Levels 1 to 3 are desk work on published factors. Level 4 and 5 need physical measurement, and the reconciliation between them is the part most programmes underestimate.
There is a clean dividing line running through OGMP 2.0 reporting, and a great deal of confusion comes from ignoring it.
Everything up to Level 3 is desk work. Published emission factors, equipment counts, throughput data. No one goes into the field.
Level 4 and Level 5 require physical measurement. That is a different trade, a different cost base and a different timeline. Blurring the two in a plan or a proposal is how programmes end up underfunded and behind.
Level 4: source-level quantification
You must quantify emissions by detailed source type using company-specific methods. In practice that means a component-level campaign across the asset.
| Purpose | Technique |
|---|---|
| Find leaks | Optical gas imaging survey. Detection only, it does not give a rate |
| Quantify what you found | High-flow sampler, calibrated bagging, acoustic quantification on valves |
| Component inventory | Count and classify valves, flanges, connectors, seals, then apply asset-specific factors |
| Vents and tanks | Engineering calculation from process data, plus measurement on significant sources |
| Flares | Combustion efficiency measurement and unlit-flare detection |
The row worth pausing on is the first. An optical gas imaging camera finds leaks; it does not measure them. It makes an invisible plume visible on a screen. Turning that into a number requires a second instrument and a second visit to the component.
Programmes that budget for an OGI survey and assume quantification is included discover the gap late, usually after the survey has produced a list of findings nobody can put a figure against.
Flares are the biggest uncertainty
A flare burning efficiently converts methane to CO2. A flare burning badly, or one that has gone out and is venting raw associated gas, emits methane directly. Methane is roughly eighty times more potent than CO2 over twenty years.
So the difference between a flare assumed to run at 98% combustion efficiency and one actually running at 92%, or unlit, changes an asset’s methane number by an order of magnitude.
Most published estimates rest on assumed efficiencies. Very few rest on measured ones. In flaring-heavy jurisdictions this is the single largest source of uncertainty in any inventory, and whoever measures it first holds the defensible number.
Level 5: site-level measurement and reconciliation
Level 5 requires independent measurement of the whole site, on a representative sample of facilities, reconciled against the bottom-up Level 4 inventory with uncertainty analysis on both.
Site-level techniques, roughly in order of cost:
- Mobile ground survey, vehicle-mounted analyser with inverse dispersion modelling
- Drone-mounted sensors, good for congested facilities and flare stacks
- Continuous point sensors or tower networks, the only option that reliably catches intermittency
- Aircraft campaigns, the standard for basin and large-site quantification
- Satellite, both public and commercial, useful for super-emitters and for cross-checking a national inventory, but too coarse for site attribution on its own
One technique done well generally beats a menu of several done shallowly.
The reconciliation is the hard part
You now have two numbers for the same asset, produced by entirely different methods. They will not agree.
Closing that gap credibly, and characterising what remains, is a statistical exercise governed by OGMP guidance on uncertainty and emissions reconciliation. It needs someone comfortable with measurement uncertainty rather than a report writer.
This is the step most programmes underestimate, and it is where a Level 5 claim either survives verification or does not. Two numbers that disagree are not a finding. They are a starting point.
The intermittency problem sits here too. A mobile survey is a snapshot. If a significant emission source operates intermittently, a snapshot may miss it entirely or catch it at its worst, and neither is representative. That is the argument for continuous monitoring on assets where intermittency is suspected, and it is a question to settle in the design rather than discover in the reconciliation.
What to get right in the plan
- Separate detection from quantification in scope and budget. They are different instruments and often different crews.
- Measure the flares. In a flaring-heavy portfolio this is where the uncertainty concentrates.
- Choose one site-level technique suited to the facility type, rather than assembling a menu.
- Secure the reconciliation capability at the design stage. Leaving it until the data arrives is how a campaign turns into an unusable pile of readings.
- Design for intermittency if the process has it, because a snapshot will not characterise it.
How ESGweise helps
We design the inventory, the source categorisation and the data architecture that a Level 4 campaign feeds into, specify what the measurement programme needs to deliver so the output is usable, and run the reporting and reconciliation logic on top of it.
To be plain about the boundary: we do not carry out physical methane measurement in house. Survey, quantification and site-level campaigns are delivered with specialist measurement partners, and accredited verification sits with an independent body. Being precise about which side of that line a deliverable sits on is the point of this article.
See our carbon and assurance services, and our oil and gas practice.
To design a measurement programme that produces usable data, talk to us.
Related guides
Frequently asked questions
What does OGMP Level 4 measurement involve?
A component-level campaign across the asset. Optical gas imaging surveys to find leaks, then quantification of what was found using high-flow samplers, calibrated bagging or acoustic methods on valves. A component inventory counting and classifying valves, flanges, connectors and seals, with asset-specific factors applied. Engineering calculation from process data for vents and tanks, with measurement on significant sources. And combustion efficiency measurement plus unlit-flare detection on flares.
What is the difference between detection and quantification?
Detection tells you a leak exists and roughly where. Quantification tells you how much is coming out. An optical gas imaging camera is a detection tool: it makes an invisible plume visible but does not give a rate. Quantifying it requires a different instrument, typically a high-flow sampler, calibrated bagging or acoustic quantification. Programmes that budget for detection and assume quantification comes with it run short.
Which site-level measurement technique should we use for Level 5?
It depends on the facility and the budget. Mobile ground survey with inverse dispersion modelling is a common entry point. Drone-mounted sensors suit congested facilities and flare stacks. Continuous point sensors or tower networks are the only option that reliably catches intermittency. Aircraft campaigns are the standard for basin and large-site work. Satellite is useful for super-emitters and for cross-checking a national inventory but is too coarse for site attribution on its own. One technique done well generally beats a menu.
Why does flare efficiency matter so much?
Because the difference between assumption and reality is large. A flare assumed to burn at 98% efficiency that is actually burning at 92%, or has gone out and is venting raw gas, changes an asset's methane number by an order of magnitude. A flare burning efficiently converts methane to CO2; one burning badly emits methane directly, and methane is roughly eighty times more potent than CO2 over twenty years. Many published estimates rest on assumed efficiencies that have never been verified in the field.