EUDR geolocation of plots of land: 4 ha threshold, polygons and practical cases

In short

Three principles frame all data collection:

• The threshold is 4 hectares. A plot of land of 4 ha or less can be located with a GPS point (a latitude/longitude pair). A plot of land of more than 4 ha must be described by a polygon tracing its perimeter - except for cattle, where the logic is different.
• Precision is regulated. Each coordinate has at least six decimal places, in WGS 84 (EPSG:4326), longitude before latitude.
• The import format of the EUDR information system is GeoJSON (a geographic data exchange format derived from JSON).

The real work is therefore not the threshold itself, but field collection from suppliers and quality control of the geometries received. That is the subject of this article.

What the regulation requires exactly

The obligation is set out in Article 9(1)(d) of Regulation (EU) 2023/1115: the DDS must contain the geolocation of all plots of land where the relevant commodities were produced. The definition of “geolocation” is given in Article 2, point 28. These two articles form the foundation of your entire data set-up.

The 4-hectare threshold, as written

Three nuances are regularly misread:

• “More than 4 hectares” means strictly greater than 4. A plot of exactly 4 hectares is not “of more than 4 hectares”: a single point is allowed. In practice, when a polygon is available, providing it secures this borderline case and makes checks easier.
• A polygon describes one production plot of land, not an approximate area covering several plots. Unconnected pieces of land are considered separate plots, even if they belong to the same owner.
• The threshold does not apply to cattle. For this sector, geolocation concerns the places where the animals are kept.

Why this threshold shapes your entire collection

The 4 ha threshold is not an administrative detail: it creates two very different levels of effort.

• A point can be recorded with a smartphone in a few seconds. A polygon requires walking the perimeter, or drawing a geometry on imagery and having it validated.
• The threshold creates borderline cases: plots close to 4 ha must be measured, not estimated from memory, or you risk choosing the wrong method.
• It determines what the authority can verify. With a polygon, the actual plot is analysed against deforestation layers. With a point, verification is done using a buffer zone, and is therefore coarser.
• It determines the format sent to the EUDR information system: a polygon geometry or a simple point.

The general framework of these obligations - product scope, due diligence, operator roles - is detailed in our guide to the EUDR.

Application dates to remember

The timetable was postponed by the amendments of December 2024 and December 2025. The application dates are now:

Geolocation must be prepared ahead of the DDS: collection lead times with distant suppliers - often several months - must be built into the backward schedule.

“[…] for plots of land of more than four hectares used for the production of the relevant commodities other than cattle, this shall be provided using polygons […]”Regulation (EU) 2023/1115, Article 2, point 28 · consolidated text, EUR-Lex

GPS point or polygon: the difference and its operational consequences

A GPS point (Global Positioning System) is a latitude/longitude pair. A polygon is an ordered sequence of vertices which, when joined, form a closed ring delimiting the plot of land. The two require neither the same work nor offer the same evidential strength.

Expected file formats: GeoJSON, WKT and KML

Three families of formats recur in geographic data exchange:

• GeoJSON: a JSON-based geographic data exchange format, readable by web tools and GIS (geographic information system) software.
• WKT (Well-Known Text): a text representation of a geometry, common in geospatial databases.
• KML (Keyhole Markup Language): an XML format popularised by consumer mapping, useful for visualising a boundary.

Common format errors

1. Reversed coordinate order: putting latitude then longitude. In GeoJSON, the rule is [longitude, latitude].
2. Wrong coordinate system: data in a local projection (UTM, Lambert) instead of WGS 84 / EPSG:4326.
3. Unclosed ring: the last point must be identical to the first.
4. Self-intersection: the geometry crosses itself and becomes invalid.
5. Coordinates in sexagesimal degrees (degrees/minutes/seconds) instead of decimal format.
6. Insufficient number of decimal places, or rounding that is too coarse.
7. Several plots merged into a single geometry.
8. File too large: the system documentation sets a limit of 25 MB per statement, GeoJSON files included.
9. Incorrect structure: mixing Feature and FeatureCollection, or Point and Polygon types.
10. Decimal separator: a comma instead of a full stop, which breaks the file's interpretation.

How to collect polygons from suppliers who have none

Many producers have neither GPS surveys nor a usable cadastral map. Collection must therefore be organised, not improvised. Here are the methods, ranked in decreasing order of reliability.

