We seek contributions from both the observational and modelling communities relating to any area of ice-sheet, ice-shelf, or glacier hydrology. This includes but is not limited to: surface hydrology, melt lake and river formation; meltwater processes within the ice and firn; basal hydrology; subglacial lakes; impacts of meltwater on ice-sheet stability and flow; incorporation of any of these processes into large-scale climate and ice-sheet models.
Large parts of the Greenland Ice Sheet are fringed by ice-marginal (or ice-contact) lakes. These lakes have increased in number and size as a result of enhanced ice melt and the retreat of the ice sheet margin over recent decades. It has historically been assumed that Greenlandic ice-marginal lakes exist at a relatively uniform temperature of around 1°C year-round, thus having minimal influence on ice dynamics and subaqueous melt rates at the ice-water interface. However, there are almost no in-situ temperature measurements to test this hypothesis, meaning their influence on future ice sheet behaviour remains unclear. Here, we present continuous time series of lake water temperatures collected between July 2024 and August 2025, within three lakes on the western margin of the Greenland Ice Sheet. The results show that lake surface temperatures reached highs exceeding 10°C, with water temperatures above 4°C throughout the entire water column of one study lake during summer months. Summer stratification often persisted for several weeks, whilst inverse stratification was observed when water temperatures fell below 4°C. During winter months, surface ice cover maintained stable inverse stratifications, with lake temperatures ranging between 0 and 4 °C. Although lake temperatures remained largely stable during winter, one lake exhibited a cooling trend and significantly higher variability, potentially indicative of continued subglacial meltwater input.
We combine our sub-hourly lake temperature measurements with meteorological, lake turbidity and ice front calving data, enabling us to investigate sub-diurnal to seasonal controls on lake temperature variability. These analyses show how neighbouring lakes can have markedly differing thermal characteristics, likely due to differences in size, localised topography and variable subglacial and supraglacial meltwater inputs. Our results highlight how uniformly cold temperature values are likely unsuitable when modelling ice-lake dynamics, and that lake terminating sectors of the ice sheet may be experiencing greater rates of frontal ablation than previously realised.
How to cite: Tuckett, P., Harpur, C., Scoffield, A., Quincey, D., Barnett, H., Abrahams, J., Mallalieu, J., Rawlins, L., Sutherland, J., Merchant, C., Woolway, R. I., Carrea, L., McCarroll, N., Wang, W., and Rippin, D.: Thermal characteristics of Greenlandic ice-marginal lakes derived from in-situ temperature data, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-5028, https://doi.org/10.5194/egusphere-egu26-5028, 2026.
Surface melt is a key control on ice sheet mass balance through meltwater runoff, and the surface-to-bed meltwater connection disturbs ice dynamics. The presence of supraglacial lakes (SGLs), a crucial component of the hydrological system, reduces the surface albedo, resulting in heightened solar radiation absorption and consequently enhancing mass loss. However, it is difficult to quantify lake-albedo feedback because little is known about the bottom ablation process, which is difficult to observe and is currently not incorporated in the regional climate models. This research mainly focuses on the simulation of SGL based on the improved GlacierLake_v2 model. Firstly, the specific albedo-depth parameterization for SGL is developed in the western Greenland ice sheet based on satellite observations, and a meteorologically driven runoff module to calculate meltwater input is also incorporated in GlacierLake_v2. Secondly, the SGL-albedo feedback (the melt rate difference between SGL and bare ice) is quantified in Lake BlueSnow, and its influencing factors are explored.
The lake albedo is calculated using narrow-to-broadband conversion, and the lake depth is extracted from ICESat-2. The albedo-depth parameterization is described by an exponential function. Compared to the measured bottom ablation, the GlacierLake_v2 achieves superior performance over the original GlacierLake model, with RMSE reduced by more than 50%. The SGL-albedo feedback exhibits an exponential decline as lake depth increases. Summer snowfall rapidly suppresses the ice sheet surface melt rate while exerting little influence on the lake bottom melting, thereby triggering strong SGL albedo feedback. We are currently developing a distributed SGL model aimed at simulating lake evolution in both horizontal and vertical dimensions and at the volume estimation of buried lakes. There is also the prospect of integrating GlacierLake_v2 into the more comprehensive hydrological model to decrease the uncertainty in surface mass loss predictions.
