Figure 2. The importance ranking of predictors. The importance of each predictor in each model was evaluated based on the random forest model fits to training datasets. (a) and (b) show the importance based on decrease in accuracy and Gini Index for elevated fuel stratum, respectively. (c) and (d) show the same but for near-surface fuel stratum, while (e) and (f) for surface fuel stratum. Colour marks different predictor types: climate- cyan; topography-brown; soil- grey; plant-red.