Severity-Driven Assessment of Greenhouse Gas Emissions from Large Mediterranean Wildfires Using Remote Sensing and Vegetation Mosaics

Estimating wildfire greenhouse gas (GHG) emissions in Mediterranean landscapes is challenging due to heterogeneous fuel mosaics and limited scalability of field-based approaches. This study presents a Geographic Information System (GIS) based framework that integrates land-cover data, pre-fire biomass estimates, fire severity mapping, and established emission factors to produce spatially explicit estimates of biomass consumption and GHG emissions. Fire severity was derived from multitemporal Sentinel-2 imagery using the differenced Normalized Burn Ratio (ΔNBR) and combined with land-cover information to define vegetation–severity classes for emission estimation. A key innovation is the identification of co-occurring vegetation types within the same spatial units, allowing emissions to be quantified across vegetation mixtures rather than single classes, providing a more realistic representation of Mediterranean forests. Applied to the 2022 Bejis wildfire, pre-fire biomass within the burned area was 673,601 tons. Coniferous forests dominated, but co-occurrence with shrubland and herbaceous layers produced the highest emission contributions, highlighting the role of vegetation interactions. Total emissions were estimated at 625,938 tons of equivalent CO2, and comparison with large-scale datasets (CAMS Global Fire Assimilation System, Global Fire Emissions Database) shows general coherence. This severity-driven, vegetation-explicit framework demonstrates robust potential for quantifying wildfire emissions across heterogeneous Mediterranean landscapes, though uncertainties remain due to pre-defined biomass, burning efficiency, emission factors, assumptions in fire severity mapping, and limited field validation. The approach can support improved regional GHG inventories and wildfire management strategies.