Contrast the environment of the Bering Sea shelf and slope from observations made during 1996, 1997, and 1998 to understand the strong interannual variability in the ecosystem.

1. seasonal pack ice extent and duration (1995, a record ice year; 1996, early advance and retreat; 1997, average; 1998, early advance and retreat)
2. wind-driven mixing over the shelf during spring and summer (1996, average; 1997, anomalously low; spring 1998, anomalously high),
3. summertime sea surface temperatures (1996, average; 1997, 4^oC positive anomaly; 1998, early summer only slightly warmer than normal, late summer not yet determined)
4. mixed layer depth (1996, >20m; 1997, 10-20m; 1998, not yet determined)
5. spring phytoplankton bloom (1996, late bloom; 1997, typical ice edge bloom in timing and magnitude; 1998, anomalously late)
6. summer nutrient reservoir concentrations (1996, normal; 1997 anomalously low; 1998, normal).

Conditions during 1997, at least in terms of forcing and lower-trophic response, are well described. The anomalously warm sea surface temperature observed in the Bering Sea resulted from regional wind mixing and heat exchange with the atmosphere, rather than propagation of an oceanic anomaly from the equator. As occurred in recent years, an early spring diatom bloom (about 12 mg m^-3) was associated with sea ice. By the end of April, chlorophyll concentrations had decreased to pre-bloom values. During April winds were unusually weak and these conditions generally persisted through August. The anomaly in wind mixing followed the same general pattern. A striking mixing event, however, did occur in mid-May. The impact of this storm was to mix the upper 40-45 m, thereby making nutrients from the lower layer available in the upper water column. This reduced the reservoir of nutrients typically found throughout the summer in the lower layer. The storm also weakened the pycnocline, which permitted further depletion of nutrients. This likely occurred through both a vertical flux of nutrients across the pycnocline to the surface and net photosynthesis below the mixed layer throughout the summer. An examination of heat content revealed that it was similar to that in the previous year. The heat, however, was concentrated in a shallow mixed layer. The extreme SST anomalies appear to be due primarily to the lack of winds rather than to increased solar radiation resulting from reduced cloud cover. This warm upper layer extended over portions of the coastal domain into waters as shallow as 30 m. In general, the coastal domain waters are mixed. One consequence was that the transition between coastal and middle shelf water was poorly defined and tens of kilometers wider than previously reported. The changes in structure likely affected the usual biophysical dynamics that result in primary and secondary production throughout summer. While biophysical processes likely account for much of the nutrient depletion on the shelf, a change in the flux from source waters may have exacerbated this situation. Observations of temperature and salinity versus depth were collected several along a slope/shelf transect. In spring 1997, transport in both the Aleutian North Slope Current (ANSC) and the Bering Slope Current (BSC) was unusually large, > 6 x 10⁶ m³ s^-1, whereas transport is typically <3 x 10⁶ m³ s^-1. Moored current records from the ANSC revealed consistent flow, supporting the inference of steady, strong flow during 1997. How the enhanced strength of these currents affects shelf/slope exchange is not known. The flux of oceanic water through Bering Canyon is a source of nutrients for the shelf. During 1997, satellite tracked drifters revealed that little or no onshelf flow occurred there also.

• Use SEBSCC model simulations to compare circulation, its effect on pollock survival, and upper-trophic-level interactions in the southeastern Bering Sea for warm and cold years to determine the influence of interannual variability in the ecosystem, and provide results of SEBSCC biophysical models.

A second, but less tested, hypothesis is that oceanographic conditions, particularly temperatures, limit the growth of some populations. In cold years with an extensive cold pool, pollock are forced off the shelf on to the outer shelf. Pollock recruitment is reduced because of lower egg hatch, and concentration of adult pollock result in increased cannibalism. In warm years with a small cold pool, pollock recruitment is enhanced because of higher egg hatch, and the dispersal of pollock results in decreased cannibalism. Within a year or so, the population catches up so that cannibalism balances egg hatch. This may have been the mechanism that caused the pollock fluctuations over the regime shift.

One hypothesis has been refuted. Because lactating northern fur seals prey primarily on pollock, it was thought that an indication of foraging success as determined from fur seal teeth would correlate with observed pollock abundance. SEBSCC research does not show a strong direct link.

Project scientists have identified several more features that are candidate indices of the status of the Bering Sea ecosystem. Among these are:

1. extent of ice and its influence on the timing of the phytoplankton bloom and hence the succession of bottom-up mechanisms that must match in time and space the needs of first feeding pollock larvae,
2. wind which influences the ecosystem through mixing and by advection as direct wind-driven flow. The latter may increase the separation between early life history stages and their cannibalistic parents, since recent evidence suggests that many larvae over the shelf may be in water depths less than 10 m. The timing of storms also play a role in the intensity of stratification over the middle shelf and the depletion of nutrients in the cold pool.
3. concentration of nutrients retained in the bottom layer of the middle shelf (essential for prolonged production at the inner front),
4. species composition of phytoplankton (e.g., rare coccolithophorid blooms),
5. location, strength and stability (eddies) of the Aleutian North Slope Current and Bering Slope Current system that affects advection of nutrients and pollock larvae onto the shelf, and
6. increased presence of previously low abundance biota (e.g., jellyfish and coccolithophores).