Rosalino L.M.,CARNIVORA Nucleo de Estudos de Carnivoros e Seus Ecossistemas |
Rosalino L.M.,University of Lisbon |
Sousa M.,Instituto da Conservacao da Natureza e da Biodiversidade |
Pedroso N.M.,CARNIVORA Nucleo de Estudos de Carnivoros e Seus Ecossistemas |
And 4 more authors.
Vie et Milieu | Year: 2010
Determination of species geographic distribution and factors constraining it is a fundamental step for wildlife management and conservation. Red fox (Vulpes vulpes Linnaeus, 1758) is a widely cited species in carnivore ecological literature, mainly due to its wide distribution, generalist behavior and commonness. Nevertheless, few data are available on distribution constraints in the south-western part of its range. Our study aims to describe what factors are constraining the local distribution of this carnivore in central-west Portugal - a mountainous Mediterranean area, with a strong Atlantic climatic influence. A presence/pseudo-absence (based on the detection of signs of presence) Logistic Regression Model (LR) and a presence-only Maximum Entropy Model (Maxent) were constructed, testing the effect of several biotic (e.g., prey distribution) and abiotic variables (e.g., land cover, distance to urban areas, distance to roads, elevation) as constraining factors in the local distribution of the fox. The resulting models, based on 30 positive fox signs (plus 30 random pseudo-absence in LR) showed that only variables directly associated with food resources (presence of agricultural patches, closeness to human settlements/structures and proximity to areas with wild rabbit occurrence) significantly influenced the presence of foxes. These results were consistent for both modelling approaches. The high model fit of the LR model (AUC = 0.808), together with that of the Maxent analysis (AUC = 0.728), gave a high degree of confidence on these results. Our results demonstrate that although subject to some criticism., the indirect census method is easy to implement and can provide reliable results on populations' distribution and limiting factors. This approach might, and should, be complemented with other methods (e.g., captures, non-invasive methods, etc.) in order to obtain more precise information on population dynamics and ecology.