Premium
MULTIVARIATE ANALYSIS OF SCALE‐DEPENDENT ASSOCIATIONS BETWEEN BATS AND LANDSCAPE STRUCTURE
Author(s) -
Gorresen P. Marcos,
Willig Michael R.,
Strauss Richard E.
Publication year - 2005
Publication title -
ecological applications
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.864
H-Index - 213
eISSN - 1939-5582
pISSN - 1051-0761
DOI - 10.1890/04-0532
Subject(s) - species evenness , ecology , species richness , habitat , spatial ecology , community structure , multivariate statistics , population , abundance (ecology) , mantel test , geography , biology , statistics , mathematics , demography , sociology , genetic diversity
The assessment of biotic responses to habitat disturbance and fragmentation generally has been limited to analyses at a single spatial scale. Furthermore, methods to compare responses between scales have lacked the ability to discriminate among patterns related to the identity, strength, or direction of associations of biotic variables with landscape attributes. We present an examination of the relationship of population‐ and community‐level characteristics of phyllostomid bats with habitat features that were measured at multiple spatial scales in Atlantic rain forest of eastern Paraguay. We used a matrix of partial correlations between each biotic response variable (i.e., species abundance, species richness, and evenness) and a suite of landscape characteristics to represent the multifaceted associations of bats with spatial structure. Correlation matrices can correspond based on either the strength (i.e., magnitude) or direction (i.e., sign) of association. Therefore, a simulation model independently evaluated correspondence in the magnitude and sign of correlations among scales, and results were combined via a meta‐analysis to provide an overall test of significance. Our approach detected both species‐specific differences in response to landscape structure and scale dependence in those responses. This matrix–simulation approach has broad applicability to ecological situations in which multiple intercorrelated factors contribute to patterns in space or time.