Where, when and why: modelling the distribution and habitat of deep-diving cetaceans incorporating variables depicting the deep oceanic layers

The use of Species Distribution Models (SDMs) has increased considerably in recent decades, notably for conservation purposes. SDMs are used particularly to characterise and predict marine top predator distributions thanks to the use of surface dynamic environmental variables (easily accessible and available at various spatial and temporal scales) as proxies for prey distribution. For oceanic species that spend most of their time in depth waters like deep-diving cetaceans (here beaked whales and sperm whales), the use of surface variables may limit the ability to correctly infer their habitats through SDMs. We combine, static variables that characterise the topography of the bottom water and dynamic variables integrated over different depth classes that characterise the water column into Generalised Additive Models to model the distribution of deep-diving cetaceans in the Bay of Biscay and to identify which variables are the most important for each species. We obtained relationships with the environment that allow predicting the highest densities of beaked whales and sperm whales near the continental slope, near canyons and seamounts and in the abyssal plain of the Bay of Biscay. We also identified different responses between beaked whales, for which surface, subsurface and static variables were selected as the most important variables, and sperm whales. For the latter only surface and depth variables were selected, which could suggest differences in foraging strategies and in the prey targeted between these species. The continuous development of ocean models and the availability of depth variables, allows as we have shown, the improvement of the tools available for the planning of human activities, especially for species that would be closely linked to processes taking place in deep waters, such as top predators