Dickson Fjord 2023: The Seismic Signal Was Never Missing

At 12:35:03 UTC on 16 September 2023, roughly 4,500 seismic stations began recording a clean monochromatic tone at 10.88 millihertz — a period of 92 seconds — and kept recording it for up to nine days. The instruments were not confused. The United States Geological Survey catalogue holds the moment as event us7000kx2u, a magnitude 4.3 landslide in eastern Greenland, computed from zero stations, with no azimuthal gap, no minimum station distance and a depth fixed at zero. Everything was heard. Nothing could be named, because naming is a property of the detection algorithm and not of the sensor.

Silence Index — D-184

Genuinely remote

Ranks 32 of 99 scored dossiers — quieter than 68% of the others we have scored. A low score is not a warning; it describes who else will be standing next to you.

How the total is weighted. These are the fixed weights of the four inputs, the same on every dossier — not this place’s scores. We do not publish per-input bars, because the per-input values were never recorded article by article and inventing four plausible numbers that sum to the total would be a fabrication dressed as a chart.

Footfall pressure — visitors measured against the ground that absorbs them30%

Access friction — permits, caps, seasons, boats, weather windows30%

Acoustic floor — how close the nearest engine is20%

Unresolved questions — what the record still cannot settle20%

This dossier was scored before 2 September 2026, when the fourth input became night darkness. It keeps the input it was reasoned with until it is re-scored at its next source re-check; the methodology page records the change.

The index is a desk instrument scored by this publication, not a field measurement. How it is calculated, and where it fails.

What this dossier is built from

This is desk research. Nobody from this publication has been to Dickson Fjord, and almost nobody has: it lies inside the North-East Greenland National Park, several hundred kilometres from the nearest settlement, and the lawful route in runs through a permit office in Nuuk rather than an airport. What follows is assembled from the peer-reviewed papers in Science and Nature Communications; the machine-readable USGS earthquake catalogue, queried directly rather than read about; the FDSN channel-code standard and the GEOFON and EarthScope network metadata that define what a seismic station is built to hear; NASA, JPL and CNES documentation for the SWOT mission; the Greenland Government's own expedition-permit machinery; the Danish armed forces' descriptions of the units that found the damage; UNESCO and Protected Planet records; Statistics Greenland and Visit Greenland's national accounts; and the GeoNames gazetteer.

It is not a story about a mystery. It is a story about the difference between an instrument and a catalogue, which is one of the more elegant distinctions in applied geophysics and one that almost none of the coverage of this event bothered to draw.

The signal, as an engineering problem

Start with what the sensors did, because they did it beautifully.

A rock-ice avalanche of 25 million cubic metres came off a peak 1,200 metres above Dickson Fjord and into 540 metres of water in a channel 2.7 kilometres wide with a sharp inner bend. The initial backsplash reached about 200 metres; subsequent waves reached 110 metres; and within minutes the whole thing settled into a transverse seiche — water sloshing from wall to wall — with a maximum amplitude of about 7.4 metres. Svennevig and colleagues model an oscillating horizontal force of 5 × 10¹¹ newtons acting perpendicular to the fjord axis, moving roughly 14 metres of water back and forth along a ten-kilometre reach.

That force is a very long period source. It radiated a tone at 10.88 millihertz with a quality factor starting near 500 and climbing towards 3,000, which is a technical way of saying the fjord was a remarkably low-loss resonator. The signal appeared on vertical-component records from about 4,500 stations out to 90° — ten thousand kilometres. It appeared at DK.SCO, the Danish network station at Scoresbysund, which the paper gives as 300 kilometres away; on our own great-circle arithmetic from the published source coordinate to the Ittoqqortoormiit gazetteer entry the figure is about 312 kilometres, and we print both rather than choosing. It appeared at II.BFO, the Black Forest Observatory at Wolfach in Germany, 29° away — our arithmetic makes that 3,232 kilometres. The International Monitoring System's infrasound arrays I37NO and I43RU registered the air moving at a range of 3,310 kilometres.

None of this is remarkable for the hardware. The Global Seismographic Network's own USGS fact sheet describes instruments sensitive with high fidelity to oscillations with periods of thousands of seconds. A 92-second tone is comfortably inside that. The sensors were doing exactly what they were built to do.

Why a network that heard it could not name it

Here is the pinch. A global seismic catalogue is not a recording. It is a data structure, and the structure has a shape.

The International Seismological Centre, which produces what it calls the definitive record of the Earth's seismicity, assembles its Bulletin from agency submissions of hypocentres, phase arrival times and focal mechanisms. Its automatic grouping software scores candidate events on azimuthal gap, station distribution and phase count. Every one of those metrics presupposes a picked arrival — a discrete moment at which energy arrives at a station, from which a travel time can be differenced against another station's arrival to triangulate a source. Impulsive P and S onsets are what the whole apparatus eats.

The Dickson Fjord seiche had none. It was not an arrival; it was a hum that faded up out of the landslide's own high-frequency signal and then simply persisted. There was no onset to pick, therefore no travel time, therefore no location, therefore — from the catalogue's point of view — no event.

You can watch this happen in the record. Query the USGS event service for the box containing the fjord on 16 September 2023 and exactly one event comes back: us7000kx2u, magnitude 4.3 on the Ms_VX scale, 72.797°N 27.115°W, depth 0 km, event type "landslide", reviewed. Now read the quality fields. Number of stations used: 0. Azimuthal gap: null. Minimum station distance: null. The catalogue caught the avalanche as a surface-wave amplitude measurement and filed it with no location solution behind it at all, because there was nothing to solve with. Query the same box for the whole of October 2023 and the count is zero — the second landslide-tsunami of 11 October, from the same gully, with the same period and radiation pattern at half the amplitude, has no catalogue entry whatsoever.

