California's Airborne AI Methane-Leak Detection — From AVIRIS-NG Plumes to Statewide Enforcement Data

The California Air Resources Board (CARB), with NASA's Jet Propulsion Laboratory, UC Riverside and Lawrence Berkeley National Laboratory, flew the AVIRIS-NG imaging spectrometer across roughly 22,000 square miles of California in 2016 and 2017, examining more than 272,000 facilities and components statewide. A trained neural network automatically detected methane plumes in the imagery, and the Geospatial Source Attribution Automated Model (GSAAM) — a hierarchical algorithm — then attributed each plume to a specific facility. Peer-reviewed results published in Environmental Research Letters report GSAAM's facility-attribution accuracy rose from 51.3% under a simple nearest-distance method to 99.6%.

CARB found that fewer than 0.2% of surveyed infrastructure was responsible for 34–46% of the state's total methane emissions, with landfills alone accounting for 41% of point-source emissions detected. The resulting Vista-CA dataset identified 15–18% more emitting facilities than existing EPA or CARB inventories. A named case followed up by CARB: an October 2016 leak at the Sunshine Canyon Landfill was detected by the airborne survey and confirmed remediated on a subsequent 2017 flight.

The program has since evolved into a satellite-based near-real-time detection pipeline that CARB credits with helping identify and stop large leaks, and its data now feeds CARB's oil-and-gas and landfill methane regulations. CARB itself cautions that no methane observing system can efficiently survey the entire state at high resolution, so coverage remains prioritized rather than comprehensive, and formal enforcement mechanisms tied directly to satellite detections were, per CARB's own documentation, still being developed for landfill rules as of the most recent reporting.

Read the full analysis: https://ww2.arb.ca.gov/our-work/programs/california-satellite-partnership/california-methane-surveys