Regional Decline of an Iconic Amphibian Associated with Elevation, Land-Use Change, and Invasive Species

Ecological theory predicts that species with restricted geographic ranges will have the highest probability of extinction, but species with extensive distributions and high population densities can also exhibit widespread population losses. In the western United States populations of northern leopard frogs (Lithobates pipiens)-historically one of the most widespread frogs in North America-have declined dramatically in abundance and geographic distribution. To assess the status of leopard frogs in Colorado and evaluate causes of decline, we coupled statewide surveys of 196 historically occupied sites with intensive sampling of 274 wetlands stratified by land use. We used an information-theoretic approach to evaluate the contributions of factors at multiple spatial extents in explaining the contemporary distribution of leopard frogs. Our results indicate leopard frogs have declined in Colorado, but this decline was regionally variable. The lowest proportion of occupied wetlands occurred in eastern Colorado (2-28%), coincident with urban development and colonization by non-native bullfrogs (Lithobates catesbeianus). Variables at several spatial extents explained observed leopard frog distributional patterns. In low-elevation wetlands introduced fishes, bullfrogs, and urbanization or suburbanization associated negatively with leopard frog occurrence, whereas wetland area was positively associated with occurrence. Leopard frogs were more abundant and widespread west of the Continental Divide, where urban development and bullfrog abundance were low. Although the pathogenic chytrid Batrachochytrium dendrobatidis (Bd) was not selected in our best-supported models, the nearly complete extirpation of leopard frogs from montane wetlands could reflect the individual or interactive effects of Bd and climate patterns. Our results highlight the importance of considering multiple, competing hypotheses to explain species declines, particularly when implicated factors operate at different spatial extents.