A novel spectral index for estimating live fuel moisture content and mapping burned area using EnMAP imagery

Accurate and reliable estimation of live fuel moisture content (LFMC) is fundamental for improving wildfire early warning and risk management. However, current LFMC retrieval approaches are constrained by the complex interactions between leaf biochemical composition and canopy structural properties, resulting in limited robustness and transferability across ecosystems and observation conditions. To address these challenges, this study establishes an integrated theoretical and methodological framework for LFMC estimation by combining radiative transfer modelling with sensitivity analysis. Specifically, the PROSAIL model was coupled with the Extended Fourier Amplitude Sensitivity Test (EFAST) to quantitatively assess the spectral sensitivity of reflectance to key biochemical and structural parameters. Based on these analyses, a novel LFMC-sensitive spectral index was developed to overcome the limitations of conventional single-parameter indices and to enhance the biochemical interpretability. The proposed method was validated using standardized leaf spectral datasets (LOPEX and CABO) and subsequently scaled up to the satellite level using EnMAP hyperspectral imagery. Applications to the 2023 Nacimiento wildfire in Chile and the 2025 Los Angeles wildfire in the United States showed that the LFMC-derived burned area maps were in good spatial agreement with MODIS fire products while providing substantially finer spatial resolution and greater detail. Overall, this study provides a physically based framework for LFMC retrieval and demonstrates the potential of hyperspectral remote sensing for wildfire monitoring, early warning, and risk assessment.