Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
THE BERING SEA ECOSYSTEM: GEOLOGY, PHYSICS, CHEMISTRY, AND BIOLOGY OF LOWER TROPHIC LEVELS 28 3 The Bering Sea Ecosystem: Geology, Physics, Chemistry, and Biology of Lower Trophic Levels MARINE GEOLOGY The Bering Sea is framed by the Seward and Chukchi peninsulas on the north, by the Kamchatka Peninsula on the southeast, and a 1,900 km long ridge and island chain that comprise the Aleutians to the south and southwest (Figure 3.1). Geographically, the Bering Sea lies between 52¹ and 66¹ north latitude, and 162¹ east and 157¹ west longitude. The narrow, 85 km long passage of the Bering Strait connects the Bering Sea on the north to the Chukchi Sea and the Arctic Ocean. This subarctic Bering Sea lies in the northern part of the Pacific Basin. It exchanges water with the Arctic Ocean (through Bering Strait) and with the Pacific Ocean, into the Bering Sea from the Gulf of Alaska through the Aleutian Islands and into the northwest Pacific Ocean through Kamchatka Strait. The Bering Sea covers almost 3 million km2 and is unusual in having an extremely wide continental shelf, ranging from 500 km wide in the southeast region to over 800 km wide in the north. The Bering seafloor is partitioned into a series of sedimentary basins, the principal of which are (1) the Aleutian Basin, just north of the Shirshov Ridge, Bowers Bank, and Aleutian Islands; (2) Komandorsky (Commander) Basin, adjacent to the Komandorsky Islands, Kamchatka Peninsula, and Shirshov Ridge; (3) Bowers Basin, enclosed by the arm of Bowers Bank; (4) Anadyr Basin, encompassing the Gulf of Anadyr; (5) Chirikov Basin, adjacent to the Chukotka Peninsula and Bering Strait; (6) Norton Basin, between Alaska's Seward Peninsula and Yukon- Kuskokwim delta; (7) Bristol Basin, between the Alaska Peninsula and mainland, and (8) the Beringian Shelf, which encompasses the remainder of the continental shelf from Chukotka Peninsula, St. Lawrence Island, and Alaska mainland southwestward to the extensive submarine canyon system described below (Hood and Kelley, 1974; Sharma, 1977). The Bering Sea region shelf is unusual from the global perspective in being extremely smooth and generally featureless (Figure 3.2), with the exception of three large and some small islands. Its gradient is among the gentlest in the world (0.24 m/km), with bathymetry less than 200 m deep and a very steep continental margin (Sharma, 1977). This continental shelf is incised by seven of the largest submarine canyons in the world. From north to south, they are Navarinsky, Pervenets, St. Matthew, Middle, Zhemchug, Pribilof, and Bering canyons. The abyssal Aleutian Basin, which lies at depths of between 2,800 and 3,600 m, constitutes the southwest portion of the Bering Sea and is adjacent to the Aleutian Islands, Komandorsky Islands, and Kamchatka Peninsula. Two submerged mountain chains, the Bowers and Shirshov
THE BERING SEA ECOSYSTEM: GEOLOGY, PHYSICS, CHEMISTRY, AND BIOLOGY OF LOWER TROPHIC LEVELS 29 Figure 3.1 Physiographic features of the Bering Sea seafloor (adapted from Hood and Kelley, 1974).
THE BERING SEA ECOSYSTEM: GEOLOGY, PHYSICS, CHEMISTRY, AND BIOLOGY OF LOWER TROPHIC LEVELS 30 Figure 3.2 A side-scan sonar image of the Bering Sea seafloor (EEZ-SCAN Scientific Staff, Atlas of the U.S. Economic Zone, Bering Sea, U.S. Geological Survey miscellaneous investigations series 1-2053 1991).
THE BERING SEA ECOSYSTEM: GEOLOGY, PHYSICS, CHEMISTRY, AND BIOLOGY OF LOWER TROPHIC LEVELS 31 ridges, divide this basin into the separate Aleutian, Bowers, and Komandorsky basins (Cooper et al, 1987a). The Bering Sea became separated from the rest of the Pacific Basin through evolution of the Aleutian Island arc during the Eocene. This arc of islands with active volcanoes was established at the current site of continental drift and collision, where the northward-moving Pacific is subducted beneath the North American Plate. Preceding the formation of this island arc, the Beringian margin was the site of plate subduction, with fault block basins forming to the southwest in response to tectonic forces. The rocks found in the Alaska-Bering Sea-Siberia region appear to have moved several thousand kilometers northward during the Mesozoic era and were accreted in their approximate current locations by the Paleocene epoch. Bedrock geology of the offshore areas of the Bering Sea is not well known; only a few rock samples have been collected or drilled, leaving seismic reflection geophysical data to provide basic information. The Aleutian Basin is believed to be underlain by Mesozoic oceanic crust, and has been a center for sediment deposition of between 2 and 9 km since formation of the Aleutian Island arc. Deposition in smaller basins, such as the Bowers and Komandorsky, consists of a thinner sedimentary section of between 1 and 3 km. Bedrock underlying the Beringian Shelf is comprised primarily of Paleozoic, Mesozoic, and Cenozoic sedimentary rocks, and Tertiary and older volcanic rocks. Formation of the submarine canyon system of the Beringian margin occurred in Cenozoic glacial periods during low stands of sea level (Carlson and Karl, 1988). For more detailed discussions of the geologic evolution and history of the Bering Sea region, see Bogdanov and Neprochnov (1984), Carlson and Karl (1988), Churkin (1972), Cooper et al. (1991, 1987a, 1987b), Dundo and Lopatin (1989), Grantz et al. (1970a, 1970b), Hood and Kelley (1972), Lisitsyn (1966), Marlow et al. (1976), Scholl et al. 1966, 1968, 1978), and Sharma (1977). Sediments The configuration of the Bering Sea continental shelf, margin, and abyssal basins, in association with weather patterns of the region, influences the basal physical oceanographic conditions present in the Bering Sea. Circulation models indicate that tides and the Bering Slope Current flow to the northwest paralleling the continental slope, although they are complicated by countercurrents and eddies developed near the canyons (Kinder et al., 1975). The breadth and size of the continental shelf contribute to the huge storm waves of the Bering, which are capable of moving sediment at outer shelf and slope depths. Principal unconsolidated seafloor sediments are gravels, sands, silts and clays derived from previous glacial action, river erosion and deposition, and ice rafting. Again, the shelf is very flat, except for a few basins and banks, with transitory swales and ridges, channels, and depressions. Sediments of the partially enclosed bays, such as Bristol Bay, Norton Sound, and Chirikov Basin, and areas adjacent to islands of the shelf have more sands and gravels than clays, but clays and silts predominate on the deeper shelf. Accumulations of organic materials in shallow sediments of Norton Sound, Chirikov Basin, and elsewhere in the northern Bering Sea have undergone chemical transformation into natural gases. Gases ranging from methane to carbon dioxide have been found to seep into the
THE BERING SEA ECOSYSTEM: GEOLOGY, PHYSICS, CHEMISTRY, AND BIOLOGY OF LOWER TROPHIC LEVELS 32 water column at a variety of seafloor locations either derived from circular blowout pits or along faults. It has been postulated that the gas has a proclivity for release during storms when shallow sediments are disturbed by wave-action. These sites can be extensive: 7,000 km2 in Norton Sound alone are underlain by areas of potential shallow gas pockets (Larsen et al., 1980). The effects of gas seepage on marine life in the Bering Sea, positive or negative, are unknown at this time. Sediments deposited annually in the Bering Sea are derived primarily from the Yukon River (88 to 100 million metric t), with lesser amounts from the Kuskokwim (4 million t) (Sharma, 1977) and the Russian coast. Plumes of sediment from the rivers draining the Russian and Alaskan coasts are easily seen in satellite photography, and these plumes concentrate suspended sediment load near the coastal regions. Sharma (1977) details the mechanisms of ice rafting of sediment: fine sediments are apparently incorporated during ice formation from suspended sediment in water and resuspension of bottom sediment by storm waves in shallow water; ice-rafted clay, silt, and biogenic materials drop into the water column after the winter months along the receding ice edge margin-amounts have not been quantified. Bottom sediments derived from the Alaskan mainland rivers and coastal erosion are deposited on the seafloor and then swept by northward currents toward the Bering Strait. Sediments from the Russian coast deposited on the Bering Shelf are swept northward and eastward toward the Bering Strait as well. Materials derived from the Aleutian Islands and Kamchatka Peninsula are mostly deposited into the abyssal basins. Figure 3.3 shows sediment distribution on the northern Bering Sea shelf. Coastal Geography Geography of the coastal areas bordering the Bering Sea is strongly influenced by the geologic forces that shaped and are shaping the region: the relentless erosional forces of the Bering itself and the subarctic climatic conditions. Alaska Regional Profiles (undated) contains detailed maps and descriptions of the coastal regions of Alaska, while Landscape Atlas of the USSR (Plummer et al., 1971) profiles aspects of the Russian coast. The local topography and resultant water, surface materials, and vegetation affect the presence of local and regional biota. The southern border of the study region is bounded by the Aleutian Islands, a chain of volcanic islands, many of which are still active, driven by tectonic forces. The islands extend more than 1,100 miles (1,770 km) and consist of more than 50 islands, in five groups, separating the Bering Sea from the northern Pacific Ocean. The Aleutian and Shumagin islands are low mountains with steep to moderate slopes and rolling topography. Plateaus and uplands occur in some places in the chain. Elevations of the islands range from sea level to nearly 5,000 feet (1,524 m). A good number have wave-derived terraces up to 600 feet (183 m) above sea level, and are bordered by lower sea cliffs from previous sea level stands. Generally broad and flat intertidal platforms derived from glacial period sea level changes surround some islands. Those islands with peaks higher than 3,000 feet (914 m) were heavily glaciated and include fjords extending up to 2,000 feet (610 m) into the sea. Frequent lakes occur in ice-derived basins on islands showing signs of glaciation, while streams move water in high-gradient descent
THE BERING SEA ECOSYSTEM: GEOLOGY, PHYSICS, CHEMISTRY, AND BIOLOGY OF LOWER TROPHIC LEVELS 33 to the ocean. Where volcanoes have erupted in the recent past, the presence of ash, cinders, or lava has either killed local vegetation or provided a poor substrate that is resistant to revegetation for a number of years. Figure 3.3 Sediment distributions of the Bering seafloor based on data from Sharma (1977) and Hood and Kelley (1974). The Pribilof Islands are five small islands in the Bering Sea that lie 200 miles (322 km) north of the Aleutian island Unalaska. St. George, one of two populated islands in the group, has hills and ridges with steep cliffs rising up to 900 feet (274 m), whereas St. Paul has a rolling plateau with some extinct volcanic peaks. The islands of St. Matthew, Pinnacle, and Hall are located in the Bering Sea north of the Pribilofs and approximately 220 miles (324 km) west of mainland Alaska. These islands have volcanic ridges separated by valleys; highest elevations of approximately 1,500 feet (458 m) are associated with volcanic cones. Shorelines of these islands are mostly quite steep, and St. Matthew has some lagoons and freshwater lakes. Islands of the Bristol Bay area of Alaska have extensive glacial outwash and morrainal deposits at the base of volcanic islands at their cores. These sediments are bisected by streams and small lakes, and include beaches with sand dunes. Other parts of the southwestern Alaska coast have unconsolidated gravel, sand, silt, and clay of glacial origin. Lowland deltas and
