The landfill is now being asked to report its own gas
the landfill is now being asked to report its own gas: a vision transformer called mapl emit reads nasa’s hyperspectral data from the space station and finds methane plumes at 24 of the world’s 25 biggest landfills, and it learned the job partly on 3.6 million physics simulated plumes.
the dump now files its own emissions ledger.
Context
Google Research's post of September 1, 2026 describes Methane Analysis and Plume Localization with EMIT (MAPL-EMIT), a deep-learning framework that detects, quantifies and localizes methane plumes from NASA's EMIT instrument on the International Space Station. It says the model was published in PNAS, and that EMIT has a 60 meter spatial resolution and an 80 km wide field of view.
The post says no global labeled dataset of real plumes exists, so the team built a physics-based simulation framework and injected 3.6 million synthetic plumes into real EMIT scenes for training. The arXiv version describes a vision transformer that uses the complete EMIT radiance spectrum. The post says the model mapped plumes at 24 of the world's 25 top-emitting landfills.
The 24 of 25 figure, the vision transformer, the ISS instrument and the 3.6 million simulated plumes match the sources. The sources say 'top-emitting landfills', where the note says 'biggest landfills', which is a different ranking since emission rate is not the same as size.
The note says the model learned 'partly' on simulated plumes, and that fits: the plumes were synthetic but injected into real EMIT backgrounds. The 24 of 25 comes from Google's own post. A search snippet from the paper's text mentions detection at 17 of the 25 top-emitting landfills, but which method that refers to was not seen, so any comparison with the 24 is unsupported, not refuted.
'The dump now files its own emissions ledger' is the author's framing. Detection of a plume is not the same as measuring annual emissions, though the model also quantifies enhancement.
Related work
- Mapping global methane emissions from space with deep learning (Google Research blog, September 1, 2026) ↗Primary source for MAPL-EMIT, the training approach and the landfill result.
- Global monitoring of methane point sources using deep learning on hyperspectral radiance measurements from EMIT (arXiv 2604.10094; PNAS) ↗Paper version with the vision transformer and synthetic plume training details.
Watch next
- Read the PNAS paper for how detection and quantification were validated. Check how detections at landfills compare with airborne measurements.
Sources
Provenance
The note above is reproduced unedited from the original post, first published on Threads on 3 October 2026 at 21:48 IST. Sources are the papers and datasets the note draws on.
View the original post ↗Embed this note
More notes
The air is now being asked to keep its own ledger
the air is now being asked to keep its own ledger: ecmwf’s aifs compo becomes the first ai model to forecast atmospheric composition globally every three hours, cleanair simulates 365 days of pm2.5 over china in ten seconds, and a unified framework maps six pollutants at one kilometer across the whole country. the air now files its own composition report.
read the note →The current is now being asked to draw its own map
the current is now being asked to draw its own map: china’s langya 2.0 predicts six ocean phenomena including internal waves and mesoscale eddies, a deep net called wenhai resolves eddies globally with air sea flux formulas built in, and scripps infers surface currents from the way temperature patterns deform in satellite images. the ocean now files its own circulation report.
read the note →The soil is now being asked to report its own carbon
the soil is now being asked to report its own carbon: a nix color sensor paired with generative data augmentation predicts soil organic carbon without a lab, random forest drives 74 percent of soil health mapping studies, and sentinel 2 tracks five year carbon change across france and italy from 922 samples. the dirt now files its own carbon account.
read the note →