Model of trawlable area using benthic terrain and oceanographic variables—Informing survey design and habitat maps in the Gulf of Alaska

Bottom trawl surveys provide fishery‐independent data on relative abundance and life history parameters for a wide range of marine taxa. Survey data are used to assess species distribution, biological interactions, and ecosystem structure and to manage marine resources. Not all bottom types or oceanographic conditions accommodate this survey method. We applied National Ocean Service hydrographic smooth sheets to evaluate physical attributes associated with habitat available to surveys. Random forests were used to evaluate the relative influence of benthic terrain and oceanographic predictors in determining accessibility to bottom trawl gear. We examined the marginal importance of each predictor, quantified the response gradient, and applied piecewise regression to determine threshold breakpoint values. Thresholds were used to develop predictive maps and distinguish untrawlable habitat at the scale of discrete towpaths and survey stations. Untrawlable habitat was associated with increased complexity in terrain, roughness, slope, surface curvature, substrate coarseness, current, and aspect. Maps of critical thresholds suggest different variables constrain the probability of a successful trawl in the nearshore, shelf, and continental slope. Overlay analysis of the model projection demonstrates the utility of archived smooth sheet data and identifies areas where higher resolution data might improve results. The model and maps produced in this analysis might be used to identify habitats available to and impacted by commercial trawl fisheries, inform the relative availability of various species and habitat types to bottom trawl surveys, evaluate bias in assessment indices and ecosystem metrics derived from survey data, and advance habitat‐specific biomass estimates.