sea were obtained from the National Ocean Survey. High and low water analysis gives the mean tide range and Greenwich high and low water luni tidal intervals HWI and LWI. Ac- cording to Schureman (1958), if the tide is semi- diurnal, the M2 amplitude can be estimated by multiplying the mean range by 0.47. The phase may be found by: M°2 = V2(HWI + LWI) X 28.984 + 90°. For current-meter data, a 29-day harmonic analysis based on Schureman (1958) was used. Constituents Oi , Ki , N2 , M2 , and S2 are derived directly, and other constituents are inferred from these on the basis of equilibrium relationships. The harmonic method was used for currents because uncertainties in data quality and seasonal variations obviated the use of the slightly more accurate response method. The results of these analyses are given in Table 8-1 (for current-meter data, the four major con- stituents in an ellipse representation) and in Table 8-2 TABLE 8-2 Results of response analyses for the new pressure-gage observations' Station Lat, N Long, W 0 I Pi Ki N2 M2 S 2 Start Date Length Days H G H G H G H G H G H G Yr JD BC20 60 26 171 05 9.8 310 5.8 323 18.1 326 6.9 121 20.5 171 2.2 249 77 260 300 BC3 55 01 165 10 26.4 304 13.3 317 40.9 319 15.8 40 41.9 89 3.2 2 76 077 73 BC13B 55 30 165 49 23.8 311 11.4 322 34.4 325 13.2 52 35.5 106 1.8 327 76 158 114 BC13D 55 47 165 23 23.0 312 11.0 324 33.4 327 15.0 56 39.0 109 1.4 314 77 252 131 BCIO 57 17 169 33 17.3 319 8.2 330 24.9 333 8.7 77 24.9 131 1.5 266 76 153 101 BC4 58 37 168 14 6.3 305 3.9 300 12.4 303 11.1 98 33.4 151 1.8 252 75 250 58 FX2 58 32 167 56 4.0 286 1.8 284 8.9 288 11.8 104 33.8 158 1.9 251 78 200 64 BC9 59 13 167 42 2.4 220 3.1 253 9.7 258 12.4 108 36.7 164 2.0 246 76 269 230 BCll 59 42 167 15 10.6 165 6.0 204 18.3 207 11.9 98 35.9 155 2.8 224 76 154 102 BC21 60 23 169 11 7.4 271 5.3 294 16.6 297 10.1 136 30.9 189 2.9 265 77 260 246 BC7 55 42 163 01 31.3 318 16.0 332 49.0 335 24.5 81 71.4 134 1.4 45 76 080 70 BC2 57 04 163 22 19.0 358 9.3 11 28.3 13 14.7 102 45.2 157 0.5 343 76 151 200 BC15 57 39 162 42 21.7 25 9.9 45 29.9 48 11.1 116 36.2 168 1.3 257 77 257 129 LDl 62 30 166 07 17.9 286 10.3 319 31.7 324 13.2 274 46.1 328 6.8 49 78 204 54 NC17 62 53 167 05 10.7 303 6.2 336 19.0 341 7.4 274 25.6 330 3.0 121 77 263 293 NC18 63 09 168 23 4.4 339 2.9 351 8.9 356 4.7 272 22.4 324 2.6 59 76 245 120 LD2 63 13 168 35 4.6 338 2.6 356 8.1 359 6.2 261 26.6 319 5.3 58 78 203 54 LD4 64 47 166 50 2.3 75 1.1 218 4.2 222 2.3 47 4.9 138 0.6 324 78 205 55 NCIO 65 45 168 27 0.5 228 0.6 277 2.7 296 2.5 147 12.0 213 3.5 284 77 202 26 GEO 64 00 165 30 9.5 23 4.8 66 14.0 71 4.4 349 13.0 44 2.5 131 77 189 60 PROBE LD5 64 08 163 00 16.9 60 10.4 106 32.1 110 LO 186 2.0 233 0.2 277 78 206 41 UN- 65 53 160 47 24.0 55 13.9 102 42.9 108 5.2 164 17.3 222 3.0 326 77 220 95 ALAKLEET ' Amplitude H in mbar of pressure. Greenwich. 1 mbar equals 1.007±.003 cm of sea water in the Bering Sea. Phases G are referred to Tides 129 (for pressure gauges, the six major constituents). Harmonic constant amplitudes H are cm /sec for currents and mbar for pressure-gauge data. For this region, 1 mbar equals 1.007±.003 cm of sea water. Phases are referred to Greenwich. ACKNOWLEDGMENTS This chapter is contribution no. 431 from the NOAA/ERL Pacific Marine Environmental Lab- oratory. The work was supported in part by the Bureau of Land Management through interagency agreement with the National Oceanic and Atmos- pheric Administration, under which a multiyear program responding to needs of petroleum develop- ment of the Alaskan Continental Shelf is managed by the Outer Continental Shelf Environmental Assessment Program (OCSEAP) Office; and in part by NOAA's Environmental Research Laboratories. The authors wish to express their appreciation to the following: L. Long and S. Wright of PMEL for data processing; B. Zetler of IGPP, S. K. Liu of Rand Corp., K. Aagaard and A. Clarke of the University of Washington, J. Schumacher, J. Larsen, and R. Preisendorfer of PMEL, T. Kinder of NORDA, and the anonymous reviewers for helpful suggestions; J. Fancher of the National Ocean Survey, D. Cacchione of the U. S. Geological Survey, and W. J. Ingraham, Jr. of the National Marine Fisheries Service for providing data; and J. Golly for drafting. Bogdanov, K. T., K. V. Kim, and V. A. Magarik 1964 Numerical solutions of tide hydro- dynamic equations by means of BESM-2 electronic computer for the Pacific Area. Trudy Inst. Okeanol. Akad. Nauk SSSR 75: 73-98. Cartwright, D. 1979 Tidal theory. In: Symposium on long waves in the ocean, Man. Rep. Series 53: 35-8. Marine Sciences Dir., Dep. of Fish, and Environ., Ottawa. Cartwright, D., W. Munk, and B. Zetler 1969 Pelagic tidal measurements. Trans. Amer. Geophys. Union 50: 472-77. Chamell, R. L., and G. A. Krancus 1976 A processing system for Aanderaa current meter data. NOAA Tech. Mem. ERL PMEL-6. Coachman, L. K., and K. Aagaard 1966 On the water exchange through Bering Strait. Limnol. and Oceanog. 11: 44-59. Coachman, L. K., K. Aagaard, and R. B. Tripp 1975 Bering Strait: The regional physical oceanography, Univ. of Washington Press, Seattle. Defant, A. 1961 Physical oceanography, II, Pergamon Press, Oxford. REFERENCES Bloom, G. L. 1964 Water transport and temperature measurements in the eastern Bering Strait, 1953-1958. J. Geophys. Res. 69: 3335-54. Bogdanov, K. T. 1961 New charts of the cotidal lines of semi-diurnal tidal waves (M2 and S2 ) for the Pacific Ocean. Sov. Oceanog. 1: 28-31. Fleet Weather Facility 1977-1978 Eastern-Western Arctic Sea ice analysis, Suitland, Md. Fleming, R. H., and D. Heggarty 1966 Oceanography of the southeastern Chukchi Sea. In: Environment of the Cape Thompson Region, Alaska., U. S. Atomic Energy Comm., Div. of Tech. Information, 697-754. Goodman, J. R., J. H. Lincoln, T. G. Thompson, and F. A. Zeusler 1942 Physical and chemical investigations: Bering Sea, Bering Strait, Chukchi Sea during the summers of 1937 and 1938. Univ. of Washington Pub. Oceanog. 3. 105-69. 130 Physical oceanography Harris, J. R. 1904 Manual of tides, Part IV. Appendix 5. Rep. of Superintendent U. S. Coast and Geodetic Survey, Washington, D. C. 394-5. Hastings, J. R. 1976 A single-layer hydrodynamical- numerical model of the eastern Bering Sea shelf. Mar. Sci. Comm. 2: 335-56. International Hydrographic Bureau 1966 Tides, harmonic constants. Spec. Pub. No. 26, Monaco. Jeffreys, H. 1921 Tidal friction in shallow seas. Phil. Trans. Roy. Soc. London, Ser. A, 221: 237-64. Leendertse, J. J., and S. K. Liu 1977 A three-dimensional model for es- tuaries and coastal Seas: IV, Turbulent energy computation. The Rand Corp. R-2187-OWRT. Leonov, A. K. 1960 Regional oceanography. Part I., Lenin- grad, (transl.) Liu, S. K. and J. J. Leendertse 1978 Three-dimensional subgridscale-energy model of eastern Bering Sea. Proc. XVI Coast. Eng., Conf. Amer. Soc. Civil Eng. Munk, W. H., and D. Cartwright 1966 Tidal spectroscopy and prediction. Phil. Trans. Roy. Soc. London, Ser. A. 259: 533-81. Munk, W. H., and G. J. F. MacDonald 1960 The rotation of the Earth. Cambridge Univ. Press. Office of Climatology and Oceanographic Analysis Division 1961 Climatological and oceanographic atlas for mariners, 2, N. Pacific Ocean. Preisendorfer, R. W. 1979 A transport formulation of the tsunamic propagation problem. Marine Geodesy 2: 67-82. Ratmanoff, G. E. 1937 On water interexchange in Bering Strait. Explorations of Seas of USSR. Hydro. Pub. Leningrad and Moscow, 119-33. (transl.) Schureman, P. 1958 Manual of harmonic analysis and prediction of tides. Coast and Geodetic Survey, U. S. Dep. Comm. Washington, D. C. Siinderman, J. 1977 The semidiurnal principal lunar tide M2 in the Bering Sea. Deutsche Hydrog. Zeitschrift 30: 91-101. 1979 A three-dimensional model for es- tuaries and coastal seas: VI, Bristol Bay simulations. The Rand Corp. R-2405-NOAA. Mamayev, O. I. 1958 The influence of stratification on vertical turbulent mixing in the sea. Izv. Geophys. Ser.: 870-5. Sverdrup, H. U., M. W. Johnson, and R. H. Fleming 1942 The oceans. Prentice-HaU, Englewood Cliffs, N. J. United States Department of Commerce 1964 United States Coast Pilot 9: Pacific and Arctic coasts Alaska, Cape Spender to. Beaufort Sea. U. S. Gov. Printing Office, Washington, D.C. Section II I Ice Distribution and Dynamics William J. Stringer, editor Recent Fluctuations in Sea Ice Distribution in the Eastern Bering Sea H. J. Niebauer Institute of Marine Science University of Alaska Fairbanks, Alaska ABSTRACT The eastern Bering Sea shelf (-1000 km long X 500 km wide) is ice covered in winter but ice free in summer. Time series of weekly percent ice coverage are presented, illustrating details of this phenomenon for the period 1973-79. Advances and retreats of the ice edge are correlated with fluctuations in sea and air temperatures, with surface winds, and with regional meteorological events. The period 1973-79 is shown to be a time of extreme fluctuations with 1973-76 character- ized by below-normal temperatures and above-normal ice cover under northerly winds, while 1976-79 was a period of strong rise in temperatures and retreat of the ice pack under wdnds shifting to southerly. INTRODUCTION The southeast Bering Sea shelf is a relatively shallov^r (shelf break ~150 m) but wide (~500 km) region (Fig. 9-1) that is seasonally covered with ice. During a typical winter, the ice advances about 1,000 km south from the Bering Strait to the shelf break (Fig. 9-1). This advance occurs primarily by freez- ing of seawater within the Bering Sea (Leonov 1960) and does not imply advective advance through the Bering Strait (Tabata 1974). In spring, about 63 percent of the ice melts within the basin (Lisitsyn 1960), and the remainder leaves by way of the various passes and straits. In addition to the strong seasonal cycle in ice coverage, large multiyear variations have been ob- served in the Bering Sea. Walsh and Johnson (1979), who considered the extent of sea ice in the Northern Hemisphere for the years 1953-77, show year-to-year ice fluctuations exceeding 5° latitude (~340 km) in nearly all the seas surrounding the Arctic Ocean. However, they show that the standard deviation of the departure from the monthly mean extent of sea ice is greater in the Bering Sea along 170°W (Fig. 9-1) than in the other high latitude seas. As an example of one of these fluctuations, Kukla and Kukla (1974) showed that the onset of the decline of sea-surface temperatures (SST) in the early 1970's on the Bering Sea shelf coincided with anomalous southward penetration of the ice pack. McLain and Favorite (1976) related the fall in SST during this period to changes in the atmospheric cir- culation which caused northerly winds over the Bering Sea. Niebauer (1978) has related a subsequent rise in SST in the late 1970's to, again, cheinges in the atmospheric circulation which caused southerly air flow over the region. This chapter analyzes the remarkable decrease in iciness as related to this recent short-term climatic fluctuation and rise in SST reported by Niebauer (1980; Chapter 3, this volume). On similair time scales, Sater et al. (1974) and Rodgers (1978) have shown that ice conditions in the Beaufort Sea are correlated with meteorological conditions such as sea-level pressure and wind direc- tion. Rodgers has related light ice summers to more frequent southerly surface winds; the winds are reversed in heavy ice summers. For a recent bibliog- raphy of authors who have described interannual sea-ice fluctuations of several hundred kilometers in other high latitude seas surrounding the Arctic Ocean, see Walsh and Johnson (1979). The following sections present a description and analysis of the seasonal ice cycle derived from weekly mean percent ice cover for the eastern Bering Sea for 1973-79. The multiyear variations are then discussed in relation to short-term climatic variations and to fluctuations in SST, air temperatures, and surface winds. 133 134 Ice distribution and dynamics U.S.S.R. ALASKA 2000 m "^ — -^ Ateutian PACIFIC OCEAN I Shelf break \ 170' 180" 170' 160' Figure 9-1. Chart of the Bering Sea showing the bathymetry of the eastern shelf. Examples of the southern ice limit for April 1976 and 1979 are indicated. Percent ice cover calculations were made as outUned in the text using the area blocked out as indicated over the eastern Bering Sea (after Niebauer 1980). DATA Weekly mean southern ice limit data for the Bering Sea were obtained from the Naval Fleet Weather Facilities in Suitland, Maryland. The estimates represent both satellite and visual observations. The weekly percentage of ice coverage was calculated as the ratio of ice coverage to total area (ice plus open water) considered in Fig. 9-1. This gives a quanti- tative estimate of ice coverage (hereafter called ice) but does not take into account ice thickness or concentrations (e.g., 1/8 concentration is weighted equally with 8/8 concentrations). Monthly mean SSTs were obtained from the Naval Fleet Numerical Weather Central, Monterey, CaUfor- nia, through D. R. McLain of National Marine Fish- eries Service in Monterey. The data represent anal- yses of temperature reports made by ships of oppor- Recent fluctuations in sea ice distribution 135 A A \ /^ ^^^ 1973-74 - y\/^ llllllllm ■iliiiiiiliiii. ^■-■^■-■■f ■■'■'V' — -^- — r""""-- n —* T"""' • '' 1 ...... Tii if 1 \ 1 1973-78 mean Oct Nov Dec Jan Feb Mar Apr May Jun Ju Feb Mar Apr May Jun Jui 1973-78 mean ,-' \„/ .1973-78 mean 197?-78 Apr May Jun Ju F .■>fx ililfcfe. ■ M i^^M iiiiiiil ,1973-78 nean ■ ^f r-^.f 'Y;I9?8~79' ■ftl. (missing dat9^ r--^^ 1 ■\ |.--^ \ Oct Nov Dec Jan Feb Mar Apr May Jun Jul Figure 9-2, a-f Percent ice cover as calculated from the area indicated in Fig. 9-1 for the eastern Bering Sea for the winters 1973-79. The shaded area is the six-year mean (seasonal cycle) while the solid line is the individual winter data (redrawn after Niebauer 1980). tunity in the area. Bottom temperatures were ob- tained from Coachman and Charnell (1979). North- ern Hemisphere 700 mb pressure chairts were ob- tained from Monthly Weather Review. Air temp- eratures and surface winds for the Pribilof Islands were taken from the Local Climatological Data published by the U.S. Department of Commerce. RESULTS Seasonal ice cycle Fig. 9-2 (a-f) illustrates the mean seasonal cycle of ice for the eastern Bering Sea. The ice generally begins its seasonal southward formation in November. It is estimated that about 97 percent of the ice in the Bering Sea (Leonov 1960) is formed within the Bering Sea. Very little ice is transported south through the Bering Strait (Tabata 1974). The ice apparently forms like a giant conveyor belt, being generated along the south-facing coasts in the Bering Sea and moving southward at as much as 0.5 m/sec before finally melting at its southern limit (Pease, this volume). Therefore, although the ice on the Bering Sea shelf is subsequently pushed around by the wind on shorter time scales (see, for example, Muench and Ahlnas 1976), only about 3 percent of the seasonal sea ice cover is actually advected onto the shelf through the Bering Strait. 136 Ice distribution and dynamics Seasonal ice formation progresses at an average rate of 12-13 percent per month of the area of the eastern shelf considered in Fig. 9-1, reaching 60-65 percent coverage in late March. The ice advance generally consists of a short, rapid advance (~24 percent per month) in November-December before slowing to ~6- 7 percent per month in December-March. With the exception of the rapid advance in November and part of December, the ice appears to dissipate faster than it forms, at about 18-20 percent per month in late March to early July. Lisitsyn (1960) has reported that during the period of ice retreat, 63 percent of the ice melts within the Bering Sea basin. The remainder leaves the Bering Sea by way of the various straits and passes (Fig. 9-1). Interannual departures from seasonal ice cycles Before we consider year-to-year departures from mean seasonal cycles, it is instructive to look at a longer time series of mean annual SST from the Pribilof Islands (Fig. 9-3). The annual SST was near the 15-year mean of 4.2 C in 1963 before rising to 5.4 C in 1967. SST then fell to 2.8 C, or 1.4 C below normal in 1975. Since then there has been a rapid rise, to 5.7 C, or 1.5 C above normal, in 1978. Sim- ilar data from Bristol Bay (Fig. 9-3) display similar characteristics. Thus, for the period of our interest (i.e., the last five years, 1974-79, of this time series) the mean annual SST of the southeastern Bering Sea shelf water has risen nearly 3 C. Consider now departures from the seasonal means of ice (Fig. 9-2) and annual SST. A maximum in ice coverage and relatively low SST occur in the spring of 1976. For these reasons January 1976 has been arbitrarily chosen as a point at which to divide the time series into two sections. The first section is characterized by lower-than-normal sea temperatures and above-normal ice cover. This period coincides with above-normal upper (3,000 m or 700 mb) air flow from the north as outlined by McLain and Favorite (1976) and Niebauer (1980). An extreme example of above-normal ice cover on the eastern shelf occurred in April of 1976 when the ice cover was ~25 percent above the mean. An example of the relatively strong upper air flow from the north during this period is shown in Fig. 9-4. Here a mean low is situated over eastern Siberia, causing southward flow off the Arctic Ocean down over Siberia and then eastwaird flow over the Bering Sea. Some indication of the severity of the condi- tions is given by the mean air temperature for April for the Pribilof Islands of -8.2 C (6.2 C below normal) under a mean northerly component of the surface winds of 2.2 m/sec (~0.4 m/sec above normal A ^A SE Bering Sea shelf bottom water Mean Annual SST TC " Pribilof Islands Mean Annual SST T^C " Bristol Bay 63 64 65 66 67 68 70 71 YEARS 72 73 74 75 76 77 78 Figure 9-3. Mean annual sea surface temperatures from the area of the Pribilof Islands and Bristol Bay and shelf bottom water temperatures (Coachman and Charnell 1979) for June for the S.E. Bering Sea (after Niebauer 1980). northerly flow). In fact, for the period 1974-76 there were only five months in which there was monthly mean southerly flow and these were all summer months. The second period (~ 1976-79) is characterized by a strong rise in SST (~1 C/yr) and a precipitous fall (~9-10 percent/yr) in ice (Niebauer 1980). These observations coincide and are probably a result of the abrupt swing of the upper air flow from northerly to southerly as reported by Niebauer (Chapter 3, this volume; 1980). Illustrated in Fig. 9-5 is an example of the strong flow from the south which probably caused the extreme below -normal (~35 percent below normal) ice extent in January 1979. This strong southerly flow persisted for most of the 1978-79 winter, resulting in over 40 percent below Recent fluctuations in sea ice distribution 137 Figure 9-4. 700-mb (~33 km) height pressure chart for April 1976 (labeled in dekameters) (after Wagner 1976). normal ice cover by April 1979. Again, some indica- tion of the mildness of conditions in the Pribilof Islands is given by the mean monthly temperature for April 1979 of 2.3 C (~4.3 C above normal) under the 2.3 m/sec southerly component of the surface wind (~3.9 m/sec above normal southerly flow). During this period of January 1976 to May 1979 there were 13 months during which there were mean southerly winds in the Pribilof Islands. Seven of these months were within the November-May ice season. Fig. 9-6 illustrates the regression of the January southern ice extent over the years 1975-79, showing that the ice edge in the eastern Bering Sea has re- treated ~550 km over this period. Table 9-1 shows that the rather dramatic rise in the mean January air temperature in the eastern Bering Sea coincided with the ice retreat and the shifting of the north-south component of the surface wind from north to south. Thus, in the period 1975-79 the mean January air temperature has risen ~9.3 C, coinciding with a northerly to southerly shift in winds of ~ 6.2 m/sec (~12 kn) over 1976-79. As mentioned previously, this coincides with the dramatic retreat of the ice edge (Figs. 9-2 and 9-6) and the rise in SST (Fig. 9-3). TABLE 9-1 January mean air temperatures with deviation from the mean and the monthly mean North-South component of the surface wind with the deviation from the mean from the Pribilof Islands for the period 1975-79 Air Deviation N-S surface Deviation temperature from mean wind from mean January (T) (T°C) (m/sec) (m/sec) 1975 -8.0 -4.7 -1.8 -0.5 1976 5.6 2.3 3.8 2.5 1977 -0.8 2.5 -2.7 -1.4 1978 0.9 4.2 1.8 3.1 1979 1.3 4.6 2.4 3.7 138 Ice distribution and dynamics Figure 9-5. Same as Fig. 9-4 ex- cept for January 1979 (after Wagner 1979). T7^ ^ lvv \_ --X— -oe ^ «^ -- — " ^ V^" — \^^^^ ^^« >^:3^V— «^ >^.---A^ \ ^>^ j^>:::i:^2:^^^^iS^ ^feS*^^ ^rfnm R fa H / ^^y^^^J^^""^^'^ )lVx [\vM ^®5i i-r-\ //y 1 \-rwM "^^J^fjMyyC/ 1]^ ^^^^^v WvV/ ^/// ^ \.^*CvV\^ ^^N/ g\W v/i Aa \ / / \ m x/ ''^i^ / no ^ lllllir 1S79\ N./ / ^"^^^^ \ --^^<:-^ -^. , ^ r '\, ACKNOWLEDGMENTS I wish to thank J. Niebauer, who did much of the initial ice data processing. I also wish to thank the publications and drafting departments of the Institute of Marine Science, University of Alaska, for their help in preparing this manuscript. This work, Contribution No. 403, Institute of Marine Science, University of Alaska, was supported by the National Science Foundation, Division of Polar Programs, Grant DPP 76-23340 A02 (PROBES). Additional support was provided by the Alaska Sea Grant Program and the State of Alaska under projects M/81-01 and R/06-06. Recent fluctuations in sea ice distribution 139 Figure 9-6. January southern ice limit for 1975-79. REFERENCES Coachman, L. K., and R. L. Charnell 1979 On lateral water mass interaction— a caise study, Bristol Bay, Alaska. J. Phys. Oceanogr. 9: 278-97. Leonov, A. G. 1960 Regional oceanography, Part 1 (in Russian). Gidrometeoizdat, Lenin- grad. (Transl. avail. Nat. Tech. Inf. Serv.) Springfield, Va. AD 627508 and AD 689680. Kukla, G. J., and H. J. Kukla 1974 Increased surface albedo in the Lisitsyn, A P. Northern Hemisphere. Science 183: I960 Recent sedimentation in the Bering 709-14. Sea. IPST Press, Jerusalem. 140 Ice distribution and dynamics McLain, D. R., and F. Favorite 1976 Anomalously cold winters in the southeastern Bering Sea, 1971-75. Mar. Sci. Comm. 2(5): 299-334. Muench, R. D., and K. Ahlnas 1976 Ice movement and distribution in the Bering Sea from March to June 1974. J. Geophys. Res. 81(24): 4467-76. Niebauer, H. J. 1978 On the influence of climatic fluctua- tions on the biological and physical oceanography of the southeast Bering Sea Continental Shelf. Presented at the Joint U.S. /Japan Bering Sea Ecosystem Seminar, Seward, Alaska, 7 August 1978. 1980 Rodgers, J. C. 1978 Sea ice and temperature variability in the eastern Bering Sea and the relation to atmospheric fluctuations. Sub- mitted to J. Geophys. Res. Meteorological factors affecting interannual variability in summertime in the Beaufort Sea. Monthly Weather Rev. 106(6): 890-7. Sater, J. E., J. E. Walsh, and W. I. Wittman 1974 Impingement of sea ice on the north coast of Alaska. In: The coast and shelf of the Beaufort Sea, J. C. Reed and J. E. Reed, eds., 85-105. Arctic Inst. N. Amer. Tabata, T. 1974 Wagner, A. J. 1976 Movement and deformation of drift ice as observed with sea ice radar. In: Oceanography of the Bering Sea, D. W. Hood and E. J. Kelley, eds., 373-382. Inst. Mar. Sci., Univ. of Alaska, Fairbanks. Occ. Pub. No. 2. Weather and circulation of April 1976— Unprecedented spring heat wave in the Northeast and record drought in the Southeast. Monthly Weather Rev. 104(7): 975-82. Weather and circulation of January 1979. Widespread record cold with heavy snowfall in the mid-west. Monthly Weather Rev. 107(4): 499- 506. Walsh, J. W., and C. M. Johnson 1979 An analysis of arctic sea ice fluctua- tions, 1953-77. J. Phys. Oceanogr. 9(3): 580-91. 1979 Remote Sensing Analysis of Ice Growth and Distribution in the Eastern Bering Sea S. Lyn McNutt NOAA Pacific Marine Environmental Laboratory Seattle, Washington Present Affiliation : Science Applications, Inc. Bellevue, Washington ABSTRACT Ice thickness distribution and ice types for the eastern Bering Sea are inferred from satellite imagery and available aircraft data. These have been combined with analyses of ice bridging and floe trajectories to estimate movement and generation of ice within the pack. The location of the ice edge has been plotted for different dates using satellite image- ry and ice analysis charts. The position of the edge and the variability of the geographic location of the ice types and thicknesses supports a theory of ice generation according to which ice forms along the leeward side of east-west -trending coasts, is advected to the south-southwest within the pack, is broken into floes near the ice edge by the effects of wave propagation, and melts at the edge when the thermodynamic limits of its stability are reached. INTRODUCTION In March, 1979, a joint experiment was conducted by NOAA, NASA, and the University of Washington in the Bering Sea. The NOAA ship Surveyor, sta- tioned along the ice edge, provided ground truth for physical oceanographic and meteorological experi- ments (Pease 1979). Personnel from the University of Washington tracked floe movement at the ice edge and obtained core samples of the ice (Martin and Bauer, this volume). Researchers from Scott Polar Research Institute installed accelerometers on floes to assess wave propagation into the pack. The NASA C130 aircraft flew one mission 14 March 1979 over the Surveyor. The remote sensing equipment on board included a laser profilometer, a step-frequency radiometer (Harrington et al. 1979), a scatterometer, and a 23-cm format, 153.12-mm focal length camera. The NASA CI 31 aircraft flew over the same location a few hours later with a Side-looking Airborne Radar (SLAR) (Schertler 1979) on March 14. An additional track was flown in the Norton Sound area March 27. The Surveyor was in the ice during this period of maximum extent and at the beginning of a period of gradual retreat (Pease 1979). This paper examines remote sensing evidence about the ice generation regime in the Bering Sea. Satellite imagery from the Defense Military Satellite Program, TIROS, and LANDSAT are used in conjunction with available ice charts from the Navy/NOAA Joint Ice Center (FWS) in Suitland, Maryland, and the above- mentioned field program, to formulate daily charts on ice conditions (Fig. 10-1). These charts were averaged to show weekly conditions. This data set is used to describe the ice regime in the eastern Bering Sea for March 1979. Because ice in the Bering Sea is not a closed pack, as in the Arctic, the ice regime is qualitatively different. In the Bering Sea all ice is first-year ice and melts completely by the end of each season. The period 1-31 March 1979 was unique in that the ice was at maximum extent and also began a rapid, complete meltback to ice-free conditions. WEEKLY AVERAGE ICE CONDITIONS Four charts of weekly ice conditions have been produced by averaging daily charts. The charts cover the following periods: (1) March 1-7; (2) March 8-14; (3) March 15-21; (4) March 22-28. (Clouds obscured conditions at the very end of the month; extrapola- tions were not made.) 141 142 Ice distribution and dynamics Bering Sea Ice Analysis March 1, 1979 .172 ]W toy Ki(l° ISS" Bering Sea Ice Analysis March 2, 1979 1B0° 17? or toy Tfti Bering Sea Ice Analysis March \ 1979 Figure 10-1. Daily ciiarts on ice conditions in tiie Bering Sea. Tliese analyses were compiled using Defense Military Satellite Program, TIROS, and LANDSAT imagery in comparison with FWS ice charts. 1. March 1-7 (Fig. 10-2). The ice is at maximum extent. Significant features include polynyas south of St. Lawrence Island and Nunivak Island and south of the western mainland in Norton Sound; floes which stream from the east of St. Lawrence Island to the south and southwest around the polynya area; the thick first-year ice (120 cm-2 m) north of St. Law- rence Island and Nunivak Island; and loose floes and streamers along the ice edge. 2. March 8-14 (Fig. 10-3). This week showed little change in ice conditions from the previous week; however, the ice continued to advance during the early part of the week and then began a gradual retreat. 3. March 15-21 (Fig. 10-4). A warming trend began during this week, accompanied by a reversal of northeast winds to a southerly direction (Pease, this volume). Breakup began and is especially noticeable in areas near the coast previously occupied by young first-year ice and polynyas. 4. March 22-28 (Fig. 10-5). This figure shows the effects of the breakup due to a warming trend. There is open water north of St. Lawrence Island and west of the mainland. The stream of floes around the south-southwest coast of St. Lawrence Island is still evident as are the floes and thicker ice west of the island. The geographic constancy of features such as floe streamers, polynyas, and ice bands along the edge supports a theory of ice generation for the eastern Bering Sea like that which was proposed for the western Bering Sea by Loshchilov (1974) during the BESEX experiment. A schematic diagram (Fig. 10-6) shows that the ice, when at maximum extent, oper- ates like a conveyor belt. The ice is formed in the leeward side of the east-west trending coasts and is advected downwind into the pack. This has also been observed by Muench and Ahlnas (1976) and Ahlnas and VVendler (1979). As the ice advances, it thickens and continues to be blown downwind to the edge, where it melts. This type of morphology for Bering Sea ice is evidenced by cores taken from the ice (Martin and Bauer, this volume, Martin and Kauffman 1979, Gloerson et al. 1974), which show ice compac- tion north of the islands and divergence around the islands and along the edge of the pack. The ice in the area to the east of St. Lawrence Island changes gradually from compaction to divergence. This is due, in part, to the shape of the ice pack, which is formed in a more constricted area than the one to which it is advected. Remote sensing data collected in March support the hypothesis that (1) ice is formed in the leeward side of east-west trending coasts under a dominant wind regime from the northeast; (2) ice is compacted north of the islands; (3) ice diverges around the islands; (4) ice at the seaward edge has reached its thermodynamic limit and is in the process of decay; and (5) ice retreat during breakup progresses most rapidly into the areas of thinner ice which had been supplying ice to the rest of the pack. 170' 175 175" 170" 165 160 155 150 65 62 59 56 AVERAGE ICE CONDITIONS March 1-7, 1979 ^'--^- J >- ■ Fast ice 65 62 59" 56" 180' 175 170 165 160' Figure 10-2. Weekly average ice conditions, 1-7 March 1979. Figure 10-3. Weekly average ice conditions, 8-15 March 1979. 143 Figure 10-4. Weekly average ice conditions, 15-21 March 1979. 59 AVERAGE ICE C^DITIONS March 22-28, 19 ■ Fast ice 56 180 175' 170" 165" 160 Figure 10-5. Weekly average ice conditions, 22-28 March 1979. 144 Remote sensing analysis of ice growth and distribution 145 165° 160° 155° — I 1 1 1 1 1 1 1 r Ice At Maximum Extent New Ice Converging Diverging 65' Figure 10-6. Schematic diagram of areas of ice generation, compaction, and divergence. This is meant as a first look for future refinement. The area of Eastern Norton Sound under certain wind conditions also behaves as an ice-formation area. ICE FORMATION In the initial stages of ice formation, grease ice is produced in the upper layer of the water and floats to the surface (Martin, Oil-Ice Interaction, this volume). This ice is blown downwind on the water's surface and piles up at the leading edge of ice stream- ers (Fig. 10-7). As the grease ice layer thickens, it damps the surface wave field (Martin, Oil-Ice Inter- action, this volume) and is compacted until it forms a surface layer which can support ice growth under- neath. This thickening ice forms pancakes and eventually larger floes which are incorporated into the pack. This replacement process appears to be continuous. Fig. 10-8 shows a LANDSAT 3 image of the polynya area west of Nome along the Seward Penin- sula. It has been enhanced to bring out detail in the first two gray levels of the image so that the grease ice forming in the lee of the mainland is visible. The wind was from the northeast (Pease, this volume). Streamers of grease ice can be seen, as well as large floes made of thin ice which was piled up and had broken free. Fig. 10-9, a mosaic of five LANDSAT 3 scenes from 12 March 1979, shows ice formation along a track from the Bering Strait to the ice edge near St. Matthew Island. Here grease ice is visible south of St. Lawrence Island. The gray signature of this thinner ice gradually appears whiter due to the thicker ice downwind. This polynya with grease ice forming within it was also noted during BESEX in 1973 (Ramseier et al. 1974), and is often evident in FWS ice charts (Eastern-Western Arctic Sea Ice Anal- yses, 1972-78). The wind was from the north (Pease, this volume). Cores taken by Martin and Kauffman (1979) at the ice edge show that the upper portion of these floes consisted of a layer of consolidated grease ice. This was also observed in cores taken during BESEX in 1973 (Gloerson et al. 1974, Ramseier et al. 1974). Figure 10-7. Grease ice piioto taken from the CV990 aircraft during BESEX, 1973. (Courtesy of S. Martin, Univ. of Wasiiington, NASA/Goddard). Tiie ice is being blown downwind wiiere it piles up at the leading edge. The grease ice streaks also damp the waves at the surface. 146 Remote sensing analysis of ice growth and distribution 147 Figure 10-8. 2 March 1979, LANDSAT Imagery. This image has been enhanced to show details of the grease-ice forma- tions in the leeward side of the shoreline. These photographs support the hypothesis that the ice is formed to the north in large polynya areas and is blown south by the wind to the ice edge. ICE COMPACTION Ice in the Bering Sea, moving under the influence of the prevailing wind, tends to compact as it con- stricts or meets an obstacle. Sodhi (1977) and Shapiro and Burns (1975) show that this type of ice bridging often occurs north of St. Lawrence Island. FWS ice charts often show areas of thick first-year ice north of St. Lawrence, Nunivak, and St. Matthew islands (Eastern -Western Arctic Sea Ice Analyses 1972-78). Fig. 10-9 shows an example of such ice compaction. The wind is from the north and ice on the north side of St. Lawrence and St. Matthew Islands appears to be thicker, consistently white in appearance, and composed of tightly compacted floes with ridging structure evident on the surface. On the next day, 13 March, the LANDSAT 3 image (Fig. 10-10) shows that the same area of compaction north of St. Lawrence Island is relatively unchanged. The ice on both days was at maximum extent (Fig. 10-3). The same phenomenon can be seen earlier in the month in the 2 March imagery from LANDSAT 3 (Fig. 10-11). The gray scale of the image differs significantly from the previous image because less light was available. However, the same areas of relative thickness can be seen north of St. Lawrence Island. The wind was from the northeast (Pease, this volume). Figure 10-9. 12 March 1979, LANDSAT. This image is a mosaic of five separate scenes. It shows convergence north of St. Lawrence Island, St. Matthew Island, the polynya behind Lawrence Island, breakout of floes through Bering Strait, floes moving around St. Lawrence Island, ice streamers and bands at the edge, and roll cloud formation. 148 Remote sensing analysis of ice growth and distribution 149 174° 172' 170' 168' LYN McNUTT Figure 10-10. 13 March 1979, LANDSAT. ICE DIVERGENCE Ice which is compacting tends to raft and ridge until shearing causes portions of the ice to break off. The freed ice then separates from these bridging areas, causing leads to be formed (Sodhi 1977). This ice moves downwind as giant floes and vast floes in a matrix of smaller floes and brash ice (an accumula- tion of floating ice made of fragments <2 m across). This accounts for the areas with large floe streams in Figs. 10-1-10-5. This phenomenon shows up well in both Fig. 10-9 and 10-10, and was noted during the BESEX experiment by Ramseier et al. (1974), Gloerson et al. (1974), and Campbell et al. (1974). When locations of the leads and floes for both days are compared, it is seen that the floes have moved, and that the orientation of the leads remained consis- tent, approximately perpendicular to the wind. r c O Pi o a O ai '-'S ClH ,-1 i50 A more detailed study of this phenomenon was undertaken using enhanced TIROS imagery for March 16, 17, 18, and 19 (Fig. 10-12), so that individual floes could be tracked for a longer period of time. Muench and Ahlnas (1976) tracked floes during spring 1974 and found that their dominant direction of travel, in the area south of St. Lawrence Island, was to the south-southwest under predominantly northeast winds. Twelve floes A-G, J-N (Fig. 10-12), were tracked during the period 16-19 March for two days, and where possible for three days. Due to the resolution of the imagery, all floes appeared to be giant (>10 km across). Floes could not be tracked Remote sensing analysis of ice growth and distribution 151 longer than three days because they tended to break up into floes smaller than the resolution of the imagery. Table 10-1 lists these floes by letter, gives their approximate size, the direction of the wind from true north (Pease, this volume), the floe direc- tion, and the floe speed. Floes A-G were followed on 16-17 March. For B-F the wind direction was estimated at 220° -225° relative to true north. The floes traveled 235° to 245°. For A and G the winds were estimated to be at 205° and the floes traveled 180°. These floes are within an area where their movement is affected by the ice shear around St. Lawrence Island. For floes Figure 10-12. Floe trajectory map. 16-19 March 1979. The dominant wind regime was from the northeast during this period; floes tended to move to the south-southwest in response to the wind around the polynya area behind St. Lawrence Island. 152 Ice distribution and dynamics TABLE 10-1 Floe trajectories, March 16 - 19, 1979 Floe Wind direction Floe direction (from true north) Floe velocity Floe size (m/sec) (km) .28 8X20 .13 16X40 .50 16X40 .23 12X20 .23 12X20 .23 16X20 .22 20X20 .22 36X32 .22 12X12 .28 16X36 .46 30X28 .41 16X32 .39 14X34 .42 32X60 A B D E F G J K L M N 205° 220°-230' 220° -230' 220° -225' 220° -225° 220° -225° 220° -225° 220° -225° 205° 230° -235° 230° -235° 230° -235° 230° -235° 210° 180 256° 218° 252° 238° 254° 237° 245° 180° 238° 238° 230° 235° 217° J-M, on the 18th and 19th, the winds were from 230° to 235° and the floes traveled 230° -238°. For N, the floe traveled 217° under winds estimated to be at 210°. All floe movements show a strong correlation with wind direction and appear to move to the south-southwest under the northeast wind regime which is dominant at that time of year (OCSEAP 1977). The ice movement is slightly to the right of the wind. This agrees with data reported by Sverdrup (1928). Floes B and C were tracked for three days and maintained the strong relationship between wind direction and direction of floe movement. The average distance traveled by the floes during a 24-hour period was 26 km with a range of speed from 0.13 m/sec to 0.50 m/sec. The average speed was 0.26 m/sec. The velocity, distance, and direction of travel of these floes imply a strong divergence of ice within the pack downwind toward the ice edge. The lead orientation for the same period (Fig. 10-13) reflects the movement of the ice to the southwest under this wind pattern and suggests a movement of the ice toward the edge to the west of St. Matthew Island. SLAR data from the NASA C131 aircraft (Fig. 10-14) also show this lead pattern east of St. Lawrence Island in an area of large floes. Winds this day were also from the northeast (Pease, this vol- ume). ICE-EDGE PHENOMENA Figs. 10-2 and 10-3 show little change in the location of the edge itself, yet floe studies indicate that volumes of ice from within the pack are being blown to the edge. In order to study the nature of the ice at the edge, ground truth data needed to be collected. The NOAA ship Surveyor was stationed at the ice edge from 2 to 14 March 1979 (Pease 1979; Pease, this volume). On 14 March, two NASA aircraft flew over the ship while ground truth data were being collected. The flightline for the C131 is shown in Fig. 10-14. Fig. 10-15 shows the flight track from the C130 aircraft. An analysis of in-situ data on ice cores (Martin and Kauffman 1979) and floe drift, along with information on wave attenuation, shows that ice along the edge consists of rotten floes which have reached a thermodynamic limit of stability and have begun to melt (Martin and Bauer, Pease, this volume). A plot of the location of the isotherms of the water surface as observed during the cruise period shows that the — 1 C isotherm moved southward with respect to the ice edge (Pease, this volume). The CTD casts taken also show a lens of less saline water extending out from the edge— evidence of melting (Pease, this volume). Remote sensing analysis of ice growth and distribution 153 Figure 10-13. Lead orientation, 16-19 Marcii 1979. Leads follow the same general trend downwind as the floes. Satellite imagery such as that seen in Fig. 10-11, TIROS, and Figs. 10-9 and 10-10, LANDSAT, were used to study small-scale variations of these ice bands. These band features are not an isolated occurrence. They are often noted on FWS charts and have been studied by Muench and Charnell (1977). They can also be seen in a LANDSAT 3 image taken 5 March 1979 (Fig. 10-16), and on the 14 March imagery (Fig. 10-14). An assessment of change in an ice band can be made using a photograph from the C130 overflight (Fig. 10-17). Although this image was taken approjj- imately three hours earlier than the C131 SLAR image, the basic shape of the formation is still recog- nizable and can be seen to be made up of small, angular floes. Similar photography from successive altitudes was used to study the characteristics of the floes within these bands. Fig. 10-18 shows a cross section of one of these bands. The distance across the band is 20 km. Darker floes in the middle of the band indicate thinner ice. The darker color and the nature of the surface patterns suggest that these floes are rotting. Ship's personnel who had occupied stations on similar floes reported that this type of ice was melting rapidly, and that it was thin enough to respond plastically to waves which were propagating through the ice (Pease 1979; Martin and Bauer, this volume). Fig. 10-19, taken from a lower altitude, shows another cross section of a band. The flightline was from west to east, the wind from the northeast. 154 Ice distribution and dynamics SLAR IMAGERY MARCH 14,1979 BERING SEA Figure 10-14. 14 March 1979, SLAR Mosaic. (Cour- tesy of R. Schertler, NASA/ Lewis.) Tiie C131 flightline is on tlie map at tiie middle riglit. Only the ice-edge lines and a portion of the long line from Nome are reproduced. Remote sensing analysis of ice growth and distribution 155 168° 174° 73° 17 2° 171° 170° 169° 168° 167° 165= Figure 10-15. 14 March 1979, C130 flightline. Smaller floes were blown downwind to the ice band. The leading edge of the band is made up of angular, thicker white ice floes. These floes are held together at the leading edge by the effects of the incoming swell; yet, as a group, they tend to move downwind at speeds of 0.3 m/sec (Martin and Bauer, this vol- ume). Fig. 10-20 shows the surface configuration of melting found on one of the thinner floes. This melting is not like that of Arctic ice, where melt puddles form on the surface of thick floes, but appears to be due to floe motions causing water to spill over the sides, and to the effects of a surround- ing area of warmer water which causes the floe to rot. The two basic types of floes found at the edge were the highly rafted, relatively thick, white, angular floes (Fig. 10-21), whose freeboard allowed them to travel rapidly under the influence of the wind; and the thinner, larger floes (Fig. 10-22) found in the interior of the bands. These larger floes were variable in size. but were generally an order of magnitude larger than the white floes, which typically measured between 30 and 60 m in diameter. These sizes compare well with those measured by Martin and Kauffman (1979). Martin and Bauer (this volume) have hypothesized that floes along the ice edge are formed from pack ice which is broken up at the edge by the effects of wind and incoming swell. First, the sheet ice is broken into vast rectangular floes whose width is proportional to the wavelength of the incoming swell (for an illustra- tion, see Fig. 12-14, Martin and Bauer, this volume). These floes are further broken down and the thicker ice is rafted while being broken into small floes, which have sharp, angular edges, unlike pancake ice, which tends to be rounded. Pease (1979, this volume) noted from CTD casts along the ice edge that there is a surface lens of less saline, warmer water extending from the ice edge— evidence of meltwater. Preliminary returns analysed from the step-frequency radiometer on board the C130 tend to support Figure 10-16. 5 March 1979, LANDSAT. This is a two-image mosaic showing ice bands to the southeast of Nunivak Island. 156 Figure 10-17. 14 March 1979, C130 photo mosaic of an ice band also seen in the 14 March SLAR data run. 157 Figure 10-18. Ice band near the NO A A ship Surveyor, showing the two types of ice in a band approximately 20 km across. The light-colored, angular floes are thicker ice; the darker ice near the center is thinner, melting ice. Figure 10-19. Ice band near the NOAA ship Surveyor. The flightline was from west to east (left to right), the wind was from the northeast. Ice fragments can be seen to be advecting downwind to the band. The leading edge of the band is loosely held together by the influence of incoming swell. 158 Figure 10-20. Two-frame mosaic of thinner ice near tiie interior of an ice band. The floes show the effects of melting along the edges and in the center, especially. i 159 160 Ice distribution and dynamics Figure 10-21. Detail of thicker, rafted white floes. Note the angular edges and the presence of snow on the surface. The floes are 30-60 m in size. this hypothesis. The return from the radiometer indicates that the ice is relatively fresh and occurs as a matrix of less-saline water than would normally be found on the surface (Swift, personal communica- tion). ICE RETREAT At the end of March 1979, the ice in the Bering Sea began a rapid retreat. This can be seen rather dramatically by comparing the 16 March TIROS Figure 10-22. Detail of thin, melting floes. Note the lack of snow on the surface and thaw holes. 161 162 Ice distribution and dynamics image (Fig. 10-11) with the 26 March TIROS image (Fig. 10-23). This period of retreat coincided with a time of maximum insolation and warm southerly winds generated by a low over the Aleutians (Pease, this volume). A comparison of Fig. 10-23 with the average weekly chart for 22-28 March (Fig. 10-5) shows the meltback proceeding in the eastern Bering Sea between the coast and Nunivak and St. Lawrence Islands, an area into which ice is continually advected (Fig. 10-6). It is interesting to note the persistence of the floes around St. Lawrence Island. The SLAR data for March 27 (Fig. 10-24) shows an open water/ thin ice area north of St. Lawrence Island (McNutt 1977), more leads present on the eastern edge of St. Lawrence Island, and open areas and new leads in Norton Sound. Figure 10-23. 26 March 1979. This TIROS enlargement shows the effects of a rapid meltback. (Compare with Fig. 10-11.) <3> oi S fS" .5 fe ■S -C hJ J2 c ^^ 5X) ^ c« a; -*-» ^jju zi (-• O ^ C/3 CO C >- -V^ ;g O O O Eh ^Pi O-Xi c/5 "^ CO r" '— ' i-X TS '* ^ es A, A-^ LU < X q: o -J < CO 2 3 O CO O f- QC O CO I Z I i63 164 Ice distribution and dynamics The ice during this period of meltback appeared to be retreating most rapidly in areas which had pre- viously been supplied by ice blown down from the north. The change of wind direction from north- northeast to south-southeast affects floe trajectories and tends to blow ice to the north instead of to the south-southwest. One would not expect to find ice bands along the edge under these conditions, and yet isolated, string-like features of ice persist in areas which axe otherwise ice free (Fig. 10-23). In some areas, e.g., the area to the southwest of St. Lawrence Island, ice persists and remains relatively motionless. More work needs to be done to study the effects of wind, insolation, and water temperature on ice movement during this retreat. REFERENCES Ahlnas, K., and G. Wendler 1979 Sea observations in the Bering, Chuk- chi and Beaufort Sea, Proceedings of POAC— 79, Norwegian Inst. Tech. Campbell, W. J., P. Gloerson, and R. O. Ramseier 1974 Synoptic ice dynamics and atmos- pheric circulation during the Bering Sea experiment. Results of the US contribution to the joint US/USSR Bering Sea experiment. NASA/ Goddard (NASA X-910-74-141). CONCLUSIONS The 1978-79 ice season was a relatively light ice year, perhaps because of short-term climatic varia- tions in the mean annual sea surface temperatures (Niebauer, this volume). March 1979 was unique, in that one month encompassed conditions of maximum extent and rapid retreat. When the Bering Sea ice is at maximum extent, it reaches a steady state in which ice is formed in polynyas on the leeward side of east-west trending coasts and advected downwind. The ice bridges along the northern coasts of islands and shear zones existing on both sides of the wedge of thicker ice. Floes are broken off at these shear zones and, under north-northeast wind conditions, floes originating on the eastern side of St. Lawrence Island drift south-southwest towards the ice edge near St. Matthew Island. Ice at the edge is broken up by the effects of wind and incoming swell and is blown downwind into ice bands. Ice in these bands reaches its thermodynamic limit and melts along the edge. During the breakup the ice retreats under the effects of increased insolation and southerly winds. The ice near the coast, which had been diverging from the north, is no longer replenished and retreats first. Under continued southerly wind conditions, indivi- dual floes may begin to migrate northward. ACKNOWLEDGMENTS This work is Contribution No. 443 from the NOAA Pacific Marine Environmental Laboratory. Writing of the manuscript was funded, in part, by NOAA/NESS/SPOC Group Project #MO-A01-78- 00-4335. Reproduction of the satellite imagery would not have been possible without the skills of Jim Anderson, PMEL. Fleet Weather Facility 1975 Western Arctic sea ice analysis, 1972- 1975. Capt. S. L. Balmforth, C. O. Suitland, Maryland. 1976 Eastern -western Arctic sea ice analy- sis. Cmdr. V. W. Roper, C. O. Suit- land, Maryland. 1977 Eastern-western Arctic sea ice analy- sis. Capt. J. A. Jepson, C. O. Suitland, Maryland. 1978 Eastern-western Arctic sea ice analy- sis. Cmdr. J. C. Langemo, C. O. Suitland, Maryland. Gloerson, P., R. Ramseier, W. J. Campbell, T. C. Chang, and T. T. Wilheit 1974 Variation of the morphology of selec- ted microscale test areas during the Bering Sea experiment. Results of the U.S. contribution to the joint U.S./ U.S.S.R. Bering Sea experiment. NASA/ Goddard (NASA X-910-74- 141). Harrington, R. F. 1979 An airborne remote sensing 4.5 to 7.2 Gigahertz stepped frequency micro- wave radiometer. Proc. IEEE Micro- wave Symposium. Loshchilov, V. 1974 Characteristics of the ice cover in the operational area of the "Bering" expedition (1973), Preliminary results of the Bering expedition, 18-30. Remote sensing analysis of ice growth and distribution 165 Martin, S., and P. Kauffman 1979 Data report on the ice cores taken during the March 1979 Bering Sea ice edge field cruise on the NOAA ship Surveyor, Univ. of Washington, Dep. of Oceanogr., Spec. Rep. No. 89. McNutt, S. L. 1977 Interpretation key for microwave discrimination of sea ice characteris- tics, M.A. Thesis, UCLA. Muench, R. D., and K. Ahlnas 1976 Ice movement and distribution in the Sea from March to June, 1974, J. Geophys. Res. 81(24). Muench, R. D., and R. L. Charnell 1977 Observations of medium-scale features along the seasonal ice edge in the Bering Sea, J. Phys. Oceanog. 7(4). Ramseier, R. O., P. Gloerson, W. J. Campbell, and T. C. Chang 1974 Mesoscale description for the principal Bering Sea ice experiment. Results of the U.S. contribution to the joint U.S./U.S.S.R. Bering Sea experiment. NASA Goddard (NASA X-910-79- 141). Schertler, R. 1979 Shapiro, L. H. 1975 Background information on copy negatives of LeRC original SLAR/ Imagery, Lewis Research Center/ NASA. and J. J. Burns Satellite observation of sea ice move- ment in the Bering Strait region. Climate of the Arctic, Rep. Univ. of Alaska, Fairbanks. OCSEAP 1977 Climate atlas of the outer continental shelf waters and coastal regions of Alaska, 2, Bering Sea. Arctic Envi- ronmental Information and Data Center, Anchorage, Alaska. Sverdrup, H. V. 1928 The wind-drift of the ice. In: The Norwegian polar expedition with the Maud 1918-1925, Scientific results, IV: 1. A. S. John Griegs Boktrykkeri, Bergen. Pease, C. 1979 Cruise report for NOAA ship Surveyor 28 Feb. - 17 March, 1979, NOAA, PMEL, Seattle, Wash. Sodhi, D. S. 1977 Ice arching and drift of pack ice through restricted channels. U.S. Corps of Eng., CRREL Rep. 77-18. Nearshore Ice Characteristics in the Eastern Bering Sea 11 William J. Stringer Geophysical Institute Fairbanks, Alaska ABSTRACT Bering Sea nearshore ice conditions are described on the basis of a compilation of fast-ice edge satellite data, ob- servations of specific ice events, and results from other studies. LANDSAT imagery at 1:500,000 scale was used to map Bering Sea ice conditions between 1973 and 1976 in nearshore areas. From these maps, secondary single-attribute maps were compiled, giving the edge of fast ice at various epochs during these four years. Seasonal maps representing midwinter, late winter to early spring, and mid-to-late spring were then compUed. The seasoned average maps were then compared to determine seasonal trends in the location of the fast-ice edge. This information was analyzed together with imagery showing specific ice events and bathymetric charts, wind data, tidal variations, and observed ice trajectories. The result is a regional description of average nearshore ice conditions along the Bering Sea coast from Cape Prince of Wales to Cold Bay on the Alaska Peninsula. Over this distance a north-south transi- tion is found from fast-ice conditions similar to those in the Beaufort and Chukchi Seas (fast ice bounded by grounded ridge systems at a depth of 20 m ) to conditions generated by large tidal variations, offshore winds, and highly mobile ice, with the result that fast ice is found only in highly protected, shallow waters. INTRODUCTION Any discussion of oceanic ice in the nearshore area necessarily centers on fast ice — ice fixed with respect to shore. Although fast ice is found along almost all ice-bound coasts, its characteristics vary from one place to another. This variation depends on many factors, including the local bathymetry, internal stresses in the adjacent ice pack, local surface winds, tides, and currents. Usually the fast ice is composed largely of annual ice, with perhaps oc- casional interfused pieces of multiyear ice. In Alaskan waters, annual ice seldom grows to a thick- ness of more than two meters. Because of the low buoyancy of ice, most of the vertical extent of fast ice is below the surface level. Obviously then, at water depths of less than two meters, the fast ice is actually bottom-fast after it grows thick enough. Changes in sea level, resulting either from tides or from weather patterns, create a "hinge" between the floating fast ice and the bottom-fast ice. The hinge usually takes the form of a crack between the two ice types. As the winter progresses and the ice grows in thickness, the active tidal crack will generally move seaward, leaving old cracks to the shoreward, often bridged by blowing snow. In areas where tidal variations are low, the pattern of these tidal cracks can be fairly simple. Where there are large tidal variations, as in the Bering Sea, there will be a tidal crack zone with the currently active tidal cracks determined largely by the instantaneous tide state as well as ice thickness. This pattern is superimposed on the ice state created during freeze-up in the nearshore area. It is possible for the ice to simply freeze in place, growing thicker with time, but this is often not the case. In reality a wide variety of ice conditions can be found, depending on the history of ice dynamics in a par- ticular freeze-up season. The original ice sheet, for instance, may freeze to a thickness of 30 cm; a storm may ensue that withdraws the ice from shore, breaks most of it up into small (1 m) plates, and then drives it to shore again. The plates may then freeze together in an extensive rubble field and form the fast ice for that year. Other initial conditions are possible. In 1973, an "ice push" event occurred at Kotzebue, Alaska: a stable sheet of moderately thick (1-2 m) fast ice was driven as much as 15 m onto the beach just south of town, carrying with it a surplus landing barge used as a salmon cannery (Mr. Albert Francis, personal communication). Kovacs and Sodhi (1979) have documented a number of these incidents in the Beaufort Sea region, as well as a related phenomenon, ice piling events: instead of an ice sheet being pushed across the beach and adjoining tundra, a large pile of broken ice is created at or near the beach. 167 168 Ice distribution and dynamics Figure 11-1. Idealized Beaufort Sea nearshore ice regime showing one of many possible configurations. Note apron of new ice adjacent to floating shear ridge in attached fast ice zone. This ice has formed as a result of a recent advection event. Offshore, the floating fast ice is often anchored by pressure and shear ridges with sufficient keel depth to be grounded on the ocean floor. Generally, few large grounded ridges are found in shallow water (up to 12 m) and ridges are seldom thick enough to be ground- ed in water deeper than 20 m. While a great deal of work has been done toward determining these limits in the Beaufort Sea (Reimnitz and Barnes 1973, Kovacs and Mellor 1974, Kovacs 1976, and Stringer 1978), relatively little work has been done in the Bering (Stringer 1978, Ray and Dupre, this volume, Duprel980). Floating fast ice is not always bounded by a zone of grounded ridges (sometimes called stamukhi, see Reimnitz and Barnes 1976). Whether or not a grounded-ridge zone exits, fast ice can extend sea- ward up to 100 km or more (Stringer 1974). If the grounded-ridge zone is present, it tends to protect the enclosed fast ice from deformation resulting from pack-ice forces, although deformations may still take place within this protected zone. The grounded-ridge (or stamukhi) zone is an important feature because it often determines the boundary between fast ice and pack ice. The "flaw Nearshore ice characteristics 169 Pack Ice Fast Ice ■o *4— O) N Li_ O Ll_ m Cracks Figure 11-2. Idealized Beaufort Sea nearsiiore ice regime siiowing one of many possible configurations. Note pile of ice on beach, large apron of attached ice starting at a shear ridge and extending seaward. lead" is often found just seaward of the deepest grounded ridge. In this zone a large amount of pack-ice energy is expended that must be accounted for when modeling nearshore ice mechanics. With the increased attention to offshore structures required in the course of petroleum development, the grounded- ridge zone has become important in relation to physical hazards. Beyond the grounded -ridge zone, an apron of floating fast ice (here called "attached" ice in order to emphasize the absence of grounded features to the seaward) can often be found extending from a few meters to many kilometers into the ocean. Since the stability of attached ice is tenuous, it can easily be converted into pack ice by an ice-breaking event. Figs 11-1 and 11-2 give some idea of the range of conditions that can be found. The situation depicted in Fig. 11-1 shows relatively undeformed bottom-fast ice along the beach with tidal cracks occurring near the 2-m isobath. Offshore in water a few meters deep, occasional piles of pressured ice may in fact be grounded. These often act as single-point anchors and are generally created as weaker ice is pressured around stronger pans. This pattern extends out to 1 70 Ice distribution and dynamics the grounded ridges. The dimensions of the pans and piles around them, as well as the distance to the grounded ridges, can vary widely. For instance, the pans could be 30 or 3,000 m in diameter and the ice piles could be 1 or 10 m above sea level. The distance to the grounded ridges could be from 1 to 30 km. Because of the necessary vertical exaggeration in these figures, the angle of repose of ice ridges shown appears to be much steeper than it is. Furthermore, the thickness of unpressured ice is exaggerated in the vertical plane, which may give a false impression of the geometry involved. Beyond the grounded ridges, the attached ice is depicted as being relatively smooth but with some hummocking. Finally a large floating ridge is en- countered which would be grounded if it were further inshore. As depicted here, this ridge was recently the edge of fast ice with active differential motion taking place along it. But the ice opened and moved seaward, forming a large flaw lead, which froze to a thickness of 10 or 20 cm and then failed in tension, forming a new flaw lead. This narrow lead now defined the edge of fast ice. Beyond the lead, the ice can truly be classified as pack ice. Fig. 11-2 shows another common nearshore ice situation. Here, a rubble pile is found on the beach with the active tidal crack beyond its base. Just off- shore, ice is piled against the beach and grounded in a few places. In some years many such ice piles are found near the beach in exposed locations (Barrow, for instance). Ice of this type makes activities involv- ing transportation across the fast ice particularly difficult. Beyond the grounded ice, a second tidal crack may be found, followed by floating fast ice with occasional minor ridges which may be grounded in relatively shallow water. Farther offshore, there are smooth pans separated by hummocked ice and, finally, near the 14-m isobath is the grounded-ridge zone. As depicted here, the pack ice in the past has been driven along the outside edge of the outermost grounded ridge, creating a shear ridge similar to the floating ridge in the attached ice zone of Fig. 11-1. Here, however, the attached ice has not recently withdrawn but has remained adjacent to the ground- ed ridges, creating a zone of thin ice. The flaw lead is found offshore from a large floating ridge in the attached ice. The pack ice begins at this point and extends seaward. Again, it should be emphasized that the vertical scale creates an inaccurate impression of horizontal dimensions: the distance between the grounded-ridge zone and the large offshore ridge could be on the order of 10 or 20 km. However, observations by LANDSAT (Stringer 1978) and laser profilometer (Tucker et al. 1980) show that these large shear ridges formed by differential motions of ice sheets are generally found in the nearshore area. Hence, their frequency diminishes with increasing distance from the grounded-ridge zone. FACTORS DETERMINING BERING SEA NEARSHORE ICE CONDITIONS The description of nearshore ice conditions and behavior presented up to this point is largely based on observations in the Beaufort and Chukchi Seas and applies to some areas of the Bering Sea. However, two factors influence ice behavior in some areas of the Bering Sea that are almost totally absent in the Beaufort: tides and ice advection. While the Beaufort coast experiences tides with a variation of only a few decimeters, tides at many locations on the Bering coast range over several meters. Also, while Beaufort Sea ice is almost always packed in against the coast, in many places along the Bering coast the ice is almost continually being pushed away from shore by winds and currents (Stringer 1978, McNutt, Pease, Ray and Dupr6, this volume). Fig. 11-3 shows an ice profile more typical of fast ice in the Bering Sea than Figs. 11-1 and 11-2. A grounded ridge is shown some distance from shore, but certainly closer to shore than the 20-m isobath. In order to be even semipermanent, this ridge must be sufficiently grounded to withstand the buoyant forces during high tides, and must have such a geometry that tidal fluctuations will not cause disintegration. Obviously, grounded ridges cannot present a continuous dam against the large forces created during tidal variations; hence, breaks and other disruptions of these ridges are common. Inshore from the grounded ridges, floating and bottom-fast ice are found. The extent of both these ice types depends greatly on the tide state, since the Bering coast has extensive mud flats covered by very shallow water. Several active and inactive tidal cracks can be found. Because of lateral motions caused by tidal currents and disruptions of the grounded ridges by tidal fluctuations, fast ice in the Bering Sea is not nearly so stable as fast ice in other areas with little tidal variation. Attached ice can occasionally be found beyond the grounded-ridge system, but because of tidal varia- tions, the flaw lead is most often found just seaward of the grounded ridges. Again, as in Figs. 11-1 and 11-2, the necessary vertical exaggeration should be considered when viewing this schematic drawing. Nearshore ice characteristics 1 71 Figure 11-3. Idealized Bering Sea nearshore ice regime showing effect of large tidal variations on fast ice. As depicted here low tide has caused the fracturing of floating fast ice and the breakup of attached fast ice. This effect tends to limit the extent and stability of Bering Sea fast ice. Advective export of Bering Sea nearshore ice £ilso contributes to the limitation of fast ice. There are many areas along the Bering coast where ice motion has a significant seaward component, and as a result, grounded ridges are seldom buUt in these locations. This condition contrasts sharply with that of near- shore ice in the Beaufort Sea, where the pack ice is nearly always present along the fast-ice boundary and is often driven along the fast ice with a shoreward component of force, thus creating the well-known shear ridges often found in that area. These two distinguishing influences, tidal fluctua- tions and ice advection, affect the Bering coast in varying degrees. At some locations they combine to severely limit the edge of fast ice to isobaths even less than 6 m. At other locations, conditions similar to those found in the Beaufort Sea prevail, with ice ridges grounded along the 20-m isobath. 1 72 Ice distribution and dynamics THE LOCATION OF BERING SEA FAST ICE CHARACTERISTICS OF NEARSHORE ICE IN THE BERING SEA In order to identify ice characteristics on a site- specific basis, maps of nearshore ice conditions were prepared from LANDSAT imagery at a scale of 1:500,000, showing the location of fast ice, pack ice, leads, ridges, hummock fields, and other identifiable features as well as shoreline and bathymetry. The techniques involved in preparing these maps have been described elsewhere (Stringer 1979) and will not be elaborated here. The individual maps of LANDSAT scenes, each covering an area of approximately 160 km X 160 km (100 X 100 nautical miles), were reduced to a scale of 1:1,000,000 and combined to produce composite single-attribute maps of the Bering Sea nearshore area at specific instances. The most important charac- teristic of sea ice for determining nearshore ice conditions was found to be the edge of fast ice. The ice season was divided into three periods, and a series of three maps was compiled showing the location of the edge of fast ice at specific times over several years. Figs. 11-4, 11-5, and 11-6 show fast-ice edges monitored during the periods of mid-winter, late winter to early spring, and mid-to-late spring between 1973 and 1976. Shown along with these ice edges is the average ice edge. In order to reveal temporal changes of location of the average ice edge, these average edges were compiled on one map (Fig. 11-7). Obviously, any significance accorded to trends apparent on this map must be tempered by consid- eration of the variability exhibited in the ice-edge data. At some locations, the edge of fast ice varies considerably in position during each period. Al- though the average edges in these locations show a temporal trend, it has only minor significance. In other locations, the variability of the fast-ice edge for each period is small compared to the changes in the average position from period to period. Strong temporal trends with year-to-year dependability are indicated. Finally, Fig. 11-8 shows the regional nearshore ice characteristics of the Bering Sea. This map was prepared on the basis of the considerations just described and by analysis of LANDSAT images for ice features such as ridges and hummock fields and of smaller scale satellite imagery for dynamic ice motions and other nearshore characteristics. The following section adds more detaU to the nearshore ice descriptions of Fig. 11-8 (items 1-33). In order to describe ice conditions along a segment of the coast, it is necessary first to describe the pack-ice conditions just offshore; the order of pre- sentation here conforms to that necessity. Ice motion through the Bering Strait The Bering Strait,. with the Diomede Islands in its center, represents a highly restricted ice passage between the Chukchi and Bering Seas. For some time, popular belief held that ice motion through the strait was generally from south to north in response to oceanic currents. This was supported by many ship-based springtime observations of broken pack ice passing northbound through the strait, often at speeds of several knots because late season lows to the south were causing southerly winds (Pease, this volume). Shapiro and Bums (1975) and Ahlnas and Wendler (1979) have shovm that occasional "break- out" events occur to the south when Chukchi Sea pack ice is extruded through the strait at fairly high velocities. These events have been examined in more detail by Prichard et al. (1979) and found to occur only during somewhat rare (two to three times annually) meteorological conditions which result in northerly winds across the ice in the strait, combined with a reversal of the usual south-to-north currents. During these events fairly large quantities of ice can pass through the strait, producing extensive grounded ice-ridge systems on Prince of Wales Shoal just to the north of the Alaskan side of the strait (Stringer 1978). It seems reasonable that during this process, relatively deep-draft first-year ice features would be created and transported into the Bering Sea and perhaps into western Norton Sound, but not eastern Norton Sound (Ahlnas and Wendler 1979). Nearshore ice conditions— Bering Strait to Yukon Delta At Wales, the edge of fast ice closely follows the coast, largely as a result of the combined effect of deep waters (>20 m) adjacent to the coast and ice motions through Bering Strait. From Wales to the Port Clarence entrance, the edge of fast ice is gen- erally inshore from the 20-m isobath in all periods. At the entrance to Port Clarence, the edge of fast ice usually bridges the narrow embayment of the 20-m isobath into the port. However, a tongue of open water occasionally follows this indentation. Just south of the Port Clarence entrance Hes a group of shoals as shallow as 4 m. These often appear to be BERING SEA MID -WINTER FAST ICE EDGE 1974 FEBRUARY 17- MARCH 6 1975 FEBRUARY 20- MARCH 4 1976 FEBRUARY 16 - MARCH 4 -X... AVERAGE "S" INDICATES SHOALS BATHYMET Figure 11-4. Map showing the location of the Bering Sea fast-ice edge during midwinter for 1974, 1975, and 1976. 173 Figure 11-5. Map showing the location of Bering Sea fast-ice edge during late winter/early spring for 1973 and 1974. 174 Figure 11-6. Map showing the location of Bering Sea fast-ice edge during mid-to-late spring for 1973, 1974, 1975, and 1976. 175 Figure 11-7. Map comparing average Bering Sea fast-ice edges for winter, late winter/early spring, and mid-to-late spring in order to determine seasonal changes. 7 76 Figure 11-8. Map summarizing Bering Sea regional nearshore ice ciiaracteristics. m 1 78 Ice distribution and dynamics anchoring the fast ice, sometimes creating a seaward bulge extending as far as the 20-m isobath. South of the Port Clarence shoals, the period edges of fast ice draw together and approach the 20-m isobath from the coastal side. At Sledge Island the 20-m isobath makes an abrupt 80° turn to the east, entering Norton Sound. The period-average edges of fast ice follow this isobath past Nome and on to Cape Nome. In this region the edge of fast ice exhibits considerable variability during at least two periods. Hence, the agreement of the period averages should not lead to the conclusion that the edge of fast ice here is stable: in fact, it can occasionally be found well toward shore or beyond the 20-m isobath. Extensive grounded ice piles were observed off Nome during the spring of 1973, well inshore of the 20-m isobath. At that time, attached fast ice extended considerably farther offshore to approximately the 20-m isobath. Although ice is generally moving away from this coast, these ice piles are evidence that at times the ice can be driven against this shore. Inside Norton Sound there is no correlation between the edge of fast ice and the 20-m isobath. From Cape Nome to Cape Darby, the edge of fast ice follows the coast rather closely in waters ciround 16 m deep. At Golovin Bay, just west of Cape Darby, the edge of fast ice shows considerable variability as it bridges the mouth of the bay. At Cape Darby, the edge of fast ice follows the headland closely, making a 90° turn from a north- west-southeast trend to a northeast-southwest trend. The bathymetric configuration here is steep and the possibility of grounded features anchoring the ice decreases rapidly with distance from shore. From Cape Darby, the edge of fast ice is indented at the mouth of Norton Bay, nearly touching the headland at Cape Denbigh. This edge of fast ice follows more or less the 12-m isobath, but as the mid-winter ice-edge map (Fig. 11-4) shows, the edge of fast ice can exhibit some variability across Norton Bay. From Cape Denbigh to Stuart Island the edge of fast ice characteristically bridges the southeastern end of Norton Sound, as far as 50 km from shore in waters very close to 20 m deep. However, the bot- tom of Norton Sound is relatively flat and it is unlikely that depth is a controlling factor in the location of the fast-ice edge here. Fig. 11-7 indicates a systematic trend in location of the average ice edge toward shore with time in this portion of Norton Sound. Analysis of the distribution in each period shows that this trend is reasonably valid. Further- more, the variation from period to period for each year also shows this effect. However, the variation within each period is such that only the midwinter average location of the fast-ice edge can be relied upon with any confidence. The late spring ice edge, for instance, can vEiry from near the shore to nearly 30 km seaward. From Stuart Island, the fast-ice edge follows a westerly course to a point south of Cape Nome, yet 50-60 km offshore from the mouth of the Yukon River. At that point, it makes a broad turn to follow the coast to the south. The seaward extent of fast ice here appears to be determined largely by the location of the Yukon prodelta. The fast-ice edges aire all in the vicinity of an underwater slope between two relatively flat plains at depths of 6 and 12 m. Fig. 11-7 shows that the average location of the fast-ice edge here builds seaward from midwinter to late winter to early spring, then erodes back even further than midwinter during the mid-to-late spring. This pattern appears to be caused by ice piling on the prodelta slope. Although many major ridges can be seen on LANDSAT imagery of the Beaufort Sea, only a few can be seen on Bering Sea images. This portion of the prodelta region is one of the places where such major ridges have been observed. Furthermore, Thor and Nelson (this volume) report a high density of ice-scour features in this area as a result of ridge- keel motions. There is shoal at the very northwestern tip of the prodelta at a reported depth of 7 m. Schertler (1978) has reported floebergs imaged on side-looking air- borne radar and in aerial photographs. Examination of the data suggests that these are rubble piles grounded on this shoal. LANDSAT imagery of this area from 1973 to the present indicates that it is likely that ice grounded on this shoal has acted to anchor fast ice in the prodelta region. From the descriptions given above, it should not be assumed that fast ice around the perimeter is flat and featureless. We have already described an ice-piling event observed off Nome in 1973. Echert (personal communication, 1979) describes three rubble piles within the Norton Sound fast-ice zone: one ap- proximately two miles east of Point Dexter, Norton Bay, 100 X 33 X 8 m in size, at a water depth of 5 m; another approximately two miles west of Stuart Island, 150 X 67 X 8 m in size, at a water depth of 5 m; and a third approximately eight miles north of the Apoon (north-flowing) mouth of the Yukon River, 230 X 100 X 5 m in size, at a water depth of about 3 m. These rubble piles were well within the fast-ice zone and were probably formed during the initial freezing process. Ice behavior within Norton Sound The central part of the Norton Sound basin is Nearshore ice characteristics 179 relatively flat, varying from 18 to 30 m in depth. Nelson et al. (this volume) report observing apparent ice keel gouges at water depths up to 22 m. Hence, at least occasionally, deep-draft ice features may be found within this basin. However, there is also evidence that a signifticant part of Norton Sound is covered by relatively thin ice not likely to pile up enough to attain such great keel depths: a large polynya in the northeastern sector is a virtually con- stant feature of Norton Sound. New ice is usually being formed here and transported westward by winds (see arrows indicating prevailing wind direc- tion on Fig. 11-8) toward the entrance to the sound. Long before it arrives there, considerable thickening takes place through compaction and normal thermal accretion. The polynya is created by almost constant north-northeasterly winds across the northern shore of eastern Norton Sound and nearly constant easterly winds across the western end of the sound (Brower et al. 1977), as shown by the wind direction arrows on Fig. 11-8. Figs. 11-9 and 11-10 are LANDSAT images of portions of Norton Sound, illustrating the range of ice conditions described above. Fig. 11-9, scene number 2399-21443 obtained on 25 February 1976, shows the entrance to Norton Sound. The Sledge Island polynya can be seen left of center. The Yukon Delta can be seen in the lower right corner with extensive fast ice on the Yukon prodelta. Several dif- ferent ages of ice are apparent on the image. The ice within the entrance to Norton Sound and that far outside the entrance are both older than the band of ice extending down the Bering coast past the entrance to the sound. In the few days before this image was taken, the Bering Sea pack moved offshore ap- proximately 20 km, allowing the band of newer ice to form. At the same time, older ice moved out of Norton Sound past the mouth of the Yukon and to the south. Hence, a band of large pans can be seen spilling out of Norton Sound to the south. Fig. 11-10, scene number 2397-21330, was ob- tained two days before the scene just described. This scene shows inner Norton Sound; there is approximately 30 percent overlap with the area of the previous scene. Of interest here is the Norton Sound polynya, which opened up during the recent ice-moving event and has since frozen over with new ice. Nearer the entrance to the sound, there is young ice which was moved out of the polynya area during the event. This ice was probably new ice at that time. This image illustrates the cyclic nature of the Norton Sound polynya. It also shows the variability of the fast ice on the eastern side of the sound: a large chunk of attached fast ice has recently been de- tached and has subsequently been broken into pans. The voids between the pans then froze. A very recent ice-breaking event resulted in a series of fractures running through this region on a southwest-northeast trend. Pack-ice behavior along the western Alaskan coast- Bering Strait to the mouth of the Yukon Various authors (McNutt, Nelson et al., and Pease, this volume) have noted that Bering Sea pack ice is generally moving from north to south past the west coast of Alaska. Although the winds at Nome have a significant easterly component, ice from the north of St. Lawrence Island has been observed to pass around the eastern side of the island and proceed along the western edge of the Yukon prodelta. Two distinct regimes of ice motion are often observed in this vicinity. Cox (personal communication) has pre- pared maps based on satellite imagery showing the Bering Sea pack-ice motion southward past the end of Norton Sound blending with the motion of ice outbound from Norton Sound along a line of shear at the entrance. A distinct line of demarcation can often be found running from north to south, dividing these two ice regimes. Ray (this volume) has also analyzed ice motions in this vicinity in detail and has concluded: In general, the seasonal pack ice in Norton Sound is largely derived in situ, and tends to flow to the west and southwest in response to the prevailing winds or to flow sluggishly (eastward) in response to relatively weak oceanic currents during periods of relatively weak winds. Along the western side of the Yukon prodelta is a region where the seasonal average ice edges nearly coincide. Their location agrees well with the edge of the prodelta where water depths change abruptly from 2 to 12 m. Occasionally grounded shear ridges have been observed along this zone, but their pres- ence has not significantly increased the extent of fast ice. Although Bering Sea pack ice is often driven into this region, the edge of fast ice has not been observed to build out to the 20-m isobath as it does in the Beaufort Sea under somewhat similar conditions. South of the Yukon Delta to Cape Romanzof , the edge of fast ice exhibits a great deal of variability. This is due at least in part to two shoals 60 km offshore at depths of 8 and 10 m and other shoals at intermediate distances. During the winter months, ice appears likely to pile on the outer shoals, forming anchors for extending the fast ice. Just south of Cape Romanzof, the average fast-ice edges are again highly coincident. Then, approaching Nelson Island, they vary significantly with time. 180 Ice distribution and dynamics ,1-^ / ^ ■X^^ '■V ■^ \ Figure 11-9. LANDSAT image obtained 25 February 1976 showing the entrance to Norton Sound. It is difficult to ascribe any particular cause to this behavior. A rather abrupt bathymetric transition from 6 to 16 m crosses this zone. The more closely coincident part of the average ice edge agrees well with this break, whereas farther south the ice edges oscillate across the break. The coincident part lies along the edge of mud flats which mark the prodelta of a former mouth of the Yukon River; the oscillating part lies across the mouth of Hazen Bay. Possibly the edge of fast ice to the north is deter- mined by ice bottom-fast to the mud flats, and to the south the irregularity is caused by tidal currents into and from Hazen Bay. The depth of the bay is be- tween 4 and 5 m and the tidal range, which is as great as 3.5 m diurnally at Cape Romanzof, is sufficient to cause high-velocity tides into and out of the bay. South of Nelson Island, around to Cape Avinof, the seasonal edges of fast ice coincide again. These boundaries appear to coincide with the 8-m isobath. Offshore from here, as far south as the southern side of Nunivak Island, there are several shoals 4-6 m deep as far as 30 km from the coast. Ice passing from north to south through Etolin Strait between the mainland and Nunivak Island often piles on these Nearshore ice characteristics 181 Figure 11-10. LANDSAT image obtained 23 February 1976 showing central portion and eastern end of Norton Sound. Note large polynya at eastern end of sound. shoals, creating relatively large (several km) islands of grounded ice. Ice around Nunivak Island Despite the large flux of ice down the Bering coast and the existence of several relatively shallow shoals (6 m) just to the north, Nunivak Island does not seem to retain an extensive expanse of fast ice. On the north side, facing the flow of ice southward along the Bering coast, fast ice bridges the embayments be- tween Cape Etolin, the peninsula at the northern tip of the island, and Cape Mohican at the western end, passing quite close to each headland and right along the bluff at Cape Mohican. To the east of Cape Etolin, fast ice extends farthest from shore and is found at its greatest depth in the waters near the island. Just to the east of Cape Etolin, the edge of fast ice in all periods comes close to the 20-m isobath. It is likely that ice passing through Etolin Strait is driven into this area, forming extensive expanses of fast ice. However, the period-average edges of fast ice on the east side of the island are all together well 182 Ice distribution and dynamics inshore from the 10-m isobath and quite close to shore. Tidal flushing through Etolin Strait probably causes this behavior. At Cape Corwin, the southeastern prominence of the island, the edge of fast ice is tangent to the curve of land forming the cape. From there, the fast ice extends across a wide bay to Cape Mendenhall, w^here it is again tangent to the coast. From Cape Mendenhall to the southwestern edge of the island, the 20-m isobath is close to the edge of the island. It is unlikely that any fast ice is grounded here. The edge of fast ice is close to shore except for an area where it bridges a wide bay just to the west of Cape Mendenhall. This ice is not likely to be ground- ed, but it remains fast because it is protected from the general north-to-south ice motion past the island. From Cape Avinof to the mouth of the Kuskokwim River, the edge of fast ice follows the edge of extensive mud flats on the north of Kuskokwim Bay. Although some variation can be seen, for the most part the fast-ice edge is consis- tently from year to year and period to period within each season on the finger-like projections of these mud flats. Individual LANDSAT images often show evidence of tidal flushing here. Large blocks of ice are broken loose and transported further offshore or set adrift. Further into the bay, there are several uncharted shoals where fast ice is frequently found. Fig. 11-11 shows LANDSAT scene 1220-21440, taken on 28 February 1978. This scene illustrates the Etolin Strait region, with Nunivak Island on the left and the mountainous Nelson Island left of top center. The scene clearly shows the motion of ice along and away from the coast in this area. There is open water along most of the coast and on the south side of Nunivak Island. Fast ice can be seen on the shoals within Etolin Strait. The southward motion of ice through the strait is illustrated by the polynyas on the south side of these shoals. Farther offshore, older and thicker pack ice can be seen embedded in a matrix of newer ice, illustrating the continuous bre£ik-up of pack ice and formation of new ice in this region. In the mouth of the Kuskokwim River is an embayment of the fast-ice edge which reaches a considerable distance upriver and is consistent in location from period to period. The Kuskokwim is navigable by oceangoing ships far past this point and has a reasonably deep channel. The tidal range here is 5 m with the diurnal range around 4 m. There is little doubt that the large tidal fluctuations are responsible for keeping this area free from fast ice. Around the east side of Kuskokwim Bay to Jacksmith Bay, the edge of fast ice is on shoals 5-10 km offshore. Here the edge of fast ice moves farther from or closer to the shore with the changing seasons. From Jacksmith Bay to Goodnews Bay, the edge of fast ice follows the coast very closely despite the presence of extensive shoals further offshore at depths of 2-4 m. The diurnal tidal variation here is approximately 3 m— probably enough to remove any ice that might be temporarily grounded on these shoals. Winds in this vicinity come characteristically out of the northeastern quadrant and tend to remove ice from this coast rather than to cause ridge-building. From Goodnews Bay to Cape Newenham, the edge of fast ice in winter and early spring (later there is none) bridges a wide embayment with depths on the order of 8 m. There is one shoal at a depth of just over 2 m north of Cape Newenham, but the ice does not appear to be anchored there. From Cape Newenham to Naknek along the northern side of Bristol Bay, fast ice is found only in well-protected embayments at water depths generally less than 4 m. This is largely the combined effect of extreme tidal variations (7 m average diurnal range at Naknek) and strong offshore winds: (1) tidal varia- tions break up and raft away ice not firmly anchored in place; (2) the pack ice is free to move towards the southwest; (3) the prevailing winds are toward the southwest. These conditions generally prevail from Naknek to Egegik Bay. From about Egegik Bay south westward there tends to be more fast ice. However, the pres- ence of ice here depends in part on the severity of the ice year and in part on meteorological events required for ice to occur here. On the first of March in 1974, a year of heavy ice (Niebauer, this volume), winds drove the Bristol Bay pack ice onto the shore of the Alaska Peninsula, creating an extensive area of fast ice, including massive ridges. These ridges were some distance inshore of the 20-m isobath, however, and were probably formed from relatively thin ice. Bristol Bay ice conditions Pack ice in Bristol Bay appears to be greatly influenced by the fact that no barrier exists to keep ice from moving to the southwest. This circum- stance, combined with the presence of strong off- shore prevailing winds around the perimeter of the bay, results in a general southwestward motion of ice out of Bristol Bay. Normally, this motion is so persistent that LANDSAT and lower-resolution satellite imagery nearly always show open water along the northern side of the bay. As explained earlier, fast ice is not extensive and is generally found only in highly protected locations. Nearshore ice characteristics 183 Figure 11-11. LANDSAT image obtained 28 February 1973 showing Nunivak Island, Etolin Strait, and the area of ex- tensive mud flats west of the mouth of the Kuskokwim River. Note ice stranded on shoals in lower Etolin Strait. Due to the nearly constant motion of ice away from the coast and the resulting open water, new ice is often forming along a broad band running east to west all across the northern side of the bay. It is often possible to clearly see the transition from open water to new ice, young ice, and first-year pack ice on a single LANDSAT image. Superimposed on this behavioral pattern is a second characteristic: as the ice moves out of Bristol Bay into a less confined area, it breaks up into large pans with dimensions on the order of 10-20 km. The voids between these pans then freeze. This new sheet may then break up. 184 Ice distribution and dynamics r>//f/*-r- vizis'. :^^ fjf 1/ . - ?>^' r^ s*/ 4S' > *— ''^ir-^" -J. n- f/ 0^ 'kip:. Figure 11-12. LANDSAT image obtained 9 March 1964 sliowing the central portion of the north side of Bristol Bay, followed by the freezing of the new leads and voids. Evidence for several cycles of this activity can often be seen. Fig. 11-12 shows LANDSAT scene 1594-21160, for 9 March 1974. This scene shows ice conditions along the northern side of Bristol Bay. Open water can be seen on the lee side of the land and adjacent islands. Farther offshore, a stepwise gradation to thicker, older ice types can be seen, illustrating that the ice moves in accordance with a series of discrete ice-moving events. This scene illustrates why buildup of extensive fast ice in this region is rare, requiring unusual circumstances. Although the characteristic motion is out of Bristol Bay, occasionally a storm can Figure 11-13. NOAA satellite image obtained 19 March 1975 showing entire Alaskan Bering Sea coast under study. The conditions seen here are typical and tend to support the description of Bering nearshore ice characteristics generated by this study. 185 186 Ice distribution and dynamics drive ice onto the coast. One such event was mentioned in describing the behavior of fast ice along the Alaska Peninsula. Beyond Fig. 11-8, there are four factors re- sponsible for the characteristic behavior of Bering Sea ice: CONCLUSIONS Nearshore ice conditions along the Alaskan Bering coast exhibit a wide range of characteristics from Cape Prince of Wales to the Alaska Peninsula. From Wales to Sledge Island, conditions are largely the same as those along the northern Chukchi Sea coast, while off Nome, fast-ice conditions are often similar to those of Beaufort Sea fast ice. However, further south along the coast, fast ice is found only in shal- low and shielded areas. Finally, along the perimeter of Bristol Bay, fast ice is found only on mud flats and the upper reaches of estuaries. Fig. 11-13 is a NO A A satellite image of the entire Bering coast obtained on 19 March 1975. This image illustrates the relationship between pack ice and fast ice described previously. Fig. 11-8 was compiled to present in one figure the general characteristics of Bering Sea nearshore ice. Although there is considerable agreement between the satellite image in Fig. 11-13 and Fig. 11-8, it should be borne in mind that Fig. 11-13 is a selected, instantaneous satellite image and that the charac- teristics of Fig. 11-8 are based on an average of ice statistics. Although the schematic diagram presented in Fig. 11-8 is generally correct, it should be stressed that specific conditions produce variations from the average. There is a general moderation of climate as one moves southward, resulting in conditions limiting the growth of ice. There is a prevailing motion of pack ice toward the ice front in the southern Bering Sea. As a result, coastal pack-ice movements are either directly away from the coast or along the coast. This motion causes large polynyas to open up in many areas. The prevailing winds are directly offshore from all south- and west-facing coasts. As a result, newly formed ice in the polynyas is transported away from shore. From Nome to King Salmon, the diurnal tidal range increases from 0.5 to 6 m. These tidal variations tend to lift ice away from the sea bottom so that it may be transported by currents and winds. One source of strong seaward currents is the large tidal variation. These currents can be particularly strong in areas of mud flats with large expanses of shallow water. REFERENCES Ahlnas, K., and G. Wendler 1979 Sea ice observations by satellite in the Bering, Chukchi, and Beaufort Sea. In: POAC 1979, Proc. 5th inter, conf. port and ocean engineering under Arctic conditions, I. Trondheim, Norway. Brower, W. A., Jr., J. L. Wise, H. F. Diaz, A. S. Prechtel, and H. W. Searby 1977 Climatic atlas of the outer continental shelf waters and coastal regions of Alaska, II: Bering Sea. AEIDC, University of Alaska, Anchorage. Duprfe, W. R. 1980 Yukon Delta coastal processes study. Final Rep. NOAA-OCSEAP RU 208. Kovacs, A. 1976 Grounded ice in the fast ice zone along the Beaufort Sea coast of Alaska. USACRREL Rep. 7632. Kovacs, A., and M. Mellor 1974 Sea ice morphology and ice as a geologic agent in the Southern Beaufort Sea. In: The coast and shelf of the Beaufort Sea: Proc. sym- posium on Beaufort Sea coast and shelf research, San Francisco. Nearshore ice characteristics 187 Kovacs, A., and D. S. Sodhi 1979 Ice pile-up and ride-up on Arctic beaches. In: Proc. 5th inter, conf. port and ocean engineering under Arctic conditions, Trondheim, Norway. Prichard, R. S., R. Reimer, and M. D. Coon 1979 Ice flow through straits. In: Proc. 5th inter, conf. port and ocean engineering under Arctic conditions, Trondheim, Norway. Reimnitz, E., and P. W. Barnes 1973 Studies of the inner shelf and coastal sedimentation environment of the Beaufort Sea from ERTS-I. NASA Rep. No. NASA-CR-132240. Schertler, R. J. 1978 Report on sea ice radar experiment (SIRE), NASA, Lewis Research Center, Cleveland, Ohio. Stringer, W. J. 1974 1978 Shore-fast ice in vicinity of Harrison Bay. Northern Engineer 5:(4). Morphology of Beaufort, Chukchi, and Bering Seas nearshore ice con- ditions by means of satellite and aerial remote sensing. Final Rep., NOAA- OCSEAP contract no. 035-022-55, Task no. 8. 1976 Marine environmental problems in the ice-covered Beaufort Sea shelf and coastal regions. Ann. Rep. NOAA- OCSEAP contract RU 6-6074, RU 205. Shapiro, L. H., and J. J. Burns 1975 Satellite observations of sea ice movement in the Bering Strait region. In: Climate of the Arctic, Geophys. Inst., Univ. of Alaska. 1979 Morphology and hazards related to nearshore ice in Alaskan Coastal areas. In: Proc. 5th inter, conf. port and ocean engineering under Arctic conditions, Trondheim, Norway. Tucker, W. B., W. F. Weeks, and M. Frank 1980 Sea ice ridging over the Alaskan continental shelf. J. Geophys. Res. (in press). Bering Sea Ice-edge Phenomena Seelye Martin and Jane Bauer Department of Oceanography University of Washington Seattle, Washington ABSTRACT Recent field work at the ice edge during March 1979 shows that the interaction of ocean swell and winds contributes to the nature of the ice edge. Our observations show that the edge divides into three zones. At the seaward edge, there is a zone approximately 10 km wide which consists of heavily rafted and ridged ice floes with heights on the order of 1 m, depths on the order of 5 m, and diameters of 10-20 m. The second zone, which is approximately 5 km wide, consists of rectangular ice floes about 20 m across and 0.5 m thick which have been broken by the ocean swell, but are not heavily rafted or ridged. In the third zone, the ocean swell propagates through the ice without fracturing it, so that the floes have horizontal length scales of km, and thicknesses of about 0.3 m. We also found that when the wind blew strongly off the ice, the outer zone, because of its increased aerody- namic drag, moved downwind ahead of the rest of the pack to form the characteristic bands observed by satellite. These bands, which are on the order of 1 km wide and 10 km long, continued to move south into warmer water until they disinte- grated under the action of warm water and waves. INTRODUCTION In this chapter, we review the ice-edge properties which we observed during our March 1979 OCSEAP Surveyor cruise. In the first part, we show that the ice adjacent to open water divides into three zones, which we call the edge, transition, and interior zones, following Squire and Moore (1980). These zones form because of the interaction of both ocean swell and wind stress with the pack ice. In the second part of the paper, we show that the increased aerodynamic roughness of the edge zone leads to the formation of long, linear bands of ice, measuring about 10 km in length and 1 km in width. These bands form nearly perpendicular to the wind during periods of off-ice winds and move southwest ahead of the pack ice. We document from observations of the formation and movement of a single band the fact that the band moved about 30 km southwest of the pack into warmer water. THE NATURE OF THE ICE EDGE Fig. 12-1 shows the location of the ice cores taken during the cruise, where the southernmost stations show the location of the ice edge. We took the ice cores using a helicopter operating from the Surveyor. In this section, we discuss the ice-edge properties based on the traverse lines labeled W, B, and C, which were respectively occupied on 6, 7, and 9 March. In the next section we first review the general properties of the ice edge, then document the specific ice properties observed on the three traverse lines. Fig. 12-2 shows a schematic diagram of the three kinds of ice, in both plan and side view. First, at the outer edge of the pack, there is open water. Then the edge zone, 5-10 km wide, consists of small broken floes which are about 10-20 m in diameter. In cross section, these floes are heavily rafted and ridged, with sail heights of up to 1 m, and keel depths of 2-4 m. Figs. 12-3a and 12-3b show from photographs taken on 9 March an aerial and surface view of the ice near the edge. The ridge in the foreground of Fig. 12-3b, which is 1 m high and made up of ice 0.1-0.2 m thick, is shown from the air in the center fore- ground of Fig. 12-3a. The reason for the heavy rafting and ridging in the edge zone is that both ocean swell and wind act on the floes to work them against one another. Squire and Moore (1980) show from data taken on the same cruise that the swell propa- gates at least 65 km into the pack. From their data, the observed predominant wave period was 8-10 sec, which yields a wavelength of 100-160 m, so that the 20 m floe diameter is a fraction of a wavelength. The second, or transition zone, about 5 km wide, is characterized by a tiled checkerboard pattern. k 189 Figure 12-1. Chart showing the location of all ice cores taken during the March 1979 Surveyor cruise (chart courtesy Carol H. Pease). EDGE ZONE 5-10KM TRANSITION ZONE INTERIOR ZONE -lOOKM OPEN WATER oooooo°oo ooooooooo SMALL BROKEN FLOES LARGE FLOES SMALL RECTANGULAR FLOES 20m 20m KM 0.2 -0.3 m Figure 12-2. A schematic diagram of the three kinds of ice which occur near the ice edge, proceeding inward from open water. The upper part of the figure shows the ice in plan view; the lower part, in side view. 190 I ^ \ Figure 12-3. Floes in the edge zone, (a) Aerial view; (b) surface piiotograph of the ridge in the foreground of (a). See text for further explanation. i-«i5 191 1 92 Ice distribution and dynamics Figs. 12-7a, 12-llb, and 12-14 show the most dra- matic examples of this patterned ice, which consists of rectangles measuring about 20 m in width and 40 m in length, with their long axes perpendicular to the direction of wave propagation. In cross section, this ice is about 0.3-0.6 m thick; the thicker floes consist of two or three rafted pieces. This zone exists because the outer ice zone reduces the swell ampli- tude to the point that the waves fracture the ice without heavily rafting or ridging it. At the inside edge of the transition zone, we observed an abrupt transition from the patterned ice to large floes measuring kilometers in extent. This third interior zone is the region of large floes where the swell amplitude is reduced so much that it propa- gates elastically without fracturing the ice. Fig. 12-8a shows an aerial view of this ice, where the floes measure kilometers in extent and 0.2-0.3 m in thick- ness. Satellite images suggest that this zone extends far to the north. In support of this general ice-edge picture, we next examine the specific properties along the three traverse lines. Line W The W-traverse took place on 6 March at the positions shown on Fig. 12-1. The wind on this day was negligible and the air temperature was about —5 C. Fig. 12-4 shows a schematic diagram of the kinds of ice observed on the traverse; the ice consisted of an edge zone 8 km wide and a transition zone 3 km wide with an abrupt transition from the rectangular floes to the interior zone at the northern edge of the transition zone. The outermost floe on this line is station Scott. We occupied this floe for a strain-gauge experiment in the morning of 6 March; we were also fortunate enough to have divers from the ship, namely Lt. Comdr. TumbuU, Lt. Williscroft, Lt. (jg.) Fox, and Mr. Kramer, survey the bottom ice profile. Fig. 12-5a shows an oblique aerial photograph of station Scott from an altitude of 60 m after the strain-gauge experi- ment and before the diving. The dark area toward the camera is called the beach, and a small ridge-crack system runs horizontally in the photograph across the floe. The floe measured about 23 m by 37 m. Fig. 12-5b shows a surface photograph of the floe sighting down the ridge-crack system toward the beach; other ridged floes are visible in the background. We de- liberately chose this floe because it was flat, unlike some of the surrounding floes which were heavily ridged. Even though the floe surface appeared flat, the under-ice topography was very rough. To illustrate. LINE W SCOTT 4 6 DISTANCE EAST Figure 12-4. A schematic diagram of the kinds of ice observed along Line W. The circles indicate edge-zone ice; the small squares, transition zone ice. Stations Scott, Wl, and W2 show coring site locations. Fig. 12-6a is a map of the surface made by Vernon Squire, showing the location of the survey lines DEC and BEA; the dashed line shows the approximate place where the floe fractured before the diving. Fig. 12-6b shows the ice thickness from the two under-ice traverses. The divers gathered this information by stretching lines with knots at 10-ft intervals beneath the ice, then recording the depth at each knot from their wrist pressure gauges. Their observations showed that the bottom was very rough, with a maximum depth of 3.5 m. Fig. 12-7a, from the transition zone, shows an aerial photograph of the floes at station Wl. These floes measured 10-25 m on a side; the one on which we landed was 0.66 m thick and showed evidence of rafting at depths of 0.23 and 0.51 m. Fig. 12-7b shows a surface view from our floe; the pressure ridges behind the helicopter were about 1 m high. When we landed on the floe, we could feel the ocean swell propagating through the ice and see the swell- ( I Figure 12-5. Floe Scott, (a) An oblique aerial photograph from 60 m. Dark area toward camera is called the beach; small ridge-crack line runs horizontally across the floe, (b) Surface view sighting down the ridge-crack line towards the beach. I, i 193 1 94 Ice distribution and dynamics HEAVILY RAFTED HEAVILY RAFTED Figure 12-6. The shape of floe Scott, (a) Surface map of Scott (courtesy Vernon Squire); (b) two under-ice cross sections of scott. Letters A, B, C, D, E on cross sections refer to survey lines on (a). induced relative motion of the surrounding floes. We also observed at least seven walruses on the ice v^^ithin 100 m of our station, which is further evidence (see Burns, Ice As a Marine Mammal Habitat, Volume 2 of this book) that these large floes surrounded by open water serve as habitat for walrus. Fig. 12-8a is an aerial photograph from an altitude of 150 m of site W2 in the interior zone, where we landed on the large floe in the center of the photo- graph. The ice further in from this station also resembled this floe; the floes were large and flat, with a low-amplitude swell propagating through them. Fig. 12-8b shows the surface view; the ice was 0.24 m thick and covered by about 5 mm of snow. Line B The B-traverse took place on the next day, 7 March, at the positions shown on Fig. 12-1. As on the previous day, the wind was negligible and the air temperature was about — 5 C. Fig. 12-9 is a sketch of the ice properties along the traverse line; again, the ice consisted of an edge zone measuring about 13 km wide and a transition zone 3 km wide with an abrupt transition from the rectangular floes to the large interior floes. For the stations along the traverse, namely the ship, B4, and B5, Fig. 12-lOa first shows an aerial view of the 90-m-long ship. In the center foreground, a party working on the ice is visible; Fig. 12 -10b shows a close-up of this party from the ship. Again, the floes measured 10-20 m across and coring obser- vations showed that the occupied floe was more than 1 m thick. Many of the floes seen near the ship had wetted surfaces, caused when the swell washes water onto the ice surface. Further into the pack. Fig. 12-lla and 12-llb, taken from an altitude of 300 m, show the ice in the edge and transition zones. In Fig. 12-llb, the rec- tangular broken ice pattern is prominent, and the ice floes are wetted along the recently broken edges. For station B5 in the edge zone, Fig. 12-12a shows an aerial view of the floe on which we landed. This floe was about 20 m across and 0.34 m thick. Fig. 12-12b shows the surface appearance of the surrounding floes, and Fig. 12-12c shows the wetted edge of the floe, with a pen on the ice for scale. Finally, at station B4 in the interior zone, the ice again consisted of large flat floes 0.26 m thick. This traverse line again shows that the smallest, thickest floes occur at the ice edge. \ ^ Figure 12-7. Site Wl. (a) Aerial photograpii from 75 m of nearby floes; (b) surface photograph. ■fc-*^ '■^r 195 Figure 12-8. Site W2. (a) Aerial photograph from 150 m; (b) surface photograph. 196 Bering Sea ice-edge phenomencL 197 or o UJ < Q 16 14 12 10 6- LINE B TRANSITION TO LARGE FLOES m° -^ ICE % EDGE^ SHIP ■ Q I I <0( 2 4 6 8 10 DISTANCE EAST (KM) Figure 12-9. A schematic diagram of the kinds of ice observed along Line B. Arrows mark the coring stations; see also legend for Fig. 12-4. Line C The C-traverse took place on 9 March 1979. On 8 March the weather deteriorated, consisting of blowing snow from the northeast with air temperatures of — 1 to —3 C. On 9 March, the wind was 5.5 m/sec from the northeast at temperatures between —4 and —5 C. The ice was beginning to rot, and there was a strong southerly swell propagating into the pack. Fig. 12-13 shows the traverse line. Because of the wind, the ice was more open along the line, and the smooth pro- gression from small floes to large floes was not evident as we flew north. Rather, the ice consisted of bands of open water and intermingled large and small floes for the first 30 km, at which point we reached the interior zone. We landed at two stations on the traverse line, C4 and C3. C4 was a large floe consisting of ice which was mushy except in the bottom few centimeters, 0.16 m thick with a minimum ice temperature of —3.1 C. On the floe surface, large amplitude 8 sec waves propa- gated through the ice on which we stood. By con- trast, station C3 was on a large floe immediately adjacent to the rectangular broken ice; Fig. 12-14 shows an aerial view from an altitude of 150 m. This picture shows the swell breaking off rectangular floes about 20 m wide from a much larger floe; again the long axis of the rectangular floes was at right angles to the direction of wave propagation. We landed on the large floe shown in the left foreground of Fig. 12-14 to observe that the ice was firm and 0.33 m thick, with a minimum ice temperature of —4.3 C. Therefore, the floe distribution shown in Fig. 12-13 is caused by the fact that when the thin floes are warm, they are more elastic, so that the swell propagates through them without fracturing them; whereas the colder thicker floes, as shown in Fig. 12-14, fracture as the swell passes through them. THE FORMATION AND MOVEMENT OF THE ICE-EDGE BANDS The most interesting effect of the ice distribution described in the previous section is that it leads to the formation of ice-edge bands. Muench and Charnell (1977) review the satellite observations of the bands of ice which form at the edge of the Bering Sea pack ice during periods of off-ice winds. They show from analysis of satellite data that these bands have lengths on the order of 10 km and widths on the order of 1 km, and that the long axes of these bands are generally oriented at 40-90° to the left of the wind. They speculate that these bands form due to surface convergences caused by atmospheric roll vortices. These bands also form at the edge of ice packs in other seas; Campbell et al. (1977, Fig. 5) document their formation in January 1974 in the Gulf of St. Lawrence from Skylab photographs; and this author has seen NOAA imagery of their formation at the edge of the Antarctic pack ice. Fig. 12-15, an aerial photograph taken from the NASA Convair-990 over the Bering Sea in 1973, Julian day 60, 00 h, 06 min, 30 s at 59.87°N, 175.063°W shows these bands (from Anonymous 1973). The aircraft heading was 002- 004° and its altitude was 10.5 km with 70° field- of-view, so that the area depicted measures 14.7 km across. r •v'^r* V^ ^ Figure 12-10. The ice near the ship, (a) Aerial view; (b) surface view of ice party. 198 7 X y / J "^ y f ' , V t \ -y / Figure 12-11. Aerial views from 300 m of the ice along Line B. (a) Edge zone; (b) transition zone. 199 ^w V ■ ixT Figure 12-12. Site B5. (a) Aerial view of floe, surface of which is marked by foot- prints, (b) Surface appear- ance of surrounding floes; (c) surface of floe B5. V ,-♦ 200 C2- 30 C3- o UJ o < 1- 20 C4- 10 s%%^ LINE C KARGE FLOES FjOPEN WATER is^ L307o OPEN WATER RECTANGULAR FLOES OOo OOo L °gg !^ SMALL FLOES OPEN WATER }_LARGE FLOES } LARGE FLOES OPEN WATER OOO oooo OOO O o o Y%o \ FLOES LARGE FLOES SMALL BROKEN V CE EDGE Figure 12-13. A schematic diagram of the l^inds of ice observed along Line C. The scale is half that of Figs. 12-4 and 12-9. The region marked 30 percent open water above Station C3 consisted of large floes with leads. Figure 12-14. Site C3. Aerial view from 150 m floes broken by swell adjacent to large floes on which we landed. Helicopter antenna is at lower right. 201 202 Ice distribution and dynamics In Fig. 12-15, beneath the cloud bands which are aligned approximately parallel to the wind, ice bands are visible with their long axes aligned at approxi- mately right angles to the wind. The bands are 10-13 km long and about 0.6 km wide at their widest point. For the same day, Gloersen and LaViolette (1974) show the surface pressure map and a U.S. Air Force Weather Service Satellite image of the Bering Sea ice. The satellite image shows numerous ice bands at the edge; the weather map shows that the winds are ap- proximately from the northeast. Coast Guard weath- er data gathered at the same time (from Campbell et al. 1975) show that the surface air temperature was — 10 C and the wind speed was 5 m/sec. How do these bands form? In the previous section, we showed that the effect of wind and waves at the ice edge was to work the ice in the edge zone so that it consists of numerous small floes, heavily rafted and ridged. Therefore the pack ice has an outer band of much thicker ice, with considerable top and bottom topography, adjacent to the open water. We there- fore carried out from the Surveyor a field experiment during a period of off-ice winds to study the move- ment of this thicker edge ice relative to the thinner ice in the transition zone. We found that because of the increased aerodynamic roughness of the ice in the edge zone it moved faster downwind than the interior ice. This relative velocity increase leads to the formation of the observed bands. The experiment On the morning of March 11, after a day of weak easterly winds which compacted the ice edge, we used the helicopter to place six targets on ice floes which were then tracked over a 23-hour period and their position determined from the Global Navigation System (GNS-500A) on the helicopter. Each target consisted of a colored nylon rectangle about 1X2 m^ , nailed to the ice. Also, on three of the floes, we set up wooden poles 2 m high with radar reflectors mounted on top. Although we found that the re- flector signal could not be distinguished from the ice on the radar of the ship, the poles served as good visual targets. All of the floes tagged were within 3 km of the ice edge. The floes tagged with yellow, purple-1, red, Figure 12-15. Aerial photo- graph from 10.5 km of the ice edges bands; north is to the top. See text for further explanation (photograph courtesy NASA). Bering Sea ice-edge phenomena 203 and green were snow-covered and rafted cakes meas- uring about 20 X 15 m^. Fig. 12-16 shows the appearance of the floe with the green target (the green floe) shortly after the target was placed. The floe had a smaU ridge running down the middle with a rafted, snow- free area to the left. The purple-2 floe also measured 20 X 15 m^ , and had a 0.6 m ridge to one side. Finally, the blue floe, shown in Fig. 12-17, measured only 3 X 5 m^ and had slightly raised edges. Although we did not measure the bottom topography of these floes, when we drilled holes 0.4 m deep for the poles, we did not reach bottom. We assume from our discussion in Section 1 that the floe thicknesses were on the order of 1-5 m. We initially placed the targets in a cross, with the yellow, purple-1, red, and purple-2 floes aligned along the bearing 140-320° over a distance of 1.4 km. On the other arm, the green, red, and blue floes were aligned along 042-222° over a distance of 1.2 km, as shown at the top of Fig. 12-18. We then overflew the targets at 4, 9, and 23 hours from the time of place- ment and recorded their position. We also deter- mined the distance and bearing from each target relative to the red floe. We lost some of the floes; purple-1 was not seen at four or nine hours, and blue disappeared after four hours. The evolution of the targeted floes was as follows: when the targets were first placed, the ice edge was compact, with only a few ice cakes in the open water to the south, proba- bly because of the weak easterly winds on the day before the experiment. Fig. 12-19 is a photograph looking from the ice toward the open water at this time, where in the original photograph the radar target mounted on the blue floe is visible in the middle of the picture. Table 12-1 shows the averaged wind data for the 23-hour period. At four hours, after the wind began to pick up from the north, the targeted floes began to move away from the edge, creating an area with a 25 percent ice concentration inside an outer band of 90 percent concentration, as shown in Fig. 12-20. As Fig. 12-18 shows, the outer ice band formed a hook outward from the pack ice. The yellow floe was in the 25 percent ice concentration region, while the other targets were in the inner high-concentration region. At nine hours (twilight so that photographs were impossible) all of the targeted floes were inside a band of ice floes 1-2 km across, several hundred meters southwest of the main pack. Our final 23-hour survey was on the morning of the next day. Figure 12-16. Aerial view of the green floe; panel measures Ix 2m . i** ^ 204 Ice distribution and dynamics Figure 12-17. Aerial view of the blue floe; pole is approx- imately 2 m high. 9hr Approximate ice edge (23 hr) , Direction of swell 270° n0.5m/sec 23 hr RED BLUE YELLOW GREEN PURPLE -2 PURPLE -t Km Figure 12-18. A schematic diagram of the observed ad- vance and relative positions of the targeted ice floes during the 23-hour observation period. TABLE 12-1 Averaged wind data for the 23-hour period Time Period Average Speed Average Bearing (m/sec) (true degrees) 0-4 4.9 021 4-9 7.5 000 9-23 10.5 006 0-23 8.8 008 At this time, we observed many bands of ice similar to those shown in Fig. 12-15. Figs. 12-2 la and b show two aerial views of a band near the ship similar to the targeted band. The targets were in the band shown in Fig. 12-18, which was now about 15 km south of the ice edge. Fig. 12-22 shows a schematic drawing of the band and the target positions at 23 hours. The band was about 2 km wide at its widest part, and 8-10 km long. It was widest at the end toward the swell, and had a long, curving tail. Fig. 12-23 shows the way the band looked toward the tail; and Fig. 12-24 shows the way the head looked, sighting against the direction of swell propa- gation. Because of the action of the swell and the wind waves, this part of the band consisted of many small fragments of ice, pancake and larger, mostly Figure 12-19. Aerial view from 75 m looking toward the ice edge as it appeared just after the placement of the targets. The blue floe is barely visible in the middle of the photograph. '-''^ "*-a<^~- . "> ..^ .->-■' Figure 12-20. Aerial view from 150 m of the ice edge as it appeared four hours after the placement of the targets. The targeted ice floes were mostly in the higher concentration of floes shown in the foreground. 205 Figure 12-21. Aerial view of a band of ice floes as it appeared the morning of March 12. to the band; (b) a section of the same band which was about 3 km east of the ship. (a) the 90-m-long Surveyor next 206 Bering Sea ice-edge phenomena 207 submerged floes which were covered with small pieces of ice. Finally, Fig. 12-25 shows the appearance of the green floe at 23 hours. To summarize, during our observational period, the red floe moved from 58°56.6'N, 170°4.7'W to 58°41.9'N, 170°22.4'W, so that it travelled 32 km in the direction 210°, or at approximately 25° to the right of the average wind, at an average speed of 0.38 m/sec, or at 4 percent of the mean wind speed from Table 12-1. The two major forces acting on the band, which account for its relative motion, are aerodynamic wind drag on the rougher floes in the edge zone, and the force which results from absorption and reflection of wind waves and swell. Due to the shape and rough- ness factors, the thicker, rougher ice in the edge zone has higher drag coefficients than the thinner, smoother ice found inside the edge. For example, Banke et al. (1976) found through measurements that sharp-edged, rough pancakes have air -ice drag coef- ficients two to four times larger than smoother, thinner floes. From a force balance, this suggests that at the same wind speed the thicker, rougher ice will move slightly faster than the smooth ice. \o-ru ' NORTH WIND IRECTION SWELL DIRECTION Figure 12-22. A schematic diagram of the band of ice in which the targeted ice floes were found after 23 hours. The location of the floes is marked using the same notation as in Fig. 12-18. The dotted portion represents the edges of the band characterized by small fragments of ice, pancakes, and larger, mostly submerged floes. Thus, for wind speeds less than 5 m/sec, like those during the first few hours of the experiment, the thicker ice might move 1-2 cm/sec faster than the thinner ice. At wind speeds greater than 10 m/sec, this value may increase to a 10-20 cm/sec difference in speeds. These numbers confirm what we observed: before the experiment the winds were low and easterly, so that due to the Coriolis force the ice motion would have been to the northwest. In ad- dition, momentum transferred from the swell to the ice by reflection and absorption would have helped to form the compact ice edge which we observed at the start of the experiment. Then during the first four hours the winds, as Table 12-1 shows, came from the north as described above, and the targeted floes moved slightly south of the main edge. A speed difference of 1 cm/sec over four hours would cause the faster-moving ice to move about 150 m ahead of the slower ice. During the next five hours, the winds remained northerly and increased in speed, thereby increasing the difference in speeds and the distance between types of ice. In the final fourteen hours, the winds remained near 10 m/sec and the targeted floes were found some 15 km south of the ice edge. This requires an average difference in speeds of about 25 cm/sec. This number appears to be quite high when only wind and water stresses are considered; we must also consider the transference of energy and momen- tum from waves to the ice. Once the bands move away from the immediate ice edge, the absorption and reflection of wind waves also exert a force on the ice. Longuet-Higgins (1977) shows that the force per-unit-width exerted on a floating mat by waves was equal to V2 pg (a^ + a'^ — b^ ), where p is the water density, g the gravitational acceleration, a the incident wave amplitude, and a' and b the reflected and transmitted amplitudes respectively. Wadhams (1973), through experiments with petri dishes in a wave tank, shows that the energy of short-period waves is almost all reflected or absorbed while that of longer-period waves is mostly transmitted with an exponential energy decay with distance into the ice. Dean and Harleman (1966) also suggest that for stationary and moored obstacles, the reflection coefficients are generally quite large when the horizontal dimensions of the obstacle are larger than half the incident wavelength. Thus, the re- flected short-period wind waves enhance the motion of the band away from the main ice edge and also help to compact the ice into a band, because this force is primarily applied to the windward edge of the ice. Similarly, the enlarged head of a band may be the result of the absorption of swell energy at the leading edge of the band. These wave effects can enhance the ice speed by as much as 10 cm/sec, which brings the speed difference roughly to the desired 25 cm/sec. Thus, wave effects will contribute to the ice motion in the vicinity of the ice edge. Figure 12-23. An aerial view from 75 m of the tail of the band in Fig. 12-22. Figure 12-24. A view from 25 m of the small fragments, pancakes, and submerged floes at an edge of the band being acted upon by wind waves and swell and repre- sented by dotted line in Fig. 12-22. 208 Bering Sea ice-edge phenomena 209 -^-^p^ Figure 12-25. An aerial view from 50 m of tiie green floe as it appeared 23 hours after the target placement. Consequences of the band motion Table 12-2, which Usts the air and water properties measured near the band during the period of our observations, shows that as the band moved south, it moved into warmer, more saline water. At the beginning of the experiment the water temperature was —1.2 C, and the salinity was 31.9°/oo. At the end of the 23 hours, the ice was in water of —0.45 C and a salinity of 32. 2*^/00. In both cases, the freezing temperature of the water was about —1.7 C. The water temperature ranged from +0.5 to +1.3 degrees above freezing and we observed the ice to melt. TABLE 12-2 Water and air properties observed from the ship near the band Elapsed Air Water Salinity Freezing Time Temperature Temperature Point (h) rc) CO 0/00 (°C) 0 -0.70 -1.20 31.94 -1.73 4 -1.09 -1.30 32.05 -1.73 9 -1.79 -0.62 32.11 -1.73 23 -4.50 -0.45 32.23 -1.74 Average -3.14 -0.60 32.11 -1.73 Since the band was relatively long and thin, lateral melting may have been of some importance, but vertical melting was most likely dominant because of the larger exposed surface area. Another form of melting is wave-induced; the waves fracture the ice floes, thereby creating for each floe a larger ratio of surface area to volume. Waves breaking over the ice can also cause melting of the upper surface. Wave- induced melting was visible along the windward and swellward edges of the band (Fig. 12-24). In sum- mary, our observations suggest that these long thin bands of small floes rapidly melt in the warmer water. From satellite images one can perceive the im- portance of this melting regime on a larger scale. In consecutive images, interior ice features are seen to move to the south or southwest throughout most of the eastern Bering Sea in late winter and early spring, as reported by Muench and Ahlnas (1976) and by McNutt in this volume. However, the fact that the southern edge does not appear to advance south during much of this time suggests that the bands carry away enough ice to the southwest, where the ice melts, to maintain the ice-edge position. If this is indeed what is occurring, then a very large amount of ice is melted near the southern ice edge during periods of strong northeast winds. 210 Ice distribution and dynamics CONCLUSIONS ACKNOWLEDGMENTS From analysis of satellite data, several investi- gators, e.g., Muench and Ahlnas (1976), McNutt in this volume, and Loshchilov (1973) conclude that much of the ice in the Bering Sea forms in the north, then is conveyed to the southwest by the strong northeast winds which accompany the anticyclonic circulation of the Siberian high-pressure system. The present investigation of the southern ice-edge proces- ses shows in part how the ice-edge position is main- tained. The large ice floes of the pack ice are advected southwest by the wind. As they approach the ice edge, first the ocean swell propagating into the pack fractures them into the characteristic rectangular pattern; then in the edge zone the floes are heavily rafted and ridged, which leads to a great increase in aerodynamic roughness. Because of this increase in roughness the ice in the edge zone moves away from the pack with a greater relative velocity, and dis- tributes itself into the characteristic bands. These bands continue to move southwest into warmer water until they disintegrate. Further, the southward movement and disintegration of the bands exposes the transition zone to higher swell amplitudes, so that this zone becomes the edge zone and in turn becomes heavily rafted and ridged, and blows southwest. In our observations, a band moved 32 km in 23 hours at an average speed of 0.38 m/sec into water that was 1.3 C warmer than its freezing temperature. Thus the formation and movement of the ice bands are one way in which the pack ice is rapidly dispersed and melts. Moreover, these bands may play an important role in the dispersal of pollutants. First, oU spilled to the north which is trapped in the ice may enter this conveyor belt and be released 30-50 km south of the pack as an ice band disintegrates. Second, the bands generally had sharp leading edges and diffuse trailing edges, which suggests that the band moves faster than small pieces of ice or oil slicks. Therefore, for ex- ample, an oil slick southwest of one of these bands may be overtaken by the band and carried south to the point where the band disintegrates, so that the slick may spread over a larger area than would be anticipated from wind action on the slick alone. DEDICATION We dedicate this chapter to the memory of Robert Lewis Charnell, lost at sea near Hawaii in December 1978, who gave us help and encouragement during the planning phase of this ice-edge experiment. We thank Capt. James G. Grunwell and the officers and crew of the NOAA Surveyor for their help in carrying out the field operations. We are particularly grateful to Lt. Bud Christman, the pilot of the Bell 206, who flew his helicopter in support of scientific operations for 10 of the 13 days which we spent at the ice front. Mr. Peter Kauffman designed and built much of the equipment used in the field operations, and participated in most of the field work described in this chapter; we greatly appreciate his support. We also thank Mr. William Abbott and Dr. Per Gloersen of the NASA Goddard Space Flight Center for use of the photograph in Fig. 12-15. Most of this work was supported by the Bureau of Land Management through an interagency agreement with the National Oceanic and Atmospheric Administration, under which a multiyear program responding to the needs of petroleum development of the Alaska continental shelf is managed by the Outer Continental Shelf Environmental Assessment Program (OCSEAP) office. S. Martin also gratefully acknowledges the support of the U.S. Department of Commerce Contract No. 78-4335 for the analysis of the ice core data, and J. Bauer also gratefully acknowledges the support of the Office of Naval Research Under Task No. NR307-252 and Contract No. N00014-76- C-0234. Publication No. 1119 of the Department of Oceanography, University of Washington. REFERENCES Anonymous 1973 U.S.-U.S.S.R. Bering Sea Expedition, Convair-990 Navigational Flight Data. Banke, E. G., S. D. Smith, and R. J. Anderson 1976 Recent measurements of wind stress on Arctic Sea ice. J. Fish. Res. Bd. Can. 33: 2307-17. Campbell, W. J., P. Gloersen, and R. O. Ramseier 1975 Synoptic ice dynamics and atmos- pheric circulation during the Bering Sea Experiment. In: U.S.S.R./U.S. Bering Sea experiment, K. Ya. Kon- dratyev, ed., 164-85. Gidrometeoiz- dat, Leningrad. Bering Sea ice-edge phenomena 211 Campbell, W. J., R. O. Ramseier, R. J. Weaver, and W. F. Weeks 1977 Sky lab floating ice experiment. Miscellaneous Spec. Pub. No. 34, Dep. of Fisheries and the Environment, Fish, and Mar. Serv., Ottawa, Canada 1977. Dean, R. G., and D. R. F. Harleman 1966 Interactions of structures and waves. In: Estuary and coastline hydro- dynamics, A. T. Ippen., ed., 341-403. McGraw-Hill, N. Y. Gloersen, P., and P. E. LaViolette 1974 Satellite imagery and weather for the BESEX area, 15 February through 10 March 1973. Longuet-Higgins, M. S. 1977 The mean forces exerted by waves on floating or submerged bodies with applications to sand bars and wave power machines. Proc. Royal Soc. London, A352: 463-80. Loshchilov, V. S. 1973 Characteristics of the ice cover in the operational area of the "Bering" expedition. In: Prelim, results of the Bering expedition, 18-30. Muench, R. D., and K. Ahlnas 1976 Ice movement and distribution in the Bering Sea from March to June 1974. J. Geophys. Res. 81: 4467-76. Muench, R. D., and R. L. Charnell 1977 Observations of medium-scale features along the seasonal ice edge in the Bering Sea. J. Phys. Oceanog. 7: 602-6. Squire, V. A., and S. C. Moore 1980 Direct measurement of the attenu- ation of ocean waves by pack ice. Nature 283: 365-8. Wadhams, P. 1973 The effect of a sea ice cover on ocean surface waves. Ph.D. Dissertation, Cambridge Univ. Eastern Bering Sea Ice Dynamics and Thermodynamics Carol H. Pease NOAA Pacific Marine Environmental Laboratory Seattle, Washington ABSTRACT During winter 1979, hydrographic, meteorological, and ice floe data were collected over the Bering Sea shelf. The ice pack extended to 59° N; however, there appeared to be little or no in-situ freezing in the study area. Hydrographic data from the marginal ice region (seaward limit of the ice) showed that less saline, cold (s— 1.4 C) waters existed in an upper layer; the lower layer was as much as 1 C warmer. Floes advected toward the south to southwest at rates as high as 0.5 m/sec during north-to-northeast wind events. Floes rotted along the margin in periods on the order of days. Little ridging of ice was observed over the open shelf. Rafting was prevalent among floes battered by wind and swell at the ice edge. We observed that in the fall, northerly winds cool the water of Norton Sound and the Bering Sea north of St. Lawrence Island until the water column is isothermal at freezing temperatures. Further cooling causes freezing. Under northerly wind conditions, ice is advected south into water where it is no longer in thermodynamic equilibrium. The resulting meltwater is mechanically mixed and is a source of cooling for the waters of the southern Bering shelf. These observations suggest that ice formation and movement in the Bering Sea can be likened to a conveyor belt: growth occurs primarily in the north, advection due to wind stress is gener- ally southward, decay occurs at the thermodynamic limit, and the limit advances somewhat as meltwater cools the upper layer. INTRODUCTION On a cruise of the NOAA ship Surveyor during the first two weeks of March 1979, many types of oceanographic, meteorologic, and ice floe data were collected in order to identify processes inherent in the distribution and condition of sea ice over the Bering Sea shelf (Fig. 13-1). These data, in con- junction with simultaneous edge-specific studies by Martin and Bauer (Chapter 12, this volume), remote sensing studies by McNutt (Chapter 10, this volume), and previous work by Muench and Ahlnas (1976) and Ahlnas and Wendler (1979), yield a well-formed description of mesoscale interaction between dynam- ics and thermodynamics in controlling the pack ice conditions in the Bering Sea. This chapter describes and presents supporting evidence for this interaction. 64''N Chukotsk \ Bering -P.eninsulay5/w//^5g^g^(j Peninsula 76° 172° 168° 164° 160° I 56°W Figure 13-1. Eastern Bering Sea ice study area including the observation area during the cruise of the NOAA ship Surveyor during March 1979. ICE-EDGE HYDROGRAPHY Forty-one CTD's (Fig. 13-2) and twenty-two arrsondes (Fig. 13-3) were taken along the ice edge during the first two weeks of March 1979 from the NOAA ship Surveyor. Oceanic and atmospheric surface temperatures were taken every hour, and surface salinity was taken every two hours. The water temperatures were at first taken as bucket temperatures with a calibrated thermometer (±0.2 C), but after the thermometer was lost, a calibrated 213 214 Ice distribution and dynamics 174° 173° 172 69° 168° 167° 166' I 74= 173° 17 2= 171° 170° 169° 168° 167° 166° 165= Figure 13-2. The location of CTD stations taken during the first two weel^s of March 1979 from the NOAA ship Surveyor. thermistor string (±0.1 C) was used. This method also provided air temperature measurements near the surface. Although temperatures were measured within a few centimeters of the air-sea interface, they are considered representative of more extensive layers since turbulence from the ship's wake provided mechanical mixing. Salinity was determined from bucket samples using an auto-salinometer. With a motorized psychrometer, hourly dry-bulb and air temperatures were taken at the bridge throughout the cruise. However, problems were encountered with the wet-bulb thermometer readings and the data were discarded as unreliable. The bridge measurements were made approximately 10 m above the surface. In general, the water temperature was above the freezing point of the ice except for that measured in CTD cast No. 1, 80 miles south of Nunivak Island in 30 m of water on March 2. Floes were rotting from the bottom west of this area. Melt puddles were not formed as in the traditional summer Arctic melt condition, but some floes were wetted when waves washed over them. Details of the mechanical effects of the swell on the ice and resulting edge processes are discussed by Martin and Bauer (Chapter 12, this volume). A major feature of the water column along the ice edge was fresh cold water due to ice melt. CTD sections perpendicular to the ice edge west of the Nunivak area showed a meltwater lens (Fig. 13-4). The pycnocline downwind of the pack edge was slightly deeper than that under the ice itself, ap- parently because ice cover protects the water column from wind-mixing; but the exact shape of the front is speculative. The ice in this figure was in surface waters of =1.0 C, although ice was observed in slightly warmer waters at other times. We also observed the effect of melting by ex- Ice dynamics and thermodynamics 215 Figure 13-3. Surveyor. The location of radiosonde launches made during the first two weeks of March 1979 from the NOAA ship amining the change in the distribution of sea surface temperature (Fig. 13-5). All surface measurements during the two-week cruise are plotted on the same chart and contoured for the first and second weeks sepEirately. The eastern end of the figure shows the change observed over a two-week period, while the western end shows almost steady-state conditions. Note that while the —1.0 C isotherm moved about 35-40 km south during the cruise, the +1.0 C iso- therm did not move appreciably. The salinity field showed the same trend; although the contouring is less significant with only half as many samples (Fig. 13-6), the advance of meltwater is strikingly apparent. It is possible that movement of meltwater toward the south could be due to advection of the water mass itself. However, the mean currents on the shelf are weak and northwesterly, generally in opposition to the advance of meltwater (Kinder and Schumacher, Chapter 5, this volume). Contours of dynamic height during the March cruise period also suggest weak flow toward the northwest (Fig. 13-7). Although this dynamic topography has less relief than that shown previously by Charnell et al. (1979), it is similar in inferred speed and direction. Also, the pattern suggested by Figs. 13-5, 13-6, and 13-7 is coherent over periods longer than a tidal cycle. The amount of ice melt required to lower temper- ature to observed values (Fig. 13-4) was estimated by assuming that the water had been +1.0 C and that the ice had already warmed to its freezing point. Under these assumptions, enough heat was extracted to melt a 60-km-long, 0.5-m-thick strip of ice of 15°/oo salinity. This melt estimate was probably too high since off-ice winds also cool the water, but it shows that the proposed physical mechanism is capable of producing the observed hydrographic features. 