Climate TRACE: Satellite-Derived Emissions Estimates

Last updated: 2026-05-20 · PlainEmissions Editorial

What Climate TRACE is

Climate TRACE is an independent coalition that produces greenhouse-gas emissions estimates using satellite observations, sensor data, and machine learning. Unlike the other three sources tracked on PlainEmissions, Climate TRACE does not rely on self-reported activity data — it observes the atmosphere and the physical infrastructure directly and infers emissions from what it sees.

The project launched publicly in 2021 with country-level annual estimates. Subsequent releases have added quarterly updates and — most importantly — facility-level resolution for major point sources: individual power plants, refineries, steel mills, cement kilns, oil-and-gas fields, and shipping vessels.

How it works (at a high level)

Climate TRACE combines several measurement modalities:

• Direct satellite observation of CO2 and methane plumes. Instruments like TROPOMI on Sentinel-5P, Carbon Mapper, and GHGSat detect methane plumes from individual facilities with sufficient spatial resolution to attribute emissions to specific sites. CO2 column measurements (OCO-2, OCO-3, GOSAT) work at coarser resolution but capture the integrated atmospheric column.
• Inference from infrastructure observation. When direct plume detection isn't possible (cloud cover, sensor revisit gaps), machine-learning models trained on satellite imagery of facility characteristics (thermal infrared from power plants, gas-flare brightness, ship AIS tracks) predict emissions based on observed activity.
• Activity-data fusion. Where Climate TRACE has limited direct observation (some agricultural sources, dispersed urban transport), the team uses external datasets (IEA energy, FAO agriculture, AIS shipping) processed through their own emission-factor models.

What it does well

Climate TRACE's biggest strength is capturing emissions from large point sources. Power plants, oil-and-gas extraction sites, cement and steel plants, large agricultural feedlots, and refineries can be observed directly. For these sources, satellite-derived estimates are now often more accurate than bottom-up inventory estimates because the observation captures actual operational emissions rather than calculated ones.

The methane signal is the breakthrough. TROPOMI and Carbon Mapper have repeatedly identified "super-emitter" events — individual gas-pipeline leaks, well blowouts, and flaring events releasing thousands of tonnes of methane in days. These events are typically absent from official inventories because the operators don't report them, but they're visible from space.

What it does less well

Climate TRACE is weakest where Climate TRACE's inputs are weakest:

• Diffuse sources. Cars, residential heating, distributed agriculture — these emit from millions of small sources that no satellite can attribute individually. For these sources, Climate TRACE essentially falls back to activity-data-times-emission-factor estimates, similar in spirit to EDGAR.
• LULUCF. Carbon stocks in vegetation and soil are not directly satellite-observable in the same way as plumes. Climate TRACE LULUCF figures are less authoritative than dedicated forestry inventories.
• Historical depth. Many of the relevant satellites only came online in the 2010s. Climate TRACE's historical time series is shorter than EDGAR's (which reaches back to 1970) — useful for recent-year cross-checks, less useful for trend analysis over multiple decades.

Why it matters for inventory verification

Until Climate TRACE, the only independent check on a country's UNFCCC inventory was another bottom-up model (EDGAR). Both modelling approaches rely on activity data that is often supplied by the same country being checked. Satellite observation breaks that circularity: the atmosphere doesn't lie, and a methane plume over a Turkmenistan gas field appears in the data whether or not Turkmenistan reports it.

For the first time, this creates pressure for inventory improvement. Countries whose self-reported figures are substantially below satellite-observed atmospheric concentrations face uncomfortable scrutiny from journalists, ESG analysts, and international counterparts. Several recent UNFCCC inventory revisions explicitly cite satellite-observed methane as the driver.

Reading Climate TRACE figures on PlainEmissions

When a country page shows Climate TRACE substantially above EDGAR or UNFCCC, the most likely explanation is methane from oil-and-gas or unreported flaring. When Climate TRACE is substantially below, the most likely explanation is incomplete sensor coverage of some sector (typically agriculture or LULUCF). Our methodology page documents which sources have the strongest claim to authority for each sector × country × year — Climate TRACE is the strongest source for large-point-source methane, weaker for distributed transport.

The future

Satellite GHG monitoring is in rapid technical advance. NASA's GeoCarb (planned for the 2020s), the EU's Copernicus CO2M (2025+), and private satellites like Carbon Mapper's commercial constellation will dramatically improve facility-level resolution and revisit frequency. Within a decade, real-time monitoring of every major point source globally will likely be operational. Climate TRACE is the leading-edge of that shift, and PlainEmissions tracks it alongside the more conventional inventory sources.

Key takeaways

• The number is downstream of the methodology — which gas, which source, which framing.
• Multi-source disagreement is informative, not embarrassing — it surfaces uncertainty.
• Production-based versus consumption-based; LULUCF-included versus excluded; per-capita versus absolute — each framing answers a different question.
• Satellite-derived measurements (Climate TRACE) are progressively rebalancing the historical reliance on self-reported inventories.
• GWP100 is convention; GWP20 dramatically reweights short-lived gases like methane.

Further reading on PlainEmissions

• Understanding CO2-equivalent and GWP100 — the unit-conversion explainer.
• EDGAR vs UNFCCC — why two sources disagree.
• LULUCF: most-disputed sector — where source disagreement is largest.
• Climate TRACE explained — independent satellite verification.
• Reading emissions time series — six framing choices to watch.
• IPCC sector taxonomy — the canonical 8-sector hierarchy.
• Methodology page — full data-source provenance and harmonization steps.

Definitions used on this site

• CO2-equivalent (CO2e): any greenhouse gas expressed as the mass of CO2 that would produce equivalent warming over a chosen time horizon, typically 100 years.
• GWP100 / GWP20: global-warming potential over 100 (or 20) years; the multiplier used to convert from native gas units to CO2-equivalent.
• LULUCF: Land Use, Land-Use Change, and Forestry — the IPCC sector covering carbon stocks in vegetation and soils.
• Production-based emissions: emissions attributed to the country where they physically occur.
• Consumption-based emissions: emissions attributed to the country where the final goods or services are consumed.
• Annex I: the group of historically-developed countries under the UNFCCC with deeper reporting obligations.