Incorporating Predation and Temperature into Multi-Species Statistical Catch-At-Age Models: an Example from the Bering Sea

As the scope and scale of ecological studies has expanded, so too has institutional knowledge of the importance of variability, physical disturbances, and food-web-, habitat-, and species- diversity to ecosystem structure. The importance of such biocomplexity to the natural world has similarly translated into the design and implementation of stock-assessment models used for fisheries management; increasingly, there is a movement away from single species management towards multi-species and/or ecosystem process-based approaches. A variety of tools have emerged to address such management needs, including multi-species age-structured statistical models (MSM). MSM combines traditional catch-at-age stock assessment models with multispecies virtual population analysis models (MSVPA) in a Bayesian framework and uses various abundance and diet data (e.g., catch-at-age data, predator diet information) to estimate fishing mortality, recruitment, stock size, suitability coefficients and predation mortality. Such an approach also provides a statistical framework to evaluate and manage both the direct and indirect effects of fisheries harvest on multiple species. However, previous iterations of the model used static predator rations to predict species interactions and were therefore unable to capture climatic driven changes in predation and fishing impacts. In this study, we modified an existing MSM for the Bering Sea to incorporate temperature dependent predator rations. The results of the model were used to compare fishing impacts during various thermal regimes and provide an example of how mutli-species models can be used to manage ecosystems in fluctuating climatic conditions.