Antarctica is losing ice at an accelerating rate. Since the early 2000s, GRACE and GRACE-FO satellite gravity measurements have recorded average net losses of approximately 150 billion metric tons per year across the continent. The rate is not constant: it has increased over the observational period, and the losses are geographically concentrated in West Antarctica and along the Antarctic Peninsula, where warming ocean waters have direct access to vulnerable glacier systems.
This is not a distant or speculative risk. It is a process already under way, already contributing to global sea level rise, and already reshaping the glaciological map of the continent.
Antarctica holds roughly 26.5 million cubic kilometers of ice, divided between two major ice sheets. The East Antarctic Ice Sheet is the larger of the two and rests primarily on bedrock at or above sea level. It has been relatively stable, though even there, recent observations have raised questions about long-term vulnerability.
The West Antarctic Ice Sheet is the principal concern. Much of it rests on bedrock that lies below sea level, in some places more than two kilometers below, and slopes downward as you move inland. This geometry creates the conditions for marine ice sheet instability.
Around the continent’s margins, floating platforms of ice called ice shelves extend over the ocean. These shelves are not merely passive features. They exert back-pressure on the glaciers behind them, slowing their flow toward the sea. When an ice shelf thins or collapses, the glaciers it was restraining can accelerate dramatically.
The most consequential process in West Antarctic glaciology is marine ice sheet instability. When a glacier retreats across a bed that slopes downward toward the interior, the margin, called the grounding line, migrates into progressively deeper water. Deeper water means more ice in contact with the ocean, more melting, faster flow, and further retreat. The process can become self-sustaining once it begins, with no obvious mechanism to arrest it short of a change in the underlying bedrock geometry.
Thwaites Glacier, sometimes called the Doomsday Glacier, is the most closely watched example. Roughly the size of Florida, it currently contributes approximately four percent of annual global sea level rise on its own. Its grounding line has been retreating. Its ice shelf, which has historically provided stabilizing back-pressure, is fracturing. Warm Circumpolar Deep Water, a relatively warm water mass that exists at depth throughout the Southern Ocean, is intruding beneath the shelf and accelerating basal melting. The processes now visible at Thwaites are consistent with the onset of marine ice sheet instability, and the scientific community has been clear that arresting them, once fully initiated, would not be straightforward.
Pine Island Glacier, to Thwaites’ east, has shown similar behavior and has been losing mass more rapidly than at any point in the observational record.
The immediate driver of accelerated ice loss is ocean warming. The Circumpolar Deep Water that reaches the ice shelves of West Antarctica has warmed significantly since the mid-twentieth century. As this water contacts the underside of ice shelves, it melts ice from below at rates far exceeding any surface melting. This sub-shelf melting thins the shelves, reduces their buttressing effect, and sets in motion the glaciological cascade described above.
Atmospheric warming matters too, particularly along the Antarctic Peninsula, which has warmed faster than almost any comparable region on Earth over the last seventy years. Warmer air accelerates surface melting, contributes to the ponding of meltwater on ice shelf surfaces, and drives the hydrofracture processes that contributed to the collapse of the Larsen B ice shelf in 2002. That event, in which approximately 3,250 square kilometers of ice disintegrated within weeks, produced a measurable acceleration in the glaciers that Larsen B had been restraining.
Complete loss of the West Antarctic Ice Sheet would raise global sea levels by several meters, commonly estimated at about 3.3 meters for WAIS alone, with additional contributions possible from vulnerable East Antarctic sectors. No responsible projection suggests this occurs by 2100, but partial loss on century timescales is an active subject of scientific analysis rather than a theoretical extreme. Current projections for the contribution of Antarctic ice loss to sea level rise by 2100 range from a few tens of centimeters to well over a meter, with the upper end of that range driven by uncertainty about whether marine ice sheet instability processes will remain bounded or will accelerate beyond current projections.
A meter of sea level rise, distributed unevenly around the globe and amplified by storm surge and tidal variability, does not affect coastlines uniformly. It affects the most densely populated and least protected coastal areas first and most severely. Understanding the pace and dynamics of Antarctic ice loss is not an academic exercise. It is a prerequisite for making sensible decisions about infrastructure, settlement, and adaptation at a global scale.
The scientific tools available for monitoring Antarctic ice have improved substantially over the last three decades. GRACE and GRACE-FO satellites track changes in Earth’s gravitational field, allowing mass balance estimates for the ice sheets as a whole. CryoSat-2 uses radar altimetry to measure surface elevation changes across the continent. IceBridge airborne surveys have mapped bed topography, a critical input for modeling where retreat is likely to accelerate. Field programs at Thwaites and other key glaciers are providing direct measurements of ocean temperatures, basal melt rates, and grounding line position.
These observations are converging on a consistent picture. The rate of ice loss from Antarctica has increased. The drivers of that loss are understood. The processes most likely to amplify it are identified. What remains uncertain is the pace at which those processes will unfold, and what combination of emissions trajectories and physical thresholds will determine that pace.