Why EDGAR and UNFCCC Disagree on the Same Country

Last updated: · PlainEmissions Editorial

Two different ways to count the same thing

Open any country page and you'll see two emissions totals that don't match. EDGAR — the EU's Joint Research Centre database — typically reports a different number than that country's official UNFCCC inventory submission, sometimes by 5%, sometimes by 30%, occasionally by more. This isn't a mistake. It's a feature of how these two systems are designed.

EDGAR is a globally-consistent bottom-up model. The European Commission's Joint Research Centre takes activity data (energy consumption from IEA, agricultural statistics from FAO, industrial production from UN Industrial Statistics) and applies its own emission factors to estimate emissions for every country using the same methodology. The strength is comparability — every country is measured the same way. The weakness is that EDGAR doesn't have local knowledge of country-specific practices, fuel quality, or regulatory measures.

UNFCCC inventories are self-reported. Each country compiles its own inventory using IPCC reporting guidelines, with country-specific emission factors, country-specific activity data, and country-specific quality-assurance processes. The strength is local accuracy — a Norwegian inventory team knows Norwegian conditions better than a Brussels modeller does. The weakness is heterogeneous methodology and (in some cases) political pressure to present favorable numbers.

Where the gap is small

For well-resourced developed countries with mature inventory systems (EU member states, the United States, Canada, Australia, Japan), EDGAR and UNFCCC typically agree within 5-10% on the national CO2 total from energy. Energy CO2 is the easiest sector to estimate — fuel consumption data is well-recorded, emission factors are well-characterized, and the chemistry is straightforward (carbon in fuel → carbon dioxide in atmosphere).

For these countries, when EDGAR and UNFCCC agree closely, you can be confident the headline national energy-CO2 number is reliable to within roughly 5%. That's tighter than most economic indicators.

Where the gap is large

Three sectors regularly produce disagreement greater than 20%:

Where the gap is enormous

For some developing countries the gap can exceed 50% on the national total. Common reasons: incomplete UNFCCC reporting (some countries submit inventories only every 4-8 years, and the most recent submission may use 5-year-old activity data); large LULUCF uncertainty; methane from extensive livestock with no country-specific factor work; and — more rarely — political incentives to under-report. The reverse can also happen: country-specific inventories sometimes report more emissions than EDGAR because the country has better data on coal-mining methane or specific industrial-process emissions than the global model.

How to read the disagreement

For policy debates, the legal record is the UNFCCC inventory — that's what countries are accountable for under the Paris Agreement. For cross-country comparison and historical trend analysis, EDGAR is more useful because the methodology is consistent. For independent verification, Climate TRACE's satellite-derived estimates are increasingly the third leg of the stool.

PlainEmissions country pages render the spread across all four sources rather than picking one. When you see EDGAR and UNFCCC within 5% you can move on. When you see them 30% apart, the methodology page documents which sectors are driving the gap. That transparency is the entire point of multi-source reconciliation.

A worked example: Russia's emissions

Russia's UNFCCC inventory has historically reported substantially lower national totals than EDGAR's bottom-up model. The gap is driven by two factors. First, LULUCF: Russia claims very large forest-sink credits that bring the national net total down, while EDGAR's land-use model is more conservative. Second, methane from oil and gas extraction: Climate TRACE satellite observations have repeatedly found more methane in Russian producing regions than the inventory reports. The result is a national-total disagreement of roughly 15-25% depending on year and sector. Reading the Russia page on PlainEmissions you can see the EDGAR + UNFCCC + Climate TRACE figures side by side and form your own view.

The takeaway

Pick one number and you're picking a methodology. Look at all four and you see methodology disagreement as a first-class feature of the dataset. PlainEmissions exists precisely to make that disagreement visible rather than hide it inside a single picked-favorite figure.

Key takeaways

Further reading on PlainEmissions

Definitions used on this site