A crevasse is a deep fracture in a glacier or ice sheet, opened by the stresses that build as ice flows, stretches, and bends over uneven terrain. These cracks can reach tens of meters deep, span several meters wide, and run for hundreds of meters along a glacier’s surface. They form wherever the tensile stress in ice exceeds its strength, and that threshold turns out to be surprisingly well defined, sitting in a range that researchers have only recently pinned down with satellite data and fracture-mechanics models. Crevasses are far more than static scars in ice, though. They channel meltwater to the glacier bed, influence how ice shelves break apart, and even shape ocean circulation beneath floating ice.
How Crevasses Form
Ice in a glacier behaves like a very slow-moving fluid under most conditions, deforming and creeping under its own weight. But when the ice is pulled apart faster than it can deform, it snaps. The pulling can come from several directions: where a glacier accelerates down a steepened slope, where it spreads laterally after leaving a narrow valley, where fast-moving ice shears against slower ice or a rock wall, or where the bed beneath drops away. In each case, the strain rate in the ice creates tensile stress that eventually exceeds the ice’s ability to stretch without breaking.
Pinning down the exact stress at which ice cracks has been a long-standing challenge. Laboratory tests on small ice samples tend to give higher strength values than what glaciologists observe in the field, because real glacier ice contains air bubbles, grain boundaries, and pre-existing flaws that weaken it. Recent work using satellite-derived velocity fields over Greenland found that crevasses begin forming at stresses around 265 kPa in steady-flow onset regions, with considerable spread depending on local conditions.1The Cryosphere. Failure strength of glacier ice inferred from Greenland crevasses A separate study calibrated against Antarctic ice shelves estimated the tensile strength of natural glacier ice at roughly 200 to 260 kPa under standard flow-law assumptions, narrowing the range produced by earlier observational work.2Journal of Glaciology. Fracture criteria and tensile strength for natural glacier ice calibrated from remote sensing observations of Antarctic ice shelves The picture these studies paint is consistent: glacier ice fractures at stresses well below what lab specimens endure, and crevassed and uncrevassed areas sit in clearly distinct stress zones.
Once a crack initiates at the surface, it propagates downward as long as the tensile stress at its tip exceeds the resistance of the ice. The crack’s own weight eventually squeezes it shut at depth, because the overburden pressure from the ice above increases with depth and acts to close the fracture. For a dry, air-filled crevasse in temperate ice, this balance limits the maximum depth to roughly 30 meters in many settings, though the number varies with the local stress field and ice temperature. Colder, stiffer ice and stronger extensional flow allow deeper penetration.
When Water Gets Involved
The depth limit for a dry crevasse changes dramatically when water enters the picture. Meltwater pooling inside a crevasse exerts an outward pressure on the crack walls that counteracts the overburden pressure trying to squeeze it shut. Because water is denser than ice, a water-filled crevasse can, in principle, propagate all the way through the glacier regardless of its thickness. This process, called hydrofracture, is one of the most consequential mechanisms in glaciology.
Seismic observations have confirmed that water-driven fracturing pushes crevasses deeper than the dry limit. Researchers comparing icequake depths to the theoretical maximum depth of dry crevasses found that some fracture events occurred well below that limit, a clear signature of hydrofracture at work.3Geophysical Research Letters. Breaking the Ice: Identifying Hydraulically Forced Crevassing Modeling studies using fracture mechanics have shown that water-filled cracks beneath supraglacial lakes are an effective mechanism for driving fractures through thick ice sheets, essentially turning a surface pond into a wedge that can split ice hundreds of meters thick.4Geophysical Research Letters. Constraints on the lake volume required for hydro‐fracture through ice sheets Poro-damage mechanics models reinforce this: hydrofracture can enhance crevasse depth and, in some cases, allow a crack to penetrate the full thickness of the ice, leading to iceberg detachment.5Journal of Glaciology. A non-local continuum poro-damage mechanics model for hydrofracturing of surface crevasses in grounded glaciers
This is why surface meltwater on ice shelves worries glaciologists so much. When melt ponds fill existing crevasses faster than the water can drain or refreeze, the hydraulic pressure drives the crack deeper. Enough ponds, in the right places, can trigger a chain reaction of fracturing that leads to large-scale ice shelf collapse. The rapid breakup of the Larsen B Ice Shelf in 2002 is the most cited example: thousands of melt ponds had formed on the surface in the preceding summers, and the shelf disintegrated in a matter of weeks.
