The LULUCF Data Hub: translating global land use emissions estimates into the national GHG inventory framework

Abstract. Global estimates of anthropogenic land use CO₂ fluxes from bookkeeping models and dynamic global vegetation models, as well as those derived from Earth observations (EO), can differ widely from those obtained by aggregation of national GHG inventories (NGHGI). These discrepancies are mostly due to conceptual differences, reflecting varying definitions (or system boundaries) of anthropogenic CO₂ fluxes, particularly on the forest carbon sink. This study presents a new online platform ‒ the "LULUCF Data Hub" ‒ which compiles various datasets of country-level CO₂ emissions and removals from land use, land-use change and forestry (LULUCF). Through a multi-disciplinary collaboration, results from two key global initiatives ‒ one based on carbon cycle modelling (Global Carbon Budget, GCB 2024) and one on Earth Observations (Global Forest Watch, GFW 2024) ‒ are "translated" to conceptually align with NGHGI definitions. We compared these global approaches to the 2024 NGHGI data submitted to the UNFCCC under the Paris Agreement reporting rules (cut-off date April 2025). The aim is to enhance comparability and foster consensus on anthropogenic land-use CO₂ flux estimates. Translating definitions between datasets considerably improves the match in land use CO₂ flux estimates, both globally and for most countries. At the global level, NGHGI indicate that LULUCF is a net sink of -3.4 Gt CO₂ yr^-1 (2001–2023 average; 2023 estimate is an extrapolation for all NGHGI), whereas GCB reports a net source of +4.7 Gt CO₂ yr^-1. This +8.1 Gt CO₂ yr^-1 discrepancy drops to +1.8 Gt CO₂ yr^-1 when GCB fluxes are translated to NGHGI definitions. When the global net forest-related flux of -5.8 Gt CO₂ yr^-1 estimated by the GFW model (sum of fluxes from Forest and Deforestation including organic soils) is limited to a proxy of managed forest area, the sink declines to -3.5 Gt CO₂ yr^-1, closely aligning with NGHGI (-3.2 Gt CO₂ yr^-1 for forest-related flux only, including Forest, Deforestation, and organic soils in Indonesia). Despite improved alignment in magnitude after translation, some differences persist in global CO₂ flux trends: in the last two decades, forests (including re/afforestation) have remained a roughly constant net sink in NGHGI and GCB but are a declining net sink in GFW; deforestation emissions are stable or slightly declining in NGHGI and GCB but increase in GFW. For other land uses, NGHGI globally reports a greater sink than GCB (the GFW model does not include other land uses). At the country-level, we identify both convergences and divergences in CO₂ flux estimates between NGHGI and the translated datasets – in sign, magnitude and trend. Notably, after translation, estimates of both GCB and GFW still differ from individual NGHGI values in 19 countries by more than ±0.1 Gt CO₂ yr^-1 on average (2001–2023), highlighting areas that deserve further investigation.

The LULUCF Data Hub was first presented in July 2024 during the IPCC Expert Meeting on Reconciling Anthropogenic Land Use Emissions, showing preliminary data. In November 2024 version 1.0 (with NGHGI pre-Paris reporting, GCB 2024 data and GFW 2024 data) was launched. Our study shows results from version 2.0, corresponding to 2024 data from all datasets: i) NGHGI data from the first Biennial Transparency Reports (BTRs) submitted under the Paris Agreement (cut-off date April 2025), complemented by previous reports (from version 1.0) for countries with no BTR submission, ii) data from GCB 2024 and iii) data from the GFW carbon flux model 2024. Data from all versions, including future annual updates, can be visualized and downloaded from the LULUCF Data Hub hosted by the European Union Forest Observatory (European Commission Joint Research Centre, 2025) and from the online repository (Melo et al., 2025; available at https://doi.org/10.5281/zenodo.17140775).