MODELING POTENTIAL DISTRIBUTIONS OF THREE EUROPEAN AMPHIBIAN SPECIES COMPARING ENFA AND MAXENT

Understanding the distribution and habitat preferences of amphibians is crucial to protecting their declining populations. It remains a challenge because most species are difficult to detect, enough data on their occurrence are needed, and the contribution of climatic and habitat factors is not well known. Various modeling approaches exist both to infer habitat preferences based on known locations, and to extrapolate species geographic distributions. We used presence-only data of three anuran species from original naturalist databases covering 34,750 km(2) in the western part of France, including 660 occurrences for the Common Midwife Toad (Afytes obstetricans), 1,910 for the Spined Toad (Bufo spinosus), and 975 for the Common Brown Frog (Rana temporaria). We compared two current modeling approaches, the Ecological Niche Factor Analysis (ENFA) and Maximum Entropy (MaxEnt) to model their potential distributions by including high resolution climate variables, and habitat and landscape variables. For each species, both analyses allowed a refined understanding of the relationships between habitat factors and the species distribution. We showed that climate and natural grasslands are key factors in explaining the species distributions and that the surrounding environment of aquatic habitats is an important driver of amphibian presence. The two models gave different predictions of distributions, which may lead to different planning of conservation areas. Our study confirms the importance of using and comparing several models, and evidenced the importance of collecting field data at a regional scale.