← Innovations

Artificial intelligence

Climate TRACE — seeing emissions

An oil refinery at dusk — the class of industrial site emissions inventories try to measure from space
Image: Trevor Harris ( CC BY-SA 2.0 )
Artificial intelligence Deployed

What it is

Climate TRACE is a coalition that publishes a global inventory of greenhouse-gas emissions built from satellites, other remote sensing, and machine learning, rather than from self-reported totals alone. Related public systems include GHGSat (commercial methane plumes) and Carbon Mapper (public-good methane and CO₂). The point is independent seeing, not a new furnace.

Problem it targets

Inventories disagree. Methane leaks from oil, gas, and coal are often higher than reported. You cannot manage what you refuse to measure.

How it works

Models learn the look of a power plant, a feedlot, or a flare and estimate activity and emission factors, then fuse that with atmospheric retrievals where they exist. Every number has uncertainty. A bright pixel is a clue, not a court verdict.

Status and players

Climate TRACE coalition (non-profits, companies, universities); GHGSat; Carbon Mapper (including NASA/JPL partnerships); IMEO (UNEP) methane science. National inventories remain the legal baseline.

Risks and limits

False plumes, political blowback, and treating a dashboard as enforcement. Satellites see some sectors better than others (a dairy lagoon ≠ a city of two-stroke bikes).

Sources

Climate TRACE methodology and inventory releases; GHGSat and Carbon Mapper public notes; UNEP IMEO methane reports.