Google and NASA map methane emissions from space

Google and NASA have released the MAPL-EMIT model, an AI system trained to detect methane emissions from satellite imagery. The model identifies 50% more leaks than human experts reviewing the same data.

This matters because methane is a potent greenhouse gas – roughly 28 times more effective at trapping heat than CO₂ over a century – yet remains difficult to quantify at scale. Oil and gas operators, landfills, agricultural facilities, and coal mines all emit significant volumes, but pinpointing and measuring them has relied on ground surveys, aerial observation, and self-reporting. Those methods miss leaks and favour verification of disclosed sources.

The satellite-based approach scales differently. Space-based detection covers geography no ground team can feasibly traverse. It creates a continuous record. It allows third-party verification of corporate emissions claims without site access. For companies reporting Scope 1 and Scope 3 emissions under GHG Protocol or SBTi frameworks, this introduces a transparency layer that doesn't depend on operator cooperation.

Google and NASA published their findings in a scientific journal, suggesting peer review preceded release. That's a meaningful signal – it differentiates this from a one-off pilot or marketing exercise.

The deployment question now: who gets access? If this becomes public infrastructure, it reshapes corporate emissions accounting and greenwashing detection. If it remains proprietary or restricted to paying customers, its impact narrows to organisations that can afford it. The tech works. The question is whether it becomes a commons or a competitive advantage.

Google, NASA launch AI model to map methane emissions from space

Report This Content