6533b854fe1ef96bd12adec5

RESEARCH PRODUCT

Incorporating classified dispersal assumptions in predictive distribution models – A case study with grasshoppers and bush-crickets

subject

description

Abstract Current and future species distributions depend on environmental conditions, but the ability of species to shift their range boundaries or to expand their distribution ranges in response to global change also depends on their dispersal capacity. Dispersal capacity, however, has often been neglected in previous studies that either assumed no-dispersal or full dispersal, both of which are unrealistic for most taxa. The aims of this study are (i) to identify the predictors of the present spatial distribution on a regional scale for 13 grasshoppers and bush-crickets, and (ii) to derive predictions of their future distributions under climate change by applying different dispersal capacity classes for different mobile species. We analysed relations between 13 grasshopper and bush-cricket species and their environment in southern Germany by calculating multivariate binary logistic regression models. Assuming unlimited dispersal, we predicted the future distribution (2041–2055) for seven grasshopper species that showed stronger sensitivity to climate variables (independent contribution to R N 2 > 60 % ) than other species. We established alternative range predictions for three out of the seven species that are predicted to expand their range in the future by assuming different dispersal capacities rather than no-dispersal or unlimited dispersal. All models showed a high discriminative power and for the seven most climate-sensitive species they predicted a marked change, under full dispersal, in their distribution range. For the three species that are predicted to expand their range in the future, unlimited dispersal strongly overestimated ranges in comparison to no-dispersal and classified dispersal. Even the most mobile of these species were unable to completely fill future suitable areas assuming classified dispersal capacity. The no-dispersal model underestimated the ranges of species in comparison to classified dispersal. Our findings illustrate how a more realistic assessment of species dispersal capacity can strongly alter model results and conservation needs based on predictions.