How to cite: Wu, J., Zheng, L., and Hui, F.: Greenland supraglacial lakes albedo-depth parameterization from multi-source remote sensing: an application of lake-albedo feedback modeling, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-5158, https://doi.org/10.5194/egusphere-egu26-5158, 2026.
Supraglacial channels play a fundamental role in efficiently transporting surface meltwater across, into, and beyond glacier systems, with implications for surface mass balance, glacier dynamics, and the hydrochemistry of glacier runoff. Although such channels have been frequently documented in ice sheet settings, the resolution of satellite imagery makes similar studies on mountain glaciers challenging, and field observations remain scarce. Here, we combine uncrewed aerial vehicle (UAV) imagery, field measurements, and historical aerial photographs to conduct high-resolution mapping of supraglacial channels on Glacier du Brenay in the Swiss Alps, which provides new insights into their distribution and characteristics. Our results reveal a dense pattern of broadly dendritic channel networks influenced by ice surface structures. We find that most first-order channels, which are only visible in UAV imagery, terminate englacially in crevasses. Consequently, englacial routing is likely more widespread at Glacier du Brenay than suggested by coarser resolution imagery, where the larger and more detectable channels often flow off the glacier terminus. Channel size and distribution are influenced by the glacier’s surface profile, which dictates the location and extent of channel catchments. Larger upstream-originating catchments are associated with increased channel dimensions, and where deeply incised channels from these catchments flow off the terminus, they are associated with rapid retreat at the proglacial margin. We find that continuous debris cover produces shallow channels and results in increased channel roughness, decreased water velocity, and reduced surface lowering rates. In contrast, discontinuous debris is associated with the highest rates of surface lowering and can produce pitted topography, which also increases channel roughness. Future research should therefore consider small-scale hydrological processes.
How to cite: Wytiahlowsky, H., Stokes, C., Hodge, R., and Clason, C.: Understanding supraglacial drainage networks on mountain glaciers using high‑resolution orthophotos, UAV imagery and field data, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-5730, https://doi.org/10.5194/egusphere-egu26-5730, 2026.
Supraglacial melt lakes have been linked to large-scale Antarctic ice shelf collapse, such as those observed at Larsen A and B ice shelves in the early 2000s. As Earth’s climate warms, surface melt and the formation of supraglacial melt lakes are expected to increase. Currently, there is no representation of the impacts of supraglacial melt lakes in large-scale ice sheet modeling. In this work, we implement physics-based parameterizations of supraglacial melt lakes in the Ice Sheet and Sea-Level System Model (ISSM) to simulate their effect on fracture propagation and calving. We described a new physics-based modeling protocol for capturing realistic interactions between surface melt and calving. We show the influence of including these interactions on benchmark ISMIP6 simulations by comparing them to simulations which either (a) do not simulate supraglacial melt lakes or (b) simulate instantaneous collapse of all floating ice in Antarctica. We further discuss caveats with this approach and directions for future research.
How to cite: Grau, D., Poinelli, M., Schlegel, N., and Robel, A. A.: Modeling interaction between supraglacial melt lakes and calving in transient Antarctic simulations, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-5759, https://doi.org/10.5194/egusphere-egu26-5759, 2026.
Moulins are vertical conduits on the Greenland ice sheet (GrIS) formed by hydrofracturing of crevasses and supraglacial lakes. They determine where, when, and how much surface meltwater can be drained to ice sheet bed, thereby controlling subglacial water pressure, developing subglacial drainage system, and eventually affecting ice motion. Therefore, obtaining moulin distribution over extended time period is essential for understanding variability in surface-to-bed meltwater connections. Using 10 m Sentinel-2 satellite imagery, we map interannual moulin distributions on the southwest GrIS during late summer from 2016 to 2021, as well as the seasonal moulin evolution throughout the warm 2019 summer. We find that widespread moulins drain the majority of meltwater into the ice sheet during 2016-2021. They are even denser in warmer years with larger proportions of meltwater drainage. Notably, a considerable number of moulins recur in multiple years and act as stable meltwater connections. During the warm 2019 summer, moulins first increase substantially at low elevations in June, then expand to high elevations in July, and remain relatively stable in August. As a result, surface-to-bed meltwater connections are spatially localized early in the melt season and become more discrete as the melt season progresses.