There is a second, prettier version of the same limit in the channel codes. The FDSN source-identifier standard defines band code H as 80–250 samples per second and B as 10–80, both specified with a low-frequency corner at ten seconds; L is about one sample per second and V is a tenth of that. The automatic pickers that feed the catalogues run on the fast channels. DK.SCO streams BH, HH, LH and VH alike, so the data existed at Scoresbysund the whole time — but a 92-second period sits an order of magnitude outside the band the detection stack is pointed at. The recording is not the detection. The sensor is not the algorithm.

What popular coverage gets wrong

The received version runs like this: a mysterious signal of unknown origin shook the planet for nine days, seismologists were baffled, and the thing was so unclassifiable that it had to be filed as a "USO" — an unidentified seismic object. That framing ran almost everywhere in September 2024, from the American science monthlies to the British and Australian broadcasters to the wire pickups that followed them, and it survives in the search results today. It puts the mystery in the Earth.

Four things in the primary record contradict it.

First, the coinage is not evidence of institutional bafflement; it is a working group's joke about its own paperwork. UCL's own release on the paper states plainly that the signal "was dubbed a 'USO'" by the scientists who found it, and GEUS's Danish-language release says the same. It is shorthand for an unlabelled row, invented by the people who went on to label it.

Second, the source was not unknown for nine days. It was not even unknown for three. On 17 September 2023 — the day after the collapse — a passenger aboard the expedition cruise ship Ocean Albatros, a former member of the Sirius Dog Sled Patrol, noticed that the patrol's station on Ella Ø did not look as it should and contacted the Danish Joint Arctic Command. Arctic Command sent a Challenger aircraft and the inspection vessel Knud Rasmussen. The Navy's aerial photomosaic of the landslide zone is dated 19 September 2023. The physical event had a name and a photograph within seventy-two hours.

Third, the USGS catalogue was never silent about the landslide. It carried it as a landslide, correctly typed, from within the same review cycle. What it could not carry was the nine days of ringing that followed, and that is a different failure with a different cause.

Fourth — and this is the part the mystery register actively obscures — the hard problem was never detection. It was mechanism. Sixty-eight researchers from forty institutions in fifteen countries spent about a year establishing that a fjord could act as a resonator with a Q of several thousand and hold a seiche for nine days with no external driver. Kristian Svennevig of GEUS, the lead author, put the difficulty where it belongs: nobody had the slightest idea what caused the signal, and the mystery was solved by interdisciplinary effort. Thomas Forbriger of the Karlsruhe Institute of Technology, one of the directors of the very observatory whose record shows the tone at 29°, said only that the signal was totally different from those we know from earthquakes. Different is not undetectable. It is unenumerated.

The only instrument that could see the water

The seismometers heard the fjord. They could not see it. For that there was precisely one machine in the sky, and its limits are worth stating with affection.

SWOT — Surface Water and Ocean Topography, a joint NASA and CNES mission — launched on 16 December 2022 carrying KaRIn, a Ka-band radar interferometer working at a wavelength of 0.86 centimetres. It flies at about 890 kilometres, inclined 77.6°, and maps sea-surface height across a 120-kilometre swath split by a ±10-kilometre gap along the nadir track. Its fast-sampling one-day-repeat commissioning phase ran through the first half of 2023; by September it had settled into the science orbit, a 20.86-day repeat with ten-day sub-cycles. Monahan and colleagues at Oxford, working the pixel-cloud product, report over 300,000 individual measurements in a single pass, at 2.5 metres along-track and 10–70 metres across.

That is an extraordinary instrument, and it caught the fjord four times. Three usable passes followed the September event, at 0.5, 1.5 and 4.8 days — and large parts of the 17 and 18 September data were compromised by noise artefacts that had to be filtered by hand. The October event got one usable pass, at 0.5 days. From those snapshots the Oxford team estimated a maximum cross-channel slope of 1.83 ± 0.59 metres per kilometre in September and 1.37 ± 0.13 in October — roughly a two-metre tilt across the width of the fjord. NASA's Earth Observatory, working the 17 September pass against a 6 August baseline, puts the north side as much as 1.2 metres higher than the south.

And then the honest sentence, which we quote because it is the whole argument of this dossier in one line: SWOT data alone cannot estimate the total amplitude of the seiche, as the observations could have occurred at any phase of the oscillation period. A snapshot of a standing wave taken at an unknown moment in its cycle tells you the slope and not the height. The Oxford team recovered an initial amplitude of 7.9 metres only by marrying the satellite's spatial precision to the temporal precision of a seismic station 1,322.9 kilometres away at Alert in the Canadian Arctic — the water's shape from orbit, the water's clock from the ground.

What the prior art actually said

Seiches were not new to seismology in 2023. Walter, Olivieri and Clinton showed in the Journal of Glaciology in 2013 that calving-induced seiches in Greenland fjords produce narrow spectral peaks between roughly 0.002 and 0.007 hertz on broadband seismometers, resonating for many hours, and argued that at regional distance the seiche was often the only reliable evidence a calving had happened, because the higher-frequency energy did not clear the noise floor even inside 100 kilometres. Svennevig and colleagues state the tsunami-seiche case more narrowly: before 2023, only short-duration loading-induced tilt, under an hour, on seismometers within about 30 kilometres.

