Environmental controls on the distribution of wildfire at multiple spatial scales

摘要: Despite its widespread occurrence globally, wildfire preferentially occupies an environmental middle ground and is significantly less prevalent in biomes characterized by extremes (e.g., tundra, rain forests, deserts). We evaluated the biophysical “environmental space” of from regional to subcontinental extents, with methods widely used for modeling habitat distributions. This approach particularly suitable biogeographic study wildfire, because it simultaneously considers patterns multiple factors controlling suitability over large areas. Maxent boosted regression tree algorithms assess wildfire–environment relationships three levels complexity (in terms inclusion variables) at spatial scales: conterminous United States, state California, five wildfire-prone ecoregions California. The resulting models were projected geographically obtain predictions also ...