How to cite: Wang, Y. and Yang, K.: Satellite mapping of the varying moulin distribution on the southwest Greenland ice sheet, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-6241, https://doi.org/10.5194/egusphere-egu26-6241, 2026.
Meltwater buffering in the firn of glaciers, ice caps and ice sheets is an important, yet relatively uncertain, process that determines their mass balance. In the firn layer, meltwater and rain percolate downwards until it is buffered in wet layers, refrozen, or runs off. The efficiency of water retention determines the ratio of refreezing to runoff, while the vertical distribution of refreezing has a long-lasting impact on the subsurface density, heat conductivity and temperature profile. An accurate representation of water buffering is particularly crucial for estimating future runoff from the Greenland Ice Sheet. There, the formation of ice lenses in the former percolation zone could dramatically reduce the buffering capacity of the firn layer. Current firn models either are empirically-based and struggle to represent the complex processes determining water buffering or are physics-based but computationally expensive due to strong non-linearities in the governing equations.
Here, we present the intermediate-complexity percolation model FirnPerc that aims to capture the relevant water processes in firn within a fast and stable computational framework. Matrix percolation through firn is represented as gravity-driven flow, neglecting capillary forces. In cases of low water content and water flux, model layers can be partially wetted and can therefore remain below the melting point on average for a while. When water flow stalls on ice lenses, the model can form a slush layer. Gradual water percolation through ice lenses is parameterised, with exponentially decreasing efficiency for increasing ice layer thickness. Finally, in the absence of an explicit horizontal flow description, slush water is assumed to run off very slowly. With these process parameterizations, FirnPerc can resemble important features of Greenlandic firn, such as fast, deep percolation at the start of the melting season, ice lenses with slush layers on top, and deep, water-rich year-round aquifers.
How to cite: van de Berg, W. J.: FirnPerc: An intermediate-complexity water percolation model for firn, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-7173, https://doi.org/10.5194/egusphere-egu26-7173, 2026.
The Queen Maud Land (QML) sector of East Antarctica comprises a complex system of grounded ice sheet and fringing ice shelves that regulate ice discharge to the Southern Ocean. Ice-sheet evolution in this region is controlled by interactions between atmospheric forcing, katabatic winds, and bedrock topography, producing strong spatial variability in accumulation, flow, and thermal regimes. While the bordering ice shelves currently act as stabilising buttresses, they are sensitive to oceanic heat intrusions, changing sea-ice conditions, and episodic surface melt. Melt–refreeze processes enhance firn compaction, weaken surface integrity, and may precondition ice shelves for hydrofracture under future warming, despite QML presently exhibiting a positive mass balance trend.
We investigate the thermal and structural evolution of snow, firn, and ice in QML using an integrated framework that combines multi-frequency passive microwave observations with physically based modelling. Passive microwave measurements provide complementary sensitivity to near-surface melt processes and deeper firn and ice structure, enabling the detection of both contemporary melt signals and long-term subsurface changes. Lower-frequency observations penetrate deep into the firn and ice column, whereas higher-frequency observations respond to surface temperature, liquid water content, and accumulation variability.
Snow, firn, and ice evolution is simulated using the Glacier Energy and Mass Balance (GEMB) model, running on the Ice Sheet and Sea Level System Model (ISSM), providing vertical profiles of temperature, density, and liquid water content driven by meteorological forcing. These profiles are used to forward-model microwave brightness temperatures with the Microwave Emission Model of Layered Snowpacks (MEMLS) across frequencies from 1.4 to 36.5 GHz, accounting for densification and refrozen ice layers. Modelled brightness temperatures are evaluated against satellite observations, providing twice-daily coverage of QML since 2010.
We present spatial and temporal patterns of grounded ice-sheet structure, surface and subsurface temperature variability, fresh snow accumulation, and ice shelf melt signatures, together with residuals between observed and modelled brightness temperatures. Our results demonstrate the value of radiometric modelling for constraining firn structure, melt processes, and ice-shelf vulnerability in regions with sparse in situ data. By integrating passive microwave observations with physical firn models, this work supports improved calibration, initialisation, and confidence in projections of mass balance and structural evolution in the Queen Maud Land sector of East Antarctica.
