Using hierarchical models to estimate stock-specific and seasonal variation in ocean distribution, survivorship, and aggregate abundance of fall run Chinook salmon

Ocean fisheries often target and catch aggregations comprising multiple populations or groups of a given species. Chinook salmon (Oncorhynchus tshawytscha) originating from rivers throughout the west coast of North America support mixed-stock ocean fisheries and other ecosystem components, notably as prey for marine mammals. We construct the first coastwide state-space model for fall Chinook salmon tagged fish released from California to British Columbia between 1977 and 1990 to estimate seasonal ocean distribution along the west coast of North America. We incorporate recoveries from multiple ocean fisheries and allow for regional variation in fisheries vulnerability and maturation. We show that Chinook salmon ocean distribution depends strongly on region of origin and varies seasonally, while survival showed regionally varying temporal patterns. Simulations incorporating juvenile production data provide proportional stock composition in different ocean regions and the first coastwide projections of Chinook salmon aggregate abundance. Our model provides an extendable framework that can be applied to understand drivers of Chinook salmon biology (e.g., climate effects on ocean distribution) and management effects (e.g., consequences of juvenile production changes).