Sentinel-5P TROPOMI methane column mapping
TROPOMI's shortwave infrared channels map the total atmospheric methane column daily at 7-by-5.5 km resolution, flagging anomalies above oil and gas fields, landfills and wetlands. Large point sources are detectable; smaller facilities require commercial sensors to confirm.
Sensors
- Sentinel-5P TROPOMI: Shortwave infrared channels at 2305 to 2385 nm retrieve total-column methane (XCH4) at 7 by 5.5 km per pixel (improved from the original 7 by 3.5 km after a 2019 processor update). Daily global coverage. Near-real-time Level-2 products available within three hours of sensing.
- GHGSat: Commercial Fabry-Perot spectrometers on a growing constellation of small satellites. Pixel size roughly 25 by 25 metres; resolves individual facility-level plumes down to approximately 100 kg/hour emission rate. Tasked on demand, not continuous.
- GOSAT-2: JAXA heritage sensor with Fourier-transform spectrometry across shortwave and thermal infrared. Coarser spatial footprint (~10 km diameter) but precise column retrievals used to cross-validate TROPOMI bias corrections.
- MethaneSAT: EDF-backed satellite launched March 2024. Targets oil and gas basins at roughly 100 by 400 metre pixels with wide swath (~200 km). Designed to quantify regional emission budgets and attribute them to specific infrastructure clusters.
What the shortwave infrared actually measures
TROPOMI does not image methane directly. It measures sunlight that has passed down through the atmosphere, reflected off the surface, and travelled back up again. Methane absorbs specific wavelengths in the 2305 to 2385 nm shortwave infrared band. The depth of those absorption features, once corrected for surface albedo and atmospheric scattering, yields the total column-averaged dry-air mole fraction of methane, written as XCH4 and expressed in parts per billion.
Two retrieval algorithms are in operational use. WFMD (Weighting Function Modified Differential Optical Absorption Spectroscopy) is computationally fast and well-suited to large-area screening. RemoTeC uses a full-physics radiative-transfer inversion and handles aerosol scattering more carefully, at the cost of processing time. Both are applied to the same Level-1B radiances; comparing their outputs is a standard quality check. Neither retrieves a vertical profile. The result is a single column-integrated number, which means a surface source and an upper-tropospheric anomaly produce the same signal. Separating them requires ancillary meteorological data or coincident profile measurements.
What a 7-kilometre pixel can and cannot resolve
The pixel footprint is the central practical constraint. At 7 by 5.5 km, TROPOMI can detect the integrated column enhancement from a large, persistent source: a major gas processing plant, a coal mine ventilation shaft complex, a large landfill, or a flooded rice-paddy district. Published studies using TROPOMI data have identified XCH4 anomalies exceeding 20 to 50 parts per billion above background over known super-emitter sites in the Permian Basin, Turkmenistan and the Hassi Messaoud field in Algeria. Those detections are real and reproducible.
Smaller facilities, a single well pad, a modest municipal landfill, a cattle feedlot, sit well below the detection floor. A source emitting a few hundred kilograms of methane per hour spread across a 38.5 square kilometre pixel will typically be lost in retrieval noise, which runs to roughly 1 percent of the background column (around 18 ppb under good conditions). This is not a failure of the instrument; it is physics. The honest answer is that TROPOMI provides a basin-scale or regional budget. Attribution to a specific facility requires a pointed, high-resolution sensor such as GHGSat or MethaneSAT tasked on the anomaly coordinates.
Cloud, albedo and the tropical data gap
Methane retrievals require reflected sunlight, which means cloud is a hard blocker. TROPOMI applies a cloud-fraction filter, typically discarding pixels with cloud radiance fraction above 0.5 in the SWIR. In persistently cloudy regions, particularly the tropics and high latitudes in winter, this removes a substantial share of daily observations. Over the Congo Basin or the Indonesian archipelago, where wetland and agricultural methane sources are scientifically important, cloud cover can render a week or more of passes unusable.
Surface albedo matters too. Dark surfaces (dense forest, open water) return little reflected light, degrading signal-to-noise. Bright surfaces (desert, snow) are preferable. This introduces a geographic sampling bias: arid oil and gas regions in the Middle East and Central Asia are well-observed; tropical peatlands and rice systems are systematically under-sampled. Any regional budget that draws on TROPOMI data without accounting for this gap will underestimate tropical source contributions.
Bias correction and the role of TCCON
Raw TROPOMI XCH4 retrievals carry systematic biases that vary with viewing geometry, surface type and season. The standard correction procedure compares satellite retrievals against the Total Carbon Column Observing Network, a ground-based network of Fourier-transform spectrometers at roughly 25 stations worldwide. TCCON instruments measure the same column quantity with sub-percent precision and traceability to the WMO methane scale.
The comparison reveals biases typically in the range of a few parts per billion, which sounds small against a background of around 1900 ppb but matters when the anomaly signal of interest is only 20 to 30 ppb. Bias correction coefficients are derived per-algorithm and updated as the TROPOMI calibration drifts over the mission lifetime. Users working with the official ESA Level-2 product (version 2.x onwards) receive a bias-corrected XCH4 field, but should still check whether the TCCON stations used for correction are geographically representative of their region of interest. A correction derived from European and North American stations may not transfer cleanly to Central Asian or equatorial scenes.
