Method · Lidar
Why a survey that only finds bumps will never find a pond barrow
A pond barrow is a deliberate hollow, not a mound. Our detector accepts positive relief only, so it cannot find one at any setting. Over Salisbury Plain it recovered 0 of 5, and no tuning would have changed that. A survey can be blind to a whole monument class and still report a respectable overall figure.
0 of 5 pond barrows, 0 of 6 disc barrows, 0 of 2 henges, 0 of 1 cursus
Our detector recovered 33.1% of the scheduled monuments on a block of Salisbury Plain. Split that by monument class and something more interesting appears: four classes at zero, and for three of them the zero is permanent.
| Class | Score | Why |
|---|---|---|
| Pond barrow | 0 of 5 | A depression, not a mound. The detector accepts positive relief between 0.20 and 1.50 metres, so the shape is outside what it can represent. |
| Disc barrow | 0 of 6 | A small mound inside a wide flat platform ringed by a bank. Mostly flat. The crest-to-mean filter we added to reject barn roofs rejects these too. |
| Henge | 0 of 2 | Too large. The area ceiling is 8,000 square metres. |
| Cursus | 0 of 1 | Linear and kilometres long. The elongation limit of 1.8 rejects it. |
| Long barrow | 2 of 6 | Elongated by definition, so half are rejected by the same elongation limit. |
Two different kinds of zero
The henge and the cursus zeros are tuning. Raise the area ceiling and the henges come into range. Relax the elongation limit and the cursus does. Both changes have a cost, in candidates you then have to triage, and both are available.
The pond barrow zero is not tuning. It is a category error built into the detector, and there is no threshold anywhere in the code that would fix it.
The detector looks for local relief above the surrounding ground. A pond barrow is below it. Not faintly, not marginally: the sign is wrong. Zero of five, and zero at every threshold, because the thing it is looking for is not the shape those monuments are.
Run 005, Salisbury Plain, scored against the National Heritage List, 6 August 2026. Detector accepts positive relief between 0.20 and 1.50 metres.
That distinction is the one worth carrying to anybody else's survey. A tuning zero says the settings were conservative. A structural zero says an entire class of archaeology is invisible to the method and will be invisible everywhere it is used.
How the blindness got there
Nobody decided to exclude pond barrows. The detector was built to find round mounds, because round mounds are what most of the monuments in the first survey block were, and the filters were written one at a time against problems as they appeared. The crest-to-mean filter went in to stop barn roofs ranking first. It worked, and it took the disc barrows with it.
Every one of those decisions was locally reasonable. The blindness is what they add up to, and it only became visible when the output was scored by class against an independent list. That is the argument for scoring by class, in one paragraph.
What we are doing about it
A hollow finder is the obvious answer: the same local relief model, thresholded on negative deviation instead of positive. It is not difficult. What it will do to precision is the open question, because a moor and a ploughed field are full of hollows, and the honest way to find out is to build it, rescore both blocks, and publish whatever comes out.
Until then the pond barrow zero stands, in public, in a table, next to a recall figure that looks acceptable without it.
The four questions, applied
The same four we put to every result on this site, turned on this method.
- How much of the corpus?
- All 121 scheduled monuments in the Salisbury Plain block, classified by type.
- What was recovered?
- A list of monument classes the method cannot see, separated into the ones that are a tuning choice and the ones that are structural.
- What did we say about what we could not do?
- That one class is unreachable at any setting, and that the fix has not been built or measured yet.
- Did anybody check it independently?
- The monument classes come from the heritage record, not from us. The scoring is ours.
Sources
- LIDAR Composite Digital Terrain Model, 1 metre. Environment Agency, Open Government Licence v3, 2026.
- National Heritage List for England. Historic England, Open Government Licence v3, 2026.
Related
Where this is written up in full
Lost Under the Canopy
Six surveys, 907 candidates, one confirmed. Then the same four questions turned on the most famous survey results in the world.
All four are written and none is on sale. Advance readers can read them first, in exchange for an honest review.