Click on figure to enlarge it.Figure 1. Seven biophysical domains of the southeastern Bering Sea; historical monitoring sites.
SEBSCC is timely in its development because of parallel activities taking place in the research community. The North Pacific Marine Science Organization (PICES) and GLOBEC International recently approved a joint research plan to study processes underlying Climate Change and the Carrying Capacity (the Four C's) of the North Pacific. PICES is composed by six member nations (Canada, China, Japan, Korea, Russia, and the United States), and the Four C's program is designed to address four central scientific issues (Table 1) in twelve regional and basin scale components, one of which is the eastern Bering Sea. The SEBSCC project will focus on selected aspects of each of the four issues. Thus, this project should be a U.S. contribution to a larger-scale multi-national research effort to improve understanding of processes controlling the carrying capacity of the North Pacific region(1).
Lower Trophic Level Response: How do primary and secondary producers respond in productivity, and in species and size composition, to climate variability in different ecosystems of the subarctic Pacific?
Higher Trophic Level Response: How do life history patterns, distributions, vital rates, and population dynamics of higher trophic level species respond directly and indirectly to climate variability?
Ecosystem Interactions: How are subarctic Pacific ecosystems structured? Do higher trophic levels respond to climate variability solely as a consequence of bottom-up forcing? Are there significant intratrophic level and top-down effects on lower trophic level production and on energy transfer efficiencies?
A. Key Goals of Coastal Ocean Program (NOAA Strategic Plan) Develop predictive capabilities to:
The SEBSCC project is designed to integrate with and leverage other ongoing programs in Alaskan waters, including those conducted by the Alaska Fisheries Science Center (AFSC), the Pacific Marine Environmental Laboratory (PMEL), and the University of Alaska Fairbanks (UAF).(1) In addition, 1994 amendments to the Marine Mammal Protection Act (MMPA) require that a scientific research program be developed and conducted to monitor the health and stability of the Bering Sea marine ecosystem. This sister program to SEBSCC is currently in its planning stages, and no funds have been identified for implementation of the plan.(6)
The SEBSCC project centers on pollock as a pivotal consumer and prey species in the southeast Bering Sea ecosystem (SEBSCC Conceptual Model and Scientific Approach)(3). Any sustained change in pollock abundance or even distribution could result in significant changes in the overall ecosystem.
The goals of the SEBSCC project(4) are to:
The SEBSCC project has a budget of $500K in FY96 and $1000K per year for the succeeding five years. Funding for 3-year modeling and retrospective studies is scheduled to commence on April 15, and funding for two-year process studies is scheduled to commence on October 1 (Calendar, RFP). Each year has NOAA ship time allocated for 1 - 2 months each in spring and summer, in addition to the regularly scheduled triennial surveys in the Bering Sea in 1997 and 2000.(4)
Seven physically and biologically distinct areas or habitats can be defined for the southeastern Bering Sea (Figure 1). These areas may be grouped into the middle continental shelf domain (50 - 100m isobath), outer shelf domain (100 - 200m isobaths), and continental slope (>200m isobath) domain.(7) The middle shelf is a two-layered system, with the upper layer mixed by winds and the lower layer mixed by tides (Figure 2). The outer shelf is a three-layered system, with a layer of variable fine-structure properties separating the mixed upper and lower layers. In general, the mean current is weak or insignificant (<1 cm s-1) in the middle shelf domain; weak but significant toward the northwest over the outer shelf; and stronger but variable in the Bering Slope Current over the slope (Figure 3).(7)
Click on figure to enlarge it.
Figure 1. Seven biophysical domains of the southeastern Bering Sea; historical monitoring sites.
Click on figure to enlarge it.
Figure 2. Vertical profiles of temperature, salinity, and sigma-T over the shelf.
The water that forms the Bering Slope Current originates in the Aleutian passes. The westernmost pass that provides significant input (~3 Sv) is Amchitka Pass. Amukta Pass provides additional transport (~1 Sv), and is the source of the relatively warm (>4° C) subsurface temperature maximum layer. The Bering Slope Current often breaks into a series of 20-200 km-wide eddies, the larger ones extending at least 500 m deep. Flow through Unimak Pass, the only major pass connecting the Gulf of Alaska to the eastern Bering Sea shelf, is weak in spring and stronger in fall. In September, 1995, much of the inflow through Unimak Pass followed the 50 m isobath toward Bristol Bay, while some water flowed northwestward along the 100 m isobath. This inflow, together with flow onto the shelf from the North Aleutian Flow component of the Bering Slope Current, and/or eddies, helps to make the Unimak Island region a separate habitat.(7)
Possible flow across the shelf north of the Pribilof Islands (Figure 3) is inferred from hydrographic observations and satellite-tracked drifting buoys. There may be some tidal rectification that increases retention of water in the Pribilof Islands region. In general, storms and tidal energy decrease northward along the shelf.(7)
Click on figure to enlarge it.
