Conservation strategies worldwide are designed around species distributions and biogeographic patterns, often under the assumption that these patterns are relatively stable. With climate change, however, distributions are now shifting as species seek more suitable conditions. There is an urgent need for credible predictions of future species distributions to guide conservation planning. Nowhere is this more evident than in the ocean where range shifts are up to an order of magnitude faster than on land. By using 50 years of observational data provided by fisheries research surveys on the continental shelves of North America, my proposal will develop validated, process-based predictions of future marine species ranges along this continent. As opposed to previous efforts, this project will directly test models against past observations, and then use the best models for prediction into the future. These models will consider climate, dispersal, and species interactions, putting them at the forefront of species distribution modeling. Products from this research will include: 1) validated models for predicting marine species distributions along the continental shelves of North America, 2) predictions of future shifts in species distributions driven by climate change, 3) forecasts of future biodiversity hotspots and ecoregional boundaries, and 4) evaluation of existing marine conservation plans in light of future species distribution shifts.