For a place so remote that almost nobody will ever see it from the ground, Petermann Glacier produced an unusually well-observed rupture. On 3 August 2026, European radar satellites recorded pronounced deterioration through the centre of its floating tongue. By 4 August, a 76-square-kilometre slab had separated from the eastern side and become an ice island.
The new iceberg is roughly Manhattan-sized in area and may be as much as 150 metres thick, according to the European Space Agency’s account of the event. It is Petermann’s largest loss of floating ice since 2012 and, by ESA’s assessment, the Arctic’s most significant calving event since 2020.
That is a large piece of ice, but the familiar city comparison needs one guardrail. “The size of Manhattan” describes surface area, not shape, mass or the amount of ice above the water. Most of a thick tabular iceberg sits out of sight below the surface.
The final break took a day, but the failure took years
The clean before-and-after satellite images can make this look like a one-day event. Mechanically, it was the end of a much longer story.
Researchers supported in part by ESA’s ARCTEX project have monitored Petermann since 2019. The group brings together scientists from the universities of Ottawa, Stirling, Lancaster and Leeds with the Canadian Ice Service. Over time, they documented expanding fractures and increasing deformation in the floating tongue.
Radar interferometry acquired in April 2026 already showed cracks and motion inside the ice. On 3 August, Sentinel-1 imagery showed deterioration along the tongue’s centreline. The eastern slab was separate the following day.
So “broke free on 4 August” identifies the moment the island became distinct. It does not mean an intact sheet of ice became unstable without warning overnight. Glacier calving is often the abrupt final frame of a slow structural change.
What exactly broke away?
Petermann is a marine-terminating glacier in northwest Greenland. Ice grounded on land flows down a broad channel toward Petermann Fjord. Near the coast it crosses the grounding line, loses contact with the bed and continues over the sea as a floating tongue.
The detached object came from that floating section. It is called a tabular iceberg because it retains the broad, relatively level form of the ice tongue, rather than the peaked profile people often imagine when they hear “iceberg.” In the Arctic, an exceptionally large, flat piece like this is also called an ice island.
The image is radar, not an ordinary photograph. Sentinel-1 sends microwave energy toward the surface and measures the return. That lets it observe through cloud and during polar darkness, two advantages that are hard to overstate when the target is a glacier near Greenland’s northern coast.
The two larger pieces are possibilities, not appointments
The new island may be only the first large loss in this phase. ESA says existing rifts outline two further sections with estimated surface areas of about 97 and 87 square kilometres. Either would be larger than the piece released on 4 August.
Those estimates explain the concern in the headline, but “could detach” is not the same as “will detach soon.” A mapped rift is not a countdown. Cracks can lengthen, branch, slow or stall as the distribution of stress changes. Tides flex the floating tongue, ice continues to flow from upstream, and contact with the sides of the fjord can redistribute forces.
In plain terms, the fractures have already divided the tongue into plausible future slabs. Scientists can identify their approximate boundaries and areas. They cannot yet assign either one a reliable calving date, and it would be premature to add 97 and 87 square kilometres to the ice already lost.
This iceberg will not directly lift the sea
There is another boundary worth drawing carefully. The 76-square-kilometre section was already floating before it separated. Its weight was therefore already displacing seawater. Breaking it off, or watching it melt later, does not directly produce the sea-level effect that occurs when grounded land ice enters the ocean.
The National Snow and Ice Data Center explains the distinction succinctly: floating shelves do not directly raise sea level when they break up, but their loss can matter indirectly if the glaciers feeding them accelerate.
Floating tongues can act as a brake. Friction along their margins and contact with fjord walls or shallow features create resistance to the flow behind them. Remove enough of that resistance and grounded ice may move faster toward the ocean. That land ice does add to sea level.
It would still be too simple to say that every calving event releases the brake by the same amount. The buttressing supplied by different parts of an ice tongue varies. Researchers now need to measure Petermann’s response to this particular loss, especially changes in flow speed and strain upstream of the new front.
