1,842 Metres Went Into The Freezer. The Analysis Has Eaten 1,109 Of Them.
A 30-metre section of one Greenland core has been called removed since 2015. This newspaper audited thirteen custody transfers and 806 sample requests: 27.4 metres were measured away, 1.8 lost to a dead alarm line, 0.6 misfiled abroad.
Analysis is written by news reporters and interprets facts they have verified. It is not opinion, and it takes no position on what should be done.
COPENHAGEN, Denmark — The core is 1,842 metres of ice. It was drilled at Havsteen Dome in East Greenland between 1994 and 1998, and 733 metres of it are left.
The rest has not been lost, withheld or destroyed. It has been measured. Cut into discs, weighed out against 806 approved requests, melted, degassed, run through spectrometers and written up.
A deep ice core is the only archive we know of that has to be destroyed in order to be read.
The section everybody asks about
Between 1,164 and 1,194 metres the core passes through the sulphate laid down in the years around 536 CE — the dust veil, the summer that did not arrive, the cold decade in the tree rings.
The standing account is that this 30-metre section was removed from the archive and is unavailable. In 2015 a research group asked for it and was refused in writing. That letter is real and we have read it.
The section is unavailable. We wanted to know what unavailable meant, so we stopped asking the archive and audited its custody record instead: thirteen institutional transfers, four countries, 806 approved requests, two freezer failures, and every mass logged between the drill tent and this month.
What the log says happened to it
Twenty-seven point four metres were consumed by sixty-one approved measurements between 1999 and 2018. Every one of them has a requestor's name, a date, an institution and a mass in grams, in a log that has been public since 2004 and that nobody, as far as we can establish, had ever added up.
Fourteen point nine metres went to melted chemistry. Six point two to gas extraction. Four point one to tephra and microparticles, and 2.2 to water isotopes.
The largest single draw was 1.9 metres in 2004, for a continuous-flow run that produced the high-resolution sulphate curve now cited by most of the people who say the ice is missing. The measurement everybody quotes is the reason the ice is not there.
Nine hours in March 2007
One point eight metres are a genuine loss and they are the only genuine loss in thirty metres.
On the night of 14 March 2007 a compressor failed in a store holding 340 metres of core. The temperature rose from −26 °C to −4 °C over about nine hours. The store's alarm worked. It dialled out, as designed, to an analogue telephone line that had been disconnected nineteen days earlier in a billing dispute between an estates department and a supplier over a line rental of £43 a quarter.
Solveig Ranum arrived at 06:40 to start the day's cutting, felt the door seal, rang the number on the sticker, got nothing, and rang her manager at his home at 06:52. The incident was logged at 07:31 and the store was back below −20 °C by that evening. Her name is on the file and she has never been named in any account of this core until today.
Drawer 11, and two transposed digits
Six-tenths of a metre are in a freezer in another country, in a drawer, correctly stored at −24 °C, catalogued as depth 1,149 metres.
It is from 1,194 metres. Two digits are transposed on a label typed in 2011, and this newspaper found it by matching logged masses against depths, which is arithmetic and not detection. It had been requested three times and refused three times as unavailable, and it was eleven metres from a person whose job included knowing where it was.
Our vault holds one comparable custody instrument, PX-1980, and one comparable set of cores, PX-2125. Neither institution logs mass at issue. This one does, which is the only reason this story exists.
What is actually gone
Two-tenths of a metre. About 1.4 kilograms.
We cannot tell you whether that is ice or arithmetic, because the log changed its rounding in 2003 and sixty-one requests of accumulated rounding gets you most of the way there. It is printed above under Unverified and it is the honest end of the audit.
The desk takes this from it. A core is a wasting asset, spent on purpose by the people who wanted it, with the paperwork complete and public throughout. The one real loss was not a decision about the sixth century. It was a compressor, an autodialler, and a quarterly invoice for forty-three pounds that nobody in a building of four hundred people was responsible for paying.
The six o'clock edition
Three stories a morning, in your inbox before the coffee is made.
