Abstract
Species distribution modeling provides a robust tool for improving knowledge about cetacean distribution and habitat use. For at-risk species, such as the short-beaked common dolphin and bottlenose dolphins in the Mediterranean Sea, discerning areas of suitable habitat is critical for conservation planning and mitigation. This study’s aim was to investigate patterns of dolphin occurrence (T. truncatus & D. delphis) in the waters surrounding Samos, Greece, by modeling species occurrence records (from 2017-2020) as a function of environmental covariates (Bathymetry, slope, distance to shore, sea surface temperature, salinity) by applying a maximum entropy (or Maxent) modeling approach. In doing this, we developed high-performing models that elucidate suitable habitat extents across three study periods (total, summer, winter). Results indicate that predicted suitable habitat changed in extent from summer to winter, both interspecifically and intraspecifically. Moreover, regularized training gains and variable contributions in model predictions identified bathymetry (depth m) as the most important variable to both study species. Response curves generated by Maxent’s output universally identified that the model's predictive ability depended on species occupancy in shallow waters. Thus, we recommend that conservation initiatives should attempt to reduce anthropogenic pressures in shallow water areas in the Eastern Aegean Sea.