THE BERING SEA ECOSYSTEM: GEOLOGY, PHYSICS, CHEMISTRY, AND BIOLOGY OF LOWER TROPHIC LEVELS 34 coastal plains have formed in the material, with extensive beaches, spits, and offshore bars, dunes, and islands. Permafrostâpermanently frozen ground, remaining frozen at least one winter to the nextâis found throughout much of Alaska and the Russian Far East, with depths of less than 1 m up to 610 m (2,000 feet) or more bordering the Arctic Ocean. In general, permafrost is mostly absent in the Aleutians and Alaska Peninsula, discontinuous in the area south of the Kuskokwim River (60¹ north latitude), and continuous to the north along the coast. Similar depths have been reported in Russia, with the southern permafrost limit found in the south- central Kamchatka Peninsula, then stretching north to the Bering Strait; the highest content of ice is found in the deposits of the northern coastal plains and islands of Novosibirsk, where ice may constitute 10 to 15 percent of the soil at depths up to 10 m (30 feet). Permafrost limits rooting depth of plants. Where an active layer that thaws in summer and refreezes in winter may be present in the permafrost, the melting results in small pools of standing water conducive to marsh and tundra development and accumulation of peat deposits. Certain kinds of features are formed in permafrost environments that border the Bering Sea coast including polygonal ground, stone nets, thaw lakes, and beaded drainage (Ferrians et al., 1969). Small areas of high brush, spruce-hardwood forest, and spruce-poplar forest occur only along drainages of the major and some medium sized rivers that drain toward the sea (the Kuskokwin and Yukon, for example). High brush may be found along the largest of the Aleutians in southerly-facing mountain slopes primarily bordering the Pacific Ocean. The southwestern part of Alaska is one of the most seismically active areas of North America and the world. Over the last 100 years, nine Alaskan earthquakes have been recorded with magnitude of 8 or greater on the Richter scale, with a large number of additional earthquakes registering 7 or more. The results of these earthquakes include faulting, cliff collapse, landslides and slumping, changes in water drainage patterns, tsunamis with their resultant wave-based erosion and destruction, underwater slumps, and increased water turbidity (Eckel, 1970). The most recent destructive earthquake in the region (magnitude 7.5) struck the Kamchatka Peninsula in June 1995. Volcanoes and volcanic ranges border nearly the entire Bering Sea coast. Nearly 30 active volcanoes are found in the southwestern part of Alaska, almost all of which have been active since 1760 (Alaska Geographic, 1976). Kamchatka has 127 volcanic cones, of which approximately 15 are active (National Geographic, 1994). Volcanic, eruptions result in earthquakes, magma and ash flows into coastal and marine areas, land and water temperature induced changes, changes in local and regional substrates, and increased ocean turbidity in the vicinity and downwind of volcanic ash or magma eruptions. The impacts on local and regional flora and fauna include die-offs from smothering or heat effects, and migration due to changed ecological conditions. In addition, the mountain ranges formed by the volcanoes, especially those along the Russian coast, intercept the landward flow of air from the Pacific, causing localized high precipitation zones. The Yukon delta consists of tidal flats and lowlands with lakes, ponds, and tundra, and also some highlands, such as the Cape Romanzoff Bluffs. The river deltas of the Yukon and Kuskokwim are significant because they provide a major nesting area for waterfowl, including geese, ducks, and swans. Raptors are not common in tundra habitats, but they are found in cliff areas near rivers and the Bering Sea. The cliffs, such as those at Nunivak, also provide suitable
THE BERING SEA ECOSYSTEM: GEOLOGY, PHYSICS, CHEMISTRY, AND BIOLOGY OF LOWER TROPHIC LEVELS 35 habitat for seabird rookeries (for kittiwakes, murres, puffins, and gulls). Passerines are less abundant on the deltas than shorebirds, and leave by mid-September. The beaches, sea cliffs, tundra, and willow-lined water courses bring in many smaller birds, including Old World species. The extensive beach environments, ranging from rocky to muddy, attract intertidal fauna that are food sources for large populations of birds and marine mammals. Sandy beaches are particularly important for harbor seals, whereas Steller sea lions are found along the rocky coast. Fur seals prefer the coasts of the Pribilofs. Sea otters inhabit the shallow waters along the islands, where offshore reefs and kelp beds provide safety and food. Erosion of beaches and coastal marches generally occurs at a slow rate across southwestern Alaska. However, storms can severely affect local or regional areas through storm surges and waves, as undercutting of cliffs and bluffs by waves results in collapse and cliff retreat. The fallen materials become available for redistribution in the sea or down current along the coasts. PHYSICAL OCEANOGRAPHIC STRUCTURE Physical Characteristics and Bathymetry of the Northern North Pacific and Bering Sea The circulation of the northern North Pacific and Bering Sea, seen in a global context, serves to transport heat and freshwater poleward. It also replenishes the nutrients in the surface layers to support biological productivity. Changes in this transport system affect heat, salt, and food supply for the Bering Sea ecosystem. The Basins The northeast Pacific Ocean contains the Gulf of Alaska, a feature bounded on the northern and eastern sides with a moderately wide (100 km) shelf. It has relatively free communication to the south and west. Water depths are in excess of 4,000 m; the deepest feature is the Aleutian Trench, in close proximity to the Aleutian Island arc (Figure 3.4). The Aleutian-Komandorsky Island arc forms a semipermeable boundary between the Gulf of Alaska and the Bering Sea. The ability for water exchange between the North Pacific and Bering Sea depends on the water depth in the passes. Since there are no deep passes east of 180¹ W, deep water exchange is restricted to the western side of this boundary. Gulf of Alaska Circulation Deep ocean circulation for the Gulf of Alaska consists of a large (1,000 km) cyclonic (counterclockwise) gyre that advects warm water northward along the British Columbia coast and southeast Alaska. In its northward flow off the British Columbia shelf break, the current
THE BERING SEA ECOSYSTEM: GEOLOGY, PHYSICS, CHEMISTRY, AND BIOLOGY OF LOWER TROPHIC LEVELS 36 is wide and unorganized (Tabata, 1982). As it is diverted westward at the apex of the gulf, it forms a western boundary current and becomes a much more coherent flow. The currents have speeds of 25 to 75 cm s-1 and transports on the order of 10 to 30 Sverdrups (Sv) (1 Sv = 1 million m3 s-1) (Dodimead et al., 1963). The current continues southwestward along the shelf break in the western Gulf of Alaska, closely following topography, until it reaches a longitude of about 180¹ W. Here, it tends to continue on a zonal, or sometimes a southwestward, trajectory (Thomson, 1972). Figure 3.4 Bathymetric map of the Bering Sea (Sayles et al., 1979) The northern North Pacific is the terminus for the world oceans deep circulation. Deep water formed in other high-latitude regions of the North Atlantic and Southern Ocean reaches the Gulf of Alaska after traveling for centuries. Gulf of Alaska deep waters are therefore some of the oldest waters in the world's ocean and have very high nutrient concentrations. Unlike most other high-latitude regions, the North Pacific is not a site for the formation of deep water. The reason for this absence of deep water is the relatively low density surface water that is created by high rates of precipitation and runoff in the region. The northeast Pacific has been thought of as analogous to an estuary (Tully and Barber, 1960), with the halocline and accompanying pycnocline serving as a cap on the ocean in the North Pacific and Bering Sea. The deep ocean circulation is accompanied by a very active coastal circulation. High rates of precipitation over the Gulf of Alaska and coastal freshwater discharge around the
THE BERING SEA ECOSYSTEM: GEOLOGY, PHYSICS, CHEMISTRY, AND BIOLOGY OF LOWER TROPHIC LEVELS 37 perimeter create a low-salinity surface layer that contributes to a cyclonic flow over the shelf, especially adjacent to the coast (Reed and Schumacher, 1986). On average, more than 23,000 m3 s-1 of terrestrial freshwater discharge enters the coastal region, creating a buoyancy-influenced alongshore flow in the Gulf of Alaska that is constrained along the coast by the cyclonic (downwelling) wind system. Current speeds in excess of 175 cm s-1 have been reported for this flow called the Alaska Coastal Current (Royer, 1982). The width of this current is generally less than 20 km. It flows westward in the northern Gulf of Alaska and along the Aleutian Island arc to Unimak Pass, where a portion of it enters the Bering Sea (Figure 3.5). This coastal flow continues around the Bering Sea in a cyclonic sense, until it flows through Bering Strait into the Arctic Ocean. The exchange of water between the North Pacific Ocean and Bering Sea through the passes in the island arc is quite uncertain. Best estimates are that there is an outflow through Kamchatka Pass (21.0 Sv) (Arsen'ev, 1967) and inflow through Near Strait (14.4 Sv), with smaller inflows through Kiska, Buldir, and Semichi passes (West Aleutian group) of 0.7 Sv and inflows through the Central Aleutian group (Tanaga and Amchitka passes) of about 4.4 Sv. There is also a net inflow below 3,000 m through Kamchatka Strait. New observational results on inflow into the Bering Sea show that flow through Amukta Pass is larger than previously expected (-1 Sv). This inflow is comparable to the flow measured through the much deeper Amchitka Pass. The flow through Amukta Pass is particularly important as a source of warm (>4° compared to -3.5° C), subsurface (-200 m) water in the southeastern Bering Sea basin. This flow is intermittent, however. The balance of all of these flows includes a net outflow through Bering Strait of about 0.8 Sv, which is quite certain (Coachman and Aagaard, 1988). Thus, there is something less than 5 percent net throughflow in the Bering Sea; it is extremely important, however to global freshwater flux. Bering Sea Circulation and Hydrographic Structure Flow descriptions for Bering Sea circulation have usually been based on inferences from water properties (Figures 3.5 and 3.6), though Stabeno and Reed (1994) use drifter climatology. There is a general cyclonic flow within the basin, with an intensified boundary current, the Kamchatka Current, and a northwestward-flowing eastern boundary associated with the eastern continental slope. The eastern boundary current flow is characterized as weak and variable, typical of eastern boundary currents elsewhere (Schumacher and Reed, 1992). There are two circulation features along this slope, one being the large, persistent salinity front, and the other being the Bering Slope Current, which transports about 5 Sv of water northwest along the shelf edge at speeds in the range of 10 to 20 cm-1 s-1 (Kinder et al., 1975, 1986), although the speed in winter may be slower (Royer and Emery, 1984). The current intensifies as a western boundary current as it bifurcates and travels north around the Gulf of Anadyr and through Bering Strait, and south along the Kamchatka Peninsula (Kinder et al., 1986). The physical basis of the Bering Slope Current is poorly known. Kinder et al. (1975) have suggested that it is driven by planetary waves. There is theoretical and observational evidence that it existsâat timesâas a series of eddies, or that it includes eddies, some possibly as permanent features that are generated by topographic interactions, baroclinic instabilities, or
THE BERING SEA ECOSYSTEM: GEOLOGY, PHYSICS, CHEMISTRY, AND BIOLOGY OF LOWER TROPHIC LEVELS 38 Figure 3.5 Bering Sea circulation schemes with water masses (Takenouti and Ohtani, 1974). Figure 3.6 Schematic of flow through the passes of the Aleutian-Komandorski chains (Stabeno and Reed, 1994).
