Suzanne Bevan

Antarctica’s contribution to global sea-level rise is accelerating, yet projections from ice-sheet models continue to span a wide range despite sustained advances in resolution and physical realism. Using a coordinated ensemble of simulations from the H2020 European PROTECT project, we assess both the ability of six state-of-the-art ice-sheet models to reproduce observed Antarctic mass loss since the early 1990s and the extent to which present-day behaviour constrains future evolution.Across most of the ice sheet, model agreement with observations is limited, reflecting strong sensitivity to model structure and internal dynamics rather than to external forcing alone. In sharp contrast, the Amundsen Sea sector of West Antarctica exhibits a persistent and robust relationship between modelled present-day mass-loss rates and projected sea-level contribution that extends to the end of the twenty-first century and beyond. This sector emerges as the only region where contemporary observations retain demonstrable predictive power for long-term outcomes, while elsewhere compensating processes dominate. Our results identify a fundamentally regional limit of predictability for the Antarctic Ice Sheet, highlighting where emergent constraints can meaningfully inform projections—and where uncertainty is likely irreducible with current models and observations.

Antarctic ice shelves typically comprise continental meteoric ice, in situ-accumulated meteoric ice, and marine ice accumulated at the shelf base. Using borehole optical televiewer logs from across Larsen C Ice Shelf, Antarctic Peninsula, we identify and report an intermediate ice unit, located between continental and in situ meteoric ice, that is tens of metres thick and formed of layers that progressively increase in dip (by ~60°) with depth. The unit’s stratigraphic position and depth, supported by flowline modelling, indicate formation at the grounding zone. We hypothesise that the unit forms due to changes in the surface slope of feeder glaciers at the grounding zone, resulting in both variable surface accumulation and intense deformation. The top of the unit also marks the depth at which lateral consistency in radar layering is lost from radargrams, which may, to some degree, mark the depth of grounding zone ice across all ice shelves.

The retreat of the Antarctic Ice Sheet is conventionally attributed to increased ocean melting of ice shelves, potentially enhanced by internal instability from grounding lines near retrograde bed slopes. Ocean melting is enhanced by increased intrusion of modified Circumpolar Deep Water (mCDW) into ice shelf cavities. Upwelling from the release of subglacial meltwater can enhance mCDW’s melting ability, though its efficacy is not well understood and is not represented in current ice sheet loss projections. Here we quantify this process during an exceptional subglacial lake drainage event under Thwaites Glacier. We found that the buoyant plume from the subglacial discharge temporarily doubled the rate of ocean melting under Thwaites, thinning the ice shelf. These events likely contributed to Thwaites’ rapid thinning and grounding line retreat during that period. However, simulations and observations indicate that a steady subglacial water release would more efficiently enhance basal melt rates at Thwaites, with melt rate increasing like the square root of the subglacial discharge. Thus, it remains unclear whether increased subglacial flooding events provide a stabilizing influence on West Antarctic ice loss by reducing the impact of subglacial water on ocean melting, or a destabilizing influence by triggering rapid changes at the grounding zone.

Mass loss from the Antarctic Ice Sheet is increasing, accelerating its contribution to global sea level rise. Projecting the future evolution of the Antarctic Ice Sheet but also better understanding the processes at play is therefore of major importance for the mitigation/adaptation of/to sea level rise. Despite considerable advancements in the initialization of ice sheet models over the last decade, challenges persist in reproducing the observed trend in global Antarctic mass loss. This discrepancy between models and reality reflects in the large range of sea level projections in the recent Ice Sheet Model Intercomparison Project (ISMIP6). As part of the European H2020 project, PROTECT, we conducted Antarctic Ice Sheet simulations with six European ice-sheet models until 2150, focusing on the ability of the model to reproduce observations. These simulations were driven by a range of ocean and atmospheric forcings derived from Earth System models or downscaled by regional climate models under various Shared Socioeconomic Pathways (SSPs). Our experimental design enables us to sample climate forcing as well as model and parametric uncertainties, ensuring a comprehensive exploration of the future evolution of the Antarctic System and its contribution to sea level riseOur simulations confirm that, regardless of the model used, the Amundsen sector is the region that will most likely dominate mass loss in the decades to come. In high emission scenarios (SSP5-8.5), a large increase in surface mass balance is also expected to temporarily overshadow acceleration in mass loss caused by ice-shelf basal melting. All the models show an acceleration in mass loss from the middle of the 22th century, following the significant increase in surface melting from the end of the 21st century for the SSP5-8.5 scenario. This emphasizes the pivotal role of surface melt in the long-term evolution of the Antarctic ice sheet and its contribution to sea level rise.

Thwaites Glacier, a large outlet glacier of the West Antarctic Ice Sheet, holds over a half meter of sea level rise equivalent. The large potential contribution to sea level is concerning given that the glacier may be vulnerable to self-sustaining processes of rapid retreat due to the retrograde bed slope that characterises much of the glacier’s bed. Such a reverse-sloping bed exists behind the relatively high ridge on which the western calving front (WCF) of the Thwaites Glacier terminus currently rests. Our study focuses on the factors that control the calving dynamics of the WCF and the ability of mélange to influence these dynamics. Employing the 3D Helsinki Discrete Element Model (HiDEM), we find that calving at this location currently occurs as rifts form and widen due to longitudinal tensile stresses associated with ice flow across the grounding line. Calving is restricted in HiDEM simulations that include a constricted mélange field that is confined within the bounds of the model domain. A thicker, constricted mélange field fully suppresses calving. These simulations show the development of robust force chains that transmit resistive forces to the Thwaites WCF. In the future, the ability for mélange to influence the calving dynamics at the WCF will depend on the degree to which it is constrained in the wide Amundsen Sea Embayment, either through binding in land-fast sea ice or jamming behind large, grounded icebergs. As such, sea-ice conditions and iceberg characteristics will need to be considered along with the presence of mélange in investigations of the future retreat of the prominently recognised Thwaites Glacier.