So the template existed at local range and nobody had looked for it at teleseismic range, because there was no reason to think a fjord could hold a tone for nine days. Once there was, the search became trivial. The team's own conference abstract, presented at the Seismological Society of America meeting in Anchorage in spring 2024, reports four previously unknown similar events since 2016, found by going back through archives that had been recording them all along. This is the ordinary, undramatic way instrument science advances: the data were never missing, only the query.

The numbers that do not agree

We are not going to smooth these. The distance from the rockslide to the Danish research base on Ella Ø is given as 72 kilometres in Science, 70 kilometres in the GEUS release, and about 65 kilometres in the team's own SSA abstract; the straight-line great circle from the published source coordinate to the island's gazetteer point is 62 kilometres, and the difference is presumably the along-fjord path the wave actually took. The seiche amplitude is 7.4 metres in Science, 5 metres in the earlier abstract, and 7.9 ± an inference in Nature Communications. The event magnitude is Ms_VX 4.3 in the USGS catalogue and 3.4 in the figure GEUS gave Greenland's public broadcaster in the week of the event. The author count is 51 by one reading of the paper's own byline and 68 in the releases issued by UCL, KIT, GEUS and Aarhus. Each of these is a measurement made against a different definition, which is the normal condition of a well-observed event and not a scandal.

Why the Silence Index is 84

Physical remoteness is near maximal: an uninhabited fjord inside a 972,000-square-kilometre national park, roughly 300 kilometres from Ittoqqortoormiit and its 345 residents, with the park's total year-round human population somewhere around forty, most of them military, meteorological or scientific. Access friction is close to as high as a lawful regime gets without an outright ban — twelve weeks' notice, DKK 1,000,000 of search-and-rescue cover made out to the Danish state, a DKK 200,000 bank guarantee, and an application fee that is not refunded if the answer is no.

What keeps it off ninety is footfall. Cruise ships pass east and west of Ella Ø every week in the ice-free season — the paper says so in as many words, while noting that none were in the fjords during the tsunami — and Greenland counted 18,957 expedition-cruise visitors in 2024 against 95,185 cruise passengers in total. This is not an unvisited coast; it is a lightly and seasonally trafficked one with a hard permit in front of it.

For the fourth input we weighted unresolved questions rather than night darkness, because this article reasons about the former throughout and because the honest answer to "what is still open here" is: a great deal. That pushes the score up rather than down, and 84 is where the four inputs land.

What the machine should have had

The Science paper's own recommendation is not a call for better seismometers. It is a call for networks of high-sampling-rate sea-level gauges in confined basins across a range of geographic settings — that is, for putting an instrument in the water, where the physics is, instead of inferring it from a tone in the crust and a radar snapshot taken at an unknown phase.

Something close to that already existed by accident. Søren Rysgaard's group at Aarhus University's Arctic Research Centre had installed a measurement station directly below the cliff that failed, days before it failed. The instruments survived. His comment on that is the least technical and most useful sentence anyone produced about the whole affair: it is frightening to think what could have happened if his team, or Sirius, or a large cruise ship had been in the area.

The institutional response has been slower than the science. The Government of Greenland's Expedition Office issued a notice dated 28 November 2025 urging extreme caution in Dickson Fjord west of Ella Ø and across the whole national park, treating the entire park as a high-risk area and stating that whether sailing restrictions will be introduced was still being clarified. The Association of Arctic Expedition Cruise Operators names Dickson Fjord alongside the Uummannaq area as a place with a history of large-scale landslides producing waves up to 200 metres and tells operators to avoid high-risk zones and conduct thorough risk assessments — sound advice with no instrument behind it, since Greenland's monthly slope-stability bulletins cover Uummannaq in the west and not this coast.

So the position in 2026 is this. A fjord in the world's largest national park rang for nine days at a frequency that four and a half thousand instruments recorded faithfully and no automatic system could name. The one machine capable of seeing the water passes on a twenty-day cycle and cannot tell you the height of a wave it photographs, only its tilt. The thing that finally identified the event was a passenger on a tourist ship who recognised a hut. There is no failure anywhere in that sequence, and there is no mystery in it either. There is only a very good detector shaped for a different question, which is what every detector always is.

Corner Codex — D-184

Place: Dickson Fjord, North-East Greenland National Park, Greenland
Rockslide source: 72.81°N, 26.95°W, from the source-side coordinate in Science (2024)
Event: 16 September 2023, 12:35:03 UTC — 11:35 local East Greenland time
Volume: 25 × 10⁶ m³ of rock and ice from a peak 1,200 m above the water
Wave: About 200 m backsplash, waves to 110 m, settling to a seiche of about 7.4 m
Signal: Monochromatic at 10.88 mHz — a 92-second period — for up to nine days
Reach: About 4,500 stations to 90° (10,000 km); DK.SCO at ~300 km, II.BFO at 29°
Catalogue entry: USGS us7000kx2u, Ms_VX 4.3, depth fixed 0 km, stations used: 0
Second event: 11 October 2023, same gully, half the amplitude — no USGS catalogue entry
Only water sensor: SWOT's KaRIn, on a 20.86-day repeat; four usable passes across both events
Protection: North-East Greenland National Park, 972,000 km², UNESCO biosphere from 1977
Access: Expedition permit, 12 weeks' notice, DKK 1,000,000 search-and-rescue cover
Silence Index: 84 / 100

The Traveler's Panel · checked 25 August 2026

Changing information. Prices, permits, seasons and road access change, sometimes at short notice. Everything here is a starting point for your own confirmation — not a quote, and not a guarantee. Verify with the operator and the relevant official body before you book.

