Overstory and Understory Fuel Type Mapping Using GEDI and Sentinel Data Fusion

Wildfires play a transformative role in the Mediterranean basin, affecting forest composition and structure. Accurate fuel mapping is essential for advancing fire risk assessments and refining fire behavior models. Wildfires typically begin in surface fuels and can escalate to canopy fuels if there is sufficient continuity in the canopy. This study addresses the need for a comprehensive approach to mapping overstory and understory fuels within an integrated classification system, incorporating forest structure and phenology in central Portugal. Fuel types were classified based on the FirEUrisk hierarchical fuel classification system (FHFCS) through a three-step approach: 1) overstory mapping using multispectral and radar data from Sentinel-1 and Sentinel-2, combined with topographic variables; 2) estimation of shrubland and grassland heights using biophysical models based on precipitation and the Normalized Difference Vegetation Index (NDVI); and 3) understory mapping using spaceborne LiDAR data from the Global Ecosystem Dynamics Investigation (GEDI), employing decision-based rules and spatial interpolation of GEDI footprints. This methodology offers a simple and efficient approach for large-scale mapping of both overstory and understory using multispectral, radar, and LiDAR data in the absence of airborne LiDAR, which could enhance fire simulation models for both surface and crown fires.