Almost every geolocation question in EUDR work reduces to the same root: what is one plot? The regulation collects geolocation per plot of land, not per farm, supplier, or shipment, so drawing the boundary wrong poisons everything downstream. This page explains the unit and the mistakes teams make when they guess at it.
A practical preparation workflow
- 1
Identify each production location as a separate plot or establishment.
- 2
Source each boundary from cadastre, concession, or field GPS evidence.
- 3
Assign one stable plot_id and one geometry per plot, then validate.
The unit is the plot, not the farm
The EUDR collects geolocation for all plots of land where the relevant commodities were produced. A plot is an identifiable piece of land, in practice the parcel or holding unit that the production records, cadastre, concession, or supplier documentation already recognises. A 200-hectare estate with twelve cocoa fields is twelve plots, not one. Submitting a farm outline where twelve polygons belong understates precision, breaks the per-plot connection to production, and confuses screening tools that evaluate each polygon independently.
Cattle use establishments instead
For cattle products the geolocation targets the establishments where the animals were kept, and a point is the expected form. Establishments play the same role plots do for other commodities: they are the per-unit record the DDS and any screening attach to. Mixing establishment points into a plot file, or converting them to polygons to look consistent, creates data that is neither legally meaningful nor analytically useful.
Where the boundary comes from
A defensible boundary traces to evidence: cadastre extracts, concession or lease documents, supplier declarations with maps, or GPS traces collected in the field. Each source has failure modes: cadastral data can be outdated, supplier maps can cover the wrong parcel or the whole property, and hand-drawn boundaries carry no provenance at all. Record which source produced each boundary and when, because reviewers and auditors ask exactly that.
Identity: one stable plot_id
Each plot needs a stable identifier that survives file conversions, screening runs, and re-exports, and the business record must say which producer, supplier, batch, or shipment depends on it. Duplicate features with different ids, one id reused for two parcels, and overlapping boundaries between plots are the issues validation should surface, because they quietly corrupt both the DDS geolocation file and any satellite screening attached to it.
How this connects to the four-hectare rule
Plot size also decides the geolocation form: a single point is allowed for plots up to four hectares, polygons are required above that threshold. Measure the real plot, not the property, when applying the rule. The full requirements, precision, formats, and conversion, are covered on the geolocation requirements page; the free GeoJSON validator checks the structural side in the browser.
What to check
- One geometry per plot of land
- Cattle recorded as establishments
- Boundaries match source documents
- No duplicate or overlapping plots
Plot classification is a factual and sometimes legal question; a technical file structure does not settle it.
Questions teams ask
Is a farm one plot or several?
Several, when production happens on multiple identifiable fields; the unit is the plot of land, not the property or the supplier.
What do cattle operators geolocate?
The establishments where the animals were kept, provided as points rather than plots.
Can two plots share one boundary file?
No; overlapping or duplicated boundaries are review findings, and each plot needs its own geometry and stable id.
Does a valid polygon prove the plot is right?
No. Structure proves readability; the boundary still needs a source and a link to the actual production and producer.