THE BERING SEA ECOSYSTEM: GEOLOGY, PHYSICS, CHEMISTRY, AND BIOLOGY OF LOWER TROPHIC LEVELS 39 Rossby waves (Karl and Carlson, 1987; Kinder et al., 1975, 1980; Paluszkiewicz and Niebauer, 1984; Royer and Emery, 1984). The eddies are apparent on some satellite altimeter data. Usually, however, it is a well- behaved eastern current. Model results (Overland et al., 1994) and observations (Stabeno and Reed, 1994) show that the Bering Slope Current separates from the shelf break at about 58.5¹ N. At most a small portion of the current continues to the head of the basin. For the Kamchatka Current, the western and southern branch of the Bering Slope Current in the western Bering Sea, eddies are also apparently a common feature of the flow regime (Verkhunov and Tkachenko, 1992). The transport of water through the Bering Strait from the Bering Slope Current is maintained by a north-south atmospheric pressure differential that tilts sea level down to the north (Coachman et al., 1975). Hydrographic structure over the eastern Bering Sea is well defined, consisting of three domains separated by physical structural fronts (Coachman, 1986; Coachman and Charnell, 1979; Kinder and Coachman, 1978; Schumacher et al., 1979). The inner front coincides approximately with the 50 m isobath, the middle front with the 80 to 100 m isobath, and the shelf break front with the 170 m isobath (Figure 3.7). The associated domains are referred to as the coastal domain, middle domain, and outer domain. Two other distinct domains exist off the shelfâa narrow, energetic shelf break domain and the oceanic domain. A mixed-layer model reproduces the general structure of the inner, middle, and outer shelves, based on wind mixing, heating, and tidal mixing (Coachman, 1986). The shelf break front is less clearly describable, but is apparent as a change in the horizontal salinity gradient. There are two water masses on the shelf: Alaska Coastal water and Central water. Coastal water is found shoreward of the 50 m isobath in the southeastern Bering Sea, whereas Central water is found in the middle domain, from the inner front to the middle front at a depth of about 100 m (see Figure 3.7). Alaska Coastal water is the result of coastal freshwater discharge combined with more saline water from the deep basin. It is generally well mixed by winds and tides. Offshore of the inner front, the Central water in the middle domain has a lower layer that is isolated from seasonal heating, producing a two-layer system. Temperatures in the lower layer reflect the prior winter conditions. Water of the outer domain is not really an identifiable water mass, but instead is a mixture of central shelf and Bering Sea basin water. It is less strongly layered because of greater tidal and advective energy (currents), but exhibits vertical fine structure that originates in the middle front and is extremely important to the flux of properties across the domain and vertically in the water column (Coachman, 1986). Circulation over the shelf is related to domain structure (Coachman, 1986). In the outer shelf, tidal currents account for about 80 percent of the flow, with the mean about 5 cm s-1 alongshore to the northwest and an onshore-offshore flow of 1 to 5 cm s-1 that is quite variable on time scales of days. Tidal mixing is very important in the outer domain. In the middle shelf, the only important flow is that due to tides and inertial currents. There is very little net motion. Vertical mixing due to tides is also important here. Tidal currents account for about 95 percent of the flow energy in the coastal zone, and there is a 1 to 5 cm s-1 mean flow following the 50 m isobath northward (Coachman, 1986). In contrast to the Gulf of Alaska circulation, the circulation of the Bering Sea shelf has relatively small net flows and relatively large tidal forcing.
THE BERING SEA ECOSYSTEM: GEOLOGY, PHYSICS, CHEMISTRY, AND BIOLOGY OF LOWER TROPHIC LEVELS 40 Figure 3.7 A cross-shelf advection model developed by the PROBES, (Processes and Resources of the Bering Sea Shelf) investigators and applied to organic matter partitioning and subsequent distributions of zooplankton and seabirds (McRoy, et al., 1986).
THE BERING SEA ECOSYSTEM: GEOLOGY, PHYSICS, CHEMISTRY, AND BIOLOGY OF LOWER TROPHIC LEVELS 41 Sea Ice An unusual physical characteristic of the Bering Sea shelf is the annual ice cover that extends more than 1,700 km at its farthest seasonal extent (Walsh and Johnson, 1979). In summer, the ice edge retreats into the Chukchi and Beaufort Seas (Figure 3.8a) (Niebauer and Schell, 1993). In winter, much of the shelf is ice- covered. The growth of ice over deep water is limited by warm water in the central basin (Figure 3.8b). Thus, the maximum extent of the ice is restricted to the shelf (Overland and Pease, 1982). The sea ice affects exchanges with the atmosphere and inhibits the transfer of freshwater (salt) and heat. It changes the coupling of the oceanic and atmospheric momentum exchanges by altering the surface roughness. The creation and melting of the sea ice alters the horizontal and vertical density gradients. Increases or decreases in the vertical density gradient affect the mixing and transport of nutrients and organisms within the euphotic zone. The ice edge also serves as both source and sink of freshwater that can affect productivity. In fall during freeze-up, freshwater is extracted from the seawater, while during the spring, melting supplies freshwater to the ice edge. Sea ice is also important in influencing bottom temperatures, and hence influences many species on the shelf. In winter, there is little stratification, and the sea is cold from top to bottom. In colder years, there is more sea ice than in warmer years. The ice helps to cause and maintain stratification when it melts, because water at the melting point of ice is lighter than slightly warmer water. After the ice has melted, the sun further causes stratification, and thus the bottom temperature changes only very slowly. Thus, in cold yearsâwith extensive sea iceâthe colder-than-normal bottom temperature is even more persistent than usual (Coachman, 1986); the colder temperatures are also attributed directly to the ice (Ohtani and Azumaya, 1995). Thus, the extent of sea ice is related to the distribution and abundance of temperature-sensitive bottom-dwelling species and some near shore species such as crabs, flatfishes, and herrings. Marine Chemistry of the North Pacific and Bering Sea Measurements of water column chemistry for the North Pacific and Bering Sea are very limited, especially in the coastal regions. Because the deep waters of the North Pacific are quite old, nutrients have accumulated in them during their passage though the deep regions of the world's oceans. An example of the nutrient distribution versus depth can be found in Geochemical Ocean Sections Study (GEOSECS) station 218 (50°27´N, 176°35´W). A layer of maximum nutrients exists at about 1,000 m, with phosphate at about 3 µmoles/kg, nitrate at about 4 µmoles/kg and dissolved oxygen at about 10 µmoles/kg. The dissolved oxygen is at a minimum at this depth, with maximum values approaching 300 µmoles/kg at the surface (Reeburgh and Kipphut, 1986). Cyclonic winds from storms cause upwelling in the deep portions of the Bering Sea and Gulf of Alaska, though the intensity is usually insufficient to bring the nutrient-rich water into the euphotic zone. Van Scoy and others (1991) have evidence from tritium measurements that on rare occasions, the deep water can penetrate the halocline and enter the upper layers.
THE BERING SEA ECOSYSTEM: GEOLOGY, PHYSICS, CHEMISTRY, AND BIOLOGY OF LOWER TROPHIC LEVELS 42 Figure 3.8a Maximum, mean, and minimum extent of ice edges for September 1 in the region of the Bering Sea and Arctic Ocean for the area centered on the Bering Strait, 1973â86 (Niebauer and Schell, 1993). Figure 3.8b Maximum, mean, and minimum extent of ice edges for March 15 in the region of the Bering Sea and Arctic Ocean for the area centered on the Bering Strait, 1973â86 (Niebauer and Schell, 1993).