216 Ice distribution and dynamics 5 8° 46 9 N I70°25,3'W 36 31 TEMPERATURE (°C) 35 34 58°56.9' N I70°068'W 33 -I 4 -■°-^ 58° 46,9 N 170= 253 W 36 31 58°469'N I70°25,3'W 36 31 - c ^ "^25,8 — ' 25< - 25,9 10 km Figure 13-4. A CTD cross section perpendicular to tiie ice edge showing the meltwater lens typical along the margin of ice. The ice existed in water of about — 1.0 C and colder. The CTD station numbers correspond to those in Fig. 13-2. DISCUSSION Ice forms in situ during a typical ice-year in late fall (November-December) in Norton Sound, in the Bering Sea north of St. Lawrence Island, along the Alaskan coast, and eventually southward into Bristol Bay (Fleet Weather Facility 1972-1975, 1976, 1977, and 1978). These areas are shallow (less than 30 m), except for the region north of St. Lawrence Island, and water is observed to be isothermal at the freezing point (about —1.7 C). Cooling is caused by a negative radiation balance and sensible and latent heat transfer due to offshore winter winds. Under the influence of northerly to easterly winds, the ice floes are trans- ported into warmer water. The leading floes melt, adding relatively cold, fresh water to the existing water column. Under continued northerly winds, the new leading floes encounter colder waters and thus have longer residence times. As the season progresses, regions of ice growth expand and the floes are con- tinuously advected beyond these areas. Satellite and aircraft photographs of the Bering Sea in winter show persistent polynyas in the lee of Cape Nome and St. Lawrence, St. Matthew, and Nunivak Islands during periods of northerly to northeasterly winds (Muench and Ahlnas 1976, and McNutt, Chapter 10, this volume). Less commonly, a polynya can also be observed in eastern Norton Sound during more easterly wind events. Gray streaks, which Martin £ind Kauffman (1979) attrib- ute to grease ice production, are often observed in satellite images of these polynyas. Floes 0.3-0.5 m thick near the ice edge but not yet rafted are formed of frazil in approximately the upper half, while the lower half grows in place as columnar crystals (Martin and Kauffman 1979). This indicates that floes con- tinued to grow for a time after the grease ice had con- solidated. Overflights show that the floes do not change much in thickness after they are advected away from shore and south of about 62°N until they are rafted and melted at the pack edge. Thus, we see areas of growth, relative thermal stability and trans- port, and thermal and mechanical destruction. These roughly correspond to the regions described by Muench and Ahlnas (1976) in the spring of 1974. Be- cause of these characteristics, the ice can advance 200-300 km beyond the fall growth region, while the growth region itself expands about 100-200 km be- yond its fall extent. Spring is heralded in the Bering Sea by shifting of the storm tracks north from the Gulf of Alaska to the southern Bering Sea (Overland, Chapter 2, this volume). Given sufficient time, the radiation balance alone would melt the ice pack, but the shifting of the winds from cold northerly to warm southerly can accelerate this procedure considerably. The entire pack can become uniformly rotten in three or four weeks after low-pressure systems penetrate the southern Bering Sea. A wind field during rotting Ice dynamics and thermodynamics 21 7 I Figure 13-5. Hourly surface water temperatures between 2 March (JD-61) and 15 March (JD-74) 1979 from the NOAA ship Surveyor, contoured for the westward and eastward trelts respectively. Note that the +1.0 C isotherm did not move signifi- cantly while the 1.0 C isotherm moved 35-40 km due to the inclusion of meltwater. conditions occurred in late March 1979 (Fig. 13-8). The accelerated rotting was observed in winter 1974, a year with extensive ice cover (Muench and Ahlnas 1976), and in 1979, a year with light ice cover (McNutt, Chapter 10, this volume). The sea ice is generally divergent on the open Bering shelf because the wind field is slightly di- vergent when winds are northeasterly and the basin is less constrictive downwind. A northeasterly wind field occurred in early March 1979 (Fig. 13-9). The divergence resulted in little or no ridging of Bering Sea ice compared to the Arctic pack. The few observed ridges formed mainly windward of St. Lawrence Island and in southern Norton Sound. These features generally had low relief because there was little ice to build ridges. Parmenter and Coon (1973) showed that the maximum height of ridges is a function of ice strength and thickness of con- tributing floes. Since the maximum ridge thickness is about 10 times that of contributing floes, we expect 3-5 m ridges from floes 0.3-0.5 m thick around St. Lawrence Island and possibly 10-20 m ridges from floes 1.0-2.0 m thick in southern Norton Sound. Vertical water-column structure over the eastern Bering Shelf is complicated. It appears that in areas where the shelf is less than 50 m deep, the entire water column is tidally mixed. In water deeper than 50 m, the structure is two-layered with a pycnocline generally at about 20 m. The upper layer is mixed by wind and the lower layer by tides. Kinder and Schumacher discuss this in detail (Chapter 4, this volume). In ice-growth regions, vertical mixing is enhanced by the extrusion of salt during the freezing process. Since typical ice salinities are 10*^/oo or greater (Martin and Kauffman 1979), about 20°/oo is extruded. A typical advection rate for floes is about 218 Ice distribution and dynamics 168° Figure 13-6. Bihourly ocean surface salinities between 2 March (JD-61) and 15 March (JD-74) 1979 from the NOAA ship Surveyor, contoured for the westward and eastward treks respectively. Note that the lighter isohalines have moved south during the two-week period while the heavier isohalines have not appreciably moved. 0.25 m/sec (McNutt, Chapter 10, this volume, Muench and Ahlnas 1976), and we assume steady- state production versus advection of ice 0.3 m thick with a moderately persistent northerly wind regime for about four months. This is equivalent to growing over 5 m of ice in the growth region during the season. Growing this much lO^/oo ice would in- crease the salinity of 20 m of water by about 5-6°/oo over the season. With different choices for seasonal extent, velocities, and persistence, the range of salinity increase over the season is from 1.25°/oo to 7.5°/oo. However, since the mean current is from the melt zone toward the growth zone, one would expect some of this salt to be recycled, thereby partially masking the salinity-enhancing process. The lower limit of these values is in agreement with those observed by Coachman et al. (1978) for the shoal region southwest of Nunivak Island. A similar calculation can be made of the number of times the ice pack replaces itself. At a floe speed of 0.25 m/sec, assuming the pack extends 300 km south of Nome and persists for four months, the pack re- places itself about eight times in the season. With various choices for seasonal extent, velocities, and persistence, the range of possible replacements is from two to ten. It should be remembered, however, that the cycling of the pack is continuous. The number of replacements is an indication of the dynamic and thermodynamic vigor of the system. SUMMARY AND CONCLUSIONS During the winter of 1979, ice was observed to form in the northern and coastal regions of the Bering Sea, to advect south-southwesterly, and to melt along Ice dynamics and thermodynamics 219 170" 169° 168° 167° 166° 165° •60° 174° Figure 13-7. Dynamic topography between 2 March and 15 March 1979 computed from CTD stations illustrated in Fig. 13-2. Note northwesterly set of lines of constant topography. This implies a relatively weak northwesterly current. the southern margin; ice processes over the Bering Sea shelf can thus be thought of as a conveyor belt. The interaction between the dynamic and thermo- dynamic processes of decay along the margin was ex- plored for March 1979 using measurements taken from the NOAA ship Surveyor. This analysis and remote sensing analysis by McNutt (Chapter 10, this volume) show that the same processes were at work during 1979, an extremely light ice year, as in 1974, a medium-heavy ice year (Muench and Ahlnas 1976). This suggests that the extent of the pack ice from one year to another is a function of both the advective scheme and thermodynamic processes. The conveyor belt process of ice growth and decay is responsible for enhancing salt content in northern and coastal waters (ice production areas) and reduc- ing salinity, resulting in stratified waters along the ice margin. This process has a considerable effect on the structure of water properties of the shelf, at least during the winter season. The vigor of the advective scheme, i.e., wind blowing the ice, also affects the extent of freezing. ACKNOWLEDGMENTS This study was funded through the Marine Services Project at Pacific Marine Environmental Laboratories. The cruise time on the NOAA ship Surveyor was requested by the late Robert Charnell of PMEL and arranged by the OCSEAP Juneau Project Office. Dr. Seelye Martin of the Department of Oceanography at the University of Washington acted as chief scientist on the cruise. The routine oceanographic measure- 0BSERVED SLP AND WINDS ^ OBSERVED SLP AND WINDS 4 OOZ 28 MRR 1979 0BSERVEO SLP AND WINDS ^ ODZ 29 MflR 1979 0BSERVED SLP RNO WINDS 4 12Z 28 MflR 1979 OBSERVED SLP AND WINDS 4 12Z 29 MAR 1979 0BSERVED SLP AND WINDS 4 OOZ 30 MflR 1979 12Z 30 MAR 1979 Figure 13-8. Sea-level pressure and wind fields calculated from digitized NWS-Alaska region surface analysis charts for 28, 29, and 30 March 1979. This pattern is the dominant spring condition driving the ice northerly and warming the basin. Longest vectors represent 20 m/sec winds. 220 SERVED SLP RND WINDS 4 0BSERVED SLP AND WINDS 4 OOZ 1 MflR 1979 0BSERVED SLP FIND WINDS 4 OOZ 2 MRR 1979 35ERVED SLP AND WINDS 4 12Z 1 MRR 1979 0BSERVED SLP AND WINDS 4 12Z 2 MRR 1979 0BSERVED SLP AND WINDS 4 OOZ 3 MflR 1979 12Z 3 MAR 1979 Figure 13-9. Sea-level pressure and wind fields calculated from digitized NWS-Alaska region surface analysis charts for 1, 2, and 3 March 1979. This pattern is the dominant winter condition driving the ice southwesterly and cooling the basin. Long- est vectors represent 20 m/sec winds. 221 222 Ice distribution and dynamics merits were carried out by the survey technicians on the Surveyor, who patiently modified their opera- tions to meet my requirements. Frances Parmenter of the NESS field office in Anchorage, Bruce Webster, ice forecaster for NWS Fairbanks, and Doris Brown of the NWS Anchorage office contributed materials for the field parameter analysis. Lt. (jg.) Daniel V. Hunger and AGC Wlliam B. Damico flew the Navy ice reconnaissance for the NOAA-Navy Joint Ice Center under the leadership of Comdr. James C. Langemo. Sigrid Salo of PMEL handled much of the data analysis and prepared the figures. Joy Golly and James Anderson completed the drafting and photog- raphy. Drs. James E. Overland and James D. Schu- macher discussed important aspects of the physics. Seelye Martin and Lyn McNutt were my colleagues in the experiment. REFERENCES Ahlnas, K., and G. Wendler 1979 Sea ice observations in the Bering, Chukchi, and Beaufort Seas. Pro- ceedings of POAC-79, Norwegian Institute of Technology. Charnell, R. L., J. D. Schumacher, L. K. Coachman, and T. H. Kinder 1979 Bristol Bay oceanographic processes. Fourth Ann. Rep. to OCSEAP, Research Units 549/141. Coachman, L. K., T. H. Kinder, J. D. Schumacher, and R. L. Charnell 1978 Bristol Bay oceanographic processes. Third Ann. Rep. to OCSEAP, Re- search Units 141/594. Fleet Weather Facility 1972-1975 Western Arctic Sea ice analysis. Capt. S. C. Balmforth, CO. Suitland, Md. 1976 Eastern -western Arctic Sea ice analy- sis. Comdr. V. W. Roper, CO., Suit- land, Md. 1977 Eastern-western Arctic Sea ice analy- sis. Capt. J. A. Jepson, CO. Suitland, Md. 1978 Eastern-western Arctic Sea ice analy- sis. Comdr. J. D. Langemo, CO., Suitland, Md. Martin, S., and P. Kauffman 1979 Data report on the ice cores taken during the March 1979 Bering Sea ice edge field cruise on the NOAA ship Surveyor (Sept. 14, 1979). Univ. of Washington, Dep. of Oceanogr. Spec. Rep. 89. Muench, R. D., and K. Ahlnas 1976 Ice movement and distribution in the Bering Sea from March to June 1974. J. Geophys. Res. 81 (24): 4467-76. Parmenter, P. R., and M. D. Coon 1973 Mechanical models of ridging in the Arctic sea ice cover, AIDJEX Bull. 19: 59-112. Anticipated Oil-Ice Interactions in the Bering Sea Seelye Martin Department of Oceanography University of Washington Seattle, Washington ABSTRACT In the lee-shore regions of the Bering Sea, ice formation is dominated by the presence of wind waves and swell. Labora- tory experiments and field observations show that the kinds of ice which form in these regions are grease and pancake ice. The laboratory experiments also show that much of the oil spilled within these kinds of ice accumulates on the surface. The chapter also summarizes the entrainment mechanisms for oil released under first -year ice, and discusses from field observations possible oil entrainment mechanisms which may occur at the ice edge. INTRODUCTION There are several ways in which oil and sea ice may interact in the Bering Sea. The interaction processes may be divided into those which capture oil and those which transport it. The capture processes include the interaction of oil with smooth, unbroken first-year ice, the entrainment of oil by rafted or ridged first-year ice and by grease and pancake ice, and the interaction of ocean swell, oil, and the ice floes at the ice edge. The transport processes include the interaction of oil with a Langmuir circulation in the lee-shore polynyas where grease ice occurs, the general advection of oil by large-scale ice movement, and the specific advection of oil by the ice bands which form at the ice edge. OIL AND FIRST-YEAR ICE Martin (1980) gives a detailed description of oil en- trainment by first-year ice, based on both OCSEAP data and a field experiment carried out for the Canadian Beaufort Sea Project. Fig. 14-1, which sum- marizes the evolution of a winter under -ice oil spill over the growing season, shows that the oil behavior divides into the following three parts : November-February. When oil is first released under smooth first-year ice, it collects there in natural undulations, caused by snowdrift -induced insulation. These undulations have an amplitude of order 0.1 m and a length scale of order 10 m. Some oil also flows a short distance up into the ice through the brine drainage channels. March-May. As the ice warms in the spring, top- to-bottom brine channels open up in the ice and the oil rises through these channels to the ice surface, where it spreads laterally, both under the snow and within the first few porous centimeters of the ice surface. The oil on the surface then absorbs solar radiation through the snow, causing the snow above the oil to melt. June-August. In the early part of the summer, the trapped oil continues to rise through the ice to the surface, where above the trapped oil melt ponds with oily surfaces form. Because the sun heats the oil and the winds blow the oil on the melt-pond surfaces against the edges of the ponds, the oiled melt ponds grow both laterally and in depth faster than the un- oUed ponds. As the ice continues to decay through- out the summer, much of this oil in a weathered form will flow back into the ocean either off the tops of the floes or by melting through the ice bottom. 223 224 Ice distribution and dynamics SNOW SEA I CE OIL SEAWATER •NEW ICE GROWTH SNOW SEAWATER OIL photographs suggested that such ponds contained as much as 30 percent of the spilled oil. Observations from the Buzzards Bay spill also show that oil trapped either on the surface or within floes will be transported a considerable distance by winds and currents. Fig. 14-4 shows an aerial view of an oUed ice floe measuring 75 m across moving through the 12-km-long Cape Cod Canal. The dark material on the floe surface is oil, and oil streaks are also visible on the open water. Observations showed that the floe moved through the canal into Massachusetts Bay, where the oil was released when the floe melted. This observation suggests that in the Bering Sea trapped oU may be carried distances on the order of hundreds of kilometers before the oil is released when the floe melts. WATER -WATERLINE OIL a. WATER CURRENTS SEAWATER Figure 14-1. The interaction of oil witii first-year ice (a) November-February; (b) March-May; (c) June-August. See text for further description. Figure 14-2. Interaction of oil in a tidal current with rafted ice. (a) Oil enters lee of raft; (b) oil fills surface pond;(c) current sweeps unsheltered oil away (adapted from Figure 2.3 of Deslauriers et al. 1977). RESULTS OF THE BUZZARDS BAY OIL SPILL The Buzzards Bay, Massachusetts, oil spill of 28 January 1977, described in Deslauriers et al. (1977), yields additional information on how oil responds to tidal currents under rafted and ridged ice. Fig. 14- 2a-c shows schematic diagrams of how oil which was initially swept under the ice came to the surface. In Fig. 14-2a and b, the strong 0.5 m/sec tidal currents sweep the oil into the lee of newly formed rafted ice, from whence it flows up into a surface pond, re- placing the denser sea water. Then (Fig. 14-2c) the tidad current sweeps away the unsheltered oil into other areas of rafted ice. Fig. 14-3 shows the appear- ance of an oil-filled pond in plan view based on coring observations; these ponds were filled to a depth of 100-150 mm with oil. Estimates from aerial ICE FLOES- Figure 14-3. Plan view drawn from field observations of the oil pond shown in side view in Fig. 14-2 (adapted from Figure 2.2 of Deslauriers et al. 1977). Anticipated oil-ice interactions^ 225 Figure 14-4. Aerial view of oiled ice floe in the Cape Cod Canal, 17 February. The highway next to the canal gives scale (from Figure 2.11, Deslauriers et al. 1977). I OIL IN GREASE ICE Grease ice is a slurry of smadl ice crystals and sea- water which grows in the marginal seas. Martin et al. (1978) give a preliminary discussion of how this ice grows, and Martin and Kauffman (1980) discuss exhaustively its general properties. Grease ice in the Bering Sea Fig. 14-5 shows a satellite image of the Bering and Chukchi Seas for 17 March 1976. The weather charts for this day and the preceding three days show that the winds were dominated by the Siberian high- pressure system, which created northeast winds over the Bering Sea. On 17 March 1976, the air tempera- ture recorded at 1800 Z at Wales was —25 C and the wind velocity was 15 m/sec. The NOAA-4 image shows that one effect of these cold winds was to cre- ate lee-shore regions of dark, new ice formation where the older ice was blown away. These areas occur in Norton Sound from Wales to Norton Bay, on the south side of both St. Lawrence and Nunivak islands, in Kotzebue Sound from Point Hope to Kivalena, and along the Siberian coast. For the same day, Fig. 14-6 is a LANDSAT image of the polynya south of St. Lawrence Island. The LANDSAT image, which covers an area 185 km across and has a 70 m resolution, shows that long linear ice plumes form approximately parallel to the wind and extend 10-40 km downwind of the island, and that there appears to be a pile-up of new gray ice downwind of the ice plumes. A close-up of these plumes is shown in Fig. 14-7, an aerial photograph of a polynya taken at a different time from the previous two images by the NASA Convair-990 on 20 February 1973 in the Bering Sea as part of the joint U.S.-U.S.S.R. Bering Sea Experi- ment. The aircraft flew at a radar altimeter height of 147 m and the camera had a 70° field-of-view, so that the field covered by the image measures 205 m across. The flight heading was 185°, and the re- corded wind speed was 11.5 m/sec from 332° or from the upper left of the photograph, so that the plumes are again lined up parallel to the wind direc- tion. The air temperature at altitude was about — 14 C. From inspection of the photograph the predominant wavelength was about 1.5 m. This photograph shows several interesting features. First, the grease-ice plumes have several different scales, ranging in width from less than 1 m to 35 m across (the plume marked a at the lower left). Se- cond, the spacing between the plumes ranges from 2 to 4 m for the smallest bands to about 170 m between the wide plumes marked a and b ; hence the size and spacing of the large plumes is such that they are visible on the LANDSAT. Third, there is an obvious pile-up of grease ice at the downward end of the plumes marked c, d, and e, accompanied by a 226 Ice distribution and dynamics Figure 14-5. Composite NOAA-4 visible satellite image of the Bering Sea for 17 March 1976. broadening of the head of the band. It is also clear, on the original negative at least, that there are no waves behind the heads c, d, and e. Dunbar and Weeks (1975) call these plumes "tadpoles" and show that pancake ice forms in the heads. They also suggest that an oceanic Langmuir circulation leads to the formation of the grease-ice plumes. Fig. 14-8 shows an oblique photograph from 150 m of a polynya 3 km wide south of Nome on 5 March 1978. The air temperature was —20 C, the wind speed was 15 m/sec from 010°, and the predominant wavelength was 6 m. The photograph clearly shows the organization of the grease ice into plumes parallel to the wind, and the pile-up of grease ice downwind against the floes. The arrow on the photograph marks the location of a detritus line, which we will discuss later. The field and laboratory work of Martin and Kauffman (1980) shows that the grease ice is a slurry of small ice platelets which individually measure approximately 1 mm in diameter and 1-10 ^im in thickness. Fig. 14-9 shows a photograph taken through crossed polaroids of the clusters of ice platelets ; the clusters which formed in our laboratory grease ice measure 1-5 mm across. Our observa- tions show that the concentration of such crystals ^ Anticipated oil-ice interactions 227 ' 1^ J ; Figure 14-6. LANDSAT image of tlie polynya soutii of St. Lawrence Island for 17 March 1976. within the grease-ice plumes ranges from 15 to 45 percent and that the thickness of the ice varies from 0.1 to 0.3 m in the laboratory to more than 1 m in the field. Also, our experiments show that because grease ice has a nonlinear viscosity, which increases as both the ice concentration increases and the shear rate decreases, it is an excellent wave absorber. Fig. 14-10, adapted from Pollard (1977) shows a schematic diagram of the Langmuir circulation which accompanies the formation of grease ice. The combi- nation of the bi-directional wind-wave spectra and the mean wind drift leads to the formation of alternating roll vortices in the ocean, where the grease ice accu- mulates in the convergence zones. This also suggests that, if oil is emulsified into small droplets by the wave breaking, these droplets may circulate around in the Langmuir rotors, as well as pile up in the conver- gence zones. Pollard (1977) states that typical con- vergence zone separations range from 5 to 300 m, and that rotor downweUing velocities are observed to be between 20 and 60 mm/sec. Further, the recent theoretical work of Craik and Leibovich (1976) sug- gests that the accumulation of a viscous substance such as oil or grease ice in the convergence zones will, by transforming the wave momentum into an additional mean current, intensify the rotor veloci- 228 Ice distribution and dynamics I i Figure 14-7. Aerial phiotograph from 147 m of grease ice formation in the Bering Sea. See text for further explanation. Photograph courtesy NASA. ties. This intensification could lead to the dis- tribution of more oil throughout the water column. The laboratory results Our laboratory tank, described in Martin et al. (1978) and shown in Fig. 14-11, measured 2 m long. 1 m wide, and 0.6 m deep. In this tank, which was filled to a depth of 0.4 m with a 34^/00 NaCl solu- tion and placed in a cold room, we grew grease ice by generating waves. Fig. 14-11 shows the appearance of the wave decay and the pUe-up of grease ice in the tank. Figure 14-8. Oblique aerial view from 150 m of grease ice herded into Langmuir streaks south of Nome on 5 March 1978. See text for further description. Figure 14-9. Photograph of clusters of frazil ice crystals taken between crossed- polaroids. The subject meas- ures 21 mm in the vertical. 229 230 Ice distribution and dynamics SIDE VIEW Figure 14-10. Schematic drawing of grease ice and the Langmuir circulation adapted from Pollard (1977, Figure 8.3). -INSULATION Figure 14-11. Schematic side-view drawing of the experi- mental tank, with inner dimensions of 2 m in length, 1 m in width, and 0.6 m in depth. To illustrate both the induced circulation and the form that oil entrainment takes in waves propa- gating into grease ice, Fig. 14-12 shows a composite image of a sideview photograph of the grease ice above a sketch of the wave-decay-induced mean ve- locity field in the grease ice ; the sketch also shows the final distribution of oU released on the surface ahead of the grease ice. In the photograph, the grease ice is white, its maximum thickness is about 0.2 m, the wave-period is 0.54 sec, and the wavelength is 0.46 m. The paddle is to the left in the photograph and a wave probe sticks into the water ahead of the paddle; the gap in the photograph is caused by a tank sup- port. Although we did not release oil into the grease ice shown in the photograph, the sketch shows the oil distribution which occurred in a similar experiment described below. For oil entrainment, the most important property of grease ice is that it changes abruptly from liquid to solid behavior; the transition is shown in Fig. 14-12 as the "dead zone." Upstream of the dead zone, the grease ice behaves as a liquid with a surface tempera- ture very nearly equal to the freezing point of sea water; within this zone Martin and Kauffman (1980) show that the wave decay is linear. Within the liquid region there is also a mean rotary circulation which at the leading edge is of order 0.3 m/sec. Be- hind the dead zone, the grease ice behaves as a solid and the surface temperature decreases rapidly. Ex- periments with both plastic chips and oil show that material released on the surface ahead of the dead zone either accumulates at the dead zone or else circulates around as small droplets within the grease ice. For possible field examples of dead zones, for example, the arrow in Fig. 14-8 marks a detritus line where small ice chunks accumulate and the wave amplitude appears to go to zero; this line is probably the local dead zone and would serve as an accumula- tion point for oU released on the open water. Again, in Fig. 14-7, we believe that the heads of the plumes marked c, d, and e, where the ice is thicker and there is a wave shadow downwind, also contain local dead zones which would serve as regions of oU accumula- tion. To verify these general points, we conducted a specific grease ice oil spill experiment. In this exper- iment, we grew grease ice to an average depth of about 100 mm at an air temperature of —25 C. We then raised the room temperature to — 2 C and gener- ated waves with a period of 0.55 sec, a wavelength of 0.48 m, and an amplitude of 35 mm. These waves were of large amplitude and on the verge of breaking. As soon as the room had warmed up, we poured 200 ml of Prudhoe Bay crude oil onto the water just ahead of the grease ice. We took two kinds of photo- graphs of the spill: first, we placed a camera on a fixed mount above the tank, looking down on the ice at an oblique angle. This camera had a motor drive which took photographs every ten seconds. Second, we used a handheld camera to take photographs of the ice and oil at various positions and times around the tank. Fig. 14-13 shows the overhead sequential photo- graphs of the spill; the time marked on each photo- graph is the elapsed time after the start of the spUl. The first frame, 0 seconds, shows the pouring of the oil on the surface. The boundary between the open water and the grease ice is visible just below the hand in the photograph, and at the lower left corner of the photograph, the line of bubbles marks the position of the dead zone. These lines are not parallel because the camera is mounted at an oblique angle. The subsequent photographs at 30, 60, and 190 seconds show that the oil was rapidly transported into the Anticipated oil-ice interactions 231 I SEAWATER DEAD ZONE Figure 14-12. Composite figure showing the wave decay and induced circulation within the grease ice. The top shows a side-view photograph of grease ice in our tank; the bottom is a slcetch of the induced mean velocity and the final distribution of oil spilled in the ice. > grease ice, then slowly pumped onto the ice beyond the dead zone. Also visible in the photographs at 60 and 190 seconds are the small oil droplets which were produced by the turbulence at the leading edge of the grease ice and are being carried around within the grease ice by the circulation shown in Fig. 14-12. The last two photographs, at 240 and 310 seconds, show that most of the oil is now on the surface beyond the dead zone, with 5-10 percent of the oU circulating as droplets within the grease ice. For another view of the same experiment. Fig. 14- 14 shows several photographs taken with the hand- held camera. Fig. 14-14a shows the oil being poured into the tank, 14-14b shows the appearance of the oil at approximately 10 seconds, and 14-1 4c the appear- ance at 250 seconds. In 14-14c, both the wave decay and the pile-up of the oil at the dead zone are clearly visible. Figs. 14-14d and 14-1 4e, close-ups of the oil droplets, show that these droplets are approximate spheres of 1-3 mm diameter, and circulate within the grease ice without wetting it. Finally, Fig. 14-14f shows the appearance of the oil on the surface after the paddle was turned off at about 30 minutes after the start. In a result similar to Metge's (1978) obser- vations, the photograph shows that because of the grease ice, the oU moves less freely on the surface than it would on open water. To summarize, the laboratory experiments show that most of the oil ends up on the ice surface be- yond the dead zone, with some oil droplets circula- ting in the grease ice ahead of the dead zone. Our field observations suggest that if oil was spilled in the various polynya zones shown in Fig. 14-5, some would end up on the grease ice surface in the Langmuir convergence zones, some would accumulate in the local dead zones, and a smaller fraction would be emulsified into oil droplets by the breaking waves; these droplets would then circulate around within both the grease ice and the Langmuir rotors. Much of this oil would accumulate downwind of the polynyas in the regions of gray ice as shown south of St. Lawrence Island in Fig. 14-6. Finally, as Pease and McNutt show in their respective chapters (this volume), because these coastal polynyas serve as regions of ice generation for the entire Bering Sea, oil which is released in the polynyas may be carried to the ice edge. OIL SPILLED IN ICE FLOES OSCILLATING IN A WAVE FIELD In the field there are two situations in which oil may be associated with ice floes oscillating in a wave field: it may be released under pancake ice, or it may 232 Ice distribution and dynamics 0 s 30 s Figure 14-13. Overhead sequential photographs of an oil spill in grease ice. be released at the ice edge in the 10-20 km zone of floes agitated by ocean swell. In the first situation, Martin et al. (1978) show that pancake ice develops from grease ice, and thus also occurs in the lee-shore polynyas; in the second, Martin and Bauer (this volume) discuss ice-edge floes agitated by ocean swell. For oil entrainment, it is important in both situa- tions that (1) the floes and cakes exist as separate units with either open water or water and grease ice around them; (2) the incident swell or wind waves cause the floes to oscillate back and forth so that crude oil can be pumped onto their surfaces. For example, in the chapter about the ice edge by Martin and Bauer (this volume), Figs. 12-lOa, 12-lOb, 12- 11a, and 12-1 lb show ice floes which have been broken by the incident swell, are oscillating in this swell, and have wetted edges caused by the pumping of seawater onto the floe surfaces. Martin et al. (1978) show that pancake ice forms when grease ice becomes so thick that the circulation within the ice is suppressed; the surface then freezes into chunks of pancake ice floating over a layer of grease ice. In the field, the pancakes form downwind of the regions of grease-ice formation. On 19 March 1978, we surveyed the ice over a region stretching from Cape Thompson in Kotzebue Sound 90 km south. Within this region, we found an area of at least 100 km^ covered with pancakes about 0.5 m in diameter. Fig. 14-15a shows the rough surface ap- pearance caused by rafted pancakes on the site at 67° 19.2'N, 166° 31.3'W, and Fig. 14-15b shows a close-up photograph of the pancakes on the surface. When we cored through the pancakes, we found that they were about 100 mm thick, over a layer of frozen grease ice 285 mm thick, so that the original grease ice thickness was 385 mm. The field evidence sug- gests that these pancakes grow over large areas under very stormy conditions. In our previous laboratory study, Martin et al. (1978) found that oil released beneath long linear pancakes came to the ice surface between the oscillat- ing cakes, where the oscillating motion pumped Figure 14-13, Cont. 60 s 190 I » •'■''•,. - -♦■> ' * • 240 s 310 s 233 234 Anticipated oil-ice interactions 235 li o CO l^BHiP^^<^^" ' ^^^^■Flift' ♦. ••. ' ^^M?^l-^^ r-f ■'■' ■ ■■■Twivl'* ^»* • .%' ,♦■"•. ,* #• • . . • . ^ . # , . ./ ^ ^ ^ •* ' • . ' * ' ' * * • * * ' . ' 4 ■ • . •- ■ * ..iv'*; . •-»*•- ■• ♦. . . , ' i ■ « * • . « « • . • • • ♦ . . ' •' 4' « ■'^■. '^ ,. t ,, * » t • i^HMHHlH * = » • ^■■k . •• f ^^^^^^^^^B IbI^^^^^^^^K ■^^^^^B H^^^^^^HL * ^^^^^^^^^^^K ■hHBBHHmB^^^ ^ "C^^Vfrl^Y:- \T'- o a 3 o c 3 O Ph "* 3 about 50 percent of the oil onto the ice surface. Because long linear cakes do not appear in nature, we repeated the experiment in a two-dimensional wave- field, so that nearly circular cakes formed in our tank. We did this experiment in the tank shown in Fig. 14-11. At the side wall of the tank we added a wedge (shown in Fig. 14-16 in plan view) so that the refrac- tion of waves around the wedge generated a two- dimensional wave field. In the experiment, we first grew a layer of grease ice 140 mm thick at an air temperature of —30 C. After the pancakes began to form on the surface, we raised the room temperature to —12 C. The pancakes were 0.1-0.3 m in diameter and 10-20 mm thick. The wave period was 0.91 seconds so that the wavelength was 1.3 m. The wave amplitude at the paddle was 35 mm, and the separation between pancakes oscillated between 0 and 20 mm. Once the pancakes were established, we released 500 ml of Prudhoe Bay crude oil beneath the grease ice through a vertically mounted discharge tube. Figs. 14-17 and 14-18 show the development of the surface appearance of the oil over 9,160 seconds or about 2.5 hours; the elapsed time from the oil dis- charge is marked on each photograph in seconds. A 12-inch ruler on the ice just below the spill gives the scale. The photographs show that the oil first appears in the cracks around the pancakes, then the oscilla- ting motion of the cakes drives the oil laterally through the cracks from its point of first appearance and pumps some of the oil onto the floe surfaces. Figs. 14-19a and b show a lateral view of the spill at about 60 sec and at 1,660 sec respectively. These lateral views show that the ridge heights around the cakes vary from 0 to 30 mm, and that the oil is pumped over the rims onto the interior of the pan- cake. We also found that, although the grease ice in the cracks around the pancakes inhibited the oil move- ment through the crack system, the oil was pumped laterally through the cracks with what appeared to be the higher fractions, or lighter-colored oil, running ahead. As the final photograph in Fig. 14-16 shows, at the end of the experiment, the oil was contained with considerable density variation within a region measuring IX 0.6 m^ . This means that on the average, 1 ml of oil covered 12 cm^ . On the day after the experiment, we scraped up the oil from the ice surface and from the cracks where the oU was accessible, to recover the following amounts of oil: from the surface, 32 ml; from the cracks, 90 ml. This meant that a total of 120 ml of \ Figure 14-15. Pancake ice in tiie field, (a) Surface appearance of large pancake-covered area; (b) close-up of the pancakes with chisel for scale. 236 Anticipated oil-ice interactions 237 LU _l Q Q Figure 14-16. Plan-view sketcii of tank used for pancake ice experiments. oil was accessible to cleanup from the surface, or about 25 percent of the total spilled. To search for the rest of the oil, we then cut up the ice into vertical sections around the oil spill. In a schematic drawing, Fig. 14-20 shows that the oil beneath the surface occurred on the pancake bot- toms, deep within the cracks, and, in the form of small droplets, within the grease ice. The importance of the grease ice, then, is that it retards both the rise of oil to the surface and its lateral spreading within the cracks. In summary, in the case of an oil spill beneath a wind-agitated field of pancakes, we can expect that a sizeable fraction of the oil will be pumped onto the ice surface, and that the remainder will be bound up in the grease ice/pancake ice system. Second, the observations of swell propagation through the fields of ice floes near the ice edge and the observations of wetted surfaces on these floes strongly suggest that a similar pumping of oil onto the floe surfaces wUl occur in this region. OTHER TRANSPORT PROCESSES Finally, we consider the possible large-scale move- ment of oil by the general ice circulation in the Bering Sea described in the chapters by McNutt, Pease, and Martin and Bauer (this volume). Fig. 14-21 is a schematic drawing of the large-scale winter ice motion during periods of northeast winds. Pro- ceeding from the upwind direction the figure shows the production of grease ice in the lee-shore polynyas and the general southward advection of the large ice floes. As these large floes near the ice edge, the propagation of ocean swell into the pack breaks them into smaller floes (Martin and Bauer, this volume). As these continue to approach the edge, the action of wind and swell causes rafting and ridging, so that sails as high as 1 m and keels as deep as 5 m form. This increase in aerodynamic roughness means that the wind moves these floes to the southwest faster than those of the unroughened pack. As Bauer and Martin show (this volume), at the ice edge these floes are organized into the bands observed on satellite photographs like Fig. 12-5; these bands move south- west at velocities of about 4 percent of the wind speed at 25° to the right of the wind. As they move into warmer water, the bands disintegrate, thereby exposing new pack ice to be broken by the swell. To summarize this process, the wind stress causes the interior pack ice to move southwest to the edge from the lee-shore polynyas. Pease shows that this movement from the polynyas to the ice edge takes 30-60 days. As the ice approaches the edge, the wave action roughens the ice so that it forms into bands which move southwest at a higher velocity than the rest of the pack. This southwest motion continues until the ice melts. This "conveyor belt" then, may transport oil which is spilled in the northern Bering Sea to a location well south of the ice edge. Also, if oil were spilled at the ice edge, this ice motion could capture the slick and spread it into open water farther south. The laboratory and field studies on oil pollu- tion and sea ice suggest that much of the spilled oil will accumulate on the water and ice surface. Be- cause our laboratory experiments are of short dura- tion, however, we do not yet have a good idea of how the oil will evolve once it is on the surface. Field studies of oil spills in temperate water— for example, the 1976 Argo Merchant (Grose and Matson 1977) and the 1979 IXTOC-1 blowout^how that, because of the combination of wind and wave mixing with evaporation and solar photo-oxidation, the oil forms a mousse-like emulsion with seawater, which then forms into pancakes with horizontal dimensions of 1-100 m and thicknesses of 0.05-0.3 m. If similar emulsions form in polar ocean spills, they may not interact with the ice in the same way as the crude oil and diesel spills described in this chapter. ACKNOWLEDGMENTS The author thanks Ms. Katherine Martz and Dr. Constance Sawyer for the photographs in Figs. 14-5 and 14-6 respectively. We also thank the Alaska of- fice of the National Weather Service for use of their archives. We thank Mr. William Abbott and Dr. Per Gloersen of the Goddard Space Flight Center for the loan of the original negative of Fig. 14-7, and Mr. James Anderson of the Pacific Environmental Labor- atory for his care in printing the image. Peter Kauff- man contributed greatly to this chapter by his work o O 01 o o 238 o CO O O u a O So fa n. 239 o O O o Csl O 240 Anticipated oil-ice interactions 241 o o m on running the oil-spill experiments; we greatly appre- ciate his help. This study was supported by the Bu- reau of Land Management through interagency agree- ment with the National Oceanic and Atmospheric Administration, under which a multiyear program responding to needs of petroleum development of the Alaskan continental shelf is managed by the Outer Continental Shelf Environmental Assessment Program (OCSEAP) Office. This is publication number 1120 of the Department of Oceanography, University of Washington, Seattle. tub a 0) u a Oi 3 cr a> M Oi .c C4-I O a .2 3 .s c o u r* •+*- ICE BANDS .V 3<=«;iQat>^3QQ-^^:,^^ ''^~7r7-rr77^rY77/T7~r77-7r7r7-777TT77rrTr. Figure 14-20. Schematic side-view drawing of the oil Figure 14-21. A sketch showing the general ice motion location within a pancake ice/grease ice mixture. and deformation during periods of northeast winds. Anticipated oil-ice interactions 243 REFERENCES Craik, A.D.D., and S. Leibovich 1976 A rational model for Langmuir circu- lations. J. Fluid Mechanics 73: 401-26. Martin, S., and P. Kauffman 1980 A field and laboratory study of wave damping by grease ice. J. Glaciology (in press). Deslauriers, P.C, S. Martin, B. Morson, and B. Baxter 1977 Bouchard #65 Oil spUl in ice covered waters of Buzzards Bay. OCSEAP, NOAA, ERL, Boulder, Colo. Dunbar, M., and W.F. Weeks 1975 Interpretation of young ice forms in the Gulf of St. Lawrence using side- looking airborne radar and infrared imagery. U.S. Army CRREL Res. Rep. 337. Grose, P.L., and J.S. Matson, editors 1977 The Argo Merchant Oil Spill, NOAA, ERL, Boulder, Colo. Martin, S. 1980 A field study of brine drainage and oil entrainment in first-year sea ice. J. Glaciology 22: 473-502. Martin, S., P. Kauffman, and P.E. Welander 1978 A laboratory study of the dispersion of crude oil within sea ice grown in a wave field. In: Proceedings of the 27th Alaska Science Conference, West, G.C., ed. AAAS, Fairbanks. Metge, M. 1978 Pollard, R. T. 1977 Oil in pack ice coldroom tests (Imper- ial Oil Ltd., Production Res. Div., Calgary, Alberta). Observations and models of the struc- ture of the upper ocean. In: Model- ling and prediction of the upper layers of the ocean, E. B. Krause, ed. 102- 17, Pergamon Press, New York. Section III Geology and Geophysics C. Hans Nelson, editor Sedimentary Processes and Potential Geologic Hazards on the Sea Floor of the Northern Bering Sea Matthew C. Larsen, C. Hans Nelson, and Devin R. Thor U.S. Geological Survey Menlo Park, California ABSTRACT A dynamic environment of strong bottom currents, storm waves, and gas-charged sediment on the shallow sea floor of the northern Bering Sea creates several potential geologic hazards for resource exploration. Thermogenic gas seeps, sea-floor gas cratering, sediment liquefaction, ice gouging, scour-depression formation, coastal and offshore storm surge and associated deposition of storm-sand, and movement of large-scale bedforms all are active sedimentary processes in this epicontinental shelf region. Interaction between the processes of liquefaction and the formation of shallow gas pockets and craters, scour depres- sions, storm-sand deposits, and slumps results in sediment instability. Liquefaction of the upper 1-3 m of sediment may be caused by cyclic storm-wave loading of the Holocene coarse-grained silt and very fine grained sand covering Norton Sound. The widespread occurrence of gas-charged sediment with small surficial craters (3-8 m in diameter and less than 1 m deep) in central Norton Sound indicates that the sea-floor sediment is periodically disrupted by escape of biogenic gas from the underlying peaty mud. During major storms, lique- faction may not only help trigger crater formation but also magnify erosional and depositional processes that create large-scale scour areas and prograde storm-sand sheets in the Yukon prodelta area. Erosional and depositional processes are most intense in the shallower parts of the northern Bering Sea and along the coastline during storm surge flooding. Ice gouges are numerous and ubiquitous in the area of the Yukon prodelta, where the sediment is gouged to depths of 1 m. Though much less common than in the prodelta, ice gouges are present throughout the rest of the northern Bering Sea where water depths are less than 20 m and at times where water is as much as 30 m deep. In the Yukon prodelta area and in central Norton Sound, where currents are constricted by shoal areas and flow is made turbulent by local topographic irregularities (such as ice gouges), storm-induced currents have scoured large (10-150 m in diameter), shallow (less than 1 m deep) depressions. The many storm-sand layers in Yukon prodelta mud show that storm surge and waves have generated bottom-transport currents that deposit layers of sand as thick as 20 cm as far as 100 km from land. Storm-surge runoff may reinforce the strong geostrophic currents near Bering Strait, causing intermittent movement of even the largest sand waves (10-200 m wavelength, to 2 m height). INTRODUCTION studies of potential geologic hazards on the Norton Basin sea floor in the northern Bering Sea have been conducted by the U.S. Geological Survey (U.S.G.S.) in the course of evaluating oil and gas lease tracts preparatory to Outer Continental Shelf (OCS) leasing. The data base for this evaluation included 9,000 km of high-resolution geophysical tracklines (Nelson et al. 1978b, Thor and Nelson 1978, Larsen et al. 1979a), 1,000 grab samples, 400 box cores, and 60 vibracores; in addition, hundreds of camera, hydrographic, and current-meter stations have been occupied during the past decade by U.S.G.S., Nation- al Oceanic and Atmospheric Administration (NOAA), and University of Washington oceanographic vessels (Figs. 15-1 and 15-2). The northern Bering Sea is a broad, shallow epi- continental shelf region covering approximately 200,000 km^ of subarctic sea floor between northern Alaska and the U.S.S.R. The shelf can be divided into four general morphologic areas: (1) the western part, an area of undulating, hummocky relief formed by glacial gravel and transgressive-marine sand sub- strate (Nelson and Hopkins 1972); (2) the south- eastern part, a relatively flat featureless plain with fine-grained transgressive-marine sand substrate (McManus et al. 1977); (3) the northeastern part, a complex system of sand ridges and shoals with fine- to medium-grained transgressive sand substrate (Nelson et al. 1978a); and (4) the eastern part, a broad, flat marine reentrant (Norton Sound) covered by Holocene silt and very fine sand (Nelson and Creager 1977). A detailed discussion of bathymetry 247 248 Geology and geophysics Figure 15-1. Sampling stations for U.S. Geological Survey research in northern Bering Sea between 1967 and 1978. and geomorphology of the northern Bering Sea is given by Hopkins et al. (1976) (Fig. 15-3). The northern Bering Sea is affected by a number of dynamic conditions: winter sea ice, sea-level setup, storm waves, and strong currents (geostrophic, tidal, and storm). The sea is covered by pack ice for about half the year, from November through May. A narrow zone of shorefast ice (sea ice attached to the shore) develops around the margin of the sea during winter months. Around the front of the Yukon River Delta, shorefast ice extends to 40 km offshore (Thor et al. 1978). During the open-water season, the sea is subject to occasional strong northerly winds, and in the fall strong south-southwesterly winds cause high waves and storm surges along the entire west Alaskan coast (Fathauer 1975). Throughout the year, there is a continual northward flow of water with currents intensifying on the east side of strait areas (Coachman et al. 1975). Although diurnal tidal ranges are small (less than 0,5 m), strong tidal currents are found in shoreline areas and within central Norton Sound (Fleming and Heggarty 1966, Cacchione and Drake 1979a). This chapter reviews basic sedimentary processes of this epicontinental shelf region and discusses certain potential geologic hazards related to these processes: thermogenic gas seepage, biogenic gas saturation of sediment and cratering, sediment liquefaction, ice gouging, current scouring, storm- sand deposition, and mobile bedform movement (Fig. Sedimentary processes and potential geologic hazards 249 65' 166' 164° 1 EX PLANATION • Grab sample D Box core A Vibracore ^ Bore hole 50 _l ••••••••••• ••••••• Kilometers •• • !*fll*t •«•• ••••••••••••••• ••••••••••• •**•••• •••• •••••• ••*••• • j_ ± Figure 15-la. Cross-hatched area of Fig. 15-1, showing closely spaced sampling grid offshore at Nome. 15-4). These geologic hazards may pose problems for the future development of offshore resources in Norton Basin. SEDIMENTARY PROCESSES Yukon Delta processes The Yukon River drains an area a little less than 900,000 km^ , providing a water discharge of approxi- mately 6,000 m^/sec and a sediment load of 70-90 million mt per year (Dupr6 and Thompson 1979, Cacchione and Drake 1979a). The sediment load, almost 90 percent of all sediment entering the Bering Sea, is composed mainly of very fine sand and coarse silt with very little clay. The Yukon Delta plain, like many deltas, is fringed by prograding tidal flats and distributary mouth bars. The delta front and prodelta are offset from the prograding shoreline by a broad platform (referred to as a subice platform) 30 km at its widest reach. This platform appears to be related to the presence of shorefast ice that fringes the delta for half the year. The term "delta front" describes the relatively steep margin of the offshore delta environment charac- terized by rapid deposition of sediment in water 2-10 m deep. The prodelta, an area of extremely gentle slopes, marks the distal edge of the deltaic sediments extending as far as 100 km offshore. Processes on the Yukon Delta and offshore operate under seasonal regimens (Dupr^ and Thompson 1979). The ice-dominated regimen begins with freeze-up in late October or November. Shorefast ice extends 10-40 km offshore, where it terminates in a series of pressure ridges and shear zones formed by the interaction of shorefast ice with the highly mobile seasonal pack ice. River breakup, typically in May, marks the be- ginning of the river -dominated regimen. Once the shorefast ice melts or drifts offshore, sedimentation is dominated by normal deltaic processes under the influence of the high discharge of the Yukon River. Increasingly frequent southwest winds and waves associated with major storms during late summer mark the beginning of the storm-dominated regimen. High wave energy and decreasing sediment discharge from the Yukon cause considerable coastal erosion and reworking of deltaic deposits. Coastal storm surge In November 1974 a severe storm moved from "O r- J <^ Sr' rv ^i ) ^ o ( ^ \1 % V^ •■4 ^ (P \ o 250 Sedimentary processes and potential geologic hazards 251 25 0 25 50 75 100 km Figure 15-3. Generalized bathymetry of northern Bering Sea in 10-m contour intervals. southwest to northeast across the Bering Sea. Peak winds were 111 km per hour from the south, and nearshore waves were reported to be 3-4 m in height (Fathauer 1975). Coastal flooding extended from Kotzebue Sound (north of Bering Strait) to just north of the Aleutian Islands (Fathauer 1975). The maxi- mum sea-level setup, measured by the elevation of debris lines along the coast of Norton Sound, ranged from 3 to 5 m above mean sea level (Sallenger et al. 1978). During this storm, extensive inland flooding occurred and erosion of coastal bluffs 2-5 m high took place near Nome. Irregular landward erosion, as much as 18 m, occurred west of Nome, where bluffs are 3-5 m high. East of Nome, where bluffs are 1.5-2 m high, landward erosion was as much as 45 m. The water level in the Norton Sound area reached its peak on 12 November, when as much as 2 m of water was standing in the village of Unalakleet and the static high-water line at Nome was 4 m above mean low water. Storm currents The transport of sediment in Norton Sound and Norton Basin can be described in terms of distinct quiescent and storm regimes (Cacchione and Drake 1979a). The quiescent regime is characterized by generally low levels of sediment transport caused mainly by tidal currents. Fine silt and clay move as "wash load," and bedload transport is negligible except in shallow areas where surface waves become 252 Geology and geophysics Figure 15-4. Potentially hazardous areas of northern Bering Sea (from Thor and Nelson 1979). dominant. Current speeds in this regime are no greater than 30 cm/sec. Calm weather conditions prevail for about 90 percent of the year in the northern Bering Sea, and less than 50 percent of the sediment transport takes place under these conditions (Cacchione and Drake 1979a). Norton Sound is commonly exposed to strong southerly and southwesterly winds generated by low-pressure weather systems in September, October, and November. A two-day storm in Sep- tember 1977 transported sediment equal to the trans- port that would occur during four months of quies- cent conditions. Current speeds were as much as 70 cm/sec during this storm. Graded storm-sand layers to 20 cm thick (Nelson 1977) occur in sea-floor strata of the northern Bering Sea, widespread evidence of major storm-surge events. The effects of storm surge are intensified by two factors: extremely shallow water depth (less than 20 m), and strong bottom return currents that may move large amounts of sediment northward to the Chukchi Sea. The thickness of Holocene sediment in Norton Sound relative to Holocene sediment input from the Yukon River indicates that significant amounts of Sedimentary processes and potential geologic hazards 253 sediment have been resuspended and transported out of Norton Sound (Nelson and Creager 1977). About 10 percent of the Yukon River input into Norton Sound may be carried as suspended sediment through the Bering Strait into the Chukchi Sea under non- storm conditions (Cacchione and Drake 1979a). As much as 40 percent of Holocene sediment discharged from the Yukon River appears to be missing from Norton Sound; this difference of 30 percent may have been resuspended and transported during storms (Cacchione and Drake 1979a). Storm currents not only resuspend and transport massive amounts of suspended sediment, but also appear to move leirge amounts of sand in bedload transport for considerable distances offshore. Graded storm-sand layers are extensive throughout southern Norton Sound; their thickening toward the Yukon subdelta, the apparent source region, suggests massive movement of bedload sediment aw^ay from the delta toward the adjacent offshore region during storms (Nelson 1977). Wave effects Waves and wave-induced currents are the dominant sedimentological agents on the inner shelf of north- western Norton Sound and in the approaches to Bering Strait (Hunter and Thor 1979). Sedimentary features common to both areas include sand and gravel patches and ribbons, wave ripples, sand waves, and ice gouges (Hunter and Thor 1979). Wave ripples with spacings to 2 m are common in both the Port Clarence and Nome areas in zones where sediment is well sorted and grain size ranges from coarse sand to pebbly gravel. Ripples in the Port Clarence airea trend northwest-southeast and can be explained as the result of storm waves from the southwest Bering Sea. Trends of ripples in the Nome area indicate dominant wave activity from south to southwest. Ribbons of sand and gravel are well developed near the entrance to Port Clarence. These bedforms may be produced by wave action or by wave-induced net water motion in the direction of wave propagation. A rich assemblage of depositional and erosional features, both wave-formed and current-formed, occupies the floor in shallow water close to the southern shore of Seward Peninsula. Wave-formed features are more common; some of the current- formed features imply considerable sediment trans- port by strong bottom tidal currents. Only the broad patterns of wave and current movement in southwestern Norton Sound are known. The major wave trains originate in the southern Bering Sea; waves move northward and refract clockwise around protruding Yukon shoals. Smaller waves with shorter periods are generated by north- easterly winds and move southwestward. Liquefaction The Yukon River sediment that covers most of the bottom of Norton Sound (McManus et al. 1977) is primarily silt with considerable amounts of very fine sand in some areas and a generally minor content of clay-size material. The sediment thickness is usually less than 3 m, except near the Yukon Delta, where accumulations are as thick as 10 m (Olsen et al. 1979). The material is generally dense; there are zones of relatively loose sediment (material of low density) in gas-charged areas. In the delta areas sampled by 6-m vibracores, relatively loose zones of sediment were observed above and between dense layers. Freshwater peaty mud beneath Yukon marine silt is somewhat over-consolidated and contains substan- tial amounts of organic carbon and gas. The presence of gas indicates that the pore pressure in the peaty muds may be high. If it is, the strength of the mate- rial could be low despite its highly consolidated state. The dominantly coarse-silt to fine-sand texture of the material, the occurrence of loose sediment zones, and theoretical calculations utilizing GEOPROBE cyclic wave loading data (Olsen et al. 1979, Clukey et al. 1980) indicate that Yukon prodelta sediment in southwestern Norton Sound is susceptible to lique- faction. Liquefaction of the prodelta deposits may be caused by cyclic loading resulting from exposure of the Yukon prodelta to large storm waves from the southwest. Water depths are shallow enough for much of the wave-generated surface energy to be imparted to the bottom sediment, so that the upper 1-2 m of sediment may be liquefied during extreme storm-surge events (Clukey et al. 1980). This lique- faction of prodelta sediment may influence storm- sand transport, formation of sediment depressions, and gas cratering. Ice scour Ice on the Bering shelf scours and gouges sur- ficial sediment of the sea floor (Fig. 15-5). The annual ice cover in this subarctic setting is generally thin (less than 2 m); thick ice capable of gouging forms where pack ice collides with and piles up against stationary shorefast ice, developing numerous pressure ridges (Thor and Nelson, this volume). A wide, well-developed shear zone forms in south- western Norton Sound as ice moving southward from the northeast Bering Sea and westward along south- ern Norton Sound converges in the shallow water of 254 Geology and geophysics 100 km Figure 15-5. Distribution and density of ice gouging, direction of movement of pack ice, and limits of shorefast ice in northern Bering Sea (from Thor and Nelson 1979). the Yukon prodelta. Consequently, numerous zones of pressure ridges are formed. In this region, where the water is 10-20 m deep, the density of ice gouges is highest. Gouges are found in water to 30 m deep, and furrows are as much as 1 m deep. Ice gouging affects the sea floor under shorefast areas only minimally or not at all (Thor and Nelson, this vol- ume). Current scour depressions Zones of large flat-floored depressions in Norton Sound occur mainly in two areas: west of the Yukon prodelta and 50 km southeast of Nome on the flank of a broad shallow trough (Fig. 15-6) (Larsen et al. 1979a). These features range from individual more or less elliptical depressions 10-30 m in diameter to large areas with irregular margins, 80-150 m in diameter. The depressions are 60-80 cm deep (Larsen et al. 1979b). Bottom current speeds in depression areas are 20- 30 cm/sec under nonstorm conditions and were measured at 70 cm/sec during a typical autumn storm (Cacchione and Drake 1979a). Both zones of depres- sions are on flanks of gently sloping shoals, where strong tidal or geostrophic currents shear against the slopes. Small-scale ripple bedforms are associated with depression areas and mean grain size ranges from 4 to 4.5 0 (0.063-0.044 mm). Depressions in the Yukon Delta area are aissociated with extensive ice gouging. The gouge furrows commonly expand into large shallow depressions (Larsen et al. 1979 and Thor and Nelson, this volume). Experiments in flumes containing fine sand and silt have shown that currents flowing over an obstruc- tion will scour material immediately downcurrent Sedimentary processes and potential geologic hazards 255 0, CH IR I K O V BASIN BERING SEA SCOUR DEPRESSION EXTENSIVE AREA OF SCOUR RIPPLE BEDFORMS ZONE OF MEAN BOTTOM CURRENT >20 cm /sec ZONE OF STORM SAND LAYERS, I- 20 cm thick 00 Km 164 162 Figure 15-6. Location of scour depressions, extensive scour and ripple zones, and strong bottom currents in Norton Sound, showing area of storm sand deposition (modified from Larsen et al. 1979b). from the obstruction (Young and Southard 1978). The large scour depressions observed in Norton Sound may be a characteristic erosional bedform developed during storms when strong currents and high wave energy are focused on silt-covered slopes where local topographic disruptions set off flow separation and downcurrent scour. Sandwaue dynamics Strong geostrophic currents prevail throughout much of the northern Bering Sea, particularly where westward land projects into the northward flow, as in the eastern Bering Strait area (Fleming and Heggarty 1966, Coachman et al. (1975). In such regions large bedforms develop and migrate, forming an unstable sea floor (Nelson et al. 1978a). These large bedforms include large-scale sand waves 1-2 m high with wavelengths to 200 m, and small-scale sand waves 0.5-1.0 m high with wavelengths of 10 m. They occupy the crests and some flanks of a series of linear sand ridges 2-5 km wide and as much as 20 km long between Port Clarence and King Island. Sand-wave movement and bedload transport take place during calm weather (Nelson et al. 1978a), but maximum change apparently occurs when severe southwesterly storms generate sea-level setup in the eastern Bering Sea that enhances northerly currents. In contrast, strong north winds from the Arctic reduce the strength of the northerly currents and thereby arrest bedform migration. Sediment gas charging The distribution of acoustic anomalies suggests that almost 7,000 km^ of sea floor in Norton Sound 256 Geology and geophysics and Chirikov Basin is underlain by sediment contain- ing gas (biogenic and/or thermogenic) sufficient to affect sound transmission through these zones (Holmes 1979a). Core-penetration rates (Nelson et al. 1978c, Kvenvolden et al. 1979c) and sediment samples from 2-6 m vibracores confirm gas saturation of near-surface sediment at several locations chEirac- terized by acoustic anomalies. The isotopic composi-_ tions of methane at four of the sites range from —69 to -800/00 (513Cpdb) (Kvenvolden et al. 1978, 1979b). This range of values clearly indicates that the methane is formed by microbial processes, possibly operating on near-surface Pleistocene peat deposits that underlie Holocene deposits throughout the northern Bering Sea. At one site in Norton Sound, near -surf ace sediment is apparently charged with CO2 actively seeping from the sea floor accompanied by less than one percent hydrocarbon gases (Kvenvolden et al. 1979c). Meth- ane in this gas mixture has an isotopic composition of — 36°/oo, a value suggesting that it is derived mainly from thermal processes, probably operating at depth in Norton Basin. Geophysical evidence indicates that the hydrocarbon gases migrate into the near-surface sediments along a fault zone (Nelson et al. 1978c). Subbottom reflector terminations on continuous seismic profiles near the fault zone outline a large zone of anomalous acoustic responses about 9 km in diameter at a depth of 100 m, caused by a thick subsurface accumulation of gas. Gas geochemistry and extensive voids due to gas expansion in vibracores suggest a high degree of gas saturation at the seep site (Kvenvolden et al. 1979b). Biogenic gas cratering Small circular pits on the sea floor are found over a 20,000-km^ area of central and eastern Norton Sound (Fig. 15-7). The craters in the northern Bering Sea are young, as shown by their presence within modern ice-gouge grooves and by the fact that relict buried craters have not been observed in seismic profiles (Nelson et al. 1979b). These craters range from 1 to 10 m in diameter, averaging 2 m, and are probably less than 0.5 m deep. They are associated with many acoustic anomalies observed on seismic profiles and with subsurface Pleistocene peaty mud that is commonly saturated with biogenic methane (Holmes 1979b, Kvenvolden et al. 1979a, Nelson et al. 1979a). The extensive reflector-termination anom- alies and peat with a high gas content in east-central Norton Sound suggest that gas-charged sediment may be the cause of crater formation. Two basic mechanisms for gas venting are possible. First, continuous local degassing may maintain craters as active gas vents on the sea floor. Second and more likely, gas may be intermittently vented, particularly during severe storms when near-surface sediment may liquefy. The occurrence of surface craters in overlying marine sediment and the presence of high quantities of methane trapped beneath cohesive marine mud in Norton Sound suggest that gas venting may be episodic in this lithologic setting. Absence of craters in the noncohesive near-surface fine-to-medium sand and gravel of Chirikov Basin indicates that gas proba- bly diffuses gradually through this more porous sediment that overlies the peaty mud. Further evidence for intermittent venting of gas is the broad, shallow shape of the craters, unlike the deep, conical, actively bubbling vents of the thermogenic seep. Lack of methane in bottom water also suggests that the craters are not continually active vents. POTENTIAL GEOLOGIC HAZARDS Thermogenic gas cap The extent of active gas seepage into the water column and gas saturation in near -surface sediment above a thick sediment section with acoustic anom- alies suggest a possible hazard for future drilling activity in the thermogenic gas seep area south of Nome. Artificial structures penetrating the large gas accumulation at 100 m or intersecting associated faults that cut the gas-charged sediment may provide direct avenues for uncontrolled gas migration to the sea floor. Shallow gas pockets Gas-charged sediment creates potentially unstable surficial-sediment conditions in Norton Sound. Approximately 7,000 km^ of Norton Sound is underlain by acoustic anomalies with potential shallow gas pockets everywhere except under the Yukon prodelta (Holmes 1979a, Nelson et al. 1979a). Pipelines built across areas of these potential gas pockets may be damaged by stress induced by the unequal bearing strength of gas-charged and normal sediment, particularly if the near-surface sediment is undergoing liquefaction caused by cyclic loading of storm waves. The gas saturation and lateral and subsurface extent of any shallow gas pockets will have to be detailed in any site investigations for platforms or pipelines. Gas craters Gas craters cover a large area of north-central Norton Sound. During nonstorm conditions, near- surface gas in this area may be trapped by a layer 1-2 Sedimentary processes and potential geologic hazards 257 Figure 15-7. Distribution and density of biogenic gas-generated craters on sea floor of Norton Sound, sliowing isopachs of Holocene mud derived from the Yukon River and deposited since Holocene postglacial sea-level rise (from Thor and Nelson 1979). m thick of impermeable Holocene mud. We postulate that the gas escapes during periodic storms, forming craters at the surface. The storm processes initiate rapid changes in pore-water pressures because of sea-level setup, seiches, erosional unloading of cover- ing mud, and possible sediment liquefaction from cyclic wave loading (Clukey et al. 1980). Gas venting and sediment craters or depressions which seem to form during peak storm periods may be a potential hazard to offshore facilities because of the rapid lateral changes in bearing strengths and collapse of sediment which form the craters. Sediment collapse may also expose pipelines to ice-gouging hazards. During nonstorm conditions, the upper several meters of sediment at many locations has reduced shear strength because of the near-surface gas satura- tion and presence of peat layers. Choosing locations for structures will require extensive testing of the substrate to determine the extent and activity of gas cratering at a given site. Liquefaction The assessment and prediction of sea-floor stability are affected by the possibility that a sedimentary deposit will liquefy under cyclic loading and behave as a viscous fluid. The liquefaction potential of Norton Sound sediment is great in central Norton Sound and in the vicinity of the western Yukon prodelta (Clukey et al. 1980, Olsen et al. 1979). Possible causes of liquefaction include upward migration of gas from thermogenic and biogenic sources, earthquakes, and ocean waves. Bottom features that may be caused in part by liquefaction include scour depressions and much sediment crater- ing where Yukon sediment is thin. Loss of substrate support by sediment liquefaction must be considered in the construction of pipelines, drilling platforms, and other types of structures resting on the sea floor. Full assessment of this problem requires extensive studies of in-situ pore 258 Geology and geophysics pressure, gas saturation, and wave cyclic loading during storms. Ice scour The maximum intensity of ice gouging occurs in central Norton Sound at depths of 10-15 m in an area surrounding the Yukon Delta. The remaining area of Norton Sound, where depths are less than 10 m or more than 20 m, has a low density of gouging, or none at all. Special studies of nearshore areas off Nome and Port Clarence were made when they became potential centers for commercial develop- ment and activity. Offshore of Nome, the focal point for logistics in the northern Bering Sea because it is an area of ice divergence, only a few gouges were found in water more tham 8 m deep (Thor and Nelson, this volume). Several gouges were found at the northern end of Port Clarence spit and inside the tidal inlet, but again none occurred in water less than 8 m deep. Ice gouging presents some design problems and potential hazards to installations in or on the sea floor. Pipelines and cables should be buried at a depth that allows for maximum ice gouging of 1 m, plus a safety factor for combined effects with current scour around the western Yukon prodelta front or gas cratering in central Norton Sound. Current scour depressions The highest density of scour depressions in Norton Sound is in two areas: west and northwest of the Yukon delta, and southeast of Nome (Fig. 15-6). In areas of high density, artificial structures that disrupt current flow may cause extensive erosion of Yukon- derived silt or very fine grained sand and create potentially hazardous undercutting of the structures. Even buried structures such as pipelines may be subject to scour because strong currents can greatly broaden and deepen naturally occurring ice gouges, thus exposing the structures. The abundance and lateral extent of scour depressions are greatest where they occur with ice gouging in the Yukon Delta areas. Replicate surveys have shown that scour depressions recur annually. Full assessment of this geologic hazard requires long-term current monitoring in specific localities of scour to predict current intensity and periodicity, especially during severe storms: measured current speeds have increased more than 100 percent even under moderate storm conditions (Cacchione and Drake 1979b). Mobile bed forms Large migrating bedforms form an unstable sea floor in the area west of Port Clarence. The actual rates of bedform movement are not known, but development and decay of sand waves up to 2 m in height has been observed during a one-year period. Pipelines could be subject to damaging stress if free spans developed where the structure crossed such areas of migrating sand waves 2 m high. Studies to this time indicate that potential for the most extreme scour exists in regions of sand ribbons and gravel plus shell pavement within the strait. Sea floor relief changes most rapidly in the Port Clarence sand-wave area, where the scour in sand-wave troughs may reach depths of 2 m. Repli- cate surveys have shown that such scour may occur each year in some areas of the Port Clarence sand- wave field. Long-term monitoring of currents and bedform movement is particularly important in determining actual rates of change in this area, the only large natural harbor on the Alaskan coast north of the Aleutians. Coastal and offshore storm-surge hazards The northern Bering Sea has a history of major storm surges accompanied by widespread changes in sea-floor sedimentation (Cacchione and Drake 1979a, Fathauer 1975, Nelson 1977); these changes compli- cate maintenance of sea-floor installations and mass transport of pollutants. The November 1974 storm was the most intense measured in historic time, but storms of 1913 and 1946 also caused considerable damage (Fathauer 1975). Severe storms, such as that of November 1974, have caused extensive flooding along the Norton Sound coast between Nome and Unalakleet and on the St. Lawrence Island coast (Sallenger et al. 1978). At Nome, storm surge and waves overtopped a sea wall, causing damage reported at nearly 15 million dollars (Fathauer 1975). The periodicity and intensity of storm surges will have to be carefully studied in planning whether and where pipelines should come ashore in this area. Rapid sedimentation of thick storm-sand layers (15-20 cm) is a problem in the Yukon Delta area. Pipelines, offshore facilities, and other structures impeding the erosion, transport, and redeposition of sediment in southern Norton Sound will require careful design. Accurate monitoring of the storm- surge process wiU require long-term deployment of an array of current meters and tide gauges in the north- ern Bering Sea. Sedimentary processes and potential geologic hazards 259 ACKNOWLEDGMENTS We thank Keith Kvenvolden for information concerning gas in Norton Sound sediment, David Cacchione and David Drake for data on current and sediment movement, WiUiam Dupre for information on deltaic processes, Abby Sallenger and Ralph Hunter for data concerning inner shelf processes, and Harold Olsen and Edward Clukey for data on geotechnical properties. We are grateful for the assistance of Phyllis Swenson, Marybeth Gerin, and Joan Esterle in drafting and preparation of figures and for the assistance of Helen Ogle in preparing the manuscript. David Drake and David Hopkins made helpful review comments. The cruises were supported jointly by the U.S. Geological Survey and the Bureau of Land Manage- ment through interagency agreement with the Na- tional Oceanic and Atmospheric Administration under which a multiyear program responding to needs of petroleum development of the Alaska continental shelf is managed by the Outer Continental Shelf Environmental Assessment Program Office, Coachman, L. K., K. Aagaard, and R. B. Tripp 1975 Bering Strait: The regional physical oceanography. Univ. of Washington Press, Seattle. Dupr6, W. R., and R. Thompson 1979 The Yukon delta: A model for deltaic sedimentation in an ice-dominated environment. Proc. Offshore Tech. Conf., Paper No. 3434: 657-61. Fathauer, T. F. 1975 The great Bering Sea storms of 9-19 November, 1974. Weatherwise Mag., Amer. Meteorological Soc. 28: 76-83. Fleming, R. H., and D. Heggarty 1966 Oceanography of the southeastern Chukchi Sea. In: Environment of Cape Thompson Region, Alaska, N. J. Wilimovsky and J. M. Wolf, eds., 679-94. U.S. Atomic Energy Commis- sion. REFERENCES Cacchione, D. A., and D. E. Drake 1979a Sediment transport in Norton Sound, Alaska: U.S.G.S. Open-FUe Rep. 79-1555. 1979b Bottom shear stress generated by waves and currents in the northern Bering Sea. In: Holocene marine sedimentation in the North Sea basin, S. C. Nio, R. T. Schattenhelm, and T. C. E. Van Weering, eds. Spec. Pub. Inter. Assoc. Sedimentologists, Blackwell Scientific Pub., London (in press). Clukey, E. C, D. A. Cacchione, and H. W. Olsen 1980 Liquefaction potential of the Yukon prodelta. Proc. Offshore Tech. Conf., Houston, Texas, 1980, Paper No. 3773, in press. Holmes, M. L. 1979a Distribution of gas-charged sediment in Norton Basin, northern Bering Sea. In: Holocene marine sedimentation in the North Sea basin, S. C. Nio, R. T. Schattenhelm, and T. C. E. Van Weering, eds. Spec. Pub. Inter. Assoc. Sedimentologists, Blackwell Scientific Pub., London (in press). 1979b Distribution of gas-charged sediment in Norton Sound and Chirikov Basin. Environmental assessment of the Alaskan Continental Shelf, Ann. Rep. Principal Investigators for the year ending March 1979, Environ. Res. Lab., Boulder, Colo., NOAA, U.S. Dep. of Commerce 10:75-94. Hopkins, D. M., C. H. Nelson, R. B. Perry, and T. R. Alpha 1976 Physiographic subdivisions of the Chirikov Basin, northern Bering Sea. U.S.G.S. Prof. Paper 759-B. 260 Geology and geophysics Hunter, R., and D. R. Thor 1979 Depositional and erosional features of the northeastern Bering Sea inner shelf. In: Holocene marine sedimen- tation in the North Sea basin, S. C. Nio, R. T. Schattenhelm, and T. C. E. Van Weering, eds. Spec. Pub. Inter. Assoc. Sedimentologists, Blackwell Scientific Pub., London (in press). Kvenvolden, K. A., J. B. Rapp, and C. H. Nelson 1978 Low molecular weight hydrocarbons in sediments from Norton Sound (abs.). Amer. Assoc. Petroleum Geol. Bull. 62: 534. Kvenvolden, K. A., C. H. Nelson, D. R. Thor, M. C. Larsen, G. D. Redden, J. B. Rapp, and D. J. Des Marais 1979a Biogenic and thermogenic gas in gas-charged sediment of Norton Sound, Alaska. Proc. Offshore Tech. Conf., Paper No. 3412. Kvenvolden, K. A., G. D. Redden, and C. H. Nelson 1979b Gases in near-surface sediment of the northern Bering Sea. In: Holocene marine sedimentation in the North Sea basin, S. C. Nio, R. T. Schatten- helm, and T. C. E. Van Weering, eds. Spec. Pub. Inter. Assoc. Sedi- mentologists, Blackwell Scientific Pub., London (in press). Kvenvolden, K. A., K. Weliky, and C. H. Nelson 1979c Submarine seep of carbon dioxide in Norton Sound, Alaska. Science 205: 1264-6. Larsen, M. C., C. H. Nelson, and D. R. Thor 1979a Continuous seismeic reflection data, S9-78-BS cruise, northern Bering Sea. U.S.G.S. Open File Rep. 79-1673. 1979b Geologic implications and potential hazards of scour depressions on Bering shelf, Alaska. Environ. Geol. 3: Nelson, C. H. 1977 Storm surge effects. Environmental assessment of the Alaskan continental shelf, Ann. Rep. Principal Investi- gators for the year ending March 1977, Environ. Res. Lab., Boulder, Colo., NOAA, U.S. Dep. of Commerce 18: 111-19. Nelson, C. H., and J. S. Creager 1977 Displacement of Yukon-derived sediment from Bering Sea to Chukchi Sea during the Holocene. Geology 5: 141-60. Nelson, C. H., M. E. Field, D. A. Cacchione, and D. E. Drake 1978a Activity of mobile bedforms on Bering shelf. Environmental assess- ment of the Alaskan Coi)tinental Shelf, Ann. Rep. Principal Investi- gators for the year ending March 1978, Environ. Res. Lab., Boulder, Colo., NOAA, U.S. Dep. of Com- merce 12:291-307. Nelson, C. H., M. E. Field, and W. R. Dupr6 1979a Linear sand bodies on the Bering epicontinental shelf. In: Holocene marine sedimentation in the North Sea basin, S. C. Nio, R. T. Schatten- helm, and T. C. E. Van Weering, eds. Spec. Pub. Inter. Assoc. Sedi- mentologists, Blackwell Scientific Pub., London (in press). Nelson, C. H., M. L. Holmes, D. R. Thor, and J. L. Johnson 1978b Continuous seismic reflection data, S5-76-BS cruise, northern Bering Sea. U.S.G.S. Open FUe Rep. 78-609. Nelson, C. H., and D. M. Hopkins 1972 Sedimentary processes and distribu- tion of particulate gold in the north- em Bering Sea. U.S.G.S. Prof. Paper 689. McManus, D. A., V. Kolla, D. M. Hopkins, and Nelson, C. H., K. A. Kvenvolden, and E. C. Clukey C. H. Nelson 1978c Thermogenic gas in sediment of 1977 Distribution of bottom sediments on Norton Sound, Alaska. Proc. Off- the continental shelf, northern Bering shore Tech. Conf., 1978, Paper No. Sea. U.S.G.S. Prof. Paper 759-C. 3354: 1612-33. Sedimentary processes and potential geologic hazards 261 Nelson, C. H., D. R. Thor, M. W. Sandstrom, and K. A. Kvenvolden 1979b Modern biogenic gas-generated craters (sea-floor "pockmarks") on the Bering shelf, Alaska. Geol. Soc. Amer. Bull. 90:1144-52. Olsen, H. W., E. C. Clukey, and C. H. Nelson 1979 Geotechnical characteristics of bot- tom sediments in the northern Bering Sea. In: Holocene marine sedimenta- tion in the North Sea basin, S. C. Nio, R. T. Schattenhelm, and T. C. E. Van Weering, eds. Spec. Pub. Inter. Assoc. Sedimentologists, Blackwell Scientific Pub., London (in press). Sallenger, A. H., J. R. Dingier, and R. Hunter 1978 Coastal processes and morphology of the Bering Sea coast of Alaska. Environmental Assessment of the Alaskan Continental Shelf, Ann. Rep. Principal Investigators for the year ending March 1978, Environ. Res. Lab., Boulder, Colo., NOAA, U.S. Dep. of Commerce 12: 451-70. Thor, D. R., and C. H. Nelson 1978 Continuous seismic reflection data. S5-77-BS cruise, northern Bering Sea. U.S.G.S Open File Rep. 78-608. 1979 A summary of interacting surficial geologic processes and potential geo- logic hazards in the Norton Basin, northern Bering Sea. Proc. Offshore Tech. Conf., Paper No. 3400: 377-81. Thor, D. R., C. H. Nelson, and R. O. Williams 1978 Environmental geologic studies in northern Bering Sea. In: U. S. Geological Survey in Alaska, accom- plishments during 1977, K. M. Blean, ed., U.S.G.S. Circular 772-B: B94- B95. Young, R. N., and J. B. Southard 1978 Erosion of fine-grained marine sedi- ments. Sea-floor and laboratory exper- iments, Geol. Soc. Amer. Bull. 89: 663-72. The Ice-dominated Regimen of Norton Sound and Adjacent Areas of the Bering Sea Vema M. Ray^ and William R. Dupre^ ^ Presently at: MGF Oil Company Houston, Texas ^ Department of Geology University of Houston Houston, Texas ABSTRACT The patterns of ice formation, movement, and deformation in Norton Sound and adjacent areas of the Bering Sea were studied using LANDSAT and NOAA satellite imagery for the years 1973-77. The results demonstrate not only the marked seasonality of marine processes, but also the significant role of bathymetric and meteorologic conditions in controlling ice movement in this high-latitude, epicontinental sea. The ice-dominated regimen of the Norton Sound region begins with ice freezeup in October and lasts through the winter to spring breakup in May. Freezeup begins as tempera- tures drop in early fall and ice starts to accumulate around the Yukon Delta; during that time water and sediment discharge from the Yukon River becomes insignificant. Oceanographic currents are relatively ineffective in transporting ice during the winter months, as strong northerly winds generally control ice movement in the winter phase. Consequently, ice di- vergence from the northern coast of Norton Sound and ice convergence along the northern margin of the Yukon Delta front are common throughout the winter phase. Ice ridging and associated gouging result from the compaction of shore- fast ice along the northern delta front. Shearing and gouging also occur along the western margin of the delta where pack ice, pushed southward by the winds, shears past the shorefast ice boundary. Periods of rapid advection of pack ice from the Chukchi Sea through the Bering Strait are also common in the winter phase. Such events cause ice floes to move as much as 45 km per day, usually along a relatively narrow band on the eastern side of the Bering Sea. During May, winds become predominantly offshore and warming temperatures begin to melt the ice and snow in the interior. The higher temperatures and increased discharge from the rivers trigger ice breakup along the coast. North- ward-flowing water currents aided by offshore winds carry ice away from the delta. By June the high sediment influx from the Yukon River dominates the coastal processes in the delta region. INTRODUCTION The prospect of oil and gas exploration in Norton Sound (Fig. 16-1) has focused increased attention on the marine processes that characterize the region. Figure 16-1. Location of study area. Because these processes vary systematically through- out the year, it is possible to recognize ice-dominated, river-dominated, and storm -dominated regimens (Fig. 16-2); each regimen consists of a characteristic suite of geologic processes and associated potential geologic hazards (Dupre and Thompson 1979). The purpose of this paper is to discuss the ice- dominated regimen of Norton Sound from freezeup, 263 264 Geology and geophysics COASTAL TEMPERATURE PERCENT ICE COVER C. STORM FREQUENCY D. RIVER DISCHARGE Figure 16-2. Diagram illustrating the seasonality of processes in the Yukon Delta/Norton Sound region of the Bering Sea. Sources of data include: (a) Summary of average monthly temperatures at Unalakleet, 1941-70 (NOAA); (b) Summary of ice observations for Yukon Delta region from Brower et al. (1977); (c) Frequency of major low-pressure centers in the northern Bering Sea region, from Brower et al. (1977); and (d) Discharge of the Yukon River at Kaltag, 1962 (U.S.G.S. Water Resources Data). typically in late October or early November, to breakup in May (Fig. 16-3). Of particular interest is the extent and variability of shorefast ice, the pat- terns and rates of movement of seasonal pack ice, and the weather conditions under which such movement occurs. Most of this work has been based on the interpretation of satellite imagery, but we hope that this study can provide background for more detailed studies based on ground monitoring and in-situ measurements. ICE-DOMINATED REGIMEN FREEZEUP I I I BREAKUP I Figure 16-3. Three phases of the ice-dominated regimen. Previous work Most of the work on ice dynamics and the possible effect of ice on offshore petroleum development has been concentrated in Arctic regions such as the Beaufort Sea; relatively little work has been published to date on ice movement and morphology in sub- arctic regions such as the Bering Sea and Norton Sound. Muench and Ahlnas (1976) were the first to use satellite imagery to study regional patterns of ice movement in the northern Bering Sea; the relatively low resolution of their imagery (NOAA weather satellite data) and the relatively short period of record (March to June 1974), however, limited the utility of their work. Shapiro and Burns (1975) used higher -resolution LAND SAT imagery to document a short-lived ice deformation event just to the north of the Bering Strait. Stringer (1977, 1978) mapped a variety of ice-related features in the Beaufort, Chukchi, and Bering Seas with the use of LANDSAT imagery. Similarly, Dupre (1978) used LANDSAT imagery to study the complex interrelationships between ice and patterns of deltaic sedimentation associated with the Yukon Delta. This present study is designed to expand on these previous studies, emphasizing the patterns and rates of ice movement in Norton Sound. In doing so, the study also pro- vides information to aid in the explanation and extrapolation of patterns of ice gouging in Norton Sound as described by Thor et al. (1978). Methods of study The data were compiled from imagery acquired by the Multispectral Scanner system of the LANDSAT and NOAA-2, 3, and 4 satellites. Meteorologic data were also taken from daily surface synoptic charts from the National Climatic Center in North Carolina. The extent of shorefast ice and ice floes was mapped from LANDSAT images (1:1,000,000) on acetate overlays that were superimposed on standard bathy- metric base maps of the northern Bering Sea. Overlays for successive days could be superimposed to chart the movement of particular ice floes over an approximate 24-hour period. Images were registered with respect to landforms in order to map the posi- tions of the ice floes on successive days. In general, sea ice movement cannot be accurately monitored by referencing the scenes to coordinates, because co- ordinates provided on the margin of LANDSAT images allow for only approximate registration (Colvocoresses and McEwen 1973). According to Colvocoresses and McEwen (1973), the systematic, root mean square error of position for points on the satellite images ranges from 200 to 450 meters, with no detectable additional error associated with image duplication. Since the sea ice is moving on the order of kilometers per day, this error should be considered insignificant for the purposes of this paper. The LANDSAT images were band 5 and 7, 9 X 9-inch positive prints. The images and the base map seemed to be perfect overlays. The band-5 images Ice-dominated regimen of Norton Sound 265 proved to be most useful in defining nilas ice (a thin, elastic crust), whereas band 7 was more useful for defining the sea-ice boundary and delineating pack-ice floes and areas of shorefast ice. There is no dis- tinction made between newly formed ice and open water on most maps in this paper, because of the difficulty in distinguishing the two . The NO A A 10 X 10-inch satellite photoprints (infrared) furnished only general meteorological information. They did provide a good overview of ice movement, but were not used for detailed measure- ments. Wind patterns and weather systems could be observed from NO A A imagery, but wind velocity and directional information was obtained from daily surface synoptic charts. Although some weather station readings are influenced by local orographic conditions, most of the information is believed to be useful for the purposes of this study. ICE-DOMINATED REGIMEN The pattern of ice formation, movement, and deformation in the Norton Sound region is sig- nificantly affected by the nearshore morphology of the Yukon Delta. The Yukon River has formed an ice-dominated delta characterized by a broad, shallow sub-ice platform crossed by sub-ice channels that extend as far as 25 km beyond the major distrib- utaries (Fig. 16-4). The platform is characterized by relatively stable shorefast ice for much of the year, whereas the sub-ice channels have more dynamic ice and sediment movement, particularly during breakup. The more steeply dipping delta front has relatively intense ice deformation and related gouging, whereas the more distal part of the delta is an area of rela- tively complex seasonal pack-ice movement. Because of the complexity of ice movement, both in space and time, the intervals of freezeup, winter, and breakup will be discussed separately. Freezeup In late October, as coastal temperatures drop below 0 C, ice crystals typically begin to form and accumulate as new ice along the shore of Norton Sound. Bottomfast ice forms along the shallow margins of the delta (e.g., on intertidal mudflats and subaqueous levees); some of the smaller sub-ice channels begin to be covered by floating fast ice. Since the larger sub-ice channels that extend beyond the main distributaries are relatively deep and continue to maintain a channelized flow of water offshore, they are the last of the nearshore areas to freeze. The shorefast ice continues to expand farther offshore in November, until the ice reaches a maxi- mum width of from 15 to 30 km, approximately coincident with the limits of the sub-ice platform. Most of the shorefast ice is floating fast ice (Fig. 16-4), separated from the bottom fast ice by active tidal cracks that coincide approximately with the 1 m isobath. The inner zone of bottomfast ice is typically covered with a sheet of ice (aufeis) (Fig. 16-5), which forms by water flowing over bottomfast ice due to the tide- or storm-induced rise and fall of floating fast ice. The shorefast ice continues to expand seaward until it encounters mobile seasonal pack ice. At this time pressure ridges develop, become grounded, and a seaward-accreting stamukhi zone develops approximately coincident with the delta front in water depths of 5-15 m. The stamukhi zone is well developed by the beginning of December and, for the purpose of this report, marks the begin- ning of the winter period. Winter The winter phase of the ice-dominated regimen is characterized by the establishment of a relatively stable band of shorefast ice fringed by a complex zone of ice deformation features that form the stamukhi zone (as defined by Reimnitz et al. 1977). The patterns of ice movement seaward of the stamukhi zone are rather complex, reflecting both local and regional meteorologic events as well as the effect of bathymetry in deflecting and grounding ice floes. Seasonal pack ice in Norton Sound is typically 0.7-1.2 m thick (Brower et al. 1977) and is largely in situ, having formed within a zone of ice divergence in the northeastern side of the sound. It usually flows to the west and southwest (Fig. 16-6) in response to the predominant northeasterly winds resulting from a relatively stable high-pressure system that develops during the winter months. Sluggish ice movement controlled by relatively weak oceanic currents occurs only during periods of weak winds. The southwest- flowing ice commonly converges against the shorefast ice fringing the Yukon Delta to form a broad stamukhi zone. The deformation of pack ice within the stamukhi zone causes pressure ridge raking which produces numerous parallel furrows as the ice keels plow through the bottom sediment (Reimnitz and Barnes 1974). As expected, the area of extensive ice gouging delineated by Thor et al. (1978) generally coincides with and parallels the stamukhi zones as delineated in this study. Ice seaward of the stamukhi zone is deflected to the west, where it leaves Norton Sound and joins the stream of rapidly moving Bering DEPOSITIONAL ENVIRONMENTS OF THE YUKON DELTA RELICT SEDIMENT KILOMETERS MHW MLW -DELTA MARGIN- DELTA PLAIN sub-ice platform tidal flat DELTA FRONT--><^'f<:^^k(^sub-ice'' channel ■rr^. .. PRODELTA _ ,- ■■_^~~- . ^ IDEALIZED PROGRADATIONAL SEQUENCE Figure 16-4. Depositional environments of the modern lobe of the Yukon Delta (from Dupre and Thompson 1979). 266 Ice-dominated regimen of Norton Sound 267 Figure 16-5. Location of shear zones around tiie Yukon delta, as determined from LANDSAT imagery, 3/13/73 to 3/6/75. Hatching delineates areas of overflow icing (auteis). Dotted area delineates the 5-10 m bathy- metric interval, which is ap- proximately coincident with the delta front (see Fig. 16-4). SHEAR ZONE SEASONAL PACK ICE STAMUKHI ZONE -SHOREFAST ICE- Floating fast Bottomfast ice ' ice Yrr^rrrrrrrT^^-^TTrr^' ,i;frr^^'^^'^- e^ "> of sub-ice channel) Sea pack ice. Pack ice rarely enters the sound from Norton Sound is relatively complex. In general the the west except during prolonged periods of strong ice flows mainly to the south in response to the westerly winds. prevailing northerly winds. Significant reversals in The pattern of ice movement to the west of the path of the flow can occur, however, as illustrated Figure 16-6. Patterns of ice movement for 14-15 March 1974. Arrows indicate the direction and amount of ice move- ment in one day. Dashed lines within the shorefast ice to the north of the delta indicate inactive shear zones. 268 Ice-dominated regimen of Norton Sound 269 on the LANDSAT and weather data for 23-27 Feb. 1976 (Fig. 16-7). During this period, winds began to blow up to 25 knots from the south. As a result the southerly flow of ice ceased and northerly movement of ice as fast as 15 km/day was charted (Fig. 16-7). The rapid reversal was the result of the passage of two large low-pressure systems (Fig. 16-8A and B) and illustrates the extent to which ice movement is responsive to meteorologic events. Shapiro and Burns (1975) noted that unusually strong northerly winds can result in a major ice deformation event in which masses of ice are de- formed and funneled out of the Chukchi Sea and through the Bering Strait. Similar deformation features were noted on NOAA (VHRR) imagery for 13-15 March 1976. During this time, ice floes west of the Yukon Delta were moving as fast as 45 km/day (Fig. 16-9), presumably in response to a major ice deformation event similar to that described by Shapiro and Burns. It is important to note that the zone of very rapid ice movement is restricted to a relatively narrow band approximately 70 km wide, which is referred to as the "racetrack." This zone is often recognized as a band of highly fractured nilas ice (Fig. 16-7) that presumably forms as the source of the pack ice (the Chukchi Sea) becomes temporarily plugged at the Bering Strait. The racetrack can be seen on LANDSAT imagery throughout much of the winter and recurs annually. The eastern boundary coincides approximately with the 20-m isobath, suggesting that this isobath may reflect the grounding of ice flows at the entrance to Norton Sound, and effectively deflects the ice to the south, parallel to the isobaths. There is often a zone of open water between the Bering Sea pack ice and the edge of the shorefast ice west of the Yukon Delta. This typically forms during periods of easterly offshore winds, but may be com- pletely closed during periods of onshore winds. Some of it may, however, be caused by relatively deeper ice keels, grounded in water depths of 15-18 m, acting as an offshore ice barrier to some onshore movement of ice. This zone of open ice west of the delta is of particular interest to the natives in the area, as it greatly facilitates winter hunting. Little, if any, sediment enters Norton Sound during the winter months, and yet the suspended sediment concentration measured beneath the ice in the west-central part of Norton Sound is essentially as high in winter as it is during much of the summer (personal communication, David Drake, U.S.G.S.). This suggests that a significant amount of sediment initially deposited during the summer months is resus- pended during the winter and hence is available to be redistributed by sub-ice currents. The unusual off- shore increase in sand off the Yukon Delta (Dupre and Thompson 1979) may come from such a process, but the exact mechanism(s) for such resuspension are unclear. Breakup Breakup along the coast is a relatively brief event that marks the transition between the ice-dom- inated and river-dominated regimens; the significance of breakup, however, far outweighs its brief duration. River breakup along the Yukon, as with most of the coastal rivers in northern Alaska, is marked by a tremendous increase in sediment and water discharge, resulting in ice jams, extensive inland flooding, and riverbank erosion. As river discharge begins to increase, floating fast ice begins to lift, both in the river and along the coast. During this time, the sub-ice channels are especially well delineated by the floating fast ice. The bottomfast ice begins to be flooded by an over-ice flow (Fig. 16-10), which has been described for the North Slope by Reimnitz and Bruder (1972) and Walker (1974). Some sediment is carried onto the ice, thereby effectively bypassing much of the inner sub-ice plat- form. Much of the sediment seems to remain in the sub-ice channels, which cross the sub-ice platform. Some of the sediment is probably deposited from suspension on subaqueous levees farther offshore; much of it, however, bypasses the sub-ice platform completely and is deposited on the delta front or prodelta. The floating ice that marks the sub-ice channels soon breaks up and is removed to sea. Much of the over-ice flow may drain through strudel holes (as defined by Reimnitz and Bruder 1972) or cause the bottomfast ice to melt in place. Large pieces of floating fast ice break off to be transported farther offshore. Grounded ice may remain in some shallow areas to the northwest of the delta; floating pack ice may remain trapped in the middle of Norton Sound because of the sluggish currents. Figure 16-11 illustrates conditions that are typical during breakup. Floe movement is to the north as is common during late April and May when northerly winds die down and northward-flowing currents become more effective in transporting ice (cf. Muench and Ahlnas (1976). The floes were moving up to 20 km/day on 7 May 1974 (Fig. 16-10). A low-pressure system moved into the area on the 8th, however, bringing a temporary restoration of winter-phase northerlies. Temporary reversals of ice movement towards the south were also noted in late May 1974 by Muench and Ahlnas (1976). These short-lived events are typically associated with {^ WIND DIRECTION / ICE DRIFT VECTOR v- 23*7S FEBfiUAflX J8 y 25-26 FfBRU»r 1/ 7B- 7 I F(BRUA 0 0 20 50 40 V3 X** BATHYMETRY IN METERS Figure 16-7. Patterns of ice movement for 23-27 February 1976. Arrows indicate the direction and amount of ice movement in one day. Dasiied lines within the shorefast ice indicate inactive shear zones. Note the reversal in ice movement between 25-26 February and 26-27 February. 270 Ice-dominated regimen of Norton Sound 271 Figure 16-8. Weather condi- tions for the period 24 February to 27 February 1976 (from NOAA Surface Synoptic weather charts). Note the reversal of winds in the Bering Strait area due to the westward migration of low-pressure systems a and b. 12:00 AM GMT 27 FEB 76 the passage of a low-pressure system crossing the usually no longer present, although some areas of Bering Sea. Southerly winds typically follow, how- unconsohdated pack-ice floes may be present, par- ever, facilitating the breakup and removal of the ticularly in the center of Norton Sound. At this time shorefast ice. By early June the shorefast ice is the distributary channels have been cleared of ice and UNALAKLEET* Cape Roman -61' if Or- - ^ OPEN WATER {} WIND DIRECTION / ICE DRIFT VECTOR / n-l4 MARCH 19 ^14-18 MARCH It 0 10 ^o SO *o VI km BATHYMETRY IN METERS 166 I *^u L. Figure 16-9. Patterns of ice movement for 13-15 March 1976. Arrows indicate direction and amount of ice movement in one day (up to 45 km/day). 272 sub-ice channeljs open OPEN WATER -c^ TT- Sub-ice flow Over-ice flow •>^>- V^~^~ >;f:p..:*-7^ V^i*'^ Figure 16-10. Diagram of sediment dispersion patterns during spring breakup. Note the role of over-ice flow and sub-ice channels in providing mechanisms by which sediment may by pass the inner part of the sub-ice platform. 273 OPEN WATER {} WIND DIRECTION ■X ^C\)e / ICE DRIFT VECTOR ^ B-S MAY 74 0 (0 20 » 40 » XU BATHYMETRY IN METERS Figure 16-11. Patterns of ice movement for 7-9 May 1974. Arrow indicates direction and amount of ice movement in one day. The pack ice is much less consolidated than earlier in the year. 274 Ice-dominated regimen of Norton Sound 275 \ are introducing an apron of sediment-laden water over much of the prodelta region, marking the beginning of the river-dominated regimen. SUMMARY The patterns of ice formation, movement, and deformation in the Norton Sound region were studied with LAND SAT and NO A A satelUte imagery for the years 1973-77. The results document not only the marked seasonality of marine processes throughout the year, but also the significant role of bathy metric and meteorologic conditions in controlling the patterns and rates of ice movement in the region. The results have been summarized in a map (Fig. 16-12) of generalized ice hazards similar in some respects to maps done for the entire Bering Sea by Stringer (1978). The following is a brief summary of the types of ice-related hazards that characterize each of the zones. Zone la is a zone of shorefast ice that extends to the outer edge of the sub-ice platform of the Yukon Delta, approximately coincident with the 2-3 m isobath. Over-ice flow occurs throughout the winter in areas of bottomfast ice near the major distrib- utaries (shown in hatched pattern). Sub-ice currents beneath the floating fast ice may result in some resuspension of sediments in the sub-ice channels and on the outer edge of the sub-ice platform. This is a relatively stable zone throughout the winter, but large sheets of ice may break off during spring breakup. Zone lb is a slightly less stable area characterized by floating fast ice during most of the winter; the area can, however, be completely ice free under some conditions (e.g. 13-15 March 1976). Zone Ic is the zone of shorefast ice that fringes most of Norton Sound. It consists largely of floating fast ice, more variable in extent and less stable than ice in Zones la and lb, because large sheets of ice may break off repeatedly throughout the winter. Zone Ila is a broad, seaward-accreting stamukhi zone formed by the convergence and deformation of ice originating mainly in Norton Sound. The con- figuration of the outer margin of this zone seems to be controlled by Stuart Island to the east and a series of offshore shoals to the west, and coincides approximately with the 14 m isobath. This zone is characterized by extensive ice shearing and a rela- tively high intensity of ice gouging of the sea floor (Thor et al. 1978). Zone lib is west of the delta in water 3-14 m deep. It is a relatively unstable area characterized by ice deformation and accretion to the shorefast ice (Zone la) during periods of westerly onshore winds, or an offshore movement of ice and the development of a large polynya (open water area) during periods of easterly offshore winds. This zone is also characterized by moderately high density of ice gouging. Zone III is an area of seasonal pack ice formed mainly in situ, within Norton Sound. The ice typi- cally moves south and west in response to the domi- nant northeasterly winds throughout the winter, but it may drift slowly in response to oceanic cur- rents during periods of low winds. The southern portion of this zone is characterized by widespread shearing of ice and coincides approximately with the area of very high density of ice gouging delineated by Thor et al. (1978). The western boundary coincides approximately with the 20-m isobath, separating pack ice formed in Norton Sound from the thicker pack ice formed farther to the north. Bering and Chukchi pack ice enter the sound only rarely when especially strong northwesterly winds blow. Zone IV consists of seasonal pack ice formed in the northern Bering and Chukchi Seas. The ice typically moves southward in response to northerly winds for most of the winter. Short periods of northerly ice movement can, however, occur during the passage of low-pressure systems. It is typical for the ice to begin to move consistently to the north in late April or early May. Zone IVa is the "racetrack," characterized by intervals of extremely rapid southerly movement of pack ice (up to 45 km/day) following major ice-deformation events north of the Bering Strait (described by Shapiro and Burns 1975). This zone is characterized by highly fractured nilas ice during periods of relative quiescence. The eastern margin of this zone is approximately coincident with the 22 m isobath. The more variable western margin appears to be controlled by the geometry of ice piling up on the northern side of St. Lawrence Island. The rapid ice movement in this zone, combined with the lack of ice gouging (Thor et al. 1978) suggests that this is a zone where ice is unlikely to affect the bottom. Zone IVb, in water depths of 20-22 m, is characterized by less rapid ice movement than in the racetrack. Some grounded ice may occur in this zone, particularly in the area of shoals southwest of the delta. Zone IVc, in water depths of 14-20 m, is characterized by open water during periods of easterly winds and onshore moving pack ice during periods of westerly winds. This zone rarely has a stamukhi zone accreted to the shorefast ice, although some grounded ice and ice gouging will occur. Zone IVd is similar to zone IVb, and was not studied in detail. 276 Geology and geophysics -62 + 0 10 20 30 40 50 KM BATHYMETRY IN METERS ^3 1 SHOREFAST ICE II STAMUKHI ZONE III NORTON PACK ICE IV BERING PACK ICE V ICE FORMING ZONE / Dominant Direction of Ice Movement Figure 16-12. Zonation of ice hazards in tiie Yukon Delta/Norton Sound region based mainly on LANDSAT and NOAA satellite imagery, supplemented information on ice gouging from Thor and Nelson (1979). Zones Va and Vb are zones of ice divergence ACKNOWLEDGMENTS formed by persistent offshore winds (cf. Muench and Ahlnas 1976). They are typically areas of open water This study is supported in part by the Bureau of where ice is actively forming for most of the winter. Land Management through interagency agreement Ice-dominated regimen of Norton Sound 277 with the National Oceanic and Atmospheric Ad- ministration, under which a multiyear program responding to the needs of petroleum development of the outer continental shelf is managed by the Outer Continental Shelf Environmental Assessment Program (OCSEAP) office. We also wish to thank Dave Hopkins (U.S.G.S., Menlo Park), who conceived and initiated the study of coastal processes along the Yukon Delta, as well as Erk Reimnitz and Peter Barnes (U.S.G.S., Menlo Park), who provided in- valuable insight into the role of ice in sedimentary processes. Reimnitz, E., and P. Barnes 1974 Sea ice as a geologic agent on the Beaufort Sea. In: The coast and shelf of the Beaufort Sea, J. C. Reed and J. E. Sater, eds., 301-354. Arctic Inst. N. Amer. Reimnitz, E., and K. F. Bruder 1972 River discharge into an ice-covered ocean and related sediment dispersal, Beaufort Sea, coast of Alaska. Geol. Soc. Amer. Bull. 83: 861-6. REFERENCES Brower, W. A. Jr., H. F. Diaz, A. S. Prechtel, H. W. Searby, and J. L. Wise 1977 Climatic atlas of the outer continental shelf waters— Coastal region of Alaska: II, Bering Sea. Arctic Environmental Information and Data Center, Anchorage. Reimnitz, E., L. J. Toimil, and P. W. Barnes 1977 Stamukhi zone processes: impli- cations for developing the Arctic Coast. In: Proc. Offshore Tech. Conf., May 2-5, 1977, OTC paper 2945: 513-18. Shapiro, L., and J. J. Burns 1975 Satellite observations of sea ice movement in the Bering Strait Regions. In: Climate of the Arctic, G. Weller and S. A. Bowling, eds., 379-86. Univ. of Alaska, Fairbanks. Colvocoresses, A. P., and R. B. McEwen 1973 Progress in cartography, EROS pro- gram. Symposium on significant results obtained from ERTS-1 NASA/GSFC, March 5-9, 1973. Dupre, W. R. 1978 Dupre, W. R., 1979 Yukon Delta coastal processes study. Ann. Rep. of principal investigators for the year ending March 1978, NOAA-OCSEAP. and R. Thompson The Yukon Delta: a model for deltaic sedimentation in an ice-dominated environment. Proc. 11th Ann. Offshore Tech. Conf., OTC paper 3434: 657-64. Muench, R. D., and K. Ahlnas 1976 Ice movement and distribution in the Bering Sea from March to June, 1974. J. Geophys. Res. 81:(24) 4467-76. Stringer, W. J. 1977 1978 Morphology of Beaufort, Chukchi, and Bering Seas nearshore ice con- ditions by means of satellite and aerial remote sensing. Ann. Rep. of principal investigators for the year ending March 1977, NOAA-OCSEAP 15: 42-110. Morphology of Beaufort, Chukchi, and Bering Seas nearshore ice con- ditions by means of satellite and aerial remote sensing. Final Rep., NOAA- OCSEAP Contract No. 035-022-55. Thor, D. R., and C. H. Nelson 1979 A summary of interacting, surficial geologic processes and potential geol- ogical hazards in the Norton Sound Basin, northern Bering Sea. Proc. 11th Ann. Offshore Tech. Conf., OTC paper 3400: 377-85. 278 Geology and geophysics Thor, D. R., H. Nelson, and R. O. Williams Walker, H. J. 1978 Potential hazards of ice gouging over 1974 The Colville River and the Beaufort the Norton Sound Basin sea floor. In: Sea: Some interactions. In: The Ann. Rep. principal investigators for coast and shelf of the Beaufort Sea, J. the year ending March 1978, NOAA- C. Reed and J. E. Sater, eds., 513-40. OCSEAP. Arctic Inst. N. Amer. Ice Gouging on the Subarctic Bering Shelf 17 Devin R. Thor and C. Hans Nelson U.S. Geological Survey Menlo Park, California ABSTRACT Ice striking the sea floor gouges surficial sediment of the shallow Bering epicontinental shelf of Alaska. Two types of ice gouge have been recognized: the single gouge, a single furrow, and multiple gouges or raking, a wide zone of numer- ous, subparallel furrows. Single gouges, the most common type, are cut by single-keeled pieces of thick ice, whereas multiple gouges are formed by multikeeled, thick, pressure- ridge ice. Gouges occur in water depths of 30 m or less, but are most dense in water 10 to 20 m deep. Although some gouge incisions are as deep as 1 m, most are 0.5 m or less. Ice gouges trend parallel to pack-ice movement, which in turn is generally parallel to isobaths and coastline configuration. Mean gouge trend in Norton Sound is west -east, in the north- eastern Bering Sea north-south. The annual ice cover in this subarctic setting is thin (less than 2 m). Ice thick enough to gouge the substrate forms in compression and in shear zones; there moving pack ice collides with and piles up against other pack ice or stationary shorefast ice to develop pressure ridges. Southward-moving pack ice in the northeastern Bering Sea and westward-moving pack ice in Norton Sound converge vidth, and shear past, a shorefast ice zone 10-30 km wide that covers the shallow water offshore of the Yukon Delta. The intensity of ice deformation in this zone causes the highest gouge density in the study area. In contrast, northeastern Norton Sound is an area of ice divergence and only minimal ice gouging. The rest of Norton Sound and northeastern Bering Sea is either in ice-divergence areas or else water depths are too great for ice to touch bottom, and thus ice gouge density is low. Gouging is extremely rare inshore of the shear zone, because shorefast ice is relatively static and protects inshore areas from the dynamics of the shear or compression zone and consequent ice gouging. INTRODUCTION Development of natural resources in northern latitudes has led to increased research on the effects of ice on shelf sediment in arctic regions like the Beaufort Sea (Reed and Sater 1974, Reimnitz et al. 1973, Reimnitz et al. 1977, Barnes et al. 1978). Until recently, however, research on ice gouging had not been done in subarctic regions like the Bering Sea. A variety of gouge features are found in many areas of the northeastern Bering Sea, even though ice conditions there are not so severe as in high-latitude arctic regions. Ice gouging into the sea floor is a potential hazard to future resource development and such sea-floor installations as pipelines and wellheads. This chapter discusses general ice conditions and ice movement in the northeastern Bering Sea, the effect of ice as an erosional and depositional agent that influences the geomorphology and depositional history of the shallow subarctic Bering Sea shelf, and ice gouging as a potential hazard to resource devel- opment in and around Norton Basin. The terminol- ogy used is adopted from Barnes et al. (1978), particularly the word "gouge" to describe the feature and the process of ice interacting with the sea floor. Geographic setting The floor of the northeastern Bering Sea is a broad, shallow epicontinental shelf (Figs. 17-1 and 17-2). Water depths in Chirikov Basin range from 20 m on the eastern side to 50 m in the central part. The shelf is generally flat and featureless except for a prominent series of ridges and swales that subparallel the coastline off Port Clarence. A large, elongate marine reentrant forms Norton Sound, bounded on the north by Seward Peninsula, on the east by the Alaskan mainland, and on the south by the Yukon Delta. Except in a broad trough in the northern part of the sound, where depths are as great as 27 m, water depths in Norton Sound range from 10 to 20 m. The offshore part of the Yukon Delta is a zone of extensive shoals covering about 8,000 km^ (Fig. 17-2). Water depths 10-30 km offshore do not exceed 3 m; beyond that zone there is a gentle break in slope and the depth increases to 10 m as far as 50-70 km from shore. The substrate of the Yukon prodelta, derived from the Yukon River, consists of 279 > c CO c O CO CO > Si M o w Vh -^ !^ ca c o CA s CQ U ai Oi •o c/3 X! C cd , , CO U « >> x: a< o 0) )aO C o -w 3 n 00 01 t— di m i-H ■Sr 73 !-; J= CO O -(^ OS CO 1-1 J2 CD t~ •n OS c T-H CO nr a c fO 3-1 B 3 X CO Cl) a* T3 C/J C rin HH C CQ G CO +J ,ii! CO S ac- 0 05 >. K 1—1 jn ■^-^ T5 cy3 CO r! :co XI CO C 'a r x; c CO 0 <; c 0) -c o 2 CO CO '^ J2 3 CJ 0 0 G OJ 3 b CT> CJ T— 1 s CO >. 0 S ^ ■5 CO c OJ 0 CO CO J= X! 3 CO C CO QJ CO OJ CO ^ C/J .^^ 0 )3£ J3 S 3 0 05 G 9-4 r- I-H cu Ii XI J3 05 CO 0 3 05 G 0 T— 1 ^C U2 ' ■ ■0 >> 0) C ^ 0 CO « •■^ CO 3 ^ K CJ •at, fO~ a Cm CD 0 9-1 05 U3 I-H c n c >, ^ w 0) h4 (i> CO tlC CO tt(l T3 4-^ e Vh ^ 0 -D n 2 c CO •Oti lf5 c s 05 1—1 evj a; r- .X4 1 \ T-l ^ CO 2 CM QI QJ) Ol IK cj_i O "2 £ ;= X O W m d -g • * <"* c« 0) T3 a, C •- C o I' ■g -S .2 « 2 g 1- <- Xi a, -C « cn rt g CO iH 5 J= O '35 vet" S^ i^ Q> o S S c "^ a (1) a> a; x: -O £ -- «H o o _* « c "3- O) O P J w CO ^ a> "" O "O C CO pH gii 2 S a. '=>^ '^ § c CO o X c:5 W a; • . -fcj T o) C 1> too O "" o c M^ 4^ .M 2S4 Ice gouging on the subarctic Bering Shelf 285 64° - -j- + - Figure 17-5. Rose diagrams representing trend and density of gouges. Division into areas I-V based on zones of similar trending gouges. Zone of shorefast ice based on evaluation of Landsat imagery (Dupre 1977, 1978; R. E. Hunter, pers. comm. 1977). Single gouge widths range from 5 to 60 m; a width of 15-25 m is most common. Gouge patterns range from straight through sinuous to sharp-angled turns (Fig. 17-4). Incision depths of gouges, as measured on the sea-floor profile of sonographs (Fig. 17-4E) and on the 200-kHz fathometer record (Fig. 17-3B), can be as much as 1 m, but most gouges range in depth from 0.25 to 0.5 m or less. These figures may be conservative because of the geometric relation between the narrow width of the gouge and the spread of the acoustic cone of the fathometer trans- ducer (Reimnitz et al. 1977). The original incision depth is impossible to determine unless the gouge is seen as the keel plows the bottom, because afterward the gouge will be filled in. Multiple gouges, or raking (Figs. 17-4F and 17-4G) are produced when multi keeled floes (such as pres- sure ridges) plow or rake the bottom sediment, creating many parallel furrows (Reimnitz et al. 1973, Reimnitz and Barnes 1974). Unlike single gouges, raking is not ubiquitous, but in the Yukon prodelta area the raking process is more prevalent than single gouging. Zones of raking are 50-100 m to several km wide. The deepest incisions caused by raking ob- served on the records are about 1 m; but raking, like 286 Geology and geophysics single gouges, usually produces incisions less than 0.25-0.5 m deep. TRENDS AND DISTRIBUTION OF GOUGES Analysis of the trend and distribution of gouges allows us to recognize five areas of gouging with similar trends (areas I-V) and two large areas almost devoid of gouges (VI and shorefast ice zone) (Fig. 17-5). Absolute direction of ice movement cannot be predicted because criteria needed to make certain distinctions, such as gouge terminations, were not seen on the sonographs. In areas I and II (Fig. 17-5), the dominant trend of gouges is distinctly subparallel to isobaths and the coastline. There is more data scatter in areas III, IV, and V, but gouges again are generally parallel to isobaths and the coastline. The greatest data scatter is seen in area V, but this may reflect the irregular bathymetry of ridge and swale topography off Port Clarence. Except for a couple of gouges off the northwestern end of St. Lawrence Island, area VI is devoid of ice gouges. Density of ice gouges is as much as 25 times higher around the Yukon Delta, where the water is 10-20 m deep, than in other areas of the northeastern Bering Sea (Table 17-1 and Fig. 17-5, areas I and II). Not coincidentally, the Yukon prodelta is the largest expanse of shallow water in the study region. Here the density of ice gouges can be as high as 75 gouges/km^ . Density of ice gouging is 60 times higher in water 10-20 m deep than in water 5-10 m or 20-39 m deep (Table 17-2). Gouging has not been seen in water shallower than 5 m or deeper than 30 m. TABLE 17-1 Gouge density by area Trackline Total number Average density Area km^ km of gouges (gouges/km^ *) TABLE 17-2 Gouge density by water depth interval I 5,500 530 1,684 3.18 II 8,000 1,005 5,080 5.05 III 9,500 1,100 917 0.83 IV 15,500 400 993 2.48 V 7,900 1,120 216 0.19 VI 50,400 766 4 0.03 Depth interval(m) km^ Trackline Total number Average density km of gouges (gouges/km^*) 0-10 16,500 480 147 0.31 10-20 24,600 2,100 8,593 4.09 20-30 32,700 1,300 143 0.11 30-40 26,000 750 0 0 40-50 12,600 450 0 0 >50 5,400 170 0 0 *Assuming that 1 km trackline of side-scan sonar is repre- sentative of 1 km^ . *Assuming that 1 km trackline of side-scan sonar is repre- sentative of 1 km^ . GEOLOGICAL SIGNIFICANCE Trend and density of gouges The interplay of geomorphology, water depth, oceanic conditions, and location of compression or of shear zones (Fig. 17-2) determines the pattern of ice gouging in the northern Bering Sea (Figs. 17-5 and 17-6). The orientation of ice gouges depends on the direction of ice drift under the influence of wind and water current. The dominant trend of ice gouges, therefore, in Norton Sound is east-west and in the Bering Sea north-south (Figs. 17-5 and 17-6). Land promontories like the Yukon Delta tend to block ice movement and to cause compression and shear zones to form. Ice ridges around the Yukon Delta formed by the collision and shearing of moving pack ice with stationary shorefast ice account for the high density of ice gouges in areas I and II (Fig. 17-5). Areas within the zone of shorefast ice, like the large area around the Yukon Delta (Fig. 17-5), are devoid of gouges. Only the edge of the shorefast ice is deformed by the pack ice, and subsequent deformation occurs continually seaward through a process of migration of the compression /shear zone through time (Dupre 1978). Areas III and IV are characterized by low density of ice gouges (Fig. 17-5). Gouging in areas III and IV is the product of ridges formed in an ice-divergence zone by intercolli- sions of pack ice. The density of ice gouges in area V is low because this area is not in a convergence zone and at most places water depth exceeds normal ice- keel depths. Area VI seems not to have any ice gouging because water depths (Fig. 17-2) exceed normal ice-keel depths (Fig. 17-5). Age of ice gouges Although no specific studies were made to deter- mine the age and longevity of gouges, the gouges Ice gouging on the subarctic Bering Shelf 287 Figure 17-6. Summary of ice gouging: density, siiorefast ice limits, and ice movements in nortiieastern Bering Sea. seem to be modern ephemeral phenomena that recur annually. West of Port Clarence and in the nearshore area of Nome, ice gouges cut through ripple and sand-wave fields that are in dynamic equilibrium with present wave or current motion (Nelson et al. 1978, Hunter and Thor 1979) (Figs. 17-4A and B). Here the juxtaposition of old gouges, highly modified by ripples or sand waves, with new gouges suggests that gouges are being formed each winter. A number of geologic processes act to destroy gouges rapidly once they have formed. The initial smoothing of ice gouges can be enhanced by: (1) the saturated, silty substrate that tends to seek a mini- mum relief equilibrium with sides of the gouge flowing or slumping toward the center of the gouge. and (2) the constant oscillatory pounding of wave motion on the sea floor that causes shear failure in the soft sediment (Henkel 1970), making gouge sides collapse toward the center. The dish-shaped profiles of most gouges (Figs. 17-4E and 4G) indicate that these processes normally occur. Repeated surveys of ice gouges in water less than 20 m deep in the Beaufort Sea have shown that gouges are frequently smoothed over completely in one season (Barnes and Reimnitz 1979). In the Bering Sea, the ice-free season is three to four months longer than in the Beaufort Sea, allowing more time for the considerably stronger open-water wave and current regimes of the Bering Sea to destroy gouges. In Norton Sound, storm waves and currents caused 288 Geology and geophysics by the advance and retreat of storm-surge water, in addition to normal tidal and geostrophic currents, resuspend and transport large quantities of surficial sediment (Cacchione and Drake 1978, Nelson and Creager 1977). Destruction of gouges is augmented by biological reworking of surficial sediment, an active process in Norton Sound (Nelson et al., volume 2). In summary, gouges tend to be either eroded or buried because they are not in equilibrium with the dynamic physical processes on the sea floor. This reinforces the hypothesis that gouges in the Bering Sea are present-day phenomena involving develop- ment of some new gouges each ice season. Ice-sediment interaction Ice acts as both an erosional and a depositional agent; it gouges, mixes, and deforms the substrate, and promotes current scour. Ice partially controls the geomorphology of the Yukon Delta (Dupre and Thompson 1979). Sediment mixing and deformation of the substrate are important processes in densely gouged areas such as the Yukon prodelta, where pressure-ridge raking can gouge 1 m into the sediment. One event of pressure-ridge raking can affect several square kilo- meters of sea floor,^ and mix or disrupt several million cubic meters of sediment. A zone of de- formed sediment in box core No. 48 (11-18 cm interval. Fig. 17-3C) may represent an ice-gouge event. The sharpness of gouge morphology is highly dependent on the type of substrate being gouged. The sediment of the Yukon prodelta is a moderately cohesive sandy silt that will hold a shape better than the coarser-grained sediment of central Norton Sound or offshore from Port Clarence (Clukey et al. 1978, Nelson and Hopkins 1972, McManus et al. 1977). The gouge shown in Fig. 17-5A and some gouges shown in Fig. 17-4A are examples of forms with sharp relief in a competent substrate. Groups shown in Fig. 17-4A are smoother in form because they cut into a cohesionless sand substrate in the Port Clarence area. Prominent broad (50-150 m wide), shallow (0.6- 0.8 m deep) depressions on the western Yukon prodelta are associated with areas of intense ice gouging and strong bottom currents (Larsen et al. 1979). Topographic disruption by ice gouges in these areas apparently causes flow separation in the strong currents, thereby initiating scour depression for ^ Area of gouging times depth of gouging. Ex.— 2,000 m (length of gouged zone) X 1,000 m (width of gouged zone) X 0.5 m (depth of gouge) = 1,000,000 m^ . extensive distances downstream. Consequently, large regions of scour may continue to expand away from intensely gouged areas (Fig. 17-4H). The extensive depositional sand shoals of the Yukon Delta front coincide with the seaward extent of shorefast ice, stamukhi (grounded pressure ridges), and zones of dense ice gouging (Figs. 17-2 and 17-6). Reimnitz and Barnes (1974) have noted this relation in the Colville Delta area of the Beaufort Sea. They postulate that pressure ridges and stamukhi act as sediment traps or dams, channelize winter currents, or bulldoze sediment to form shoals. Thus, a cycle is formed in the sense that shoal areas determine the extent of shorefast ice and the location of a shear zone and pressure ridges, which in turn cause shoals to develop. Dupr6 (1978) hypothesizes that the geomorphology of both onshore and offshore parts of the Yukon Delta is controlled by ice. RESOURCE DEVELOPMENT: POTENTIAL HAZARDS To summarize, gouges are ubiquitous throughout the northeastern Bering Sea in water depths of 5-30 m. Ice-gouge density varies from rare to sparse in the northeastern Bering sea and northern Norton Sound; the maximum density is around the Yukon Delta (Fig. 17-6). The depth of ice gouges is fairly uniform throughout the northeastern Bering Sea and seems to be independent of gouge density. Although the maximum observed ice-gouge depth is about 1 m and maximum observed current scour about 1 m, the combination of these forces could affect the bottom to depths of several meters, thus presenting some design problems and potential hazards to installations in or on the sea floor. Pipelines and cables should be buried below the combined effective depth of ice gouging and current scour, plus a safety factor. Special studies of nearshore areas off Nome and Port Clarence were conducted because both are potential centers for commercial development and activity. Nome, already a well-established small city, is the focal point for barge traffic in the north- ern Bering Sea. Port Clarence, the only naturad harbor in the northern Bering Sea, has high potential for development as a site for future shipping activity. Offshore Nome, being an area of ice divergence, is not heavily gouged. Although some gouges were found offshore, none were in water shallower than 8 m, and several are probably not related to ice. They are narrower (less than 1 m) than typical ice gouges (more than 5 m wide) and may have been produced by anchor, anchor chain, or cable drag from the tugs and barges that frequent the port of Nome. Ice gouging on the subarctic Bering Shelf 289 Several gouges were found near Port Clarence at the northern end of the Port Clarence spit and on the northern side of Port Clarence inside the tidal inlet, but none occurred in water less than 8 m deep. ACKNOWLEDGMENTS We thank William Dupre, University of Houston, for data concerning pack ice movement and shorefast ice limits; Ralph Hunter, U.S. Geological Survey, for data on shorefast ice limits; and David Drake, U.S. Geological Survey, for data on ice thickness. Jim Evans and Ron Williams compiled data on gouges from sonographs. Valuable discussions on ice proces- ses and interpretation of sonographs were held with Peter Barnes, Erk Reimnitz, and Larry Toimil, U.S. Geological Survey. Marybeth Gerin helped with figure layout and drafting. Erk Reimnitz and Harry Cook, U.S. Geological Survey, made helpful com- ments on the manuscript. The officers, crew, and technical staff of the R/V Sea Sounder made data collection a successful and enjoyable endeavor. The cruises were supported jointly by the U.S. Geological Survey and the Bureau of Land Manage- ment through interagency agreement with the Na- tional Oceanic and Atmospheric Administration, under which a multiyear program responding to the needs of petroleum development of the Alaska continental shelf is managed by the Outer Continen- tal Shelf Environmental Assessment Program (OCSEAP) Office. REFERENCES Arctic Research Laboratory 1973 Ice character in Bering and Chukchi Seas. Naval Oceanic Systems Center, Dep. of Navy, San Diego, Calif. Barnes, P. W., D. McDowell, and E. Reimnitz 1978 Ice gouging characteristics: Their changing patterns from 1975-1977, Beaufort Sea, Alaska. U.S.G.S. Open File Rep. 78-730. Barnes, P. W., and E. Reimnitz 1979 Ice gouge obliteration and sediment redistribution event; 1977-1978, Beaufort Sea, Alaska. U.S.G.S. Open File Rep. 78-848. Belderson, R. H., N. H. Kenyon, A. H. Stride, and A. R. Stubbs 1972 Sonographs of the sea floor. Elsevier Pub. Co., New York. Brower, W. A., H. F. Diaz, A. S. Prechtel, H. W. Searby, and J. L. Wise 1977 Climatic atlas of the outer continental shelf waters — coastal region of Alaska: 2, Bering Sea, AEIDC, Anchorage, Alaska. Cacchione, D. A., and D. E. Drake 1978 Sediment transport in Norton Sound, Northern Bering Sea. Environmental Assessment of the Alaskan Continen- tal Shelf, Ann. Rep. Principal Investi- gators for the year ending March 1978, Environ. Res. Lab., Boulder, Colo., NOAA, U.S. Dep. of Commerce 12: 308-450. Clukey, E. C, H. Nelson, and J. E. Newby 1978 Geotechnical properties of northern Bering Sea sediment. U.S.G.S. Open File Rep. 78-408. Coachman, L. K., K. Aagaard, and R. B. Tripp 1975 Bering Strait: The regional physical oceanography. Univ. of Washington Press, Seattle. Duprfe, W. R. 1977 1978 Yukon Delta coastal processes study. Environmental Assessment of the Alaskan Continental Shelf, Ann. Rep. of Principal Investigators for the year ending March 1977, Environ. Res. Lab., Boulder, Colo., NOAA, U.S. Dep. of Commerce 14: 508-53. Yukon Delta coastal processes study. Environmental Assessment of the Alaskan Continental Shelf, Ann. Rep. of Principal Investigators for the year ending March 1978, Environ. Res. Lab., Boulder, Colo., NOAA, U.S. Dep. of Commerce 11: 384-446. Dupre, W. R., and R. Thompson 1979 The Yukon Delta: A model for deltaic sedimentation in an ice domi- nated environment. Proc. Offshore Tech. Conf. 1: 657-64. 290 Geology and geophysics Fathauer, T. F. 1975 The great Bering Sea storms of 9-19 November, 1974. Weatherwise Mag., Amer. Meteorological Soc. 28: 76-83. Fleming, R. H., and D. Heggarty 1966 Oceanography of the southeastern Chukchi Sea. In: Environment of Cape Thompson region Alaska, M. H. Willimovsky and J. M. Wolfe, eds., 697-754. Washington, D.C., U.S. Atomic Energy Commission. Flemming, B. W. 1976 Side-scan sonar: A practical guide. Side Scan Sonar, A comprehensive presentation: EG&G Environmental Equipment Division, Waltham, Mass., A-l-A-45. Goodman, J. R., J. H. Lincoln, T. G. Thompson, and F. A. Zeusler 1942 Physical and chemical investigations: Bering Sea, Bering Strait, Chukchi Sea during the summers of 1937 and 1938. Univ. of Washington Pub. in Oceanography, 3 (2): 105-169 and appendix 1-117. Henkel, D. J. 1970 The role of waves in causing sub- marine landslides. Geotechnique 10: 75-80. Hunter, R. E., and D. R. Thor 1979 Depositional and erosional features of the northeastern Bering Sea inner shelf (abs.). Amsterdam, Inter. Assoc. Sedimentologists, Program and Ab- stracts, Eleventh Inter. Cong, in Sedimentology (in press). Husby, D. M. 1969 Report of oceanographic cruise U.S.C.G.C. NORTHWIND, northern Bering Sea-Bering Strait-Chukchi Sea, July 1969. U.S. Coast Guard Oceano- graphic Rep. 24. 1971 Oceanographic investigations in the northern Bering Sea and Bering Strait, June-July 1969. U.S. Coast Guard Oceanographic Rep. 49. Larsen, M. C, C. H. Nelson, and D. R. Thor 1979 Geologic implications and potential hazards of scour depressions on Bering shelf, Alaska. Environmental Geology 3: 39-47. McManus, D. A., V. Kolla, D. M. Hopkins, and C. H. Nelson 1977 Distribution of bottom sediments on the continental shelf, northern Bering Sea. U.S.G.S. Prof. Paper 759-C, C1-C31. McManus, D. A., and C. S. Smyth 1970 Turbid bottom water on the conti- nental shelf of northern Bering Sea. J. Sedimentary Petrology 40: 869-77. Muench, R. D., and K. Ahlnas 1976 Ice movement and distribution in the Bering Sea from March to June 1974. J. Geophys. Res. 81 (24): 4467-76. National Oceanic and Atmospheric Administration 1974 Local climatological data—Ann. sum- mary with comparative data for Nome, Unalakleet, Shismaref, and Wales, Alaska. Nelson, C. H., and J. Creager 1977 Displacement of Yukon-derived sedi- ment from Bering Sea to Chukchi Sea during Holocene time. Geology 5: 141-6. Nelson, C. H., M. E. Field, D. A. Cacchione, and D. E. Drake 1978 Areas of active large-scale sand wave and ripple fields with scour potential on the Norton Basin sea floor. Envi- ronmental Assessment of the Alaskan Continental Shelf, Ann. Rep. of Principal Investigators for the year ending March 1978, Environ. Res. Lab., Boulder, Colo. NOAA, U.S. Dep. of Commerce 12: 291-307. Nelson, C. H., and D. M. Hopkins 1972 Sedimentary processes and distribu- tion of particulate gold in the north- ern Bering Sea. U.S.G.S. Prof. Paper 689. Ice gouging on the subarctic Bering Shelf 291 Pratt, R., and F. Walton 1974 Bathy metric map of the Bering shelf. Geol. Soc. Amer. , MC-7. Reed, J. C, and J. E. Sater, eds. 1974 The coast and shelf of the Beaufort Sea: Arctic Inst. N. Amer., Arlington, Virginia. Reimnitz, E., and P. W. Barnes 1974 Sea ice as a geologic agent on the Beaufort Sea shelf of Alaska. In: The coast and shelf of the Beaufort Sea, J. C. Reed and J. E. Sater, eds., 305-351. Arctic Inst. N. Amer., Arlington, Virginia. Reimnitz, E., P. W. Barnes, and T. R. Alpha 1973 Bottom features and processes related to drifting ice: U.S.G.S. Misc. Field Studies Map MF -5 32. Reimnitz, E., P. W. Barnes, L. J. Toimil, and J. Melchior 1977 Ice gouge recurrence and rates of sediment reworking, Beaufort Sea, Alaska. Geology 5: 405-8. Shapiro, L. H., and J. J. Burns 1975 Satellite observations of sea ice movement in the Bering Strait region. Climate of the Arctic, Rep., Univ. of Alaska, Fairbanks, 379-86. Stringer, W. Thomson 1977 J., S. A. Barrett, N. Blavin, and D. Morphology of Beaufort, Chukchi, and Bering Seas nearshore ice condi- tions by means of satellite and aerial remote sensing. Environmental Assessment of the Alaskan Continen- tal Shelf, Ann. Rep. of Principal Investigators for the year ending March 1977, Environ. Res. Lab., Boulder, Colo., NOAA, U.S. Dep. of Commerce 15: 42-150. Thor,D. R., and C. H. Nelson 1978 Continuous seismic reflection profile records, SEA 5-77-BS Cruise northern Bering Sea. U.S.G.S. Open File Rep. 78-608. The Role of the Kaltag and Kobuk Faults in the Tectonic Evolution of the Bering Strait Region Mark L. Holmes' and Joe S. Creager^ * U.S. Geological Survey, Seattle, Washington ^ Department of Oceanography University of Washington Seattle, Washington ABSTRACT The generally latitudinal trend of the major basins and mountain belts in northwestern Alaska and northeastern Si- beria places severe constraints on palinspastic reconstructions of the tectonic features of this region. The essential continu- ity of these linear trends since Precambrian time precludes the possibility that there has been any significant north-south relative movement between Alaska and Siberia. The geological similarity between the Seward and Chu- kotsk Peninsulas suggests that they probably acted as a single unit, bounded on the south by the Kaltag fault and on the north by the Kobuk fault. This tectonic block moved east- ward more than 100 km during the Tertiary, partly as a result of sea-floor spreading in the North Atlantic and Arctic Oceans. An important assumption in this model is that the Kobuk Trench represents a major left -lateral transcurrent fault. INTRODUCTION In any attempt at palinspastic reconstruction or mobUistic modeling of the major tectonic features of northwestern Alaska, serious consideration must be given to the long-term similarities in the geologic his- tories of northern Alaska and northeastern Siberia. The generally latitudinal distribution of the major mountain ranges and geosynclinal basins has led sev- eral writers (Gates and Gryc 1963; Smirnov 1968; Churkin 1969, 1970, 1972, 1973; Lathram 1973; and Meyerhoff 1973) to point out that these features have had an essential continuity since Precambrian time, and that the close relationship of adjacent ele- ments precludes the possibility that there have been any major relative movements between Alaska and Siberia. This view holds that the 1,350-km-wide Ber- ing-Chukchi Shelf is an integral part of both North America and Eurasia which has accreted between the Canadian and Siberian shields since Proterozoic time. The continuity of many of the major geologic fea- tures certainly does seem to preclude the possibility that there has been any major movement at right angles to their roughly east-west trends. However, there is abundant evidence of significant amounts of late Mesozoic and Tertiary thrusting and transcurrent faulting that imply differential east-west tectonic transport; this type of tectonic activity would main- tain an apparent continuity in the older geologic trends. GEOLOGICAL SETTING Figs. 18-1 and 18-2 show the geography and gen- eralized geological and structural elements of the Arc- tic area of western Alaska and eastern Siberia, respec- tively. Rocks ranging in age from Precambrian to Pleistocene occur in the several mountain belts and extensive coastal plain zones of both Alaska and northeastern Siberia. Two sedimentary basins occupy the continental shelf areas north and south of Bering Strait. These large and complex structural troughs contain rock sequences possibly as old as Late Creta- ceous. Quaternary nonmarine sediments forming the low- land 75 km wide on the northern side of the Chu- kotsk Peninsula rest on Paleozoic sedimentary, meta- morphic, and igneous rocks and early Tertiary intru- sives (Nalivkin 1960, Suslov 1961). These rocks also form the main spine of the Chukotsk Range. Moun- tains of the Tenilny Range consist of Precambrian and Paleozoic sedimentary and metamorphic rocks, with some Mesozoic intrusive and Cenozoic extrusive rocks (Nalivkin 1960, Markov and Tkachenko 1961). 293 294 Geology and geophysics 170 180 170 -? 7 7 7 7 7 7 7 7 1 7 — 7 1 ? 