Thwaites Glacier is not expected to collapse all at once, and no scientist can put a single date on it. What researchers can say is that the glacier is losing about 50 billion tons of ice per year more than it gains from snowfall, its floating ice shelf could break apart within the next several years to decades, and the full loss of the glacier’s ice would play out over centuries. The distinction between these timescales matters enormously for understanding what “collapse” actually means.
What “Collapse” Means for Thwaites
When scientists talk about Thwaites collapsing, they’re usually referring to two separate events on very different timescales. The first is the breakup of the glacier’s floating ice shelf, the massive slab of ice that extends out over the ocean and acts like a cork holding back the land-based ice behind it. The second is the long-term, irreversible retreat of the entire glacier into the sea. Confusing the two is the main reason headlines about Thwaites often sound more alarming than the science warrants, or occasionally less alarming than it should.
The floating ice shelf is the more immediate concern. Satellite imagery has revealed widening fractures and structural weaknesses across the eastern ice shelf. Researchers at Oregon State University warned that at the current rate of change, these weaknesses could cause the ice shelf to collapse in as little as five years, though they emphasized that extrapolating such nonlinear processes is highly uncertain. That warning was issued in late 2021, meaning we are now within that projected window. The shelf hasn’t collapsed yet, but it continues to weaken.
Why Warm Water Is the Main Driver
Thwaites isn’t melting primarily because of warmer air. The real threat comes from below: warm ocean water called Circumpolar Deep Water is reaching the base of the glacier where it meets the seafloor, a zone known as the grounding line. This warm water melts ice from underneath, reducing the friction that keeps the glacier anchored to the bedrock. As the ice thins and lifts off the ground, the glacier speeds up and the grounding line retreats inland.
The problem gets worse because of the shape of the land beneath the ice. The bedrock under Thwaites slopes downward as it goes inland, which means that as the grounding line retreats, it encounters progressively thicker ice sitting on progressively deeper ground. This creates a feedback loop: retreat exposes more ice to warm water, which causes more melting, which causes more retreat. Scientists call this marine ice sheet instability, and it’s one of the most well-established concerns in glaciology.
Recent measurements show that melt rates vary dramatically across the glacier. On the Thwaites Glacier Tongue, the freely floating western portion, ice is melting from below at roughly 50 meters per year. The eastern ice shelf melts at about 10 meters per year at its grounding zone entrance. That fourfold difference helps explain why the tongue has retreated much faster in recent decades. Researchers also found that warm seawater may be penetrating several kilometers upstream of the official grounding line through a tidal pumping process, meaning the zone of active melting extends further inland than previously mapped.
A poleward shift in westerly winds has increased the intrusion of this warm deep water onto the continental shelf. Along parts of Antarctica’s continental slope, mid-depth ocean temperatures have warmed by 0.8 to 2.0°C between the mid-20th century and recent years. This is not a process that reverses quickly even if atmospheric warming slows.
The Ice Cliff Scenario Looks Unlikely
For years, one of the most alarming possibilities was something called marine ice cliff instability. The idea, first formalized in 2012, was that if the floating ice shelf disintegrated, it would expose towering ice cliffs at the glacier’s edge. Cliffs taller than about a kilometer should, in theory, collapse under their own weight, exposing even taller cliffs further inland and triggering a chain reaction of rapid ice loss.
This hypothesis drove some of the most dramatic projections of sea level rise. But the most comprehensive investigation to date, conducted as part of the International Thwaites Glacier Collaboration, found that this runaway scenario is unlikely at Thwaites. Researchers ran simulations using three different ice sheet models, even with parameters designed to produce aggressive cliff failure. In every case, the ice cliff retreated very little over 100 years. The exposed cliffs simply weren’t tall enough in enough places to sustain a chain reaction.
Even when the models forced Thwaites to retreat at its current rate for 50 years before removing the ice shelf entirely, the cliff instability still didn’t take hold. The calving rate would need to be at least 25 times higher than current best estimates for the runaway process to occur. This is one of the most important findings in recent Antarctic science: the worst-case, rapid-collapse scenario appears to be off the table for Thwaites, at least through the mechanisms we understand today. Researchers do note, however, that damaged or fractured ice could lower the threshold for cliff failure in ways not yet captured by models.
The Tipping Point May Already Be Close
A 2025 study in Nature Climate Change mapped tipping point thresholds for Antarctic ice basins and found that a global temperature increase as low as 1 to 2°C above pre-industrial levels could trigger the long-term collapse of roughly 40 percent of marine ice volume in West Antarctica. The world has already warmed about 1.3°C above pre-industrial levels, which places us at the doorstep of this range.
“Long-term” is the key qualifier. Crossing a tipping point doesn’t mean the ice vanishes overnight. It means the process becomes self-sustaining and effectively irreversible, even if temperatures stabilize. The glacier would continue retreating for centuries. Marine-based sectors in East Antarctica, holding ice equivalent to about 5 meters of sea level rise, face similar tipping risks at 2 to 5°C of warming.
How Much Sea Level Rise Thwaites Could Cause
Thwaites currently accounts for about 4 percent of all global sea level rise. That may sound modest, but it’s concentrated in a single glacier, and the rate is accelerating. Antarctica as a whole is shedding roughly 135 billion tons of ice per year, raising sea levels by about 0.4 millimeters annually, and Thwaites and its neighbor Pine Island Glacier are responsible for the lion’s share of that loss.
If the ice shelf breaks apart, the land-based ice behind it will flow faster into the ocean. The ice shelf’s collapse alone won’t raise sea levels directly, since floating ice already displaces its weight in water. But losing that buttressing effect would accelerate the discharge of ice that is currently sitting on land. Over decades and centuries, Thwaites alone holds enough ice to raise global sea levels substantially, and its retreat could destabilize neighboring glaciers across West Antarctica.
The most extreme projections, which assumed runaway ice cliff instability, suggested sea level rise severe enough to submerge most of coastal Florida, leaving only an interior strip of high ground from roughly Gainesville to north of Lake Okeechobee. A 2024 Dartmouth study found that these highest projections are unlikely, which is consistent with the finding that ice cliff instability probably won’t occur at Thwaites. But even the moderate scenarios, with a meter or more of rise over the coming century or two, would reshape coastlines worldwide.
A Realistic Timeline
The floating ice shelf could break apart within the next decade or two. It is already deeply fractured, and each season of warm water intrusion and surface melting weakens it further. Once the shelf goes, the glacier’s flow into the ocean will accelerate, but not catastrophically in a single event.
The broader retreat of Thwaites, the kind that commits the world to meters of sea level rise, will unfold over centuries. The process may already be irreversible given current warming levels, but “irreversible” and “imminent” are not the same thing. The ice will not suddenly slide into the sea. It will thin, accelerate, and retreat year after year, with the pace depending heavily on how much additional warming occurs and how much warm water continues reaching the glacier’s base.
What scientists are most confident about is the direction: Thwaites is losing ice, the loss is accelerating, and no mechanism currently observed is slowing it down. The glaciers in this region of West Antarctica are retreating at rates of a few kilometers per year, and the retreat clusters precisely where the geography funnels warm water toward deep, inland-sloping bedrock. The question is no longer whether Thwaites will lose a significant portion of its ice, but how fast.