What was heard	A monochromatic 10.88 mHz tone — 92-second period — for up to nine days
Where it was heard	About 4,500 stations to 90°; DK.SCO at ~300 km, II.BFO in Germany at 29°
What the catalogue holds	USGS us7000kx2u, Ms_VX 4.3, landslide, zero stations, no gap, no dmin
The October repeat	11 Oct 2023, same gully, half amplitude — no USGS catalogue entry at all
Who identified it	A cruise passenger and ex-Sirius man, 17 Sept 2023, by looking at a hut
Only water instrument	SWOT KaRIn, 890 km altitude, 20.86-day repeat, four usable passes
Legal status	North-East Greenland National Park, 972,000 km², Executive Order no. 7 of 1992
Getting in	Expedition permit, part one twelve weeks before departure, part two eight

Getting there — Iceland → Nerlerit Inaat (Constable Pynt) → chartered vessel or expedition ship north into Kong Oscar Fjord

There is no route to Dickson Fjord in the sense that word usually carries. The fjord is an uninhabited arm of the Kong Oscar Fjord system inside the North-East Greenland National Park, roughly 300 kilometres north of Ittoqqortoormiit and its 345 residents, and the park's entire year-round population is around forty people who are there because a government or a university put them there.
The permit machinery is the route. The Government of Greenland's Expedition Office in Nuuk requires part one of an application — fee receipt, detailed itinerary, route and camp map — twelve weeks before departure, and part two — insurance statement, expedition leader's documented experience, participant list, safety plan, and any radio, firearm or medical permits — eight weeks before. Cruise operators file their own category; individuals do not slip in under one.
Insurance is the real gate, and it is written for the state's convenience rather than yours: DKK 1,000,000 of search-and-rescue cover for the group, made out to the Danish state and the Government of Greenland, plus DKK 280,000 per person for evacuation south of 78°N (DKK 600,000 north of it), plus a DKK 200,000 bank guarantee against the cost of taking you to the nearest town if you decide you want to leave and are not actually in danger. The cover must hold whether the incident was accident or negligence.
Physically, the approach is by air to Nerlerit Inaat, 70°44′35″N 22°39′01″W, and then by water. Everything else — Mestersvig's airstrip, Daneborg, Zackenberg — is a state or research facility and not a public airport. Nothing in this fjord system has a road.
The park rules themselves are short and old: Executive Order no. 7 of 17 June 1992, amended in 1999. No disturbance of mammals or birds or their nests. All waste out. No flying below 500 metres, no aircraft landings and no depot air-drops without specific permission. The Sirius Patrol holds police authority over expeditions in the area, which is a fact worth internalising before you argue with anyone on a dog sled.
Since 28 November 2025 the Expedition Office has urged extreme caution in Dickson Fjord west of Ella Ø and across the whole park, treating the entire park as high-risk on the strength of the 2023 event, and has said that whether it will restrict sailing in the fjord is still being clarified. Check that page before anything else, because it is the one thing here that may have changed since we wrote.

Indicative costs — verify before booking

Application fee	DKK 4,000, categories A–C; DKK 25,000 for a research station. Non-refundable even if refused
Search-and-rescue cover	DKK 1,000,000 for the expedition group, to the Danish state
Evacuation cover	DKK 280,000 per member south of 78°N; DKK 600,000 per member north of it
Bank guarantee	DKK 200,000, against non-emergency transport to the nearest town
The fjord itself	Nobody should be pricing this. There is no landing site, no facility and nothing at Dickson Fjord to buy. What you are paying for is the state's cost of coming to get you
What we will not quote	Expedition cruise fares. Every figure we found came from operators selling the voyage, none from a published tariff

What to pack

A realistic assessment of your own rescue timeline. The Joint Rescue Coordination Centre in Nuuk covers all of Greenland, the nearest assets are a Challenger aircraft and an inspection vessel, and the Expedition Office's own warning names long response times as a specific hazard.
Everything for full self-sufficiency in continuous permafrost with no resupply. Zackenberg, the nearest research station of any size, runs 1 May to 31 October and is about 25 kilometres from Daneborg — neither is a shop.
Containers and a plan for taking every gram of waste out, because the 1992 executive order requires it and there is no alternative anywhere in 972,000 square kilometres.
A satellite communications set-up you have tested, and the radio permit if your kit needs one — the Expedition Office lists it among the part-two documents.
Sober expectations about the slopes. The 2023 failure came off a peak 1,200 metres above the water and put a wave 200 metres up the opposite side. AECO's advice to its own members is to avoid high-risk zones and assess before entering, and Greenland's monthly slope-stability bulletins do not cover this coast.

Where to stay

Nothing stays at Dickson Fjord. There is no settlement, no hut and no landing anywhere near the slide, and the nearest occupied structures are the Sirius Patrol's summer station and the Danish research base on Ella Ø, both of them working facilities and neither of them accommodation. In practice everyone who sees this coast sleeps aboard the vessel that brought them, which is exactly the arrangement that shapes the Silence Index here: the fjord is empty, but ships pass east and west of Ella Ø every week of the ice-free season, and the Science paper notes with visible relief that none of them were in the fjords on 16 September 2023. Further north, Daneborg at 74°18′16″N 20°13′20″W is the Sirius Patrol's Greenland base and Zackenberg, 25 kilometres away, is Aarhus University's monitoring station, open May to October under Greenland Government ownership. Both are closed to visitors. The nearest place with beds you can book is Ittoqqortoormiit, about 300 kilometres south, and even that is a settlement of 345 people at the mouth of the largest fjord system on Earth rather than a base for anything.