How Others Covered This
The same events, as reported elsewhere on the same day. We list what each outlet had that we did not, as well as what we had that they did not — including where we come off worse. Why we print this.
-
The Hollow PostThe Thirty Metres That Cover The Sixth Century — And Who Ordered Them Out
Treated the archive's inability to produce the section on request as an instruction rather than an inventory.
Had that we did not
The 2015 refusal letter in full, which is a real document and does say no.
Left out
The sampling log sitting behind that letter, which is public, is 806 rows long, and explains 27.4 of the 30 metres by name and date.
-
Signal & LedgerCold Chain, Wasting Asset: What A Core Costs To Keep And What It Costs To Read
Read the archive as an inventory problem and priced the freezer estate for readers who run cold stores.
Had that we did not
The depreciation framing, which is correct and which we have borrowed from them.
Left out
The nine hours in March 2007, which appear in their model as shrinkage and in the incident file as a woman ringing a dead number at 06:44.
-
They Buriedthis newspaper1,842 Metres Went Into The Freezer. The Analysis Has Eaten 1,109 Of Them.
Audited the whole custody record — thirteen transfers, four countries, every logged mass — instead of asking the archive one more time for the section.
Had that we did not
All 806 requests, the two freezer failures, the transposed label, and the 0.2 metres we cannot place.
Left out
Our headline says the analysis ate the ice, which is a verb the curator did not use and objected to twice before we published, and we ran it anyway because it is a better verb. — V. Ashcombe-Doyle, standards editor
Right of Reply
They Buried contacted Dr. Ingrid Sørbø-Achebe, curator of the ice-core archive, Copenhagen on 8 June 2026 in writing, with the full audit, all 806 coded rows and every figure, and ten days to answer. Replied on 18 June 2026 and asked that it run entire. Printed unedited, including the parts about us.
Your arithmetic is right and I have checked it, which took my colleague and me two days, and I would like that on the record before I complain about everything else.
You have used the word eaten. I asked you not to and you have used it anyway, so let me explain what it costs you. An ice core is not a painting. It is not held for its own sake and it does not become more valuable by remaining whole. It is the only archive I know of that has to be destroyed in order to be read, and every gram of it that has gone has gone through a form, signed by me or by my predecessor, because somebody made a case that the measurement was worth more than the ice. That is the entire purpose of the building I work in. When you write that the analysis ate it, you make the ordinary operation of an archive sound like an accident that happened to us.
And the ice is not gone. This is the part I cannot get any of you to print. Those 27.4 metres are now 41 published measurements, with their calibrations, their uncertainties and in most cases their raw traces, and those will exist after every physical core in the northern hemisphere has sublimated in some future power cut. We converted a wasting asset into a permanent one. I would do it again tomorrow and I would sign the same forms.
On March 2007 I will not fight you. It was a failure, it was our failure, the alarm was ours to maintain and the line was ours to pay for, and I have never once said otherwise in nineteen years. But I want to say what you have done with it. You have found an invoice for forty-three pounds and you have made it the villain, because forty-three pounds is a number a reader can hold. The failure was not an invoice. It was that no person in an organisation of four hundred was responsible for testing whether the alarm reached a human being, and we tested the sensors monthly and the telephone line never, and it took losing 1.8 metres of the sixth century to notice. Print the system, not the bill.
One further thing. You gave me ten days to answer eight months of your work. I answered inside ten days because if I had not, you would have printed that I did not answer. I understand why you work this way. I would like you to notice that it is a form of pressure, and that you describe it in your method as though it were a courtesy.
On Drawer 11 you are simply right and I am glad you found it. It has been requested three times since 2011 and refused three times as unavailable, and the people who were refused were entitled to it, and I have written to all three.
Published unedited under our right-of-reply guarantee.
How was this story?
We publish the result, whatever it is. Reader verdicts appear on the front page and in our newsroom metrics.
Readers' Letters 0
Printed at once under the name you give and read by the desk afterwards; anything unfit is removed, with a note saying so, and nothing else is ever deleted — only corrected. Letters that changed something in the story carry a mark saying so, and there are 5 of those across the archive.