Wait, what? A biodiversity database shows a plant record at one map point. Beside the coordinates it says Coordinate uncertainty: 1000 m. A student measures from the dot and says, “So the plant was exactly one kilometre away from this point.”
That is not what the field means. In Darwin Core, the standard used widely for biodiversity records, coordinateUncertaintyInMeters is a horizontal distance from the stated latitude and longitude describing the smallest circle intended to contain the whole recorded location. In other words, it is a radius of uncertainty around the point, not an exact error arrow telling you where the organism really was.
Quick answer
“Coordinate uncertainty = 1000 m” means the true recorded location is represented as lying somewhere within an uncertainty circle of radius 1000 m around the stated coordinates. It does not mean the real location is exactly 1000 m away. It does not mean the centre point is known to be wrong. And it does not mean uncertainty can be replaced by zero when the value is unknown.
The owned learner job — not a new mapping or GPS owner
This Reality Lab owns one communication-object problem: how to evaluate a biodiversity occurrence record that displays a coordinate plus an uncertainty radius. It does not replace lessons on maps, GPS physics, measurement uncertainty, sampling, populations or coordinate systems. Those belong to their existing owners.
For the general observation-versus-inference boundary, route to How to Tell Observation, Inference, Prediction and Explanation Apart in PSLE Science. For sampling and occurrence evidence, use Sampling, Field Surveys, Populations & Ecological Evidence for PSLE.
Draw a circle before you draw a conclusion
Imagine an original record:
Observation EK-476
Species: River Fern
Latitude: 1.3500
Longitude: 103.8200
Coordinate uncertainty: 1000 m
Date: 12 April 2026
Do not imagine the organism sitting exactly on the coordinate point. Instead, draw a circle with radius 1000 m around the point. The record is telling you that the location is represented with that uncertainty area.
The centre is a coordinate reference. The circle is the spatial uncertainty. The real occurrence could be near the centre, near the edge, or anywhere else inside the described location area if the metadata are correct.
Three distances that sound similar but are not the same
| Distance | Meaning | Common mistake |
|---|---|---|
| Coordinate uncertainty | Radius describing the uncertainty area around the stated coordinates. | Treating it as the exact error. |
| Actual error | The true distance between the stated point and the real occurrence location, if the real location were known exactly. | Assuming it equals the uncertainty radius. |
| Distance to another feature | Distance from the coordinate or uncertainty area to a road, stream, school or habitat boundary. | Ignoring the uncertainty circle when deciding whether the occurrence was “inside” or “outside”. |
A good map reader keeps these three quantities separate.
Why a radius is not an error measurement
Suppose the record’s uncertainty radius is 1000 m, but the true organism location happens to be only 120 m from the stated coordinates. The metadata can still be consistent. The uncertainty value describes the size of the location area, not a claim that the organism must sit on the outer boundary.
Likewise, if the true location happened to be very close to the centre, the record would not suddenly have “zero uncertainty”. The uncertainty describes what the record can justify, not the distance we discover later by luck.
Observed, recorded, represented and inferred
- Observed: an organism was recorded at some real place.
- Recorded: a coordinate and uncertainty value were entered into the data record.
- Represented: a mapping system places a marker at the coordinate, often without showing the full uncertainty circle by default.
- Inferred: a reader claims the organism was exactly at the marker, exactly 1000 m away, definitely inside a habitat patch, or definitely outside a boundary.
The map marker is useful, but it can visually hide the uncertainty field. This is a classic Reality Lab problem: the interface looks more precise than the metadata actually are.
Why zero is not the right answer when uncertainty is unknown
The current Darwin Core definition states that zero is not a valid value for coordinateUncertaintyInMeters. If the uncertainty is unknown, cannot be estimated or does not apply because there are no coordinates, the field should be left empty rather than filled with zero.
This is a valuable scientific habit: unknown is not the same as none. A missing uncertainty estimate does not prove perfect location accuracy.
Worked case 1: the plant and the stream boundary
A plant record has a coordinate uncertainty radius of 800 m. The point marker appears 200 m inside a protected wetland. A website headline says, “This species was definitely found inside the wetland.”