Safety & responsible travel

Do not repeat the 'mystery signal' framing. The instruments recorded the event faithfully from the first minute; what failed was an automatic classification stack built for impulsive arrivals. Saying the Earth baffled science gets the causation backwards and makes the actual engineering lesson invisible.
Treat the whole park as a hazard area rather than treating Dickson Fjord as the one bad fjord. That is the Expedition Office's own position since November 2025, and the Science team's conference abstract identified four previously unrecognised similar events since 2016.
If you are on an expedition ship, ask whether the operator is an AECO member and whether it has read the rockslide addition to the association's operational guidelines, which names this fjord explicitly. It is a fair question and an easy one to answer.
Leave research instruments alone. Automated stations run year-round through these fjords and one of them was sitting directly beneath the cliff that failed. They survived, and their data is part of why anyone knows what happened.
Do not go looking for the slide scar as a sight. It is a slope that recently discharged 25 million cubic metres of rock and ice into 540 metres of water and then did something similar again three weeks later.

Nearby, and quieter

Ella Ø, 62 kilometres out-fjord on our arithmetic and 72 by the paper's reckoning of the water route, is where the 2023 wave came ashore four metres high and ran eighty metres inland, and it is the only place in this story where the tsunami left marks a person could photograph from the ground. South is Mestersvig, a two-person research base with an airstrip at 72°14′N 23°55′W, and beyond it Nerlerit Inaat, the airport at 70°44′35″N 22°39′01″W that is the region's only real door. The whole of it sits inside the North-East Greenland National Park — 972,000 square kilometres in the World Database on Protected Areas, designated in 1974, a UNESCO Man and the Biosphere reserve since 1977, and the largest protected area on the planet by a margin that makes comparison pointless. Greenland counted 18,957 expedition-cruise visitors in 2024 across the entire country; the number who set foot anywhere in this park is a small fraction of that, and nobody publishes it.

Getting there, in order

1Iceland
↓
2Nerlerit Inaat (Constable Pynt)
↓
3chartered vessel or expedition ship north into Kong Oscar Fjord

Each leg is a separate booking and a separate chance to be weathered out. Build spare days into the chain, not onto the end of it.

Where it is

Dickson Fjord and Ella Ø, North-East Greenland National Park · 72.8300°N 26.3000°W · Open in Google Maps · OpenStreetMap

Key locations

Dickson Fjord rockslide source — the source-side coordinate published in Science (2024); the peak itself, Hvide Støvhorn, has no gazetteer entry we could find
72.8100°N 26.9500°W
Dicksons Fjord — the single GeoNames record for the fjord, 72°50′00″N 26°50′00″W
72.8333°N 26.8333°W
Ella Ø — the Danish research base and Sirius summer station; 4 m runup, 80 m inundation
72.8478°N 25.0556°W
Mestersvig — research base and airstrip, population 2; about 120 km from the slide
72.2333°N 23.9225°W
Nerlerit Inaat Airport — Constable Pynt, elevation 14 m — the region's only public airport
70.7431°N 22.6503°W
Ittoqqortoormiit — population 345; the settlement at the DK.SCO station, ~312 km from the slide
70.4844°N 21.9619°W
Daneborg — Sirius Dog Sled Patrol base and meteorological station, 44 m elevation
74.3044°N 20.2222°W

Coordinates are for orientation and are accurate to roughly the width of the feature named, not to a doorway. Several of these points are inside protected areas or need a permit, a guide or a boat — read the access notes above before you plan a route to any of them.