Figure 3. Bering Sea flow inferred from hydrographic observations and tracks of drifting buoys.³W² indicates weak and variable flow.
The three types of vertical structure on the shelf (Figure 2) are important features to include in physical models, as are inflows along the Aleutian chain and outflows to the north and west, shelf-slope exchange, meanders and eddies 20 - 200 km in extent, and the spatial extent of sea ice and the cold pool.(8)
Climatic Regimes and Sea Ice
The southeastern shelf environment has shifted interannually between warm, "mezzo," and cold modes (Figure 4), with intervals of six to seven consecutive years forming warm, "mixed," and cold intervals. Transitions between these modes have coincided with changes in climatic conditions in the North Pacific in 1976-77 and 1983-84.(9)
Click on the figure to enlarge it.
Figure 4. Climatic thermal modes of the southeastern shelf.(9)
Three categories have been proposed by PICES/GLOBEC to characterize long-term variations in climatic conditions in the North Pacific Ocean and Bering Sea:(1)
Click on the figure to enlarge it.These climate modes and regimes are reflected in the extent of sea ice cover in winter, and indicate that sea ice extent, which is easily monitored, can be used as a predictive index for climatic mode. Sea ice forms in polynyas in the northeastern Bering Sea and moves southward until it melts, transporting cold, fresh water southward and into deeper water.(9) The southernmost extent of the ice edge varies widely from year to year (Figure 5), extending south of the Pribilofs in cold years such as 1995.(7) Sea ice extent, in turn, determines the southeastern extent of the "cold pool," a layer of water less than 2°C that resides on the bottom over the middle shelf during summer (Figure 6). The 2°C isotherm is important because it represents a physiological boundary for subarctic species such as pollock.(9)
Click on the figure to enlarge it.
Figure 6. Annual formation of sea ice forms a permanent bottom pool of cold water (<2°C) of varying extent.
Primary production on the Bering Sea shelf varies spatially and temporally, e.g., annual primary production can vary depending on the location, frequency and intensity of storms that reset conditions favorable for pioneer diatom species that nourish herbivorous zooplankton (10). Existing data on the spatial pattern of primary production in the Bering Sea (Figure 7) show that production is highest along the shelf break, relatively high (with low variance) on the outer shelf, lower (but with a high variance) on the middle shelf, and low over the inner shelf and the Aleutian Basin. The spring bloom is an important phenomenon over the middle shelf, and is most intense at the retreating ice edge. Physical processes, such as eddies and tidal upwelling, seem to prolong the spring bloom through the summer at the shelf break, maintaining high phytoplankton production and standing stock and abundant populations of higher trophic level organisms. Eddies either stimulate or spatially concentrate primary production.(10) High concentrations of chlorophyll in eddies over the slope (Figure 8) and off the shelf suggest that these features may be as productive as the middle shelf front.
The ice edge melt zone spring bloom over the middle shelf north of the Pribilofs has been studied over seven seasons between 1975 and 1989, and was monitored in 1995 using continuous chlorophyll fluorescence along transects and at moorings south of the Pribilofs (Figure 1). Chlorophyll values declined over the middle shelf by late May and were greater over the outer shelf and along the Alaska Peninsula.(7) These results agree with previous findings that this bloom lasts only two to three weeks at the most.
Click on the figure to enlarge it.
Figure 7. Interpreted generalized pattern of primary production in the Bering Sea.(10)
Click on the figure to enlarge it.
