Testing remotely-sensed predictors of meso-carnivore habitat use in Mediterranean ecosystems

Maria J Santos,S L Ustin,M Santos-Reis,Luis M Rosalino

doi:10.1007/s10980-016-0360-3

Abstract

The legacy of human use of Mediterranean ecosystems results in spatial and temporal heterogeneity of resources for wildlife. Understanding wildlife use of these ecosystems may be improved by including information on ecosystem type, structure, and function extracted from remote sensing data. To assess whether we can improve our understanding of wildlife-habitat use by including information on ecosystem type, structure and function. We tested whether remote sensing derived descriptors of ecosystem type, structure (tree cover and patch size) and function (productivity and stress) determine the habitat of stone martens (Martes foina), common genets (Genetta genetta), and European badgers (Meles meles) in southern Portugal. We linked radio-tracking data from five stone martens, five genets and eight badgers with aerial photography, and some spectra-selectivity to classify vegetation, its structure, productivity and drought stress. Statistically-derived generalized linear mixed regression models using combinations of remotely sensed descriptors of ecosystem type, structure and function, performed better than single ecosystem type descriptors. Inclusion of information on ecosystem functioning in predictive models of habitat use is more informative than ecosystem type alone, suggesting functional relationships between wildlife and their habitat. However, inclusion of both ecosystem type and function maybe limited to finer spatial resolutions. Our results illustrate the untapped potential of remote sensing to provide detailed descriptors of habitat at adequate spatial scales, now that they are freely available and are systematically collected over space and time. This information adds useful insights on wildlife-habitat relationships under changing patterns of land use and climate.

Highlights

Understanding the degree to which species can cope with the effect of land use and climate changes on its habitat is fundamental in ecology, conservation, and sustainability (Morrison et al 1998; Scott et al 2002; Boyce 2006)
Inclusion of information on ecosystem functioning in predictive models of habitat use is more informative than ecosystem type alone, suggesting functional relationships between wildlife and their habitat
Our results illustrate the untapped potential of remote sensing to provide detailed descriptors of habitat at adequate spatial scales, that they are freely available and are systematically collected over space and time

Summary

Introduction

Understanding the degree to which species can cope with the effect of land use and climate changes on its habitat is fundamental in ecology, conservation, and sustainability (Morrison et al 1998; Scott et al 2002; Boyce 2006). The last decades have seen a proliferation of modeling studies on species niche (Pulliam 2002; Guisan and Thuiller 2005), resource selection (Boyce 2006), habitat suitability (Hirzel et al 2006; Zielinski et al 2006), and how these relationships dictate species geographic distributions (Guisan and Zimmermann 2000; Rushton et al 2004; Moisen et al 2006) These inferences have shown that habitat can be described in many forms, from descriptors of ecosystem type (Scott et al 2002), spatial and temporal structure (Zielinski et al 2006; Santos 2010), and ecosystem functioning (Wulder et al 2004), to inclusion of multiple (cumulative and dynamic) effects between ecosystems and climate (Tingley et al 2012; Santos et al 2014). These data require matching with adequate environmental context information

Objectives

Methods

Results

Discussion

Conclusion