Satellite Methane Monitoring: How Space-Based MRV Is Auditing Corporate Emissions

For thirty years, corporate greenhouse gas reporting has rested on a quiet assumption: that the number a company calculates is the only number anyone has. Regulators, buyers and auditors could check the arithmetic and the activity data, but they could not independently check the result. That assumption is expiring.

The two ways to count emissions

Everything in this article turns on one distinction.

	Bottom-up	Top-down
Method	Activity data × emission factor	Inferred from measured atmospheric concentrations
Question answered	"How much should this activity emit?"	"How much was emitted, given what is in the air?"
Used by	Corporate inventories, BRSR, CDP, national reporting	Atmospheric science, regulators, increasingly NGOs and media
Fails when	Emission factors are wrong, activity data is incomplete, sources are unlisted	Spatial resolution is coarse, cloud cover blocks retrieval, sources are hard to separate
Typical output	A precise-looking number	A range with explicit uncertainty

Bottom-up is how virtually all reporting is produced, including everything filed under SEBI's BRSR framework. It is auditable and cheap. Its weakness is structural: if the emission factor is wrong, every number built on it is wrong in the same direction, and the inventory contains no mechanism for noticing.

Top-down inverts the problem. Measure methane in the atmosphere, model how air moves, and solve for the emissions that must have produced the observed pattern. Its weakness is attribution — the atmosphere does not label plumes with a company name.

The instruments doing this today

• TROPOMI (aboard ESA's Sentinel-5P, launched 2017) is the workhorse of global methane surveying — daily near-global coverage at roughly 7 km resolution, freely available. It cannot isolate a single factory, but it reliably flags regions and very large point sources.
• GOSAT (JAXA, launched 2009) has the longest continuous record of satellite column methane. Its 14-year archive is what made the Siberian trend analysis possible at all — you cannot detect a decadal trend with an instrument launched last year.
• GHGSat is a commercial constellation at roughly 25 m resolution, designed for facility-level attribution of leaks.
• MethaneSAT, developed by the Environmental Defense Fund, was built to sit between the two — regional coverage with enough resolution to quantify diffuse area sources rather than only large point plumes.
• Sentinel-2 and Landsat were not designed for methane at all, but researchers have shown that very large plumes can be detected in their shortwave infrared bands, extending the effective historical record backwards.

A worked example: how a top-down estimate is actually built

The 2026 Science study on Siberian methane is a clean illustration of the full method, and worth walking through because the same architecture is being applied to industrial sources.

1. Gather observations. Column methane retrievals from GOSAT, plus surface measurements from NOAA's global network — 217 locations in this case.
2. Start from a prior. Begin with a conventional bottom-up inventory: EDGAR for anthropogenic sources, GFED for biomass burning, WetCHARTs for wetlands.
3. Run the transport model. GEOS-Chem simulates how methane released at those assumed rates would disperse, and predicts what each instrument should observe.
4. Correct the prior. An Ensemble Kalman Filter adjusts the emissions until the model reproduces what was actually measured, weighting by the uncertainty of each observation.
5. Validate against held-back data. Check the result against observations not used in the inversion — TCCON ground spectrometers, flask samples from Japan Airlines aircraft, TROPOMI, and flux towers.

That last step is what separates a credible inversion from a curve fit. In the Siberian study, the correlation between model and independent Siberian surface observations rose from 0.23 to 0.72 after inversion, and model bias fell by 70%. The corrected emissions explain reality better than the inventory did.

Where the gap opened

The headline number for MRV practitioners: the study's eastern Siberian fire-methane estimate came out 6.7 to 41 times higher than GFED, FINN, GFAS and QFED — the standard global biomass-burning inventories.

The authors' explanation is instructive, because none of the three causes is exotic:

1. Emission factors resting on very few studies. The methane emission factor for boreal peat fire is derived from a small number of field and laboratory measurements. It was applied globally anyway, because it was the best available.
2. Uncertainty in the activity data. Satellite-derived burned area and fire radiative power are themselves estimates with error bars, and they feed directly into the inventory.
3. Missing sources entirely. The inventories count methane from combustion. They do not count methane subsequently released from permafrost that the fire thawed.

Every one of those failure modes has a direct corporate analogue: a default factor from a thin evidence base, activity data that is partly estimated, and an emission source nobody thought to include in the boundary.

What this means for Indian companies

India's methane profile is distinctive. Unlike the US or EU, where oil and gas dominates, India's methane comes principally from rice cultivation, livestock, landfill and wastewater, and coal mining — and the first two are diffuse area sources that are genuinely difficult to measure by any method. That cuts both ways: harder to verify from orbit, but also harder to defend when someone tries.

Practical exposure, roughly in order of how likely you are to be quantified externally:

• Landfill and wastewater operators. Large landfills are among the most reliably detectable methane sources from space, and Indian sites have already appeared in published global plume surveys. If you operate one, assume it is visible.
• Coal mining. Ventilation air methane from large operations is detectable and increasingly reported by third parties.
• Gas distribution and industrial users. Detectable at the larger end, and directly in scope of EU methane rules if you export into that supply chain.
• Agriculture and food supply chains. Not attributable to an individual farm, but regional emissions are quantifiable — which matters for companies making supply-chain claims about rice or livestock sourcing.
• Carbon project developers. Methane-avoidance projects face the same scrutiny in reverse: satellite data can support your claim as easily as it can undermine it.

What to do about it

1. Map your methane sources honestly. Before improving numbers, know where methane physically occurs in your operations and value chain. Many companies have never asked, because methane is a small line in a CO₂-dominated inventory.
2. Find out what your factors rest on. For every methane emission factor in your inventory, establish where it came from and how much evidence sits behind it. A factor traceable to a handful of measurements from another climate is a risk you are carrying without knowing.
3. Measure where it is material. For significant sources, replace defaults with site-specific measurement — flux chambers, continuous monitors, drone or aerial survey. Cost has fallen substantially; a landfill or large plant can now be surveyed for a fraction of what it cost five years ago.
4. Check yourself before someone else does. Public TROPOMI data and IMEO's plume notifications are freely available. Looking up your own sites is a cheap exercise with a high information value.
5. Build the discrepancy into your process. Decide now who investigates and how you respond when an external estimate contradicts your reported number. Improvising that under media or investor pressure goes badly.
6. Document everything. When your number is challenged, your defence is the audit trail: which factor, which source, which measurement, which date, reviewed by whom.

The direction of travel

Satellite MRV is not going to replace bottom-up inventories — attribution is too hard, and regulation is built around entity-level reporting. What it does is end the monopoly. Your reported figure is becoming one estimate among several, and it will increasingly be the one that has to justify itself against an independent measurement.

Companies that get ahead of this treat measurement as an investment in defensibility rather than a compliance cost. Those that do not will eventually find out what their real emissions are from someone else's press release.

Carbon Credit Consulting designs MRV systems that stand up to independent scrutiny, and builds greenhouse gas inventories on evidence rather than defaults. Explore our GHG accounting and carbon offset project development services, or talk to us about your methane exposure.