1 } 1 1 — I 1 TT i — ( 1 T ■^WRANGEL ISLAND CHUKCHI SEA NUNIVAK ISLAND ^^--^ fill VJ 74 160 Figure 18-1. Geographic features of northwestern Alaska and northeastern Siberia. The geology of Seward Peninsula is strikingly simi- lar to that of the Chukotsk Peninsula. Paleozoic and some Precambrian sedimentary and metamorphic rocks underlie most of the peninsula, with extensive occurrences of Tertiary extrusives near Cape Espen- berg and in the eastern part of the peninsula. The rocks of Seward Peninsula are cut by numerous zones of thrust faults, along which the movement has been eastward (Sainsbury 1969). From Cape Krusenstem northwest to Cape Lis- burne, the coastline consists for the most part of cliffs formed by rocks of the Brooks Range, the De Long Mountains, and the Lisburne Hills. The Brooks Range is formed by a belt of predominantly Paleozoic rocks, and Early Cretaceous orogenic rocks have been thrust northward, forming tectonic sheets of Jurassic mafic and ultramafic rocks and Devonian and Missis- sippian carbonate rocks unlike the Devonian schists and Mississippian rocks which they override. North of this trend is a narrow zone of chaotically deformed lower Mesozoic and upper Paleozoic rocks that has been called the Disturbed Belt (Tailleur and Brosge 1970, Brosg6 and Tailleur 1970). The southern margin of the Brooks Range (Fig. 18- 2) is formed by the Kobuk Trench (Grantz 1966). South of this feature, the rocks of the Yukon-Koyu- kuk basin consist of strongly deformed Permian to Jurassic volcanic and Cretaceous sedimentary and vol- canic rocks (Payne 1955, Patton 1973). The southern part of the Yukon-Koyukuk basin is cut by the Kal- tag fault (Fig. 18-2), a major northeast-trending trans- current fault showing right-lateral displacement (Pat- ton and Hoare 1968). The Lisburne Hills, extending from Cape Lisburne on the northwest to Cape Thompson in the southeast, show structures similar to those of the Brooks Range; here the Paleozoic rocks have been thrust eastward in imbricate thrust sheets over the Cretaceous rocks of the Colville Geosyncline (Campbell 1967, Martin 1970). The marine geology of Norton basin (Fig. 18-2) has been discussed by Moore (1964), SchoU and Hopkins (1969), Grim and McManus (1970), Tagg and Greene (1973), Nelson et al. (1974), Holmes and Cline (1978), Holmes and Fisher (1979), and Fisher et al. (1979). The basin consists of a trough 5 km deep (Holmes and Fisher 1979) that may be filled with Upper Cre- taceous and Paleogene nonmarine rocks and Neogene marine deposits. The basin is underlain by an acous- tic basement characterized by velocities of 4.5 to 6.9 km/sec, probably consisting of metamorphosed Paleo- zoic and upper Mesozoic rocks which have been lo- cally intruded by upper Mesozoic plutons similar to those occurring on St. Lawrence Island and which al- so form the several islands in northern Norton basin (Holmes and Fisher 1979). North of Bering Strait another sedimentary basin underlies much of the southern Chukchi Sea (Fig. 18- 2). Hope basin (Grantz et al. 1970, 1975; Eittreim et al. 1978), is similar in many respects to Norton basin. It contains approximately 3-3.5 km of sedimentary fill consisting of possible Upper Cretaceous and Paleo- gene coal-bearing nonmarine deposits and Neogene marine sediments. The acoustic basement beneath this sedimentary basin is characterized by compressional velocities ranging from 4.3 to 5.2 km/sec (Grantz et al. 1975), indicating that basement may consist of rocks similar to the Mesozoic and Paleozoic carbonate and clastic rocks and Precambrian metamorphic rocks which have been mapped onshore around the margins of the basin . The northeastern and northern margin of the basin is formed by Herald and Wrangel arches (Grantz et al. 1975), two uplifted areas where the Mesozoic and Pa- leozoic basement rocks have been thrust northward and northeastward over younger Cretaceous rocks of the ColvUle Geosyncline. These arches form a struc- tural and stratigraphic continuation of the Lisburne Hills upHft. Grantz et al. (1975) have shown that the Herald Fault Zone along the northeast side of Herald Arch is an offshore extension of the thrust front along the Lisburne Hills (Martin 1970). Northwest of the Herald Fault Zone, Grantz et al. (1970, 1975) were able to trace the westward off- shore extension of the Colville Geosyncline structures 170 180 7 7 7 7 7 7 170 160 150 7 7 7 7 7 — 7 7 1 1 1 — I 1 — 1 1 — 1 — 140 T — \ — r - 74 - 70 FAULTS "7"^ TRANSCURRENT I I I I NORMAL T T f THRUST - 62 170 160 150 mm CENOZOIC NON-MARINE DEPOSITS CENOZOIC EXTRUSIVE VOLCANICS INT EXT lESOZOIC VOLCANICS rMH7vNi ESOZOIC GEOSYNCLINAL DEPOSITS PALEOZOIC GEOSYNCLINAL DEPOSITS PRECAMBRIAN ROCKS Figure 18-2. General geological and structural elements of northwestern Alaska and northeastern Siberia. Compiled from Nalivkin (1960), King (1969), Sainsbury (1972), Patton (1973), Beikman (1978). 295 296 Geology and geophysics and show that they appear to be truncated by the faults and folds along the front of the Herald Fault Zone. THE KOBUK FAULT The Kobuk Trench, or Trough, has been described by Grantz (1966) and Patton (1973) as a possible strike-slip fault which extends eastward for over 480 km across northern Alaska to the point where it inter- cepts the northeast-trending Kaltag fault (Fig. 18-2). The southward deflection of major structural trends on the north side of the Kobuk Trench was thought by Patton (1973) to represent large-scale drag folding due to major left -lateral movement. ERTS imagery studies have revealed a series of cross faults which could be tensional features indicating left-lateral dis- placement along the Kobuk strike-slip fault (Lathram 1972, 1973; Baker 1974). In the valley of the Kobuk River the deformational zone is approximately 20 km wide. The age of faulting is thought to be pre-middle Tertiary, but some local rupturing of Pleistocene drift (Patton 1973) indicates that movement has probably continued through Tertiary and into Pleistocene time. Grantz et al. (1975), Holmes et al. (1968a), Holmes (1975), and Eittreim et al. (1978) show a prominent west-trending ridge formed by the acoustic basement surface in the southeastern part of Hope basin (Fig. 18-2). The ridge appears to be approximately 150- 200 km long and 20-22 km wide with a rehef of 250- 275 m. Several faults can be seen cutting the basin fill along the flanks of the ridge (Grantz et al. 1975) and the upper Neogene (?) unit in the basin has been uplifted and eroded over the crest of the ridge (Holmes 1975), forming a pronounced angular uncon- formity with the base of the Holocene section (Holmes et al. 1968b). The basement uplift and associated west-trending structural features are situated in such a way as to appear to be the offshore extension of the Kobuk Fault Zone. Marine gravity measurements (Ruppel and McHendrie 1976) also indicate a correlation be- tween the Kobuk Trench and this offshore feature. THE KALTAG FAULT The Kaltag Fault (Fig. 18-2) is another major trans- current fault which trends in a northeastern direction for over 440 km across north-central Alaska from Norton Sound. SchoU et al. (1970) have traced the offshore continuation of the Kaltag fault for over 320 km to the southwest beneath St. Matthew and Hall basins. Patton and Hoare (1968) mapped several pro- vince boundaries and geologic trends which have been displaced up to 130 km right-laterally since Creta- ceous time; bending of the structural grain near the fault was ascribed by Patton (1973) to large-scale drag folding. Marine seismic refraction data suggest that the old- est sedimentary unit in Norton basin may consist of Upper Cretaceous rocks similar to those in the Yuk)n- Koyukuk basin (Holmes and Fisher 1979). If this interpretation is correct, it could indicate that move- ment along the Kaltag fault began in late Cretaceous time. Deformation of sedimentary units in St. Mat- thew basin (Fig. 18-2) suggests movement along the fault at least by early-middle Tertiary time (Scholl et al. 1970). Grantz (1966) and Patton and Hoare (1968) describe offset drainage patterns and displaced strata along the onshore segment of the fault, indica- ting right-lateral movement along the fault of as much as 2.4 km during the Holocene. MODEL FOR MAJOR HORIZONTAL DISLOCATIONS A possible evolutionary model may explain the present configuration and relationship of the major tectonic features observed in the Bering Strait region, including western Alaska and eastern Siberia. Ele- ments of the late Mesozoic and Cenozoic geologic his- tory are presented in outline form to weave together a coordinated account of a complex sequence of events which relates the major horizontal movements (with the exception of the longitudinal faulting of the Brooks Range and De Long Mountains) to a relative eastward movement of Siberia toward Alaska. We do not intend to imply that a crustal plate boundary ex- ists between these two continental areas, even though others have proposed such a feature (Le Pichon 1968, Hamilton 1969). Two basic assumptions are made: (1) Seward Pen- insula represents a direct extension of the geologic trends of the Chukotsk Peninsula rather than of the Alaskan Yukon-Koyukuk basin (Sachs and Strelkov 1961, Sainsbury 1972); (2) the Kobuk Trench does, in fact, represent a major left-lateral transcurrent fault (Lathram 1973, Baker 1974), which has been active at least since early Tertiary time (Patton 1973). A corollary assumption is that the east-west ridge/ trough structure in southeastern Hope basin (Fig. 18- 2) is the offshore extension of the Kobuk fault. The central feature of the model is the late Creta- ceous development of a single continuous arcuate belt of deformation and thrusting extending from eastern Seward Peninsula northwest across the southern Chukchi Sea to Wrangel Island. The contemporaneity Kaltag and Kobuk faults 297 of various segments of this arc has already been pointed out by Grantz et al. (1975), Holmes (1975), and Patton and Tailleur (1977). Herald and Wrangel arches and the Lisburne Hills (Fig. 18-2) comprise a fault and fold belt which is younger than the Brooks Range /De Long Mountains; these younger structures are superimposed on and truncate the older ones (Grantz et al. 1975). According to the model, this zone represents a juncture between fold systems of differing ages, the major transport having taken place along the Herald Fault Zone. The sequence of steps by which this model ac- counts for the present-day structural relationships is as follows (Fig. 18-3 through 18-6): (1) In Late Cretaceous time, northward travelling overthrusts culminating in the Brooks Range and De Long Mountains resulted in aggregate longi- tudinal shortening of at least 130 km. Gravity sliding from a southerly uplifted area (Martin 1970), or backthrusting of the Colville Geosyncline beneath the geanticline (Tailleur 1973) are possible causative mechanisms. Sachs and Strelkov (1961) show that the Pacific and Arctic basins were connected by a seaway across eastern Seward Peninsula at this time (Fig. 18-3). Figure 18-3. Late Cretaceous tectonic framework of tiie Bering Strait region. (2) Easterly directed compressional deformation beginning in latest Cretaceous time produced east- ward and northeastward overthrusts along an arcu- ate front extending from eastern Seward Peninsula to Wrangel Island (Martin 1970, Sainsbury 1972, Patton 1973). Minimum aggregate shortening in the Lisburne Hills uplift was 18 km (Martin 1970). During this phase, the transcurrent faults devel- oped along the eastern and western ends of the De Long Mountains (Martin 1970) in response to the eastward compression, causing the De Long Moun- tains to assume an arcuate northward bulge (Fig. 18-4). Closure of the seaway between Seward Pen- insula and mainland Alaska occurred along the thrust belt (Sachs and Strelkov 1961) at this time. SIBERIA Figure 18-4. Latest Cretaceous tectonic framework of the Bering Strait region. (3) In late Cretaceous and early Tertiary time (Fig. 18-5) thrusting ceased, and a trans-current faulting phase began. Strike-slip movement occurred on the Kobuk and Kaltag faults, possibly localized along pre-existing faults or zones of weakness, and the wedge-like region between these faults (Seward Peninsula and Yukon-Koyukuk basin) moved rela- tively eastward into north-central Alaska. The thrust front extending from Seward Peninsula to Wrangel Island was offset along the offshore exten- sion of the Kobuk fault (Fig. 18-5). The oroclinal pair in the Richardson Mountains far to the east may have been formed at this time as a result of Figure 18-5. Late Cretaceous/early Tertiary tectonic frame- work of the Bering Strait region. 298 Geology and geophysics right-lateral movement along the porcupine linea- ment and Tintina fault (Tailleur 1973). Total dis- placement along the Kaltag fault is approximately 130 km (Patton and Hoare 1968)— enough to ac- count easily for the projected left-lateral offset of the thrust belt north of Seward Peninsula, if we as- sume a compensating right-lateral offset of similar magnitude along the Kobuk fault. (4) Subsidence of Norton and Hope basins (Fig. 18-6) probably began in earliest Tertiary time (or z3 x^ -^--^-C:^ \ '% c:^M^ r*>y [ ^ \ \ X^ X / SIBERIA W V •^ ALASKA 0 200 oV \ n n ) K m Figure 18-6. Early -middle Tertiary tectonic framework of the Bering Strait region. shortly before), during the late stages of the trans- current faulting phase. The subsidence could have begun after the cessation of the major phase of thrusting evidenced in the Lisburne Hills or Herald Arch; the basins could also be a negative tectonic element resulting directly from the compression. Early Tertiary thrust faulting along the western tip of Seward Peninsula (Sainsbury 1972) created a fractured zone along which the new seaway, Bering Strait, would be established in late Miocene time. (5) Vertical and horizontal adjustments have recur- rently affected the region since early Tertiary time. Hope and Norton basins continued to subside, fin- ally accumulating from 3 to 5 km of Paleogene and Neogene sediments in their deepest parts (Grantz et al. 1975, Holmes and Fisher 1979), and some minor Paleogene thrusting took place in eastern Seward Peninsula (Sainsbury 1972). Mild Quater- nary uplift occurred in areas of northern Al- aska and northeastern Siberia (Stovas 1965). These vertical movements around the margins of Hope basin probably contributed to the formation of some of the steep normal faults which cut the Ter- tiary basin fill. The presence of the angular uncon- formity on top of the Neogene (?) unit over the crest of Kotzebue anticline in southeastern Hope basin indicates that this ridge also experienced rela- tive uplift and erosion in Quaternary time, and some right-lateral movement continued along the Kaltag fault during the Holocene (Grantz 1966, Patton and Hoare 1968). This discussion has interesting implications for the- ories of Arctic Basin evolution which require a leirge suture or fault to extend southward from the Canada basin continental margin across the Chukchi and Ber- ing Sea shelves. Freeland and Dietz (1973) have pro- posed such a model, stating that evidence for such a fault is provided by the prominent north-south belt of conspicuous magnetic anomalies (Bassinger 1968) recorded in eastern Herald basin. But Grantz et al. (1975) have traced the Albian and younger Mesozoic beds of the Colville Geosyn- cline and Herald basin and the Late Cretaceous to ear- ly Tertiary deformational zone of folds and faults of the eastern and western structural provinces across this magnetic anomaly belt— evidence which does not support the hypothesized fault zone. The magnetic anomalies first mapped by Bassinger (1968) evidently are related to pre-Cretaceous intrusive bodies similar to those found at many other locations in Alaska and Siberia. The hypothesis of Freeland and Dietz (1973) would also require that slip along the Kobuk fault be right- lateral. Although evidence to the contrary is con- vincing (Lathram 1972, Baker 1974), the possibility exists that this broad and complex feature has experi- enced two episodes of movement in opposite direc- tions. This is only one (Freeland and Dietz 1973) of a number of hypothetical reconstructions proposed in papers about postulated relative motions of Ameri- can, Pacific, and Eurasian plates which have been written since the general acceptance of the paradigm of "new global tectonics" (Isacks et al. 1968). Not all the authors of these models seem to have recog- nized the fact that sea-floor spreading along mid- ocean ridges has global rather than local and regional effects. The various Atlantic, Arctic, and Pacific spreading models produce an interesting, if somewhat confusing, array of possible crustal movements in the area of northern Alaska and northeastern Siberia because of its position with respect to the proposed major plate boundaries (Fig. 18-7). Most of these proposed tectonic movement models are mutually exclusive; many others are completely conjectural and were proposed only in order to solve a problem in some other remote ocean basin. Kaltag and Kobuk faults 299 120 130 EXPLANATION Sea - Floor Spreading Motion Young Subduction Zone AAA Older Subduction Zone T c> Transform Fault a Megasheor 140 180 170 160 150 ^ Drifting Motion = =^ Rifting Motion I ^ ^"^ Oroclinal Bending I ^--^ — Syntaxial Bending ■^~~' ^ Thrust Fault tzz, Strike - Slip Fault Figure 18-7. Compilation of proposed crustal movements affecting Alaska and northeastern Siberia (after Lathram 1973). SUMMARY The long-term similarity in the geologic histories of northern Alaska and northeastern Siberia and the close relationship between adjacent elements pre- cludes the possibility that any large relative move- ments have occurred, especially in a north-south di- rection, between these two land masses. Within the restrictions imposed by this observation and the es- tablished chronologies and character of the major tec- tonic features of this region, a tectonic model can be formulated which accounts for the observed displace- ments and deformational styles. Two important assumptions have been corrobor- ated by field observations: the Seward Peninsula is more closely related to the Chukchi Peninsula than to adjacent parts of Alaska, and the Kobuk Trench rep- resents a major transcurrent fault along which the most recent sense of movement has been left-lateral. The model requires eastward movement of the Siber- ian block toward Alaska, and involves a net transport of approximately 100-150 km. After the Late Cretaceous development of north- ward-directed thrust and slide sheets in the De Long Mountains and Brooks Range, strong eastward-direc- ted compressive forces resulted in the formation of a long arcuate thrust belt extending from eastern Se- ward Peninsula across the Lisburne Hills and Lis- bume-Wrangel Arch to Wrangel Island. A long-estab- lished seaway connecting the Arctic and Pacific basins across eastern Seward Peninsula was closed by this thrusting and uplift, which involved a minimum east- ward displacement of approximately 20 km. Trans- current faulting occurred at the eastern and western ends of the De Long Mountains in response to this compression. Continued compression in Late Cretaceous to early Tertiary time resulted in large offsets of approximate- ly 130 km along the Kaltag (right -lateral) and Kobuk (left-lateral) faults, and northeastward bowing and 300 Geology and geophysics thrusting of the Lisburne-Wrangel Arch. Subsidence of Hope basin and Norton basin was initiated in the Late Cretaceous or earliest Tertiary. The eastward and northeastward movements in the Lisburne Hills and along the Lisburne-Wrangel Arch produced along the front of the thrust zone a belt of folds and faults (western structural province), which truncated similar features established as a result of earlier northward movements in the De Long Mountains and Brooks Range. Subsequent horizontal and vertical movements since the early Tertiary have resulted in continued subsidence of Hope basin in the southern Chukchi Sea and Norton basin in the northern Bering Sea, voth minor transcurrent motion along the Kobuk and Kaltag faults. ACKNOWLEDGMENT Chapman, R. M., and E. G. Sable 1960 Geology of the Utukok-Corwin region northwestern Alaska. U.S.G.S. Prof. Paper 303-C. Churkin, M., Jr. 1969 Paleozoic tectonic history of the Arc- tic basin north of Alaska. Science 165: 549-55. 1970 Fold belts of Alaska and Siberia and drift between North America and Asia. In: Proceedings of the geologi- cal seminar on the north side of Alas- ka, W.L. Adkison and M.M. Brosge, eds., G1-G15. Amer. Assoc. Petro- leum Geologists, Los Angeles, Calif. This study is Contribution No. 1152, Department of Oceanography, University of Washington. The re- search was supported by National Science Founda- tion Grants GA-808, GA-11126, and GA-28002. REFERENCES Baker, R. N. 1974 ERTS updates geology. Geo times 19: 20-22. Bassinger, B. G. 1968 Marine magnetic study in the north- east Chukchi Sea. J. Geophys. Res. 73: 683-7. Beikman, H. 1978 Brosge, W. P., 1970 Preliminary geologic map of Alaska. U.S.G.S., scale 1 :2,500,000. and L L. Tailleur Depositionad history of northern Alas- ka. In: Proceedings of the geological seminar on the north slope of Alaska, W.L. Adkison and M.M. Brosge, eds., D1-D18. Amer. Assoc. Petroleum Geologists, Los Angeles, Calif. Campbell, R. H. 1967 Areal geology in the vicinity of the Ohariot site, Lisburne peninsula, northwestern Alaska. U.S.G.S. Prof. Paper 395. 1972 Western boundary of the North Amer- ican continental plate in Asia. Geol. Soc. Amer. BuU. 83: 1027-36. 1973 Geologic concepts of Arctic Ocean ba- sin. In: Arctic geology, M.G. Pitcher, ed., 485-99. Amer. Assoc. Petroleum Geologists, Tulsa, Okla. Eittreim, S., A. Grantz, and O. T. Whitney 1978 Isopach maps of Tertiary sediments, Hope Basin, southern Chukchi Sea, Alaska. U.S.G.S. Misc. Field Studies Map MF-906. Fisher, M. A., W. W. Patton, Jr., D. R. Thor, E. W. Scott, C. H. Nelson, and C. L. Wilson 1979 The Norton basin of Alaska. Oil and Gas J. 21 May, 96-8. Freeland, G. L., and R. S. Dietz 1973 Rotation history of Alaskan tectonic blocks. Tectonophysics 18: 379-89. Gates, G. O., and G. Gryc 1963 Structure and tectonic history of Alas- ka. In: Backbone of the Americas, O.E. Childs and B.W. Beebe, eds., 264- 77. Amer. Assoc. Petroleum Geolo- gists Memoir 2, Tulsa, Okla. Grantz, A. 1966 Strike-slip faults in Alaska. Open File Rep. U.S.G.S. Kaltag and Kobuk faults 301 Grantz, A., M. L. Holmes, and B. A. Kososki 1975 Geologic framework of the Alaskan continental terrace in the Chukchi and Beaufort Seas. In: Canada's contin- ental margins and offshore petroleum exploration, C.J. Yorath, E.R. Parker, and D.J. Glass, eds., 669-700. Cana- dian Soc. of Petroleum Geologists, Calgary, Can. Grantz, A., S. C. Wolf, L. Breslau, T. C. Johnson, and W. F. Hanna 1970 Reconnaissance geology of the Chuk- chi Sea as determined by acoustic and magnetic profiling. In: Proceedings of the geological seminar on the north slope of Alaska, W.L. Adkison and M.M. Brosge, eds., F1-F28. Amer. As- soc, of Petroleum Geologists, Los An- geles. Grim, M. S., and D. A. McManus 1970 A shallow seismic-profiling survey of the northern Bering Sea. Marine Geol. 8: 293-320. Hamilton, W. 1969 The Uralides and the motion of the Russian and Siberian platforms. Geol. Soc. Amer. Bull. 81: 2553-76. Holmes, M. L. 1975 Tectonic framework and geologic evo- lution of the southern Chukchi Sea continental shelf. Univ. Washington Ph.D. Dissertation. Holmes, M. L., and J. D. Cline 1978 Geological setting of the Norton basin gas seep. Proc. Offshore Tech. Conf., 73-80. Houston, Texas. Holmes, M. L., J. S. Creager, and D. A. McManus 1968a Structure and history of the Chukchi Basin. Program, Geol. Soc. Amer. Ann. Meeting, Mexico City, 141 (ab- stract). 1968b High frequency acoustic profiles on the Chukchi Sea continental shelf. Proc. Amer. Geophys. Union 49: 207 (abstract). Holmes, M. L., 1979 and M. A. Fisher Sonobuoy refraction measurements in Norton basin, northern Bering Sea. Program, Amer. Assoc. Petroleum Geologists Ann. Convention, Houston, 104 (abstract). Isacks, B., J. Oliver, and L. R. Sykes 1968 Seismology and the new global tecton- ics. J. Geophys. Res. 72: 5855-5900. King, P. B. 1969 Tectonic map of North U.S.G.S., scale 1:5,000,000. America. Lathram, E. H. 1972 Nimbus IV view of the major struc- tural features of Alaska. Science 175: 1423-7. 1973 Tectonic framework of northern and central Alaska. In: Arctic geology, M. G. Pitcher, ed., 351-60. Amer. As- soc. Petroleum Geologists Memoir 19, Tulsa, Okla. Le Pichon, X. 1968 Sea-floor spreading and continental drift. J. Geophys. Res. 73: 3661-97. Markov, F. G., and B. V. Tkachenko 1961 The Paleozoic of the Soviet Arctic. In: Geology of the Arctic, G.O. Raasch, ed., 31-47. Univ. Toronto Press, Tor- onto, Can. Martin, A. J. 1970 Meyerhoff , A. 1973 Structure and tectonic history of the western Brooks Range, De Long Mountains and Lisburne Hills, north- ern Alaska. Geol. Soc. Amer. Bull. 81: 3605-22. Origin of Arctic and north Atlantic Oceans. In: Arctic geology, M.G. Pitcher, ed., 562-82. Amer. Assoc. Petroleum Geologists Memoir 19, Tul- sa, Okla. 302 Geology and geophysics Moore, D. G. 1964 Acoustic reflection reconnaissance of continental shelves: Eastern Bering and Chukchi Seas. In: Papers in ma- rine geology, Shepard commemora- tive volume, R.L. Miller, ed., 319-62. MacMillan Co., N.Y. Nalivkin, D. V. 1960 The geology of the U.S.S.R. mon Press, N.Y. Perga- Nelson, C. H., D. M. Hopkins, and D. W. Scholl 1974 Tectonic setting and Cenozoic history of the Bering Sea. In: Marine geology and oceanography of the Arctic seas, Y. Herman, ed., 119-140. Springer- Verlag, N.Y. Patton, W. W., Jr. 1973 Reconnaissance geology of the north- ern Yukon-Koyukuk province, Alaska. U.S.G.S. Prof. Paper 774-A. Patton, W. W., Jr., and J. M. Hoare 1968 The Kaltag fault, west-central Alaska. U.S.G.S. Prof. Paper 600-D: D147- 153. Patton, W. W., Jr., and I. L. Tailleur 1977 Evidence in the Bering Strait region for differential movement between North America and Eurasia. Geol. Soc. Amer. Bull. 88: 1298-1304. Payne, T. G. 1955 Mesozoic and Cenozic tectonic ele- ments of Alaska. U.S.G.S. Misc. Geol. Investigations Map 1-84. Ruppel, B., D,, and G. McHendrie 1976 Free-air gravity anomaly map of the eastern Chukchi and southern Beau- fort seas. U.S.G.S. Misc. Field Studies Map MF-785. Sachs, V. N., and S. A. Strelkov 1961 Mesozoic and Cenozoic of the Soviet Arctic. In: Geology of the Arctic, G.O. Raasch, ed., 48-67. Univ. Toron- to Press, Toronto, Can. Sainsbury, C. L. 1969 The A.J. CoUier thrust belt of the Seward Peninsula, Alaska. Geol. Soc. Amer. Bull. 80: 2595-6. 1972 Geologic map of the Teller quadrangle, western Seward Peninsula, Alaska. U.S.G.S. Misc. Geol. Investigation Map 1-685. Scholl, D. W., and D. M. Hopkins 1969 Newly discovered Cenozoic basins, Bering Sea shelf, Alaska. Amer. Assoc. Petroleum Geologists Bull. 53: 2067- 78. SchoU, D. W., M. S. Mario w, J. S. Creager, M. L. Holmes, S. C. Wolf, and A. K. Cooper 1970 A search for the seaward extension of the Kaltag fault beneath the Bering Sea. Program, vol. 2, Amer. Geol. Soc. Cordilleran Sect. Meeting, Hay ward, Calif., 141-2 (abstract). Smimov, A. M. 1968 Role of the Precambrian basement in structural evolution of the Pacific mo- bile belt (particularly its northwestern section). In: Pacific geology: 1, M. Minato, ed., 145-65. Tsujika Shokan Publishing Co., Tokyo. Stovas, M. V. 1965 Recent uplift of the coasts of the Kara, Laptev, East Siberian, and Chu- kotka Seas. Doklady Akad. Nauk SSSR161: 28-9. Suslov, S. P. 1961 Physical geology of Asiatic Russia. W.H. Freeman and Co., San Francisco. Tagg, A. R., and H. G. Greene 1973 High resolution seismic survey of an onshore areas neair Nome, Alaska. U.S.G.S. Prof. Paper 259A:A1-123. Tailleur, I. L. 1973 Tailleur, I. L. 1970 Probable rift origin of Canada basin. In: Arctic geology, M.G. Pitcher, ed., 526-35. Amer. Assoc. Petroleum Ge- ologists Memoir 19, Tulsa, Okla. and W. P. Brosg6 Tectonic history of northern Alaska. In: Proceedings of the geological sem- inar on the north slope of Alaska. W. L. Adkison and M.M. Brosge, eds., E1-E19. Amer. Assoc. Petroleum Geologists, Los Angeles, Calif. Section I¥ Chemical Oceanography Donald W. Hood, editor Some Geochemical Characteristics of Bering Sea Sediments D. C. Burrell, K. Tommos, A. S. Naidu, and C. M. Hoskin' Institute of Marine Science University of Alaska Fairbanks, Alaska ^ Present address : Harbor Branch Foundation Fort Pierce, Florida ABSTRACT Biogeochemical data are presented for surficial Bering Sea sediments; most are for separate single collections on the southeastern shelf and Norton Sound. Sand-sized sediment predominates, gravel occurs in certain nearshore areas, and mud-sized material is usually a minor component except in Norton Sound adjacent to the Yukon River discharge. In general, the distribution of size fractionation conforms to the present physical environment as this is currently understood and relict and palimpsest sediment is of minor distribution. Southeastern shelf infauna demonstrates a reciprocal re- lationship: individual organisms are at a maximum in fine sand-sized sediment, but (wet weight) biomass increases in sediment finer and coarser than this. Illite and a glycol- expandable component are the dominant shelf clay minerals (<2 iim), together with chlorite and minor kaolin. Heavy metal contents — especially defined extractable fractions — correlate with fineness of mean grain size; hence contents are, in general, relatively reduced over these shelf areas. Increases in near-bottom particulate contents may be attributed to sediment resuspension. INTRODUCTION The work reported in this chapter relates primarily to samples collected on OCS-sponsored cruises: to the southern Bering Sea/Bristol Bay region in 1975, and to Norton Sound in September 1976. These were chemical survey cruises, and, of necessity, the sample grid spacing employed was far too coarse to permit detailed geochemical characterization of any one region. Furthermore, our primary long-term goal has been to contribute to an understanding of the dynamics both of the shelf sediments and of chemicals across the benthic interface. In this re- spect, not only are the present descriptive data embarrassingly sparse, but interpretation is limited by insufficient extant information about the physical regime. We can report only inferences about trends and variations. Nevertheless, the sample sets have been collected from little-studied regions of the Alaskan shelf, and the data should form a useful framework for more detailed investigations in the future. PREVIOUS WORK The physical oceanographic regime The general circulation of the Bering Sea was initially described by Ratmanoff (1937), who re- ported that most of the water flowing in from the north Pacific entered through passes in the central and western Aleutian Islands and exited northwards by way of the Bering Strait. Later workers have emphasized that the induced Pacific circulation is restricted to the region adjacent to the Aleutian chain. Thus, from observations in the vicinity of Unimak Pass, Barnes and Thompson (1938) described a current structure northward from the passes basi- cally defined by the shelf topography. It may be inferred from this and other early work that the north-northwest flow over the outer continental shelf tends to isolate the southeasterri shelf region from the surface waters of the southern and central Bering Sea. Bristol Bay has received more attention than any other region of the Bering Sea. Dodimead et al. (1963) summarized circulation investigations 305 306 Chemical oceanography made over the previous 30 years or so. Hebard (1959) and Takenouti and Ohtani (1974) have described a cyclonic circulation on the southeastern shelf supposedly in response to regional meteor- ological, tidal, and topographic control. Intensive recent investigations have revised these concepts somewhat. Through the ice-free season Bristol Bay waters can be basically classified into three water types: coastal water lies shoreward of an oceano- graphic front located near the 50-m isobath; shelf water occupies the region approximately between the 50-m and 100-m isobaths; Alaskan Stream/Bering Sea water intrudes into the southeastern basin region (Coachman and Charnell 1979). Kinder (1977) and Schumacher et al. (1979) have described the structural front which, coinciding approximately with the 40-50-m depth contours, separates the well-mixed coastal and two-layered central shelf domains. This feature is characterized by marked vertical property gradient changes, has a typical width of 5-10 km, and is generally located where the input of buoyant energy balances tidally generated mixing energy; the front occurs where the height of the bottom mixed layer closely ap- proximates the water depth so that tidal mixing occurs throughout the (almost vertically homoge- neous) column. Recent research (especially by the PROBES group) has provided a basic description of the outer shelf region (between Unimak Island and the Pribi- lofs) between the shelf break— around 150 m— and approximately the 100-m depth contour. This region, bounded by two oceanographic fronts, appears to constitute a transition zone between the Bering Sea source and deep shelf water. The inner boundary front (80-100 m depth, around 100-130 km from the shelf break) separates the two-layer shelf region from the transition zone which is basically three -layered: upper and bottom zones subject to wind and tidally generated mixing respec- tively, separated by a region of relatively low energy. This front extends from surface to bottom and appears to migrate into deeper water in the winter, when wind mixing operates to greater depths (50-60 m). The outer front is located mainly over the shelf break and forms a partial barrier between oceanic source and shelf waters. Since this structure does not extend to the bottom, Bering Sea water flows land- ward along the bottom of the outer shelf. Schu- macher et al. (1979), Coachman and Charnell (1979), and Coachman (1979) report no appreciable net circulation in the central shelf region; but tidal mixing operates to 40-50 m above the bottom and currents of the order of 50 cm /sec have been ob- served. The other effective mixing mode— wind energy— operates to 10-20 m in spring and summer and, as noted, considerably deeper in winter. Surficial sediment studies Creager and McManus (1967) briefly described the bottom sediments of the northern Bering Sea and summarized previous cruises and work to around 1965. McManus et al. (1974) thoroughly charac- terized a central-northern region. There have been many limited regional studies, particularly in the southeastern Bristol Bay region (see summary of Askren 1972) and the Gulf of Anadyr (considerable Soviet literature). Sharma et al. (1974) have con- sidered the use of satellite imagery to delineate sediment transport within Bristol Bay. A compre- hensive bibliography has recently been compiled by Naidu (1977). PROCEDURES Surficial sediment samples from the southern Bering Sea for which size fractionation, clay miner- alogy, heavy metal, and infauna data are given in this report were mostly collected in June 1975, using a Haps (Kanneworff and Nicolaisen 1973) stainless steel corer and van Veen grab. Additional samples for the size analysis program were collected on subse- quent NOAA/OCS-sponsored cruises in the same area. Suspended sediment samples were obtained on the former cruise by in-line filtering water from rosette-mounted drop-top Niskin bottles through a Nuclepore membrane. Since this would have been an unsuitable procedure for obtaining column samples to determine soluble trace constituents, a separate sampling program for these was conducted aboard the U.S.S.R. hydromet vessel Volna in July-August 1977. The stations occupied on this joint U.S. -U.S.S.R. expedition did not, unfortunately, coincide with stations of the earlier cruises. Bottom sediment samples were collected in Norton Sound (September, 1976) using the same techniques as for the south- eastern Bering/Bristol Bay region. Sources of the additional clay mineralogy data for the northern Bering Sea have been given by Naidu et al. 1977. Granulometric analysis has been by conventional sieving-pipetting procedures (Folk 1968) with genera- tion of Inman (1952) statistical parameters through a modified computer routine after Creager et al. (1962). For the southeastern Bering Sea shelf sam- ples the phi-grade scale of Krumbein (1936) and the Wentworth (1922) size limits and class terms have been used; also the size terms of Folk (1954), based on a classification scheme that emphasizes gravel, Geochemical characteristics of sediments 307 I since the proportion of gravel is in part a function of the highest current velocity at the time of deposition. Only major categories (gravel, sand, mud) were determined for the surficial sediment samples col- lected from Norton Sound. Clay mineral analysis essentially followed the procedures given by Naidu et al. (1971) and Mowatt et al. (1974). All clay mineral quantifications in this study are based on the method of Biscaye (1965) and are, at best, semi-quantitative. Total analysis for elemental constituents of de- posited and particulate sediment has been performed by D. E. Robertson using neutron activation: Mn, V, and Al by "rabbit" irradiation, and the other ele- ments using techniques given by Robertson and Carpenter (1974). Chemical extracts of sediments were obtained using 25 percent v/v acetic acid as mandated by the contracting agency. Metal concen- trations thus determined are lower than would have been given by the more conventional mixed acid -reducing treatment; but the latter has only the virtue of more universal use, since several studies (e.g., Luoma and Jenne 1976) have shown little actual correlation between "bioavailability," as evidenced by measured uptake, and various chemical leaching treatments. Soluble copper and lead con- tents of seawater were obtained using anodic strip- ping voltammetry (Heggie and Burrell 1976). Identification and quantification of the southern Bering Sea infauna have been performed by the University of Alaska Marine Sorting Center. SEDIMENT DISTRIBUTION The distributions of mean phi, sorting (standard deviation), grain-size mode for sand, and the percent abundance of gravel, sand, silt, and clay across the southeastern Bering Sea continental shelf are shown in Figs. 19-1 to 19-7. Such maps of distributions of (surficial) sediment characteristics can be used to indicate source areas, processes of transportation and deposition, and environments of deposition (K. Tommos, unpublished data); only a brief summary is given here. Sand -sized sediments, in the classification scheme of Folk (1954), dominate this region of the Bering Sea shelf; in Wentworth class terms, coarse silt is also regionally significant. The relative abundance of sand ranges from about 20 percent to almost 100 percent (average content, 68 percent). Samples collected along the Alaska mainland, at an average depth of about 30 m, consist of >90 percent sand, and an 80- percent sand isopleth generally follows the 45-m isobath. The silt content in this region is low— less than 10 percent. It was found that isopleths of MEAN (ph Figure 19-1. Southeastern Bering Sea shelf: Grain-size mean (0). c^^' SORTING Figure 19-2. Southeastern Bering Sea shelf: Sediment sorting. sediment textural parameters tend to follow bathy- metric contours on the inner shelf: sediment finer than 125 nm (3 0) does not occur in depths of <35 m and generally the mean size decreases uniformly below this depth. Sediments coarser than 250 lim (2 0) are restricted to depths of <50 m, sug- gesting that resuspension and transport of fine-grain sediments occur in this zone— conclusions which support the findings of Askren (1972). This sedimen- tological pattern is compatible with recent concepts of the local physical oceanographic regime. The high 308 Chemical oceanography V + + 63 ^: J ^ ^ 711 .^It^ ^m9 ^149 *'l X4>-\f- 7U +'^^ H>\?' +'" 300 Vy + * S3 „ ,^'" '■^' + TU 74 4^H^ + ^300 ^74 ^ ^]H9 V"*^ ^149 149 + '" ^,4S1^30O BO ^ + / J 119 ^U9 +'^ ^,2S ^U9 yO _^1 25 fci. £^\ _oS^°'_ 172. • 170. • GRAIN SIZE NODE. jUM, FOR SAND Figure 19-3. Southeastern Bering Sea shelf: Grain-size mode (jum) for the sand fraction. GRAVEL C/.l Figure 19-4. Southeastern Bering Sea shelf: Abundance of gravel in surface sediment (%). 11^. '0. SPND (7.) IS". ■ 54. ■(). 'a^«6,16 4-'"-"/ ^-^ J^ / . "39.43 19 84 y ^ V ~?s 60. ■ ^-..Jg^." ^^^t^ i - -^\ + ' V X / « ^^..S ,^0.7, -,' i S9. • 1 39 SD ^ 1^-^'' 4 ' ''* W _^ 0.G3 0 26 + ° ^^ Xfl 1 58. • ^33.BS 4-"-°= ? QQ 4 0.02 4. ■f 87 410.'6§2, ■*" _^ 0.61 + 23. S3 3j j3 _j_ 11.16 + ^'-^^ 3.. ^ 3.75 cw;0^ ^ 24. la ^ 30,33 ^ 33. J8 ^ 5.S6 y^ 1 .fr.o /^ rA 10 ^, . 0.3p^ JZ} ^ p. us ^19. 9? XV ^yjjfS^f*- ss. ■ _^ 28.01 ^ 6 4 3° ^ Jo • f K -_.-^.- - Figure 19-5. Southeastern Bering Sea shelf: Abundance Figure 19-6. Southeastern Bering Sea shelf: Abundance of sand-sized sediment (%). of silt-sized sediment (%). sand content, well-sorted character, and consistent increase in mean size in depths shallower than ap- proximately 50 m reflect the proximity to the mainland sediment sources and the subsequent influence of wind waves and tidal, storm, and coastal currents. The sandy gravel, gravelly sand, and sand found in this southeastern shelf region are apparently being supplied by the bordering rivers: the Egegik, Naknek, Kvichak, Nushagak, and Kuskokwim Rivers are major contributors. Poorly to very poorly sorted sediment is present only close to shore. Gravel-sized material was obtained mostly in nearshore areas, especially in sediment from Bristol Bay proper, Kuskokwim Bay, and Unimak Pass. These sediments probably represent reworked fluvial or beach deposits, or, in isolated cases, relict or palimpsest deposits. The relative amounts of the silt component in the sediments over the entire south- eastern Bering shelf region range from 0.02 to 50 percent with a mean content of around 20 percent. This component is dominant in the low-energy environment region of this southern shelf; the largest Geochemical characteristics of sediments 309 CLflf C/.l Figure 19-7. Southeastern Bering Sea shelf : Abundance of clay -sized sediment (%). concentrations, however, occur in a number of isolated patches (K. Tommos, unpubhshed data). The mean concentration of clay-sized sediment is around 7 percent, with greater (but never dominant) contents north of Unimak Pass and St. Matthew Island, adjacent to St. George Island, and generally in the northwest part of this shelf region. Large gradients in grain size, sorting, and sand and silt content which are evident adjacent to the 40-50 m zone are believed to be the result of, or at least influenced by, the presence of the seasonally stable oceanographic front which here separates the coastal and central shelf domains. The pronounced gradients in turbulence and hence transport efficiency result in deposition of coarser-grained sediment landward and finer-grained sediment seaward of this zone. Across the remainder of the shelf, the relative amount of the sand component decreases discontinuously. Sedi- ments having a larger mean 0 (very fine sand) and improved sorting were collected near the shelf edge at around 150 m. This supports the presence of the permanent current proposed by Barnes and Thomp- son (1938). More recent investigations (Coachman and Charnell 1979) have emphasized the presence of a very low velocity northwesterly drift along the outer shelf. But Coachman (1979) has noted that sporadic eddies along the shelf edge may produce local transient flow of basin water onto the shelf. The effect of this on the energy regime could account for the increasing mean grain size and improved sort- ing in this zone. At this time it is possible to give only a general description of the textural variations in the surface sediments of the southeastern Bering Sea shelf. The preliminary assumption can be made that this sedi- ment is in dynamic equilibrium with the present energy environment. Emery (1968) considered that relict sediment covers most of the shallower portions of the Bering Sea, but this has been generally dis- puted by more recent workers (e.g., Askren 1972, Knebel 1972, Sharma 1972). There are as yet no satisfactory explanations for those restricted regions still defined as relict or palimpsest. These are fre- quently interpreted as fluvial material deposited during periods of lowered sea level. McManus et al. (1974) summarized contemporary studies indicating that the Yukon River previously flowed south of St. Lawrence Bank, so that fine sediment from this source would have been deposited in the present southern Bering Sea. The areal plots of sediment texture given here generally show a decrease in grain size seaward across the shelf agreeing with the graded shelf model of Swift (1970). The sediment finer than 250 nm (2 0) and the deterioration of sorting across the shelf reflect the decrease in the energy of dispersal mechanisms with water depth and distance from shore. There are, however, significant interruptions in this trend, suggesting that the surfi- cial sediments of the southern Bering Sea continental shelf are still approaching equilibrium with the pre- sent physical environment. Only primary size fractionation data (mud, sand, and gravel categories) are presently available for the Norton Sound surface sediment samples as shown in Figs. 19-8 to 19-10. Sand-sized material predomi- Figure 19-8. Norton Sound: Abundance of mud-sized sediment (%). 310 Chemical oceanography Figure 19-9. Norton Sound: Abundance of sand-sized sediment {%). Figure 19-10. Norton Sound: Weight (%) abundance of gravel in surficial sediment. nates in the outer reaches of the sound, as on the southeastern shelf discussed above. Sand is also the major component of the surficial sediment within Norton Bay. Within the embay ment adjacent to Unalakleet, however, and over a considerable portion of the central region of Norton Sound, mud (clay- plus silt-sized sediment) constitutes the dominant category. The latter deposit is clearly attributable to Yukon River discharge, although most of the im- mense quantity of sediment derived from this source is transported northwards through the Bering Strait. Gravel is dominant adjacent to the shoreline off Port Clarence. INTERACTIONS BETWEEN BENTHOS AND SEDIMENT SUBSTRATE Scatter diagrams for total number of individuals and wet weight (m~^ ) for each of the 25 southern Bering Sea stations having paired benthos-sediment size fractionation values are given as Figs. 19-11 and 19-12. These data demonstrate a reciprocal relation- ship with the largest total number of individuals (on a unit area basis) and the lowest total wet weight occurring at localities having grain-size modes be- tween 125 and 149 iJim (fine sand); larger weights of macrobenthos tended to occur where either coarser or finer sediment predominated. Since sediment particles between 100 and 200 ^m are among the first to move under current stress (Sternberg 1972) and tend to be well sorted as a consequence, it would appear likely that the large number of small indivi- duals at fine-sand stations is the result of tidal current action. Correlation coefficients between numbers and weights (m~^ ) of each species of sorted macrobenthos (where present at >2.5 percent) have been deter- mined (Hoskin 1978). Of 138 species, 16 have a correlation coefficient of 0.5 or greater and these may be conveniently considered by feeding mode. Scavengers (e.g., the amphipod Paraphoxus sp.. Fig. 19-13) increase in abundance with increasing coarse- ness of grain-size, and there appears to be a threshold in substrate grain size since neither this organism nor Echinarachnius parma were found where grain-size modes were <149 iim. For three predators (polychaetes, Nephtys caeca, and N. longasetosa) both numbers and weights increased with increasing coarseness of grain-size modes in their substrate. This trend was reversed for the gastropod predator Cylichna alba. The deposit-feeding polychaetes Ophelia limacina and Spio filicornis (Fig. 19-14) were also observed to increase in number and biomass with increasing grain size, and there may be a particle-size threshold for the former since it was not found in this sampling in substrates having grain-size modes <105 /um. For other deposit-feeding polychaetes identified, either the data were inconclusive (Ampharete arctica and Trauisia forbesii), or no relationship between abun- Geochemical characteristics of sediments 31 1 m (D 3 > c e n E 5000 p 4500 4000 - 3500 3000 2500 2000 1500 1000 500 0 L _L I J_ _L _L J 63 74 88 105 125 149 177 Grain-size mode t^m 210 250 300 Figure 19-11. Scatter plot for total number of macroben- thos individuals (m"^) vs. grain-size mode for southeastern Bering Sea shelf. 2500 r- 1000 - CM E ^ 800 ~ 600 400 200 L _L I _L J 63 74 88 105 125 149 177 Grain-size mode pm 210 250 300 Figure 19-12; Scatter plot for total wet weight of macro- benthos (m"^) vs. grain-size modes for southeastern Bering Sea shelf. dance and substrate modal grain size was apparent (Myriochete heeri and the Terebellids). The filter- feeding amphipod Haustorius eous was not found in substrates with grain-size modes finer than 149 /xm. The abundance of this organism decreased with increasing sediment coarseness, but the relation- ship between weight per unit area and substrate grain-size modes is unclear. It seems clear that characterization of the substrate by primary (dispersed) grain-size analysis can provide at best only an imperfect description of habitat. Johnson (1974) has well emphasized the importance of mixed-phase aggregates to benthos distributions- parameters which cannot be determined using con- ventional size fractionation techniques. Nevertheless the data summarized here (Hoskin 1978) provide a partial first-order description of animal-sediment relationships in the southern Bering Sea. CLAY MINERALOGY The clay minerals of the southeastern Bering Sea shelf consist predominantly of an expandable com- ponent (40-70 percent: Fig. 19-15) and illite (20-50 percent: Fig. 19-16) together with chlorite (30-40 percent). Kaolinite is present generally in amounts less than 10 percent. There are also indications of amorphous material in concentrations greater than have been found elsewhere on the Alaskan shelves. Paraphoxus sp. r = 0.688 200 r- 150 100 50 0 L 4- 4 k—l 1. X 63 74 88 105 125 149 177 210 Grain-size mode =pm 250 300 Figure 19-13. Scatter plot for number of individuals of the scavenger amphipod Paraphoxus sp. (m"^) against grain-size mode for southeastern Bering Sea shelf. Spio f ilicornis r = 0.506 r .06 CM 05 fc ^^ O) 04 £ O) • 03 * *^ o $ .02 .01 0 L -i- _L _L J_ ^ _L ■4 63 74 88 105 125 149 177 210 250 300 Grain-size mode /im Figure 19-14. Scatter plot for wet weight of the deposit feeding polychaete Spio filicornis (m"^) against grain-size mode for southeastern Bering Sea shelf. 312 Chemical oceanography 170' 168° 164' 160* 158* Figure 19-15. Distribution of expandable mineral component (%) in <2 /im size fraction of southeastern Bering Sea sedi- ments. The results of potassium and magnesium saturation of the (glycol) expandable mineral component suggest the latter to be degraded illite. The <1 ^m (e.s.d.) fraction shows less illite but an increased expandable component as compared with the <2 //m fraction (Naidu and Mowatt 1977). Expandable material may represent contributions from bordering volcanic regions; the kaolinite component presumably is introduced by way of the Kuskokwim-Nushagak and other river systems. Fig. 19-17 illustrates the primary distribution of clay minerals in the north Bering Sea region, re- presenting pooled information from a number of cruises and sources (A. S. Naidu, unpublished data). Naidu et al. (1977) have also recently published the results of a survey of some 80 surficial samples from this area. Again illite is invariably the predominant clay mineral present in the <2 iim e.s.d. size class. Similarly, subordinate amounts of a gly col-expand- able mineral (10-40 percent), chlorite (18-32 per- cent), and kaolinite (6-18 percent) are present. Significant amounts of the expandable component appear to be degraded illite, and enhsmced amounts (30-40 percent) were observed south of St. Lawrence Geochemical characteristics of sediments 313 170' 168° 164 160° 158 Figure 19-16. Distribution of illite (%) in <2 idm size fraction of surficial sediments of southeastern Bering Sea. Island. The region extending off the Yukon River north up to the eastern margin of the Chirikov Basin, the western portion of Norton Sound, and the Shpanberg Strait demonstrate significantly higher kaohnite/chlorite ratios. This is attributable to a local terrigenous source and current dispersal. The presence of notably high concentrations of expand- able component south of St. Lawrence Island is difficult to explain. Possibly clays of the region are relict and have been formed by alteration of local volcanic debris. The distribution patterns of clay minerals, together with corresponding information from north of the Bering Strait not considered above, support the common thesis that the Chukchi Sea is the major depositional site for clay -size sediment from the Yukon River. SEDIMENT HEAVY METALS Since the specialized sampler required to obtain contamination-free surficial sediment samples could not operate in sandy or coarser sediment, only a limited set of samples was available to us for this phase of the study. Both total and extractable (i.e.. 314 Chemical oceanography 180° 170° 160° 180 170° 180° 170° 160" 160' ~W CHUKOTSK PENINSULA GULF OF "anadyr\„ "5 ''~r.,f'.-:- SEWARD PENINSULA '^?:::-:::::SKuzitrin River ^--,Koyuk River 180° VA}A)^**'''*'*-''St. Lawrence Island.*... ••''"■■'•■•• -v :-:-.vW eFo;*'«%--'^-'^N-L':-:-:-:v:-:-" Norton sound | 'I '' a**"? .•>■.■:■■.■:■;' -.4^) Yukon River .n...i.TC-:-:-.-.T^ ... 170° 180° 170° 160° 160° /CHUKOTSK PENINSULA I ! .S- ] GULF OF ANADYR i^>*^^*-a Kobuk River SEWARD PENINSULA Koyuk River ,'^!*!^'ll!|JPr?'"'-"'oveas St. Lawrence Island ^ 66° XPANDA3LE % ^ 20 30 0 12 19 m 0-11 66" 180 170° 160° Figure 19-17. Distribution of illite and expandable clay mineral component and kaolinite/chlorite ratio of <2 yum size fraction of northern Bering Sea region. "available") heavy metal contents were determined for the south Bering Sea and Bristol Bay. Whole- rock sample concentrations of a wide range of trace metals have been given by Robertson (1977). Mean extrac table and total contents, and correlations of the latter, are given in Table 19-1. Fig. 19-18 illustrates the areal distribution of total vanadium (iug/g) as an example. These heavy metal contents are low— much lower than mean values reported for polluted coastal regions, but also lower than has been determined for surficial Gulf of Alaska sediments (as part of the same survey program: Burrell 1978). Robertson (1977) has drawn attention to the negative cor- relation with calcium (Table 19-1): calcareous sedi- ments demonstrate very low heavy -metal contents. But it is suggested here that the generally coarser character of the sediment as compared with, for example, the eastern Gulf of Alaska, is also a major factor: finer -grained sediments generally contain higher concentrations of sorbed metals. This is less readily apparent from these data because the metals for which both extractable and total data exist— manganese and iron— occur as primary oxide pre- cipitates and coatings. Figs. 19-19 and 19-20 show distributions of iron in this region: extractable con- tents (25 percent v/v acetic acid treatment) are some order of magnitude less than totals. The coefficients given in Table 19-1 primarily appear to demonstrate highly significant (0.1) correlations for structurail alumina-silicate elements: Al, Fe, Ca, Co. Although lack of suitable facilities prevented comprehensive collection of water-column samples for trace elemental analysis, a limited number of par- ticulate sediment samples were collected concurrently with the bottom sediment samples described above and these have been analyzed by D. E. Robertson (various unpublished reports). Mean values (MgA) ^or surface and near -bottom samples Eire given in Table 19-2. Robertson (1977) has suggested that the particulate values for manganese and aluminum are relatively high and for vanadium relatively low for an open-ocean environment. Near-bottom total Al, Fe, Mn, and V values are also significantly higher than those obtained for surface samples. This characteristic should perhaps be considered in conjunction with the mean soluble data for copper and lead in central Bering Sea waters Figure 19-18. Distribution of total vanadium (ppm) in surficial sediments of soutiieastern Bering Sea shelf. Geochemical characteristics of sediments 315 TABLE 19-1 Mean extractable and total metal concentrations and correlation coefficients of total contents for surficial sediments of the southern Bering Sea Contents Correlation Coefficients Extractable Total Al Ca Fe Mn V Cr Co As 6.09 ± .96 % Al 2.86 ±.9 % Ca .847 0.14 ± .05% 2.84 ± .55 % Fe .915 .782 494 + 199 ppm Mn .733 .361 .715 93 ± 17 ppm V .880 .569 .811 .948 64 ± 21 ppm Cr .570 .774 .648 .337 .435 10 ± 2 ppm Co .863 .738 .950 .629 .745 .596 3.6 ± 1.5 ppm As .466 .338 .713 .642 .582 .534 .707 0.6 ± .15 ppm Sb .500 .544 .658 .495 .474 .679 .706 .661 < 2.5 ppm Ni 11 ± 6 ppm Zn ) also given in Table 19-2 (these latter data are consid- ered more fully by Burrell, Chapter 21, this volume; Burrell 1978; and especially Heggie, 1980, and unpublished data). Near-bottom enhancement of particulate Al (in w/v units) would suggest bottom sediment resuspension, and such a mechanism could account for apparent local increases in the concentra- tions of Fe, Mn, and V at the base of the water column (Robertson, unpublished report). Such could also be extended to include enhancements in "solu- ble" contents— for example, the copper and lead of Table 19-2; i.e., enhancements due to the presence of particulate material less than the nominal 0.4 nm filter pore size (Burrell 1978). Fig. 19-21 illustrates the distribution of surface soluble (passing 0.4 //m membrane) copper and gives data for near -bottom samples from within the 100-m shelf contour. The mean bottom-water copper concentration given in Table 19-2 is for this shelf suite. It has been argued (Heggie, 1980) that the shallow-water copper gradient (negative away from the benthic boundary into the column) may be attributable to a flux of, in this case, remobilized copper from the sediments. We have previously demonstrated such a transport under estuarine conditions for copper (Heggie and Burrell 1980) and manganese (Owens et al. 1979), and such a process over a highly productive shelf region is certainly feasible. Data for the distribution of particulate manganese over the shelf are unfortun- ately too sparse either to support or refute this argument, and this important topic awaits specific investigation. EXTRflCTflBLE FE Figure 19-19. Content (%) of extractable iron from southeastern Bering Sea sediment. 316 Chemical oceanography Figure 19-20. Total iron content (%) of southeastern Bering Sea sediment. The distribution of size-fractionated sediment and the mineralogy of the clay-sized component within Norton Sound has been briefly noted above. The character of the surficial sediment in this area is influenced to a considerable degree by the Yukon ne'..' ITT* ■ ncT* m.' u7.'' m.' n!. • 171. 153, ' IH Figure 19-21. Soluble (<0.4 jum) copper contents of surface waters of central Bering Sea (D. T. Heggie, unpub- lished data). The second value given for shelf samples shows corresponding near-bottom values. TABLE 19-2 Some mean elemental particulate and soluble contents of Bering Sea water (from Robertson, Heggie, unpublished data) Element n Surface n Bottom A. Particulate sediment Al Mg/1 7 17.9 ± .2 17 39.0 ±.3 Fe Mg/1 4 4.13 ±.17 10 15.8 ±.5 Mn Mg/1 7 0.33 ±.03 17 0.83 ±.07 V Mg/1 - <16 16 92 ±19 As Mg/1 4 12.3 ±1.8 10 11.9 ±1.7 Co Mg/1 4 3.7 ± .09 10 5.4 ± .08 Sb Mg/1 4 1.03 ±.21 10 0.88 ± .27 B. Soluble concentrations Cu Mg/1 23 0.41 ± .22 13 0.69 ± .38 Pb Mg/1 23 0.20 ± .04 13 0.35 ± .10 River discharge. Table 19-3 demonstrates the close correlation of extractable contents of heavy metals from the sediment with the weight fraction of fine- grained sediment— i.e., with available solid -phase surface area. ACKNOWLEDGMENTS This work, Institute of Marine Science Contribu- tion No. 405, was supported largely by the Bureau of Land Management through interagency agreement with the National Oceanic and Atmospheric Adminis- tration, under which a multiyear program responding to the needs of petroleum development of the Alas- kan Continental Shelf is managed by the Outer Continental Shelf Environmental Assessment Program Office, and by the State of Alaska. Sediment samples for clay minerals studies were kindly supplied by Drs. J. S. Creager and C. H. Nelson. We are grateful for permission to include unpublished data by Dr. D. Robertson. Geochemical characteristics of sediments 31 7 TABLE 19-3 Mean extraction heavy metal concentrations and correlation coefficients for Norton Sound surficial sediments Coachman, L. K. 1979 Component 1: Water circulation and mixing in the southeast Bering Sea. Prog. Rep., PROBES Phase I, 1977-78. Metal Correlation coefficients %Mud Fe Zn Ni 0.2 % Fe 0.86 0.01 % Mn 0.53 0.27 5.3 ppm Zn 0.67 0.71 2.2 ppm Ni 0.73 0.98 0.76 0.7 ppm Cn 0.62 0.84 0.84 0.85 < 0.1 ppm Cd Coachman, L. K., and R. L. Charnell 1979 On lateral water mass interaction— A case study, Bristol Bay, Alaska. J. Phys. Oceanogr. 9(2): 278-97. Creager, J. S., and D. A. McManus 1967 Geology of the floor of the Bering and Chukchi Seas— American studies. In: The Bering land bridge, D. M. Hopkins, ed., 7-31. Stanford Univ. Press, Stan- ford, Calif. Creager, J. S., D. A. McManus, and E. E. Colhas 1962 Electronic data processing in sedimen- tary size analysis. J. Sedimentary Petrology 32:833-9. REFERENCES Askren, D. R. 1972 Holocene stratigraphic framework: southern Bering Sea continental shelf. M.S. Thesis, Univ. of Washington, Seattle. Barnes, C. A., and T. G. Thompson 1938 Physical and chemical investigations in Bering Sea and portions of the North Pacific Ocean. Univ. of Washington Publications in Oceanography 3(2): 35-79. Biscay e, P. E. 1965 Mineralogy and sedimentation of recent deep-sea clay in the Atlantic Ocean and adjacent seas and oceans. Geol. Soc. Amer. Bull. 76: 803. Burrell, D. C. 1978 Distribution and dynamics of heavy metals in Alaskan shelf environments subject to oil development. Unpub. Ann. Rep. to NO A A, Contract No. 03-5-022-56. Inst. Mar. Sci., Univ. Alaska. Dodimead, A. J., F. I'avorite, and T. Hirano 1963 Salmon of the North Pacific Ocean, Part II: Review of oceanography of the subarctic Pacific region. Inter. N. Pac. Fish. Comm. Bull. 13: 1-11; 177-87. Emery, K. O. 1968 Relict sediments on continental shelves of world. Bull. Amer. Assoc. Petrol. Geologists 52: 445-64. Folk, F. L. 1954 The distinction between grain size and mineral composition in sedimentary- rock nomenclature. J. Geol. 62: 345- 51. Folk, R. L. 1968 Petrology of sedimentary rocks. Univ. of Texas, Austin, Texas. Hebard, J. F. 1959 Currents in southeastern Bering Sea and possible effects upon king crab larvae. U.S. Fish and Wildlife Service, Spec. Scientific Rep.— Fisheries, No. 293. Heggie, D. T. 1980 Copper in surface waters of the Bering Sea. Presented at spring meeting, A.G.U., Toronto. 318 Chemical oceanography Heggie, D. T., and D. C. Burrell 1976 Differential pulsed anodic stripping voltammetry: Application to measiire- ment of copper in seawater. Proc. 27th Alaska Science Conf. 1980 Depth distributions of copper in the water column and interstitial waters of an Alaskan fjord. In: The dynamic environment of the ocean floor, K. A. Fanning and F. T. Mannheim, eds. D. C. Heath and Co., Lexington, Ky., in press. Hoskin, C. M. 1978 Benthos-sedimentary substrate inter- actions. Unpub. rep. Inst. Mar. Sci., Univ. of Alaska, Fairbanks. Inman, D. L. 1952 Measures for describing size of sedi- ments. J. Sedimentary Petrology 22: 125-45. Johnson, R. G. 1974 Particulate matter at the sediment- water interface in coastal environments. J. Mar. Res. 32:313. Kanneworff, E., and W. Nicolaisen 1973 The "Haps": a frame supported bottom corer. Ophelia 10: 129. Luoma, S. N., and E. A. Jenne 1976 Factors affecting the availability of sediment-bound cadmium to the es- tuarine, deposit feeding clam Macoma baltica. In: Radioecology and energy resources, C. E. Gushing, ed. Ecology Soc. Amer. Spec. Pub. No. 1. McManus, D. A., K. Venkatarathnam, D. M. Hopkins, and C. H. Nelson 1974 Yukon River sediment on the northern- most Bering Sea shelf. J. Sedimentary Petrology 44: 1052-60. Mowatt, T. G., A. S. Naidu, and N. Veach 1974 Glay mineralogy of the lower Golville River delta, north arctic Alaska. State of Alaska, Div. Geol. Geophys. Survey, Open File Rep. 45: 39. Naidu, A. S. 1977 A bibliography of the available litera- ture on recent sediments of the con- tinental shelves of Alaska. Unpub. rep., Inst. Mar. Sci., Univ. of Alaska, Fairbanks. Naidu, A. S., D. G. Burrell, and D. W. Hood 1971 Glay mineral composition and geo- logical significance of some Beaufort Sea sediments. J. Sedimentary Pet- rology 41: 691. Kinder, T. H. 1977 The hydrographic structure over the continental shelf near Bristol Bay, Alaska, June 1976. Univ. of Wash- ington, Dep. of Oceanogr. Tech. Rep., Ref: M77-3. Knebel, J. J. 1972 Holocene sedimentary framework of the east-central Bering Sea continental shelf. Ph.D Dissertation, Univ. of Washington, Seattle. Krumbein, W. G. 1936 Application of logarithmic moments to size frequency of sediments. J. Sed- imentary Petrology 6: 35-47. Naidu, A. S., J. S. Greager, and M. D. Sweeney 1977 Characteristics and dispersal pattern of clay minerals in the north Bering Sea. G.S.A. meeting, Seattle. Naidu, A. S., and T. C. Mowatt 1977 Composition, source and dispersal pat- tern of clay minerals in southeast Bering Sea, Alaska. G.S.A. meeting, Seattle. Owens, T. L., D. C. Burrell, and H. V. Weiss 1979 Reaction and flux of manganese within the oxic sediment and basin waters of an Alaskan fjord. Proceedings Fjord Oceanographic Workshop, Victoria, B. G. (in press). Geochemical characteristics of sediments 31 9 Ratmanoff, G. E. 1937 Exploration of the seas of Russia. Publications of the Hydrological Insti- tute, Leningrad No. 25. Robertson, D. E. 1977 Natural distribution of trace metals in three Alaskan shelf areas. Unpub. rep. to NOAA. Sternberg, R. W. 1972 Predicting initial motion and bed-load transport of sediment particles in the shallow marine environment. In: Shelf sediment transport: Process and patterns, D. J. P. Swift, D. B. Duane, and O. H. Pilkey, eds., 61-82. Dowden, Hutchinson and Ross, Stroudsberg, Pennsylvania. Robertson, D. E., and R. Carpenter 1974 Neutron activation techniques for the measurement of trace metals in en- vironmental samples. NAS-NS-3114. Schumacher, J. D., T. H. Kinder, D. J. Pashinski, and R. L. Charnell 1979 A structural front over the continental shelf of the eastern Bering Sea. J. Phys. Oceanogr. 9: 78-87. Sharma, G. D. 1972 Graded sedimentation on Bering shelf. Proc. 24th International Geological Conference, Section 8: 262-71, Mon- treal. Sharma, G. D., F. F. Wright, J. J. Burns, and D. C. Burbank 1974 Sea surface circulation, sediment trans- port and marine mammal distribution, Alaska continental shelf. Unpub. rep., Inst. Mar. Sci., Univ. of Alaska, Fair- banks. Swift, D. J. P. 1970 Quaternary shelves and the return to grade. Marine Geology 8: 5-30. Takenouti, A. Y., and K. Ohtani 1974 Currents and water masses in the Bering Sea: A Review of Japanese work. In: Oceanography of the Bering Sea, D. W. Hood and E. J. Kelley, eds., 39-57. Inst. Mar. Sci., Occ. Pub. No. 2. Univ. of Alaska, Fairbanks. Wentworth, C. K. 1922 A scale of grade and class terms for clastic sediments. J. Geology 30: 377-92. \ ft The Distribution and Elemental Composition of Suspended Particulate Matter in Norton Sound and the Northeastern Bering* Sea Shelf: Implications for Mn and Zn Recycling in Coastal Waters Richard A. Feely, Gary J. Massoth, and Anthony J. Paulson Pacific Marine Environmental Laboratory Environmental Research Laboratories, NOAA Seattle, Washington ABSTRACT The distribution and elemental composition of suspended particulate matter in Norton Sound and the northeastern Bering Sea shelf were studied in July 1979. Samples were analyzed for total suspended matter and particulate C, N, Mg, Al, Si, K, Ca, Ti, Cr, Mn, Fe, Ni, Cu, and Zn. The results show that the bulk of suspended material in Norton Sound consists of sedimentary material discharged from the Yukon River and resuspended bottom sediments. The Yukon River material enters the sound from the southwest, is transported north and northeast around the perimeter of the sound, and exits from the northwest. The concentrations of the major and trace elements in the particulate matter and their elemental ratios with aluminum indicate that: K, Ca, Ti, Cr, Fe, Ni, and Cu are primarily associated with aluminosilicate material derived from the Yukon River and resuspended sediments, and C and N are primarily associated with terrestrial organic material in estu- arine samples and marine organic material in offshore samples. Significant enrichments of Mn and Zn, observed in the off- shore samples, are attributed to Mn recycling in the sediments followed by precipitation of Mn onto particulate phases in the water column, with the Mn oxyhydroxides scavenging Zn. INTRODUCTION Particles suspended in seawater play a major role in regulating the chemical forms, distributions, and deposition of many of its constituents. This is partic- ularly true in coastal waters where dissolved and particulate matter in runoff from rivers interacts with seawater. Particles in coastal waters are the result of continuing physical, chemical, biological, and geologi- cal processes and are the precursors of marine sedi- ments. These processes may include the supply of inorganic and organic substances from river runoff, aeolian fallout and coastal erosion, resuspension of previously deposited sediments, biological produc- tion of organic materials, and chemical adsorption- desorption and flocculation processes. Variations in the composition of particles in. suspension can be sensitive indicators of such processes. In this chapter we present the results of a survey of the distribution and elemental composition of suspended material in Norton Sound and the northeastern Bering Sea shelf. The results are related to known patterns of water circulation and previously pub- lished information on the chemical composition of suspended material from the Yukon River. We present arguments to support the position that Mn and possibly Zn undergo dissolved-to-particulate phase changes as a result of interactions between suspended matter and dissolved species within Norton Sound. 321 322 Chemical oceanography HISTORICAL BACKGROUND Previous work on suspended matter in Norton Sound has been limited to studies of LANDSAT photographs and distributions of suspended matter. Sharma et al. (1974) used density-shced LANDSAT photographs and sea-truth measurements to study distributions of suspended matter in Norton Sound during the late summer of 1973. Concentrations of suspended matter were highest near the mouth of the Yukon River (range: 2-8 mg/1) and in Norton Bay (range: 3-4 mg/1), in the northeast corner of the sound. The authors postulated that the general pattern of cyclonic circulation in the sound caused suspended material to be transported to the north and northeast along the coast. The authors also noted that unusually high concentrations of partic- ulate matter (>9.0 mg/1) were observed through- out the water column in the region approximately 30 km south-southwest of Nome. They suggested that this plume could have been a detached portion of the Yukon River plume which was isolated by tidal pulsation. Cacchione and Drake (1979) combined surveys of suspended matter during September and October 1976 and July 1977 with deployments of a tripod (GEOPROBE) containing instruments designed to measure bottom currents, pressure, temperature, and light transmission and scattering to study dis- persal patterns of suspended matter in Norton Sound. They described the transport of suspended materials as dominated by distinctly different quiescent and storm regimes. The quiescent regime was charac- terized by relatively low levels of sediment transport caused by tides and mean flow to the north and northeast, augmented by surface waves during spring tides. The authors stated that in this period, much of the fine-grained suspended matter present over the prodelta was resuspended at shallow depths during spring tide and transported northward with the mean current. The storm regime, which occurs during the months of September through November, was characterized by strong southerly and southwesterly winds which generate waves with heights of 1-3 m and periods of 8-11 sec. The storm events cause near-bottom shear velocities in excess of that required for resuspension of bottom sediments and, as a result, more than 50 percent of the sediment trans- port occurs during this regime. Although there is no background information on the chemistry of suspended matter in Norton Sound, extensive studies of trace-metal partitioning in various phases of Yukon River materials were conducted by Gibbs (1973, 1977). He concluded that transition metals associated with oxyhydroxide coatings and crystalline phases comprised the major fraction (72-91 percent) of riverine transition metal trans- port to the sea. Particulate organic phases contained the next largest fraction (3-16 percent of the total). Metals in solution and metals sorbed to particulate materials made up the remainder (5-15 percent of the total). 164° 162° 160° 170* 168° I66°W 164° 162° NORTON SOUND, ALASKA 0 20 40 60 niice o( Woles ' ' ' ' Noulicol Miles 65° Figure 20-1. Physiographic setting of Norton Sound showing: (a) locations of suspended matter stations, 7-18 July 1979 (b) bathymetry in meters. Suspended particulate matter 323 I I THE STUDY REGION Norton Sound is a shallow embayment of the Bering Sea in the central region of the west coast of Alaska (Fig. 20-1). It extends east-west about 220 km and north-south about 150 km. The Yukon River, which flows into the southwest quadrant of the embayment, is the major source of fresh water and suspended matter to the sound as well as to the entire eastern Bering Sea shelf. Its annual load of suspended matter, 88 X 10^ tons, ranks 18th among the major rivers of the world (Inman and Nordstrom 1971). The annual discharge curve for the Yukon River (Fig. 20-2) is unimodal with peak flow occurring during June and low flow condi- tions persisting throughout the winter months. Additional lesser freshwater influx into the sound occurs along the coastline east of the Yukon River Delta and along the northern coast. Water circulation in the vicinity of Norton Sound has been described by several authors (Coachman et al. 1975, Muench and Ahlnas 1976, Muench et al. Chapter 6, this volume). The shelf water west of Norton Sound, the Alaskan coastal water, has a net northward flow of about 1.5 X 10^ m^ /sec. About one-third of this flow passes between St. Lawrence Island and the mouth of Norton Sound. This flow induces the cyclonic water circulation inside Norton Sound. The intensity of the cyclonic flow appears to be affected by local winds and by freshwater runoff. The eastern half of the sound is characterized by two vertically well-mixed layers. The upper layer contains runoff water from the coastal rivers; the lower layer contains cold, dense residual water formed during the previous winter. Both water masses follow the general pattern of cyclonic flow in the region, although much more sluggishly than surface and bottom waters further to the west. The distribution of sediments in Norton Sound has been summarized (McManus et al. 1974, Sharma 1974, Nelson and Creager 1977, and McManus et al. 1977). In the central and southern regions, the sediments consist of very fine grained sands and silts which are modern. In the northern region, silty sands predominate everywhere except for a narrow strip along the coast between Cape Nome and Cape Douglas. Here, coarse sands and gravels predominate because bottom currents have caused almost com- plete erosion of the fine-grained sediments. Approx- imately one-half to two-thirds of the sediment load of the Yukon River is deposited as a band of sedi- ments extending from the Yukon River Delta north- ward and eastward around the perimeter of the sound. The remaining sediment load of the Yukon River is transported to the north through Bering Strait and deposited in the Chukchi Sea. ro E o < X o (/) Q CL UJ > 0 N D Figure 20-2. Monthly means and ranges for Yukon River disciiarge. Data compiled from U.S.G.S. streamflow ob- tained at Pilot Station (located approximately 200 km up- stream from the river mouth) for period of record: 1975-78. EXPERIMENTAL PROCEDURES Sampling methods In order to obtain information about the distribution and composition of suspended matter in Norton Sound, samples were collected as part of an interdisciplinary survey of the region (7-18 July 1979). Fig. 20-1 shows the locations of the stations occupied during the survey. Water samples were collected in General Oceanics Model 1070 10-1 PVC Top Drop Niskin bottles from the surface and 5 m above the bottom, and from intermediate depths along two vertical sections spanning the length of the 324 Chemical oceanography sound (Transect I: Stations 3, 4, 8, 11, 15, 20, 29, 38, 43, and 48; Transect II: Stations 1, 9, 10, 17, 18, 22, 25, 27, 39, 42, and 49). To avoid loss of rapidly settling particles (Gardner 1977, Calvert and McCartney 1979), aliquots from each Niskin bottle were rapidly withdrawn (within 10 to 15 minutes of collection) and vacuum-filtered through preweighed 0.4-/im pore-size Nuclepore polycarbonate filters (47 mm in diameter for suspended matter concentration determination and 25 mm in diameter for elemental analyses other than C and N) and precombusted 0.45-ium pore-size Selas silver filters (25 mm in diameter for C and N analyses). All samples were rinsed with three 10-ml aliquots of deionized membrane-filtered water (adjusted to pH 8.0), placed in individual polycarbonate petri dishes with lids slightly ajar for a 24-hour desiccation period over sodium hydroxide, and then sealed and stored for subsequent laboratory analysis. Temperature and salinity data were obtained with a Plessey Model 9040 CTD system equipped with a Model 8400 data logger. This system sampled several times per second for simultaneous values of conduc- tivity, temperature, and depth. The data were averaged to provide 1-m temperature and salinity values, from which sigma-t was computed. The sigma-t values are accurate to ±0.02 sigma-t units. Analytical methods Total concentrations of suspended matter were determined gravimetrically. The weighing precision (2a = ±0.011 mg) and volume reading error (±10 ml) yield a combined coefficient of variation in sus- pended matter concentration of approximately 1 percent. This variability is probably overshadowed, however, by that associated with the sampling preci- sion as reported by Feely et al. (1979) for other Alaskan coastal waters, where the relative standard deviation ranged from 5 to 25 percent. The major (Mg, Al, Si, K, Ca, Ti, and Fe) and trace (Cr, Mn, Ni, Cu, and Zn) elements in the sus- pended matter were determined by x-ray secondary- emission (fluorescence) spectrometry using a Kevex Model 0810A-5100 x-ray energy spectrometer and the thin-film technique (Baker and Piper 1976). A silver x-ray tube (operated at 50 kV, 40 mA) was used to excite a sequence of secondary targets (Fe target for Mg through Cr; Se target for the remain- ing elements) which efficiently fluoresced the range of elements in each sample. Standards were pre- pared from suspensions of finely ground U.S.G.S. Standard Rocks (W-1, BCR-1, AGV-1, and GSP-1; 90 percent by volume were less than 15 iim in diam- eter as determined by scanning electron micros- copy) collected on Nuclepore filters identical to those used for sample acquisition. At a filter loading of 325 iigjcm^ the determination limits (three times the minimum detection limits) were less than 0.25 percent and 13 ppm for the major and trace ele- ments, respectively. The relative standard deviations resulting from 10 replicate analyses of a sample with a similair weight distribution were less than 3 percent for major elements and less than 8 percent for trace elements. The amorphous Mn and Zn in the poorly structured oxyhydroxide phase of selected suspended matter samples were determined by the method of Bolger et al. (1978). Desiccated samples were leached with 5 ml of 25 percent (v/v) Ultrex acetic acid at room temperature for two hours. The resulting supernate was filtered through an acid-cleaned polypropylene-glass apparatus containing a 0.4- iim. Nuclepore filter. The residue was rinsed with quartz-distilled water, then filtered; the supernate was combined with the original supernate, acidified with 0.5 ml of concentrated Ultrex HCl, and stored in an acid-cleaned polyethylene bottle. The Mn and Zn in this solution (weak-acid-soluble) were analyzed by flameless atomic absorption procedures using standard addition methods. The remaining solid suspended matter (weak-acid-insoluble) was dis- solved in an Ultrex HCI-HNO3-HF matrix following Eggimann and Betzer (1976) and analyzed for Mn and Zn in a similar manner. Analysis of total particulate carbon and nitrogen in the suspended matter was performed with a Hewlett Packard Model 18 5B CHN analyzer. In this procedure, particulate carbon and nitrogen compounds were combusted to CO2 and N2 (micro Pregl-Dumas method), chromatographed on Poropak Q, and detected sequentially with a thermal conduc- tivity detector following a modification of the procedure outlined by Sharp (1974).' NBS acetanilide was used for standardization. Analyses of replicate surface samples yield relative standard deviations ranging from 2 percent to 10 percent for carbon and 7 percent to 14 percent for nitrogen. ' Because the silver filters used for sample acquisition could not be accurately weighed, carbon and nitrogen data were determined on a mass-per-volume basis. Weight percent data for carbon and nitrogen were obtained by comparison with suspended matter loadings obtained with the 47-mm Nucle- pore filters. Suspended particulate matter 325 I66'W a. I I I66'W 164' 170° 168° 166° 164° 162° 160° 6 2° b, TOTAL SUSPENDED MATTER(mg/l) <1.0 168' C. d. Figure 20-3. Distribution of: (a) salinity (b) temperature (c) sigma-t (d) total suspended matter at the surface in Norton Sound, 7-18 July 1979. RESULTS AND INTERPRETATIONS Distribution and transport of suspended matter Figs. 20-3 and 20-4 show the distributions of salinity, temperature, sigma-t, and total suspended matter at the surface and 5 m above the bottom for the July 1979 cruise in Norton Sound. As shown in Fig. 20-3, surface distributions of particulate matter were dominated by the discharge of sedimentary ma- terial from the Yukon River. Surface concentrations of suspended matter were highest near the mouth of the Yukon River, where values ranging between 100 and 154 mg/1 were observed. The Yukon River plume (as indicated by the 5.0-mg/l isopleth) ex- tended to the north and northeast across the length of the sound. Another portion of the plume with lesser concentrations of suspended matter (1.0-2.7 mg/1) extended north and northwest to a point about 20 km southwest of Cape Rodney. Both portions 326 Chemical oceanography 170* Btring Stroil Diomedes 02.0 166° a. 168' 63- Bering Sfroil 0, DiomMe. 66' b. I66'W ISA- TOTAL SUSPENDED MATTER(mg/l) < 1.0 1-5.0 5-1 5 15-50 160* c. Figure 20-4. Distribution of: (a) salinity (b) temperature (c) Norton Sound, 7-18 July 1979. appear to have originated from the Yukon River, and their trajectories tend to follow the general pattern of cyclonic circulation in the sound (i.e., Yukon River material enters the sound from the southwest, is transported north and northeast around the inside perimeter of the sound, and exits from the north- west). These data are supported by the salinity and temperature measurements, which indicated movements of low-salinity (12-240/oo), relatively warm (10-11 C) water to the northeast along the coast. These results are consistent with the general conclusions of Sharma et al. (1974) from suspended matter data obtained in August 1973. They are also sigma-t (d) total suspended matter at 5 m above the bottom in consistent with dispersal patterns of the Yukon River plume inferred from LANDSAT satellite photographs (Nelson et al. 1975). For example, Fig. 20-5 shows a LANDSAT photograph of the Yukon River plume taken on 20 July 1979. The plume, which appears in lighter grey tones than the less turbid water, can be traced as far north as approximately 70 km from the Yukon River Delta and as feir east as 50 km from Stuart Island. These features are also consistent with the data of Cacchione and Drake (1979) from surveys made during quiescent periods in September 1976 and July 1977. Thus, it appears that the transport processes described above predominate throughout Suspended particulate matter 327 MSS Band 5 of images E-21640- Figure 20-5. LANDSAT 21360-5 and E-21640-21363-5 taken on 20 July 1979, showing evidence of transport of sus- pended matter (appearing ligiiter in tone than the less turbid water) into Norton Sound. r I the region, at least during periods of calm weather in the summer. The near-bottom distribution of salinity, tem- perature, and total suspended matter also gives evidence of cyclonic movement of low-salinity (20-220/00), warm (-10 C) water to the northeast along the coast. This water mass can be traced as far north as Cape Darby. Near-bottom concentrations of suspended matter were highest near the mouth of the Yukon River and in the region about 20-30 km south- southwest of Nome. The near-bottom plume just seaward of the Yukon River extended to the north- east along the coast in a manner very similar to the surface plume. The near-bottom concentrations were 328 Chemical oceanography a. DISTANCE FROM THE EASTERN SHORE I68°59 30 48 43 40- KILOMETERS 300 200 I I STATION S 58 29 20 15 100 b, DISTANCE FROM THE EASTERN SHORE Kl LOMETERS 400 300 200 100 SALINITY (7oo) Contour Interval = 1.0 Transect I TEMPERATURE (°C) Contour Interval = 1.0 c, DISTANCE FROM THE EASTERN SHORE KILOMETERS 400 300 200 100 DISTANCE FROM THE EASTERN SHORE Kl LOMETERS 400 300 200 100 I68°59'30" 48 43 40- SIGMA- t Contour Interval = 1.0 40 TOTAL SUSPENDED MATTER (mg/L) Contour Interval = 2.0 Figure 20-6. Vertical cross section of the distribution of: (a) salinity (b) temperature (c) sigma-t (d) total suspended matter for transect I in Norton Sound. generally higher than surface concentrations, indicat- ing that: (1) some fraction of the Yukon River material had settled to the near-bottom region during transit, and/or (2) a portion of the bottom sediments had been resuspended and remained in suspension. Figs. 20-6 and 20-7 show cross sections of total suspended matter, salinity, temperature, and sigma-t for two east-west transects across the length of Norton Sound. Transect I is near the middle of the sound and Transect II is in the northern half, approxi- mately 20-30 km from the coast. The two transects show very similar water mass characteristics. On the eastern side of the sound, both transects showed evidence of a two-layer system with the pycnocline varying in depth between 8 and 14 m (e.g.. Stations 3 and 8 from Transect I and Stations 1, 9, and 10 from Transect II). Concentrations of suspended matter in this region showed a steady increase from about 6-8 m to the bottom. In the middle region (Stations 15, 20, and 29 of Transect I and Stations 22, 25, 27, and 39 of Transect II), the water column was virtually unstratified and concentrations of suspended matter increased to values greater than 30 mg/1 in the near-bottom waters (e.g.. Station 25 in Transect II). These water properties suggest that there is a tidally induced frontal zone with intense enough vertical mixing to break down the stability structure, with subsequent resuspension of sediments throughout the water column in some locations. This feature has the same characteristics as the anomalous plume of suspended matter described by Sharma et al. (1974) from data obtained in August 1973 at the same location. These data support the general conclusion of Cacchione and Drake (1979): during quiescent periods in the sound, resuspension of bottom sediments occurs as a result of increased tidal mixing during spring tide. Further seaward, the water column was moderately stratified and concentrations of suspended matter decreased to values below 0.5 mg/1 in surface waters. In near-bottom waters, however, concentrations of suspended matter were generally greater than 2.0 mg/1. The enriched concentrations of suspended matter near the bottom were probably caused by a combination of factors, including advective trans- port of particle-laden water to the northwest from Suspended particulate matter 329 a, DISTANCE FROM THE EASTERN SHORE KILOMETERS 400 300 200 100 I I I69°0I 00 10 20 30 40F I STATIONS 39 27 25 22 18 17 lez-oz'oo" SALINITY (%< Contour Interval = 1.0 DISTANCE FROM THE EASTERN SHORE Kl LOMETERS 300 200 100 TEMPERATURE CO Contour Intervol = 1.0 DISTANCE FROM THE EASTERN SHORE KILOMETERS 400 300 200 100 I69°0I 00 SIGMA-t Contour Interval = 1.0 I i69°oroo" DISTANCE FROM THE EASTERN SHORE Kl LOMETERS 400 300 200 100 I I STATIONS 39 27 25 22 18 17 TOTAL SUSPENDED MATTER (rT>q/L) Contour Interval = 2.0 0 I I62»02'00" Figure 20-7 Vertical cross section of tiie distribution of: (a) salinity (b) temperature (c) sigma-t (d) total suspended matter for Transect II in Norton Sound. k within Norton Sound (Muench et al., Chapter 6, this volume) and local resuspension of bottom sediments. Particulate elemental composition In order to determine regional variations of the chemical composition of suspended material in Norton Sound, the particulate samples from the July 1979 cruise were analyzed for their major and trace element content by the methods described previously. The resulting data have been separated into five regions: Yukon River estuary with salinities of less than 15°/oo, Yukon River estuary with salinities between 15 and 25°/oo, eastern Norton Sound, central Norton Sound, and western Norton Sound/northeastern Bering Sea shelf. The averaged chemical data, along with published data for the Yukon River, are given in Tables 20-1 and 20-2. Table 20-3 shows C/N and element/Al ratios for the averaged data. The elemental concentrations and elemental ra- tios illustrate some differences in composition be- tween the suspended material discharged from the Yukon River and suspended matter in the sound. These differences can be viewed in terms of relative percentages of aluminosilicate and organic matter. Since most of the aluminum in marine particulate matter is in aluminosilicate material (Sackett and Arrhenius 1962), the Al concentrations in the sus- pended matter multiplied by 10 can be used to estimate percentages of aluminosilicate in the partic- ulate matter. Similarly, Gordon (1970) suggested that particulate carbon multiplied by a factor of 1.8 may be used to estimate the amount of organic ma- terial in the suspended matter. On the basis of the particulate Al and particulate C concentrations, the composition of the suspended matter from the Yu- kon River estuary was determined to be approxi- mately 88 percent aluminosilicate material and 6 percent organic matter. In like manner, samples from eastern and central Norton Sound are deter- mined to contain about the same percentage of aluminosilicate material (88-92 percent). These results illustrate the predominance of the detrital 1—1 >. O ^ 'So CM O CO CO "O -" O S " p a. XJ tu T3 C o ^ 3 ° <; .2 O 5 t; .s O 1X1 O) - — - CO -c ? > 3 C en o =M .id O 3 C >^ .2 +1 c S o S a +1 S to O a, +1 6§ H ^ +1 • to CO ;e rH O ^ +1 65 ^ ^ +1 ■ to — S '-' c« > +1 <; ^ +i 6^ ^ -^ ^ k.^+1 6§ tw -2 o "a d 12; S CO c ,2 _aj a. a. ^ s o U5 CQ (1> W Q o CSI CO 05 O 05 CM "* 00 CM Tt ^ T-H co lO CO ^ 05 o 00 00 00 o CO ^ M "a CO H 00 CM 00 > C o s CO -t^ C W o o 3 >H s c o -ft; 3 iH a> +1 0> 00 +1 Oi 00 +1 U5 O +1 (N rH Oi 00 03 rH o cr> iq o o o +1 rH O +1 CM o +1 CO rH O CD rH O lO C£) 00 ^_ lO o +1 05 CM 05 03 CM rH 05 LO CM rH LO CO O o O +1 CD CO o rH o +1 CM O +1 lO t- r-i CO +1 05 rH +1 1— 1 00 CO o +1 o o +1 ■^ rH +1 ?^ !>. O 3 CD •*-j to K) I. 'o' o c5 S js; O ,=« CU (M ■«; 3 3 0) toti c OX) c _c 3 X! -O Ol o c« X! > 'Si X! C CO O c Ol ■>! CD >. "o X3 s CO 0) 0) (Tl CO "V^ u n l^ 3 a. C/3 & « CO o c- i>- CD c- 05 ,.— ^ r-i tr- "• — ' t~ c« 05 rH CO Q U5 c/i XI O) ^ CJ >-l o 3 O s C/3 0) <§ Pi ««-l Vh CO q; 4^ 4-3 CO CO Q ^ CO i "o -kd 3 »-i :z a; 4^ •a cH c C8 a; on ci _C -4J 'So Cm CO 3h S > o 3-1 C4-I U3 CO Xi iS a> C a> cc ■c« t/3 X5 CD 0 n. CD >. 9h (U S-i ^ a> -M j:: 775 CJ c«-i -tJ 2 c ^ g; >-i w a> c 0) ^ S O _g a; > 3 0 C3 3 £f W 0 XI c "« CU CJ c a> 'lac 1 x: O -^ T3 ^ O^ %— t .t-3 ° C C/3 CO CJ S >-i m o ^ 3 530 a „ o (35 cd t- o o> _o 1-1 TJ _>. OJ 3 -4J 1-3 u £ 00 "3 tH U3 t>- s T3 o W ^ ^^ CO • r-4 )H a> "a CO S -1 1 xi CO O O) ^1 1 c c ,2 "S -o 5 o O CO o 1 1 B z 3 _c CC ^ s to Ni a. +1 ^ S to O a +1 a to Z a, +1 • to r"^ S '-' !ii > +1 c S to S a, i-H S & +1 s to )-< a T-H o a, +1 65 to h^ 55 T-H +1 to CO O ■4J fe5 tH +1 to -1-J 1-1 +1 5§ to M 6§ 1-1 +1 to 1-1 +1 5§ to 1—1 +1 ^ to iH +1 55 to 6 4-i +1 o a, c c o c to o ^ +1 +1 in +1 CO o +1 (M O +1 LO id +1 CD CT) c^ 00 +1 (N O CD i-l +1 C~ lO lO +1 in o +1 in o +1 CD '^ CO 00 +1 00 1-1 00 CD +1 iH +1 1-1 00 00 +1 CD O CO lO 00 q O O +1 o" o" +1 "*. co_ '^ (M_ i-I o +1 i-I O +1 1-1 o +1 (M o +1 1—1 c- q c- o CO co' +1 i-i CO +1 oq ^. 1-1 CO ai 1-i 05 1—1 csi o +1 o +1 cd' t3 C O CO c o CD O 'st* Oi CSI +1 CD (Ji iH in CM +1 o +1 t~ o CD iH (M tH +1 00 CD Tj< CO 00 00 Tt q o o +1 o +1 CO o +1 CD in 0> CD 00 CD LO +1 in +1 00 q id o +1 05 00 +1 00 c^ CO CM +1 in q o o +1 o +1 1-1 o +1 (>q o +1 00 00 '^ lO o CO +1 CO +1 q t- 00 CSI 05 i-I +1 oo' l-I +1 q q co' o +1 1-1 1-1 +1 o o +1 CD CD q q 1-1 00 +1 id co' +1 ca CO C3 ^ > s c; i s "o 01 Si B s $ B . "o -4^^ 3 :-l 2 a> X! ^ c CO 0) oi COD S-4 c a> to ■0 _cO > ki -bj Ol ut ^ 0) X! 772 0 «M •i-J 2 c ^ 0) ^ s- C/^ Ol -(■J cw C4— 1 C o; ■5 s o _aj ^ "O) -^ 0 QC _c s CO 'co CJD 3 %-i C/2 T3 0) C £ 0 c s: O 0) 0 a> CtH c -c XJ M S 1) CO ^c 2 CD '5 ■"7" ^ o 2 0 tH Si T3 C CO -2 CO ■a % O a> X! C4-1 s: OJ 0 c CJ CO CO 0) O) x: be o CO o s^ £ C O) 0 a> c/3 40 percent by weight), which contains less Mg, Al, K, Ti, Mn, Fe, Ni, and Cu than aluminosilicate material (Martin and Knauer 1973). It is important to note, however, that on the average the samples from this region contain about 88 percent more Mn than the Yukon River estuarine samples. Similarly, Zn concentrations in the suspend- ed matter from this region are about the same as in the estuarine samples even though there is a signifi- cant drop in relative amount of aluminosilicate material in the suspended matter. These findings indicate that Mn and Zn concentrations in the sus- pended matter are controlled by distinctly different chemical processes. In an attempt to determine the chemical nature and source of the enriched Mn and Zn in the off- shore suspended matter, selected surface and near- bottom samples were treated with 25 percent (v/v) acetic acid to separate poorly structured oxyhy- droxides from the more crystalline phases. This procedure has been shown to selectively dissolve trace elements precipitated in acid-soluble metal oxides and those adsorbed onto mineral surfaces without affect- ing highly oxidized ferromanganese minerals or the lattice structure of clays (Hirst and NichoUs 1958, Chester and Hughes 1967, and Bolger et al. 1978). The results of these experiments are given in Table 20-4. The data show higher amounts of weak-acid- soluble Mn in the offshore samples than in the estuarine samples, which are significant at the p <0.05 level. These higher amounts, computed by taking the differences between the offshore and estuarine samples as a ratio to the estuarine samples, range between 134 percent and 351 percent and Suspended particulate matter 333 TABLE 20-4 Partitioning of Mn and Zn between weak-acid-soluble (WAS) and weak-acid-insoluble (WAI) fractions of suspended material from Norton Sound and northeastern Bering Sea Shelf (Data presented as a percentage of total suspended matter) Sample location No. of samples WAS Mn ±la WAI Mn ±la WAS Zn ±lo WAI Zn ±la Yukon River Estuary Eastern Norton Sound Central Norton Sound Western Norton Sound 0.066 0.052 0.0059 0.0140 ±0.017 ±0.006 ±0.0028 ±0.0031 0.155 0.040 0.0095 0.0099 +0.038 ±0.011 ±0.0031 ±0.0021 0.184 0.054 0.0108 0.0114 ±0.085 ±0.017 ±0.0056 ±0.0026 0.298 0.074 0.0107 0.0125 ±0.092 ±0.041 ±0.0053 ±0.0070 I I account for all of the excess Mn in the suspended matter. Similarly, the data for Zn in the weak-acid- soluble fraction show enrichments ranging between 61 percent and 83 percent in the offshore samples which are significant at the p <0.20 level. These results indicate that in the offshore waters Mn and Zn are being concentrated in the weak-acid-soluble fraction of the particulate matter, which in these samples probably consists of poorly structured oxyhydroxides of Mn. The probable sources of this material will be discussed below. DISCUSSION Probable source of excess Mn in the suspended matter There are several possible sources of the excess Mn in the suspended matter of Norton Sound. These include (1) differential settling of particles of various sizes; (2) resuspension of Mn-enriched sediments; and (3) reductive dissolution of Mn within recent sediments followed by oxidative precipitation of Mn onto particulate phases in the water column. The first mechanism is unlikely in view of Gibbs's (1977) data for the chemical variations in the various size- fractions of Yukon River suspended material. The mean particle size distribution of suspended material in the sound would have to be about an order of magnitude smaller (i.e., a decrease from an average size of about 20 nm to about 2 idm) for the two- to-threefold increases in total Mn to occur. Unless some unusual chemical interactions were occurring in the estuary, this would necessarily be accompanied by a similar enrichment of total Fe and Cu in the suspended matter. No enrichments of that mag- nitude were observed in the Fe and Cu data. Further- more, the particle-size data of Cacchione and Drake (1979) indicate that suspended matter in Norton Sound is primarily composed of fine-to-medium silt in the range between 4 and 32 ^^m. These data indicate that if differential settling occurs in Norton Sound, it is definitely not of the magnitude required to produce the observed Mn enrichments in the suspended matter. The resuspension mechanism can also be refuted using a similar argument. While the distributions of suspended matter indicated that bottom sediments were being resuspended, the Mn content of the bulk sediments has been reported to be only in the range of 600-1,650 ppm (Larsen et al. in press). This means that the Mn content of the resuspended material would have to exceed the concentration observed within the sediments by a factor of about 2-4 to ac- count for the observed Mn concentrations in the suspended matter. This would occur only if the clay size-fraction of the sediments were being preferen- tially resuspended. Since the particle-size data of Cacchione and Drake (1979) do not show any evi- dence for a decrease of this kind, this mechanism does not seem likely. Reduction of Mn after burial in recent sediments with accompanying upward transport of dissolved 334 Chemical oceanography Mn into the overlying water, followed by precipita- tion onto suspended matter best explains the ob- served data. Efflux of Mn from rapidly accumulating sediments has been reported for several estuarine and coastal environments (Elderfield 1976, Graham et al. 1976, Aller 1977, Trefry 1977, Yeats et al. 1979, and Massoth et al. 1979). From studies of the sediments extending seaward of the Mississippi River, Trefry (1977) found that Mn fluxes from recent sediments varied directly with sedimentation rate. High Mn fluxes (i.e., ~2.7 jug/cm^/d) were observed in sediments that accumulate at a rate of about 2.0 g/cm^ /yr, whereas low Mn fluxes (0.71 Mg/cm^ /d) were observed in sediments that accumu- late at a rate of 0.08 g/cm^ /yr. In Norton Sound modern sediments with accumulation rates ranging from 0.05 to 0.17 g/cm^ /yr cover an area of approxi- mately 22,000 km^ (Nelson and Creager 1977). Assuming an average sedimentation rate of 0.1 g/cm^ /yr for these sediments and using linear interpo- lation of Trefry's (1977) Mn flux data (i.e., 0.68 jug/cm^/d), approximately 1.5 X 10^ g Mn would be released daily into Norton Sound from this source. At this rate it would require approximately 21 days to account for all of the estimated excess Mn in the particulate matter (approximately 3.1 X lO"* g Mn, assuming a total area of 45,000 km^ , an average depth of 16 m, an average concentration of sus- pended matter of 4.0 mg/1, and an average concen- tration of excess Mn of 1,079 ppm). If it is assumed that the rate of Mn oxidation is fast relative to an accumulation time of 21 days, then contact periods approximately equal to this time would be required for the chemical interactions to occur. Circulation in the sound is not completely understood, but studies conducted in summer indicate relatively sluggish circulation (Muench et al., Chapter 6, this volume). Net currents, with speeds varying between 10 and 15 km/d in surface waters and between 1 and 4 km/d in deep water, have been measured for short periods of time. Using a mean current of 8 km/d and a mean travel distance of 400 km, it is estimated that about 50 days are required for water to pass through the sound— a little more than twice the time required for the Mn from the sedi- ment to accumulate on the suspended matter. Thus, if the underlying assumption that the kinetic rate of Mn oxidation in coastal waters is relatively rapid is correct, then the sediments could easily be the major source of the excess Mn in the suspended matter. The assumption of a rapid rate for Mn oxidation is supported by the recent findings of VVollast et al. (1979), to the effect that Mn oxidation in the Rhine and Scheldt estuaries is essentially complete within 10 days and the process is mediated by several strains of marine bacteria indigenous to coastal environ- ments. Implication for the geochemistry of Zn in Norton Sound This discussion of the geochemical behavior of Mn in the sound is also important for understanding the chemical behavior of Zn in the suspended matter. As noted earlier, both Zn and Mn are enriched in the weak-acid-soluble fraction of the particulate matter. This is probably due to adsorption and/or coprecipitation of Zn on or in the newly formed Mn oxy hydroxides. Fig. 20-8 shows a plot of the relationship between total Zn and total Mn and for both surface and near-bottom samples. The plot of total Zn versus total Mn is roughly linear (r = 0.60), indicating an association between these two metals in the particulate matter. These results suggest that as the Mn oxyhydroxides form on the particulate matter, Zn is scavenged from solution. In similar fashion, the relationship between weak-acid-soluble Zn and weak -acid-soluble Mn is also linear (r = 0.39). This process effectively concentrates Zn and Mn in the suspended matter, which eventually either settles to the bottom of the sound or is transported to the northwest into the northeastern Bering Sea shelf and beyond. 700 - • TOTAL PARTICULATE Zn and Mn 600 500 o WEAK-ACID SOLUBLE Zn and Mn — • Y= 123.4 +0.038X n 400 - • . / /r = 0.60 n = 36^^ CL ^ o z M 300 200 1 00 - • XJif W oo_ __ -. • • *\Y= 65.6 +0.01 ex "" ° 1 1 1 1 r=0.39 n=26 1 1 1 1000 2000 3000 4000 5000 6000 7000 MANGANESE (PPM) Figure 20-8. Plot of the relationships between total partic- ulate Zn vs. total particulate Mn and weak-acid-soluble Zn vs. weak-acid-soluble Mn for selected surface and near- bottom samples from Norton Sound and northeastern Bering Sea shelf. Implications for the geochemistry of suspended materials and sediments in regions beyond Norton Sound The physical, chemical, and biological processes affecting the distribution and chemical composition Suspended particulate matter 335 I of suspended matter in Norton Sound and the north- eastern Bering Sea shelf may also affect the composi- tion of suspended materials and sediments in oceanic waters beyond the Bering Sea. As we stated earlier, the work of Nelson and Creager (1977) indicates that as much as one-third of the sediment load of the Yukon River bypasses the northern Bering Sea to accumulate in a thick deposit of Holocene sediment in the southern Chukchi Sea. These authors suggest that resuspension and transport of previously depos- ited sediments from Norton Sound during storms may contribute significantly to this phenomenon. Further support for this mechanism is indicated by the recent findings of Cacchione and Drake (1979) that more than 50 percent of the sediment transport in Norton Sound occurs during storm activity. Since the resuspended sediments in Norton Sound become enriched in Mn and Zn in oxygenated waters, it is reasonable to expect that the particulate matter will remain enriched in these metals until the particles are buried within the sediments and the recycling process is reinitiated in the reducing zone (Elderfield 1976). If it is further assumed, as sug- gested by Yeats et al. (1979), that the Mn precipi- tating in the water column will tend to be preferen- tially associated with small-sized particles, then the Mn and Zn content of the suspended material could be further enriched if some of the larger particles settle out. Therefore, Mn and Zn could be continually enriched in the particulate matter that is transported past the Bering Sea into the Chukchi Sea, where it forms a major fraction of the suspended matter and the recent sedimentary deposits. Thus, Mn and possible Zn recycling in the Chukchi Sea may be even more pronounced than in Norton Sound, because the incoming particulate materials are significantly enriched in these metals and the accumulation rates for recent sediments in the Chukchi Sea are as great as or greater than in Norton Sound (Naidu and Sharma 1972, Nelson and Creager 1977). If this is true, then Mn and Zn can undergo a number of recycling processes before they are ultimately buried in continental shelf, slope, or deep ocean sediments. ACKNOWLEDGMENTS The authors wish to express their appreciation to Captain Nixon and the crew of the Discoverer, without whose help this work would not have been possible, and to Ms. M. Lamb and Mr. R. Dyer for assisting in sample collection and data reduction. This study, Contribution No. 444 from the NOAA/ ERL Pacific Marine Environmental Laboratory, was supported in part by the Bureau of Land Manage- ment through interagency agreement with the Na- tional Oceanic and Atmospheric Administration, under which a multiyear program responding to needs of petroleum development of the Alaskan continental shelf is managed by the Outer Con- tinental Shelf Environmental Assessment Program (OCSEAP) office. REFERENCES AUer, R. C. 1977 The influences of macrobenthos on chemical diagenesis of marine sedi- ments. Ph.D. Dissertation. Yale Univ., New Haven, Conn. Baker, E. T., and D. Z. Piper 1976 Suspended particulate matter: collec- tion by pressure filtration and ele- mental analysis by thin film x-ray fluorescence. Deep-Sea Res. 23: 181-6. Bolger, G. W., P. R. Betzer, and V. V. Gordeev 1978 Hydrothermally-derived manganese suspended over the Galapagos Spread- ing Center. Deep-Sea Res. 25: 721-33. Cacchione, D. H., and D. E. Drake 1979 Sediment Transport in Norton Sound, Alaska, U.S.G.S. Open File Rep. 79-1555. Calvert, S. E., and M. J. McCartney 1979 The effect of incomplete recovery of large particles from water samples on the chemical composition of oceanic particulate matter. Limnol. Oceanogr. 24(3): 532-6. Chester, R., and M. J. Hughes 1967 A chemical technique for the separa- tion of ferro-manganese minerals, carbonate minerals and adsorbed trace elements from pelagic sediments. Chem. Geol. 2: 249-62. Coachman, L., K. Aagaard, and R. B. Tripp 1975 Bering Strait: The regional physical oceanography. Univ. of Washington Press, Seattle, Wash. 336 Chemical oceanography Eggimann, D. W., and P. R. Betzer 1976 Decomposition and analysis of Refractory Oceanic suspended mate- rial. Anal. Chem. 48: 886-90. Elderfield, H. 1976 Manganese fluxes to the oceans. Mar. Chem. 4: 103-32. Feely, R. A., E. T. Baker, J. D. Schumacher, G. J. Massoth, and W. D. Landing 1979 Processes affecting the distribution and transport of suspended matter in the northeast Gulf of Alaska. Deep-Sea Res. 26(4A): 445-64. Gardner, W. D. 1977 Incomplete extraction of rapidly settling particles from water samples. Limnol. Oceanogr. 22(4): 764-8. Gibbs, R. J. 1973 Mechanisms of trace metal transport in rivers. Science 18: 71-3. 1977 Transport phases of transition metals in the Amazon and Yukon Rivers. Geol. Soc. Amer. Bull. 88: 829-43. Gordon, D. C. 1970 Some studies of the distribution and composition of particulate organ- ic carbon in the North Atlantic Ocean. Deep-Sea Res. 17: 233-44. Graham, W. F., M. L. Bender, and G. P. Klinkhammer 1976 Manganese in Narragansett Bay. Lim- nol. Oceanogr. 21: 663-73. Hirst, D. M., and G. D. NichoUs 1958 Techniques in sedimentary geochem- istry, I. Separation of the detrital and non-detrital fractions of lime- stones. J. Sediment. Petrol. 28: 461-8. Inman, D. L., and C. E. Nordstrom 1971 On the tectonic and morphologic classification of coasts. J. Geology 79: 1-21. Larsen, B. R., C. H. Nelson, C. Heropoulos, and S. S. Patry Distribution of trace elements in bottom sediments of northern Bering Sea. U.S.G.S. Open File Rep. (in press) . Loder, T. C, and D. W. Hood 1972 Distribution of organic carbon in a glacial estuary in Alaska. Limnol. Oceanogr. 17(3): 349-55. Martin, J. H., and G. A. Knauer 1973 The elemental composition of plank- ton. Geochim. Cosmochim. Acta 37:1639-53. Massoth, G. J., R. A. Feely, P. Y. Appriou, and S. J. Ludwig 1979 Anomalous concentrations of partic- ulate manganese in Shelikof Strait, Alaska: An indicator of sediment- seawater exchange processes. EOS Trans. Amer. Geophys. Union 60 (46): 852. McManus, D. A., V. KoUa, D. M. Hopkins, and C. H. Nelson 1974 Yukon River sediment of the north- ernmost Bering Sea Shelf. J. Sedi- ment. Petrol. 44: 1052-60. 1977 Distribution of bottom sediments on the continental shelf. Northern Bering Sea. U.S.G.S. Prof. Paper 759-C: C1-C31. Muench, R. D., and K. Ahlnas 1976 Ice movement and distribution in the Bering Sea from March to June 1974. J. Geophys. Res. 81:(24)4467-76. Naidu, A. S., and G. D. Sharma 1972 Geological, biological and chemical oceanography of the eastern central Chukchi Sea. In: WEBSEC-70: An ecological survey in the eastern Chukchi Sea, 173-95. United States Coast Guard Oceanographic Unit, Washington, D.C. Suspended particulate matter 337 Nelson, C. H., and J. S. Creager 1977 Displacement of Yukon-derived sedi- ment from Bering Sea to Chukchi Sea during Holocene Time. Geology 5: 141-6. Sharma, G. D. 1974 Contemporary depositional environ- ment of the eastern Bering Sea. In: Oceanography of the Bering Sea, D. W. Hood and E. J. Kelley, eds. 517-52. Inst. Mar. Sci. Occ. Pub. No. 2., Univ. of Alaska, Fairbanks. Nelson, C. H., D. M. Hopkins, and D. VV. Scholl 1974 Cenozoic sedimentary and tectonic history of the Bering Sea. In: Ocean- ography of the Bering Sea, D. W. Hood and E. J. KeUey, eds., 485-510. Inst. Mar. Sci. Occ. Pub. No. 2, Univ. of Alaska, Fairbanks. Sharma, G. D., F. F. Wright, J. J. Burns, and D. C. Burbank 1974 Sea surface circulation, sediment transport and marine mammal distri- bution. Alaska continental shelf: ERTS Final Rep. Nat. Tech. Service Rep. E74-10711. Sharp, J. H. 1974 Nelson, C. H., B. R. Larsen, and R. W. Rowland 1975 ERTS Imagery and dispersal of the Yukon and Kuskokwin River Plumes. In: Principal sources and dispersal patterns of suspended particulate matter in nearshore surface waters of the northeast Pacific Ocean, P. R. Curlson, ed., 26-40. U.S. Dep. of Commerce Nat. Tech. Inf. Serv. E75-10266. Sackett, W. M., and G. Arrhenius 1962 Distribution of aluminum species in the hydrosphere, I. Aluminum in the oceans. Geochim. Cosmochim. Acta 26:955-68. Trefry, J. H. 1977 Improved analysis for particulate organic carbon and nitrogen from seawater. Limnol. Oceanogr. 6(19): 984-9. The transport of heavy metals by the Mississippi River and their fate in the Gulf of Mexico. Ph.D. Disserta- tion, Texas A & M Univ. Wollast, R., G. Billen, and J. C. Duninker 1979 Behavior of manganese in the Rhine and Scheldt Estuaries. Estuarine and Coastal Marine Science 9: 161-9. Yeats, P. A., B. Sunby, and J. M. Bewers 1979 Manganese recycling in coastal waters. Mar. Chem. 8: 43-55. b Some Heavy Metal Contents of Bering Sea Seals David C. Burrell Institute of Marine Science University of Alaska Fairbanks, Alaska ABSTRACT The Cd, Cu, Ni, and Zn contents of liver, kidney, and muscle tissue of a suite of seal samples taken from the central Bering Sea are given. Heavy metals, particularly cadmium, are highest in livers and kidneys. Spotted seals have the lowest contents of heavy metals, and, tentatively, bearded seals the highest. However, no clear correlation between the contents of these metals and feeding habits or age could be discerned from the data base. INTRODUCTION This chapter presents basehne data obtained as part of the BLM/NOAA Outer Continental Shelf Environmental Assessment Program for Alaska. The samples were originally collected for other purposes. Nevertheless, we have obtained a number of precisely analyzed heavy metal data for an area— the central Bering Sea— which is of considerable economic importance and current oceanographic interest, but for which very little chemical characterization exists. The major portion of this work concerns heavy metal contents of various seal tissue samples. These distributions are considered in the context of the environmental baselines determined for water and sediments within coincident and adjacent portions of the Bering Sea. ANALYSIS TECHNIQUES All biota and sediment samples considered here have been analyzed spectrophotometrically for selected heavy metals. The major limitations on analytical precision for both classes of samples are connected with the dissolution step and inter-element matrix effects. The problem of oxidizing the tissue samples without loss of volatile metals was addressed here by low-temperature ashing in an oxygen plasma furnace. The residues from this step were then treat- ed with Ultrex nitric acid in a teflon digestion bomb prior to graphite furnace atomic spectrometric analy- sis. Standard curves run with each batch were pre- paired by adding standsirds to a matrix prepared in bulk for each type of tissue sample. NBS standards were carried through with each batch to monitor accuracy. Soluble seawater trace metal data are only a very minor component of this particular report. The strin- gent sampling strategy needed has been discussed in detail by Burrell (1978). Final analysis was by differential pulse anodic stripping voltammetry. ENVIRONMENTAL DATA By means of the careful sampling procedures de- scribed by Burrell (1978; Heggie, unpublished data), a number of water samples were collected from the U.S.S.R. hydromet vessel Volna on the stations shown in Fig. 21-1 in July-August 1977. Coinciden- tally, many of these stations are in that part of the Alaskan shelf west of Nunivak Island from which the seal specimens discussed below were collected the previous spring. Soluble copper and lead values have been deter- mined and Fig. 21-2 shows the mean ranges for over 100 samples collected through the water column (Heggie, unpublished data). These closely conform to distributions given previously for the Gulf of Alaska and elsewhere. Since mean soluble concen- trations vary only between narrow limits in the open ocean, this similarity was expected. However, nano- gram ranges have been suggested only recently for these metals; our data support such levels for 339 340 Chemical oceanography 175° 180' 175 170 165 160 Figure 21-1. Stations occupied by U.S.S.R. hydromet vessel Volna July-August 1977. unpolluted open-ocean water and at the same time add confidence to their accuracy. Although we have good data on the soluble portions of copper and lead only, there is no evidence of any anomalous regional trends, and it is expected that the concentrations of other trace metals in solution will show the same low ranges given in earlier publications for the Gulf of Alaska (e.g., Burrell 1978). The geochemistry of the surface sediments is discussed in Chapter 19 and need not be considered further here. HEAVY METAL CONTENTS OF MARINE MAMMALS Fig. 21-3 shows the localities of sacrificed seal samples collected on two separate cruises in March- April and May-June of 1977. The original objective of this project was to look for statistical differences between the heavy metal contents of four species of seal which were thought to have distinctive feeding habits. Our data are o cc LU o cc Figure 21-2. Mean ranges for soluble (<0.4 jum) and lead concentrations for samples collected from the surface and adjacent to the bottom at the stations given in Fig. 21-1 (D. T. Heggie, unpublished data). Heavy metal contents of seals 341 Figure 21-3. Localities of seal samples collected on O.S.S. Surveyor cruise March-April 1977 (open symbols) and O.S.S. Discoverer cruise May -June 1977 (closed symbols). largely for ribbon, spotted, and bearded seals (Tables 21-1 and 21-2), which were believed to feed pre- dominantly on demersal fish and benthos, pelagic fish, and invertebrate benthos respectively. Un- fortunately it was not found feasible to obtain representative food species when the mammals were collected. Nor were stomach contents suitable for either identification or analysis: in many cases it was found that the animals had starved for a number of days. The biological investigators also found the seals to be largely opportunistic feeders, so that diet differentiation by species, where it occurred, was observed only where there was adequate choice. For most of the individuals collected we have analyzed for cadmium, nickel, copper, and zinc in muscle, liver, and kidney tissue and these data, as means of duplicate determinations, are given in Tables 21-3 and 21-4. Table 21-5 gives accuracy and precision information relating to this batch of num- bers. 342 Chemical oceanography TABLE 21-1 Bering Sea O.S.S. Surveyor 31 March-27 April 1977 Seal samples collected for heavy metal analysis Sample Latitude Longitude Species Sex Weight Age # (N) (W) (kg) (yrs) 1 58°51.0 173°08.0 Ribbon F 39.5 1 2 58°51.0 173°08.0 Ribbon M 102.0 — 3 58°56.0 172°40.0 Ribbon M 81.8 4 4 58°45.6 172°55.4 Spotted M 35.0 1 5 59°00.6 173°15.0 Bearded F 181 6+ 6 58°53.0 173°07.0 Ribbon M 107.3 15 7 58°43.9 169°32.9 Bearded M 232 12 8 58°48.5 169°41.0 Bearded F 227 12+ 10 59°06.3 169°41.3 Spotted F 41.8 1 11 58°24.7 164°52.3 Ribbon F 98.6 3 16 58°21.3 164°49.7 Bearded F 204.5 2 28 58°54.2 169°13.6 Spotted M 89.9 6 29 58°40.1 169°40.3 Ribbon M 59.9 3 30 58°34.8 169°28.8 Spotted M 84.0 7 32 59°22.5 173°43.0 Spotted M 118.0 17 As noted above, marked differences in heavy metal contents as a reflection of a particular type of food would not be expected. Nevertheless, Table 21-6 lists some possible trends based on this very limited sample batch. It appears likely that the spotted seals have the lowest metal contents and— even more tentatively— bearded seals the highest. The former are considered to be largely consumers of finfish, whereas the bearded seals eat large quantities of benthic invertebrates. Since more metals are concen- trated in the benthos than in pelagic communities, this possible trend is of interest. Clearly more posi- tive correlation would require not only a considerably larger number of specimens of one species but also concurrent food species or, preferably, fresh stomach contents. Liver and kidney contents show generally elevated levels of these metals, as would be expected. Cad- mium in these organs is notably high, but the general lack of comparable reference data does not permit comment as to whether it is unusually so. Olafson and Thompson (1974) have reported on the isolation of metallothioneins from seal livers, and have sug- gested that the biosynthesis of such cadmium-binding protein acts primarily as a detoxification mechanism in these, as in terrestrial mammals. Assuming that these complexes serve no other metabolic function, then their presence implies toxic ambient marine levels of these metals (notably cadmium, but also mercury and zinc). The mean kidney content of cadmium for all the analysis samples is around 24 ppm dry weight: a "critical" liver concentration in man has been esti- mated at around 200 ppm. Kerfoot and Jacobs (1976) have echoed earlier models in which a daily intake of some 50 jug/day of cadmium will produce the observed mean body burden of around 30 mg, which would approximately correspond to a critical organ concentration of around 50 ppm. Esti- mates such as these assume a long residence time for cadmium in the mammalian tissue: this is one of the chief health hazards of this metal. There appears to be no correlation of liver and kidney contents of metals with the age of the animals given in this report, however. In the muscle tissue, contents of these metals, especially zinc, may increase with age; but, in general, the assumption of relatively short residence times with varying organ contents reflecting recent eating habits is attractive. (Note also that contents of cadmium in livers, for example, of the May samples are generally lower than in samples taken in April.) We have here, however, far too few individuals of any one species or age group to permit anything approaching a rigorous statistical analysis. TABLE 21-2 Bering Sea O.S.S. Discoverer 25 May-June 1977 Seal samples collected for heavy metal analysis Sample Latitude Longitude Species Sex Weight Age # (N) (W) (kg) (yrs) 1 60°37.7 174°27.2 Ribbon M 80.0 10 2 60°36.3 174°37.5 Spotted F 57.7 10 3 60°36.3 174°37.5 Spotted M 45.5 0.3 5 60°36.3 174°37.5 Spotted M 49.1 2 7 Ribbon M 73.6 7 8 60°26.5 168°55.8 Spotted F 55.0 4 10 60°24.2 169°49.8 Spotted M 68.6 5 13 60°35.9 168°10.1 Ringed M 8.7 1 15 60°56.6 170°48.3 Spotted F 37.7 <1 TABLE 21-3 Bering Sea O.S.S. Surveyor 31 March-27 April 1977 Heavy metal contents of seal tissue (Mg/g dry weight) Sample Species Tissue Cd(a) Ni(a) Cu(a) Zn(a) 01 Rib muscle 0.13 + 0.01 2.5 + 0.4 6.7 + 0.5 37 + 12 liver 6.4 + 0.1 2.6 + 0.9 16.5 ± 0.2 169 + 7 kidney 53.0 + 0.3 2.8 + 0.2 19.4 + 0.1 149 + 1 02 Rib muscle 0.3 (c) 1.3 (c) 4.4 + 0.6 50 + 0 liver 8.7 + 1.2 1.3 (c) 16.5 + 3.5 8 + 2 kidney 20.1 + 4.5 0.5 + 0 16.5 + 3.5 102 + 22 03 Rib muscle 0.25 + 0 8.3 + 1.0 5.5 + 0 54 ± 2 liver 6.4 + 0.1 2.2 + 0.4 26.9 + 1.3 140 + 15 kidney 34.5 (c) 3.0 (c) 16.0 (c) 98 (c) 04 S muscle 0.14 + 0.01 3.0 + 0.2 6.7 + 1.3 51 ± 9 liver 0.4 + 0.1 2.5 + 0.2 25.0 + 2.5 116 + 51 kidney 16.7 + 0.2 2.4 + 0.4 44 + 13 113 ± 23 05 B muscle 0.8 + 0.1 2.8 + 0.3 7 + 2 175 + 25 liver 21.0 + 0.5 1.3 + 0.1 36.5 (c) 182 + 2 kidney 16.0 + 2.0 2.0 + 0.1 28.5 + 0 163 + 3 06 Rib muscle 0.47 + 0.01 5.8 + 0 6.8 ± 0.8 18 + 6 liver 11.4 + 0.1 0.9 + 0.2 29.0 + 2.5 100 + 10 kidney 33.4 + 0 1.2 + 0.2 17.5 + 0.3 48 + 0 07 B muscle 0.57 + 0.04 2.8 + 0.2 <5 (b) 147 + 20 liver 22.2 + 1.3 0.8 + 0 22.8 + 0.2 170 + 17 kidney 22.5 + 2.0 1.3 + 0.3 40 + 6 160 + 15 08 B muscle 1.26 ± 0 1.2 + 0.1 7.6 + 0 40 + 1 liver 41.1 + 0.9 0.5 + 0 44.1 + 2.4 87 + 3 kidney 17.4 + 0.1 1.5 + 0.1 28.1 + 0.6 110 + 1 10 S muscle 1.5 + 1.0 2.4 + 0.2 11 + 3 52.5 + 0 liver 0.6 + 0.1 8.1 + 1.6 18 + 4 213 + 7 kidney 44.3 + 0.7 <1.0 (b) 24.5 (c) 183 (c) 11 Rib muscle 0.24 + 0.02 <0.5 (b) 16.5 + 3.5 140 + 5 liver 35 + 15 <0.5 (b) 13.8 + 1.2 15 + 10 kidney 16 (c) <0.5 (b) 19 + 6 8 + 3 16 B muscle 1.13 + 0.07 4.8 + 0.2 6.9 + 0.6 60 + 10 liver 5.3 + 0.6 5.4 + 1.1 37.6 + 1.6 228 + 3 kidney 108.7 + 0.5 6.2 + 0.2 34.7 + 0.2 110 + 20 17 Walrus muscle 1.50 + 0.25 2.4 + 0.5 6.3 + 1.3 43 + 8 liver 26.0 + 1.0 1.6 + 0.3 41.6 + 0.4 99 ± 1 kidney 26.5 ± 1.0 <1.0 (b) 27.7 + 0 96 + 1 28 S muscle 0.11 + 0.01 <0.5 (b) 6.4 + 0.05 68 + 1 liver 4.5 + 0.05 <0.5 (b) 16.4 + 0.1 136 + 1 kidney — — — - - - - - 29 Rib muscle 0.70 + 0.03 <0.5 (b) 1.7 + 0 74 + 2 liver 17.0 + 0.3 <0.5 (b) 5.2 + 1.4 130 + 3 kidney 37 + 5 <0.5 (b) 5.2 + 1.3 92 + 14 30 S muscle 0.16 + 0.04 <0.5 (b) 4.0 + 0.1 133 + 0 liver 7.3 + 0.2 <0.5 (b) 18.5 + 0.2 160 + 0 kidney 34 + 5 <0.5 (b) 15 + 2.5 120 + 17 32 S muscle 0.17 + 0.01 <0.5 (b) 1.1 + 0.1 62 + 9 liver 15.9 + 0.8 <0.5 (b) 17.1 + 0.9 125 + 23 kidney 49 + 7 <0.5 (b) 4 + 2.5 110 + 32 (a) mean of duplicate determinations (b) duplicate determinations (c) single determinations 343 1—1 CM tH 1— I CSI JOB '53 Qj IDX) 3 ^^ 1-3 Ol •^ CO CO TO U5 M '^ Z- CUD l^ =« o ."2 c o c a o M "cS O q CO lO CDiOC~ o6<^(N t>-*CO tr-'oco CMi-HCO COCXJCO lOlMC^ O^CO CDlOOO CT>oO(T3 ^-||^^(^J ooi-HOi i— i>-i ^ooo CO lO m lO O O T-H V o o o VV V LO LO UO o o o VV V O 03 +1 +1 '^^ o o 'r^ •r^ 03 CM CM CO 00 CO ^ lo 05 LO CM 00 O CD 05 LO LO LO lO o o o V V V o o o 00 CM O 0> LO ^. °°. '^ cm' t-I CO o o o V "* T-H 00 th cm' CO LO lO lO o t-' o V -^ V T-H Tj< o o LO o "* T-H T-H O CM o -^o o o o O O O +1 +\ +1 +\ -H +1 +i +: O CO o V CM O V o o o V CM lO_ O CO LO xj c ^H c Si c 3 > T3 2 3 E > ■a 3 > s 2 3 ^ -O 55 i- C 3 ^ TJ 5 >- C 3 g T3 £ = 2 W3 ^ C 3 ^ TO £ ~ S ^^ c 3 > •o £ M t/3 o CO c F ty2 c 0) o T3 CO c c o f^ CO £ CO •3 u a> c D. HI £ 3 XI -o CIJ o CO C) T3 c OJ CO a. en £ 3 C<1 o i 344 Heavy metal contents of seals 345 TABLE 21-5 ACKNOWLEDGMENTS Bering Sea Marine mammals analysis program— precision and accuracy (/ig/g dry weight ± one standard deviation) a. NBS Standard #1571 orchard leaves Element n This study NBS certified Cd 3 0.15 ± 0.06 0.11 ± 0.02 Ni 4 1.4 ± 0.03 1.3 ± 0.2 Cu 7 10.5 ± 3 12 ± 1 Zn 7 25 ±5 25+3 b. NBS Standard # 1577 bovine liver Element n This study NBS certified Cd 3 0.31 ± 0.07 0.27 ± 0.04 Ni 4 0.9 ± 0.08 — Cu 3 150 ± 30 193 ± 10 Zn 8 120 ±20 130 ± 10 TABLE 21-6 Heavy metal distributions in seal tissue from Bering Sea-Spring 1977 Cadmium 1. Concentrations in the kidneys of ribbon seals greater than in spotted or bearded 2. Higher cadmium contents in the muscle tissue of bearded seals than in ribbon or spotted 3. Liver contents of spotted seals relatively low 4. Liver contents of bearded seals relatively high Nickel This work was supported by BLM/NOAA Contract No. 03-5-022-56. The soluble copper and lead data were determined by D. T. Heggie. F. Fay and L. Shultz collected the seal samples. T. Manson and D. Weihs performed the chemical analyses. This is Institute of Marine Science Contribution No. 406. REFERENCES Burrell, D. C. 1978 Annual report to OCSEAP. Unpub. Rep. Inst. Mar. Sci., Univ. of Alaska, Fairbanks. Kerfoot, W. B., and S. A. Jacobs 1976 Cadmium accrual in combined waste- water treatment-aquaculture system. Environ. Sci. Tech. 10: 662. Olafson, R. W., and J. A. J. Thompson 1974 Isolation of heavy metal binding pro- teins from marine vertebrates. Mar. Biol. 28: 83. Muscle contents generally higher than liver or kidneys Copper Zinc 1. Spotted seal kidneys generally higher than liver, but reverse trend for ribbon and bearded seals Muscle contents of bearded (and possibly spotted) seals higher than ribbon Concentration of zinc higher in livers than in kidneys of all species I I Preliminary Observations of the Carbon Budget of the Eastern Bering Sea Shelf Donald W. Hood Friday Harbor, Washington ABSTRACT Preliminary studies of the CO2 system of the PROBES area of the Bering Sea shelf in May of 1976 and May and June of 1978 have shown partial pressures of CO2 in the surface water as much as 250 juA (ppm) less than overlying air, which av- eraged about 329 ju A (ppm). These low pressures are evidently produced by photosynthesis, but their persistence long after nutrients are depleted and photosynthesis is low indicates a large sink for CO2 under earlier bloom conditions accom- panied by limited respiration in the water column which would recycle the fixed organic carbon back to the inorganic carbon pool. If transfer of CO2 through the sea surface from the atmosphere is on the order of grams of carbon per day, many months would be required to reach equilibrium without the help of respiration, which apparently occurs later in the season than the field work to date. A high correlation between deficiency of CO2 in the water with respect to air (— APCO2) and nitrate was found when values of CO2 were greater than — 130 mA. When CO2 was be- low — 130 mA, nitrate tended to be zero while PCO2 was found as low as —230 txA. Total CO2 measurements made during the phytoplankton bloom period (May-June 1978) gave values significantly lower (1.5-1.8 mM) than those found under non-bloom conditions (2.00 mM). The CO2 represented by the difference in total carbon dioxide between bloom and non-bloom conditions is apparently being held in the system as fixed carbon, either in the form of detritus avaUable for consumption, or as the portion of primary production which is stored in the body tissues of flora and fauna. INTRODUCTION It is well known from many previous observations (Park et al. 1974, Gordon et al. 1973, Kelley et al. 1971, Park et al. 1958) that the PCO2 of euphotic regions of the ocean becomes depressed with respect to the overlying air under conditions of high primary productivity. Although photosynthesis is the only known sink for molecular carbon dioxide, the carbon dioxide respired by mammals and birds directly into the atmosphere is an effective sink which may be of importance in some regions. The supply of CO2 to the euphotic zone during the season of maximum primary productivity is derived primarily from flux through the sea surface from the 347 overlying air, although additions from vertical trans- port from deeper water, respiration, and horizontal convection may contribute heavily at other seasons and in other circumstances. It is clearly established, however, that under positive net carbon fixation, stability in the water column structure conducive to phytoplankton growth limits vertical transport; and transport of properties by horizontal convection becomes negligible if surrounding waters are similar to the study area. Thus, under conditions of high primary productivity, the product of the difference in PCO2 (in fxA) between air and surface water and the exchange rate of CO2 through the surface is directly related to the net photosynthesis (see equation 1). In the carbon budget net photosynthesis is represented by the sum of the flux of CO2 through the sea surface and the decrease of total inorganic CO2 in the water column, corrected for the direct precipitation of carbonate salts (e.g., carbonate exoskeleton formation). If there is vertical transport, as in regions of up- welling, the stability of the water column is disrupted and the deep water rich in carbon dioxide reaches the surface. This water, which has been enriched with molecular carbon dioxide through biological pro- cesses at depth, is supersaturated with respect to air, and a net flux of CO2 from the water to the air oc- curs. Hood and Kelley (1976) have used a measure of this flux to estimate vertical transport. Carbon dioxide, a fundamental component in all metabolic processes, is closely coupled to many chemical and physical events in the ocean. It is easy to monitor the partial pressure of carbon dioxide in the surface water continuously at sea; this is a power- ful qualitative tool for mapping the ocean for such features as upweUing, biological activity, currents, and freshwater influx. Moreover, since the concen- tration of all the components of the carbon di- 348 Chemical oceanography oxide system in sea water is sensitive to most oceano- graphic events, a detailed understanding of carbon dioxide dynamics in the environment will provide ad- ditional signals useful in evaluating the initiation, path followed, and fate of some otherwise poorly understood elements of the ecosystem. This study of the carbon budget was undertaken in the eastern Bering Sea to assist the PROBES (Pro- cesses and Resources of the Bering Sea Shelf) project in understanding biological energy flow and trophic dynamics in the important biological resource area often known as the "golden triangle." The base of this triangular region extends from just north of Unimak Pass running east to the 80-m bathymetric contour. The two sides join near St. George Island, forming the triangle. The PROBES project seeks to follow the sequence of events involved in energy flow and efficiency of food utilization in the biota at the lower trophic levels of the biological community by following the early life cycle of the walleye pollock (Theragm chal- cogramma), in the hope that it will function as a tra- cer of other organisms in this trophic level. This study requires the detection of the early changes in the system that indicate the conditions for onset of zooplankton spawning, phytoplankton productivity, and the hatching of the widely scattered pollock eggs (see chapters by Hattori and Goering, Cooney, and Nishiyama, in Volume 2 of this book). More conven- tional methods for following ecosystem changes that affect the development of the biological community- monitoring temperature, light, nutrient regime, water- column structure, and the presence of biological com- munity components— may not be adequate to define the finestructure on which these processes depend for initiation. Data on the changes in the carbon dioxide system that occur during the course of ecosystem development may provide additional information needed to explain how the system functions. The carbon dioxide cycle To do budgetary studies of carbon in the ma- rine environment it is first necessary to determine the pathways and fates of carbon dioxide in the carbon cycle. This compound is required for photosynthesis, is one of the products of respiration, and is a compo- nent of the sedimentary process in the form of car- bonates, shells of organisms, or carbon in organic detritus. The last may provide energy for the benthos long after primary production has ceased, be buried to enter later into the longer geochemical cycle, or perhaps even be transported far from its place of ori- gin by ocean currents to play a significant role in the global carbon cycle. Fig. 22-1 is a simplified carbon Figure 22-1. Simplified flow of carbon in the ecosystem. The vertical shaded bar represents the division between the inorganic and organic part of the carbon dioxide system. cycle, showing only the essential features. The cycle can be divided into several compartments: gas ex- change through the sea surface as controlled by Henry's Law; the dissolved inorganic components as controlled by the chemical equilibrium of the carbon dioxide system, including the carbonate minerals; photosynthesis or primary production, the major mechanism for carbon dioxide transfer from the in- organic to the organic carbon compartment as con- trolled by physiological requirements of plants and the chemical and physical structure of the water column; respiration, which returns organic carbon consumed by marine biota to the inorganic system as CO2, as controlled by biological energy transfer processes; formation of detritus as derived from the excess of primary production over consumption, fecal pellets, and animal remains (much of this detritus settles to the bottom as the organic component of particulate matter and provides food for the benthic biomass or enters into other sedimentary processes); and, finally, the alkaline earth carbonate secretions and precipitates which may form long-term deposits of carbon dependent upon solubility equilibrium rela- tionships for recycling to the inorganic pool. Since until now I have had only two opportunities to make carbon dioxide measurements in this region- one in May 1976 and one in June and July of 1978— the results reported here are preliminary in nature. And yet these observations have revealed some facts and raised some questions that should be reported now. For 1980 and subsequent seasons, more defini- tive and extensive studies are planned which it is hoped will permit a more rigorous evaluation of the The carbon budget 349 carbon budget, particularly that of the inorganic system. METHODS Samples taken on the Acona cruise in 1976 were stored and frozen, and nutrient determinations were made by Technicon Autoanalyzer techniques at the Seward Station of the Institute of Marine Science. On the T. G. Thompson cruise analysis was made aboard ship by the five-channel Technicon Auto- analyzer using the methods of Patton and Whitledge (PROBES Progress Report 1978). The techniques for pCOg measurement in surface water have been described earlier (Ibert and Hood 1963, Gordon et al. 1973, and Kelley et al. 1971). During these experiments a stream of seawater pumped from a bow intake system at a rate of 10-20 1/min. was split so that 5 1/min. passed through an equilibrator system to provide a gas sample in equili- brium with the sea water phase (Hood and Kelley 1976). The large volume of sea water minimized tem- perature excursions of the sample to less than +1 C as it passed through the ship to the laboratory. Three to four liters of this water were run through the multi- phase equilibrator (Ibert and Hood 1963, Hood and Kelley 1976) to assure complete equilibration with the gas stream without distorting the CO 2 equili- brium of the water. The CO2 concentration in the dry air and in the dry air after equilibration with sea water is expressed as a volume fraction (mixing ratio) in ppm by volume when the total pressure exerted by the air passing through the infrared analyzer is at one atmosphere. In an approximation sufficiently ac- curate for this investigation, the barometric pressure was assumed to be one atmosphere. The concentra- tion of CO2 in the air phase was measured by relating the signal of the output of the dispersed beam IR analyzer (Beckman model 310) to that obtained from two standard gases (one above and one below analysis values). The standard gases were referenced to those of Keeling of Scripps Institute of Oceanography, La JoUa, California, accurate to ±0.2 ppm. The accuracy of the technique employed here, on the order of ±5 ppm, could be improved easily by eliminating some technical problems of instrument stability and the difficulty of reading concentrations from strip-chart recorder records. The carbon dioxide exchange rates of CO2 in- vading the sea surface were determined by using a free exchange curvette technique similar to that of Sugiura et al. (1963) and further described in Hood and Kelley (1976). The transfer rates obtained by this technique are probably on the low side since the direct effect of wind on the sea surface is influenced by the curvette cover. Since actual transfer is a mo- lecular diffusion process through an air-water film and a water-water film, the rate is probably deter- mined by the liquid boundary, according to the theo- ry of Bohr (1899). The experimental design of the measurements made here attempts to avoid influ- encing water turbulence and thus the thickness of the water-water layer. If results obtained by this method are indeed low, how low remains to be determined (see Skirrow 1975 for further discussion). Measurements of pH were made according to the method of Smith and Hood (1964) using a single probe calomel-glass electrode and a Coleman null- point pH meter. Tris-buffer made up in sea water and adjusted to the temperature of samples in a water bath kept at sea-surface temperature was used as a pH standard. Total CO2 measurements were made by infrared measurement of the quantity of CO2 released from acidification of 3 ml of a sample of sea water. Both peak height and peak integration techniques were used to estimate the concentration as compared to a sodium carbonate standard. Bottled nitrogen was used as the carrier gas to sweep the generated CO2 from the sample through the IR analyzer. A preci- sion of ±3 per cent, based on replicate analysis, was obtained with this technique. RESULTS AND DISCUSSION The salinity, temperature, and density profile of a typical station in the study area is shown in Fig. 22-2. The detailed data for this station, including oxygen, nutrients, and pH, are shown in Table 22-1. The PCO2 pressure of the surface water (2 m) was 162 m A and the air was 332.6 //A to give a ApCOg of -170.7. The stability of the water column was apparently suf- ficient to support primary production in the euphotic zone (above 35 m) as indicated by the low nutrients, high pH, and low PCO2 , but this zone was underlain with unusual finestructure with water properties gradually changing to those expected below a well- defined pycnocline. Kinder and Schumacher (Chap- ter 4, this volume) have reviewed the hydrography of this region ; they observed that the outer Bristol Bay transition zone (PROBES study area) contains three ocean fronts between which the horizontal salinity gradient is near zero. The outer front, centered over the shelf break, has a horizontal salinity gradient of about 9.4 X 10"^ g/kg/km, similar to that of the mid- dle front, centered over the 100-m contour. The inner front is at about the 40-m contour and is typi- fied by a homogeneous vertical water mass; waters inside the front are heavily influenced by coastal 350 Chemical oceanography Temperature (° C) 0 1 2 3 4 5 6 Q. Q 23 0= 30 0 31 Salin 32 ity (O/oo) 33 34 35 - ' < [i: 1 60 - 1^\ 90 - \\\ 20 en - \ 11 T SO, I 24 25 26 27 28 Sigma-t Figure 22-2. Temperature, salinity profile of typical station (No. 28, Cruise No. 226, R/V Acona, May 1976) in PROBES study area. processes. The fronts appear to act as flux boundaries across which only limited amounts of materials are exchanged. The finestructure in the deeper water be- tween the middle and outer fronts consists of layers of waters of significantly different sigma-t values that originate from the other fronts and penetrate the in- ner area as fingers. The significance of this phenome- non in providing the euphotic zone with nutrients has recently been examined by Coachman and Walsh (1980). Its importance to the CO2 budget is being examined during the 1980 field season. As seen in the results of the surface pCOg distribu- tion study for May 1976 (Fig. 22-3), in the region of the shelf north of 55° and east of the 200-m contour the water was depleted of CO2 by more than 70 pi A and in most cases was more than 170 fiA lower than the air. In general, west of the 200-m contour and south of 55° the PCO2 of the water was within ±20 M A of that of the air. While the surface water of the region just north of Unimak Pass had slightly higher PCO2 values than the air, which may indicate up- stream upwelling or mild vertical mixing in the pass, no values on the order of +200-300 nA, typical of waters welled up from below the discontinuity layer, were observed (Hood and Kelley 1976). To compare the surface PCO2 values with other parameters measured in the euphotic zone, average values obtained to a depth of 35 m in a cross section of the widest part of the shelf at 55°50'N were plotted (Fig. 22-4). The observations made by Coachman and Charnell (1977) found the middle front well developed at the 100-m isobath and the outer front at about the 150-m isobath. The data shown in Figs. 22-3 and 22-4 for PCO2 and nutrients, particularly nitrate, from the 1976 cruise clearly reflect the outer front but show no clear indication of the inner front. A similar situation prevailed for 1978 (Fig. 22-5); a slight increase in surface PCO2 was seen on the eastern side of the front at a depth of approximately 100 m. Perhaps if data had been collected further shoreward the influence of the inner front would have been more evident; however, it is apparent from these data that the productivity regime which influences nutrient and pCOg values occurs throughout the shelf domains. TABLE 22-1 Depth Temperature Salinity Oxygen PO4 NH3 NO3 SiOg m °C o/oo Sigma-t ml /I MgA/1 MgA/1 MgA/1 MgA/1 pH 5 2.69 32.148 25.67 9.87 0.40 0.3 0.0 15 8.44 10 2.62 32.153 25.68 9.24 0.44 0.3 0.0 17 8.40 15 2.60 32.157 25.69 20 2.57 32.154 25.69 9.09 0.56 0.3 0.0 16 8.42 25 2.52 32.152 25.69 30 2.47 32.152 25.69 9.08 0.44 0.3 1.7 16 8.37 35 2.42 32.120 25.67 40 1.85 32.120 25.71 45 2.06 32.363 25.89 50 2.07 32.560 26.05 8.45 0.86 1.2 6.2 18 8.36 75 2.67 32.563 26.01 8.17 0.94 1.7 8.3 22 8.32 100 2.81 32.872 26.24 6.79 1.72 1.1 23.6 51 8.18 The carbon budget 351 169° 165' 161° 56^ 55^ 54^ 169° 165° 161' Figure 22-3. Distribution of pC02 in surface water in PROBES study area, May 1976. Nitrate-chlorophyll-pC02 relationships A plot comparing the average nitrate values of the top 35 m of the 1976 study area with PCO2 is pre- sented in Fig. 22-6. The circled dots in the top figure show zero nitrate values at the surface where PCO2 was measured, but higher values at depth. The clus- tering of points of near-zero nitrate values at ApC02 values of less than —135 juA in the top figure, which represents 1976 data, and —110 juA for the middle figure, which represents 1978 data, is of considerable physiological interest. These data indicate that the photosynthesizing plants are capable of removing all the nitrate from the water column relatively uninflu- enced by the concentration of molecular carbon dioxide. It may be that other forms of CO2 (i.e., HCO3 ) are being utilized; if so, the stoichiometric ef- fect on the PCO2 is the same as for direct utilization (Fig. 22-1). The chlorophyll a content appears to bear little direct relation to PCO2 content, probably because these two physiologically important compo- nents are functioning on different time scales. A high chlorophyll content is associated with a low PCO2, but the recovery time for CO2 is slow (because of limited atmospheric exchange capacity and depen- dence on respiration of fixed carbon for recovery) and the photosynthesis which caused its decrease had ceased to occur long enough before the analysis was made that the plants and their chlorophyll a had disappeared. The utilization of ammonia, present in 5. H "> 170 W O 3 32.5 -70 2 32.0 -140 1 31.5 -200 O 31.0 55 12 8 46 44 42 41] Station No. o' o a z (m9 A/i) 1.2 12 48 1.0 10 40 0.8 8 32 (0 o 0) 0.6 6 24 ^ 0.4 4 16 3 Figure 22-4. Cross section 55°50'N eastern Bering Sea shelf in June 1976. All parameters av. 0-35 m depth except ApC02 , which is 2 m. 352 Chemical oceanography 100 fm 50 fm 0 * t • -100 38 • 39 . "* • 4*3 Leg 5 40 42 • 44 45 46 -200 [ t 54°39.7N t 56°15.0N 166°24.6 W 164°49.9W 100 fm 50 fm 0 * ♦ -100 37 3*6 3*5 • 34 . 