Integration of landscape metric surfaces derived from vector data improves species distribution models

Abstract

A species’ distribution across the landscape is not random, but it is affected by distribution, size, abundance and connectivity of landscape patches. This spatial configuration of the landscape shapes ecological processes, for example the location of home ranges, migration routes and migration ability. Landscape metrics describe the configuration of a landscape quantitatively. While traditional approaches in habitat modelling only consider environmental attributes at a specific location, the integration of landscape metrics adds more functional information. In this paper we evaluated a method of directly incorporating a set of landscape metrics as covariates into a Maxent habitat model. Specifically, we used hexagons as statistical units for the calculation of landscape metrics. With this method also landscape metrics calculated with vector data sets can be used for SDM. We tested this approach for the smooth snake (Coronella austriaca) in the Austrian Alps. The experimental designs resulted in an improvement of the habitat models. Moreover, the results demonstrated the benefits of landscape metrics for the model outcomes at different scales.