Habitat-based predictive mapping of rockfish density and biomass off the central California coast

Abstract

Understanding the association between components of habitat and fish distribution and abundance is important in order to achieve accurate stock assessments. We developed gen - eralized additive models (GAM) and spatially predictive maps of rockfish abundance at the individual species level using habitat descriptors collected from visual surveys and fine-scale bathy metry. We advanced beyond presence/absence and presence only models to create predictive maps of fish density (100 m-2) and biomass (kg 100 m-2) for Sebastes rosaceus (rosy rockfish) and S. constellatus (starry rockfish), both common species in commercial and recreational fisheries along the central coast of California. Selected models included co-variables of seafloor depth, complexity, substratum type, and heterogeneity. Predicted density and biomass of both species were highest in areas of complex rock on the continental shelf off Points Lobos and Sur at 50-90 (S. rosaceus) and 80-120 m (S. constellatus) water depth. Our results will be useful both in stock assessments of these data-poor species as well as in allocation of fishing effort, catches, and other space-based management decisions.