The Butte Fire: A Case Study in Using LIDAR Measures of Pre-Fire Vegetation to Estimate Structure Loss Rates

The Butte Fire occurred in the northern Sierra Nevada foothills of California in September, 2015, resulting in the loss of two lives and an estimated 1,565 structures in 70,000 acres (28,000 ha.) burned. This study evaluates the utility of vegetation measures derived from pre-fire LIDAR data in predicting structure loss for the Butte Fire. Additional explanatory variables such as elevation, topography, structure density, and structure access are also examined to determine their impact on structure loss rates. Loss estimates based on LIDAR-derived vegetation measures are compared to estimates derived from infrared aerial imagery to evaluate the relative effectiveness of using LIDAR for this purpose. LIDAR-derived vegetation density in the 50-foot (15 m.) zone around each structure was found to be the most significant variable associated with structure loss. Elevation was the second most statistically significant predictor of structure loss. On average, a 10% increase in vegetation density in the 50-foot (15 m.) zone around each structure led to an estimated 10.2% increase in structure loss. A 1000 foot (300 m.) rise in elevation was associated with a 15% increase in structure loss. Topographic variables such as slope, aspect, and topographic position did not appear to have an important effect on structure losses. Measures of structure density and structure access also were not statistically significant predictors of structure loss rates. Vegetation cover derived from infrared aerial imagery proved nearly as accurate as LIDAR-derived vegetation density in estimating the probability of structure loss.