How to cite: Colliander, A., Schlegel, N.-J., Hossan, A., Walker, C., Harper, J., Lemmetyinen, J., and Riihelä, A.: Integrating Microwave Remote Sensing with Physical Models to Reveal Melt Dynamics and Structural Variability in Queen Maud Land, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-5264, https://doi.org/10.5194/egusphere-egu26-5264, 2026.
Dynamic hydrologic networks on and within ice shelves are increasingly recognized as critical controls on ice shelf stability and ice sheet mass loss. To date, ice shelf hydrology has been primarily framed in terms of surface-derived meltwater, including supraglacial pond formation, firn infiltration and refreezing, and meltwater aquifer development. Recent firn modeling suggests that a substantial fraction of Antarctic ice shelf firn lies below sea level, providing extensive pore space that is susceptible to seawater intrusion and brine aquifer formation. Despite this potential, seawater infiltration into porous firn has remained largely unexplored beyond a small number of spatially limited observations, many of which were collected decades ago.
Here, we present the first continental-scale observational assessment of seawater-derived aquifers within Antarctic ice shelves. We analyzed ~145,000 km of airborne radar flightlines and identified diagnostic brine signatures along ~4,500 km of profiles spanning more than 30 Antarctic ice shelves, demonstrating that seawater infiltration occurs wherever suitable observations exist. By reconciling radargrams with available digital elevation models and lidar data, we show that the dominant infiltration mechanisms vary regionally: Antarctic Peninsula ice shelves commonly exhibit large infiltration zones in thin or damaged ice, whereas East Antarctic ice shelves are characterized by localized intrusion along rifts and basal crevasses. We further map the depth of the brine water table across these systems, revealing spatial variability in aquifer geometry linked to ice shelf structure and infiltration mechanism. Climate model projections indicate that regions currently hosting brine aquifers are projected to experience larger future increases in surface meltwater inputs than ice shelf regions without detected brine, highlighting the potential for the development of mixed aquifer systems. Our results demonstrate that seawater infiltration represents a widespread and previously underappreciated hydrologic pathway within Antarctic ice shelves and highlight the need to incorporate these systems into emerging frameworks of ice shelf hydrology and stability.
How to cite: McDowell, I., Culberg, R., Nordahl, C., Villagomez, V., Scambos, T., and Miller, J.: Continental-scale observations of seawater infiltration in Antarctic ice shelves reveal an overlooked hydrologic system, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-13687, https://doi.org/10.5194/egusphere-egu26-13687, 2026.
The transfer of surface meltwater to the base of the Greenland Ice Sheet modulates basal sliding, which controls ice loss into the ocean. The way in which water reaches the glacier bed, organizes itself near the ice-bed interface, and affects ice dynamics remains poorly understood due to limited observations. In this study, we present and interpret a multi-physical observational dataset from Isungnuata Sermia, West Greenland, acquired during the 2024 and 2025 melt seasons as part of the REASSESS and SLIDE projects.
We show how Distributed Fibre Optic Sensing (DFOS) measurements combined with meteorological observations and models, and Global Navigation Satellite System (GNSS) measurements of surface ice motion, provide observational constraints on and improve our conceptual understanding of the surface-to-bed hydrological connection. We deployed fibre optic cables at the surface and in four 600 m boreholes, providing measurements of temperature, strain, and seismicity at thousands of sensing points at the surface, and through the full depth of the glacier.
We observe that seismic activity is temporally linked with surface melt on a daily and weekly scale and evolves over time in alignment with the coalescence of surface channels. We apply Matched-Field Processing (MFP) beamforming to detect and locate seismic sources, aiming to resolve the vertical extent of fracturing related to water input. We observe the migration of some high-frequency seismic events to greater depths during periods of high runoff. This process is coincident with ice surface acceleration and subsequent deceleration events. Temperature and strain measurements from the DFOS system indicate highly variable distributions of temperature and deformation, which enable exploration of the prevalence and importance of englacial fracturing in surface-to-bed water transport.