Figure 8. Chlorophyll patterns detected by satellite during June 1983.11)
Coupling of Primary and Secondary Production
Since the PROBES of the late 1970's, lower trophic level research has focused on the coupling between primary and secondary production and higher trophic levels, especially pollock. On the middle shelf and parts of the outer shelf,less than half of the annual primary production is thought to be grazed by zooplankton, and a large proportion apparently sinks to the bottom. During warm years, approximately half of the annual primary production on the shelf is thought to be Phaeocystis, which is not grazed by zooplankton. In cold years (e.g., 1975-1977), Phaeocystis is absent. It is not clear whether the fraction of primary production that sinks to the bottom ungrazed is dominated by Phaeocystis. Nearly the entirety of the remaining primary production, mostly by diatoms, is thought to be consumed by zooplankton. As evidence of this close coupling, growth rates and production of the large outer shelf copepod Neocalanus are proportional to the amount and pattern of primary production.(10)
The loss of primary production from the pelagic system is a vital segment of the production budget to clarify.(10) The degree of coupling between the benthos and the plankton varies across the shelf. Primary production is highest at the shelf break (Figure 7), but grazer biomass and percent of production consumed are also relatively high (although still low in absolute terms). The percent of primary production consumed is lowest over the middle shelf. This is where one finds higher organic carbon concentrations in the sediments, and higher benthic biomass. Classifying consumer guilds in the different shelf domains (Figure 10) reflects differences in benthic-pelagic coupling. It shows that the guild that utilizes benthic infauna is centered over the middle shelf, where one finds the highest flux of carbon to the benthos. The guild that feeds on small pelagic fish is centered over the shelf break and outer shelf, where their zooplankton prey have the highest biomass.(7)
Click on figure to enlarge it.
Figure 9. Benthic-pelagic coupling across the Bering Sea shelf.
Click on figure to enlarge it.
Figure 10. Consumer guilds of the Bering Sea shelf.
Catch of pollock stocks in the eastern Bering Sea has been fairly constant at about 1.0 to 1.3 million metric tons per year for the last 20 years (Figure 11). Increased catches in the "Donut Hole" area are thought to have come from a separate stock. In the 1990's, the exploitation rate has risen, however, as adult (age-3+) biomass has decreased (Figure 12). The fishery is supported by large year classes, which were observed between 1978 and 1984, in 1989, and possibly again in 1992 (Figure 13).(5)
Click on figure to enlarge it.
Figure 11. Catch of pollock in the eastern and central Bering Sea.
Click on figure to enlarge it.
Figure 12. Eastern Bering Sea shelf pollock biomass and exploitation rate.
Click on figure to enlarge it.
Figure 13. Recruitment to the eastern Bering Sea shelf pollock population.
Recruitment of pollock into the fishery at age-3 and -4 is one of the most difficult and important parameters to assess for fishery management purposes. Once pollock enter the fishery, the abundance and distribution of year classes can be monitored using survey and catch data. While knowledge of the trends in biomass and recruitment may be sufficient for pollock management, estimates of absolute exploitable biomass and numbers of incoming recruits would provide more accurate assessments and predictions for establishing fishery quotas. Other important questions remain regarding pollock stock structure and larval exchange both on the eastern Bering Sea shelf itself and between the eastern shelf and the Aleutian Basin.(5)
It is argued that fishery recruitment is stabilized by population dynamics, while the major cause of signal destabilization appears to be the physical environment. Evidence of both processes is visible in the pollock data. Between 1964 and 1994, recruitment of age-3 pollock varied by a factor of 12, from a high of about 17 billion in 1978 to a low of just over 1 billion in 1988, with no clear temporal trend.(4)
The wide variability in pollock recruitment appears to result from fluctuations in the physical environment. However, compensatory processes such as cannibalism may stabilize recruitment. That is, low mortality at one life cycle stage is offset by high mortality at another stage. This pattern suggests that a series of life stage "switches" regulate year class size; if mortality is high or low at all switch points, a small or large year class, respectively, results; but usually high mortality at one or two of the stages produces moderate year classes.(4)
There may be reason for concern about stability of the eastern Bering Sea pelagic population. We know that over most of the last two decades, almost all of the assessed higher trophic level biomass in offshore waters, ranging from 6Ð16 million metric tons, has been pollock. Accordingly, the "effective number of species" in the pelagic domain is only 1.1, compared to 5.5 in the benthic domain. This low diversity raises the possibility that the ecosystem may be easily altered by regime shifts or other environmental perturbations.(4)
Larval Transport and Juvenile Distribution
When seeking mechanisms explaining variations in pollock recruitment, it may be important to study juvenile pollock for several reasons(12):
