Improving Forest Land Cover in Economic Models using Recent Remote Sensing Products

Forest plays an increasingly critical role in the global carbon cycle, yet no consistent globally harmonized dataset exists that distinguishes intact from modified forest across regions and is compatible with integrated assessment frameworks that are used in global climate and socioeconomic analysis. Here we compile and compare forest area estimates from four widely used global datasets, including FAO FRA, MODIS, ESA-CCI, and HILDA+ at the country and Agro-Ecological Zone (AEZ) level over 2001–2020. While all four datasets show a strong agreement in total, intact, and modified forest area shares (pairwise r > 0.95) at country level, substantial disagreement emerges in forest area estimates in certain regions, particularly in ecologically heterogeneous biomes such as tropical savanna-forest transition zones and sparse woodland margins, driven by differences in canopy cover thresholds and classification schemes. Building on these findings, we developed a country- and AEZ-level dataset of modified and intact forest area covering 2001–2020, compatible with the Global Trade Analysis Project (GTAP) framework. This dataset provides a valuable resource for quantifying forest biomass and carbon inventories, modelling land-use-based greenhouse gas emissions, and supporting climate policy analysis under international frameworks.