Before any satellite imagery is classified, a REDD+ baseline has already been largely determined by two decisions: which area counts as the reference region, and which years count as the reference period. Both are choices, both are defensible in multiple ways, and historically both were where baselines got inflated.
The Two Levers
| Choice | Pushes the rate up when… | Pushes it down when… |
|---|---|---|
| Reference region | Drawn to include high-pressure areas | Drawn conservatively around similar conditions |
| Reference period | Ends on a high-deforestation year | Covers a longer, averaged span |
Neither lever is illegitimate to pull, as the region genuinely should resemble the project area, and the period genuinely should be recent enough to be relevant. The problem is that the defensible range for each is wide, and a developer choosing at the favourable end of both compounds the effect.
How the Reference Region Should Be Defined
The principle is similarity, meaning the reference region is meant to represent what would have happened in the project area, so it should match on the variables that drive clearing, including accessibility, slope, distance to roads and settlements, forest type, tenure, and legal status.
A region that matches on forest type but not on road access is not a valid analogue. Roads are among the strongest predictors of deforestation in most tropical settings, and a reference region threaded with logging roads will produce a rate the roadless project area was never going to experience.
The practical test an assessor applies asks whether you can show the distribution of the key drivers in the reference region resembles the distribution in the project area. Not the average, but the distribution. Matching averages while differing in spread is a known way to look similar and behave differently.
How the Period Should Be Defined
Standards typically specify a minimum span, often around a decade, precisely to prevent short windows centred on unusual years. A single El Niño driven fire year, or a policy gap before an enforcement change, can dominate a three year average.
Longer is generally more defensible, with one caveat: if a genuine structural change occurred, such as a moratorium, a road built, or a concession revoked, then pre change years may no longer be representative. That argument can be made, but it has to be made explicitly and evidenced, not achieved silently by choosing a start date.
Why This Is Being Taken Out of Developers' Hands
The jurisdictional shift in newer methodologies exists because of exactly this. When a government or standard sets a single allocated baseline for a whole jurisdiction, the reference region and period stop being project-level choices, and the compounding disappears.
For developers, the practical consequence is that baseline-setting expertise is becoming less of a differentiator and measurement quality more of one. A project cannot argue its way to a larger baseline; it can only measure its actual performance accurately.
What Still Needs Doing at Project Level
Even under allocated baselines, projects must map their own area accurately, demonstrate that the allocation was applied correctly, and monitor actual change. The historical analysis becomes an input to a jurisdictional process rather than a project deliverable, but the mapping accuracy underneath it remains entirely the project's responsibility
Turn climate goals into a verified portfolio
TREEO connects the full carbon cycle, encompassing eligibility, simulation, real time monitoring, and registry ready reporting, thereby combining expert consulting with dMRV technology so the evidence exists before anyone asks for it.



