Fig. 5

Variable importance plot ranking the explanatory variables (y-axis) based on their predictive power and contribution to the accuracy of the predicted B. pfeifferi presence and absence in the study area by the RF model. A higher MDG value indicates that the variable was significant in distinguishing between classes in the model. A variable with a higher MDA contributed more to the predictive accuracy of the model, as its absence (or randomization) lead to a significant drop in model performance.