THE BERING SEA ECOSYSTEM: GEOLOGY, PHYSICS, CHEMISTRY, AND BIOLOGY OF LOWER TROPHIC LEVELS 43 However, the central portions of these seas are less relatively productive the shelves and coastal regions. Coastal freshwater discharge could be a source of nutrients for near shore production, but during periods of peak runoff, the upper layers are devoid of nutrients (Reeburgh and Kipphut 1986). They and others suggest that the deep, nutrient-rich waters must be carried from the deep ocean onto the shelves, where they are mixed upward to supply the nutrients required for high productivity. Because the coastal regions are usually influenced by converging wind systems (downwelling), some other mechanism or mechanisms, such as wind mixing, tidal energy, or eddies near to the shelf break in the southeastern Bering Sea, must be invoked to achieve a vertical movement of the nutrient-rich water over the shelf. Reed and Stabeno (1990) speculate that some of the flows in the Bering Sea are chaotic; thus, there are mesoscale processes that can move the deep nutrients upward into the surface layers, most commonly along the shelf break. Sources and Magnitude of Variability There are various theories about interdecadal changes in the earth's atmosphere and oceans. Potential causes for climate changes include increases of greenhouse gases, El Niño Southern Oscillation (ENSO) events, changes in solar, volcanic, or magnetic activity, collisions with meteors, tidal fluctuations, and anthropogenic effects. Many of these possible influences are believed to be amplified at high latitudes. For example, Spelman and Manabe (1984) and Bryan and Spelman (1985) predict that there will be a ''polar amplification" in ocean temperature changes at high latitudes due to increases in greenhouse gases. Roberts and Olson (1973) found that the atmospheric pressure over the Gulf of Alaska responds to geomagnetic storms that are caused by sunspots. Royer (1989) suggests that changes in sea surface temperatures in the high-latitude North Pacific might be caused by changes in atmospheric turbidity as influenced by volcanic eruptions. Interdecadal sea surface temperature changes at high latitude are thought to be caused by lunar nodal tide forcing (Loder and Garrett, 1978; Maksimov and Smirnov, 1965; Royer, 1993). Thus, the high latitudes seem to be regions that have great potential for interdecadal variability. Unfortunately, the environmental records are usually too short to be able to identify, with confidence, the cause or causes of the observed variability, and predicting future atmospheric and oceanic climates is uncertain. Atmospheric Forcing and Ocean Circulation Atmospheric pressure systems propagate across the North Pacific from west to east, often encountering the coastal mountain range of Alaska, where they stagnate and fill. Storm activity in the Gulf of Alaska has a seasonal pattern, with the storm track found south of the Aleutian Islands in winter (October-April) (Wilson and Overland, 1986). This track moves northward into the Bering Sea in spring, ending ultimately in the Arctic Ocean in summer (May-September). It progresses southward in fall, once again entering the North Pacific in winter. Storms pass over the gulf every few days in winter (Hartmann, 1974). These storms
THE BERING SEA ECOSYSTEM: GEOLOGY, PHYSICS, CHEMISTRY, AND BIOLOGY OF LOWER TROPHIC LEVELS 44 result in nearly continuous cloud cover, winds up to 40 m s-1, and precipitation rates of more than 800 cm y-1 in the coastal mountain ranges (Wilson and Overland, 1986). The seasonal progression of the storm tracks over the North Pacific and Bering Sea could lead to a large seasonal variability in the ocean circulation, especially over the shelf and near shore. In winter, the Gulf of Alaska has strong upwelling winds over its central (deep water) region, with downwelling winds near shore. High rates of coastal precipitation accelerate the alongshore flow through buoyancy forcing, especially in fall. In summer, the influence of a high-pressure system leads to diminished winds, possibly even upwelling along the northern boundary of the Gulf of Alaska, and relatively low rates of precipitation. Thus, the coastal flow is greatest in fall and winter (Johnson et al., 1988). High sea level pressure can develop over the region in winter, creating ridges that can deflect storms to the north or south (Wilson and Overland, 1986). These ridges play an important role in year-to-year variability in regional atmospheric conditions, as will be discussed later in this report. The position of these storm tracks is critical to the coupling between ocean and atmosphere in the Bering Sea and Gulf of Alaska, which affects the deep mixing and advection from the deep water of nutrient-rich water onto the shelf of the Bering Sea, and hence the rate of biological productivity. Interannual variability in the physical parameters in the upper ocean can occur with changes in the rates of exchange with the atmosphere and through changes in ocean advection, both horizontal and vertical. Increased wind stress should increase the vertical exchanges through upwelling-downwelling processes associated with Ekman pumping, and through increased mixing and exchange rates. Increases in storm activity over the Gulf of Alaska or Bering Sea should increase the cyclonic circulation and upwelling from Ekman pumping in the central (deep) regions of the Gulf of Alaska and Bering Sea. Concurrent with an increase in upwelling in the offshore regions, there would be an increase in the downwelling in the coastal regions as a result of coastal convergences. Increased cyclonic atmospheric activity will increase precipitation rates, enhancing the baroclinic coastal flow. The exchanges through the Aleutian Islands and Bering Strait are also altered with changing atmospheric conditions. Surprisingly, evidence for seasonal changes in ocean circulation is not plentiful. The Alaska Coastal Current has an annual signal (Johnson et al., 1988; Royer, 1982), but similar measured fluctuations of the deep ocean currents in the North Pacific (Musgrave et al., 1992; Tabata, 1991) and Bering Sea are lacking. Nevertheless, a model of the Bering Sea circulation indicates that seasonal increases in circulation of the system should take place (Overland et al., 1994). In addition to the few measurements of seasonal variability, no measurements of interannual fluctuations in the deep ocean currents exist for this region. Short-term fluctuations exist in the flow, through passes into the Bering Sea (Reed, 1990), but current-meter deployments are not of sufficient duration to address seasonal and interannual variability. Fortunately, the interannual variability of the flow out of the Bering Sea through Bering Strait has been determined (Coachman and Aagaard, 1988). In addition to responding to ENSO events and lunar nodal tide forcing, the upper ocean in the Bering Sea and Gulf of Alaska might be responding to even longer period fluctuations in atmospheric forcing. Several authors (Ebbesmeyer et al., 1991; Francis and Hare, 1994; Graham, 1994; Miller et al., 1994; Trenberth and Hurrell, 1994) have suggested that an atmospheric regime shift has occurred over the North Pacific. An index of a composite of 40
THE BERING SEA ECOSYSTEM: GEOLOGY, PHYSICS, CHEMISTRY, AND BIOLOGY OF LOWER TROPHIC LEVELS 45 environmental variables made a sudden jump to a new state after 1976. Trenberth and Hurrell's analysis shows that after 1976 the atmospheric circulation of the North Pacific remained in an "abnormal" state until the late 1980s (Trenberth and Hurrell, 1994). Winter sea level pressure for the North Pacific from 1946 to 1992 illustrates the unusual conditions from about 1976 through 1988 and variations since then (Figures 3.9 and 3.10). There is supporting evidence of major interannual changes occurring in the large-scale atmospheric patterns that can cause shifts in the storm tracks. From 1950 to 1970, maximum blocking ridge activity over the North Pacific was in winter (Table 3.1). From 1980 to 1990 this maximum in activity shifted to fall. Though the number of blocking ridges remains nearly constant, there is an interdecadal shift in the marine ridge activity over the North Pacific. This should in turn influence the marine environment by altering the storm tracks over the North Pacific and Bering Sea. Evidence of the interannual changes is the altering of the upper layer structure of the Gulf of Alaska due to changes in Ekman pumping (Lagerloef, 1994); Sambrotto (1985) has also identified changes in productivity due to changes in mixing in the Bering Sea. Ocean Temperature Interdecadal changes in the sea surface temperature of the Gulf of Alaska and Bering Sea since 1947 (Figure 3.11) demonstrate a relatively warm period (>1° C variance) around 1958 in both basins. Temperatures in both ocean basins decline until about 1973, when they are about 1° C below normal. From 1973 until 1980, there was a relatively rapid increase (about 0.1° C y-1) for both basins. After 1980, the two temperature histories are more divergent, with the gulf warming until 1984 and the Bering Sea showing a general decline to 1989. Since about 1984, the two temperature anomaly variations have displayed the same temporal pattern, though the Bering Sea has remained about 0.5° C less than the Gulf of Alaska. This pattern differs from the earlier portion of the temperature anomaly record (1956â82); with few exceptions, the sea surface temperature anomalies in the Gulf of Alaska have been lower or similar to those in the Bering Sea. The lower temperature anomalies in the Bering Sea since 1982 could be caused by either the Gulf of Alaska's receiving greater quantities of warmer water, or less warm water entering the Bering Sea. In either case, the differences between the temperature anomalies continued through the end of 1993. The best correlation (0.49) between the monthly mean sea surface temperatures at these two sites was in phase, indicating no propagation of these anomalies. Additionally, there are no long-term water column measurements in the Bering Sea. The closest such time series is for the Gulf of Alaska, and so these data are used as a proxy for the Bering Sea. The low frequency sea surface temperature fluctuations in the Bering Sea follow a similar trend for those at Gulf of Alaska (GAK 1) (59°50.7' N, 149°28.0' W) near Seward, Alaska (Figure 3.12). Upper-layer interannual thermal fluctuations with similar amplitudes have been reported for other regions of the North Pacific Ocean (Lie and Endoh, 1991), and they are usually associated at lower latitudes with decadal changes in wind stress (Qui and Joyce, 1992). A tantalizing choice for the possible forcing of interannual upper ocean temperature fluctuations in the northern North Pacific is the teleconnection from tropical forcing, that is, ENSO events with periodicities of three to six years. They are often associated with interdecadal climate fluctuations in the North Pacific (Philander, 1990). Forcing of the high-
THE BERING SEA ECOSYSTEM: GEOLOGY, PHYSICS, CHEMISTRY, AND BIOLOGY OF LOWER TROPHIC LEVELS 46 Figure 3.9 Time series of mean North Pacific sea level pressures averaged over 30°-65° N, 160° E to 140° W for the months of November through March beginning in 1925 and smoothed with the low pass filter (Updated from Trenberth and Hurrell, 1994).
THE BERING SEA ECOSYSTEM: GEOLOGY, PHYSICS, CHEMISTRY, AND BIOLOGY OF LOWER TROPHIC LEVELS 47 North Pacific Index - Actual and Intervention Model Fit (1900-92) Figure 3.10 Two indicators of large-scale long-term climate variability over the North Pacific and Bering Sea in the twentieth century. Top panel shows Trenberth and Hurrell's (1994) North Pacific index (normalized) and a time series of intervention model fits similar to that of Francis and Hare (1994). Bottom panel shows time history of catch (dashed line), intervention model fit (thin solid line), and estimated interventions (thick solid line) for western Alaska (Bering Sea) sockeye salmon (Francis and Hare, 1994).
THE BERING SEA ECOSYSTEM: GEOLOGY, PHYSICS, CHEMISTRY, AND BIOLOGY OF LOWER TROPHIC LEVELS 48 Table 3.1 Decadal frequency of occurrence of marine ridges during fall and winter, 1950â89 Frequency of Occurrence Season 1950â59 1960â69 1970â79 1980â89 Fall (Sept.-Oct.) 4 1 4 10 Winter (Nov.-Mar.) 16 17 14 8 Source: Salmon (1992). Figure 3.11 Interdecadal changes in the sea surface temperature of the Gulf of Alaska and Bering Sea since 1947 (Committee on the Bering Sea Ecosystem).
