Generalized Linear Models (GLM’s)
For A. microtis , the global model including all predictor variables was the most adequate model to explain the variation in the occurrence of this species (Appendix-Table 3). When evaluating the isolated effect of each variable, we observed that only Nat_open_areas did not affect the distribution of A. microtis (Table 1). The variables, Forest, Water, and Anthr_open_areas had a positive effect on the occurrence of the species (Table 1) (Figure 5A). We observed an increase in the probability of A. microtis by 74.4% for areas with higher percentages of these variables, especially Forest and Water. The Urban_areas variable negatively affected the occurrence of A. microtis (Figure 5A). We estimated that this variable reduces the occurrence of the species by about 50%.
Table 1. Predictive variables selected for Atelocynus microtis by the ANOVA, using the most ajusted model: A. microtis~Forest + Anthr_open_areas + Urban_areas + Nat_open_areas + Water.