Spatial Modeling with Integrated Age-Structured Assessment Models in a Changing Environment

Climate change forcing can affect the spatial distribution of fish populations, which would subsequently impair the accuracy of spatially aggregated age-structured assessment models. In this study spatially aggregated models were compared to spatially explicit models under different climate change scenarios. The climate change scenarios that were evaluated influenced the spatial distribution of walleye pollock (Theragra chalcogramma) in the Eastern Bering Sea by shifting the population northward. Overall, accurate estimates of total biomass resulted from both spatially aggregated and spatially explicit models, while the uncertainty in biomass estimates was reduced with a spatially explicit model. Spatially explicit models that estimated ontogenetic and climate change parameters gave smaller Deviance Information Criterion (DIC) values than those that set such parameters, although the range in DIC across replicated datasets could be large. Further, if the functional form of climate change effects on movement is unknown, using a random walk to estimate movement is feasible.