How Accurate Must EUDR Coordinates Be?

Six decimal digits do not prove physical accuracy

EUDR defines geolocation using latitude and longitude coordinates with at least six decimal digits. That is a numerical precision requirement. It does not establish a universal survey-accuracy tolerance in metres, and it does not prove that a coordinate identifies the correct production location. The data must also be reliable enough for the operator's substantive assessment.

Avoid describing six decimal digits as a guarantee of centimetre-level or sub-metre accuracy. Decimal places can be generated by software even when the source position is coarse or wrong. The meaningful question is how the location was obtained and whether the evidence supports its relationship to the production land.

Precision describes the recorded number

A coordinate with many digits expresses a fine numerical position. If the original observation was taken at the wrong farm gate, however, extra digits only specify the wrong position more precisely. Adding zeros to a rounded coordinate does not establish newly observed information about the land.

Keep the original source record and distinguish formatting from collection. An export tool may produce the required decimal representation without improving the underlying observation. The reviewer should be able to see whether the value came from field collection, an imagery interpretation, a cadastral dataset or an unexplained spreadsheet supplied by an intermediary.

Accuracy concerns correspondence with the real place

A location record is useful when it corresponds to the relevant production plot or establishment. Field collection can be affected by device quality, obstructed reception, operator error and the position from which the measurement was taken. Imagery-based mapping depends on suitable imagery, interpretation and a clear understanding of the land boundary.

For example, a point captured at a producer's house may be geographically accurate for the house but inadequate for identifying a separate production plot. The supplier should explain what the observation represents. Accuracy at a recorded feature and correctness of the chosen production feature are related but different checks.

Inspect the acquisition method

Ask who collected the coordinates, when, using which source or method, and how the producer confirmed the location. For a polygon, establish whether it came from a boundary walk, recognised land records, imagery digitisation or another defensible source. Record limitations rather than replacing them with an unsupported accuracy label.

A desktop mapping exercise may be adequate for some settings while needing further evidence in others. Dense tree cover, unclear boundaries or inconsistent land records can justify additional checks. There is no reason to demand one expensive method from every supplier if another reliable method answers the actual location question, but the choice should be understandable.

Coordinate systems can create large errors

The current official GeoJSON specification uses WGS 84 decimal degrees with longitude before latitude in coordinate arrays. Data recorded in a projected national system require a proper coordinate transformation. Relabelling easting and northing columns as longitude and latitude does not transform their values.

Axis order can also cause an apparently valid but misplaced location. Where both swapped numbers fall within possible numerical ranges, a range check alone will not catch the problem. The invalid-coordinate checks explain why country, coastline and source comparisons need to accompany basic number validation.

Compare the geometry with independent context

Inspect the record against suitable basemaps, the declared production country, known settlement and farm references and any supporting boundary evidence. Look for systematic shifts, duplicated points assigned to distant producers and polygons covering unrelated features. Such checks are practical indicators, not independent proof of lawful land rights.

When two sources disagree, investigate rather than selecting whichever produces a favourable risk result. A satellite layer and a producer map may use different dates or define different areas. Preserve the discrepancy and its resolution. The operator should know why it concluded that the chosen geometry identifies the land associated with the product.

Geometry quality can affect risk assessment

Article 10 includes the source, reliability and validity of information among risk criteria. Uncertain location data can prevent a meaningful historical land assessment or obscure whether a concern affects the production plot. The uncertainty should therefore be reflected in the risk decision rather than dismissed after successful upload.

For a small plot near a forest-change boundary, additional perimeter evidence may be useful even where a point is legally permitted. For a larger non-cattle plot, a point cannot replace the required polygon. The legal representation requirement and the evidential sufficiency of the particular geometry must both be considered.

Correct errors with an audit trail

Keep the original value, corrected value, source, reason, reviewer and affected products. If a correction changes the assessed land, repeat the relevant substantive checks and examine whether earlier conclusions or submissions are affected. A silent movement of the point can conceal an important change in the origin evidence.

The EUDR geolocation guide sets the broader context. Use Articles 2(28), 9 and 10 for the legal information and reliability requirements and the current Commission GeoJSON documentation for technical representation. A defensible record explains the location evidence and its limitations, instead of treating decimal formatting as proof of physical accuracy.

An internal quality category can help organise follow-up, provided its meaning is explicit. For example, the team might distinguish a producer-confirmed boundary from an unresolved approximate location. Such categories are management choices rather than EU accuracy grades, and they should not imply official certification or replace the evidence review needed for the actual product.