The Wave That Rang Earth Like a Bell — And We Finally Saw It

On September 16, 2023, at 12:35 UTC, something strange happened in a remote Greenland fjord. Seismographs around the world began picking up an unusual signal — not a typical earthquake jolt, not a volcanic rumble, but a smooth, repeating pulse arriving exactly every 92 seconds. Clean. Monotonous. Persistent.

And it didn’t stop.

For nine days, seismometers on six continents registered this ghostly, almost musical vibration, emanating from somewhere in the Arctic. Seismologists were baffled. The signal had the wrong frequency for an earthquake, the wrong character for any known geological process, and the wrong duration for anything they’d seen before. Then, a month later, it happened again — another week-long, 92-second heartbeat, from the same location.

For over a year, scientists theorized about what had caused it. Now, a team from Oxford has done something no one had managed before: they’ve actually seen the wave — using a next-generation satellite that can photograph the shape of the ocean’s surface from space.

The Cascade That Started It All

To understand what the wave was, you first have to understand what caused it.

Dickson Fjord is a remote, 35-kilometer arm of seawater cutting through the mountains of northeastern Greenland, in the Northeast Greenland National Park — one of the most isolated places on Earth. Like many Arctic fjords, it’s flanked by steep mountain slopes that were, for millennia, buttressed by glacial ice. But Greenland’s glaciers are retreating, and as the ice thins and pulls back, the rock walls it once pressed against are left unstable, like a stack of books with its bookend removed.

On that September morning, the bookend gave way.

A block of metamorphic rock — 150 meters thick, 480 meters wide, 600 meters long — detached from the mountainside. Together with a glacier it struck on the way down, the avalanche amounted to roughly 25.5 million cubic meters of rock and ice (about 10 times the volume of the Empire State Building), reaching a peak velocity of 42 meters per second before slamming into the fjord below.

The impact generated a tsunami. But not a gentle coastal wave — a 200-meter-high wall of water surging up the opposite side of the fjord, with wave runup observable as far as 100 kilometers away. The force of it didn’t just crash on the shore and dissipate. Instead, because the fjord is narrow and enclosed, the wave bounced. It reflected off the fjord walls. It became trapped.

What you get when a wave bounces endlessly back and forth in a confined basin is called a seiche — a standing wave, a rhythmic oscillation of the entire water body, like the water in a bathtub when someone disturbs it. The frequency of a seiche is determined by the geometry of the basin: its length, depth, and shape. For Dickson Fjord, the math worked out to a period of almost exactly 92 seconds.

That oscillating water — thousands of tons of it, sloshing transversely across the fjord — applied a rhythmic force to the bedrock. The simulations, published in Science in September 2024, showed the seiche was applying a maximum force of roughly 5 × 10¹¹ Newtons — about 100 billion pounds — pushing sideways against the fjord walls with every single oscillation. And because the fjord is long, and because the water’s momentum kept the seiche going with remarkably little decay, that force transferred efficiently into the ground and propagated outward as seismic waves.

Earth rang like a bell. For nine days.

The Science paper by Kristian Svennevig and colleagues (DOI: 10.1126/science.adm9247) laid out this chain of events in meticulous detail, combining satellite imagery, global seismic networks, and tsunami simulations. It was a triumph of geophysical detective work. But there was one thing missing: direct observation of the seiche itself. No one had seen it. No one had been in the fjord. The Danish military vessel that reached the area three days into the first event saw nothing unusual — the seiche had already attenuated below visual threshold, though its seismic signature continued.

The theory was solid. The math was right. But the actual wave had slipped away unseen.

A Satellite Built for Exactly This Moment

The SWOT mission — the Surface Water and Ocean Topography satellite — launched in December 2022 aboard a SpaceX Falcon 9. Built jointly by NASA and France’s CNES, with contributions from the Canadian and UK space agencies, it carries an instrument unlike any previous ocean-observing satellite: the Ka-band Radar Interferometer, or KaRIn.

KaRIn consists of two radar antennas mounted at the ends of a 10-meter boom extending from each side of the spacecraft. By comparing the return signals from both antennas, it can triangulate the height of the water surface below — not just along the narrow track directly beneath the satellite (as older altimeters could do), but across a 50-kilometer-wide swath of ocean, at a resolution of 2.5 meters. It’s the difference between a single thread of sonar and a wide-angle camera.

Previous satellites had tried to find the seiche and failed. The problem was geometry and timing: conventional altimeters sample only a thin line beneath their track, and revisit intervals are too long to catch a rapidly-decaying oscillation in a narrow fjord in a remote corner of Greenland. SWOT’s wide-swath interferometry changed the calculation.

Thomas Monahan, a doctoral student at Oxford’s Department of Engineering Science, and his colleagues realized that SWOT had overflown the fjord during the aftermath of both events. They set about analyzing the data with new techniques — including Bayesian machine learning to fill in the temporal gaps and connect satellite snapshots to seismic records — and found what they were looking for.

