Résumé
Species richness is widely used by animal ecologists as a biodiversity metric. Modelling landscape variation in species richness is, however, subject to strong statistical constraints when reliable richness estimates are restricted to few sampling sites. In this study, we assessed the efficacy of some richness surrogates whose computation is based on the relative abundance of relevant species groups. Available from any single sample, abundance estimates are usually adequately modelled as a function of landscape configuration, and as such offer considerable advantages over the direct modelling of species richness. When applied to a complex bat assemblage in a fragmented neotropical rainforest, most candidate surrogates were tightly correlated with observed species richness (r= 0.71 to 0.88). These surrogates can be used as reliable tools to compare the efficiency of different landscape management scenarii or landscape restoration priorities with regard to biodiversity.