Cross-checking a satellite-flagged methane plume against Subpart W reporting
An alert lands in your inbox: a satellite picked up a methane plume somewhere near a facility you track. Your first instinct is to pull the GHGRP Subpart W submission for that facility and see if the numbers line up. They won't, not directly, and knowing why is most of the job.
Start with the plume, not the inventory
Before you open EPA's Envirofacts or FLIGHT data, nail down what the detection tells you. A plume flag gives you a location, a timestamp, and usually a mass emission rate estimate in kilograms per hour. That's an instantaneous or short-window reading. Subpart W, by contrast, is an annual facility-level total, built mostly from engineering calculations and emission factors applied to equipment counts, not from a sensor watching the stack that day.
A detection's location has to match a facility boundary you can identify before any of the rest of this works. If the plume map only gives you a basin or county, you're stuck cross-referencing well pad locations, gathering compressor stations, and processing plants by hand, hoping you land on the right GHGRP ID. A detection that resolves to an actual site footprint, not just a region, is what makes the next steps possible at all.
Matching the detection window to a Subpart W reporting year
Once you have a facility, check the reporting lag. Operators submit Subpart W data by March 31 for the prior calendar year, and EPA typically publishes it that October. If your plume was flagged last week, there's no current-year GHGRP record to check it against yet. You're comparing against last year's submission, or the one before, depending on how current the published data is.
That gap matters more than it sounds like it should. A facility's pneumatic controller population, its flare efficiency assumptions, its equipment leak survey results, all of that is baked into a number that describes a prior year, not the day your satellite passed overhead. If the site added compressors or changed out high-bleed controllers since that filing, the reported number is already stale relative to what's in the ground now.
Where the two numbers won't line up, and that's fine
Even with the facility and the year matched correctly, don't expect the plume mass rate to reconcile cleanly against the annual tonnage in Subpart W. The units don't translate without a lot of assumptions about duration, frequency, and whether the detected event is representative of normal operations or a one-off. A single large detection, especially one that falls into EPA's "other large emission events" category under the revised Subpart W rules, is supposed to get reported separately from routine equipment-based estimates. If the facility didn't report an event like that for the relevant period and your plume suggests something well above background, that's the discrepancy worth flagging.
The useful comparison isn't "satellite kg/hr equals reported metric tons CO2e." It's whether the pattern of detections at this site over months looks consistent with what the facility told EPA its major emission sources are. Repeated plumes at a site that reported minimal flaring and no large release events is a pattern worth a closer look, not a single apples-to-oranges subtraction.
Building an audit trail for the discrepancy
Document the facility ID, the GHGRP reporting year you checked against, the detection date and estimated rate, and the specific source category you'd expect it to fall under (equipment leaks, pneumatics, liquids unloading, a large release event). If you're going to raise this with an operator, an auditor, or a fund's ESG desk, that chain needs to hold up without someone having to redo your work.
This whole process gets a lot faster when the plume detection already resolves to a site boundary instead of a basin-wide bulletin. Methane Plume Map is built for that step specifically: a per-site plume map updated daily where tasking allows, so you know which pad or facility to pull the Subpart W record for in the first place.
If facility-level reconciliation is part of your regular MRV workflow, take a look at what a daily per-site plume map can add to it.