Sources & further reading — checked 25 August 2026

Svennevig and 67 others — 'A rockslide-generated tsunami in a Greenland fjord rang Earth for 9 days', Science 385, 1196–1205 (13 September 2024), author PDF. The full paper: 25 × 10⁶ m³ rock-ice avalanche from Hvide Støvhorn 1,200 m above the fjord at 72.81°N 26.95°W on 16 September 2023 at 12:35:03 UTC; ~200 m backsplash, 110 m waves, 7.4 m seiche; a monochromatic 10.88 mHz signal on ~4,500 stations to 90°; DK.SCO at 300 km and II.BFO at 29°; a 540-m-deep, 2.7-km-wide fjord; the 11 October 2023 repeat at half amplitude; and the note that cruise ships pass east and west of Ella Ø weekly but none were in the fjords
Monahan, Tang, Roberts and Adcock — 'Observations of the seiche that shook the world', Nature Communications 16, 4777 (3 June 2025). The SWOT analysis: SWOT_L2_HR_PIXC version 2.0 pixel clouds at 2.5 m along-track and 10–70 m across, over 300,000 measurements per pass; a 20.86-day repeat orbit with 10-day sub-cycles; usable passes at 0.5, 1.5 and 4.8 days in September and one at 0.5 days in October; maximum cross-channel slopes of 1.83 ± 0.59 and 1.37 ± 0.13 m/km; an initial amplitude of 7.9 m; station II.ALE at 1,322.9 km; and the statement that SWOT alone cannot give the seiche's total amplitude because a pass could fall at any phase of the cycle
US Geological Survey — FDSN event service, query for 70–76°N, 32–20°W, 16–17 September 2023. A single event returned for the box and day: us7000kx2u, magnitude 4.3, magnitude type ms_vx, 72.797°N 27.115°W, depth 0 km, place 'eastern Greenland', status reviewed
US Geological Survey — FDSN event service, full record for event us7000kx2u. The catalogue entry's quality fields: origin 2023-09-16T12:35:00Z, event type 'landslide', title 'M 4.3 Landslide - eastern Greenland', nst 0, gap null, dmin null, rms 0.28, tsunami flag 0, depth fixed at 0 km, last updated 18 November 2023 — a magnitude with no location solution behind it
US Geological Survey — FDSN event service, query for 70–76°N, 32–20°W, all of October 2023. A count field of zero: the second Dickson Fjord landslide-tsunami of 11 October 2023, documented in Science as originating from the same gully with the same fundamental period, has no entry at all in the USGS catalogue
International Seismological Centre — The ISC Bulletin. The Bulletin described as the definitive record of the Earth's seismicity from 1900; built from agency hypocentres, arrival times and focal mechanisms; automatic grouping scored on azimuthal gap, station distribution and phase count; and the reviewed Bulletin running about 24 months behind real time, currently to 1 August 2024
International Federation of Digital Seismograph Networks — SEED channel codes, band code table. The band-code definitions used in this dossier: H is 80–250 samples per second and B is 10–80, both specified with a corner period of ten seconds or longer; L is about 1 sample per second, V is 0.1 to 1, U is 0.01 to 0.1 — the bands a 92-second signal actually lives in
US Geological Survey — Fact Sheet 2011-3021, 'The Global Seismographic Network'. More than 150 stations in over 80 countries; seismometers described as extremely sensitive across a wide band, recording oscillations with periods of thousands of seconds as well as motions with periods under a tenth of a second, with high fidelity; data free and open
GEOFON, GFZ Potsdam — DK, the Danish Seismological Network. The network operated by GEUS and archived at GFZ; station SCO at Scoresbysund running 1999–2001 and again from 6 August 2010 to the present, streaming BHE/BHN/BHZ, HHE/HHN/HHZ, LHE/LHN/LHZ and VHE/VHN/VHZ — the slow channels on which a 92-second tone is recorded
GFZ — Black Forest geomagnetic observatory record. The observatory's geographic position given as 48.33°N, 8.33°E at Wolfach — the coordinate used for this dossier's great-circle calculation of 3,232 km, or 29.07°, from the rockslide
Karlsruhe Institute of Technology, Geophysical Institute — Black Forest Observatory. The observatory at Heubach 206, Wolfach, jointly run by KIT and the University of Stuttgart since 1971, with Thomas Forbriger, Rudolf Widmer-Schnidrig and Walter Zürn named as its directors — two of whom are co-authors of the Science paper
Karlsruhe Institute of Technology, via EurekAlert — 'How a Fjord in Greenland generated a global seismic signal' (13 September 2024). Thomas Forbriger of KIT quoted that the signal was totally different from those known from earthquakes; Kristian Svennevig on the volume of falling material; 68 researchers from 40 institutions in 15 countries; and detection at nearly 20,000 km, in Antarctica
University College London — 'Climate-change-triggered landslide caused Earth to vibrate for nine days' (12 September 2024). The statement that the signal was dubbed a 'USO' — unidentified seismic object — by the scientists who found it; that it contained only a single vibration frequency, so standard seismic analysis could not pin down its origin; Dr Stephen Hicks quoted on being completely baffled; and the Danish military sailing into the fjord days afterwards to photograph it
GEUS — 'Fjeldskred udløst af klimaændringer fik Jorden til at vibrere i 9 dage' (12 September 2024, Danish). The survey's own account: the USO label; Kristian Svennevig quoted that nobody had the slightest idea what caused the signal and that it was solved only by a massive interdisciplinary and international effort; and Ella Ø given as 70 km from the source, against 72 km in the paper
Aarhus University, Arctic Research Centre — 'Climate change triggered a landslide in Northeast Greenland…' (12 September 2024). Professor Søren Rysgaard's group having installed a measurement station directly beneath the 1.2-km cliff days before it collapsed, and the instruments surviving; his remark that it is frightening to think what could have happened had his team, Sirius or a large cruise ship been in the area; and the seiche described as oscillating about 10,000 times over nine days
GEUS publications — Svennevig and others, 'An Extraordinary Tsunamigenic Rockslide Into a Greenland Fjord Rang the Earth for 9 Days', SSA Annual Meeting 2024, Anchorage. The team's earlier conference abstract, giving the damaged base as about 65 km away and the stabilised seiche as 5 m rather than the 7.4 m later published, and reporting four previously unknown similar events identified since 2016
US Geological Survey publications catalogue — 'A rockslide-generated tsunami in a Greenland fjord rang Earth for 9 days'. The full author list as catalogued by the USGS, running to 68 names beginning Svennevig, Hicks, Forbriger, Lecocq and Widmer-Schnidrig, with journal, year and DOI 10.1126/science.adm9247
Walter, Olivieri and Clinton — 'Calving event detection by observation of seiche effects on the Greenland fjords', Journal of Glaciology (2013). The prior art: three Streckeisen STS-2 broadband stations within 100 km of calving fronts in north-west Greenland, 2009–2011; narrow spectral peaks between about 0.002 and 0.007 Hz; seiches resonating for many hours; and the finding that surface-wave and high-frequency energy often failed to clear the noise floor even inside 100 km
