Jump to Content

Mapping trace gases from space with machine learning

Applying machine learning to satellite observations to detect, quantify, and map greenhouse and trace gas emissions across scales — from broad-area monitoring from geostationary orbits to high-resolution point-source plume localization across diverse satellite instruments.

Explore methane plumes on Earth Engine Read the blog
variant_editorial_zoom_captioned (1)

Applying machine learning to satellite observations to detect, quantify, and map greenhouse and trace gas emissions across scales — from broad-area monitoring from geostationary orbits to high-resolution point-source plume localization across diverse satellite instruments.

Explore methane plumes on Earth Engine Read the blog

Multi-scale global monitoring from space

Greenhouse gas monitoring requires comprehensive, multi-scale Earth observation. Emissions operate across dramatically different spatial and temporal regimes: facility-level super-emitters (such as pipeline leaks, compressor blowouts, and landfill vents) require high-spatial-resolution pinpointing, whereas regional and continental carbon cycles demand high-cadence, contiguous monitoring to track diurnal flux and diffuse trends.

Google Research applies advanced deep learning techniques to diverse spaceborne instruments to bridge these observational gaps. In partnership with NASA's Jet Propulsion Laboratory (JPL), our MAPL-EMIT framework leverages vision transformers on imaging spectrometer data from the International Space Station to detect and localize point-source (emissions occurring from a small spatial footprint on the order of a few tens of meters) methane plumes globally with unprecedented sensitivity.

Simultaneously, our work on geostationary weather satellites — such as GOES-East ABI — demonstrates that physics-guided neural networks can extract contiguous, 10-minute dry-air column CO₂ (XCO₂) enhancements across entire continents.

Together, these advances contribute toward a multi-tier monitoring approach aimed at supporting mitigation efforts and independent emissions verification. This work aligns with Google’s broader effort behind Google Earth AI, our collection of geospatial models and datasets to turn planetary data into actionable intelligence. By applying deep learning to satellite imagery at scale, we aim to complement broader planetary AI initiatives with specialized tools for targeted environmental monitoring.

Featured publications

Global monitoring of methane point sources using deep learning on hyperspectral radiance measurements from EMIT
Vishal Batchu
Michelangelo Conserva
Alex Wilson
Anna Michalak
Varun Gulshan
Phil Brodrick
Andrew Thorpe
Christopher Van Arsdale
Proceedings of the National Academy of Sciences (2026)
Preview abstract Anthropogenic methane (CH4) point sources are critical drivers of near-term climate forcing, safety hazards, and system inefficiencies. Global tracking with imaging spectroscopy is just becoming feasible, but still largely relies on manual tracking. Here we present the Methane Analysis and Plume Localization with EMIT (MAPL-EMIT) model, an end-to-end vision transformer framework that advances the state of the practice by directly utilizing the complete radiance spectrum from the Earth Surface Mineral Dust Source Investigation (EMIT) instrument to jointly retrieve methane enhancements across all pixels within a scene. This approach joins spectral information content - where the methane signature resides - with essential spatial context to significantly lower detection limits. MAPL-EMIT simultaneously supports quantification, plume delineation, and source localization, even for multiple overlapping plumes. The model was trained on 3.6 million physics-based synthetic plumes injected into global EMIT radiance data. Evaluation against synthetic observations confirms the model’s ability to identify plumes with high recall and precision and to capture weaker plumes relative to existing matched-filter approaches. On real-world benchmarks, MAPL-EMIT captures 79% of known hand-annotated NASA L2B plume complexes across a test set of 1084 EMIT granules, while identifying twice as many plausible plumes than identified by human analysts. Further validation against coincident airborne data, top-emitting landfills, and controlled release experiments confirms the models efficacy at identifying previously uncaptured sources. By incorporating model-generated metrics such as spectral fit scores and estimated noise levels, the framework can further limit false-positive rates. Overall, MAPL-EMIT enables high-throughput implementation on the full EMIT data catalog, shifting methane monitoring from labor-intensive workflows to a rapid, scalable paradigm for facility-level accountability. View details
Quantification of atmospheric carbon dioxide from the Geostationary Operational Environmental Satellite (GOES East)
Aarón Sonabend
Sean Campbell
Christopher Van Arsdale
Anna Michalak
arXiv (2026)
Preview abstract The need for characterizing global variability of atmospheric carbon dioxide (CO2) is quickly increasing, with a growing urgency for tracking greenhouse gasses with sufficient resolution, precision and accuracy so as to support independent verification of CO2 fluxes at local to global scales. The current generation of space-based sensors, however, can only provide sparse observations in space and/or in time, by design. While upcoming missions may address some of these challenges, most are still years away from launch. This challenge has fueled interest in the potential use of data from existing missions originally developed for other applications for inferring global greenhouse gas variability. The Advanced Baseline Imager (ABI) onboard the Geostationary Operational Environmental Satellite (GOES-East), operational since 2017, provides full coverage of much of the western hemisphere at 10-minute intervals from geostationary orbit at 16 wavelengths at an approximately 2km2 spatial resolution. Here, we leverage this high temporal resolution by developing a single-pixel, fully connected neural network to estimate dry-air column CO2 mole fractions (XCO2). The model employs a time series of GOES-East's 16 spectral bands, ECMWF ERA5 lower tropospheric meteorology, MODIS surface reflectance, solar angles, and day of year. Training used collocated GOES-East and OCO-2/OCO-3 observations. The model is able to capture realistic XCO2 variability when compared against held-out years of OCO-2 observations and against observations from the TCCON network. We also present case studies illustrating the use of the model to observe XCO2 enhancements over urban areas, drawdown over agricultural regions, and enhanced ability to observe regions with persistent cloud cover thanks to GOES-East’s full spatial coverage and high observational frequency. Overall, although GOES-East derived XCO2 precision cannot rival that of dedicated instruments, its unprecedented combination of contiguous geographic coverage, 10-minute temporal frequency, and multi-year record offers the potential to observe aspects of atmospheric CO2 variability currently unseen from space. View details
×