Landscape Ecology of Mammals: Relationships between Density and Patch Size

A much discussed issue in landscape ecology is how processes that operate within spatially subdivided subpopulations scale-up to create a larger, landscape-level dynamic. A first step in answering this question is to ask to what degree subpopulations within a landscape vary in performance. Here we test the null hypothesis that densities of mammalian populations are constant over patches of varied size, i.e., that performance, as estimated via density, does not covary with patch area. Using a composite database from published studies, we found that densities of 20 of 32 species did not vary with patch area, while five showed increasing and seven decreasing density-area relationships. Studies reporting significant density-area relationships tended to include a greater number of patches of a greater range of sizes than those that reported no relationship, suggesting that statistical power may be an issue. Landscapes comprised of smaller, less-isolated patches tended to have negative density-area relationships and landscapes with larger, more isolated patches tended to have positive density-area relationships. Our results suggest that no consistent density-area relationship operates over all systems of patches. Instead, the patterns appear to be scale-dependent: frequent movement of individuals in the process of selecting habitats (patches) over smaller-scaled landscapes produced negative density-area relationships; movement of individuals among more isolated patches appeared to involve larger- and longer-scale population processes involving colonization and extinction and positive density-area relationships. Despite the fact that patches represent a central focus in landscape ecology, they appear to be a construct of human convenience rather than biological entities with a set number and kind of processes.