33 3*2 • 31 • • 30 29 Leg 4 -200 SS^SO.ON T 56°24.QN 167°10.1'W 165''35.2'W < 100 fm 0 * 50 fm 6 o Q. § o -100 -200 • • 19 18 • t ^° 55°01.9'N 167°50.0W • 21 • 22 • ^"^ 23 • 24 « 2-e . 27 Leg 3 • 28 t 56°59.1N 165"'54.0'W o a II O o a < 0 -100 -200 100 fm 50 fm t ■ '' 1*6 * t 15 55°33.8'N *. • • 14 . 13 12 • 11 • • 9 10 ^ 56°56.0'N Leg 2 168°10.4'W 1 66''54.0W +100 100 fm 50 fm 0 t t • -100 • 2 ( • 3 • • 4 5 • 6 • 7 f Leg 1 -^°° 55°5'l.0N 57°08.6'N 168°48.0'W 167°36.0W (statiofi numbers indicated by numerals) Figure 22-5. Distribution of PCO2 in surface water in five cross sections of PROBES study area in 1978. the euphotic zone in concentrations of 0.1-1.0 )UgA/l, is associated with high chlorophyll a, but the biolog- ical processes which released the ammonia from organic matter would be expected also to release carbon dioxide to the water column in about the same proportion as it was fixed by photosynthesis. The PCO2 data, as well as all other data (see Hat- tori and Goering, Volume 2), suggest that nitrogen limits phytoplankton growth in this system and that the recovery of CO2 concentrations is relatively slow compared to the dynamics of photosynthesis. Main- tenance of the low PCO2 values for the periods found in these studies must be related to a slow resupply of CO2 to the euphotic zone, controlled by the ex- change rate of CO2 through the sea surface, vertical transport or respiration. To maintain low PCO2, then, photosynthesis must be supported by nitrate, not ammonia, through resupply from below the discontinuity layer. Seasonally, of course, nitrate is resuppHed from the mixing associated with thermo- haline convection. +60 +90 +90 -\ 10 O) E (0 £ s 0 -210 -180 -150 -120 -90 -60 +30 +60 +90 A PCO2 at 2 m depth (ppm CO2) Figure 22-6. Relation between nitrate-nitrogen concentra- tions and difference between water and air PCO2 values. Top: data for 1976 R/V Acona cruise. Middle and bot- tom: data for 1978 Leg III R/V T. G. Thompson. Total carbon dioxide During the Thompson cruise in 1978, a preliminary investigation of the total carbon dioxide in the water column was undertaken. The vertical distribution at three stations in the PROBES area is shown in Table 22-2. Station 3016 was in the outer frontal area (55°33.8'N, 168°10.4'W), station 3044 (55°53.0'N, 165°13.l'W) in the central domain, and station 3086 (56°56.8'N, 166°54.0'W) in the middle frontal area in a region of high chlorophyll a. All the values found were low compared to the concentration of 2.05 mM normally found in surface sea water during unproductive periods in this region of the world's ocean (Park et al. 1974). Station 3044 was in an area The carbon budget 353 TABLE 22-2 Total CO2 at stations 3016, 3044, and 3086 mMCOa/l Depth Sta. 3016 Sta. 3044 Sta. 3086 0 1.94 1.45 1.63 5 1.96 1.37 1.61 10 1.85 1.35 1.63 15 1.93 1.50 1.64 20 1.91 1.54 1.58 25 1.85 1.61 1.59 30 1.99 1.64 1.68 40 1.94 1.69 1.66 50 2.05 1.75 1.70 75 1.95 1.71 1.65 100 1.99 1.74 — of low nutrients in which the primary production peak had already passed. The surface values found were significantly lower than at station 3086, in an active primary production area. At station 3016, in the outer frontal zone, a higher level of total carbon dioxide appears to be retained. Few conclusions can be drawn from these limited data regarding the distri- bution of total CO 2 on the eastern Bering Sea shelf, but it is clear that the quantity of total CO 2 varies with environmental conditions and ecosystem acti- vity. It is also clear that the loss of approximately 450 g of inorganic carbon from the water column be- tween the outer front, suggested to be at near pre- bloom concentrations of total carbon dioxide, and the middle domain, representing conditions after the spring bloom, must be accounted for. (See Kinder and Schumacher, Chapter 4, this volume, for location of these domains.) A detailed study of the total CO2 in the PROBES area has been undertaken during this 1980 field sea- son (Hood and Codispoti, personal correspondence) to clarify the questions raised here concerning the loss of total carbon dioxide and relate its chemistry to the ecosystem processes of the shelf region. Diurnal changes On one occasion during the Acona cruise in 1976, a 24-hour station was occupied at 167° 05' W, 55°8'N. Measurements of PCO2 were made at 30-minute in- tervals except for two-to-three-hour periods when the equipment was being used for measuring the rate of exchange of CO2 through the sea surface. It was ex- pected that during the hours of darkness (between 2200 and 0300 local time at this latitude in May) the large negative ApC02 between water and air would be reduced by the continued flux of CO2 from atmosphere to the water and by respiration of the in- digenous organisms during a non-photosynthetic per- iod. Extreme diurnal shifts in PCO2 in the water were earlier observed by Park et al. (1958) over sea- grass meadows in Texas bays and by Kelley (1975) over eelgrass meadows in Izembek Lagoon, Alaska. The results of this study (Fig. 22-7) do not support the diurnal shift concept. The sudden increase in PCO2 which occurred between 0430 and 0500 was probably caused by an influx of a different water mass at this moment, as evidenced by an accom- panying drop in temperature, a slight increase in ni- trate, and a slight change in the fine spectrum of the water column (Nishiyama 1976). Moreover, a diurnal change would probably have been gradual rather than sudden and would probably have coincided with the hours of darkness rather than coming well after dawn. During the third leg of the T. G. Thompson cruise in May-June 1978 several stations were revisited after a time lapse of 9-13 days. The results of these obser- vations are given in Table 22-3. The first four paired stations listed (6, 7, 8, 9 and 98, 97, 95, 84) were in the northeast section of the outer shelf domain and across the middle front into the central shelf domain (see Kinder and Schumacher, Chapter 4, this volume). At the time of the first visit these stations all had sig- nificant quantities of nitrate-nitrite in the euphotic zone; this was depleted during the period between the visits, and yet only moderate decreases in PCO2 were observed at three paired stations while at the fourth (7 and 97) pCOg remained essentially the same. At stations 10, 11, 12, and later 83, 81, and 80, nearly zero nitrate-nitrite was observed on both visits and II -=-180 O © Average pH measurement s to 30m depth ^^— Ave Prot age pCOj able error of PCO2 measurement '^ _«_f_ ■"■■' .-■■ — ■ ■ J ■ ■ ■ ■ ® May 22 1976 I (local time) May 23 1976 Figure 22-7. Surface water PCO2 and pH values, averaged to 35 m depth, at a 24-hour station in PROBES area in May 1976. 354 Chemical oceanography TABLE 22-3 A comparison of ApC02 ^^ the same geographic locations after a time lapse of several days NO3+NO2 N03+N0£ Time ApC02 AtgA/1 Sta. Location Time ApC02 MgA/1 Elapsed time Sta. Location 10 11 12 16 18 56 38'9N; 168°03'0W 56°50'9N; 167°5l'9W 57°08'6N; 167°36'6W 56°56'0N; 166°54'0W 56°56'0N; 167°4'0W 56°3l'6N; 167°16'9W 56°2l'lN; 167°27'lW 55°33'8N; 168°10'4W 55°20'0N; 168°20'0W 0100 May 28 0330 May 28 0730 May 28 1200 May 28 1500 May 28 1830 May 28 2000 May 28 0731 May 29 1658 May 29 -110 -114 -76 -121 -142 -143 -144 -96 20.3 17.8 5.8 11.8 0.1 0.0 0.1 6.3 3.8 98 97 95 84 83 81 80 117 118 56 38'IN; 1100 168°05'0W June 8 56°5l'lN; 0800 167°5l'9W June 8 57°10'0N; 0430 167°33'0W June 7 56°56'0N; 0600 166°53'9W June 7 56°44'0N; 0300 167°3'9W June 7 56°3l'4N; 2300 167°16'9W June 6 56°2l'0N; 2000 167°27'8W June 6 55°34'0N; 1030 168°09'0W June 11 55°39'9N; 1130 168°13'lW June 11 113 -110 -88 -140 -136 -131 -121 -100 -108 0.0 0.0 3.0 0.0 0.0 0.0 0.0 3.6 1.9 11.5 da. 11.3 da. 10.9 da. 9.8 da. 9.5 da. 9.2 da. 9.0 da. 13.2 da. 13.0 da. after approximately nine days elapsed the PCO2 did increase, although not dramatically. Stations 16 and 18 and later 117 and 118 show a decrease in nitrate- nitrite as well as PCO2 after the 13-day time lapse. In general the PCO2 of the surface water tends to fall with the utilization of nitrate and nitrite (bloom conditions) and increase in waters with no nitrate and nitrite nutrients. It is clear, however, from these studies that a depressed PCO2 in the water column re- quires considerable time to regain equilibrium with the atmosphere. The slow recovery is directly related to respiration and the exchange of CO 2 between the atmosphere and the ocean, since these are the pri- mary sources of carbon dioxide to the ocean. Carbon dioxide exchange The environmental conditions found in late May 1976 and again in late May and early June 1978 pro- duced very low PCO2 values in the surface waters, perhaps the lowest ever measured in the open ocean. From our general knowledge of the carbon cycle these low values could only be produced by an excess of photosynthesis over respiration, the rate of which is related to the flux of CO2 into the sea surface. This rate is expressed by Henry's Law: F = aApC02A (1) where F is flux of gas in moles/m^ /day, a is the ex- change coefficient for CO2 through the sea surface in moles/m^ /day, ApC02 is the difference in atmos- pheres between the PCO2 of the water and the overlying air, and A is area in square meters. The exchange coefficient, a, is very difficult to measure in the field because of the changing sea-surface condi- tions and the requirement that the complex surface, active in exchange, not be disturbed by the experi- mental conditions. Direct measurements of a. have been attempted by Sugiura et al. (1963), Park and Hood (1963), and Hood and Kelley (1976). The The carbon budget 355 basic technique used in these three papers, reviewed by Skirrow (1975), employs a canopy-type capsule placed on the sea surface, allowing free movement of the water phase while the capsuled air phase is sampled for incremental concentration changes which are related to the exchange coefficient. The fact that the method limits the immediate wind effect on the sea surface probably lowers the exchange rate and thereby renders minimum values for the exchange rate a. Broecker and Peng (1974) have estimated the exchange rates for the world ocean to be 792 g/m^ / d/atm (1700 m/yr piston velocity) based on radio- active carbon distribution data. A summary of the exchange data reported in the literature is given in Table 22-4, along with an esti- mate of the flux in grams of carbon exchanged per day under Bering Sea conditions. The time required for the Bering Sea shelf waters to reach equilibrium with the atmosphere can be estimated if the exchange rates are known. Chosen for consideration are the conditions represented by the time series stations 10, 11, and 12 (occupied 8 May 1978) and 83, 81, and 80 (occupied 6 June 1978) (see Table 22-3). These six stations represent three pairs at the same geo- graphical locations occupied about nine days apart. The water conditions, including nutrients at near- zero concentrations, were essentially the same on both visits, indicating post-bloom status. The initial PCO2 of the surface water averaged 187 ixA or 143 //A below that of the atmosphere. The total CO2 de- TABLE 22-4 Exchange rate coefficients and flux of CO2 through the sea surface Reference a Flux* ,2 /J„,W„4^™\ «/™2 Media (g/m^/day/atm) g/m /day Bohr 1899 Suguira et al. 1963 Hoover and Berkshire 1969 Miyake and Hamanda 1960 Guyer and Tobler 1934 Seven-year atmospheric residence time Suguira et al 1963 (1) Suguira et al. 1963 (2) Station 3086, June 1978* Station 3118, June 1978* Station 3114, June 1978* Station 3114, June 1978* Lab 3740 0.60 Lab 4610 0.74 Lab 2160 0.35 Lab 290 0.05 Lab 94320 15.0 Fid 7920 1.3 Fid 28220 4.5 Fid 13100 2.1 Fid 11000 1.8 Fid 8500 1.4 Fid 3450 0.6 Fid 2020 0.32 *Used a PCO2 pressure difference between water and air of 160 juA, the average for the cruise period of 1978. ficiency in the water column is assumed to be the same as given in Table 22-2 or about 37.5 M of CO2 less than in pre-bloom conditions. At an estimated invasion rate of CO2 based on the 792 g/m^ /d/atm exchange rate and a pCOg of 143 //A, the flux rate is 1.1 g/mVd. The deficit in total CO2 for the 100 m of the water column from Table 22-2 is about 490 g. Since equilibrium will be approached in a logarithmic manner characteristic of the changing partial pressure, the time to reach one-half equilibrium would be 490 g deficiency divided by 1.1 g flux/d multiplied by 0.692 or 308 days, assuming that water chemistry remains the same. During the nine days between the observations at the time series stations, 10 g of CO2 (based on seven- year residence time) should have invaded the sea sur- face, representing a 2-percent recovery of the deficit in total CO2 . The change in measured PCO2 was from an average of —143 /iA at the beginning to — 130 juA after nine days, or a change of 8 percent, for part of which respiration must be responsible. Even considering the many assumptions and errors in experimental measurements, these data nevertheless indicate that short-term changes in the CO 2 system should not be expected since the rate of supply from the atmosphere is slow in comparison to the deficit in CO2 caused by the spring bloom and low rates of res- piration in the water column. The absence of a short- term shift in PCO2 which should be brought about by respiration must be the result of relatively small ani- mal populations and little microbial activity under these conditions, in comparison to the deficit in total carbon dioxide concentrations created by the earlier heavy phytoplankton bloom. Ultimately animals in the water column or in the benthos will consume most of the fixed organic car- bon and return the inorganic carbon components to the system, thereby greatly speeding up the rate at which equilibrium with the atmosphere is reached. Only fixed carbon lost to the sediments, stored in biomass, or transported from the area would be re- placed by CO2 from the air. Carbonate precipitation would also influence the balance but its importance has not been evaluated here since alkalinity measure- ments, useful for estimation of changes in carbonate ion concentration, were not available in these studies. Influence of birds and mammals The influence of the large numbers of birds and mammals which consume fixed carbon from the sea and release much of it directly to the air through res- piration, or to land and sediments through defeca- tion, can at the present time only be estimated, since population data for the PROBES area are not yet 356 Chemical oceanography available. It has, however, been estimated that 450,000 tons of marine mammals which gain 1 million tons of body weight per year inhabit the eastern Bering Sea shelf (PROBES 1973). If mam- mals are 10 percent efficient, then 10 million tons of food is consumed to achieve this gain in body weight. If food items are 30 percent solid matter and 50 percent of that is carbon then 15 percent of the food items or 1.5 million tons (1.5 X 10'^ g) of carbon or 5.5 million tons of carbon dioxide are respired by marine mammals annually. If one-half of this is transported directly to air or land, for each m^ of the shelf (1.22 X 10'^ m^ ) 4.5 g of COg is transferred in this way to be replaced by atmospheric carbon dioxide through sea-surface exchange. Birds effect carbon dioxide transfer in a similair way, but are estimated to have only 10 percent of the impact of mammals. Together these two groups may transfer 5.0 g C02/m^ /yr. By comparison, the com- mercial fish catch of about 2 million tons per year on the Bering Sea shelf removes only about 1 g CO2 /m^ / The carbon budget in estimating the flow of fixed energy The gross primary production in a system may be expressed as Pg = F-H (2002(0 - 2C02(t)) + R (2) where Pg is gross productivity; F is a A p CO 2 or flux of CO 2 through the sea surface, as given in expression (1); 2CO2 is total inorganic carbon in the water col- umn initially (i) and at time of computation (t) (i;C02 in both cases, here and in following computa- tions, must be corrected for carbonate precipitation, which is based on changes in alkalinity between initial and final time); and R is respiration. The net productivity by one definition Pn(i) is that production which occurs in excess of what is re- quired to maintain the photosynthetic apparatus, or: P„(i) = F + (2:C02(t))-Rp (3) where all terms of the expression are the same as in (2) except Rp, which represents respiration by plants and associated organisms. This value is probably not obtainable in the field, but must be measured by laboratory bottle experiments. This net productivity is that normally measured by the conventional bottle techniques used in primary productivity measure- ments. A second net production is represented by : P„(2) = F + (2C02(i)-2C02(t)) (4) This is that production which occurs in excess of res- piration of the total biomass of the water column at time t. The CO 2 component which is derived from respiration is effective in increasing total CO2 and in- creasing PCO2 of the water, thereby reducing the flux. To estimate the energy flow to respiration, data are needed for different times during the growing season, in which case: Rb = (Ft2 - Fti ) + (2C02t, - 2C02t, ). (5) Such data were not available from the 1976 and 1978 field season, but will be collected in future work. Some indication of respiration rate can be gained, however, from the exchange data for the time series at stations 10, 11, and 12 and 83, 81, and 80 (Table 22-3). During this period 10 g of CO2 was exchanged through the sea surface based on average exchange rates for the world ocean and the PCO2 of the water column increased by 8 percent. Since this percent increase is the same as that for total CO2 (Broecker 1974), the deficit of CO2 found at station 3044 of 1620 g/m^ would have changed by 130 g C02/m2 (35 g C/m^ ). This amount plus the 10 g of carbon exchanged gives an estimate for respiration for the nine days involved of 5.0 gC/m^/day. This value appears high and should not be used as an estimate of respiration for budgetary studies until more data are available and the method is refined by direct meas- urements of both exchange rates and total CO2 at the times of interest. It is given here only to indicate the potential of this technique in helping to understand energy flow in the ecosystem. Another net productivity Pn(3) that is more easily estimated and that is also valuable for energy flow considerations is the productivity or fixed carbon that remains in the system after grazing by animals in the water column. This might be considered stored energy to be used later by all higher trophic levels of the system. It is represented by the expression : P„(3) = EC02(i)-2C02(t) (6) and can be determined for any time when the total carbon dioxide content of the water column is known. For example, using data from Table 22-2 for station 3016 as 2C02(i) and station 3044 as 2:C02(t), the difference in integrated values for these two stations (Pn(2)) is 37 moles; about 450 g of fixed carbon/m^ is held in the water column for future utilization. To provide reliable data for this param- eter, total carbon dioxide measurements before and after the spring bloom at enough stations to allow es- timates of net productivity for the whole PROBES area will be needed. The partitioning of P^o) into compartments of energy for use by the various components of the biological community will require that the distribu- The carbon budget 357 tion of particulate carbon over the study area be de- termined. These numbers coupled with inorganic carbon studies will provide data on the amount of fixed carbon reaching the benthos as well as how much might be horizontally transported from the area. SUMMARY AND CONCLUSIONS The data reported here are of a preliminary nature and, except for PCO2 data in the surface water, do not permit us to draw reliable conclusions. It is well established that a large area of the eastern Bering Sea shelf annually becomes extremely depleted in gaseous CO2 with respect to the overlying air mass. This de- pression is associated with primary production and is accompanied by a decrease in nutrients and total car- bon dioxide components. Once the depression is established, recovery to equilibrium is slow. The defi- cit in total CO2 caused by the spring bloom (on the order of 450 gC/m^ ) can only be replaced by atmos- pheric exchange (on the order of 1.5-4.0 gC/m^/d) and respiration (on the order of 2-5 gC/m^/d) or between 1 and 2 percent per day. On the basis of the information obtained thus far, it is clear that understanding the carbon budget can be valuable in determining energy flow in the course of PROBES ecosystem studies and will lead to a much better understanding of the amount of carbon fixed in the system and where this carbon is going within the system. ACKNOWLEDGMENTS Some of the work described in this chapter was sponsored by the Institute of Marine Science, Univer- sity of Alaska, and the PROBES program, which is funded by the National Science Foundation, Division of Polar Programs, under grant DPP 7623340 to the University of Alaska. Special acknowledgment is given to Dr. John J. Kelley for providing standard gases and supplying much of the equipment for this work, and to Dr. Lou Codispoti, A. Hafferty, and G. Friederich for help with working up some of the data and reviewing the manuscript. REFERENCES Bohr, C. 1899 Definition und Methode zur Bestim- mung der Invasions- und Evasions- coefficienten bei der Auflosung von Gasen in Flussigkeiten. Ann. Physik u. Chemie, n.F., 68: 500-25. Broecker, W. S. 1974 Chemical oceanography. Brace Jovanovich, Inc. Harcourt Broecker, W. S., and Peng, T. H. 1974 Gas exchange rates between air and sea. Tellus 26: 19-35. Coachman, L. K., and R. L. Cham ell 1977 Fine structure in outer Bristol Bay Alaska. Deep-Sea Res. 24:869-89. Coachman, L. K., and J. Walsh 1979 A diffusion model of cross-shelf nutrients. PROBES Prog. Rep. VI: 38-105. Inst. Mar. Sci., Univ. of Alaska, Fairbanks. Gordon, L. I., P. K. Park, J. J. Kelley, and D. W. Hood 1973 Carbon dioxide partial pressures in North Pacific surface waters. Mar. Chem. 1: 171-98. Guyer, A., and B. Tobler 1934 Zur Kenntnis der Geschwindigkeit der Gasexsorption von Flussigkeiten. Helv. Chim. Acta 17:257-71. Hood, D. W., and J. J. Kelley 1976 Evaluation of mean vertical transports in an up welling system by CO 2 mea- surements. Mar. Sci. Comm. 2: 387- 411. Hoover, T. E., and D. C. Berkshire 1969 Carbon dioxide exchange between the atmosphere and the ocean. J. Geo- phys. Res. 74: 456. 358 Chemical oceanography Ibert, E. R., and D. W. Hood 1963 The distribution of carbon dioxide be- tween the atmosphere and the sea. Tech. Rep. 63-9-7, Dep. of Oceanogra- phy, Texas A & M Univ., College Station. Kelley, J. J. 1975 Dynamics of the exchange of carbon dioxide in Arctic and Subarctic re- gions. Ph.D. Dissertation. Inst. Mar. Sci., Univ. of Alaska, Fairbanks. Kelley, J. J., L. L. Longerich, and D. W. Hood 1971 Effect of upwelling, mixing and pri- mary productivity on CO2 concentra- tions in surface waters of the Bering Sea. J. Geophys. Res. 76: 8687-93. Miyake, Y., and A. Hamanda 1960 General Assembly Assoc. d'Oceano- graphie Physique, Union Geodesique et Geophysique Internationale. Hel- sinki. Ref. in Sugiura et al. 1963. Park, K. P., L. I. Gordon, and S. Alvarez-Borrego 1974 The carbon dioxide system of the Bering Sea. In: Oceanography of the Bering Sea, D. W. Hood and E. J. Kel- ley, eds., 107-47. Occ. Pub. No. 2. Inst. Mar Sci., Univ. of Alaska, Fair- banks. Park, K., D. W. Hood, and H. T. Odum 1958 Diurnal pH variation in Texas Bays, and its application to primary produc- tion estimation. Inst. Mar. Sci., V: 48-64, Univ. of Texas, Port Aransas. PROBES 1973 Processes and resources of the Bering Sea shelf. E. J. Kelley and D. W. Hood, eds. Pub. Info. Bull. 74-1, 23- 28. 1978 Progress report, Inst. Mar. Sci., Univ. of Alaska, Fairbanks. Skirrow, G. 1975 The dissolved gases— carbon dioxide. In: Chemical oceanography, J. P. Riley and G. Skirrow, eds. 2: 1-192. Academic Press, New York. Nishiyama, T. 1976 STD and water bottle casts. Cruise re- port, R/V Acona Cruise 227, Inst. Mar. Sci., Univ. Alaska, Fairbanks. Park, K., and D. W. Hood 1963 Radiometric field measurements of CO2 exchange from the atmosphere to the sea. Limnol. Oceanogr. 8: 287- 9. Smith, W. H., Jr., and D. W. Hood 1964 pH measurement in the ocean: A sea water secondary buffer system. In: Recent researches in the fields of hydrosphere, atmosphere and nuclear geochemistry. Sugawara Festival Vol- ume. Maruzen Co. Ltd. 185-202. Sugiura, Y., E. R. Ibert, and D. W. Hood 1963 Mass transfer of carbon dioxide across sea surfaces. J. Mar. Res. 21: 11-24. Organic Matter in the Bering Sea and Adjacent Areas N. Handa and E. Tanoue Water Research Institute Nagoya University Nagoya, Japan ABSTRACT Particulate matter collected in the Bering Sea and its adjacent areas during the cruises of R/V Hakuho Maru in the summers of 1975 and 1978 was analyzed for organic carbon and nitrogen and chlorophyll a. The concentrations of partic- ulate organic carbon and nitrogen were measured with the ranges of 16-420 MgC/1 and 1-85 MgNA, 19-185 MgC/1 and 1-25 MgN/1, 46-1,040 jugC/l and 6-160 /ugN/l, and 82-380 jugC/l and 16-70 MgN/1 in the Oyashio, the deep Bering Sea, the continental shelf, and the Qiukchi Sea areas respectively. In view of the concentration of particulate carbon and ni- trogen, the most productive areas in the Bering Sea are the Oyashio area and the transition areas of the continental slope to the continental shelf, where inorganic nutrient supply systems are well operated by the East Kamchatka Current and upwelling respectively. The ratios of particulate organic carbon to chlorophyll a were in agreement with these con- clusions. The particulate matter was also analyzed to determine the amino acid, monosaccharide, and fatty acid composition. Organic compounds such as essential amino acids, mannose and glucose, and saturated fatty acids are suggested as valuable tracers for analysis of vertical transportation of organic matter in the marine environment. Polyunsaturated fatty acids were implicated as useful diagnostic tools in determining the biolog- ical activity of phytoplankton in oceanic areas. Vertical variability of the organic composition is briefly discussed with relation to the biological degradation of particulate organic matter. INTRODUCTION The Bering Sea is one of the largest confined seas of the world ocean, bounded on the west and east by the land masses of Siberia and Alaska respectively and on the south by the Alaska Peninsula and Aleutian Islands arc (Shore 1966). It has been reported by several workers (Zenkevitch 1963, Hood and Kelley 1974) that the Bering Sea and its adjacent northern North Pacific Ocean are areas where some of the highest values of both primeiry and secondary bio- logical productivity have been observed in the world oceans. This is thought to be due to the abundant supply of inorganic nutrients provided by the influx of nutrient-rich water masses from the North Pacific Ocean into the Bering Sea and by extensive vertical mixing during winter. Advancement of the hydrography of the Bering Sea has been made by extensive studies conducted by several workers over the last decade (Favorite 1967, 1974; Ohtani et al. 1972). In the Bering Sea the warm eastern subarctic Pacific water is transformed into cold western subarctic water (Ohtani 1970). Seventy-five percent of the water entering the Bering Sea originates from the Alaskan Stream (eastern subarctic Pacific water) and the rest from water of the subarctic gyre (Favorite 1974). After entering the Bering Sea, water is involved in a large counter- clockwise gyral movement, which causes upwelling of deep water as observed in the temperature profile of the deep Bering Sea basin. Another remarkable characteristic of the vertical structure of water throughout the Bering Sea is the dichothermal water which occurs in a well-defined intermediate cold layer below the surface (Uda 1955, Takenouti and Ohtani 1974). The surface water is cooled in winter and a dichothermal layer of uniform salinity and low temperature is formed by thermal convection. The upwelling of more marine water associated with the counterclockwise eddies increases the salinity and temperature gradients underneath the surface layer. During the summer the temperature near the surface increases, but the warming does not reach to the bottom of the surface layer where the cold water remains as a dichothermal layer. Apart from the above studies, little is known con- cerning hydrographic features of the Bering Sea 359 360 Chemical oceanography which influence the distribution of organic matter. Only limited data are available on the distribution of dissolved and particulate organic matter in the Bering Sea. Loder (1971) and Hood and Reeburgh (1974) report some data for a region north of Unimak and Unalaska Islands in the eastern Aleutian arc and Nakajima (1969) gives data for a region of the eastern deep Bering Sea. The extent of horizontal and vertical variability in the concentration of dissolved and particulate organic matter is not well established. The aims of this chapter are: first, to show the distribution profiles of dissolved and particulate organic carbon and nitrogen in the Bering Sea and its adjacent areas, including the Oyashio, the deep Bering Sea, the eastern continental shelf, and the Chukchi Sea, in order to find the factors which control the regional variability of these organic elements; and, second, to describe the biochemical constituents of particulate matter to gain a better understanding of the biological and chemical aspects of the dynamic processes which control the distribution of organic material in these oceanic areas. MATERIALS AND METHODS Five to ten liters of sea water samples were filtered through a glass fiber filter (Whatman GF/C) with 47 mm diameter, precombusted at 450 C for four hours. The particulate matter collected on the filter was kept frozen at — 20 C until analysis. Before anal- ysis, the filters were allowed to stand in a chamber filled with HCl vapor for one hour to remove carbonate-carbon, after which they were dried in an oven for two hours at 80 C. Particulate organic carbon and nitrogen were determined by CHN- Corder, Yanaco Model MTS-2 carbon-nitrogen analyzer. Menzel (1966) proposed that dissolved organic matter was adsorbed on a glass fiber filter during the filtration processes. Thus, an estimate of the organic carbon and nitrogen adsorbed was conducted to correct the concentration of the particulate organic carbon and nitrogen. A sea water sample collected in the Mikawa Bay near Nagoya was filtered through glass fiber filters in various volumes separately. Each of the filters was analyzed for total organic carbon and nitrogen. The results obtained indicated that 38 )ug and 11 Mg of dissolved organic carbon and nitrogen respectively were adsorbed on each of the glass fiber filters during the filtration. The concentration of chlorophyll a was determined by the fluorometric method described by Yanagi and Handa (1970). The fluorometric method (Udenfriend et al. 1972) was used for the determination of particulate amino acid after hydrolysis of combined amino acid with HCl (6N) at 110 C for 24 hours. This method was standardized on the composite sample of amino acids consisting of a bulk of Chlorella vulgaris (Fowden 1954). Content of carbon and nitrogen in amino acid was calculated by multiplying the amino acid value obtained by 0.475 and 0.139 respectively. The amino acid composition of the particulate samples collected at stations 11 and 33 was deter- mined by the amino acid analyzer, Hitachi Model KLA-5. Particulate matter collected from 40-50 1 of sea water was collected on the glass fiber filter by filtration. The glass fiber filter was treated with 10 ml of HCl (6N) at 110 C for 24 hours to hydrolyze combined amino acids. The hydrolysate was evap- orated to dryness at 40 C with a rotary evaporator. The residue was dissolved in 0.05N HCl, and applied to the chromatographic column. Acidic and neutral amino acids were separated by the column (9 mm in outside diameter, 250 mm in length) packed with Hitachi custom resin #2618 and eluted with sodium citrate buffer (0.2N), pH 3.25 and then pH 4.25. Basic amino acids were eluted from the column (9 mm in outside diameter, 100 mm in length), packed with Hitachi custom resin #2618, and eluted with sodium citrate buffer pH 5.28. For the determination of fatty acid composition, 400-450 1 of sea water were collected from various depths from the surface through 2,000 m and filtered through a precombusted glass fiber filter (Whatman GF/C) (285 X 420 mm). The glass fiber filter with particulate matter was cut into small pieces, which were transferred into an Erlenmeyer flask. To the flask was added chlo- roform-methanol (2:1 v/v, 20 ml) and then a cer- tain amount of methyl heneicosanoate (C2i:o) (usually 10 Mg)- The mixture was stirred vigorously under nitrogen at room temperature overnight. After removal of solid residues by filtration through a glass fiber filter, the filtrate was reduced to a small volume under reduced pressure by the rotary evaporator at 40 C. Distilled water (20 ml) was added and lipid materials were extracted with petroleum ether (45-60 C b.p., 20 ml) four times. The combined extracts were dried over anhydrous sodium sulfate and evap- orated by the rotary evaporator and then dried in vacuo. To the dried material was added benzene (1 ml) and 0.5N KOH in absolute methanol (4 ml). The solution was refluxed under nitrogen at 100 C for two minutes to saponify lipid materials. Free fatty acids generated were methylated with 14 percent BF3 in absolute methanol (5 ml) at 100 C for two minutes (Metcalfe et al. 1966). After cooling, distilled water I I i Organic matter 361 (10 ml) and then petroleum ether (10 ml) were added to the reaction mixture successively and methyl esters of the fatty acids were extracted with petroleum ether four times. The combined extracts were washed with distilled water three times and dried over anhydrous sodium sulfate. The fatty acid methyl esters were cleaned up by thin-layer chromatography (TLC) using siUca Gel G (Merck, type 60) and petro- leum ether/ethyl ether/acetic acid (95: 5:1, v/v) as a solvent system for the development of the TLC plate. After extraction of the methyl esters from the silicic acid with chloroform -methanol (9:1 v/v, 10 ml), the extract was dried over anhydrous sodium sulfate and evaporated to dryness under reduced pressure. The fatty acids methyl esters were dissolved in ethyl ether, and a few microliters of the solution were injected into the gas liquid chromatograph (Hitachi, Model K53) equipped with hydrogen flame ionization detector under the following conditions: glass column (2 m in length, 3 mm in inner diameter) packed with 1 percent OV-1 on Chromosorb W (80/100 mesh); carrier gas, nitrogen (flow rate of 30 ml/min.); oven temperature, 120-310 C programmed at 5 C/min.; injection port temperature, 300 C. To the residue after extraction of lipid materials from particulate samples collected from various depths of the Bering Sea (as stated before) was added sulfuric acid (72 percent v/v, 5 ml). The glass fiber filter soaked with sulfuric acid was allowed to stand at room temperature for three hours and then hydro ly zed at 105 C for 20 hours after dilution to IN sulfuric acid by the addition of distilled water. An aliquot of the hydrolysate was used for determining total carbohydrate by the phenol sulfuric acid meth- od (Handa 1966). A certain amount of inositol was added to the hydrolysate as an internal standard for quantitative determination of carbohydrate and then the hydrolysate was neutralized with barium hydroxide. The barium sulfate formed was removed by centrifugation. The precipitate was washed with distUled water four times, and the combined supernatants were treated with sodium borohydride (ca. 0.05 g) to reduce sugars to corresponding sugar alcohols (Abdel-Akher et al. 1951). After 16 hours the excess sodium borohydride was decomposed by the addition of acetic acid until gas evolution ceased. The solution was evaporated to dryness. To the residue was added 10 ml of methanol and then the methanol was removed by evaporation under reduced pressure. This procedure was repeated four times to remove boric acid as a complex of boric acid with methanol. The residue was dehydrated completely in vacuo and then sugar alcohols were acetylated with acetic anhydride (15 ml) in the presence of anhydrous sodium acetate (ca. 0.1 g) at 100 C for two hours. The reaction mixture was poured into ice water and allowed to remain at room temperature overnight with vigorous stirring. The sugar acetate was extracted four times with chloroform. The combined extracts were washed with distUled water, dried over anhydrous sodium sulfate and evaporated to dryness by the rotary evaporator. The residue was dissolved in a small volume of methylene chloride. A portion of the methylene chloride was injected into a gas liquid chromatograph (Yanaco Model G8) equipped with hydrogen flame ionization detector under the following conditions: glass column (1.5 m in length, 3 mm inner diameter) packed with 3 per- cent ECNSS-M on chromosorb W (100-120 mesh); carrier gas, nitrogen (flow rate of 30 ml/min.); oven temperature, 120-210 C programmed at 2 C/min.; injection port temperature, 300 C. RESULTS Distribution profiles of particulate organic carbon and nitrogen in the Bering Sea and its adjacent areas Sea water samples were collected from oceanic areas indicated in Fig. 23-1 during the cruise of the R/V Hakuho Maru of the Ocean Research Institute, University of Tokyo, from 21 June to 18 August 1975. The Oyashio area Hydrographic stations were occupied at five loca- tions off the Kuril Islands extending into the western basin of the Bering Sea. The concentration of partic- ulate organic carbon in the surface water was meas- ured in the range of 70 to 422 MgC/1, with 169 jugC/l as an average value. High values of more than 200 jugC/1 of POC were found in the surface water at stations 4, 5, and 7, while low values of POC were measured in the surface water at station 6 (Fig. 23-2). Marked differences in the concentrations of POC among the stations were observed in the intermediate waters, where high values of POC were found at stations 4 and 7, but low values at station 6. These data clearly indicate that the concentration of POC in the surface water has an effect on that of the inter- mediate waters. The distribution profile of PON is almost identi- cal with that of POC (Fig. 23-3). The concentration of PON in the surface water was measured in the range of 17-85 /xgN/l as an average value. These values in the surface water tended to decrease with depth to one-half to one-third in the intermediate and deep waters, although regional variations in values were still evident in the deeper waters. One of the most remarkable features of the 362 Chemical oceanography 120" 140° 160° 180 160° 140° 120° 120" 140° 160° 180" 160° 140° 120° Figure 23-1. The cruise track of R/V Hakuho Mam in summer of 1975. Numbers represent stations occupied. Oyashio/east Kamchatka current areas is the cold and less saline dichothermal layer at intermediate depth. The formation of the dichothermal water has been considered to occur first in the northeastern margin of the Bering Sea basin between Cape Olyutorsky and the Bay of Anadyr (Uda 1955, Ohtani et al. 1972). Strong clockwise eddies and currents along the coast and the continental shelf augment the sinking of the surface water to depth, resulting in the formation of a homogeneous water layer down to 500 m with very small temperature and salinity gradients. Precipitation and radiation in the summer cause the temperature of the surface waters to rise, leading to the formation of a cold water layer at intermediate depth in the Bering Sea and adjacent areas. This dichothermal water flows from Anadyr Bay southwestward along the Kamchatka Peninsula as the east Kamchatka current and then along the Kuril Islands. Here it splits and part flows eastward as the subarctic current while the rest continues to flow southwestward as the Oyashio. A well-developed dichothermal layer with low temperature (—0.6-1.5 C) and low salinity (32.9-33.80/oo) was found to occur in the intermediate layer at 50-311 m depth at stations 4 and 7, whUe higher values of temperature (2.0-3.7 C) and salinity (33.l-33.90/oo) were found at station 6. Temperature, salinity, and depth of the dichothermal water have been defined to be 1.0-2.0 C, 33.7-33.80/oo (salinity at the bottom of the halocline) at 100-300 m depth for the east Kamchatka current and 1.0-2.0 C, 33.80/oo at 100-140 m depth for the Pacific subarctic water (Ohtani et al. 1972). These results clearly indicate that stations 4 and 7 are in the east Kamchatka current, while station 6 is in Pacific subarctic water. The vertical profiles of particulate organic carbon and nitrogen, chlorophyll a and some oceanographic elements such eis temperature, salinity, and density are shown in Fig. 23-4 and 23-5. The concentra- tion maxima of chlorophyll a were found at 20 and 30 m at stations 4 and 6 respectively. These depths coincide with the maximum concentrations of particulate organic carbon and nitrogen and with the minimum of POC/chl. a and C/N ratios at these stations. These data suggest that phytoplankton are actively growing at these depths. Values for POC/chl. a of 200 and C/N of >6 for the particulate matter Organic matter 363 Stn4 StnS Stn6 Stn7 ~-10o 300 - E 500' I a 1000 UJ o 2000 P O C (;jgC / I ) Figure 23-2. Distribution profile of particulate organic carbon in the Oyashio area. Stn 4 Stn S Stn 6 Stn 7 5000 Figure 23-3. Distribution profile of particulate organic nitrogen in the Oyashio area. found in surface water layers at these two stations were observed. These data strongly suggest that the particulate organic carbon at these two stations consists largely of phytoplanktonic organic matter. In the subsurface through deep waters, C/N values of particulate matter tended to increase with depth at station 6, while no significant change in the values with depth was found at station 4. These data indicate that the degradation of particulate organic matter by biological agents in the deep water is more active at station 6 than at station 4. A characteristic feature of the waters in this area was found in the distribution profile of NH4-N (Fig. 23-4 and 23-5). The concentration maximum of NH4-N was found to be below the euphotic zone at all the stations in this area. Higher values were found at stations 4 £ind 7, in the cold water masses, than at station 6, in the warm water masses. Total inorganic nitrogen consisting of NOj-N, NO^-N, and NH^-N was measured within the range of 20-30 /JgN/1 in the waters above 100 m depth at both stations 4 and 7, but no significant difference in the concentration of total inorganic nitrogen between these stations was observed. Thus, the increment of NH^-N in the cold water masses might be due to the retardation of the nitrification process required to oxidize NH4-N to NO^-N. The deep Bering Sea area The distribution of water temperature was observed at five stations in the deep Bering Sea. The surface layer above 20 m consisted of warm water at all stations, because of precipitation and radiation in summer as observed in the Oyashio area. The temperature then tended to decrease rapidly toward the deep to form an extremely sharp thermocline in the water layers between the surface and 50 m. Cold water with a minimum value of —0.57 C formed the intermediate layer of 50-200 m depth at stations 7 and 8. This indicated that the east Kamchatka current extended to station 8 at that time. The dichothermal layer was much less developed at stations 9 and 10, while intermediate warm water was observed at depths between 250 and 500 m in the area centered at station 9. The 3.5 C isotherm showed a domelike upward curvature. Kitano 1970 reported that the warm water mass was introduced by the infiltration of the Alaska Stream into the Bering Sea across the Aleutian Island chain at Kamchatka Pass and west of the Attu Islands. We found interme- diate water of temperature higher than 3.8 C in the area of 55-58°N and 173-176°E, where we also found the intermediate warm water mass. These data indicate that warm water derived from the Alaskan Stream occurs consistently in the intermediate layer of the Bering Sea. The distribution profile of salinity in the deep Bering Sea indicates that the surface and subsurface layers of stations 7 and 8 consist of waters character- istic of the east Kamchatka current and those of station 11 are affected by the intrusion of less saline water from the continental shelf area. The water between the surface and 250 m at stations 9 and 10 is not affected by the continental shelf water, but upwelling of the deep water is likely to occur in the area centered at station 9 because of the upward 364 Chemical oceanography P O N ( pgN/l ) 20 40 60 P O C ( >jgC/l ) 50 100_ 150 80 100 ~1 250 300 NH4-N (>jgat N/l) 0 12 3 4 I I 1 \ r~ CHLOROPHYLL a (ug/ 1 ) 0 1 2 3 4 TEMPERATURE( "C ) 10 12 3 4 5 6 — I 1 r NH4-N sal!nity( •/..)' 32 33 34 35 R Q C/N RATIO 200 400 600 800 1000 25 POC/CHL a RATIO 28 Figure 23-4. Vertical profiles of particulate organic carbon and nitrogen, C/N, chlorophyll a, POC/chl. a, NH4-N, tem- perature, salinity, and sigma-t at station 4. curvature of the salinity profile from the surface to 500 m. These observations are in agreement with the observed temperature profile in this area. The distribution profile of particulate organic carbon along 57° N in the deep Bering Sea is shown in Fig. 23-6 (see Fig. 23-1 for locations). The concen- tration of particulate organic carbon rEinged from 80 to 240 //gC/1 in the euphotic layer (0-50 m depth) (Otabe et al. 1977). The values tended to decrease rapidly with depth below the euphotic zone at all stations except those on the continental shelf. The vertical distribution of particulate organic carbon in the intermediate water is rather complicated in this oceEinic area. High concentrations of particulate organic carbon were measured in the intermediate cold waters at stations 7, 8, and 10, while low con- centrations of particulate organic carbon were found in the intermediate warm waters centered at station 9. The distribution profile of particulate organic nitrogen in the deep Bering Sea is shown in Fig. 23-7. More than 20 MgN/1 of particulate organic nitrogen was found in the euphotic layers at stations 7, 8, 12, 14, and 15. More than 10 was found at stations 9,10, and 11. All these values tended to decrease to one-half at depths of 75-100 m. High values of par- ticulate organic nitrogen were found in the cold waters of the intermediary depths at stations 7 and 8 as observed in the distribution profile of particulate organic carbon. These data indicate that low tem- perature retards the rate of decay of the particulate organic carbon materials by biological agents, result- ing in high concentrations of particulate organic carbon in these waters. Increased values of particulate organic nitrogen were found in the deep waters of stations 10 and 11. The transportation of particulate matter from the continental shelf along the continental slope may result in these high increments of particulate organic carbon and nitrogen in the deep waters. A stepwise increase in the C/N value of the par- ticulate matter was observed to occur in the water layers between 50 and 125 m and between 300 and 3,500 m (Fig. 23-8) at station 9. The C/N value was PON (pgN/l ) 20 40 60 80 I PO C (>jgC/l) 5Q 100, 150 200 100 NH4-N(pgatN / I) 0 1 2 3 4 I I T \ r CHLOROPHYLL i ( ;jg / | ) 0 _^_^^1 . 2 3 4 ^?^ Organic matter 365 TEMPERATURE ( °C ) -1 0 1 2 3 4 5 6 7 1 \ I I — r SALINITY ( °/o. 32 4 6 8 C/N RATIO 200 400 600 800 1000 25 POC/ CHLa RATIO 26 27 28 Figure 23-5. Vertical profiles of particulate organic carbon and nitrogen, C/N, chlorophyll a, POC/chl. a, NH4-N, salinity, and sigma-t at station 6. found to be 6 in the euphotic layer, tending to increase rapidly to 10 with depth in the underlying waters, where an extremely rapid increase of the POC/chl. a value was also observed. The C/N value of the particulate matter tended to increase with depth to over 16 in the intermediate warm water at which depth the dissolved oxygen decreased to less than 0.5 ml/1. These data indicate that decay of particulate organic matter occurs in the intermediate as well as the deep waters of this region. No significance has yet been determined for the stepwise changes in POC or C/N values, but they are thought to be related to seasonal or yearly differences in production and distribution of organic matter. The continental shelf area The distribution profile of temperature in the transect along 57° N in 1975 over the continental shelf area indicates that a well-developed thermocline occurred in the water layers at approximately 30 m. Homogeneous distribution of temperature was observed in the layers above the thermocline through- out the stations, while a domelike cold water mass was found below the thermocline to the bottom in the areas centered at station 14 where the core temperature was —0.5 C. The formation and occurrence of the cold water mass has been described by several workers (Fleming 1955; Ohtani 1969; Kitano 1970a, b; Kinder and Schumacher, Chapter 4, this volume). The extremely cold water mass with core temperatures ranging from — 1.0 to —1.7 C is formed in the area between the Gulf of Anadyr and St. Matthew Island south of St. Lawrence Island as a result of severe winter cooling. The cold water extends southeast while increasing its core temperature mainly by lateral mixing with surrounding warm water. The bottom water at station 14 may be the extreme southern front of the cold tongue since this feature was not found to the south at stations 19 or 20. The vertical profile of salinity in the transect along 57° N indicates that water with relatively high salinity intrudes on the bottom of the continental shelf. However, the deep water at station 14 may not be affected by the intrusion of deep Bering Sea water. The waters of the eastern part of the continental shelf area were much affected by less saline coastal water and/or river water (Shore 1966, Roden 1967, Ohtani 366 Chemical oceanography P OC (>jgC/ l) Figure 23-6. Distribution profile of particulate organic carbon in the deep Bering Sea and the continental shelf areas. 170 E 175 E E180 W 175 W 170 W 165 W Figure 23-7. Distribution profile of particulate organic nitrogen in the deep Bering Sea and the continental shelf areas. 1969). Water with salinity less than 31.5°/oo oc- curred in the surface layers of stations 16, 17, and 18. The distribution profiles of particulate organic carbon and nitrogen are shown in Fig. 23-9 . Partic- ulate organic carbon values higher than 100 MgC/1 were observed at all shelf stations with an increase eastward from station 12 to station 15 in the surface and subsurface layers, but not in the bottom waters. The distribution profile of particulate organic ni- trogen in this area was almost identical to that of particulate organic carbon. C/N values of the partic- ulate matter ranged from 6.2 to 8.5 at stations 12 through 16. Values of POC/chl. a ranging from 52 to 248 were observed in particulate matter at stations 12 through 16. These data indicate that the particulate organic matter consists of phytoplanktonic material to a large extent; however, no characteristic features of C/N and POC/chl. a were observed in core waters with low temperatures, as found centered at station 14. Particulate organic carbon and nitrogen were found to be very abundant at station 17 and the Secchi disk depth was only 7 m (Otabe et al. 1977). The particu- late matter of this station had C/N and POC/chl. a values of more than 15 and 700 respectively even in the euphotic zone. These data indicate that the standing stock of particulate matter was much affected by allochthonous materials at this station. Station 18 showed less effect of allochthonous materials on the standing stock of particulate matter than station 17. One of the most characteristic features of the continental shelf areas was that the standing stock of inorganic nutrients was very low. Si02 -Si and NO^-N were measured within the ranges of only 3-10 and 0.1-1.0 fig/l respectively and almost undetectable amounts of NO^-N and NH4-N were found in the water layers from the surface to 10 m at stations 12-18. High concentrations of particulate organic carbon and nitrogen were observed throughout this region. The inorganic nutrient concentrations tended to decrease from the oceanic areas (stations 11 and 19) to station 17, 10 miles south of Nunivak Island, with a steep gradient at the boundary between the continental shelf and the continental slope areas; conversely, concentrations of particulate organic carbon and nitrogen tended to increase toward the continental shelf. Several authors (Koblentz-Mishke 1965, Karohji 1972, McRoy et al. 1972, Sanger 1972) have meas- ured high productivity rates for the continental shelf area of the Bering Sea. It is clear that primary productivity of phytoplankton in the continental shelf area of the Bering Sea is nutrient dependent. During the summer of 1978 a second cruise was made to the Bering Sea (Fig. 23-10). Distribution profiles of various materials along 173°-168° 30'W are shown in Fig. 23-11. Particulate matter was determined within the range of 0.25-4.38 mg/1 in the i Organic matter 367 CHLOROPHYLL a(Aig/l) 0 0 25 0.5 075 10 125 1 I I I I I PON (jugN/l ) 0 10 20 30 r OXYGEN (ml /I ) 0 2 4 6 I \ 1 r- TEMPERATURE (°C ) 0 2 4 6 8 ~I SALINITY (Voo) 33 34 35 838 100 200 300 400 500 POC/CHLOROPHYLLa RATIO Figure 23-8. Vertical profiles of particulate organic carbon and nitrogen, C/N, chlorophyll a, POC/chl. a, temperature, saUnity, and dissolved oxygen at station 9. continental shelf area through the Chukchi Sea. Regional variability was evident. High values of particulate matter were detected in the surface and subsurface waters between the continental slope and St. Matthew Island areas, while low values were observed between St. Matthew Island and the Bering Strait. Slightly higher values than in the Bering Strait were again observed in the Chukchi Sea. Almost identical distribution profiles were obtained for particulate organic carbon and chlorophyll a. More 368 Chemical oceanography POC / CHLOROPHYLL a RATIO 0 200 400 600 800 2000 I — I — 1 I r~i I \ r" I CHLOROPHYLL a (^g/| ) n M 1 f ? I C/N RATIO 0 4 8 12 16 P0N(^9N/l) SALINITY( °/..) 0 20 40 60 80 31 32 33 T I I I \ \ Figure 23-9. Vertical profiles of particulate organic carbon (—•—•—), and nitrogen (~o— o— ), C/N (-D D ), chlorophyll a {-^ ^-), POC/chl. a {—^- — ^—), temperature, and salinity at stations of the continental shelf area. definite boundaries for the distribution of these materials were found to occur on the continental slope, St. Matthew Island, and Bering Strait areas. From these data, it can be seen that high productivity values due to phytoplankton occur in the area be- tween the continental slope and St. Matthew Island and in the Chukchi area. Low productivity would be predicted in the area extending from north of St. Matthew Island to the Bering Strait. No significant differences in the concentration of SiOa-Si, NOi"-N, NOg-N, NH^N, and PO|--P be- tween the areas were observed; thus the possibility that inorganic nutrient concentrations are the direct cause of regional variability in phytoplanktonic productivity is discounted. Vertical profiles of at, however, suggest that upward transport of inter- mediate waters rich in inorganic nutrients occurred along the continental slope toward station 30. Dis- tribution profiles of dissolved organic carbon (DOC) clearly show the occurrence of upwelling in the transition area between the continental slope and the continental shelf. From these data, it can be con- cluded that inorganic nutrients are being transported into the euphotic zone in the region of station 30, thus providing the high productivity observed. Poor phytoplanktonic productivity observed in the area between St. Matthew and St. Lawrence Islands is likely to be associated with the cold water mass in the bottom layer here. The water mass has a very steep temperature gradient (4 C/10 m) at 30-40 m, which limits mixing of the water and causes surface water nutrients during the spring bloom to be depleted with no later supply providing for continued primary productivity. High productivity of phytoplankton was observed at station 21 in the Chukchi Sea, most likely as a result of the regional upwelling of sea water indicated by vertical distribution profiles of temperature and salinity. A well-defined cold water mass centered at 100 m at station 6 was found to occur along the continental slope (Fig. 23-12 a and b); it was characterized by higher values of the ratio of particulate organic carbon to particulate matter (POC/PM) than were found in ambient waters. Salinity and temperature profiles indicate that the cold water mass is identical with the type D water mass described by Ohtani et al. (1972), formed by the mixing of Alaskan Stream waters flowing along the Aleutian Ridge and the continental slope. Considering that more than 30 percent of POC/PM was found in the Alaskan Stream waters off the Aleutian Islands arc (Tanoue and Handa 1979), it is conceivable that the particulate matter of the cold water mass at station 6 may be Organic matter 369 Figure 23-10. Track chart of the KH-78-3 Cruise of the Hakuho Mam. transported from the oceanic area south of the Aleutian Islands without significant modification in its chemical composition. Since there are no data on the chemical and biochemical nature of the organ- ically rich particulate matter, it would be premature to discuss the processes of its formation. DISCUSSION AND SUMMARY Oceanic areas of the Bering Sea and its adjacent areas were divided into four regions on the basis of the hydrographic (Kitano 1970a and b, Takenouti and Ohtani 1974) and geographical (School et al. 1968) features. In this discussion similar areas were established and divided into the surface, intermediate, and deep water layers to provide for a more precise understanding of the characteristic features of the distribution of particulate organic matter. The results are shown in Table 23-1. Regional variability was evident in particulate organic carbon and nitrogen determined to be in the ranges of 34-1,038 MgC/1 and 5-79 jugN/1 in the surface waters of the Bering Sea and its adjacent areas. The average values of carbon and nitrogen tended to decrease in the following order: Chukchi Sea > the continental shelf > the Oyashio > the deep Bering Sea. The world's highest values of particulate organic carbon and nitrogen were found in the continental shelf and Oyashio areas. Such high values have never been detected in the open oceans except in the upweUing area of the South Pacific off Equador (Menzel 1967). The concentrations of particulate organic carbon and nitrogen in the Bering Sea and its adjacent areas were found to be two to five times higher than those obtained in the North Pacific Ocean (Holm-Hansen 1969, Gordon 1971, Handa et al. 1972), the North Atlantic Ocean (Chester and Stomer 1974, Banoub and Williams 1972, Gordon 1977), and the Indian Ocean (Menzel 1967, Chester and Stomer 1974). These high concentrations of particulate matter are considered to be among the most remarkable characteristics of the Bering Sea and its adjacent areas. High concentrations of particulate organic carbon and nitrogen were not found in particulate matter collected from all of the oceanic areas of the Bering Sea, but were observed at stations 4, 5, and 7 in the 2001— Figure 23-11. Distribution profiles of particulate matter, chlorophyll a, POC, and DOC along 173 -168 30 W. 370 Organic matter 371 Station 8 9 10 11 NOME 200 Figure 23-12a. Temperature distribution in eastern Bering Sea in summer of 1978. Station 8 9 10 11 POC/PM (%) 200 B Figure 23-12b. Particulate organic carbon in percentage of total particulate matter in eastern Bering Sea. Oyashio area and in the continental shelf area south of St. Matthew and Nunivak islands. The Oyashio area has a well-developed dichothermal in the inter- mediate layer of 50-250 m and the region south of St. Matthew is affected by upwelling of intermediate water along the continental slope from the deep Bering Sea. From these data, it can be seen that high values of particulate organic carbon and nitrogen were found only in areas where transport of inorganic nutrients to the euphotic zone occurred either because of vertical instability of the nutrient-rich intermediate water as observed in the Oyashio area or because of upwelling of the intermediate and deep waters as observed in the continental shelf area. To estimate the carbon content of living organisms in particulate matter, the relationship between chlorophyll a and organic carbon and nitrogen was examined. Linear relationships between these mate- rials were found to occur and the regressions found for samples in the euphotic layers of the four areas are shown in Table 23-2. The independent terms of each of the equations are the concentration of the particulate organic carbon when chlorophyll a is equal to zero; this is the concentration of organic carbon and nitrogen in detritus or allochthonous materials. The ratios of the detrital organic carbon 372 Chemical oceanography TABLE 23-1 Summary of the ranges and average values (± 1 SD) of particulate organic carbon and nitrogen and C/N in the Bering and Chukchi Seas 0-100 m 100-500 m 500 m Particulate organic Particulate organic Particulate organic carbon nitrogen C/N carbon nitrogen C/N carbon nitrogen C/N Area AigC/l /jgN/1 MgC/1 AigN/l MgC/1 A/gN/1 Oyashio area Range 70-422 10-85 3.3-8.4 28-142 2.2-21 5.4-26 16-145 1-21 5.5-20 N.S.* 8 28 28 43 42 42 41 37 37 Average 169±81 29±16 6.1+1.0 62±29 7.4±3.6 8.8±2.2 53±27 5±4 11±2.8 Deep Bering Sea area Range 34-185 5.7-26 4.5-9.5 26-126 2.4-8.7 5.9-15 19-61 1.1-9.1 5.3-19 N.S.* 20 20 20 30 28 28 43 35 35 Average 91±39 13±6 7.1±1.5 46±21 4.5±1.5 9.9±1.9 33±10 3.4+1.9 11±3.7 Continental Shelf area Range 50-1038 7.5-157 4.4-15 46-260 5.7-39 6.3-9.6 — — — N.S.* 113 108 108 18 17 17 — — — Average 204±178 31±25 6.4±1.9 65±44 7.4±4.9 8.3±4.9 — — — Chukchi Sea area Range 82-376 16-72 48-59 — — — — — — N.S.* 15 15 15 — — — — — — Average 220±94 42±19 5.3±0.28 — " " ~ *Number of samples. TABLE 23-2 Relationship between chlorophyll a and particulate organic carbon in the euphotic layers of various ocean areas Number of Regression Area Samples equation r Oyashio area 21 POC=210 Chi. a + 64 0.714 21 PON= 37 Chi. a+ 8 0.814 Deep Bering 12 POC= 87 Chi. a + 48 0.671 Sea area 12 PON= 13 Chi. a + 10 0.866 Continental 32 POC= 72 Chi. a + 94 0.458 Shelf area 39 PON= 12 Chi. a + 11 0.616 Chukchi Sea 12 POC= 60 Chi. a +103 0.894 area 12 PON= 12 Chi. OH- 17 0.917 and nitrogen to the particulate organic carbon and nitrogen were calculated with the average values of 34.2 and 27.5 percent, 52.7 and 76.9 percent, 46.0 and 35.5 percent, and 46.8 and 40.5 percent for the Oyashio area, the deep Bering Sea area, the continen- tal shelf area, and the Chukchi Sea area respectively. These data indicate that 50-70 and 25-70 percent of particulate organic carbon and nitrogen consist of phytoplanktonic carbon and nitrogen in these oceanic areas. Holm-Hansen (1969) reported that the ratios of living organic carbon to particulate organic carbon were 62-73 percent and 55-73 percent on the basis of the biomass estimated by the measurements of ATP and chlorophyll a respectively. From these data, it can be seen that detrital organic carbon is abundant in the deep Bering Sea area, and phytoplanktonic organic carbon is more abundant in the particulate organic carbon of the Oyashio and Chukchi Sea areas. Extremely high values of detrital organic carbon were found in the particulate matter from stations 17 and 18 off Nunivak Island on the continental shelf. Since less saline waters were observed at these sta- tions, it can be assumed that drainage from land brings about the increase in detrital organic carbon of the particulate matter at these stations. BIOCHEMICAL CONSTITUENTS OF PARTICULATE MATTER Free and combined amino acids Particulate samples collected at stations 11, 13, 14, and 33 in the deep Bering Sea, continental shelf, and northern North Pacific areas during Hakuho Maru cruise KH-75-4 were analyzed for free and combined amino acids. The particulate amino acids Organic matter 373 were found to be within the ranges of 10.3-78.0, 104-156, and 10.4-96.4 ^g/\ in the deep Bering Sea, continental shelf, and northern North Pacific areas respectively (Fig. 23-13). These values indicate that regional variations of particulate amino acids are identical with those of particulate organic carbon and nitrogen. The concentration of particulate amino acids tended to decrease in a steep gradient with depth in the surface and subsurface layers at all stations. No significant vertical gradient was observed in the particulate amino acid of the deep waters at stations 11 and 33, but slightly higher values were found in the intermediate waters at station 33. The ratios of particulate amino acid carbon (PAC) and nitrogen (PAN) to particulate organic carbon (POC) and nitrogen (PON) in the surface and subsur- face layers (0-70 m) were found to have an almost identical range of values throughout the stations. Slightly lower values, however, were obtained at stations 13 and 14. It is assumed that this effect is CONCENTRATION OF AMINO ACI D ( ^J 9 / I ) 50 100 150 Stn.14 5000L Figure 23-13. Vertical profiles of particulate amino acids at various stations in the deep Bering Sea, continental slielf , and northern North Pacific areas. largely due to contamination by terrigenous materials since these stations are in Bristol Bay. PAC/POC and PAN/PON at these stations were found with the ranges of 24.6-31.3 percent and 47.3-62.0 percent respectively. These values do not conflict with the ratios obtained in the high latitude areas (30-50° N) of the North Pacific Ocean, but much lower values were found in tropical and sub- tropical areas of the Pacific Ocean (Handa et al. 1972). Amino acid composition of the particulate matter at stations 11 and 33 is shown in Table 23-3. Serine, glycine, and alanine were found to be the dominant components of particulate amino acids in the samples from £l11 depths, whereas aspartic acid and glutamic acids dominated from the euphotic zone (50 m) down. Valine, leucine, and lysine were found in intermediate concentrations in all of the samples analyzed. The amino acid composition of the par- ticulate matter collected from the surface was similar to that of marine phytoplankton, but the composi- tion varied with depth. The molar ratios of alanine and arginine tended to decrease with depth, while those of serine, glutamic acid, and glycine increased. Arginine, lysine, aspartic acid, glutamic acid, serine, glycine, and alanine have been reported as major amino acids in marine phytoplankton and in zooplankton and fecal pellets (Parsons et al. 1961, Siegel and Degens 1966, Starikova and Korshikova 1969, Daumas 1976). Arginine nearly disappeared from the water column in the deep waters of station 11, but was found at 2,000 m, the only depth sam- pled, at station 33. Rittenberg et al. (1963) and Degens et al. (1964) reported that arginine tended to decrease rapidly with depth in particulate matter and produce ornithine and urea (Clark et al. 1972, Degens 1970). We also observed 1.05-1.54 /ug/l of urea in particulate matter from the deep water at stations 11 and 33; samples were not analyzed for ornithine. Thus, these data may suggest that such a rapid decrease in arginine in particulate matter is due to bacterial use. Relative abundances of the essential amino acids were much higher in the surface waters than in the deep waters, but significant amounts of the essential amino acids including arginine, lysine, histidine, isoleucine, and methionine were found in the par- ticulate samples from deep water. In the particulate samples from the deep water of station 33, the relative abundance of the essential amino acids was found to be about the same as in the surface samples of particulate matter at station 11. These data strongly suggest that particulate amino acids in the deep water layers must be rapidly transported from 374 Chemical oceanography TABLE 23-3 Distribution of amino acids in particulate matter collected from various depths (in residues per 1,000) Statior I Depth m Basic Acidic Asp Glu Thr Ser Neutral Pro Gly Ala Val He Lue Sulfur Met Aromatic Phe Total Arg Lys His mmol/1 11 30 31 63 21 12 11 7 96 42 125 104 68 44 68 8 36 307 50 32 86 29 92 75 61 170 35 150 95 58 35 23 6 23 278 2,250 tr 60 17 106 129 53 132 40 182 83 59 40 69 3 30 121 2,750 tr 69 20 109 132 59 125 43 164 86 63 39 63 3 26 122 33 2,000 34 73 26 87 97 49 171 24 191 73 46 101 12 34 24 234 the surface layers with only limited modification in transit. It appears likely that fecal pellets of zoo- plankton are an important carrier of the particulate amino acids from the surface to the deep water layers. Carbohydrates Particulate matter was collected for carbohydrate analysis (PCC) at several stations in the Bering Sea. Composite samples of this material were analyzed for carbohydrates by gas chromatographic techniques and found to be in the range of 15.3-32.1 /igC/1 or 13.1-27.2 percent of POC, in the continental shelf area (Table 23-4). These values were about the same as the PCC/POC values (0.12-0.40) obtained in the Pacific Ocean from 50°N to 68°S along 170°W (Handaetal. 1972). Regional variability in the concentration of partic- ulate carbohydrate was evident. High values were found in the continental shelf area, whereas low values of PCC/POC were obtained in the deep Bering Sea. Parsons and Strickland (1962) reported that detrital samples from marine environments sometimes gave unexpectedly high concentrations of protein, while carbohydrate was only a minor component. McAllister et al. (1960) speculated that oceanic TABLE 23-4 The monosaccharide composition of carbohydrate in particulate matter from various areas during the cruise of KH-75-4. Monosaccharide Deep Bering area Surface Deep Layer Layer (Stns. 9,10,11) (Stns. 9,10,11) 0-175 m >175m Continental Shelf area Surface Bottom Bottom Layer Layer Resusp. (Stns. 12,16,17,19) (Stns. 12,19) (Stns. 19) 0-50 m 50-80 m 100-190 m Phamnose Fucose Ribose Arabinose Xylose Mannose Galactose Glucose Total carbohydrate (/igC/l) Carbohydrate carbon/ Org.C. (%) — mol. % 6.24 1.55 3.01 6.31 5.02 19.1 7.03 3.33 7.81 5.37 2.42 tr. tr. tr. tr. 5.71 3.23 5.68 2.00 6.87 3.90 2.55 2.80 5.80 11.0 20.8 19.0 14.5 7.14 8.11 5.48 1.25 5.39 65.2 10.6 36.2 65.4 65.3 53.0 3.2 3.91 3.0 8.42 32.1 13.5 17.6 27.7 15.2 13.1 Organic matter 375 detritus from the northeastern Pacific originated from fragments of zooplankton. These data suggest that the particulate matter of the deep Bering Sea may be largely composed of the bodies of protein- rich bacteria. As shown in Table 23-4, glucose and mannose were found to be dominant species of monosaccharides after hydrolysis of the particulate carbohydrate with dilute sulfuric acid. Fucose was also found to be abundant, but only in the particulate samples from the surface and subsurface layers of the deep Bering Sea area. Comparable values of monosaccharides were obtained in the particulate samples from the continental shelf and the deep sea area although lower v£ilues of mannose were found in the shelf samples. Handa and Yanagi (1969) reported that particulate carbohydrate of the northwestern Pacific Ocean was composed mainly of glucose and mannose, which accounted for 21.8-54.6 percent and 9.5-24.4 percent of the total carbohydrate. The authors found that fucose, also abundant, accounted for up to 12.8 percent of the particulate carbohydrate in those samples. On the basis of these data, there do iiot appear to be significant differences in the chemical nature of carbohydrates between the particulate samples from the Bering Sea and those from the northwestern north Pacific Ocean. According to detailed analyses of carbohydrates of marine diatoms conducted by Handa (1969) and Hang and Myklestad (1976), cell wall polysaccharides which are soluble in alkali consist mainly of mannose with smaller concen- trations of fucose, glucose, galactose, rhamnose, xylose, and arabinose. Water-extractable polysac- charides, however, consist mainly of glucose, which accounts for 90 percent of this polysaccharide fraction. Arabinose, mannose, rhamnose, and ribose were detected as minor monosaccharide constituents. These data indicate that the high proportion of mannose to total carbohydrate in the particulate samples of the Bering Sea must be due to cell wall polysaccharides of diatoms which have been reported to be the main primary producers in the Bering Sea at the time when the particulate samples were collected (Ishimaru and Nemoto 1977). Glucose, another main component of particulate carbohydrate, is also found in the intermediate and deep waters, but is unlikely to be derived from water- extractable polysaccharides of diatoms since it is quickly consumed by microbiological agents (Handa 1969). Handa and Yanagi (1969) reported that glucose still accounted for 30-41 percent of residual carbohydrate after removal of water-extractable carbohydrate in the particulate samples collected from the surface and subsurface waters of the north- western North Pacific Ocean. Thus these data suggest that a certain portion of the glucose found in the hydrolysate of particulate matter is derived from cell wall polysaccharides of the diatom cells. Another source of glucose is terrigenous cellulose fibers which have been observed in the detritus of the deep waters even at Station "P" in the North Pacific Ocean (Parsons and Strickland 1962). It is likely that particulate matter in the continental shelf area is more affected by these terrigenous cellulose materi- als than that in deep water. However, with the present state of knowledge, the source of the glucose cannot be clearly determined because of the limited information available on the chemistry of polysac- charides of particulate matter. Fatty acids Intensive studies have been conducted to analyze for fatty acids in all kinds of biological materials found in the ocean in order to determine the nutri- tional value of various items in the marine food chain. In this study, emphasis was placed on determining the fatty acid content of particulate matter found in the Bering Sea and evaluating the biological activity of these particulate materials, especially those collected from the surface and subsurface waters. Some effort is also given to the utilization of fatty acid distribu- tion data in vertical transport studies. Particulate samples were collected from station 4 of the cruise of KJ-78-3 in the Bering Sea. After con- version of fatty acids extracted from the particulate matter to corresponding methyl esters, the methyl ester composition was determined by gas chromato- graphy or by combined gas chromatography and mass spectrometry. The fatty acid composition of the particulate matter at station 4 of KH-78, shown in Table 23-5, ranged from 1.4 to 6.3 MgC/1, which accounted for 4.9-11.7 percent of the total organic carbon. The concentration tended to decrease with depth while the ratio of fatty acid carbon to partic- ulate organic carbon tended to increase with depth. High proportions of C14.0 , Ci6:o , and Ciea acids to total fatty acids were found in the particulate samples from surface and subsurface waters. The particulate matter collected here consisted mainly of diatoms (75 percent in cell number) with some dinoflagellates (25 percent). This distribution agrees well with the fact that the Bacillariophyceae have been distinguished from other classes of phytoplankton having a high level of C14.0 and C^q^ acids, but a low level of Cig-o acid (Ackman et al. 1968). One of the characteristic features of the particu- late matter from station 4 is the high values of poly- 376 Chemical oceanography TABLE 23-5 Fatty acid compositions of particulate matter at station 4 of KH-78-3. Depth (m) Fatty acid 1 50 110 177 500 2,000 iso^ 14:0^ tr-^ 0.6 2.2 1.2 0.5 0.6 14:1 tr 1.0 1.3 0.9 0.4 0.4 14:0 17.7 13.7 10.1 9.6 6.7 7.4 anteiso^ 15:0 0.3 1.9 1.5 1.5 0.8 0.7 iso 15:0 1.3 3.1 3.0 2.4 1.2 1.3 15:0 0.9 2.7 6.0 4.5 2.9 3.6 16:2,16:3,16:4 3.0 tr tr 1.1 ND ND 16:1 12.3 15.5 12.4 8.0 4.4 4.8 16:0 26.7 41.4 31.1 36.2 34.0 38.4 anteiso 17:0 0.8 0.9 2.2 2.2 1.0 1.1 iso 17:0 0.5 1.1 2.2 1.4 0.9 0.9 17:0 0.4 0.7 1.5 1.4 1.8 1.8 18:2,18:3,18:4 13.5 3.8 tr 1.6 tr 1.6 18:1 17.8 10.5 8.8 6.3 4.4 4.3 18:0 5.0 1.3 13.0 18.7 37.9 27.5 19:0 tr 0.8 0.8 0.3 0.4 0.4 20:0 tr 0.6 0.9 0.9 1.2 1.5 22:0 0.8 0.4 1.6 0.5 0.8 0.8 23:0 tr tr tr 0.2 tr 0.3 24:0 tr tr 1.3 0.8 0.8 1.5 25:0 tr tr tr tr tr 0.3 26:0 tr ND^ tr 0.2 tr 0.9 27:0 ND ND ND ND ND ND 28:0 ND ND tr tr tr tr Total (MgC/1) 6.3 3.4 1.4 2.9 2.6 3.3 FA-C/OC^(%) 4.9 5.8 6.6 12.0 11.7 10.5 N/B* 34.5 14.2 11.2 13.3 25.6 25.0 ^Branched fatty acids ^All fatty acids are given by normal chain length: '^ Trace •^None detected ® Fatty acid carbon/organic carbon ^Normal acid/iso- and anteiso-acid number of double bond. unsaturated Cjg acids, which are important bio- chemically in connection with the photosynthetic processes of diatoms— for example, in the stimulation of chloroplast formation (Rosenberg and Gouanx 1967) and as an active intermediate of photosyn- thetic carbon transfer (Kates and Volcani 1966). Schultz and Quinn (1977) observed that in coastal waters the total concentration of fatty acids increased as did the relative concentration of polyunsaturated Ci8 acids as the phytoplankton bloom progressed. The polyunsaturated acids varied from 7 percent at the start to 12 percent at the peak of the phytoplank- ton bloom, then dropped to 3 percent at the end of the bloom in coastal waters. These data are compati- ble with the work of Jeffries (1970), who found that the Cig:2 acid accounted for 4 percent and 7 percent of total fatty acids in the summer-fall and winter- spring phytoplankton assemblages respectively. From the data in Table 23-5, it may be deduced that phytoplankton of the particulate matter from station 4 were actively growing cells sampled during the period of the phytoplankton bloom. Of the total fatty acids of the surface particulate matter, approxi- mately 30 percent were found to be unsaturated Cig Organic matter 377 acids, which concentration decreases rapidly with depth. These data, coupled with the literature cited above, point to the possible utility of this kind of observation as a diagnostic tool in assessing the biological activity of phytoplankton in marine environments. Several branched (iso- and anteiso-) fatty acids of Ci4, Ci5, and C17 were found in the particulate samples from the surface through deep waters with lower values than those obtained in the sedimentary samples from various sources. Extensive work by Leo and Parker (1966) and Cooper and Blumer (1968) indicates that iso- and anteiso-acids are potential indicators of bacterial contribution to sedimentary organic matter because branched acids predominate over the straight chain acids in bacteria (Kates 1964, Kaneda 1967). Blumer (1970) reported that the ratio of normal acid to iso- and anteiso-acids was calculated to be 3-7 and 50-100 for marine sediments and marine plankton, respectively, and low values of the ratio in the sediments were due to bacterial contami- nation. In view of these data, the particulate matter from the surface water (1 m) sampled at station 4 was almost free from bacterial contamination as indicated by the relatively high value (34.5) of the normal acid/iso- and anteiso-acids observed. Bacterial con- tamination of the particulate matter is, however, evident in the intermediate waters from 50 to 177 m depth, where low values of the normal acid to branched acid ratio were found. Such phenomena may be expected to occur in the intermediate waters since the particulate organic matter is available to bacteria once it sinks from the euphotic zone to the underlying intermediate waters. The selective degra- dation of normal acids, especially unsaturated acids, may also affect the acid/iso- and anteiso-acid ratios to some extent. All of the fatty acid components decrease in abundance in the particulate matter with depth except the saturated C^g acid, which increased from 0.32 MgC/1 at 1 m to 0.91 MgC/1 at 2,000 m. Since saturated C^g acid is only a minor component of the fatty acids of phytoplankton, zooplankton (Hinch- cliffe and Riley 1972, Lee et al. 1971), and fishes (Bishop et al. 1976), these biological materials are not likely to be a direct source of this accumulation. Selective degradation of the particulate fatty acids is a more likely source of this material, but more complex processes may cause the increment increase in the particulate matter found at depth. Sorption of dissolved fatty acids by mineral particles in sea water may be another process functioning to increase the concentration of saturated Cig acid in the deep waters. The sorption process is controlled by physi- co-chemical factors (temperature, pH, and salinity) (Meyers and Quinn 1973), the carbon number of fatty acids (Barcelona and Atwood 1979), and the size of the mineral particles (Morris and Calvert 1975). The observation that saturated Cig acid is the most abundant of dissolved fatty acids (Williams 1965, Ackman and Hooper 1970) in suspended particles of 4-10 nm in diameter (Tsunogai et al. 1977) suggests that the sorption processes of dis- solved fatty acids with mineral particles may be functional in causing the increased concentration observed. REFERENCES Abdel-Akher, M. J., J. K. Hamilton, and F. Smith 1951 The reduction of sugars with sodium borohydride. J. Amer. Chem. Soc. 73:4691-4. Ackman, R. G., and S. N. Hooper 1970 Analyses of fatty acids from New- foundland copepods and sea water with remarks on the occurrence of arachidic acid. Lipids 5:417-21. Ackman, R. G., C. S. Tocher, and J. McLachlan 1968 Marine phytoplankter fatty acids. J. Fish Res. Bd. Can. 25:1603-20. Banoub, M. W., and P. J. Williams 1972 Measurements of microbial activity and organic material in the western Mediterranean Sea. Deep-Sea Res. 19:433-43. Barcelona, M. J., and D. K. Atwood 1979 Gypsum-organic interactions in the marine environment: sorption of fatty acids and hydrocarbons. Geochim. Cosmochim. Acta 43:47-53. Bishop, D. G., D. G. James, and J. Olley 1976 Lipid composition of slender tuna (Allothunnus fallai) as related to lipid composition of their prey (Nycti- phanes australis). J. Fish. Res. Bd. Can. 33:1156-61. 378 Chemical oceanography Blumer, M. 1970 Dissolved organic compounds in sea water. Saturated and olefinic hydro- carbons and singly branched fatty acids. In: Organic matter in natural waters, D. W. Hood, ed., 153-67. Inst. Mar. Sci., Occ. Pub. No. 1. Univ. of Alaska, Fairbanks. Chester, R., and J. H. Stomer 1974 The distribution of particulate organic carbon and nitrogen in some surface waters of the world ocean. Mar. Chem. 2:263-75. Clark, M. E., G. A. Jackson, and W. J. North 1972 Dissolved free amino acids in southern California coastal waters. Limnol. Oceanogr. 17:749-58. Cooper, W. J., and M. Blumer 1968 Linear, iso and anteiso fatty acids in recent sediments of the North Atlantic. Deep-Sea Res. 15:535-40. Daumas, R. A. 1976 Variations of particulate protein and dissolved amino acids in coastal seawater. Mar. Chem. 4:225-42. Degens, E. T. 1970 Molecular nature of nitrogenous com- pounds in sea water and recent marine sediments. In: Organic matter in natural waters, D. W. Hood, ed., 77-106. Inst. Mar. Sci., Occ. Pub. No. 1. Univ. of Alaska, Fairbanks. Degens, E. T., J. H. Reuter, and K. H. Shaw 1964 Biochemical compounds in off shore sediments and sea water. Geochim. Cosmochim. Acta 28:45-66. Fleming, R. H. 1955 Review of the oceanography of the Northern Pacific. Inter. N. Pac. Fish. Comm. Bull. 2:1-43. Fowden, L. 1954 A comparison of the composition of some algal protein. Ann. Bot. 18:257-66. Gordon, D. C. 1971 Distribution of particulate organic carbon and nitrogen at an oceanic station in the Central Pacific. Deep- Sea Res. 18:1127-34. 1977 Variability of particulate organic carbon and nitrogen along the Halifax- Bermuda Section. Deep-Sea Res. 24:257-70. Handa, N. 1966 1969 Examination on the applicability of the phenol sulfuric acid method to the determination of dissolved carbohy- drate in sea water. J. Oceanogr. Soc. Japan 22:79-86. Carbohydrate metabolism in the marine diatom Skeletonema costatum. Mar. Biol. 4:208-14. Handa, N., and K. Yanagi 1969 Studies on water-extractable carbohy- drates of the particulate matter from the northwest Pacific Ocean. Mar. Biol. 4:197-210. Handa, N., K. Yanagi, and K. Matsunaga 1972 Distribution of detrital materials in the western Pacific Ocean and their biochemical nature. Mem. 1st. Ital. Idrobiol. 29 Supp.: 53-71. Favorite, F. 1967 1974 The Alaskan Stream. Bull. Inter. N. Pac. Fish. Comm. 21:1-20. Flow into the Bering Sea through Aleutian Island passes. In: Oceano- graphy of the Bering Sea, D. W. Hood and E. J. Kelley, eds., 3-37. Inst. Mar. Sci., Occ. Pub. No. 2. Univ. of Alaska, Fairbanks. Hang, A., and S. Myklestad 1976 Polysaccharides of marine diatoms with special reference to Chaetoceros species. Mar. Biol. 34:217-22. Hinchcliffe, P. R., and J. P. RUey 1972 The effect of diet on the component fatty acid composition of Artemia salina. J. Mar. Biol. Assoc. U.K. 52: 203-11. Organic matter 379 Holm-Hansen, O. 1969 Determination of microbial biomass in ocean profiles. Limnol. Oceanogr. 14:740-7. Kitano, K. 1970a A note on the thermal structure of the eastern Bering Sea. J. Geophys. Res. 75: 1110-15. Hood, D. W., and E. J. Kelley, eds. 1974 Oceanography of the Bering Sea with emphasis on renewable resources. Inst. Mar. Sci., Occ. Pub. No. 2. Univ. of of Alaska, Fairbanks. Hood, D. W., and W. S. Reeburgh 1974 Chemistry of the Bering Sea: An overview. In: Oceanography of the Bering Sea, D. W. Hood and E. J. Kelley, eds., 191-204. Inst. Mar. Sci., Occ. Pub. No. 2. Univ. Alaska, Fairbanks. Ishimaru, I., and T. Nemoto 1977 Preliminary report of the Hakuho Maru Cruise KH-75-4. Ocean Res. Inst., Univ. Tokyo. Jeffries, H. P. 1970 Seasonal composition of temperate plankton communities: fatty acids. Limnol. Oceanogr. 15:419-26. Kaneda, T. 1967 Karohji, K. 1972 Fatty acids in the genus Bacillus. I. Iso- and anteiso fatty acids as charac- teristic constituents of lipids in 10 species. J. Bact. 93:894-903. Regional distribution of phytoplank- ton in the Bering Sea and western and northern subarctic regions of the North Pacific Ocean in summer. In: Biological oceanography of northern North Pacific Ocean (Moto- da Commemorative volume), A. Y. Takenouti, ed., 96-116. Idemitsu- Shoten, Tokyo. Kates, M. 1964 Bacterial lipids. In: Adv. Lipid Res., R. Paoletti and D. Kritchevsky, eds., 2:17-90. Academic Press, N.Y. Kates, M., and B. E. Volcani 1966 Lipid components of diatoms. Bio- chim. Biophys. Acta 116:264-78. 1970b A note on the salinity structure of the eastern Bering Sea. Bull. Tohoku Reg. Fish. Lab. 30:79-85. Koblentz-Mishke, O. L. 1965 Velichina pervichnoi produktsii Tik- kogs okeano (Primary production in the Pacific). Okeanologiya 5:325-37. Lee, R. J., J. Hirota, and A. M. Barnett 1971 Distribution and importance of wax esters in marine copepods and other zooplankton. Deep-Sea Res. 18: 1147-65. Leo, R. F., and P. L. Parker 1966 Branched chain fatty acids in sedi- ments. Science 152:649-50. Loder, T. C. 1971 Distribution of dissolved and partic- ulate organic carbon in Alaskan polar, subpolar and estuarine waters. Ph.D. Dissertation, Univ. of Alaska, Fair- banks. McAllister, C. D., T. R. Parsons, and J. D. H. Strickland 1960 Primary productivity and fertility at Station "P" in the North-East Pacific Ocean. J. Cons. Int. Explor. Mer. 25:229-38. McRoy, C. P., J. H. Goering, and W. E. Shiels 1972 Studies of primary production in the eastern Bering Sea. In: Biological oceanography of the northern North Pacific Ocean (Motoda Commemora- tive Volume), A. Y. Takenouti, ed., 199-216. Idemitsu-Shoten, Tokyo. Menzel, D. W. 1966 Bubbling of sea water and the produc- tion of organic particles: a re-evalua- tion. Deep-Sea Res. 13:963-6. 1967 Particulate organic carbon in the deep sea. Deep-Sea Res. 14:229-38. 380 Chemical oceanography Metcalfe, L. D., A. A. Schmitz, and J. R. Pelka 1966 Rapid preparation of fatty acid esters from lipids for gas chromatographic analysis. Anal. Chem. 38:514-15. Meyers, P. A., and J. G. Quinn 1973 Factors affecting the association of fatty acids with mineral particles in sea water. Geochim. Cosmochim. Acta 37:1745-59. Morris, R. J., and S. E. Calvert 1975 Fatty acid uptake by marine sediment particles. Geochim. Cosmochim. Acta 39:377-81. Nakajima, K. 1969 Suspended particulate matter in the waters on both sides of the Aleutian ridge. J. Oceanogr. Soc. Japan 25:239-48. Ohtani, K. 1969 On the oceanographic structure and the ice formation on the continental shelf in the eastern Bering Sea. Bull. Fac. Fish. Hokkaido Univ. 20:94-117. 1970 Relative transport in the Alaskan Stream in winter. J. Oceanogr. Soc. Japan 26:271-82. Ohtani, K., Y. Akiba, and Y. Takenouti 1972 Formation of western subarctic water in the Bering Sea. In: Biological oceanography of the northern North Pacific Ocean (Motoda Commemora- tive Volume), A. Y. Takenouti, ed., 31-44. Idemitsu-Shoten, Tokyo. Otabe, H., T. Nakai, and A. Hattori 1977 Underwater irradiance and secchi disk depth in the Bering Sea and the northern North Pacific in summer. Mar. Sci. Comm. 3:255-70. Parsons, T. R., K. Stephens, and J. D. H. Strickland 1961 On the chemical composition of eleven species of marine phytoplank- ters. J. Fish. Res. Bd. Can. 18: 1001-16. Parsons, T .R., and J. D. H. Strickland 1962 Ocean detritus. Science 136:313-14. Rittenberg, S. C, K. O. Emery, J. Hulsemann, E. T. Degens, R. C. Fay, J. H. Reuter, J. R. Grady, S. H. Richardson, and E. E. Bray 1963 Biogeochemistry of sediments in experimental Mohole. J. Sed. Petr. 33:140-72. Roden, G. 1967 Rosenberg, A. 1967 Sanger, G. A. 1972 On river discharge into the North- eastern Pacific Ocean and the Bering Sea. J. Geophys. Res. 72:5613-29. and J. Gouanx Quantitative and compositional changes in monogalactosyl and diga- lactosyl diglycerides during light- induced formation of chloroplasts in Euglena gracilis. J. Lipid Res. 8:80-3. Fishery potentials and estimated bio- logical productivity of the subarctic Pacific region. In: Biological ocean- ography of the northern North Paci- fic Ocean (Motoda Commemorative Volume), A. Y. Takenouti, ed., 562-74. Idemitsu-Shoten, Tokyo. School, D.W., E.G. Edwin, D. Buffington, and D.M. Hopkins 1968 Geological history of the continental margin of North America in the Bering Sea. Mar. Geol. 6:297-330. Schultz, D. M., and J. G. Quinn 1977 Suspended material in Narragansett Bay: fatty acid and hydrocarbon composition. Org. Geochem. 1:27-36. Shore, G. G. 1966 Bering Sea. In: The encyclopedia of oceanography, R .W. Fairbridge, ed., 134-41. Reinhold Pub. Co., N.Y. Siegel, A., and E. T. Degens 1966 Concentration of dissolved amino acids from saline waters by ligand- exchange chromatography. Science 151:1098-1101. Starikova, N. D., and R. I. Korshikova 1969 Amino acids in the Black Sea. Ocean- ology 9:509-18. r Organic matter 381 \ Takenouti, Y., and K. Ohtani 1974 Currents and water masses in the Bering Sea. A review of Japanese work. In: Oceanography of the Bering Sea, D. W. Hood and E. J. Kelley, eds., 39-57. Inst. Mar. Sci., Occ. Pub. No. 2. Univ. of Alaska, Fairbanks. Tanoue, E., and N. Handa 1979 Distribution of particulate organic carbon and nitrogen in the Bering Sea and northern North Pacific Ocean. J. Oceanogr. Soc. Japan 35:47-62. Tsunogai, S., M. Minagawa, S. Konishi, M. Kusakabe, T. Shinagawa, and M. Nishimura 1977 North Pacific. Preliminary Rep. the Hakuho Maru Cruise KH-75-4. Ocean. Res. Inst., Univ. of Tokyo, 54-6. Uda, M. 1955 The fluctuation of Oyashio Current in relation to the atmospheric circulation and to the distribution of the dicho- thermal waters in the North Pacific Ocean. Records Oceanogr. Wks. Japan 2(2):43-55. Udenfriend, S., S. Stein, W. Dairman, W. Leimgruber, and M. Wigele 1972 Fluorescamine: A reagent for assay of amino acids, peptides, protein and primary amines in the picomole range. Science 178:871-2. Williams, P. M. 1965 Fatty acids derived from lipids of marine origin. J. Fish. Res. Bd. Can. 22:1107-22. Yanagi, K., and N. Handa 1970 Some examinations on determining methods of chlorophyll a in marine particulate matter. Bull. Plankton Soc. Japan 17:31-41. Zenkevitch, L. 1963 Biology of the seas of the U.S.S.R. George Allen and Unwin, London. i Hydrocarbons of Animals of the Bering Sea D. G. Shaw and E. R. Smith Institute of Marine Science University of Alaska Fairbanks, Alaska ABSTRACT Samples of biological materials including pelagic animals and marine birds and mammals have been collected in the Bering Sea for hydrocarbon determination. Pristane, a com- pound derived from chlorophyll of zooplankton, was both abundant and ubiquitous, being detected in 91 percent of the tissues analyzed in concentrations up to 220 /Jg/g. Heneicosa- hexaene was also common, detected in 32 percent of the samples in concentrations up to 88 Mg/g- However, the latter compound, which is produced by marine algae, was essentially confined to organisms low in the food web. No hydrocarbons from petroleum or terrigenous plant sources were detected in any of the animal tissues analyzed. Our analyses show that the source of hydrocarbons in the Bering Sea pelagic environment is biosynthesis in that environment. This is in keeping with the current understanding of productivity and carbon flow in this area. INTRODUCTION In many coastal environments the hydrocarbon constituents of marine animals include important contributions from anthropogenic and natural terri- genous sources. However, all of the hydrocarbons identified in organisms collected in the southeastern Bering Sea in the spring of 1976 and 1977 appear to have had their origin in the marine pelagic system of that region. We base this conclusion on the results of the analyses of 34 samples of plankton, fish, bird, and marine mammal tissues carried out as part of a survey of the kinds and amounts of ambient hydrocarbons in the Bering Sea environment. These are the first reported determinations of hydrocarbons in a suite of biological materials collected in the Bering Sea. METHODS Plankton, fish, and bird samples which were col- lected in 1977 expressly for these hydrocarbon 383 analyses were immediately frozen in specially cleaned (500 C, 24 hours) glass containers. Marine mammal tissue samples had been collected in 1976 as part of a study of general biology and life history of Alaskan marine mammals. These tissues had been wrapped in aluminum foil, placed in plastic bags, and frozen. In the laboratory approximately 10 g of tissue was extracted for two hours at 90 C in a capped centri- fuge tube containing 10 ml 4 N aqueous KOH and 2 ml hexane. After the extract had cooled to room temperature, an additional 10 ml of hexane was added. The tube was then shaken well and centri- fuged for 15 minutes at 2,500 rpm. The hexane (upper) phase was removed by pipette and the aqueous phase extracted twice more with hexane in the same manner. The hexane extracts were com- bined and dried over anhydrous sodium sulfate. The hexane was concentrated to about 5 ml and an aliquot evaporated and weighed to determine the total nonsaponifiable lipids. Extracts were column chromatographed on silica gel (5 percent water). Saturated hydrocarbons were eluted first with hex- ane. Next unsaturated hydrocarbons were eluted with 40 percent benzene in hexane. Eluates were concentrated to approximately 1 ml for angilysis by gas chromatography (GC). Each fraction was first analyzed by flame ionization GC (Hewlett Packard 5710) using a 50 m by 0.7 mm support coated open tubular column with OV-101 serving as stationary phase. Quantification was accomplished by a digital integrator (Hewlett Packard 3380 or 3385) and corrected for percentage recovery. Hydrocarbons were identified through the use of internal and external standards and by mass spectrometry (MS) 384 Chemical oceanography using a Hewlett Packard 5390/5933 computerized GCMS system. RESULTS Table 24-1 shows the concentrations of hydro- carbons found in Bering Sea animals; Table 24- 2 gives the dates and locations of sample collections. Seventeen samples of planktonic invertebrate material were analyzed from the following taxa: Neomysis rayi (a mysid), Parathemisto pacifica (an amphipod), Medusae (jellyfish), Chaetognatha (arrow worms), and Euphausiacea (euphausiids). The mysid sample was essentially devoid of hydrocarbons. However, the remainder of this group all contained substantial concentrations of pristane and several also contained heneicosahexaene (21:6). One sample of fish, Mal- lotus uillosus (the common capelin), was analyzed. Its principal hydrocarbon constituent was pristane. Livers from three species of sea birds were also investigated. These included Uria aalge (Common Murre), Uria lomvia (Thick-billed Murre), and Rissa tridactyla (Black-legged Kittiwake). The murre livers were quite low in hydrocarbons, but the kittiwake livers had a very high concentration of pristane. Samples of kidney were analyzed from four species of seal, Erignathus barbatus (bearded seal), Histriophoca fasciata (ribbon seal), Phoca vitulina richardsi (harbor seal), and P. vitulina largha (spotted seal). These species showed low to moderate concentrations of hydrocarbons, including pristane. Additional spotted seal tissues, including liver, muscle, and blubber, were also studied. While the hydrocarbon contents of these tissues were qualitatively similar to those of the kidneys examined, the liver and blubber tended to have substantially higher hydrocarbon concentra- tions. Ten of the thirteen seal tissues investigated also contained phthalate esters, synthetic organic compounds used in the manufacture of some plastics. DISCUSSION Pristane, which was found in all but three of the samples analyzed in this study, is a well-known marine biogenic hydrocarbon. Pristane has been identified in numerous species of zooplankton and is particularly abundant in Calanus (Blumer et al. 1963). Avigan and Blumer (1968) demonstrated that pristane is derived in calanoids from dietary chloro- phyll. Blumer (1967) has also shown that pristane can be retained by higher marine organisms which feed on zooplankton. A sizable body of subsequent work has confirmed the large extent to which pris- tane is produced and transferred without chemical alteration by marine animals (National Academy of Sciences 1975). Our finding that pristane is both ubiquitous and abundant in Bering Sea biota is further confirmation of this. The three highest concentrations of pristane, 220, 104, and 101 Mg/g, were observed in the liver of the black-legged kittiwake and in two samples of spotted seal blubber. This finding is consistent with the idea (Teal 1977) that higher marine animals without gills (birds and mammals) may accumulate higher hydro- carbon concentrations since they lack the efficient mechanism that gills afford for partitioning lipids into seawater. However, it must be noted that the highest concentration of pristane observed, 220 jUg/g in the kittiwake, is only a factor of three higher than the highest observed in an invertebrate. In the tissues of spotted seal which were examined individually, pristane concentration tends to be high in the blubber, intermediate in the liver, and low in muscle and kidney, consistent with general solu- bility. But for individual seals there axe exceptions to this trend. For instance, in animal 30-76 the lowest pristane concentration is found in the liver, while in animal 20-76 the concentration in the liver is higher than in the blubber. These apparent anomalies may reflect the different hydrocarbon turnover times of the tissues and the nutritional states of the individ- ual seals. Heneiocosahexaene, the all cjs-3,6,9,12,15,18 isomer (21:6), has been identified in marine phyto- plankton by Blumer et al. (1970) and independently by Lee et al. (1970). They showed that considerable variation exists in the concentrations of this com- pound produced by various species of algae and also in the extent to which 21:6 is accumulated by various species of zooplankton fed on phytoplankton rich in this compound. Subsequent work has shown that 21:6 occurs widely in marine algae (Blumer et al. 1971, Youngblood and Blumer 1973). Because 21:6 is labile chemically (Blumer et al. 1970) and un- doubtedly also biochemically, it is not accumulated by consumer organisms nearly to the extent that pristane is. Our finding that among the animals of the Bering Sea 21:6 is largely restricted to the zoo- plankton is consistent with these generalizations. Most of the seal tissues analyzed contained phtha- late esters. Phthalates have been reported in marine water, sediment, air, and biota (Giam et al. 1978); however, we cannot exclude the possibility that the phthalates in these tissues were introduced as contaminants after sample collection. These seal samples were not collected for hydrocarbon deter- mination; rather they were collected and saved as part of a study of the general biology and life history of I 1 i TABLE 24-1 Hydrocarbon concentrations (Mg/g, Wet Weight) in Bering Sea animals Total Total Material Sample number saturated unsaturated Pristane 21:6 Neomysis rayi 38-50 <0.01 <0.01 — — Parathemisto pacifica 50-60 70 20 70 — Parathemisto pacifica 21-36 68 4.1 68 3.4 Parathemisto pacifica 21-37 20 13 20 — Medusae 50-62 2.3 7.5 ■ 2.3 1.2 Chaetognatha 11-28 8.3 0.9 8.3 0.9 Chaetognatiia 2-4 60 10.3 60 9.0 Ciiaetognatha 5-18 53 2.3 53 2.3 Chaetognatha 7-24 34 8.9 34 7.6 Chaetognatha 11-29 7.5 0.6 7.5 — Chaetognatha 11-30 11.1 <0.01 11.1 — Euphausiacea 2-1 52 31 52 27 Euphausiacea 7-21 66 150 58 — Euphausiacea 7-23 58 100 49 88 Euphausiacea 21-38 9.5 34 5.7 26 Euphausiacea 44-52 3.6 34 1.2 — Euphausiacea 50-61 1.0 11 1.0 10.9 Mallotus villosus 38 10 1.1 10 — Uria aalge 77-30 <0.01 <0.01 — — Uria lomvia 77-37 0.4 0.6 0.4 — Rissa tridactyla 77-31 220 0.8 220 — Erignathus barbatus kidney 22-76 <0.01 0.25 — — ■ Histriophoca fasciata kidney 24-76 0.21 0.21 — Phoca vitulina richardsi kidney 12-76 0.95 0.52 — Phoca vitulina largha kidney 30-76 4.5 — 4.5 — muscle 30-76 0.85 — 0.85 — liver 15-76-1 66 — 66 — liver 15-76-2 65 1.7 65 — liver 20-76-1 30 2 30 0.1 liver 20-76-2 43 — 43 — liver 30-76 0.22 — 0.22 — blubber 15-76 104 — 104 — blubber 20-76 23 140 23 — blubber 30-76 140 — 101 — Phthalate 0.88 1.1 5.5 20 18 9.3 14 19 2.2 37 385 386 Chemical oceanography TABLE 24-2 Date and location of collection of samples for hydrocarbon determination Material Neomysis rayi Parathemisto pacifica Parathemisto pacifica Parathemisto pacifica Medusae Chaetognatha Chaetognatha Chaetognatha Chaetognatha Chaetognatha Chaetognatha Euphausiacea Euphausiacea Euphausiacea Euphausiacea Euphausiacea Euphausiacea Mallotus villosus Uria aalge Uria lomvia Rissa tridactyla Erignathus barbatus Histriophoca fasciata Phoca vitulina richardsi Phoca vitulina largha Phoca vitulina largha Phoca vitulina largha Sample Number Date Latitude N Longitude W 38-50 2 June 1977 60°27' 170°0l' 50-60 8 June 1977 57° 59' 168°49' 21-36 28 May 1977 59°3l' 174°48' 21-37 28 May 1977 59°3l' 174°48' 50-62 8 June 1977 57°59' 168°49' 11-28 23 May 1977 58°39' 172°15' 2-4 22 May 1977 54°42' 165°59' 5-18 22 May 1977 55°33' 168°09' 7-24 23 May 1977 56°05' 169°40' 11-29 23 May 1977 58°39' 172°15' 11-30 23 May 1977 58° 39' 172°15' 2-1 22 May 1977 54°42' 165°59' 7-21 23 May 1977 56°05' 169°40' 7-23 23 May 1977 56°05' 169°40' 21-38 28 May 1977 59°3l' 174°48' 44-52 6 June 1977 60°22' 169°07' 50-61 8 June 1977 57°59' 168°49' 38 2 June 1977 60°27' 170°0l' 77-30 25 May 1977 60°37' 174°38' 77-37 25 May 1977 60°37' 174°38' 77-31 25 May 1977 60°37' 174°38' 22-76 28 March 1976 56°12' 165°29' 24-76 19 April 1976 57°2l' 172°41' 12-76 25 March 1976 56°08' 164°20' 15-76 26 March 1976 56°05' 164°3l' 20-76 27 March 1976 56°05' 164°3l' 30-76 24 April 1976 56°05' 162°47' Alaskan marine mammals. The samples had been wrapped in aluminum foil, placed in plastic bags (a potential contamination source), and frozen. But we tend to believe that post-collection contamination is not the only source of these phthalates, since the smaller samples analyzed did not show the higher concentrations one might expect if various-sized pieces of tissue were contaminated from a constant source such as an outer plastic wrapper. Moreover, seals have been shown to accumulate chlorinated hydrocarbons (Holden and Marsden 1967). Considerable significance also lies in the groups of hydrocarbons which were not found in this suite of samples. The hydrocarbons associated with higher terrigenous plants were not observed. This group of hydrocarbons would include odd chain length normal alkanes with 23 to 31 carbon atoms. Their absence indicates that, at least in the spring, terrigenous Hydrocarbons of animals 387 hydrocarbon sources are minor compared to marine sources, in accord with the well-known high primary productivity of the Bering Sea. But this relationship may not continue throughout the year, since marine primary production is highly seasonal in northern waters. Shaw and Baker (1978) have documented that at roughly the same latitude, in Port Valdez, Alaska, pristane dominates the hydrocarbons of intertidal biota during the summer but terrigenous normcd alkanes dominate in winter. Another group of hydrocarbons not observed in the animals of the southeastern Bering Sea is the fossil hydrocarbons. This indicates that neither natural petroleum seepage nor pollution is resulting in significant accumulations of petroleum hydrocarbons in marine animals of the area. It should be noted, however, that the analytical techniques used in this study would not have detected low levels of poly- cyclic aromatic hydrocarbons. Since evidence has been presented that atmospheric transport has brought this class of compounds to the sediments of the Gulf of Alaska (Laflamme and Hites 1978) and the arctic (Shaw et al. 1979), their presence in the Bering Sea in trace amounts cannot be excluded. ACKNOWLEDGMENTS We thank F. Fay for seal tissues, G. Divoky for bird livers, and D. Mcintosh for mass spectral analy- ses. This study, Contribution No. 407, Institute of Marine Science, University of Alaska, was supported under contract number 03-5-022-56 between the Uni- versity of Alaska and the National Oceanic and Atmospheric Administration, to which funds were provided by the Bureau of Land Management. REFERENCES Avigan, J., and 1968 M. Blumer On the origin of pristane in marine organisms. J. Lipid Res. 9: 350-2. Blumer, M. 1967 Blumer, M., R. 1971 Blumer, M., M. 1970 Hydrocarbons in the digestive tract and liver of a basking shark. Science 156: 390-1. R. L. Guillard, and T. Chase Hydrocarbons of marine phytoplank- ton. Mar. Biol. 8:183-9. M. MuUin, and R. R. L. Guillard A polyunsaturated hydrocarbon (3,6, 9, 12, 15, 18 heneicosahexaene) in the marine food web. Mar. Biol. 6: 226-35. Giam, C. S., H. S. Chan, G. S. Neff, and E. L. Atlas 1978 Phthalate ester plasticizers : A new class of marine pollutant. Science 199: 419-21. Holden, A. V., and K. Marsden 1967 Organochlorine pesticides in seals and porpoises. Nature 216: 1274-76. Laflamme, R. E., and R. A. Hites 1978 The global distribution of polycyclic aromatic hydrocarbons in recent sedi- ments. Geochim. Cosmochim. Acta 42: 289-303. Lee, R.F., J. C. Nevenzel, G. A. Paffenhofer, A. A. Benson, S. Parson, and T.E. Kavanagh 1970 A unique hexaene hydrocarbon from a diatom (Skeletonema costatum). Biochim. Biophys. Acta 202: 386- 88. Blumer, M., M. M. MuUin, and D. W. Thomas 1963 Pristane in zooplankton. Science 140: 974. National Academy of Sciences 1975 Petroleum in the marine environment. Washington, D.C. 388 Chemical oceanography Shaw, D. G., and B. A. Baker 1978 Hydrocarbons in the marine environ- ment of Port Valdez, Alaska. Envir- on. Sci. Technol. 12: 1200-05. Shaw, D. G., D. J. Mcintosh, and E. R. Smith 1979 Arene and alkane hydrocarbons in nearshore Beaufort Sea sediments. Estuar. Coastal Mar. Sci. 9:435-49. Teal, J. M. 1977 Food chain transfer in hydrocarbons. In: Fate and effects of petroleum organisms in marine ecosystems and organisms, D. A. Wolfe, ed., 71-77. Pergamon Press, N. Y. Youngblood, W. W., and M. Blumer 1973 Alkanes and alkenes in marine benthic algae. Mar. Biol. 21: 163-72. Organic Geochemistry of Surficial Sediments from the Eastern Bering Sea M. I. Venkatesan, M. Sandstrom, S. Brenner, E. Ruth, J. Bonilla, I. R. Kaplan, and W. E. Reed Institute of Geophysics and Planetary Physics and Department of Earth and Space Sciences University of California Los Angeles, California I ABSTRACT The distribution and concentration of hydrocarbons in surficial sediments from the continental shelf of the eastern Bering Sea were determined as part of an environmental survey. Gravimetric and gas chromatographic analyses of the aliphatic fractions indicate that the hydrocarbons are predom- inantly allochthonous detritus, probably transported by rivers, with minor autochthonous elements. Gas chromatographic/ mass spectrometric characterization of aromatic fractions indicates that the source of aromatic compounds may be pyrolytic. The fact that autochthonous hydrocarbons in the sediments from this region noted for its high biological productivity are found in relatively small amounts suggests that rapid and efficient recycling of marine lipids has occurred within the water column or at the sediment-water interface. Comparison of hydrocarbon distribution patterns of open-shelf sediments with those of eelgrass and sediments from Izembek Lagoon indicates that coastal lagoons are probably not a major source of hydrocarbons in the south- eastern Bering Sea sediments. However, carbon isotopic composition of humic acids and protokerogens isolated from the lagoon and shelf sediments suggests a possible source relationship. It is possible that the terrigenous hydrocarbons in the Norton Sound region come mainly from the Yukon River. Nine samples from Norton Sound collected over a period of three years near suspected petroleum seepage show alkane distribution patterns which are not characteristic of weathered petroleum. Comparison of data from the Bering Sea region with those from the Southern California Bight, known to be petro- leum-contaminated, clearly shows that the Alaskan sediments are "clean" and generally free of petrogenic material. INTRODUCTION The eastern Bering Sea is noted for its high biolog- ical productivity and renewable biological resources. It is also a potentially important petroleum-producing area. And yet little is known about the distribution of organic matter in its sediments. In general, organic geochemical data from the Alaskan outer continental shelf are relatively scant except for the hydrocarbon studies of Peake et al. (1972), Kaplan et al. (1977, 1979) in the Gulf of Alaska and Bering and Beaufort seas, Shaw et al. (1978) on Beaufort Sea sediments, Chester et al. (1976) on sediments from Prince William Sound, the northeast Gulf of Alaska, Kinney (1973) in Cook Inlet and Port Valdez, and Wong et al. (1976) in the Beaufort Sea, along the Canadian border. In our laboratory, we have undertaken a study of hydrocarbon distribution and concentrations in the surficial sediments of the eastern Bering Sea. The analyses provide details of the natural hydrocar- bon background for the study of marine processes and for monitoring hydrocarbon pollution in con- junction with an increase in offshore petroleum production. In this chapter, we present results of the analyses of the aliphatic and aromatic fractions (along with other supplementary geochemical data) of approximately 20 sediment samples from the south- eastern Bering Sea area, a sample of eelgrass and the sediment within the eelgrass environment of Izembek Lagoon on the northwest side of the Alaska Penin- sula, and about 40 sediment samples from the Norton Sound region (northeastern Bering Sea). EXPERIMENTAL PROCEDURES Sampling The locations of the sampling stations are given in Figs. 25-la and b. The surface sediments were collected with steel Van Veen grab sampler during the summer 1975 cruises of the NOAA ship Discoverer, 389 390 Chemical oceanography Kuskokwim Bay 24,..i mmf- :**■ Bristo 160 Figure 25-la. Location of southeastern Bering Sea stations. Leg III, and with modified aluminum Van Veen grab sampler (Soutar 1976) in the 1976 and 1977 cruises of the R/V Sea Sounder. About 20 sediment samples, including two box -core samples, were collected in the summer 1979 cruise of the Discoverer. Care was taken during sampling to avoid contamination from the sampler or the ship. Except for a hinge in the sampler that could not be easily cleaned, all materials that came in contact with the sediment were cleaned and rinsed with organic solvents. Sediments were placed in prewashed glass jars and frozen until analy- sis. Carbon and sulfur analysis Elemental analysis of sulfur was carried out on freeze-dried sediment samples combusted in a LECO (Laboratory Equipment Corporation) Model No. 523 induction furnace and the resulting sulfur gases were titrated according to ASTM procedure E30-47, using a LECO Model No. 517 titrator. Total carbon and organic carbon (carbon remaining after treatment with 3N HCl) were determined by combustion of dry sediment samples in a LECO Model No. 572-100 semiautomatic, acid-base carbon determinator. De- tails of these procedures are given in the pertinent instruction manuals of LECO. Extraction of hydrocarbons The frozen sediment (southeastern Bering Sea) was placed in precleaned cellulose extraction thimbles and washed with distilled water to remove salts. The wet sediment was freeze-dried for 48 hours and extracted in a Soxhlet extractor for 100 hours with toluene: methanol (3:7) with one solvent change after 24 hours. Hexane extract of the water wash was com- bined with these extracts, reduced to a small volume on a rotary evaporator at 38 C, and treated with activated copper to remove sulfur (MacLeod et al. 1976). The extract was saponified by refluxing for four hours with 0.5 N KOH in a 1:1 mixture of water and methanol; the nonsaponifiable portion was then extracted into hexane and fractionated on a glass column packed with silica gel beneath neutral alum- ina (activity grade 1). A column with a length-to- Organic geochemistry of surficial sediments 391 169 167 165 163 161 159 8)) ^33 .70 .34 ^40 "^3 ^"^^^fAOn- ^44 298 ^0.154 ! • 166 ,47A 33A n-n 1 1 1 1 1 0 0 10 20 30 40 50 r niles niw _.i=-:l_ 1 1 1 65 61 169 167 165 163 161 Figure 25-lb. Locations of Norton Sound stations. •1976 stations; + 1977 stations; M979 stations. internal diameter (I.D.) ratio of 20:1 was packed with a weight ratio of 100 parts alumina and 200 parts silica to one part sample. Elution with two column volumes each of hexane, benzene, and methanol was carried out. The benzene fractions were further purified by thin-layer chromatography on silica gel plates developed in CH2CI2 to remove the methyl esters. The gas chromatographic analysis of the hexane fractions was carried out using a Hewlett- Packard Model No. 5830A instrument, equipped with a linear temperature programmer, FID detector, and electronic integrator. A glass SCOT column, 50 m X 0.4 mm I.D., coated with OV-101 (SGE Scientific, Inc.) was used. The column was tempera- ture-programmed from 100 C to 275 C at 2 C/min. with helium carrier gas flow of 4 ml/min. and held isothermally for 60 min. Norton Sound sediment samples were extracted with methanol for 24 hours to avoid freeze-drying, then with toluene: methanol for 76 hours and proc- 392 Chemical oceanography essed as discussed earlier for column chromatography after partitioning the methanol extract with hexane. Only silica gel was used for column fractionation and aliphatic and aromatic fractions free of methyl esters were eluted (Venkatesan et al. 1980). Glass capillary column coated with OV-101 (J&W), 30 m long and 0.25 mm I.D. was used in the Hewlett-Packard Model 5840A gas chromatograph. It was programmed from 35 C to 260 C at 4 C/min and then held isothermal for two hours. The flow rate of helium carrier gas was 3.6 ml/min. Selected samples were analyzed by gas chromato- graphy/mass spectrometry on a Finnigan model 4000 quadrupole mass spectrometer directly interfaced with a Finnigan Model 9610 gas chromatograph. The GC was equipped with a glass capillary column like the one described above for the analysis of aliphatic fractions. The aromatic fractions were analyzed using an SE 54 (J&W) column. The mass spectrometric data were acquired and processed using a Finnigan Incos Model 2300 data system. Solvents were of high purity grade (Burdick and Jackson "distilled in glass" grade). Trace organic compounds in double-distilled water were removed by passing through a column of Chromosorb 102. NaCl was heated at 500 C overnight and KOH fused at 500 C for two hours. Silica gel and alumina were sonicated twice with methylene chloride-methanol mixture and once with hexane prior to activation. RESULTS The eastern Bering Sea shelf has been arbitrarily divided in this study into two regions, southeastern Bering Sea and Norton Sound (northeastern area), for the sake of convenience in discussing the data. Results of analysis of 8 samples from about 20 samples collected in summer of 1979 from Norton Sound are also included. The range of organic carbon (0.14-1.3 percent) given in Tables 25-1 and 25-2, and sulfur (0.014-0.1 percent) are low compared to many fine-grained Recent marine sediments (Degens 1967, Didyk et al. 1978). Sample 48 from Norton Sound collected in 1977 has an anomalous organic carbon content (4.23 percent) and the total carbon content is also unusually high (9 percent). However, a sample collected very close to this station (160) in 1976 had a more normal carbon content of 0.7 percent. In summary, the data are typical of other unpolluted, relatively coarse marine sediments. The total hydrocarbons (Tables 25-1 and 25-2) range from 1 to 29 /xg/g of dry sediment with the exception of EBBS 35, which has a value of 241 Mg/g)- This range of values is low compared to unpoUuted Recent marine sediments from other environments (30 and 100 ppm) and is 10-100 times lower than the values of total hydrocairbon in sediments contaminated by petroleum (Palacas et al. 1976, Farrington and Tripp 1977, Crisp et al. 1979, Venkatesan et al. 1980). Representative histograms of alkane distributions and gas chromatographic traces of a few of the hexane fractions are illustrated in Figs. 25-2, 25-3, and 25-4. The range in total n-alkanes resolved by gas chromatography is from 0.1 to 9 pig/g dry sediment and reflects the variability in sediment grain size, dis- tance from the source, and organic carbon content. Although the chromatograms differ in detail, the hydrocarbon distributions are similar in many re- spects. An n -alkane series with carbon numbers of 16-34 is present in all sampling areas corresponding to those generally reported for marine sediments. The pronounced odd carbon preference (Tables 25-1 and 25-2) with /2-C27 or n-C2Q predominating (Fig. 25-3), is not characteristic of n-alkanes in petroleum. Pristane, phytane, and unidentified branched or unsaturated components are present generally in low amounts. Gas chromatographic traces of hexane fraction in eelgrass and sediment from Izembek Lagoon and isotopic analysis data on humic substance and proto- kerogens are presented in Fig. 25-5 and Table 25-3, respectively. Multiple homologous series of extended diterpanes and triterpanes were identified by GC/MS analyses and relative distribution histograms of a few stations are shown in Fig. 25-6. An example of a histogram of a sample from Southern California Bight is included for comparison. Concentrations of specific aromatic compounds in samples analyzed by GC/MS are given in Table 25-4. DISCUSSION Elemental analysis The organic carbon values are low considering the relatively high biological productivity of this conti- nental shelf region (Simoneit 1975), apparently as a result of a combination of oxic conditions at the sediment surface and a high-energy depositional environment (Sharma 1974). When the sediment type in samples from the south- eastern Bering Sea is classified according to grain size (sand, silt, and clay) after Shephard (1954), it is clear that sediments having lower organic carbon (0.3-0.4 percent) are sandy and those with higher organic carbon content (0.6-0.8 percent) are either silty or rich in silt-clay. Thus, the distribution of organic carbon here supports the findings of Bordovskiy V ea S o to S ' CO OJ ifl n C .C 2 05 H o S CO *^ b£ a; s u 0) CO o PhIcl, X X c o ^^ CO X3 pS O w 0) CO — « 3 CO .2 ttfi O CO 3. ■^1 O XI oi (D CO t- a> oi [> ■* r-i a> o csi i-i 00 CO cm' t^' o CM c~ 1-1 05 o C~; CO (TS 00 CO C^J CM Co' CO LC CO o OOiHiHOOOCT5COC~OOCOmU5C~ CMTfTtOOOi-OLOlOC^COOOt-C- CeOrJfCOCOCOCOCSlCSlCOCOCSlCSlCOrH cDooc^ .cDoa>ocooTt Tj; 00 a> 00 C-; cm ■*_ t>; r-I LO CO C CM* f-' CO* 00 cm' ■** r-i cm' •_ CD p in 1-1 i-| O tH -rf -^^ p CO C~ I>; CD p CD ■^' co' o •^' 1-i t-" co' co' TtH in t-' 00 lo lO o i-i cd o6 •<*' i-I COCOCMCOCOCMOOCMCMt-I 1— ICMCMCM-rtcoo>i-ii-icDCMr~ocM .ooi:~ihc~c-c~cd CM i-| t>; CO CO LO -^^ Tt<_ p CO CO CO -rj^ TJ p ■*_ CO CM C~; CD l> O O O CD O O CD O O CD O O O C CD O O O O O CD CD CO CJ5 C~ CD CO lO CO O lO CD Oi CD-rt cm' iH LO ■^' in Tt<' o ^' o cm' o cm' t-' o os oo' cm' cd' oJ oJ ^" CD 1-1 !>■ '^ O ^_ iH t- iH 00 c:i 05 ■* -* CO 00 ■<* CD 00 -*_ CO p p in co' co' t~' CD 00 o" in ■^' i-I i-i cm' t}<' cm' c~' o co cd' cm cd co' iH 00 l-( iH OOCMt-OS-rtiOOLnt^OOOrHCOCDi-lTtiCOOOOS-rt^ineQ i-ii-ii-icMCMcococo-^-*-<*TtiioinininincDcDLn 393 73 >i a ^ ^ oE o a. a> .a ^ o CJ '-w £ CO O o H CO ^ a •a to a c O 1) a> CO a X a a) =^ V. o ^ w 3 Si a CO C CO „ js a o -U .— ^J -O CO g CO O p c <1> c „ s: P5 o in T3 a; CO .S K^ to a &< o o a^ o " < X <1> c " CO Oi J2 > II -o T3 o c II 394 Chemical oceanography TABLE 25-2 Gravimetric and gas ciiromatographic data of Norton Sound sediment samples station Aliphatic Aromatic /?-alkanes Organic HC X ^q4 n-alkanes 4 Pr Odd No. fraction (Mg/g) fraction (/Jg/g) (Pg/g) carbon (%) OC OC Ph Even 1976 47 9.6 7.5 3.28 0.93 18.4 3.7 2.00* 5.38 49 24.8 4.1 5.69 1.12 25.8 5.1 1.50* 6.06 70 2.2 6.2 0.01 0.31 27.1 0.1 8.00 1.65 88B 3.9 5.7 0.69 0.53 18.2 1.3 2.14 4.11 105 1.8 0.9 0.07 0.93 2.9 0.1 n.d. 11.21 125 0.1 2.4 0.69 1.18 2.6 1.3 2.00 4.02 131 9.0 2.9 7.18 0.96 27.3 6.3 n.d. 2.80 137 17.8 4.5 8.69 n.d. n.d. n.d. 3.00* 4.07 147 6.8 2.3 2.24 0.33 27.5 6.8 1.80* 2.85 154 16.3 4.2 5.45 0.99 20.7 5.5 3.60* 5.69 156 7.1 5.5 5.06 1.30 9.7 3.9 2.67* 5.57 162 2.3 2.3 0.45 0.92 5.0 0.5 2.00* 4.75 166S 1.1 0.8 0.16 1.16 1.6 0.1 4.00* 5.16 168S 3.2 2.2 1.48 1.10 4.9 1.4 3.50* 5.26 169S 2.6 4.0 0.95 0.33 20.1 2.9 3.60 4.47 170S 4.4 2.2 2.57 0.52 12.8 4.9 n.d. 5.80 172S 10.9 3.8 2.89 0.87 16.9 3.3 n.d. 4.70 174S 3.9 2.0 1.79 0.82 7.2 2.2 6.00* 4.50 1977 34 0.8 0.7 0.09 0.12 12.4 0.8 2.0* 4.55 35 2.2 1.1 0.57 0.59 5.7 0.9 7.0* 5.15 39S 0.6 0.2 0.09 0.38 2.3 0.2 2.5* 5.35 4 IS 2.5 0.8 0.23 0.44 7.5 0.5 4.0* 4.78 42S 4.4 1.9 0.83 0.32 19.9 2.6 5.5* 4.78 43 1.0 1.7 0.38 0.60 4.5 0.6 6.5* 4.22 44 2.1 0.9 0.25 0.52 5.7 0.5 6.0* 3.21 48S 5.8 5.0 1.60 4.23 2.6 0.4 5.0* 6.37 147r 5.4 1.3 1.22 0.28 23.8 4.4 3.1* 5.26 1777 5.5 2.7 1.75 0.24 18.9 2.0 1.3* 5.12 17t 14.1 2.2 2.18 0.86 26.3 6.4 3.0* 5.67 iv§ 3.2 0.9 0.95 0.50 8.0 1.9 4.0* 5.34 1979 13 4.5 5.1 2.08 0.38 25.2 5.4 1.65 4.32 15 4.0 1.8 1.63 0.48 11.9 3.4 2.47 3.81 18 3.2 1.5 2.22 0.48 9.8 4.6 1.86 4.04 20 5.4 1.5 1.50 0.40 17.3 3.8 n.d. 4.03 22 8.9 5.5 2.83 0.86 16.8 3.3 2.44 3.35 25 5.2 1.9 1.57 0.66 10.8 2.4 n.d. 4.11 33A 1.7 0.9 0.29 n.d. n.d. n.d. n.d. 4.09 47A 1.4 1.1 0.44 n.d. n.d. n.d. n.d. 4.29 Aliphatic = eluted by hexane; Aromatic = elutcd by hexaneibenzene (3:2 by volume); * = approximate values based on peak heights. The rest are the same as those given in Table 25-1. Samples are 0-2 cm except B = bulk; S = surface; 77 = 0-3 cm; t = vibracore, 0-3 cm; § = 160 cm vibracore. Samples 14-17 belong to a different program (USGS). (1965) and Sharma (1974), that organic carbon increases with decreasing mean grain size of the sediment. Sediments in this region are reported to become progressively finer grained from nearshore to the edge of the shelf (Sharma 1974). The total hydrocarbon content of these sediments follows the same trend, with low concentrations of total hydro- carbons in coarse-grained sediments close to shore and higher concentrations in fine-grained sediments near the shelf edge (Table 25-1). The sulfur content of these sediments (0.01-0.13 percent) is quite low compared to other marine sediments (Didyk et al. 1978) and indicates relatively oxidizing conditions within the sediments. The organic carbon content of samples collected in 1977 (mostly offshore) from Norton Sound is gener- ally lower than those collected in the previous year (Table 25-2). The organic carbon and hydrocarbons are higher in nearshore sediments. Apparently the organic carbon content in this region, unlike that of the southeastern Bering Sea sediments, is generally related to the distance from the presumed terrigenous source, the Yukon River. The total hydrocarbon/organic carbon (HC/OC) ratio of these samples varies between 0.0002 and 0.005 (Tables 25-1 and 25-2) and is in the range re- ported for other unpolluted sediments (Pcdacas et al. 1976). The only exception is EBBS 35 which has an HC/OC ratio of 0.059, similar to those obtained in the Gulf of Mexico (Gearing et al. 1976) and South- m 80 60 40 20\- ^ oLil EBBS 8 S to T3 cc o o o 15 24 /50 130 //O 90 10 50- 50- 10- 0 EBBS 57 50- 40- 50- 20 \0 0 15 24 EBBS 58 ^ 55 55 15 15 24 500 200 m 0 EBBS 64 ■ 1 1 1 ii 55 \J \5 24 55 500 .EBBS 65 200 - m - n . . M i: : L /5 24 55 " \5 24 55 ^ \b CARBOM NUMBER 24 55 Figure 25-2. Histograms of n-alkanes (-) and Cjg and C20 isoprenoids (— -) distributions in selected southeastern Bering Sea surficial sediments. 395 vu EBBS 37 IS NS 174 EBBS 35 IS. Station 823 San Pedro Basin Station 727 Tanner Bank, West Figure 25-3. Gas chroma- tograms of hexane fraction from southeastern Bering Sea (EBBS) and Norton Sound (NS) sediments. Station 823 (contaminated witli weath- ered petroleum) Venkatesan et al. 1980; Station 727 (contaminated with fresh petroleum), Kaplan et al., 1976. Numbers 15-33 refer to carbon-chain length of n-alkanes. Pr: pristane. UCM: unresolved complex mixture. I.S.: Internal Standard hexa- methyl benzene. 396 a> CO c CC o 15 2^ 55 CARBON NUMBER Figure 25-4. Histograms of n-alkanes (-) and C19 and C20 isoprenoids (— -) distributions in selected Norton Sound surficial sediments. 1976 stations = 47, 88, 131, 156, and 166; 1977 stations = 41 and 14; 1979 stations = 13 and 15. 397 nC|5 nC|7 nC|9 EELGRASS ZOSTERA MARINA HEXANE PR. J^-J uu Um. ^ 100 200 TEMPERATURE (°C) 275 IZEMBEK LAGOON HEXANE FR. C21 C23 C25 ^27 •29 ^^-VJ '31 100 ^ 200 275 TEMPERATURE (°C) Figure 25-5. Gas chromatograms of hexane fractions extracted from eelgrass (Zostera marina) leaves and surficial sediments from Izembek Lagoon. Numbers 15-31 refer to carbon-ciiain length of «-alkanes. 