Can we disentangle predator–prey interactions from species distributions at a macro-scale? A case study with a raptor species


P. Aragón, Dept of Ecology and Evolution (DEE), Biophore, Univ. of Lausanne, CH-1015 Lausanne, Switzerland. E-mail:


There is a debate on whether an influence of biotic interactions on species distributions can be reflected at macro-scale levels. Whereas the influence of biotic interactions on spatial arrangements is beginning to be studied at local scales, similar studies at macro-scale levels are scarce. There is no example disentangling, from other similarities with related species, the influence of predator–prey interactions on species distributions at macro-scale levels. In this study we aimed to disentangle predator–prey interactions from species distribution data following an experimental approach including a factorial design. As a case of study we selected the short-toed eagle because of its known specialization on certain prey reptiles. We used presence–absence data at a 100 km2 spatial resolution to extract the explanatory capacity of different environmental predictors (five abiotic and two biotic predictors) on the short-toed eagle species distribution in peninsular Spain. Abiotic predictors were relevant climatic and topographic variables, and relevant biotic predictors were prey richness and forest density. In addition to the short-toed eagle, we also obtained the predictor's explanatory capacities for 1) species of the same family Accipitridae (as a reference), 2) for other birds of different families (as controls) and 3) artificial species with randomly selected presences (as null models). We run 650 models to test for similarities of the short-toed eagle, controls and null models with reference species, assessed by regressions of explanatory capacities. We found higher similarities between the short-toed eagle and other species of the family Accipitridae than for the other two groups. Once corrected by the family effect, our analyses revealed a signal of predator–prey interaction embedded in species distribution data. This result was corroborated with additional analyses testing for differences in the concordance between the distributions of different bird categories and the distributions of either prey or non-prey species of the short-toed eagle. Our analyses were useful to disentangle a signal of predator–prey interactions from species distribution data at a macro-scale. This study highlights the importance of disentangling specific features from the variation shared with a given taxonomic level.