THE BERING SEA ECOSYSTEM: GEOLOGY, PHYSICS, CHEMISTRY, AND BIOLOGY OF LOWER TROPHIC LEVELS 49 latitude North Pacific by ENSO events can occur through the atmosphere and ocean. The atmospheric connection is through adjustments of global circulation, that is, zonal displacement of atmospheric high- and low- pressure cells. It is manifested as changes in the Pacific North American (PNA) index teleconnection (Wallace and Gutzler, 1981). El Niño accounts for approximately 20 percent of the variance in the centers of action of the PNA pattern (Philander, 1990). A primary feature of the PNA is the Aleutian Low, and during ENSO events it deepens and moves southwestward during the North Pacific winter (Bjerknes, 1969). ENSO forcing in the ocean at high latitudes is primarily through the poleward propagation of Kelvin waves (Jacobs et al., 1994). This conclusion is supported by data of Enfield and Allen (1980), who found poleward-propagating, coastal-trapped disturbances along the west coast of North America correlated with equatorial disturbances. Not surprisingly, the temperature fluctuations at depth at GAK 1 are well correlated with ENSO events (R = -.36, >99% confidence interval) (three-to-six year intervals) (Royer, 1994). Figure 3.12 Sitka air temperature anomalies (solid curve), GAK 1 sea surface temperature anomalies (dotted curve), and GAK 1 temperature at 250m. (GAK 1 is located at approximately 60° N, 149° W in 263 m of water at the mouth of Resurrection Bay; Sitka air temperature anomalies are smoothed by using a fifth-degree Butterworth filter with a five-year cutoff.) (After Royer, 1993). In addition to the fluctuations associated with ENSO forcing, however, the water column temperature variations at GAK 1 have been found to be associated with the lunar nodal tide
THE BERING SEA ECOSYSTEM: GEOLOGY, PHYSICS, CHEMISTRY, AND BIOLOGY OF LOWER TROPHIC LEVELS 50 signal (R = -.6, >99% confidence interval) (18.63-year period) (Royer, 1994). The temperature variability at 250 m at GAK 1 associated with the nodal tide is about three times that of ENSO. This lunar nodal tide component is the twelfth-largest tidal component and is related to the changes in lunar declination over the 18.6-year periodicity. Equilibrium tide theory predicts that this tidal component will vary with latitude, with larger amplitude at high latitudes than at mid-latitudes (Parker et al., 1994). Because the interdecadal sea surface temperature variability seems to occur simultaneously in the Gulf of Alaska and Bering Sea (see Figure 3.11), it is expected that the lunar nodal tide forces Bering Sea parameters in a similar fashion as it does in the Gulf of Alaska. Again, the temperature anomaly patterns are similar, with no phase shift, suggesting simultaneous forcing. Using air temperatures at Sitka, Alaska, as a proxy for the Gulf of Alaska and Bering Sea surface temperatures, (R = 0.21, N = 564), the data record can be extended back to 1828 (Figure 3.13). Semiregular variations of about 1° C in the air temperatures have occurred at Sitka for more than 150 years. Note that the decade of the 1970s was the coolest in the entire record, suggesting that perhaps water temperatures at that time were unusual. Changes in atmospheric circulation can cause changes in the exchange of water between the North Pacific and Bering Sea. This has been demonstrated on a seasonal basis in numerical models (Overland et al., 1994) (Figure 3.14), but it has yet to be directly measured, although, Reed (1990) has reported fluctuations in the exchanges through Near Strait. Generally, seasonal ocean circulation responses to wind forcing are surprisingly small. For example, although wind energy increases considerably in winter over the eastern Bering Sea, Schumacher and Reed (1992) found that wind forcing accounted for only a small part of the current fluctuations. Seasonal changes in the Alaska Stream represent at most 10 percent of mean flow (Royer, 1993); some studies question whether the seasonal signal even exists (Musgrave et al., 1992). Tabata (1991) found the ocean's baroclinic transport response (along Line P, at about 50° N) to seasonal forcing to be insignificant. Hollowed and Wooster (1992) have suggested that the intensity of the Alaskan Gyre can change on an interannual basis in response to changes in atmospheric forcing, but there is no direct evidence that large-scale circulation reacts to seasonal changes in atmospheric forcing. Thus, seasonal responses to wind forcing remain uncertain. The heat, salt, and nutrient supply for the Bering Sea depends on the exchange of water between the Bering Sea, North Pacific, and Arctic Ocean. Coachman and Aagaard (1988) found that the flow through Bering Strait (at a mean 0.8 Sv) has considerable variability. The variability from 1946 to 1969 was such that 70 percent of the transports into the Arctic Ocean through Bering Strait were greater than the mean, whereas from 1969 to 1985 76 percent of the transports were less than the mean. Smallest transports occurred in 1974, 1976, and 1980 representing three of the four lowest transports in this century. Bering Strait transports apparently depend on large-scale atmospheric pressure patterns, in contrast to deep Bering Sea circulation. Significant interannual changes in local ocean circulation might be caused by mesoscale eddies that themselves cause significant larger-scale changes in circulation. In a numerical model of the exchange of water between the North Pacific and Bering Sea, Overland et al. (1994) found that the Alaska Stream west of 175° E became unstable, with considerable meandering as it separates from the Aleutian Islands. As the flow becomes unstable, mesoscale
THE BERING SEA ECOSYSTEM: GEOLOGY, PHYSICS, CHEMISTRY, AND BIOLOGY OF LOWER TROPHIC LEVELS 51 Figure 3.13 Sitka air temperature anomalies with least squares fit of 18.6-year lunar nodal tide cycle (smooth curve, with anomalies smoothed by using a fifth-degree Butterworth filter with a five-year cutoff) (after Royer, 1993). Figure 3.14 Variability in the Bering Sea as determined from eddy and mean kinetic energies (Overland et al., 1994).
THE BERING SEA ECOSYSTEM: GEOLOGY, PHYSICS, CHEMISTRY, AND BIOLOGY OF LOWER TROPHIC LEVELS 52 eddies form, as detected in satellite altimetry measurements (Okkonen, 1993). These eddies could block the flow of this relatively warm, high-salinity North Pacific water into the Bering (Reed et al, 1993; Stabeno and Reed, 1992). The length of time of the change in this flow is on the order of months. Mesoscale eddies could also play an important role in the vertical advection of nutrient-rich water into the euphotic zone. More recently, Jacobs et al. (1994) have observed poleward-propagating Kelvin waves at the shelf break that extend from the equator into the Bering Sea. These waves excite Rossby waves, which propagate westward very slowly at speeds of a few centimeters per second. Their speeds are inversely related to latitude, and it takes up to a decade for the disturbances to travel across the mid-Pacific, more than 5,000 km. Although the zonal expanse of the Bering Sea basin is less than that of the entire mid-latitude Pacific, the Rossby phase speeds are also less, so that interannual time scales are likely. These phenomena contributes to the interdecadal temperature fluctuations. These same Rossby waves that are series of mesoscale eddies enhance the vertical motion that can bring nutrient-rich water into the surface layers, as mentioned in the previous paragraph. Ice There is considerable interannual variability in the ice cover, with values for 1976 being 10 to 15 percent below normal and those for 1979 being 10 to 15 percent above normal (Niebauer and Day, 1989). Large variability in the maximum extent of the ice edge occurred between 1974 and 1976, when the edge was at 56°â 56.5° N, and 1978 and 1979, when it was at 58°â58.75° N (T. Wyllie-Echeverria, personal communication). Ohtani and Azumaya (1995) showed how summer bottom temperatures are affected by the existence of ice cover in winter. This interannual variability is attributed to the extent of winter northerly winds, which in turn can be related to winter atmospheric circulation patterns referred to above (perhaps related to ENSO events). Manak and Mysak (1987) have associated changes in the Bering Sea ice cover with atmospheric circulation changes caused by changes in heat and moisture advection. Summary Interannual variability in the North Pacific and Bering Sea exists on a variety of time scales and has multiple causes (Table 3.2). Forcing by seasonal and shorter time scale events (tides and storms) is well documented, though the ocean responses are less well known. Of interannual periodicities, the best known are ENSO events that radiate from the equatorial region at irregular intervals, although intervals most commonly range from three to seven years. This accounts for approximately one-third of the ice and sea surface temperature variability in the Bering Sea (Niebauer and Day, 1989). An interannual periodicity of six to seven years has been found along Line P (about 50° N), but whether this holds north of the line is questionable (Tabata, 1991). Sunspots with periodicities of 11 and 22 years (Hale cycle) have been implicated in upper atmospheric variability over the Gulf of Alaska (Roberts and Olson, 1973), but there is no evidence of their influence on ocean processes. Interdecadal variability of ocean
THE BERING SEA ECOSYSTEM: GEOLOGY, PHYSICS, CHEMISTRY, AND BIOLOGY OF LOWER TROPHIC LEVELS 53 temperatures in the Gulf of Alaska has been associated with the 18.6-year lunar nodal tide, which accounts for more than one-third of the variability throughout the water column at the coast (Parker et al., 1994; Royer, 1993). Over this tidal cycle, the relative importance of diurnal and semidiurnal tidal components is altered. Influences from this effect in the Bering Sea are undetermined, but responses like those in the Gulf of Alaska are likely. The role of mesoscale eddies in the ocean is not determined, but they may play an important part in controlling the exchange of water between the North Pacific and Bering Sea (Okkonen, 1993). They might also have an important part in mixing water masses and providing microclimates for enhanced productivity. Changes in global atmospheric pressure patterns can create persistent blocking ridges over the North Pacific, leading to decadal and longer shifts in pressure patterns and storm tracks. These changes might be considered "regime shifts", which alter the timing of critical storm events that can affect biological productivity; they are discussed further below. Table 3.2 Atmosphere-ocean variability time scales and forcing mechanisms Period Forcing Mechanism Diurnal/semidiurnal Tides 3â10 days Storms Seasonal Solar declination Interannual (years) 1+ Mesoscale ocean eddies 3â7 ENSO events 6â7 Mid-latitude atmospheric events 10+ ''Regime shift" 11 Sunspots 18.6 Lunar declination 22 Sunspots Source: Committee on the Bering Sea Ecosystem The interactions of several or all of the potential long-period forces acting over and in the Bering Sea should be considered. They might have nonlinear or even chaotic interactions (Parker et al., 1994). Some combination of annual, 3-to-7 year, 6-to-7 year, 11-year, 18.6-year, and 22-year cycles could lead to the atmospheric and oceanic variability that takes place at these latitudes. Interannual changes of the flow into the Bering Sea and subsequent outflows could influence the biota in the Bering Sea, Arctic Ocean, and North Pacific. The transfer of nutrients, salt, and heat are unquestionably important to biological productivity. Upwelling and
THE BERING SEA ECOSYSTEM: GEOLOGY, PHYSICS, CHEMISTRY, AND BIOLOGY OF LOWER TROPHIC LEVELS 54 mixing caused by winds will have an interannual response, as suggested by Sambrotto (1985). For example, Brodeur and Ware (1992) show a doubling of zooplankton biomass, pelagic fish, and squid abundance in the northeast Pacific from the periods of 1956 to 1962, and 1980 to 1989. Venrick and others (1987) have observed changes in the epipelagic ecosystem in the North Pacific. Total chlorophyll and phytoplankton increased along with calanoid copepods after 1976, according to McFarlane and Beamish (1992). They attribute the improved primary production in the central North Pacific to increased nutrient supply associated with a more intense low- pressure system in winter (see Figure 3.9). The concept of "regime shifts" in this region has been suggested by Francis and Hare (1994). The atmospheric forcing for such regime shifts has been noted by Trenberth and Hurrell (1994), Salmon (1992) (see Table 3.1), and Schwing (1994). The change in 1976 is just one of many that have occurred over the North Pacific (Namias et al., 1988), although its influence might have been enhanced by an abnormally cold period in the early 1970s and it seems to be important in terms of the current organization of the Bering Sea ecosystem, as described in Chapter 6. Although atmospheric circulation appeared not to have changed back to its previous condition by 1989 (Trenberth and Hurrell, 1994), similar changes are likely to occur in the future. PRIMARY AND SECONDARY PRODUCTION Water Column Primary Production As discussed earlier in this chapter, the Bering Sea is distinguished by a very broad continental shelfâ approximately half of the geographic area of the Bering Sea is shelf, with an average water depth of only 50 to 75 m. The remainder is an oceanic basin 3,000 to 4,000 m deep. Continental shelves are in general more productive than most other regions of the ocean, because tidal and wind energy can occasionally interact with sufficient strength to resupply nutrients to the euphotic zone from reservoirs beneath the thermocline (Walsh, 1988). The wide shelf in the Bering Sea provides an expansive area over which such physical processes can occur. Beyond this general setting, other processes lead to important mesoscale variability in patterns of primary production in the Bering Sea and very high production in several regions. These include the flow of Alaska Stream water from the Gulf of Alaska into the Bering Sea through passes in the Aleutian Island arc, upwelling at the edge of the continental shelf and along the Aleutian Island arc, and currents at the shelf edge and across the northern shelf. Each process supplies nutrients to the euphotic zone during summer, and thus extends the growing season well beyond the period of the spring bloom characteristic of this and other high-latitude seas. The sections that follow highlight information on phytoplankton and zooplankton, including the predominant species known from the Bering Sea. Unfortunately, detailed studies on the distribution, dominance, and fluctuations of these organisms are few.