Crevasses from Below
Not all crevasses start at the surface. Basal crevasses form on the underside of floating ice shelves, where the ocean meets the ice. They open upward rather than downward, driven by tensile stresses at the base, and can penetrate a significant fraction of the ice thickness. Because they are hidden from above, they are harder to detect and have historically received less attention, but they play an outsized role in ice shelf dynamics.
A striking example of their importance comes from direct observations made with an underwater vehicle in a basal crevasse at the Ross Ice Shelf grounding zone. The survey revealed asymmetric melting along the lower crevasse sidewalls and freezing in the upper reaches of the crevasse. Freshwater released by melting at depth and salt rejected from freezing above drove an overturning circulation inside the crevasse itself, layered on top of a throughflow jet running parallel to the coastline. The data showed that basal crevasses influence ocean circulation and mixing at ice shelf grounding zones to a degree that had not previously been recognized.6Science Advances. Direct observations of melting, freezing, and ocean circulation in an ice shelf basal crevasse
This finding matters because it means basal crevasses are not just passive voids in the ice. They actively modify how warm ocean water reaches and melts the underside of ice shelves. A crevasse that concentrates warm water flow toward the grounding zone, the point where the ice transitions from resting on bedrock to floating on the ocean, can accelerate retreat in ways that surface observations alone would miss.
The Firn Layer Complication
Glaciers and ice sheets are not solid ice from top to bottom. The upper meters to tens of meters consist of firn, a transitional material between fresh snow and dense glacial ice. Firn is less dense, more porous, and mechanically weaker than the ice below it. This layered structure turns out to matter for crevasse propagation in ways that simple uniform-ice models underestimate.
Modeling work focused on depth-dependent firn properties has shown that the firn layer can increase the vulnerability of ice shelves to fracture and calving. Because firn has different elastic properties and lower density than the underlying ice, crevasses propagating through it behave differently than a uniform-ice model would predict. Ignoring this layered structure leads to underestimates of how deep crevasses penetrate and how far rifts propagate across ice shelves.7The Cryosphere. The influence of firn layer material properties on surface crevasse propagation in glaciers and ice shelves In practical terms, the spongy top layer of an ice shelf is both a structural weakness and a reservoir that can hold meltwater, feeding the hydrofracture process described above.
Crevasses and Ice Sheet Stability
Crevasses connect to ice sheet stability through several paths. The most direct is calving: when crevasses penetrate deeply enough, they define the fracture planes along which icebergs break free. Surface crevasses propagating downward and basal crevasses propagating upward can meet in the middle, severing a block of ice entirely. This is how most tabular icebergs detach from Antarctic ice shelves.
Damage accumulation adds a subtler dimension. Ice that has been partially fractured but not yet broken all the way through is weaker than intact ice. Over long timescales, this damage enhances the viscous flow of ice, softening the glacier and allowing it to deform more readily.8The Cryosphere. Combining damage and fracture mechanics to model calving The practical consequence is that crevassed regions of an ice shelf can thin and weaken long before any calving event occurs, priming the shelf for a sudden collapse when conditions finally tip over a threshold.
Crevasses also contribute to ice loss by routing surface meltwater to the glacier bed. On the Greenland Ice Sheet, meltwater draining through crevasses and moulins (vertical shafts melted through the ice) lubricates the bed, temporarily speeding up glacier flow. While this speedup is often seasonal and self-limiting in some areas, it is another pathway through which surface warming translates into ice dynamics.
The Mountaineering Hazard
For anyone traveling on a glacier, crevasses are the primary safety concern. The danger is compounded by snow bridges, thin coverings of wind-blown or accumulated snow that can completely hide a crevasse from view. A snow bridge that supports one person’s weight in the cold of early morning may fail under the same person in the afternoon, when solar warming has weakened it.
Field monitoring on an Alpine glacier found that snow bridges are complex structures combining different layers and cornices of snow with varying physical properties, all of which change with weather conditions and glacier dynamics. A closing crevasse can actually improve the stability of its snow bridge by compressing the snow, while an opening crevasse lowers it.9Cold Regions Science and Technology. Multiparameter monitoring of crevasses on an Alpine glacier to understand formation and evolution of snow bridges This means the same bridge can strengthen or weaken over time depending on what the crevasse beneath it is doing, a subtlety that makes simple rules of thumb unreliable.
