Multivariate Spatial Regressions Help Explain Wildfire Hot Spot Intensities in Washington, USA

Abstract

Wildfires have become increasingly prevalent in the western United States, posing threats to human communities and the built environment. This study builds upon previous research by investigating the factors influencing wildfire hot spot distribution in Washington State. Using spatial regression models (generalized linear regression and geographically weighted regression), we examine the relationships between wildfire hot spots and various geographic features, including climate variables, human-caused ignitions, land use, population density, road density, and the wildland-urban interface. Our results indicate that lightning-caused fires and road density are significant factors contributing to hot spot intensity in central Washington, while human-caused ignitions play a crucial role in eastern Washington. Surprisingly, precipitation shows varied correlations with hot spots, with some areas experiencing an unexpected positive relationship between precipitation and hot spot intensity due to increased fuel growth. The study highlights the importance of localized approaches to wildfire mitigation, emphasizing the need for tailored risk reduction strategies based on regional factors.

Keywords: Wildfire, spatial analysis, hazard mitigation, spatial statistics, Washington state