NASA Earth Observatory — 'Sizing Up a Greenland Tsunami'. SWOT water-elevation data over Dickson Fjord from 17 September 2023 against a 6 August 2023 baseline, showing the north side up to 1.2 m higher than the south; SWOT flying at about 900 km; and Josh Willis of JPL on seeing the shape of the wave being something impossible before SWOT
CNES / AVISO+ — SWOT mission parameters. Launch on 16 December 2022; orbit altitude 890 km; inclination 77.6°; a 21-day repeat cycle with a 1-day cycle during the CalVal phase; KaRIn operating at a wavelength of 0.86 cm; and a 120-km swath with a ±10-km nadir gap
NASA Earthdata — 'Preparing for Oceans of Data' (SWOT calibration and validation). The engineering checkout phase running January to March 2023 and in-situ validation through spring and into June 2023, with the first 90 days of the science mission flown fast-sampling — establishing that the daily-revisit phase preceded September 2023, though no exact transition date is given on this page
KNR, Greenland's public broadcaster — 'Mindre tsunami i nationalparken: Siriuspatruljens station er ramt' (21 September 2023). The contemporaneous account: a former Sirius patrol member aboard a cruise ship noticing the Ella Ø station looked wrong and alerting Arctic Command; the Challenger aircraft and the inspection ship Knud Rasmussen sent to investigate; Arctic Command's statement that parts of a mountain appeared to have fallen into the water; equipment washed to sea from an unoccupied station; and a magnitude of 3.4 attributed to GEUS
Danish Armed Forces — Arktisk Kommando, tasks and units. The command headquartered in Nuuk; its sovereignty, fisheries-inspection, environmental and hydrographic tasks; the Joint Rescue Coordination Centre coordinating rescue in Greenland; the inspection vessels Triton and Hvidbjørnen and the air force's Challenger aircraft; and the Sirius sledge patrol operating in the north-east national park
Danish Armed Forces — Om Slædepatruljen Sirius. Sirius founded in 1950 as Operation Resolut; its sovereignty enforcement, search and rescue, police authority over expeditions and environmental enforcement in the North-East Greenland National Park; trained by Jægerkorpset and deployed under Arctic Command
Government of Greenland, Expedition Office — 'Extreme Caution in Dickson Fjord and the National Park in North and East Greenland' (28 November 2025). Naalakkersuisut urging extreme caution in Dickson Fjord west of Ella Island and across the national park; the 16 September 2023 landslide-tsunami and nine days of wave activity cited as the basis; the whole park treated as high-risk; and the statement that whether restrictions on sailing in Dickson Fjord will be introduced was still being clarified
Government of Greenland, Expedition Office — National Park in North and East Greenland, rules. Executive Order no. 7 of 17 June 1992 as amended by Executive Order no. 16 of 5 October 1999; the ban on disturbing mammals and birds and their nests; the requirement to remove all waste; and the need for special permission to fly below 500 m or to land aircraft and lay depots
Government of Greenland, Expedition Office — Insurance and bank guarantee. DKK 1,000,000 of search-and-rescue cover for the whole expedition group; DKK 600,000 per member north of 78°N inside the national park and DKK 280,000 per member south of it; a DKK 200,000 bank guarantee for non-emergency transport; and cover required regardless of whether an incident arises from accident or negligence
Government of Greenland, Expedition Office — Before you apply. A non-refundable application fee of DKK 4,000 for categories A, B and C and DKK 25,000 for category D research stations, covering administrative processing only, with separate permits required where activities or groups differ
Government of Greenland, Expedition Office — Deadlines. Part one of the application, with the fee receipt, itinerary and route or camp map, due twelve weeks before departure; part two, with insurance statement, leader experience, participant list, safety plan and any radio, firearm or medical permits, due eight weeks before
AECO — Operational Guidelines, rockslide addition (updated 13 June 2025). The cruise industry's own guidance naming the Uummannaq area and Dickson Fjord as places with a history of large-scale landslides generating waves up to 200 m, telling operators to avoid high-risk zones and conduct thorough risk assessments, and pointing to the Greenland Mineral Authority's monthly stability updates for Uummannaq
Visit Greenland — Explore Northeast Greenland National Park. The park at 972,000 km², the world's largest, designated in 1974 with UNESCO Man and the Biosphere status from 1977; approximately 40 inhabitants, chiefly military, research and meteorological personnel; and permits to be applied for at least 12 weeks before departure
Protected Planet / WDPA — North-East Greenland, WDPA ID 2065. The World Database on Protected Areas record giving a reported area of 972,000.0 km² of terrestrial and inland-waters protected area in Greenland, with no management-effectiveness evaluation recorded
Visit Greenland — Tourism Satellite Account 2024. An estimated 149,000 international tourists in 2024, 64% of them cruise tourists; 76,228 conventional and 18,957 expedition cruise passengers for a total of 95,185; tourism at 4.9% of GDP and 6.2% of employment
INTERACT station catalogue — Zackenberg Research Station. The Aarhus University station at 74.47°N 20.58°W in the southern part of the national park, owned by the Government of Greenland, open 1 May to 31 October, running ClimateBasis, GeoBasis, BioBasis, MarineBasis and GlacioBasis programmes across about 500 km², about 25 km from Daneborg
GeoNames gazetteer — Dicksons Fjord, East Greenland. The single gazetteer record for the name: Dicksons Fjord, feature type fjord, 72°50′00″N 26°50′00″W in Østgrønland
GeoNames gazetteer — Ella Ø. Ella Ø, also given as Ella Island, feature type island, 72°50′52″N 25°03′20″W — the point used for this dossier's 62 km great-circle figure to the rockslide source
GeoNames gazetteer — Ittoqqortoormiit. Ittoqqortoormiit, populated place, population 345, at 70°29′04″N 21°57′43″W, with its heliport and the Ittoqqortoormiit Kommuniat division — the settlement nearest the DK.SCO station
GeoNames gazetteer — Nerlerit Inaat Airport. Nerlerit Inaat Airport, 70°44′35″N 22°39′01″W at 14 m elevation — the airfield through which the whole Scoresby Sund and Kong Oscar Fjord region is reached
GeoNames gazetteer — Mestersvig. Two records: Mestersvig as a scientific research base with a population of 2 at 72°14′00″N 23°55′21″W, and the Mestersvig airfield at 72°14′15″N 23°55′58″W
GeoNames gazetteer — Daneborg. Daneborg, scientific research base and meteorological station, 74°18′16″N 20°13′20″W — the Sirius Patrol's Greenland base, about 269 km from the rockslide on our arithmetic