THE BERING SEA ECOSYSTEM: GEOLOGY, PHYSICS, CHEMISTRY, AND BIOLOGY OF LOWER TROPHIC LEVELS 55 Continental Shelf Production everywhere in the Bering Sea is limited by both nutrients and light. From fall to spring, light levels are sufficiently low that even in the presence of abundant nutrients phytoplankton cannot grow. Over the shelf, seasonal sea ice extends the period of light limitation in spring, particularly in cold years when sea ice extends all the way to the shelf edge and retreats slowly. During winter, a combination of wind mixing and cross- shelf diffusion and tidal mixing resupply nutrients to the depleted waters on the shelf from the reservoir in the basin of the Bering Sea (Coachman and Walsh, 1981; Whitledge et al., 1986). As ice melts in spring, the freshwater floating on top of the denser seawater establishes a strongly two- layered water column with a sharp but shallow halocline that retards vertical mixing. In the presence of sufficient nutrients, the plankton cells bloom (McRoy and Goering, 1974; Niebauer and Alexander, 1985). The ice-edge bloom follows the retreating sea ice north, and can be short-lived if the upper stable layer is shallow, i.e., low initial nutrient inventory, or if winds break down the density stratification caused by meltwater from the ice on the seawater. Under these conditions, the ice-edge bloom contributes only a small amount, perhaps 10 percent, to the total annual production budget in the Bering Sea (McRoy and Goering, 1974). In years when wind mixing is low, the density stratification of the ice-edge bloom might be subsequently reinforced by surface heating and the ice-edge bloom could become in essence the spring bloom over a portion of the shelf (J. Niebauer, unpublished data). After the retreat of sea ice and the termination of the ice-edge bloom, the physical zonation of the shelf plays a fundamental role in nearly all aspects of the production regimes of the different domains (Iverson et al., 1979a; Walsh and McRoy, 1986). The fronts retard across-shelf advective and diffusive processes that would otherwise distribute nutrients from the basin reservoir over the shelf. This, together with the breadth of the shelf, greatly slows the rate of nutrient renewal in the coastal domain, and the water column rapidly becomes depleted of nutrients in spring during the bloom. Because of slow renewal and a shallow water column, and hence small initial nutrient inventory, total production in the coastal domain is in the range of only 50 to 100 g C m-2 y-1, of which less than 20 g C m-2 y-1 is new production (Figure 3.15) (Hansell et al., 1993; Springer and McRoy, 1993). Production in the middle domain is greater than in the coastal domain, because it is nearer the basin, the water column is deeper, and there is a larger initial nutrient inventory. The spring bloom is commonly initiated by thermal stratification of the water column after sea ice dissipates, which again reduces wind-mixing depth and allows plankton cells to remain in favorable light conditions (Sambrotto et al., 1986). Production is probably in the range of 150 to 200 g C m-2 y-1, of which 30 to 50 g C m-2 y-1 is new production (Hansell et al., 1993). Walsh and McRoy (1986) considered their estimate of 166 g C m-2 y-1 to be conservative, because measurements of production had never been taken throughout the entire growing season. Production in the vicinity of the Pribilof Islands and St. Matthew Island, which lie in the middle domain, might be somewhat higher than average for the domain because of episodic upwelling during summer. Production in the outer domain was estimated to be similar to that in the middle domain, about 160 g C m-2 y -1, which is also considered to be conservative (Walsh and McRoy, 1986). Vertical fine structure in the water column of the outer domain is a feature throughout summer,
THE BERING SEA ECOSYSTEM: GEOLOGY, PHYSICS, CHEMISTRY, AND BIOLOGY OF LOWER TROPHIC LEVELS 56 and should lead to higher supply rates of nutrients depleted by phytoplankton growth (Whitledge et al., 1986). New production was estimated to fall in the range of 40 to 70 g C m-2 y-1 (Hansell et al., 1993), and total production is probably between 200 to 250 g C m-2 y-1. Figure 3.15 Generalized pattern of primary production in the Bering Sea and northern Gulf of Alaska (Springer and McRoy, in review). Estimates of total production are commonly based on nitrate-nitrogen uptake rates (new production) that are converted by using the f-ratio, or ratio between new production and total production (Epley and Peterson, 1979). Sambrotto and others (1993) have recently demonstrated that this approach typically underestimates total production by as much as about 50 percent. Current estimates of production over the shelf and basin of the Bering Sea therefore suffer from the combined effects of short-term measurements, that is, measurements that do not span the entire growing season, as well as conservative methods for converting nitrogen uptake rates to carbon assimilation.
THE BERING SEA ECOSYSTEM: GEOLOGY, PHYSICS, CHEMISTRY, AND BIOLOGY OF LOWER TROPHIC LEVELS 57 Basin Few estimates have been made of primary production in the Bering Sea basin. New production is apparently only about 10 g C m-2 y-1 (Hansell et al., 1993), and total production is probably similar to the inner domain, or 50 to 100 g C m-2 y-1 (e.g., Saino et al., 1979; Sapozhnikov et al., 1993; Taniguchi, 1969). Shelf Break The seaward side of the outer shelf domain lies approximately at the shelf edge. It is a comparatively narrow region with distinct physical characteristics and, apparently, a bountiful production regime. The shelf edge domain, or "green belt," apparently extends around the entire perimeter of the Bering Sea continental shelf and across the northern shelf, through Bering Strait, and into the Chukchi Sea. The eddies along the shelf edge are energetic and might be important to mesoscale nutrient dynamics in the upper water column, and thus to production budgets (e.g., Karl and Carlson, 1987). The interaction of strong tidal currents with the abrupt, steep shelf break promotes upwelling at the front (Coachman, 1986), which also helps supply nutrients to the euphotic zone. As a result, primary production apparently remains elevated throughout summer, long after the termination of the spring bloom (Iverson et al., 1979b; Karl and Carlson, 1987). Up to 110 g C m-2 y-1 of new production was estimated near the shelf break (Hansell et al., 1993), compared to the estimates of less than 10, 40 to 70, 30 to 50, and more than 20 g C m-2 y-1 of new production over the basin and outer, middle, and inner shelf domains, respectively. Assuming an f-ratio of 0.3, and remembering the caution noted above about f-ratios, 110 g C m-2 y-1 of new production would convert to approximately 365 g C m-2 y-1 of total production in the green belt. Deposition of phytodetritus from in situ production in this zone might account for high sediment carbon on the slope (Lisitsyn, 1966) and contribute to off-shelf transport of carbon from the outer domain as has been proposed (Walsh, 1983; Walsh et al., 1981). In the northern branch of the green belt, new production in the Anadyr Stream across the Bering-Chukchi shelf is estimated at up to 288 g C m-2 y-1, or 700 to 800 g C m-2 y-1 of total primary production (Hansell et al., 1993; Springer and McRoy, 1993). These values for production on the northern shelf and along the shelf edge in the eastern Bering Sea agree well with estimates of production at the shelf edge along the Kamchatka Peninsula, the western extension of the green belt, which fall in the range of 300 to 800 g C m-2 y-1 (Sapozhnikov et al., 1993). Aleutian Island Arc The Alaska Coastal Current, a baroclinic coastal current flowing to the southwest over the shelf of the western Gulf of Alaska, and the Alaska Stream are both important to property distributions across the arc (Favorite, 1974; Reed, 1990; Schumacher et al., 1982) and apparently to local production budgets (Koike et al., 1982). In addition, upwelling at some
THE BERING SEA ECOSYSTEM: GEOLOGY, PHYSICS, CHEMISTRY, AND BIOLOGY OF LOWER TROPHIC LEVELS 58 locations further contributes to mesoscale nutrient and production budgets (Kelley et al., 1971; Park et al., 1973; Swift and Aagaard, 1976). Depending on spatial variability in the distribution and effects of physical processes, production probably ranges between about 200 and 300 g C m-2 y-1 across the arc (McRoy et al., 1972; Sanger, 1972) and along the northern side of the Alaska Peninsula (Schell and Saupe, 1989). Water Column Secondary Production The fronts dividing the continental shelf into domains affect patterns of secondary production also. The distribution of zooplankton communities that results from the physical partitioning of the shelf determines the fate of carbon produced in the water columnâthat is, the way carbon is divided between pelagic and benthic food webs. Four general groupings of zooplankton have been described for the Bering Sea: oceanic and outer shelf, mixed oceanic-shelf, middle shelf and coastal, and near shore (Cooney, 1981). Among the numerous species of herbivorous zooplankton in the Bering Sea, large calanoid copepods are a particularly important link in the food web between phytoplankton and higher trophic levels. Neocalanus cristatus, N. plumchrus, Eucalanus bungii, and Metridia pacifica are oceanic species found in deep water across the North Pacific. They are large, efficient grazers of large diatoms typical of the spring bloom flora, and are in turn important prey of most species of planktivorous fishes, birds, and mammals. They do not occur over the shelf inshore of the middle front, that is, in the middle or coastal domains, however, because of the effect the fronts have in restricting cross-shelf advective and diffusive exchange. There they are replaced by Calanus marshallae, the only large copepod found higher on the shelf, and by various species of small copepods. Zooplankton biomass changes correspondingly, with a much larger biomass over the outer domain and decreasing biomass over the middle and coastal domains. Likewise, grazing stress is greater in the outer domain than in the middle or coastal domains, resulting in a much larger pelagic shunt of primary production in the outer domain (Cooney, 1981; Cooney and Coyle, 1982). In the middle and coastal domains, a considerably larger proportion of the annual primary production settles to the bottom, where it fuels important benthic food webs. In all domains, the ice-edge production is ungrazed, because it precedes the development of zooplankton populations in spring (Coyle and Cooney, 1988). The greatest biomass of zooplankton apparently occurs in the green belt from the southeastern Bering Sea into the Chukchi Sea (Cooney, 1981; Springer et al., 1989). Zooplankton biomass is probably high along the coast of the Kamchatka Peninsula as well. Enhanced primary production likely accounts for the high zooplankton biomass along the shelf edge, while advection in the Anadyr Stream is the major factor across the northern shelf (Springer and Roseneau, 1985; Springer et al., 1989). Estimates of annual zooplankton production in the southeastern Bering Sea are as high as 64 g C m-2 y-1 for the shelf edge, 40 g C m-2 y-1 (range 16 to 50 g C m-2 y-1) for the outer shelf domain, 19 g C m-2 y-1 (range 8 to 30 g C m-2 y-1) for the middle shelf domain, and just 4 g C m-2 y-1 (range, 2 to 6 g C m-2 y-1) for the coastal domain (Cooney, 1981; Vidal and Smith, 1986).
