The first presentation focused on a three-dimensional model of the circulation on Georges Bank that was used to track the transport of cod and haddock larvae. The primary focus of this particular modeling study is to determine to what extent the retention of the larval stages of these fish on the bank is physically determined. The simulations indicated that surface waters (the upper 15 m) on the southwestern part of Georges Bank are advected offshore relatively rapidly; conversely, most of the water at greater depths (near 30 m) is recirculated and remains within the 100 m isobath on the Bank. This implies that passively transported larvae have a better chance of remaining on the Bank, and surviving to recruit if they remain in the deeper portions of the water column. Simulation experiments indicated that interaction of vertical migration behavior and advective transport could be important in determining the fate of individual larvae in this ecosystem.

Interactions between the physical environment and organisms at very small scales were addressed by two presentations. The first included a model of turbulence, which described three-dimensional turbulent fields and provided an indication of the spatial (length) and temporal (persistence) scales of turbulence. A second presentation considered the effects of turbulence on the feeding currents of copepods. Turbulent intensity can markedly affect feeding success and feeding behavior. For example, in experiments different levels of turbulence led to different allocations of time among various feeding behaviors (e.g., slow swimming, fast swimming, etc.). More studies are needed to better document the effects of variable turbulence on feeding and animal behavior more generally. Scaling the results of single individual turbulence experiments (which document cm scale interactions) to populations of organisms operating on much larger spatial scale and longer temporal scales is a major difficulty that needs to be addressed. Parameterizing the effects of animal-turbulence interactions may be a way to incorporate their effects into larger-scale models.

Three presentations focused on biological aspects of secondary production measurement and modeling. The first related to the relatively recent recognition that protozooplankton are a major grazer in most ocean ecosystems. Clearly, if much of the primary production is being consumed by this group, this will have an impact on the overall energy available to the mesozooplankton, including larval fish. Protozooplankton are also consumers of smaller producers (nannophytoplankton and bacteria) that are directly unavailable to mesozooplankton. Thus, the protozooplankton may be an important trophic link between small producers and larger consumers. For example, models that ignore the grazing impact of protozooplankton will underestimate food available for metazooplankton and overestimate mortality of young fish larvae. The second presentation highlighted the importance of including inter-individual variability in making estimates of and modelling production. Examples from the literature on copepod weights, gut contents and feeding rates were used to illustrate the extreme variability that could exist between individuals subjected to presumably similar environmental conditions. Also emphasized was the behavioral flexibility that permits individual organisms to react to environmental change. The third presentation described a fluid mechanics model of the filtering current of a large calanoid copepod.

The scales of the biological and physical processes that are presently modeled are mismatched. For example, existing circulation models cover a large range of space and time. They tend to be developed for advective processes that occur over long (month to year) time scales and large (regional to basin) spatial scales. Spatial resolution of a typical ocean circulation model might be 30 km in the horizontal and 25 to 250 m in the vertical dimension. Biological processes, such as trophic interactions, vertical distributions and vertical migrations, typically occur at much finer spatial and shorter time scales in the ocean. Examples are the <1 day doubling times of phytoplankton and the strongly heterogeneous vertical distributions, on scales of cm to m, of both phytoplankton and zooplankton. This mismatch in time-space domains of the two disciplines (biological and physical) creates difficulties in developing coupled bio-physical models. One strategy, albeit an expensive one, to couple biological and physical processes in models is to increase the spatial resolution and decrease the time step of circulation models to match more closely the biological requirements. An alternative approach is to understand the biological interactions at the smaller scales, but parameterize their effects, treating them as subgrid scale processes, for inclusion in physical circulation models.

Some of the issues involved in coupling a biological-biooptical model to a physical circulation model were discussed in the specific context of a coupled circulation-biooptical model of the California Current Transition Zone. Biological features represented in the model included nutrient concentrations, nutrient uptake and growth of two size classes of phytoplankton, and the dynamics, including growth and reproduction, of three classes of zooplankton. The model simulations indicated the importance of including the microbial loop and the stage (or size) dynamics of the zooplankton in the model structure. Examples of how to model zooplankton by including information on their size or stage were presented and the advantages and disadvantages of the various approaches discussed. Three questions were then posed for general discussion.

• If models like the ones discussed were made generally available to the oceanographic community, would they have wide applicability and would they be used?

• What level of support or documentation would be needed to make such models generally useful?

• What basic model structures (e.g., size-structured or stage-structured herbivores; number and complexity of the trophic web; single-species vs. multispecies) are the most useful for producing model estimates of zooplankton secondary production?

Next Section