Modelling the species richness distribution of French dung beetles (Coleoptera, Scarabaeidae) and delimiting the predictive capacity of different groups of explanatory variables

Aim: To predict French Scarabaeidae dung beetle species richness distribution, and to determine the possible underlying causal factors. Location: The entire French territory has been studied by dividing it into 301 grid cells of 0.72 x 0.36 degrees. Method: Species richness distribution was predicted using generalized linear models to relate the number of species with spatial, topographic and climate variables in grid squares previously identified as well sampled (n = 66). The predictive function includes the curvilinear relationship between variables, interaction terms and the significant third-degree polynomial terms of latitude and longitude. The final model was validated by a jack-knife procedure. The underlying causal factors were investigated by partial regression analysis, decomposing the variation in species richness among spatial, topographic and climate type variables. Results: The final model accounts for 86.2% of total deviance, with a mean jack-knife predictive error of 17.7%. The species richness map obtained highlights the Mediterranean as the region richest in species, and the less well-explored south-western region as also being species-rich. The largest fraction of variability (38%) in the number of species is accounted for by the combined effect of the three groups of explanatory variables. The spatially structured climate component explains 21% of variation, while the pure climate and pure spatial components explain 14% and 11%, respectively. The effect of topography was negligible. Conclusions: Delimiting the adequately inventoried areas and elaborating forecasting models using simple environmental variables can rapidly produce an estimate of the species richness distribution. Scarabaeidae species richness distribution seems to be mainly influenced by temperature. Minimum mean temperature is the most influential variable on a local scale, while maximum and mean temperature are the most important spatially structured variables. We suggest that species richness variation is mainly conditioned by the failure of many species to go beyond determined temperature range limits.