Featured Products

Source Cooperative is a Radiant Earth project

methaneset · Cesar Aybar · Source Cooperative | Source Cooperative

methaneset

Product Details

Visibility: Public
Owner: Cesar Aybar
Created: 16 Dec 2025
Last Updated: 16 Dec 2025

Product Contents

root

l89

README

MethaneSET-L89: Global Landsat 8/9 Dataset for Methane Emission Detection

methaneset-l89 is a Landsat 8/9 subset of the MARS-S2L dataset, accessed on December 5, 2025. MARS-S2L is an AI-driven operational methane emitter monitoring system for Sentinel-2 and Landsat imagery, developed and maintained as a live system by the United Nations Environment Programme's International Methane Emissions Observatory (UNEP-IMEO). The dataset compiles a large, expert-verified global collection of methane emission events spanning January 2018 to December 2024. The system uses a U-Net architecture to produce per-pixel probabilistic plume masks and scene-level detection scores, enabling near-real-time methane monitoring for climate mitigation. This subset provides Landsat 8 (OLI) and Landsat 9 (OLI-2) imagery with TOA reflectance bands B02, B03, B04, B05, B06, B07, including manually verified methane plumes across distinct emitter sites globally.

Note: As MARS-S2L is an actively maintained operational system, this snapshot represents the dataset state as of December 2025. For the most current data, refer to the official UNEP-IMEO MARS repository at https://huggingface.co/datasets/UNEP-IMEO/MARS-S2L

Dataset Information

Version: 0.1.0

License: CC-BY-NC-SA-4.0

Keywords: methane, plume-detection, remote-sensing, Landsat-8, Landsat-9, OLI, CloudSEN12, ERA5-Land, GEOS-FP, MBMP, PlumeViewer, GHG, earth-observation, deep-learning

Tasks: segmentation, regression

Dataset Overview

Partitions: 37 files Spatial coverage: [-121.91, -50.75, 151.42, 51.35] (WGS84) Temporal coverage: 2018-01-05 to 2024-12-31

Dataset Structure

Root: FOLDER (23,474 samples)

Hierarchy:

Level 1: FILE → FILE → FILE → FILE → FOLDER (117,370 samples)
Level 2: FILE (23,474 samples)

Metadata Fields

LEVEL0

Field	Type	Description
`id`	`string`	Unique sample identifier within parent scope. Must be unique among siblings.
`type`	`string`	Sample type discriminator (FILE or FOLDER).
`stac:crs`	`string`	Coordinate reference system (WKT2, EPSG, or PROJ string)
`stac:tensor_shape`	`list<item: int64>`	Raster dimensions e.g. [bands, height, width]
`stac:geotransform`	`list<item: double>`	GDAL affine transform [origin_x, pixel_width, rot_x, origin_y, rot_y, pixel_height]
`stac:time_start`	`timestamp[us]`	Acquisition start timestamp (microseconds since Unix epoch, UTC)
`stac:centroid`	`binary`	Raster center point in EPSG:4326 (WKB binary format)

LEVEL1

Field	Type	Description
`id`	`string`	Unique sample identifier within parent scope. Must be unique among siblings.
`type`	`string`	Sample type discriminator (FILE or FOLDER).
`geotiff:stats`	`list<item: list<item: float>>`	Per-band statistics (List[List[Float32]]): categorical mode returns class probabilities, continuous mode returns [min, max, mean, std, valid%, p25, p50, p75, p95]
`taco:header`	`binary`	Binary TACOTIFF header (35 bytes + tile counts) for fast reading without IFD parsing
`internal:parent_id`	`int64`	Foreign key referencing parent sample position in previous level (ZIP, FOLDER, TACOCAT).
`internal:relative_path`	`string`	Relative path from DATA/ directory. Format: {parent_path}/{id} or {id} for level0 (ZIP, FOLDER, TACOCAT).

LEVEL2

Field	Type	Description
`id`	`string`	Unique sample identifier within parent scope. Must be unique among siblings.
`type`	`string`	Sample type discriminator (FILE or FOLDER).
`plume:geometry`	`binary`	Plume geometry in WKB (Well-Known Binary) format
`internal:parent_id`	`int64`	Foreign key referencing parent sample position in previous level (ZIP, FOLDER, TACOCAT).
`internal:relative_path`	`string`	Relative path from DATA/ directory. Format: {parent_path}/{id} or {id} for level0 (ZIP, FOLDER, TACOCAT).

Usage

Python

R

Julia

Data Providers

UNEP International Methane Emissions Observatory (IMEO) - producer
Research Institute of Water and Environmental Engineering (IIAMA), UPV - processor
Department of Engineering, University of Cambridge - contributor
Vector Institute, University of Toronto - contributor

Dataset Curators

Anna Vaughan (UNEP IMEO) - anna.vaughan@un.org
Gonzalo Mateo-Garcia (UNEP IMEO) - gonzalo.mateogarcia@un.org
Claudio Cifarelli (UNEP IMEO) - claudio.cifarelli@un.org
Itziar Irakulis-Loitxate (IIAMA, Universitat Politècnica de València) - itiloi@upv.es

Publications & Citations

If you use this dataset in your research, please cite:

DOI: 10.48550/arXiv.2408.04745

Vaughan, A., Mateo-Garcia, G., Irakulis-Loitxate, I., Watine, M., Fernandez-Poblaciones, P., Turner, R. E., Requeima, J., Gorroño, J., Randles, C., Caltagirone, M., & Cifarelli, C. (2024). AI for operational methane emitter monitoring from space. arXiv preprint arXiv:2408.04745.

Introduces MARS-S2L system with 6 months operational results: 457 detections and 62 formal notifications across 22 countries.

DOI: 10.48550/arXiv.2511.21777

Allen, A., Mateo-Garcia, G., Irakulis-Loitxate, I., Montesino-San Martin, M., Watine, M., Requeima, J., Gorroño, J., Randles, C., Mokalled, T., Guanter, L., Turner, R. E., Cifarelli, C., & Caltagirone, M. (2024). Artificial intelligence for methane detection: from continuous monitoring to verified mitigation. arXiv preprint arXiv:2511.21777.

Extended operational deployment demonstrating 1,015 stakeholder notifications across 20 countries and verified permanent mitigation of six persistent emitters.

BibTeX

Generated with TacoToolbox