This is the same reason a 2024 ScienceBlog report on severe thinning at Greenland’s 79° N-Glacier focused on what was happening beneath and behind the floating tongue, rather than treating the visible edge as the whole system.
Petermann’s front has been redrawn before
Petermann has a long record of producing ice islands. In August 2010, it released a piece measured by NASA at about 251 square kilometres, roughly four times the area of Manhattan. Attention then turned to its likely route toward Nares Strait, a path that is relevant again now.
A second major calving in July 2012 released roughly 32 square kilometres farther upstream. NASA later concluded that the 2010 and 2012 events together reduced the floating tongue by about one-third. Petermann then remained comparatively stable for more than a decade, though smaller losses continued.
By 2023, NASA satellite analysis showed that a newer rift first observed in 2017 had converged with an older fracture. The agency also described Petermann as thinning, retreating and accelerating over the longer term.
Rifting and calving are normal parts of an outlet glacier’s life cycle. Ice keeps flowing seaward, so the front cannot simply advance forever. Yet “normal process” does not mean “unchanging system.” The location, frequency and scale of the losses, along with thinning and flow speed, are the measurements that reveal whether the glacier is moving into a different state.
The ocean is working on the ice from below
A satellite sees the surface brilliantly, but some of the most important action is hidden beneath it. Warmer ocean water can enter the fjord and melt channels into the underside of the floating tongue. Close to Petermann’s grounding zone, research summarized by NASA found melt rates reaching as much as 80 metres per year in some locations.
Surface air temperature, ocean heat, glacier flow, tides and the existing geometry of fractures all interact. The 4 August calving should not be reduced to one simple cause on the strength of a before-and-after image. The current ESA report documents the fracture sequence; it does not present a single-event climate attribution.
Greenland’s internal plumbing can complicate that picture further. A 2025 ScienceBlog report described a subglacial lake flood that burst upward through the ice elsewhere in northern Greenland. Researchers thought the downstream water pulse may have contributed to a separate calving event. That is not evidence for the same mechanism at Petermann, but it is a useful warning against treating every break as a surface-only process.
Calving can also change the water around a glacier. As ScienceBlog reported in 2025, measurements at another Greenland fjord showed that falling icebergs generated deep internal waves that mixed the water column and carried heat toward ice. That result does not automatically describe Petermann, but it illustrates why the front, the ocean and the fractured ice beyond it need to be studied as one coupled environment.
Why one-day radar repeats changed the view
The timing of this calving gave researchers an unusually dense record. During commissioning of the newly launched Sentinel-1D, it flew in tandem with Sentinel-1C. Images of the same place could be acquired one day apart.
By comparing the phase of radar signals in successive observations, interferometry can reveal subtle surface motion and deformation. At Petermann, the rapid repeat cycle let the team examine fracture propagation, the ice tongue’s response to tides and the approach to failure in near-real time.
A weekly image could show that a crack had grown. Daily interferometry can begin to show how it grew. That difference matters when deformation accelerates immediately before calving, and it should help researchers test which observable changes are useful warning signs at other glaciers.
The ice island is now a moving target
Detachment ended one scientific question and began another. Environment and Climate Change Canada is monitoring the island’s trajectory and assessing risks to ships and offshore infrastructure.
Large Arctic ice islands can survive for years. They drift, rotate, ground in shallow water and break into fragments. A single vast slab is conspicuous in satellite data; smaller descendants can be more numerous and harder to track. Their route through Petermann Fjord and toward Nares Strait will depend on winds, currents and sea-ice conditions.
The research team plans to combine satellite imagery, aerial observations and tracking data. For the glacier itself, the most telling next measurements will include the speed of ice upstream, the opening and propagation of the two remaining rift systems, and any further retreat of the calving front.
Petermann’s new outline is now an observed fact. The fate of the two larger sections behind it remains a well-founded possibility, not a scheduled event. That combination, a documented break followed by a genuinely uncertain next act, is precisely why continuous radar coverage matters.
Discussion about this post