Multi-scale approach to disentangle the small mammal composition in a fragmented landscape in central Brazil

Abstract

Habitat quality and its availability affect the structure of communities at a variety of different spatial and environmental scales, including habitat and landscape levels. Analyses conducted at multiple scales have demonstrated the importance of landscape patterns with regard to community structure. We used a multi-scale approach to better understand the composition of small mammal communities in central Brazil. We delimited 24 sampling units for the purpose of investigating the response of small mammal composition at the habitat, patch, and landscape scales, while evaluating quality and availability of habitat in each patch, as well as landscape configuration and landscape composition. We used a Mantel correlogram to check for spatial autocorrelation, and partial redundancy analysis to correlate the matrix of species composition with the matrix of environmental variables. Small mammal composition was spatially structured and correlated with variables at all environmental scales: vegetation structure (fallen logs, lianas, and canopy cover), habitat patch (normalized difference vegetation index [NDVI]), and landscape composition (total core area—a forest cover metric). Our findings show evidence of the limited dispersal capacity of small mammals and the importance of habitat quality at habitat and patch scales, and habitat availability at the landscape scale, for small mammal composition. This study highlights the importance of a multi-scale approach for understanding community structure in a human-altered landscape.