To check an emission reduction claim, follow its records: the baseline, the case for additionality, the operating data and the scope of any verification. A photograph of equipment proves none of them. If one link in that trail breaks, the headline number inherits the break.
A drone shot of a flare at golden hour shows that the hardware exists. It cannot tell you whether the flare burned last Tuesday, whether a meter was calibrated or whether the result would have happened without the project. Each of those questions needs a different kind of record.
The public figure sits at the end of decisions about boundaries, reference scenarios, measurements and calculations. A weak decision early on can survive all the way into a polished report with a forest on the cover.
A 2020 European Commission study found that 53% of the environmental claims it examined were vague, misleading or unfounded, and 40% were unsubstantiated. In the Commission’s sample, two claims in five arrived with no evidence to back them.
What is a baseline?
An emission reduction is measured or estimated against something. The baseline describes a credible situation without the intervention, and the comparison has to name its boundary and period. Nobody can observe that alternative directly, so somebody has to justify it.
If the baseline assumes an implausibly high release, even accurate project measurements produce an inflated benefit. Picture a slimming club that measures your progress against the weight you would have reached eating a whole cake every day. Its results would look spectacular, and every gram of them would come from the cake.
Additionality adds a second test. Did the intervention change what would happen? Existing legal duties, likely investment and common practice all count. For a funder, the question becomes whether the activity would go ahead anyway, and the answer needs a reasoned assessment, not a badge.
Boundaries leak as well. In accounting, leakage means an increase in emissions outside a project’s boundary caused by the project itself, such as land clearing that shifts elsewhere once one area gains protection.
What records should a methane project keep?
For methane work, the records start with gas flow, gas composition and whether the treatment equipment was running. Monitoring has to cover the variables and periods the method needs. Uptime matters because installed equipment and operating equipment are different states, and a flare that sat unlit through a wet February destroyed nothing that month, however good it looked in March.
Even a lit flare can let some methane through unburned, so a calculation that assumes perfect destruction overstates the benefit. At a landfill, capture and destruction are separate measurements: how much gas reached the system, and how much of that gas the system destroyed.
Meters need paperwork of their own. A gas-flow meter may require periodic checks against a traceable standard, and calibration still cannot prove the device was installed correctly or ran continuously. A leak programme needs proof of repair as well, because finding a leak does not demonstrate that anyone fixed it.
Gaps in the data need a stated treatment. So does measurement uncertainty, because meter precision, calibration and sampling all shape a methane estimate, and a narrow-looking number can hide a wide range. Then date everything. Source files and calculation versions let a reviewer follow how a figure was produced and what changed along the way.
Read the verifier’s scope
Validation and verification sound alike and do separate jobs. Validation assesses a proposed activity or design against specified requirements, often before any results exist, so a validated plan proves nothing about later operating performance.
Verification assesses reported information against defined criteria. An independent reviewer may examine monitoring data and calculations for one reporting period, or a single calculation, without endorsing every sentence a campaign later builds on the result. A statement covering 2025 is silent about 2026. A useful assurance statement names its criteria, its scope and the level of confidence it provides. Independence and competence need evidence of their own, and assurance has limits.
Who gets to count the tonne?
Two annual reports can celebrate the same tonne. Double counting arises through duplicate issuance, multiple uses or overlapping claims, and it turns one reduction into two.
Registries give each unit a unique serial number to track it through a programme. Retirement removes a credit from circulation for a stated use, which confirms a registry action and falls short of confirming project quality. A credit’s label alone is no evidence of the result behind it. Check the vintage too, because the date a credit was bought and the date its reduction happened are separate facts.
The ICVCM’s Core Carbon Principles set conditions for a credible carbon credit, and additionality, permanence, robust quantification and no double counting are among them. Permanence asks how long a benefit lasts and what could reverse it, such as a fire releasing carbon stored in trees.
Company accounts draw one more line. Under the GHG Protocol, a company’s inventory counts its own emissions across Scopes 1, 2 and 3, and supporting reductions elsewhere does not automatically shrink it.
Make questions possible
No ClimateCurrent record appears in this article. Follow the projects we co-finance and our first site. We’ll publish what the work achieves, with its method and sources.
We run a membership platform for climate action, started on 24 October 2026, the International Day of Climate Action. Companies can take part now and pay for places for their people under the company terms. For yourself, keep your place and follow the work. We report openly on how contributions are used and what they help achieve. A contribution is not an offset and makes nobody climate neutral or carbon neutral.
In the meantime, practise on everyone else’s numbers. Ask for the meter log before you admire the drone shot.