Science1 distinct publisher3 min readUpdated
Ground-penetrating radar mapped four wings of St Clare's Convent beneath a sports field earmarked for an underground car park, overturning the assumption that its remains were gone.
The Scientist · Science desk

Compiled by The ScientistSomething wrong?How this is made
In January 2025, archaeologists pushed a ground-penetrating radar unit across a snow-covered sports field at Byparken in Roskilde, Denmark, and watched the outlines of St Clare's Convent appear in the returning data [1]. The survey was commissioned because the municipality plans an underground parking facility beneath that field, and the museum organisation ROMU wanted to know what construction would hit before it started [5].
The answer was: a lot. The scans indicate that substantial portions of the complex, which dates to 1256, survive roughly a metre below the grass [4]. The radar traced four wings around a central cloister courtyard, with a church in the southern wing [11]. The northern and eastern wings look especially well preserved, and in the east the data show room divisions and foundations for the pillars that once carried vaulted ceilings [12]. Further readings suggest buildings extending beyond the central monastery square [13].
That matters because the working assumption had been the opposite. Historians had taken it that little remained after the convent's post-Reformation dissolution in the 16th century and subsequent demolition [2][8]. The paper trail supported that view: around half a million of its bricks were sold off to Selsø Castle, which reinforced the belief that the site had been stripped [8]. The lesson is narrow and useful. Salvaging a building's brick is not the same as removing its footprint, and a documented demolition is not a survey.
The method was cheap relative to what it replaced. ROMU, the National Museum of Denmark and the Norwegian Institute for Cultural Heritage Research (NIKU) ran the work together, with NIKU's Arne Abel Stamnes and the National Museum's Magnus Hjorth Jørgensen collecting the radar data and Stamnes processing and interpreting it [9]. Instead of opening trenches, the team drove an ATV-mounted scanner over the turf, sending electromagnetic waves into the soil and reading the returns for changes that mark masonry, ditches and other buried features [10]. The field had escaped later building work, which is why the ground could be read without being broken [c9b]. Stamnes said in a press release that this was the first time he had documented such a well-preserved monastery complex with GPR, and that in his assessment the results are among the clearest from a monastery complex in Scandinavia [3].
The administrative consequence is the part operators should note. St Clare's is the last of Roskilde's five known medieval monasteries to be located [14][7]. The other four sites are protected; St Clare's is not, because authorities had believed its remains were gone [15]. By derivation, it is the only one of the five carrying no protection [16] - a designation gap created by a desk-based conclusion that a day's fieldwork has now contradicted.
Next is a decision by Roskilde municipality on whether archaeologists may dig trial trenches in the section of the field proposed for the car park [17]. Those trenches would test what the radar detected, shape any larger excavation, and identify which parts of the convent need protection [18]. Archaeologist Jesper Langkilde of ROMU said the survey gave a solid basis for planning further investigations [6], and NIKU building archaeologist Regin Meyer said the dataset shows the building-archaeological potential beneath the surface and the value of geophysics for identifying and protecting building remains [19].
Follow any of these and your For You feed starts watching them — no settings page required.
Ranked by verification strength, evidence, and original report placement.
The scans showed that substantial portions of the convent, which dates to 1256, survive about 3 feet (1 meter) underground.
In January 2025, archaeologist Arne Abel Stamnes and colleagues ran a ground-penetrating radar scanner across the snow-covered sports field at Byparken in Roskilde, Denmark, and shapes outlining parts of St Clare's Convent appeared in the data.
Historians had assumed that little remained of the medieval complex after its demolition centuries earlier.
Stamnes said in a press release: "This is the first time I have documented such a well-preserved monastery complex with GPR, and in my assessment these are among the clearest GPR results from a monastery complex in Scandinavia."
Plans to construct an underground parking facility beneath the sports field prompted the investigation; before the project proceeded, the museum organisation ROMU wanted a clearer picture of what construction might encounter.
Archaeologist Jesper Langkilde said in the press release: "We knew that St Clare's Convent had been located here, but not how much of the complex from 1256 had been preserved. The GPR survey has produced very clear results and a solid basis for planning further investigations."
Evidence-backed comparisons of source perspectives and observed adoption signals. Read the methodology
Which Builder, Operator, and Investor concerns the observed source mix emphasized—not a truth score.
Evidence, demonstrated adoption, hype gap, incentives, and confidence are assessed independently, each on its own current evidence. How these are measured.
Single-outlet account of a press release, detailed but unverified by excavation
The cluster contains one publisher, whose article explicitly cites a NIKU press release as its source. The reporting is internally specific — named surveyors, named institutions, method description, wing-by-wing preservation detail, depth estimate — and quotes three specialists by name, which lifts it above a bare announcement. But all quotes come from survey participants, no survey parameters or dataset are disclosed, there is no peer-reviewed publication, and the interpretation has not been ground-truthed because trial trenching is still awaiting a municipal decision.
One completed survey at one site; follow-up work not yet authorised
There is a real, completed deployment: a multi-institution GPR survey executed over the Byparken field in January 2025 that produced an actionable plan used to argue for further investigation. Uptake stops there. No trenching has occurred, no protection designation has been granted, no change to the parking project has been reported, and there is no evidence in the cluster of the approach or the result being taken up at other sites or by other authorities.
Superlatives outrun unexcavated radar interpretation
The framing of a 'lost' convent 'discovered' and of results 'among the clearest' from any Scandinavian monastery complex is stronger than what the underlying evidence currently carries: the location was already known, the assessment is the surveyor's own, and nothing has been dug. The gap is moderate rather than severe because the reported specifics — four wings, a southern church, eastern room divisions and pillar foundations — are concrete, the survey team is named, and the article itself flags that trenches would be needed to confirm the reading.
All voices are survey participants promoting geophysical survey value
Every substantive statement in the cluster comes from an organisation that performed or commissioned the work, routed through a NIKU press release: NIKU's Stamnes rates the results as among Scandinavia's clearest, NIKU's Meyer argues the benefit of geophysical surveys for identifying and protecting remains, and ROMU's Langkilde frames the output as a basis for further investigation it would carry out. ROMU also commissioned the survey in the context of a development project whose scope may be affected, and the finding supports a case for protected designation that the same heritage bodies would administer. No developer, municipal official or independent specialist is quoted to offset this.
Coherent single-source account with no independent corroboration
The factual spine — who surveyed, when, where, with what method, and what was mapped — is reported consistently and attributed to named people and institutions, and the article is transparent that it draws on a NIKU press release. Confidence is held near the middle because there is exactly one publisher, zero independent verification, no methodological parameters, and the key interpretive claims are unexcavated and self-assessed.
invest
Sarsia bets EUR 3.7M that the cheapest low-carbon concrete is the concrete already standing1 distinct publisher
Distinct publishers with included, body-backed reporting in this cluster.
1 article · August 20, 2026