398 I Organic geochemistry of surficial sediments 399 TABLE 25-3. Carbon isotopic analysis of humic acids isolated from eastern Bering Sea and Izembek Lagoon surficial sediments Sample 613C PDB Source EBBS 17 EBBS 24 EBBS 35 EBBS 65 Izembek Lagoon Eelgrass -21.00/00 -21.50/00 -21.10/00 -21.50/00 -I8.40/00 -10.30/00 Peters (unpublished) Peters (unpublished) Peters (unpublished) Peters (unpublished) Stuermer et al. (1978) McConnaughey and McRoy (1979) em California (Venkatesan et aL 1980), suggesting pollution of the sediment. The total n-alkanes to organic carbon (n-alkanes/OC) ratio is less than 0.0007 for all samples (Tables 25-1 and 25-2). Much higher ratios would be expected if unweathered petroleum were present in the sediments (Palacas et al. 1976). It should be stressed that this criterion is only useful in indicating the presence of unweathered petroleum in sediments. The absence of resolvable n-alkanes in sample 35 (Alk/OC = 0) could, however, indicate the presence of weathered petroleum, as in sediments in nearshore basins of Southern California reported by Venkatesan et al. (1980). Gas chromatographic data Southeastern Bering Sea The gas chromatographic data indicate that alloch- thonous lipids are the predominant source of hydro- carbons in shelf sediments (Fig. 25-2). This material may be derived from the Kuskokwim or Nushagak Rivers or possibly the coastal lagoons along the shores of the Alaska Peninsula. The drainage basins are in mixed spruce-alder woodlands, which are possible sources of the high molecular weight hydrocarbons in the shelf sediments. The contribution of organic detritus and associated hydrocarbons from the coastal lagoons will be discussed later. The abundance of aUochthonous hydrocarbons is generally related to the amount of total sedimentary organic material rather than distance from the presumed source— the two rivers. This is evident from the correlation of total n-alkanes with organic carbon. Both organic carbon and n-alkanes occur at their highest concen- trations in the fine-grained sediments near the center and edge of the shelf (i.e., stations 54, 64, and 38). It is surprising to find the aUochthonous alkanes distri- buted throughout the shelf sediments, since the surface currents in the eastern Bering Sea sweep Kuskokwim River water north to Norton Sound and Bering Strait. The deposition of aUochthonous alkanes in the shelf sediments may in part be related to processes involving ice formation and spring breakup. The n-alkane distributions of the samples analyzed from the Bering Shelf are quite different from sedi- ments contaminated with weathered or fresh petro- leum (Fig. 25-3). The homologous series of isopre- noids is not found in the samples. The fact that pristane is much more abundant than phytane (Pr/Ph ratios range from 2 to 18, Table 25-1) suggests that the isoprenoids are derived from varying amounts of biogenic materials rather than from petroleum. The relatively high concentration of pristane in Station 65 may have been derived from the presence of Calanus species, which are rich in pristane (Blumer et al. 1964), in the overlying water column. These subarc- tic oceanic copepods have been reported to be predominantly distributed in the shelf region, north of the Pribilof Islands (Motoda and Minoda 1974). These samples lack UCM, unlike sediments known to have been contaminated by petroleum (Farrington et al. 1977, Crisp et al. 1979, Venkatesan et al. 1980). The only exception is sample 35 as presented in Fig. 25-3, whose gas chromatogram is characterized by a broad UCM in the entire elution range, demon- strating no resolved hydrocarbons. This pattern is typical of weathered petroleum contamination and is similar to the sediments from Southern California nearshore basins (Venkatesan et al. 1980), a conclu- sion consistent with the anomalous HC/OC ratio of this sample. The source of these hydrocarbons may be natural submarine seepage, although none have been reported in the southern Bering Sea shelf. However, faults in this area (Marlow et al. 1976) could allow leakage of petroleum from underlying reservoir rocks. The extensive fishing operations around this area make an anthropogenic origin for these hydrocarbons possible; it seems unlikely, however, since the chromatogram from station 37 (Fig. 25-3) and other samples from this area of 400 Chemical oceanography TABLE 25-4 Poly cyclic aromatic hydrocarbons in sediment samples analyzed by GC/MS (ng/g) Southeastern Bering Sea Norton Sound (1976) Norton Sound (1977) Norton Sound (1979) Station 59 Station 64 Station 131 Station 166 Station 35 Station 43 Station 25 O-Xylene - - - T T - 0.2 Isopropylbenzene - - - - -- - - n-Propylbenzene - - - T T - - Indan - - - - - - - 1,2,3,4-Tetramethylbenzene - - - T - - - Naphthalene T T T T T - T 2-Methylnaphthalene T T T T T T T 1-Methylnaphthalene T T T T T T T Biphenyl - - T T T T - 2,6-Dimethylnaphthalene - - T T - - -- Dimethylnaphthalenes' - - T T - - 0.1 Trimethylnaphthalenes - - T T T T 0.3 Fluorene - - T - T T 0.1 Dibenzothiophene - - -- - -- - - Phenanthrene 0.3 1.1 4.2^ 0.2 1.9^ 0.8^ 0.4 Anthracene - - - - T T -- Methylphenanthrenes 0.5 1.3 0.6 T - 0.2 0.8 Fluoranthene - 3.0 T 0.1 0.5 T T Pyrene 1.4 5.0 2.2^ 0.1^ 0.2 T 0.2 Benz(a)anthracene - - - - T - - Chrysene 1.2 0.9 3.0^ 0.6 T 0.4 1.0 Benz(e)pyrene - T T T - - T Benz(a)pyrene - -- -- -- T T -- Perylene T T 9.8 T T 1.3 9.4 Simonellite - - -- See Pyrene T T T Cadalene T - 0.3 T - -- T Retene 1.8 7.3 2.8^ T 3.1^ 1.4 T = trace ' Excludes 2,6-diniethylnaphthalene when identified ^ Coelutes with unknown compound ^ Coelutes with simonellite intensive fishing do not show evidence of such petroleum contribution. The hydrocarbon distribution of samples from stations 12 (Fig. 25-2) and 45B resembles the two- component mixture of sources found in western Gulf of Alaska sediments (Kaplan et al. 1977), namely an allochthonous component represented by the odd C25-C31 n-alkanes, a nonselective distribution of /t-alkanes from C17 to C25 with a UEirrow UCM maximizing in the C22 region. The homologs < n-C could be of marine origin and consist of residues from primary production {n-Cn and n-Cig : Han and Calvin 1969) and from microbially-altered algal detritus (Johnson and Calder 1973, Cranwell 1976, Hatcher et al. 1977). These biolipids could also have been derived from the bacteria themselves (Lijmbach 1975). Further, these two stations at the southern edge of the study area are far from the influence of surface currents transporting terrigenous material from the Kuskokwim River. This may explain why these samples have predominant autochthonous alkane distribution. The predominance of allochthonous biolipids at stations 17, 37, and 38 demonstrates the importance of depositional environment in determining the distribution of hydrocarbons that ultimately accumu- late in the sediments. The stations near the edge of the continental shelf and south of the Pribilof Islands are in an area of high biological productivity and would be expected to have a high contribution of marine-derived autochthonous lipids as in Walvis Bay (Boon et al. 1975) and Cariaco Trench (Simoneit 1975). Yet, unlike these anoxic environments, in the oxidizing and high-energy depositional envi- ronment of the eastern Bering Sea (Sharma 1974) efficient recycling of autochthonous lipids takes place within the water column or at the sediment-water interface. This degradation and recycling may be the result of physical processes, such as strong currents, waves, or ice-gouging, as well as biological processes, such as burrowing by benthic fauna, which cause extensive reworking of the deposited lipids. The fact that this process apparently leaves the sediment enriched in allochthonous alkanes suggests that these I Organic geochemistry of surficial sediments 401 hydrocarbons could be associated with biologically refractory organic material. Eelgrass The dominant biota of the coastal lagoons along the shores of the southeastern Bering Sea is eelgrass, which could contribute nutrients and organic detritus to the shelf (Barsdate et al. 1974, McConnaughey and McRoy 1979). Samples of eelgrass and the sediments within eelgrass lagoons were analyzed to characterize their hydrocarbon distribution for comparison and correlation with those of outer shelf sediments. Gas chromatograms of hexane fraction from surface (0-10 cm) sediment and eelgrass leaves from Izembek Lagoon are illustrated in Fig. 25-5. The hexane fraction from eelgrass consists almost exclu- sively of a simple mixture of n-alkanes, n-Cis , n-C-^-j , and n-Ci9 and only small amounts of hydrocarbons beyond n-C^^. The distribution is quite different from that found in most other higher plants (Eglinton and Hamilton 1963); it may be the result of the age of the plants and abundance of heterotrophic bac- teria, or it may simply mean that this species of subtidal vascular plant does not produce the cuticular wax. It is interesting that the n-alkane distribution is similar to that reported for many species of algae; this may indicate that the alkanes are derived from epiphytic and/or macrophytic algae living on the surface of eelgrass leaves, although they are reported to be seldom above 6 percent of the eelgrass biomass in the main eelgrass beds (McConnaughey and McRoy 1979). Hydrocarbon distribution of the surface sediment within the lagoon (Fig. 25-5) indicates that detritus derived from the decomposition of eelgrass leaves cannot be correlated with the allochthonous hydrocEirbon distributions found in the shelf sedi- ments. The chromatogram is characterized by a mixture of n-alkanes from C17 to C31 , with predomi- nant alkanes n-C2i and n-C23 and a slight UCM in the n-C22 region, which could be a result of microbial decomposition of the eelgrass detritus (Johnson and Calder 1973). This is distinctly different from the hydrocarbon distribution in shelf sediments, which are characterized by a predominance of C25-C31 odd carbon-numbered n-alkanes. Carbon isotopic composition of humic acids and protokerogen isolated from surface sediments of the Bering Sea measured by K. Peters (University of California at Los Angeles, personal communication) and that of humic acids isolated from the surface sediment of Izembek Lagoon by Stuermer et al. (1978) are listed in Table 25-3. The humic acids and protokerogens of surface sediments should have carbon isotopic values chEiracteristic of land plants, since the hydrocarbons appear to be predominantly terrigenous (assuming the same source for all of them). However, the shelf sediment humic acids have carbon isotopic values from —21 to — 22°/oo (Table 25-3), in between those expected for marine and terrigenous organic carbon (Nissenbaum and Kaplan 1972, Gearing et al. 1977, Craig 1953). Humic acid from Izembek Lagoon has a carbon isotopic value of — 18°/oo. This is isotopically lighter than the values of — lO^/oo reported by Smith and Epstein (1970) for Zostera marina (whole plant) from Point Mugu Lagoon, California, and — 10.3°/oo reported by McConnaughey and McRoy (1979) for eelgrass from Izembek Lagoon; it probably indicates a large contribution of planktonic algal and bacterial carbon to the sediment. Indeed, the 5 i3C values of lagoon animals including zooplankton, infauna, and nekton range from — 12°/oo to — 23.1°/oo heavier than the average Si^c value for Bering Sea phyto- plankton (— 24.4°/oo: McConnaughey and McRoy 1979). These data thus indicate that eelgrass and microalgal carbon may be selectively consumed by a wide variety of organisms. Moreover, the particulate organic carbon (POC) in the lagoon ranges from 5 = -19.9 to -I2.6O/00, and POC collected at the mouth of the lagoon from -22.1 to -17.30/oo. Thus, the isotopic composition of humic acids from the lagoon appears to represent a mixed contribution of eelgrass, plankton, and epiphyte detritus to the lagoonal sediments. Therefore, the values for 6 i^C of organic carbon between —21 and — 22*^/oo measured in the shelf sediments could also be the result of mixing a terrigenous component and a component from the lagoon containing eelgrass detritus. Thus, although hydrocarbons characteristic of the eelgrass lagoons are not found in the eastern Bering Sea shelf sedi- ments, the more resistant humic substances and protokerogens derived from the eelgrass detritus could be transported to the shelf environment. Norton Sound A representative gas chromatogram and histograms of alkane distribution given in Figs. 25-3 and 4 illustrate the predominance of 23, 25, 27, 29, and 31 carbon n-alkanes with a maximum at n-C^-j , indicat- ing that land plants contribute most of the organic matter (Table 25-2). Stations 49, 131, and 137, in the sound and closer to shore, have the highest content of lipids and n-alkanes (Table 25-2). The lipid and n-alkane contents progressively decrease along the northeast to southwest transect (i.e., from stations 131 to 147, Fig. 25-lb). The southwest stations 154 and 156, near the mouth of the Yukon River, are richer in organic content and n-alkanes. These may have been derived from terrigenous silt resuspended from the Yukon prodelta which extends 402 Chemical oceanography across the mouth of the sound. Hydrocarbons from marine plankton in this region are low. Further south, around stations 162 to 168 and west, offshore in stations 39 to 44 and 47A, the terrigenous detri- tus is diluted by open-ocean sedimentation. The M-alkane content of the 1977 samples is also lower than that of 1976 and 1979 samples as expected because of the increased distance of the former stations from the shore. Station 49, near Pastol Bay, is richer in lipid content and has nearly four times the n-alkane content of station 48, even though these stations are equidistant from the shore. This might be explained by the fact that because station 48 is in a higher- energy region than station 49, dispersion of organic matter into the open ocean may occur more readily. The northern region of Norton Sound seems to be poorer in lipids and also in n-alkanes (Stations 70, 88, 105), possibly because there are no major sediment sources near the area and the longshore current movement and sediment transport are directed north toward the Bering Strait. Pristane and phytane are found in trace amounts at a few stations. The very low quantities detected indicate that zooplankton species like Calanus cope- pods, which contribute substantial pristane to a few stations north of the Pribilof Islands in the south- eastern Bering Sea, may be impoverished in this area. Nine samples from stations 47, 172, 174 (1976), 14-17 (1977), and 18 and 22 (1979), collected near the location where petroleum seepage was suspected (Cline and Holmes 1979) show alkane distribution patterns (Fig. 25-4) uncharacteristic of petroleum. Gas chromatographic /mass spectrometric data In stations 8, 41, 59, and 65 from the southeastern Bering Sea, the compounds eluting between the normal alkanes from C21 to C27 have been identified as C21 polyolefins and C23-C27 mono- and diolefins. These compounds probably represent a contribution from marine organisms. Branched olefins have also been identified in a few samples. Stations 25, 43, 131, and 169 in Norton Sound contain alkenes such as CigHae and a cyclic alkene, C25H46. The predom- inant alkene at station 169 is a C17 H34 alkene eluting between n-Cig and n-Cn ; pristene is also found in significant amounts. In general, stations from off- shore contain more alkenes than those nearshore, indicating that these alkenes are of marine origin. Retene (I, Appendix) and simonellite (II) are present at levels below 3 ppb in all these stations, the former more abundantly than the latter. These diterpenoids are molecular markers derived from terrigenous resinous plants (Simoneit 1977). Cada- lene (III), a sesquiterpenoid residue found in minor amounts in the sediments at these stations, can be of mixed marine and terrigenous origin. This compound has also been identified in petroleum (Bendoraitis 1974), and in shales and siltstones from the North Atlantic Ocean (Simoneit and Mazurek 1978a, b). Multiple homologous series of cycloalkanes (e.g., diterpanes, steranes, and triterpanes) could be of biogenic or petrogenic origin. The extended diter- panes and most triterpanes exhibit a strong fragment ion (base peak) at m/z 191 in their mass spectra. Examples from the eastern Bering Sea and an exam- ple from Southern California Bight are shown in Fig. 25-6. The inferred skeleton for the homologs is structure IV, based on the structure proposed for the m/z 191 fragment ion of the series (Anders and Robinson 1971). The tricyclic diterpanes range from C19H34 to C26H48 in stations EBBS 35, NS 43, 131, and 169 and also in EBBS 65 (Simoneit and Kaplan 1980). In a few other stations, such as EBBS 8 and NS 17, diterpenes are more dominant than diter- panes, where only one or two homologs were de- tected. In most of the stations, except EBBS 35, the triter- penoids appear to be derived mainly from bacteria or algae (DeRosa et al. 1971, Cardoso et al. 1976), and consist of 17/3(H)-hop-22(29)-ene (diploptene, V), hop-17(21)-ene (VI), 17|3(H)-22,29,30-trisnorhopane (VH, R=H, 17/3), and 17/3(H)-hopane (VII, R=C3H7) and the series of extended 17i3(H)-hopanes ranging from C31 to C33 , with only minor amounts of the 17a(H)-hopanes (VIII). The Alaskan sediments contain predominantly C27 , C30, and C31 17/3, 21/3 triterpanes. Several C30 triterpenes with a double bond in addition to diploptene have also been de- tected. Most of the C29 triterpanes found are not hopanes and their identity has not yet been deter- mined. A C30H50 triterpene with base peak 69 and a molecular ion 410 a.m.u. (not a hopene) is present at most of the stations and is also as abundant as n-C^o ■ The reason for its occurrence is not known. The extended hopanes (>C3i) are present as single C-22 diastereomers. The presence of predominantly 17i3(H) stereomers and of the triterpenes which are present in living organisms indicates that these compounds are of recent biogenic origin. The pres- ence of only small quantities of 17a(H) stereomers suggests that there is no input from petroleum components (Dastillung and Albrecht 1976, Simoneit and Kaplan 1980, Venkatesan et al. 1980). For purposes of comparison, an example of Recent sediment from San Pedro Basin, Southern California Bight, is given in Fig. 25-6, where the triterpanes are predominantly the 17a (H) homologs, and the ex- Organic geochemistry of surficial sediments 403 cn ^ E - C O c CD O c o O (D > Triterpenoids -|7a(H) -170(H) - mono-ene Extended diterpanes I I I I I I I 20 EBBS 8 III ;:: •ii III •II III •II •ii tir^ 25 30 I I I I I I I I I I 35 40 Triterpenoids . -|7a(H) ' •-170(H) •- mono-ene NS 17 Extended diterpane f I I I I I I I I I I I Y I frt 15 20 25 30 I I I I I I I 35 40 'C t- Triterpenoids :-^l7a(H),22 RandS --I7/3(H) Extended EBBS 35 diterpanes Triterpenoids -|7a(H),22 RandS -170(H) San Pedro Basin California Extended diterpanes 'C °c Figure 25-6. Relative distribution histograms of diterpenoids and triterpenoids based upon m/z 191 mass chromatograms. EBBS = southeastern Bering Sea; NS = Norton Sound; San Pedro Basin = 20-25 mm core, Venkatesan et al. 1980. Dia- stereomers indicated by dotted and continuous Unes. I is 17a; (H), 18a(H), 21j3(H)-28, 30-bisnorhopane. tended hopanes (VIII) are present as 1:1 mixtures of the C-22 diastereomers (Venkatesan et al. 1980). There, the dominant homolog is 17a(H), 18i3(H), 21i3(H)-28, 30-bisnorhopane (C28H48) which has been proposed to be a molecular marker of Southern California petroleum (Seifert et al. 1978, Simoneit and Kaplan 1980). Stations from the eastern Bering Sea contain no C28 triterpenoid. The triterpenoidal distribution of station NS 17 (Fig. 25-6) clearly indicates that it is not contaminated with petroleum, although it is believed to be near a suspected petro- leum seep. The only exception is station EBBS 35 (Fig. 25-6), which shows a triterpenoidal distribution very similar to petroleum-contaminated Southern California sediments. This is consistent with the observed «-alkane distribution pattern of this station, typical of weathered petroleum. This sample consists predominantly of 17a(H) homologs and the extended hopanes are present as 1:1 mixtures of the C-22 diastereomers. The C28 bisnorhopane found in 404 Chemical oceanography Southern California petroleum is also found in this sample, but is much less abundant than in Southern California Bight sediments (Venkatesan et al. 1980). The resolved polynuclear aromatic compounds (PAH) are <1.0 iug/g sediment in these stations, much less than those found in Beaufort Sea sediments (Kaplan et al. 1979). Concentrations of selected PAH compounds are presented in Table 25-4. Generally, biphenyl, naphthalenes, phenanthrene, fluorene, fluoranthene, pyrene, chrysene, and traces of their alkyl substituted homologs have been detected by GC/MS analyses. Benzo(E)pyrene has also been found in all the samples. In general, the parent PAH compounds are more abundant than their alkylated homologs (Table 25-4), indicating that they come from pyrolytic sources rather than from crude oil shales (Coleman et al. 1973, Youngblood and Blumer 1975). A more detailed study is required before any positive conclusions can be reached on sources and mechanisms of distribution. Perylene is more abundant (~10 ppb in station 131) in nearshore sediments than in offshore sedi- ments (traces in stations 35 and 166). Perylene could be generated in situ by transformation of some terrestrial precursor compound (Bergmann et al. 1964). Quinones transported from soils and originat- ing from plants could form perylene (Aizenshtat 1973) in reducing micropaleoenvironmentary condi- tions as described by Welte and Ebhardt (1968), even though the overall sediment is oxidizing. Coronene and 1,12 benzoperylene have also been detected in these samples. CONCLUSIONS The alkanes in sediments of the study area gener- ally show a biomodal distribution typical of a mix- ture of allochthonous and autochthonous sources. In short, the absence of unresolved complex mixture and the distribution pattern of n-alkanes and ex- tended triterpanes in most of the stations studied can be attributed to recent biogenic sources characteristic of unpolluted environments. The allochthonous lipids, the primary source of hydrocarbons in the surface sediments, are probably transported to the continental shelf by river discharge and erosion and redistribution of surface sediments. Correlation of the hydrocarbon distribution of the eastern Bering Sea with the hydrocarbons extracted from eelgrass (Zostera marina) and sediments from within Izembek Lagoon indicates that the latter environment may not be a significant source of hydrocarbons in the outer shelf sediments. However, carbon isotopic analysis of humic and kerogenous substances from the lagoon and shelf sediments indicates that these biologically refractory organic materials may be transported to the shelf environ- ment. The presence of relatively smaU amounts of auto- chthonous hydrocarbons in the sediments, in spite of the high biological productivity of the region (espe- cially the southeastern Bering Sea), suggests rapid and efficient recycling of marine lipids within the water column or at the sediment-water interface. Presence of higher concentrations of relatively labile hydrocar- bons derived from the autochthonous sources, identified in only a few stations on the Bering Sea shelf, may be important in an assessment of the fate and effects of petroleum products introduced into this marine environment. Any petroleum contam- ination and deposition in those environments may last a long time. The stations in Norton Sound suspected to be near natural gas seeps do not show hydrocarbon distribu- tion patterns characteristic of petroleum. The distribution of PAH compounds is more complex than that of n-alkanes and appears to be predominantly pyrolytic, possibly derived from forest fires. ACKNOWLEDGMENTS We would like to thank Dr. B. R. T. Simoneit for the identification of the triterpenoids from GC/MS analyses and D. Meredith for technical assistance. This study was performed with the support of NOAA contract No. 03-6-022-35250, and is Contribution No. 2069, Institute of Geophysics and Planetary Physics, U.C.L.A. ADDENDUM: Recently Kvenvolden et al. (1979) have reported the seep near Station 17, Norton Sound, to be thermogenic in origin with CO2 as its major component. APPENDIX: Structures cited in the text I. retene, CiqHiq II. simonellite, C19H24 III. cadalene, C15H 18 IV. extended diterpanes R=C2H5-C|2H25 < V. diploptene, C30H50 VI. hop-l7(2l)-ene,C3oH5o rT^R VII. 17^8 (H), 2l/3(H)-hopanes R= H, C2H5, CjHy VIII. extended 17a (H), 2l/3(H)-hopanes R-CH3-C5H11 405 406 Chemical oceanography REFERENCES Aizenshtat, Z. 1973 Perylene and its geochemical signifi- cance. Geochim. Cosmochim. Acta 37: 559-67. Cardoso, J., P. W. Brooks, G. Eglinton, R. Good- fellow, J. R. Maxwell, and R. P. Philip 1976 Lipids of recently -deposited algal mats at Laguna Mormona, Baja California. In: Environmental biogeochemistry, J. O. Nriagu, ed., 149-74. Ann Arbor Science Publishers, Ann Arbor, Mich. Anders, D. E., and W. E. Robinson 1971 Cycloalkane constituents of the bitu- men from Green River shale. Geo- chim. Cosmochim. Acta 35: 661-78. Barsdate, R. J., M. Nebert, and C. P. McRoy 1974 Lagoon contributions to sediments and water of the Bering Sea. In: Oceanography of the Bering Sea, D. W. Hood and E. J. Kelley, eds., 553- 76. Inst. Mar. Sci., Occ. Pub. No. 2. Univ. of Alaska, Fairbanks. Bendoraitis, J. G. 1974 Hydrocarbons of biogenic origin in petroleum— aromatic triterpanes and bicyclic sesquiterpanes. In: Advances in organic geochemistry, 1973, B. Tissot and F. Bienner, eds., 209-14. Editions Technip, Paris. Bergmann, E. D., R. Ikan, and J. Kashman 1964 The occurrence of perylene in Huleh Peat. Israel J. Chem. 2: 171-2. Blumer, M., M. M. Mullin, and D. W. Thomas 1964 Pristane in the marine environment. Helgolander Wiss. Meeresuntersuch. 10:187. Boon, J. J., J. W. deLeeuw, and P. A. Schenk 1975 Organic geochemistry of Walvis Bay diatomaceous ooze— I. Occurrence and significance of the fatty acids. Geochim. Cosmochim. Acta 39: 1559-66. Chester, S. N., G. H. Gump, H. S. Hartz, W. E. May, S. M. Dyszel, and D. P. Enagonio 1976 Trace hydrocarbon analysis: The NBS Prince William Sound /NE Gulf of Alaska Baseline Study, NBS Tech. Note 889. Cline, J. D., and M. L. Holmes 1979 Submarine seepage of natural gas in Norton Sound, Alaska. Science 198: 1149-53. Coleman, H. J., J. E. Doley, D. E. Hirsch, and C. J. Thompson 1973 Compositional studies of a high- boiling 370-535° C distillate from Prudhoe Bay, Alaska, crude oil. Anal. Chem. 45: 1724-37. Craig, H. 1953 The geochemistry of stable carbon isotopes. Geochim. Cosmochim. Acta 3: 53-92. Cranwell, P. A. 1976 Decomposition of aquatic biota and sediment formation: Organic com- pounds in detritus resulting from microbial attack on the alga, Ceratium hirundinella. Freshwater Biol. 6: 41- 8. Crisp, P. T., S. Brenner, M. I. Venkatesan, E. Ruth, and I. R. Kaplan 1979 Organic chemical characterization of sediment-trap particulates from San Nicolas, Santa Barbara, Santa Monica and San Pedro Basins, California. Geochim. Cosmochim. Acta 43 (in press). Bordovskiy, O. K. 1964 Accumulation and transformation of organic substances in marine sedi- ments. Mar. Geol. 3:3-82. Dastillung, M., and P. Albrecht 1976 Molecular test for oil pollution in surface sediments. Mar. Poll. Bull. 7: 13-15. w I Degens, E. T. 1967 Diagenesis of organic matter. In: Diagenesis in sediments, G. Larsen and G. V. Chilinger, eds., 343-89. Elsevier, N.Y. DeRosa, M., A. Gambacorta, L. Minale, and J. D. Bu'Lock 1971 Bacterial triterpenes. Chem. Comm. 619-20. Didyk, B. M., B. R. T. Simoneit, S. C. Brassell, and G. Eglinton 1978 Geochemical indicators of paleoen- vironmental conditions of sedimenta- tion. Nature 272: 216-22. Eglinton, G., and R. J. Hamilton 1963 The distribution of alkanes. In: Chemical Plant Taxonomy, T. Swain, ed., 187-96. Academic Press, N.Y. Farrington, J. W., N. M. Frew, P. M. Girchwend, and B. W. Tripp 1977 Hydrocarbons in cores of Northwes- tern Atlantic coastal and continental margin sediments. Est. Coastal Mar. Sciences: 793-808. Farrington, J. W., and B. W. Tripp 1977 Hydrocarbons in Western North At- lantic surface sediments. Geochim. Cosmochim. Acta 41: 1627-41. Gearing, P., J. N. Gearing, T. F. Lytle, and L. S. Lytle 1976 Hydrocarbons in 60 northeast Gulf of Mexico shelf sediments: A prelimi- nary survey. Geochim. Cosmochim. Acta 40:1005-17. Gearing, P., E. F. Plucker, and P. L Parker 1977 Organic carbon stable isotope ratios of continental margin sediments. Meir. Chem. 5: 251-66. Han, J., and M. Calvin 1969 Hydrocarbon distribution of algae and bacteria, and microbiological activity in sediments. Proc. Nat. Acad. Sci, U.S.A. 64: 436-43. Organic geochemistry of surficial sediments 407 Hatcher, P. G., B. R. Simoneit, and S. M. Gerchakov 1977 The organic geochemistry of a Recent sapropelic environment: Mangrove Lake, Bermuda. In: Advances in organic geochemistry, 1975. R. Campos and J. Goni, eds., 469-84. Enadimsa, Madrid. Johnson, R. W., and J. A. Calder 1973 Early diagenesis of fatty acids and hydrocarbons in a salt marsh envi- ronment. Geochim. Cosmochim. Acta 37: 1943-55. Kaplan, I. R., P. Mankiewicz, W. E. Reed, E. Ruth, M. Sandstrom, and M. I. Venkatesan 1976 High molecular weight hydrocarbon analysis in sediments. Ann. Rep. to B.L.M. Year 1, Vol. III. Rep. 3.1. NTIS. Kaplan, I. R., W. E. Reed, M. W. Sandstrom, and M. I. Venkatesan 1977 Characterization of organic matter in sediments from Gulf of Alaska, Bering and Beaufort Seas. Ann. Rep. to U.S. Dep. of Commerce, NOAA. Kaplan, I. R. M. I. Venkatesan, S. Brenner, E. Ruth, J. Bonilla, and D. Meredith 1979 Chciracterization of organic matter in sediments from Gulf of Alaska, Bering and Beaufort Seas. Ann. Rep. to U.S. Dep of Commerce, NOAA. Kinney, P. J. 1973 Baseline hydrocarbon concentration. In: Environmental studies of Port Valdez, D. E. Hood, W. E. Shiels, and E. J. Kelley, eds., Inst. Mar. Sci., Occ. Pub. No. 3. Univ. of Alaska, Fair- banks. Kvenvolden, K. A., K. Weliky, H. Nelson, and D. J. Des Marais 1979 Submarine seep of carbon dioxide in Norton Sound, Alaska. Science 205: 1264-6. Lijmbach, G. W. M. 1975 On the origin of petroleum. Proc. 9th World Petrol. Cong. 2: 357-69. 408 Chemical oceanography MacLeod, W. D., D. W. Brown, R. G. Jenkins, L. S. Ramos, and V. D. Henry 1976 A pilot study on the design of a petroleum hydrocarbon baseline investigation for northern Puget Sound and Strait of Juan De Fuca. NOAA Tech. Mem. ERL MESA-8: 40. Marlow, M. S., H. McLean, T. L. Vallier, D. W. Scholl, J. V. Gardner, and R. Powers 1976 Preliminary report on the regional geology, oil and gas potential and environmental hazards of the Bering Sea Shelf south of St. Lawrence Island, Alaska. U.S.G.S. Open File Rep. 76-785. McConnaughey, T., and C. P. McRoy 1979 ^^C label identifies eelgrass (Zostera marina) carbon in an Alaskan estu- arine food web. Mar. Biol. 53: 263-9. Motoda, S., and T. Minoda 1974 Plankton of the Bering Sea. In: Oceanography of the Bering Sea, D. W. Hood and E. J. Kelley eds., 202- 42. Inst. Mar. Sci., Occ. Pub. No. 2. Univ. of Alaska, Fairbanks. Nissenbaum, A., and I. R. Kaplan 1972 Chemical and isotopic evidence for the in situ origin of marine humic substances. Limnol. and Oceanogr. 17: 570-82. Palacas, J. G. Roberts 1976 P. M. Gerrild, A. H. Love, and A. A. Baseline concentrations of hydrocar- bons in barrier-island quartz sand, northeastern Gulf of Mexico. Geol- ogy 4:81-4. Peake, E. , M. Strosher, B. L. Baker, R. Gossen, R. G. McCrossan, C. J. Yorath, and G. W. Hodgson 1972 The potential of Arctic sediments: Hydrocarbons and possible precursors in Beaufort Sea sediments. Proc. Inter. Geol. Cong. 24th Session, Montreal. 5: 28-37. Seifert, W. K., J. M. Moldowan, G. W. Smith, and E. V. Whitehead 1978 First proof of Cgg -pentacyclic triter- pane in petroleum. Nature 271: 436-7. Sharma, G. D. 1974 Contemporary sedimentary regimes of the Eastern Bering Sea. In: Oceano- graphy of the Eastern Bering Sea, D. W. Hood and D. J. Kelley, eds., 517-40. Inst. Mar. Sci., Occ. Pub. No. 2. Univ. of Alaska, Fairbanks. Shaw, D. G., P. Else, R. Gritz, G. Malinky, G. Mapes, D. Mcintosh, J. Schwartz, E. Smith, and J. C. Wiggs 1978 Hydrocarbons: Natural distribution and dynamics on the Alaskan outer continental shelf. Ann. Rep. to U.S. Dep. of Commerce, NOAA. Shephard, F. P. 1954 Nomenclature based on sand-silt-clay ratios. J. Sed. Petrol. 24: 151-3. Simoneit, B. R. T. 1975 Sources of organic matter in oceanic sediments. Ph.D. Dissertation, Univ. of Bristol, England. 1977 Diterpenoid compounds and other lipids in deep-sea sediments and their geochemical significance. Geochim. Cosmochim. Acta 41:463-76. Simoneit, B. R. T., and I. R. Kaplan 1980 Triterpenoids as molecular indicators of paleoseepage in Recent sediments of the Southern California Bight. Mar. Environ. Res. 3: 113-28. Simoneit, B. R. T., and M. A. Mazurek 1978a Lipid geochemistry of Cretaceous sediments from Vigo Seamount, DSDP/IPOD Leg 47B. In: Initial reports of the Deep Sea Drilling Project, 47B, W. B. F. Ryan et al., eds. 565-70. U.S. Govt. Printing Office, Washington, D. C, in press. 1978b Search for eolian lipids in the Pleisto- cene off Cape Bojador and lipid geochemistry of a mudstone, DSDP/ IPOD, Leg 47 A. In: Initial reports of the Deep Sea Drilling Project, 47 A, W. B. F. Ryan et al., eds. U.S. Govt. Printing Office, Washington, D. C, in press. Smith, B. N., and S. Epstein 1970 Biogeochemistry of the stable isotopes of hydrogen and carbon in salt marsh biota. Plant Physiol. 46: 738-42. Organic geochemistry of surficial sediments 409 Soutar, A. 1976 Collection of benthic sediments for chemical analysis. Rep. to B.L.M. Southern California benthic and water column baseline research, III. Rep. 1.1.NTIS. Welte, D. H., and G. Ebhardt 1968 Distribution of long chain n -paraffins and n -fatty acids in sediments from the Persian Gulf, Geochim. Cosmo- chim. Acta 32: 465-6. Stuermer, D. H., K. E. Peters, and I. R. Kaplan 1978 Source indicators of humic substances and proto-kerogen. Stable isotope ratios, elemental compositions and electron spin resonance spectra. Geo- chim. Cosmochim. Acta 42: 989-97. Wong, C. S., W. J. Cretney, R. W. MacDonald, and P. Christensen 1976 Hydrocarbon levels in the marine environment of the southern Beaufort Sea. Tech. Rep. No. 38 to Beaufort Sea Project. Dep. of the Environ- ment, Victoria, B. C. Venkatesan, M. I., S. Brenner, E. Ruth, J. Bonilla, and I. R. Kaplan 1980 Hydrocarbons in age-dated sediment cores from two basins in the Southern California Bight. Geochim. Cosmo- chim. Acta 44:789-802. Youngblood, W. W., and M. Blumer 1975 Polycyclic aromatic hydrocarbons in the environment: Homologous series in soils and Recent marine sediments. Geochim. Cosmochim. Acta 39: 1303-14. Hydrocarbon Gases in Near-surface Sediment of the Northern Bering Sea Keith A. Kvenvolden, George D. Redden, Devin R. Thor, and C. Hans Nelson -U.S. Geological Survey Menlo Park, California ABSTRACT Methane, ethane, ethene, propane, propene, n-hutane, and isobutane are common in bottom sediment of the northern Bering Sea. At eight sites the content of methane rapidly increases with depth within the first four meters of sediment. These concentration gradients and absolute methane concen- trations indicate that the interstitial water of the near-surface sediment at these sites may be gas saturated. These gas-charged sediments may be unstable, creating potential geologic hazards and, in certain areas, causing the formation of seafloor craters. The isotopic compositions of methane at four of the sites range from —69 to 80** /oo (e'^CpuB)- This range of values clearly indicates that the methane derives from microbial processes, possibly within the near-surface Pleistocene peat deposits that are common throughout the northern Bering Sea. At one site in Norton Sound, near-surface sediment is charged with CO2, accompanied by minor concentrations of hydro- carbons, that is seeping from the seafloor. Methane in this gas mixture has an isotopic composition of — 36**/oo, a value that suggests derivation from thermal processes at depth in Norton Basin. The presence of sediment charged with methane or CO 2 cannot in general be predicted from analyses of surface sedi- ment, which usually contains hydrocarbon gases and CO2 at low concentrations. Sampling beneath a sediment depth of about 0.5 m is generally required to detect high concentrations of gas. Acoustic anomalies detected on high-resolution seismic records indicate the presence of gas-charged sediment, but gas analyses of sediment samples from areas with these anomalies do not always confirm that high concentrations of gas are there. Conversely, high concentrations of methane are some- times found at sites where no acoustic anomalies are obvious on high -resolution records. INTRODUCTION About twenty years ago Emery and Hoggan (1958) described the occurrence of hydrocarbon gases in near-surface marine sediment from Santa Barbara Basin, off southern California. These anoxic sedi- ments contain methane, ethane, propane, butanes. pentanes, and hexanes, with methane being one to almost five orders of magnitude greater in concentra- tion than any of the other hydrocarbons. Geochemi- cal studies that followed have generally focused on methane and the processes that can account for its occurrence and distribution in a variety of aquatic sediments (Reeburgh 1969, Whelan 1974, Martens and Berner 1974, Claypool and Kaplan 1974, Orem- land 1975, Barnes and Goldberg 1976, and Kosiur and Warford 1979). Recently Bernard et al. (1978) described the distribution of methane, ethene, propane, and propene in shelf and slope sediment in the Gulf of Mexico, and Kvenvolden and Redden (1980) reported on the occurrence of these gases in sediment from the outer shelf, slope, and basin of the Bering Sea. The present study examines inner shelf areas of the Bering Sea and considers the hydrocar- bon gases methane (Ci ), ethane (C2 ), ethene (C2:i ), propane (C3), propene (C3:i ), isobutane (i-C4), and n-butane (n-C^ ) in sediment of the inner Bering Shelf in Norton Sound and the adjacent eastern Chirikov Basin (Fig. 26-1). Norton Sound is an elongate, east-west trending bay in the western coast of Alaska bounded on the north by the Seward Peninsula, on the east by the Alaskan mainland, on the south by the Yukon Delta, and on the west by the Chirikov Basin. The floor of the sound is very flat, and the water averages about 20 m deep. To the west in the Chirikov Basin water depths increase to about 50 m, especially in the northern part of the basin at the Bering Strait. When sea level fell in late Pleistocene time, the floor of Norton Sound and the eastern Chirikov Basin became exposed (Nelson and Hopkins 1972). During 411 + -1^ j; „ CD A CM T— f^ 03 " p cn cn'^ io 00 b ,_ i'^ct), 6 '^ CM CO f^ o CM ",n <>o '■"- SS v mm •i^y^'?- ^~ IT) CD CM CO CD ininvvoo p ^oiooC^Jco fl^S?;22 ?t 2t^i^V^ o • .^•c^co, 7 V o9°co CO 6 s CD o Or O 6 pO 66 7 7-18(0 5(0-44 8-4(0 3(110- 8-2(0- (12-22 N 6 o cb -« CO CD -*7 CD COfficD p CO • CD 6 • £ 3 ^ CM CO H^^ in 3 ~ o '^S <^ 3; Uj CO CD C ai a; x: >i o cu " si 0) % ^ T-l "^ OJ — c S ca ^ c" (- .ti o CO .*- ^ i >.S o _ 'S > ■S Id §-^ 0) > CQ QJ c — c S c« — 3 0) C == « S O a _ o CQ Xi U3 CS J= O O) «« = N o o >> S "O c .2 0) OS I'll o « & J § s cd T^ "I ti-i ^D ^ O O) -S S s -I 4i2 Hydrocarbon gases in near-surface sediment 413 this time fluvial processes and tundra vegetation characterized the area (Hopkins 1967), and peaty mud was deposited over much of the region. This mud contains 2-8 percent organic carbon. As sea level rose during latest Pleistocene time, mcirine sedimentation resumed. In Holocene time, fine- grained, sandy silt derived mainly from the Yukon River blanketed the area with a cover up to 10 m thick (McManus et al. 1977, Nelson and Creager 1977). In contrast to the nonmarine sediment, the organic content of the overlying sediment ranges only from 0.5 to 1.0 percent (Nelson 1977). METHODS The procedures used for this work consisted of analyses of sediment samples for hydrocarbon gases and carbon dioxide on board ship and measurements of the carbon isotopic compositions of organic matter, Cj, and CO2 in shore-based laboratories (Nelson et al. 1978, Kvenvolden et al. 1979a, b). Sediment samples were recovered by vibracores and surface grab-samplers during three summer field seasons in 1976, 1977, and 1978. Gases were ex- tracted from sediments in the following manner: volumes (about 0.5 1) of wet sediment were extruded into 0.95 1 double-friction-seal cans that had two septum-covered holes near the top. Helium-purged distilled water was added to each can until a 100-ml headspace remained. Each can was closed with a lid, and the headspace was purged with helium through the septa. The cans were vigorously shaken for ten minutes to release gases into the headspace. Exactly one millimeter of the headspace gas mixture was analyzed by gas chromatography using both flame ionization (for hydrocarbons) and thermal conducti- vity (for Ci at high concentrations and for CO2 ) detectors. Concentrations of gases were determined by measurements of peak heights on chromatograms and with quantitative standards. Partition coeffici- ents were used to correct for the different solubilities of the hydrocarbon gases. Concentrations are re- ported in nl. 111, or ml per liter of wet sediment. This method of extraction yields semiquantitative results, but, because all samples were processed in the same manner, the results can be compared. The carbon isotopic composition of organic matter, Ci , and CO2 were determined by mass spectrometry. Before mass spectrometry Ci and CO2 were separated in a vacuum line-combustion appara- tus modified after Craig (1953). Results are reported in parts per mil (°/oo) as: si3r^ _ "C/'^C sample - ^^C/^^C standard v innn C/'^C standard where the standard is Peedee belemnite (PDB). There were insufficient concentrations for carbon isotopic determinations of C2 and higher hydrocarbons. RESULTS Ci is the most abundant hydrocarbon gas found in the first five meters of sediment in Norton Sound and the eastern Chirikov Basin. Fig. 26-2 shows the geo- graphic distribution of maximum concentrations of Ci. At eight sites this concentration exceeds 1 ml/1, and at five of these sites (8-4, 8-8, 8-15, 8-21, and 8-22) concentrations exceed 10 ml/1. At the other stations the maximum amount of C^ measured was less than 100 /il/1 with two exceptions, at 7-33 and 8-6, where it was about 200 //l/l at each site. The amount of Ci found depends to some extent on the depth of core, for concentrations of Cj generally increase with depth. Thus, surface samples and short cores usually have lower amounts of C^ than do samples taken from greater depths. The vertical distribution of Cj at the eight sites mentioned above is shown in Fig. 26-3. For these sites the concentra- tion of Ci increases by four or five orders of magni- tude within the top four or five meters of sediment. Also shown is the distribution of methane at sites 7-17 and 8-3. The data for these sites are combined because the sites are at essentially the same location sampled during two different field seasons. Here the Ci concentrations, especially in deeper samples, aire much lower than at the eight sites. The major gas at 7-17/8-3 is not Cj but CO2. The second most abundant hydrocarbon gas in this area is C2 ; C3 concentrations are usually slightly lower and generally parallel C2 concentration profiles with depth. The geographic distribution of C2 + C3 is shown in Fig. 26-4. The maximum concentrations of these two hydrocarbons are usually less than 1 //l/l. At two sites, 8-4 and 8-17, maximum C2 + C3 con- centrations are slightly higher (1.2 and 1.1 ixljl, respectively), but at site 7-17/8-3, C2 + C3 maximum concentration is almost 8 jul/l- At the eight sites where Cj shows 4-5 orders of magnitude increase in concentrations with depth, C2 + C3 increased by about two orders, but the profiles of concentration are more variable (Fig. 26-5) than the Ci concentra- tion profile (Fig. 26-3). The concentrations of C2 + C3 in samples from 7-17/8-3 are much higher than in all other samples. Both C2:i and C3:i are present in all samples analyzed. Concentrations are variable, but as a gen- eral rule C2:i exceeds C2 in surface samples and with increasing depth the reverse is observed. At site 7-17/8-3, C2 is much more abundant than C2:i , with 414 o LL o CO -z. o < LU o o o c o 1-1 O 2 c S3 Q. fl OJ ITl > U3 3 f ) c (1) a; S c o X! n -4^ c )H •^ o T1 Cu 9^ CS c/l > 3 <1> s 3h U) . r. F c o -o <4^ if. -4^ O o C4-I >-c o ^ x: u a c w )-l h (U o eft CO Ol CO •o cc c CM CO ?* ■n 3 c 3 U Uh M 416 CONCENTRATIONS OF C2 + C3 nUL wet sediment 50 100 150-- 200 E o H 250-- Q. UJ O 300-- 350-- 400-- 450-- 500-- H \ — I I I I I H \ \ 1 I I I I Figure 26-5. Graph of concentrations of C^ + Cj in nl/1 and m1/1 of wet sediment vs. depth (cm) for sediment samples from cores taken at nine sites in Norton Sound and eastern Chirikov Basin. 417 418 Chemical oceanography the C2/C2:i ratio reaching a maximum of 340 at a depth of 60 cm. A similar relation holds for C3.1 and C3. At the surface C^-i is usually more abun- dant than C3, and the reverse is true for deeper samples. At 7-17/ 8-3, C3 is always more abundant than C3.1— by a factor of 47 at a depth of 200 cm. The concentrations of i-C4 and n-C4are lower than the lighter hydrocarbon gases and in many cases reach the limit of detection of the method, about 2 nl/1. In general the concentrations of i-C4 + n-C4 are less than 100 nl/1, and the distribution with depth is variable. In samples from 7-17/8-3, concentrations of i-C4 + A2-C4 reach a maximum of 14 jul/1 at a depth of 200 cm. Hydrocarbons from C5 to C7 are measured in a single backflush peak from chromatography and are designated C5+. C5+ hydrocarbons commonly occur in high concentrations in surface samples from this area and particularly in core samples from 7-17/8-3. BIOGENIC METHANE The occurrence in anoxic sediment of high concen- trations of Ci resulting from microbial decomposi- tion of organic matter is well established (Emery and Hoggan 1958, Barnes and Goldberg 1976, Reeburgh and Heggie 1977, and Kosiur and Warford 1979). Cj is both produced and consumed by microorganisms, and models for these processes have been devised (Claypool and Kaplan 1974, Martens and Berner 1974, Barnes and Goldberg 1976, Kosiur and Warford 1979). In less reducing sediments of open marine en- vironments, Ci is also present but at concentrations as much as five orders of magnitude less than those observed in anoxic marine sediments (Bernard et al. 1978, Kvenvolden and Redden 1980). Although there is much less Ci in sediments of open marine environments, the processes that generate gas are probably similar to those in anoxic sediments, but much slower. At eight sites in Norton Sound and the eastern Chirikov Basin abundances of C^ increase by four or five orders of magnitude within the first five meters of sediment, reaching concentrations near or ex- ceeding saturation of the interstitial water. These shallow sediments are probably anoxic, and the Ci probably is being generated by the decomposition of peaty mud that contains 2-8 percent organic carbon, and is buried under marine sediment of lower carbon content (0.5-1.0 percent). This sediment cover is thickest near the front of the Yukon Delta and thins to the north (McManus et al. 1977, Nelson and Creager 1977, Nelson 1977). The depth of burial of the peaty mud may account for the groupings of the Ci concentration profiles shown in Fig. 26-3. Seven of the sites (6-121, 6-125, 6-131, 8-4, 8-8, 8-15, and 8-21) have profiles that group together. These seven sites are located in the eastern and northern parts of Norton Sound and in the Chirikov Basin near Port Clarence (Fig. 26-2). In these areas, peaty sediment is buried under about 2 m of sandy silt. The seven profiles show maximum concentrations below about 1.5 m. Therefore, if peaty mud is the source of the methane, the depth of its burial accounts for the depth at which high Ci concentrations are found. In contrast, one C^ concentration profile (8-22) reaches maximum values below 3 m. This site was northwest of the Yukon Delta in the southern part of Norton Sound, where the sediment cover is thicker and the peaty mud is more deeply buried. Thus there is a correlation between the depth of buried organic matter and the depth at which Cj concentrations reach high values. At site 7-17/8-3, Ci concentra- tions follow a different trend to be discussed below. That the high concentrations of Ci at eight sites result from microbiological processes is supported by both chemical and isotopic data. Higher molecular weight hydrocarbons accompany Ci , and the ratio Ci /(C2 + C3 ) can be used as a guide to interpret mode of formation. Likewise, the carbon isotopic composition of Cj can be used to interpret process of formation (Bernard et al. 1976, 1977). Microbial degradation of organic matter produces hydrocarbon gases with Ci/(C2 + C3) ratios greater than 1,000, and with S'^Cipdb lighter than -6OO/00. Table 26-1 shows these parameters for samples from the eight sites. On the basis of the criteria stated above, it is clear that the Cj at the eight sites was derived from microbiological processes, and the buried peaty mud, in which the organic carbon has an isotopic composition of —28^/00 (Kvenvolden et al. 1979a, b), is the likely source. Other sites may exist in Norton Sound and eastern Chirikov Basin where Ci concentrations exceed 1 ml /I. Finding these locations will require sampling below about 1 m of sediment (3 m or more off the Yukon Delta), because the occurrence of high amounts of Cj at shallow depths is not manifest at the surface. Either the surface layer of sandy silt seals the Ci , preventing its migration to the surface, or the rate of consumption or diffusion of C^ in the upper meter is very rapid, leading to low concentra- tions of Ci at the surface. At two sites, 8-6 and 7-33 (Fig. 26-2), maximum concentrations of Ci of 224 and 196 jul/1 respectively, may hint that much higher concentrations are present at greater depths. At site 7-33, the deepest sample (Fig. 26-1) came from 70 I TABLE 26-1 Ci/(C2+C3) ratios and 5' ^Ci values for samples containing Ci concentrations in excess of 1 ml /I Maximum Site Ci/CCa+Ca) 8-4 24000 8-8 71000 8-15 28000 8-21 440000 8-22 88000 6-121 6500 6-125 28000 6-131 5400 5'^Ci *(0/oo) -80^ nd nd nd nd -72^ -69^ -75^ *relative to the PDB standard ^ Kvenvolden et al. (1979a, b) 2 Nelson etal. (1979) cm. If this sample, containing 196 lul/l, were plotted on Fig. 26-3, it would fall within the envelope of Cj - concentration profiles of cores in which C^ exceeds 1 ml/1 at depth. The case for site 8-6 is not so clear. The sample containing 224 n\/\ comes from a depth of 220 cm (Fig. 26-1). If plotted on Fig. 26-3, this value would fall below the envelope of Ci -concentra- tion profiles. At site 8-6, peaty mud may be more deeply buried than at other sites in northern Norton Sound. Only deeper sampling can directly verify the presence of higher amounts of Cj . At other sites in Norton Sound and eastern Chirikov Basin, Ci concen- trations are below 100 ij.\/l and at many sites below 10Ml/l(Fig. 26-2). POSSIBLE BIOLOGIC ORIGIN OF OTHER HYDROCARBONS Besides Ci , other hydrocarbon gases are present in these sediments, but in much lower quantities than Ci . The maximum concentrations of C2 + C3 are in the same range as the minimum concentrations of Cj . At sites where Ci increases rapidly with depth (Fig. 26-3), C2 + C3 also generally increases (Fig. 26-5), but much more slowly than Cj . Concentrations of i-C4 and n-C^ are even lower than concentrations of C2 + C3, and the i-C4 + n-C4 concentrations are erratic with depth. As a generalization, however, the abundances of the higher hydrocarbons, C2 , C3 , and C4 , are greater in samples where concentrations of Ci are greater. Therefore, the processes that produce Ci may also be responsible in part for the generation of the higher hydrocarbons. Microbiological production and consumption provides a reasonable mechanism to Hydrocarbon gases in near-surface sediment 419 account for Cj at the eight sites where Ci concentra- tions increase beyond 1 ml /I. Therefore, microbial processes may also explain the occurrence of the higher molecular weight hydrocarbons, although evidence for this process remains circumstantial. Laboratory experiments have demonstrated the microbial formation of C2 and C3 (Davis and Squires 1954). Thus there is support for the suggestion that the C2 and the C3 hydrocarbons at these sites can come from microbial processes, but there is no precedent in the microbiological literature for the production of C4 hydrocarbons. The presence and distribution of C2:i are probably controlled by biological processes, but these processes likely differ from those which account for the very high Ci concentrations. These unsaturated hydro- carbons have been formed by microbial action in the laboratory (Davis and Squires 1954), and C2:i is produced in soils by bacteria (Primrose and Dilworth 1976). In the sediment the process seems to take place uniformly, because there is no obvious concen- tration gradient with depth. In surface samples concentrations of C2:i and C3.1 are higher re- spectively than concentrations of C2 and C3 . With depth, concentrations of C2 and C3 increase slightly so that below the surface, ratios of C2/C2:i and C3/C3.1 are usually equal to or greater than one. THERMOGENIC HYDROCARBONS The above discussion focused mainly on sites where Ci concentration increases rapidly with depth and consideration has been given to the heavier hydrocarbons associated with this Cj . At one site, 7-17/8-3, however, Ci concentrations are not unusu- ally high— less than 100 [xljl (Fig. 26-3)— but con- centrations of C2 + C3 (Fig. 26-5) and i-C4 + n-C4 are unusually large relative to concentrations seen else- where in the sediments of this area, or for that matter, anywhere else in marine sediments off Alaska. Site 7-17/8-3 has been studied in great detail since anomalous hydrocarbon concentrations were first discovered in the water column at the site (Cline and Holmes 1977). Nelson et al. (1978) showed that the sediments here also contain anomalous hydro- carbon concentrations. Kvenvolden et al. (1979a, b) confirmed the hydrocarbon chemistry and discovered that the major gas component within the sediment and escaping into the water column is COg. The Ci /(C2 + C3 ) ratios in sediment at this site are less than 10 and the S'^C^ is — 36°/oo. These num- bers differ greatly from those discussed earlier where microbiological processes were inferred. The hydrocarbons at site 7-17/8-3 are probably 420 Chemical oceanography derived from thermochemical processes, judging from the molecular distribution of Ci , C2 , and C3 and the isotopic composition of Cj . In addition, anoma- lously high concentrations of i-C4 , n-C^^ , and C5+ (gasoline-range hydrocarbons) support the mechanism of thermochemical processes (Kvenvolden and Claypool 1980). Heat for this process must be available at depth in Norton Basin. The hydro- carbons resulting from the thermal decomposition of organic matter within the basin must migrate along with CO 2 up fault zones to the surface and escape as a seep. Although the hydrocarbon chemistry indi- cates that the hydrocarbons at site 7-17/8-3 probably migrate from depth, the concentration profiles (Figs. 26-3 and 26-5) indicate that special conditions of migration must exist. The fact that hydrocarbons are leaking into the water column suggests that surface sediments should contain large amounts of these hydrocarbons. On the contrary, surface samples at this site contain very low concentrations of hydro- carbons. In fact, surface samples (0-10 cm) at this site show no evidence of the high concentrations of hydrocarbons deeper in the sediment. The gradients of Ci (Fig. 26-3), C2 + C3 (Fig. 26-5), and i-C4 + n-C4 decrease rapidly toward the sediment surface. This rapid decrease and lack of significant quantities of hydrocarbons at the sediment surface can be explained either by rapid diffusion of hydrocarbons into the water column from the first few centimeters of sediment or by the presence of discrete gas vents that pipe the hydrocarbons through the sediment, leaving few hydrocarbons remaining in the sediment. The second explanation is more reasonable, because active gas vents were seen by television in 1978 (Kvenvolden et al. 1979a). The concentration profiles (Figs. 26-3 and 26-5) at site 7-17/8-3 reach a maximum value at a depth of about 1-2 m and then decrease. This profile suggests that migration from greater depths does not involve diffusion of hydro- carbons within the underlying sediment but rather that the hydrocarbons are following distinct conduits such as faults. Near the surface the hydrocarbons are dispersed into the sediment where they eventually vent along with CO2 into the water. That the hydro- carbons from this seep are present in the water column has been documented by Cline and Holmes (1977). The waters of Norton Sound and the eastern Chirikov Basin also contain a regional distribution of hydrocarbon gases (Cline et al. 1978), the sources of which, in part, may be the underlying surface and near-surface sediment. GEOPHYSICAL EVIDENCE The presence of gas in near-surface sediments can cause acoustic anomalies on high-resolution geophysi- cal records where the gas is no longer in solution in the interstitial water but takes the form of bubbles. Schubel (1974), for example, demonstrated how high concentrations of gas affect acoustic properties of sediments. At the eight sites where Cj concentra- tions exceeded 1 ml/1 and may have reached and exceeded interstitied water solubility, bubble-phase Ci may be present. Acoustic anomalies would be expected on high-resolution records from these sites if free gas is indeed present. Geophysical transects, using 800-J boomer, 3.5 kHz sub bottom profiler, and 120 kJ sparker systems, indicate that near-surface acoustic anomalies are widespread in Norton Sound. Fig. 26-6 shows those sites, sampled for hydrocarbon gases, at which acoustic anomalies are seen on geophysical records. At three sites (6-125, 8-4, and 8-21), acoustic anoma- lies correspond to samples having high concentrations of Ci . The acoustic anomaly and associated Ci at 8-4 were discussed in detail by Kvenvolden et al. (1979 a, b). The characteristics of the acoustic anomaly at 8-21 suggest that the cause may be controlled more by the presence of glacial till de- posits than by gas. At sites 7-17/8-3, geophysical records show both near-surface and deeper acoustic anomalies. There the sediment is charged with CO2 rather than Cj , and the CO2 is escaping from the seafloor as a submarine seep, observed acoustically and by television (Kvenvolden et al. 1979a). At five sites (6-121, 6-131, 8-8, 8-15, and 8-22) on Fig. 26-6, where high concentrations of C^ were measured, no acoustic anomalies were detected. At sites 6-121 and 6-131 no geophysical records were obtained; it is uncertain, therefore, whether or not acoustic anoma- lies are present at these sites. At sites 8-8, 8-15, and 8-22, high-resolution geophysical records show no evidence of acoustic anomalies, although the geo- chemical measurements indicate high concentrations of Ci (Figs. 26-2 and 26-4). Apparently the gas concentrations at these sites were not in the proper range to produce anomalies on the records of the geophysical systems employed. On the other hand, acoustic anomalies were observed on records at sites 8-1, 8-9, and 8-10, but maximum Cj abundances in cores at these sites were only 6, 4, and 15 iJil/l, respectively. Concentrations this low are not ex- pected to produce acoustic anomalies. In addition, acoustic anomalies were found in geophysical records at sites 6-168, 7-22, and 7-25, but sampling at these sites was not deep enough (Fig. 26-1) to test the presence of high Cj or CO2 concentrations at depth. High concentrations of Ci in near-surface sediment may cause instability and may lead to crater forma- H ^ H _, - E A E . .^iiiii;^ z o ° - . ■ •.-, (iiiiiCl" :.'■ r-'-^0^^^ ""^^ ' A . . ■ v V*' viv- ■■" ^.iiiiii^S ^^< CO < z a: ^w < < \ < 5 z O 1- < S ■ ■"■"■^^ \ ^^ 3 'C o z o o cr z o . -i.-iji ? V- o o h^ -^ •.;:# ^ o CM Eo z < o o -J ' ■'.- \ ■^ /r'^SJ;.,'., o X ^ ::5 Vft CO CD Kiii - < 5 o o o S CO •'■• vT\ - ::) ^ s.^ff' ^^ %.- •' »i/V- "- CO iSSi # o CO J. -w w 4- CO J, \ CO O CM I S o 4- N^ ; ^ • 'v 00 O 00 CD •J® Vv -- y CO # ' -j- CJQ 4- % llo ^•■J \^/ ^ ^ :^ o ^- Co r^ + Uj ^p CO M§^ V^ U 0i M C3 »> )'.. V :y- ^^ ^ ^^ <5> f^-'p f # ( t¥ Uj QQ .■'/Ji^ f^y^^ V \ •m' Jj^^^i" -■'■] v'' iT "' 1 ^ 15 s d" ^ "^ •=!" '^ ^ ^ 200 STA 46 Figure 27-3. Dissolved methane (nl/1, STP) along a north- south section through Unimak Pass. Observations were made in Sept.-Oct. 1975. shown in Figs. 27-4a and b. At the beginning of the observations (Sta. 46), the distribution of methane was veriiically homogeneous, or nearly so. At approx- imately 4 hours, methane concentration in the sur- face layers increased to over 200 nl/1 and remained high for the following 20 hours. Similar trends were observed at depth although there appeared to be a sig- nificant lag period (four to five hours). Because the only reasonable source of the methane is St. George Basin, it is assumed that during the observational peri- od water was advecting south through Unimak Pass. To the north at Station Ebb 37, just east of the Pribilof Islands, the concentration of methane at the surface and at depth was uniform over the entire ob- servational period. This suggests that methane-rich water from St. George Basin did not move onto the shelf during this period. On the contrary, the mean current trajectory was probably parallel to the iso- baths and would result in advection of dissolved methane toward the northwest (Coachman and Char- nell 1979). In contrast to conditions observed the previous fall, concentrations of all LMW hydrocarbons were elevated in the surface waters during July. Concen- tration of dissolved methane in the surface waters is shown in Fig. 27-5a. Port MoUer once again was a significant source, as it was the previous fall. A con- centration of methane greater than 1300 nl/1 was ob- served at Station 28, or a value approximately 26 times the saturation amount. The source of this methane is believed to be sediments rich in organic matter inside Port MoUer, where decomposition of vegetable matter could result in the vigorous produc- "04 < 12 16 20 24 TIME (hrs) STA Ebb 37 12 16 20 24 28 32 36 TIME (hrs) Figure 27-4. Short-term variability of dissolved methane at stations 46 (a) and Ebb 37 (b) in Sept.-Oct. 1975. tion of methane. In the absence of significant micro- bial oxidative processes, methane dissolved in the shallow brackish waters of the lagoon would be transported through the entrance by tidal pumping (Barsdate et al. 1974). Because the air-sea exchange of methane is relative- ly slow, advection of methane-rich water could be traced east along the Alaska Peninsula for nearly 200 km. Although a cyclonic mean flow in the coastal water does not appear to dominate the shelf salt budget (Coachman and Charnell 1979), a weak coastal current exists along the peninsula (Kinder and Schumacher, Chapter 5, this volume). The surface distribution of methane in July, not unlike temperature, delineates quite sharply the hy- drographic domains over the shelf. Near the shore in water depths less than about 50 m, the concentra- tions of methane were everywhere greater than 100 nl/1 and reflected increased production relative to the previous fall. The source of the methane is not known precisely, but would include coastal sources as well as in-situ production from both bottom sedi- ments and the water column. The Kuskokwim River may also be a significant source of methane as suggested by the high concen- trations (> 400 nl/1) found near Cape Newenham. The Kuskokwim River was not sampled for dissolved hydrocarbons, but the Yukon River plume was sam- pled in July 1979 and found to contain concentra- tions of methane in excess of 2,000 nl/1. We would expect the lower reaches of the two rivers to be similar in dissolved methane content, since their lower drainage basins are similar. As mentioned 430 Chemical oceanography ® Figure 27-5. Surface (a) and near-bottom (b) distribution of dissolved methane (nl/I, STP) in July 1976. Near- bottom samples were taken within 5 m of the bottom. above, however, the benthic production in Kusko- kwim Bay must be considered on the basis of the available data. Methane concentrations in the surface waters of the middle shelf during July ranged from 50-100 nl/1 (Fig. 27-5a). Vertical stratification and low produc- tion of methane result in the observed distribution. The near-bottom distribution of methane is reflected in Fig. 27-5b. Previously identified sources near Port MoUer and Cape Newenham are evident, although concentrations at depth are somewhat reduced. This observation would be expected if the rivers in the area are a significant source of the methane. The bulk of the middle shelf water contained rather uniform levels of methane (50-100 nl/1), ex- cept for the outer shelf region north of Unimak Pass, where concentrations in excess of 400 nl/1 were found. These concentrations are slightly less than the values observed the previous fall (see Fig. 27-2b) and suggest that methane production is seasonal. Water temperature, carbon production rate, and quality of the organic matter would all be expected to strongly influence the production rate of methane, hence the in-situ concentration, if dispersive factors remain constant. The influence of circulation and mixing on the dis- tribution of methane is depicted in Fig. 27-6a, a zonal section through Bristol Bay (see Fig. 27-1; Sec. II). Methane is vertically homogeneous in the coastal water (Sta. 25A-17), but shows vertical structure over the middle shelf (Sta. 17-42). Surface waters in the middle shelf region are close to saturation with respect to methane in the atmosphere, whereas DISTANCE (km) 0 200 400 600 < ID to ro CM (\J CM 400 DISTANCE (km) 200 in (i3 r- 00 '^ ^ ^ ^ 200- Figure 27-6. (a) Vertical distribution of dissolved methane (n/nl, STP) along Sec. II in Bristol Bay (see Fig. 27-1). (b) Vertical distribution of dissolved methane (nl/1, STP) along Sec. I terminating in Unimak Pass. Observations were made in July 1976. Dissolved LMW hydrocarbons 431 concentrations increase to near 200 nl/1 at depth. Analogous to the situation predicted for salt (Coach- man and Chamell 1979), some of the methane found over the middle shelf probably originates by lateral diffusion from the outer shelf. Undoubtedly, there is an indigenous source, but its significance cannot be evaluated at this time. A core of methane-rich water is evident at Station Ebb 48 (Fig. 27-6a). It presumably originates from the southeast in water depths near 120 m (see Fig. 27-6b; near Sta. 44). The core properties of the methane suggest a mean flow at depth toward the northwest. Current-meter observations of Kinder and Schumacher (Chapter 5, this volume) taken over the outer shelf for the period June-July 1976 show a net current speed of 1-2 cm/sec at 100 m to the north- west (see their Sta. BC-13B), supporting the observed methane trajectory. The distribution of methane along a north-south section between Nunivak Island and Unimak Pass is shown in Fig. 27-6b. General structural features shown in Fig. 27-6a are preserved in this section, ex- cept for the complexity of multiple sources near Unimak Pass (Sta. 45-48). In contrast to the condi- tions observed in the previous year (see Fig. 27-3), there was a large accumulation of methane at depth south of Unimak Pass. The highest concentration of methane (630 nl/1) was observed at 400 m (Sta. 48), rather deep penetration for methane presumably originating from shallow shelf waters. However, since the concentration of methane at 100 m was 550 nl/1, methane-enriched water found offshore probably originated from the broad shelf ailong the southern Aleutian Peninsula. Whereas the data in Fig. 27-6b suggested a northward movement of water through Unimak Pass during the observational period, it probably was not sustained for any significant peri- od of time. This is exemplified by our observations of methane in the vicinity of Unimak Pass during Sep- tember and October 1975 (see Fig. 27-4a). LMW hydrocarbons The concentrations of the C2+ hydrocarbons (C2 to C4 ) are governed largely by seasonal processes in- volving biological activity or increased levels of insolation. Frontal dynamics appear to play a lesser role in controlling the distribution of these than for methane. A summary of the hydrocarbon concentra- tions (means and standaird deviations) for the vairious hydrographic domains is shown in Table 27-1. For the purpose of clarity and to delineate possible source regions, the data were organized according to hydro- graphic domains as described by Coachman and Char- nell (1977). Distinctions are also made between surface waters and near-bottom waters. Because of analytical problems encountered during the fall cruise, the concentrations of ethene and pro- pene include those of ethane and propane. Many observations conducted in the coastal waters of Alaska show that the concentration of ethene is ap- proximately three times that of ethane (see Fig. 27- 8). The relationship of propene and propane is similar. Since concentrations of butanes (iso- and n-) were always below the detection threshold of 0.05 nl/1, they are not included in Table 27-1. They do, how- ever, occur in measurable amounts near well-defined gas seeps in Norton Sound and Cook Inlet (Cline and Holmes 1977, Cline 1977). Both ethene and propene show a significant sea- sonal signature in surface waters. For example, the surface distribution of ethene during the summer of 1976 is shown in Fig. 27-7. Localized sources of ethene are not well defined, although the eastern por- tion of Bristol Bay appeared to be more productive than the outer shelf region in July. The distribution of ethene, like methane, does not correlate well vdth the hydrographic parameters. The reason for this probably lies in the relative rates of production and the decoupling of ethene production from hydro- graphic domains: water-column production of ethene is not influenced strongly by the 50-m frontal system, but is largely controlled by photochemical and bio- chemical processes in the surface and near-bottom waters. During the fall, in all hydrographic domains, the concentration of ethene was less than 1 nl/1, in- creasing to concentrations greater than 2 nl/1 in sum- mer. A similar trend was noted for propene, although the concentrations were systematically less. As expected, the vertical distribution of the Ci -C4 hydrocarbons in the coastal domain is invariant, al- though the seasonal component is evident in the depth -averaged mean concentrations (Table 27-1). This fact suggests that hydrocarbon production oc- curs in the water column or possibly at the sediment- water interface. If hydrocarbon production occurs principally at the sediment-water interface, tidally induced turbulence appears to be strong enough to homogenize the coastal water mass. A similar phe- nomenon obtains for heat and salt (Kinder and Schu- macher, Chapter 4, this volume). In the middle shelf domain (50 m < z < 100 m), the influence of both surface production and vertical stability come into play. Methane showed little seasonal difference (average), suggesting that water- column production was minimal and whatever verti- cal structure was present can be attributed to changes pq < — O a =" ~ — ' O) o a •X3 o C/3 Ol C > O C3 C C ^ a* ° S c ^ 2 a O a; -" S g I Q .r CO _tj 2 > c w P-, Pi « PL, OJ HI C - Pi c ca ^ d:; w ;z; w K H <-■ w rt Oil a <: c w CO 2 ff; < K H W ^ c as o Q CO Pi tH CD ^ CO '^ ■* lO en c~ o tH 05 1— 1 o T-l r-l o o csi l-H T— 1 1—1 1—1 o csi tH T-H CO CO csi o c^ CO (M lO tH o o o o o o 1—1 o o O o o LO -^ ^ CD CO CO '^ c>q 1-1 LO t- CO o o o o o o 1-1 1-1 iH o o o CO CO CO CO C- tS* o o o o o o ■ ' ■ ■ ' ' CO c^ CM CO CM CM o o o o o o ' ' ' • ' • -* Tt CO LO ■^ CO o o o o o o r^ 00 CD c- 00 CD C- '^ c- o 00 00 1—1 1-1 1-1 CM o iH '^ '^ ^ Tt< CM 1-1 CO c~ CO CM CM t^ O CO OS 1-1 00 00 o o o 1—1 o o CO CM 1—1 iH iH O CT3 O O i-H 00 t- O i-I LO o 00 ^ 05 CM CO CM CO CO CM CM Csl tH LO lO LO lO 1-1 1-1 1—1 CM 1-1 CM CM iH CD lO CO LO ^ CD o O o O O O ' ' ■ ■ ■ 05 O CD CO T-H 05 O iH O 1-1 1-1 o 05 00 05 00 00 CO CO 1—1 o o CM 1-1 CD CO CO lO LO 1-1 1-1 CO CD 1-1 iH CO o CM -^ LO LO CM lO CO o O O -* CO CM CM lO CD CO Tj< lO CD ^ 05 o cr> CD O CD LO CO 03 c^ CO CO "3 s w O cS lO o V JH o -O 1— I :s 6 ic-< lO CI O LO n o r/1 o >H CM 0) -1 O O O 1—1 CM -^ CO Si 1^ O ° V lO LO o 05 00 1—1 ^ CD tH 1-1 CM 03 Si u CO jn H CO o a; o -o 1-1 T3 o § LO >. 3 CO i-s r- m o> a iH 3 "-a n o rn o S-l CM -1 o o o 1-1 T3 C CO c CO x: X! 3 C c a, o ■a c lO 05 J2 3 ca >> a> .^^ > >. •*^ CO o 100 m), the in- crease in the concentration of ethane is presumably related to the increased carbon flux to the sediments (Sharma 1979) and a concomitant drop in redox po- tential. Decaying plankton lying on the bottom (R. Reeburgh, Univ. of Alaska, personal communication) would be expected to give rise to an anoxic stratum at the bottom. This condition would favor increased production of ethane as was shown to occur in the anoxic waters of the Black Sea (Hunt 1974). It is not proposed that the near-bottom waters of St. George Basin are anoxic, only that microenvironments or microstrata near the bottom represent a significant source of ethane to the water column. Because the sediments of Izembek Lagoon are rich in organic mat- ter (Barsdate et al. 1974), and are presumably anoxic, it is not surprising to find elevated concentrations of ethane near the entrance of the lagoon. The relatively high concentrations of ethane found in the surface waters of Unimak Pass are the result of strong vertical mixing from below, but the possible influence of shipping, fishing, and transportation activities on the hydrocairbon distribution cannot be assessed at the present time. As stated previously, hydrocarbons of different sources reflect distinct compositional patterns. Gaseous hydrocarbons generated in well-oxygenated waters or from shallow horizons in bottom sediments are usually characterized by high concentrations of methane (Brooks and Sackett 1973) and relatively high concentrations of alkenes (Bernard et al. 1978). Hydrocarbons derived from thermal processes are characteristically enriched in C2+ alkanes relative to methane and devoid of the alkenes (Frank et al. 1970, Clark and Brown 1977). Frank et al. (1970) have used these compositional characteristics to pro- pose the ratio [Ci ] /[C2] +[C3 ] as a possible indica- tor of a source. Ratios in excess of 500 suggest a biological source, whUe ratios less than 50 indicate a thermogenic origin. In an attempt to elucidate the origin of gaseous hydrocarbons, the ethane/ethene [C2]/[C2:i] ratio was plotted against the methane/ethane plus propane [Ci]/[C2]+[C3] ratio. Fig. 27-9 shows such a plot for distributions in Bristol Bay and includes two con- trasting marine environments. Comparisons are made with the hydrocarbon distributions in the vicinity of the Norton Sound gas seep (Cline and Holmes 1977) and with the anoxic waters of the Black Sea (Hunt 1974). Including all the analyses from Bristol Bay, the [Ci]/[C2]+[C3] ratio ranges from 30 to 500, while the [C2]/[C2:i] ratio ranges between 0.1 and 10- ,'?\0^ ^ 10' - 10' 10" 10° 10' 102 [C2]/:C2:|] 10= 10" Figure 27-9. Compositional hydrocarbon trajectory dia- gram for Bristol Bay. For comparison, the compositional fields of the Norton Sound gas seep (Cline and Holmes 1977) and the anoxic waters of the Black Sea (Hunt 1974) are indicated. Mixing of Bristol Bay hydrocarbons with various thermogenically derived hydrocarbon mixtures results in a family of mixing trajectories, two of which are shown here (see text). Along one of the dry gas trajectories, the fraction (%) of ambient water is given, assuming that the hydrocarbons in the source are each about 100 times the ambient levels. 1 . These compositional ratios obtain for most of the Alaskan shelf waters and are probably typical of high latitude, pristine coastal waters. On the other hand, anoxic waters in the Black Sea (Hunt 1974) reflect a narrow compositional field characterized by high con- centrations of methane, [Ci ] /[C2 ] +[C3 ] = 500, and relatively high concentrations of ethane, 20 < [C2 ] / [^2:1 ] < 50. Provisionally, it is assumed that the Black Sea represents a model of the hydrocarbon composition to be expected from anoxic marine sources. At this point it becomes useful to consider the compositional trajectories that might be observed if gases of thermogenic origin were mixed with biologi- cally derived hydrocarbons. To accomplish this, four distinct petroleum /natural gas sources were consid- ered from an analysis of gas and oil well data (Moore et al. 1966). The equations that govern the mix- ing trajectories in the [Ci]/[C2] + [C3] and [C2]/ [^2:1 ] plane are: [Ci] [Ci]a+((1-x)/x)[Ci]t [C2]+[C3] [Q]a+[C3]a+((1-x)/x)[C2]t+[C3]t (1) [C2] [C2]a [Calx [Q:i]=[C2:x]a"'^^1-^^/^)[Q:i]a ^2) The fraction of ambient water is x, whereas con- centrations of the individual hydrocarbons are shown in brackets. The two sources, between which mixing 436 Chemical oceanography is assumed to occur, are indicated by the subscripts A (ambient) and T (thermogenic source). The imphcit assumptions are that the source ratios are constant and that the concentration of ethene in the natural gas source is zero. To develop possible mixing scenarios, it was neces- sary to evaluate the possible range of LMW hydrocar- bon mixtures that might occur as the result of offshore production. The [Ci]/ [CaJ+LCg] fre- quency diagram for 366 terrestrial gas wells (Moore et al. 1966) was plotted, and by this means three dis- crete compositions, identified on the basis of their [Ci]/ [C2]+[C3] ratio, were defined, ranging from what is described as a "very" dry gas (methane rich) to a "typical" wet gas (methane poor). Some of the wet gases were associated with petroleum. A sum- mary of the calculations is shown in Table 27-2. TABLE 27-2. Mean mole fraction of methane, ethane, and propane, calculated for three arbitrarily defined natural gas fractions (Moore et al. 1966). Number of samples used in each statistic calculation is shown in column 5. Standard deviation about the mean is shown in parentheses. Mole Fraction [Ci ] ^ ^ *^3 [Cal+LCg]" Definition .79(.ll) .075(.04) .033(.026) 7 313 .96(.02) .011(.005) .002(.002) 61 47 .98(.02) .001(.0006) Tr 672 6 Wet Gas Dry Gas "Very" Dry Gas The actual concentration ratio observed in the water will depend on the component partial pressure (or mole fraction), the Bunsen coefficient, which de- pends upon salinity and temperature, and the depth of water at which gas injection occurs. Implicit in the previous statement is that equilibrium is achieved be- tween the gaseous and aqueous phases and that the rate of solution of the gases is a function of the thickness of the stagnant film boundary layer (Broecker and Peng 1974). Actually, equilibrium is probably not achieved. However, because the Bunsen coefficients and diffusion coefficients are similar for methane, ethane, and propane, the solubility ratio will not be significantly different from the equilib- rium ratio. If these minimum conditions hold, the following expression relates the component partial pressure in the gas phase to the equilibrium solubility ratio : r ^ . / ^ [C2]' + [C3]'-/32D2P2 +/33D3P3 ^^^ where |3i, Dj, and p; represent the Bunsen coefficient, diffusion coefficient, and mole fraction of (1) meth- ane, (2) ethane, and (3) propane. Assuming a mean temperature and salinity of 5 C and 30° /oo. Di = 0.85 X 10' cm^ /sec Dj = 0.69 X 10^ cm^ /sec D3 = 0.55 X 10' cm^/sec (3, = 0.04 ml CH4 (STP)/ml H^ O )32 = 0.06 ml C2 He (STP)/ml H^ O /33 = 0.06 ml C3 Hg (STP)/ml H^ O Bunsen coefficients (jSj), corrected for the "salting out effect" and the molecular diffusion coefficients (Dj), not corrected for the ionic strength of seawater, were estimated from the data of Bonoli and Wither- spoon (1968). The solubility ratio, [Ci]'/[C2]'+ [C3]', for gas well data calculated from equation 3, is shown in Table 27-2 and is also shown as mixing end members in Fig. 27-9. Because it has been as- sumed that the concentration of ethene is zero (i.e., [C2 ] '/[C2:i ] ' "^ °°), the mixing end members are lo- cated at the extreme right margin. The remaining end member we wish to consider is a wet gas associated wdth petroleum. For this pur- pose, the volatile fraction from the reservoir fluid of the Sadlerochit formation, Prudhoe Bay, was selected (Anon. 1971). The so-called volatile fraction, Ci to C4 hydrocarbons plus air gases, was normalized to 100 percent and the partial pressures of methane, ethane, and propane were calculated. As in the ex- ample given earlier, a gas of this composition is equil- ibrated with a parcel of Bristol Bay water. Table 27- 3 gives the mole fraction composition of the gas phase and the resulting equilibrium solubility ratio. Bunsen coefficients and diffusion coefficients were the same as before. TABLE 27-3. Mole fractions of methane, ethane, and propane from the reservoir fluid of the Sadlerochit formation, Prudhoe Bay (Anon. 1971), and the resulting equilibrium solubility ratio. Gas phase was assumed to contain only Ci -C4 hydrocarbons, nitrogen, carbon dioxide, and helium. Mole Fraction C2 C3 [Ci]'/[C2]'+[C3]' 0.44 0.051 0.030 3.7 As expected, the gas is relatively rich in C2+ hydro- carbons and its composition is similar to that calcula- ted earlier for a wet gas derived from gas wells. Of the gases present in the normalized mixture, methane, ethane, and propane constituted 81.6 percent by vol- I ume, C4 hydrocarbons amounted to 3.4 percent, CO2 was 14.2 percent, and the remaining 0.7 percent was divided between Ng and He. Having estimated some possible petroleum and natural gas end members, tra- jectories that would result from mixing these end members with Bristol Bay water can be examined. The mixing trajectory resulting from the injection of a typical dry gas will be considered first. To simplify the calculation, it is assumed that am- bient Bering Sea water contains approximately the following concentrations of LMW alkanes: [Ci] ~ 100 nl/1, [C2] =1 nl/1, [C2:i] = 3 nl/1, and [C3] = 0.4 nl/1, giving a [Ci ] /[C2] +[€3] ratio of 71 and a [C2 ] /[C2:i ] ratio of approximately 0.33. It is appar- ent that this locus may be biased toward low values of the [Ci]/[C2]+[C3] ratio (Fig. 27-9). If we assume that the local source of hydrocarbons is 100 times the ambient levels, the resulting mixing trajectory, computed from equations 1 and 2, is shown in Fig. 27-9. The values above the solid circles represent the percentage of ambient water at those points. Mixing a gas of this composition with the am- bient water results in little change in the [Ci]/ [C2]+[C3] ratio, but a large shift in the [Cgj/ [C2:i] ratio. This results directly from the assump- tion that the concentration of ethene was zero in the natural gas source. It is interesting that the mixing line passes through the compositional field observed in the region of the Norton Sound gas seep (Cline and Holmes 1977). Conclusions as to the source of ther- mogenic hydrocarbon gases in Norton Sound on the basis of this diagram would be misleading, however, since the ambient [Ci ] /[C2] +[€3] ratio in Norton Sound is significantly higher (~500) than the average value assigned for Bristol Bay (~71). One additional feature of the compositional field diagram (Fig. 27-9) is that the mixing line between two defined sources is invariant with respect to the source concentrations of the hydrocarbons, as long as the ratios are fixed. This means that regardless of the source strength, compositional changes due to mixing will occur along the mixing line between the two sources. With the relative concentrations (i.e., 100:1) chosen in the above example, seepage or leakage of a dry gas in Bristol Bay would be observable to approx- imately 99 percent dilution. If the relative concen- tration ratio is increased to 1000:1, the effect is ob- servable to 99.9 percent dilution. Finally, we calculate the hypothetical mixing line, assuming a wet gas composition similar to that found in the Sadlerochit formation. This trajectory is also reflected in Fig. 27-9 and lies significantly below the previously calculated dry gas curve. General conclusions drawn from Fig. 27-9 are that Dissolved LMW hydrocarbons 437 typical wet and dry gases should be readily distin- guishable from biologically produced hydrocairbons using a combination of the [C^ ] /[C2 ] +[C3 ] and [C2]/[C2:i] ratios. The composition of LMW hy- drocarbon mixtures characterized by [Ci]/[C2] + [C3] < 20 and [C2]/[C2:i] > 100 strongly suggests that they are thermally derived. The exception is a methane-rich dry gas, such as that produced currently from several wells in upper Cook Inlet (Kelly 1968). These gas wells contain methane in excess of 98 mole percent, only traces of ethane, and no measurable propane. An injected gas of this composition, in all likelihood, will not be distinguishable from the suite of hydrocarbons formed biologically under anoxic conditions (e.g., in sediments, lagoon environments, anoxic waters). Consequently, a dry gas of this composition would be interpreted as biogenic on the basis of its [Ci ] /[C2 ] +[C3 ] and [C2 ] /[C2:i ] ratios. Additional useful tracers include the [C2]/[C3] ratio (Nikonov 1972) and the 5^^C composition of the seep methane (Brooks et al. 1974, Bernard et al. 1976). In the latter study, Bernard and coworkers have plotted the 5 ^^ C of the dissolved methane against the [C^ ] /[C2] +[C3] ratio for a number of gas seeps investigated along the Texas shelf. Biogenic methane was found to be isotopically light (— 60<^/oo to —70^/00 vs. PDB), whereas methane derived from thermogenic sources was significantly heavier at — 40°/oo to — 50°/oo. In a similar study, Kvenvolden et al. (1979) found the pore fluid methane at the locus of the seep in Norton Sound to be isotopically heavy at — 36^/oo, suggesting a thermal source. There is no doubt that the isotopic composition of dissolved methane is an excellent diagnostic param- eter in identifying the origin of natural gas. However, usually the concentration of methane is so low as to preclude the use of isotopic fingerprinting, except where gas venting is vigorous (e.g., Bernard et al. 1976). At low concentrations, the compositional ratios should be useful in identifying sources of hy- drocarbons and their spatial trajectories. Because these hydrocarbons are dissolved, they readily iden- tify the trajectories of other dissolved components, which may possess greater toxicity (e.g., benzene). In the following discussion, attention will be given to the dispersion field of LMW hydrocarbons origi- nating from a localized source and some estimates of the areal diffusion and advection scales applicable to Bristol Bay. Dispersion model The major advantages of using LMW hydrocarbons as in-situ tracers of petroleum are related to their relatively high abundance in crude oil and natural gas 438 Chemical oceanography and the low concentration levels at which they can be measured. In the event of a spill, well blowout, or subsurface seep, it becomes useful to define the trajectory and areal impact of the dissolved and sus- pended hydrocarbons. Because the LMW hydrocar- bons are moderately soluble, these compounds become useful tracers of the soluble fraction. It is not the purpose of this report to model actual gas or oU seeps in Bristol Bay, since none were found, but rather to provide general guidelines on the usefulness of LMW hydrocarbons as tracers of the dissolved frac- tions of petroleum. For simplicity we assume that the petroleum source is continuous and steady in time, and results in a vertically homogeneous distribution within a selected layer (e.g., surface mixed layer). The model describing these minimum conditions is a two-dimen- sional advection-diffusion equation given by Csanady (1973): terline distributions (y = 0) will be discussed. Hence: ay yayj up dx kC-0, (4) where C is the concentration of the dissolved hydro- carbon, X and y are space coordinates, Ky is the scale- dependent horizontal eddy diffusivity, U is the mean horizontal velocity in the x-direction, and k is a first- order decay constant. In this case, we will use the first-order decay term to describe the air-sea exchange process and assume biological oxidation and produc- tion to be negligible. Because the source of hydrocar- bons is presumably small, the horizontal eddy diffu- sivity depends on the mixing scale (Okubo 1971). The solution to equation 4 for a line source is given by Csanady (1973) and is reproduced here with- out derivation : C(x,y) = i/2Coexp(-kx/U)[erf((L/2+y)/v^)+erf((L/2-y)/v^)] (5) where Co is the vertically integrated concentration at X = 0, L is the length of the line source, and Oy is re- lated to Ky through the equation Ky = (U/2)(day^/ dx). Equation 5 is given in terms of the concentra- tion at X = 0 rather than the production rate. This was done because the input rate for a seep would be difficult to quantify. Moreover, to our knowledge the production rate of known submarine gas seeps has never been documented. Thus, the model is ex- pressed in terms of the concentration field, which should be easily measured. To simplify the discussion in terms of maximum trajectory scales, only the cen- C(x,o) = Co exp(-kx/U) [erf(L/V8a7 )] (6) Using empirical relationships presented by Okubo (1971), where o^^^ = 0.0108t'-3\ U = x/t, and a^^ - 2o^Oy, equation 6 becomes C(x,o) = Coexp(-kx/U)[erf(L/0.208(x/U)'-i')] (7) after substitution. In Bristol Bay, as in any body of water, the source of gas may be at the surface or at depth. To dis- cuss dispersion scales under both these conditions we will assume that (1) the source is at the surface and is consequently influenced by air-sea exchange and (2) the source is at depth and physically isolated from the surface by a strong pycnocline (k = 0). We fur- ther assume that the component is biologically un- reactive, which is a reasonable assumption at low concentrations. Methane will be selected as the model component but any of the LMW hydrocarbon species would suffice. Equation 5 is valid for all dis- solved hydrocarbon species that possess chemical and biochemical reactivities similar to that of methane. Under the assumptions of the model, methane in- jected at the surface will be dispersed by diffusion and advection, and lost from the system through air- sea transfer. The flux of methane across the air-sea boundary can be described by the stagnant film boun- dary layer model (Broecker and Peng 1974), Fi =(D/h)(Ci-Ci'), (8) where D is the molecular diffusion coefficient, h is the thickness of the stagnant film, C^ is the average concentration in the mixed layer, and C^ ' is the equi- librium solubility concentration. The thickness of the molecular film is a function of sea-surface rough- ness or wind velocity (Emerson 1975). From the derivation of Fick's second law, it can be readily shown that transport across the sea surface is : Ti =(D/Az-h)(Ci-Ci'), (9) where Az is the depth of the surface mixed layer. Thus from equation 9 and by analogy to equation 4: k = D/Az-h. (10) In order to assess the importance of air-sea ex- change on the methane dispersion scale, the value of k must be estimated. To accomplish this, nominal seasonal values for surface temperature and mean sea- Dissolved LMW hydrocarbons 439 lar surface winds were used. All other parameters were computed from these data. Parameters used in the calculation of k are summarized in Table 27-4. TABLE 27-4 Parameters used in the estimation of the first order decay constant (equation 10) governing the escape rate of methane across the sea surface. Wind Temp.^ Speed'' C m/sec cm^ /sec m hd jum k /sec Summer Winter 10 0 6.3±0.8 9.9±0.5 1.1X10-^ 0.65X10"^ 20 75 80 15 6.6X10-'^ 5.8X10-'' ^Kinder and Schumacher, Chapter 4, this volume ^Broweretal., 1977 •^Bonoli and Witherspoon, 1968 98 mole percent), similar to that currently being pro- duced in Cook Inlet, represents an exception to the foregoing generalizations. In all likelihood, a meth- ane-rich gas of thermogenic origin would appear to be of biological origin in the [C^ ] /[C2 ] +[C3 ] and [C2 ] /[C2:i ] compositional fields. FinEilly, an estimate was made of the possible dis- persion scale of LMW hydrocarbons arising from a line source, using a steady-state two-dimensional model. On the basis of typical mean velocities of 2 and 10 cm /sec with the inclusion of a sink term for air-sea exchange, it was estimated that dissolved LMW hydrocarbons would be identifiable in surface waters over distances of 40-150 km, assuming a local hydro- carbon concentration of 100:1 above ambient. At depth below a strong pycnocline, the horizontal mixing scale increases to 100-500 km. Actual trajec- tory scales expected in Bristol Bay will depend on a number of factors. Among them are the strength of the source, degree of solution, diffusive mixing scales, air-sea exchange, and biological oxidation rates. In this report the usefulness of LMW hydrocarbons as a diagnostic indicator of petroleum hydrocarbon is described. Not only are these gases excellent tracers of dissolved or emulsified petroleum hydrocarbons, but their ease of measurement allows real-time obser- vations to be made— a valuable adjunct to a moni- toring program. gram responding to the needs of petroleum develop- ment of the Alaskan shelf is managed by the Outer Continental Shelf Environmental Assessment Pro- gram (OCSEAP) Office. REFERENCES Anonymous 1971 Prudhoe Bay data are revealed at Alaska for the first time. Gas and Oil J., 57-64. Barsdate, R. J., M. Nebert, and C. P. McRoy 1974 Lagoon contributions to sediments and water of the Bering Sea. In: Oceanography of the Bering Sea, D.W. Hood and E.J. Kelley, eds., 553-78. Inst. Mar. Sci., Occ. Pub. No. 2, Univ. Alaska, Fairbanks. Bernard, B. B., J. M. Brooks, and W. M. Sackett 1976 Natural gas seepage in the Gulf of Mexico. Earth Planet. Sci. Lett. 31: 48-54. ACKNOWLEDGMENTS I wish to thank Lee Ohler, Anthony Young, and Susan Hamilton, who participated in the cruises to the Bering Sea. Special thanks are due to Charles Katz, who diligently scrutinized the data for errors and organized it into a usable form. Criticism of the manuscript was provided by Drs. H. Curl, R. Feely, J. Lavelle, and J. Schumacher of PMEL/NOAA, Dr. K. Kvenvolden of USGS, and Dr. R. Gammon of the University of Washington. To all these people I am grateful for their helpful suggestions. Finally, I wish to express gratitude to the captains and crews of the NOAA research ships Discoverer and R/V Moana Wave, without whose assistance in the field this work could not have been carried out. This study, Contribution No. 456 from the NOAA/ ERL Pacific Marine Laboratory, was supported by the Bureau of Land Management through interagency agreement with the National Oceanic and Atmos- pheric Administration, under which a multiyear pro- 1978 Light hydrocarbons in recent Texas continental shelf and slope sediments. J. Geophys. Res. 83: 4053-61. Bonoli, L., and P. A. Witherspoon 1968 Diffusion of paraffin, cycloparaffin and aromatic hydrocarbons in water and some effects of salt concentra- tion. In: Advances in organic chem- istry, P. A. Schenck and I. Havenaar, eds., 373-84. Pergamon Press, New York. Broecker, W. S., and T. H. Peng 1974 Gas exchange rates between air and sea. Tellus 26: 21-35. Brooks, J. M., J. R. Gormly, and W. M. Sackett 1974 Molecular and isotopic composition of two seep gases from the Gulf of Mexi- co. Geophys. Res. Lett. 1: 213-16. 442 Chemical oceanography Brooks, J. M., and W. M. Sackett 1973 Sources, sinks, and concentrations of light hydrocarbons in the Gulf of Mexico. J. Geophys. Res. 78: 5124- 58. 1977 Significance of low-molecular-weight hydrocarbons in marine waters. In: Proc. 7th Inter. Meeting on Organic Geochemistry. 455-68. Brower, W. A., Jr., H. W. Searby, J. L. Wise, H. F. Diaz, and A. S. Prechtel 1977 Climatic atlas of the outer continental shelf waters and coastal regions of Alaska: II, Bering Sea. AEIDC, Univ. Alaska, Anchorage. Carlisle, C. T., G. S. Bayliss, and D. G. Vandelinder 1975 Distribution of light hydrocarbons in seafloor sediments: Correlations between geochemistry, seismic struc- ture, and possible reservoired oil and gas. OTC2341: 65-8. Clark, R. C, and D. W. Brown 1977 Petroleum: Properties and analyses in bio tic and abiotic systems. In: Effects of petroleum on Arctic and Subarctic marine environments and organisms, 1-90. Academic Press, N.Y. Cline, J.D. 1977 Distribution of light hydrocarbons, Ci -C4 , in the northeast Gulf of Alaska, Lower Cook Inlet, southeast- ern Chukchi Sea. In: Environmental assessment of the Alaskan conti- nental shelf, 8: 180-268. DOC/NO AA. Cline, J. D., and R. Feely 1976 Distribution of light hydrocarbons, C1-C4, in the Gulf of Alaska and southeastern Bering Shelf. In: En- vironmental assessment of the Alas- kan continental shelf, 9: 443-547. DOC/NOAA. Cline, J. D., R. Feely, and A. Young 1978 Identification of natural and anthro- pogenic petroleum sources in the Alaskan shelf areas utilizing low molecular weight hydrocarbons. In: Environmental assessment of the Alaskan continental shelf, 8: 73-198, DOC/ NOAA. Cline, J. D., and M. L. Holmes 1977 Submarine seepage of natural gas in Norton Sound, Alaska. Science 198: 1149-53. Coachman, L. K., and R. L. Charnell 1977 Finestructure in outer Bristol Bay, Alaska. Deep-Sea Res. 24: 869-89. 1979 On lateral water mass interaction— A case study, Bristol Bay, Alaska. J. Phys. Oceanogr. 9: 278-97. Csanady, G. T. 1973 Turbulent diffusion in the environ- ment. D. Reidel Publishing, Boston. DOI/DOC (Department of the Interior /Department of Commerce) 1975 Environmental assessment of the Gulf of Alaska, Southeastern Bering and Beaufort Seas, Boulder, Colo. Dunlap, H. F., J. S. Bradley, and T. F. Moore 1960 Marine seep detection— A new recon- naissance exploration method. Geo- physics 25:275-82. Ehhalt, D. H. 1974 The atmospheric cycle of methane. Tellus 26: 58-70. Emerson, S. 1975 Gas exchange rates in small Canadian Shield lakes. Limnol. Oceanogr. 20: 754-61. Feely, R., and J. D. Cline 1977 The distribution, composition, and transport of suspended particulate matter in the northeastern Gulf of Alaska, southeastern Bering shelf, and lower Cook Inlet. In: Environ- mental assessment of the Alaskan continental shelf, 8: 89-179, DOC/ NOAA. Frank, D. J., W. M. Sackett, R. Hall, and A. D. Fred- ericks 1970 Methane, ethane, and propane concen- trations in the Gulf of Mexico. Amer. Assoc. Petrol. Geol. Bull. 54: 1933-8. I Harrison, H., J. E. Johnson, and J. D. Cline 1980 Light hydrocarbons in the atmospher- ic boundary layer over the North Pacific. (Submitted to J. Geophys. Res.) Hunt, J. M. 1974 Hydrochemistry of Black Sea. In: The Black Sea, geology, chemistry, and biology, E.T. Degens and D.A. Ross, eds., 499-504. Amer. Assoc. Petrol. Geol. Mem. 20. Kelly, T. E. 1968 Gas accumulations in non-marine strata. Cook Inlet, Basin, Alaska. In: Natural gases of North America, B.W. Beebe, ed., 49-64. Amer. Assoc. Petrol. Geol. Mem. 9, Part I. Kinder, T. H. 1977 The hydrographic structure over the continental shelf near Bristol Bay, Alaska, June 1976. Dep. Oceanogr. Tech. Rep. M77-3: 1-61. Univ. of Washington. Kinder, T. H., and L. K. Coachman 1978 The front overlying the continental slope in the eastern Bering Sea. J. Geophys. Res. 83: 4551-9. Kvenvolden, K. A., K. Weliby, C. H. Nelson, and D. J. Desmarais 1979 Submarine carbon dioxide seep in Norton Sound, Alaska. Science 205: 1264-6. Dissolved LMW hydrocarbons 443 Moore, B. J., R. D. Miller, and R. D. Schrewsbury 1966 Analyses of natural gases of the United States, 1964. Bureau of Mines IC-8302: 1-144. Nikonov, V. F. 1972 Distribution of methane homologs in gas and oil fields. Doklady Akad. Nauk SSSR226: 234-6. Okubo, A. 1971 Oceanic diffusion diagrams. Sea Res. 18: 789-802. Deep- Oremland, R. S. 1975 Methane production in shallow-water, tropical marine sediments. Appl. Microbiol. 30: 602-8. Reeburgh, W. S., and D. T. Heggie 1977 Microbial methane consumption reac- tions and their effect on methane distributions in freshwater and marine environments. Limnol. Oceanogr. 22: 1-9. Sackett, W. M. 1977 Use of hydrocarbon sniffing in off- shore exploration. J. Geochem. Explor. 7: 243-54. Schumacher, J. D., T. H. Kinder, D. H. Pashinski, and R. L. Charnell 1979 A structural front over the continental shelf of the eastern Bering Sea. J. Phys. Oceanogr. 9: 79-87. Lamontagne, R. A., W. D. Smith, and J. W. Swinner- ton 1975 C1-C3 hydrocarbons and chlorophyll a concentrations in the equatorial Pacific Ocean. In: Analytical methods in oceanography, T.R.P. Gibbs, Jr., ed., 163-71. Advances in Chemistry Series 147, Amer. Chem. Soc. McCarty, P. L. 1964 The methane fermentation. In: Principles and applications in aquatic microbiology, H. Heukelekian and N.C. Dondero, eds., 314-43. John Wiley, N. Y. Scranton, M. I., and P. G. Brewer 1977 Occurrence of methane in the near surface waters of the western subtrop- ical North Atlantic. Deep-Sea Res. 24: 127-38. Scranton, M. I., and J. W. Farrington 1977 Methane production in the waters off Walvis Bay. J. Geophys. Res. 82: 4947-53. Sharma, G. D. 1979 The Alaskan shelf. N. Y. Springer-Verlag, 444 Chemical oceanography Smith, A. M., and R. J. Cook 1974 Implications of ethylene production by bacteria for biological balance of soil. Nature (London) 252: 703-5. Swinnerton, J. W., and R. A. Lamontagne 1974 Oceanic distribution of low-molecular- weight hydrocarbons. Environ. Sci. andTechnol. 8: 657-63. Swinnerton, J. W., R. A. Lamontagne, and J. S. Hunt 1977 Field study of carbon monoxide and light hydrocarbons production related to natural biological processes, NRL Rep. 8099: 1-9. Wilson, D. F., J. W. Swinnerton, and R. A. Lamon- tagne 1970 Production of carbon monoxide and gaseous hydrocarbons in seawater: Relation to dissolved organic carbon. Sci. 168: 1577-9. Wolfe, R. S. 1971 Microbial formation of methane. In: Advances in microbial physiology, A.H. Hose and J.F. Wilkinson, eds., 6: 107-46. Academic Press, N. Y. Swinnerton, J. W., and V. J. Linnenbom 1967 Determination of the Cj to C4 hydrocarbons in seawater by gas chromatography. J. Chromatogr. Sci. 5: 570-3. Yamamoto, S., J. B. Alcauskas, and T. E. Grozier 1976 Solubility of methane in distilled water and seawater. J. Chem. Eng. Data 21: 78-80. I Bection 1 Fisheries Oceanography Felix Favorite, editor Overview of Fisheries Oceanography Felix Favorite Northwest and Alaska Fisheries Center Seattle, Washington ABSTRACT In the past, specific fisheries oceanography investigations designed to explore or explain the complex interrelationships of fish and the various elements of the eastern Bering Sea environment have not been carried out in spite of the fact that early investigators consistently pointed out the inadequacies of casual environmental observations obtained during fishing cruises and recognized the need for more complete studies. Although today the acquisition and processing of environ- mental (physical and chemical) data are rapid and straight- forward, using STD and autoanalyzer systems and shipboard computers, integrated studies are still rare. This overview puts the subsequent chapters in this section in the perspective of fisheries oceanography and also suggests that although annual stock assessments may satisfy management criteria, conserva- tion practices wall require information on the year-round distributions of juveniles as well as adults, multispecies inter- actions, and rather specific relations within the ecosystem. This is a large task, but more emphasis on fisheries oceano- graphy studies will accelerate acquisition of the required knowledge. INTRODUCTION The rapacious exploitation of whales in the Bering Sea to obtain whale oil in the mid-1 9th century was accomplished without any consideration of conserva- tion principles or of impact on the ecosystem; it is gratifying to see the extensive efforts funded by the Outer Continental Shelf Environmental Assessment Program (OCSEAP) to define conditions and pro- cesses in this area as in this century we seek to obtain petroleum from below the sea floor. Because of the purpose of the OCSEAP program, to assess real and possible effects of oil exploration and exploitation, many of the individual grants are largely disciplinary (e.g., physical oceanography, fishery biology, ich- thyoplankton), and area specific (e.g., Norton Sound, Navarin Basin, St. George Basin, Bristol Bay Basin). The purpose of this section is to consider the conti- nental shelf of the eastern Bering Sea as an integral unit and briefly evaluate environmental and biological continuities and relationships of certain fishes throughout the area. Although we have far exceeded the five chapters called for in the initial guidelines for this study, the subject is still not adequately covered. The term "fisheries oceanography" stems from fisheries hydrography, which originated at the turn of the century, and essentially embraced any and all conditions influencing fisheries. Since the word "ecology" apparently was avoided, one suspects that originally the term meant the study of those imme- diate conditions that would help to catch fish, or perhaps explain why fish were not caught, rather than the broad studies of the ocean or even very specific but basic research. The shift from fisheries hydrography to fisheries oceanography in mid-cen- tury was largely semantic, and did not represent a change in activities; those in the burgeoning field of ocean studies could not decide whether they would march under the banner of oceanography or ocean- ology. Such a discussion may seem trivial, but Chapman (personal communication) canvassed the world in an attempt to obtain an acceptable defini- tion of fisheries oceanography, without success. The actual scope of studies embraced by fisheries oceanography is virtually unlimited— from external forces, global or extraterrestrial, that alter the phy- sics, chemistry, biology, or geology of the sea, to internal forces that redistribute, modify, consume, or guide this energy, including interactions of all life, life-support systems, and the effects of man. Cur- rents, temperatures, and density have long been considered a part of fisheries studies. Although one may consider dismissing studies of nutrient chemistry because the absence or abundance of nutrients is reflected in primary (plant) production, without these studies there will be no evidence of why phyto- plankton are scarce or abundant. In a study of pelagic fish one can argue that even primary produc- tion can be dismissed because it is reflected in sec- 447 448 Fisheries oceanography ondary production (herbivores), but primary pro- duction cannot be ignored in discussing demersal or bottomfish, which consume benthos, whose devel- opment depends in large part upon that part of the plant production which reaches the bottom. How- ever, secondary production and the subsequent trophic level, tertiary production (carnivorous plank- ton), are consumed in various stages of fish develop- ment and must be considered. Recently the role of starvation— the lack of abun- dance of the required kind and size of forage at the appropriate time for fish larvae, juveniles, and adults— has become an important consideration in fisheries oceanography studies. Studies of larval fish in the Bering Sea raise many questions because our knowledge of early life stages of fish is extremely fragmentary. Although some information is avail- able on age-one pollock and juvenile yellowfin sole distributions, there are vast gaps in our knowledge of the distributions and movements of these juveniles and those of other commercial species for a year, or in some instances even several years, before they are recruited to the fishery, and we are almost totally ignorant about juveniles of non-commercial species. This has not been a primary concern in managing adult stocks, but if one considers the eastern Bering Sea as a potential area for shelf ranching, a concept approaching reality as a result of the Fisheries Con- servation and Management Act (FCMA), an entire spectrum of fisheries and oceanographic information will be required to define this ecosystem. Although fisheries oceanography studies are not limited to adult commercial fish and their environ- ment, discussions in this section are restricted in this way because studies on physical and biological oceanography, benthos, ice, and other aspects of the Bering Sea ecosystem are presented in other sections of this volume. However, in studies of adult fish the term fisheries oceanography does not mean merely matching climatic or casual marine observations obtained during fishing or other cruises with catch data in an attempt to establish one-to-one relation- ships; it carries the connotation of controlled studies of relationships among environmental conditions, processes, or events, and species behavior and multi- species interactions. It signifies carefully designed experiments to define relationships or verify sus- pected associations. Such experiments, with the possible exception of parts of the Processes and Resources of the Bering Sea Shelf (PROBES) pro- gram, have not been conducted in the eastern Bering Sea even though a great deal of general information has been obtained from foreign fishing fleets, speci- fically Japanese and Soviet. Because of management demands, much of the U.S. research effort in the eastern Bering Sea has been devoted to what is present, rather than why, and what factors can and do cause fluctuations in distribution and abun- dance of stocks. But this trend is changing slowly as more attention is directed toward ecosystem models. BACKGROUND During a discussion of fisheries oceanography in the eastern Bering Sea at Fairbanks, Alaska, in 1974, participants noted the dissimilar spawning and migration patterns of halibut, yellowfin sole, and Pacific salmon and pointed out that PROBES offered an excellent opportunity to conduct multidisciplinary research in relation to or in concert with fisheries (Favorite 1974a). Although an excellent group of investigators joined PROBES, little effort was focused on any fishery but pollock, and only the egg and larval stages have been studied. Five years later, this opportunity to investigate in greater detail these and other patterns of multispecies interactions and re- source-environment relations was welcomed. It might be said that nature abhors an equilib- rium—natural forces do not necessarily operate to maintain the status quo, and often man's intervention is all that is required to tilt the system, sometimes drastically. In our haste to evaluate present condi- tions we should not ignore the past or assume that steady-state balances between primary production and biomass exist. Except for folklore, our early knowledge of fish components stems largely from Cook's voyage in 1778, during which halibut, cod, salmon, and other fishes and mammals were encoun- tered (Munford 1963), but of course relative abun- dances were unknown. For roughly a quarter of the 19th century, the shelf area south of St. Matthew Island, with the exception of inner Bristol Bay, was considered a major whaling ground; and the area between Nunivak and Unimak islands was still in use in 1875. The exploitation of whales and fur seals in this airea and the subsequent exploitation of cod in the same general area in the late 19th and early 20th centuries must have altered several niches and extant transfers of energy between trophic levels, but there are no data to show cause and effect. Did the reduc- tion in whale populations augur well for fish popula- tions because of an increase in availability of plank- tonic forage or a decrease in consumption of fish? Did the extensive removal of cod by the fishery permit the current massive crab biomass to establish itself? Will the extensive present fishery on pollock result in a change in the feeding habits of fur seals that could trigger changes in the ecosystem? Overview 449 The information on recent events is better but still fragmentary. The recent sharp decline of halibut, although not necessarily caused by overfishing, is certainly related to the extensive bottom trawling effort during the last decade or two. The apparently large pollock resource may be due to reductions of older age groups by the fishery, thereby reducing the effects of cannibalism, or to the fact that there appears to be a substantial stock seaward of the shelf serving as a reservoir and source of replenish- ment. Yellowfin sole abundance was severely re- duced after the initial fishing on this stock in the 1960's. Herring abundance has always been variable and attempts today to estimate stock size are subject to close scrutiny. And of course for years catches by Japanese salmon motherships have affected salmon returns to various river systems ; the forecast estimates of these returns are based largely on counts of down- stream migrants supplemented with data from ocean- ic monitoring. There is little need to be defensive about the incompleteness of knowledge concerning environ- mental processes affecting fisheries and stock assess- ment in the eastern Bering Sea; such studies and evaluations are still being conducted in the North Sea, where fisheries investigations have been carried out independently and cooperatively by many nations for centuries. It was not until the late 18th century that Vancouver showed the eastern side of Bering Strait to be part of North America and provided a map of coastal features. And it was not until the early part of the 19th century that maps and charts indicated that the name Kamchatka Sea had been changed to Bering Sea. Wilimovsky (1966) has summarized early voyages in this area. Typically exploitation has come before scientific exploration. By the mid-1 9th century, whaling grounds over the southeastern Bering Sea had largely been abandoned and whalers were entering the Arctic Ocean. At the turn of the century, the cod on Slime and Baird banks north of the Alaska Peninsula were heavily fished and numerous salmon canneries were in operation in Bristol Bay and surrounding areas. Yet it is only today that an understanding of oceanographic conditions and processes is being attained, largely through studies by OCSEAP, PROBES, NWAFC, and other agencies. Knowledge of bathymetry has also been fragmen- tary. Although at the end of the 19th Century U.S. Coast and Geodetic Survey and U.S. Bureau of Fisheries charts revesiled the existence of the broad shelf and its abrupt termination at the edge of the Bering Sea basin, it has been only within the last decade that a large 300-m depression shown on Japanese and Soviet charts between the Pribilof Islands and St. Matthew Island was found to be nonexistent (Favorite 1974b). And only within the last few years has some of the complexity of the bathymetry at the shelf edge been discovered as a result of fishing operations. FISHERIES OCEANOGRAPHY There are several periods during which fisheries and oceanographic data have been collected with wide- ranging degrees of synopticity and thoroughness. The whalers in the mid-1 9th century recorded casual environmental data that when combined with data from geographical explorations provided insight into resources and conditions (Dall 1882— see also In- graham, this volume). At the end of the century the U.S. Bureau of Fisheries conducted studies from aboard the steamer Albatross, but after that, fisheries or oceanographic studies were not resumed until the 1930's; except for the World War II period field efforts have markedly accelerated up to the present time. Steamer Albatross Rathbun (1894) reported that the Albatross ex- plorations were noteworthy in that they constituted an innovation in support of the fishing industry beyond anything ever attempted by any other nation. In a summary of cruises from 1888 to 1892 in the eastern Bering Sea he noted that although three cod banks had been recognized north of the Alaska Peninsula— Slime, Baird, and Kulukak— these banks were not necessarily separate and distinct, except that the westernmost. Slime Bank, was characterized by immense numbers of a large jellyfish, brownish or rusty in color, measuring from 15 to 45 cm across, with long tentacles having great stinging powers. Although not observed upon the sea surface, they appeared abruptly around July 1 in subsurface concentrations which interfered with lowering fishing hooks, using cod trawls, and even raising anchors. Today the general location may be construed as a frontal zone between oceanic and shelf water, but other circumstances are certainly involved in the phenomenon. The largest and best cod during this summer period were taken 11-14 km north of the peninsula; those inshore were of smaller size and inferior quality. Even today there is little specific information concerning the nature or structure of a front along the coast. Rathbun noted also that problems concerning the habits of fur seals on the Pribilof Islands, near the outer edge of the bank. 450 Fisheries oceanography made it very important that the physical and biolog- ical features of the surrounding area be thoroughly studied, and that the distribution of cod was greatly influenced by the movements of capelin, herring, and sand lance but scarcely anything was known regarding the habits of these species on the Alaska coast. However, shortly after the turn of the century the Albatross cruises in this area ceased. Only in the last few years have such studies been resumed. Salmon studies, 1938-41 Except for some tagging studies in the 1920 's (e.g., Gilbert 1924) near Port Moller of salmon returning to Bristol Bay streams, there is little evidence of marine research in the area until the 1930's, when the Japanese established a crab fishery and a small operation for bottomfish on the shelf. When it became apparent in 1937 that these operations were to extend to salmon fishing, funds were made avail- able to the U.S. Fish and WUdlife Service (FWS) to conduct investigations of the migration routes and availability of salmon in western Alaska, particularly salmon of the Bristol Bay region. Oceanographic studies were conducted aboard the U.S.C.G.T. Redwing in 1938 and salmon fishing studies aboard chartered fishing vessels in 1939; both operations continued until 1941, when the outbreak of World War II interrupted them and delayed the reports of results (Bamaby 1952, Favorite and Pedersen 1959, and Favorite et al. 1961). Although salmon were caught in all areas fished over the southeastern part of the shelf, they were apparently more abundant in the area 54-108 km offshore along the north side of the Alaska Peninsula. The cause of this abundance is still unexplained, although it has been ascertained that seaward-migrating smolts occur inshore of the shoreward migration of adults (Straty 1974). Fur- ther, there was a marked shift in dominance from sockeye to chum salmon north of the Pribilof Islands, in an area believed to represent a migration path of the latter to more northern coastal streams. Prophet- ically, Barnaby noted that there were large popula- tions of bottomfish and shellfish in the area and that it was only a matter of time before exploitation would occur. Crab studies, 1941 In 1940 Congress approved a special appropriation authorizing the FWS to conduct a king crab study off the coast of Alaska and operations in the Bering Sea were conducted from April to September in 1941 (Fishery Technology Laboratory 1942). Bottom isotherms indicated a tongue of cold water between Cape Newenham and the Pribilof Islands in the direction of Port Moller; sharp gradients along the north side of the Alaska Peninsula were believed to cause rapid changes in water temperature relatively close inshore, particularly between Amak and Seal Islands, where the greatest concentrations of crabs were found. But since similar conditions did not occur in other productive areas, the significance of this possible relationship could not be evaluated. However, it was pointed out that the incidental catch of edible flatfish was phenomenal. Larger than usual cod were found in offshore waters at depths deeper than usually fished commercially; pollock and yel- lowfin sole were found in great abundance, and the possibility of a commercial fishery for halibut was indicated. These studies also ceased at the outbreak of the war. Bottom trawling and oceanography, 1948-49 Crab studies resumed in 1947 aboard the charter vessel Alaska (King 1949) in a small area about 75 km northwest of Port Moller. When conditions per- mitted, air, surface, and bottom temperatures were obtained. The maximum crab catch occurred at a bottom temperature of 3 C and catches were made at temperatures from 1.65 to 7.25 C. Unusually large incidental catches of pollock and yellowfin sole were roughly an order of magnitude larger than cod catches and undoubtedly this prompted the chartered mothership operation (Pacific Explorer, with a fleet of 10 fishing vessels) from April to July 1948 in the Amak Island and Black Hills area (Wigutoff and Carlson 1950); but no environmental data are re- ported. Perhaps the first fishery cruise that seriously con- sidered oceanographic conditions was the chartered Deep Sea from Unimak Pass to Norton Sound during the last week in June and the first week in July 1949. The report of the cruise (EUson et al. 1950) indicates awareness of previous oceanographic studies (Rat- manoff 1937, Barnes and Thompson 1938, Goodman et al. 1942, and others), and results indicated: there was a positive correlation between depth and bottom temperatures except for the corridor of subzero bottom water extending past St. Matthew Island to the southwestern portion of St. Lawrence Island; bottom water temperatures, more than any other condition, were positively correlated with the abun- dance of fish taken over the entire area; temperatures at the surface and bottom were several degrees lower than those observed in the previous September from aboard the U.S.F.W.S. Washington; summer warming influenced the northward migration of many fish; the cold corridor effectively acted as a faunistic barrier to the movement of many species through the strait between Siberia and St. Lawrence Island; the fauna in this area consisted mainly of eelpouts. Overview 451 I I liparids, sculpins, and Tanner crabs; the corridor was believed to extend past St. Matthew Island and the PribUof Islands into the southeastern Bering Sea, and to cause species entering Bering Sea to be shunt- ed through the Aleutian passes eastward before continuing northward; cod and flatfish in particular were assumed to "lead" along the edges of the cold water barrier, an assumption which would explain the concentrations found in the approaches to Bristol Bay and northward along the warm water channel adjacent to the coast. These cruises indicate that a number of years ago there was ample evidence of relations between these resources and the environment to justify conducting specific fisheries oceanography studies that would enhance conservation and management practices. Recent studies Commercial exploitation of resources, largely by the Japanese, intensified in the 1950's. Scientific studies and conservation measures were enhanced by the formation in 1953 of the International North Pacific Fisheries Commission (INPFC), whose reports, documents, and bulletins are readily accessible and describe Japanese and U.S. investigations, in progress even today, of crab and fish resources. And the NWAFC has conducted an extensive annual survey of resources in the southern part of the area since the early 1970's. But environmental observations are limited and only incidental to fishing effort. The extensive Soviet Bering Sea Comprehensive Scientific-Commercial Expedition began in 1958; it resulted in numerous compilations of environmental data summarized by Favorite et al. (1976) and fishing reports (Moiseev 1963-72) cited throughout the following chapters. Even though this Soviet activity prompted extensive analyses of existing oceanograph- ic and fisheries data, and environmental observations obtained from aboard fishing vessels contributed to the knowledge of resource-environment relations, such observations were incidental to the fishing effort, fragmentary, and not available through normal channels. Furthermore, no major Soviet oceano- graphic vessel participated in the investigations. Recently there has been a trend toward coopera- tive international fisheries research that becomes more coordinated and successful each year. Emphasis is still, however, largely on fishing at predetermined station locations that provide assessments of resource abundance; little funding, personnel, and vessel time are given to specific fisheries oceanography studies that may provide the key to more efficient and effective resource assessment and improved manage- ment and conservation. SECTION SUMMARY Many of the physical and biological characteristics of the eastern Bering Sea are directly related to fisheries oceanography. First, it has an unusually broad, relatively flat continental shelf the slope of which abruptly increases mainly near the 150-m isobath. It is relatively isolated from the currents of the North Pacific basin by the Alaska Peninsula, the Bering Sea basin by the abrupt shelf edge, and the Arctic basin by the narrow and shallow Bering Strait. It has three major embayments— Bristol Bay, Norton Sound, and the Gulf of Anadyr. It is basically a subarctic environment with ice cover present and expanding seaward from October to a maximum in April, and largely absent by June. Although surface and coastal water temperatures in summer are not unlike those as far south as the Washington coast, the —1.8 C water temperatures under the ice in winter and near the bottom at mid-shelf during summer identify the Bering Sea shelf regime. High river runoff in late spring markedly alters coastal water properties and flow, and oceanic perturbations greatly affect flow at the outer portion of the shelf; whereas conditions at mid-shelf are largely influenced by tidal currents. Warming and the formation of a surface layer of ice melt in summer result in positive stability; during the winter, ice formation prod- uces slightly negative stability, which causes water overturn. The sluggish circulation and nutrient replenishment as a result of winter overturn are conducive to high production and standing stocks in spring and summer; the ice in winter causes extensive fish migrations to the deeper, warmer water at the shelf edge and upper slope. Pollock are the dominant semidemersal species, yellowfin sole the dominant demersal species, and herring a dominant pelagic species although the anadromous salmon are abundant from June to September. Although the largely discipline-oriented approaches by individual OCSEAP research units and the gener- ally independent field surveys make it difficult to make multispecies or interdisciplinary assessments, the opportunity to bring the fragmentary information on the various life stages of major biological com- ponents up to date, to assess dynamic aspects of the ecosystem, and to extend the available knowledge through the use of simulation techniques presented a great challenge; and that is what we as a group have attempted to accomplish— in spite of frustrations associated with the lack of continuity of data in space and time, the multiplicity of sampling techniques, and large gaps in basic knowledge. We have had to be selective in our approach and limit discussions of 452 Fisheries oceanography resource-environment relations to several commercial fish which obviously prey on zooplankton, other fish, and benthos, and are preyed upon by fish, birds, and mammals. Even though such interactions are poorly documented, numerical assessments have been made using simulation techniques. First, background information on environmental conditions (Ingraham) and distributions of ichthyo- plankton (Waldron) is presented. Next, distribu- tions, abundance, migrations, spawning activities, and environmental relations of five selected species are summarized: Pacific halibut (Hippoglossus steno- lepis), which spawn seaward of the shelf in winter and migrate shoreward in spring (Best); Pacific herring (Clupea harengus pallasi), some of which migrate inshore across the shelf in spring and spawn along the coast from April to July, and others winter inshore (Wespestad and Barton); walleye pollock (Theragra chalcogramma), which spawn perhaps ubiquitously in subarctic waters but certainly in large numbers inshore in spring along the shelf edge (Smith); yellow- fin sole (Limanda aspera), which migrate inshore across the shelf in late spring and early summer and spawn in the southeastern part of the shelf (Bakkala); and salmon (Oncorhynchus spp.), which as adults migrate shoreward across the shelf to various coastal areas in late spring and early summer to spawn and, as juveniles, migrate seaward across the shelf to oceanic areas in spring and summer (Straty). Then, having provided a general, descriptive assessment of temporal and spatial variations in fish distributions, migrations, and spawning locations in relation to environmental conditions or events (Favorite and Laevastu), we evaluate the total biomass of ecosystem compo- nents and their interactions, using simulation tech- niques (Laevastu and Favorite). After having perused the following chapters, one cannot help feeling that OCSEAP studies, rather than having completed baseline data acquisitions for the eastern Bering Sea shelf and achieved basic or prelim- inary understanding of processes, have actually only permitted us to define the nature of the tasks that have to be accomplished if we are to obtain a suffi- cient understanding of the eastern Bering Sea ecosys- tem to assess the impact of oU exploration and exploitation and protect the extant living marine resources. Up to this point little new information has been obtained and previous knowledge has been only marginally extended. One talks about warm and cold years, but nothing is known about what occurs under the ice. One attempts to measure primary and secondary production without concern about patchi- ness, predation, and reproduction, or evidence of biomass balances. One conducts casual studies of ichthyoplankton without regard to the limitations in our knowledge of their identification or location and never pursues organisms much beyond the egg stages. One knows little or nothing about the distributions or migrations of juveniles for the years before they enter the fishery as adults. One notes that cold conditions appear to delay shoreward migration of adults but also that in some instances migrations of more northerly stocks do not appear to be influenced at all by equivalent temperatures. And, in most cases, knowledge of year-round distributions and migrations of adult fishes and specific causes for aggregations or movements is still severely limited. It is easy to say that the ocean is a turbulent regime, that each year is different, and that only gross assessment can be made and only general relations ascertained. However, there is an apparent order to the Bering Sea ecosystem that implies in many instances some very specific fish responses that we do not understand; only dedicated, multidisciplinary research (fisheries oceanography) will provide the answers. ACKNOWLEDGMENTS I wish to express my appreciation to all partici- pants, not one of whom failed to complete his task on schedule in spite of the pressures of other work. I also thank H. Larkins and M. L. Hayes (NWAFC) for comments, M. Gregory (NWAFC) for editorial assistance, E. Zweifel (NWAFC) for document preparation, and Robert Peterson (SAI) for figure coordination. REFERENCES Barnaby, J. T. 1952 Offshore fishing in Bristol Bay and Bering Sea. U.S. Bur. Comm. Fish., Spec. Sci. Rep. Fish. 89. Barnes, C. A., and T. G. Thompson 1938 Physical and chemical investigations in Bering Sea and portions of the North Pacific Ocean. Univ. Wash. In: Ocean 3(2): 35-79 (Append. 1-64). Overview 453 Dall, N. H. 1882 I I EUson, J. G. 1950 Report on the currents and tempera- tures of Bering Sea and adjacent waters. Append. 16— Rep. for 1880. U.S.G.S. Gov. Print Off.: 297-340. D. E. Powell, and H. H. Hildebrand Exploratory fishing expedition in the northern Bering Sea in June and July 1949. U.S. Fish. WUd. Serv., Fish. Leaf. 369. Favorite, F. 1974a Physical oceanography in relation to fisheries, pp. 157-179. In: Bering Sea oceanography: an update 1972- 1974, Y. Takenouti and D. W. Hood, eds. Inst. Mar. Sci. Rep. 75-2, Univ. Alaska, Fairbanks. 1974b Riddle of Bering Sea soundings resolved. Mar. Fish. Rev. 36(22): 30-32. Favorite, F., A. J. Dodimead, and K. Nasu 1976 Oceanography of the Subarctic Pacific Region, 1960-72. Inter. N. Pac. Fish. Comm. Bull. 33. Favorite, F., and G. Pedersen 1959 Bristol Bay oceanography, August- September 1938. U.S. Fish. Wild. Serv. Spec. Sci. Rep. Fish. 311. Favorite, F., J. W. Schantz, and C. R. Hebard 1961 Oceanographic observations in Bristol Bay and the Bering Sea 1939-41 (U.S.C.G.T. Redwing). U.S. Fish. Wild. Serv. Spec. Sci. Rep. Fish. 381. Fishery Technology Laboratory 1942 Report of the Alaska crab investiga- tion. Fishery Market News 4 (5a). Gilbert, C.H. 1924 Experiments in tagging adult red salmon, Alaskan Peninsula Fisheries Reservation, summer of 1922. Bull. U.S. Bur. Fish. 39: 39-50. Goodman, J. R., J. H. Lincoln, T. G. Thompson, and F. A. Zeusler 1942 Physical and chemical investigations: Bering Sea, Bering Strait, Chukchi Sea during the summers of 1937 and 1938. Univ. Wash. In: Ocean 3(3): 81-103 (Append. 1-48). King, J. E. 1949 Experimental fishing trip to Bering Sea. U.S. Dep. Int. Fish Leafl. 330. Moiseev, P. A., ed. 1963-1972 Soviet fisheries investigations in the northeast Pacific. Munford, J. K., ed. 1963 John Ledyard's journal of Captain Cook's last voyage. Oregon St. Univ. Press, Corvallis. Rathbun, R. 1894 Summary of the fishery investigations conducted in the North Pacific Ocean and Bering Sea from July 1, 1888 to July 1, 1892, by the U.S. Fish Com- mission Steamer Albatross. In: Marshall McDonald, Comm., Bull. U.S. Fish Comm. 12:127-201. Ratmanoff, G. E. 1937 Explorations of seas of Russia. Pub. Hydrol. Inst., Leningrad. No. 25. Straty, R. R. 1974 Ecology and behavior of juvenile sockeye salmon (Oncorhynchus nerka) in Bristol Bay and the eastern Bering Sea. In: Oceanography of the Bering Sea, D. W. Wood and E. J. Kelley, eds., 285-319. Inst. Mar. Sci. Occ. Pub. No. 2, Univ. Alaska, Fair- banks. Wigutoff, N. B., and C. B. Carlson 1950 S. S. Pacific Explorer, part V. 1948 operations in the North Pacific and Bering Sea. U. S. Dep. Int. Fish. Leaf. 361. Wilimovsky, N. J. 1966 Synopsis of previous scientific ex- plorations, 1-5. In: N. J. Wilimovsky and J. N. Wolfe, Environment of the Cape Thompson region, Alaska. U.S. Atomic Energy Comm., Oak Ridge, Tenn. Shelf Environment W. James Ingraham, Jr. Northwest and Alaska Fisheries Center Seattle, Washington ABSTRACT Monthly mean environmental conditions of ice, tempera- ture, runoff, and salinity for winter (January -March), May, July, and September are presented for the eastern Bering Sea shelf area. Mean geostrophic flow (0/50 db) in summer (July) reflects the northwesterly flow at the shelf edge and a west- ward flow from the Yukon River area seaward south of St. Lawrence Island. Tidal currents as derived from a hydrody- namical-numerical model reflect a NW/SE flow over the outer shelf and NE/SW flow north of the Alaska Peninsula, between Nunivak and St. Matthew Islands, and in the southern Gulf of Anadyr. Although data are fragmentary, environmental conditions over the shelf are highly variable and examples of N-S surface wind components (1946-75), ice cover (1976-78), and bottom temperatures (1969, 1973, 1976, and 1977) reflect departures from mean conditions. INTRODUCTION The eastern Bering Sea shelf is characterized by several features that have pronounced effects on its hydroclimate. Sloping gently westward over 500 km, the shelf terminates abruptly as the continental slope drops somewhat precipitously into the Aleutian basin, resulting in a coastal-oceanic water interface that largely divides the surface area of the Bering Sea in half. The shelf is isolated from the direct effects of circulation in the Pacific Ocean by the Alaska Penin- sula and sheltered from the Arctic Ocean by the narrow, shallow Bering Strait. Seasonal advance and retreat of ice cover in winter and spring and the extensive discharge of runoff in spring and summer dominate conditions and processes in the hydroclimate. Since physical proc- esses are discussed in another section of this book, we are concerned here primarily with conditions over the entire shelf area, insofar as data are available. For increased perspective on recent studies I will make brief reference to the considerable knowledge gained from early historical studies, present mean conditions for January-March, May, July, and September, and discuss variations in ice, winds, and bottom tempera- tures. 455 Although even today the Bering Sea is considered a remote and hostile area, it must have been consider- ably more treacherous and forbidding to early ex- plorers and particularly whalers with limited re- sources and technology. The loss of 33 whaling vessels north of Bering Strait in 1871, only four years after the United States acquired Russian America, resulted in environmental studies in the northern shelf area (e.g., a survey of Bering Strait from U.S. Schooner Yukon in 1880) and an assessment of logs and reports from whaling and other vessels crossing this area (Dall 1882); considerable information was derived. First, on the subject of ice and surface temp- eratures: (1) the southern limit of ice was shown first to extend to 56 °N in the central Bering Sea and well below the Pribilof Islands in the shelf area; (2) in cold years ice remains around St. Paul Island until late May; (3) ice opens to the west of St. Lawrence Island first, for the sea is frequently navigable there and north of the island into Norton Sound at a time in the season when the passage between Nunivak Island and St. Lawrence Island is still blocked with decaying ice; (4) water opens first where the south- erly cold set away from the ice is strongest; (5) the extent of ice in different years varies and depends on winds, with prolonged winds from the north bringing southward loose flow ice which is prevented by the formation of new ice from returning northward when the winds change; (6) the water retains nearly its normal temperature to within a very short distance from a large field of ice; (7) the Yukon discharge during summer months gives rise to definite local currents; (8) in summer at depths of 18 m (10 ftm) a higher temperature is found than in adjacent deeper water; (9) surface temperatures in Norton Sound obtained at anchor on 7 July over a 24-hour period varied from 6.7 to 12.2 C, a range of 5.5 C. Recent 456 Fisheries oceanography cruises and satellite imagery have largely borne out these early observations. Second, on the subject of surface currents it was recognized that: (1) discharges from numerous large rivers and the westerly and northerly marching tide create a distinct set in the vicinity of their mouths more or less directed by local winds; (2) the Kusko- kwim and Bristol Bay rivers have a southwesterly or southeasterly set depending on prevailing wind and tide; (3) discharge from the Yukon River in calm weather proceeds in a northwesterly direction, but under northerly winds or ebb tide, or both, part or all of it passes to the south of St. Lawrence Island; (4) away from the influence of the rivers the current is influenced a great deal by the direction and force of the wind; (5) north of Nunivak Island (60°4l'N, 166°04'W) the U.S.S. Corwin reported NW/SE tidal currents of 0.5-2 kn (~ 25-100 cm/sec) while in the ice; (6) at St. George Island a steady current of 1-2 kn (~ 50-100 cm /sec) from the west is sometimes ob- served for several days; (7) currents across Bering Strait (0705-1415 hrs, 5 Sept. 1980 from aboard U.S.S. Yukon under fresh northwesterly wind, transit W to E) varied as follows: E21.8°S, 3.4 kn; N24.3° W, 1.5 kn; W26.3°S, 3.8 kn; S1.3°E, 2.5 kn; W31.0° N, 2.2 kn. Numerous reports of drifts of whalers under various conditions of ice, winds, and sea were presented for analysis and provide considerable insight not only into tidal flow but also into general drift. All this information reflects the complexity in flow and the difficulty even today of evaluating flow from a limited number of current meters. Third, although subsurface data were sparse be- cause of limited instrumentation, there were reports of conditions across Bering Strait and of environ- mental divisions over the shelf of considerable value. The U.S.S. Yukon hydrothermal cross section of Bering Strait in September 1880 clearly indicated: the 6 C temperature difference across the strait^vertically isothermal conditions of 2.6-3.3 C (36.7-38 F) in the western third, the temperature gradient (3.3-7.2 C) regime in the central third, and the slight stratification of water of 7.8-9 C (46-48 F) in the eastern third with maximum values near the surface, offshore of the eastern side of the strait. However, in spite of the temperature differences implying that opposing currents exist in the strait, the fact that strong northerly and southerly currents had been experienced on both sides of the strait by various navigators suggested that there was not sufficient evidence to report that a southerly cold current occurs on the west side of the strait, and a warm northern current on the east ; but it was recog- nized that the northerly flow on the east side of the strait nearshore was composed largely of river runoff. In addition, on the basis of temperature data only, the Bering Sea was separated into three divisions: (A) the shallows (e.g., Bristol Bay, Norton Sound, etc.) and the area seaward to the 46 m (25 ftm) contour; (B) the moderate depths, between the 46 m (25 ftm) and the 137 m (75 ftm) contours, from Bering Strait to the Alaska peninsula; and (C) the deep waters to the south and west. Attempts to improve on this scheme continue even today but no significant deviation from this early assessment has been accept- ed. Favorite (1974), prior to OCSEAP and PROBES studies, distinguished four subdomains on the basis of temperature-salinity relations: Gulf of Anadyr, West Alaska Coast, Mid-Shelf, and Shelf Edge (Fig. 29-1). Subsequent PROBES studies have focused on the frontal zones between these subdomains only in the southern part of the shelf (e.g.. Coachman and Charnell 1979). Although most investigators are familiar with the University of Washington oceanographic studies of the 1930's (e.g., Barnes and Thompson 1938) and those by the U.S. Bureau of Fisheries in 1938-41 (e.g.. Favorite and Pedersen 1959; Favorite et al. 1961), few are aware of the studies by the U.S. Naval Electronics Laboratory in 1949, because portions of them were classified. Some results are available (Saur et al. 1952) that describe general summer conditions, and temperature and salinity relations were used to identify seven water masses in the eastern Bering and Chukchi Seas. Oceanographic studies conducted from 1953 to the Figure 29-1. Shelf sub-domains ascertained by T-S rela- tions in 1 X 1° quadrangles. Shelf environment 457 present time during fisheries investigations under the aegis of the International North Pacific Fisheries Commission (INPFC) and other studies during this period have been summarized by Dodimead et al. (1963) and Favorite et al. (1976); and recent OCSEAP and PROBES studies are summarized in Chapters 1-8 of this book. MONTHLY MEAN CONDITIONS Ice information presented is based on data from Potocsky (1975) and recent satellite imagery; temp- erature and salinity data are based on the current (June 1979) NODC geofile (including OCSEAP data) and Japanese and NWAFC fisheries cruises. These last data have been computer-processed into monthly means by V2° (lat.) and 1° (long.) quadrangles, and anomalies are derived from these means; however, only summaries of conditions for winter (January- March), May, July, and September are presented and discussed. There are several limitations to the presen- tations of mean conditions in this area: first, data are not sufficient for summaries by V2 X 1° quadrangles (nor even 2 X 2°), and thus there are numerous gaps in each month and only limited data from October to March. Second, equivalent data were not acquired in warm and cold years; thus in some instances the values represent mean conditions, in others (in warm or cold years) possibly extremely anomalous conditions. (The numbers of observations and years are not presented here, but are available at NWAFC.) Third, the paucity of data can also result in isolated anomalous values and tonguelike protrusions that may merely reflect a preponderance of data from a single year rather than a significant environmental feature. Nevertheless, the data reflect a high degree of continuity of conditions throughout the area that cannot be obtained in any other manner. Ice In the northern Bering Sea, monthly mean air temperatures vary from 10 C in summer to — 20 C in winter and extreme conditions extend this range over 10 C higher in summer and lower in winter. Tem- peratures are mostly below 0 C from October to May and thus sea ice plays a dominant role in the ocean environment. Freezing occurs at about —1.8 C in this area, and the formation of ice limits winter cooling of sea water to this temperature. The actual nature of the ice edge in any one location— compact or spread out— is also influenced by advection of ice due to wind drift. Much of the increasing solar radiation in spring is consumed in the melting of ice and the seasonal increase of sea surface temperatures is delayed. However, when ice is no longer present, dilution and abrupt warming at the sea surface result in a shallow, stable surface layer which retains most of the incom- ing radiation and permits nearly subtropical condi- tions to occur, particularly in the inshore areas. The area is free of ice for only a short period, usually from June or July to August or September. Usually by September ice begins to form near Cape Dezhnev and along the western shore of Bering Strait; by October it extends into the Gulf of Anadyr, across Bering Strait, and even into the eastern part of Norton Sound. Mean conditions (Fig. 29-2) indicate a progressive southwesterly advance through March or April to the edge of the continental shelf, but a rapid retreat from May to June, from north of the Pribilof Islands to Bering Strait. Potocsky (1975) presents an excellent summary of mean ice conditions and indicates wide deviations from the monthly means. (See also the ice section of this volume.) Surface and bottom temperature Mean surface temperatures primarily reflect the annual cycle of insolation; secondary effects are observed in nearshore areas affected by river runoff and reduced cloud cover, and in areas of ice cover where spring warming is delayed until the ice is melted. Winter (January-March) surface temperatures of less than —1.5 C (Fig. 29-3) reflect the approxi- mate extent of ice cover from Bristol Bay to Cape Navarin. The warmest water, 2 C, is offshore in the ice-free area, and intermediate temperatures of 1-2 C Figure 29-2. Monthly mean location of ice edge. 458 Fisheries oceanography 170' 175" 180° 175° 170° 166° 160" 155" 150° Figure 29-3. Long-term mean sea surface temperature (C) for January to March, May, July, and September. are found just south of the ice on the southern portion of the shelf. By May, temperatures over the southern shelf have increased to 3-5 C, but over most of the shelf relatively low temperatures between —1 and 1 C prevail, marking the initial stage of the heating cycle after the ice melt. Surface temperatures for July show that most of the area has warmed rather quickly to 6-8 C and inshore temperatures have reached a maximum except in the extreme north, with Bering Strait temperatures near 5-6 C and negative temperatures still present on the east coast of Siberia north of the strait. Local warm spots are seen with temperatures of 12 C in Anadyr and Bristol Bays, and in excess of 16 C in Norton Sound. August is the warmest month for surface water temperatures; the majority of the area reaches 8-10 C, thus reducing the contrast between the offshore areas and the coastal areas, where temperatures are nearly the same as in July. A dramatic change does occur, however, in Bering Strait with the apparent south- ward movement of cold east Siberian coastal water and the apparent northward movement of the warm Alaskan coastal water, producing a pronounced gradient of nearly 10 C between the cold western and Shelf environment 459 warm eastern sides of the strait. By September the effects of the beginning of the cooling cycle are evident; most of the area temperatures are between 7 and 9 C and coastal temperatures have decreased to about 10 C or less. The gradient across Bering Strait has decreased to about 6 C. Bottom temperatures result primarily from the processes of vertical mixing and diffusion and to a lesser extent from the processes of horizontal mixing and advection. Mean temperatures lower than — 1.5 C in winter (January-March) were distributed over most of the shelf, extending seaward as far southward as mid-Bristol Bay and as far offshore as St. Matthew Island, but close to shore in Anadyr Bay (Fig. 29-4). Surface cooling, the formation of ice, and overturn have produced a vertically isothermal water column which extends to a depth of at least 50 m and some- times as much as 75 m. Warmer water of 2-4 C is found at depth near the shelf break associated with the general cyclonic (counterclockwise) circulation over the Bering Sea basin and the northward exten- sion of Alaskan Stream water from the North Pacific Ocean. Thus, isotherms tend to be roughly parallel to the bathymetric contours. The presence of 0-1 C water just north of the Alaska Peninsula is probably correct in the mean sense, but the serpentine nature Figure 29-4. Long-term mean bottom temperature (C) for January to March, May, July and September. 