Nature paper reports a one-year blind test of satellite methane detection over 746 events

A paper in Nature reports a year-long single-blind test of how well satellites detect and measure methane leaks.
The test covered 746 events: 481 controlled methane releases ranging from 9.5 to 1,183.7 kg/h and 265 blank controls.
The releases were made in Arizona and timed with overpasses from 12 government, commercial and non-profit satellites, and 15 analysis teams from nine countries produced over 2,250 estimates.
Instruments built for the purpose achieved a 92 to 100 per cent true positive rate with a 0 to 11 per cent false positive rate, while repurposed instruments detected 62 to 88 per cent of releases.
The authors note that 79 per cent of false positives came from methods based on artificial intelligence, and that ground-measured cloud cover above 20 per cent degraded performance across systems.

Events tested: 746 Controlled releases: 481, from 9.5 to 1,183.7 kg/h Satellites involved: 12 Analysis teams: 15, from nine countries
Why it matters for exams: Methane monitoring from space connects climate policy with remote sensing, a favourite cross-over area.
UPSC
15 October 2026 Source: Nature

In plain words

Single-blind test Single-blind test — A test in which the analysts do not know the true answer they are being measured against.

False positive False positive — A reported detection where in fact nothing was released.