The point alone is not enough. The uncertainty circle may extend well beyond the wetland boundary. The record could still be excellent evidence that the species occurred near that area, but the exact inside-versus-outside claim is stronger than the spatial metadata support.
Worked case 2: two records that look like duplicates
Two insect records have coordinates 300 m apart. Each has coordinate uncertainty of 1000 m. A learner says, “They are two different sites because the dots are different.”
Not necessarily. Their uncertainty areas overlap heavily. They could represent different observations from different places, or they could be two records whose true locations were much closer than the map dots suggest. The correct interpretation requires date, observer, event and provenance information, not just marker spacing.
Worked case 3: the point beside a school
A sensitive species appears as a point 400 m from a school, with uncertainty 2 km. A social post says, “Rare animal found 400 m from the school.”
The statement treats the central coordinate as an exact location. A more careful claim would preserve the uncertainty: the occurrence record is centred about 400 m from the school but carries a 2 km spatial uncertainty radius. Whether the animal was actually closer or farther cannot be resolved from that record alone.
Worked case 4: old specimen, broad locality
A museum label from 1910 says only “Upper Valley”. Decades later, someone georeferences the locality to a central point with a 12 km uncertainty radius. A map now shows one crisp pin.
The pin is a convenient representation of a broad historical locality. It is not new evidence that the collector stood at that exact coordinate. Georeferencing can turn text into spatial data, but the uncertainty must travel with the converted location.
Worked case 5: rounding creates uncertainty
Suppose a precise coordinate is rounded heavily before publication. The displayed point may shift from the original position, and the rounding itself increases spatial uncertainty. GBIF recommends recording uncertainty and documenting generalisation when coordinates are deliberately made less precise.
A published point that looks tidy can therefore be a deliberately generalised location. Precision of display is not always precision of knowledge.
Worked case 6: a 30 m value is not a promise
The Darwin Core examples include values such as 30 m as a reasonable lower-limit estimate for some GPS records under good conditions when actual precision was not recorded. If a learner sees “30”, they should not convert it into “the GPS made an error of exactly 30 m”. It remains an uncertainty description.
Representation check: the marker is usually smaller than the uncertainty
Mapping interfaces often show a pin only a few pixels wide. At city scale, a 5 km uncertainty radius might be hundreds of times larger than the marker symbol. The display therefore creates a precision illusion unless the uncertainty field is inspected.
When location matters to the claim, do not stop at the dot. Ask for the uncertainty.
Comparison check: can these two records answer the same spatial question?
Imagine Record A has uncertainty 20 m and Record B has uncertainty 10 km. Both show the same species. They may both be valid occurrence records, but they are not equally useful for every spatial question.
- For country-level occurrence, both may be useful.
- For deciding which side of a narrow river the organism was on, Record B may be too uncertain.
- For identifying a specific tree or nest site, even Record A may be insufficient.
Data quality is not one universal score. It is fitness for a particular question.
Method check: where can uncertainty come from?
- GPS measurement limits.
- Old maps or imprecise locality descriptions.
- Rounding coordinates.
- Georeferencing a named place to a representative point.
- Obscuring or generalising sensitive locations.
- Unknown original collection precision.
Different causes can produce similar-looking uncertainty values. Provenance tells you how the radius was obtained and how to interpret it.
What strengthens a fine-scale location claim?
- A small, justified coordinate uncertainty relative to the spatial claim.
- Clear georeferencing method.
- Original GPS or field records.
- Consistent datum and coordinate information.
- No evidence of deliberate spatial generalisation unless accounted for.
- Independent map or field evidence supporting the location.
What weakens it?
- A large uncertainty circle crossing the boundary central to the claim.
- Unknown uncertainty.
- Coordinates derived from a broad place name.
- Rounded or obscured coordinates without checking metadata.
- A screenshot that hides location accuracy.
- Treating a map marker as more precise than the underlying record.
How far can the conclusion travel?
If the uncertainty radius is 1000 m, you can say the record represents a location within the documented uncertainty area around the stated coordinates. If the entire uncertainty circle sits inside one habitat patch, an inside-patch claim may be stronger. If the circle crosses several habitats, a precise habitat assignment requires more evidence.
The rule is simple: spatial claims should not be finer than the spatial evidence.