THE BERING SEA ECOSYSTEM: GEOLOGY, PHYSICS, CHEMISTRY, AND BIOLOGY OF LOWER TROPHIC LEVELS 59 Sources and Magnitude of Variability Primary production in the outer and middle domains apparently can vary by as much as 30 to 50 percent between years, depending on the frequency and intensity of summer storms that mix the water beneath the seasonal pycnocline and resupply nutrients to the euphotic zone (Sambrotto and Goering, 1983; Sambrotto et al., 1986). Production in the coastal domain likely varies little between years because the water column is sufficiently shallow and well mixed that nutrients are depleted during the bloom in all years. Little is known about variability in the physical system or productivity at the shelf edge. Across the northern shelf and in the Alaska Stream, two highly advective regions, variability in transport has been linked to variability in biomass of phytoplankton and zooplankton. Annual fluctuations in transport in the Anadyr Stream across the northern shelf of the Bering Sea (Coachman, 1993) can lead to differences in primary productionâproduction appears to be proportional to transport (Springer and McRoy, 1993). The fact that Anadyr Water derives its physical and chemical characteristics (i.e., high nutrients and low oxygen) from the subsurface Bering Sea suggests that variability in transport of the Bering Slope Current (of which the Anadyr Stream is the northern component) is related to large- scale variability in the circulation of the Bering Sea. This variability might be forced by basin-scale changes in the Aleutian Low. Intensified Aleutian Low wind stress curl during 1976 to 1988 could have resulted in increased northward Sverdrup transport in the central and eastern Bering Sea, producing increased southward flow in the western boundary current circulation of the subarctic Pacific (Salmon, 1992). This finding is also consistent with the results of Sekine (1988, 1991), who showed that the Oyashio Current intrudes farther south than normal during years of intensified Aleutian Low wind forcing. However, the possible link between changes in Bering Sea circulation and variations in large-scale wind forcing are tenuous. The Bering Slope Current system should respond to basin scale changes in wind forcing over the Bering Sea, although direct evidence of this is not available. Similarly, it might respond indirectly to changes in the Gulf of Alaska through the effects of wind-forced variations in the flow of the Alaska Stream through the Aleutian Island arc. An analysis of the flow regime along the shelf break using satellite altimeter data revealed coherence between the Bering Slope Current and the Kamchatka current and significant interannual variability in the frequency and intensity of eddies (Fox and Leben, 1993). Eddies were particularly pronounced in 1986 compared to 1985, 1987, and 1988, a pattern consistent with the anomalous circulation and production on the northern shelf observed in 1986 compared to the other years. Interannual fluctuations in the Alaska Stream are related to transport in zooplankton biomass in the western Bering Sea (Wickett, 1967). On the order of 15 to 20 Sv of water from the Alaska Stream flood through the Aleutian Island Arc into the Bering Sea (Favorite, 1974; Reed, 1990), apparently carrying a great amount of zooplankton. This notion seems plausible in light of evidence that zooplankton are carried to the shelf from within the Alaska Gyre through Ekman transport generated by the Aleutian Low (Cooney, 1986), and that interannual fluctuations in zooplankton abundance on the northern Gulf of Alaska shelf are related to the strength of the Aleutian Low wind field (Brodeur and Ware, 1992; Francis and Hare, 1994). In particular, during years when the Aleutian Low is strong, the shelf of the northern Gulf of Alaska is characterized by high zooplankton abundance due to increased onshore Ekman
THE BERING SEA ECOSYSTEM: GEOLOGY, PHYSICS, CHEMISTRY, AND BIOLOGY OF LOWER TROPHIC LEVELS 60 transport, while a weak Aleutian Low corresponds to decreased zooplankton abundance on the shelf in conjunction with weakened onshore transport. Tabata (1991) has shown that oceanographic properties such as temperature, salinity, and dissolved oxygen in the Gulf of Alaska undergo oscillations in periods of two to three and six to seven years (see Table 3.2), while Salmon (1992) has suggested that changes in the strength of transport in the Alaska Gyre might be closely related to changes in wind forcing within the Aleutian Low on these same time scales. Thus, variations in the strength and position of the Aleutian Low might lead to variations in circulation and biomass in the Alaska Stream and account for the observations of Wickett (1967). Venrick and others (1987) found a significant change in water column chlorophyll concentrations north of Hawaii at the time of the regime shift in the mid-1970s (Figure 3.16). Most of the change occurred below 100 m and was hypothesized to be due to an increase in the winter production of deep phytoplankton species caused by a deepening of the winter mixed layer. Likewise, Brodeur and Ware (1992) found a significant shift in zooplankton biomass in the Alaska Gyre, which also corresponded to the recent regime shift (Figure 3.17). Similar oceanographic time series, both physical and biological, are lacking in literature for the western Bering Sea. Decadal-scale changes in the mean position of the Aleutian Low pressure field and resulting wind fields over the Bering Sea might produce patterns in annual primary and secondary production, with stormy eras characterized by higher production than calmer eras, as the data of Sambrotto and Goering (1983) suggest. Elsewhere, long-term patterns of phytoplankton and zooplankton production do track changes in the mean physical conditions of the marine environment (e.g., Aebischer et al., 1990). BIOLOGY OF LOWER TROPHIC LEVELS Benthic Production The benthic communities of the Bering Sea are dominated by a large variety of polychaetes, amphipods, gastropods, and bivalves. They vary tremendously in space due to food supply, bottom depth, disturbance (e.g., ice gouging and marine mammal feeding), sediment type, and predation. The distribution, abundance, and biomass of benthic fauna coincide with physical domains (Feder et al., 1985; Grebmeier et al., 1988; Haflinger, 1981; Stoker, 1981). For example, the middle shelf of the Bering Sea south of St. Lawrence Island is dominated by bivalves and polychaetes (Grebmeier and Cooper, 1995; Haflinger, 1981; Stoker, 1981), the region north of St. Lawrence Island to Bering Strait is dominated by ampeliscid amphipods (Grebmeier et al., 1989; Highsmith and Coyle, 1992). In comparison, the continental slope areas in both Bering and Chukchi seas are dominated by a variety of polychaetes (Feder et al., 1994; Grebmeier, 1993, unpublished data; Stoker, 1981; Zenkevitch, 1963). The highest benthic biomass is attained in the middle shelf of the southeastern Bering Sea and in the northern Bering and southern Chukchi seas as a result of the large amount of food delivered in a shallow water column over the five-to-seven month ice-free period (Grebmeier, 1993; Grebmeier and McRoy, 1989; Haflinger, 1981). In the southeastern middle shelf, carbon flux to the benthos is
THE BERING SEA ECOSYSTEM: GEOLOGY, PHYSICS, CHEMISTRY, AND BIOLOGY OF LOWER TROPHIC LEVELS 61 Figure 3.16 Index of phytoplankton production in the central North Pacific (observations of integrated chlorophyll a in the CNP). Bars indicate the 95 percent confidence intervals of the mean x = r :n; the number of observations is shown above each bar. Winter values (open squares) and values before 1968 are excluded from our analyses. Definitions: t, 97.5 percentile of t statistic with n - 1 df; , the variance of the observations (Venrick et al., 1987). Figure 3.17 Relationship between wind stress (Ekman transport) at 60° N, 149° W in the northern Gulf of Alaska and yearly zooplankton biomass values for the periods 1956â62 (circles) and 1980â89 (squares) (Brodeur and Ware, 1992).
THE BERING SEA ECOSYSTEM: GEOLOGY, PHYSICS, CHEMISTRY, AND BIOLOGY OF LOWER TROPHIC LEVELS 62 high, in spite of only moderate primary production, because of inefficient grazing by the herbivorous zooplankton community. In the northern Bering and southern Chukchi, benthic carbon supply results from prodigious primary production that far surpasses the ability of the zooplankton to control. The latitudinal gradient in benthic biomass mirrors the high production on the northern shelf (Figure 3.18). Benthic infaunal biomass in the southeastern Bering Sea averages 1 to 10 g C m-2 in the inner domain, 1 to 20 g C m-2 in the middle domain, and 1 to 11 g C m-2 in the outer domain (Table 3.3; Figure 3.19; Feder et al., 1980; Haflinger, 1981; Stoker, 1981). It is lowest along the Alaska coastline in the northern Bering and Chukchi seas, where it is limited by ice gouging, river runoff, high current scouring, and low overlying water production (Figure 3.20; Grebmeier et al., 1988, 1989; Stoker 1981). Benthic biomass increases to the west to 20 to 40 g C m-2 southwest of St. Lawrence Island and in regions of the Gulf of Anadyr (Table 3.3; Figure 3.21; Sirenko and Koltun, 1992; Grebmeier, 1993; Grebmeier and Cooper, 1995). The infaunal community structure is generally dominated by bivalves and polychaetes (Grebmeier, 1993; Stoker, 1981), although it becomes dominated by amphipods in the Chirikov Basin, north of St. Lawrence Island (Table 3.3), with biomass ranging from 10 to 30 g C m-2 (Figure 3.20; Feder et al., 1985; Grebmeier et al., 1988, 1989; Highsmith and Coyle, 1992; Stoker, 1981). Infaunal biomass peaks at 50 to 60 g C m-2 in the southern Chukchi Sea, where a downstream deposition of organic materials occurs (Grebmeier and McRoy, 1989; Grebmeier, 1993). Epifaunal benthos on the eastern Bering Sea shelf (Figure 3.19) is numerically dominated by bivalves, then anthropods and echinoderms; however, nearly 70 percent of the total epibenthic biomass represents sea stars (Jewett and Feder, 1981). The most important epifauna on the southeastern shelf include four species of commercially harvested crabs: king crabs Paralithodes camtschatica (red) and P. platypus (blue), and the Tanner (snow) crabs Chionoecetes opilio and C. bairdi (Figure 3.22). In comparison, sea stars (Asterias amurensis, Evasterias echinosoma, Leptasterias polaris acervata, and L. nanimensis) dominate in the northeast Bering Sea. Limited studies of fauna of the shelf break and deep basin in the eastern portions of the Bering Sea indicate the benthic fauna is dominated by polychaetes of low biomass, about 78 g wet wt m-2, or less than 5 g C m-2 (Grebmeier, 1993; Sirenko and Koltun, 1992). Benthic populations in the basin are limited by depth and the reduced carbon supply to the bottom. Few studies are available in western literature related to the western Bering Sea, but it is reasonable that the high primary production reported for this region, similar to the upper estimate for production in the northern branch of the green belt, would directly influence benthic faunal populations and sediment processes on underlying shelf systems that it traverses. Petersen and Curtis (1980) hypothesized that polar regions exhibits a stronger benthic-pelagic coupling than temperate and tropical areas, and a number of field studies have confirmed this hypothesis (e.g., Grebmeier, 1993; Grebmeier and McRoy, 1989; Stoker, 1981). Regions of high overlying water column production in the Bering and Chukchi seas have a direct influence on underlying benthic biomass (Grebmeier, 1993; Grebmeier et al., 1988; Rowe and Phoel, 1991). Additional studies support these patterns of direct coupling of water column production and underlying benthos in the southeast Bering Sea (Rowe and Phoel, 1991), and on the continental shelf of the Barents Sea, where sediment bacterial growth rates were higher where carbon deposition was greater than in higher-latitude regions (Pfannkuche and Thiel,
THE BERING SEA ECOSYSTEM: GEOLOGY, PHYSICS, CHEMISTRY, AND BIOLOGY OF LOWER TROPHIC LEVELS 63 Figure 3.18 Variation of benthic biomass with latitude on the Bering and Chukchi sea shelves (vertical lines indicate standard deviation and brackets indicate the coefficient of variation around the mean) (Stoker, 1981.)