Official and scientific bodies are the authority for safety, regulatory and scientific claims. Commercial operators are used only for practical detail such as tour length, meeting points and indicative pricing — never for a safety or scientific statement.

How this was made

Desk research. Compiled from official bodies, scientific and conservation sources, established journalism and licensed operators. Not based on a personal visit.

Editor’s method note — what we checked, corrected and could not verify

Desk research only. Nobody from this publication has been to Dickson Fjord or anywhere in the North-East Greenland National Park, and this dossier makes no claim to have seen any of it.
Every URL cited was opened and read before citation. Sources that could not be opened are listed at the end of this note and are not cited anywhere in the piece.
The catalogue findings are ours and are reproducible. We queried the USGS FDSN event service directly for a box covering 70–76°N and 32–20°W: one event on 16 September 2023 (us7000kx2u, ms_vx 4.3, type 'landslide', nst 0, gap null, dmin null), and zero events across the whole of October 2023. We did not take these from any secondary account. Catalogues are revised, and a later query may return something different.
Distances marked as ours are great-circle calculations from the rockslide coordinate of 72.81°N 26.95°W published in Science to gazetteer or observatory coordinates, on a 6,371.0088 km mean Earth radius. They give about 312 km to Ittoqqortoormiit, 3,232 km (29.07°) to the Black Forest Observatory at 48.33°N 8.33°E, and 62 km to Ella Ø. Where a published figure differs — 300 km for DK.SCO, 72 km for Ella Ø — we print both and name both.
Ella Ø's distance from the source is given as 72 km in Science, 70 km by GEUS and about 65 km in the team's own SSA conference abstract. The seiche amplitude is 7.4 m in Science, 5 m in the abstract, and 7.9 m as inferred by Monahan and colleagues. The event magnitude is 4.3 (Ms_VX) in the USGS catalogue and 3.4 in the figure GEUS gave KNR in the week of the event. We reconcile none of these.
The author count is inconsistent across sources we read. The USGS publications catalogue lists 68 names, and the UCL, KIT, GEUS and Aarhus releases all say 68 researchers from 40 institutions in 15 countries. One machine reading of the paper's own byline returned 51. We use 68 and flag the discrepancy rather than silently choosing.
The claim that the detection stack is pointed away from a 92-second signal is an inference we make from two cited primary documents — the FDSN band-code table and the GEOFON channel listing for DK.SCO — and not a statement quoted from either. Neither the Science paper nor any agency page we opened sets out the automatic-catalogue failure in those terms. UCL's release says only that standard analysis could not pin down the origin because the signal lacked frequency diversity. Treat the channel-band argument as our reading, plainly labelled.
We do not use the word 'mystery' in our own voice about this event, and we do not treat 'USO' as a scientific classification. Both the UCL and GEUS releases describe it as a label the investigating scientists themselves applied to an unidentified row.
The Expedition Office notice urging extreme caution in Dickson Fjord carries a date of 28 November 2025 on the page as we read it, and the same site's news index carries an item of the same date. We report the date as printed and note that we could not corroborate it against a second government source.
Sources we could not read, and therefore did not cite: the Nature news feature that first reported the 'USO' coinage, which is paywalled beyond its opening lines; the DR.dk interview with the Ella Ø base staff, which returned 403; the IRIS/EarthScope MDA station page for DK.SCO and the IRIS FDSN station web service, both robots-blocked; the Project IDA FDSN station page for II.BFO, robots-blocked; the NOAA NCEI Global Historical Tsunami Database event record 5911 and its API, which failed on robots and then timed out; the Greenland Mineral Resources Authority's landslide monitoring and situation-assessment pages at govmin.gl, which failed certificate verification; and the Naalakkersuisut news item on communication failures around the Dickson Fjord tsunami, robots-blocked in both Danish and English.
Things we could not verify: how many times SWOT's swaths actually cross Dickson Fjord in a calendar month, no source we opened publishing a revisit count for that latitude and location; the exact calendar dates and times of the four usable SWOT passes, which Monahan and colleagues give only as intervals after the event; the exact date SWOT left its one-day-repeat calibration orbit for the 20.86-day science orbit; whether any automatic detection system anywhere flagged the 10.88 mHz tone before a human looked, which no source we opened states either way; the true source-to-Ella-Ø distance, three official figures being in circulation; whether sailing restrictions in Dickson Fjord have since been introduced, the Expedition Office's own notice saying only that the question was being clarified; the identity and coordinates of Hvide Støvhorn, which appears in the Science paper but has no GeoNames entry we could find; how many of the four previously unknown seiche events reported in the 2024 conference abstract have since been published in full; and the current condition of the remaining unstable slope above the fjord, which is not covered by Greenland's monthly stability bulletins.

AI is used to draft this text, and a human checks it. Every figure, date and quotation above was verified by Tony Hall against the source cited for it in the reference list; where sources disagree, the disagreement is shown rather than resolved, and where a claim could not be confirmed it is either marked unverified or left out. The illustrations are AI-created and labelled as such — no photograph on this site is presented as a record of a real place. Full disclosure · Editorial policy.

Access details change — recheck before you travel. Found an error? Tell us — corrections are dated and shown in place.

Nothing Was Missing From the Record: Dickson Fjord and the Shape of the Detector