Numerical ocean models (NOMs) are used in stock assessment models to link climate to fisheries management, but there are fewer options to link NOMs to ecological data and processes for upper trophic levels at smaller scales (1-1000km2). I introduce three high-resolution options using Alaska case studies: whole-of-ecosystem spatio-temporal models for decadal forecasts; synthesizing tags, surveys, and fishery data to understand seasonal Pacific cod movement; and mechanistic movement models for reconstructing tracks from archival tags. These examples all show a growing scope for data assimilation to connect ocean physics to fisheries management.Keywords: Regional Ocean Modelling System; Animal movement; Fisheries Management