THE BERING SEA ECOSYSTEM: GEOLOGY, PHYSICS, CHEMISTRY, AND BIOLOGY OF LOWER TROPHIC LEVELS 64 Table 3.3 Comparison of phytoplankton and zooplankton production, benthic biomass, and sediment oxygen uptake rates for different regions on the continental shelves of Bering and Chukchi seas Area Phytoplankton Primary Zooplankton Benthic Biomass Sediment Oxygen Production (g C m-2 y-1) Secondary (g C m-2) Uptake Rates Production (g C m-2 (mmol O2 m-2 d-1) y-1) Southeastern Bering Sea Inner domain 50â80 (1) - 10 (2) 3â8(3) Middle domain 166 (1) 8â30 (4, 5) 10â20 (2,6) 3â8 (3) Outer domain 162 (1) 30â50 (4, 5) 10 (2,6) 5â10 (3) Oceanic domain 50 (7) 20 (4) <5 (2) - Northern Bering Sea Alaska Coastal Water 80 (8) 5 (9) <10 (11, 12) <10 (10, 11) Bering Shelf Water/ 80â480 (8, 13) 9 (9) 10â30(10, 11) 10â30 (11, 12) Anadyr Water Southern Chukchi Sea Alaska Coastal Water 80 (8) - <10 (11, 12) <10(11, 12) Bering Shelf Water/ 470â720 (8, 14) - 10â60 (10, 11) 10â40 (11, 12) Anadyr Water Northern Chukchi Sea ''Southern group" 50â100 (15, 16) - 1â11 (17) 5â10 (11, 18) "Northern group" 50â100 (15, 16) - 2â20 (17) 5â10 (18) Source: Modified from Grebmeier et al. (1995).
THE BERING SEA ECOSYSTEM: GEOLOGY, PHYSICS, CHEMISTRY, AND BIOLOGY OF LOWER TROPHIC LEVELS 65 Figure 3.19 Total infaunal and epifaunal biomass in the southeastern Bering Sea (Haflinger, 1981).
THE BERING SEA ECOSYSTEM: GEOLOGY, PHYSICS, CHEMISTRY, AND BIOLOGY OF LOWER TROPHIC LEVELS 66 Figure 3.20 Distribution of macrofaunal benthic biomass on the shelves of the northern Bering and Chukchi seas (Grebmeier, 1993).
THE BERING SEA ECOSYSTEM: GEOLOGY, PHYSICS, CHEMISTRY, AND BIOLOGY OF LOWER TROPHIC LEVELS 67 Figure 3.21 Distribution of macrofaunal biomass on the northern shelf of the Bering Sea, June 1990 (Grebmeier and Cooper, 1995).
THE BERING SEA ECOSYSTEM: GEOLOGY, PHYSICS, CHEMISTRY, AND BIOLOGY OF LOWER TROPHIC LEVELS 68 Figure 3.22 Distribution of king and Tanner crabs in the eastern Bering Sea (vertical lines indicate areas of high abundance) (Otto, 1981).
THE BERING SEA ECOSYSTEM: GEOLOGY, PHYSICS, CHEMISTRY, AND BIOLOGY OF LOWER TROPHIC LEVELS 69 1987). However, it is not known whether temporal patterns of vertical flux (continuous versus pulsed) give rise to different benthic communities in polar systems. In the northern Bering and southern Chukchi seas, benthic standing stock is correlated with water column production (Grebmeier, 1993; Grebmeier and McRoy, 1989). Highest values occur under the Anadyr Water and Bering Shelf Water in regions influenced by high water column production, low zooplankton grazing, and presumably high carbon flux of phytoplankton to the benthos. This conclusion is also invoked for the relatively high benthic biomass occurring in the middle domain of the southeastern Bering Sea (Haflinger, 1981; Jewett and Feder, 1981). Sediment respiration experiments on shallow continental shelves provide an indication of organic carbon deposition to the benthos, which in the "hot spots" of the northern Bering and southern Chukchi seas ranged from 20 to 30 mmol O2 m-2 d-1, or an average carbon requirement of 0.5 to 1.0 g C m-2 d-1. These values are similar to that obtained in a sediment trap experiment in the Chirikov Basin, where carbon flux averaged 0.5 g C m-2 d-1 in an area where sediment respiration rates indicated an average benthic requirement of 0.48 g C m-2 d-1 (Fukuchi et al., 1993; Grebmeier and McRoy, 1989). Sediment oxygen uptake under the waters near the Alaska coastline in both the Bering and Chukchi seas indicates reduced carbon flux to the benthos, with values averaging less than 10 mmol O2 m-2 d-1 (Grebmeier, 1993; Grebmeier and Cooper, 1995). Sources and Magnitude of Variability Benthic populations provide a long-term, integrated value for processes occurring in the overlying water column. Both physical and biological parameters directly influence benthic populations. For example, changes in current transport and water mass conditions influence water temperature, nutrient influx, and oxygen concentrations to various regions of the Bering Sea ecosystem, thus influencing subsequent water column production and potential food supply to the sediments and underlying benthic fauna. Changes in physical transport mechanisms also influence sediment transport and deposition; sediment composition determines benthic community structure, whereas food supply drives benthic biomass (Grebmeier et al., 1988, 1989). Sediment processes can provide indications of variability in the water column. Sediment oxygen uptake rates reflect short-term, seasonal carbon supply to the benthos. Cycling of organic carbon back into inorganic carbon dioxide and nutrients and/or organic compounds (e.g., dissolved organic carbon) can influence both overlying and downstream production in the system. Therefore, any climatic or human-induced changes that effect the ecosystem will have long-term impacts on benthic populations and carbon turnover on these shallow shelves; the actual magnitude is as yet unknown. The green belt regime, described in the previous section (see Figure 3.15), extends along the shelf break of the entire Bering Sea continental shelf and across the northeastern shelf through Bering Strait and into the Chukchi Sea. Sediment carbon is higher on the slope than in the shelf system (Lisitsyn, 1966), and it is proposed that high production in the green belt may account for this accumulation. Benthic production reaches its highest levels in both the Bering and Chukchi seas under the northern branch of the green belt in regions influenced by nutrient-rich Anadyr Water (see Table 3.3). If changes in the Bering Sea ecosystem were to occur that directly influenced green
THE BERING SEA ECOSYSTEM: GEOLOGY, PHYSICS, CHEMISTRY, AND BIOLOGY OF LOWER TROPHIC LEVELS 70 belt production, whether via natural or anthropogenic effects at either primary or secondary production level in the water column, one would expect a direct impact on the carbon being deposited to the benthos and subsequent impact on underlying benthic populations. Changes in benthic populations, particularly fauna that are a food source to commercially harvested demersal fish and crabs, could have a long-term impact on these fisheries that are currently so productive in the Bering Sea. Marine mammal populations (e.g., gray whales, walruses, and bearded seals) that feed on the productive benthic fauna could also be affected. Studies of benthic infauna populations have been done since the early 1930s in the Bering Sea (Grebmeier and Barry, 1991; Sirenko and Koltun, 1992, and references therein), but few have been extended long enough to investigate the population dynamics that are needed to differentiate natural environmental impacts from anthropogenic ones. However, a few studies over multiple years do indicate possible ecosystem-level changes and are especially informative because they occur in the northern Bering Sea, where commercial fishing pressure is minimal. Variability in benthic macroinvertebrate epifauna (crabs) is discussed in the following chapter. A recent three-year study of the dominant ampeliscid amphipods in the northern Bering Sea, one of the most productive benthic communities in the world (Grebmeier et al., 1988; Highsmith and Coyle, 1990), found a decrease in production from 1986 to 1988 (Highsmith and Coyle, 1992). The authors indicate, however, that further long-term studies are needed to determine whether this decrease in amphipod production is due to natural predation cycles by benthic-feeding gray whales or to a long-term climate trend (Coyle and Highsmith, 1994). Their modeling results predict that perturbations in predation on the amphipod populations will have long-term effects (tens to hundreds of years) because these invertebrates are slow-maturing species, a common characteristic of high-latitude organisms. The large species of this amphipod community, Ampelisca macrocephala , requires a high carbon flux to the detrital pool it utilizes for food, and with low predation rates, which limits its occurrence to the northern cold, productive waters (Highsmith and Coyle, 1991). The modeling study by Coyle and Highsmith (1994) indicates that any decrease in carbon flux would be detrimental to such large species, shifting the population to smaller species that would out compete the large amphipods, thus changing the benthic community structure. In addition, the low bottom temperatures of the northern Bering Sea exclude most bottom-feeding fish (Bakkala, 1981; Jewett and Feder, 1980); thus, any shift to higher temperatures would extend their range and could increase predation pressure by fish, gray whales, and crabs on these amphipod populations. Another issue is the apparent spatial shift and population decline of a bivalve community, dominated by nuculanid, nuculid and tellinid bivalves, south and southwest of St. Lawrence Island in the northern Bering Sea. The single most common bivalve, Nuculana radiata, is a surface deposit feeder associated with finer silt sediments, which is consistent with the higher total organic carbon (and lower C/N values) observed in these surface sediments (Grebmeier and Cooper, 1995). Although benthic sampling in this region has not been extensive, apparent shifts in species dominance over the past few decades suggest that other processes are influencing the biomass and community structure of these Bering Sea benthic communities. Abundant benthic populations and high biomass have been documented near the St. Lawrence Island polynya and westward into the Gulf of Anadyr, including both infauna (Feder et al., 1985; Grebmeier, 1993; Grebmeier et al., 1988, 1989; Neiman, 1963; Sirenko and
THE BERING SEA ECOSYSTEM: GEOLOGY, PHYSICS, CHEMISTRY, AND BIOLOGY OF LOWER TROPHIC LEVELS 71 Koltun, 1992; Stoker, 1981) and epifauna (Jewett and Feder, 1981). The dominant benthic fauna in the Gulf of Anadyr region until the late 1980s were reported to be the common arctic tellinid bivalve Macoma calcarea and the nuculid bivalve Nucula belloti. Since then there has been a change in the dominance and distribution patterns of the major bivalves, however with the surface deposit feeding N. radiata (a more cold-water tolerant species than Macoma) moving west into the Gulf of Anadyr (Grebmeier, 1993; Grebmeier and Cooper, 1995; Sirenko and Koltun, 1992). Both N. belloti and M. calcarea declined in abundance and were observed at more westerly locations (Sirenko and Koltun, 1992). In contrast, the Nuculana community was the same in 1960 as in the 1930s, and then started to expand in the late 1970s and early 1980s, reaching its current population structure and position in 1988 (Sirenko and Koltun, 1992). Sirenko and Koltun (1992) suggest that the change in bivalve community dominance in the Gulf of Anadyr may be due either to a northward shift of cold water since the 1930s or to preference by the dominant bivalve fauna for a specific sediment type or associated chemical composition (Nuculana prefers a muddier sediment regime than Macoma ). Grebmeier and Cooper (1995) hypothesize that variable transport conditions, such as a decrease in the size of the Gulf of Anadyr gyre or the intensity of circulation around the gyre, could have influenced carbon supply rates to the benthos, leading to increased settling of finer-grained sediments over a larger area, or at least further to the west. A continuing decline in both abundance and biomass values of the dominant bivalve species and associated benthic fauna in the St. Lawrence island Polynya and Gulf of Anadyr region may indicate that a larger, ecosystem-scale change is responsible (Grebmeier and Cooper, 1995, unpublished data). Studies over both interannual and decadal time scales are required to elucidate factors influencing observed changes in benthic populations in the northern Bering Sea.