Standard glacier travel protocols reflect this uncertainty. Roped teams, spaced far enough apart that no two people stand on the same crevasse, are the baseline precaution. Probing the snow surface with a pole helps detect hidden voids, but offers no guarantee. Route-finding relies on reading the glacier surface for clues: sagging depressions, slight discoloration, and linear patterns in the snow. Experienced mountaineers also consider the glacier’s flow pattern, since crevasses tend to open perpendicular to the direction of greatest stretching. A glacier turning a corner, for example, will have crevasses fanning out on the outside of the bend.
Detecting Crevasses from Above
On larger scales, satellite remote sensing and ground-penetrating radar are the primary tools for mapping crevasses. Satellite radar imagery, particularly from missions like TerraSAR-X, can reliably identify crevasse locations by balancing the penetration depth of the radar signal into snow with horizontal resolution. Ground-penetrating radar provides direct measurements of snow bridge thickness and crevasse width, and has been validated against physical excavations at sites like the McMurdo Shear Zone and the Brunt Ice Shelf.10Cold Regions Science and Technology. Crevasse and rift detection in Antarctica from TerraSAR-X satellite imagery
Mapping crevasses across entire ice sheets is about more than safety. The spatial pattern of crevasses is a direct record of the stress field within the ice, which in turn reflects the speed, direction, and acceleration of glacier flow. A comprehensive review of glacier structures noted that crevasses relate well to measured strain rates, making them a visible proxy for the forces driving ice motion.11Reviews of Geophysics. Structures and Deformation in Glaciers and Ice Sheets Changes in crevasse patterns over time can signal that a glacier is speeding up, thinning, or shifting its flow direction, all useful indicators for monitoring ice sheet health.
Crevasses Recorded in the Landscape
Crevasses are temporary features of a living glacier, but they can leave a mark long after the ice is gone. When sediment fills a crevasse while the glacier is still active, and the ice later melts away, the sediment infill remains as a raised ridge on the landscape. These features, called crevasse fill ridges, have been documented in the forelands of surging glaciers in Iceland, where ridges up to two to three meters high with steep slopes and sharp, irregular crests preserve the geometry of the original cracks. In some cases, basal clay was pressed upward through crevasses to the glacier surface in marginal zones where the ice was thin.12IntechOpen. Perspective Chapter: What Sort of Ice Dynamics Are Crevasse Fill Ridges Connected with? – Research Overview
These landforms give geologists a window into the behavior of glaciers that disappeared thousands of years ago. The orientation, spacing, and geometry of crevasse fill ridges record the stress state of ice that no longer exists, allowing researchers to reconstruct flow patterns, identify past surge events, and map the extent of former glaciers. In regions that were covered by ice sheets during the last glacial period, like Scandinavia and northern North America, these ridges are part of the toolkit for understanding how ice retreated.
Crevasses on Other Worlds
Fractures in ice are not unique to Earth. Several moons in the outer solar system, most famously Jupiter’s Europa and Saturn’s Enceladus, have surfaces made of water ice overlying liquid oceans. The dramatic lineaments crossing Europa’s surface have long been interpreted as crevasse-like fractures driven by tidal stresses from Jupiter’s gravity. A key question is whether any of these fractures could penetrate entirely through the ice shell, creating a direct connection between the surface and the ocean below.
Fracture-mechanics modeling applied to icy moons has found that full-thickness penetration is much harder than earlier estimates suggested, once the interactions between neighboring fractures are accounted for. Surface fractures alone are unlikely to reach the ocean, even under generous stress assumptions. Basal fractures, analogous to the basal crevasses found under Earth’s ice shelves, propagate further upward into the shell but still fall short of full penetration in most scenarios. The combination of tensile cracking and shear failure, connecting a surface crack with a deep basal fracture, appears necessary to open a pathway through the entire ice shell, and this is only possible in thinner regions under favorable stress conditions.13The Planetary Science Journal. Propagation of Vertical Fractures through Planetary Ice Shells: The Role of Basal Fractures at the Ice–Ocean Interface and Proximal Cracks
Whether or not through-going fractures exist on Europa matters for astrobiology: a crack connecting the ocean to the surface could bring ocean material, and any potential biosignatures, to where a lander could reach them. Enceladus already demonstrates this possibility with its south polar geysers, where water from the subsurface ocean erupts through fractures in the ice shell. The physics governing these fractures shares a common foundation with terrestrial crevasse science, and the models developed for Antarctic ice shelves are being adapted directly to study fracture on icy moons. It is one of the more unexpected connections in planetary science, where understanding a hazard on an Alpine glacier feeds into the search for life beyond Earth.