Wildfire Hazards: A Model of Disaster Response

Disaster managers allocate resources to accomplish a set of objectives in a highly uncertain environment. The allocation decisions made throughout the course of a disaster inevitably impact final outcomes such as the damage and cost of a disaster. We develop a unique stochastic dynamic model of disaster response in which managers face a temporal tradeoff between disaster containment and the protection of valuable assets, and focus on wildfire response. Comparative dynamics indicate that when the value of concentrated threatened assets increases, wildfire managers divert response resources away from wildfire suppression toward asset protection at the expense of re growth. This leads to a rise in the expected duration, size and cost of a wild re. Based on our theory, we motivate and develop an econometric model that contributes to the literature on multivariate frailty (hazard) model estimation. We use this model to jointly estimate wildfire duration, size, and cost in a way that exploits the temporal variation in a unique dataset of U.S. wildfires from 2001 to 2008. Our results suggest, among other things, that 100 threatened residential structures leads to an increase in the expected wildfire duration by 8.6%, expected acreage burned by 26%, and the expected response cost by 22.2%. As the wildland urban interface continues to grow, wildfire managers will increasingly face the tradeoff between containment and protection. Consequently federal and state agencies can expect longer, larger, and more expensive wildfires in the future.