Together, these data offer potential insights into the mechanism of fracture-driven water transport, basal pressure increase, and subsequent regulation and dissipation via subglacial-water-channel development, and the relationship of these processes to glacial dynamics.
How to cite: Fabbri, L., Gimbert, F., Paris, N., Chauve, T., Michel, A., Le Bris, T., Picard, L., Duphil, R., Fontaine, F., Freche, G., Doyle, S. H., Thorpe, S., Livingstone, S. J., and Sole, A.: Distributed, Multi-Physical Fibre Optic Sensing of the Isunnguata Sermia Englacial Hydraulic System and its Impact on Glacial Dynamics, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-8117, https://doi.org/10.5194/egusphere-egu26-8117, 2026.
Annual ice flow along the land-terminating margins of the Greenland Ice Sheet has been negatively correlated with surface melt over recent decades, a trend commonly attributed to the seasonal evolution of efficient subglacial drainage in response to larger melt volumes. However, there remains scant observational evidence of ice flow behavior at higher elevations in the accumulation zone, despite increasing surface melt and runoff. Here, we employ in-situ GPS measurements to analyse multi-decadal (1996-2023) variations in ice motion and surface slope along the land-terminating K-Transect, West Greenland, between ~ 1400 and ~ 1900 m.a.s.l.
We show that below the equilibrium line altitude (ELA), annual ice motion is negatively correlated with surface melt, consistent with the self-regulation of ice flow previously reported across the ablation zone. In contrast, above the ELA, we observe a small but persistent ice flow acceleration, punctuated by slowdowns in the large melt years of 2012 and 2023. We find that this ice flow acceleration has largely been driven by surface steepening, with the resultant increase in driving stress calculated to account for 70.0 ± 26.7% of the acceleration observed at ~ 1700 m.a.s.l. between 2009 and 2021. However, together with continued muted seasonality in ice flow at ~ 1900 m.a.s.l. already identified during 2009-2012, we also find clear evidence of direct surface meltwater access to the ice sheet bed at ~ 1700 m.a.s.l for the first time in our observation record. It is therefore possible that the drainage of surface meltwater has driven some of the observed acceleration, through increasing the basal sliding component of ice flow.
With the rate of ice surface lowering at the ice sheet margin predicted to continue to exceed that in the ice sheet interior, we expect that ice surface steepening will likely persist, thereby driving sustained ice flow acceleration across the higher elevations in the coming years. The direct impact of this ice flow acceleration on mass loss by drawdown will likely be modest; whilst increased ice will be advected to lower elevations where air temperatures are higher, downstream self-regulation of ice flow is expected to constrain the resultant increase in ice flux. However, our findings show that surface-to-bed connections can form above the ELA, which has implications for the volume and timing of runoff from the high elevation regions which undergo summer melt increasingly often.
How to cite: Picton, H., Tedstone, A., Nienow, P., Machguth, H., Goldberg, D., Jullien, N., Clerx, N., Posch, C., Gastaldello, M., and van As, D.: Ice surface steepening drives multi-decadal acceleration of the Greenland Ice Sheet interior, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-16195, https://doi.org/10.5194/egusphere-egu26-16195, 2026.
The subglacial drainage system of the Greenland Ice Sheet plays a central role in modulating basal sliding and ice flow dynamics. While the seasonal evolution of this system has been widely studied during the melt season onset, its late and post-melt season behavior remains poorly constrained due to limited spatial and temporal observational coverage. Here, we use Sentinel-1 double-difference interferometric SAR data to document spatially continuous patterns of vertical and horizontal ice sheet motion across the late melt season in a predominantly land-terminating region of western Greenland. Our observations reveal recurrent, localized vertical subsidence features, generally coinciding with subglacial troughs, that persist several weeks into the post-melt season (late August to October) after surface melt inputs have largely ceased. The localized subsidence signals are accompanied by ice flow deceleration at a regional scale, extending more than 100 km inland, to ice thicknesses above 1300 m. We interpret this pattern as a dynamic response to the gradual drainage of water stored in weakly-connected cavities or, alternatively, englacial or sedimentary components, which drives a large-scale water pressure decrease and hence widespread flow deceleration. Although prior studies have suggested that weakly-connected cavities are drained following de-pressurization of efficient channels, the multi-week time scale and far-inland extent of the observed dynamic response suggest that channels may not be the sole driving mechanism. The magnitude of both subsidence and slow-down scales with melt season intensity, suggesting a mechanistic link between meltwater drainage efficiency and late-season ice dynamics, in line with previous observations of ice flow self-regulation. Our findings offer new insights into the seasonal evolution of Greenland's basal hydrology with continuous spatial coverage, highlighting how localized corridors of meltwater evacuation affect ice motion over much larger scales.
