Random forest model results.

Higher values of the “mean decrease in accuracy” and the “mean decrease in Gini index” indicate higher predictor variable importance. Variables in bold are the ones included in the random forest model that minimizes the number of variables used as well as the out-of-bag error rate after applying the model improvement ratio approach (see Material and Methods). Conditional inference variable importance is shown for a conditional inference random forest model, which corrects for potential biases due to correlations between predictor variables. Variables marked by ‘*’ are the five most important variables according to the conditional inference. The variables Jan-Dec represent the seasonal timing of rainfall; the remaining variables are representative of spatial differences in habitat. Also see Tables 1 and 2.