A spatial modeling approach for evaluating impacts of climate-driven species movement on biomass estimation methods.

Fishery stock assessments typically rely on biomass estimates derived from stratified random sampling, where a key assumption is a consistent spatial biomass distribution over time. However, climate-driven movements of marine species may be violating this assumption, potentially introducing biases into biomass estimates. To address this, we develop a spatially explicit data-driven mathematical modeling framework where species-specific movement is driven by environmental variables such as water temperature and geographic habitat preferences. To demonstrate this modeling approach we develop spatial simulations for three Atlantic fish species under several temperature scenarios and population trends. We then compute biomass estimates derived from the stratified random samples of the model output, and compare estimates derived from design-based stratified mean to those estimated from a spatio-temporal model-based approach that allows inclusion of environmental covariates. Our modeling framework produces spatial models that include climate-driven changes in biomass distributions, and resulting biomass estimates are impacted by species shifting their spatial densities over time. This framework has broad uses including evaluation of survey designs, management strategy evaluations, climate-driven biomass predictions, and conducting a rigorous statistical assessment for climate-induced bias of specific biomass estimation approaches.