A Comparison of Multi-temporal Airborne Laser Scanning and the Fuel Characteristics Classification System for Estimating Fuel Load and Consumption

Abstract:

Characterization of pre-fire fuel loading and subsequent fuel consumption are important for scientists and managers looking to assess fire behavior, fire effects, and smoke emissions. Quantification of these processes can occur at multiple scales from ground-based, in-situ measurements to global fuel models and satellite observations. Errors in translating fuel estimates from different scales stem from the high spatial heterogeneity of fuels, suggesting that comparison of multiple data sources is needed to calibrate fuel estimation methods and identify sources of uncertainty. To address this need, we evaluated two wildland fires in the Inland Northwest (USA), comparing post-fire field measurements, pre- and post-fire airborne laser scanning (ALS) data, and pre-fire Fuel Characteristics Classification System (FCCS) fuel model estimates. Smoke plumes were sampled at both fires as part of the Biomass Burning Fluxes of Trace Gases and Aerosols (BB-FLUX) project, suggesting further applications for the fuel consumption estimates derived in this study. Field measurements and ALS data were used to estimate pre-fire fuel load and fuel consumption at fine-scale (5m resolution), while FCCS data were used to estimate pre-fire fuel load at 30m resolution and generate fuel consumption estimates in Fuel and Fire Tools (FFT). Results at the Tepee fire (Oregon, USA) indicated good agreement in total fuel consumption estimates between the ALS (9,385 Mg) and FCCS (9,010 Mg) methods. However, at the Keithly fire (Idaho, USA) consumption estimates were much lower using ALS (10,090 Mg) than FCCS (33,990 Mg). The largest contributor to this difference was the Black cottonwood-Douglas-fir-quaking aspen forest class, which based on field observations was largely misclassified by FCCS, comprising 9% of the burned area and accounting for 25,181 Mg of consumed fuel. Among non-forested areas, ALS and FCCS estimates were more similar, accounting for 7,233 Mg and 8,455 Mg of fuel consumption respectively. This study demonstrates that fuel consumption estimates from ALS and FCCS are generally in good agreement, which suggests minimal loss of accuracy due to scaling. However, the generality of FCCS also makes it prone to classification errors as observed at the Keithly fire.