Exploring urban vegetation type and defensible space's role in building loss during wildfire-driven events in California

Abstract

The role of building characteristics and survival during wildfires are well studied. Less so is the role of urban vegetation type, condition, and location on building loss in fire events. We mapped and statistically modeled parcel-scale urban vegetation characteristics across different Defensible Space Buffers (DSBs) and their role in predicting building loss in shrub and forest dominated urban ecosystems. Using 3.0 m resolution PlanetScope imagery, geospatial data, and eCognition we mapped parcel-scale vegetation types and building characteristics in fire affected neighborhoods in Ventura and Paradise, California US. Classification and Regression Trees predicted building loss according to three different DSBs in two different ecoregions. An urban-chaparral model predicted higher bare ground cover and higher moisture content trees were significant predictors of building survival in DSBs 0–2 m from buildings. While in DSBs 10–20 m from buildings, percent bare ground, distance to herbaceous, building density, and tree distance were predictors of building loss. The urban-forest model predicted percent bare ground, distance to bare ground and herbaceous cover were significant predictors of buildings loss, while percent overhanging tree cover was less influential in predicting in building loss in DSBs less than 2 m. In DSBs 2–10 m from buildings, low shrub and tree moisture, and building densities were the most important predictors of building loss; while distance to scattered trees and building density were significant predictors in DSBs 10–20 m from buildings. Results can be used to understand the tradeoffs between vegetation-related benefits and fire hazard and for developing home insurance and municipal ordinance requirements.