The western region of the wide Thwaites Glacier terminus is characterized by a near‐vertical calving front. The grounding line at this western calving front (WCF) rests on a relatively high ridge, behind which exists a reverse‐sloping bed; retreat of the grounding line into this over‐deepening basin could therefore expose deep calving faces that may be subject to ice‐cliff failure. Here, we use the 3D Helsinki Discrete Element Model to identify the factors that control the calving dynamics in this location. We then focus on the ability of mélange to influence these dynamics given the wide embayment in which Thwaites Glacier terminates. We find that calving along the WCF is currently influenced by ice flow across the grounding line and consequent longitudinal tensile stress and rift formation. Calving is slowed in simulations that are initiated with a highly constricted mélange, with a thicker mélange suppressing calving entirely. We liken the constrained simulations to a scenario in which mélange piles behind a large grounded iceberg. In a future which may see calving become a more dominant control on the retreat of Thwaites Glacier, this type of blockage will be necessary for robust force chains to develop and transmit resistive forces to the terminus. The ability of the mélange to hinder calving at this location will be determined by the presence and rigidity of binding land‐fast sea ice and iceberg keel depths. Therefore, it is necessary to represent calving, mélange and sea ice in a single framework to predict the fate of Thwaites Glacier.

Rift propagation, rather than basal melt, drives the destabilization and disintegration of the Thwaites Eastern Ice Shelf. Since 2016, rifts have episodically advanced throughout the central ice-shelf area, with rapid propagation events occurring during austral spring. The ice shelf's speed has increased by ~70% during this period, transitioning from a rate of 1.65 m d⁻¹ in 2019 to 2.85 m d⁻¹ by early 2023 in the central area. The increase in longitudinal strain rates near the grounding zone has led to full-thickness rifts and melange-filled gaps since 2020. A recent sea-ice break out has accelerated retreat at the western calving front, effectively separating the ice shelf from what remained of its northwestern pinning point. Meanwhile, a distributed set of phase-sensitive radar measurements indicates that the basal melting rate is generally small, likely due to a widespread robust ocean stratification beneath the ice–ocean interface that suppresses basal melt despite the presence of substantial oceanic heat at depth. These observations in combination with damage modeling show that, while ocean forcing is responsible for triggering the current West Antarctic ice retreat, the Thwaites Eastern Ice Shelf is experiencing dynamic feedbacks over decadal timescales that are driving ice-shelf disintegration, now independent of basal melt.

The Amundsen Sea Embayment in West Antarctica is experiencing the most rapid mass loss and grounding line retreat in Antarctica. Its glaciers are vulnerable to retreat through marine ice sheet instability. There is uncertainty over the timing and magnitude of retreat and in particular the response of Thwaites Glacier to thinning of its ice shelf and to ocean forced retreat of its neighbouring glaciers. We find that the response of Thwaites to melting of its ice shelf is limited. However, retreat of its neighbours can drive substantial retreat in Thwaites. We examine the impact of ice shelf buttressing on the stability of the grounding line. Further experiments show that extreme ice shelf forcings are required to trigger retreat in Thwaites in isolation. We also demonstrate that long-term stability is sensitive to the treatment of basal stress near the grounding line.

We investigate the physical basis of the crevasse-depth (CD) calving law by analysing relationships between glaciological stresses and calving behaviour at Sermeq Kujalleq (Store Glacier), Greenland. Our observations and model simulations show that the glacier has a stable position defined by a compressive arch between lateral pinning points. Ice advance beyond the arch results in calving back to the stable position; conversely, if melt-undercutting forces the ice front behind the stable position, it readvances because ice velocities exceed subaqueous melt rates. This behaviour is typical of self-organising criticality, in which the stable ice-front position acts as an attractor between unstable super-critical and sub-critical regimes. This perspective provides strong support for a ‘position-law’ approach to modelling calving at Sermeq Kujalleq, because any calving ‘rate’ is simply a by-product of how quickly ice is delivered to the critical point. The CD calving law predicts ice-front position from the penetration of surface and basal crevasse fields, and accurately simulates super-critical calving back to the compressive arch and melt-driven calving into the sub-critical zone. The CD calving law reflects the glaciological controls on calving at Sermeq Kujalleq and exhibits considerable skill in simulating its mean position and seasonal fluctuations.

Mass loss from the Amundsen Sea Embayment of the West Antarctic Ice Sheet is a major contributor to global sea-level rise (SLR) and has been increasing over recent decades. Predictions of future SLR are increasingly modelled using ensembles of simulations within which model parameters and external forcings are varied within credible ranges. Accurately reporting the uncertainty associated with these predictions is crucial in enabling effective planning for, and construction of defences against, rising sea levels. Calibrating model simulations against current observations of ice-sheet behaviour enables the uncertainty to be reduced. Here we calibrate an ensemble of BISICLES ice-sheet model simulations of ice loss from the Amundsen Sea Embayment using remotely sensed observations of surface elevation and ice speed. Each calibration type is shown to be capable of reducing the 90% credibility bounds of predicted contributions to SLR by 34 and 43% respectively.