Predictions of species distributions derived from correlative models can help to understand the spatial patterns of biodiversity, refugia and identification of possible threats for relict populations. The amount of available data and software is steadily increasing as well as the number of studies applying niche model techniques. However, a discrepancy between increasingly complex studies and the understanding of underlying processes, derivation of valid assumptions, and the development of conceptual backgrounds is still a problem (Jimenez-Valverde et al. 2008). One problem might be that models predicting potential distributions of species are easily computed and that test statistics are sometimes misleading (Lobo et al. 2008) facilitating misinterpretations. However, properly applied and critically evaluated, during each step, from model building to the final interpretation, CEMs can be powerful tools and very helpful to generate and test hypotheses.
Acknowledgements We are indebted to Bernhard Misof and Stefan Lotters for fruitful discussions and Michael Raupach for logistic support. Robert Hijmans helped us with palaeoclimatic scenarios. DR was funded by the 'Graduiertenforderung des Landes Nordrhein-Westfalen' and JD by the 'Deutsche Forschungsgemeinschaft' (DFG).
Was this article helpful?