How to cite: Andersen, J. K., Gimbert, F., and Davison, B. J.: Meltwater Drainage in Troughs Controls end-of-summer Regional Ice Flow Deceleration in Western Greenland, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-17759, https://doi.org/10.5194/egusphere-egu26-17759, 2026.
Subglacial water flow plays a critical role in basal sliding and, consequently, in glacier and ice-sheet dynamics. However, modelling the coupled evolution of subglacial drainage and ice flow remains challenging. This study investigates the evolution of the basal ice–water interface by analysing heat and fluid flow in idealised englacial channels. We extend the classical Röthlisberger model for circular channels to elliptical channel geometries. A hybrid turbulent–laminar melt scheme captures heat generation from both viscous and turbulent dissipation, while a viscous flow law models the creep closure of the surrounding ice. The flow and temperature profiles in elliptical channels are solved for with differential melting between the roof and walls of the channel. We find that elliptical channels tend towards a circular shape when laminar melting dominates, whilst the flow of ice tends to increase the eccentricity of the channel. Our hybrid laminar-turbulent melt model permits variations in the distribution of melting along the ice-water boundary and the existence of stable, non-circular cross-sections. These stable channels obey pressure-flux relationships that we use to explore the evolution and dynamics of hydrologically interacting channels in a wider subglacial drainage network, working towards a simplified and scalable subglacial hydrology model.
How to cite: Brown, I., Warburton, K., and Neufeld, J.: Elliptical Röthlisberger Channels: Modelling Flow, Heat and Evolution, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-21093, https://doi.org/10.5194/egusphere-egu26-21093, 2026.
A significant part of the uncertainty of the future contribution of the Antarctic ice sheet to sea-level rise stems from unknown subglacial conditions such as geology and water content. Sliding at the base of the ice sheet depends on geological conditions, as water routing differs between soft and hard beds regions. However, ice sheet modelling studies usually do not distinguish between these different types of bed properties and instead assume it to be constant across the model domain. Also, current subglacial hydrology models have a high computational cost and their application to ice sheet models is often limited to the initialization stage.
A recent subglacial hydrology model by Kazmierczak et al. (2024) overcomes these challenges by introducing simplifications that allow to simultaneously account for hard, soft, and mixed beds, as well as efficient and inefficient drainage. The hydrology model is computationally fast, hence it can be easily fully coupled to any ice sheet model.
In this study, we apply this new subglacial hydrology model on the Ronne-Filchner basin, which includes one quarter of all ice in Antarctica and the second-largest ice shelf world-wide. The Ronne-Filchner basin exhibits sharp contrasts in subglacial topography, geology and hydrology, making it an interesting test case for the hydrology model. Furthermore, a sharp increase in sub-shelf temperatures under the Ronne-Filchner ice shelf has been suggested in coming centuries, yet this basin has been comparatively understudied thus far. We present projections under different warming scenarios and subglacial conditions, aiming to provide a sensitivity analysis of these different factors, and their interactions, on future sea level change. Our results provide insight on the importance of a detailed inclusion of subglacial geology and water evolution in large-scale ice sheet models.
How to cite: Scheiter, M., Coulon, V., and Pattyn, F.: A mixed-bed subglacial hydrology model applied to the Antarctic Ice Sheet, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-14928, https://doi.org/10.5194/egusphere-egu26-14928, 2026.
