EUDR Enforcement — Interpretation Guide

How to interpret the downloadable files from the dashboard. For full methodology details see the Deforestation Check documentation.

1. Plot GeoJSON Export

Downloaded via the "Download Plot GeoJSON" button in the plot details view.

Filename: {assent_clientid}_{supplierid}_{partid}_{ingestion_date}_{status}.geojson

Attributes

PropertyDescription
uuidUnique plot identifier.
plotidPlot ID from the original upload.
area_ha_commodityPlot area in hectares.
noncompliance_areaDeforestation area in natural forest (ha). See methodology.
noncompliance_area_percDeforestation percentage (natural forest only).
deforestation_confidenceDetection confidence level.
deforestation_alertAlert classification (critical / none).
deforestation_startFirst date of detected deforestation (null if none).
deforestation_endLast date of detected deforestation (null if none).
monitoring_startMonitoring period start date.
monitoring_endMonitoring period end date.
ingestion_dateProcessing date.
thumbnail_urlURL to the satellite thumbnail image.
partidPart identifier.
supplieridSupplier identifier.
assent_clientidClient identifier.

2. Audit Trail ZIP

Downloaded via the "Download DDS" button in the aggregate table.

Filename: {assent_clientid}_{supplierid}_{partid}_{ingestion_date}_{status}.zip

2.1 ZIP Contents

FileDescription
audit_trail_{hash}.pdfConsolidated PDF audit trail with maps, statistics, and compliance determination.
dds_geolocations_{hash}.geojsonLightweight GeoJSON for EU Traces upload.

2.2 Audit Trail PDF Sections

SectionContent
EUDR Compliance SummaryDocument owner, client email, supplier name, processing date, DDS reference/verification numbers, submission status.
SummaryCountry-level and plot-level maps, summary statistics (total plots, area monitored, deforestation area/rate, high/moderate risk plots, compliance bar).
Document InformationCommodity, HS codes, country of production, monitoring periods.
Information RequirementsDetailed compliance information per EU regulation requirements.
Risk AssessmentRisk mitigation details and assessment outcomes.
Non-Compliant PlotsIndividual plot details for plots flagged as non-compliant, with satellite thumbnails.

2.3 EU Traces GeoJSON

A lightweight GeoJSON intended for direct upload to the EU Traces system.

PropertyDescription
ProducerNameName of the producer / supplier.
ProducerCountryCountry code of the production location.
ProductionPlaceName or description of the production place.
AreaPlot area.
Only these four attributes plus geometry are included to keep the file within the EU Traces 25 MB upload limit.

3. Facility Validation Results

When screening facilities via the Producer Facility Screening tab, the response includes per-facility validation metrics and, optionally, deforestation, Article 10 procedural risk, and legality / adverse-media intelligence.

3.1 Deforestation Check (Optional)

Enabled via the "Include deforestation check" checkbox. Runs a 30 m resolution analysis within a 50 km radius of each facility, masked to natural forest only.

PropertyDescription
deforestation_area_haTotal deforestation area (ha) within 50 km radius on natural forest, post-EUDR cutoff (2021+).

3.2 Article 10 Procedural Risk (Optional)

Enabled via the "Include Article 10 risk check" checkbox. This is the canonical Article 10(2) procedural-risk assessment, scored from the criteria a–j rubric (country indices, presence of forest, indigenous/land-tenure conflict, anti-corruption, circumvention, and so on). The same rubric is applied on both the plot and facility paths.

PropertyDescription
article10_risk_score0.0–1.0 Article 10 procedural-risk score.
article10_risk_levellow / medium / high / not_assessed

3.3 Legality / Adverse-Media (Optional)

Enabled via the "Include legality / adverse-media" checkbox. Uses AI web search to find recent reports linking the producer to legality risks. This is the Article 9 adverse-media signal and is independent of the Article 10 procedural score above.

PropertyDescription
legality_risk_scoreBanded score: 0.20 (low) / 0.55 (medium) / 0.85 (high). On the facility path this is Article 9 adverse-media only (no deterministic blend).
legality_risk_level / risk_levellow / medium / high, derived from the relevance tiers of the sources found.
adverse_media_risk_score / adverse_media_risk_levelAdditive aliases of the two fields above (same values), retained for backward compatibility.
legality_sourcesArray of sources, each with title, URL, date, snippet, relevance tier (facility/local/regional/national), and risk signal type.
legality_summaryOne-sentence synthesis of the adverse-media signal.

How the legality level is derived from source relevance

The legality risk level is set by the most relevant source found (not by adding per-source deltas), then mapped to the banded score:

Strongest evidence foundRisk levelScore
A source naming the specific company/facility (facility tier)high0.85
A source covering the same district/municipality or province/state (local / regional tier)medium0.55
Three or more country-level (national) sources, with nothing more specificmedium0.55
Fewer than three national-only sources, or nonelow0.20
No adverse signal found folds to low (a clean result), never "not assessed". Sources are ranked by relevance then freshness, and adverse-media results are cached for 24 hours.
The EU Article 29 country tier (country_tier: low / standard / high) is a separate signal. It is a routing gate that selects the due-diligence regime for the country of production — it is not folded into the Article 10 or legality scores.

4. Plot Validation — Collection-Level Article 10

For plots, Article 10 procedural risk is assessed at the collection level (not per-plot), similar to cherry-picking and circumvention risk. The system searches for reports covering the entire plot collection area. The result appears in the collection properties.article10 object:

KeyDescription
riskOverall tier: low / medium / high / not_assessed.
criteriaPer-criterion breakdown (a–j plus a supplementary protected_area_overlap), each with its assessment text, risk, deterministic inputs, and sources.
avg_article10_risk_scoreThe numeric score that fed the overall tier.
gemini_summaryLLM-generated narrative summary of the Article 10 assessment.
all_sourcesConsolidated list of grounding sources cited across the criteria.