Google and NASA’s Jet Propulsion Laboratory have developed an artificial intelligence model called MAPL-EMIT that tracks global methane emissions from space, according to an announcement from the tech giant. The tool monitors greenhouse gas leaks using NASA’s Earth Surface Mineral Dust Source Investigation instrument, an advanced imaging spectrometer attached to the International Space Station, as detailed in a study published in the Proceedings of the National Academy of Sciences.
How the MAPL-EMIT AI Model Detects Methane Leaks
The system tackles a critical bottleneck in finding methane sources by automating detection and localization through deep learning, according to the study. Researchers trained the model on 3.6 million physics-simulated methane plumes to help it see through complex, cluttered landscapes. Google and NASA state that this training allows the tool to detect 50% more methane leaks than human experts can uncover. The model has identified over 23,000 hidden methane plumes globally, including 24 of the 25 largest-emitting landfills in the world, according to the research findings.
Pro Tip: Researchers, policymakers, and operators can access open-source models on Kaggle and inference tools on GitHub, while the global plume database is available on Earth Engine with an accompanying visualization app, according to Google.
Comparing Global Mappers and Point Source Instruments
Measuring methane from space requires balancing field of view, spatial resolution, and spectral resolution, according to research published on the Google Research blog. Built to identify subtle shifts in background methane levels, global mappers like TROPOMI combine broad coverage—featuring a swath width of approximately 2,600 kilometers—with a coarse spatial resolution of roughly 5.5 kilometers by 3.5 kilometers and a precise spectral sampling of 0.1 nanometers.
In contrast, point source mappers like EMIT excel at measuring emissions at the facility scale by combining an 80-kilometer-wide field of view with a very high spatial resolution of 60 meters and a moderate spectral resolution of 7.4 nanometers. However, fully unlocking this data globally presents challenges because varied landscapes provide a complex backdrop where surface materials can masquerade as methane, making deep learning necessary to understand the broader visual context, according to Google.
Did You Know? As reported by the Environmental Protection Agency, municipal solid waste landfills stand as the third-largest contributor to human-caused methane emissions within the United States. Furthermore, methane has a warming potential 30 times greater than that of carbon dioxide over a 100-year timeframe, according to Google and NASA.
Broader Corporate Initiatives and Carbon Credits
The collaboration with NASA builds on Google’s broader work to manage and curb superpollutants that warm the atmosphere more than carbon dioxide on a per-ton basis. Earlier this year, Google announced a partnership with AI-powered recycling and waste sorting company AMP. That deal generates carbon credits for Google through its recycling project and enables the company to remove 200,000 tons of carbon dioxide by 2030, alongside curbing methane emissions generated by landfill waste, according to company statements.
Frequently Asked Questions
What is MAPL-EMIT?
MAPL-EMIT is an artificial intelligence model developed by Google and NASA’s Jet Propulsion Laboratory that tracks global methane emissions from space using NASA’s EMIT instrument on the International Space Station, according to the PNAS study.
How many methane plumes has the model found?
The AI model has uncovered over 23,000 hidden methane plumes globally, including 24 of the 25 largest-emitting landfills in the world, according to Google and NASA.
Where can researchers access the methane data?
Google has released the global plume database on Earth Engine alongside an Earth Engine app, with open-source models available on Kaggle and inference tools on GitHub, according to Google Research.
Explore more environmental technology stories or subscribe to our newsletter for the latest updates on space-based climate monitoring.
Related reading