When the model and the measurement disagree, publish both
Across 40 sites the measured figure sat above the modelled one about two thirds of the time. The interesting cases are the other third.
Resources
Insights that move the market
Method notes, field notes and the occasional argument with a coefficient table. Written by the people who take the readings.
Six things worth reading first.
Uncertainty bands are a feature, not an apology
A number without a band is a claim about precision that nobody checked. Here is how ours are derived and what widens them.
Spend-based Scope 3 punishes the companies that improve
If your reported emissions track invoices, sourcing better looks identical to sourcing the same. The incentive is backwards.
Two weeks to instrument a site, one season to trust it
What actually happens between signing and the first defensible figure, including the parts that involve helicopters.
Restatement by append, and why editing is disqualifying
A record that can be quietly corrected is not a record. Corrections belong on the chain, next to what they correct.
Forty sites in, the coefficient tables are the weak link
Regional coefficients were built from a handful of studies and then reused for two decades. Measurement is starting to show where.
A demo library — the articles are titles, not files.
Questions
The four we are asked every time.
Do you sell carbon credits?
No. We measure, and we publish what we measure. Selling the thing you also verify is the conflict this industry already has.
What happens if the measurement is worse than the model?
It gets published. That has happened, and it is the reason a measured figure is worth more than a modelled one.
Can our auditor access the raw series?
Yes, in full, with the provenance chain. Re-derivation by a third party is the point of recording it that way.
How small a site is worth instrumenting?
Below roughly 500 hectares a tower rarely pays for itself, and inversion alone carries a wide band. We will say so rather than sell you one.
Want the method note that is not written yet?
Tell us what you are trying to prove and we will send the closest thing we have.
Let's talk