Conclusions
Relying on the variable importance of predictors of ecological phenomena
is intimately connected with management and policy. The statistical
identification of the most influential predictors for a given phenomenon
(e.g., invasive species success) enables better environmental
policy-making and conservation decisions. Integrating the evaluation of
variable importance into statistical analyses can be crucial for
interpreting results in ecological and evolutionary studies. Future
research with RF and HP techniques, possibly among use of other modern
developments, holds promise in integrating drivers from both coarse and
fine scales, and variable importance measures can be applied to address
other biodiversity questions, particularly those arising from the
impacts of global change.