460 Fisheries oceanography of the contours may be misleading due to the absence of data during extreme winters when this area may be less than —1 C, or mild winters when temperatures are probably near 1-2 C; for this reason, the presence of a northeastward flow which is implied if one interprets the contours as warm advection from the southwest cannot be substantiated. Another area that appears to be influenced by some cross-shelf, warm advection, however, is Anadyr Bay. By May there has been little change in the bottom temperature distribution in the outer shelf area between the 100 and 200-m isobaths, but the shallow area from inside the 50-m isobath to the Alaskan coastline, from Bristol Bay to the mouth of the Yukon River, has warmed from the winter condition of less than —1.5 C to between 0 and 2 C, with highest values occurring closest to shore. Therefore, the dominant feature is the area of remnant negative temperatures which appears as a large southeastward tongue located about mid-shelf. Apparently the dominant process in the warming of coastal bottom water at this time is downward mixing of surface- warmed water rather than horizontal warm advection from the south. The negative isotherms associated with the tongue appear discontinuous and in isolated areas; again, this is probably the effect of inadequate temporal distribution of data. Because the best data coverage is for midsummer, mean bottom temperatures for July show the most complete picture of shelf temperatures. The coastal area has warmed to more than 10 C from Bristol Bay to Norton Sound, thus greatly accentuating the inshore edge of the cold tongue, where a large tem- perature gradient (10-12 C) occurs. Temperatures at the seaward side of the tongue remain about the same north of the Pribilof Islands. The cold core of the tongue appears to warm slightly as the portion south of 58 °N is now more clearly defined by the 1 C isotherm; whereas to the north small remnant patches of <— 1.5 C water still occur. An increase of about 1 C is seen in a separate tongue advecting wairm ( >4 C) water northwestward from Unimak Island. Similar to surface temperatures, bottom temperatures across Bering Strait reflect about a 4-C west-to-east gradient. By September the warm coastal water (8-10 C) reaches its greatest seaward extent, thus sharpening the already high gradient on the eastern side of the cold tongue and shifting it slightly seaward to about 100 km offshore. A division appears to develop in the cold core near 58° N as 2-3 C water influenced by the westward movement of the warmer coastal water appears to separate the tongue into two areas of temperatures lower than 1 C, to the north and south. The reappearance of some negative temperatures in the southern core reflects additional September data in cold years, because July temperatures are higher (0-1 C), and it is unlikely that the apparently colder water in September was advected from the north across the intermediate warm area. In Norton Sound bottom temperatures increase to 6-8 C. And on the northern edge of the cold core, a tongue of 2- C water intruding from the west replaces the negative temperatures of July in Anadyr Bay. River runoff Although data on river runoff are sparse, general characteristics of the area may be deduced from data for the Yukon River, the largest individual source of fresh water. The Yukon River drains 45 percent, the Anadyr River 20 percent, the Kuskokwim 12 percent, the Nushagak 3 percent, and the remaining rivers 16 percent of the total land area draining into the Bering Sea. Because of regional differences in precipitation, however, it may be misleading by as much as a factor of two to estimate the magnitude of discharge from the major rivers by comparing their drainage areas alone. The only data for an extended period of time are for the Yukon River at Ruby where 22 years of monthly mean discharge data are available. Long- term monthly averages (Fig. 29-5) indicate a pro- nounced seasonal pattern of runoff; low during the freezing period, decreasing from 2 X 10^ m^ /sec in November to 1 X 10^ m^ /sec in April, it increases rapidly to a peak of 13.5 X 10^ m^ /sec in June and then steadily decreases to low levels by Novem- ber as the freezing period begins. Some examples of variability are evident in the last four years, which are characterized by alternating high and low values. Beginning in 1975, above- normal spring and summer runoff was followed by low runoff in spring and summer of 1976. This pattern was repeated with above-normal runoff in spring 1977 and below-normal values during spring and summer 1978. The lowest values in the last four years occurred in 1978, and a shift in timing of the peak discharge occurred from June to July. This shift also occurred in two other low-discharge years, 1960 and 1963. For three consecutive summers (1976, 1977, and 1978) during the period after the annual peak discharge, runoff has been below normal. Surface and bottom salinity Mean surface salinities increase from November to April by the processes of salt exclusion during freez- ing and vertical mixing and then decrease from May to October by precipitation and river runoff, which delivers a large pulse of fresh water to the surface layer. Maximum values ( >33.3°/oo) occurring in Shelf environment 461 MONTHLY MEAN YUKON RIVER RUNOFF AT RUBY o o o o c 3 cr 10 12 5 6 7 Month Figure 29-5. Monthly mean Yukon River runoff (m^ /sec) at Ruby 1975-78 and long-term (22-year) mean. northern Anadyr Bay and north of St. Lawrence Island (Fig. 29-6) are beheved to be local results of the freezing process inasmuch as they are separated by large areas of lower salinity. Values lower than 32.0^/00 occupy the majority of the shelf area south of St. Lawrence Island and out to about the 100 m isobath. By May several changes are evident. A decrease of l-2°/oo is evident in inner Bristol Bay, in Norton Sound, and north of St. Lawrence Island, and a long tongue of values less than 31.0°/oo extends eastward from the Yukon River mouth to just south of St. Lawrence Island. By July ice is no longer present and the river runoff has resulted in a pro- nounced dilution of surface water along the west coast of Alaska and in Anadyr Bay; a minimum value of less than 16.0°/oo occurs in Norton Sound. Despite this marked dilution inshore and lesser changes offshore (evident in the westerly displace- ment of the 33.00/00 isohaline— nearly 200 km), the 32.0°/oo isohaline appears to be in a consistent position roughly paralleling the 100-m isobath along an approximate line from Unimak Island to Cape Navarin. By September the effect of further dilution seaward of the shelf is evident as very little water of greater than 33.0°/oo remains over the basin, but conditions in the shelf area are similar to those of July except for slightly higher salinities near the Yukon River mouth and in Norton Sound, probably the result of mixing and decreased runoff. Mean bottom salinities below 50 m generally increase with depth. In January-March values of bottom and surface salinity are largely the same in water less than 50 m deep^essentially isohaline conditions (Fig. 29-7). In water deeper than 50 m, however, stratification appears to develop, as shown by the displacement of both the 32.0 and 33.0 bottom isohalines about 100 km farther shoreward (northeastward) than the corresponding surface isohalines. This salinity increase with depth allows the higher temperatures at these depths to exist in a vertically stable density field. There is little change in bottom salinities in May except for the presence of some slight dilution to less than 31.0°/oo in inshore Bristol Bay and south of St. Lawrence Island. By July, however, strong dilution has occurred in Norton Sound with values of less than 24.0<^/oo and to a lesser extent in Bristol Bay (< 28.0O/oo). A notable feature is the absence of dilution in Anadyr Bay at the bottom despite marked dilution at the surface; this condition persists through September as the area with salinities greater than 33.0*^/oo continues to enlarge. Also in September dilute water extends farther offshore in Norton Sound and bottom salini- ties increase slightly in inshore Norton Sound and Bristol Bay, probably because of horizontal mixing. Water masses as reflected by surface and bottom temperatures There have been numerous attempts to describe water masses in the Bering Sea using the classical method of T-S (temperature vs. salinity) diagrams to recognize groups of data representing water of simi- lar characteristics, which are put into envelopes based on dominant processes and/or the resulting water structure. Recent studies (e.g., Takenouti and Ohtani 1974) have considerably more detail than the early ones, and OCSEAP studies have refined consid- erably our knowledge of water masses and their inter- actions over the southern shelf offshore from Bristol Bay (Kinder 1977, Kinder et al. 1978); however, four major distinct environments stand out— Gulf of Anadyr, West Alaska Coast, Mid-Shelf, and Shelf Edge (Favorite 1974). Although salinity characteristics influence the water mass definitions in the Gulf of Anadyr (where 462 Fisheries oceanography 175° ISO" 175° 170° 165* 160* 155" Figure 29-6. Long-term mean sea-surface salinity ("/oo) for January to Marcii, May, July and September. high salinities are associated with low temperatures due to the salt exclusion process) and also near the shelf edge area (from the influence of Bering Sea basin water), the most striking feature of the shelf environment is the temperature regime, particularly in the mid-shelf and coastal waters. The coastal water mass is clearly distinct from the others because of its relatively constant vertically homogeneous nature, whereas marked stratification occurs in sum- mer at mid-shelf. Thus, to illustrate the extent and seasonal changes in mean water mass conditions, it is expedient to look at temperature differences between surface and bottom. During January-March, mean temperature dif- ferences (surface minus bottom) are near zero over most of the shallow shelf as ice cover and deep vertical mixing result in isothermal conditions (Fig. 29-8). The most significant feature is the oblong area of relatively large negative differences (—2 to —4 C) which occurs all along the outer shelf, where seasonally cold surface temperatures override a deep, warm, slightly more saline water mass. By May there Shelf environment 463 I is little change except west of St. Lawrence Island where two extreme values of — 2 C and 2 C indicate local conditions at the beginning of spring warming after completion of ice melting. Dramatic changes are seen by July and continue through September as the dominant summer feature, a difference of 6-8 C, develops over the remnant tongue of cold mid- shelf bottom water. The eastern side of this area is clearly defined by the sharp decrease in temperature difference, which approaches zero in coastal water; the western side, near the outer shelf, is less well defined since the difference is due to the only slightly warmer bottom temperatures. Flow Although the most reliable data on flow come from direct current measurements, few data have been taken and only in local areas of individual interest such as Bering Strait and Bristol Bay (Coachman and Charnell 1979), so that indirect methods are still required to infer net flow or mean circulation; these methods have proved effective in gaining knowledge of the general flow in the clima- tological sense which concerns us here. One must keep in mind that since tidal components on the order of 50 cm /sec occur daily which reverse either diumally or semidiurnally, any estimate of mean flow Figure 29-7. Long-term mean bottom salinity (°/oo) for January to March, May, July, and September. 464 Fisheries oceanography 55 150 Figure 29-8. Long-term mean surface minus bottom temperatures (C) for January to March, May, July, and September. will probably be at best about one order of magni- tude less than the maximum of the instantaneous flow. An NWAFC tidal model study (Hastings 1975) indicated that the reversals of flow are largely NE/SW in inner Bristol Bay, between Nunivak and St. Mat- thew islands, and between St. Matthew Island and the Gulf of Anadyr; and NW/SE over much of the re- in the absence of direct measurements, the geo- strophic approximation is widely used to calculate mean currents from the vertical distributions of temperature and salinity relative to an assumed level of no motion, usually below 1,000 m depth. This method has been used to establish the existence of northwestward mean flow of the Transverse Current (Favorite et al. 1976) seaward of the shelf edge, but is invalid in shallow water over the shelf except for showing anomalies of specific volume from a geo- metric level. The mean data for July did cover sufficient area to warrant examination, and contours Shelf environment 465 Figure 29-9. Tidal currents (cm/sec) calculated from numerical tidal model for (a) four hours before high tide, and (b) four hours before low tide at the eastern boundary (from Hastings 1975). of anomaly of dynamic height computed from the surface relative to 50 db (0/50 db) suggest several major features of flow (Fig. 29-10). Northwestward flow is indicated over the basin between Unimak Island and Cape Navarin and a component veers to the right around Anadyr Bay, eastward to St. Law- rence Island, and finally northward through Bering Strait. There is a suggestion of a weak westward cross-shelf flow just north of St. Matthew Island and also a northeastward flow into Bristol Bay along the north coast of the Alaska Peninsula, with a westward meandering return flow out of northern Bristol Bay. Although the action of surface winds could result in markedly different flows in the upper and lower portions of this 0-50 m layer, which are well isolated in summer by thermal and haline stratification, these results are interesting and should form the basis for further investigations. VARIABILITY McLain and Favorite (1976) have presented monthly anomalies of air and sea surface temperature in the vicinity of the Pribilof Islands (57°N, 170°W) from 1967 to 1975, which reflect not only a long- term downward trend (which reversed itself in 1977) but also considerable monthly and annual variability. Mean monthly north and south geo- strophic wind components computed for 1946-75 for the grid point 57°N, 170°W (Fig. 29-11) indicat- ing potential periods of cooling (and ice advance in winter and spring) and warming (and ice retreat in winter and spring) reflect the extreme variability of meteorological conditions; and the 12-month running mean indicates an approximate three-year cycle and a long-term trend of increasingly dominant northerly ANOMALY OF DYNAMIC HEIGHT (0/50db) July Ji_L_J I l_ 170' Figure 29-10. Long-term mean anomaly of dynamic height (0/50 db) for July. 466 Fisheries oceanography I960 1963 1966 1969 1972 1975 Figure 29-11. Monthly mean wind (m/sec) components at 57°N, 170°W, 1946-75. (Positive values indicate northward component, negative values indicate southward component.) Twelve-month running mean showing annual mean trends. winds (cold conditions). Kihara (1971) has reported the extreme variability of the extent of Alaskan Stream Extension water over the southern part of the outer shelf. Although polar waves considerably alter air temperature, wind, and precipitation, annual variability in shelf conditions is most easily shown by the extent of ice cover and by bottom temperatures. Ice In recent years (1976-78) the extent of ice cover for November (when ice first extends seaward) was minimal in 1976 and maximal in 1977; whereas for April (usually the month of greatest ice cover), the maximum extent occurred in 1976 and the ice edge has been farther north each succeeding year, with the greatest changes occurring on the eastern side of the shelf (Fig. 29-12). Thus, there is no correlation between spring conditions and subsequent fall con- ditions (i.e., summer conditions can obliterate cold trends established in the spring). Furthermore, although the normal southerly winds in spring result in a compact ice edge and hasten its northerly retreat, northerly winds can cause anoma- lous conditions. In April 1976 an unusual and large southward displacement of part of the ice edge in the Bristol Bay area resulted in ice immediately north of the Alaska Peninsula which forced vessels out of the area (Northwest Fisheries Center 1976).' This displacement not only caused water in the open area in the northern Bristol Bay to be subjected to early insolation (and thus warming), but because spring warming at the surface had already begun, considera- bly altered the temperature structure of the water in the area into which the ice was advected, resulting in extremely anomalous conditions in spring 1976. Bottom temperatures, 1973 and 1977 Because of the scarcity and fragmentary nature of bottom temperature data, I have chosen to illustrate * Northwest Fisheries Center, 1976, Ice conditions in the eastern Bering Sea, April 1976. Northwest Fisheries Center Monthly Report, April 1976, 7-9. Northwest and Alaska Fish. Cent., Nat. Mar. Fish. Serv., NOAA, Seattle, Wash. Shelf environment 467 Figure 29-12. Mid-month position of the ice edge, April 1976-79 and November 1976-78. variability in the months of July 1973 and July 1977, for which time data were available for Norton Sound and Bristol Bay, even though it is known that ex- tremes of temperature occurred in other years (warm in 1968-69 and 1977-79, cold in 1971 and 1976). Bottom conditions reflect general surface conditions (warm or cold, not specific temperatures); the reverse is not true. These data indicate that in summer a general shelf categorization of warm or cold years may be unreliable for determining bottom condi- tions—in July 1973 positive anomalies of 3 C oc- curred in Norton Sound and negative anomalies of 2 C in Bristol Bay. However, in July 1977 negative anomalies in excess of 4 C occurred in Norton Sound and positive anomalies of 2 C in Bristol Bay (Fig. 29-13). Thus, anomalies of bottom temperature of different signs and magnitudes may occur over relatively short space scales within the shelf area. It is apparent that before one can consider movements of fish throughout the shelf, one needs to know the environment over the entire shelf. 170° 165 160° 155 150 <.i...::::2 '^^^^^^^ ;>ffOTTOM TEMPERATURE f.^.:-'^ cruiiteii listed In Table 30-1. *Bolhyloiim ichiiildii Includes Lfuroghsaui sllltiiut, L scluiUtlll, and B. slllbiui. * Prolomyclophiis llioinploni Includes Klecirona orrllca. fiipsoruj liifludM /oiii/>D'ivrfnji »p,, and /., Icucoptarut. ''Sllchaeldiv lncludF» Lumpenldav. ** Aminvdylei hexaplerut inK\»A«% A loblanut penonalut, '* l.imonda probatriilea inchidev/. puncloliai't " Plfurifiiccle* quailnlubc'CuloUn iiivliidcv I'lal ported). Reports of Soviet studies have usually been from sources which were originally published in Russian and translated into English; for the most part, station positions had to be estimated from small charts contained in the original publications, and in some instances it was impossible to determine catch by station. On the other hand, information collected during OCSEAP cruises, obtained from processed reports of the Northwest and Alaska Fisheries Center (NWAFC) or from the data file of the National Oceanographic Data Center (NODC), is accurate with respect to position and catch. Because of the varying degrees of accuracy in position and for convenience in presentation, all data have been presented on the basis of geographic areas measuring ¥2° of latitude by 1° of longitude. These areas are designated by the coordinates of the lower right corner in west longi- tude (i.e., the area from 57 ¥2° N to 58° N and from 165°W to 166°W is designated 57y2°N, 165°W) and by the lower left corner in east longitude. Because it was not always possible to determine the number of net hauls made at one position, a station as used here is defined as one or more net tows made at a geographic location within a few hours. Samples separated in time by a few days or more were counted as separate stations. This treat- ment tends to minimize effort, and in terms of net hauls sampling effort is much greater than shown here. The total sampling effort east of 174°E (Fig. 30-1) was estimated using cruise reports that listed station positions; the total includes only the cruises listed in Table 30-2. Sampling covered the entire Bering Sea but with a concentration of effort along the conti- nental slope south of the Pribilof Islands. Division of Figure 30-1. Number of stations at which ichthyoplank- ton samples have been collected, 1955-78. Ichthyoplankton 473 1 ICHTHYOPLANKTON SAMPLES W^^T' H March-May kUT/ — 5 I ! « jiH|M| yy^-^ \ i ^"^ * VT v-^-'^^^B B ^ ^- 1 E rmT^siHi^H^H irr 1 JI^HP^TJIlII I ki I ""^ 1 i^^^^^^^^^^^^^^^H 1 ^ , )i^^^^^^^^^^^^^^^V^ I ?^S?fiS?Tf6 ^ r* uSsQjSH^ttlT^ ^ S ^■i ^3^^^^^^^^^^^^^UK^\ 5 IT ■■* T ^T xL>^4-^¥i^^^^^^l^^^^FA '~y'~~lTr~4~Jl~tTr Ir 12 2 1 1 1 1 L-4---WwRWCIwJ^^^BfeS ■ftxr~httH~fi^4^4i^^I^ 1 1 If 11 T*4---r'T^T ' f -^^^^3r\^"B hfjjj^^ T ~t 1 R 7TTlT|lJLuW=5tTr ^^H\^^ iU 3 t 4 el 2Ti4!Vplli-^^iVrTiT^^ ' 11 4 ^4^ i 1 1 3igTTTTr , ^^+-2^ \ JiBFt.; v X^ * 4 1 5 6.?: 8 6 1. ^, -^.;. , 3I ^irv- r\UFH 1 S: r 1 116 ^'13 i'';i2 yj^^^4iA Cjr7T--CtrTW~l4--]— 1 216 i:*tOl-i..^ \ VA-^-V'-T \ \ 3 3' b 1* i"*^^',^^ \_l---4-A — \ A-- ~ 1 liii-Z-lii^f^^— T — 1 — \ilX^r\\Z^ 1 ~r ~ "^^i^j^ ^~-T~i^r\\---T\\rT t/*r"^f3lj — J — r?~r-f--I— "■. ^ 1 jl?*^ — i-T — 1 — \Z^^^X--T\CX ' 7 — r^'rlZT iTr^-f—l—Ul -i ^•^ .■ , _L- 1 — 1 — \ lJu-Vm — \\ J -L ^i' ~^ — 4-— ■""! — 1 — r^Jh-rT'L^ j[ — 1 AA^\\\\^S^X\\\\- Figure 30-2. Number of stations at which ichthyoplank- ton samples have been collected in spring, March-May. 63' 1 ICHTHYOPLANKTON SAMPLES ^^P-f I Sept ember-No vember SUrT-f- 1 1 1 ■ '''SPBT"""! 6 ™ ^^-^ > > -li J" !• 1 1 J — i — r )^^H 1 I^^^^^^^^^^^^^^^H~jL7 tt T2 ■k ^^^^^^^^^^^I^mJI/ 1 fir ^^^^^^^^^^^P^^^^iLJ '" 1 i 1 . I^^^^^^^^^^^^^H rrr 1 i ^1 A l^^^^^^^^V^ 59' sr 55' 53 51 1 1 1 ^ '^^^H^^^Hr ^ X 1 ^, "uA^B^H^^^^ jl 1 LJt^-O i 3 F 1-3 iiTT i-^^-^^-wiVf^^r^ rill 1 Ul^U-W^^^^fc^ ^ T 1 1 1 aHU—VsTTl iU^^Fua '1^ ■ 1 TTT4iVr"-rTT J»^i^ 5 1 2 2 iSlV^vrTXA^Bruvn f ■ V 1 niiinn^Vin^pFtT-j^ '^P^^f4 "~ 3 J ^ 443fflVrjiW-W "" 1 " iJ. ^ 4j44mV^iiK^Pria^ S. fj" /^^^/^ir^(^-~?J^ -^1 iniii4iyp^Fi3-A-j^^ i.5iiti^gS^+:E^ J~'T~~fj^ 1 ±1+ il^PWVnilPrAniC- ^Tj~~l~~~Lj~~i — /t^v/JJLJZI 11- ^S^4-HxPrTX^ ■ ' ~ ~~ __ "" ' T ZX— i— V — Tl \ Ji--WT\ J ^rjf^-^^ --^ •'■ TS" IIL- ---V-l^^T^V^ 1 1 J 1 L ^ __ Figure 30-4. ton samples November. Number of stations at which ichthyoplank- have been collected in fall, September- the effort into quarter-years (Figs. 30-2-30-5) shows that 35 percent of the sampling was done in spring (March-May), 55 percent in summer (June-August), only 9 percent in fall (September-November), and less than 1 percent in winter (December-February). Thus, fish whose eggs or larvae are planktonic during the six months from September through February were undersampled in comparison to those that are plank- tonic during spring and summer. Pleuronectidae, for example, show considerable diversity in time of spawning. Yellowfin sole spawn in midsummer and thus are not available for capture with plankton nets in spring or early summer. Halibut and turbot, on the other hand, spawn in winter and by the following spring and summer their larvae are difficult to capture with plankton nets. Pollock, one of the most abun- Figure 30-3. Number of stations at which ichthyoplank- ton samples have been collected in summer, June-August. 1 ICHTHYOPLANKTON SAMPLES WK^T- 5 ? 1 1 1 11 6 H Uecember-February UTt~-L~h-i 1 ^" ^^^^B* t„. KTTrn^ - - - 1 +44-1^ 11 ^^^^^^^^^^^^^^^HmLj rttTT" ^ k* Xh-h^I^^^ ^ -^^^^^^^^H ' ^ 1 ,>1^^^^^W^» 1 -1 4-t{^^HiiiiiMI-' ~'f~~~^tjji~Ji~f!^^ ~^' ■ ^^, i T~-4-sr~iT-~i^^ ~^ ^ 1* -L^r^^^^B^H^ 7 |i|i ■*, ' -T wrnj^^^^^Rr 5 31 2 U -44-HTmprmBrl ^ .'^1 1 ' 44-FHAxi+-\^Klll ~'~ ^XlXV^^^r^^WtjM Clu^~T~~f^^ \ ■ 'I 4-4-H=rEyyBrXrM V zCPrV\-rxMP^Llr ... _ . -444-HiaFrXPr 1 ^■- u^V. .^ j- J lEP; Viiifflfni^rS: I~Zj iT-~J-^3l~4^ — r--l-~l~XLL_ ' — T^ \ 1 1 i-v4— ;V2^*\lt'^^7u4r-^ \ \- " ^'l If^^l*^^ (^-"-^l^ 1 ~ __. , i—[-A—\--Wv\^r\,--\--^^r'\^J^ i~^Jr^^ ■ ' > y?^ — 1 — \\ \--\^ — \ \ ■i r^ tiJKMMi — TjI--i---rrL]ji- ■" " ^ — Iljl4---T\0\- W~r~Hv— i_ rir^~t~-l~JIL i ~ ~" __. si-. ' i_X— i--+ — T — 1 — T \I-Wt — T \ ^ -^ ■~^ ^- -V-H-ttXV-V^i; ^ ~j — / — LI / j~^f^ H&Si ■ 1 zti ^^44djp^Vr^ =±=J — iT—t — /- 1"--/ J-TTr~1 1 1 h -H 11 I L-4— V- T \ 1 — V- Figure 30-5. Number of stations at which icthyoplank- ton samples have been collected in winter, December- February. dant fish in the Bering Sea, spawn mainly during spring and their larvae are relatively scarce in samples collected during late summer and fall. A further obstacle to inter-cruise comparisons is the difference in types of tow used in making the collections. Vertical net tows, which usually strain a relatively small volume of water, capture only small numbers of fish larvae, and the more uncommon taxa are seldom found in collections made by this method. A horizontal tow at the surface— a neuston tow for example— may produce large numbers of eggs and larvae, but these may belong to only a few taxa whose larvae characteristically inhabit this zone. Greenlings and Atka mackerel, family Hexagram- midae, and sculpins of the genus Hemilepidotus are often abundant in neuston collections, whereas TABLE 30-2 Ichthyoplankton cruises in tiie Bering Sea east of 174°E, 1955-1979 Cruise Year Vessel Lat. N. Long. Number Type Date From To From To Stations" of tow' Net Reference 1 1955 Oshoro Maru 8-20 July 52° 58° 175'^ E 163' W 17 s FLN 2 1956 Oshoro Maru 23 Jul-13 Aug 52 63 174 E 168 W 34 S FLN 3 1957 Oshoro Maru 7-11 July 52 54 174 E 172 W 9 S FLN 4 1957 Brown Bear 3-29 Aug 52 56 177 E 166 W 20 0 IKMT 5 1958 Oshoro Maru 8-27 June 52 62 175 E 175 W 21 g FLN 6 1958 Zhemchug 29 Jun-5 Aug 53 64 174 E 160 W 124 s,v IKS-80 7 1958 Zhemchug 19 Aug-18Sep 55 63 174 E 162 W 95 s,v IKS-80 8 1959 Alazeya 13-19 March 54 59 174 E 165 W 60 s,v IKS-80 9 1959 Oshoro Maru 21 Jun-12Jul 52 58 174 W 167 W 17 s FLN 10 1960 Oshoro Maru 19Jun-13Jul 52 60 175 E 163 W 13 s FLN, PC 11 1961 Oshoro Maru 1-14 July 53 60 174 E 172 W 11 s FLN 12 1962 Ogon 21-31July 58 63 174 W 163 W 88 s,v IKS-80 F/r.80/113 13 1963 (Halibut Comm. charter) May 54 55 166 W 165 W 10 14 1963 Oshoro Maru 14 Jun-14 Jul 54 59 177 E 167 W 7 S FLN, IKMT 15 1965 Seskar March-May 54 61 179 W 160 W 95 S,V,T IKS-80 16 1965 Sesl^ar June-July 54 65 180 163 W 86 S,V,T IKS-80 17 1966 Oshoro Maru 13 Jun-31 Jul 53 65 177 E 164 W 32 S FLN 18 1967 Oshoro Maru 12 Jun-20 Aug 53 65 175 E 161 W 22 s FLN 19 1968 Oshoro Maru 12 Jun-8 Aug 52 64 177 E 162 W 28 S FLN 20 1969 Oshoro Maru 10 Jun-2 Aug 53 58 180 160 W 35 S FLN 21 1970 Oshoro Maru"* 6 Jun-16 Aug 52 58 172 W 161 w 50 S FLN 22 1971 George B. Kelez 26 May-9 Jun 53 56 173 W 168 w 6 0 B-60 23 1971 Oshoro Maru^ 17 Jun-23Jul 52 58 174 E 161 w 31 S FLN 24 1972 Oshoro Maru^ 11 Jun-12 Aug 52 65 175 W 161 w s FLN 25 1972 Professor Deryugin March -Jun 54 62 174 E 164 w 0,H IKMT (CI.) 26 1973 Oshoro Maru^ 11 Jun-23Jul 52 59 180 160 w 36 s FLN 27 1974 Oshoro Maru'' 7 Jun-22 Jul 52 58 178 W 167 w 41 s FLN 28 1975 Shunyo Maru 4 Mar-3 Jul 54 61 161 W 179 w 121 V 1.4mLN 29 1975 Discoverer 20May-15Jun 53 60 158 175 w 108 v,o Im.NIO 30 1975 Oshoro Maru"' 13 Jun-5 Jul 52 59 180 160 w 78 s FLN 31 1975 Discoverer 15-26 Aug 54 60 159 172 w 76 v,o Im.NIO 32 1975 Miller Freeman 13-24 Nov 54 62 162 170 w 38 v,o lm,NI0 33 1976 Surveyor 17 Mar-26 Apr 54 58 162 174 w 57 v,o Im.NIO 34 1976 Miller Freeman 26 Apr-31 May 54 59 175 W 158 w 56 S.O Sam. .B-60 35 1976 Oshoro Maru'' 14 Jun-25 Jul 52 64 174 E 162 w 39 s FLN 36 1976 Discoverer 5-17 Aug 55 66 161 168 w 80 v,o Im.NIO 37 1977 Miller Freeman 16 Apr-13May 53 58 174 W 163 w 134 S.O Sam.. B-60 38 1977 Oshoro Maru'' 12Jun-24 Jul 52 60 178 W 163 w 51 S FLN 39 1978 Miller Freeman 11 Feb-18Mar 54 60 164 W 177 w 32 S,0 Sam., B-60 40 1978 T. G. Thompson 11-29 Apr - 157 S,0 N, B, NIO 41 1978 Ogon 10 Apr-20 May 54 62 162 W 180 349 S,V X.80 42 1978 Oshoro Maru June-August 66 S FLN 43 1979 Hakuho Maru 23-30 July 55 65 168 W 173 w 5 S,0 ORI-100 Faculty of Fisheries, Hokkaido Univ. Faculty of Fisheries, Hokkaido Univ. Faculty of Fisheries, Hokkaido Univ. Aron. 1960 Faculty of Fisheries, Hokkaido Univ., Musienko, 1963 Musienko, 1963 Musienko, 1963 Faculty of Fisheries, Hokkaido Univ. Faculty of Fisheries. Hokkaido Univ. Faculty of Fisheries, Hokkaido Univ. Kashkina, 1965, 1970 1957a 1957b 1958 1959 1960 1961 1962 Dunlopetal, 1964 Faculty of Fisheries, Hokkaido Univ., 1964 Serobaba, 1968 (report only pollock) Serobaba, 1968 (report only pollock) Faculty of Fisheries, Hokkaido Univ.. 1967 Faculty of Fisheries, Hokkaido Univ., 1968 Faculty of Fisheries, Hokkaido Univ., 1969 Faculty of Fisheries, Hokkaido Univ., 1970 Faculty of Fisheries, Hokkaido Univ., 1972 Dunn and Naplin 1973 Faculty of Fisheries. Hokkaido Univ., 1973 Faculty of Fisheries. Hokkaido Univ., 1974 Serobaba 1974 (pollock-vertical sections) Faculty of Fisheries, Hokkaido Univ.. 1975 Faculty of Fisheries. Hokkaido Univ., 1976a Wakabayashi, Mito and Nagai, 1977 NODC Data" Faculty of Fisheries. Hokkaido Univ., 1976b NODC Data' NODC Data' NODC Data' Waldron and Favorite, 1977 Faculty of Fisheries, Hokkaido Univ., 1977 NODC Data' Waldron and 'V'inter, 1978 Faculty of Fisheries, Hokkaido Univ., 1978 Waldron. 1978 PROBES, 1979" Moiseev and Bulatov, 1979^ Faculty of Fisheries, Hokkaido Univ., 1979 Haryu, Endo and Nishiyama, 1979 Cruise No.— an arbitrary number assigned for the purposes of this summary. ^Type of tow: S=horizontal surface tow probably to a maximum depth of 1 m; 0 = oblique tow from the surface to a depth and back to the surface; V = vertical net tow; H-a horizontal net tow at some depth below 1 m; T-Total; definition of this not known, but the terminology appeared in Serobaba (1968). 'Net: FLN— fish larva net. See references for specifications which varied slightly between 1955 and 1977. IKMT-Isaacs-Kidd Midwater Trawl (Isaacs and Kidd 1953). IKMT(CI.)— an IKMT with a closing type cod end. IKS-80—". . .nets of 140 mesh with a mouth diameter of 80 cm." (Kashkina 1970). B/r-80/113— a reverse cone net with dimensions and mesh the same as the IKS-80. B— bongo net. size not specified. B-60— bongo net with a mouth diameter of 60 cm and mesh of 0.505 mm. Im— a simple one-meter ring net. N— a neuston net. Sam.— A Sameoto type sampler (Sameoto and Jaroszynskl 1969). NIO— a 2-meler National Institute of Oceanography trawl. PC— underway plankton catcher. '' Reports of the Oshoro Maru cruises since 1969 do not list station positions or catch, but give only total stations sampled. ' Catch by station for these cruises was obtained from the OCSEAP data file at NODC. 'Cruise report PROBES 78. Leg I, 8 April to 1 May 1978. ^Moiseev, E. I. and O. A. Bulatov 1979. The state of pollock stocks in eastern Bering Sea. Pacific Research Institute of Fisheries and Oceanography (TINRO), Vladivostok 1979. 474 Ichthyoplankton 475 flatfish larvae are almost completely absent from such tows. In general, an oblique tow from near bottom to the surface appears to capture a large number of species as well as large numbers of speci- mens. Examination of Table 30-1 shows that many of the collections, including most of those done aboard the Oshoro Maru, were made with surface tows. Because those cruises represent a majority of sampling efforts west of the continental slope, the species composition in that area is biased towards those taxa with surface-dwelling planktonic larvae. A third point to consider in comparing samples col- lected during different cruises is the area coverage and station spacing. Economics, logistics, and program objectives usually dictate that a particular sampling effort be confined to a single vessel often operating within a small area and /or sampling at relatively few stations. The Oshoro Maru cruises often covered a large area but with widely spaced stations. Most of the effort by the Miller Freeman consisted of samp- ling at closely spaced stations within a relatively small area. Cruises by the Zhemchug and Seskar covered large areas within which stations were spaced at moderate intervals, but most sampling was restricted to the continental shelf and slope. As a result of restricted areal coverage combined with uneven seasonal distribution of effort and the limitations of the type of tow that predominated in certain areas, the distribution of many species is not well defined. Yellowfin sole larvae are not reported in very many collections; this may be because their spawning area appears to be well up on the continen- tal shelf in moderately shallow water, whereas much of the sampling effort was devoted to the outer con- tinental shelf and continental slope. Pacific herring spawn in intertidal areas and their larvae must remain close to shore while they are planktonic, for they are reported in very few collections. Pollock spawn along the continental slope and outer shelf and their eggs and larvae are seldom captured in water of less than 75 m depth. Myctophids and bathylagids are prob- ably undersampled because of the predominance of surface tows in the central basin where they inhabit the deeper waters. Much of the sampling effort, then, has been focused on the distribution and abundance of a few species, in particular pollock and yellowfin sole; collections of other species have been incidental to this main objective. It is likely that definitive infor- mation about the ichthyoplankton community in the Bering Sea can be gained only through a concerted cooperative effort by Japanese, Soviet, and U.S. biologists to promote at least one year-round samp- ling effort covering the entire Bering Sea, with sampling at moderately spaced stations, and including at least one oblique and one neuston tow at each station. AVAILABLE FAUNA AND CATCH There are generally considered to be about 300 species of fish in the Bering Sea, divided among 150 genera and 40-45 families (Quast and Hall 1972, Wilimovsky 1974). Of these, four families— Petro- myzontidae, Squalidae, Rajidae, and Acipenseridae— with 6 genera and 12 species, are extremely unlikely to appear in plankton samples. In addition, larvae of Salmonidae and Gasterosteidae, with 7 genera and 16 species, are seldom found in marine plankton, al- though they do occur in samples from estuaries and brackish water areas. There remain 34 families with 137 genera and 270 species that can reasonably be expected to occur in plankton samples either as eggs or larvae, or both. Collections resulting from the cruises discussed here contain 87 separate taxa, of which 60 are species, 13 are genera (9 of these are also included in the specific listing), and 14 are families. All told, they include representatives of 52 genera in 24 families; the various taxa and the collections from which they were obtained are presented in Table 30-2. Thus, larvae of over one-third of the genera, over half of the families, but of only about one-fifth of the species of fish present in the Bering sea as adults have been collected and identified. Most of the species not included in the list are in the large families Cottidae, Zoarcidae, Cyclopteridae, and Stichaeidae, which account for some 170 species. The 14 families not included in the collections contain only 23 species. Because of the variations in spawning time and the logistics of vessel cruises, it is seldom possible to sample the entire spectrum of ichthyoplankton during any single cruise. Examination of Table 30-2 shows that of a total of 87 taxonomic categories of larvae, no report lists more than 38 taxa and four reports list 10 or fewer taxa. Certain of the reports in Table 30-1 (e.g., Serobaba 1968, reporting on the 1965 cruise of the Seskar) deal with single species and hence do not include listings of other larvae and eggs that were almost certainly present in the collections. DISTRIBUTION AND ABUNDANCE OF ICHTHYOPLANKTON Clupeidae^ The Pacific herring, Clupea harengus pallasi, is the only member of this family in the Bering Sea; as an ^The arrangement of families presented here follows that proposed by Greenwood et al. (1966). 476 Fisheries oceanography adult it is abundant and commercially valuable, but larvae were caught at only 10 stations on three cruises in the summer and fall of 1975 and the summer of 1976. Only 24 herring larvae were caught, all at shallow water stations in Bristol Bay and Norton Sound (Fig. 30-6), and the largest catch was 9 lEirvae. A partial explanation of the scarcity of herring larvae is that spawning takes place in inter- tidal areas and larvae apparently remain close to shore during the period when they would be susceptible to capture with plankton nets. Most of the sampling effort shown in Fig. 30-1 was in areas outside those occupied by herring larvae. Herring eggs are demersal and none were reported. Osmeridae Four species of smelt occur in the Bering Sea as adults and two of these were present in plankton samples. Capelin, Mallotus villosus, were caught at 76 stations on 10 cruises in spring, summer and fall (Fig. 30-7). With one exception, all occurred south of 60° N almost exclusively over the continental shelf and extending into the easternmost part of Bristol Bay. The single exception was a capelin larva col- lected just south of Bering Strait in August 1976. Collections made aboard the Miller Freeman in the spring of 1977 (Waldron and Vinter 1978) indicate that capelin are more easily caught with neuston nets than with deeper fishing bongo nets; more specimens were collected at night than during the day. A single juvenile Rainbow smelt, Osmerus mordax, was caught in quadrangle 61V2°N, 167°W in Novem- ber 1975. Unidentified osmerid smelt were reported at five stations on two cruises in spring and summer. Osmerid smelt are demersal spawners and no eggs were reported. Bathylagidae The deepsea smelts are represented in the Bering Sea by five species (one name is of questionable validity), and larvae of three of the species were found in plankton samples. The most frequently collected species was the northern smoothtongue, Bathylagus schmidti, which was caught at 69 stations on 10 cruises in spring, summer, fall, and winter. Catches of B. schmidti were generally fewer than 5 larvae per tow with a maximum of 22 at one station. The Pacific blacksmelt, Bathylagus pacificus, was collected at 51 stations on six cruises in spring and summer. Catches of B. pacificus were usually fewer than five larvae per tow with a maximum of eight at one station. The stout blacksmelt, Bathylagus milleri, was collected at only two stations on one cruise in summer; the two specimens should be described as large juveniles or small adults even though they were collected with a plankton trawl. In contrast to Osmeridae, deepsea smelts were collected over the outer continental shelf, the continental slope, and the deep central basin south of 60° N and from 165°W to 178°E (Fig. 30-8). Since there did not appear to be any marked difference in distribution between the northern smoothtongue and the Pacific blacksmelt, their distribution was plotted as one figure. The two stout blacksmelt were collected at two stations in quadrangle 52y2°N, 178°E. As far as can be determined, none of the bathylagid larvae were captured with surface nets. Because most of the sampling effort over the deep central basin was with Figure 30-6. Number of stations at which larvae of various taxa have been caught. Each symbol represents one station. Figure 30-7. Number of stations at which larvae of Os- meridae have been caught. Ichthyoplankton 477 surface tows, the lack of specimens from that portion of the Bering Sea may be a sampUng artifact. Although eggs of deepsea smelt are pelagic and identifiable, none were reported. Chauliodontidae Of this family of bathypelagic fish, a single species exists in the Bering Sea, the Pacific viperfish, Chaul- iodus macouni. Larvae and juveniles were col- lected at only three stations on two cruises in sum- mer, all of which were close to the Aleutian Islands between 172°W and 178°E (see Fig. 30-6). Eggs of this species are pelagic but none were re- ported. Myctophidae The family of lanternfishes includes seven species in the Bering Sea; larvae of three of these were reported from plankton collections. Larvae of the northern lampfish, Stenobrachius leucopsarus, were collected most frequently of the three species and were caught at 35 stations on nine cruises in spring, summer, and fall. Northern lampfish larvae were distributed from the edge of the continental shelf across the slope and along the Aleutian Islands westward to 177°E (Fig. 30-9). Only one larva of this species was caught north of 57° N, and there were no large catches made at any station. Larvae of the bigeye lanternfish, Protomyctophum thompsoni, were collected at 10 stations on three cruises in spring and summer. They were caught at two stations along the continental shelf and at eight stations close to the Aleutian Islands westward to 177°E. Catches were small, consisting of only one or two larvae at any one station. California headlightfish, Diaphus theta, were caught at only four stations on one cruise in summer. These specimens were juveniles or young adults found close to the Aleutian Islands from 174° W to 178° E. Almost all myctophid larvae were caught with obliquely or vertically towed nets that fished well below the surface. As with bathylagids, the apparent absence of myctophid larvae from the central deep- water region of the Bering Sea may be an artifact caused by surface sampling rather than an actual distribution pattern. Little is knovni about eggs of myctophids and although they are assumed to be pelagic, none were reported. Gadidae There are four or possibly five species of codfish in the Bering Sea, and larvae of three of these were found in 21 plankton collections made during all seasons. Larvae and juveniles of saffron cod, Eleginus gracilis, were caught at 13 stations on two cruises in August 1975 and 1976. All of those classified as larvae were caught in Norton Sound and Bering Strait; the juveniles were caught in outer Bristol Bay (Fig. 30-10). Most samples contained only one or two specimens, but one sample from quadrangle 63y2°N, 164°W yielded 54 saffron cod larvae. Eggs of saffron cod are demersal and none were re- ported. Larvae of Pacific cod, Gadus macrocephalus, were caught at five stations on three cruises in spring and summer. The total catch was only 10 larvae, five from a station in quadrangle 52° N, 174°W in the Aleutian Islands and five from four stations on the 1 LARVAE OF BATHYLACIDAE I^HI^J-J! m 11 [ ! jjm ! ^IF'^^H - «.-.. 1 W"M • 1 <. *^ 1 1 i PI^^HK^I'-f+TXri In ' I ^^^^^^^^^^^^^^^^^H ^^^^^^^^^^^^^^^B' ^7^JfV/~~^-~/// rTY~-i-4-u-J_- J- ^^^^^^^^^^^^^^HP < Li^flM^^^HT^ 77-~-L 7^~^A~LJ /TT"^Nj^-4l-i_j_ ^ ^ lAl^V^^^^^Uf^ t r j_V-4-ri«^^^^^H^^^V^ ~T^-Ir/T^~J~I//TH^A-MaE-L t L mW^^mM T--Zl3/~~r-XjrT~rHW-+-l3E - ' .~ - ~ *"^ T Sf2jT7~fit^ -^. ^ —J a4::4-kfTlA«rY,\^ ( ^ \ L— 4— ^ — r \ \^lflf^-! Ti' 1-- -*: ^ -1 LJi-4 — \-^\ iHPV^' ^■■- ^ DS^^^ttrf^^JrF ~^ ^ ■*^ X -0 \ ]-\\^SKr^^r-\^''' \ '^ ' 1 1 ■^■\ lxK-l-^Knrk*i4^ \ 1 T-+j!jT^^XJr- ~^ 1 2 T^'j \ 1 -Si^^TPrJrrvi '3 ,3!7 : 4 : _lip^-H-x?VriJ 1 1 5 8 B^A - +-+=tT_L4--rTju 1 t-.^H *. tS^i- -''M^\li5-TTTlX (~~-JJj~7-~^f^^ 1 V / ' ■"" ■ ^ f-' ' 1 _J — r'TllJU--r'n\^ - at » -* flT tffi4iu3pf5: ~tj~~f~4^^4^Xlr^ 1 — 165 170' 175- 180- 175- i7{: 165 160 155" ISO' 1 LA. Buar ,1F MvrT,->p»Tn:,c ^K^^J^L i. ^^^B 1 r i Hi n Protomvctophum thompsoni H ^^^UmMl M It 1 - fc ''i ^^^^^^B^^VlIi I--' 1 ~ ~ "^ ' ^ ■ I?^~t4-XJ /rT+^-i—L ~ ~ • 1 ~ -1= ~ ■i c. -I]7T-X3rr7H^--i-3-- J ' - ~ fS~r^Srrf-- ^^ ■" 'fT ' ~ ~ J , ' ~ , (^"i m - ^ _. '1~~~j-~~J-^^7f^ — f — h~l--~l— ^■' mI [ ■01 qX3 'i~T — /— jLTt — 1 — t- 1 (1 2H~4^4^ i; t - ^^^F* ^cj~fiteiip~f~ --i-- 'i) i ' *r fJZZrrT-^wZijMif J_g ■ -T t^^TT- 1 'zh 1 -1^^ Figure 30-8. Number of stations at which larvae of Bathylagidae have been caught. Figure 30-9. Number of stations at which larvae of Myctophidae have been caught. 478 Fisheries oceanography continental shelf south of Nunivak Island. They were caught in surface, oblique, and vertical net tows. Eggs of Pacific cod are demersal or benthonic and none were reported. Larvae identified only to the family Gadidae were caught at 41 stations on eight cruises in spring and summer. Catches at individual stations usually amounted to fewer than five larvae, but at one station just west of Unimak Pass, 59 larvae of this taxon were caught during a spring cruise. The distribution of larvae identified only as Gadidae extended from the Aleutian Islands to Bering Strait and from 163° W to 176°W. Stations at which these larvae were caught were scattered from well up on the continental shelf across the slope and, rarely, over deep water. It seems likely that the unidentified gadids from north of St. Lawrence Island belonged to a different taxon from those to the south. Walleye pollock, Theragra chalco gramma, have probably been the primary objective of more ich- thyoplankton surveys than any other fish in the Bering Sea. As a result the distribution and abun- dance of their eggs and larvae are quite well known and have been described in various reports (Musienko 1963; Serobaba 1968, 1974; Waldron and Favorite 1977; Waldron and Vinter 1978; Waldron 1978; Nishiyama, Volume 2 of this book). These reports show larvae to be most abundant between Unimak Pass and the Pribilof Islands along the continental slope, with total distribution extending to 59° N in spring. During summer, larvae show a more widespread distribution from the Aleutian Islands to 64V'2° N, and from well up on the continental shelf in Bristol Bay across the central basin to 177°E. Because of the large number of samples containing ^65 170 175' 180' 175' 170 165- 160 155 ISO H LARVAE OF GADIDAE ^^99Bf~44-Jv ^H 1 — p ^ ^ hk t.^M^^^^^^^^^^^^^^^^^^H ■" 3 \ i'^^^^^^^^^^^^^^^^^ frT^^B^Pr^ 1 t'\\^~\\ \ \ ^ \ ^^^^^^^^^^^^^^^^^^^^^K p^^^^^^^p 'i'^!f^-UrtT~i-~4~J~\^ -L ^^^^^^^^H^ -* ■i \X9^^^^^Bm^ \ '■* n ^^-''^W^f^^^rt 111 \TLX-wf^w^^ml ■ _ ^:LX.»i^\Tr\x^^x\m t \r --y^nxuvH^»\ \^ L^/ ' 7 7 — f~~ I I 1 7 — j- — J — 1 1 yT vft L JIUv- H-Jr T Gmta. Vd -^J^tdrr^^ "E , ^mLuu4-+- rTijR^:\- v.. i i^^^:^^ ^"ii 1 mi-W ++t imFr:TY-\'.] 4-J^~t~^~i~4Jyt~i~^~~[~-\--\~-^» V ■^' 5i™ 1 -ir4 "^ " "1 '■fHr\^-r^ V' \ ^ lZj^j~~T-4-43'^T~iT "' 'iiAiXiLU«CTJHt^^ ■ 'Cvi (iJkU&w^'Tr \ T^L'A' \ \ J ~ ^ r- ©lJ*-J#P^T \ }lW^ \^ 2w~~7----fJ~~Tf~~^| I . exasJM-hTT X A-v-r u 4 ^P H-HjrtVHrl^ ■1: r JKSl --144-TiJ^-A-T J , " ^7 '4-W4-4-tnnil4--V-TiJ T^-A-J. W ■ rMTr~j~-^L^ ^T~ .-J-- ji_" ^ ■\ u^XX\\\'\XX\\'l^ _ ittffHip;^ larvae it was possible to show distribution for two periods, spring and summer, and for two types of net tows, surface tows and combined oblique and vertical tows. Larvae were caught in surface tows made at 40 stations on two cruises in March-May, and at an undetermined number of stations on one other spring cruise for which the catch by station could not be determined. These surface samples showed the larvae to be distributed along the continental slope in a somewhat discontinuous pattern (Fig. 30-11). During the same two cruises larvae were caught in oblique net tows at 124 stations and on another spring cruise at 15 stations. Distribution shown by these oblique tows covers essentially the same areas as shown by the surface tows but there was more continuity (Fig. 30-12). During June-August, larvae were caught in surface tows at 76 stations, mainly on nine cruises of the Oshoro Maru, and were shown to be widely distrib- uted over the entire Bering Sea (Fig. 30-13). Fewer vertical and oblique tows were made during summer and larvae were caught at only 31 stations; yet these showed a wide distribution from the Aleutian Islands to 63i/2°N and from 164°W to 178°E (Fig. 30-14). Sampling for larvae has been intensive along the continental slope and outer shelf, but there is a gap in our knowledge of their distribution and abundance over the deep central basin. Comparison of catches in surface and oblique net tows shows that relatively fewer larvae are caught in surface tows and at fewer stations. Most of the sampling over the central basin was with the surface tows (see Figs. 30-1 and 30-3) and most of those were during summer. It is rea- sonable to assume that sampling by means of oblique tows would reveal a larger population of pollock 1 THERAGRA CHALCOCRAMMA 5ilpTTT^ 1 1 1 ^ llWi.-,.J V fl 1 Ij-+-Vi^--__J ^^■^^laaii^^^^^^^HMijT 1 ^i ^ 1» XuMJ^^^ma^^ ^^^^^^^^^^^^^^^^■-L il ^^^^^^^^^B^-^SUzK 1 ^ 1 ^^^^^HP^/ #' "M—T Jl / 1 N "1 , ^^^^HW~ Lit r-j ■L-L^l r \ u 1I^^^^^^^^^^^^^^^^^^M\ / JB^y f L 1 1 ir\^^~^X-L_. 1 \ 1 ""^^^^^^^^^^^^^^^^^V ^ / fntiurtiiSSx. ' i \ ^ ( J IM-^^l^^^^^H^ r I t /r~/-ij/7r^St ^ wrT\^^^^t^m:A " r-J t ii^-U-LlTi — h-K^lL 1 1 1 -^'-^'HyF 'T^t^^'jj^^^p^^ r~/-r7/r^-tLrT ~fl-- -4-^ ' i 1 LL-l--|--y ' "A 1— \J^nP\iJi r P IT r^l-~4ljiri~~r~ -L \£. ^ ^ 1 r_L4 — \ — pn "" K'^+'-d^^^ y^ Lj^ r rf-IIrTT Tt— ^-I- L~ ^^ *i 1 n__L-4— r \ \ AJ^Pr'^ '■^ -J~7~ r 7 -/JTr 7 / Tt^- i f^. V 1 L_T-4-4— T" \ ^^F*> iV ^^ ~~f Tt't i~Iif r / T I -I A 1 ^ ,i X — \ — 1 \ ^HKU^^*' V ^^ 7-2~r T--/--tr / r rT~ X ~^ ^^ '^ _. . > - 1 1 'TiiMirin'Ai'^ ^ r Cj /'/~~f~Itr^w — TT— i—LL ■ ■ T^ 111 |-v4-'T3^^yk*L-MVH \ \ --X/T~-f -tjK T~"r-t--4-— 1— ■" ^ "* T^^^rf^^ ' T-ViM \-^ -iT~r~T--trTir^-T-t . ^ ^\s^ljf^^^\\-\-\\_\-- 2"^~7 — j~--~LJ^r' ^ T\ I ~4— f ^ j ^ IsjLfii: — i — V^^'^t \" -\ — \ — X^ - T^^T^r-~l~^^f ~i~~T ^-~+~~ ^L_a :*jiS3— Vn^ — T \ Ju-i-'A — \ \~ r^p^f-uljjj — f — 1~ , 2 zsS---X-\\X\X\\X\ rF$Qi>f44^^^^ ^ ^^ eu v^. ^^^\\\\X^a\\a ZLiZ&4-rTX-Prr: JTr-F-l-Jdrirf-f^^^-l^ _ -^U-^^r-r^iX^r^r^:^^ Figure 30-10. Number of stations at which larvae of Gadidae have been caught. Figure 30-11. Number of stations at which larvae of Theragra chalcogramma have been caught with surface tows in spring, March -May. Ichthyoplankton 479 175- 170" 165° 1 IJUtVAE OF l^^^f ffll I ■? ■ THERAGRA CHALCOGRAMMA Jr^~-l } \ "+^ ■ Oblique or Vertical Tows mlj]~\\ ~\T f -- H^c^B "— -■"-■ < M ILiZL-Hia^^H^I^^H^^^I hL T 1 1 ^kj^^^^^^^^^^^^^^^^^^H ^^^^^^^^^^^B^^^tlA^l r LA^^^^^^^^^^^^^^^^^^I 1 ^^^^^^^^^^^H^^^^^^K ^W^Wiy 1 t~t~h-L-Lj I \ I l^^^^^^^^^^l^^^^V^ 1 ^^^^^^^^^^^^l^^^^^v '~i~~~fS~ir~J~^^ ^. (•* I \^^^^^^^^^^^^B^^ -"s ialsS^H^^^^^^^^a^^i '^?C7T~l'^+-J^7[r~~TH--^^^ ^ l1>^H^^^^^^^^^^^^^\ T_l-4--»B^^K^^BK^ T-^7"~~fT^4lJ^/Tr~T-4--5^ti: ~£ :l_ 1 1 Li — U-^?^'\ tiL-V^^SR^jJI /-^/r^-4li r~H-+4^ 1 liM^'~^^^^rV'^^Pr^ Crt^ 4 1 1 1 rT_-U-^— r- \ \'\'^fY^ f S 2 1 1 1 L-i — i — r \ A J^^^L V U' -flS. 1 3tiMiLU-+-H:;T:^B?fmi2^ izFf~4^^^ ^^ OTfmija-+-+-ao!Esyr^-^sr\,#^ tjj^^ 1 fffl^siAiiUJHyiEflifia^'^t:^ 2 1 Ea « 6 •'-fc^^f'y 1 1" Jf-^i^ \ J rqj^-A^^ 2^ 6 ]3S5P^i*T l32v---V\ ^A-^ V ^=5 '^^\^f^\'\yP^'r\\\Sr- 1 '2£(-^2JK2— ^ — T — \jiL-+--T\t-V lTv~?-3ZJ/ — f?~r--f— t ^ _ 2L1^ -ilf— — I — \ — iTLJr-rni^ '~7r~^+--XjTiT--f--l— -■ ^^^^S^-TT^^^Jri^ / — r~4^-~h'r\^'~r*r-~i~~A~=L^ ^t\J ^«. ^ t — ^—4-^1 / T~T^— f— t^i^^^Dl' ^ r J^ _1— 4—4— -Lt — 1 — 1J-^A--t\L- ^ / 7 / — \—-L WAT'T-^^ — 1 — t- Figure 30-12. Number of stations at which larvae of Theragra chalcogramma have been caught with oblique or vertical tows in spring, March-May. 63" er 59 57- 55 53' 51 165'- 170- 175- ISO' 175' 170' 165- 160° 156° 150° 6T 61° 59" 57° 55- 53 51- 1 LARVAE OF ■ THERAGRA CHALCOGRAMMA H Surface Tows, June-August ^Tt-ZJ]/ J — t — LXTt Jim ■ r ^^B iT :i t J?"" 1 ] ^*^^n ^.. '^ / 1 5 ^ #< ^r ^ [TTi 1 Dz 1 ~ ^ 1 14 ' ^- _Lj 41^ 1 f^ h ^ tTTL ] 1 u^ A| TS ( '\ 1, 1 ii T_r rfsT 1 1 iL 'ff^ \ 1 ^ '1 \ ,i- ', 31 1 (V ~ ^ 1 1 Ti h" ■ t J , ■^^ Sj; i-^J- al 1 ■ iifV 12 >s _Tn 1 .' P ^ nnr- ri^ III " .p>i f ij* r^'TT^ 1 ■ . i^jij^ ^ ~^ ^ Ti -4 175 180 175 170 165 160 Figure 30-13. Number of stations at which larvae of Theragra chalcogramma have been caught with surface tows in summer, June-August. Figure 30-14. Number of stations at which larvae of Theragra chalcogramma have been caught with oblique and vertical tows in summer, June-August. larvae in the deep central basin than is shown by surface tows. Since pollock eggs are among the few routinely identified in plankton samples, their distribution can be shown in some detail. They are present generally along the outer continental shelf and continental slope from the Aleutian Islands to 60° N (Fig. 30-15). Within this area they occur from late February to late July but are most abundant from about mid -March to mid-May. Abundance of eggs has been treated in greater detail than that of larvae, probably because eggs are caught more readily than larvae in the vertical net tows which were used on most of the So^'iet cruises. Quantitative estimates of Figure 30-15. Area within which eggs of Theragra chalco- gramma have been caught, and centers of abundance in five years. abundance are shown for cruises made in 1959 (Musienko 1963), in 1965 (Serobaba 1968), and in 1972 (Serobaba 1974), although the last shows only vertical distribution along section lines. Esti- mates of egg abundance based on oblique net tows are given for 1976 (Waldron and Favorite 1977), for 1977 (Waldron and Vinter 1978), and for 1978 (Waldron 1978). Maximum abundance of eggs beneath 1 m^ of sea surface, as shown by vertical net tows, was 598 in March 1959 between Unimak Pass and the Pribilof Islands and over 2,000 eggs at some time from March to May 1965 immediately north of Unimak Pass. Maximum numbers of eggs beneath 1 m^ shown by oblique net tows was 1,268 in May 480 Fisheries oceanography 1976 east of the Pribilof Islands, 1,041 in April 1977 north of Unimak Island, and 958 in March 1978 northwest of Unimak Pass. Since most of the samples of ichthyoplankton were collected with open nets that caught all plankton between the maximum tow depth and the surface, there is little information on the vertical distribution of the various fish larvae and eggs. During a Soviet cruise in 1972 (Serobaba 1974), samples were col- lected with a closing net fished at several different levels. Results of this survey show pollock eggs and larvae present to depths of 1,000 m, but catches below 100 m were small, and below 200 m they were fewer than one per 1,000 m^ . Zoarcidae Eelpouts are one of the larger families with 29 species present as adults in the Bering Sea, yet larvae have been recorded for only three stations (see Fig. 30-6). All eelpouts were caught at stations over the continental shelf during spring, summer, and winter. The lack of eelpouts in plankton collections probably reflects spawning characteristics, behavior of young, and configuration of the net tow used to collect most plankton samples. According to Hart (1973), reproduction among eelpouts may be ovip- arous or ovoviviparous ; Clemens and Wilby (1961) state that eelpouts may be viviparous. Eggs of some eelpouts are as large as 7 mm (Ly codes palearis, in Hart 1973) and hence the newly hatched young might be quite large, say on the order of 20 mm, and active. If large size and active behavior were coupled vdth dwelling on or close to the bottom, larvae of eelpouts would be difficult to capture with standard plankton nets. In order to determine the distribution of larval eelpouts it is likely that a special sampler will have to be used to fish the layer immediately above bottom. Macrouridae RattaUs or grenadiers are represented in the Bering Sea by nine species, and Coryphaenoides pectoralis is thought to be abundant enough to support a com- mercial fishery (Novikov 1970). Larvae of grenadiers were caught at eight stations on five cruises in spring and summer, all close to or over the continental slope between the Aleutian Islands and 56V2°N, and between 167°W and 180° (see Fig. 30-6); only one or two larvae were caught at any station. Macrourid eggs are pelagic and identifiable to family, and were caught in a vertical tow near 55° N, 177°E (Musienko 1963). Melamphaeidae This family of bathypelagic fish is represented in the Bering Sea only by the highsnout melamphid, Melamphaes lugubris; two small adults were caught at one station in quadrangle 52y2°N, 178°E in the western Aleutian Islands (see Fig. 30-6). No melamphid eggs were reported. Scorpaenidae Rockfish are a commercially important group with 12 species of Sebastes and 3 of Sebastolobus in the Bering Sea. At present it is difficult, if not impossible, to identify larval rockfish from the Bering Sea to species although it is possible to separate genera. Larval Sebastolobus were not reported from any of the plankton collections, but Sebastes sp. larvae were caught at 89 stations on 10 cruises made in spring and summer. Individual catches were not large, generally fewer than 20 larvae per station, but a few samples contained 30-40 larvae each. Rockfish larvae were distributed mainly over the continental slope, the adjacent shelf, and deep water, although a few were caught well up on the continen- tal shelf. They were caught from the Aleutian Islands north to about 63°N, although only two stations were north of 60°N, and from 164°W west- ward to 175°E (Fig. 30-16). Rockfish larvae were collected with surface, oblique, and vertical net tows, but most appear to have been caught in oblique tows. Rockfish of the genus Sebastes are ovoviviparous and no eggs were reported. Hexagrammidae The greenlings and Atka mackerel were one of the most commonly occurring groups of larvae in all plankton samples. In the Bering Sea the family contains two geneia—Pleurogrammus, with one species, the Atka mackerel, P. monopterygius, and Hexagrammos or greenlings, with five species. Since identification to species of the greenlings is not always possible, these larvae have been treated here only as a genus. Greenling larvae were collected at 251 stations on 21 of the 26 cruises; only one other genus occurred as frequently and that was the sculpin genus Hemilepidotus. Greenling larvae appeared in catches made during all seasons of the year from late February to November. Catches at individual stations were moderate, usually fewer than 10 larvae, and the maximum catch was 253 larvae at a station in quadrangle 53°N, 178°E in June 1960. Larval Hexagrammos sp. were distributed widely across the Bering Sea from the Aleutian Islands to 62° N and from 161°W in Bristol Bay to 174°E, but appeared to Ichthyoplankton 481 be caught more frequently south of the Pribilof Islands along the edge of the continental shelf and to the west in the vicinity of Bowers Ridge (Fig. 30-17). Larvae of Atka mackerel were caught at 118 stations on 11 cruises in winter, spring, and summer from late February to August. Catches at individual stations were generally smaller than for greenling larvae and the maximum catch was 37 larvae at a station in quadrangle 55° N, 170°W. Atka mackerel larvae were distributed over much the same area as were greenlings, occurring from the Aleutian Islands to 62°N and from 160°W in Bristol Bay to 175°E, but they were present at less than half as many stations as greenlings (Fig. 30-18). Both greenlings and Atka mackerel larvae are caught more frequently with surface than with oblique and vertical net tows. Eggs of Hexagrammidae are demersal and none were reported. m ■ ■ 1 w 'm -1 1' ^ ttt? e -■«■. -1 ^^^B -° - '' \ f-t^~^ *i iy \ \X\4^^^^m^^m ^^^^^^^^^^^B^T^VIi. -- 1 '- rTT L l^^^^^^^^^^^^^^^^^^l j~ ~ ~' ~" , 1 i^^^^^^^^^^^^^^^H\ • l^^^^^^^^^^^^^^^B \ -t 1 ^iSB^^^^^^^^BKT 1 ^■To^v^^^^^^^^^m^^ rfi^ 1 ■^ ^ L_^4-^7^^^^^H^^^Vrl 1 1 -L-v- ^^f^l^Hj^Kfji . ^1 " 4iH-V-trcWr^R----?'IR i K iuUH4Hi3,«rtv^ ~~-l^ Ji~~r~J^~~tr~r~l-~l-l-L i 1 L__l — W— ^ \ \Jfr^- \ \ 2 3 - ||~'r_i--+--r-\ i^^+-^ >■' ^ 'Tj~~J^^ ' 2 ^~, i 1 % vWmMtvV^^ \ 'L. ■i 1 1 ■■t 1 l1 \-\-\:im^tf\,iy^-\^ \ 1 1 a 2 1 1 ^'i^^f'^'V^V' 'v^''T^ ^ 1 ^ ^ ^i^H^TT^ \ -LV-A — \ L- :^^jj^ 6 2 ^JAiu-i — x-A — \Xr\-\ — \\i j^j2l2J|^l_,__l — 'r\"L-X-\ \CX 1 1 "" S\^'^\-+—~\ — iii-V-A — \"u-\- ri~4~JlItiTr^-4^lJx "~ ", ;• • [ZlIILlI \ — \ — ^ZCP'r-^Ty^ r~4~4ij — Ptw-^— Err . ■ \,ji - ■* ZiHi— i-"4 — 1 — \ lZX-A-'m' \ J:rH-44J:ln~r^^^^^^ ^^H--rTirtW--T Figure 30-16. Number of stations at wliich larvae of Se- hastes sp. have been caught. 63' 61' 59 57 55- 53' 51 1 PLEUROGRAMMUS MONOPTERYGIUS 1 \ 1 ^1 1 \ 11 ■ '-- 3 ^^H :_. I ;. 1 k m III , 1 cJ^^^^^^^^^^^^^^^H 4 ' ^ 1 T^^^^^^^^^^^^^^^^^^b] 1 i^^^^^^^^^^^^^^^^^^MX ~ "~/~TC/T~T^-/~ir/rr+4-43II l^= 1 L^^^^^^^^^^^^^v S^ ' • 1 ^^F-^^^^^^^^i^^^ r~/-3l7rH-IirTri4-^ '*. WTJ^^^^^l^^^ >-XT r~/-Ct7 rT-Ry~ 1 u -W-VrlB^'i'^^BF^ TsoSSsttxrH^^^ XlPrTrixA^KJi /QjSsQurn44iK^ 3^ ■ ' ~ II_LU-H-i'Tri«i-^Bl.J^ ^CZTrTSQlTTH— Mw— ^ ^ C t — i_ 44-WVcn44«!u-\^ ^/-4j~Tr7-ijLT~rrT- - t ' 1 « iIUiH-iT^IJ»VXV 1 1 1 z '* ^ 4-4HTr1u:^rrX5pf 2rrT~+-ilrrTTQ3tti 1 1 1 ■r IUUiH«KEPrVC5: 1 ~ ~ 1 1 %'*i^\^^^^^^~Vjc\'\\\V-' 1 1 ,-■ 2i~2^AlAl-3jJ^Pi*M Vjl--\''\\^\-- 5bSsQjT77tW4^ ■ ~ ~ 1 ?±*IX^H — 1 — T I ^i-A-n\ i- fjn-ttjJrHFPt^ ~ " lUEW-v-^-rr j^-M-^ u Q^W--J-_4j7~7^ L ~ ~ - 1 ^!i^l_4— I \ — I'^t^^^A— \''\"TIA ~ ^-.i* • ZJHI— 4 — \ — \ — \X-X--\'\\ \ i ' ■* 1 31— 4--l-n — 1 — \ jHWt — TV 7 — -i~—Ill T — rT»^--f--Ejyp*1 T ■ "" 1 I TI3--5— T~T — T tlv-TT ll __ ^ The family of sablefishes is composed of two species, both of which are present in the Bering Sea as adults. One larva tentatively identified as skilfish, Erilepis zonifer, was caught at 57°41'N, 174°25'W in July 1960. Larvae of sablefish, Anoplopoma fimbria, were caught at 28 stations on 11 cruises in spring and summer. Usually only a few were caught at any one station, but an unusually large catch of 45 larvae was made in quadrangle 55°N, 164°W near Unimak Island in 1968. Larvae of sablefish were distributed gen- erally from the Aleutian Islands to about 59° N and from 162° W to 175°E, but were found most fre- quently along the edge of the continental shelf and in the vicinity of Bowers Ridge in the central Aleutian Islands (Fig. 30-19). More appear to have been I LARVAE OF HEXAGRAMMOS SP. ^KfLIf^ s 1 ^1 ^ -+ P VKH :- !' 1 44-.^^ -f" |t -TTUJ^-^™ ■— ■ ' 1 ttc^ % .JMd^^^ TTT 1 . l^^^^^^^^^^^^^^^H B|^^BpH^>'f-f-/~Ilr 1 lA i^^^^^^^^^^^^^^^^K l^^^^^^^^^^^^^^^^^^m^. 1 1 \ i^^^^^^^^^^^^^^^V \ Yr^jST--/2lj^~ff--P ^1= ' 1 ^^^^^^^^^^^^^^B^Pv ^■i'^^^^^^^^^^^^^Q^^V /-~jC/~r7--/-XT7rT~r^^ t ^ n^U^^^^^^H^^^^PA 1 2 Jll4--+-™^S^0^B|^BfeB ■7^j^'7~+J!rnM--/-5Jlt -^ 1 1 ^ L_J — +-T^ F \ v.Jv-^^^^\ ii9l AttW^irrffl^t >5 1 1 I jJ—l— ^^ V'\-^B^ ^^r /JtxTsQlj^^ ' '^ 2 1 1 U44-ht^TXA-4aKll'\^ -~l~Jr~h4~JI~fr~h4~4~4^ 2 T [vlil Ttrxp4^v+i2iPVVx^ J]j~-/---jL/_T^^ 1 2 1 2 p 1 2 1 1 L\ — \jY^ \ ^P^*-^ \' *"- 7ljrT'--CrT/~T-4--J-ItlLZ ' 1 2 t 2 9. 3 aj '^lM-^-^^2ibP^--V'^' \ 'V Q~ lT-4XtfTr4-4-4J^ 2 2 ■ is 9' 5i?i-^ A3^HrjjlL-+-1' L-V 1 rT"i^6""7 7 ^j^^C*^f~V^- ^-V^ \ i 1 1 2 2' 7 iu^ti^|^:*rr' \\J\'-\ — \ \^- ^7t~?JII]H~^-W3I^ 1 ii 1 I^ lAil-i^^^t— r""i — \AlX^^r\CX-- ti'^-^^^JJiirM^UL ^ 1 ~ p-tziziKiii— +— 1 — 1 — \CW-T\ 7 — f^'f—Uj- — tr^—i-liSIL 1 JTnlJIIl \ — 1 — 1 Tju-i — T\j. uSH^tlzW-^ifflt 1 "TT 1 vt' -HH-TTi3J--+-^ru - • -T- ^^X+\\xX\X\Z IjwW-H^-i/rT — (M^?*^ 1 iJ44-HrG-Vr iLJ — 1^ 1 — 1-, 1- 1 ■! f~t h i^r 1 1 1 1 1 1 1 1 — H-i 1 1 u-t 1 Figure 30-17. Number of stations at which larvae of Hexagrammos sp. have been caught. 170 175 175" 170- 1 LABVAE OF ^^^d p 1 1 tn = 63^ ANOPLOPOMA FWIRIA Mfi^ j !| t if -|^^^^;^^ ,__. , „..- 1 < # 1 Li— Urad^^^^^^^^^^^^l 1 1 ^*'\ ^1 N i ^^^^^^^^^^^^^^^^> 59 57 55 53- 51 ~^ ~i ;'£^^^^^^^^^^^^^^^^^^b\ ~r'^ml::tr'~i~-~l-~J^ T~fT~v4\-- 1 i^^^^^^^^^^^^^^^^^V \ 1 1 ~ hiAI^^^^^^^^BK^ ctfiS^ (^ % ^t ^iJ^^^^^^^^^^^^d^^^ Tti~-f~Tjj~]-4^ i^ s \i ri>4-w^^^^^^H^^^^F^ T ^ 1 1 1 L-i-^r-W^9^^^^^tlkA 5 ■4i3j'~~j~^~J'\~~ij — /--i— XXm^ ^\ " \ TLJ — \ — Y'?^ i^-^v^^3p^'^"l /-st~7r~/H--XrrT~f-4-4— LcS^ r : 1 " 1 LJ— WWr-is . Ui^K eH W^SS^^irTT-- ^^ 5~ ] I -U^HitUAMfA-V^ 1 ( \ _ZJX\-+WTjm^^jT 7~ t "TI_UW-+-tT^Pferiil^ 1 ■ ~ ^ r^ c , ■■ A—V-V^SKPfR^^'''"^'-^ "^ 5 rrt~r~~i-~^iin~'T^^ '- "t^ V 1 Ti4-4=^n**»iA^=''V^ \ V 1 " ^ iikm^wsx^ . JTt r~~r—~L^ / JiPVnr-r — 1— 1 ■' ^ ' "^ ^J^'H-nTXWrXi^ ^ T^4'dJXrr~nTM^TiiZ ' ■ ^ r: 3^^4^4-V-n3Vri; n'^-ijT^rJBlJtltl " v/" fnz^-^-H-tlXVrVtPt 1 ,~ "■ -, !*#r IIin-4-rTiJ^Vri ~/477T-4-447?4~vrT~r~[— r-r ,- > •" < =+-' " -1 — HlCXVH:^ -1 rr zt ^--Hrri^-^-rT^ - — — — h- — r^ 1 L_V— (— m L^V- Figure 30-18. Number of stations at which larvae of Pleurogrammus monopterygius have been caught. Figure 30-19. Number of stations at which larvae oi Ano- plopoma fimbria have been caught. 482 Fisheries oceanography caught with surface than with obhque or vertical net tows. Eggs of sablefishes are reportedly pelagic, but none were found. Cottidae With some 75 species in 36 genera, sculpins are by far the largest family of fish in the Bering Sea and were present in aU 26 collections. Larvae of this group can usually be identified as cottids but rela- tively few are identifiable to species. In these collec- tions only seven genera with six species were listed, the remainder coming under the heading of unidenti- fied cottids (see Table 30-2). Of these, three species and one genus each were captured on only one cruise. The most commonly occurring and widespread sculpin larvae were the Irish lords, Hemilepidotus sp., which were caught at 208 stations on 21 cruises during all seasons of the year from late February to November. Irish lords include five species in the Bering Sea, one of which is found only on the west- em side. Identification to species is possible for larvae with countable finrays but the smaller larvae are usually identified only to genus; thus distribution is shown only for the genus. Irish lords have one of the widest distributions of all larvae in the Bering Sea, extending from the Aleutian Islands to 63° N and from 160°W in Bristol Bay to 174°E (Fig. 30-20). Although there are insufficient quantitative data for the entire region, collections made aboard the Oshoro Maru during summer were all made by approximately the same type of surface tow, and these show no areas of consistently high abundance. Actual catches of Irish lord larvae for a 10-minute surface tow range from T to 3.709. with onlv 5 of 114 collections containing more than 100 larvae. The catch of 3,709 larvae was made in quadrangle 58V2°N, 165°W. Only the northeastern area of the Bering Sea, roughly the Nunivak Island /St. Matthew Island /St. Lawrence Island area, appears to be devoid of Irish lord larvae. This genus of sculpins appears to share the surface layer with the hexagrammid larvae, for more are caught with neuston and other surface-towed nets than with nets sampling deeper waters. Irish lords differ from the hexagrammids in that there is a greater difference in day and night catches of the sculpins than of hexagrammids, with larger catches made at night. A second genus of sculpins common in these collections was Myoxocephalus, great sculpins, with 10 species present in the Bering Sea. Since the taxonomy of the adults is unclear, no attempt was made in any of the collections to identify larvae to species. Larvae of great sculpins did not occur as frequently as those of Irish lords but were caught at 53 stations on nine cruises in the summer. They have one of the widest north-south distributions, extending from the Aleutian Islands to 65y2°N, but appear to be limited mainly to the continental shelf, with east-west distribution extending from 161°W to 174°E (Fig. 30-21). Although the catches were small, one or two larvae per station, the great sculpins are present in the northeastern area in which Irish lords were absent. The area of greatest abundance is in Bristol Bay, where catches of up to 408 larvae per station were reported in quadrangle 57°N, 161°W, and catches of more than 50 larvae were reported by several stations. Most of the great sculpin larvae were caught with surface tows. 63 61 59 57 56 S3 1 LAKVAE OF ^^^^^Hl^ f 1 ^ 1 1 1? ^■M| ^-. r .- 1 ■ HEMlLtPllXJlUb SP. BBfj^^i' 1 1 ("(T^ ft k. f( \ rx4— v-V'^^H ■' I i ^ ka 1 I LiJi^M^^^^HI^^I^^^I 1 1 f^ T^i^^^^^^^^^^^^^^^H r L '' l^^^^^^^^^^^^^^^^^^l 4 1 ^ . l^^^^^^^^^^^^^^^^K r^2rr-?JIrrT-R— U- i^^^^^^^^^^^^^^^By 1 1 ^^^^^^^^^^^^^^^^^^p ^ -fc 1 ^^^^^^^^^^^^^^V?v /^/-~ZrrriQirrT~M>Si Jm ij^V^^^^^^^^U^^ •7^ -» ^r [ ^iv^^^^^^^^^^^^^f-k - 1 1 Li— I— Sf^R^^^^Bv'jS ~rj~?r-/JjT]M^ J- .vi 1 i| i_i-HpT'A yAl^H^aai Q^j^r-t-Xj/TTH-JJI h\^'\ r 1 1 _L4— 4vv--Y .\ -^^P) ^^ "i vum . J' ^ 1 1 "jTTni^--yM-nPy|^^TT^ -V-Xj-f^Wlrrrf^ i\ 1 l| \ 1 H^Tv'l'* l^jff^V' \' 1 1 a 1 1 J ^ 1 1 1 1 ui A^-X- \ -JHRy^t y, Jp^ > 5 1 " "* ' ^lik^^\-''\'y \ r" r--jL 7^7 — r — i-JT rTT i — r — I — 2 3 1 '1' ' 4 J 4 . 3 , 2 . ^ . ^ ,J^rVi^\^\\ 1 1 3 1 fe s 1 .'■'^^P't' \ \\-\\\^ 2 J 3 1 1 ' 1 y ' ^|p4«-r \_LA-— \ — \ \^ 23-TyjjliVKP4^ 3 J 3 sj^^' -i— 4 — \\\V\~\'\'\i ' ?^HW4i?l»H^ff 1 ... ," ' 2aJK_I— 4 — \\\\-\\\Z\ rp4-4J~j~i-HM~niiT~j • Jjj? jL_4 \ — \\\lX-\\\ZX TffSffS^^^^ ^ 1 .* '''^—~~\\Al^ZX-^r^\I^P\-^ ifWrrFffl Ht- ^= — - 5i ^^H-rSivS-" 165 1W 175' 180' 175' ITO-^ ISS- 160 155 ISO- .^MM\ ^ llta' ■1MM|M 63" 1 mrOXOCEPHALUS SP. Ktf-H-^^ !>«. IDI^ - ^«,vi£ . ■-. I HIrtflflHt v^ I'' 1 \mi^m ■ ^B^EtllTT ^ M 1 ssBsni ^■%mmmr^ fit ^^■LdHI 2 fi ^ S7 ifri^H-^ 1 1 i rTr^4~f~i~- ^ ■ IIW a r-fy-UJlIZ . ■°ESk M jihRJf -- ► -.A ■■■!S^' ^A 1 ^ (1 r ^v .^ ' "9 7/T^^tw— k r^^^^^d^m tflr Tr~r-4^i;lL' T^JJ-i-H--*! l^lit ^,M T — r~-r-4-^3E ^^ ^" ;rkj^^^ 57 rr—t— Ct iiS 1 SiSRS ikV/ i T~R— f- ^^ -Ht-PttSw ^^VBrv - — f~~i— i— L ^■, ill L _1 — 4^^T \ 1 J^S^^. V \-:- rr— f- 1— -v^ ^Tiiil^V-^vf 1 J^SV-T V ¥-'- 55 r~i~~J-~i~~. ^■-- tHM^ Mr^r^Ai^ \ ^ -j—lJUlI "" Y-i ^^^^rW\ \ Tt— f—*— ' SEl-Ua^^P» •m^tZA-JVt^JlJ Tr^-4-— qixp4#pv^ r\Ij^-+--T\j^ r\r-+— 1— iW^^f*-^ 4-^'*'*^^ r~T— f— L ~~rn— 1— ^ VM TTZC-V-TTil-A i-Ii ll_ ii4__|li — \ — t- \^\\-\-\yJ\ J___;J_[^ T- ji '" •* ijZ-vA — \ t- ~f-F¥¥^^^ - . ^itn-u-ft^ zszX^-ykX- Figure 30-20. Number of stations at which larvae of Hemilepidotus sp. have been caught. Figure 30-21. Number of stations at which larvae of Myoxocephalus sp. have been caught. Ichthyoplankton 483 In addition to Hemilepidotus sp. and Myoxoceph- alus sp., six other taxa of sculpins, plus a category called unidentified cottids, occurred in the samples. Distribution of these miscellaneous species is shown in Fig. 30-22. Artediellus pacificus, the hookhorned sculpin, was collected at one station east of the Pribilof Islands on the continental shelf during late spring. Icelinus borealis, the northern sculpin, was caught at two stations between Unimak Island and the Pribilof Islands on one cruise in spring. Gymnocanthus sp. was caught at one station northeast of St. Lawrence Island during summer. Malacocottus sp., probably M. kincaidi, the black- fin sculpin, was caught at four stations between Unimak Pass and the Pribilof Islands during a spring cruise. Triglops pingeli, the ribbed sculpin, was caught at one station southeast of Nunivak Island during a fall cruise, and Triglops sp. was caught at three additional stations, one northeast of St. Lawrence Island on a summer cruise, one between Unimak Pass and the Pribilof Islands, and one near the Alaska Peninsula during a spring cruise. In addition to these identifiable taxa, there were numerous cottids which could only be identified to family. These were distributed mostly over the continental shelf and slope south of 60° N, but a few were collected as far north as Bering Strait and west along the Aleutian Islands to 17 7° E. Unidentified cottids were collected at 83 stations on 11 cruises in spring and summer. Eggs of cottids are demersal and none were found in any of the plankton collections. Agonidae The poachers are represented in the Bering Sea by 16 species and larvae of seven of these were present in the plankton samples. Larval poachers were caught at 45 stations on 14 cruises in spring and summer (Fig. 30-23). Agonus acipenserinus, the sturgeon poacher, was caught at 18 stations on seven cruises, all of which were in summer. They were distributed mainly in Bristol Bay and the adjacent continental shelf except for one larva caught at a station north of Nunivak Island. Odontopyxis trispinosa, the pygmy poacher, was caught at four stations on one cruise in summer, all north of St. Lawrence Island and in Bering Strait. Asterotheca infraspinata, the spinycheek starsnout, was caught at two stations just north of Unimak Island in summer. Aspidophoroides sp. larvae were caught on two cruises, A. bartoni on one cruise, and A. olriki on another, all in summer. Larvae of this genus were found north of St. Lawrence Island (A. olriki) and along the continental shelf edge south of the Pribilof Islands. Hypsagonus quadricornis, the fourhorn poacher, was caught on two cruises in spring and summer, with one specimen from the Gulf of Anadyr and the other specimen north of Unimak Island. Occella dodecaedron, the Bering poacher, was caught at one station in summer on the edge of the continental shelf between Unimak Pass and the Pribilof Islands. Larval agonids identified only to family were caught at 22 stations on six cruises in spring and 170° 165° Figure 30-22. Number of stations at which larvae of various cottids have been caught. Each unnumbered symbol represents one station. 165" 170" 175" 180" ITS" 17C 165° 160^ 155 150' ^^^^^B O ARonus acipenserinus 1 !| 1 Ijl ^^HHP • Asterotheca Infraspinata ■ ^f^^Y^r^ A Aspidophoroides sp. ■ 11 \-\- A A. bartoni ■ # !Li-U4-V:2*.-?^^,„,... , 1 ^^^^^^^^^^^^^^^^f^^'^^t^J.^i-^'i >; L'l^^H ▼ Odontopyxis trispinosa 1 ^^^^^^^Kl^l l""T-f— T-L 1 1 k ^ fy, [^^^1 O IJnldentifled agonids ^J ^^^HW-J / ll—h-iJZI 1 1 \ i^^^^^^^^^^^^^^MX 7) ^^^^^^^^^^^^^^m \ ~h~r~ISir~~h-]~SrTr-^^ ^-^ !S 1 ^^^^V^^^^^^^^^Mi^^n '% { liLi-T^'^^^^^^B^^^^^ -++-\-TWS^mmlKrm y-~tjr~l~~ljriri~4^ ■- ^ ru34-^MnxA^KJl fttri^Jj^^ \£ V' ' ^^^-kHrfiS-V^R-^ l-~cri-4~Ilirj^^ r^ V- TI iCS-^rH-iSMrtv^ i\ < ^ 1 L— 1—^^*5^ \^Pr^' \' ^ . 'i' y\ \ A I —MA — iTi iKIr^^' '* ^ Ag5u, S'^^cQpK^vB rjr~~7~-~jL/_]7~j;:rf — f— /— L_I_ TT i^ ^ &^,j3im^g^^r\XA,r\-TiA *r L^-_ ^iX!3SWFM-T l-X-\-r U tT:"EliIJS^4-+-rrL4--+^U y:i-Jl7~j7 — j — PIlJIJZ! ■ ■ .fs^ i, SPUll— 4 — 4 — t" \ \ ^U-A"^ \ tT"*T~^f3=J — Tf — r-f--I— ^i— ^ — i-J — \ — 1 — \l\^^X-\'\\^ / / ~/^-^J~J^ T^'^~t~^^f^^ — J— i_ 1 1 j_ , — ^^''^llI^I^A^^tA-\iX^X^ '~~h~f^^ — '- Zt ziJ^^--PrTXi4-n Figure 30-23. Number of stations at which larvae of various agonids have been caught. Each unnumbered symbol represents one station. 484 Fisheries oceanography summer. They were distributed from the Aleutian Islands north to 58° N, and from 161° W in Bristol Bay to 174°W along the Aleutian Islands and along the continental slope. Eggs of agonids are demersal and none were reported. Cyclopteridae The snailfish and lumpsuckers, with 49 species, are the second largest family (after Cottidae) in the Bering Sea and one of the least known with respect to both adult and larval identification. Cyclopterids were present in samples from 20 cruises, but only four taxa were identified, in addition to larvae identified to family level. Aptocyclus ventricosus, the smooth lumpsucker, was caught at five stations on four cruises with a total catch of six specimens. All were caught with surface nets during spring and summer, four of them near the Aleutian Islands between 166°W and 175°W, and two specimens at a station in quadrangle 55V2°N, 171°W (Fig. 30-24). Nectoliparis pelagicus, the tadpole snailfish, was caught at 10 stations on four cruises during spring and summer, with a total catch of 20 specimens. Most were caught in nets fished below the immediate surface, but a few were taken with surface nets. The tadpole snailfish was found in a relatively small area between the Aleutian Islands and 56° N, and between 168°Wandl75°W. Larvae of snailfish of the genus Liparis, of which there are 13 species in the Bering Sea, were caught at 41 stations on 12 summer cruises. Most samples contained relatively few larvae but at one station in quadrangle 57V2°N, 162° W, 671 larval Liparis sp. were caught (see Cruise 20, Table 30-1). Larvae of this genus were distributed widely from the Aleutian Islands to Bering Strait and between 162°W and 175°E. They were caught most frequently over the continental shelf but were also found over the conti- nental slope and adjacent deep water. Most of them were caught with surface nets. Unidentified snailfish were caught at 46 stations on 10 cruises during spring and summer, between the Aleutian Islands and 60°N, and from 162°W to 174°E. Within these limits they were found over the same areas as were Liparis sp. Cyclopterid eggs are demersal and none were reported. Ba thy masteridae The searchers and ronquils are represented in the Bering Sea by three species in two genera, with one genus and one species present in the samples. Bathymasterids were caught at 76 stations on 18 cruises in spring, summer, and fall— only three speci- mens were caught in fall. The family is widespread in the Bering Sea; it occurred at stations from the Aleutian Islands to 63° N and from 163° W to 174° E (Fig. 30-25). Specimens were foimd over the conti- nental shelf, slope, and deep water of the central basin; more were collected with surface nets than with nets that fished deeper. Catches at individual stations were generally small— fewer than five speci- mens—but at two stations (in quadrangles 53° N, 174°E and 55i/2°N, 167°W) 262 and 245 were caught (Cruise 18, Table 30-1). Only six specimens of Bathymaster signatus, the searcher, were caught, three each at two stations in quadrangles 54y2°N, 169°Wand55°N, 167°W. 165' 170 175' 180- 175- ITO" 155' 160" 155' 150" 63- 59 55 53 51" 63' 61 59' 57 55 53 51 1 LARVAE OF BATHYMASTERIDAE ^^^~L1.J:^^^M I ■ •-- 1^^! 1 H f( ^Im/tfwSs ttt —- ^ *1 m\ 1 1 ^ \^^^ :n^^M 1 ^M 'fTpi'H4itl ¥hPU^ '^ ' ''^/l7irr^~/j^/[~r~f-4— ^ y f^^ ? A -TCTTfTLnW-T'^-f^^ ^_L-U-4--V-« TWClfftQST4J-H-±^ X -Hiutt /jrff^Q^y-f^^ >^ " '^-ZlJ~TW-X2~r7H~f--M3^ ' i!.^ xEuPrV (\ t^^^Xt^^ 1 ^' LLUWH-t ''""" ^-. 2 1 JiM-i /J// 7^ — i-~-J- 1 tC 1 sLSSi?' f li~. lS^PV sW^Tjijr^^^W— tir r^ - _.. ^ I3^Ili^^^++- J^i^JJJ^^ ^aSTv 1 iS. 1 , ~ ' „ ^m^ -* Jl -4-H-tt Zrr~r^^ t= =PM^-H 175 ISO ■ 175 170 165 160 Figure 30-24. Number of stations at wiiich larvae of Cyclopteridae have been caught. Each unnumbered symbol represents one station. Figure 30-25. Number of stations at which larvae of Bathymasteridae have been caught. Ichthyoplankton 485 Eggs of bathymasterids are demersal and none were reported. Anarhichadidae There are one or possibly two species of wolffishes in the Bering Sea; they were represented in the collections by five larval Anarhichas orientalis, the Bering wolf fish, caught at five stations on one cruise in spring (see Cruise 37, Table 30-1). All were caught with neuston nets at stations near the outer edge of the continental shelf between Unimak Pass and the PribUof Islands (see Fig. 30-6). Eggs of wolffish are demersal and none were reported. Stichaeidae The pricklebacks, a fairly large family with 21 species present in the Bering Sea, were represented in these collections by six species in four genera and stichaeid larvae identified only to family. Prickle- backs, present in 23 of the collections, had one of the widest distributions of all the fish larvae, from the Aleutian Islands to 65°N, and from 158°W in Bristol Bay to 174°E (Fig. 30-26). Most stichaeid larvae were caught with surface tows, although some were taken with oblique and vertical tows. One species, Lumpenus maculatus, was caught more frequently in oblique tows. Alectridium aurantiacum, the lesser prickleback, was collected at only a single station (see Cruise 1, Table 30-1) in quadrangle 54i/2°N, 165°W. Chirolophis polyactocephalus, the decorated warbonnet, was caught at 27 stations on two cruises in spring in the general area between the Aleutian and Pribilof Islands from deep water to the outer conti- Figure 30-26. Number of stations at which larvae of Sti- chaeidae have been caught. Each unnumbered symbol represents one station. nental shelf. Catches were moderate, usually fewer than 20 larvae per station, with a maximum catch of 78 at a station in quadrangle 53 1/2° N, 168° W. The genus Lumpenus included those pricklebacks identified only to genus, plus three species: L. fahricii, the slender eelblenny; L. maculatus, the daubed shanny; and L. medius, the stout eelblenny. Lumpenus sp. was caught at 43 stations on 10 cruises in spring and summer over an area almost as wide as for the family, from the Aleutian Islands to 63°N and from 161°W to 174° E at stations far up on the continental shelf, over the continental slope, and over the deep central basin. Catches at individual stations were generally small and the maximum catch was only 22 larvae. Larvae of Lumpenus fahricii were caught at 12 stations on three cruises in summer over a wide area from the vicinity of Unimak Island to 65°N and from 160°W in Bristol Bay to 178°W in the Gulf of Ana- dyr. All catches of this species were made over the continental shelf and individual catches were only one or two larvae per station. L. maculatus larvae were caught at 10 stations on three cruises in spring and summer. More restricted in distribution than L. fahricii, they were found from the Alaska Peninsula to only 58V2°N and from 161°W to 173° W. All larvae of this species were caught over the continental shelf and, in contrast to most of the other stichaeids, more larvae of L. maculatus were caught in oblique and vertical tows than in surface tows. Individual catches usually consisted of only one or two larvae, with a maximum catch of 19 at a station in quadrangle 58V2°N, 170°W. A single larva of L. medius was caught in summer at a station in quadrangle 58° N, 166° W. Larvae of Stichaeus punctatus, the arctic shanny, were caught at 18 stations on four cruises in spring and summer. They had a fairly wide distribution over the continental shelf, continental slope, and deep water from 54° N to 62y2°N and between 164°W and 177°E. Individual catches were small and variable, ranging from a few to moderate numbers with a maximum of 131 larvae at a station in quadrangle 62i/2°N, 172°W. Stichaeid larvae identified only to family were caught at 69 stations on 11 cruises in spring and summer. They were found over a range extending from the Aleutian Islands to 60° N and from eastern Bristol Bay at 158°W to 174° E, and over the conti- nental shelf, continental slope, and deep water of the central basin. Individual catches ranged from one or two to a maximum of 556 per station, with the maximum caught at a station in quadrangle 54y2°N, 170°W. 486 Fisheries oceanography Stichaeids are demersal spawners and no eggs were reported. Cryptacanthidae Of the two species of wrymouths in the Bering Sea, one, Lyconectes aleutensis, the dwarf wrymouth, was represented in these collections by a larva caught during an early spring cruise at a station in quadrangle 55°N, 165°W (see Fig. 30-6). The wrymouths are demersal spawners and no eggs were reported. Ptilichthyidae This family contains a single species, Ptilichthys goodei, the quillfish; although it is not a common species, quillfish larvae were caught at 11 stations on five cruises in spring and summer. Most of the larvae were captured in Bristol Bay and out to the shelf edge, but a single specimen was caught in quadrangle 53°N, 175°E (see Fig. 30-6). All of the quillfish larvae were caught in surface tows. Nothing is known concerning the eggs or spawning habits of quillfish. Pholidae The gunnels, with five species in the Bering Sea, are one of the less well known families in the area and all larvae of this taxon have been identified only to family. Gunnel larvae were caught at 40 stations on six cruises in spring and summer; most were caught during summer. They have a wide east-west distribu- tion extending from 161° W in Bristol Bay to 174° E, but are limited in their north-south distribution from the Aleutian Islands to 58°N (Fig. 30-27). They were taken at a few stations over the continental shelf but the majority were caught over the deep central basin. Almost all larvae were caught with nets towed at the surface; catches were often of moderate size, exceeding 50 larvae at several stations in the vicinity of Bowers Ridge; a maximum catch of 131 gunnel larvae occurred at a station in quadrangle 53V2°N, 179°E. Eggs of this family are demersal and none were reported. Zaproridae There is only one species of prowfish, Zaprora silenus, in the Bering Sea. Larvae were caught at five stations on three cruises in spring and summer with a total catch of eight larvae. Most were caught near the Aleutian Islands over the continental slope, but one was caught near the Pribilof Islands (see Fig. 30-6). It is not known if the eggs of this species are pelagic or demersal. Ammodytidae The family of sand lances is represented in the Bering Sea by one species, Ammodytes hexapterus, the Pacific sand lance. Larvae were caught at 170 stations on 19 cruises in spring and summer; a few were caught in early fall. They were distributed widely from the Aleutian Islands to 65y2°N and from 159°W in eastern Bristol Bay to 174°E, with a majority of collections being made over the continen- tal shelf (Fig. 30-28). Sand lances were one of the most numerous larvae in the collections, with many samples containing more than 100 larvae and a maximum catch of 1,337 larvae at a station in quad- Figure 30-27. Number of stations at which larvae of Pho- lidae have been caught. 175° ITO" 160° 155° 63 59 57 55 53 51 1 LARVAE OF H^BPP+-XXI ■ AMMODYTES HEXAPTERUS iK^^iLT/i*' i 1 1 1 :±JH|H :-/-:: 1 6 ^gi^^^_^^|liTr 1 --w 1 1 1 _i — l--T"\^^H 1 ^ k i |Limiiki2l^HiiiHii ^^^^^IRIvri ^ f 1 T L l^^^^^^^^^^^^^^^^^H ^■■^TwT/1 1 V T l^^^^^^^^^^^^^^^^B 1 ^^^^^^^^^^^^^^^^^^K\ -r^J^7"~/--jL// T~fy~l-d-J_ \ ^^^^^^^^^^^^^^^V ^ -t. 1 '^^^^^^^^^^^^^^^^P^ S . ial^S^V^^^^^^^^Q^w\ ^-A^/jS^4iirTTT4-T^^ > ~ 1 1 l_^4^:^^^^^^I^^^^Kni T-TQ^W--/-IlrT7W-f^43l ^ '1 !I_wWWbStISk^^^Kv:4 r~sCj/T^~/Zi/ — ' — r~r~4— 3E- ^ 1 1 i_LL-Wri f '" \ tl^^^SP^ jiai ^L/TTt— f-jJJ frh-f — i-J-_ ^ ) N "■ ^ 1 '■irilI_|iVi^TTK,,y-^^m cW 5 '^^J^T^-^Jjrrn'-- T h, ~ 1 1 1 _4_4-^— W^TiLA-tJB^A-A'I 1< ' 2. 2 1 1 -^-n"T^i 1 1 \ ^^^M • W ^trrTW~ST-H44f -J T ' 1 2 t 'j-?l^+^ \ '■BP^' ' ^ Iln^-QrnH-44lII _■■- :+- V ~ 1 \^-A^r^\Jlff^--\'^^\^ & /~jrr~M2rTfTHH--f— C " 1 "■ T 'i 2 2 *'-ft\-^Tg^^*iMi,+'*V\ A- -V _j~~/-~jLjj~7:nr^ —i 1 ' ■ "rrSwT^^I \T^\ \^ 1 1 L- 2^TVi93|iPW\ \__U-VA^ A-^ i/">T~-A--L/^ — r^R— C-L ■-^rSEEA— 4 — \\^\\-\\\^ fi. jCljr-/--iJj TTrfhrhrl — i 1 ^ i^ 8KI-pj — \ — x^ l1-V^l ■■ 1 J *^ — 4— -1 — \ — 1 — Tj-^V-A — TL-X '~r~~'-~4J^'i^^-h^-lMIl — ~1 i f'- • -: ZII— --4 — iT _LlX-4 — T — \JA » ■* e^ —4-^^ — \ — TiL-I-A — 1^ \ , ~/^7~7 i — f— L PT~T~frJzrJ— r-- ~ ■ jUL-^-f— 1 — 1 — vCL-V-tT •~ — _=cx_L_j__i__i L-t— r 1 1 ^jl: Figure 20-28. Number of stations at which larvae of Ammodytes hexapterus have been caught. Ichthyoplankton 487 I \ rangle 56V2°N, 164° W. During the summer, larvae appear to be concentrated in an area just north of the tip of the Alaska Peninsula and Unimak Island, where several catches amounted to more than 400 sand lance larvae per tow. Most of the larvae were caught in surface tows, but during the spring oblique tows seemed to catch more larvae although the catches were still not large. Sand lance are demersal spawners and no eggs were reported. Pleuronectidae The right-eyed flounders, one of the commercially important groups, are represented by 17 species in 14 genera in the Bering Sea. Larvae of this group were collected in 20 of the 26 cruises during spring, summer, and fall, most abundantly in spring. In general, fewer flounder larvae were caught in surface tows than in oblique or vertical tows. The genus Atheresthes contains two species: A. stomias, the arrowtooth flounder, generally considered an eastern Bering Sea species, and A. evermanni, the Kamchatka flounder, considered a western Bering Sea species but possibly extending to the Pribilof Islands. Since larvae of these two species have not been described adequately to permit specific identification, the records of capture are combined and presented here as distribution of the genus. Larval Atheresthes sp. were collected at 74 stations on eight cruises in spring and summer. Most of these were caught over the outer continental shelf and the continental slope but a few were taken over deep water west of the slope and from shallower water in Bristol Bay (Fig. 30-29). The distribution extended from the Aleutian Islands to 58y2°N and from 162°W to 177°W. Individual catches of this species were small and usually consisted of only a few larvae per tow. Eggs of Atheresthes sp. are pelagic but none were reported. The genus Hippoglossoides includes H. elassodon, the flathead sole, and H. robust us, the Bering floun- der, two very similar species which occur in the Bering Sea with overlapping ranges. Differentiation of the adults to species is difficult and larval descrip- tions are not adequate to permit identification to species; data are presented here for the genus. Larval Hippoglossoides sp. were caught at 22 stations on seven cruises in summer and on one cruise in fall. As with all flounders, most larvae of this genus were caught in oblique or vertical tows, and catch per tow was small. Hippoglossoides sp. larvae were distrib- uted from the vicinity of Unimak Pass north to 60° N and from 162°W to 173°W (Fig. 30-30). Most were caught over the continental shelf, only a few over the deeper water of the continental slope. As with most other flounders, the eggs of Hippo- glossoides sp. are pelagic and are readily identifiable to genus and possibly to species on the basis of diameter. Hippoglossoides sp. eggs were generally distributed along the outer continental shelf between about Unimak Island northwest to west of St. Mat- thew Island; they were found in samples collected between April and August (see cruises 6, 7, 34, and 37, Table 30-1). Larvae of Pacific halibut, Hippoglossus stenolepis, were caught at 22 stations on six cruises in spring and early summer. They were usually found at stations over the continental slope or deeper water, but a few were caught on the edge of the continental shelf; 170" 165° 160^ 155° 63" 61- 59 57 53 51 1 LARVAE OF ^^^HFMI ' 1 ■ HIPPOGLOSSOIDES SP. WB^ILT'T^^ ■ 1 5; ^^ JHH . zl p^ ^ _ JQ- ' ■' XU^^^^^^^J ■; ^ s -U-^ T ^A^^^^^^^^H i ^^^^^Hi5^ f-i^T^W-ii 1 K ,^^^^^^^^^^^^^^^^K \2 i^^^^^^^^^^^^^^^Vt \^^^^^^^^^^^^^^^^^^^ ^ /f-~~f^J/~tl~--~h-^L^^ — f-^' N _i i '1 1 \^^^^^^^^^^^^B^^i \ fe \ i f-^W^^^^m^" -% 1 k T^-Tx^^^^^^^mrx u _U-4-+-W,f53P^^BP^ 1 ^.i jii—^— ^— ^"\ J.-^;^^gf\J^m i~--icJ~l'T~4^JSli — P~r~4--I~J-JLrf\ \ ^\-\ — 4— rvT — VA'^^^ *i"P i T 1 1 — bv4--r^ \ A Ij^Br' \ A ** \ 1 ^ jlll--i— r~\ \J^v^ \'' '-— 45,. V ■ ^■-C T |— tri •^'' JHp^^^^'V''-'^'- ^ -1-1: -^ 3 1 H l__i I — ai^p^v^AM^ \ '^- lZj2r~f-~-f^^ ■^^ ^V 1 -^^^^\~^^^^^Srr^^W ~Jl~ir'~f~~C^^ ii. \ ^ \',-^S^\\ vv'%^ \\ r ~ ' ^ • ~.'^^^^3^^-- :^^S^^'T~^L\\'\j^ Hx; >. ^•\yt.'^^^^'\Zr \--^'-'\v\ [T - 'J^i it ^ ^ -— ^~~i — 1 — \ L— i— W^^ — \\ ^-'^ *'T ■^ i_i-4 — rj^TIIl--+-M^ ^35 _i- ^ I^XH t — \ — vxIL-\-A^\ -J^Iii~4-~4~^S^iZj^:^^ ' 4 ^T^tTIIIK-tTLC Figure 30-29. Number of stations at which larvae of Atheresthes sp. have been caught. Figure 30-30. Number of stations at which larvae of Hippoglossoides sp. have been caught. 488 Fisheries oceanography these were distributed in a narrow band extending from the vicinity of Unimak Pass to quadrangle 58° N, 174°W, northwest of the Pribilof Islands (Fig. 30-31). Individual catches usually consisted of only one larva but five were caught in a one-hour tow in quad- rangle 54°N, 165°W (Dunlop et al. 1964). Eggs of Pacific halibut are pelagic but are generally found at depths greater than 200 m; no halibut eggs were reported. Larvae of Lepidopsetta bilineata, the rock sole, were caught at 41 stations on four cruises in spring and summer, but the catches on three of these cruises consisted of a single larva. The remaining larvae were caught in the area between Unimak Pass and the Pribilof Islands (see cruise 37, Table 30-1). Except for one rock sole larva reported from Norton Sound, this species was restricted to the continental shelf from the Aleutian Islands to 56V2°N and between 163°W and 172°W (Fig. 30-32). Catches at individual stations were small; the largest catch was 12 larvae. All were caught in vertical or oblique tows. The eggs of rock sole, unlike those of most other flounders, are non-planktonic and none were re- ported. Larvae of Limanda aspera, the yellowfin sole, were caught at 29 stations on four summer cruises and one fall cruise. Catches were generally small, usually fewer than 20 larvae per station. Different from those of most other flounders, larvae of yellowfin sole were found only on the inshore shallow portion of the continental shelf, where they were distributed from 57y2°N to 64°N and between 162°W and 170°W (Fig. 30-33). Eggs of yellowfin sole are pelagic and large quanti- ties were collected during several cruises (see cruises 165 170 175 160 175 170 165 160 155 150 | 1 LARVAE OF PLEURONECTIDAh ^^^il l^^l p '^ '^ K^^ /\ Limanda proboscidea 1 63 61- 59 I !=■«. \ j-I^^PO Platichthys stellat.is | ^^^^^^^^^^^^^^^H-J.. fl i \\\A D Pleuronectes quadricuberculatiis 1 s ^' 4 --Vei^ 1 i N S9 57 55 53 5t ^P^n^H-fl^ni --^' • T 1 '^^^^^^^^^^^^V \ ' 1 J^^^^^^^^^^B^P'i TSsCrrf^Q^^ \ « I %ljSstjTi~H^# ± " ■*, ^ 1^ 1 ^U-f^l^^^^^^^^^^^V^ ~/JTtL^^ -4 ~ \'f\ ^ -1 l_JfcM — lS"r^ , L--l^^^^ Wi 57 rTT-HCrrT~/-4-IEj7ri — rrrr f+nnn_ut4M^d^^rTA»r\>M 55 S3 51- TptiBj-niSz ■• ■ ' '-3Xi4-W-H-n:^\JB12sU:-' ^cm^^-'Wtrsiirv^ra^ hZr"i — r~4~-JIjT''Ti — I--4--4-I k ■ '^.^iPffB-vEp -^ ; ' '^ JIJi — r~~4^I]yTl — f — r-l~~~ 1 1 1 ■ ;^^. -i^^HrtXWcl^ rjW~-JL!^7T — r~'r~l~~ ji ¥^\-[-fjn^^ ,.- ■■« ^ 'I ' ZX— 4— 1 — 1 — T \ A-'V-n'l '~T'~i'~~~l-—LJ t~^f^ — t»^^S3^ i - ~ it TX\\\\^sX\\^ 175 180 175 170 165 160 6, 7, and 12, Table 30-1). In 1958, both eggs and larvae were caught, but in 1962 only eggs. Musienko (1963) reported maximum densities of 500-1,000 eggs/m^ south of Nunivak Island in July 1958, and Kashkina (1965) reported maximum densities of 1,684 eggs/m^ northwest of Nunivak Island and 336 eggs/m^ south of Nunivak Island in July 1962. These are the only reports of quantitative distribution of yellowfin sole eggs. A single postlarva of Limanda proboscidea, the longhead dab, was collected in early September in Bristol Bay (see Fig. 30-31). No eggs of this species were listed in any report. Only three larvae of Platichthys stellatus, the starry flounder, were caught, one in the vicinity of the Pribilof Islands and two in Norton Sound (see Fig. 30-31); all were caught during summer. No eggs of this species were listed in any of the reports. Larvae of Pleuronectes quadrituberculatus, the Alaska plaice, were caught at a single station north of Nunivak Island in July 1962 (see Fig. 30-31). Eggs of Alaska plaice were caught at 57 stations on four cruises in spring and summer although during two of the cruises eggs were caught at only three stations. Eggs of this species were distributed from 55°N off Unimak Island to 59y2°N and from 159°W in Bristol Bay to 175° W, and all were found over or very close to the continental shelf (Fig. 30-34). On one cruise (see cruise 34, Table 30-1) eggs of Alaska plaice were widely distributed, with centers of abundance near the Alaska Peninsula in quadrangle 55y2°N, 162°W, just south of Nunivak Island in quadrangle 59°N, 167°W, and northwest of the PribUof Islands in quadrangle 57i/2°N, 171°W. All the ■1 1 ff' ■ LARVAE OF R^^^^M^^V ^^1 1 LEPIDOPSETTA BILINEATA fc/f?— j I f T^ ^^B \ ■ ^HH 1 •™ \ iTJjy-^B — ■ ■ ■ ■ s M B^VrTrikll^^^^^^H T IJ^^^^^^^^^^^^^^^H ^ 1 L^^^^^^^^^^^^^^^^^^B 1 il^^^^^^^^^^^^^^^^^^m\ '^^i-Jmyhi^^ \ H^^^^^^^^^^^^^^^B ' ^T~r^irrr4Jirr -*. ' 1 v^^^^^^^^^^^^^^pi r~-4~Ji~T~~4~~J^ J — r — r-~h^rl~l T !S JiBj^^V^^^^^^^K^^^' ^~~4^Jij--4-Jjl~^^ * ^TT ^-Xi^^^^^^^^^^^^r \ '~~r~~Oi^[-0 \-\]Wi^^r^W^^r^^ -i7 77--~/--Xj /T^--t--4--- ^^^j — 1 * ■ \ rni 1 n^^^^ *"^ /r7~tfn4~nxr ■ r iT-t— ^-^^^T-'T -^^F T^ ■! ^ 1 _l--l-4^-TM^|^^'T \J* -^ 1 t__i 1— i^m^rti'M'^ \ .it- 'i ^'' \ ~| !^^-i'-4dKnW"k»^^pV^ v-^ 1 mM \-^r^^\\^^^\'^\ A .'' t~ - -YW^^^X-k^rXXV > ^H^-4^+nH^n-VrX ■^-^ ^/jP'lzii— \ — \ — 1" \ \^ — \^\ \. ~. A^'J^^^Ui-A — \ — \ — \\Lli--Vnji_JV ^^^ "" gsj-ci.' --\--yW — \y^-\\\\^ A=^'-3 N» ^ jj^ — 4--4~^ — \\jZ\'\'^r\^ C- L^ ™ _J A^-^WixX^WiX- the unidentified larvae component present in almost aU plankton collections. Synaphobranchidae (1) Notacanthidae (4) Gonostomatidae (4) Alepocephalidae (1) Alepisauridae (1) Anotopteridae (1) Moridae (1) Oneirodidae (2) Zeidae (1) Gasterosteidae (2) Pentacerotidae (1) Trichodontidae (2) Scytalinidae (1) Bothidae (1) Groupings of taxa by habitat Vertical distribution of larvae could not be de- termined for any taxon except pollock, but certain larvae were present either exclusively or mainly in one type of tow or another. Table 30-3 shows taxa 165' 170" 175'' 180- 175- 70-' 165' 160" 155- 150- 1 EGGS OF PLEURONECIES 1 . .^Mit mmmt ^a^imm ■ OUADRITUBERCULATUS 1 wmw »i.. WIKmlL^_^i^^^ - ' - 1 1 m LLU-+-Vi^ mnt 1 1 < k« 'MIB'BSr \ f* "^^BBB* " ~ ^ T- hbbbw ^ imBir ^ rftK ^idasmsB •Hs-LIJZ \ hmh jj r^^^Kl^rT^'4~4iJ if v~~-WulZ T HHW1l> M /~^rr~7T--+JrrT ~r4— HJZ -1= ■HW '~* — ^ffli TtA-X^Tj^CTjtt" r l» !&>-.«. ^ *^i ^iQ^f^f-Xi/r '•4 ~ m^^i^^ m "ff-X^/^iQ-ij/' TT—KEtl ~ 1 _WHiMPK^^^^»sa r-~7j7TT^~Zlirr/~ 1 ~ "■ .. JH-v'tt /~J-Z~lr~Z~tiM ^+J4^~'i% " 1 "' 1 "T^i?-— 1— VtKT -rVliJ^^K SR '^^AJ/r-~fjr7 — Tr -|4lII ft 4 1 \ 1 1 ipiHx3;i] \ AJ^BF '\ /A^ "-X3~7T^-tjrT \ w 1 IIl.rrH"Ti .\ ^Kp^,: w Y— ^Z^j/t^ClT — J — r -iJIILl 1 *' 2 . r _IIJ4MtV? \ J^^KV^' ' Y 2^ T7~7~-/-JlT~r~r--/- "T-I— i— 1 "--J ^^W i 1 l-\=\ — hdj jUB^r-V^, L3t- Z^I/ Tr-^CL^rWr^ ~ ' K<^ 1 T^ IM^y ^^AnO^ ~i/T^f-i3yrT~ ' - 5:::T;iL-iyiytfP^ -jrir~^-ijTT~ " ~ _J— P-^'-t^T '\j\iX-W\i- Sy^T^—tiirT ' ~^ ■^r~^^S--44 ^I13--t\lA' S^*^'J---tlL7 — M '"1■^J^ -^i^^EP^""!! ^i-P4\tX trr^^r~;r$jr/7H ■/'■ 22— -P — Ti — tlt-V4\TCA ^^/■''^--IIj/Tt 4441 ." -' ^•'L r' '^ ZIf-4 — \ — \ TjZ--v-A4\I^ /T~~+^J]Df/W'^ — f— ^i«P5t;-=:i -- J » ^ ' J-ZX--4— 1 — 1 — 1 LJ^--\ — \\ J^^^^tQ7y ^r#4 __ ±tm^4+4+4- ttVnuv Figure 30-34. Number of stations at which eggs Pleuronectes quadrituberculatus have been caught. of n 11 ■ tKiMi ^1^-J 1 LARVAE OF REINHARDTIUS ^BPw-J- - v ^^| I "tfl ■ HIPPOGLOSSOIDES ^T/ i_l ^5^ '1 -'.. H H^^^I^^Kfr * < 1 . f< k. 1 1 \^L^^^^^^^m < \ I'J^^^^^^^^R IPj^^^Hlffj' f-n—h-f-L 1 X M ' ^ , 1 :fl^^^^^^^^^^^^^^^H\ """^i^^^jRy ^-/W'^-/— X / i — rV~-f-^ii- i ' i^^^^^^^^^^^^^^^^^^^^B \ V ~ -^ \ !S -Mr ' ^^i^^^^Mh^ 1 "*. ~ ~ T w^ sTM^^^i^^rrt \ \ i"^"^ — vr^^^^H^^^^^^^ 1 " 1 Li-Vy v'\ v^^^y^'™" Mr^ \ ~' 1 t-l.T_l— Tp^ — v-v-^^^r ^*Pi Q~i~liQitjT~rTM-+ JTJK ^ 1 ~~ '\\\i\>X--r\\K^^Kr\\^ '/^ /rH^--/r/ i — f — T~~i~-Lin V 1 1, 2 .L-i — — x\\ \^P^^" V W -*' ^ ■vl tX4-4-tV'i JJItirT' V fe -3 *; %x iji 1 m_J — 1 — litfffv--^^ \ '^ Lrjr~-T~-4~Ijr>'f~~r~T — t—l~~l^ 1 i-- tuiji;;t4--yiws'ixs-rr\--y 1 ISMi-aBy^P''m:a-rV^ UJ ? / 1 1* * ^^"^ WftvtlJr4--rlJr' < ^ L« 5JA2-yr4TTlVrX SjlW3jIj4rT-+-ti 1 ^ > ^>l-5Ll''J Ph— ^M^'V'I L-i--l'^V''nA' rr*T-~23r7 — I — r — M— iz - f- 2^^ 1J4 4 — H^lZJt-VmL^- CTJJ^C^ -~ '^' ^ ; r^ P" ^^ 2niIl---4--4 — 1 tZV-^V'TuIi it-f-i' ^» '* ■ ^X-' IILX-V--1 — T \ L-V-A^l l~~~h~--l---~Li — Mv-f---^^^&®! H' 1- - 4" 1 i__i___i__-i I — \ — \ \ i— 4 — \^ J 7-^1 rr~-i~~UJ^-'I^Z~T~^ \-^ — — 1 — r Figure 30-33. Number of stations at which larvae of Li- manda aspera have been caught. Figure 30-35. Number of stations at which larvae Reinhardtius hippoglossoides have been caught. of 490 Fisheries oceanography TABLE 30-3 Taxa of larvae caught either exclusively or predominantly in surface net tows; or in combined oblique and vertical net tows. TABLE 30-4 Taxa of larvae caught either exclusively or predominantly in various habitats Habitat Taxa Surface Net Tows Oblique or Vertical Net Tows Osmeridae Hexagrammos sp. Pleurogrammos monopterygius Anoplopoma fimbria Hemilepidotus sp. Myoxocephalus sp. Aptocyclus ventricosus Liparis sp. Bathymasteridae Anarchichas orientalis Stichaeidae (except Lumpenus maculatus) Ptilichthys goodei Pholidae Ammodytes hexapterus Bathylagidae Myctophidae Eleginus gracilis Gadus macrocephalus Theragra chalcogramma Zoarcidae Nectoliparis pelagicus Lumpenus maculatus Lyconectes aleutensis Pleuronectidae present in two types of collections, surface tows and combined oblique and vertical tows. It was more difficult to group larvae by water depth at stations where they were caught. In Table 30-4, six different areas are identified in which various taxa were found either exclusively or pre- dominantly. RESEARCH RECOMMENDATIONS Previous cruises had some serious shortcomings: either each cruise or series of cruises focused on a particular species or area, or else the ichthyoplankton study was of secondary importance to some other primary objective. Before continuing with additional surveys of this type, it is important to obtain a comprehensive view of the total ichthyoplankton community of the Bering Sea. Lack of such informa- tion in the past has resulted in surveys with limited objectives, carried out at the wrong time of the year or in the wrong areas. The knowledge gained from a comprehensive survey would not only clarify inter- pretations of past fragmentary data but would also allow more economical and efficient planning of future surveys. Such a survey would almost certainly have to be a cooperative effort involving Japanese, Soviet, and United States vessels and biologists, conducted over a period of at least one year. Inshore Inner Continental Shelf Total Continental Shelf Continental Shelf and Slope Continental Slope and Deep Water Continental Shelf, Slope, and Deep Water Clupea harengus pallasi Limanda aspera Osmeridae Eleginus gracilis Myoxocephalus sp. Cottidae, unidentified Agonidae Lumpenus fabricii L. maculatus L. medius Zoarcidae Ptilichthys goodei Lyconectes aleutensis Lepidopsetta bilineata Pleuronectes quadrituberculatus Platichthys stellatus ■ Theragra chalcogramma (in spring) Hippoglossoides elassodon Hippoglossus stenolepis Bathylagidae Myctophidae Anoplopoma fimbria Aptocyclus ventricosus Pholidae Atheresthes sp. Gadidae, unidentified Theragra chalcogramma (summer) Sebastes sp. Hexagrammos sp. Pleurogrammus monopterygius Hemilepidotus sp. Bathymasteridae Cyclopteridae Stichaeidae (except those above) Ammodytes hexapterus Reinhardtius hippoglossoides Ichthyoplankton 491 REFERENCES Aron, W. 1960 » The distribution of animals in the eastern North Pacific and its relation- ship to physical and chemical condi- tion. Univ. Wash., Dep. Oceanogr. Tech. Rep. 63. Bailey, R. M., J. E. Fitch, E. S. Herald, E. A. Lachner, C. C. Lindsey, C. R. Robins, and W. B. Scott 1970 A list of common and scientific names of fishes from the United States and Canada (third edition). Amer. Fish. Soc, Washington, D.C., Spec. Pub. No. 6. Clemens, W. A., and G. V. Wilby 1961 Fishes of the Pacific coast of Canada. Fish. Res. Bd. Can., Bull. No. 68 (second edition). Dunlop, H. A., F. H. Bell, R. J. Myhre, W. H. Hardman, and G. M. Southward 1964 Investigation, utilization and regula- tion of the halibut in southeastern Bering sea. Inter. Pac. Halibut Comm., Rep. No. 35. Dunn, J. R., and N. A. Naplin 1973 Planktonic fish eggs and larvae col- lected from the southeastern Bering Sea, May-June 1971. Nat. Oceanic Atmos. Admin., Nat. Mar. Fish. Serv., Northwest Fish. Cent., Seattle, Wash. MARMAP Survey I, Rep. 5. Faculty of Fisheries, Hokkaido University 1957a IV. 1955 cruise of the Oshoro Maru to the Bering Sea and northern North Pacific (NORPAC Project). 15. Data on fish larvae collected with fish larvae net. Data Rec. Oceanogr. Obs. Expl. Fish. 1: 112-15. 1957b V. 1956 cruise of the Oshoro Maru to the Bering Sea. 14. Data on fish lar- vae collected with a fish larva net. Data Rec. Oceanogr. Obs. Expl. Fish. 1:204-7. 1958 I. 1957 cruise of the Oshoro Maru to Aleutian waters. 15. Data on fish larvae collected with fish larva net. Data Rec. Oceanogr. Obs. Expl. Fish. 2:70-4. 1959 III. The Oshoro Maru cruise 42 to the Bering Sea in May -July 1958 (IGY Programme). 12. Data on fish larvae collected with fish larva net. Data Rec. Oceanogr. Obs. Expl. Fish. 3: 122-7. 1960 I. The Oshoro Maru cruise 44 to the Bering Sea in June-July 1959. 12. Data on fish larvae collected with a fish larva net. Data Rec. Oceanogr. Obs. Expl. Fish. 4: 80-6. 1961 IV. The Oshoro Maru cruise 46 to the Bering Sea and North Pacific in June-August 1960. 17. Data on fish larvae collected with a larva net. 18. Data on fish larvae collected with an underway plankton catch V. Data Rec. Oceanogr. Obs. Expl. Fish 5: 202-11. 1962 II. The Oshoro Maru cruise 48 to the Bering Sea and northwestern North Pacific in June-July 1961. 10. Data on fish larvae collected with a larva net. Data Rec. Oceanogr. Obs. Expl. Fish. 6: 78-80. 1964 IV. The Oshoro Maru cruise 4 to the Bering Sea and northwestern North Pacific in May- July 1963. 6. Data on fish larvae collected with a larva net and a small planned Isaacs-Kidd mid- water trawl net in the Bering Sea. Data Rec. Oceanogr. Obs. Expl. Fish. 8: 257-60. 1967 III. The Oshoro Maru cruise 19 to the northern North Pacific and Bering Sea in June-August 1966. 8. Data on fish larvae collected with a larva net. Data Rec. Oceanogr. Obs. Expl. Fish. 11: 219-26. 1968 VI. The Oshoro Maru cruise 24 to the northern North Pacific and Bering Sea in June- August 1967. 10. Data on fish larvae collected by surface tow and midwater tow with a larva net. Data Rec. Oceanogr. Obs. Expl. Fish. 12: 375-82. 492 Fisheries oceanography 1969 I. The Oshoro Maru cruise 28 to the northern North Pacific and Bering Sea and Gulf of Alaska, June-August 1968. 15. Data on fish larvae col- lected with a larva net. Data Rec. Oceanogr. Obs. Expl. Fish. 13: 70-5. 1970 I. The Oshoro Maru cruise 32 to the northern North Pacific, Bering Sea and Bristol Bay in June-August 1969. 12. Data on fish larvae collected with a larva net. Data Rec. Oceanogr. Obs. Expl. Fish. 14: 68-73. 1972 The Oshoro Maru cruise 37 to the northern North Pacific, Bering Sea and the Gulf of Alaska in June- August 1970. Data Rec. Oceanogr. Obs. Expl. Fish. 15: 1-95. 1977 The Oshoro Maru cruise 61 to the Bering Sea and Bristol Bay in June- August 1976. Data Rec. Oceanogr. Obs. Expl. Fish. 20: 1-87. 1978 The Oshoro Maru cruise 65 to the Bering Sea and Bristol Bay in June- August 1977. Data Rec. Oceanogr. Obs. Expl. Fish. 21: 1-79. 1979 The Oshoro Maru cruise 70 to the Bering Sea and the North Pacific Ocean in June- August 1978. Data Rec. Oceanogr. Obs. Expl. Fish. 22: 1-71. Greenwood, P. H., D. E. Rosen, S. H. Weitzman, and G. S. Myers 1966 Phyletic studies of teleostean fishes, with a provisional classification of living forms. Bull. Amer. Mus. Nat. Hist. 131:341-455. 1973 The Oshoro Maru cruise 41 to the northern North Pacific, Bering Sea and Bristol Bay in June-August 1971. Data Rec. Oceanogr. Obs. Expl. Fish. 16: 1-93. 1974 The Oshoro Maru cruise 45 to the northern North Pacific, Bering Sea, Bristol Bay and Chukchi Sea in June- August 1972. Data Rec. Ocean- ogr. Obs. Expl. Fish. 17: 1-129. Hart, J. L. 1973 Pacific fishes of Canada. Bd. Can., Bull. 180. Fish. Res. Haryu, T., Y. Endo, and T. Nishiyama 1979 Ichthyoplankton collected in the Bering and Chukchi Seas July 23-30, 1979. Inst. Mar. Sci., Univ. Alaska, Fairbanks. (Prepared for the Prelimi- nary Report of the R/V Hakuho Maru Cruise KH-78-3.) 1975 The Oshoro Maru cruise 49 to the Bering Sea and Bristol Bay in June- August 1973. Data Rec. Oceanogr. Obs. Expl. Fish. 18: 1-105. 1976a The Oshoro Maru cruise 53 to the Bering Sea and Bristol Bay in June- July 1974. Data Rec. Oceanogr. Obs. Expl. Fish. 19: 1-91. 1976b The Oshoro Maru cruise 57 to the Bering Sea, Bristol Bay and the northern North Pacific Ocean in June-July 1975. Data Rec. Oceanogr. Obs. Expl. Fish. 19: 92-151. Isaacs, J. D., and L. W. Kidd 1953 Isaacs-Kidd midwater trawl. Final Report. Univ. Calif., Scripps Inst. Oceanogr. SIO Ref. 53-3, Oceanogr. Equip. Rep. 1. Kashkina, A. A. 1965 On the reproduction of yellowfin sole (Limanda aspera) in the eastern Bering Sea and the changes in its spawning stock (from samples of ichthyoplank- ton). In: Soviet fisheries investiga- tions in the northeast Pacific, P. A. Moiseev, ed., 4:182-90. U.S. Dep. Comm./NTIS. (Transl. by Israel Prog. Sci. Transl., 1968.) Ichthyoplankton 493 ft 1970 Summer ichthyoplankton of the Bering Sea. In: Soviet fisheries investigations in the northeastern Pa- cific, P. A. Moiseev, ed., 5:225-45. U.S. Dep. Comm./NTIS. (Transl. by Israel Prog. Sci. Transl., 1972.) Musienko, L. N. 1963 Ichthyoplankton of the Bering Sea (data of the Bering Sea Expedition of 1958-1959). In: Soviet fisheries investigations in the northeastern Pacific, P. A. Moiseev, eds., 1:251-86. U.S. Dep. Comm./NTIS. (Transl. by Israel Prog. Sci. Transl., 1968.) NODC Data File 1979 (A computer print-out of data main- tained in the OCSEAP data files at the National Oceanographic Data Center, Washington, D.C., Tracks 450, 532, 558, 561, and 1335.) Novikov, N. P. 1970 I Biology of Chalinura pectoralis in the North Pacific. In: Soviet fisheries investigations in the northeastern Pacific, P. A. Moiseev, ed., 5:304-31. U.S. Dep. Comm./NTIS. (Transl. by Israel Prog. Sci. Transl., 1972.) Quast, J. C, and E. L. Hall 1972 List of fishes of Alaska and adjacent waters with a guide to some of their literature. NOAA, Nat. Mar. Fish. Serv., Spec. Sci. Rep. Fish. 658. Sameoto, D.D. 1969 Serobaba, 1. 1. 1968 , and L. O. Jaroszynski Otter surface sampler: A new neuston net. J. Fish. Res. Bd. Can. 26:2240-4. Spawning of the Alaska pollock Theragra chalcogramma (Pallas) in the northeastern Bering Sea. (Transl. in Probl. Ichthyol. 8(6) 789-98.) 1974 Spawning ecology of the walleye pollock (Theragra chalcogramma) in the Bering Sea. (Transl. in J. Ich- thyol. 14(4):544-52.) Wakabayashi, K., S. Mito, and T. Nagai 1977 Report on the biological research of groundfish in the Bering Sea and the Gulf of Alaska by Shunyo Maru in 1975. Far Seas Fish. Res. Lab., Shimizu 424, Japan. Waldron, K. D. 1978 Ichthyoplankton of the eastern Bering Sea 11 of February to 16 March 1978. Nat. Mar. Fish. Serv., Northwest and Alaska Fish. Cent., Seattle, Wash., Proc. Rep. Waldron, K. D., and F. Favorite 1977 Ichthyoplankton of the eastern Bering Sea. In: Environmental assessment of the Alaskan continental shelf. U.S. Dep. Comm., NOAA, and U.S. Dep. Int., BLM, 9:628-82. Waldron, K. D., and B. M. Vinter 1978 Ichthyoplankton of the eastern Bering sea. Nat. Mar. Fish. Serv. Northwest and Alaska Fish Cent., Seattle, Wash., Proc. Rep. Wilimovsky, N. 1974 Fishes of the Bering Sea: the state of existing knowledge and requirements for future effective effort. In: Ocean- ography of the Bering Sea, D. W. Hood and E. J. Kelley, eds., 243-56. Inst. Mar. Sci., Occ. Pub. No. 2, Univ. Alaska, Fairbanks. Halibut Ecology E. A. Best International Pacific Halibut Commission Seattle, Washington ABSTRACT A small but important halibut fishery exists in the Bering Sea. The distribution of halibut within the area is seasonal and dependent upon climatic conditions. The fish migrate to deep water for spawning during the winter and return to shallow areas for summer feeding. The timing and extent of the summer movement are controlled by oceanographic condi- tions. Spawning is known to occur between Unimak Island and the PribUof Islands at depths of 250-550 m and probably occurs at other locations within the Bering Sea. All stages of life history forms from larvae to spawning adults have been taken. Tagging studies have shown a movement of halibut from the Bering Sea into the Gulf of Alaska. A mixing of stocks of halibut within the Bering Sea is also suggested. Halibut in the commercial landings range from 7 years old and 4.5 kg to over 30 years old weighing over 100 kg. Commercial landings from the southeastern Bering Sea reached a peak of more than 7,000 mt in 1963. Current landings are about 250 mt. Historically management of the halibut re- source has been the responsibility of the International Pacific Halibut Commission. Between 1963 and 1976 management was assumed by the International North Pacific Fisheries Commission, with recommendation from the International Pacific Halibut Commission to the Canadian and United States governments. With the advent of the Fisheries Conservation and Management Act of 1976, management responsibility was returned to the International Pacific Halibut Commission subject to approval by Canada and the United States. INTRODUCTION The Pacific halibut, Hippoglossus stenolepis Schmidt, the largest member of the family of floun- ders called Pleuronectidae, occurs in waters of the North Pacific Ocean and Bering Sea. This large, commercially valuable flounder has been the object of an intense commercial fishery in the northeastern Pacific Ocean since 1888 and was used for subsistence before that. A journal of Captain Cook's voyage of exploration reports the use of "holybret" by the crews of his vessels, natives, and Russian fur traders at Unalaska in 1777 (Munford 1963). An intensive commercial fishery in the early 1900 's off the continental United States and British Colum- bia and in the Gulf of Alaska depleted the resource and created the political climate which brought into existence the convention between the United States and Canada "For the Preservation of the Halibut Fishery of the Northern Pacific Ocean Including Bering Sea," which was ratified 21 October 1924. This original convention provided for the establish- ment of the International Fisheries Commission (IFC), renamed International Pacific Halibut Com- mission (IPHC) in 1953, to provide for the manage- ment of the halibut resource. This convention, which was amended in 1930, 1937, 1953, and 1979,^ still provides the means of managing the halibut fishery on behalf of Canadian and United States citizens. During this 55-year period there have been some trying times, particularly for the management regime established for the Bering Sea. The International Convention for the High Seas Fisheries of the North Pacific Ocean between Canada, Japan, and the United States came into force in 1953. This convention created the International North Pacific Fisheries Commission (INPFC) to manage the stocks of fish within the convention area, including the Bering Sea. A notable feature of this treaty was the principle of abstaining from fishing certain fully exploited stocks of fish. Under this provision Japan was required to abstain from fishing halibut in waters of the North Pacific Ocean and Bering Sea east of 175°W provided that stocks of that species con- tinued to meet the qualifications requiring absten- tion. No determination regarding abstention could be made until the convention had been in effect for five years (BeU 1969). From 1958 through 1961, INPFC annually re- viewed the requirements of the halibut stocks for continued abstention. In 1962, it was agreed by the three contracting parties that halibut in the Bering Sea east of 175°W no longer qualified and halibut in that area was removed from the abstention list. Japanese nationals were permitted to fish halibut beginning in 1963. Other countries not signatory to ^ The 1979 amendment had not been ratified by either Canada or the United States at the time of writing. 495 496 Fisheries oceanography these conventions were free to fish halibut in the Bering Sea without limitation outside the fishing zone established by the United States. With the enactment of the Fisheries and Conserva- tion Management Act of 1976 (FCMA), the control of fishing within 200 miles of the coast of the United States became the exclusive concern of the United States Government. But recommendations to the government for the management of the halibut fishery in the Bering Sea are still the responsibility of IPHC. The development of a halibut fishery in the Bering Sea was inhibited by the distance from established processing plants and transportation facilities and the presence of fishing grounds closer to the home ports of the fishing fleet. The relatively large landings from the Gulf of Alaska have dictated that much of IPHC's research activity be directed to that area to fulfill management directives. LIFE HISTORY Samples of all stages in the life history of halibut from larvae to mature adults have been taken from the waters of eastern Bering Sea. The complex process of evaluating the contribution of each stage to the ecosystem has not yet been accomplished. Description Early ichthyologists describing fish of the eastern Pacific Ocean believed the hadibut there to be identi- cal to the Atlantic halibut, Hippoglossus hippoglos- sus. Schmidt (1904, 1930) noted some differences in the shape of the scales, length of the pectoral fin, and shape of the body and described it as a separate species, H. stenolepis. Vernidub (1936) in further investigations did not consider these differences to warrant specific status and considered the Pacific form to be a sub-species, H. h. stenolepis; Schmidt (1950) later agreed with Vernidub. Since that time North American ichthyologists have detected other differences great enough to separate species and at the present time the American Fisheries Society (1970) accepts H. stenolepis. Halibut are compressed laterally and, except in the larval stages, have both eyes on one side of the head. Adults can be distinguished from other North Pacific flounders by their large size; large, almost symmet- rical mouth with conical teeth; a pronounced arch of the lateral line above the pectoral fin; and the cres- cent-shaped or lunate tail. Hahbut are usually dex- tral, that is, both eyes are on the right or colored side, which is directed toward the surface. Rare specimens have the eyes and color on the left side of the fish. Color varies from olive to dark brown or black, with lighter, irregular blotches. The color pattern of the ocean floor often influences the intensity of the color. The left or blind side is white, and faces the ocean bottom. Halibut are thinner than most other flatfishes, averaging about one-third as wide as they are long. The small scales are well buried in the skin, giving the fish a smooth look. Distribution In the eastern Pacific Ocean halibut are found from northern California into the Bering Sea. The distribu- tion continues across the Bering Sea and off the Asiatic Coast between Hokkaido Island and the Gulf of Anadyr, as well as in the Sea of Okhotsk. Pres- ently the largest concentrations are in the Gulf of Alaska, with a smaller population in the Bering Sea. Halibut are demersal, living on or near the bottom, usually in water of 2-8 C, although they will tolerate colder temperatures. Younger ages are found in relatively shallow inshore areas while adults have been reported as deep as 1,000 m and adult halibut are regularly taken at depths of 300-500 m in the Bering Sea during the spring fishing season. In trawl hauls made in the eastern Bering Sea, the youngest halibut have been found in the southern part but larger, older juveniles and adults are known to range into the more northern part. The younger fish may not have the strength and endurance to make the longer migration and compete with the larger individuals even though the environment is suitable during the summer. The fairthest north that halibut have been taken during IPHC surveys was 62°30'N. Two juveniles were caught in three hauls at this latitude during July. September surveys made by National Marine Fisheries Service (NMFS), when the waters are near their warmest temperatures, have reported a "small" halibut at 65°15'N, and a single adult halibut at 66°02'N, 168°02'W (Best 1977). Seasonal movements During the winter months, ice covers much of the area and water temperatures near bottom drop to 0 C or less; this condition forces the halibut to concen- trate in the deeper, warmer water along the continen- tal edge. With the advent of spring the ice retreats and the water gradually warms to temperatures that permit the halibut to disperse over the expanse of shallow flats, which provide a suitable environ- ment for a nursery for young halibut as well as feeding grounds for the larger juveniles and adults. Halibut ecology 497 Juvenile halibut were abundant at depths of 330-370 m between Unimak Pass and the Pribilof Islands during an IPHC survey in March; few were taken at shallower depths. Novikov (1964) reported concentrations of halibut in the same area in March. In April a concentration of halibut was reported near the northern entrance of Unimak Pass (ca. 54°40'N, 165°09'W) at depths of 80 and 104 m where bottom temperatures were 3.1 and 3.8 C. On the same cruise east of 164°W, when bottom temperatures were —1.0 to 1.4 C, no halibut were taken (Best 1977). As the warming progresses, young halibut move eastward along the north side of the Alaska Peninsula and usually are found throughout Bristol Bay in June. By late June, they spread northward toward Nunivak Island. An IPHC survey in the vicinity of St. Mat- thew Island (60°30'N) during the latter part of June found bottom temperatures to be 0 C or less and only 10 halibut were caught in 14 hauls. Temperatures of 4 C or greater were encountered south of 59°30'N and catches increased to 86 fish in five hauls. An IPHC charter found a concentration of adults in the vicinity of St. Matthew Island in August and Septem- ber, but they left the area by mid-October. The occurrence of small halibut near shore in Norton Sound during July, August, and September was reported by Turner (1886). Many of the tagged halibut released in the Bering Sea have been recovered within the area, confirming the annual migration from deep areas in winter to shallow areas in summer and return to deep water in the winter. Novikov (1970) reported that a halibut tagged in deep water north of Unalaska Island in April was recovered 32 days later on the flats. The fish had travelled 184 km, an average of 5.7 km per day. IPHC tagged nearly 3,600 halibut near St. Matthew Island during August and September and 221 were subsequently returned. Although most (114) were recovered near the release site, 56 were recaptured from deep water along the continental slope between Unimak Pass and the Pribilof Islands, and west to 180°. Most of the recoveries from the deeper areas were made by the North American setline fleet during the April fishing season, although some were from the Japanese trawl fleet during the winter. An additional 51 recoveries were made in the Gulf of Alaska. Several dense concentrations of adult hali- but, such as the above, have been found. These concentrations have provided some excellent short- term fishing, but after the initial exploitation the concentrations no longer exist. A schematic diagram of the winter and summer distribution of halibut in the Bering Sea is given in Fig. 31-1. 170° ^7S' 18ff 175" 170° 165° 160" 155" 150" 170° 175° 180" 175' 170' Figure 31-la, b, and c. Generalized pattern of halibut distri- bution in the eastern Bering Sea. 498 Fisheries oceanography Environmental Factors Water temperature recorded near bottom during the IPHC summer surveys in the Bering Sea ranged from a low of —1.7 C at a station near mid-shelf in 1975 to a high of 10.5 C at a shallow station in Bristol Bay in 1965. During the 1960 's, IPHC surveys caught few halibut at stations with temperatures of less than 2 C and catches were largest when the temperature was 4-5 C. Soviet research trawlers in the southeastern part reported the highest catch per hour at water temperatures of 3.5-5.5 C (Novikov 1964). The Fisheries Agency of Japan (1972) re- ported increased catches of juvenile halibut when bottom temperatures were about 5 C. Commercial halibut fishing in the Gulf of Alaska is generally conducted in water of 3-8 C (Thompson and Van Cleve 1936). A period of cooler water temperatures occurred between 1970 and 1976. Some of the IPHC stations at which halibut had been taken in previous years could not be sampled during June of 1972 and 1975 because of drifting ice. No halibut were taken at several ice-free stations in the survey area where water temperatures were 0 C or lower. A comparison of bottom water temperatures during June of 1967 and 1972 (Figs. 31-2A, B) indicates that under the cooler conditions halibut were found to be concentrated in a smaller portion of the southeastern Bering Sea than in warmer years. The limiting effect of the 2 C isotherm was less noticeable in the colder years. Warm water conditions returned to the Bering Sea again in 1977 and 1978 and juvenile halibut were distributed throughout the survey area much as they were in 1967. The timing and extent of the annual movement from deep to shallow areas seem to depend upon the severity of the preceding winter. If the winter is mild and warming occurs early, the suitable temperature range will extend farther north. These conditions permit the halibut to migrate from the edge grounds onto the flats earlier in the year and open a large portion of the Bering Sea for feeding during the summer. If conditions are favorable, halibut may disperse as far north as Bering Strait. After a cold winter, warming is delayed and the area suitable for halibut is restricted to the more southern portion of the Bering Sea. When conditions inhibit northward dispersion, halibut tend to form more concentrated schools in the southern portion of the Bering Sea. Subpopulation structure The tagging studies conducted by IPHC, Japan, and the U.S.S.R. have shown some indication of the movement of stocks between the eastern and western Bering Sea as well as into the Gulf of Alaska. The most notable movement was made by a fish released 5 July 1975 at 61°18'N, 175°12'E during a joint IPHC-U.S.S.R. research cruise off Kamchatka, and recovered 19 May 1977 near the Shumagin Islands in the Gulf of Alaska. The rate of the interchange has not been calculated due to the small number of tags recovered and the lack of effective fishing effort on adults in the western areas. Migration from the Bering Sea into the Gulf of Alaska, documented during the first tagging experi- ment by IPHC in 1930, has been confirmed by all tagging studies since. Many of the recoveries of tagged halibut released in the Bering Sea were made off southeastern Alaska and British Columbia, and as far south as northern California (Fig. 31-3). One of the longer movements was from St. Matthew Island to Hecate Strait, British Columbia, a distance of over 3,200 km; the fish was captured within one year after tagging. This is an average movement of about 8.8 km per day. Dunlop et al. (1964), after making corrections for fishing effort, calculated that 24 percent of the halibut tagged in the Bering Sea in 1959 emigrated to the Gulf of Alaska. Best (1977) reported that three tagged juveniles released in the southeastern Bering Sea were recovered by the setline fishery in the Gulf of Alaska. The foreign trawl fleet made substantial recoveries from a group of tagged juveniles (mean length about 30 cm) released by IPHC near Unimak Island. These fish were tagged and released in July at 54°45'N, 164°45'W in depths of less than 50 m and were recovered from an area centered about 54°50'N, 165°30'W at depths in excess of 200 m, a distance of about 50 km from the release site, the following winter. Two longer-term recoveries from this release have been recorded, one from west of the Pribilof Islands (58°35'N, 177° 18' W) and the other from the Aleutian Chain near Amukta Island (52°50'N, 171°25'W). If the theory of compensatory emigration accord- ing to which young halibut migrate to "counteract the drift of the natant eggs and larvae and to maintain the species in its habitat" (Dunlop et al. 1964) is valid, the two longer-term tag recoveries would suggest that some of the juveniles in the southeastern Bering Sea originate from spawning grounds to the west, while the three recoveries from the Gulf of Alaska suggest spawning to the east. Certainly these few tag recoveries are insufficient to reach any conclusion on migratory patterns, but they do make for some interesting speculation. A logical explana- tion is that the shallow areas of the southeastern Y I 174 172 170 16 166 164 162 160 158 156 1967 Figure 31 -2a. Distribution of halibut in number per 60-minute iiaul and approximate location of bottom isotherms (C) in June 1967. 174 172 170 168 166 164 162 160 158 156 154 59 1 iiiiii ::-:--:::v:S:-:':-:-:v:':-:-; Wii mmm iii m ■:■:•;■::;■■::.■:::::::,■:■:■:■ 59 58 ..# 0 •■i> ■■ftmsy 0 /'' IF" ■■■1 liii If Wlf, iii; ms 58 57- ■¥■■ ■■:::■■ 0° 0 2 1° ' 1 0 0 -5-— — ■ 15 0 0 0 t 1 1 \\ ..liiiii -iiiiii iiiiiiP 1 J mmm mi 57 2- .. T- 56 55 6\^11 S-—""^ 1 ~~4^\\2 .::Jii mm t ■■ •Sis- ■■■ 55 54 1972 %. 15 0 20 40 ao 80 1 Wkm B ,.:■:.. % HJUT t^ T Figure 31-2b. Distribution of halibut in number per 60-minute haul and approximate location of bottom isotherms (C) in June 1972. 499 500 Fisheries oceanography 170 165 60° rn 1 I I I — r^" 100 50 0 100 58° ^ /wash.] \ Cal. ) 160° Figure 31-3. Recoveries of halibut tagged in the Bering Sea during 1959. box. The number of fish tagged is shown in the black Bering Sea may be a nursery area for more than one stock of halibut. Two tagged juveniles released on the Pacific Ocean side of Unimak Island at approximately the same time as the above Bering Sea releases were recovered in the Bering Sea the following winter by the same trawl fleet. This movement of two small fish was only about 125 km, but it is the only record of any movement of halibut from the Pacific Ocean to the Bering Sea. These scattered fragments of infor- mation only open the door to speculation and do not define the interchange of halibut throughout the Bering Sea and the adjacent areas. Maturity When all information on maturity collected from research cruises in the Bering Sea east of 175°W was taken into account, the age at 50 percent maturity for female halibut was calculated to be 13.8 years, at a length of 122 cm. Males averaged 7.5 years and 72 cm. This is older than the age (approximately 11 years) but about the same size reported for female halibut from the Gulf of Alaska (Schmitt and Skud 1978). In the Bering Sea fewer than 1 percent of the nine-year-olds and about 1 percent of the 90-cm females were observed to be mature. These data fit well with the information reported by Novikov (1964) for the eastern Bering Sea. Novikov (1964) also reported an average fecimdity of 1,164,000 eggs for 12 females between 100 and 120 cm in length and 2,338,000 for females between 120 and 140 cm from the Bering Sea during the period 1957-60. IPHC has no information on fecun- dity of halibut from this area. However, a 1973 study in the Gulf of Alaska (Schmitt and Skud 1978) reported only about half as many eggs for fish of the same size. The increased fecundity may be the result of an evolutionary process to ensure survival in the environment of the Bering Sea; or there may be a discrepancy in the methods of calculation used in the two studies. Spawning Studies conducted in the Gulf of Alaska have provided information on halibut spawning (Thomp- son and Van Cleve 1936). Information to date indicates that halibut spawning in the Bering Sea is similar in nature. An IPHC research cruise during the winter of 1963-64 found mature unspawned fish at depths of 250-550 m between Unimak Island and the PribHof Islands in December. The same general area was fished again in January and most of the fish were spawned out. Spawning undoubtedly takes place at other locations along the continental edge west of the Pribilof Islands as well as along the Aleutian Islands. Halibut ecology 501 The temperature of water at that depth remains a fairly constant 3-4 C the year round; consequently the developing halibut ova probably experience nearly the same conditions as in the Gulf of Alaska. Van Cleve and Seymour (1953) estimated that hatching would take about 23 days at 4.7 C. Under controlled laboratory conditions Forrester and Alderdice (1973) observed hatching after 20 days at 5 C but at 4 C ova did not survive to hatching. Under conditions found in the Bering Sea hatching would probably require at least three or four weeks. The accepted pattern of circulation in the Bering Sea indicates that any spawning southeast of the Pribilof Islands should be carried in a northwesterly direction along the continental slope to the Asiatic Coast (Favorite 1974). Larvae have been captured west of the Pribilof Islands (see Waldron, this vol- ume), providing a measure of circumstantial evidence to support this theory, and fish-of-the-year have been taken from shallow areas of the western Bering Sea. Halibut larvae also have been collected from locations near Unimak Pass (see Waldron, this volume). These collections suggest either that spawning takes place along the Aleutian Islands or that the larvae drift through the Aleutian Island passes from the Gulf of Alaska. Thompson and Van Cleve (1936) reported newly hatched larvae at greater depths than the eggs in the Gulf of Alaska. However, by the age of three to five months all larvae were in the upper 100 m. Young halibut completed metamorphosis and were on the bottom by May or June at a length of 22-29 mm. Musienko (1957) reported taking young-of -the -year halibut 34-42 mm long in September and October off the Kamchatka Peninsula. This period of time for development through metamorphosis seems reasona- ble for the temperature regime found in the Bering Sea. Although the early larval development takes place in rather deep water, the growing larvae gradually rise towards the surface waters. If conditions of currents and food are satisfactory the young halibut should find itself in relatively shallow water close to shore when metamorphosis occurs and the larvae change from a pelagic to a bottom habitat. Although meta- morphosed young-of-the-year have not been reported from the eastern Bering Sea, they probably exist; IPHC regularly captures one-year-old halibut during late June from water less than 15 m deep along the northern shore of the Alaska Peninsula and from the bays on the Bering Sea side of the Aleutian Islands, using special small -mesh nets. Size and age composition The size and age of halibut captured depend to a large extent upon the type of fishing gear used. The juvenile halibut survey by IPHC and fishery resource surveys by the NMFS are conducted with a standar- dized trawl net which catches small halibut mostly in the range of 25-60 cm. The minimum size limit for commercially caught halibut is 81 cm, which is about equal to 4.5 kg. Setline gear will catch some small halibut but the undersized fish are discarded at sea. The commercial landings sampled by IPHC contain halibut from 80 to more than 200 cm. Some fish as young as seven years old will begin entering the setline fishery (a year-class is not fully recruited until age 10) and a few fish are caught each year that are over 30 years old and weigh more than 100 kg. The size distribution and age composition from the juvenile survey and commercial landings taken east of 175°W in 1977 are shown in Fig. 31-4. 50 40 30 20 10 H Trawl N=859 X=3.9 years I I Setline N = 1,298 X = 13.8 years I Lfl flfln n n n n r-i [-1 1-1 2 4 6 8 10 12 14 16 18 20 22 24 26 Age /; ; \ - Trawl i '■ N=871 X=36.1 cm " ' ' i\ Setline N=1,655 X=110.8 cm ■li V <--/---... ^'"^-'^^ 18- 16- 14- 12- 10- 8- 6- 4- 2- 20 40 60 80 100 120 140 (60 180 200 CM Figure 31-4. Length and age distribution of halibut from survey trawl fishing and commercial setline landings. 