Understanding how and why fuel reduction treatments are effective, and when they're not


A fuel reduction treatment, where flammable material is removed from a forested system, is one of the prominent forest management actions to contend with increasing wildfire hazard in the Western US. As wildfires burn through previously treated forest stands we have an opportunity to evaluate how well the fuel treatment meets management goals. This is difficult quantitatively because there are often multiple management goals that may be in conflict, and because fire is a contagious disturbance with a strong spatial autocorrelation structure. This violates the requirements of most standard statistical methods. Often we fall back on finding ways to control for the autocorrelation (e.g., through sub-sampling), rather than mining the information that the autocorrelation structure provides. We are exploring the application of spatially explicit models to improve assessment of fuel treatment efficacy based on measured and remotely-sensed data.