Document: MIC-3-65-42

Species-specific and scale-dependent responses to landscape configuration: A simulation model based on correlation matrices.

WILLIG, M.R.*, M.GORRESEN and R.E.STRAUSS

Texas Tech University, Lubbock, TX 79409-3131 USA ¹

Abstract:
The response of populations and communities to landscape configurations, including aspects of habitat fragmentation, is a major focus of ecology and conservation biology. Nonetheless, responses may be both taxon-specific and scale-dependent. We used a matrix of correlations and partial correlations between the density of a species or a measure of community structure, and a suite of eight landscape characteristics, to represent biotic responses to habitat configurations. We then developed a simulation model to evaluate similarities in landscape responses among species at the same spatial scale, or among scales for the same species (or assemblage). Correlation matrices can be similar based on the magnitude or direction of association between landscape features, and a biotic response variable, such as those related to the densities of species or the structure of assemblages (e.g., richness, evenness, and diversity). Attributes of the magnitude and direction of association may subsequently be combined via a meta-analysis to provide an overall test of significance. Moreover, the overall result can be decomposed into constituent parts pertaining to particular landscape characteristics. We illustrate the utility of the approach with reference to the demographic and community-level responses of bat species to habitat configurations at three spatial scales (areas corresponding to circles of 1, 3, and 5 km radius) in Interior Atlantic Rainforest of eastern Paraguay. Landscape features focused on forested habitat, and included 8 indexes: relative area of forest cover, mean patch size, patch density, edge density, mean nearest neighbor distance, mean proximity, and mean shape index. Both species specific-differences in response to landscape configuration, and scale-dependence in those responses were detected by our model. This matrix-simulation approach has broad applicability to any ecological situation in which multiple intercorrelated factors contribute to patterns in space or time.

Keywords: scale-dependence, landscape configuration, habitat fragmentation, simulation model

This abstract is being presented at: 10:45 AM in session:
Oral Session #58: Landscape Ecology.