Quantifying the multiscale environmental controls on wildfire using species distribution models

Despite its widespread occurrence globally, wildfire preferentially occupies an environmental middle-ground and is significantly less prevalent in biomes characterized by environmental extremes (e.g., tundra, rainforests, deserts). We evaluated the biophysical “environmental space” of wildfire from regional to continental extents, using methods developed for modeling species distributions (“niche models”). This approach is particularly suitable for the biogeographical study of wildfire, because it simultaneously considers patterns in multiple factors controlling wildfire suitability over large areas. We used the Maxent algorithm to assess wildfire-environment relationships for three levels of complexity in variable inclusion at three spatial scales: the conterminous United States, the state of California, and five wildfire-prone ecoregions of California. The resulting models were projected geographically to obtain spatial predictions of wildfire suitability and also projected to other regions to assess their generality and spatial “transferability.” Predictions of the potential range of wildfire had high classification accuracy but also predicted the suitability of areas where no wildfires have been recently observed, hinting at their potential (or past) suitability. The models identified several important variables that were not suspected to be important in the large-scale control of wildfires. Models projected to different areas were useful only when they overlapped appreciably with the target area’s environmental space. Application of this approach should allow us to explore the global range of wildfire in a changing climate, the potential for wildfire restoration where it has been “extirpated,” and, conversely, the “invasiveness” of wildfire following changes in plant species composition. To our knowledge, it is also the first application of niche models to characterize environmental controls on a disturbance process.