Urban amphibian assemblages as metacommunities

Urban ecosystems are expanding throughout the world, and urban ecology is attracting increasing research interest. Some authors have questioned the value of existing ecological theories for understanding the processes and consequences of urbanization. In order to assess the applicability of metacommunity theory to urban systems, I evaluated three assumptions that underlie the theory - the effect of patch area, the effect of patch isolation, and species-environment relations - using data on assemblages of pond-breeding amphibians in the Greater Melbourne area of Australia. I also assessed the relative impact of habitat fragmentation, habitat isolation, and changes to habitat quality on these assemblages. Poisson regression modelling provided support for an important increase in species richness with patch area (pond size) and a decrease in species richness with increasing patch isolation, as measured by surrounding road cover. Holding all other variables constant, species richness was predicted to be 2.8-5.5 times higher at the largest pond than at the smallest, while the most isolated pond was predicted to have 12-19% of the species richness of the least isolated pond. Thus, the data were consistent with the first two assumptions of metacommunity theory evaluated. The quality of habitat at a pond was also important, with a predicted 44-56% decrease in the number of species detected at ponds with a surrounding vertical wall compared with those with a gently sloping bank. This demonstrates that environmental differences between habitat patches were also influencing amphibian assemblages, providing support for the species-sorting and/or mass-effect perspectives of metacommunity theory. Without management intervention, urbanization may lead to a reduction in the number of amphibian species persisting in urban ponds, particularly where increasing isolation of ponds by roads and associated infrastructure reduces the probability of re-colonization following local extinction.