Subglacial water pressure time series from instrumented boreholes provide widespread evidence for the formation of hydraulically disconnected regions at the bed. These do not exhibit the typical diurnal pressure oscillations that indicate a connection to the melting glacier surface during summer. Importantly, the spatial extent of hydraulic disconnection can evolve over time. This is a feature of borehole data sets from both, sub-Arctic glaciers where a cold surface layer quickly seals water-filled boreholes from the surface through freezing, and from temperate mid-latitude glaciers. Current two-dimensional subglacial drainage models do not allow for the connectivity of the drainage system to evolvein time, and in fact, do not allow for the complete shut-down of the system anywahere. In this presentation, we review the observational evidence, focusing on a data set obtained at a small polythermal valley glacier in the southern Yukon Territory, Canada. We document the rapid switching of sizeable (~ 1 ice thickness in extent) parts of the bed from disconnected to connected and back again, driven by the resumption of surface melt resumes following periods of summer snow cover. We show how current drainage models need to be modified to account for suc switching behaviour, and discuss the wider implications of these modifications on drainage system behaviour and glacier dynamics. In particular, we show that switching behaviour can explain the typically observed high water pressures under glaciers during winter, and how this creates conflict with widely used friction laws. We also show how hydraulic switching may limit the ability to pump water out of the bed, which has been suggested elsewhere as a technically plausible mechanism for artificially slowing glacier flow.
How to cite: Schoof, C. and Racz, G.: Patchy subglacial drainage systems: observations and modelling, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-2281, https://doi.org/10.5194/egusphere-egu26-2281, 2026.
Seasonal acceleration of Antarctic Peninsula outlet glaciers is commonly attributed to surface melt–driven basal lubrication and ocean forcing. However, the role of internally generated basal meltwater, independent of seasonal surface inputs, remains poorly constrained by observations, despite its potential to introduce year-round dynamic sensitivity that is not captured by seasonally forced frameworks. Here we investigate how non-seasonal subglacial hydrology influences the dynamics of Fleming Glacier, a rapidly evolving outlet glacier draining into Wordie Bay on the western Antarctic Peninsula.
Using time-resolved digital elevation models (2011–2024) together with satellite-derived surface velocity and basal drag estimates, we identify an actively evolving subglacial water reservoir beneath the fast-flowing lower trunk of the glacier. Episodic filling and drainage of this reservoir produce multi-meter surface height anomalies that are temporally coincident with changes in ice velocity and basal drag. These signals occur predominantly during austral winter and exhibit weak or absent annual periodicity, indicating that they are not driven by seasonally varying surface meltwater input.
Energy-budget considerations and spatial patterns of hydraulic potential suggest that the reservoir is sustained primarily by internally generated basal meltwater produced through frictional heating, rather than by surface or oceanic meltwater sources. Episodic drainage events transiently reduce effective pressure at the ice–bed interface, promoting short-lived acceleration and spatial reorganization of basal drag upstream of the grounding line.
Our results demonstrate that internally driven, non-seasonal subglacial hydrology can modulate ice–bed coupling on multi-year timescales, highlighting an additional mechanism of outlet glacier variability that operates independently of seasonal climate forcing.
How to cite: Dong, Y., Zhao, J., Wolovick, M., Helm, V., Franke, S., Wuite, J., Krieger, L., Floricioiu, D., Kleiner, T., Jansen, D., Höyns, L.-S., Rückamp, M., and Zhong, Y.: Frictional Melt–Sustained Subglacial Hydrology Modulates Ice–Bed Coupling at an Antarctic Peninsula Outlet Glacier, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-8700, https://doi.org/10.5194/egusphere-egu26-8700, 2026.
Beneath the Antarctic Ice Sheet, active subglacial lakes act as dynamic storage–release nodes that modulate basal water pressure, influence ice flow, and regulate freshwater delivery to grounding zones. However, inventories and time series derived from single sensors remain incomplete due to limited spatial coverage, short mission lifetimes, and cross-sensor inconsistencies in sampling geometry and accuracy. Here we present a continent-wide, multi-mission framework that integrates three decades of satellite altimetry with high-resolution REMA strip DEMs to detect and monitor active subglacial lakes. We harmonize reprocessed ERS-1/2 and Envisat radar altimetry, ICESat and ICESat-2 laser altimetry, and CryoSat-2 swath measurements through local annular co-registration and a crossover-zone representativeness correction, enabling internally consistent, lake-scale elevation change estimates across sensors. The approach resolves subtle elevation variability, refines outlines delineation, and reconstructs multi-decadal (>30 years) filling–drainage histories. Applied to the Byrd Glacier basin, the framework increases the number of confirmed active lakes from 23 in existing inventories to 74, including 51 newly identified systems, and delivers refined outlines and time series for each lake. The resulting records reveal structured lake–lake interactions, motivating a process-based classification of cascading behavior into serial cascades, parallel co-variation, and terminal confluence. Our results indicate that observational incompleteness remains a primary limitation on Antarctic subglacial hydrology, and demonstrate that systematic multi-mission fusion can substantially improve detection, connectivity inference, and quantification of storage–release variability, providing stronger observational constraints for linking basal hydrology to ice dynamics and improving ice-sheet projections.
