Before you rely on a plot record: 518 control blocks in five regions

A per-plot record is only as useful as its behaviour on ground nobody disputes. Before an operator files with it, they need to know two things: does it ever give a finding to natural forest, and does it reach findings on the estates it is meant for.

We ran pre-registered control campaigns to answer both, in the five regions where the record issues findings: Côte d'Ivoire, Ghana, Malaysia, Indonesia, and East Africa for coffee. The controls are mapped boundaries of known cover, not operators' plots.

Each control boundary was read by the plot evidence record exactly as a customer's plot would be: every look in the open archive over the seasons before the cutoff, from Copernicus Sentinel-1 and Sentinel-2 and the Commission's 2020 forest maps, with national maps as the local reference where they are published.

The record either issued a finding, an established stand before 31 December 2020, or withheld one and said why. The pass lines were written down before any block was read, and every count is printed with its denominator.

For an operator, these counts are the measured behaviour behind every plot record, printed with their denominators in the document's own calibration section. The finding applies to estates and sun-grown crops from four hectares in these five regions; smaller plots and plots elsewhere receive the same document as supporting evidence.

The determination stays with the operator. What the record gives them is a document per plot they can file beside the land title and the purchase records, and a measured answer when an inspector asks what stands behind it.