Google, NASA launch AI model to map methane emissions from space

Google and NASA’s Jet Propulsion Laboratory have launched MAPL-EMIT, an advanced artificial intelligence model designed to track global methane emissions from space. Unveiled in September, the deep learning tool monitors greenhouse gases using the International Space Station’s EMIT instrument, detecting thousands of hidden methane plumes worldwide.

Google and NASA introduced their new methane-tracking artificial intelligence model in a study published earlier this month in the scientific journal Proceedings of the National Academy of Sciences (PNAS), according to reporting by Wastedive. The collaboration aims to tackle what the agencies describe as a critical bottleneck in identifying the origins of greenhouse gas emissions by automating the detection and localization of methane point sources.

How the MAPL-EMIT AI Model Tracks Plumes From Space

The system, officially named MAPL-EMIT, monitors emissions by leveraging NASA’s Earth Surface Mineral Dust Source Investigation instrument. EMIT itself is an advanced imaging spectrometer mounted on the exterior of the International Space Station. To help the software cut through cluttered and complex landscapes, the researchers trained the deep learning framework on 3.6 million physics-simulated methane plumes—effectively teaching the algorithm to recognize clouds of the gas released into the atmosphere.

According to the published study, this training regimen allows the technology to automate enhancement quantification, plume segmentation, and source localization, including situations involving multiple overlapping plumes. By streamlining these processes, the system makes methane sources significantly easier to find at scale.

Uncovering Landfill Leaks and Surpassing Human Experts

The deployment of MAPL-EMIT has already yielded substantial detection results on a global scale. Researchers found that the artificial intelligence model can identify 50% more methane leaks than human experts are able to uncover on their own. In total, the tool has pinpointed over 23,000 hidden methane plumes around the world.

Among those discoveries are 24 of the 25 largest-emitting landfills in the world. This capability addresses a major environmental challenge, given that municipal solid waste landfills rank as the third-largest source of methane emissions generated by human activities in the United States, based on data from the Environmental Protection Agency cited by Wastedive.

Targeted Climate Mitigation and Corporate Superpollutant Strategy

Methane is classified as a superpollutant, meaning it warms the atmosphere far more intensely than carbon dioxide on a per-ton basis. The launch of MAPL-EMIT is intended to enable faster and more targeted climate mitigation efforts by giving regulators and researchers precise data on where leaks originate.

From Instagram — related to google nasa model methane, Google NASA methane AI

This space-based monitoring initiative extends Google’s broader corporate work to curb superpollutants. Earlier this year, the tech conglomerate entered into a partnership with AI-powered recycling and waste sorting company AMP. That recycling project is designed to generate carbon credits for Google, removing 200,000 tons of carbon dioxide by 2030 while simultaneously working to curb methane emissions generated by landfill waste.

Photo of author

Sophie Lin - Technology Editor

Sophie is a tech innovator and acclaimed tech writer recognized by the Online News Association. She translates the fast-paced world of technology, AI, and digital trends into compelling stories for readers of all backgrounds.

US, Denmark, and Greenland Reach Permanent Security Agreement

Leave a Comment

This site uses Akismet to reduce spam. Learn how your comment data is processed.