Estimation of Live Fuel Moisture Content From Multiple Sources of Remotely Sensed Data

The live fuel moisture content (LFMC) is a crucial factor that influences both the ignition and behavior of wildfires. Previous studies have utilized remote sensing data-derived vegetation indices, meteorological indicators, and soil moisture to estimate LFMC. However, an attempt to estimate LFMC using all of these data sources collectively has not been made until now. In this study, we have integrated these remotely sensed data into the construction of LFMC estimation models. Employing the Extreme Gradient Boosting (XGBoost) algorithm, we have constructed an empirical model that provides reasonable LFMC estimates (R2 = 0.56, RMSE = 27.16%) across the western U.S. states. Among different vegetation cover types, the best LFMC estimate was found for the closed shrublands (R2 = 0.66, RMSE = 24.38%), followed by open shrublands (R2 = 0.60, RMSE = 30.04%), grasslands (R2 = 0.58, RMSE = 27.02%), savannas (R2 = 0.50, RMSE = 24.23%), and woody savannas (R2 = 0.44, RMSE = 25.94%). This study advances from previous research as it involved the integration and analysis of multiple indicators from various sources. Furthermore, it necessitated the utilization of long-term meteorological data to improve the accuracy of LFMC estimation across different vegetation cover types in a large-scale regional context.