The central obligation of the EU Deforestation Regulation is deceptively simple: prove that your commodity was not produced on land deforested after 31 December 2020. In practice, building that proof for complex, multi-tier supply chains sourcing from smallholders in Peru, Brazil, or Colombia requires a structured technical process. Certifications like Rainforest Alliance or RSPO are helpful evidence — but they are not substitutes for geolocation-based verification.
This guide walks through the technical verification process for the three most common Latin American commodities facing EUDR scrutiny: cacao, soy, and palm oil.
The Forest Baseline: What You Are Comparing Against
The EUDR requires showing that no deforestation occurred after 31 December 2020. This means you need two things: a reliable forest cover map for 2020 and a reliable deforestation detection method for the period from January 2021 to the present.
The most widely accepted datasets are:
- Global Forest Watch (Hansen/UMD) — annual 30m resolution tree cover loss, updated annually. The standard reference for 2020 forest cover.
- PRODES (Brazil's INPE) — higher resolution deforestation alerts specifically for the Amazon and Cerrado. More precise than GFW for Brazilian soy and cattle.
- SERNANP/MINAM (Peru) — official forest cover maps for Peruvian Amazon, authoritative for cacao and palm in Loreto, Ucayali, San Martín.
- MapBiomas — annual land use and land cover classification covering all of South America at 30m. Useful for cross-validation.
Cacao from Peru and Ecuador
Peru is the world's second largest producer of fine-flavour cacao, with major production in San Martín, Amazonas, and Ucayali — all regions with active deforestation fronts. The verification challenge is that Peruvian cacao frequently comes from smallholder plots of 1–5 hectares, aggregated through cooperatives and intermediary buyers (acopiadores) before reaching the export port.
The Verification Steps
- Obtain GPS coordinates for each producer plot. The minimum requirement is a single centroid coordinate per plot. Best practice for high-risk areas is a full polygon boundary. This data must come from the cooperative, the field agent, or a mobile data collection system (ODK, KoBoToolbox, or Terralyr's field census module).
- Run a spatial intersection against the 2020 forest baseline. For each plot coordinate, check whether the location was classified as forest in December 2020. Any plot where the centroid falls inside a forest polygon as of 2020 must be further assessed to confirm the land was already converted before that date.
- Run a deforestation alert check from January 2021 to present. Use GLAD alerts (University of Maryland, distributed via GFW) or PRODES alerts for Brazil. Any alert intersecting a producer plot within your supply chain must be investigated.
- Generate a per-plot compliance determination. For each plot: GREEN (no forest 2020, no alerts), AMBER (forest 2020 but clearance pre-dates 2020 with evidence), or RED (deforestation alert post-2020 — do not source until resolved).
Soy from Brazil and Argentina
Brazilian soy is produced in Mato Grosso, Pará, Bahia (Matopiba), and other states. The Soy Moratorium (2008) established a no-deforestation commitment for the Brazilian Amazon, but the EUDR extends this logic to Cerrado, which is not covered by the moratorium.
Key point: A farm that was deforested in the Cerrado before 2021 but after the moratorium baseline is still potentially non-compliant with EUDR if that deforestation occurred after 31 December 2020.
For soy verification:
- Use PRODES and the Cerrado Deforestation Monitoring System (PMDB Cerrado from INPE) as primary references.
- For large farms, plot-level verification using Sentinel-2 or Planet imagery provides the highest confidence level.
- The Brazilian Rural Environmental Registry (CAR) provides farm boundaries and declared land use — a useful starting point but not sufficient alone.
Palm Oil from Colombia and Peru
Colombia is the largest palm oil producer in Latin America. Peruvian palm is concentrated in Ucayali and San Martín. The challenge here is that palm is a perennial crop: a plantation planted in 2018 on previously deforested land will appear as existing agricultural land in 2020 — making the 2020 cut-off less straightforward.
The verification approach must go back further: using satellite time-series from 2015–2020 to confirm that the land had already transitioned away from forest before the EUDR cut-off date.
What Makes Evidence "Audit-Proof"?
Competent authorities auditing EUDR compliance will look for:
- A documented methodology for how you identified and verified each producer plot
- The specific datasets used and their dates
- The spatial analysis methodology (intersection algorithm, buffer distances used)
- A per-plot or per-supplier determination with supporting imagery or data exports
- Evidence of follow-up where AMBER cases were investigated
- A documented decision to exclude or include each supplier based on the analysis
Screenshots and PDF maps are insufficient. Authorities expect structured data: GeoJSON or shapefile exports of the producer plots, query logs from the forest databases, and a traceable reference between specific invoices and the geolocation analysis.
Scaling the Verification Process
Manual verification is feasible for importers with fewer than 50 direct suppliers. For importers dealing with cooperatives representing hundreds of smallholders, automation is not optional — it is the only way to complete verification at scale within the annual procurement cycle.
Terralyr automates the spatial intersection pipeline: ingest supplier GPS data via API or spreadsheet upload, run parallel intersection against GFW, PRODES, and SERNANP baselines, flag all AMBER and RED cases for analyst review, and export the complete verification package in formats accepted by EU customs authorities.