Coupling the Advanced Regional Prediction System and the Discrete Event Specification Fire Spread Model to Predict Wildfire Behavior

The cost of wildfire suppression in the United States has risen dramatically over the last 20 years. As the interface between wildland and urban areas expands, increased emphasis is being placed on rapid, efficient deployment of firefighting resources. Various numerical models of wildfire spread have been developed to assist wildfire management efforts over the last several decades; however, the use of coupled fire-weather models to capture important feedbacks between the wildfire and the atmosphere is a relatively new development. This research evaluates a coupled system consisting of the Advanced Regional Prediction System (ARPS) atmospheric model and the raster-based Discrete Event Specification Fire Spread model (DEVS-FIRE). After the theoretical basis of coupled fire-atmosphere modeling and the basic design of previous vector-based models are outlined, idealized tests, verification using data from the FIREFLUX experiment, and case studies of the September 2000 Moore Branch Fire and the April 2011 Rock House Fire are presented. The current version of ARPS/DEVS-FIRE produces mixed results; broader-scale feedbacks appear to be represented somewhat skillfully, but the model also exhibits systematic flaws, which are exacerbated by efforts to depict fine-scale feedbacks or fire spread in high-wind cases. These results demonstrate the importance of coupled modeling and suggest improvements that must be made to ARPS/DEVS-FIRE before reliable results may be obtained.