Predicting suitable habitat for deep-water gorgonian corals on the Atlantic and Pacific Continental Margins of North America

Mapping marine habitats and species distributions is essential in conservation and resource management. The generation of such maps, however, is particularly challenging for the poorly sampled deep-sea species. In this study, we explored the spatial suitability of deep-water coral (Families Paragorgiidae and Primnoidae) habitat on both the Pacific and Atlantic Continental Margins of North America (PCM and ACM) using Biomapper, a modeling program which can determine habitat suitability using presence-only data. The PCM study area was divided into 2 regions to limit the geographic size of the modelled area: PCM:AK, which encompasses Alaska and PCM:BC-CA, which encompasses British Columbia, Washington, Oregon, and California. Suitable habitat was determined based on quantitative relationships between physical seascape factors and biological data. For the PCM study area, the most accurate model for Paragorgiidae in PCM:AK combined temperature, slope, current and chlorophyll (chl) a concentration (Spearman's rho = 0.79), whereas in the PCM:BC-CA it combined depth and chl a concentration (rho = 0.66). For Primnoidae, in the PCM:AK the most accurate combination included depth, slope, current and chl a concentration (rho = 0.90), and in the PCM:BC-CA, it included depth, temperature, slope and current (rho = 0.85). In the ACM study area, the most accurate model for Paragorgiidae combined temperature, slope and chl a concentration (rho = 0.71), whereas the one for Primnoidae combined temperature, slope, current and chl a concentration (rho = 0.74). In both study areas, corals were predicted to occur in areas of complex topography, mainly along the continental shelf break and on seamounts. Sensitivity analyses indicated that predicted mean values of seascape factors, in coral habitat as well as niche breadth, varied with number of coral locations, but to a much lesser extent with spatial resolution. To our knowledge, this is the first study to use Biomapper for the prediction of suitable habitat in marine species.