Species distribution modeling of deep‐diving cetaceans Marine Mammal Science

Species distribution models (SDMs) have been developed and extensively validated for diverse cetaceans within the California Current Ecosystem off the West Coast of the United States. These studies have recognized the challenges associated with developing robust models for deep‐diving cetaceans—sperm whales and beaked whales—thus limiting the accuracy of predictions for management and ecological understanding. In this study, we explore whether additional biologically relevant predictor variables can improve models for deep‐divers. These variables are related to the oxygen minimum layer and phytoplankton and micronekton biomass and could influence prey availability for cetacean top predators. We found that the addition of these variables improved the performance of SDMs for sperm whales, as well as for some more common baleen whale and dolphin species, but that the accuracy of deep‐diver models was nevertheless poor. The sightings data sets for deep‐diving cetaceans have small sample sizes compared to other cetaceans, and sightings are distributed nearly randomly across the study area and model domain. These factors hinder the development of useful environmentally driven models of spatial distribution.