Satellites Accurately Detect Methane Emissions in Largest Controlled Test Yet

Researchers have completed the largest-ever controlled test of space-based methane monitoring, published October 14 in Nature. This study found that dedicated methane sensors correctly detected nearly all methane plumes released on clear days, providing a benchmark for satellite systems increasingly used by regulators, environmental groups, and companies to identify leaks and super-emitters. Senior study author Adam Brandt, a professor at the Stanford Doerr School of Sustainability, noted that the results should instill confidence given past skepticism about methane tracking from space.

Brandt and Ph.D. student Frances Reuland, along with collaborators, timed 481 methane releases at a Stanford research site in Casa Grande, Arizona, to coincide with overpasses by 11 government, nonprofit, and commercial satellites throughout 2025. Fifteen research teams from nine countries independently analyzed the observations. Instruments specifically designed for methane monitoring achieved detection rates of 92%–100%, with most estimates being up to 50% higher or lower than actual amounts, though these discrepancies largely balanced out. Brandt emphasized that detection is often the most important part, allowing for follow-up fixes.

Under ideal conditions, these dedicated instruments reliably detected plumes at rates down to about 50–200 kilograms per hour, which is close to the 100 kilograms per hour threshold for some "super-emitter" programs. However, satellite systems, which detect methane using infrared light reflected off Earth's surface, struggled to detect plumes through clouds. This poses problems in key methane-emitting regions such as tropical oil and gas producers Nigeria and Venezuela, and cold, cloudy climates like Canada and Russia. Reuland expressed hope that these results would build confidence in the field overall.

Instruments repurposed for methane sensing, such as Germany's EnMAP, detected 62%–88% of plumes on clear days, while general-purpose land-scanning systems from U.S. and European government satellites detected only 29%–37% and were error-prone for releases below 1,000 kilograms per hour. Testing for false positives involved 265 overpasses with no methane released; these were rare except for emerging artificial intelligence approaches, with 79% of false positives originating from commercial teams applying AI to land-scanning data. Researchers stated that AI methods need improvement, with Reuland noting the study provides orders of magnitude more controlled-source data for testing methane detection from space. The releases, using compressed natural gas, ranged from 10 to 1,200 kilograms per hour, totaling less than 50 metric tons over the year.

What to watch: Further research on satellite system performance over non-desert terrains and various weather conditions.

Editor's note: Comprehensive report covering almost all concrete details and stats from the extensive source text.