Field tracing with a smartphone (reference method)

The producer or a technician walks around the plot with a GPS survey app, closes the loop and saves. The file comes out in WGS 84, to be converted to GeoJSON if needed. This is the most reliable method, but it requires a smartphone, a minimum of training and network coverage or an app that works offline.

Satellite imagery and joint validation

The plot is delineated on a satellite image, then the outline is validated by the producer: on-screen review, signature or confirmation photo. This approach covers large areas quickly, provided a traceable validation is obtained to avoid later disputes.

Land registries, cadastre and cooperatives

Retrieving plots already mapped by a cadastre, a cooperative or a producer organisation avoids starting from scratch. The geometries must nevertheless be converted and checked before integration.

Going through the aggregator or trader

When the direct supplier is an aggregator, it is often responsible for upstream collection. Clarify contractually who provides which geometry and in what format, rather than discovering the gaps at DDS time.

Assisted manual tracing, as a last resort

Without a field survey, the outline is drawn on a satellite background with the producer, then the resulting area is checked. To be reserved for simple, well-delimited plots.

A minimum collection protocol

Whatever the method, require for each plot of land: a unique identifier, the identity of the producer, the crop, the declared area, the geometry in WGS 84 (≥ 6 decimal places), the survey method, the date, and proof of validation (photo, signature). Without these fields, the geometry is unusable.

Quality control of received geometries

A raw geometry as received is almost never ready to use. These checks, to be automated as far as possible, avoid discovering a problem when submitting the DDS.

Systematic upstream quality control avoids three costly consequences: a rejected import, a mislocated plot that weakens the proof, and a late correction that is impossible to obtain from a distant supplier.

Linking plot, batch and supplier record

An isolated geolocation is useless: it must be linked to the commodity it supports. The complete data model is detailed in our guide EUDR and supplier data in the ERP. This is the role of the linking key between three objects.

Two practical rules:

• A batch can come from several plots, and a plot can supply several batches. Linking is done by identifier, never by free-text description.
• Version your geometries. A corrected plot must keep its history, with the date of the correction and its justification.

In the event of an inspection, this mesh lets you answer a simple question: does this specific batch really come from these specific plots? The other expected evidence is listed in our EUDR audit checklist.

Practical cases by supplier type

Two cross-cutting lessons: data quality deteriorates further upstream in the chain, so the further downstream you are, the more you need to frame requirements by contract; and the smaller the supplier, the more support collection needs (training, tools, time). This is typically the case in the cocoa and palm oil sectors.

Common mistakes (and how to avoid them)

1. Confusing point and polygon at the threshold, particularly for a plot of exactly 4 ha.
2. Focusing on format before data: a perfectly formatted GeoJSON based on an approximate boundary remains unusable.
3. Accepting geometries without a plot identifier, making linking to batches impossible.
4. Not requiring the coordinate system, and discovering a local projection at import.
5. Starting collection at DDS time, when it takes months.
6. Concatenating several plots into a single geometry.
7. Ignoring overlaps between plots of different suppliers.
8. Not versioning corrected geometries.
9. Not tracing the provenance of each data item (survey method, date, validator).
10. Letting each site collect with its own conventions, at the expense of master data consistency.

Operational checklist

• Before collection
• List the suppliers and plots of land within the scope of the regulation.
• Measure areas to apply the 4 ha threshold.
• Define a unique plot identifier and a standard collection form.
• Choose collection methods according to supplier type.

• During collection
• Record in WGS 84 (EPSG:4326), longitude before latitude, ≥ 6 decimal places.
• Close each polygon ring.
• Document the method, date and validator.

• After collection
• Check self-intersection, ring closure and area consistency.
• Check there is no overlap between plots.
• Convert to compliant GeoJSON and test the import.
• Link each plot to batches and the supplier record.

• Governance
• Version geometries and trace corrections.
• Plan regular updates, as a plot can change.

Key takeaways

The 4-hectare threshold determines the method, the format and the workload: a GPS point for small plots, a polygon for large ones, with standardised precision and coordinate system. Your added value does not lie in the final import, but in three links: collection from suppliers who sometimes lack data, quality control of geometries, and the plot-batch-supplier link.

In practice, the next step is to map your suppliers concerned, measure areas, then industrialise a collection and control protocol.