Resumen de Statistical modelling of habitat selection
Shaykhah Aldossari, Jason Matthiopoulos, Dirk Husmeier
To understand the impact of habitat destruction or modi cation on biodiversity there is increasing demand on predictive models that reliably forecast future changes in species distributions. In the present paper, we build on an existing model, the Generalized Functional Response, whose predictions about habitat preferences and species distribution are robust to changes in habitat availability. We improve upon this model in two distinct ways by using Gaussian mixtures to approximate habitat availability and Gaussian basis functions to describe habitat preferences. The proposed model is found to improve descriptive and predictive performance when applied to realistic simulated data and real species abundance data.
