Methane is a potent greenhouse gas responsible for approximately 25% of human-induced warming since the industrial era. With over 125 countries committing to a 30% reduction by 2030 under the Global Methane Pledge, tracking localized emissions is critical. NASA’s Earth Surface Mineral Dust Source Investigation (EMIT) instrument on the International Space Station, originally designed for mineral mapping, has been leveraged to detect methane.

Scientists use its advanced hyperspectral capabilities to record hundreds of light bands per pixel, allowing researchers to identify the unique chemical fingerprints of invisible gases. In the study "Global monitoring of methane point sources using deep learning on hyperspectral radiance measurements from EMIT," published in PNAS, researchers introduced MAPL-EMIT, a deep-learning framework that automates the detection, enhancement prediction, and source estimation of methane plumes globally. MAPL-EMIT achieves a high recall of 84% on expert-annotated plumes and demonstrates a high signal-to-noise ratio compared to existing methods.

To support the scientific community, the team released a global plume database on Earth Engine, the trained model on Kaggle, and an inference library on Github. Balancing field of view, spatial resolution, and spectral resolution remains the central challenge in space-based methane monitoring. While global mappers like TROPOMI focus on background changes, point source mappers like EMIT excel at facility-scale measurements through moderate coverage and very high spatial resolution.

This work aligns with Google Earth AI efforts to turn planetary data into actionable intelligence for climate mitigation.