Potential distribution and areas for conservation of four wild felid species in Mexico: Conservation planning

Abstract

Knowing the potential distribution of species helps to focus conservation efforts more effectively, mainly when dealing with endangered species. The aim of this study was to generate potential distribution models for four species of small wild felids in Mexico (Leopardus pardalis, Leopardus wiedii, Lynx rufus and Puma yagouaroundi). The models were generated based on felids presence records, and topographic, anthropic and vegetation drivers. We used 473 records (171 for L. pardalis, 140 for L. wiedii, 86 for L. rufus and 76 for P. yagouaroundi) to build eleven models per species to then select the three with the best performance and included them in ensemble models. These were based on the formula of the weighted average, which considers the performance of the algorithms evaluated with a subsample of testing records, from which the area under the curve is calculated. In this way, in the ensemble model the consistent zones between algorithms are included, but the one with the best performance predominates. The species with the largest potential distribution area was L. pardalis with 34.3% of the national territory, while L. rufus had the smallest area (14.3%). In the four species a unique set of variables was identified that influence the probability of presence, however the altitude, the arid vegetation and the population density were important variables for three of the four species. We verified our models with recently published presence records. The results of this study reflect a robust analysis of the current and potential distribution of four species of wild felids in Mexico. In addition to being the first step to develop effective conservation strategies at national and local levels.