The GFLI database update includes a major change in land use change values, as well as a new version of the database:
- GHG Protocol & SBTi FLAG-aligned database
- GFLI database, PEF-aligned
Some terminology used in this article may be difficult for newer database users. Please consult our FAQ for a better understanding of what Land use change and its various terms are.
What did the work entail?
We’ve integrated Orbae’s jurisdictional direct land use change (jdLUC) emission factors across 1,834 feed crop and start material inventories in the GFLI database.
This change is expected to significantly alter LUC-related GHG emission profiles compared to existing sLUC-based inventories. Unlike sLUC, which allocates land conversion emissions across all crops produced within a country, jdLUC reflects the direct spatial overlap between crop expansion and observed land conversion within a given jurisdiction. As a result, jdLUC emission factors can diverge substantially from sLUC values, both in magnitude and direction.
The transition from statistical to jurisdictional direct land use change accounting within the GFLI database represents a substantial methodological improvement for LUC representation in feed LCAs. By introducing spatially explicit, satellite-based jdLUC data across all feed crop inventories, GFLI enhances the accuracy, consistency, and comparability of LUC-related GHG emissions across feed supply chains. This integration enables feed producers and traders to develop branded datasets that better reflect sourcing practices and verified deforestation- and conversion-free (DCF) achievements, while facilitating robust and timely GHG reporting under SBTi FLAG. More broadly, the approach supports improved land-use-related decision-making across the feed, food, retail, and pet care sectors.
The methodology behind the data
jdLUC uses satellite-based imagery with 20-year historical land cover change to quantify direct LUC emissions attributable to specific crops and jurisdictions.
With the update, 262 of the 1,834 crop production datasets in the GFLI database integrate jdLUC emissions factors from 94 crop-country combinations. The remaining 1,572 — a total of 612 crop-country combinations are modelled using a jdLUC proxy approach.
The starting point of jdLUC proxy is the global cropland layer of Liao et al., 2024 the first global 30 m annual cropland extent dataset (2000–2024) of annual and perennial (herbaceous and tree) cropland extent. From this, we derive cropland expansion and overlay it with land use change datasets to identify the associated land conversion, resolved country by country — producing a layer of “unspecified crops”. The gridded crop distribution dataset CropGrids (Tang et al., 2024), which maps the distribution of annual and perennial crops in 5 km resolution for the year 2020, is then used to allocate the conversion from the category of unspecified crops to specific crops (e.g., barley, wheat). The allocation follows a shared-responsibility approach and considers both annual and perennial crops.
What has changed?
We compared GFLI database version 3.1 based on the change from statistical LUC and peatlands to jurisdictional direct LUC and related peatlands. A median deviation indicates a total shift of +4.7%, while looking only at the LUC the impact roughly doubles. For most ingredients, jdLUC raises the footprint where deforestation is real and lowers it where national sLUC averages are overstated (1,262↑ / 548 ↓ of datasets). This is also the case for peat emissions (1,079↑ / 431↓).
For more details about which types of ingredients were affected and further explanation how, a recording will be made available from the launch webinar of September 8.
What isn’t updated yet:
- Newly integrated datasets for the U.S. on state level: Sorghum and wheat still use sLUC values.
- Updated datasets for the U.S. on state level: The soybean and corn from the new projects are modelled with an older version of Orbae data and therefore the integration is not fully aligned with the rest of the database.
- Branded data: those that contributed data through branded data are responsible for the update of their data towards this new standard if a (jurisdictional) direct LUC approach was not already undertaken. Step-by-step guidance is provided to facilitate that.
GFLI database (PEF-aligned)
The database will continue to look as it currently does. The following columns have been updated (climate change = CC):
- RECIPE_GWP_TOT (Global warming – Including LUC & Peat)
- RECIPE_GWP_LUC (Global warming – LUC only)
- RECIPE_GWP_PEAT (Global warming – Peat only)
- RECIPE_LUTOT (Land use – Total)
- RECIPE_LUTR (Land use – Transformation)
- EF_GWPtotal (Climate change)
- EF_GWPfossil (Climate change – Fossil)
- EF_GWPluluc (Climate change – Land use and LU change)
- EF_GWPpeat (Climate change – Fossil (only peat))
- EF_SQP (Land use)
GHG Protocol: SBTI-FLAG aligned database
With the update, this new version of the database will become available to license. This version of the database will consist of the forest, land and agriculture (FLAG) breakdown and its non-FLAG emissions, in all GFLI-existing impact assessment methods and its allocations. This allows full alignment with GHG Protocol Land Sector and Removals Standard.
The major difference between the PEF and SBTi-FLAG version is the different breakdown of emissions. The FLAG breakdown consists of the following categories:
- FLAG land-use change (LUC) — CO₂ from land conversion (direct LUC, 20-yr amortisation) + peat oxidation
- FLAG land management — non-CO₂ — N₂O, CH₄, NH₃ & NOₓ from soils & manure, plus CO₂ from lime & urea
- FLAG crop production (upstream) — input of fertilizer & pesticide production, energy, transport, etc.
- FLAG removals— fossil energy for processing — drying, milling, extraction
- FLAG land management — net CO₂ — net soil-carbon (SOC) change
- Non-FLAG emissions
Licensing options
Licensing this version of the database is possible for either only licensing this version of the database, or signing a ‘package deal’ to gain access to both the GFLI PEF-aligned and FLAG-aligned version.
Find out more about the costs.↗
What are the consequences of these decisions?
With the shift of the baseline methodology, we recommend data providers to prioritize a direct land use change approach. For data projects that include commodities available in Orbae, they can be directly used. For those not yet available, the jdLUC proxy method, which will become available in Orbae by the end of 2026, may be used. You can use the Orbae web app or any other tool or source that uses jdLUC to get this data. The GFLI methodology will phase out statistical land use change, relying on it only if no other data is available.
For all data projects, the request to have PEF-compliant data will still be required (equal amortization of land use change according to PAS-2050). Aligning the data with GHG Protocol’s FLAG data is optional.