Google and NASA turn to AI to hunt methane plumes from space
Google and NASA JPL just introduced MAPL-EMIT, a deep learning model tracking global methane emissions from space via NASA's EMIT instrument. Trained on 3.6 million simulated plumes, it detects 50% more plumes than human experts and identified over 23,000 extra plumes worldwide. The model is now open-sourced on Kaggle and GitHub.
MAPL-EMIT is an open-source deep learning model built by Google and NASA JPL that tracks global methane emissions from space using NASA's EMIT instrument.
How it works and what it finds
Spotting methane plumes against noisy terrain has long been a slow, manual chore. MAPL-EMIT bypasses that bottleneck by training on 3.6 million physics-simulated methane plumes. In practice, it detects 50% more plumes than human experts and identified over 23,000 additional plumes across the globe, successfully flagging 24 of the world's 25 largest-emitting landfills.
Google has published the full global plume database alongside an interactive visualization app on Earth Engine. The model weights are open-sourced on Kaggle, and inference tools are live on GitHub for anyone to run.
Why it matters
For climate researchers, geospatial developers, and environmental monitors, MAPL-EMIT delivers an out-of-the-box satellite detection pipeline without the headache of manual annotation.
Training on synthetic physics simulations is a smart way to beat the real-world data bottleneck in earth observation. It turns painstaking manual imagery analysis into an automated, planetary-scale system, giving operators a much faster way to pinpoint and plug massive leaks.
Sources (1)
- blog.google(opens in a new tab)Originalblog.google