Wednesday, August 4, 2010: 3:40 PM
310-311, David L Lawrence Convention Center
Background/Question/Methods Changes in the abundance and ranges of species are not only determined by climate, but also by land-use. However, land-use is only rarely considered in species distribution models (SDMs). We tested whether SDMs of European birds that included climatic and land-use variables fitted current distributions better than climate-only models. Additionally, we compared predictions for future bird distributions using climate-land-use models and climate-only models. We selected land-use variables for each species from a database assessing the importance of 17 land-use types for birds and we applied a backward selection algorithm to reduce model complexity. Models calibrated on a European scale were projected into the future using climate change scenarios for Germany. Two future land-use scenarios were applied: At one extreme, we left current land-use constant; at the other extreme, we assumed that land-use adjusts instantaneously to climatic conditions. We further calculated future bird distributions assuming no and full dispersal ability of the species.
Results/Conclusions For most species, SDMs could be improved by including land-use variables in the models. Predictions of range shifts from climate-land-use models differed substantially from climate-only models. In general, climate-land-use models predicted less range loss and less range gains. Nevertheless, differences in predicted future distributions between the no and the full dispersal models were more pronounced than between climate-only and climate-land-use models. These results demonstrate that land-use modifies climate change impacts on future bird distributions and that current climate-only models overestimate future changes. Nevertheless, most uncertainty in SDMs is caused by the difficulty in judging adequately the dispersal ability of the species.