Point or Polygon? The EUDR 4-Hectare Rule Explained
When a single GPS point satisfies the EUDR and when a full polygon is mandatory: the 4-hectare rule, six-decimal precision, edge cases and the math.
Buried in Article 9 of the EU Deforestation Regulation is the single technical rule that decides most of the mapping workload in coffee, cocoa and palm supply chains. Plots of land larger than 4 hectares must be described with polygons. Plots of 4 hectares or less may be described with a single point of latitude and longitude, at six decimal places. Cattle gets its own logic, geolocation of establishments rather than fields. One threshold, and on either side of it sit two completely different amounts of fieldwork.
Simple on paper. In practice the rule generates three questions that this post answers properly: how do you know which side of 4 hectares a plot falls on, what exactly does a compliant point or polygon look like, and, the question almost nobody asks out loud, is the point option actually a good idea even where it is legal?
First, the math of 4 hectares
A hectare is 10,000 square metres, so 4 hectares is 40,000 square metres: a square of 200 by 200 metres. Most smallholder coffee and cocoa plots sit comfortably under that, which is why the point option exists at all. But eyeballing is unreliable near the line, so measure.
The formula for a roughly rectangular plot:
area in hectares = (length in metres × width in metres) ÷ 10,000
Worked example, a plot measuring about 250 by 180 metres.
Step 1: multiply the sides: 250 × 180 = 45,000 square metres. Step 2: convert: 45,000 ÷ 10,000 = 4.5 hectares. Step 3: compare with the threshold: 4.5 is more than 4.
Result: this plot exceeds 4 hectares, so a polygon is mandatory.
Irregular plots do not do arithmetic this politely, which is the honest argument for capturing the boundary by GPS anyway: walk it once and the app computes the area for you, settling the point-or-polygon question and producing the polygon in the same trip.
What a compliant point looks like
A point is a single coordinate pair inside the plot, in WGS84, with at least six decimal digits. The precision requirement is not bureaucratic decoration: six decimals resolves to roughly 11 centimetres, while four decimals is a blur of about 11 metres, enough to place a coordinate in the neighbour's field or the forest edge. Capture the point well inside the plot, not at the gate, not on the road where the phone got signal, and record which farmer and plot it belongs to at the moment of capture, because reconstructing that mapping later is where registries quietly rot.
What a compliant polygon looks like
A closed ring of coordinates tracing the outer boundary: longitude first in each pair, first vertex repeated as the last, no self-crossing, no reliance on interior holes, since the system reads outer boundaries only. Walking the boundary slowly in one clean loop produces this naturally; stitching it together afterwards from scattered waypoints produces the bowties and open rings that fail validation.
The question worth asking: point where allowed, or polygon everywhere?
Legally, a point suffices at 4 hectares and below. Operationally, the market is drifting toward polygons for everything, and the drift has logic behind it. The deforestation check runs your geometry against forest cover maps anchored to the 31 December 2020 cutoff. A polygon states exactly which land you claim; a point leaves the surrounding area to interpretation, and interpretation is precisely what you do not want at a forest edge, where a cleared patch 80 metres away may or may not be inside your plot depending on a boundary you never provided. A polygon that visibly excludes the contested pixel is a defence. A point next to it is a conversation.
There is a real cost asymmetry on the other side: polygons mean walking boundaries, and across ten thousand smallholders that is a programme, not an afternoon. The pragmatic middle path many supply chains have landed on: points as the legal floor where plots are small and far from any forest signal, polygons wherever the plot is near the threshold, near forest, or supplying a buyer who demands them, which increasingly means most buyers. Notably, the system's own file specification treats very small polygons as points during processing, so mapping a 2-hectare plot as a polygon costs nothing on the filing side and preserves the boundary evidence.
Edge cases the rule does not spell out
A farmer with three separate plots is three features, each with its own geometry and its own side of the 4-hectare line; merging them into one blob misstates all three. Plots straddling exactly 4.0 hectares should be treated conservatively: the polygon obligation attaches above 4, measurement error is real, and a polygon is never wrong. And neighbouring farms sharing a boundary should be mapped against the same line, ideally in the same field visit, because two independently walked versions of one fence produce the overlaps that make traceability reviews unpleasant.
Frequently asked questions
Where inside the plot should the point be?
Anywhere clearly inside it, and the practical answer is near the middle. Avoid edges, access paths and anywhere chosen for phone reception rather than geography. A centroid-ish point survives scrutiny; a roadside point invites it.
Can I submit polygons for plots under 4 hectares?
Yes, and it is quietly encouraged: the regulation asks for geolocation as granular as possible, buyers prefer it, and small polygons are processed without penalty. The point option is a floor for feasibility, not a recommendation.
My plot is exactly 4.0 hectares. Point or polygon?
The obligation reads more than 4 hectares, so a point is defensible at exactly 4.0. Given that no field measurement is exact to the square metre, treating anything within measuring error of the line as polygon territory is the position that never needs defending.
Does the rule differ for cattle?
Cattle works on establishments, the locations where animals were kept, rather than crop plots, so the point-versus-polygon area logic applies differently. If you are in the cattle chain, work from the specific guidance for establishments rather than the crop framing here.
Related reading
Validating plot data by hand doesn't scale past a handful of suppliers
Qelvyn builds the internal tools traders and cooperatives use to validate due diligence statements before submission, track plot-to-container traceability, and catch a rejected GeoJSON file before the EU Information System does. If your EUDR data pipeline has outgrown manual checks, tell us what you're tracking and we'll say plainly whether a system pays for itself.