The Shape of an Earth-Shaking Wave

The results, published June 3, 2025, in Nature Communications (DOI: 10.1038/s41467-025-59851-7), are as clear as they are striking. SWOT’s elevation maps showed the fjord with cross-channel slopes — one side measurably higher than the other, with height differences of up to two meters. And crucially, when the researchers compared maps from different overflights, the slopes reversed: what was high on one side became low, and vice versa. The water was sloshing. Back and forth, back and forth, exactly as the seiche theory predicted.

The team ruled out tides, winds, and other oceanographic processes as explanations — none of them matched the timing or spatial pattern. What remained was the seiche, confirmed for the first time by direct observation. Using Bayesian methods to combine the altimetry with seismic data, they independently estimated the seiche’s initial amplitude at 7.9 meters — remarkably close to the 7-meter estimate from the Science paper’s simulations.

“SWOT is a game changer for studying oceanic processes in regions such as fjords which previous satellites struggled to see into,” Monahan wrote. It’s a satisfying scientific closure: a theory built from seismographs half a world away from the event, finally confirmed by a satellite that could look directly at the water.

What This Means for the World

There’s something almost beautiful about the chain of causality here. A warming climate thins a glacier. That glacier’s absence destabilizes a mountain. The mountain falls, 25 million cubic meters at once. A wave 200 meters high crashes into a fjord no human witnessed. The wave begins to slosh, trapped in its stone basin. The sloshing applies a force to bedrock. The bedrock shudders. And that shudder propagates outward through the mantle and crust until, 4,000 kilometers away, a seismometer in Germany, or Japan, or Antarctica, twitches every 92 seconds for nine consecutive days.

Planet Earth is connected in ways we’re still learning to read.

The Svennevig Science paper called this a “cascading, hazardous feedback between the cryosphere, hydrosphere, and lithosphere” — a five-dollar phrase that means, in plain language, that the melting of ice can set off a chain of events involving ocean, rock, and the deep structure of the Earth itself. These feedback loops are becoming more frequent and more powerful as Arctic temperatures rise — the region is warming roughly four times faster than the global average.

Dickson Fjord isn’t unique. Greenland has seen several large tsunamigenic landslides in recent years, including major events in the Karrat and Kalaallit Nunaat regions. Norway has documented similar instabilities in its fjords. Alaska’s steep, glacially-carved coastline harbors numerous potential failure zones. What the 2023 Dickson event showed — and what the 2025 Nature Communications paper confirmed — is that these events can produce effects measurable on a planetary scale, and that we now have the tools to detect and interpret them.

The Instruments That Make It Possible

There’s another thread running through this story worth pulling on: the emerging power of space-based Earth observation.

For decades, studying events like this required researchers to either be in the right place at the right time (essentially impossible for a remote Greenland fjord) or to reconstruct events indirectly from networks of instruments scattered around the world. Both approaches have limits. Ground networks miss events in data-sparse regions; indirect reconstruction involves uncertainty at every step.

SWOT represents a different approach: a satellite that can map the height of virtually every significant water body on Earth, repeatedly, with enough precision to catch a two-meter slope difference in a narrow fjord in the Arctic. It launched only a few months before the Dickson event, and it was ready.

The researchers note that even SWOT has limitations — its revisit intervals mean it doesn’t continuously watch any given location, and the temporal sparsity required them to develop new analytical methods to reconstruct the wave’s behavior between satellite passes. But those methods, combining remote sensing with Bayesian inference and seismic data, form a template that will serve science well as extreme events become more common.

“Climate change is giving rise to new, unseen extremes,” Monahan said. “These extremes are changing the fastest in remote areas, such as the Arctic, where our ability to measure them using physical sensors is limited. This study shows how we can leverage the next generation of satellite Earth observation technologies to study these processes.”

A Mystery Closed, and One Still Open

The second seismic event — the one that began about a month after the first — remains a partial mystery. A second large landslide did occur in the same general area, but the details of exactly where it happened and how it related to the first event are still being investigated. The first event may have further destabilized the slopes, making a second failure more likely. Or the two events may simply reflect how fragile these glacially-debuttressed walls have become.

What’s no longer mysterious is the fundamental mechanism: water, trapped in a fjord, sloshing for days, shaking the entire Earth. It sounds like something from a science fiction novel — a hidden wave that hums to itself in the Arctic dark, felt by seismometers on six continents without a single human observer present to see it.

It took an orbiting radar instrument, a team of engineers and oceanographers in Oxford, and some elegant Bayesian statistics to finally catch it in the act.

The wave was always there. We just needed eyes sharp enough to see it.

---

The key papers:

• Svennevig et al. (2024). A rockslide-generated tsunami in a Greenland fjord rang Earth for 9 days. Science, 385, 1196–1205. DOI: 10.1126/science.adm9247
• Monahan et al. (2025). Observations of the seiche that shook the world. Nature Communications, 16. DOI: 10.1038/s41467-025-59851-7