Evaluation of microwave soil moisture data for monitoring live fuel moisture content (LFMC) over the coterminous United States

Live fuel moisture content (LFMC), which is the ratio of water in the fresh biomass to the dry biomass, is a key variable that affects wildfire behaviour. Previous studies have assessed soil moisture as a predictor of LFMC over small areas with limited data, but a comprehensive evaluation at sub-continental scale is still lacking, and the explanatory utility has not been evaluated under different aridity conditions. In this study, the utility was evaluated using microwave soil moisture data from the ESA ECV_SM product from 1979 to 2018 and LFMC data from over 1000 sites in the coterminous United States. A time-lagged robust linear regression model was adopted, and the results were compared with analysis from in situ soil moisture measurements at adjacent sites. The results suggested that at most sites the LFMC correlates best with soil moisture within 60 days prior to LFMC sampling, and that the correlation is lower in areas with complex terrain. LFMC can be estimated from soil moisture with a mean RMSE of around 20%. The correlation between LFMC and soil moisture is significant (p<0.01) in most regions, and is mostly stable in different years. The major fuel types with a high response to soil moisture include pine, redcedar, sagebrush, oak, manzanita, chamise, mesquite and juniper, depending on the region. The LFMC ~ soil moisture correlation varies with the aridity condition, and soil moisture has a higher explanatory utility on LFMC under dry conditions. An analysis using SMAP Level-4 product indicated that the surface and root-zone soil moisture perform similarly in LFMC estimation. This study suggests that microwave soil moisture data contain sufficient information on LFMC, and may serve as a reference for the development of more sophisticated LFMC estimation methods.