Tempting reasoning that fails
- “1000 m uncertainty means exactly 1000 m error.” No: it describes an uncertainty radius.
- “The dot is the actual location.” Not necessarily: it is the stated coordinate reference.
- “No uncertainty value means zero uncertainty.” No: unknown should remain unknown.
- “Two different dots prove two different exact sites.” Not if their uncertainty and provenance overlap.
- “A more precise-looking map is better evidence.” Display precision cannot create missing location knowledge.
Original PSLE-style transfer case: the mangrove boundary
This is original practice, not an examination-board question.
A crab record is plotted 300 m inside a mangrove reserve. Its coordinate uncertainty is 900 m. The reserve boundary is narrow and the uncertainty circle extends outside it. A student writes: “The crab was definitely found inside the reserve because the map dot is inside.”
Question 1: What does the 900 m value represent?
Explained answer: It is the radius of the location uncertainty area around the stated coordinates, not the exact error distance.
Question 2: Why is “definitely inside” too strong?
Explained answer: The uncertainty circle crosses outside the reserve, so the record alone does not resolve which side of the boundary contained the true occurrence.
Question 3: What evidence would strengthen the inside-reserve claim?
Explained answer: A more precise original location or field record whose justified uncertainty area lies fully inside the reserve would strengthen it.
Question 4: Is the record useless?
Explained answer: No. It may still be useful for broader claims, such as occurrence in the surrounding landscape, even if it cannot resolve the exact boundary question.
Delayed independent return
- Why is coordinate uncertainty usually drawn as a radius rather than an exact arrow?
- Can the real location be near the centre when uncertainty is 1000 m?
- Why is zero not a valid replacement for unknown uncertainty?
- What happens when an uncertainty circle crosses the habitat boundary being studied?
- Why can an old museum locality produce a precise-looking modern pin?
Self-check: The radius describes a possible location area; yes, the true point may be anywhere within it; unknown is not zero; boundary claims become uncertain when the circle crosses the boundary; georeferencing can turn broad historical text into a representative coordinate without making the original location exact.
Explained practice: centre, radius, claim
For each record below, decide whether the claim is supported:
- Point 2 km from a coastline, uncertainty 50 m: “The record is inland.” Usually well supported if the coastline data are appropriate.
- Point 100 m inside a park, uncertainty 3 km: “The organism was definitely inside the park.” Too strong.
- Point at a school, uncertainty unknown: “The organism was exactly on school grounds.” Too strong.
- Point 20 km inside a large island, uncertainty 500 m: “The record is on that island.” Often well supported at island scale.
Parent and tutor teaching guide: use a coin and a paper circle
Place a coin on a map as the stated coordinate. Cut out a paper circle representing 1 km uncertainty. Hide a small bead somewhere under the circle without the learner seeing where. The coin is the coordinate; the bead is the unknown true location; the paper circle is the uncertainty area.
Ask the learner to explain why the bead is not necessarily on the edge, why the coin can still be useful, and why removing the circle does not improve the data. Then place a habitat boundary across the paper circle and ask whether the bead is definitely on one side.
For three students, use roles: map reader, metadata reader and claim limiter. The map reader sees only the dot, the metadata reader reveals the uncertainty, and the claim limiter revises the original sentence. Rotate roles.
Authoritative sources and curriculum frame
- SEAB — 2026 PSLE Science syllabus: current assessment includes interpreting and analysing information, evaluating observations, information and methods, and communicating explanations and reasoning.
- MOE — 2023 Primary Science Teaching and Learning Syllabus: supports healthy scepticism, evidence evaluation and careful communication.
- Darwin Core List of Terms — coordinateUncertaintyInMeters: defines the field as a horizontal distance describing the smallest circle containing the whole location and states that zero is not a valid value.
- GBIF — Data quality requirements for occurrence datasets: explains coordinate uncertainty as a radius around the stated point and strongly recommends the field for occurrence data.
- GBIF Technical Documentation — Data quality recommendations: discusses coordinate uncertainty, methodology and deliberate spatial generalisation.
The quiet habit to keep
Maps tempt us to believe the dot. Metadata tells us how much of the dot is knowledge and how much is representation. When location matters, ask for the radius before you make the claim.
