Methane 3D Satellite Tracker
A 3D globe of satellite methane readings that you can scrub by week and filter by likely source. Every plume links to the readings and filters behind it.
Drag to turn · select a plume · illustrative render, example data
How it works
What it does, and what it does not
- 01
Reads public Level 2 methane columns from Sentinel-5P TROPOMI every day, regrids them onto a 0.1° grid, and removes cloudy, low-quality and over-water pixels.
- 02
Flags a cell when it stands above the median of its neighbours by a set multiple of the local spread, and shows a plume only when the flag repeats on two days within a week.
- 03
Tags each plume with the likely source type from public facility maps. The tag is a lead to check, not an attribution.
- Input
- Sentinel-5P TROPOMI L2 CH4
- Grid
- 0.1° daily
- Update
- daily, 06:00 UTC
- Flag rule
- 2 days in 7, above local median
- Coverage
- global land, ±70° lat
- Latency
- ~ 36 h after overpass
Known limits
- Finds unusual readings. It does not estimate emission rates or name a facility.
- Bright surfaces (deserts, snow) bias the retrieval upward; those regions are shaded on the globe.
- Coastal cells are less reliable: cloud masks fail more often there. Two plumes are marked threshold-sensitive after public review.
- Persistent tropical cloud leaves gaps that can last for weeks.
Version and changelog
v2.1.0, and how it got here
Semantic versions. Changes that came out of review name the review that asked for them.
- v2.1.02026-09-28Current
- AddedThreshold-sensitive marker on plumes that disappear when the cloud threshold is halved (from public review).
- AddedAttribution note beside every plume: a source tag is a lead, not a finding.
- FixedWeek scrubber skipped the last day of a month with 31 days.
- v2.0.02026-09-02
- ChangedBackground is now the median of neighbouring cells, not the mean (from community review).
- ChangedPlumes require a repeat flag on two days within a week.
- AddedSource-type filter and per-plume reading table.
- v1.3.02026-07-14
- AddedBright-surface regions shaded on the globe.
- FixedOver-water pixels leaked through the land mask near river deltas.
- v1.0.02026-05-20
- AddedFirst public release: daily regridding, flagging and the 3D globe.
Review record
Internal → Community → Public
Every stage, its dates and what reviewers said, including the requests for changes. Each comment names the version it read.
Internal review (Institute)
2026-06-02 → 2026-06-09Passed to community- Editorial BoardInstitute
Pipeline rebuilds from the repository on a clean machine and every input is public. Globe matches the regridded files for three sampled days.
✓ Accepted
Community review (Community)
2026-06-10 → 2026-09-043 accepted, after 2 changes- R. SetiawanReviewer · GIS
Grid, projection and land mask are consistent with the sources. Accepting.
✓ Accepted - T. AdityaReviewer · statistics
The background uses a mean, which the plume itself pulls up. A median over the same radius would be more robust.
Changes requested - L. ChandraReviewer · remote sensing
Median background and the two-day repeat rule fix the false plumes I saw in v1.3.0. Accepting.
✓ Accepted
Public review (Public)
2026-09-05 → 2026-09-27No open objections · Verified- N. RahmanPublic
The attribution caveat should sit next to every plume, not only in the paper.
Comment - P. WidodoPublic
Reran with the cloud threshold halved: two Java coast plumes disappear. Worth marking them on the globe.
Comment
Cite this
Cite the version you used
Results can change between versions. Cite the version and the proof ID so a reader can open exactly what you saw.
Santoso, D., Lestari, M., Halim, K. (2026). Methane 3D Satellite Tracker (v2.1.0) [Software]. Adaptive Institute. AI-PRF-0007. https://adaptive-institute.id/products/methane-3d-satellite-tracker