Bertrand Rouet-Leduc: Large scale automatic detection of methane emissions in Sentinel 2 data using AI

Abstract

Curbing methane emissions is among the most effective actions that can be taken to slow down global warming. However, monitoring emissions remains challenging, as detection methods have a limited quantification completeness due to trade-offs that have to be made between coverage, resolution, and detection accuracy. In this talk, I will 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. We compare our detections with airborne methane measurement campaigns, which suggests that our method can reliably detect methane point sources in Sentinel-2 data down 200 to 300 kg/h. Our model shows an order of magnitude improvement over the state-of-the-art on the same data, providing a significant step towards the automated, high resolution detection of methane emissions at a global scale, every few days, down to the asset level. In addition to giving an overview of our approach and showing the results on evaluating our method against airborne detections and controlled releases, I will show a few use cases on oil and gas assets, mining operations and landfills, and talk about some of the next steps of our work.

Bio

Bertrand obtained his doctorate from the University of Cambridge, UK in 2018, in ML and materials science. Since then, Bertrand has been developing ML methods for geophysics at the Los Alamos National Laboratory (LANL), where he held a permanent staff position, at Kyoto University where he is an assistant professor, and at Geolabe where he is chief scientist officer. His research is at the interface between machine learning and geophysics, specifically bridging data science, remote sensing and geophysics. Bertrand is also Geophysical Journal International's editor for ML-related research.