Abstract

We introduce a model for the expectation of occupancy as a function of spatial scale variables and variables related to characteristics of sites for avian and bat habitat quality assessment. We utilized a unique functional representation for the joint density based on the estimated modes of the univariate distributions to model the expectation of occupancy. Unlike binary classification methods, the proposed construction does not require a clear distinction of used versus unused habitats. It also allows extending variable ranges for spatial generalization. We demonstrate how one could utilize the expectation of occupancy to reliably predict habitats a particular avian species may choose for nesting or roosting, to compare various tree species and sites, and to identify prominent characteristics that govern the selection of a site by avian and bat species. The methods can be utilized in the management of suitable sites and the conservation of endangered species.