Sand and fire: applying the sandpile model of self-organized criticality to wildfire mitigation

Abstract: Prescribed burns have been increasingly utilized in forest management in the past few decades. However, their effectiveness in reducing the risk of destructive wildfires has been debated. The sandpile model of self-organized criticality, first proposed to model natural hazards, has been recently applied to wildfire research for describing a negative linear relationship between the logarithm of fire size, in area burned, and the logarithm of the fire incidence number of that size. In this study, we leverage this powerful sandpile model to perform a series of simulations demonstrating that prescribed fires indeed suppress the incidence of destructive wildfires. The same conclusion was obtained when this method was utilized to analyze historical data of forest fires from three American states: Florida, California and Georgia. Our study justifies the application of the sandpile model to wildfire research and establishes a novel method based on the analysis of slope estimated from the sandpile model, related to the leverage described in the literature, for facilitating the investigation of potential risk factors of wildfires. The new method may be utilized for the development of optimal strategies for prescribed burning. An R-script for sandpile model simulation is available for further wildfire investigation.