Surveys during September 1995, showed juvenile pollock congregating in the vicinity of the fronts over the 50 - 70 m isobaths around the Pribilofs. Gelatinous zooplankton were also extremely abundant in the same area. Underwater video showed that juvenile pollock congregated near the tentacles of large medusae during daylight hours. At night the pollock moved away, possibly to feed. These observations suggest that the pollock may be seeking refuge from larger visual predators. Historical trawl data since 1982 show increasing catches of jellyfish starting about 1990, after relatively constant catches during the 1980's. The trophic relationship between these jellyfish and age-0 pollock is unclear - they are certainly competitors, and small pollock may also be prey of large medusae.(12)
Predation and Cannibalism
The structural fronts around the Pribilof Islands may be important aggregation sites for animals at higher trophic levels. Acoustic, trawl, and observation data from around St. Paul Island in 1994 indicated that foraging seabirds and juvenile and adult pollock were much more abundant in the vicinity of the front over the 70 m isobath surrounding the island than elsewhere.(7) Vertebrate predators consume both age-0 and age-1 juvenile pollock (Figure 14), but there may be considerable differences in selection on the two age classes by different predator life stages at different times and locations. Trawl surveys indicate that juvenile pollock dominate the forage fish community, with all forage fish exhibiting large interannual variability in abundance.(9)
One of the most important "switches" controling pollock recruitment may be cannibalism of age-1 pollock by adults.(4) Data on cold pool extent, age-1 juvenile pollock abundance, and incidence of cannibalism from 1985Ð1989 suggest that cannibalism of juvenile pollock by adults, as well as predation by arrowtooth flounder and Pacific cod, increases in cold years. This difference may result from exclusion of the larger fish and the majority of juvenile pollock from the cold pool, forcing them to share the narrow outer shelf domain thereby increasing the rates of predation and cannibalism. Regime shifts also appear to be evident in changes in the distribution of age-2 pollock and the distribution, feeding behavior, and reproductive success of thick-billed murres and Steller sea lions.(9)
By retrospective comparison, we can make inferences about the processes affecting recruitment. The spawning population was about the same size in both 1982, when a year class producing high recruitment was spawned, and 1987, when a year class producing low recruitment was spawned. Reconstruction of surface currents from April through July of these years using the OSCURS model suggests that onshelf and northward transport was stronger in 1982 than in 1987. In 1983, the strong 1982 year class was highly abundant over the central and northern shelf, while in 1988, the weak 1987 year class was found mainly on the outer shelf. One possible explanation for these differences is that strong onshelf and northward flow in 1982 may have transported age-0 pollock out of areas inhabited by adult pollock and reduced the incidence of cannibalism compared to 1987.(5)
The largest populations of pollock in the Bering Sea apparently occur on the shelves, especially to the southeast. Some recent studies showed that prey populations for pollock (Figure 14) are sparse over the basin and slope outside of eddies, so the deep waters routinely may not support large pollock populations.(4) Mean age-0 pollock size increases in deeper water, suggesting numerous hypotheses, such as age-related movements and mortality, and depth-related growth rate differences.(12)
Click on figure to enlarge it.
Figure 14. Food web of walleye pollock to 200 mm FL on the Bering Sea shelf.
Food chain budget calculations suggest that observed primary production is marginally sufficient to support pollock production in the southeastern Bering Sea. The following results emerge from simple estimates of lower trophic level and fish production based on recent results:(10)
Food Chain Modeling
When modeling the eastern Bering Sea ecosystem and its components, the features that are important to incorporate into a biological model include nutrientÐphytoplanktonÐzooplankton (NPZ) dynamics such as blooms, sinking, mesoscale variability, and both crustacean and gelatinous zooplankton; pelagic and benthic predator guilds, with pollock dominating the pelagic guild; and the migration, spawning, and cannibalistic behavior of adult pollock.(8)
Click on figure to enlarge it.
Figure 15. An annual carbon budget (g C m-2 yr-1) of the outer shelf food web. [After Walsh and McRoy, 1986.]
Past models have provided a track record of performance that can help guide future choices of modeling approach. Researchers (Walsh and McRoy) derived a carbon flow budget for the southeastern Bering Sea shelf (Figure 15) and used a simple physical model to predict pollock larval distributions under different conditions. Laevastu modeled the ecosystem without water movements, but included animal migrations. Inspired by these past efforts, a tentative framework for a biophysical modeling scheme has been proposed. A number of outstanding general modeling issues remain. Considerable thought has gone into how to "capture the best parts" of different types and different spatial and temporal scopes of models.(8)