502 Fisheries oceanography A curious phenomenon of alternating strong and weak year-classes was reported from the juvenile surveys (Best 1977). The pattern of relative year- class strength generally has continued into the older ages in the 1977 commercial landings, where ages 10, 12, 14, and 16 have maintained greater abundance than the intervening ages (see Fig. 31-4). The 1961 year-class, which produced the largest number of 2- year-olds recorded in the juvenile surveys, was still above average as 16-year-olds. Growth Although female halibut are larger than males at any given age, the difference is very small until about five or six years of age. After that time the difference becomes significant and continues throughout the life of the fish. The rate of growth slows noticeably for males at about nine years of age and a length of 90 cm. The decrease in growth rate does not occur in females until age 13 and a size of about 125 cm. The size and age at which the decrease in growth rate occurs in the two sexes suggests that it may be tied to the onset of maturity. A comparison of the length and weight of male and female halibut is given in Table 31-1. These data are actual measurements of fish examined on IPHC research or observer cruises during the period 1961-78. A comparison of growth rates for halibut from different areas shows that fish in the southeastern Bering Sea grow at a faster rate than halibut from the western Bering Sea (IPHC 1976); but according to Southward (1967) they do not grow as fast as fish in the Gulf of Alaska. He also reported an increase in growth of Bering Sea halibut beginning during the mid-1 9 50 's that apparently has continued until the present. Best (1977) reported that a series of cold years (1972-75) in the eastern Bering Sea had reduced the growth of juveniles. The year-classes affected by the reduced growth are only now entering the com- mercial fishery and the short-term reduction in growth would likely be masked by the long-term average used in calculating the size at age. Food habits Since halibut are opportunistic feeders, utilizing whatever is readily available, a wide variety of food items has been found in halibut stomachs. Generally, the small halibut feed on small crustaceans; as they increase in size they begin to eat larger prey, includ- ing more fish. Little information is available on the food of adult halibut. Examination of the stomachs of juvenile halibut from the Bering Sea indicated that they were feeding heavily on shrimp, hermit crabs, small shore crabs, and sand lance (Table 31-2). Novikov (1964) re- ported that halibut over 60 cm feed extensively on pollock and yellowfin sole. TABLE 31-1 Average length in cm and round weight in kg at age for halibut caught east of 175°W Males Females Males Females Age cm kg cm kg Age cm kg cm kg 1 11 .01 12 .01 16 92 9.62 133 31.76 2 20 .07 20 .07 17 94 10.32 137 34.97 3 29 .23 30 .25 18 105 14.77 142 39.27 4 37 .50 39 .60 19 100 12.61 147 43.93 5 45 .95 49 1.25 20 98 11.81 147 43.93 6 56 1.93 63 2.82 21 111 17.68 154 51.05 7 67 3.44 78 5.64 22 102 13.44 161 58.99 8 77 5.41 87 8.03 23 99 12.20 158 55.50 9 86 7.73 97 11.42 24 107 15.70 164 62.63 10 89 8.64 103 13.88 25 106 15.23 167 66.41 11 98 11.81 111 17.68 26 111 17.68 165 63.87 12 98 11.81 118 21.56 27 — — 178 81.66 13 98 11.81 128 28.06 28 — — 182 87.76 14 95 10.68 131 30.24 29 — — 164 62.63 15 89 8.64 131 30.24 >29 — — 197 113.43 Halibut ecology 503 TABLE 31-2 Food items observed in halibut stomachs, southeastern Bering Sea, 1976-1977. Numbers of stomachs in which each item was observed. Food item I 1-10 11-20 21-30 Fish length in cm 31-40 41-50 51-60 61-70 71-80 Total Shrimp 9 47 36 7 Sand lance 1 11 12 7 4 Hermit crab 6 12 6 Shore crab 6 4 2 Isopods 7 Tanner crab 1 2 Pollock 2 1 1 Cottids 2 Smelt 2 Cod 1 1 Annelid worms 1 Euphausiids 1 Blenny 1 Sandfish Sea poacher 102 36 24 13 7 4 4 3 2 2 1 1 1 1 1 In some areas of the Gulf of Alaska (although not significantly in the Bering Sea), halibut were found to feed heavily on Tanner crab. Halibut also have been found occasionally to feed on large king crab, partic- ularly during the soft-shell stage (Gray 1964). In earlier years when the population of halibut was larger it is possible that halibut played some part in keeping the king and Tanner crab populations at lower than present levels of abundance. Reduced predation may have been partially responsible for the increase in the abundance of crabs. Some predation upon halibut must take place, particularly at the smaller sizes, although there is little documented evidence. IPHC has, in the course of the food habit studies, found rare cases of canni- balism upon young halibut. By the age of five or six years halibut are one of the dominant preda- tory fishes and as such must be immune from preda- tion except by the larger marine mammals. Halibut hooked on setline gear are easy prey for sea lions, which can render a portion of a catch unmarketable. STOCK BIOMASS The limited and sporadic nature of the commercial fishery has precluded any reasonable estimate of the biomass of adult halibut. However, IPHC was able to calculate a maximum sustained yield from the southeastern sector (between Unimak Island and the Pribilofs) of 2,268 mt from data available be- tween 1958-63 (Dunlop et al. 1964). This yield was exceeded in 1962 and 1963, and abundance, as indi- cated by catch-per-unit-effort, dropped sharply. With the added mortality of juveniles by the trawl fleets in southeastern Bering Sea, the fishery has never again reached that level. Trawl surveys conducted by NMFS in 1975 and 1976 each produced estimates of nearly 31,000 mt of juvenile halibut (Bakkala and Smith 1978). The addition of adults, not available to the trawls used in this survey, would substantially increase this figure. The removal of small halibut as incidental catch of the foreign trawl fleets during the 1960 's and early 1970's has reduced the potential for recruitment and rebuilding of the adult halibut stocks. Time and area closures on the foreign trawl fleets have reduced the incidental catch of halibut in recent years. Although the FCMA will provide for even greater control of fishing in the southeastern Bering Sea, the halibut stock probably will never return to a size that will permit catches as large as those of 1962 and 1963, which were made from a virtually unfished stock. The strength of year-classes observed as juveniles has fluctuated in alternating years through the 1960's. Recent surveys indicate that the 1973 year-class is large, but this group of fish will not become available to the setline fishery until 1981. A single large year-class cannot significantly improve the size of a population which is ordinarily made up of 504 Fisheries oceanography over 20 year-classes. Several above-average years, preferably successive, will be needed to improve noticeably the supply of halibut. HISTORY OF COMMERCIAL UTILIZATION AND REGULATION The development of a fishery for halibut in the eastern Bering Sea was slow. The lack of processing facilities, the great distance from home ports, often adverse weather conditions, small stock size, and profitable fishing in the Gulf of Alaska all combined to keep fishing effort in the Bering Sea at low levels. Fishery A few U.S. vessels conducted a small fishery in the southeastern Bering Sea between 1930 and 1934. In those years the Bering Sea was open to fishing at the same time as the Gulf of Alaska. No further fishing occurred in the area until 1952, when U.S. vessels began taking about 100 mt annually from fishing grounds east of 175°W (Fig. 31-5). To en- courage fishing in the Bering Sea, the fishing season 7,000 6,000 5,000 - 4,000 3,000 2,000 III Japan I I Canada H United States Figure 31-5. Landings of halibut reported by Canadian, Japanese, and U.S. fishermen from the Bering Sea east of 175°W. was opened one month earlier than in the Gulf of Alaska beginning in 1958. The catch began to in- crease, reaching nearly 4,400 mt in 1962 (Myhre et al. 1977), about equally divided between U.S. and Canadian vessels. INPFC determined that halibut in the Bering Sea no longer qualified for abstention and Japan was allowed to enter the fishery in 1963 (For- rester et al. 1978). Before this time, removals were limited only by the length of the fishing season, but in 1963 a three-nation catch limit of 5,000 mt was established by INPFC for that portion of the Bering Sea along the edge of the shelf between Unimak Island and the Pribilof Islands (roughly between 165 and 170°W), despite the fact that IPHC had calculated the maximum sustained yield from this 5° of longitude to be 2,268 mt for the period of 1958-63 (Dunlop et al. 1964). The 1963 catch from the quota area reached 4,974 mt and the total catch for the area east of 175°W was 7,254 mt. In 1964, a catch limit of 2,900 mt was set for the area between Unimak Island and the Pribilof Islands, but only 972 mt could be taken from the depleted stocks. Japem discontinued longlining for halibut after the 1964 season. Catches declined after that time, and since 1970 have averaged about 250 mt, largely caught by U.S. vessels. Regulations After the poor catches of the 1964 fishing season the open period was limited to only seven days in 1965 on the advice of IPHC. Since then the open period for the area east of 175°W has been gradually increased to its present 20 days in April and Septem- ber, respectively. West of this line continuous fishing is permitted from April 10 to November 15. Poor catches in the eastern sector stimulated interest in fishing farther west, particularly along the Aleutian Islands. Canadian interest in the Bering Sea fishery has de- clined with the growth of their domestic herring fishery, which occurs at approximately the same time as the spring season. The new "Protocol for Regula- tion of the North Pacific Halibut Fishery" eliminated Canadian participation in the Bering Sea, beginning in 1979. IPHC adopted a minimum size for commercially caught halibut of 2.2 kg (5 pounds) dressed weight with head off in 1940. This minimum size was further specified in 1944 as 66 cm (26 in.) in length. The minimum size was increased to 81 cm (32 in.) in 1974, about equal to a 4.5-kg (10-pound) fish, dressed with the head off. In addition to the directed setline fishery for adult halibut, very large numbers of young halibut have Halibut ecology 505 been taken incidentally in the fisheries for other species. Hoag and French (1976) estimated the incidental catch of halibut by foreign fleets from the entire Bering Sea at 11,519 mt in 1971. No estimate of the additional catch by the traps of the domestic crab fishery has been made. Incidental catches of mostly young fish have reduced the recruitment into the adult stocks. International negotiations initiated time-area closures beginning in 1974 for locations and times of greatest incidental catches of halibut. These restrictions have significantly reduced the incidental catches of small fish, which should improve recruitment to the adult stock in the future. The first indication of improvement has been noted in the increased numbers of juveniles in the annual surveys conducted by IPHC (Fig. 31-6). HISTORY OF RESEARCH The U.S. Bureau of Fisheries Steamer Albatross reported a few small halibut taken in the course of a survey of the codfish resource of the Bering Sea in 1890 (Rathbun 1894). The Albatross returned to the Bering Sea in 1911 in the course of a survey of the halibut grounds of the Pacific Coast (Alexander 1912). Only a few small halibut were caught off Akun and Akutan Islands, however; two sets with longline gear made near Unimak Pass caught no halibut and the vessel proceeded to the Pacific Ocean. The next research in this area was the tagging of setline-caught halibut in Makushin Bay off Unalaska Island by IFC in 1930. A small U.S. fishery began about this time and lasted for three or four years. With the rebuilding of the stocks in the Gulf of Alaska the fleet was able to secure profitable catches from grounds closer to port than the Bering Sea. The increased fishing in the Gulf of Alaska required that IPHC direct its research and management efforts to that area. Research in the Bering Sea had a very low priority until funds were appropriated that permitted IPHC to initiate a large tagging program in the 1950 's. In 1956 an IPHC-chartered setline vessel found substan- tial numbers of commercial-sized fish along the shelf-edge between Unimak and the Pribilof islands. The purpose of chartered trips was primarily to tag halibut, but fish unsuitable for tagging were a source of much life history information. The accumulated information was summarized by Dunlop et al. (1964). In addition to the investigations of the halibut stocks available to setline gear, IPHC began a trawl survey of the flats (continental shelf) in 1963 which has continued to date. This study, primarily of the juvenile halibut in the area, has been summarized 1978 Figure 31-6. Abundance of juvenile iialibut (< 65 cm) in the soutiieastern Bering Sea. through 1977 by Best (1977). The NMFS has also collected information on halibut numbers and dis- tribution in conjunction with king crab and ground- fish surveys (Pereyra et al. 1976, Bakkala and Smith 1978). Japan also contributed to hahbut research as part of its commitment to INPFC. Most of the data collected by Japan and IPHC contributed to the management decisions of the INPFC and are included in Pereyra et al. (1976). The U.S.S.R. made a com- prehensive trawl survey between 1957 and 1961 and the information on halibut collected during this period was summarized by Novikov (1964). A joint IPHC-U.S.S.R. research program to tag halibut began in 1975 with a cooperative cruise in the western Bering Sea by a U.S.S.R. vessel with IPHC and U.S.S.R. scientists on board (IPHC 1976). IPHC has continued releasing tags with legends in English, Japanese, and Russian since 1975. The agreement called for the U.S.S.R. to conduct a similar program in the western regions. Information gained from this cooperative research will provide a better assessment of the halibut resource. The role that the Bering Sea plays in the life history and distribution of Pacific halibut, although of great importance, is not well understood. Looking at a map of Alaska, one first sees the Bering Sea as a large water mass isolated from the Pacific Ocean by the chain of the Aleutian Islands. In reaUty, the impression of isolation is false, for the many passes between the islands provide passage for Pacific Ocean 506 Fisheries oceanography water, immigration routes for the pelagic eggs and larvae of halibut, and pathways for later emigration of adults. RESEARCH RECOMMENDATIONS The research carried out in the Bering Sea to date has been concentrated in the eastern part. Even the research efforts of Japan and the U.S.S.R. have been conducted in a disproportionate amount in the eastern part with little if any research reported for the western part. There is a basic need to develop standardized catch and effort statistics on all fishing by all nations for the entire Bering Sea. Forrester et al. (1978) did an excellent job of compiling the available information through 1970. A coordinated tagging program should be de- veloped, stressing areas of the western Bering Sea and Aleutian Islands. Emphasis should be placed on tagging the juvenile halibut, which apparently make extensive movements from the nursery areas to the grounds where they become available to the com- mercial fisheries. Adults also need to be tagged to provide information on the seasonal exchanges of the commercial-sized halibut between winter spawning grounds and summer feeding grounds. The return of recaptured tags must be encouraged, in order to provide the greatest information from these efforts. In conjunction with the tagging program much basic biological information on halibut can be collected from fish unfit for tagging. A simple but neglected project is the comparison of specimens from the eastern and western Bering Sea and the eastern and western Pacific Ocean areas by serological or electrophoretic techniques to determine popula- tion similarities or differences. Any future research should be coordinated among scientists of all nations fishing in the Bering Sea. Communication among researchers is necessary for the best utilization of collected information. In a tag recovery program, communication with the actual fishing fleet is essential if much of the information is not to be lost or distorted before it reaches the issuing agency. SUMMARY Pacific hcdibut are found from northern California into the Bering Sea and across to the Asian coast. Their center of abundance is in the Gulf of Alaska, with a smaller but important population in the Bering Sea. An inshore feeding migration in summer and an offshore spawning migration in winter have been observed. The timing and extent of the inshore migration are influenced by environmental condi- tions, vdth cooler water tending to restrict the movement. A period of relatively cool water tem- peratures altered the summer feeding migration between 1970 and 1976. A directed movement of young fish, probably to counteract the drift during the prolonged egg and larval stages, has also been noted. Movement by tagged young halibut in both an easterly and westerly direction away from the tagging location suggests that there is a mixing of stocks in the Bering Sea. Based on returns of tagged fish, a 24-percent immigration into the Gulf of Alaska was calculated. Female halibut in the Bering Sea mature at about 13 years of age and a size of 122 cm. A female will produce from one to two million eggs or more depending upon its size. Males mature at a smaller size of about 72 cm and at an age of about seven years. Scientific observation indicates that spawning occurs at depths of 250-550 m in the area between Unimak Island and the Pribilof Islands between December and February. Capture of a few larval halibut suggest that spawning also occurs at locations west of the Pribilof Islands and along the Aleutian Islands. Halibut larvae have a pelagic life for at least five to six months and then settle in shallow water near shore. Halibut in the commercial landings are from 7 to (occasionally) over 30 years old and range in size from 80 to over 200 cm. Female halibut are larger than males at each age with the difference becoming quite large at the older ages. Examination of the stomachs of young halibut indicates a preference for shrimp, small crabs, and small fish, with larger fish utilizing larger prey items. The diet is varied; locally abundant prey of the proper size is heavily utilized. Development of a commercial fishery for halibut in the Bering Sea was hampered by lack of facilities, distance from the home port, and adequate fishing in the Gulf of Alaska. Special seasons stimulated interest in the halibut of the area and landings slowly increased during the early 1950's, accelerated as the fleet developed a knowledge of the area, and reached a peak of over 7,000 mt in 1963. Catches have averaged about 250 mt during the 1970 's. Rebuilding of the adult stocks has been hampered by the massive incidental catch of small halibut by the trawl fisheries for other species. Restrictions on trawling activities give promise of improved stock condition in the future. Halibut ecology 507 Regulation of halibut fishing by U.S. and Canadian fishermen has been the responsibility of the IPHC since 1932. The INPFC managed the halibut resource in the Bering Sea from 1963 to 1976, while IPHC made recommendations to the Canadian and United States Governments for presentation to INPFC. After 1976, management of the fish resources of the area became the responsibility of the United States under the FCMA. However, the responsibility for halibut remains with IPHC, a joint Canadian-U.S. agency. Research on halibut has been strongly influenced by the need for stock assessment and management of the resource. Between 1950 and the present a considerable amount of research has been conducted in the southeastern Bering Sea by IPHC and fishery agencies of the United States, Japan, and the U.S.S.R. This research has been largely uncoordinated and of an exploratory nature to develop and monitor the commercial fisheries in the area. Future research should be coordinated to provide a broad spectrum of information and to reduce costs of research by preventing duplication of effort. Best, E. A. 1977 Distribution and abundance of juv- enile halibut in the southeastern Bering Sea. Inter. Pac. Halibut Comm. Sci. Rep. 62. Dunlop, H. A., F. H. Bell, R. J. Myhre, W. H. Hardman, and G. M. Southward 1964 Investigation, utilization and regula- tion of the halibut in southeastern Bering Sea. Inter. Pac. Halibut Comm. Rep. 35. Favorite, F. 1974 Physical oceanography in relation to fisheries. In: Bering Sea oceanog- raphy: An update 1972-1974. D. W. Hood and Y. Takenouti, eds., 157-79. Univ. Alaska. Inst. Mar. Sci. Rep. 75-2. Fisheries Agency of Japan 1972 Report on research by Japan for the International North Pacific Fisheries Commission during the year 1970. III. Groundfish research in the Bering Sea. Inter. N. Pac. Fish. Comm. Ann. Rep. 1970:56-60. REFERENCES Alexander, A. B. 1912 Preliminary examination of halibut fishing grounds of the Pacific coast. Bur. Fish. Doc. 763: 13-56. American Fisheries Society 1970 A list of common and scientific names of fishes. Amer. Fish. Soc, Spec. Pub. 6. Forrester, C. R., and D. F. Alderdice 1973 Laboratory observations on early de- velopment of the Pacific halibut. Inter. Pac. Halibut Comm. Tech. Rep. 9. Forrester, C. R., H. J. Beardsley, and Y. Takahashi 1978 Groundfish, shrimp, and herring fish- eries in the Bering Sea and northeast Pacific— Historical catch statistics through 1970. Inter. N. Pac. Fish. Comm. Bull. 37. Bakkala, R. G. 1978 Bell, F. H. 1969 and G. B. Smith Demersal fish resources of the eastern Bering Sea: Spring 1976. Nat. Mar. Fish. Serv., Northwest and Alaska Fish. Cent., Seattle, Wash. Proc. Rep. Agreements, conventions, and treaties between Canada and the United States of America with respect to the Pacific halibut fishery. Inter. Pac. Halibut Comm. Rep. 50. Gray, G. W., Jr. 1964 Halibut preying on large Crustacea. Copeia 1964:590. Hoag, S. H., and R. R. French 1976 The incidental catch of halibut by foreign trawlers. Inter. Pac. Halibut Comm. Sci. Rep. 60. International Pacific Halibut Commission 1976 Annual Report 1975. 508 Fisheries oceanography Munford, J. K. (ed.) 1963 John Ledyard's journal of Captain Cook's last voyage. Oregon St. Univ. Press, Corvallis. Musienko, L. H. 1957 Young flatfishes (Pleuronectidae) of the far eastern seas. Trans. Int. Oceanology, 20: 254-83 (Trans. Amer. Inst. Biol. Sci., V/ashington, D.C.). Myhre, R. J., G. J. Peltonen, G. St-Pierre, B. E. Skud, and R. E. V/alden 1977 The catch, effort, and CPUE, 1929- 1975. Inter. Pac. Halibut Comm. Tech. Rep. 14 Supplement. Shmidt, P. Yu. 1904 Pisces marium orientalium. St. Petersburg, 224-25 (in Russian). 1930 On the Pacific halibut. Doklady Akad. Nauk SSSR, A, 203-8. 1950 Fishes of the Sea of Okhotsk. Trans. Acad. Sci. U.S.S.R., no. 6. (trans. Smithsonian Inst, and Nat. Sci. Found.). Schmitt, C. C, and B. E. Skud 1978 Relation of fecundity to long-term changes in growth, abundance and recruitment. Inter. Pac. Halibut Comm. Sci. Ptep. 66. Novikov, N. P. 1964 Basic 1970 elements of biology of the Pacific halibut (Hippoglossus hippo- glossus stenolepis Schmidt) in the Bering Sea. In: Soviet fisheries investigations in the northeastern Pacific, P. A. Moiseev, ed., 2:175-219. U.S. Dep. Comm./NTIS. Results of marking Pacific halibut in the Bering Sea. Izvestia TINRO Vol. 74: 328-329. Pereyra, W. T., J. E. Reeves, and R. G. Bakkala 1976 Demersal fish and shellfish resource of the eastern Bering Sea in the baseline year 1975. Nat. Mar. Fish. Serv., Northwest and Alaska Fish. Cent., Seattle, Wash. Proc. Rep. Rath bun, R. 1894 Summary of the fishing investigations conducted in the North Pacific Ocean and Bering Sea from July 1, 1888 to July 1, 1892 by the U.S. Fisheries Commission Steamer Albatross. U.S. Fish. Comm. Bull. 12: 27-201. Southward, G. M. 1967 Growth of Pacific halibut. Halibut Comm. Rep. 43. Inter. Pac. Thompson, W. F., and R. Van Cleve 1936 Life history of the Pacific hahbut (2) Distribution and early life history. Inter. Fish. Comm. Rep. 9. Turner, L. M. 1886 Contributions to the natural history of Alaska. Arctic Ser. Pub. 2, Part IV— Fishes, 87-113, Washington, D.C. Van Cleve, R., and A. H. Seymour 1953 The production of halibut eggs on the Cape of St. James spawning bank off the coast of British Columbia, 1935- 1946. Inter. Fish. Comm. Rep. 19. Vernidub, M. F. 1936 Data concerning the Pacific white halibut. Proc. Leningrad Soc. Nat. 65 143-182. I Distribution, Migration, and Status of Pacific Herring Vidar G. Wespestad Northwest and Alaska Fisheries Center Seattle, Washington Louis H. Barton Alaska Department of Fish and Game Anchorage, Alaska ABSTRACT Pacific herring are an important part of the Bering Sea food web and form the basis of a major commercial fishery. Until recently Japan and the U.S.S.R. have been major exploiters of herring. Catch peaked in the early 1970's at 145,579 mt, and then declined in response to overfishing and poor recruitment. But recently herring abundance has increased, and the United States has become the dominant exploiter. Most herring are harvested in coastal waters during the spawning period, which begins in late April /mid-May along the Alaska Peninsula and Bristol Bay and progressively later to the north. Spawning occurs at temperatures of 5-12 C and the time of spawning is related to winter water temperatures, i.e., early in warm years and late in cold years. During spawn- ing, eggs are deposited on vegetation in the intertidal zone of shallow bays and rocky headlands. Eggs hatch in two to three weeks as planktonic larvae and metamorphose to juveniles after six to ten weeks. Little is known of larval and juvenile stages in the eastern Bering Sea. Sexual maturity begins at age two, but most herring mature at ages three and four, the ages of recruitment to the fishery. Herring as old as 15 occur, but very few beyond age 10 are present in commercial catches. Age-specific mortalities are unknown but the average instantaneous natural mortality rate is estimated to be 0.46-0.47. The rate of growth is generally greater in the Bering Sea than in the Gulf of Alaska, but within the Bering Sea growth decreases to the north. Herring feed on the predominant larger zooplankton^euphausiids and cope- pods. Three major stocks occur in the Bering Sea: northwest of the Pribilof Islands, the Gulf of Olyutorski, and Cape Navarin. These have been identified as individual stocks based on differences in growth and maturation rates, and dissimilar age structures. Herring wintering northwest of the PribUof Islands migrate to the Alaska coast in spring and spawn in Bristol Bay and between the Yukon and Kusko- kwim Rivers. Although some may also spawn in the eastern Aleutian Islands, Alaska Peninsula, and Norton Sound , herring in these areas may also winter inshore near spawning grounds. North of Norton Sound some herring wdnter in brackish lagoons and estuaries. Herring wintering northwest of the Pribilof Islands arrive on the winter grounds in October and concentrate in waters of 2-4 C at depths of 105-137 m through the winter. In March schools migrate to the coast for spawning. After spawning they must remain in coastal waters, for few are found on the shelf or slope. In late August, they reappear in offshore waters in the areas of Unimak and Nunivak islands, and the seaward migration to the winter grounds continues through September and October. Although assessments of eastern Bering Sea herring have ranged from 374 thousand mt to 2.75 million mt, the current estimate of spawning biomass is 432-864 thousand mt. Fish- eries data indicate that herring declined rapidly after peak harvests in the early 1970's and that peak catches were sup- ported by a few strong year-classes. Weak year-classes oc- curred through the early 1970's, and recruitment appears to be normalized in recent years. Research is needed to refine estimates of abundance and biological characteristics of stocks, to improve the ability to predict changes in resource abundance, composition, and availability, and to identify the origin and distribution of herring in offshore areas. INTRODUCTION The Pacific herring (Clupea harengus pallasi) is a member of the family Clupeidae, which has global distribution and includes about 50 genera and 190 species found mostly in tropical and temperate waters (Svetovidov 1952). Herring occurring in the North Atlantic and North Pacific areas are similar in appear- ance, but differ mainly in number of vertebrae (55-57 vs. 52-55). Pacific and Atlantic species also differ biochemically, with significant differences 509 510 Fisheries oceanography observed in the frequencies of eight genes (S. Grant, Northwest and Alaska Fisheries Center, Seattle, Washington, personal communication). Svetovidov (1952) believes that Pacific herring arrived from the Atlantic some time between the Pliocene and the "post-glacial 'transgression' " via the Asian Arctic. Pacific herring also differ from Atlantic herring in spawning and migrational behavior: the former are spring spaviTiers, whereas the latter may be spring, winter, summer, or autumn spawners. Pacific herring spawn between the intertidal zone and about 20 m and deposit eggs on vegetation, whereas Atlantic herring spawn in deep water on a gravel bottom. Pacific herring generally remain near the spawning ground the year round and do not make extensive seasonal migrations as many Atlantic stocks do. In the North Pacific Ocean, herring are distributed along the Asiatic and North American continental shelves (Fig. 32-1); in Asia they range from Taksi Bay, near the mouth of the Lena River, to the Yellow Sea (Andriyashev 1954), and in North America from Cape Bathurst in the Beaufort Sea to San Diego Bay, California (Hart 1973). Herring are an important part of the eastern Bering Sea food web. They are pelagic planktivores, highly adapted with large mouths and numerous fine giU rakes for efficient utilization of euphausiids, cope- pods, and other zooplankton. In turn, herring are important prey for marine mammals, birds, and roundfish. Mathematical simulations of the eco- system in this area by Laevastu and Favorite (1978) indicate that annual total mortality amounts to one half of herring biomass production and that 95 percent of total mortality is by predation. Given this level of mortality, it is understandable that herring stocks exhibit strong fluctuations in abundance with apparently small changes in fishing or environmental factors. The abundance of herring declined sharply in the early 1970's and only recently has an increase be- come apparent. Although several hypotheses could be advanced to explain the strong population fluctua- tions observed, data are insufficient to conclusively establish a cause. Since rapid, marked changes in abundance are expected to occur in the future, it will be necessary to be able to identify the causes and predict their occurrence and magnitude. Present knowledge is rudimentary and inferences Figure 32-1. Geographic range of Pacific herring (Clupea harengus pallasi) in the North Pacific Ocean. Pacific herring 511 about many phases of life history must be drawn from other more thoroughly studied populations. Research is needed on all aspects of herring biology, especially interspecies interactions and environmental effects on herring. FISHERIES Archeological excavations in the area indicate that net fisheries had been developed as early as 500 B.C. (Hemming et al. 1978), and subsistence fishing for herring continues today in many native villages, especially in villages between the Yukon and Kusko- kwim rivers where alternative food resources (i.e., salmon, moose) £ire absent or in low abundance (Barton 1978). Commercial herring fisheries devel- oped in the northern Bering Sea around the turn of the century. Marsh and Cobb (1910) reported that a small fishery in Grantley Harbor on the Seward Peninsula in about 1906 supplied salt herring to Nome. Before 1909, another small herring fishery de- veloped in Golovnin Bay, Norton Sound (Rounsefell 1930). Although the Grantley Harbor fishery appar- ently was short-lived (the last reported catch was in 1917, when 300 barrels were packed), the Golovnin Bay fishery operated until 1941. The first large-scale herring fishery began in 1928, when a purse-seine fleet fished at Unalaska. Salteries were established at Dutch Harbor in subsequent years, and the catch increased to a peak of 2,277 mt in 1932 (Barton 1978). Catches ranged between 1,000 and 2,000 mt until 1937, and thereafter declined until 1946, when the fishery ended. Lack of demand and accompanying low prices for cured herring were the principal reasons for the demise of this fishery (Wespestad 1978a). A herring fishery resumed in 1959, when Soviet exploratory trawlers found wintering concentrations along the continental slope northwest of the Pribilof Islands (Dudnik and Usoltsev 1964). During the first season, 10,000 mt were harvested (Fig. 32-2). Catches increased in later years as effort increased but declined sharply in 1965 and 1966 when herring could not be located and effort was greatly reduced. Japanese vessels began fishing for herring in the late 1960's. A trawl fishery was established on the winter grounds from November to April and a gillnet fishery, which operated off the western Alaska spawning grounds, from April through June (Wes- pestad 1978b). In 1977, the area east of 168°W and north of 58° N was closed to foreign herring fishing to protect native subsistence fisheries, and in 1978 the 168°W closure line was extended to the Alaska Peninsula. This greatly limited the gUlnet fishery; no I I USSR. ^H Japan I I m I Hh- I 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 Year Figure 32-2. Eastern Bering Sea herring catch 1960-79. 8-1 r-150 0 |llM|llll|llll|llll|llll|llll|llll'| 0 1967 1968 1969 1970 1971 1972 1973 1974 Figure 32-3. Catch and catch per unit effort (CPUE) relationship for large Japanese stern trawlers in the eastern Bering Sea, 1967-74. foreign gillnetting occurred in 1978 and only a very limited amount in 1979. Catch and effort peaked in the late 1960 's and early 1970's and then decUned (Fig. 32-3). The peak catch occurred in 1970 when 145,579 mt were harvested (see Fig. 32-2). Catch dropped abruptly the next year, increased slightly in 1972, and then declined until 1976, when an increase occurred. In 1977, an allowable catch of about 21,000 mt was established by the U.S. when the 200-mile Fishery Conservation Zone was established. Herring harvests have been maintained at this level to the present. U.S. herring fisheries resumed on a small scale in Norton Sound and northern Bristol Bay in the late 512 Fisheries oceanography 1960's for herring roe and herring roe-on-kelp for the Japanese market. Harvests were small, generally under 100 mt, until 1977, when the catch increased to 2,550 mt. The fishery expanded further in 1978, and the catch rose to 7,025 mt. In 1979 the catch increased again to approximately 12,000 mt. Most of the U.S. harvest is taken with purse seines and gillnets in northern Bristol Bay between Cape Con- stantine and Cape Newenham; smaller fisheries also occur in Goodnews Bay, Security Cove, Cape Ro- manzof, and Norton Sound (Fig. 32-4). GENERAL BIOLOGY Spawning Herring spawn along the western Alaska coast in late spring to midsummer (Fig. 32-4). In most years, spawning occurs first along the Alaska Peninsula and in Bristol Bay in late April to mid-May and progres- sively later to the north. In Kotzebue Sound, it may extend from July through mid-August (Barton 1978). Spawning usually commences soon after the spawning grounds become ice-free; it has been noticed to begin 170" 175" 180° 175" 170° 165" 160° 155" 150" 65 56' 53° luu 200 km 50 100 miles I I LATE JUty- EARLY AUGUST LATE 4UNE- ' EARLY JULY MID JUNE-EARLY JU^tV COASTAL SPAWNING GROUNDS OF PACIFIC HERRING 1976-1978 Size of dot indicates relative abundance on spawning ground 65° Figure 32-4. Distribution, time of spawning, and relative abundance of Pacific herring on coastal spawning grounds ob- served during aerial surveys, 1976-78. Pacific herring 513 when water temperatures are approximately 3-5.5 C (Scattergood et al. 1959), although it has been recorded over a range of 6-10 C in northern Bristol Bay (Warner and Shafford 1977) and in a range of 5.6-11.7 C on spawning grounds between Norton Sound and Bristol Bay (Barton 1979). Prokhorov (1968) found that the approximate time of spawning in the western Bering Sea is related to winter and spring water temperatures with early maturation in warm years and delayed development in cold years. The past two years (1978 and 1979) have been mild winters, 1979 especially so, and herring have arrived at the spawning ground several days to two weeks earlier than average; in 1976, a cold year, spawning herring were not evident until mid -June. Svetovidov (1952) believes that the shore-spawning behavior of Pacific herring is caused by a lack of high water temperature in deeper water. Spawning may last from a few days to several weeks. Generally, the older herring are the first to spawn, followed by successively younger fish. Eggs are deposited on vegetation in intertidal and shallow subtidal waters, predominantly on rockweed (Fucus spp.) and eelgrass (Zostera spp.) (Barton 1979). There are two types of spawning habitats: rocky headlands and shallow lagoons and bays (Barton 1978). South of Norton Sound most spawn is found in the intertidal zone on rockweed, while from Norton Sound northward most spawn is found in the shallow subtidal zone on eelgrass. Barton believes that these differences are partly due to smaller tide changes in the northern areas. Eggs take 10-21 days to hatch, depending on water temperature. In northern Bristol Bay, hatching takes 13-14 days at 8-11 C (Barton 1979). However, Alderdice and Velsen (1971) have suggested that optimum temperatures for Pacific herring egg devel- opment are 5-9 C and that below 5 C eggs die. Little is known of the magnitude or causes of egg mortality on eastern Bering Sea spawning grounds, but studies in British Columbia revealed wave action, exposure to air, and bird predation as major causes (Taylor 1964). Wave action may be a major cause in the Bering Sea; it has been observed that a severe storm after spawning activity destroyed both depos- ited eggs and rockweed in the upper intertidal zone along the south shore of Cape Romanzof (Gilmer 1978). Predation of fish may also be an important cause of egg mortality. Concentrations of flatfishes, particularly yellowfin sole, have been observed on the spawning grounds in northern Bristol Bay (John Clark, ADF&G, personal communication). Stomachs of flatfish examined in spawning areas by both authors have revealed a high rate of egg consumption. Larval and juvenile development: General background information from Barkley Sound, B.C., studies Herring hatch as larvae averaging 8 mm in size, and remain in this planktonic stage for approximately 6-10 weeks, at which time the larvae have grown to approximately 30 mm and begin to metamorphose into juveniles (Taylor 1964). During the larval stage, they are subject to high and variable mortality rates. An important reason for mortality may be failure to obtain proper food after the yolk sac is absorbed. Another may be passive transport away from the coast by prevailing currents (Outram and Humphreys 1974); Stevenson (1962) found that when larvae in Barkley Sound were transported to the open sea, few survived. He did not find temperatures and salinity to be important in inshore areas but believed that the high mortality of larvae offshore might be connected to the increased salinity of the open sea. There are indications that the direction and magnitude of surface stress may affect the survival of larvae; northward wind stress (along a N/S coast) will result in net onshore flow and water piling up against the coast, causing onshore retention of larvae and good year-classes. Offshore movement is associated vdth poor year-classes (Outram and Humphreys 1974). When metamorphosis is complete, juveniles are free-swimming and begin to form schools that enlarge and move out of the bays as summer progresses (Taylor 1964). Hourston (1959) found that juveniles moved from the spawning grounds on the northwest side of Barkley Sound to rearing grounds on the southeast side. No specific reason could be found for migration to the southeast other than a preference for calmer, sheltered water found there. Juveniles in Barkley Sound actively feed at depths of 0.6-5 m at dawn and dusk. No sampling was done at night, but some inactive schools were observed near the surface. Juvenile schools were found in a range of salinities, but most were found at 25*^ /oo, which corresponds to Fujita and Kokudo's (1927) point of best fry survival. Bering Sea Little is known about the juvenile stage in the Bering-Chukchi Sea region from the time herring leave the coast in their first summer until they are recruited to the adult population. Rumyantsev and Darda (1970) indicate that juveniles feed in coastal waters in summer and move to deeper water in winter (juvenile herring in British Columbian and southern Alaska waters winter offshore and reappear in bays the following summer— Taylor 1964; Rounse- fell 1930). In the western Bering Sea, herring aged 0 514 Fisheries oceanography and 1 inhabit areas nearer shore and at lower tem- peratures than adults (Prokhorov 1968). Of juveniles found in the Port Clarence area in 1977, more than 50 percent were captured in Imuruk Basin, the brackish forebay of the Port Clarence/ Grantley Harbor complex. Apparently herring in the northern Bering Sea and Chukchi Sea may have a tolerance for much lower salinity, for eggs and fry were found in Imuruk Basin near Port Clarence in water of 4*^/oo (Barton 1978). Although juveniles were present in the spring spawning period (late June-early July), significant numbers were not captured until mid-August. Their presence in Hotham Inlet in November was indicated by stomach contents of sheefish (Stenodus leucichthus). Sub- stantive numbers of age-one herring were captured in June 1978 in Hagemeister Strait of northern Bristol Bay (Barton 1979). Wolotira et al. (1977) found both mature pre- spawning and immature herring in autumn (Septem- ber-October) trawl catches made in the offshore waters of the northern Bering Sea and southern Chukchi Sea; however, age-0 herring were only found in the pelagic area of Norton Sound between Cape Douglas and Golovnin Bay. Maturation and fecundity Herring spawn for the first time at ages two to six but the majority do not spawn until ages three (50 percent mature) and four (78 percent mature). By age five, 95 percent of the population has matured (Rumyantsev and Darda 1970). Sexual maturity of eastern Bering Sea herring coincides with recruitment into the fishery, primarily at ages three and four. In the herring's southern range, the onset of sexual TABLE maturity occurs earlier (stocks mature between ages three and four in British Columbia and ages two and three in California— Hart 1973, Rabin 1977). In mature herring, fecundity increases with body length and latitude (Nagasaki 1958); fecundity appears also to be higher in the eastern Bering Sea area than in the Gulf of Alaska or western Bering Sea (Table 32-1). Age and growth Herring have been found to live as long as 15 years (Barton 1978), and generally occur in substantial numbers from ages three to six, but when strong year-classes occur, ages seven to ten may comprise a substantial portion of the catch. Stocks grow at about the same rate £is those in the Gulf of Alaska and British Columbia until ages three to four, but growth is greater in the Bering Sea for older fish, and they achieve a greater maximum length and weight than the more southern stocks (Fig. 32-5). Rounsefell (1930) reported many herring of 380 mm in the catch at Unalaska (compared to a maximum of 330 mm reported for British Colum- bia—Hart 1973). In more recent investigations, Rumyantsev and Darda (1970) and Warner (1976) have found Bering Sea herring of 340-345 mm. Barton (1978) found that size-at-age in spawning concentrations along the western Alaska coast from Norton Sound northward is significantly smaller than in stocks to the south (Fig. 32-5). A general growth curve was derived for eastern Bering Sea herring by applying von Bertalanffy's equation to data reported by Shaboneev (1965) from the wdnter trawl fishery : /t = L^(l-e-^ (*-to)) 32-1 Fecundity (in thousands of eggs) of Pacific herring in different areas of the northern Pacific Ocean Area Age 4 5 6 7 8 Source E. Bering Sea 1963 1964 26.6 26.6 34.4 32.1 46.1 52.4 59.5 53.5 70.8 77.8 Shaboneev (1965) Rumyantsev & Darda (1970) Alaska Peninsula 1976 Mean: 26.4 Range: 12.6-84.8 Ages IV-VI Warner (1976) Karaginskii Bay 1963 26.4 30.1 37.4 Kachina^ W. Bering Sea 1964 39.2 43.3 50.6 Kachina Vancouver 1955 19.9 23.8 29.6 38.2 30.4 Nagasaki (1958) * Cited in Rumyantsev & Darda (1970) Pacific herring 515 LENGTH VS AGE OF PACIFIC HERRING 300-1 250- 200 150 100 Port Clarence Norton Sound Togiak SE Alaska "1 r- r 4 6 Age 10 Figure 32-5. Size-at-age comparisons of Pacific iierring from selected areas in the eastern Bering Sea and tiie nortlieastern Pacific Ocean. The parameters of von Bertalanffy's equation are: L„ (maximum length in mm) = 324.5, K (growth rate) = 0.35, and to (age in years, fish was 0 length) - 0.0261. Warner (1976) computed a von Bertalanffy curve for fish captured in trawl samples in Bristol Bay. His coefficients were L„ = 299, K= 0.18, and t^ = 2.10. These estimates, although lower, do not differ signifi- cantly from Shaboneev's data, given the VEiriances reported by Warner. Mortality Beyond the larval stage the rate of natural mortal- ity decreases sharply and continues to decrease slightly until about age five, when it begins to in- crease from senility, disease, and spawning mortality (Fig. 32-6). The magnitude of mortality in the juvenile stage is actually unknown but inferences from other fish populations suggest that it is highest in years of high egg and larval survival because of intense intraspecies food competition (Ricker 1975). A general estimate of the natural mortality rate for eastern Bering Sea herring was derived by Wespestad (1978a), using the procedure of Alverson and Carney (1975). This method estimates the natural mortality rate for eastern Bering Sea herring to be 0.47. Nat- ural mortality can also be determined by an analysis of catch curves using regression techniques (Ricker 1975). Applying this method to data presented by Laevastu and Favorite (1977), we have determined the instantaneous mortality rate for fully recruited ages of eastern Bering Sea herring to be 0.46. Age- specific natural mortality estimates are unavailable for the eastern Bering Sea stocks; however, they are TABLE 32-2 Instantaneous rate of natural mortality (M) for herring in the northeastern Pacific Area/Age 3 4 5 6 7 8 Source S. E. Alaska .20 .30 .46 E. Vancouver Is. .40 W. Vancouver Is. .46 .59 .64 .61 .72 .77 .72 .85 .85 .79 Skud (1963) Tester (1955) Tester (1955) probably similar to natural mortality rates estimated for herring stocks in southeastern Alaska and British Columbia (Table 32-2). Food and feeding The first food of herring larvae is usually limited to small and relatively immobile plankton organisms that the larvae must literally nearly run into to notice and capture. Microscopic eggs sometimes make up more than half of the earliest food; other items include diatoms and nauplii of small copepods. Herring do not have a strong preference for certain food species but feed on the comparatively large organisms that predominate in the plankton of a given area (Kaganovskii 1955). Feeding generally occurs before spawming and intensifies afterward (Svetovidov 1952). During the winter, feeding declines; it ceases in late winter (Dudnik and Usoltsev 1964). During November and December, in Kam- chatka waters of the western Bering Sea, Kachina and 240n o UJ Q. o o c •5 c (0 (A c (0 « Age (years) Figure 32-6. Total mortality of Pacific herring expressed as percent annual cohort biomass loss. 516 Fisheries oceanography Akimova (1972) found that juvenile herring con- sumed small and medium forms of zooplankton (chaetognaths, copepods, tunicates) and bentho- plankton (mysids). Euphausiids, amphipods, mol- lusks, and other organisms were found rarely, and usually in small quantities. In the demersal zone, herring stomachs contained quantities of tubes of polychaete worms, bivalve moUusks, amphipods, copepods, juvenile fish, and detritus. In the eastern Bering Sea, stomachs in August were 84 percent filled with euphausiids, 8 percent with fish fry, 6 percent with calanoids, and 2 percent with gammarids (Rumyantsev and Darda 1970). Fish fry, in order of importance, were walleye pollock, smelt, capelin, and sand lance. In spring, food was mainly Themisto (Amphipoda) and Sagitta (Chaetognatha). After spawning, the main diet was euphausiids, Calanus spp., and Sagitta spp. (Dudnik and Usoltsev 1964). Nearly 75 percent of herring stomachs examined in the spring from Bristol Bay to Norton Sound either were empty or contained only traces of food (Barton 1978). Only 25 percent of the stomachs examined were at least 25 percent or more full, and only 3.4 percent were completely full. Major food items were cladocerans, flatworms (Platyhelminthes), copepods, and cirripeds. DISTRIBUTION Stock distribution Three major herring wintering grounds have been identified within the Bering Sea: northwest of the Pribilof Islands, in the Gulf of Olyutorski (Prokhorov 1968), and near Cape Navarin (N. Fadeev, TINRO, Vladivostok, U.S.S.R., personal communication) (Fig. 32-7). Differences in the pattern of migration be- tween the coast and the outer continental shelf have effectively isolated Asian and North American herring in the Bering Sea. The different grovi^h and matura- tion rates and dissimilar age structures reported by Kachina (1978) of those wintering near Cape Navarin and those wintering northwest of the Pribilof Islands suggest that although these groups winter in close proximity there is little or no mixing between them. Most herring which winter near the Pribilof Islands are believed to spawn in Bristol Bay and in areas between the Yukon and Kuskokwim rivers. This conclusion is based on Soviet research, similarities in age composition, and the distribution of Japanese trawl catches during the spawning migration (Wespestad 1978b, Barton 1979). ADF&G aerial surveys indicate that the greatest abundance of spawning herring occurs in the Bristol Bay area Figure 32-7. Range of Pacific herring in winter and spring. and smaller spav^Tiing concentrations occur to the north and south (see Fig. 32-4). The relation of herring spawning in Norton Sound to spawning stocks to the south is unclear. Those in Norton Sound are genetically similar to spawning stocks to the south (Grant 1979) and appear in inshore waters in late May to early June (Barton 1978), which suggests they may winter offshore. However, it is possible that some or all herring remain in Norton Sound the year round. Barton (1978) reported that an autumn, non-spawning run occurs in Golovnin Bay in northern Norton Sound, and herring have been caught through the ice by local residents jigging for cod in this area and near Nome. Moreover, herring have been found in ringed seal (Phoca hispida) stomachs collected near Nome in November. To the north of Norton Sound, herring occur in Port Clarence, and in inlets from the Bering strait to areas within Kotzebue Sound. Many, if not all, stocks found north of Nome remain in the im- mediate area the year round and winter in coastal lagoons and brackish bays, even though in several locations (e.g., Port Clarence, Shishmaref Inlet, and inner Kotzebue Sound) ice may cover the surface (Barton 1978). Herring may also occur along the Alaska Peninsula and throughout the Aleutian Islands. Marsh and Cobb (1911) reported a large spawning at Atka Island in 1910, and spring and autumn runs (the latter presumably non-spawning) at Unalaska and Port Heiden. The fishery which operated at Unalaska in Pacific herring 51 7 0- (-1.8 ■=-0°) (4° -6°) (6_ -11°) age 25- (3' .", J J 7 spent 50- - immature ,2» -3°) age 4-5 (2 0_4O) £■ . (2 -3°) X ^5- (-l">-0°) h- Q_ Ld Q 100- - - (2° -3°) (go 125- 1 Rn- •^ '- - MAY JUNE JULY AUG SEPT OCT NOV DEC MONTHS Figure 32-8. Monthly distribution of Pacific lierring by temperature and depth in the eastern Bering Sea. May- November data from Rumyantsev and Darda (1970) December data from Shaboneev (1965). the 1930's and 1940's harvested herring in summer and early autumn, averaging 1,337 mt between 1929 and 1937. The current status of these stocks and their relationship to other eastern Bering Sea stocks are unknown. Recent aerial surveys by ADF&G have found small spawning concentrations on the north shore of Unimak Island, in Heredeen Bay, and in Port Heiden (Warner and Shafford 1977). Catches by Japanese trawlers just north of Unimak Pass in winter indicate that this may be the wintering area of herring spawning on the Alaska Peninsula (Wespestad 1978b). Seasonal distribution— Pribilof stock Temperature may influence seasonal distributions more than anything else. Soviet scientists found spring herring moving through sub-zero water tem- peratures on the way to spawning grounds, but during summer months they were found on the shelf in the warmer, upper layers of the water column (Fig. 32-8). As we said before, Svetovidov (1952) believes that migrations to the coast for spawning developed because of the lack of sufficiently warm water in spring and summer in the North Pacific Ocean. Furthermore, earlier warming of coastal waters provides earlier development of phytoplankton and zooplankton and hence better feeding conditions. Winter The major wintering ground of eastern Bering Sea herring is northwest of the Pribilof Islands, approxi- mately between 57 and 59° N lat., representing an area of 1,600-3,000 km^ (Shaboneev 1965) that shifts in relation to the severity of the winter. In mild winters herring concentrate farther north and west, and in severe winters they move south and east (Fig. 32-7). Dense schools are found during the day a few meters off the bottom at depths of 105-137 m and at water temperatures of 2-3.5 C (Dudnik and Usoltsev 1964). Very few were found shallower on the continental shelf where lower temperatures prevailed. Distinct diurnal vertical migrations occur in early winter; however, as the season progresses, diurnal movements diminish and herring remain on bottom during the day and slightly off bottom at night (Shaboneev 1965). Spring Soviet scientists investigating herring distribution in the mid-1960 's found that herring left the winter- ing grounds in late March : they believed that herring followed two routes to the coast, and Japanese trawl catches in April and May indicate one major and one minor path (Fig. 32-7). The past two years (1978-79) have been mild winters, 1979 especially. In these years, herring arrived on spawning grounds along the coast several days to two weeks earlier than average; in 1976, a cold year, they were not found until mid-June. Summer The failure of Soviet surveys in the mid-1960 's, using gillnets and trawls, to find herring concentra- tions on the Bering Sea slope or shelf in the summer suggests that most herring apparently remain tempo- rarily in coastal waters after spawning. Annual NWAFC summer trawl surveys covering much of the continental shelf of the eastern Bering Sea support this conclusion, for very few herring have been taken in summer surveys (Pereyra et al. 1976, Bak- kala and Smith 1978). A hydroacoustic survey conducted along the outer shelf between Unimak Pass and the U.S.-U.S.S.R. convention line in June-July 1979 found only one herring in 2,558 nautical miles and 35 midwater trawl hauls. These results indicate that only a small number of herring may remain or return offshore in summer; most remain in coastal waters. Rumyantsev and Darda (1970) concluded that herring remained in coastal waters during the summer because heavy phytoplankton blooms (1-3 g/m^ ) and poor feeding conditions exist on the outer shelf. Herring captured on the outer shelf during the summer were in poor condition, perhaps because they had been feeding on items of low nutritional value— a diet other than their preferred 518 Fisheries oceanography zooplankton. Other researchers have found herring to avoid areas of heavy bloom because of low nutri- tional value of phytoplankton and because the gill-clogging properties of certain phytoplankton species interfere with respiration (Henderson et al. 1936). Concentrations began reappearing in offshore waters in the areas of Nunivak and Unimak islands in August (Rumyantsev and Darda 1970). The move- ment offshore in the area of Unimak Island appears to occur annually in August, for it is then that U.S. fishery observers on foreign vessels in that area first encounter herring in trawl catches in greater than trace amounts. The distribution of herring between the time they leave the spawning grounds and the time they reap- pear in offshore waters is unkno^m, but salmon fishermen report catching large herring frequently in salmon gUlnets in coastal areas of Bristol Bay in late June and July. Furthermore, Dudnik and Usoltsev (1964), using drift nets, found commercial quantities of herring only in littoral areas along the northern portion of the Alaska Peninsula. The reappearance of seaward migrants in late summer in two locations suggests a summer migration along the coast (Fig. 32-9). Migration to winter grounds continues through September with the herring progressively moving to deeper water and concentrating in the 2-4 C temperature stratum. Fall Concentration in the winter grounds begins in October and continues into vdnter. Mature fish were found to arrive at the wintering grounds before immature fish (Rumyantsev and Darda 1970). Im- mature fish had a tolerance or preference for colder, less saline waters of the shelf than adult fish (Fig. 32-10). Seasonal distribution— north of Norton Sound The annual cycle of herring to the north of Norton Sound appears to be markedly different from that in the central and southern Bering Sea. In these areas herring appear to move into brackish bays and estuaries for spawning and wintering, presumably finding temperatures rendered suitable by freshwater rivers and streams. Barton (1978) found herring spawning in Imuruk Basin, a brackish forebay of Port Clarence, in 4°/oo salinity. The herring dispersed after spawning and reappeared in Imuruk Basin in mid-autumn. Large numbers of herring were found concentrated just outside Kotzebue Sound in 6-8 C water in September and October (Wolotira et al. 1977). It is SUMMER/AUTUMN HERRING MIGRATION C* Soviet data ^ Japanese data and Soviet -.- Suspected summer migration ■-• ^- ■'■» •' "^ Figure 32-9. Summer and autumn migration routes to winter grounds. Large solid arrow: area of reappearance in offshore waters as determined by Soviet research and Japanese catches. Large clear arrow: area of autumn reappearance in offshore waters reported from Soviet research. Small arrows: possible summer feeding routes and autumn migration routes. possible that these herring move south vnth the advancing edge of the ice field, but some evidence suggests that they remain in the eirea through the winter. The phenomenon of herring moving into coastal waters in winter and offshore in summer has also been reported from Asia; Andriyashev (1954) reported that populations occur in Kamchatka, Sakhalin, and Honshu which winter and spawn in brackish lakes and lagoons. Barton (1978) cites ADF&G records of herring being found in sheefish (Stenodus leucichthus) stomachs which were col- lected in Hotham Inlet, Kotzebue Sound, in late November. To the south of Kotzebue, on the northwest side of the Seward Peninsula, herring occurred in 11 of 14 stomachs of spotted seals (Phoca largha) collected at Shishmaref in October, and in 13 of 30 ringed seals collected in January and February. Marine mammal biologists who analyzed the stomachs indicated there was little doubt that herring were ingested in the area where seals were captured. ABUNDANCE TRENDS AND STOCK STATUS Estimates of absolute abundance are rare and even relative abundance data are rather limited. Attempts Pacific herring 51 9 I Figure 32-10. Distribution of Pacific herring in October and the relationship of adults (mature) and juveniles (immature) to salinity gradients. (Modified from Rumyant- sevand Darda 1970.) have been made to estimate herring biomass by a Soviet hydroacoustic trawl survey, ecosystem model- ing, and aerial surveys of spawning biomass. In 1963, three years after the fishery began, the eastern Bering Sea herring biomass was estimated to be 2.16 million mt, based on a Soviet hydroacoustic survey of the wintering grounds (Shaboneev 1965). A recent paper by Kachina (1978), using the same data, reduced this earlier estimate to 0.374 million mt by using a lower mean school density— 0.5 fish/m^ — than the original estimate, which used 3.38 fish/m^ . According to Shaboneev, schools were surveyed at night and the area and height of schools were mapped acoustically; school composition and age distribution were determined by trawling. The density used in the original estimate (3.38 fish/m^ ) was determined by comparing acoustic echograms from the eastern Bering Sea to echograms of schools sampled by purse seines in western Bering Sea coastal water. The revised estimate of 0.5 fish/m^ is based on densities observed in surveys of herring concentra- tions in the winter grounds northwest of the Pribilofs during 1969-71 (N. Fadeev, TINRO, Vladivostok, U.S.S.R., personal communication). The densities derived are questionable but cannot be fully eval- uated because few specific details of Soviet survey methods and accuracy are available. However, data reported in the literature and from people involved with herring hydroacoustic surveys indicate that the range of densities used by the Soviets may be ex- treme; an intermediate value may be more realistic. Recently, a numerical ecosystem model was applied to estimate the biomass of eastern Bering Sea herring (Laevastu and Favorite 1978). This model based herring abundance on the amount needed to sustain the diet of herring predators at reported rates of consumption. The accuracy of input parameters, such as predator population size and consumption rates, has not been sufficiently evaluated, but model results show that a minimum stock size of 2.75 million mt of herring is required to maintain com- ponents of the ecosystem, including predators, at a level observed in the mid-1 960 's before the start of intensive fishing. Aerial surveys have been flown in the past several years along the western Alaska coast during the spawning period to record the number of schools by surface area (Barton 1979). Estimates of biomass were obtained by converting estimated school surface area, using densities of 0.1 and 0.2 mt/m^ of surface area. The estimated spawning biomass along the coast from Bristol Bay to Norton Sound in 1978 was 432-864 thousand mt. These estimates include various errors— in determining surface areas and volumes of schools, in recording schools of other fish such as capelin, smelt, or sand lance as herring, and in recording the same school more than once during the season. They may therefore greatly overestimate actual spawning biomass (Barton 1979). Biomass estimates are rather rudimentary, but CPUE data of the Japanese trawl fishery and ADF&G aerial surveys indicate that herring abundance de- clined sharply in the early 1970 's and increased in the late 1970 's. The CPUE (mtAir) for Japanese large stern trawlers decreased from a high of 6.80 in 1969-70 to 0.77 in 1973-74 (Table 32-3). The CPUE of small stern trawlers also declined. The CPUE of the Japanese gillnet fishery exhibited no trend, presumably because vessels were targeting on spawn- ing concentrations, which may not reflect population abundance (Wespestad 1978b). ADF&G aerial surveys have indicated an increase in herring abundance in all major spawning areas during 1976-78 (Table 32-4). Preliminary assessment of observations in 1979 indicates an abundance similar to that of 1978, or slightly greater. The longest series of aerial counts, from southern Norton Sound, extends back to 1968; like the trawl CPUE data, it indicates a decrease in abundance during the early 1970's. Because of changes in the fishery, the current level of herring abundance cannot be related to former 520 Fisheries oceanography TABLE 32-3 Herring catch per unit effort data for the Japanese trawl and gillnet fisheries in the eastern Bering Sea. Stern trawl Gillnet Small trawlers Larg e trawlers Year (mt/hour) (n- it/hour) (mt/10 tons) 1967 1.25 2.09 _. 1968 1.75 4.63 .28 1969 0.81 6.80 .39 1970 1.06 6.74 .24 1971 0.56 1.52 .34 1972 _. 1.84 .04 1973 - 0.77 .14 1974 0.29 0.17 .35 1975 . -- -- .16 1976 -- -- " levels ; the catch and CPUE of foreign trawlers are no longer useful as indicators of abundance. Because of increased targeting on pollock, herring are not largely incidental catches to other fisheries, and the allowa- ble catches have been low in recent years (Wespestad 1978b). Furthermore, the absence of aerial surveys in major spawning areas before 1976 precludes the use of this method for determining the relationship of past to present biomass levels. Length and age frequency data indicate that catches in the late 1960's and early 1970's were composed of larger and older herring than in the past few years (Table 32-5). These data suggest that recruitment was poor until recently, a fact which may have contributed greatly to a lower herring abun- dance. Recruitment appears to have increased beginning with the 1972 year-class (Fig. 32-11). Age-four herring comprised 54 percent of the catch in 1976, 50 percent in 1977, and in 1978, 65 percent of the purse seine catch. In 1979, preliminary results suggest that the recruitment of age-four herring has decreased from that observed in 1976-78; however, it appears that age-three fish are present in greater numbers than in the recent past. Naumenko (1979) recently presented data showing several years of relatively weak year-class strength in the 1960's and early 1970's (Fig. 32-12). These data suggest that the peak catches of the fishery were sustained by a few strong year-classes and that future yields of this magnitude are likely only in the event of a series of much above average year-classes, or of a change in the ecosystem. 60 40 20 1976 n r-i 40 r- 1977 20 - 0 U — ■ 1 ' rj 1 ^-^ 1 171 r7 [71 2 3 4 6 7 60 40 20 - 1978 4 A G Figure 32-11. Age frequency of Pacific herring in the Bristol Bay herring roe fishery, 1976-78. TABLE 32-4 Relative abundance indices of spawning herring standardized to 1976 in major spawning areas of the eastern Bering Sea/ Pacific herring 521 1968 1972 1974 1975 1976 1977 1978 Bristol Bay Goodnews Bay/Security Cove Nelson Island Norton Sound: St. Michaels to Unalakleet 20.8 8.4 2.9 -- 1.0 2.1 20.0 9.5 1.0 20.9 61.6 0.5 1.0 1.0 3.2 0.0' 1.0 4.2 * Relative abundance indices are corrected school counts weighted by surface area obtained from aerial surveys. ■^Minimal survey effort. RESEARCH REQUIREMENTS Research is needed to refine estimates of abun- dance and define biological characteristics of stocks; to improve the ability to predict changes in resource abundance, composition, and availability; and to identify the origin and distribution of herring in offshore areas. Estimates of biomass of specific groundfish re- sources have been obtained through resource surveys using bottom trawls; but since herring are not gener- ally available to bottom trawls, other gear and meth- ods must be used to assess biomass. Hydroacoustic surveys, spawn deposition surveys, and aerial surveys of schooled fish are under consideration for this purpose. TABLE 32-5 Mean length of herring taken in the fisheries by all gear in all months in the eastern Bering Sea and Alaska coastal water. ^ Foreign trawl fishery Coastal fishery Mean Probable Mean Length Sample Average Length Sample Location of Year (cm) size ages (cm) size sample 1964 26.60 3,101 7 23.30 339 Norton Sound 1965 29.83 155 8-9 1966 27.16 48 6-7 1967 26.20 99 5-6 1968 29.04 4,771 8-9 28.60 350 Bristol Bay 1969 30.66 3,951 9-10 1970 30.81 3,813 9-10 1971 29.21 4,299 8-9 1974 1975 1976 3-4 20.11 791 Bristol Bay 1977 23.40^ 1,981 4-5 23.00 2,847 Bristol Bay 1978 23.28^ 3,607 4-5-6 23.27 1,031 Bristol Bay ' standard length for all coastal samples; fork length for foreign samples prior to 1978. ^Fork length (Nov. 1976-Feb. 1977) estimated standard length is 22.40 m. ^ Standard length (Dec. 1977-Jan. 1978). Sources: Foreign fishery: Fisheries Agency of Japan, Rumyantsev and Darda (1970). U.S. observers on Japanese and Soviet vessels. Coastal fishery: Alaska Dept. of Fish & Game, 1964 Annual Report; Bristol Bay Data Report No. 17; Barton et al. (1977). Warner & Shafford (1977). 522 Fisheries oceanography 100 1950 1955 I960 1965 Year-class 1970 1975 Figure 32-12. Abundance of Pacific herring year-classes in the eastern Bering Sea relative to the 1975 year-class. (From Naumenko 1979). Hydroacoustic surveys in the nearshore areas just before or during spawning are probably not practical because of the many widely scattered schools that are constantly moving through the shallow waters. Optimum results can be expected on the winter grounds, when herring are relatively stationary and concentrated. Results of surveys conducted during late winter to early spring could be applied in time for management of the roe fisheries. Spawn surveys convert the amount of spawn deposited to size of adult population, using age-sex- size composition and fecundity data. Such surveys would have to be conducted immediately after spawning so as not to be affected by losses from predation and storms. The vast size of the area, including distances between spawning areas, lack of subtidal spawning information, and various logistical problems currently render this method impractical for the eastern Bering Sea. In spite of limitations due to weather and narrow time-area coverage, aerial surveys may be one of the most cost-effective ways of measuring the abundance of spawning herring. School distribution within a limited area should be intensively studied to deter- mine if surveys are more effective at particular times and to assess the variability of schools along sighting tracks. Aerial biomass estimation procedures and species identification procedures need to be developed. If a model of spawning school distribu- tion could be developed, then statistical procedures could be used to overcome some of the weather and time limitations. Satellite technology may be used to augment aerial surveysMarge schools may be ob- servable at distances from the coast or spawn deposi- tion (milt) may be observable. A combination of low-level aircraft and satellite observations may provide solutions to the problems of effective cover- age of tracklines and time-space distribution of schools. Long-term fisheries management requires reliable forecasting of stock conditions. Until now, forecasts have been based mainly on past events, such as trends in abundance indices (CPUE) and size and age com- position of specific resources, without any considera- tion of the interactions of these resources among themselves and with the environment. Studies need to be continued to determine, for predictive pur- poses, major influences on the abundance, composi- tion, and distribution of resources. Monitoring certain oceanographic and climatological conditions (temperature, currents, etc.) in both the nearshore spawning and rearing grounds and the offshore wintering grounds may help us to understand fluctua- tions in herring abundance. There is a critical need for annueil measurements of the abundance of young fish before they enter the fisheries, in order to forecast later contribution to the exploitable stock. Assessment of pre-recruit abun- dance could be made of juveniles in nearshore nursery areas or at a later age in offshore waters. The major limitation of this method is the virtual absence of information about the distribution of eastern Bering Sea herring during the first two to three years of their life cycle. Basic biological research is needed to systematic- ally investigate population parameters, such as age-specific mortality rates, growth rates, and re- cruitment rates. Investigations are also needed to establish the degree of use of herring in the diet of marine mammals, salmon, and other predators so as to evaluate the ecological effects of harvesting. Finally, stock distribution needs to be investigated so that individual stocks within the eastern Bering Sea can be monitored in relationship to other stocks and occurrence in fisheries. REFERENCES Alderdice, D. F., and F. P. J. Velsen 1971 Some effects of salinity and tem- perature on early development of Pacific herring (Clupea pallasi). J. Fish. Res. Bd. Can. 18:1545-62. Pacific herring 523 Alverson, D. L., and M. J. Carney 1975 A graphic review of the growth and decay of popiilation cohorts. J. Cons. Int. Explor. Mers. 36(2): 133-43. Andriyashev, A. P. 1954 Fish of northern Soviet Seas. Izdatelstvo Akad. Nauk, S.S.S.R., Moscow. (Transl. 1964 Israel Prog. Sci. Transl.) Bakkala, R. G., and G. B. Smith 1978 Demersal fish resources of the eastern Bering Sea: spring 1976. U.S. Dep. Comm. Nat. Mar. Fish. Serv., North- west and Alaska Fish. Cent., Seattle, Wash., Proc. Rep. Grant, S. 1979 Biochemical genetic variation among populations of Bering Sea and North Pacific herring. Nat. Mar. Fish. Serv., Northwest and Alaska Fish. Cent., Seattle, Wash., Proc. Rep. Hart, J. L. 1973 Pacific fishes of Canada. Bd. Can. Bull. 180. Fish. Res. Hemming, J. E., G. S. Harrison, and S. R. Braund 1978 The social and economic impacts of a commercial herring fishery on the coastal villages of the Arctic /Yukon/ Kuskokwim area. Unpub. rep. Barton, L. H. 1978 Finfish resource surveys in Norton Sound and Kotzebue Sound. OCSEAP, Final Rep. (March 1976- September 1978), ADF&G, Comm. Fish. Div., Anchorage. 1979 Assessment of spawning herring and capelin stocks at selected coastal areas in the eastern Bering Sea. Ann. Rep. to N. Pac. Fishery Manage- ment Council, Contract 78-5, ADF&G, Comm. Fish. Div., Anchor- age. Barton, L. H., I. M. Warner, and P. Shafford 1977 Herring spawning surveys— southern Bering Sea. Alaska Mar. Environ- mental Assessment Project, Project Rep. Alaska Dep. of Fish and Game, Unpub. MS. Dudnik, Y. I., and E. A. Usoltsev 1964 The herring of the eastern part of the Bering Sea. In: Soviet fisheries investigations in the northeast Pacific. P. A. Moiseev, ed.. Part II: 225-9. (Transl. 1968. Israel Prog. Sci. Transl.) Fujita, T., and S. Kokudo 1927 Studies on herring. Bull. Sch. Fish. Hokkaido Imp. Univ. 1(1): 1-127. Gilmer, T. 1978 Cape Romanzof herring project. May 22-June 20, 1978. Prelim. Rep., ADF&G, Comm. Fish. Div., Anchor- age, unpub. Henderson, G. T. D., C. E. Lucas, and J. H. Eraser 1936 Ecological relations between the her- ring and the plankton investigated with the plankton indicator. J. Mar. Biol. Assoc. U.K. 21(1): 277-91. Hourston, A. S. 1959 Effects of some aspects of environ- ment on the distribution of juvenile herring in Barkley Sound. J. Fish. Res. Bd. Can. 16:283-308. Kachina, T. 1978 The status of the eastern Bering Sea herring stocks. Pac. Sci. Inst. Fish. Oceanogr. (TINRO) Vladivostok, U.S.S.R. Rep. submitted at U.S.- U.S.S.R. scientific meetings, Dec. 20-24, Northwest and Alaska Fisheries Center, Seattle, Wash, Kachina, T. F., and R. Ya. Akimova 1972 The biology of the Korfo-Karaginski herring in the first year of life. Izv. Tikhookean. Nauchnoissled. Inst. Rybn. Khoz. Okeanogr. 82: 309-20. In Russian, Eng. abstract. Kaganovskii, A. G. 1955 Basic traits of behavior of pelagic fishes and methods of scouting and forecasting them in Far Eastern waters. Akad. Nauk S.S.S.R., Tr. Soveshch. Ikhtiol. Kom. 5 (Trudy soveshchaniya po voprosam pove- deniya i razvedki ryb 1953): 26-33. In Russian. (Transl. Nat. Mar. Fish. Serv., Biol. Lab., Honolulu, Hawaii.) 524 Fisheries oceanography Laevastu, T., and F. Favorite 1977 Ecosystem model estimations of the distribution of biomass and predation with age for five species in eastern Bering Sea. Nat. Mar. Fish. Serv., Northwest and Alaska Fish. Cent., Seattle, Wash., Proc. Rep. 1978 Fluctuations in Pacific herring stock in the eastern Bering Sea as revealed by ecosystem model (DYNUMES III). I.C.E.S. Symposium on Biological Basis of Pelagic Fish Stock Manage- ment No. 31. Rabin, D. 1977 Marsh, M. C, and J. N. Cobb 1910 The fisheries of Alaska in 1909. Bur. Fish., Rep. Comm. Fish. 730. U.S. Doc. 1911 Nagasaki, F. 1958 The fisheries of Alaska in 1910. U.S. Bur. Fish., Rep. Comm. Fish. Doc. 746. The fecundity of Pacific herring (Clupea pallasi) in British Columbia coastal waters. J. Fish. Res. Bd. Can. 15: 313-30. Naumenko, N. I. 1979 The state of stocks of the eastern Bering Sea herring. Pac. Res. Inst. Fish. Oceanography (TINRO), Vladi- vostok, U.S.S.R. Unpub. rep. Outram, D. N., and R. D. Humphreys 1974 The Pacific herring in British Colum- bia waters. Fish. Mar. Serv. Can. Pac. Biol. Sta., Nanaimo, B.C., Circ. 100. Pereyra, W., J. Reeves, and R. Bakkala 1976 Demersal fish and shellfish resources of the eastern Bering Sea in the baseline year 1975. Data appendices. Nat. Mar. Fish. Serv., Northwest and Alaska Fish. Cent., Seattle, Wash., Proc. Rep. Prokhorov, V. G. 1968 Winter period of life of herring in the Bering Sea. Proc. Pac. Sci. Res. Inst. Fish. Oceanogr. 64:329-38. (Transl. 1970, Fish. Res. Bd. Can., Transl. Ser. 1433.) Status of fisheries in California. In: Proceedings of third Pacific Coast Herring Workshop, June 22-23, 1976. D. Blankenbeckler, ed.. Fish. Res. Bd. Can. MS Rep. Ser. No. 1421. Ricker, W. E. 1975 Computation and interpretation of biological statistics of fish popula- tions. Fish. Res. Bd. Can. Bull. 191. Rounsefell, G. A. 1930 Contribution to the biology of the Pacific herring, Clupea pallasi, and the condition of the fishery in Alaska. Bull. U.S. Bur. Fish. 45: 227-320. Rumyantsev, A. I., and M. A. Darda 1970 Summer herring in the eastern Bering Sea. In: Soviet fisheries investigations in the northeastern Pacific, P. A. Moiseev, ed.. Part V: 409-41. (Transl. 1972, Israel Prog. Sci. Transl.) Scattergood, L. W., C. J. Sindermann, and R. E. Skud 1959 Spawning of North American herring. Trans. Amer. Fish. Soc. 88(3): 164-8. Shaboneev, I. E. 1965 Biology and fishing of herring in the eastern part of the Bering Sea. In: Soviet fisheries investigations in the northeastern Pacific, P. A. Moiseev, ed.. Part IV: 130-54. (Transl. 1968. Israel Prog. Sci. Transl.) Skud, R. E. 1963 Herring tagging experiments in south- eastern Alaska. U.S. Fish Wildl. Serv., Fish. Bull. 63 (l):19-32. Stevenson, J. C. 1962 Distribution and survival of herring larvae (Clupea pallasi Valenciennes) in British Columbia waters. J. Fish. Res. Bd. Can. 19:735-810. Svetovidov, A. N. 1952 Clupeidae, Fauna of U.S.S.R., Fishes II, 1. Acad. Sci. U.S.S.R., Moscow. Taylor, F. H. C. 1964 Life history and present status of British Columbia herring stocks. Fish. Res. Bd. Can. Bull. 143. Pacific herring 525 Tester, A. L. 1955 Estimation of recruitment and natural mortality rate from age composition and catch data in British Columbia herring populations. J. Fish. Res. Bd. Can. 12:649-81. Wespestad, V. G. 1978a A review of Pacific herring studies with special reference to the Bering Sea. Nat. Mar. Fish. Serv., Northwest and Alaska Fish. Cent., Seattle, Wash. Unpub. MS. Warner, I. M. 1976 Forage fish spawning surveys, Unimak Pass to Ugashik River. Outer Conti- nental Shelf Environmental Assess- ment Program, Quart. Rep. (July- Sept.) 61-96. Warner, I. M., and P. Shafford 1977 Forage fish spawning surveys- southern Bering Sea. Alaska Marine Environmental Assessment Project, Project Completion Rep. 1978b Exploitation, distribution, and life history features of Pacific herring in the Bering Sea. Nat. Mar. Fish. Serv., Northwest and Alaska Fish. Cent., Seattle, Wash., Proc. Rep. Wolotira, R. J., T. M. Sample, and M. Morin 1977 Demersal fish and shellfish resources of Norton Sound, the southeastern Chukchi Sea and adjacent waters in the baseline year 1976. Nat. Mar. Fish. Serv., Northwest and Alaska Fish. Cent., Seattle, Wash., Proc. Rep. The Biology of Walleye Pollock Gary B. Smith Northwest and Alaska Fisheries Center Seattle, Washington ABSTRACT Walleye pollock, Theragra chalco gramma, occur broadly distributed over the eastern Bering Sea outer continental shelf, and at their presently estimated population size of five million mt are a large source of organic production. Within the food web pollock are an important food resource for a wide variety of other fish species, marine mammals, and avifauna. Pollock also represent a major source of predation directed toward zooplankton and cannibalistic behavior. In addition, trawl fisheries harvest approximately 950,000 mt of pollock an- nually. Although the genetic identity of eastern Bering Sea pollock is apparently distinct from populations in the western Pacific, genetic differentiation observed within the eastern Bering Sea is low, and the entire eastern Bering Sea population may be regarded as effectively a unit stock. The population is com- posed of approximately fifteen year-classes that show differ- ences in geographical distribution and behavior based upon age. The overall rate of population mortality is high, approx- imately 51 percent dying annually, and age-class abundance is variable between years. Spawning occurs predominantly along the southeastern Bering Sea outer shelf, just west and northwest of Unimak Island. Migratory movements that accompany ontogeny include northwest drift and broad dispersal by age one. In general, the adult fraction of the population undergoes season- al movements to deep water in wdnter, to spawning sites in spring, then to the outer and central shelf in summer. Two apparent influences of ocean climate, warm or cold years, affect the timing and extent of these seasonal changes in geographical range. There are fundamental needs for better understanding of important factors affecting the population dynamics of eastern Bering Sea pollock, to ensure collection of adequate time series data, and further studies of interspecific relationships and population exchange. INTRODUCTION The gadoid fish species walleye pollock, Theragra chalcogramma (Pallas 1811), is now recognized as perhaps one of the most important components of the Bering Sea biological system. Pollock have been found to represent a large fraction (ca. 20-50 percent: Pereyra et al. 1976) of the total standing stock of eastern Bering Sea demersal fishes, and the annual production of this organic pool is large. In 1978, commercial trawl fisheries operating in the eastern Bering Sea harvested 977,700 mt of pollock (Pileggi and Thompson 1979). This amount represented 56 percent of all foreign fish catches in the fishery conservation zone (i.e., inside the 200-mile limit) of the United States, and a total dockside (ex-vessel) value of approximately $92.4 million. Because of their large standing stock, wide geo- graphical distribution, and large rates of total food intake and production (growth), pollock represent an important part of the Bering Sea food web. During different life history stages pollock can be a major source of predation, feeding upon a broad spectrum of primarily zooplankton and fish prey. In turn, pollock themselves serve as an important food source for a variety of other demersal and pelagic fish species, pinnipeds, cetaceans, and avi- fauna. POPULATION CHARACTERISTICS Nomenclature The walleye pollock, T. chalcogramma, is the only recognized member of the genus Theragra and is endemic to the North Pacific (Fig. 33-1). The overall species range extends continuously around the continental shelf areas of the northern Pacific Ocean and Bering Sea. In the northwest Pacific the species extends from the Sea of Japan, through the Okhotsk Sea and Kurile Islands, and along the Kamchatka Peninsula and Anadyr Gulf. In the northeast Pacific the range extends from St. Lawrence Island through the eastern Bering Sea, through the Gulf of Alaska, and along the northeastern Pacific coast to central California (Hart 1973). Pelagic populations have also been recently found over deep water in the central Bering Sea (Okada 1977, 1978). 527 528 Fisheries oceanography Figure 33-1. The worldwide distribution of walleye pollock, Theragra chalcogramma (shaded area). In the eastern Bering Sea, T. chalcogramma is by far the most abundant of the four common gadid species. These other species include Gadus macro- cephalus (Pacific cod), Boreogadus saida (Arctic cod), and Eleginus gracilis (saffron cod). Genetic structure Although pollock are distributed (and perhaps spawn essentially continuously) along the northern rim of the Pacific Ocean and Bering Sea, genetic differentiation might reasonably be expected due to the isolating effects of both geographical distance and barriers (Kimura and Ohta 1971). This is because individual migrations will nearly always be consider- ably less than the total distributional range of the species, and instead, effectively form regional breed- ing populations. Similarly, diffusive exchange of eggs and larvae due to advective transport may be expect- ed to be higher within regional current systems than between, although progressive transport of successive generations may be significant. Consistent with these hypotheses, recent studies of biochemical genetic variation have found relatively large genetic differences between pollock sampled from extremes of the species range, but relatively small genetic differentiation within regions (Iwata 1973, 1975, 1977; Johnson 1977). Based upon differences in allelic frequencies for the protein locus tetrazolium oxidase. Grant et al. (1978) con- cluded that the widely separated populations of the western and eastern Pacific are relatively genetically isolated and distinct. Only weak regional differentia- tion was observed, however, among pollock collected from north and south areas of the eastern Bering Sea and from the Gulf of Alaska. Other investigators have proposed a more complex population structure in the Bering Sea. From anal- yses of morphometric characters, Serobaba (1977) recognized four principal Bering Sea pollock popula- tions: southeastern, northern, western, and Aleutian Island. Maeda (1972) proposed two distinct eastern populations based upon apparent migrations to separate spawning areas northwest and southeast of the Pribilof Islands. After comparisons of the age structure of commercial catches in northwest and southeast regions from 1963 to 1970, however, Chang (1974) concluded that the pollock population exploited by fisheries in the eastern Bering Sea is a unit stock. Description of habitat In the eastern Bering Sea, pollock occur widely distributed over the continental shelf, but show Walleye pollock 529 180' ne'\ ieo° 17S' Figure 33-2. The apparent density distribution of walleye pollock during August-October 1975, determined by a research vessel survey (Pereyra et al. 1976). Figure 33-3. The apparent density distribution of walleye pollock during April- June 1976, determined by a research vessel survey (Bakkala and Smith 1978). highest densities along the shelf edge (Figs. 33-2 and 33-3). The overall distribution pattern varies season- ally and between years, dominated by movements between deep and shallow water. Over the contin- ental shelf pollock show a semidemersal behavior, tending to form schools near the bottom during daytime, then dispersing up into the water column at night. Along the outer continental shelf and slope, dense shoals are formed that may exceed 20-50 km in length and have mean densities of up to 9 mt per hectare. Figs. 33-4 and 33-5 show two examples of the apparent depth distribution of pollock on the eastern Bering Sea continental shelf, determined from re- search vessel surveys. During August-October 1975, pollock were relatively symmetrically distributed around a median depth of approximately 200 m. During April -June 1976, the vertical pattern was less symmetrical but similar, although apparent densities were lower among all depth intervals except 400-450 m. The domain within which pollock occur demersally Apparent Density (kg/ha) 50 100 150 200 100 -g 200 a Q 300 400 1 ' ' 1 ' III sflii MEAN POLLOCK ABUNDANCE August-October 1975 500 Figure 33-4. Apparent vertical distribution of pollock densities on the continental shelf during August-October 1975 (based upon the data shown in Fig. 33-2). Apparent Density (kg/ha) 50 100 150 100 E 200 Q. V Q 300 400 500 200 — I MEAN POLLOCK ABUNDANCE Aprif-June 1976 Figure 33-5. Apparent vertical distribution of pollock densities on the continental shelf during April-June 1976 (based on the data shown in Fig. 33-3). 530 Fisheries oceanography Figure 33-6. Temperature and salinity ciiaracteristics of the environment of walleye pollock in the eastern Bering Sea (shaded area). Pollock T-S data replotted from Kihara and Uda (1969); North Pacific Ocean water mass data are from Sverdrup et al. (1946). in the eastern Bering Sea, then, is a broad section of the outer continental shelf and slope primarily within the 100-300 m depth range. The bottom morphology of the continental shelf in most of this region is featureless and level, with sand and silt surface sediments (Sharma 1974). In deeper water, the continental margin is steep and cut by large sub- marine canyons that substantially affect the flow and mixing of water currents along the shelf edge (Kinder etal. 1975). In comparison to temperature and salinity charac- teristics of water masses in other regions of the North Pacific Ocean, pollock occur within a relatively extreme seawater environment in the eastern Bering Sea (Fig. 33-6). These distinctive conditions— cold and low salinity— result from mixing of shelf and oceanic waters within apparently broad zones of interaction along and over the eastern continental shelf (Takenouti and Ohtani 1974, Coachman and Charnell 1979). Source waters, mixing characteris- tics, and associated hydrographic features must all be important in determining the composition of both pollock prey and predators. Stock size Seven well-documented estimates of the absolute Bering Sea pollock are A confusing feature of bulk biomass of eastern summarized in Table 33-1 these results is the use of different, and sometimes unspecified, geographical boundaries among the different analyses. In addition, because all of the estimates are based upon selective sampling (namely, demersal trawl nets, and in the case of estimates based upon commercial data, targeted fisheries), the relationships between statistical populations included in the analyses and true field populations are unclear. Management policies for eastern Bering Sea trawl fisheries assume a present pollock stock size of approximately 5 million mt (NPFMC 1978). Table 33-2 and Fig. 33-7 show the observed variability and trends for five indices of the relative annual abundance (biomass) of pollock in the eastern Bering Sea. During the sixteen years of observations 1963 to 1978, apparent pollock densities have varied approximately ± 50 percent of the long-term mean values. Rather than showing large random year-to- year differences, most variations in index values seem to have represented longer-term trends. Judging by the longest time series (Bakkala et al. 1979b), pollock are relatively less abundant now in the eastern Bering Sea region than during the period 1965 to 1970. The research survey data shown (see Table 33-2) as NMFS Crab-Groundfish are results determined from demersal trawl surveys conducted by the U.S. Na- tional Marine Fisheries Service (NMFS). A central 159,100 km^ core area (Fig. 33-8) has been surveyed Walleye pollock 531 TABLE 33-1 Summary of estimates of absolute population size for eastern Bering Sea walleye pollock. Source Region and time period* Method^ Estimated population (X 10^ mt) A. Based upon research survey data Pereyra et al. 1976 Bakkala and Smith 1978 Okada 1978;Nunnallee 1978 B. Based upon commercial fisheries data Chang 1974 Chang 1974 Low 1974 C. Based upon model estimates Laevastu and Favorite 1977 Eastern Bering Sea shelf, Unimak Pass to 61° N (August-October 1975) Eastern Bering Sea shelf, Unimak Pass to 59° N (April-June 1976) Aleutian Basin (June-July 1978) Eastern Bering Sea shelf, INPFC areas 1 and 2 (1969-1970) Eastern Bering Sea shelf, INPFC areas 1 and 2 (1970) Eastern Bering Sea, primarily INPFC areas 1 and 2 (1964-1971) Eastern Bering Sea shelf 2.426 0.679 0.840 2.3-2.6 2.3-2.4 3.45-5.83 8.235 * A description of INPFC (International North Pacific Fisheries Commission) statistical areas is given in Forrester et al. (1978). ■^Estimation methods: 1 = "area swept" (Baranov 1918; Alverson and Pereyra 1969); 2 = "cohort analysis" (Pope 1972), 3 = "model fitting," based upon commercial fisheries data. Research Survey Data (kg/ha trawled) Commercial Fistieries Data (mt/tir pair trawl) 1966 1968 1970 1972 Year 1974 1976 Figure 33-7. Annual variations in apparent relative abun- dance, based upon the indices of Pereyra et al. (1976), updated to 1978, and Bakkala et al. (1979b) summarized in Table 33-2. i ! 1/" *-'';A|^^ M U*^ »: ^ [ °'""" ^ using uniform methods during approximate- ly June 1 to August 15 of each year (Pereyra et aL 1976: Section VIII). Figure 33-8. NMFS index area (shaded) for annual eastern Bering Sea crab and demersal fish population assessment surveys. Dots indicate routine sampling locations. 00 05 iH c~ t> 05 ^ 1-1 (30 C~ CD 05 05 T-t O -w CO lO iD t- 05 05 tH tH sa 0) ^ en t- bUD o> C T-H •^^ vi CO CO CO W 05 o tH o o 0) 05 >^ tH ^ CS o> ^ CO Cl-I 05 O T-l a> % (X) CO 05 S 1-1 3 CD o 05 a s s 05 CD CD 05 CD 05 CD 05 CO CD 05 O eo_ O o CO t-' CO CD CO LO 00 CO CO CO 05 C3 T3 C 3 O X! (J3 <; 2 . a 5 > >. c £ .2 >> .. 0) CD 3 2 00 2 0> cS 1-1 ^ O 'OD PQ c- .5 w 1— I g >- C CJ C«-l W' -^ O) U3 S ~ "5 O § ^ « 1-3 CO 3 05 CD 05 o O o 00 CM 05 o co' CD 05 1—1 O d iH O0_ C~' d ^. LO_ ^' ^. oi 1 CD 00 1 ■*. oi 1 CO oi 1 '*. O 1— ( 1 CD 00 1 cq 1 eg CM 05 r-i ^' tr-_ 1-1 00 iH lO LO 05 1—1 CO ^' 00 1-1 00 CO CO CO CO oo_ l-I LO CD CD CO Csl co' CO CM o 00 '^' LO oi co' CM d ^. cm' 1-1 00 oo' 05 CO 05 CD 00 ^-C^ c a Oi B O o o rH -1^ a _o ^4^ 1-1 C« 05 V a> T3 73 CM ki O T3 OI •C ^ .ti -^ J- CO 3 >-i ^ « "t «♦-( ^ ' ^ ^ %4 CD w ;^ 'c3 S .H ?! O. C -g 05 _£ 1-3 « ca c c/3 05 c/3 2 1-1 C^ oj g C ^ x: o *= -c U2 -1 c/1 a> « 1— I cu CO .— ; a. CO 1-3 C3 OI c CO a eo ►-3 CO CQ 6§ f— « o ■o a> II ^ 4^ 0) CO T3 CO U3 2 'c 3 ^— ^ M 'c >-i Ut U5 CJ O 3 ex U2 t^' 05" CD s: -!3 CO 4^ 05 05 05 B CO > c CO XI CO 532 Walleye pollock 533 TABLE 33-3 Indices of age-class abundance measured by NMFS Crab-Groundfish research vessel surveys within a central area of the eastern Bering Sea during June to mid-August, 1973-78 (units = 10^ individuals per 159,100 km^).^ Survey Number of Number of Age-classes (yr) year trawls otoliths 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 1973 111 490 756.9 518.2 146.0 193.4 99.4 34.4 91.5 117.8 18.9 17.9 4.51 0.401 0.072 - 1974 111 954 2840.6 850.3 287.9 56.1 67.7 38.0 39.1 29.0 29.5 5.12 3.00 0.176 0.190 — — 1975 111 766 758.4 402.4 614.2 108.0 24.6 27.5 14.7 14.2 7.98 5.99 0.52 0.594 0.009 — — 1976 107 1990 729.0 500.4 479.7 1014.1 132.5 35.0 38.8 46.2 41.0 22.3 8.28 1.850 0.607 0.032 0.171 1977 112 944 2241.9 630.2 145.8 245.3 231.9 72.0 29.8 23.7 23.6 13.1 11.9 3.200 0.491 0.180 -- 1978 116 1256 1170.9 400.4 806.9 507.0 139.5 92.3 29.1 24.2 29.1 19.0 6.26 5.100 — — — OveraU mean (1973-78) 1416.3 550.3 413.4 354.0 115.9 49.9 40.5 42.5 25.0 13.9 5.75 1.82 0.280 0.050 0.030 Standard deviation (1973-78) 906.7 169.9 267.7 359.4 71.0 26.0 26.5 38.4 11.2 7.11 4.02 2.01 0.260 0.070 0.070 Coefficient of variation (1973-78) 0.64 0.31 0.65 1.02 0.61 0.52 0.65 0.90 0.45 0.51 0.70 1.10 0.93 1.40 2.33 * Based upon estimates of nominal sampling effort, uncorrected for differences in effective fishing power. Italicized values were apparently affected by in- or out-migrations. 3r 1973 Size and age composition Rather than being a homogeneous pool, the eastern Bering Sea pollock population exhibits a size and age structure that reflects pulsed annual birth inputs, birth and survival rates, and a differentiation of distributional patterns based upon age. Pollock are generally considered to spawn during only one period each year. In the eastern Bering Sea, the spawning period has been reported to extend from the end of February through June, with peak activity in May (Serobaba 1968). This concentration of spawning within a four-month period, loosely synchronized with the timing of spring plankton production, provides for a large annual pulse of eggs and larvae. Table 33-3 and Fig. 33-9 summarize the apparent age frequency distributions observed for pollock within the central region of the eastern Bering Sea (shovm in Fig. 33-8), determined from NMFS demer- sal trawl surveys during the period 1973-1978. The data represent an overall range in body size (fork length) of approximately 6-90 cm. Ages were deter- mined from readings of saccular otoliths. Features of the observed age structure include: (1) a maximum age of 15 years; (2) a relatively high overall average mortality rate; (3) variable age-class abundance between years; and (4) large, dominant year-classes were apparently not maintained over successive years. Fig. 33-10 shows the overall pattern of survivorship, based upon the mean apparent abundance of each age group (i.e., column means) during the six years of observations. Assuming that Age (yr) Figure 33-9. Annual variations in apparent age frequency distributions, based upon the data in Table 33-3. 534 Fisheries oceanography 10,000 p- -\ 1,000 (0 3 > c T- >< d> o c (0 3 < 0) (0 jO O 0) O) < c (0 a> 10 OVERALL SURVIVORSHIP (1973-1978) \ \ \ \ \. \ .\ \ \ \ZL 1 - 6 8 10 Age (yr) 12 14 Figure 33-10. Overall survivorship curve, based upon the data in Table 33-3. rates of in-migration to, and out-migration from, the index area were equal, and that: N2 -Nie-zt (1) where Ni N2 Z t e = the number of individuals at any moment, = the number of individuals at some later time, - an instantaneous rate of total mortality, = the intervening time interval, and = the natural constant 2.71828..., (Z) was 0.72/yr. This rate corresponds to 51 percent of the pollock population dying each ye£ir. Rather than being constant, mortality was relatively high for ages 1, 4, 5, and 10 or more years (Fig. 33-11). Fac- tors that may have contributed to apparently higher -2.00 -1.80 MEAN AGE-SPECIFIC L MORTALITY ^ -1.40 then the overall observed instantaneous mortality rate Figure 33-11. Age-specific instantaneous mortality rates, based upon the data in Table 33-3. mortality at ages four and five include the possibili- ties of increased selective removal by fisheries, increased migratory activities associated with onto- geny, and effects of reproductive stress. Senescence was presumably a major cause of higher mortality at ages 10 and older. Life table characteristics are summarized in Table 33-4. Contrary to the assumption of a closed population, or that in- and out-migrations were in equilibrium, the indices of abundance observed in 1975 for all age-classes except the age-one population were inconsistently low (relative to cohort abundances observed in 1976), apparently indicating a distribu- tional shift out of the index area during that survey period. Another interesting result was the indication of density-dependent mechanisms possibly regulating age-class size (Fig. 33-12). Although the relative Walleye pollock 535 year-to-year variation in abundance of one-year-old populations was slightly less than the overall average of age-classes 1-12 years, the apparent sizes of two- year-old populations were relatively constant. The lack of evidence for persistence of strong year-classes —such as the 1973 and 1976 year-classes that were unusually abundant at age one— was also unexpected, particularly since cohort dominance seems to be a feature of the dynamics of at least some simple ani- mal populations with a large cannibalistic adult frac- tion (Fox 1975). The NMFS Crab-Groundfish and other research vessel surveys of the eastern Bering Sea continental shelf have also found a geographical differentiation of age structure (Pereyra et al. 1976: Fig. IX-34; Bak- kala and Smith 1978: Fig. 38). In summer, low- density populations in central and inner areas of the continental shelf— particularly northeast of the Prib- ilof Islands— have tended to be almost exclusively composed of one-year-old individuals. Along the outer continental shelf and slope, populations have been composed of a complex mix of age-classes 2 to 10 or more years. Growth Although gross characteristics of the growth of eastern Bering Sea pollock have been described (Yamaguchi and Takahashi 1972, Chang 1974, Pereyra et al. 1976, Bakkala and Smith 1978), no studies have yet attempted to analyze or mathemati- cally model details of the growth curve(s) or the energetics of growth processes. The approach taken here will be to analyze original data to examine TABLE 33-4 Life table characteristics for walleye pollock, based upon NMFS Crab-Groundfish research vessel surveys, 1973-1978. Agei Mean apparent^ Mortality'* (yrs) abundance (X 10^) Survival^ (yr^M X "x Ix ^x 0.2-1.2 (5100.) 1.0000 1.2-2.2 1416.3 0.2777 -0.95 2.2-3.2 550.3 0.1079 -0.29 3.2-4.2 413.4 0.0811 -0.15 4.2-5.2 354.0 0.0694 -1.12 5.2-6.2 115.9 0.0227 -0.84 6.2-7.2 49.9 0.0098 -0.21 7.2-8.2 40.5 0.0079 (+0.05) 8.2-9.2 42.5 0.0083 -0.53 9.2-10.2 25.0 0.0049 -0.59 10.2-11.2 13.9 0.0027 -0.88 11.2-12.2 5.75 0.0011 -1.15 12.2-13.2 1.82 0.0004 -1.87 13.2-14.2 0.28 0.00005 -1.73 14.2-15.2 0.05 0.00001 -0.51 15.2-16.2 0.03 0.000006 — ' Assuming a median birthday of April 1. ^ Number of individuals reaching age x at time t. The value for values of ages 1.2 to age 0.2 yr was extrapolated from the n 15.2 yr, assuming constant mortality Probability that an individual survives to age x. Age-specific instantaneous mortality rate, computed as m^ = ln(nx+i)~ In(nx)- The value in parentheses at age 7.2 yr may indicate sampling error or effects of migration. 1.00 (0 ■= 0.801- (9 > ■£ 0.60- « o § 0.40 O 0.20 Mean 1 2 3 4 5 6 7 8 9 10 11 12 Age-classes (yr) Figure 33-12. Comparisons of relative year-to-year varia- bility in age-class size, based upon the data in Table 33-3. The coefficient of variation is used as the ratio of standard deviation to mean. fundamental characteristics of growth in length, growth in weight, and growth and decay of cohort weight. Table 33-5 summarizes the mean lengths of indivi- duals in each age-class represented in Table 33-3, determined from annual NMFS demersal trawl surveys of the index area (shown in Fig. 33-8). To examine growth in length, a generalized form of the von Bertalanffy (1938) growth model was fitted to the overall mean lengths (Table 33-5— column means) of ages 1 to 13 years. The form of the generalized growth model used was: Y^ =yoo[l-e-K(x-xo)]d (2) where yoo K the length or weight at age x, an asymptotic value of length or weight, a relative growth completion rate, 536 Fisheries oceanography TABLE 33-5 Summary of mean lengths-at-age determined from NMFS Crab-Groundfish research vessel surveys conducted during June to mid-August, 1973-78 (units = cm fork length) Survey year Age -classes (yr) 7 8 9 10 11 12 13 14 15 1973 1974 1975 1976 1977 1978 15.7 25.7 32.9 39.2 42.0 43.4 46.0 47.3 49.5 52.4 56.1 - 63.5 66.0 - 15.5 25.2 34.9 38.5 44.3 44.3 47.7 49.6 51.0 53.4 51.7 57.6 57.8 - - 14.3 24.3 32.5 39.6 44.3 47.1 50.5 53.4 55.3 56.2 55.6 63.4 64.0 - - 11.6 23.9 30.6 37.1 41.5 44.2 46.4 48.2 48.6 51.0 55.1 57.8 58.5 66.1 72.0 15.9 23.9 33.7 41.7 44.8 47.5 50.0 51.7 52.0 53.3 54.4 54.5 52.0 60.0 - 13.7 25.2 33.7 37.7 44.0 46.7 51.6 53.3 53.1 53.0 55.6 53.0 - - - Xjj = a hypothetical age of zero size, and d = a dimensionless exponent reflecting ab- solute growth rate. The model vi^as fitted to the data by a computer program using a Taylor-series method for least squares approximation (Draper and Smith 1966). The results are shown in Figure 33-13, and the parameter values of best fit describing growth in length (sexes combined) were: y°o - 64.81 cm K = 0.132/yr Xo = 0.78 yr d = 0.520. In summary, the pattern of growth in length observed was fast growth between ages 1 and 5, relatively rapid decrease in annual length increments with increasing age, and an asymptotic body length of approximately 65 cm. To examine characteristics of growth in weight, the mean weights -at-age were approximated by applying a power relationship to the length data in Table 33-5. The relationship used was (Pereyra et al. 1976): W = 0.0075 L2-977 (3) where : W = estimated body weight in grams, and L = body (fork) length in cm. After converting the mean length of each age-class to a computed "mean weight," for all years, the overall mean weights (i.e., column means) were de- termined, and the generalized growth model (Equa- tion 2) was fitted to the results (see Fig. 33-13). The parameter values of best fit describing growth in esti- mated weight were: yoo = 7057 g K = 0.016/yr Xo = 1.18 yr d - 0.944. 60 50 E o £ 40 •Jf 30 O 20- 10- -1200 GROWTH Age (yr) Figure 33-13. Growth in length and weight, based upon the data in Table 33-5. Symbols indicate the overall mean at each age. Walleye pollock 537 In contrast to growth in length, growth in body weight was nearly linear within the age range 1 to 13 years, with almost constant increments in weight of 114 g/yr. Table 33-6 summarizes the overall mean body sizes at each age and, combining the observed survivorship from Table 33-4 with mean weigh ts-at-age, illustrates the apparent growth and decay of cohort weights. According to this model, a pollock cohort attains maximum biomass at approximately four years of age. Reproduction Because seasonal spawning aggregations and roe production have been of particular interest to com- mercial fisheries, major features of the reproductive biology of pollock have been fairly well described (Gorbunova 1954, Zverkova 1969, Serobaba 1968, 1971, 1974, Maeda and Hirakawa 1977). The attain- ment of sexual maturity is related to body size. On the basis of a classification of macroscopic gonad conditions observed during a large demersal trawl survey of the eastern Bering Sea during April to June 1976, Bakkala and Smith (1978) found that the POLLOCK MALES April-June 1976 Fork Length (cm) i_l ! I I I I I I ■ I I I 20 25 40 45 Fork Length (cm) Figure 33-14. Length-maturity relationships observed from pollock taken during the April-June 1976 demersal trawl survey (Bakkala and Smith 1978). TABLE 33-6 Characteristics of the grovi^th of walleye pollock, based upon NMFS Crab-Groundfish research vessel surveys, 1973-1978. Age^ Mean body^ Mean body^ Relative age class'* (yrs) length (cm) weight (g) weight (g) X Yx Wx IxWx 0.2-1.2 1.2-2.2 14.4 21.9 6.08 2.2-3.2 24.7 105.2 11.35 3.2-4.2 33.0 251.0 20.36 4.2-5.2 39.0 409.6 28.43 5.2-6.2 43.5 565.0 12.83 6.2-7.2 45.5 650.9 6.38 7.2-8.2 48.7 796.6 6.29 8.2-9.2 50.6 892.6 7.41 9.2-10.2 51.6 945.4 4.63 10.2-11.2 53.2 1034.0 2.79 11.2-12.2 54.8 1125.0 1.24 12.2-13.2 57.3 1298.0 0.52 13.2-14.2 59.2 1436.0 0.07 14.2-15.2 (64.0) (1800.0) (0.02) 15.2-16.2 (72.0) (2537.0) (0.02) * Assuming a median birthday of April 1. ^Overall observed mean fork lengths. The lengths in paren- theses were based upon limited observations. ^Overall mean predicted body weights (wet tissue). '* Computed as the product of "survival" times "mean body" weight. relationship between "fork length" and "proportion of individuals sexually mature" was best described by a sigmoidal curve of the model: P = e be -cL (4) where P is the proportion of the apparent population mature at size L cm, e is the natural constant 2.71828..., and b and c are constants. The fitted model for male pollock was found to be (Fig. 33-14): p = g— 725.947e" .224L (5) and the length at which 50 percent of the individuals were mature was 31.0 cm. Female pollock matured at longer lengths than males, and the fitted model was: .209 L p = g— 867.088e (6) with the length at 50 percent maturity equal to 34.2 cm. In the eastern Bering Sea, the sexually mature fraction of the pollock population undergoes a 538 Fisheries oceanography seasonal cycle of gonad development and repro- ductive activity (Yamaguchi and Takahashi 1972). Despite a relatively extended spawning period of February through June, the majority of the breed- ing population appears to ripen and spawn during April to mid-May. Within this spawning period, it is not yet clear whether a mature female releases eggs in one short pulse, or as multiple releases over a period of days or weeks. As for many other fish species, attempts to mea- sure individual female fecundity (i.e., number of viable eggs released per year) have largely been premature to an adequate understanding of the biological processes of oocyte development, oocyte release, and the extent of resorption (Foucher and Beamish 1977). This has resulted in uncertainty over which sizes of oocytes to include in estimates of annual zygote production, uncertainty of spawning model (namely, single- or multiple-spawning), and failure to assess the number of oocytes remaining in the ovary after spawning (Shew 1978). The ovaries of female pollock contain oocyte populations often composed of two or three distinct size-classes (Gorbunova 1954, Serobaba 1974, Shew 1978). Although details of the production and fates of these different size groups remain to be studied, it seems likely that the small classes (oocyte diameters ca. 60-550 iJ.m) may represent a reserve fund (Fou- cher and Beamish 1977) from which only a fraction develop each season to be spawned. Assuming that only oocytes in the largest size class (ca. 600-1500 //m diameter) are actually released during spawning and viable, the relationship Shew (1978) found between body size and potential fecundity was: F = 0.29 L3-462 (7) where: F = estimated number of oocytes in the 600- 1500 jum diameter size-class, and L = body (fork) length in cm. By applying Equations 6 and 7 to the survivor- ship and growth data in Tables 33-4 and 33-6, esti- mates could be made of individual and age-class fecundity (Table 33-7). Interesting features of the results are that age-classes four and five years appar- ently contribute most to the potential reproduction of the population, two- and three-year-olds contri- bute little because of their small sizes, and the cumu- lative contribution of rare old (six years and older) individuals is large. The behavior of pollock during the spawning period (February to June) apparently includes migrations to, and aggregations along, the outer continental shelf and slope (Fig. 33-15). Although Figure 33-15. Locations of individuals observed in spawn- ing condition during April-June 1976. Dots siiow exact positions, siiading indicates inferred range of spawning activities. spawning probably occurs along the entire eastern Bering Sea shelf edge, the outer shelf region just west and northwest of Unimak Island has been repeatedly identified as a major reproductive site (Maeda 1972, Maeda and Hirakawa 1977, Serobaba 1968, 1974). Trawl fisheries target upon pre- spawning and spawning populations, obtaining high catch rates. Some fishing vessels harvest pollock only for roe, discarding their catches after stripping the ovaries from ripe females. Schools of pre-spawning and spawning pollock apparently move high into the water column, forming dense midwater layers (Takakura 1954, Serobaba 1974). Mating presumably involves pairing, and broadcasting of the pelagic non-adhesive eggs (ca. 1.5 mm in diameter) with no parental care. Consis- tent with reported locations of principal spawning areas, pollock eggs and larvae have been observed in highest concentrations along the outer continental shelf between the Pribilof Islands and Alaska Penin- sula (Serobaba 1968, Waldron and Vinter 1978). Although some studies of the early life history of pollock in the eastern Bering Sea have presupposed a simple counterclockwise current transport through the central shelf region during the first one or two years of age, recent data suggest a different and more complex pattern of movement (Figs. 33-16 and 33-17). As opposed to the previous impression of predominantly counterclockwise water motion over the eastern Bering Sea shelf, the present concept of Walleye pollock 539 TABLE 33-7 Reproductive characteristics of walleye pollock, based upon NMFS Crab-Groundfish research vessel surveys, 1973-1978. Agea Pivotal Proportion of Individual Potential age- (yrs) length (cm)'' females mature*^ fecundity (X 10^ )d class fecundity^ X Vs Ps K IsPsbs 0.2-1.2 1.2-2.2 13.7 0. 0. 0. 2.2-3.2 23.5 0.008 16.2 0.02 3.2-4.2 31.4 0.289 44.1 1.09 4.2-5.2 37.0 0.641 77.9 3.59 5.2-6.2 41.3 0.842 114.0 2.72 6.2-7.2 43.2 0.901 133.0 1.39 7.2-8.2 46.3 0.947 169.0 1.33 8.2-9.2 48.1 0.963 193.0 1.54 9.2-10.2 49.0 0.970 206.0 1.10 10.2-11.2 50.5 0.978 229.0 0.69 11.2-12.2 52.1 0.984 255.0 0.35 12.2-13.2 54.4 0.990 296.0 1.17 13.2-14.2 56.2 0.993 331.0 0.03 14.2-15.2 60.8 0.997 435.0 0.004 15.2-16.2 68.4 0.999 654.0 0.004 ^Assuming a median birthday of April 1. ''The size at time of spawning ("April 1") was computed as 95% the y^ value. <=Values for ages 1 to 4 from Bakkala and Smith (1978), Table 26. *^Number of oocytes (> 600 nm diameter) produced per year, using the relationship from Shew (1978). ^Potential age-class fecundity = product of survival (at time of spawning), proportion mature, and individual fecundity; 1^ values were computed using the age-specific mortality rates, assuming equal survival of males and females. Units = 10^ oocytes. • 1978 Midwater sighting ■ 1979 Midwater sigtiting t^ Overall apparent range i of Itsti predato Figure 33-16. Observations of 0-group (two- to four- month -old) pollock during June to mid-August, based upon NMFS research surveys during 1978 and 1979. Figure 33-17. Overall apparent range of one-year-old pollock during June to mid-August, based upon NMFS research surveys, 1975-79. 540 Fisheries oceanography Black-legged Kittiwake Red-legged Kittiwake UIca Myoxocephalus spp. bolini Northern (Steller) Sea Lion POLLOCK EGGS Copepod Eggs POLLOCK LARVAL -STAGES (5-20 mm) POLLOCK JUVENILES (2-20 cm) POLLOCK - ADULTS (20-90 cm) Copepods Figure 33-18. Apparent food web based upon pollock in the eastern Bering Sea. Dotted lines indicate pollock ontogeny. Solid lines show important feeding relationships, with arrowheads and line weights indicating the direction and relative magnitude of carbon flows. dominant long-term mean circulation is an extremely slow (1 cm/sec) drift to the northwest approximately parallel to the bathymetry (Coachman 1979). If spawning occurs predominantly along the southeast outer shelf, then several months of north- west drift would carry larvae to the vicinity of the Pribilof Islands. In fact, during NMFS Crab-Ground- fish trawl surveys conducted from June to mid- August, 0-group pollock (approximately two to four months old, and 35-80 mm fork length) have been observed over a large area of the northwestern outer shelf (Fig. 33-16). Highest concentrations of 0-group juveniles, showing as dense mid water or near -bottom layers, have been noted directly west of the Pribilof Islands. Inspections of the stomach contents of demersal fish species throughout the eastern conti- nental shelf region have found 0-group pollock (lengths 50-80 mm) to occur as an abundant food item only in this same area. By one year of age, pollock are broadly distri- buted over the entire central and outer continental shelf, completely overlapping and extending inshore beyond the adult range (Fig. 33-17). By two years of age, pollock remain more restricted to deep water, and with growth, begin to recruit to the trawl fisher- ies. FOOD-WEB RELATIONSHIPS One of the most striking features of the eastern Bering Sea food web is the extent to which pollock are represented in feeding relationships and overall energy exchange (Fig. 33-18). During different life history stages, pollock feed upon a broad spec- trum of prey (Takahashi and Yamaguchi 1972, Mito 1974, Clarke 1978, Feder 1978, Smith et al. 1978, Bailey and Dunn 1979). In turn, pollock serve as a major food resource for a wide variety of primarily high trophic level predators (Mito 1974, Feder 1978, Hunt 1978, Lowry et al. 1978, Smith etal. 1978). Food and feeding The feeding characteristics of pollock are appar- ently largely a function of body size, regional loca- tion, and time of year. Body size sets certain mor- phological and behavioral limits— such as size of the feeding apparatus, extent of dentition, and swimming speeds— that affect the types of prey that can be captured and ingested. Body size is also strongly related to vertical migratory behavior within the water column and geographical location— factors that also determine available prey. Although regional, seasonal, and yearly variations in prey densities must be important in determining food composition and feeding rates, these have not yet been well described in the eastern Bering Sea. Larval pollock (5-20 mm) have been reported by Clarke (1978) to undergo shifts in selection of prey sizes during ontogeny. The gut contents of small larvae (5-10 mm) consisted mainly of copepod eggs and nauplii. Larger larvae showed progressively increasing fractions of cyclopoid copepods, cala- noids, and larval euphausiids. Juvenile pollock (2-20 cm) presumably feed on a corresponding series of primarily zooplankton prey of increasing sizes, perhaps foraged from areas high in the water column. Although not well studied, important food items probably include copepods, amphipods, juvenile and adult euphausiids, eggs, and larval fishes. As adults (lengths 20-90 cm) pollock are general- ized predators that appear to follow a "diumally schooling-nocturnally active" behavior pattern, although freshly ingested food items can be found in the gut at any time of day. Large eyes are apparently adaptive for visual feeding, hunting at night, and foraging during low vdnter light intensi- ties. The sleek, fusiform body indicates capabili- ties of fast swimming speeds. A large mouth and abundant, fine needle-like dentition provide good apparatus for seizing and ingesting large, active zooplankton (primarily Thysanoessa spp. and Parathemisto spp.) and fish (mainly juvenile pollock) prey (Fig. 33-19). c 0) c o o o (0 E o (A 100 80 [D Pollock Hi Shrimp and crab Walleye pollock 541 I — I Euphausiids H Amphipods H Copepods M CM CO to Fork Length (mm) Figure 33-19. Changes in prey selection with size (from Takahashi and Yamaguchi 1972). To place the role of pollock in the eastern Bering Sea food web in perspective, estimates were made of the food requirements for three overall average stock densities and related to primary production (Table 33-8). At the average pollock densities that can be inferred from research vessel TABLE 33-8 Comparisons of estimated food intake rates of walleye pollock in the eastern Bering Sea to primary production. Pollock density' (mg wet wt/m^) Pollock density^ (mgC/m^) Food intake rate^ (mg C/m^ /day) Food intake rate as percentage of primary production rate'* 2,500 5,000 10,000 133 266 532 * Overall absolute density; the italicized value was the mean density observed by Pereyra et al. (1976), assuming 100% trawl efficiency. ^Carbon equivalents were computed assuming a dry wt. of 14% wet tissue wt., and carbon fraction of 38% of dry wt. (Parsons et al. 1977: Table 12). ^Assuming a mean feeding rate of 1.5% of body weight per day (Daan 1973). '^ Assuming a mean primary production rate (water column) of 200 mg C/m^ /day (McRoy et al. 1972). 542 Fisheries oceanography surveys (50-100 kg/ha), the estimated daily food requirements represented approximately 2-4 per- cent of average photosynthetic production. Predation Pollock are fed upon by a wide variety of other species, and cannibalism (or "intraspecific preda- tion") is also high. To a certain extent, the pattern of predation upon pollock shows regional varia- tions reflecting the range overlaps and activity areas of the different predator populations. In deep water along the outer continental shelf, juvenile and adult pollock are a major food item of Pacific cod, Pacific halibut (Hippoglossus steno- lepis), Greenland halibut (Reinhardtius hippoglos- soides), arrowtooth flounder (Atheresthes spp.), and sablefish (Anoplopoma fimbria). In the vicinity of the Pribilof Islands where large colonies of seabirds and pinnipeds occur, predation upon larval and juvenile pollock from the near-surface layer by seabirds— the Red-faced Cormorant, Phala- crocorax urile; kittiwakes, Rissa tridactyla, R. brevi- rostris; murres, Uria aalge, U. lomuia; puffins, Frater- cula corniculata, Lunda cirrhata; and Parakeet Auk- let, Cyclorrhynchus psittacula (Hunt et al., volume 2)— is significant, and larger pollock are actively dived and fed upon by the northern fur seal (Callorhinus ursinus) (Harry, volume 2). Other widely ranging predators apparently feed upon pollock throughout their eastern Bering Sea distribution: cottids (particularly Myoxocephalus spp. and Ulca bolini), the Steller sea lion (Eumeto- pias jubata), other seals (Phoca spp.) and toothed whales (perhaps particularly Phocoenoides dalli and Orcinus orca). Although cannibalism upon eggs and larvae during the post-reproductive period is a common behavior of many fish species, pollock are unusual due to the extent that intraspecific predation is a normal activity and represents a large fraction of total food intake. It is not unusual to find 35-cm pollock feeding upon 5 to 10-cm juveniles, and 60 to 80-cm adults feeding upon pollock 10-30 cm. This peculiar feeding strategy— by which cannibal- istic adults gain a meal and eliminate potential competitors at the expense of their juveniles— may result in sensitive density-dependent processes for regulating both the age structure and total size of the population (Fox 1975). FISHERY CHARACTERISTICS Present status and history Detailed reviews of the history and character- istics of the different eastern Bering Sea trawl fisheries are given in Pruter (1973, 1976), Forrester et al. (1978), Low and Akada (1978), NPFMC (1978), and Bakkala et al. (1979a). In summary, large trawl fisheries for walleye pollock have been a relatively recent development (Table 33-9), repre- senting a shift from targeting upon other fish species— primarily Pacific ocean perch (Sebastes alutus) and yellowfin sole (Limanda aspera)— prior to 1964. This shift was a consequence of declining catch rates (i.e., apparent overfishing) of the other species, and the development of new processing techniques and end products designed to improve the world market demand for pollock. During the period 1966-70, annual pollock catches increased steeply (see Table 33-9). Peak yields were taken during 1971-73, then catches declined to the present levels of 950,000-980,000 mt/yr. These recent catch levels represent manage- ment quotas established under the authority of the U.S. Fishery Conservation and Management Act of 1976, Public Law 94-265. During 1978, 84 percent of the total eastern Bering Sea pollock catch was taken by Japanese, 9 percent by U.S.S.R., 6 percent by Republic of Korea, and 0.3 percent by Taiwanese vessels. The present fishing fleets include factory ships, factory stern trawlers, independent stern trawlers, pair trawlers, and refrigerator transports. Selective characteristics Contrary to general impressions, trawl fisheries in the eastern Bering Sea are composed of special- ized units (i.e., national and company fleets, and individual fishing vessels) that selectively target upon specific fish populations (particular species and size mixes) so as to maximize the economic return from the particular catching, processing, and storage capabilities of each vessel. The market products that a fishing vessel is oriented towards— frozen products (fillets or blocks), minced fish ("surimi"), meal, or roe— largely determine fishing strategies. Depending upon seasonal variations in the availabilities of fish populations, different fishing units may: (1) exclusively target their trawling activities upon pollock; (2) switch the direction of trawling among several higher -valued species, taking pollock incidentally; or (3) at times, even discard mistakenly -caught pollock as a trash item. Fig. 33-20 shows the overall geographical distribu- tion of pollock catches (total annual yield) by Japan- ese fishing vessels in the eastern Bering Sea during 1977. In general, vessels targeting specifically upon pollock would probably have mainly fished along the Walleye pollock 543 TABLE 33-9 Summary of annual removals of pollock from the eastern Bering Sea by trawl fisheries, 1964 to 1979 (metric tons).^ Republic of China, Year Japan U.S.S.R. Korea Taiwan Poland Total 1964 174,792 0 0 0 0 174,792 1965 230,551 0 0 0 0 230,551 1966 261,678 0 0 0 0 261,678 1967 550,362 0 0 0 0 550,362 1968 700,981 0 1,200 0 0 702,181 1969 830,494 27,295 5,000 0 0 862,789 1970 1,231,145 20,420 5,000 0 0 1,256,555 1971 1,513,923 219,840 10,000 0 0 1,743,763 1972 1,651,438 213,896 9,200 0 0 1,874,534 1973 1,475,814 280,005 3,100 0 0 1,758,919 1974 1,252,777 309,613 26,000 0 0 1,588,390 1975 1,136,731 216,567 3,438 0 0 1,356,736 1976 913,279 179,212 85,331 0 0 1,177,822 1977 868,732 63,467 45,227 944 0 978,370 1978^ 821,306 92,714 62,371 3,040 0 979,424 1979^ (774,630) (60,000) (85,000) (5,000) (25,000) (950,000) * Data from Bakkala et al. (1979b), unless otherwise cited. ^Preliminary estimates. ^ Foreign fishing allocations (Pileggi and Thompson 1979). Figure 33-20. Geographical distribution of the total annual Japanese pollock catch during 1977, summarized by V2° latitude and 1° longitude squares. outer shelf. Vessels taking pollock incidentally to other species, such as yellowfin sole, Alaska plaice (Pleuronectes quadrituberculatus), and deep pleuro- nectids (Reinhardtius hippoglossoides and Ather- esthes spp.), are represented by the extended distri- butions of low-level catches over deep water and in the central shelf. Although pollock are taken in all months of the year, a majority of the annual catch is taken during the summer from June through September. Fishing efforts directed toward pollock are highly size-selective (Chang 1974: Fig. 12), a consequence of the fish densities (more specifically, catch rates) available at different body sizes, trawl mesh sizes, and processing requirements of the different vessels. Fig. 33-21 shows the size frequency distributions ob- served from "survey" and "fishery" populations during 1978. Compared to the size distribution of the pollock population determined from the more evenly weighted and inshore survey coverage of the eastern shelf, trawl fisheries were selecting relatively larger individuals at the shelf edge. A disturbing recent trend of eastern Bering Sea pollock fisheries, however, has been the selection 544 Fisheries oceanography 1.00 10 20 30 40 50 60 Fork Length (cm) Figure 33-21. Comparison of the size frequency distribu- tions of survey and fishery pollock populations. Survey results are from the 1978 NMFS Crab-Groundfish sampling survey of the index area shown in Fig. 33-8. Fishery results are from 1978 NMFS Foreign Fisheries Observer Program sampling of the commercial catches of Japanese mother- ships, Japanese large stern trawlers, and Korean large stern trawlers in INPFC area 1. of an increasingly larger proportion of small and young fish. The size distribution of the 1978 fishery catch shown in Fig. 33-21 represents an age composition of approximately 2 percent one-year- olds, 20 percent two -year -olds, 40 percent three- year-olds, 18 percent four -year-olds, and 20 percent five or more years of age. In the perspec- tive of the apparent growth and reproductive schedules for the population (see Tables 33-6 and 33-7), it would appear prudent to reduce fishing pressures upon age-classes 1 to 3 years. SEASONAL AND INTERANNUAL VARIABILITY Seasonal variations The eastern Bering Sea environment is character- ized by strong seasonality. Temperatures are seasonally cyclic and extremely cold during winter. Mixing and overturn of the water column occurs as a result of heat loss from surface layers during the winter, followed by increased water column stability during summer warming. Biological production processes follow a cycle of pulsed high activities during spring and summer, followed by a long winter period of low activity. Apart from these generalizations, distinctive biological domains (i.e., geographical subregions and water mass layers) are apparently formed in the eastern Bering Sea characterized by differ- ences in mean and extreme values of environmental conditions, relative variability, and timing of events. These different domains, or habitats, are presumably used by biological species according to their resource value (in terms of Darwinian fitness) at specific times. If we broadly define migration as "an act of moving from one spatial unit to another" (Baker 1978), then walleye pollock can be seen to exhibit a continuum of movement types. At short time scales (hours or days), the movements of indivi- duals reflect the daily activity pattern— forming and dispersing from schools, searching for food, and avoiding predators. At longer time scales (weeks and months), individuals and large subpop- ulations (such as certain age-classes) show move- ments between different biological domains— from deep water up onto the continental shelf and back, along the shelf, and between major regions. A dominant behavior of pollock appears to be seasonal movements on- and off-shelf (i.e., move- ments normal to bathymetry). During winter months, shelf populations appear to retreat to deep water along the outer shelf edge and may extend pelagically out over deep water. During the summer, vertical and onshore movements occur that result in high concentrations of adults along the outer shelf and a dispersal of primarily juven- iles (age-classes one and two years) into the central and inner shelf regions. These seasonally-related bathymetric shifts are reflected in the trawling depths of the commercial fishery (Fig. 33-22). Since pollock are ectotherms (body tempera- tures in equilibrium with their surroundings) on- and off-shelf migrations appear to be an adaptive response to the extremely cold temperatures (0.0 to —1.7 C) of the shelf domain during winter. Along the shelf edge at depths of 200-300 m, water temperatures are relatively constant— 3-5 C throughout the year, providing a warm winter refuge (i.e., freezing avoidance) layer. Dispersal from this layer out onto the continental shelf during summer presumably maximizes the exploi- tation of different food resources by different size- and age-classes. Along-shelf movements (i.e., parallel to bathy- metry) have been found from tagging studies and inferred from month-to-month changes in the patterns of commercial fisheries catches (Serobaba, 1970, Maeda 1972, Takahashi and Yamaguchi 1972). These have been suggested to represent the directed, seasonally-related movements of subpopula- Walleye pollock 545 I I tions between exclusive wintering, spawning, and summer feeding areas. Although the outer continental shelf area just west and northwest of Unimak Island does appear to be a major reproductive site, other resource areas (including winter refuge, other low- level spawning areas, and summer ranges) now appear to be broadly distributed, and there do not seem to be clear long-distance migration routes to particular seasonal activity centers. Two large subsurface domains that appear fundamentally different are the outer continental shelf areas northwest and southeast of the Pribilof Islands. Temperature conditions are colder in the northern area, winter sea ice extends seaward of the continental shelf, and winter overturn pene- trates the water column to depths below the shelf edge (Favorite et al. 1977). As a result, in this northern region the timing of spring and summer events— movements related to spawning, the spawning period, and dispersal onto the shelf— appear to follow the southeastern region by a lag of approximately 1 to 2 months. A characteristic progression of the commercial fisheries from south to north along the outer shelf during spring to mid- summer may represent targeting upon time-lagged spring aggregations along the shelf edge, following a northward migrating subpopulation, or perhaps partly result from restrictions due to ice and weather condi- tions. Evidence has been presented for a progressive de- crease in the size composition of pollock in the southeast area during the post-spawning period, presumably due to movements of large adults from the reproductive site (Serobaba 1970, Yamaguchi 1979). Whether these same, migrating, spent individ- uals account for increases in population size composi- tion that occur in the northwest area during the same time period, will need to be verified by mark and re- capture studies. The responses of pollock to winter ice cover may be expected to be largely determined on the basis of thermal characteristics of the subsurface (particularly bottom water) environment. Since an arctic distribu- tion is not shown, and there is no evidence of the development of biochemical mechanisms for adapting to freezing conditions, then it must be assumed that pollock are excluded from shallow regions of the continental shelf during winter ice cover. Along the outer continental shelf, the penetration of relatively warm (1-5 C) subsurface circulation under ice cover may still provide an adequate thermal refuge. As suggested by genetic similarities and some tagging studies, interregional exchange is apparently an important process that may also vary seasonally. u _.. , 1 1 1 1 1 T T 1 - 100 1 r 1 r - 1 Q. Q 200 n i i 1 ' ' 300 ■ J [ J L - i - 400 - 1 1 1 1 1 I - JFMAMJ JASOND 1978 Figure 33-22. Monthly variations in fishing depths for pol- lock, based upon 1978 NMFS Foreign Fisheries Observer Program data from Japanese and U.S.S.R. independent stern trawlers in INPFC area 1. Dots indicate the median fishing depth during each month, vertical bars indicate the monthly range. The shaded area, depths 135-245 m, represents the principal range of values during the year. Only gear operations in which pollock represented 25 percent or more of the total catch (by weight) were in- cluded in the analysis. A relevant research problem is the question of deter- mining the origins and relationships (to shelf popula- tions) of pelagic walleye pollock found over deep water in the central Bering Sea. During an acoustic survey of the Aleutian Basin during June and July 1978, Okada (1978) observed "spotted type" echo patterns from large pollock throughout essentially the entire region, mainly from mid water layers at approx- imately 30-80 m depths during daytime. Pollock in these layers were extremely uniform in size, with 82 percent of all individuals sampled between 44 and 50 cm. Variations between years Even if the eastern Bering Sea were a more con- stant physical environment, important properties of the walleye pollock population (e.g., abundance, size and age structure, growth, food requirements) might still be expected to undergo long-term fluctuations resulting from (1) internal processes of population regulation, and (2) the external, and variable, influ- ences of predators (including fisheries) and prey. Density-dependent processes of population control may be expressed as variations in life table character- istics. Because the population is composed of ap- proximately 15 age-classes, and because inter -age class predation (and perhaps competition) is significant, interactions between age-classes may alone result in inherent variations in age-class abundances (and 546 Fisheries oceanography hence, total population density) between years. Similar control mechanisms must also result from the size-selective activities of predator populations, producing varying patterns of age-specific mortalities. Particular properties of the population are appar- ently sensitive to certain age-classes. The weight density, or biomass, of the population will largely be dependent upon the abundances of age-classes two to five (see Table 33-6). Potential fecundity will primar- ily be dependent upon the abundances of age-classes four to nine (see Table 33-7). The actual high year-to-year variability of the eastern Bering Sea physical environment adds other components that contribute to, and perhaps accen- tuate, fluctuations in population properties. Poten- tially important variations in physical environment include changes in general climate (McLain and Favorite 1976), anomalously extensive ice cover and formation of cold bottom water during winter, and changes in the flow and characteristics of eastern Bering Sea source waters. Climate is apparently important in determining the timing and extent of seasonal migrations of pollock both on- and off- shelf. During warm years, spring movements into shallower water have appeared to occur earlier, and the density distribution of the population has appeared to shift farther inshore than during cold years (Pereyra et al. 1976: Figs. VIII-11 and VIII-12). The direct importance of thermal conditions in determining distributional characteris- tics of the population may be overemphasized, however, and other factors that vary with tempera- ture (e.g., particular food items) may be a more direct cause. Hydrographic conditions and short-term climate (weeks) presumably have large influences upon annual reproductive success by contributing high variability to the survival of eggs, larvae, and early juveniles. The effects of variable early survival upon year-class abundances at subsequent ages, however, may be dampened or obscured by later density -dependent mortality processes. of biological interactions and process rates. Single- species age-structured models are needed to examine detailed processes of internal regulation. Multispecies models are needed to study effects of interactions between species. Yield-oriented models are needed to enable more objective evaluation of potential effects of different harvesting levels by fisheries, and to test consequences of feedback and non-feedback man- agement policies. Simulations of environmental randomness should be included in all three types of models. 2. We need to ensure the collection of more uniform population data, and in particular, to strive for developing longer observational time series based upon consistent procedures. This information is needed for use in mathematical models, and will provide the basis for scientific management of the fisheries. Fisheries surveys of pollock in the eastern Bering Sea would be more productive in the long run if the participating agencies better coordinated their research efforts, standardized sampling methods and gear, and agreed upon uniform statistical areas. There is also a need to improve the uniformity and quality of data reported from the commercial fisheries. 3. There is a need to study relationships between spawning population size, reproductive success, and subsequent age-class abundances and survival rates. Size- and age-specific mortality rates need to be evaluated as potentially important density-dependent mechanisms regulating population age structure and size. 4. The interrelationships of pollock with other species, particularly marine mammals, need to be studied. 5. We need a better understanding of seasonal movements and population exchange: (1) along the eastern Bering Sea continental shelf, (2) between populations associated with the eastern Bering Sea continental shelf and pelagic populations over deep water in the Aleutian Basin, and (3) between major geographical regions. FUTURE RESEARCH Because of the importance of pollock in the eastern Bering Sea and other regions of the North Pacific, there are needs for substantially increased research activities studying their basic biology, population dynamics and ecological relationships. In order of priority: 1 . There is a need to develop mathematical simula- tion models to examine the behavior of the popula- tion dynamics of pollock under different assumptions Walleye pollock 547 I REFERENCES Alverson, D. L., and W. T. Pereyra 1969 Demersal fish explorations in the northeastern Pacific Ocean— an evalua- tion of exploratory fishing methods and analytical approaches to stock size and yield forecasts. J. Fish. Res. Bd. Can. 26: 1985-2001. Bailey, K., and J. Dunn 1979 Spring and summer foods of walleye pollock, Theragra chalcogramma, in the eastern Bering Sea. Fish. Bull., U.S. 77: 304-8. Baker, R. R. 1978 The evolutionary ecology of animal migration. Holmes and Meier, N.Y. Bakkala, R. G., and G. B. Smith 1978 Demersal fish resources of the eastern Bering Sea: spring 1976. Nat. Mar. Fish. Serv., Northwest and Alaska Fish. Cent., Proc. Rep., Seattle, Wash. Bakkala, R., W. Hirschberger, and K. King 1979a The groundfish resources of the eastern Bering Sea and Aleutian Islands region. Mar. Fish. Rev. 41(11): 1-24. Bakkala, R., L. Low, and V. Wespestad 1979b Condition of groundfish resources in the Bering Sea and Aleutian area. Nat. Mar. Fish. Serv., Northwest and Alaska Fish. Cent., Seattle, Wash., Unpub. MS. Baranov, F. I. 1918 K voprosu o biologicheskii osnovani- iakh rybnogo khoziaistva (On the question of the biological basis of fisheries). Nauchnyi issledovatelskii iktiologisheskii Institut, Izvestiia, 1(1): 81-128. Izvestiia otdela rybo- vodstva i nauchnopromyslovykh is- sledovanii, T. I, 1. (Inst. Sci. Ichthy. Invest., Proc, 1(1): 81-128. Rep. Div. Fish Management Sci. Study of the Fish. Industry, Vol. I, 1). In Russian. (Transl. 1945, avail. Dep. of Environ., Biol. Sta., Nanaimo, B.C.) Bertalanffy, L., von 1938 A quantitative theory of organic growth. Hum. Biol. 10: 181-213. Chang, S. 1974 An evaluation of the eastern Bering Sea fishery for Alaska pollock (Thera- gra chalcogramma, Pallas): population dynamics. Ph.D. Dissertation, Univ. of Wash., Seattle. Clarke, M. E. 1978 Some aspects of the feeding biology of larval walleye pollock, Theragra chalcogramma (Pallas), in the south- eastern Bering Sea. M.S. Thesis, Univ. of Alaska, Fairbanks. Coachman, L. K. 1979 Water circulation and mixing in the southeast Bering Sea. In: C.P. McRoy and J.J. Goering, eds., Prog- ress report, PROBES phase I, 1977- 78, 1-46. Inst. Mar. Sci., Univ. of Alaska, Fairbanks. Coachman, L. K., and R. L. Charnell 1979 On lateral water mass interaction— a case study, Bristol Bay, Alaska. J. Phys. Oceanogr. 9: 278-97. Daan, N. 1973 A quantitative analysis of the food intake of North Sea cod, Gadus mor- hua. Neth. J. Sea Res. 6: 479-517. Draper, N. R., and H. Smith 1966 Applied regression analysis. Wiley, N.Y. John Favorite, F., T. Laevastu, and R. R. Straty 1977 Oceanography of the northeastern Pacific Ocean and eastern Bering Sea, and relations to various living marine resources. Nat. Mar. Fish. Serv., Northwest and Alaska Fish. Cent., Proc. Rep., Seattle, Wash. Feder, H. M. 1978 Survey of the epifaunal invertebrates of the southeastern Bering Sea. Environmental assessment of the Alaskan continental shelf. U.S. Dep. Comm., NOAA/OCSEAP, Ann. Rep. 4:1-126. 548 Fisheries oceanography Forrester, C. R., A. J. Beardsley, and Y. Takahashi 1978 Groundfish, shrimp, and herring fish- eries in the Bering Sea and northeast Pacific- - historical catch statistics through 1970. Bull. Inter. N. Pac. Fish. Comm. 37: 1-147. Foucher, R. P., and R. J. Beamish 1977 A review of oocyte development in fishes with special reference to Pacific hake (Merluccius productus). Canada Fish. Mar. Serv. Tech. Rep. 755. Fox, L. R. 1975 Cannibalism in natural populations. Ann. Rev. Ecol. Syst. 6: 87-106. Gorbunova, N. N. 1954 Razmnozhenie i razvitie mintaya Theragra chalco gramma (Pallas) (The reproduction and development of the walleye pollock, Theragra chalco- gramma (Pallas)). Tr. Inst. Okeanol. 11: 132-195. In Russian. (Transl. 1972, avail. Nat. Mar. Fish. Serv., Northwest and Alaska Fish. Cent., Seattle, Wash.) Grant, W. S., M. S. Alton, and F. M. Utter 1978 Biochemical genetic variation in wall- eye pollock: stocks of the Bering Sea and the Gulf of Alaska. Nat. Mar. Fish. Serv., Northwest and Alaska Fish Cent., Seattle, Wash., Unpub. MS. Hart, J. L. 1973 Pacific fishes of Canada. Bull. Fish. Res. Bd. Can. 180: 1-740. Hunt, G. L., Jr. 1978 Reproductive ecology, foods and foraging areas of seabirds nesting on the Pribilof Islands. Environmental assessment of the Alaskan continent- al shelf. U.S. Dep. Comm., NOAA/ OCSEAP, Ann. Rep. 1:570-775. Ikeda, I., Y. Takahashi, T. Sasaki, K. Mimura, K. Ima- mura, and L. L. Low 1977 Report of the working group on average density index computation for pollock in the eastern Bering Sea to the INPFC Biology and Research Committee. Inter. N. Pac. Fish. Comm., Vancouver, Can. Unpub. MS. Iwata, M. 1973 1975 Genetic polymorphism of tetrazolium oxidase in walleye pollock. Japan. J. Genetics 48: 147-9. Population identification of walleye pollock, Theragra chalcogramma (Pal- las), in the vicinity of Japan. Mem. Fac. Fish. Hokkaido Univ. 22: 193- 258. 1977 Isozyme research and population identification. Res. Inst. N. Pac. Fish., Hokkaido Univ., Spec. Vol: 519-34. Johnson, A. G. 1977 A survey of biochemical variants found in groundfish stocks from the North Pacific and Bering Sea. Anim. Blood Grps. Biochem. Genet. 8: 13-19. Kihara, K., and M. Uda 1969 Studies on the formation of demersal fishing grounds. I. Analytical studies on the mechanism concerning the formation of demersal fishing grounds in relation to the bottom water masses in the eastern Bering Sea. J. Tokyo Univ. Fish. 55: 83-90. Kimura, M., and T. Ohta 1971 Theoretical aspects of genetics. Princeton, N.J. population Kinder, T. H., L. K. Coachman, and J. A. Gait 1975 The Bering Slope current system. Phys. Oceanogr. 5: 231-44. J. Laevastu, T., 1977 Low, L. L. 1974 and F. Favorite Preliminary report on dynamical nu- merical marine ecosystem model (DY- NUMES II) for eastern Bering Sea. Nat. Mar. Fish. Serv., Northwest and Alaska Fish. Cent., Seattle, Wash., Proc. Rep. A study of four major groundfish fisheries of the Bering Sea. Ph.D. Dissertation, Univ. V/ashington, Seattle. Walleye pollock 549 Low, L. L., and J. Akada 1978 Atlas of groundfish catch in the northeastern Pacific Ocean. Nat. Mar. Fish. Serv., Northwest and Alaska Fish. Cent., Seattle, Wash., Proc. Rep. Lowry, L. F., 1978 Maeda, T. 1972 K. J. Frost, and J. J. Bums Trophic relationships among ice in- habiting phocid seals. Environmental assessment of the Alaskan continental shelf. U.S. Dep. Comm. NOAA/ OCSEAP, Ann. Rep. 1:161-230. Fishing grounds of the Alaska pollock. (In Japanese, English summary.) Bull. Japanese Soc. Sci. Fish. 38: 362-71. Nunnallee, E. 1978 Okada, K. 1977 Report on observations aboard the Japanese research vessel Tomi Maru 52 during a Bering Sea (Aleutian Basin) pollock survey conducted in June-July, 1978. Nat. Mar. Fish. Serv., Northwest and Alaska Fish. Cent., Seattle, Wash., Unpub. MS. Preliminary report of acoustic survey on pollock stocks in the Aleutian Basin and adjacent waters in summer of 1977. (In Japanese, English summary.) Fishery Agency of Japan, Tokyo. Unpub. MS. » Maeda, T. and H. Hirakawa 1977 Spawning grounds and distribution pattern of the Alaska pollock in the eastern Bering Sea. (In Japanese, English abs.) Bull. Japanese Soc. Sci. Fish. 43: 39-45. McLain, D. R., and F. Favorite 1976 Anomalously cold winters in the southeastern Bering Sea, 1971-75. Mar. Sci. Comm. 2:299-334. McRoy, C P., J. J. Goering, and W. E. Shiels 1972 Studies of primary production in the eastern Bering Sea. In: Biological oceanography of the northern North Pacific Ocean, A. Y. Takenouti, ed., 199-216. Idemitsu Shoten, Tokyo. 1978 Mito, K. 1974 Food relationships among benthic fish populations in the Bering Sea on the Theragra chalcogramma fishing ground in October and November of 1972. (In Japanese.) M.S. Thesis, Hokkaido University, Hokkaido, Ja- pan. NPFMC 1978 Fishery management plan and draft environmental impact statement for the groundfish fishery in the Bering Sea/ Aleutian Island area. N. Pac. Fish. Man. Coun., Anchorage, Alaska, Proc. Rep. Preliminary report of acoustic survey and mid-water trawl on pollock stock on the Aleutian Basin and adjacent waters in summer of 1978. (In Japanese, English summary.) Fishery Agency of Japan, Tokyo. Unpub. MS. Yamaguchi, T. Sasaki, and K. Waka- Trends of groundfish stocks in the Bering Sea and the northeastern Pacific based on additional prelimi- nary statistical data in 1978. Fishery Agency of Japan, Far Seas Fisheries Research Laboratory. Unpub. MS. Parsons, T. R., M. Takahashi, and B. Hargrave 1977 Biological oceanographic processes. Pergamon, N.Y. Okada, K., H. bayashi 1979 Pereyra, W. T. 1976 J. E. Reeves, and R. G. Bakkala Demersal fish and shellfish resources of the eastern Bering Sea in the baseline year 1975. Nat. Mar. Fish. Serv., Northwest and Alaska Fish. Cent., Seattle, Wash., Proc. Rep. Pileggi, J., and B. G. Thompson 1979 Fisheries of the United States, 1978. U. S. Dep. Comm., NOAA, Nat. Mar. Fish. Serv., Current Fish. Statistics No. 7800. 