16:15 〜 16:30
[AAS05-10] Detection of methane emissions in satellite data at large scale using deep learning
★Invited Papers
キーワード：Methane, AI, Satellite
We show that deep learning can overcome the trade-off in terms of spectral resolution that comes with multi-spectral satellite data, resulting in a methane detection tool with global coverage and high temporal and spatial resolution.
Comparisons with airborne methane measurement campaigns suggest that our method can detect methane point sources in Sentinel-2 data down to plumes of 200 to 300 kg CH4 h−1 source, a one order magnitude improvement over the state of the art.
Our model provides a significant step towards the automated, high resolution detection of methane emissions at a global scale, every few days, and opens the possibility to bridge the gap between area mapping of methane fluxes (TROPOMI, MethaneSAT) and targeted point source imagers (GHGSat, airborne detections, etc.).
In addition to describing our method and how we assessed it, we show the results of systematically detecting methane sources over the Permian Basin, resulting in the largest methane point sources emissions catalogue to date.
Comparisons with airborne methane measurement campaigns suggest that our method can detect methane point sources in Sentinel-2 data down to plumes of 200 to 300 kg CH4 h−1 source, a one order magnitude improvement over the state of the art.
Our model provides a significant step towards the automated, high resolution detection of methane emissions at a global scale, every few days, and opens the possibility to bridge the gap between area mapping of methane fluxes (TROPOMI, MethaneSAT) and targeted point source imagers (GHGSat, airborne detections, etc.).
In addition to describing our method and how we assessed it, we show the results of systematically detecting methane sources over the Permian Basin, resulting in the largest methane point sources emissions catalogue to date.