Attributing an anomaly: source types and their signatures
Once an XCH4 enhancement is identified, the next question is what produced it. Three broad source categories dominate anthropogenic signals. Oil and gas infrastructure tends to produce persistent, spatially compact anomalies that track with production activity. Landfills produce steady, year-round signals with little diurnal variation. Wetlands and rice paddies produce seasonally modulated signals that peak during inundation periods and correlate with surface water extent derived from SAR or optical sensors.
Distinguishing these in a TROPOMI time series requires combining the methane column data with wind fields (typically from ERA5 reanalysis), land-cover maps, and ideally a coincident NO2 or SO2 column from the same instrument. A methane anomaly co-located with elevated NO2 and a known refinery is almost certainly fossil-fuel derived. The same anomaly over a river delta with no industrial infrastructure points to wetlands. The attribution is probabilistic, not certain, and the 7 km pixel means the inferred source location carries a positional uncertainty of several kilometres even after wind-back trajectory analysis. For regulatory or enforcement purposes, that uncertainty is too large; commercial high-resolution follow-up is the standard next step.
Pairing TROPOMI with commercial sensors for operational monitoring
The practical workflow is two-stage. TROPOMI provides continuous, free, global screening. When it flags a persistent or sudden column anomaly, a commercial sensor is tasked to the coordinates to resolve the source at facility scale. GHGSat has published detection limits of roughly 100 kg/hour for a single overpass; MethaneSAT targets lower thresholds across wider areas. Neither replaces TROPOMI's global daily cadence; TROPOMI cannot be replaced by either of them for basin-scale budget work.
Satellize runs this kind of layered analysis on open TROPOMI products and can add commercial tasking on client licence. The workflow is well-suited to national regulators who need to demonstrate emissions accounting against NDC commitments, or to asset owners who want independent verification of reported figures. The Sentinel-5P archive extends back to May 2018, giving more than six years of baseline against which current anomalies can be assessed. That depth is genuinely useful: a facility that has been a persistent emitter for years will show up clearly against its own history.
Typical figures
| Spatial resolution (pixel footprint) | 7 × 5.5 km (post-2019 processor update; original 7 × 3.5 km) |
| Swath width | 2600 km, enabling daily global coverage |
| Revisit / temporal cadence | Daily global; equatorial overpass ~13:30 local solar time |
| Spectral band for CH4 retrieval | Shortwave infrared, 2305 to 2385 nm (SWIR Band 7) |
| Retrieval algorithms | WFMD (fast screening) and RemoTeC (full-physics); both applied to Level-1B radiances |
| Retrieval precision (single pixel, clear sky) | Approximately 0.6 to 1% of background column (~11 to 19 ppb at ~1900 ppb background) |
| Cloud filter threshold | Cloud radiance fraction ≤ 0.5 in SWIR; higher cloud fraction pixels discarded |
| Bias correction reference network | TCCON (~25 ground stations); correction applied in ESA Level-2 v2.x product |
| Archive depth | May 2018 to present (Sentinel-5P launched October 2017) |
| Data access and delivery format | Free via Copernicus Data Space Ecosystem; NetCDF-4; Level-2 near-real-time within ~3 hours of sensing |
Analytics Satellize can run
| Basin-scale XCH4 anomaly map | TROPOMI Level-2 XCH4 compositing with ERA5 wind-back trajectory filtering to reduce transport ambiguity | Monthly GeoTIFF or GIS polygon layer showing XCH4 enhancement above seasonal background, flagged by source-type probability |
| Persistent super-emitter alert | Time-series change detection on TROPOMI daily passes; threshold exceedance over rolling 30-day baseline | Near-real-time alert (within 24 hours of overpass) with coordinates, estimated column enhancement and confidence flag |
| Seasonal emission budget estimate | Mass-balance inversion using TROPOMI XCH4 columns and ERA5 wind fields; WFMD and RemoTeC cross-validated retrievals | Quarterly PDF report with regional emission totals in tonnes CH4, uncertainty ranges and TCCON-corrected bias assessment |
| Source attribution classification | Co-analysis of TROPOMI CH4, NO2 and SO2 columns with land-cover and infrastructure databases; Bayesian source-type classification | GIS layer with per-anomaly source-type probability (oil and gas / landfill / wetland / agriculture) and supporting metadata |
| Commercial follow-up targeting package | TROPOMI anomaly coordinates combined with wind-back plume modelling to define optimal tasking window and look angle for GHGSat or MethaneSAT | Tasking brief with coordinates, priority ranking, recommended overpass window and expected detection probability |
| Multi-year trend analysis | Linear regression and Mann-Kendall trend test on TROPOMI archive from 2018 to present, cloud-gap-filled using temporal interpolation | Annual trend report showing per-basin XCH4 trajectory with confidence intervals, suitable for NDC reporting or investor disclosure |
Who does the work
We can get this done for you. Satellize runs its own analyst desk and a strong science team. You do not buy a data feed and work out what it means; our people source the imagery, run the analysis described on this page, and hand you the answer with its confidence limits stated. Discuss this requirement.