Résumé
The Ocellated Lizard, Timon lepidus (Daudin 1802) occupies the Mediterranean regions of southwestern Europe (Portugal, Spain, France, and the extreme northwest of Italy). Over the last decades, a marked decline in its population has been observed, particularly on the northern edge of its distribution. As a result, it is currently considered a threatened species, especially in France and Italy. In France, a national action plan for its conservation has been put in place. In this study, ecological niche modelling (ENM) was carried out over the entire area of France in order to evaluate the species' potential distribution, more accurately define its ecological niche, guide future surveys, and inform land use planning so this species can be better taken into consideration. The modelling used data representing 2,757 observation points spread over the known range of the species, and 34 ecogeographical variables (climate, topography, and vegetation cover) were evaluated. After removing correlated variables, models were fitted with several combinations of variables using eight species distribution model (SDM) algorithms, and then their performance was assessed using three model accuracy metrics. Iterative trials changing the input variables were used to obtain the best model. The optimized model included nine determining variables. The results indicate the presence of this species is linked primarily to three climate variables: precipitation in the driest month, precipitation seasonality, and mean temperature in the driest quarter. The model was checked by a sample dataset that was not used to fit the model, and this validation dataset represented 25% of the overall field observations. Of the known occurrence locations kept aside to check the results, 94% fell within the presence area predicted by the modelled map with a presence probability greater than 0.7, and 90% fell within the area with a presence probability ranging from 0.8 to 1, which represents a very high predictive value. These results indicate that the models closely matched the observed distribution, suggesting a low impact of either geographical factors (barriers to dispersal), historical factors (dispersal process), or ecological factors (e.g., competition, trophic resources). The overlap between the predicted distribution and protected areas for this species reveals that less than 1% of the potential distribution area is protected by strong regulatory measures (e.g., national parks and natural reserves). The knowledge obtained in this study allows us to recommend some guidelines that would favor the conservation of this species.