Effect of Climate on Wildfire Size: A Cross-Scale Analysis

Theory predicts that wildfires will encounter spatial thresholds where different drivers may become the dominant influence on continued fire spread. Studying these thresholds, however, is limited by a lack of sufficiently detailed data sets. To address this problem, we searched for scale thresholds in data describing wildfire size at the Avon Park Air Force Range, south-central Florida. We used power-law statistics to describe the “heavy-tail” of the fire size distribution, and quantile regression to determine how the edges of data distributions of fire size were related to climate. Power-law statistics revealed a heavy-tail, a pattern consistent with scale threshold theory, which predicts that large fires will be rare because only fires that cross all thresholds will become large. Results from quantile regression suggested that different climate conditions served as critical thresholds, influencing wildfire size at different spatial scales. Modeling at higher quantiles (≥75th) implicated drought as driving the spread of larger fires, whereas modeling at lower quantiles (≤25th) implicated that wind governed the spread of smaller fires. Fires of intermediate size were negatively associated with relative humidity. Our results are consistent with the idea that fire spread involves scale thresholds, with the small-scale drivers allowing fires to spread after ignition, but with further spread only being possible when large-scale drivers are favorable. These results suggest that other data sets that have heavy-tailed distributions may contain patterns generated by scale thresholds, and that these patterns may be revealed using quantile regression.