Predictive mapping of two nearshore, demersal fish species on northern California reefs using scuba-based visual surveys and remote sensing

California\u27s shallow rocky reefs provide critical habitat for a diverse assemblage of fishes. Effective management strategies for these species require both accurate stock assessments as well as a spatially explicit understanding of the relationship between fishes and characteristics of their habitat. We used a generalized additive model framework to create spatially predictive maps of the abundance and biomass of two demersal fish species prevalent on northern California reefs: lingcod (Ophiodon elongatus) and kelp greenling (Hexagrammos decagrammus). These models incorporated data from SCUBA-based fish and habitat surveys at depths from 12-26 meters as well as measures of seafloor topography derived from remotely sensed bathymetric surveys. Topographic position index, a measure of a location’s elevation relative to its surroundings, was an important predictor for all chosen models. Percentage of rocky substrate and rugosity, a metric describing habitat complexity, were also important predictive variables in many of the chosen models. These findings indicate that these species have complex associations with specific habitat features and that they may select these features of their environment at multiple spatial scales. The results presented here highlight the utility of combining remotely sensed habitat data with SCUBA-based visual surveys to aid in stock assessments and marine spatial planning