550 Fisheries oceanography Pope, J. G. 1972 An investigation of the accuracy of virtual population analysis using co- hort analysis. Inter. Comm. Northw. Atlantic Fish., Res. Bull. 9: 65-74. Sharma, G. D. 1974 Pruter, A. T. 1973 1976 Development and present status of bottomfish resources in the Bering Sea. J. Fish. Res. Bd. Can. 30: 2373-85. Soviet fisheries for bottom fish and herring off the Pacific and Bering Sea coasts of the United States. Mar. Fish. Rev. 38(12): 1-14. Serobaba, 1. 1. 1968 Spawning of the Alaska pollack [sic] Theragra chalco gramma (Pallas) in the northeastern Bering Sea. J. Ichthyol. 8: 789-98. 1970 Distribution of walleye pollock Ther- agra chalcogramma (Pallas) in the eastern Bering Sea and prospects of its fishery. In: Soviet fisheries in- vestigations in the northeastern Pacif- ic, P. A. Moiseev, ed., 5:442-51. U.S. Dep. Comm./NTIS. (Transl. Israel Prog. Sci. Transl., 1972.) 1971 O razmnozhenii mintaya (Theragra chalcogramma Pallas) v vostochnoi chasti Beringova morya (On the reproduction of walleye pollock (Theragra chalcogramma Pallas) in the eastern part of the Bering Sea). Izv. Tikhookean. Nauchno-issled. Inst. Morsk. Rybn. Khoz. Okeanogr. 75: 47-55. In Russian. (Transl. Dep. of Sec. of State of Canada, 1973.) 1974 Spawning ecology of the walleye pollock (Theragra chalcogramma) in the Bering Sea. J. Ichthyol. 14: 544-52. 1977 Data on the population structure of the walleye pollock, Theragra chalco- gramma, from the Bering Sea. J. Ichthyol. 17: 219-31. Shew, D. M. 1978 Contemporary depositional environ- ment of the eastern Bering Sea. In: Oceanography of the Bering Sea, D.W. Hood and E.J. Kelley, eds., 517-40. Inst. Mar. Sci., Occ. Pub. No. 2, Univ. Alaska, Fairbanks. Pollock fecundity study; preliminary report. Nat. Mar. Fish. Serv., North- west and Alaska Fish. Cent., Seattle, Wash., Unpub. MS. Smith, R. L., A. C. Paulson, and J. R. Rose 1978 Food and feeding relationships in the benthic and demersal fishes of the Gulf of Alaska and Bering Sea. Envi- ronmental assessment of the Alaskan continental shelf. U.S. Dep. Comm., NOAA/OCSEAP, Final Rep. 1:33- 107. Sverdrup, H. U., M. W. Johnson, and R. H. Fleming 1946 The oceans. Prentice-Hall, N.Y. Takahashi, Y., and H. Yamaguchi 1972 Stock of the Alaska pollock in the eastern Bering Sea. (In Japanese, English summary) Bull. Japanese Soc. Sci. Fish. 38: 389-99. Takakura, T. 1954 The behavior of the spawning pollock schools recorded by fish-detector. (In Japanese, English abs.) Bull. Japanese Soc. Sci. Fish. 20: 10-12. Takenouti, A. Y., and K. Ohtani 1974 Currents and water masses in the Bering Sea: a review of Japanese work. In: Oceanography of the Bering Sea, D. W. Hood and E. J. Kelley, eds., 39-57. Inst. Mar. Sci., Occ. Pub. No. 2, Univ. Alaska, Fairbanks. Walleye pollock 551 Waldron, K. D., and B. M. Vinter 1978 Ichthyoplankton of the eastern Bering Sea. Nat. Mar. Fish. Serv., Northwest and Alaska Fish. Cent., Seattle, Wash., Proc. Rep. Yamaguchi, H. 1979 Report of multi-vessel trawl survey on bottomfishes in the eastern Bering Sea continental shelf in 1978. Far Seas Fish. Res. Lab., Shimizu, Japan, Data rep. Yamaguchi, H., and Y. Takahashi 1972 Growth and age estimation of the Pacific pollock, Theragra chalco- gramma (Pallas), in the eastern Bering Sea. Bull. Far Seas Fish. Res. Lab., Shimizu, Japan 7: 49-69. Zverkova, L. M. 1969 Spawning of the Alaskan pollock {Theragra chalcogramma (Pallas)) in the waters of the west coast of Kam- chatka. Prob. Ichthyol. 9:205-9. L Population Characteristics and Ecology of Yellowf in Sole Richard G. Bakkala National Oceanic and Atmospheric Administration, National Marine Fisheries Service Northwest and Alaska Fisheries Center Seattle, Washington ABSTRACT Yellowfin sole is a major component of the demersal fish community of the eastern Bering Sea continental shelf; in 1975, it made up 64 percent of the total flounder biomass and 23 percent of the total sampled fish biomass. It has been fished commercially on a regular annual basis since 1954. Intense exploitation in 1959-62 (when 1.6 million mt were harvested) was the apparent cause of a severe decline in abundance; an estimated virgin exploitable biomass of 1.6- 2.0 million mt fell to 0.8 million mt in 1963. The biomass remained at this lower level through the early 1970's, but since 1972 it has shown a marked increase and may have reached 1.4 million mt in 1978. This increase stems from the recruitment of a series of abundant year-classes originat- ing in the years 1966-70. Yellowfin sole migrate seasonally from outer continental shelf and slope waters (>100 m) occupied in winter and early spring to inner shelf waters (15-75 m), where spawning occurs in summer. Offshore migrations by adults in fall and winter are an apparent response to the ice cover and cold water temperatures that characterize inner and central shelf waters in winter. Unlike the adults, the young remain in shallow nearshore nursery areas throughout their first few years of life. They begin to disperse to more offshore waters at three to five years of age, and by five to eight years of age they occupy much the same waters as older fish. Yellowfin sole grow slowly and at eight to nine years of age reach a length of 25 cm, which is about the average size of fish taken in the commercial fishery. They are first recruited to the fishery at four to five years and are fully re- cruited at seven years, the age when the exploitable stock reaches its maximum biomass. Growth declines between ages seven and eight, when a high proportion of the population reaches sexual maturity. The rapid increase in mortality after the ninth year is apparently due to spawning stress. Climatic variations in the eastern Bering Sea since the mid-1960's have produced some apparent changes in year-class abundance and in distributions and migrations of yellowfin sole. Abundant year -classes were produced in years of rela- tively warm climatic conditions; year-classes produced in cold years had relatively low abundance. Evidence also suggests that extensive ice cover may delay the start of spring inshore migrations, and residual cold water in central shelf regions may alter patterns of summer migrations and distributions. INTRODUCTION Yellowfin sole (Limanda aspera) is a right-eyed flounder of the family Pleuronectidae (Fig. 34-1), one of two species of Limanda found in the eastern Bering Sea, the other being the longhead dab (L. proboscidea). Yellowfin sole is a major component of the Bering Sea ichthyofauna; the longhead dab is a minor nearshore species. Yellowfin sole is a relatively small flounder; commercial catches in recent years reflect an average size of 25 cm and 175 g. Fish of 25 cm are eight or nine years old; however, maximum lengths of 53 cm (Ellson et al. 1950) and ages in excess of 20 years (Wakabayashi 1975) have been reported. The yellowfin sole is limited in distribution to continental shelf and slope waters of the North Pacific Ocean, the Bering Sea and, to a limited extent, the Chukchi Sea (Fig. 34-2). It ranges along the Pacific Coast of North America from Barclay Sound (about 49° N), Vancouver Island, British Columbia, northward into the Chukchi Sea (Hart 1973), and along the Asian coast from the Gulf of Anadyr southward to the east and west coasts of Hokkaido Island, Japan, and along the Asian mainland in the Okhotsk Sea and the Sea of Japan to about 35°N off South Korea (Fadeev 1970a). Its bathymetric range is from about 5 to 360 m, although in some regions (e.g., the Gulf of Alaska) it is limited to continental shelf waters of generally 100 m or less. The deepest recorded occurrence (360 m) is in the eastern Bering Sea. Yellowfin sole reach their maximum abundance in the eastern Bering Sea; they are by far the most abundant flounder in this region. An extensive trawl 553 554 Fisheries oceanography Figure 34-1. Yellowfin sole (illustration from Hart 1973, courtesy Environment Canada, Ottawa). survey in 1975 (Pereyra et al. 1976) found the biomass to be approximately 1.0 million mt, repre- senting 64 percent of the total flounder biomass on the eastern Bering Sea continental shelf and 23 percent of the total sampled fish biomass. After walleye pollock (Theragra chalcogramma), it was the next most abundant demersal fish. Although it is the predominant flounder in the eastern Bering Sea, yellowfin sole becomes a rela- tively minor species among the flounder complex in the Gulf of Alaska. It ranked eighth in relative abundance in flounder catches in the northern Gulf of Alaska (Hughes 1974) and is only occasionally taken in southeastern Alaska waters (Schaefers 1951, Ellson and Livingston 1952). The abundance of yellowfin sole is lower in U.S.S.R. waters than in the eastern Bering Sea, but concentrations are large enough in certain regions to attract commercial fisheries (see Fig. 34-2). Yellowfin sole and other groundfish of the eastern Bering Sea were first exploited commercially by JapEin, initisJly by exploratory vessels in 1930 and then by a moth ership -catcher boat operation off Bristol Bay in 1933-37 and 1940-41 (Forrester et al. 1978). In the 1933-37 period poUock and flounders were taken for reduction to fish meal; catches ranged up to 43,000 mt annually. In the second period the fishery targeted on yellowfin sole for human con- sumption, with catches of all species ranging from 9,600 to 12,000 mt in the two years. After World War II, Japanese distant water fisheries resumed operations in the eastern Bering Sea in 1954 with mothership and independent trawler fleets targeting on yellowfin sole. In the initial period of this fishery (1954-57), catches ranged from about 12,500 to 24,700 mt (Table 34-1). In 1958 the U.S.S.R. also entered the fishery, and from 1959 to 1962 the exploitation of yellowfin sole was very intense with annual catches ranging from 185,000 to 554,000 mt. There is evidence that catches of this magnitude severely reduced the abundance of the stock (Wakabayashi et al. 1977). In the subsequent period of 1963-71, catches were much lower, ranging from 53,800 to 167,100 mt. Catches declined even further in the period 1972-77 as the U.S.S.R. discontinued its target fishery for yellowfin sole, presumably because of low catch rates. Catches in this period, primarily by Japan, ranged from 42,200 to 78,200 mt. In 1978 the U.S.S.R. resumed its target fishery for yellowfin sole; preliminary data indicate that all-nation catches rose to about 139,100 mt. Yellow fin sole 555 120" 140° 160° 180° 160° 140° 120° 120" 140° 160° 180 160° 120° Figure 34-2. Overall distribution and areas of commercial fishing for yellowfin sole. POPULATION CHARACTERISTICS Stock biomass The virgin biomass of exploitable yellowfin sole (age six and older) in the eastern Bering Sea has been estimated to range from 1.3 to 2.0 million mt (Alver- son et al. 1964, Wakabayashi 1975). A virtual population analysis (Wakabayashi 1975) revealed that by 1963 the exploitable population was reduced to approximately 40 percent (801,500 mt) of the maximum estimate of virgin size. This decline occurred during the period of intense exploitation by Japan and the U.S.S.R., when 1.6 million mt were harvested in 1959-62. The analysis also showed some recovery of the resource in the mid-1960's, when the exploitable population was estimated to reach about 950,000 mt (Table 34-2). This was followed by another decline to about 790,000 mt in 1970. A cohort analysis covering the period 1964-75 (Wakabayashi et al. 1977) showed trends in abun- dance similar to those indicated by the virtual popu- lation analysis, but yearly biomass estimates varied to some degree between the two studies (see Table 34-2). The cohort analysis showed the exploitable population reaching a low of 604,000 mt in 1969 and then increasing to 910,000 mt in 1975. This increase in abundance has also been shown by indices of relative abundance from the commercial fisheries and from research vessel surveys (Bakkala et al. 1979). The estimated biomass in 1975 based on the cohort analysis was similar to that obtained from an exten- sive trawl survey of the eastern Bering Sea in the same year. Indices of relative abundance and biomass esti- mates from trawl surveys have shown that abundance continued to increase through 1978 (Bakkala et al. 1979). The exploitable biomass may have reached 1.4 million mt by 1978, at least 70 percent of the estimated virgin population size. The primary reason for the increase has been the recruitment of a series of relatively strong year-classes originating in the years 1966-70 (Fig. 34-3). The cumulative contribu- tion of these year-classes may have resulted in a doubling of the overall exploitable biomass over the period of 1973-78. Stock structure Yellowfin sole form dense concentrations on the outer continental shelf of the eastern Bering Sea in 556 Fisheries oceanography TABLE 34-1 Annual catches of yellowfin sole in the eastern Bering Sea (east of 180° and north of 54°N) in metric tons. Year Japan USSR ROK' Total 1954 12,562 0 0 12,562 1955 14,690 0 0 14,690 1956 24,697 0 0 24,697 1957 24,145 0 0 24,145 1958 39,153 5,000 0 44,153 1959 123,121 62,200 0 185,321 1960 360,103 96,000 0 456,103 1961 399,542 154,200 0 553,742 1962 281,103 139,600 0 420,703 1963 20,504 65,306 0 85,810 1964 48,880 62,297 0 111,177 1965 26,039 27,771 0 53,810 1966 45,423 56,930 0 102,353 1967 60,429 101,799 0 162,228 1968 40,834 43,355 — 84,189 1969 81,449 85,685 — 167,134 1970 59,851 73,228 — 133,079 1971 82,179 78,220 — 160,399 1972 34,846 13,010 — 47,856 1973 75,724 2,516 — 78,240 1974 37,947 4,288 — 42,235 1975 59,715 4,975 — 64,690 1976 52,668 2,908 625 56,201 1977 58,139 284 55 58,478 1978 62,736 76,300 69 139,105 ^ 0 indicates no fishing, — indicates fishing but no reported catch. winter. The largest of these is formed in the vicinity of Unimak Island and the second largest west of St. Paul Island; other lesser winter concentrations have also been recognized— one may be south or east of St. George Island and the other, consisting of small fish, in Bristol Bay (Fadeev 1970a, Wakabayashi 1974). Japanese tagging studies have shown that the winter- ing concentrations near St. George Island and Unimak Island combine in spring before onshore migration to areas off Bristol Bay (Wakabayashi et al. 1977). The wintering concentration of small fish in Bristol Bay is also thought to be a part of the Unimak Island /St. George Island group, perhaps representing a segment of the juvenile part of the population. 300 150 0 150 0 150 0 150 0 150 - 0 150 0 150 - 150 AGE 6 n ^ 69 ^ n n ^ n ^""^'V fi 68 n 66 n n rn II n AGE 8 n ^« fi r 66 67 n n II AGE 9 67 68 ^ 66 n AGE 10 67 68 66 _ n n n AGE 11 66 67 n n n n ^ t-i n r-i n AGE 13-20 n n n n n r 1,200 - 900 600 - 300 - ALL AGES COMBINED 1973 1974 1975 1976 Year 1977 1978 Figure 34-3. Biomass estimates for the exploitable popu- lation of yellowfin sole in the southeast Bering Sea. Year- classes for certain ages are shown with appropriate bars. The second largest wintering concentration, located west of St. Paul Island, appears to remain relatively independent of the Unimak Island/St. George concentrations throughout the year. This group probably migrates inshore between St. Paul and St. Matthew islands and forms concentrations in the vicinity of Nunivak Island in summer. The apparently independent movements and distributions of the St. Paul Island and Unimak Island /St. George Island populations have suggested the existence of independent spawning stocks of yellowfin sole in the eastern Bering Sea— a northern stock (St. Paul group) and a southern stock (Unimak Island/St. George Island/ Bristol Bay group) (Waka- Yellowfin sole 557 bayashi 1974, Wakabayashi et al. 1977). Japanese tagging studies through 1973 indicated only limited mixing of the two groups (Wakabayashi 1974) and some studies suggested differences in biological characteristics (such as growth rates, length-weight relationships, and egg diameters) between populations in the proposed north and south stock areas (Kash- kina 1965, Wakabayashi 1974). Other studies of biological characteristics, however, have not supported the two-stock concept (Fadeev 1970a, Wakabayashi 1974). In addition, returns from Japanese tagging studies since 1973 have shown greater intermixing of fish between the proposed stock areas (see Fig. 34-4 for results of Japanese tagging studies through 1977). Results of biochemical genetic studies using electrophoretic techniques have provided no evidence of major genetic differences between fish from the proposed north and south stock areas; however, the samples on which the studies were based were not taken during the spawning season, when the two stocks (if they exist) would be most clearly segre- gated (Grant et al. 1978). In addition, the finding of similar gene frequencies between fish from the two areas is not positive evidence that differentiation has not occurred. For example, if the populations had become segregated in fairly recent times, genetic differences might not be easily detected with present electrophoretic methods. Thus, although accumulating evidence from tagging and genetic studies tends to suggest the presence of a single spawning stock of yellowfin sole in the eastern Bering Sea, this question has not been entirely resolved. Size composition Yellowfin sole taken by U.S.S.R. research vessels in the eastern Bering Sea in the period 1958-64 ranged from 5 to 49 cm in length (Fadeev 1970a). The average size of fish taken by the commercial fishery was 26-27 cm in 1958 and 1959, years immediately preceding the period of intense exploitation. The dominant sizes according to Fadeev (1970a) were 24-30 cm. In recent years, the average size has been about 25 cm in the commercial fishery. Length data taken on large-scale Northwest and Alaska Fisheries Center (NWAFC) trawl surveys in 1975, 1976, and 1978 (Fig. 34-5) show fish ranging from 5 to 44 cm and averaging 22.2-22.6 cm for combined sexes. More than 96 percent of the sur- veyed population were less than 30 cm long. A relatively high proportion of fish in the 12-16 cm range in 1978 represents the recruitment of the apparently strong year-class of 1973. TABLE 34-2 Estimates of the exploitable biomass (age six and older) of yellowfin sole in the eastern Bering Sea in 10^ metric tons. Virtual population Cohort Trawl Year analysis' analysis^ surveys^ 1959 2035.1 1960 1924.0 1961 1521.6 1962 1054.6 1963 801.5 1964 856.2 912.5 1965 872.8 960.7 1966 948.1 969.0 1967 946.5 879.0 1968 862.7 635.4 1969 880.8 604.0 1970 786.7 720.8 1971 648.2 1972 660.0 1973 849.1 1974 761.4 1975 910.2 991.9 1976 1977 1978 1414.6 * Wakabayashi (1975) ^Wakabayashi et al. (1977) ^Pereyra et al. (1976); Bakkala et al. (1979) Age composition The observed age range of yellowfin sole since the early 1970's has been 2-19 years for males and 2-21 years for females (Pereyra et al. 1976, Bakkala and Smith 1978); however, about 97 percent of the sampled population were younger than 13. Age data collected during NWAFC research vessel surveys and from the commercial fishery show that the population underwent marked changes in age structure during 1973-78 (Fig. 34-6). In 1973, a high proportion (86 percent) of the population sampled by research vessel surveys was made up of fish aged seven or younger. By 1978 these age groups repre- sented only 36 percent of the sampled population, and fish of ages 8-11 represented 54 percent. These changes are the result of the recruitment and ad- vancement through the population of the series of 558 Fisheries oceanography 175 170 165w 160 OWFIN SOLE O Released in 1970 # Released in 1971 Recovery location -58N 175 170 165w Figure 34-4. Release and recovery locations of yellowfin sole tagged by Japanese research vessels in 1970 and 1971 and recovered by commercial fishing vessels (from Wakabayashi et al. 1977). Numbers of fish released are shown at each tagging location; a few fish (as indicated by light dashed lines) were released by commercial vessels. The heavy dashed line between St. George Island and Cape Avinof represents the line separating proposed north and south stock areas. relatively strong year-classes originating in the years 1966-70. Research vessel data indicate that the 1971 and 1972 year -classes may not be as abundant as those of 1966-70 but the 1973 year-class may be of above average strength. Size and age at maturity and recruitment to the fishery Male yellowfin sole begin to mature at a length of 10.5 cm, females at 18.5 cm (Wakabayashi 1974; see also Fig. 34-7). The length at which 50 percent of the population is mature is 12.8 cm for males aind 25.2 cm for females. All males are mature at 25 cm and most females at 30 cm. On the basis of samples collected in 1959-64, Fadeev (1970a) reported that males first begin to mature at 12 cm and females at 16-18 cm but that 50 percent of the population reached maturity at 16-18 cm for males and 30-32 cm for females. Wakabayashi (1974) suggests that fish may have matured at a smaller size in 1973 because the population was less abundant than in 1959-64. Females begin to mature at age six and 50 percent reach maturity at about age nine; all the females were mature at age 15 (Fig. 34-7). Yellowfin sole first become recruited to the fishery at 13 or 14 cm, which corresponds to age four or five. They become fully recruited to the fishery at age Yellowfin sole 559 SEXES COMBINED 20 30 40 10 20 30 40 Length (cm) 10 20 30 40 Figure 34-5. Length composition of yellowfin sole of the eastern Bering Sea as determined by research vessel surveys in 1975, 1976, and 1978. 100 2 75 (0 S *. 50 c 0) u 5 25 o. r MALES mT 10 15 lOOr V 3 m S +.» c 0) u « a. 75- 50- 25 10 15 l7Tr*''!Mf:WMt W^. 20 25 Length (cm) ':'f*ft'-ft''f%>>l%^ 30 35 ■ FEMALES r r-T' m 20 25 Length (cm) 30 35 I 9 1000 U.S. RESEARCH VESSEL SURVEY DATA U.S. OBSERVER DATA 1973 67 66 Innnnr 67 Hnnnn, by Jl nnnnnnr DO JD nnnn„ ^ Q Hnnrnr Hr-ii-in 00 I Innr-ir^,-, 70 69 r-i IL -^n bb nn___ xiD bb On. n be -„n no- lo 12 14 2 4 6 Age 10 12 14 Figure 34-6. Age composition of yellowfin sole as shown by data from NWAFC research vessel surveys in June- August and by U.S. observer samples from the Japanese flounder fishery in September-December. Year-classes for certain ages are shown above appropriate bars. 100 = 75 (0 S *- 50 c 0) u O 25 Q. FEMALES J] 10 15 20 Age (yr) 25 Figure 34-7. Maturity composition of yellowfin sole from the eastern Bering Sea in May-June 1973, by length for males, by length and age for females (after Wakabayashi 1974). seven (Laevastu and Favorite 1978), about the age (seven or eight years) when the exploitable stock reaches its maximum size in weight (Wakabayashi 1975). Length-weight relationships, growth, and mortality Length-weight relationships obtained in 1975 and 1976 show that females are only shghtly heavier than males (approximately 1-10 percent) from lengths of 20-30 cm (Table 34-3). Von Bertalanffy growth curves (Fig. 34-8) and their parameters (Table 34-4) illustrate the similarity of growth characteristics in males and females. Six-year means of observed lengths and calculated weights (Table 34-5) indicate 560 Fisheries oceanography TABLE 34-3 Length-weight relationships for yellowfin sole of the eastern Bering Sea (data of Pereyra et al. 1976, Bakkala and Smith 1978). Sex Year Time Period Sample Size Length range (cm) Length-weight coefficients Predicted 10 cm weight (g) 20 cm at length 30 cm Male Female 1975 1976 1975 1976 Aug-Oct Apr-Jun Aug-Oct Apr-Jun 701 213 844 293 7-40 9-33 8-40 10-37 0.0128 0.0208 0.0102 0.0168 2.956 2.779 3.035 2.870 11.6 12.4 11.1 12.4 89.8 85.6 90.6 91.1 297.6 264.2 310.2 291.9 TABLE 34-4 Parameters of the von Bertalanffy growth curve for yellowfin sole of the eastern Bering Sea (data of Pereyra et al. 1976, Bakkala and Smith 1978). Year Number of age readings Age range Length range (cm) Standard error of curve fit Parameter Sex L„ K to Male Female 1975 1976 1975 1976 609 507 707 600 0,8-15 0,4-12 0,8-15 0,4-14 0,17-30 0,13-33 0,19-39 0,13-36 1.50 0.31 0.80 0.66 40.79 31.88 40.28 32.23 0.11 0.17 0.11 0.18 0.22 -0.02 -0.09 0.10 that annual increments of length increase from age four to age six and then decline at older ages. Annual increments of weight also increase from age four to age six, but are relatively constant from seven to thirteen. Laevastu and Livingston (1978) have also shown a decline in the growth rate between ages seven and eight. They attribute this decline to the fact that a large portion of the population reaches sexual maturity at this age and energy is diverted from growth to development of sex products. Mortality decreases rapidly v\rith age in juvenile year-classes and reaches a minimum between ages five and nine, about the age of entry into the fishery (Laevastu and Livingston 1978). Mortality increases rapidly after the ninth year, when a high proportion of the population reaches sexual maturity. The higher mortality at ages 10 and above may be due to spawning stress (Laevastu and Livingston 1978). Wakabayashi (1975), using the method of Alverson and Carney (1975), estimated instantaneous natural mortality (M) for fish age four and older as 0.25, corresponding to an annual mortality rate of 22 percent. POPULATION ECOLOGY Species interactions An important predator of yellowfin sole in the eastern Bering Sea according to Novikov (1964) is the halibut (Hippoglossus stenolepis). He found a close relationship between the distribution of halibut and yellowfin sole during the summer and autumn and suggested that the movements of halibut are governed to a large degree by the movements of its principal prey, yellowfin sole. The incidence of yellowfin sole in halibut stomachs in his studies ranged from 33 to 70 percent and, in terms of composition by weight, from 30 to 55 percent over various areas of the southeastern Bering Sea. Although there must be other predators of yellowfin sole, particularly during larval and juvenile stages, they have not been docu- mented. Yellowfin sole are capable of feeding on a variety of animals, from strictly benthic forms such as clams and polychaete worms to zooplankton (mysids and euphausiids) to pelagic fish (capelin and smelt) Yellowfin sole 561 TABLE 34-5 Six-year means of observed lengths and calculated weights at age for yellowfin sole of the eastern Bering Sea from NWAFC survey data of 1973-78. Mean Annual Mean Annual length increment weight increment Age (cm) (cm) (g) (g) 3 12.0 19.5 4 14.1 2.1 32.3 12.8 5 16.6 2.5 51.7 19.4 6 19.3 2.7 81.5 29.8 7 21.0 1.7 103.9 22.4 8 22.4 1.4 126.4 22.5 9 23.9 1.5 154.2 27.8 10 25.2 1.3 181.0 26.8 11 26.4 1.2 206.3 25.3 12 27.4 1.0 230.3 24.0 13 28.7 1.3 265.2 34.9 14 29.1 0.4 278.3 13.1 15 29.4 0.3 286.6 . 8.3 40-1 YELLOWFIN SOLE 1976 Spring Trawl Survey Age (yr) Figure 34-8. Von Bertalanffy growth curves for yellowfin sole taken during spring 1976 (Bakkala and Smith 1978). Symbols indicate the mean length at each age. (Pereyra et al. 1976). About 50 different taxa have been found in stomachs of yellowfin sole in the eastern Bering Sea (Skalkin 1963). The kinds of organisms consumed vary by season, area, and size of fish. Although feeding generally stops in winter, instances of fairly intense winter feeding have been recorded (Fadeev 1970a). During the onshore migrations in May and June 1971, 73 percent of the fish that had wintered near Unimak Island were feeding, but feeding intensity was lower for fish that had wintered near St. George Island (0.05 percent), St. Paul Island (19 percent), and in Bristol Bay (0 percent) (Wakabayashi 1974). They feed more intensely as they move onto the central shelf; diet varies by region, apparently depending on the availa- bility of food organisms. Fadeev (1970a) suggests that they depend more on zooplankton when benthic organisms are scarce. Contents of 2,357 stomachs taken over a broad area of the eastern Bering Sea (Table 34-6) show that the primary food items, representing 65 percent of stomach content by weight, were bivalves, amphi- pods, polychaete worms, and echiuroid worms. Polychaete worms and amphipods were the principal food items in smaller fish (10-20 cm), polychaete worms and bivalves and then echiuroid worms and amphipods in larger fish (20-30 cm), and bivalves and echiuroid worms in fish longer than 30 cm. Recurrent group analyses (Fager 1957, 1963; Fager and Longhurst 1968) have been used to demonstrate species associations within the demersal community of the eastern Bering Sea (Kihara 1976, Mito 1977, Pereyra et al. 1976, Bakkala and Smith 1978). The procedure identifies species relationships on the basis of co-occurrence within samples. When joint occur- rences are equal to or exceed 0.50, the species are considered to show affinity. These four studies showed wide variation in the species that co-occurred with yellowfin sole. This variation might arise from differences in areas, time periods, and years of the surveys from which the data were analyzed. However, certain species co-occurred more consistently with yellowfin sole than others (Table 34-7): Alaska plaice were found to co-occur in all studies and rock sole and Pacific herring in more than one study. Pacific halibut was not one of the species showing close association with yellowfin sole based on recurrent group analysis as suggested in the food habits study of halibut by Novikov (1964). Distribution and seasonal movements Adults Wakabayashi (1974) summarized Japanese com- mercial catch data from December 1967 to October 562 Fisheries oceanography TABLE 34-6 Stomach contents (in grams) by size group of yellowfin sole collected in the eastern Bering Sea in 1970 (Wakabayashi 1974). Size Group Food Item 101-200 mm 201-300 mm >300 mm Tota Gadidae _ 60.3 26.8 87.1 Osmeridae — 12.7 — 12.7 Ammodytidae — 12.5 56.4 68.9 Other Pisces — 36.4 18.4 54.8 Amphipoda 22.6 180.2 45.3 • 248.1 Euphausiacea 1.6 61.7 38.1 101.4 Macrura 3.7 22.8 24.1 50.6 Mysidacea 0.4 — 0.2 0.6 Brachyura — 39.4 7.8 47.2 Anomura 1.7 51.5 18.3 71.5 Crangonidae 0.3 10.0 0.8 11.1 Polychaeta 31.4 360.9 63.1 455.4 Cephalopoda — — 1.7 1.7 Bivalvia 6.5 403.6 553.1 963.2 Gastropoda — 1.3 10.4 11.7 Ophiuroidea 2.3 36.1 1.4 39.8 Scutellidae 6.4 33.2 17.2 56.8 Echiurida 9.2 245.4 398.4 653.0 Ascidia 0.2 13.7 4.2 18.1 Holothuroidea — 34.8 3.5 38.3 Sand' 7.2 91.9 13.7 112.8 Others 12.8 108.2 163.6 284.6 Indistinct 12.9 64.2 27.7 104.8 Total 119.2 1,878.3 1,496.7 3,494.2 No. of Stomachs 275 1,708 374 2,357 * Possibly tubes of Polychaeta. 1968 to illustrate seasonal changes in fishing grounds. These data probably also generally illustrate seasonal changes in distribution of major concentrations (Fig. 34-9). In winter months (December-March), catches were concentrated near the 200 m isobath in the area west of the PribUof Islands; but from the Pribilof Islands to Unimak Island, they occurred between 100-200 m. In May, major catches were made near the 100 m isobath or in even shallower areas. By June, the largest catches occurred between 50 and 100 m and in July and August near the 50 m isobath. In September the main catches again shifted to deeper water (50-100 m) and in October the greatest proportion of the catches near Unimak Island were taken at the 200 m isobath. More recent data from a large-scale NWAFC trawl survey in April-June 1976 illustrate the apparent routes of movement during spring onshore migrations of the two major wintering concentrations north of Unimak Island and west of St. Paul Island. The spring of 1976 was unusually cold (Bakkala and Smith 1978) and the ice edge in April 1976 approxi- mated an extreme southern location based on obser- vations of Potocsky (1975) over the seventeen-year period 1954-70. Because of the extensive ice cover, research vessel fishing operations in April were mainly limited to the outer shelf southeast of the Pribilof Islands. The Unimak Island wintering con- centration was clearly evident north of Unimak Island immediately offshore of the ice edge (Fig. 34-10). Yellowfin sole were highly concentrated in this area with catch rates ranging up to 6,800 kg/km trawled. Bottom temperatures near the leading edge of this concentration were near 0 C; the densest portion of the concentration was in bottom temperatures of 0.1-2.0 C. In May, as the ice edge receded, the Unimak Island wintering concentration shifted eastward towards Bristol Bay, apparently moving with the receding ice edge and 0 C isotherm (Fig. 34-10). In June (Fig. 34-11), a number of independent concentrations were observed east of the location of the principal concen- tration observed in May. The nearshore concentra- tion in Bristol Bay was made up of small fish not encountered in offshore waters. Fadeev (1970a) believed that a concentration of juvenile fish he observed in this area in spring indicated over- wintering in Bristol Bay. Survey data in August to October 1975 (Fig. 34-11) and from earlier years (see Figs. VIII-23 to 29 of Pereyra et al. 1976), indicate that yellowfin sole form a large summer concentration in outer Bristol Bay between the 40 and 100 m depth contours. Tagging data suggest that these fish winter off Unimak Island and St. George Island (see Fig. 34-4). The second major wintering concentration, located west of St. Paul Island, may have been under the ice in April 1976 (see Fig. 34-10). This assumption is based on its location in May, in an area covered by ice in April. In June, the concentration of fish between St. Paul and St. Matthew islands may repre- sent a migration of this wintering group towards Nunivak Island; Japanese tagging data show that most of these fish are distributed in the vicinity of Nunivak Island in summer (see Fig. 34-4). Yellowfin sole 563 TABLE 34-7 Species showing close association with yellowfin sole as indicated by recurrent group analysis. Authority Season Years of study Species showing affinity with yellowfin sole Kihara (1976) Summer 1966-71, 1974 Mito (1977) Pereyra et al. (1976) Winter Summer 1972, 1974-75 1975 Bakkala and Smith (1978) Spring 1976 Alaska plaice (Pleuronectes quadrituberculatus) Rock sole (Lepidopsetta bilineata) Flathead sole (Hippoglossoides elassodon) Pacific cod (Gadus macrocephalus) Walleye pollock (Theragra chalcogramma) Cottidae Agonidae Alaska plaice (P. quadrituberculatus) Rock sole (L. bilineata) Yellow Irish Lord (Hemilepidotus jordani) Plain sculpin (Myoxocephalus jaok) Alaska plaice (P. quadrituberculatus) Pacific herring (Clupea pallasi) Alaska plaice (P. quadrituberculatus) Pacific herring (C. pallasi) Sturgeon poacher (Podothecus acipenserinus) Capelin (Mallotus villosus) Early life stages Knowledge of the location and timing of spawning of yellowfin sole is based mainly on catches of eggs during plankton surveys; spawning adults have rarely been observed. Spawning begins in early July and probably ends in September (Musienko 1963). Eggs were observed over a broad area of the eastern Bering Sea shelf from off Bristol Bay to off Nunivak Island (Fig. 34-12), and densities indicated that spawning was most intense south and southeast of Nunivak Island. Depths of spawning ranged from 15 to 75 m. Kashkina (1965) also observed spawning between Nunivak Island and St. Lawrence Island (see Fig. 34-12). My observations of females with "running" eggs in July 1979 indicate that spawning also occurs in northern Bristol Bay. Eggs are pelagic, and laboratory studies of eggs from the Okhotsk Sea indicate that hatching occurs in about four days at 13 C (Pertseva-Ostroumova 1961). The lower threshold temperature for success- ful egg development was 4 C. Eggs have been en- countered in the eastern Bering Sea at temperatures of 6.4-11.4 C (Kashkina 1965, Musienko 1963). At hatching, the prolarvae are small (2.2-3.1 mm) and transparent, have large yolk sacs, and are capable of swimming (Pertseva-Ostroumova 1961, Musienko 1963). About three days after hatching, the yolk sac is absorbed and the larvae begin to feed; at this stage they range in length from 3.3 to 3.8 mm. The time between hatching and metamorphosis to the juvenile stage is unknown. Partially metamor- phosed young have been found in plankton hauls at 16.5-17.4 mm in length; since larger juveniles have not been collected in plankton tows they are assumed to have begun a bottom existence in shallow near- shore waters. Juvenile yellowfin sole of 5-10 cm (two- and three- year-olds) are first observed in research vessel bottom trawls in inshore regions (Fig. 34-13). These age groups have been observed in low abundance off Kuskokwim Bay, in Bristol Bay, and along the Alaska Peninsula. They subsequently disperse to the more offshore waters occupied by adults (Fig. 34-13), and at lengths of 16-20 cm (mainly five- to eight-year- olds) occupy much the same waters as the larger fish. ENVIRONMENTAL INFLUENCES ON THE RESOURCE The eastern Bering Sea continental shelf is subject to wide seasonal and annual fluctuations in environ- mental conditions. In winter and early spring much S S 5 1 1 1 s I 1 ^1 IS it 4- • 1 1 JUU 1^ • 1 ^^ T • ' - f -^yt ^, ;' > '_ •-'• - - - — _ _ _ ^ _ X _ -.« 3 ■^ :^ - -- - <1 X- -* ^- • zw j j T s •« S s s s s mm H^eLa— \^^K- 1 HkiLjlS"' ' r54!_l-4rl-' 's S i i^ • ■•^ ■1 1 b ♦t - i: • • 9 J flJ" o •- — - y — "~ " V > • • 1 .^ k> ■- : 1 ,.■- • & •j ,' — t--m^-f-4_ I Si -U-tLL O"^ s 1 -f-Ua. .uT; s is s 1 tou 6 O o 01 c T3 0) it c CO XI c/l V j= CO CC C7i C o Q.T-1 o " 0 CO 1-5 b 15! C f - 3 "* >.J= O 0) Q -A o CO > c CO i^s U3 r-( CO CO pM „ LU _l o CO z u_ o _J _l III ID C CO 59 O CM c CD c O CO a; c« op CO 3 c D O r 1968 (Wakabayas rcentages of the thin the figure) ta longitude statistica > • • • ■ tt, j; a; ai oca O. ^^^ 564 Yellow fin sole 565 Figure 34-10. Distribution and relative abundance of yellowfin sole in April and May as shown by a NWAFC trawl survey in 1976 (after Bakkala and Smith 1978). As indicated in the text, concentrations of yellowfin sole can be noted in the vicinity of Unimak Island, west of St. Paul Island, and east of St. George Island. \ 180° 178°W 176"W 174-W I72°W 170°W 168"W 16G'W 164-W 162'W 160*W of the continental shelf may be covered by pack ice, but it is ice-free by late spring or early summer. Water temperatures under the pack ice are near the freezing point of seawater (—1.8 C) while bottom water temperatures on the shelf in summer may reach 10 C or more in shallow nearshore areas. Potocsky (1975) has documented the wide annual variation in distribution of pack ice in the eastern Bering Sea for the 17-year period 1954-70. Annual variations in surface and bottom water temperatures have also been observed (McLain and Favorite 1976, Maeda 1977, Bakkala and Smith 1978). Climatic conditions (see Ingraham, this volume) in the eastern Bering Sea appear to vary by multiyear cycles rather than randomly (Kihara 1977, McLain and Favorite 1976, Maeda 1977). These cycles are governed by the direction of prevailing winds. In years of northerly prevailing winds, the distribution of pack ice extends farther south and Alaskan Stream water entering the southeastern Bering Sea from the North Pacific Ocean is prevented from intruding onto the continental shelf. Southerly prevailing winds, on the other hand, limit the southward ad- vance of pack ice in winter and allow Alaskan Stream 566 Fisheries oceanography 180" 178' YELLOWFIN SOLE auG-OCT 1975 CATCH IN kg/km | + NO CATCH <25 25-100 100-250 >250 ■ Figure 34-11. Distribution and relative abundance of yellowfin sole in June 1976 and August-October 1975 as shown by NWAFC trawl surveys (after Bakkala and Smith 1978). As mentioned in the text, concentrations of yellowfin sole can be noted in Bristol Bay in June 1976 and on the inner shelf off Nunivak Island and in the southeastern Bering Sea in August-Septem- ber 1975. 180° 17e°W I76*W 174-W 172-W 170°W I68'W I66*W 164°W 162-W 160-W 158-W water to intrude onto the continental shelf in summer to replace residual cold water (0 C or less). One multiyear temperature cycle in the eastern Bering Sea ended in 1976. Maeda (1977) has shown that temperatures increased from 1961 to 1967 and then decreased from 1968 to 1976. The 1971-76 period has been recognized as a particularly cold period in the eastern Bering Sea (McLain and Favorite 1976, Bakkala and Smith 1978). In 1977 and 1978, temperatures were again warmer than in the pre- ceding six years. Research vessel surveys to assess the condition of crab and groundfish resources in the eastern Bering Sea have been carried out by the NWAFC annually since 1971; a more comprehensive series of data is available from 1973. The 1973-78 series provides information on the distribution and abundance of yellowfin sole during the last four years of the cold phase of the recent climatic cycle and the initial two years (1977-78) of a warmer phase. Age data col- lected during the surveys also provides information on the strength of year-classes originating as far back Yellow fin sole 567 as the early 1960's. These data provide some evi- dence that the fluctuations in environmental condi- tions observed over the past several years may have influenced the distribution, migrations, and abun- dance of yellowfin sole. Distribution and migration Possible influences of pack ice on the distribution and migration of yellowfin sole are illustrated by data from an NWAFC survey in April-June 1976. Tem- perature conditions were unusually severe during the survey; bottom water temperatures in June of the survey period were the lowest recorded in the south- eastern Bering Sea since 1966 (Bakkala and Smith 1978), and during April the ice edge approximated an extreme southern location based on 17 years of observations of pack ice distribution, 1954 to 1970 (Potocsky 1975). In April 1976, the leading edge of the large Unimak Island wintering concentration 1 74 1 72 MO 168 166 164 162 160 158 56 HH ' 59 , T \t (^ IH l^'«HB| B 59 \ \ A^ vT (|fV ''^"i s 58 \V ^^-l— ^ \ 1 ' •V 1 V|P 58 57 A ^ • i^ 1 57 56 B YELLOWFIN SOLE - AUG-SEP 1958 Eggs/m' • • t(^p^\ * 56 55 54 4 101-500 5 501-1000 6 > 1000 a *»■-,-' ^ 54 ^ « . . » '^ ■^jf^-~ 172 170 168 166 164 162 160 58 156 1 72 1 70 YELLOWFIN SOLE - JULY 1958 Eggs/m' Figure 34-12. Distribution of yel- lowfin sole eggs as shown by plankton surveys (A) in July 1962 (after Kashkina 1965) and (B, C) in July and August -early September 1958 (after Musienko 1963). 568 Fisheries oceanography Figure 34-13. Distribution of yellowfin sole by size group, August-October 1975. of yellowfin sole was near the ice edge, with ex- tremely dense concentrations of fish immediately off the ice edge (Fig. 34-14). These data give the impres- sion that the highly concentrated fish were waiting for favorable conditions to start moving onshore. As shown by later survey data (see Fig. 34-10), this concentration moved inshore in May, appsirently following the receding ice edge. The apparent se- quence of behavior in relation to the ice edge suggests that the Unimak Island wintering group avoided migrating under the ice and that ice cover in spring 1976 may have delayed inshore migrations. The timing of spring inshore migrations is not well defined, although Fadeev (1970a) has observed them starting from late April to mid-May over the three- year period 1959-61. Ice-induced delays to spring migrations are probably infrequent and of relatively short duration. The ice must reach an extreme southern location in spring months, as it did in 1976, to interfere with migrations beginning in late April and early May. In contrast to this apparent avoidance of ice cover by the Unimak Island concentration, other concen- trations of yellowfin sole may at times inhabit waters covered by pack ice. As shown earlier (see Fig. 34-10), the location of the St. Paul Island wintering concentration in May suggests that it was under the ice in April. That young yellowfin sole may winter under the pack ice in Bristol Bay has also been noted earlier. Water temperatures may also influence seasonal movements and distributions. Offshore movements in fall and winter may be an avoidance response to the cold bottom temperatures existing over the eastern Bering Sea shelf in winter. Most of the population winters on the outer continental shelf and slope in temperatures of 3.5-6.0 C (Fadeev 1970b). Spring onshore migrations, however, are not re- stricted by relatively cold water on the shelf; they commonly occur from relatively warm waters (up to 3-4 C) of the outer continental shelf to colder waters of the central shelf (Fadeev 1970b). The non- avoidance of cold water is illustrated by the survey data in April 1976 (see Fig. 34-14). Although warmer water in the range of 2-3 C was accessible to fish in the Unimak Island wintering concentration farther offshore, they were mainly found in tempera- tures of 0-2 C. Data from the series of NWAFC Crab-Groundfish surveys in June to mid- August 1973-78 illustrate the relationship of the distribution of yellowfin sole to temperature isotherms in early to mid -summer (Figs. 34-15 and 34-16). During this series of years, bottom water temperatures encompassing the surveyed distribution of yellowfin sole were variable, ranging from relatively low in 1974, 1975, and 1976 to relatively high in 1973, 1977, and 1978. Highest catch rates (>200 kg/10,000 m^ ) occurred in a rather wide range of temperatures from — 1 C to 7 C. Fadeev (1970b) has also commented on the wide tempera- ture range (0-10 or 11 C) inhabited by yellowfin sole in summer. His data showed highest catch rates in bottom water temperatures of 1-6 C. Although high concentrations of yellowfin sole were observed in a wide range of water temperatures during the 1973-78 Crab-Groundfish surveys, there was also some indication of differences in distribution relative to temperature conditions. In summer 1978, a relatively warm year, the main concentrations occurred in northern Bristol Bay. In the coldest years of the data series, 1975 and 1976, a number of widely separated concentrations were observed, some far to the south and west of northern Bristol Bay. In 1973, 1974, and 1977, years of temperature conditions intermediate in the series, concentrations appeared to be located between the more northern waters occupied in 1978 and the more southern areas where some of the concentrations were found in the Yellowfin sole 569 170 168" 164' 162" 160" 158" 156" 59' 57° 56" '►t^»»». 54' r 1^2'.''^.. "t.r.„^,^2: 1" °? CATCH (Kg/Km) LiJno catch E]<25 1125-100 H 100-250 ■ >250 AAA Ice edge jS^ 170 156 Figure 34-14. Distribution of yellowfin sole in April 1976 relative to bottom temperature contours and the edge of the pack ice. colder years of 1975 and 1976. These data imply that the rate of migration to more northern waters of the inner shelf may be slower in cold years than in warm years, or that summer distributions differ with temperature conditions. Abundance Accumulating evidence suggests that variations in water temperatures observed in the eastern Bering Sea may also affect the year -class strength of yellowfin sole. Maeda (1977) has shown predominant year- classes originating in years of relatively high tempera- tures in the eastern Bering Sea, and year-classes of below-average abundance originating in years of relatively low temperatures. More recent data appear to confirm this relationship. Bottom water tempera- tures taken in the southeastern Bering Sea in June of 1966-78 are used as a measure of temperature varia- tion and to relate to year-class abundance (Fig. 34-17). Year-classes originating in the warmer years of 1966-70, when June bottom temperatures ranged from 2.0 to 4.5 C, were relatively strong. In 1971 and 1972, when June bottom temperatures were colder (near 1.0 C), abundance of year-classes was much lower than in the previous five years. On the basis of this relationship, temperatures in succeeding years indicate that the 1973 year-class will be rela- tively strong, but that the 1974, 1975, and 1976 year-classes may be relatively weak. Preliminary data suggest that the 1973 year-class is, in fact, relatively strong; recruitment of the 1974-76 year- classes to research vessel gear is not yet adequate to measure their abundance. Environmental conditions would be expected to have their greatest influence on survival of yellowfin sole during early life history stages. Yellowfin sole spawn from July to September; and although June temperatures would therefore not be expected to influence the survival of eggs, larvae, or juvenile fish directly, they may provide an index to the environ- mental conditions that will prevail during the subse- quent spawning and early life history stages. 56 154 A YELLOWFIN SOLE ■ 1974 Calch 200 6. Hi? 1— 1 1 1 1 172 t70 168 166 164 162 160' 56- 156 I I Figure 34-16. Distribution of yellowfin sole in relation to isotherms in June-August during the relatively warm years of 1973, 1977,andl978. YELLOWFIN SOLE - 1977 Calch {Kg/ 10,000 m') Hwocalch E3 1-100 ^ 101-200 168 166 162 160 156 154 V^ — "1 "HHI ^g f w^ ^■1 f 6 \ 1 WUr 4 1 [p' ""*■""■"-"*— ——"■^^j^ Wm^ 1 ■ ^^^ ^^^p K~ /" ^.,,-— .— ..^. ^^^k^' , I'X' ^^•^^Hp^^^o 4 c YELLOWFIN SOLE - 1978 Calch (Kg/lO.OOO m') ■ ■ ■ ■ i,J ^^f^fK Q No calch EH 1-100 ^ 101-200 %.. « "M* "L? f" 164 162 158 156 571 572 Fisheries oceanography O 0) a E 1966 1968 1970 1972 1974 1976 1978 Year ^ 1800r 1966 1968 1970 1972 Year-Class Figure 34-17. Mean bottom water temperature in June at 34 standard survey stations in tiie southeastern Bering Sea (International Pacific Halibut Commission, 1976a, b; 1977; 1978) and year-class strengths of yellowfin sole. Year-class strength is based on estimated recruitment of age-6 fish measured by NWAFC research vessel surveys. Alverson, D. L., A. T. Pruter, and L. L. Ronholt 1964 A study of demersal fishes and fish- eries of the northeastern pacific ocean. H. R. MacMillan Lect. Fish., Univ. B.C. Bakkala, R., L. Low, and V. Wespestad 1979 Condition of groundfish resources in the Bering Sea and Aleutian area. Nat. Mar. Fish. Serv., Northwest and Alaska Fish. Cent., Seattle, Wash., Unpub. MS. Bakkala, R. G., and G. B. Smith 1978 Demersal fish resources of the eastern Bering Sea: spring 1976. Nat. Mar. Fish. Serv., Northwest and Alaska Fish. Cent., Seattle, Wash., Proc. Rep. EUson, J. G., and R. Livingston, Jr. 1952 The John N. Cobb's shellfish explora- tions in certain southeastern Alaskan waters. Spring 1951. Comm. Fish. Rev. 14(4): 1-20. EUson, J. G., D. E. Powell, and H. H. Hildebrand 1950 Exploratory fishing expedition to the northern Bering Sea in June and July, 1949. Fish. Leafl. 369. Fadeev, N. S. 1970a The fishery and biological characteris- tics of yellowfin sole in the eastern part of the Bering Sea. Tr. Vses Nachno-Issled. Inst. Morsk. Rybn. Khoz. Okeanogr. 70 (Izv. Tikhookean. Nauchno-Issled. Inst. Rybn. Khoz. Okeanogr. 72): 327-90. 1970b The distribution pattern of yellowfin sole (Limanda aspera. Pall.) in the northern part of the Pacific Ocean. Izv. Tikhookean. Nachno-Issled. Inst. Rybn. Khoz. Okeanogr. 74: 3-21. In Russian. REFERENCES Alverson, D. L., and M. I. Carney 1975 A graphic review of the growth and decay of population cohorts. J. Cons. Int. Explor. Mer., 36:133-43. Eager, E. W. 1957 Determination and analysis of recur- rent groups. Ecology 38: 586-95. 1963 Communities of organisms. In: The sea, M. N. Hill, ed., 2: 415-37. New York Inter-science, N.Y. I I ¥ Fager, E. W., and A. R. Longhurst 1968 Recurrent group analysis of species assemblages of demersal fish in the Gulf of Guinea. J. Fish. Res. Bd. Can. 25: 1405-21. Forrester, C. R., A. J. Beardsley, and Y. Takahashi 1978 Groundfish, shrimp, and herring fish- eries in the Bering Sea and Northeast Pacific -historical catch statistics through 1970. Inter. N. Pac. Fish. Comm. Bull. 37. Grant, S., R. Bakkala, and F. Utter 1978 Examination of biochemical genetic variation in yellowfin sole (Limanda aspera) of the eastern Bering Sea. Nat. Mar. Fish. Serv., Northwest and Alaska Fish. Cent., Seattle, Wash., Unpub. MS. Hart, J. L. 1973 Pacific fishes of Canada. Fish. Res. Bd. Can. Bull. 180. Hughes, S. E. 1974 Groundfish and crab resources in the Gulf of Alaska— based on Interna- tional Pacific Halibut Commission trawl surveys. May 1961 -March 1963. Nat. Mar. Fish. Serv., Northwest and Alaska Fish. Cent., Seattle, Wash., Proc. Rep. International Pacific Halibut Commission 1976a International Pacific Halibut Commis- sion, Ann. Rep. 1975. 1976b Items of information on the halibut fishery in the Bering Sea and the northeastern Pacific Ocean. Unpub. MS. 1977 Items of information on the halibut fishery in the Bering Sea and the northeastern Pacific Ocean. Unpub. MS. 1978 Items of information on the halibut fishery in the Bering sea and the northeastern Pacific Ocean. Unpub. MS. Yellowfin sole 573 Kashkina, A. A. 1965 Reproduction of yellowfin sole (Limanda aspera, Pallas) and changes in its spawning stocks in the eastern Bering Sea. Tr. Vses. Nauchno-Issled. Inst. Morsk. Rybn. Khoz. Okeanogr. 58 (Izv. Tikhookean. Nauchno-Issled. Inst. Morsk. Rybn. Khoz. Okeanogr. 53): 179-89. Kihara, K. 1976 Studies on the formation of demersal fishing ground. 3. Recurrent group analysis of demersal fish in the eastern Bering Sea. Bull. Soc. Franco- Japanese Oceanogr. 14(l):ll-22. 1977 Influence of abiotic marine environ- ment upon structure of demersal fish community in the eastern Bering Sea. Hokkaido Univ., Res. Inst. N. Pac. Fish., Spec. Vol., 199-204. Laevastu, T., and F. Favorite 1978 Fish biomass parameter estimations. Nat. Mar. Fish. Serv., Northwest and Alaska Fish. Cent., Seattle, Wash., Proc. Rep. Laevastu, T., and P. Livingston 1978 Some heretofore unknown numerical episodes of yellowfin sole in the eastern Bering Sea. Nat. Mar. Fish. Serv., Northwest and Alaska Fish. Cent., Seattle, Wash., Unpub. MS. Maeda, T. 1977 Relationship between annual fluctua- tion of oceanographic conditions and abundance of year classes of the yellowfin sole in the eastern Bering Sea. In: Fisheries biological produc- tion in the subarctic Pacific region, Hokkaido Univ., Res. Inst. N. Pac. Fish., Fac. Fish., Spec. Vol. 1977: 259-68. McLain, D. R., and F. Favorite 1976 Anomalously cold winters in the southeastern Bering Sea 1971-1975. Mar Sci. Comm. 2:299-334. 514 Fisheries oceanography Mito, K. 1977 Food relationships in the demersal fish community in the Bering Sea. I. Community structure and distribu- tional patterns of fish species. In: Fisheries biological production in the subarctic Pacific region. Res. Inst. N. Pac. Fish., Fac. Fish., Hokkaido Univ., Spec. Vol. 1977:205-58. Musienko, L. N. 1963 Ichthyoplankton of the Bering Sea (data of the Bering Sea expedition of 1958-1959), Tr. Vses. Nauchno-Issled. Inst. Morsk. Rybn. Khoz. Okeanogr. 48 (Izv. Tikhookean. Nauchno-Issled. Inst. Morsk. Rybn. Khoz. Okeanogr. 50): 239-69. Potocsky, G. J. 1975 Alaskan area 15- and 30-day ice forecasting guide. U.S. Naval Oceano- graphic Office, Spec. Pub. 263. Schaefers, E. A. 1951 The John N. Cobb's shellfish explora- tion in certain southeastern Alaskan waters, spring and fall of 1950 (a preliminary report). Comm. Fish. Rev. 13(4):9-19. Skalkin, V. A. 1963 Diet of flatfishes in the southeastern Bering Sea. Tr. Vses. Nauchno-Issled. Inst. Morsk. Rybn. Khoz. Okeanogr. 48 (Izv. Tikhookean. Nauchno-Issled. Inst. Morsk. Rybn. Khoz. Okeanogr. 50):223-37. Novikov, N. P. 1964 Basic elements of the biology of the Pacific Halibut (Hippoglossus hippo- glossus stenolepis Schmidt) in the Bering Sea. Tr. Vses. Nauchno- Issled. Inst. Morsk. Rybn. Khoz. Okeanogr. 49 (Izv. Tikhookean. Nauchno-Issled. Inst. Morsk. Rybn. Khoz. Okeanogr. 51). Pereyra, W. T., J. E. Reeves, and R. G. Bakkala 1976 Demersal fish and shellfish resources of the eastern Bering Sea in the baseline year 1975. Nat. Mar. Fish. Serv., Northwest and Alaska Fish. Cent., Seattle, Wash., Proc. Rep. Pertseva-Ostroumova, T. A. 1961 The reproduction and development of far eastern flounders. Izdatelstvo Akad. Nauk SSSR. Wakabayashi, K. 1974 Studies on resources of the yellowfin sole in the eastern Bering Sea. I. Biological characters. Japan Fishery Agency, Far Seas Fish. Res. Lab., Shimizu, Japan, Unpub. MS. 1975 Studies on resources of the yellowfin sole in the eastern Bering Sea. II. Stock size estimation by the method of virtual population analysis and its annual changes. Japan Fishery Agency, Far Seas Fish. Res. Lab., Shimizu, Japan, Unpub. MS. Wakabayashi, K., R. Bakkala, and L. Low 1977 Status of the yellowfin sole resources in the eastern Bering Sea through 1976. Nat. Mar. Fish. Serv., North- west and Alaska Fish. Cent., Seattle, Wash., Unpub. MS. Trans-shelf Movements of Pacific Salmon Richard R. Straty Northwest and Alaska Fisheries Center Auke Bay, Alaska ABSTRACT Five species of Pacific salmon are produced in the rivers and streams tributary to the Bering Sea shelf. All spend a portion of their juvenile and adult lives as residents of the shelf. Although this residence is transitory, salmon compose a significant and highly variable portion of the total pelagic fish biomass of the shelf. Knowledge of the distribution and abundance of salmon on the shelf is vital to our future at- tempts to assess the impact of exploiting them upon other living components of the shelf and the impact of man's modi- fying the shelf environment on the salmon resource. Both maturing salmon and juvenile salmon are present on the shelf from May through September, but their migration routes do not overlap appreciably. Juvenile salmon migrate seaward along the coast, eventually moving to offshore waters as their size increases. Maturing salmon remain in the offshore waters of the shelf until they are near their home river sys- tems. The shelf distribution of maturing salmon appears similar for all species migrating to rivers in the same geographic areas. Chinook salmon are earliest to enter the shelf during both spawning and seaward migration; later come sockeye, chum, pink, and coho salmon, in that order. Annual and seasonal variabUity in sea temperature and type and abundance of food appear to influence the distribution, growth, and, indirectly, the survival of salmon while they remain on the shelf. Significant gaps exist in our knowledge of the shelf distribution and dynamics of Pacific salmon. The greatest contribution to this knowledge can accrue from studies conducted between 168° W and the shelf edge from 56° N to 66° N. INTRODUCTION Five species of Pacific salmon, Oncorhynchus spp., are produced in the river systems tributary to the Bering Sea shelf. The sockeye salmon, O. nerka, is the most abundant species; next is chum salmon, O. keta; and then, in order, pink salmon, O. gorbuscha; Chinook salmon, O. tshawytscha; and coho salmon, O. kisutch; {Table 35-1). Salmon are anadromous, i.e., they mature in the ocean and spawn in fresh water. All salmon spend a portion of their juvenile* and adult lives as residents of the Bering Sea shelf. Although they spend only part of their lives there, salmon compose a significant and highly variable portion of the total pelagic fish biomass of the Bering Sea shelf during spring and fall. Knowledge of the seasonal movements, migration routes, and magnitude of annual variations in the biomass of salmon while they are in this area is vital to future attempts to assess the impact of exploiting them upon other living components of the shelf ecosystem and of physically modifying the shelf environment upon the salmon resource. BIOLOGY OF PACIFIC SALMON All species of Pacific salmon have similar life histories but they differ in fecundity (Table 35-2), food habits, growth rate, migration patterns, fresh- water and ocean age, age and size at maturity, and time and location of spawning. From early summer to early fall, salmon return from the sea to spawn in the rivers and streams from which they originated. When they enter fresh water, they cease feeding and derive their nourishment from body stores. All Pacific salmon die after spawning. Salmon eggs are deposited in gravel beds of rivers, streams, or lakes, and the eggs hatch during the winter. The young, known as alevins, remain in the gravel until their large yolk sacs have been absorbed and emerge from the gravel in the spring as fry. The greatest natural mortality occurs in fresh water during * The term "juvenile" refers to seaward-migrating salmon that have entered estuarine or marine waters but have spent less than one year in this environment. 575 576 Fisheries oceanography TABLE 35-1 Relative abundance (in thousands of fish) of Pacific salmon (Oncorhynchus spp.) produced in river systems tributary to the Bering Sea shelf as indicated by average of U.S. commercial catches 1961-77, inclusive^'^ and available Soviet catch statistics.*^ Species Percentage of total of Area Sockeye Chum Pink^i Chinook Coho Total all areas Bristol Bay 7,349.2 496.9 1,181.6 97.6 43.0 9,168.3 82.1 Gulf of Anadyr 150.0^ 410.5* 9.9^ scarce scarce 570.4 5.1 Yukon River 0.002 315.6 0.54 98.1 14.2 428.4 3.8 Alaska Peninsula (north side) 285.3 75.2 8.8 4.7 32.8 406.8 3.6 Kuskokwim Bay 9.0 90.6 26.3 38.7 78.4 243.0 2.2 Norton Sound 0.02 114.5 59.5 3.0 5.8 182.8 1.6 Kotzebue Sound^ 0.006 168.1 0.004 0.003 — 168.1 1.5 Arctic coast"^ i J J i — — — TOTAL 7,793.5 1,671.4 1,286.6 242.1 174.2 % by species 69.8 14.9 11.5 2.2 1.6 ^No reported catches for Kotzebue Sound in 1961. ''Source Alaska Department of Fish and Game, Juneau, Alaska. «=Pravdin (1940) and data available at TINRO. •^Even years only for U.S. waters. ^Estimated possible annual catch (Pravdin 1940). ^Average 1974-78 inclusive. ^Average 1975, 1977, and 1978. '^No estimates of number available. ^Unreliable reports of this species in American arctic rivers (Walters 1955). ^Documented reports of this species in Siberian and American rivers (Walters 1955). the early life stages, and is greatly influenced by environment. The fry of some species of salmon proceed immedi- ately to sea; fry of other species reside in fresh water for a few weeks or for one or more years. They migrate seaward across the shelf to the open ocean regions of the central and western Bering Sea and North Pacific Ocean, where they spend most of their marine life. Depending upon the species, salmon usually spend from a few months to several years at sea (see Table 35-2). Growth is slow during freshwater life, very rapid during the first summer at sea, and slower thereafter. The growth rate of salmon in both fresh water and the ocean varies among species, brood years, and stocks of the same species. Regardless of the amount of time spent in fresh water, however, salmon attain most of their growth in the ocean. ABUNDANCE OF PACIFIC SALMON ON THE SHELF According to commercial fishery catch statistics, the rivers and streams flowing into Bristol Bay (see Table 35-1) at the southeastern terminus of the Bering Sea shelf produce the greatest biomass of salmon (Fig. 35-1); information from the early 1900 's (Pravdin 1940) and recent data available at TINRO indicate that those flowing into the Gulf of Anadyr in the northwestern shelf area are a distant second. Pacific salmon are produced in certain rivers of the Siberian and American arctic coasts (Walters 1955), but estimates of their relative abundance are not available. Apparently chum salmon, Oncorhynchus keta, are sufficiently abundant in the lower Lena River, which enters the Siberian arctic, to support a commercial fishery. Salmon produced in arctic rivers must traverse the Bering Strait and the Bering Sea shelf. Trans-shelf movements of Pacific salmon 577 TABLE 35-2 Life history of the five species of Pacific salmon in Alaska— exceptions are frequent (adapted in part from Bailey 1969, Merrell 1970, and Hartman 1971). Estimated average juvenile weight at Time spent time of Time Average Average Species in fresh water seaward spent Age at adult number eggs of after emergence migration at sea spawning weight^ per female salmon Freshwater habitat from gravel (g) (yr) (yr) (kg) (thousands) Sockeye short streams and lakes 12-36 months 8.7^ 1-4 3-6 2.27 3.5 Chum short and long streams 1 month 0.37'' 2-4 3-5 2.96 3.0 Pink short streams 1 day, usually 0.26"= 1 2 2.008 2.0 Chinook large rivers 3-12+ months 4.3^ 1-6 3-6 7.26 4.0 Coho short streams and lakes 12-14 months 29.0^ 1-3 3-4 3.04 3.5 ^Average for age 1.0 and 2.0 Bristol Bay sockeye salmon (National Marine Fisheries Service Auke Bay Laboratory 1979, unpub- lished data). ''Hooknose Creek, British Columbia (from Table 5 of Bams 1970). '^Southeastern Alaska pink salmon (Bailey and Pella 1976). •^Yukon River (Salcha River) chinook salmon (Trasky 1974). ^Karluk Lake (Kodiak Island) coho salmon (Drucker 1972). ^Mean weight for Alaska Peninsula (north side), Bristol Bay, and Yukon River stocks combined. Source: INPFC Statistical Year- book (1961). ^Mean for Alaska Peninsula (north side) and Bristol Bay only. Estimates of the annual abundance of maturing sockeye, chum, pink, chinook, and coho salmon returning to Bristol Bay between 1951 and 1979, inclusive, have been made by Rogers (1977 and personal communication). Estimates of the average total abundance of each species returning to rivers and streams in Bristol Bay and along the northern side of the Alaska Peninsula have also been made by Stem et al. (1976). No similar estimates have been made for maturing salmon returning to other western Alaska rivers and streams entering the shelf north of Bristol Bay, nor are estimates available for the total abundance of each species of salmon returning to rivers and streams entering the Gulf of Anadyr or for rivers and streams along the Siberian and North American arctic coasts. If the catch statistics in Table 35-1 indicate the relative contribution of each area to the total population of maturing salmon entering the shelf each year, the areas north and west of Bristol Bay contribute 14.2 percent of this popula- tion. The reliability of this figure, however, can be determined only through improved estimates of the commercial catches, subsistence catches, and escape- ment of maturing salmon in these areas. On the basis of very limited data and several Fig. 35-1. Index map showing place-names and locations of salmon-producing regions. assumptions listed below, we can estimate the pos- sible range in the total number and biomass (weight) of maturing salmon entering the shelf each year to be 578 Fisheries oceanography between 5.6 and 65.9 million fish or 13,776-162,114 mt. The assumptions in making this estimate are as follows : 1. The catch statistics given in Table 35-1 indi- cate the contribution of each area of the shelf to the total shelf population of maturing salmon. 2. Annual fluctuations in the abundance of salmon produced in all rivers and streams entering the shelf follow annual fluctuations in abundaince for Bristol Bay between 1951 and 1979, inclusive (Rogers 1977 and personal communication ) . 3. The weighted mean weight of mature salmon of all species combined is 2.46 kg per fish. This weight was calculated by summing the products of the mean weight for each species (see Table 35-2) multiplied by its contribution to the total shelf population of salmon (see Table 35-1). The estimate is derived by dividing the low and peak estimates (4.6 and 54 million fish) of the numbers of maturing salmon returning to Bristol Bay between 1951 and 1979, inclusive, by the mean proportion of salmon (0.82) Bristol Bay contributes to the total shelf population of salmon each year (see Table 35-1). Estimates have also been made of the numbers of juvenile sockeye, chum, pink, chinook, and coho sal- mon entering Bristol Bay each year between 1950- 1976, inclusive (Rogers 1977 and personal communi- cation). In addition, the average and peak estimates of all species of juvenile salmon entering the shelf from Bristol Bay and from streams along the northern side of the Alaska Peninsula for 1955 through 1975 have been made by Stem et al. (1976). Estimates of juvenile salmon entering the shelf from areas north and west of Bristol Bay have not been made. The low and peak estimates of juvenile salmon migrating seaward from Bristol Bay for 1955-76 numbered 119 and 770 million fish, respectively. If the same rationale used to calculate the abundance of maturing salmon is used for juvenile salmon, the annual prob- able low and peak numbers of juvenile salmon enter- ing the shelf are 145 and 939 million, respectively. If 6.72 g is the weighted mean weight for all species of juvenile salmon at the time they enter the shelf, the estimated total weight of juvenile salmon on the shelf in a given year is between 974 and 6,310 mt. The weighted mean weight for all species of juvenile salmon was calculated by summing the products of the mean weight of each species (see Table 35-2) multiplied by its contribution to the total shelf population of salmon (as indicated by the adult catch in Table 35-1). SEASONAL MOVEMENTS OF SALMON ON THE SHELF The marine life of Pacific salmon may be divided into three phases: ocean life, spawning migration, and seaward migration. The seasonal timing and length of each phase is species and stock specific and fluctuates, within limits, because of environmental variation. Ocean life When salmon aire distributed throughout the vast area of the Bering Sea and North Pacific Ocean (French et al. 1975), they undergo rapid growth and attain sexual maturity. This period of ocean life may last from a few months to several years depending on the species of salmon (see Table 35-2) and ocean growing conditions. More than one generation of sockeye, chum, chinook, and coho salmon are present on the shelf during ocean life. Immature and matur- ing salmon of different races and ocean ages and from different continents or geographic areas may, depend- ing on the season, become segregated from or inter- mix with one another (Royce et al. 1968). Immature chinook salmon (age 0.1)^ apparently are abundant in the central Bering Sea near the edge of the conti- nental shelf in June and July (Major et al. 1978). Immature chum salmon, older than age 0.1, have been captured in research gillnets in July in the western Bering Sea shelf and in the slope area near Cape Navarin and the Gulf of Anadjrr (Nishiyama et al. 1968). A few immature chum and sockeye salmon have also been taken during research fishing on the Bering Sea shelf as far north as 62° N in late July (Yonemori 1967). Research fishing and salmon tagging have been conducted on the Bering Sea shelf only during the spring, summer, and early fall— periods when juvenile salmon are migrating seaward and maturing salmon are migrating to spawning grounds. The species composition and proportion of immature salmon of age 0.1 or older, of maturing salmon, and of seaward- ^ Number of scale annuli acquired in fresh water is indicated by the figure preceding the decimal, and the number of annuli acquired in the ocean is indicated by the figure following the decimal. Thus, an 0.3 fish is one whose scales reflect 3 annuli at sea (freshwater age unspecified), a 2.0 fish is one whose scales show 2 annuli in fresh water (ocean age unspecified), and an age 2.3 fish is one with 2 annuli in fresh water and 3 at sea. Total age (year of life) is obtained by adding one to the sum of the freshwater and ocean annuli. For example, an age 2.3 fish is six years old. Trans-shelf movemen ts of Pacific salmon 5 79 migrating juvenile salmon residing in the shelf area between November and April are unknown but the numbers are probably negligible. Young chinook salmon (judged by their 26-35 cm fork lengths to have been in their first year in the ocean) have been taken by a Japanese trawler fishing in mid-winter for walleye pollock, Theragra chalco gramma, in the eastern Bering Sea at the shelf edge near 56° N, 168°W and 56°N, 173°W (Major et al. 1978). General spawning migration The time of the spawning migration, from depar- ture of sexually maturing fish from the high seas of the North Pacific Ocean and western and central Bering Sea until arrival at the mouths of their respec- tive home streams or river systems, is species and stock specific. Maturing salmon are most abundant on the Bering Sea shelf from mid-May to early September (Table 35-3). Maturing chinook salmon enter the Bering Sea shelf earliest ; later in order come sockeye, summer chum, pink, fall chum, and coho salmon. The length of time a given salmon stock is present and its distribution in the shelf area during spawning migration depend upon the geographic location of its home river system, the size and/or ocean age of fish composing the population, and environmental conditions that influence the rate and direction of migration. Sockeye, chum, and pink salmon have been cap- tured in varying numbers during U.S. and Japanese research fishing and tagging studies at many places throughout the shelf during spawning migration. Fewer chinook and coho salmon, however, have been caught than other salmonid species because chinook and coho salmon are less abundant and research fishing did not take place when they were at their peak. Research fishing and tagging experiments have taken place mainly between mid-June and late July, when sockeye and chum salmon are most abundant on the shelf. Chinook salmon begin entering the shelf area in mid- to late May, coho in mid- to late July. As a result, we have less information about the distribution and direction of movement of chinook and coho than of sockeye, chum, and pink salmon. The distribution and direction of migration and relative abundance of all five species of Pacific salmon on the Bering Sea shelf are depicted in Figs. 35-2 to 35-6. These figures were prepared by plotting the locations of capture of each species on a chart of the Bering Sea shelf for all years for which data are available. For sockeye, chum, and pink salmon, certain areas of the shelf consistently yielded larger catches or larger catch-per-unit-effort values than 175^ 170' 160' 155" SOCKEYE SALMON Mid-JL ne to lale July m Location ot sockeye producing river systen ^ Direction ol migralion »- Probable direction ot (degr relat e ot stiading indicates ve abundance) A Fig. 35-2. Distribution of sockeye salmon during spawn- ing migration, mid-June to late July. Fig. 35-3. Distribution of chum salmon during spawning migration, mid-June to early August. other areas. For chinook and coho salmon, similar data were available only for Bristol Bay, where these species have been captured most frequently. Direc- tion of migration of each species was derived from the published results of tagging experiments and direction-of-movement studies. The probable direc- tion of migration was based on the geographic loca- tion or proximity of the home river system of the species and the verified direction of movement of 580 Fisheries oceanography TABLE 35-3 Inclusive dates of peak abundance of maturing Pacific salmon on the Bering Sea shelf during spawning migration (from commercial fishery and test fishing catches and migration rates). Sockeye Chum Pink Chinook Coho (summer) (fall) Estimated at shelf edge* Estuaries or river entrances: Alaska Peninsula'' Bristol Bay^9^oy^^o Figure 36-8. Schematic diagram of September distribu- tion of halibut, herring, pollock, and yellowfin sole. (Long dashes indicate probable extension of shelf range; short dashes indicate juvenile distribution, where known or different from adults.) Salmon occur throughout the shelf area. 604 Fisheries oceanography Figure 36-9. Schematic diagram of October distribution of halibut, herring, pollock, and yellowfin sole. (Long dashes indicate probable extension of shelf range; broad Hne indicates ice edge.) ice cover data are available, one can presume that the effects of the wind mixing and subsequent ice cover will quickly dominate any downward diffusion of residual heat in the water column. Both cooling effects will reach depths exceeding 100 m; the former takes place within a day or so, the amount of cooling depending on water temperatures in the surface layer, and the latter requires one to several weeks. However, it is prolonged ice cover that gradually drives bottom temperatures on the shelf down to —1.8 C and the freezing-out of fresh water results in higher salinities. Nevertheless, temperatures of 3-4 C still occur on the bottom at the shelf edge and upper slope. The ice edge in November generally extends from Bristol Bay seaward of Nunivak and St. Lawrence Islands into the western Gulf of Anadyr. By December it extends from outer Bristol Bay past St. Matthew Island to Cape Navarin. Herring and yellowfin sole are believed to reach offshore wintering grounds in November (Fig. 36-10). Herring overwintering in the Norton Sound area under the ice may experience temperatures higher than —1.8 C in the deeper areas if wind mixing did not establish neutral stability in the water column before ice cover was formed. Pollock retreat to the 100-m isobath by the end of December; at this time halibut have largely retreated seaward of the 200-m isobath. RESOURCE ENVIRONMENT RELATIONS The life histories of the dominant fish (except for halibut) selected for discussion in this section are quite diverse. Harden Jones (1977) suggests that fish movements can be summarized in the form of trian- gular patterns in which adult fish move between their spawning, feeding, and wintering areas in a sequence which depends on the season in which they spawn: spring spawners (here, pollock and herring) follow a clockwise sequence of feeding-wintering-spawning, and autumn spawners (no representatives among the species studied here) a counterclockwise sequence of feeding-spawning-wintering. Such a scheme does not account for halibut, which spawn in winter in the wintering area, or yellowfin sole, which spawn in the summer far from the wintering area. Plausible ranges and time scales of advection (passive transport) and migrations are shown schema- tically in Figure 36-11. The average advection (~1 km/hr) and average migration (~3 km/hr) are indi- cated, relating time to distance; periodic migrations such as semidiurnal and seasonal migration are also shown. The expected relative magnitudes of changes of biomass of a given species as a result of different periodic migrations as well as possible long-period "drift" within the system (which could result in long-period changes in abundance depending on transgressions of pelagic fish on and off the shelf and demersal fish into and out of the areas along the shelf) are shov^oi in Fig. 36-12. Migration speeds are J I I I I I L Figure 36-10. Schematic diagram of November-December distribution of halibut, herring, pollock, and yellowfin sole. (Long dashes indicate probable extension of shelf range; broad line indicates ice edge.) Fin fish and the environment 605 not known, except perhaps for Bristol Bay sockeye salmon, but some estimates of magnitudes of shelf distances traversed and required speeds for the eastern Bering Sea have been compiled (Table 36-2). Data for similar species in the North Sea indicate that these estimated speeds are considerably lower than might be expected and Bering Sea stocks may have considerably more mobility over the shelf than heretofore believed. Although environmental conditions are variable and stock distributions fluctuate, there are apparently no environmentally related indices, other than generally defined warm or cold years, that can be used to forecast successful or unsuccessful spawning or year-class strengths, anomalous distributions or migration paths, or fluctuations in annual abun- dances. Even though in the foregoing summaries of monthly events only mean conditions were con- sidered, and events could be advanced or delayed as much as a month as a result of warm or cold condi- tions, the initial primary considerations for deter- mining distributions were largely depth and season: for example, halibut occur in winter seaward of the 200-m isobath; adult pollock are generally seaward of 100 m in winter regardless of the position of the ice edge and seaward of 50 m in summer; juvenile pol- lock in summer are found shoreward of the 100-m isobath; the yellowfin sole wintering area is largely between the 100- and 200-m isobaths but juveniles overwinter in inner Bristol Bay under the ice, where the shallow depths assure that the water temperature will be —1.8 C; herring overwinter largely between the 100- and 200-m isobaths again apparently without Average adveclion^ "Life cycle" lO" - E 10-^ Year '/■ Yeai Migrations Dav % Day Average migration (max. adv.) 10" 10= Distance (meters) Figure 36-11. Ranges of advection and migration of stocks in marine ecosystems. "Life cycle" SemiannuaK Reversible components 0.1 1.0 10 10-^ lO-" 10' Time (hours) "Nonreversible" component 10= 10° Figure 36-12. Relative ciianges in abundance of stocks due to advection and migration. any regEird to the ice edge and some inshore stocks overwinter in ice-covered coastal regime. In only a few instances is there specific evidence of environ- mental influences. Yet if one approaches these conditions or events from a fishery oceanographer's or environmentalist's point of view rather than from that of a management biologist, another perspective is evident. TABLE 36-2 Speeds of migrating fish Species Speed (km/d) Sole" Plaice" Herring" Salmon (sockeye)^ Salmon (chum)^ Halibut^ Herring^ Yellowfin sole^ 7-16 1- 7 4-30 54 48 6 25 3- 7 " Harden Jones (1977)— North Sea area ^Kondo et al. (1965)-E. Bering Sea shelf ^Noviko (1970)-E. Bering Sea shelf "* Estimated from spawning migration— E. Bering Sea shelf Winter Most discussions of productivity deal with condi- tions on the roughly one million km^ of shelf, yet it is obvious that most of the biomass (excluding mammals) is restricted to a small fraction (less than a fifth) of the total shelf area for nearly half the year, November to April. Although one might argue that feeding in winter is usually drastically reduced, it should be obvious that bottom temperatures at the 606 Fisheries oceanography shelf edge (3-4 C) during the winter are several degrees higher than in the mid-shelf area throughout the summer (— 1 to 2 C), even though summer inshore temperatures may reach 10-15 C. Predator and prey are crowded together and there is evidence at other times and places that halibut consume yellowfin sole; pollock are primarily cannibalistic (Takahashi and Yamaguchi 1972). (1) What effects do changes in the continuity and extent of the anomalous 3-4 C bottom temperatures on the outer shelf and upper slope have on species and biomass distributions, and to what extent does predation and growth occur during winter in this narrow zone? Halibut spawn in January at depth; environmental conditions in the southeastern comer of the Bering Sea basin can be considered relatively constant. The prolonged period from egg release to the time when food is no longer provided by the yolk sac eliminates the availability of forage as a critical factor until early spring. (2) Does the success of a halibut year-class in this area depend on a predominance of onshelf flow along the shelf edge, possibly entrainment of larvae in local eddies, or is this stock maintained by transport of eggs and larvae from the North Pacific Ocean through Aleutian Island passes, with spawning in this area providing the basis for stocks on the Asian (Cape Navarin/Cape Olyutorski) coast? Of course for management purposes one can wait until individual year-classes are recruited into the fishery, but unless this question is resolved we vnll not be able to assess the effect of oil or oil spills on halibut (and other groundfish with similar behavior) recruitment in specific areas. Herring are shown to have possibly three offshore wintering grounds with equivalent bottom topog- raphies, but these represent three areas with distinct environmental conditions. The stock of the area south of Cape Navarin spends a large part of the winter, if not all of it, under the ice at —1.8 C; the major wintering ground north of the Pribilof Islands may have ice cover only in March or April, insuffi- cient time to establish —1.8 C conditions, so that temperatures of 0-2 C prevail; the possible winter concentration between Unimak Island and the PribUof Islands will encounter temperatures of 2-4 C and experience ice cover and colder conditions only in extremely cold years. (3) Are the three different winter temperature regimes critical to the isolation of these three stocks and if so, do the stocks have broader distributions than indicated and do the locations of the stocks change with environmental conditions? Attempts by NWAFC to sample herring in the major wintering ground in February 1978 failed (winter 1978 can be classified as warmer than normal). Spring The herring are the first to leave the area of con- centrated biomass in winter, and spawning migrations begin in March. This may be an instinctive move- ment, because ice cover is relatively constant, at depths in excess of 100 m temperature regimes do not change (—1.8 C, 0-2 C, and 2-4 C), and relative darkness at depth does not provide any signal or impetus. The primary factors here must be increasing light evident in upper layers and the absence of ice. (4) Since the extent of mean ice cover usually reaches a maximum in April, are shoreward migrations from wintering grounds near the shelf edge that begin in March conducted under the ice, or do herring migrate southward (or await retreat of the ice edge) and proceed to the coast in open water? If the former were true, herring would arrive in spawning areas on schedule but in cold years they would have to wait until the coast was free of ice to spawn. Although rafting can occur, the ice edge is usually composed of soft ice (vessels have easily penetrated 50-100 km) and such observations could be made; spring distribu- tions and shoreward migration routes could then be ascertained by analyses of satellite imagery of ice cover. Certainly pollock spawning in spring is a dominant event, although at this time it is difficult to compare its effects to those of sculpin spawning. The abun- dance of adults increases in spring north of Unimak Island (Low and Akada 1978). (5) Does the progres- sive cooling of shelf edge and upper slope water north of the Pribilof Islands into April as a result of pro- longed ice cover cause a southweird retreat of adult pollock or is this a spawning migration? As might be expected of a successful species, spawning is protracted. Waldron (1978) found that pollock larvae were present near the shelf edge north of Unimak Island in March; if development occurred locally, spawning must have begun in February. However, the larvae were found in a northward- protruding tongue of temperature maxima (4 C) and it is possible that they were advected into the area from south of the Alaska Peninsula, where conditions can be much warmer at this time of the year. Unlike halibut, pollock eggs and larvae occur in the surface layer and are exposed to the vicissitudes of weather from March to July. In spring 1976, eggs were found within the ice edge when an unusual southward thrust of ice occurred, and violent spring storms are the general rule rather than the exception. (6) What are the effects on pollock year-classes of stormy vs. calm weather, ice vs. no ice, and early (February- March) vs. late (May-June) spawning? A simple device that would not only assist in determining Finfish and the environment 607 answers but also permit prompt verification and evaluation of conditions would be a moored biologi- cal buoy that would sample organisms in the water column (5 m to bottom) at weekly (or shorter) time intervals from February to late spring, when the data would be recovered and evaluated. Furthermore, pollock spawning appears to be largely restricted to the area seaward of the 100-m isobath, thus very near the Pribilof Islands and the million or so fur seal predators. (7) What conditions, environmental or otherwise, appear to confine adult pollock seaward of the 100-m isobath before, during, and after spawning? It is not clear whether halibut or yellowfin sole are first to begin shoreward migrations, but the latter have a spawning incentive. (8) Do halibut follow yellowfin sole shoreward to be sure of a source of food or is their shoreward migration and subsequent consumption of benthic organisms simply a seasonal feeding behavior pattern? In respect to both inshore migrations Harden Jones's (1977) selective tidal transport hypothesis may be applicable. His studies indicate that plaice in European waters are able, because of their shape, which provides lift forces allowing them to glide (or sailplane), to move into midwater, where they spend half their time, taking advantage of tidal flow in the direction of purposeful movement, but returning to the bottom at the turn of the tide. Such a behavior in the stock between the Unimak and Pribilof Islands would permit considerable enhancement of migra- tions; current data (Dodimead et al. 1963; see Chap- ter 5, this volume) and tidal model studies at NWAFC (Hastings 1975) indicate a general northeast -south- west oscillation north of the Alaska Peninsula, particularly in Bristol Bay, that would aid migrations from the shelf edge into Bristol Bay and a change in oscillation that would aid a subsequent general northward movement over the shelf. The benefit from such tidal sequences also would aid return movement to the southwestern part of the shelf in fall. (9) Does selective tidal transport give halibut, yellowfin sole, and other flatfish considerably more mobility and migration speed than now believed (-3-6 km/d)? The onshore delay or aggregation of halibut amd yellowfin sole at the cold front north of the Alaska Peninsula is a convincing example of resource- environment relations. But the northward movement of halibut proposed by Natarov and Novikov (1970) and the eastward movement of the yellowfin sole stock north of the Pribilof Islands indicated here raise some doubts as to the absolute validity of this phenomenon; U.S.S.R. studies in the Grand Banks off the east coast have indicated that hunger transcends discomfort and fish will at times intrude into areas having environmental conditions normally avoided. (10) What are the roles of selective tidal transport and the ability of halibut and yellowfin sole not only to move quickly but to rise easily into the water column in their penetration of the cold, midshelf zone? Studies of benthos and possibly relationships to fish in this area are the subject of section XII of this book. In regard to salmon movements, over 40 years ago Moulton (1939), in the foreword to a symposium concerning the migrations of salmon, noted that in the problem of accounting for the migration of salmon one enters a field in which science and ro- mance appear to meet. In spite of environmental and tagging studies, the story of what takes place in the apparently supernatural migrations, wherein mature fish return after several years at sea to precisely the stream of their birth, was far from being com- plete, and explanations of such migrations were all equally open to question. Although considerable information on oceanic distributions and migrations has come from studies conducted since 1953 under the aegis of the INPFC, how salmon find their way in the ocean remains a major and interesting fisheries oceanography problem. Explanations of coastad migrations have the same inadequacies as those of ocean migrations. Less than a decade ago at another S£ilmon symposium Larkin (1975), after reviewing information on how salmon navigate, postulated that they might have a system of collectivizing judgment that could provide a school with remarkable naviga- tional accuracy; if they could detect their confreres by scent, adults could school effectively in the ocean as well as detect coastal regimes carrying the scent of juveniles that had commenced seaward migrations. Although such ideas are perhaps more intuitive than scientific, there are numerous examples of remarkable salmon behavior other than the fact that mature individuals return to natal streams annually and arrive at the coast with consistent and appropriate timing. Such punctuality might be considered supernatural if it were not for their mobility; for example, king salmon migrate over 2,400 km up the Yukon River in 3.5 weeks, advanc- ing at an average speed of 110 cm/sec or about 85 km/d in spite of opposing river flow (Gilbert and O'Malley 1922). Apart from any potential to swim shorter distances at higher speeds, they could transit the shelf area in much less than a week although shelf migration speeds of only 50 km/d are reported (see Table 36-2). Although knowledge of the nature of such movements is scant, there is an apparent order to the sequence of arrivals at river mouths of all species. 608 Fisheries oceanography Lcirge numbers of king salmon in trawl catches in winter 1979 near Zemchug Canyon came as a surprise to biologists who have studied salmon migrations for decades. Since this canyon is believed to have been cut across the shelf during periods of lower sea level (land bridge) by the Yukon River, if these salmon c£in be shown to be of Yukon River origin one is inclined to believe that salmon homing instincts have strong environmental imprints. But it is apparent that movements of Yukon River, Norton Sound, and Gulf of Anadyr salmon and those that eventually pass through Bering Strait are largely unknown. Questions concerning the movement of salmon in the eastern Bering Sea in relation to environmental factors are too numerous to list but obviously little is known about movements northward of the Pribilof Islands. (11) What are the migration routes over the shelf of adult and juvenile salmon from Gulf of Anadyr, Arctic Ocean, Norton Sound, and Yukon River areas? Obviously extensive environmental studies must accompany any distributional or tagging problems. There is much evidence that Pacific salmon re- spond to environmental conditions (Favorite et al. 1977) and it is apparent that migrations of various species and stocks through the various OCS oil development lease areas (e.g., St. George Basin, Navarin Basin, Norton Sound) cannot be ascertained or predicted until more information is obtained on effects of environmental conditions on shoreward migrating adults and seaward migrating smolts (juve- niles) in this area. Oil contamination of river mouths could affect upstream migrations of sockeye, coho, and king salmon and coastal spawning of chum and pink salmon; and spills at offshore sites could affect migrations of salmon of U.S. and U.S.S.R. origin (spills in spring and summer could also be locally harmful to herring, because spawning beds are in- tertidal). and also possibly to the effects of coastal dilution, which is also delayed from that in more southern areas. (13) Can the time of herring spawning at various points along the coast, which can vary from a week to a month, be defined by temperature or salinity conditions, and do advanced or delayed spawning periods affect the success of year-classes, or contribute to reduced or excessive predation on spawners? The spawning of yellowfin sole in the area off Nunivak Island is an important summer event. It occurs inside one of the sharpest frontal zones on the shelf, between the mid-shelf and coastal sub-domains. Temperatures in the frontal zone at this time increase sharply shoreward, from —1 to 10 C. Thus, spawning occurs in the warmest area of the eastern part of the shelf that lacks the excessive dilution and high turbidity produced by local rivers (unlike inner Bristol Bay and Norton Sound), in an area that has a broad extent of shallow depths (unlike the northern coast of the Alaska Peninsula). The length of time required for larvae to attain mobility or drop to the sea floor is not known, but a northerly flow in this area of 25 cm/sec (~0.5 kn) would transport them through Bering Strait in about 20 days unless they were caught in eddies in Norton Sound. (14) What mechanism prevents the dispersal of yellowfin sole larvae into the Arctic Ocean? It is apparent in the monthly mean distributions of properties for September that there is a marked seaward intrusion of coastal water in the Nunivak Island area that could have a major effect not only on circulation, but on movements of juvenile pollock and herring and the larvae of yellowfin sole. (15) Is this trans-shelf intrusion real or an artifact of averag- ing data and what is its effect on conditions and events in this area? Summer It has been pointed out that adult pollock are cannibalistic, preying extensively on juveniles; in summer, juveniles occur largely within the area between the 100-m isobath and the coast, whereas the adults are largely confined seaward of 100 m. (12) Is the presence of juveniles in inshore areas where warm conditions occur a response to environ- mental conditions including forage or are they found there because it is the only area where predation by adults does not occur; what confines adult pollock seaward of the mid-shelf area? Herring spawning occurs in northernmost areas in summer, reflecting a northerly sequence with time and perhaps a direct relation to temperature and ice. Fall Although little is known concerning precise fall dis- tributions of fishes, it is generally believed that most begin to retreat southward and seaward to the shelf edge by October. Although air temperatures are lower, mean sea surface temperatures have not decreased substantially. Water temperatures in coastal areas have lowered a few degrees but are still high (~8 C); and ice appears only in eastern Norton Sound and along the shore. Thus there is little evidence of drastic environmental changes other than in the duration and intensity of light, until Novem- ber. (16) What triggers fall seaward movements of fish to the shelf edge and what are the nature and relative timing of movements of individual stocks? Finfish and the environment 609 CONCLUSIONS The eastern Bering Sea ecosystem encompasses a vast renewable resource of great potential and value that should be carefully protected and developed but of which we have very little basic understanding. The NWAFC and the North Pacific Fisheries Council are expending great efforts to estimate the various individual resources for management purposes, but there should be an equal effort to understand condi- tions and processes as well as periodic and aperiodic events and their causes and effects. This can come about only through cooperative, interdisciplinary research. Basic research and resource assessment should continue, but fisheries oceanography studies should combine both of these and provide an integra- tive focus that is badly needed as exploitation of individual resources of commercial value accelerates. Nearly a decade ago, when congressional support for fisheries studies was wanting, one argument put forth to justify limited support for marine research was that marine fisheries were a common property resource, could not be protected, and posed real problems in regard to development. The ideal exam- ple given was a plot of trees that could be clearly defined spatially and contained a crop that could be protected (from man and diseases), selectively har- vested on demand, and selectively reseeded to provide a continuing resource. The eastern Bering Sea shelf, relatively isolated from the broad sweep of ocean currents, is a unique place to consider development of a major marine habitat that can be protected and developed when we have a basic understanding of requirements of individual species and stocks, interactions among them, and their place in the ecosystem. The economic justification is obvious; the value of fisheries products extracted in 1978 was roughly half a billion dollars. The problem of protec- tion is answered by the recent FCMA. Even from the limited discussion presented, the merit in multispecies and ecosystem approaches to investigating and understanding conditions and processes is readily apparent; such a focus should be considered in all future studies. Because this requires extensive accountability of species interactions and environmental effects, we are at a loss to handle the multiplicity of factors involved without formulating some sort of modeling or simulation techniques. A great deal of preliminary work has been accom- plished along these lines at NWAFC and a summary of an existing simulation model (DYNUMES) that presents the quantifications of biomasses omitted in this chapter is presented in the next. Such studies also permit independent assessments of movements in relation to environmental conditions such as temperature and forage requirements. REFERENCES Dodimead, A. J., F. Favorite, and T. Hirano 1963 Salmon of the North Pacific Ocean- Part II. Review of oceanography of the subarctic Pacific region. Inter. N. Pac. Fish. Comm. Bull. 13. Favorite, T., T. Laevastu, and R. R. Straty 1977 Oceanography of the northeastern Pacific Ocean and eastern Bering Sea, and relations to various living marine resources. Nat. Mar. Fish. Serv., Northwest and Alaska Fisheries Cent., Seattle, Wash., Proc. Rep. Gilbert, C. H., and H. O'Malley 1922 Investigations of the salmon fisheries of the Yukon River. In: W. T. Bower, Alaska fishery and fur-seal industries in 1930. Append. VI to Rep. of U.S. Comm. Fish. 1921, 128-54. Hansen, D. V. 1978 Lagrangian surface current measure- ments on the outer continental shelf. In: Environmental assessment of the Alaskan continental shelf. U.S. Dep. Comm., NOAA/OCSEAP, Ann. Rep. 9:590-603. Harden Jones, F. R. 1977 Performance and behavior on migra- tion. In: Fisheries mathematics, J. H. Steele, ed., 145-70. Academic Press, N.Y. Hastings, J. R. 1975 A single-layer hydrodynamical- numerical model of the eastern Bering Sea shelf. In: Ocean variability: Effects on U.S. marine fishery re- sources, J. R. Coulet, Jr. and E. D. Haynes, eds., 197-212. NOAA Tech. Rep. NMFS Cir. 416. 610 Fisheries oceanography Kondo, H., Y. Hirano, N. Nakayama, and M. Miyake 1965 Offshore distribution and migration of Pacific salmon (genus Oncorhynchus) based on tagging studies (1958-1961). Inter. N. Pac. Fish. Comm. Bull. 17. Natarov, V. V., and N. P. Novikov 1970 Oceanographic conditions in the southeastern Bering Sea and certain features of the distribution of halibut. In: Soviet fisheries investigations in the northeastern Pacific, P. A. Moiseev, ed., Part 5, 292-303. U.S. Dep. Comm./NTIS. Larkin, P. A. 1975 Some major problems for future study of Pacific Salmon. Inter. N. Pac. Fish. Comm. Bull. 32:3-9. Novikov, N. P. 1970 Results of marking Pacific Halibut in the Bering Sea. Izvestia TINRO 74:328-9. Low, L. L., and J. Akada 1978 Atlas of groundfish catch in the northeastern Pacific Ocean, 1964- 1976. Nat. Mar. Fish. Serv., North- west and Alaska Fisheries Cent., Seattle, Wash., Proc. Rep. Potocsky, G. J. 1975 Alaska area 15- and 30-day ice fore- casting guide. Naval Ocean. Off., NOOSP-263. Washington, D.C. Takahashi, Y., and H. Yamaguchi 1972 Stock of the Alaska pollock in the eastern Bering Sea. Bull. Jap. Soc. Sci. Fish. 38(4) :383-99. Moulton, R. (ed.) 1939 The migration and conservation of salmon. Amer. Assoc. Adv. Sci. Pub. 8., Sci. Press Print Co., Lancaster, Pa. Waldron, K. D. 1978 Ichthyoplankton of the eastern Bering Sea, 11 February to 16 March 1978. Nat. Mar. Fish. Serv., Northwest and Alaska Fisheries Cent., Seattle, Wash., Proc. Rep. Ecosystem Dynamics in the Eastern Bering Sea Taivo Laevastu and Felix Favorite Northwest and Alaska Fisheries Center Seattle, Washington ABSTRACT The spatial and temporal dynamic aspects of the marine ecosystem in the eastern Bering Sea were investigated with the ecosystem model DYNUMES III. Equilibrium biomasses of major species and ecological groups are presented together with their annual consumption. The latter represents a major portion of the natural mortality, which removes a major part of the annual production of biomass, especially of forage species and faster- growing juveniles of other (larger) species. Examples of the dynamics of the biomasses of some species, with certain restraints on recruitment variations, over a four- year period are also given. Marine mammals, as apex predators, remove a considerable part of the fish production and the effect of their predation on the marine ecosystem is in many respects similar to that of the fishery. Some validation of the results of simulation model compu- tations is provided by comparisons with survey results, ad- justed by catchability factors. INTRODUCTION In the past it has been customary to consider the marine ecosystem and the living resources in it as a quasi-steady-state system. This conclusion has been reached by averaging the quantity of resources over large regions, such as the eastern Bering Sea, and over long time units, such as a year or more. However, conditions in marine ecosystems are not stable in space and time, but are affected by various seasonal changes (e.g., feeding and spawning migrations) and by environmental anomalies (e.g., varying extent of ice cover), all of which are rather pronounced in the Bering Sea. The objectives of this study were to investigate the effects and spatial and temporal magnitudes of the dynamics of the Bering Sea ecosystem, including the source and sink areas of biomasses of various species and the seasonal changes of these sources and sinks as well as the changing predator-prey relations. The eastern Bering Sea contains large numbers of marine mammals as apex predators whose numbers vary seasonally. Their effects on the ecosystem were also investigated and are briefly reported. Supporting data and details are being reported currently in various NWAFC Processed Reports. MAJOR PROCESSES IN THE MARINE ECO- SYSTEM AND A REVIEW OF THE ECOSYSTEM SIMULATION MODEL DYNUMES III The major dynamic processes in any marine ecosystem which affect the abundance and distribu- tion of species are listed in Fig. 37-1 together with the major factors affecting these processes and their results. The most difficult to reproduce quantitatively are the results relating to the spawning process, i.e., the larval recruitment and the early mortalities of larvae. After many decades of intensive research on stock- recruitment problems, we now know that no simple relation exists between the size of the spawning stock and the amount of recruitment to an exploitable stock. It can only be stated in general terms that proportionally larger recruitment can result from a small spawning stock and proportionally smaller recruitment can result from a large spawning stock. The numerous hypotheses on proper food availability to larvae and subsequent effects on recruitment size have not been proved either. Many fisheries scientists are inclined to believe now that recruitment numbers to exploitable stocks are controlled mostly by preda- tion in larval and juvenile stages (Hempel 1978, Rothschild and Forney 1979) and that recruitment might be largely influenced by environmental anoma- lies (Garrod and Colebrook 1978). The conditions affecting growth and reduction of species biomasses are generally known, although refinement of knowledge on the small variations in the growth of individual species is continuing. 611 612 Fisheries oceanography A. Processes affecting biomass abundance Main process SPAWNING (reproduction, larval recruitment) GROWTH PREDATION (mortality) MORTALITIES B. ■ (SPAWNING) Processes affecting biomass distribution SOURCE-SINK AREAS (Differences in abundance affecting factors in space and time) MIGRATIONS Seasonal Life cycle dependent Environment dependent Major affecting factors Spawning biomass size (including fecundity) Predation on eggs and larvae Availability of food (including starvation) Environmental factors (including advection) Age Temperature (anomalies) Food availability Vulnerability (sp. size) Predator abundance Senescent mortality Spawning stress mortality Disease mortality Growth Predation Other mortalities Feeding migration Search for optimum environment Spawning migrations Predation avoidance migrations Feeding migrations Search for optimum environment Advection by currents (APEX PREDATORS AND FISHERY) Figure 37-1. Major dynamic processes in the marine ecosystem. Predation has been included traditionally in the all- encompassing term "natural mortality," of which pre- dation constitutes the major portion in most species. In order to remedy this shortcoming the predation mortality is computed within ecosystem models in great detail, using data on food requirements for growth and maintenance and space- and time-variable food composition. Very few fish stomach analysis studies have been accomplished on Bering Sea stocks. Those available usually list only the frequency of occurrence of food items in the stomach, making these studies of little quantitative value. Fortunately recent European studies of fish feeding habits (e.g., Daan 1973) have established that feeding in the marine environment is largely dependent on the size of the fish and the size and availability of the food; this knowledge has been incorporated in our DYNUMES (Dynamical Numerical Marine Eco- system) model. In addition, special studies at NWAFC have quantitatively ascertained the age- Ecosystem dynamics 613 dependent spawning stress and senescent mortalities (Granfeldt 1979a), and these also have been included in the DYNUMES model. Because of spatial changes in factors affecting growth and in predator and prey distributions, considerable spatial and temporal differences in the growth and decline of biomasses occur (these are described later). All of these interactions are non- linear and therefore can be evaluated quantitatively only with three- or four-dimensional ecosystem models such as DYNUMES. Few epibenthic organ- isms are sessile during most of their lives. Most species undertake migrations for various reasons; seasonal migrations are connected most often with seasonal changes in the environment, both in relation to the availability of food and to the search for optimum environmental conditions (e.g., tempera- ture). In addition, pelagic organisms are advected with seasonally changing currents. Life-cycle depen- dent migrations are for spawning, feeding, and predation avoidance. Very little is known of the migrations of fish and especially of the distribution of juveniles in the eastern Bering Sea. Only the seasonal migrations of some flatfishes are approximately known in the North Pacific (Alverson 1960). The static conditions usually implied in past stu- dies of ecosystem productivity and of the effects of fishing will not give quantitative answers when prey is quasi-stationary and predators migrate or vice versa. Some of these migration effects are shown schemati- cally in Fig. 37-2. Consider that there is a given benthos biomass as a fish food resource at the section a-b. Under stationary conditions this benthos bio- mass is grazed by the stationary predator biomass (a-b) at this location. However, if the predator moves with a speed C through this section, the "upmigra- tion" of biomass A can prey on the same standing stock of benthos at the section a-b during migration. If the speed is doubled, the biomass A+B can prey on the same benthos biomass (time factor must also be considered above). This applies also to fishing: in the stationary case, the fish is caught (and sampled) as representing the biomass present at all times. However, with varying migration speeds and quanti- ties of biomass passing through the section, the effect of a constant fishery for different segments of the population would be different. This difference becomes more complex if the biomass age composi- tion also varies with time as the fish biomass moves through section a-b. The second part of the figure shows the self-explanatory effect of predation caused by separation and /or overlap of predator and prey. These concepts are simulated numerically in the model. Distance Predator Figure 37-2. Schematic presentation of two effects of migration. (I) Witii a given migration speed c the biomass A passes through the section a-b. Doubling either the speed or time, biomasses A+B pass through the same section. (II) Schematic example of the migration of a predator into the region of a given prey. The extensive and relatively complex ecosystem simulation model DYNUMES has been described elsewhere (Laevastu and Favorite 1978a, and in press). All necessary species-specific data, such as growth rate, temperature preference, and food item preference, are obtained from available empirical data and introduced into the model. The initial distributions of groups of species are prescribed at each grid point in a geographically oriented grid (Fig. 37-3A), together with other necessary space- and time-dependent data such as depth, monthly surface and bottom temperatures, ice cover, and migration speed (U and V components). The numerous computations are accomplished at each grid point at each time step (e.g., monthly). There are four basic formulas and a great number of auxiliary computations. The first basic formula is the biomass balance formula: Bi,t,n,m ~Bit_i n,m(2-e-^i,t,n,m)e ™ Cit-i,n,r (1) where Bj t-i,n,m is the biomass of species i in previous time step (t— 1) at the location n,m; g is the growth coefficient which is a function of temperature, sea- son, and availability of food; m is a mortality coeffi- cient (senescent mortality and spav^oiing stress mor- 614 Fisheries oceanography 175° E 180° 175° W 170° 165° 160° 155° 1 T - 65° N 165° 175° E 175° W 165° 155° 145° 135° Figure 37-3. Computation grid of DYNUMES for the eastern Bering Sea (A) and the fisheries management and statistical areas (B). tality), and C is predation (mortality) as computed in previous time step: Q,t,n,m ~ j Q,j,n,m ^^id (2) MJ ~ Rj,t,n,m nj4,t,n,m (3) where Rj,t,n,m is the total food requirement of species j and riji^tnm is the fraction of food of species j which consists of species i. This fraction varies in space and time (t,n,m) depending on the availabihty of suitable food items. The migrations are computed with an "upmigra- tion" interpolation finite difference formula: Bt,„,m =Bt-i.n.„, - (tdlUt,n,„|UT„,^)-(tdlVt,„,„|VT„,„)(4) where: t^ is the length of time step, U and V are mi- gration speed components, and UT and VT are migra- tion gradients of biomass. EQUILIBRIUM BIOMASSES IN THE EASTERN BERING SEA AND THEIR LONG-TERM DYNAMICS If the growth of all individual biomasses in a given ecosystem equals their removal by predation, fishery, and other causes of mortality, we can talk about equilibrium biomasses which can be sustained in this given ecosystem. These equilibrium biomasses can be determined under certain conditions: by ignoring migration effects and using the Bulk Biomass Model for given areas (Laevastu and Favorite 1978b), and by finding a unique solution for equations (1) to (3), provided part of the consumption (C) is predeter- mined. The predetermined consumption for the eastern Bering Sea has been obtained by computing the consumption by marine mammals as the first step in the computations, using predescribed monthly numbers of mammals and keeping their food require- ments and food consumption constant. Despite a number of other minor limitations of the Bulk Biomass Model (e.g., the food composition can vary only to a limited extent), the error of the results of this model does not exceed ± 30 percent of the given value of the biomass. The equilibrium biomasses in NWAFC Statistical and Management Areas 1, 2, and 3 (Fig. 37-3B) are given in Table 37-1. These biomasses are the "minimum sustainable equilibrium biomasses" (Laevastu and Favorite 1978a) obtained by using the highest plausible growth rates and the lowest plausible food requirements. The equilibrium biomasses (Fig. 37-4) are grouped by three different regimes (pelagic, semidemersal, and demersal). The semidemersal species dominate all other ecologi- cal groups (12 milhon mt), mainly because of their more flexible feeding habits. The biomasses of pelagic and demersal species are about equal (ca. 7.5 million mt each). The most abundant species is pollock (ca. 9 million mt), followed by cottids and other smaller, noncommercial demersal species (4 million mt), and capelin, other smelts, and sand lance (3.5 million mt). Salmon, which occur seasonally, are not included in Table 37-1 and Fig. 37-4. The biomasses of demersal stocks decrease more than one order of magnitude from Area 1 to the deep ocean regime. Area 3, whereas there is little relative change in the biomasses of the pelagic species be- tween the continental shelf and deep ocean regimes. The biomasses of semidemersal stocks also decrease from the continental shelf regime to the deep ocean regime, mainly because of the disappearance of the benthic food resource. The semidemersal species live a pelagic life and consume pelagic food over the deep water. Ecosystem dynamics 615 0.1 0.2 Biomass ( x 10 tons) 0.4 0.6 1.0 2 4 6 8 10 I I I I I I n I I I I I I I Capelin, sand lance Herring Squid Atka mackerel 1 Salmon ? Pollock Rockfishes Cods Sablefish Cottids and others Yellowfin sole, rock sole Alaska plaice Other flatfishes Flathead sole Arrowtooth flounder Halibut, turbot Shrimps Crabs Figure 37-4. Equilibrium biomasses of three different regimes in tiie eastern Bering Sea. Very little information has been available about the benthos. However, recent work, much of which is reported in Section XII of this book, improves this situation considerably. The total equilibrium bio- masses require about 50 g/m^ standing stock of benthos. The existence of this standing stock is entirely possible if we compare the eastern Bering Sea with the well-investigated Barents Sea. Quantitative zooplankton data from the eastern Bering Sea have been sparse; recent information is given in Section X of this book. Soviet studies in the early 1960 's were quantitatively deficient, giving only the minimum standing stocks of copepods and no quantitative data on abundant euphausiids (Laevastu et al. 1976). Since the total equilibrium biomasses consume about 50 g/m^ of zooplankton, the annual production of zooplankton must be at least this amount. The annual turnover rate (last column in Table 37-1) (turnover rate - annual consumption /mean standing stock) provides information on the preda- tion and other mortalities of the species or groups of species. In the marine ecosystem, the younger, smaller organisms are most vulnerable to predation. The growth rate of the biomass is highest at the younger ages. Thus the growth rates determine the length of the period during which a given species is most vulnerable to predation. The distribution of biomass with age and the predation-vulnerability (Fig. 37-5) are shown for two species as an example. The long-term dynamics of the biomasses in the marine ecosystem can be studied with the DYNUMES and BBM models after determining the equilibrium biomasses by introducing a cause of change in beha- vior or abundance of any species in the ecosystem. The results of such studies have limited reliability beyond a few years because of the uncertainty in predicting the survival to recruitment (spawning success). An example of the change in biomasses of a few species in Area 1 (see Fig. 37-3B) over four years, assuming that larval recruitment was directly propor- tional to the spawning biomass present, rather than a 50 r. WALLEYE POLLOCK -Age /size at which highly vulnerable to predation S 30 Age (years) Figure 37-5. Mean biomass and its annual production distribution with age in pollock and yellowfin sole. The portion of biomass highly vulnerable to predation is indi- cated. c o o o o OX) c PQ O t^ OJ CO ^ W a> J H PQ o < U3 ^ o c o o O) C 3 cr w 9i n 0^ CO CM CO C rt o O o o 0) -4-> t/1 lac c CO 3^ CO w (1) o PQ -»^ o CO o «! o q; .2 ., CV CO X (H OD t/1 i-1 CO c 0 CO PQ -^ 0 CO 0 0 0) CM 00 ■* CO -rfi in T}< CM Tt ts< o o o o o 03 CM to CD CD C- 10 CM o" O O T-i CO OS CO 10 era c- CD 00 !>■_ CD CM CM 03 ^' 00" O lO o CD CD T-H CO tH CD 1—1 T-H CM T-i in o uo tH 1-1 C~ i-H 03 00 CM (3^ c- CD CD in CM CM CM c- 00 t~ c^ 0 ^ CO in CM in 0 03 t> •^ in ^ 00 cd" 0 03 in CO CM in 00 in CO 0 CO CD ^ ^ in 00 CM 0 0 0 0 in 0 in in in ^ CM C- -^ t~ CO CO '^ '^ CD ■rti in CO 00' o CM 03 CO 00 CO T-H 00 o o in C~ ^ CD tH CM CM CD O O C<1 CO t—' CO C-' CO co" T-H CO CM T-H C- O CD 03 ^ 03 in ^ CO 03 c- c~ CO CM CO 03 CO ^ CM 1-1 0 0 CM CO CM 1-1 0 1— 1 CM CM 0 in CO CM 1-1^ CM 1—1 0 in 0 CO 0 03 03^ co'~ T-H CM in co^ 1—1 03 1-1 in_^ cm' 03 CD 03 in 1—1 03 CM C<1 in in in CO in CM T}< CM "^„ '^^ 00' 0' CM 1-1 00 0 CM CO 00 in ^ ca 00 CO CM CO iH 03 CM 00 '^ 0 CM D~ 03 1—1 in 1-1 1-1 1—1 CO 1—1 CO in CO in CM 1-1 03 1-1 CM 00 00 CD CD in 1-1 0 CM CO in ^ 0 in 00 03 03 co c- CO tr- 00 in in cr3 -* 00 0 03 00 CO 00 •<^ CO t~ I— 1 CO 1—1 1—1 0 00 03 C- 1-1 CM CM CM CO 03 CM 03 c~ 03 in 0 0 0 0 00 CO CD 1-1 1-1 0 CM CM CO_^ 1-1 0 CO 00 1—1 0 03 CO__ cm'^ in CM CO 00 0 in in i-T 00 CO 1-1 CM CD 0 00 1-1 0 in CO 00 03 CO 1-1 1-1 c~ oo_^ '^^ co'~ co' CO_ o o CD OJ cm' 00 CO 03 00 co_^ 00^ CM 1-1 c~ CO o 1-1 CO CM Tjt ^ in w CO 1-; ^. t- in 0 in c- co_ co_ in 00 -* ^. 03 03 0 Tt; 03 s £ CM co' -^' co' 0 i-i t-' c-" t~- 00' c~" 0 00' 03 03 00' C-' co' CO < '^ CO •^ CD in 03 CO 03 C- c- rj< CO '^ 0 c- c~ I-I 0 -IJ 1—1 CM -* CO ^„ CO 03_^ ^ in 0^ CO CM CO ^ CO in^ 0^ 0 iH 1— i cm"^ i-T 0" t-"^ 1— 1 CO in 1-1 00 CO tH 03 CO ^ 03 1—1 00 CO CO CM CO CD CO CO t 0 in T-I cm' 00' 00' tH co' 03 03 in 1-i co' CQ c~-' co' 1-i co' < CD '^ CD 03 C- CJ3 C^ r-l 0 tH CO 0 CM CM CM CO 0 CO CM 03 CO "* "^^ in in 0^ CM_^ T-H^ C- in in '^ ^ ^„ '^^ cm'^ in 1-1 1—1 cm' co' 0' o c O CO c « C 3 CO 3 CO T3 CO ?ti &: .s H 616 Ecosystem dynamics 61 7 function of the relation between equilibrium biomass and actual biomass (i.e., as conventionally assumed that large spawning biomasses produce proportionally smaller recruitment and vice versa) is shown in Fig. 37-6. The interactions of the biomass fluctuations are rather complex and would take too much space to analyze here, but computer printouts are available for this study. It should, however, be pointed out that there are "natural," quasiperiodic fluctuations of biomasses of considerable magnitude in the marine ecosystem. The period can be from a few yeeirs to more than a few decades and the magnitude can be considerable (e.g., the biomass can be a fraction of a few tenths to several times its long-term mean value). These fluctuations can have manifold causes in addition to man (fishing). They can, to a certain extent, be studied with ecosystem models. -4 L Yellowfin sole Atka mackerel Pollock Herring Figure 37-6. "Natural" changes of biomasses of yellowfin sole, Atka mackerel, pollock, and herring in the fisheries management Area 1 in the eastern Bering Sea. SEASONAL DYNAMICS OF THE BERING SEA ECOSYSTEM There are two basically different causes of seasonal spatial and temporal changes in the biomasses. The changes in abundance are caused by seasonally changing growth, predation (and other causes of mortality), and production and release of eggs and milt; the changes in distribution are caused by sea- sonal migrations of species and environmental inter- actions. The results of both major causes of seasonal dynamics must be viewed spatially. Unfortunately, little consideration has been given to the spatial aspects of biomass (and ecosystem) dynamics in the past, mainly because of difficulties in empirical study by nonsynoptic resource surveys. However, the gridded ecosystem models with spatial resolution such as DYNUMES make such studies possible. Examples of spatial and temporal aspects of biomass dynamics (Figs. 37-7 and 37-8) depict the biomass sources and sinks in February and August of two different-sized groups of pollock (juveniles, <22 65*- 60^ -N'leo^'-^' 170* <22cni February J 60" I 65* 60^ -iz- ^ :,-^^^ ,■-. .v?- -ViBO^'-^" 170- S— s 22-45 cm February _J :•- -^-: ."-c?*- -S'leo^-*^' 170- S— s 22-45 cm Auguit 160' _j Figure 37-8. Sources and sinks of maturing pollock (22-45 cm long) in February and in August in the eastern Bering Sea (in lOOkg/km^). cm, and maturing pollock, 22 to 45 cm). Source refers to the area where biomass growth in a given time interval (month) exceeds its losses by predation, fishery, and other causes of mortality; sink refers to the opposite condition, i.e., losses exceed growth. The sources and sinks of older pollock (>45 cm) are not shown because this size group has only sinks (large senescence and spawning mortality, fishery, and small growth rate) which are roughly proportion- al to biomass present. The sources and sinks of all species change because of spatial and temporal changes of the processes which cause them. There is usually a sink at the periphery of the distribution of the biomass. This sink is usually compensated for by outmigration from the center of main distribution (spreading). There is a nearly continuous source of pollock off the conti- nental slope over the deep water; during winter this source area is displaced southwest where the tempera- ture of the water is higher, allowing higher growth rates. The distribution of the three different age groups of pollock in August is shown in Figs. 37-9, 37-10, and 37-11. A partial separation of juvenile and old pollock is brought about by cannibalistic predation of older pollock on its own juveniles. The highest concentration of biomass of older pollock is found off the continental slope, whereas the juveniles are found mainly on the continental shelf. The effects of seasonal depth-migrations of yellow- fin sole on its distribution changes are shown in Figs. 37-12, 37-13, and 37-14. The seasonal depth migra- tions of flatfish were investigated by Alverson (1960) and on the basis of his work it was assumed that yellowfin sole migrate from deep water into shallow water during May and June (Fig. 37-12) and back into deep water in October and November (Fig. 37-13). A migration speed of 3 km/day was assumed in the model. It is not always possible to determine a single cause of seasonal migrations, which can be spawning, search for food, or search for acceptable or optimum environmental conditions. This resulting monthly change of biomass can be considerable in an area where active and extensive migrations occur (see Figs. 37-12 and 37-13). These seasonal migrations have profound effects on other biota as well as on the evaluation of fishery resources by trawling surveys. For example, flatfish are de- pendent on benthos as a food source and migrations cause heavy grazing of benthos in some areas during some seasons, allowing a recovery period during other seasons. A proper trawling survey evaluation must account for seasonal migration to avoid erroneous results. Ecosystem dynamics 619 -3-' 180' Figure 37-9. Distribution of juvenile polloclt (<22 cm long) in August in the eastern Bering Sea (tons/km ). Figure 37-10. Distribution of "maturing pollock" (22-45 cm long) in August in the eastern Bering Sea (tons/km^). Seasonal migrations are affected by various envir- onmental anomalies; water temperature anomalies, especially at high latitudes as in the Bering Sea, are usually the most pronounced and the easiest to observe. Furthermore, more is known of the effects of temperature on the species than of the effects of other environmental variables. Two dynamic effects of temperature anomalies are included in the DYNUMES model: the "forced" migration of most species out of areas with subzero bottom tempera- tures (including a slightly increased mortality), and the effect of temperature on food uptake and growth. An example of the effect of water temperature anomalies on the growth of the herring biomass is given in Fig. 37-15, showing the sources and sinks of the biomass for a February with average or normal water temperature, and for a February with a 1.5-C positive temperature anomaly. The growth of bio- mass is considerably enhanced in the latter case, especially in the southern, warmer part of the area. Cold anomalies affect growth less than warm anoma- lies, as growth is nearly arrested at low temperatures. However, temperature anomalies of cold bottom water have considerable effect on seasonal migrations of flatfishes to feeding and spawning grounds in shallower water. In years with extensive cold bottom water formation on the shelf in the winter, the spring migrations of flatfishes to shallower water can be considerably delayed (Best, this volume). 65"*- ^^S^^^~^ Figure 37-11. Distribution of "old pollock" (>45 cm long) in August in the eastern Bering Sea (tons/km ). Figure 37-12. Change of yellowfin sole biomass distribution due to migration in May and June in the eastern Bering Sea (tons/km^ ). ^ \^^ 65* >'l80^'-'' 1 70* Y-M October 160* _l '<^ ^^** 65* 60- ..• .•^-■r-'.'i'-^'" -V'leo^-^" 170* Y-M November 160* __i Figure 37-13. Change of yellowfin sole biomass distribution due to migration in October and November in the eastern Bering Sea (tons/km ). 620 Ecosystem dynamics 621 Figure 37-14. Distribution of yellowfin sole in August in the eastern Bering Sea (tons/km^). EFFECTS OF SPACE AND TIME VARIATIONS IN THE DISTRIBUTION OF PREDATOR AND PREY Predation in any marine ecosystem is largely controlled by predator-prey size relations and the availability and suitability of prey. Thus if some "normal" or preferred prey items are not available, other food items of comparable size are substituted. This process is included in the DYNUMES model. In areas and times when not enough food is available, starvation will occur with several of its consequences, the first of which is usually a delay of sex products development. Some of the dynamic aspects of predator-prey "overlap," pertaining especially to benthos and to seasonal migrations of flatfish, have already been presented (see Fig. 37-2). The effects of spatial distribution of different prey items on the composi- tion of food of a predator are schematically shown in Fig. 37-16, which depicts a vertical section with predator-prey distribution. Not only does the food composition of the predator vary in space, but the predation pressure on the prey varies as well. The various possible types of responses of predator and prey bio masses are shown in Fig. 37-17; although these responses can be classified into some defined types, there is considerable overlap of these responses in nature due to various predator-prey dynamic p^'teo"*-''-^'' 170' I I 160° _j Figure 37-15. Effect of temperature anomaly on the source and sink of herring in February in the eastern Bering Sea (A: normal February. B: February with a + 1.5 C temperature anomaly.). 622 Fisheries oceanography Predator and relative consumption of prey (PI, P2, P3) Prey 1 p- // and relative ^ predation (P) Prey 2 ^ Prey 3 Distance Figure 37-16. Schematic presentation of the predation by one predator on three prey items with different spatial distributions (presented as a section). processes (i.e., changes in space and time). Some of the response types are as follows: when the predator biomass decreases (e.g., because of outmigration), local prey increases in abundemce (type A), and when predator biomass increases, prey decreases (type B). When the secondary prey is predator to primary prey (i.e., secondary prey is competitor to primary preda- tor), the predator biomass decreases (types C and F). Type D Prey Predator Predator Prey . Prey 2 (predator to prey 1) Predator Prey 1 Time - Predator dynamic, prey stationary Prey dynamic, predator stationary Figure 37-17. Types of responses to biomass changes in predator-prey controlled ecosystem. Prey (as competitor) Predator If the prey is mobile and decreases (e.g., by outmigra- tion), the predator biomass might decrease as well (e.g., because of starvation or forced outmigration in search of food) (type D) and vice versa (type E). The percent of mean monthly zooplankton stand- ing stock consumed per month in the eastern Bering Sea is given in Fig. 37-18. Zooplankton and benthos are the main food resource buffers in the marine ecosystem. The monthly mean zooplankton standing crop was simulated in the model, using quantitative knowledge of its past abundance and seasonal changes. The simulated zooplankton standing stock was made to vary spatially and temporally between 400 and 800 mg/m^ . The areas of high zooplankton consumption changed from month to month as affected by the distribution of consumers. This spatial and temporal change of high consumption would allow replenishment by growth and advection in previously heavily grazed areas. It can be assumed that fish eggs and larvae are consumed at the same rate as zooplankton. Thus, if a high zooplankton consumption area coincides with high abundance of pelagic eggs and larvae, low larval survival and low recruitment can be expected of the given species. Since the zooplankton consumption in the northern part of the area is low, a great part of the deceased zooplankton can settle to the bottom, providing detrital food for the abundant benthos. The con- sumption of benthos is also low in the northern part of the area, although its standing crop is high ("ac- cumulation of generations" in the sense of Spark). Thus the benthic biomass (or zooplankton biomass) cannot be proportional to the production of fish in all areas. The consumption of zooplankton near the shelf edge is fairly high (see Fig. 37-18) and this raises several questions about the possible source of zoo- plankton there. First, the zooplankton simulation in the model for this area might be too low despite the fact that considerably higher quantities were simula- ted than indicated in the literature, because present quantitative zooplankton catching methods are ob- viously deficient (Laevastu et al., 1976). Second, the zooplankton, especially euphausiids, might be trans- ported by currents to the convergence at the conti- nental slope. EFFECTS OF MAMMALS AND BIRDS AS APEX PREDATORS ON THE ECOSYSTEM During summer the Bering Sea contains more mammals per unit area than ciny other ocean area. Furthermore, during summer there are more marine birds in the region (over 40 million) than in the rest of the northern hemisphere. Both mammals and Ecosystem dynamics 623 l^ \J^ 65*- I60« _J Figure 37-18. Percent of monthly mean zooplankton standing stock consumed in February and in August in the eastern Ber- ing Sea. birds are apex predators, their effect on the ecosys- tem being similar to the effect of fishing by man. In addition to the fact that the occurrence of mammals is seasonal and their distribution uneven, their mobility has various temporal and spatial dynamic effects on the rest of the ecosystem. Esti- mates of numbers of marine mammals present vary considerably from one source to another. It is apparent that some of the estimates have little to do with reality . There are three types of seasonal occurrence of mammals— the winter visitors (e.g., "ice" seal and bowhead whale), summer visitors (e.g., fur seal, sea lion, and sperm whale), and year-round residents (e.g., beluga whale). The estimates of daily food requirements of marine mammals are also variable in the literature. The more conservative estimates are between 4 and 8 percent body weight daily, depending on the species, and these estimates were used in the model. Seasonal changes of some marine mammals and birds are shown in Fig. 37-19; these estimates are believed to be the best and most realistic estimates, found in various more reliable sources in the litera- ture. The consumption of fish by marine mammals and birds in the area is about 3 million mt a year, which is about twice the total area catch by the fishery of aU nations. The effect on the ecosystem of consumption by marine mammals is similar to the effect of the fishery— continual removal of biomass without pro- viding much return in the form of food for other biota in the ecosystem. Considering the magnitude of consumption by mammals in relation to harvest by Murres 4r Rinyed and ribbon seals I I iT I I I I I I I I I 1 4 8 12 Month Figure 37-19. Monthly numbers of some mammals and birds in the eastern Bering Sea. 624 Fisheries oceanography the regulated fishery, it becomes obvious that fisher- ies management measures will have a limited effect on fishery resources without simultaneous management of marine mammals as well. Furthermore, the sea- sonal and selective predation by many mammal species has considerable adverse effects on several fishery resources important for man (e.g., the preda- tion on returning salmon by the beluga whale, fur seal, and sea lion). VERIFICATION AND VALIDATION OF ECOSYSTEM MODEL RESULTS Verification of the ecosystem simulation model (DYNUMES III) has been accomplished by ensuring that the formulas used in the model reproduce known effects and behaviors. Further verification has been done by simulating events which are known to produce given changes in the ecosystem. Thus the evaluation of sensitivity in large ecosystem models becomes a continuing study of the response of the ecosystem to changes in various rate and state param- eters. Although the dynamic aspects of the Bering Sea ecosystem are difficult to validate empirically because (Granfeldt 1979b). Examples of this validation are given in Table 37-2 for the species which are most reliably reported quantitatively in the resource surveys. In general, the results of resource surveys are considered reliable only to about ±50 percent (Grosslein 1976), whereas the error in the model computation results does not exceed ±30 percent of the reported value. Qualitative validation of the simulation models can be provided by occasional special fisheries surveys. In the early stage of the Bering Sea ecosystem modeling, it became obvious that there must be considerable numbers of pollock (and some other fish species) over the deep water in the Bering Sea. However, pollock were never caught over deep water and the model results were severely criticized until a recent Japanese survey showed considerable numbers of older (larger) pollock over deep water. Furthermore, the deep- water areas turned out to be the source of biomass of many pelagic and semipelagic species at least part of the year. The abundant euphausiids in this area provide ample food. However, no extensive schooling occurs over deep water, making the fishery less profitable there than over the continental shelf. TABLE 37-2 Comparison of exploitable biomasses of some species in the eastern Bering Sea as obtained by surveys and as computed with PROBUB Model (in 1,000 mt) Mean 1975, 1976 Equilibrium surveys (con- exploitable verted) from biomass from Species/groups Bakkala and PROBUB Catch of species Smith (1978) Model 1975 Greenland turbot, halibut 176 222 65 Flathead sole, arrowtooth flounder 206 377 26 Yellowfin and rock sole, Alaska plaice 2,716 509 74 Pollock 3,698 6,449 1,285 Cod 233 746 57 of the absence of data, particularly from extensive time-series and behavioral studies, it is possible to validate general abundance and distribution results by comparing the computed abundance and distribution of biomasses with independently obtained empirical data such as those obtained from fisheries surveys. This comparison is rewarding only provided the latter are properly converted using catchability coefficients REFERENCES Alverson, D.L. 1960 A study of annual and seasonal bathymetric catch patterns for com- mercially important groundfishes of the Pacific northwest coast of North America. Pac. Mar. Fish. Comm. Bull. 4. Bakkala, R.G., and G.B. Smith 1978 Demersal fish resources of the eastern Bering Sea: Spring 1976. Nat. Mar. Fish. Serv., Northwest and Alaska Fish. Cent., Seattle, Wash., Proc. Rep. Daan, N. 1973 A quantitative analysis of the food intake of North Sea cod, Gadus morhua. Netherlands J. Sea Res. 6:479-517. Ecosystem dynamics 625 Garrod, D.J., and J.M. Colebrook 1978 Biological effects of variability in the North Atlantic Ocean. Rapp. Proc- verb. Reun. Cons. Int. Explor. Mer. 178:128-44. Granfeldt, E. 1979a Interactions between biomass distri- bution, growth, predation and spawn- ing stress mortality. Nat. Mar. Fish. Serv. Northwest and Alaska Fish. Cent., Seattle, Wash., Proc. Rep. 7919. 1979b Estimation of vulnerability, availabili- ty, and catchability factors and exploitable biomasses in the Bering Sea and western Gulf of Alaska. Nat. Mar. Fish. Serv., Northwest and Alaska Fish. Cent., Seattle, Wash., Proc. Rep. 7912. Grosslein,M.D. 1976 Some results of fish surveys in the midAtlantic Bight, important for assessing environmental impacts. Amer. Soc. Limnol. Oceanogr. Spec. Symp. 2:312-28. Laevastu, T., J. Dunn, and F. Favorite 1976 Consumption of copepods and eu- phausiids in the eastern Bering Sea as revealed by a numerical ecosystem model. Paper L:34, ICES CM. 1976, Plankton Comm. Laevastu, T., and F. Favorite 1978a Numerical evaluation of marine eco- systems. Part I. Deterministic Bulk Biomass Model (BBM). Northwest and Alaska Fish. Cent. Seattle, Wash., Proc. Rep. 1978b Numerical evaluation of marine eco- systems. Part II. Dynamical Numerical Marine Ecosystem Model (DYNUMES III) for evaluation of fishery re- sources. Nat. Mar. Fish. Serv. North- west and Alaska Fish. Cent., Seattle, Wash., Proc. Rep. Holistic simulation of marine eco- system. In: Analysis of marine ecosystems, A. R. Longhurst, ed. Academic Press, Inc., London (in press). Hempel, G. 1978 Synopsis of the symposium on North Sea fish stocks-recent changes and their causes. Rapp. Proc. -verb. Reun. Cons. Int. Explor. Mer. 172:445- 9. Rothschild, B.J., and J.L. Forney 1979 The symposium summarized. In: Predator-prey systems in fisheries management, H. Clepper, ed. Sport Fishing Inst., Washington, D.C. ■&G P.O. 796-495 I ADD0Dm3ED33D