How to cite: Yang, T., Liang, Q., Li, T., and Cheng, X.: A novel framework for detecting and monitoring Antarctic active subglacial lakes using multi-source remote sensing data, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-9771, https://doi.org/10.5194/egusphere-egu26-9771, 2026.
Subglacial lakes are important hydrological reservoirs within the Greenland Ice Sheet, primarily fed by meltwater from the surface that reaches the bed. Episodically, these lakes drain, releasing vast volumes of water into the subglacial drainage system and thereby altering basal hydrology and local ice dynamics. A warming climate is expected to intensify surface meltwater production and potentially increase the frequency of subglacial lake drainage events. However, direct observations of such events remain scarce, limiting our understanding of their driving mechanisms and impacts on the subglacial system.
Here we document a subglacial lake at Eqalorutsit Kangilliit Sermiat, a major tidewater outlet glacier in Southwest Greenland. This subglacial lake typically drains once or twice per melt season. We observed one such event using time-lapse imagery, GNSS measurements, terrestrial radar interferometry, and digital elevation models (DEMs). From these observations, we estimate a drainage volume of approximately 0.3 km3, and local ice surface lowering that exceeds 50 m. Additionally, we delineated the lake outline and were able to estimate lake drainage rates and refilling rates. Our results further indicate a (partial) re-routing of the subglacial drainage system following the lake drainage event. As the lake drained, we observed an acceleration of the glacier’s terminus region and the emergence of a large sediment-rich plume. This highlights the profound influence of episodic subglacial lake drainage on glacier dynamics as well as fjord circulation and ecosystems.
How to cite: Dachauer, A., Kneib-Walter, A., Welty, E., and Vieli, A.: Subglacial lake drainage event impacts basal hydrology and dynamics of Greenland outlet glacier, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-7976, https://doi.org/10.5194/egusphere-egu26-7976, 2026.
Through August-September 2015, a surface depression was observed in ArcticDEM data near the south lateral margin of Isunnguata Sermia (IS), 0.88 km2 in area and with 30 m maximum deflection. This ‘surface anomaly’, alongside flooding and sediment deposition in the IS foreland, was interpreted as the surface expression of a draining subglacial lake, with subsequent multi-year recovery of the surface elevation occurring as the lake recharged was inferred. Geophysical surveys were conducted through 2023-2025 to investigate lake dynamics over and around the surface anomaly. Reported here is the combined interpretation of active- and passive-source seismic analyses, aiming to image basal topography and determine the presence of a recharged subglacial lake.
The active-source array comprised 48 vertical-component geophones installed in a 2D profile, recording seismic energy made at a surface impact source; the passive acquisition included a dense grid of Fairfield 3-component seismic nodes, buried in a shallow augur-hole to optimise coupling. By combining active- and passive-source seismic reflectivities, amplitude-versus-angle (AVA) curves can be populated across 0-50° angle range, thus allowing the material properties either side of the glacier bed to be characterised. Subglacial water, either as a deep lake body or in saturated sediment, would produce negative-polarity reflectivity, given the acoustic softness of water versus the overlying ice.
Rather than the planar and specular reflectivity often associated with subglacial lakes, seismic profiles show a sloping and rugose bed. Seismic reflection polarities suggest hard basal conditions. AVA analysis, extending from near-zero incidence to the ~50° critical angle, shows consistent positive polarities across all identified bed reflections, suggesting a substrate that is acoustically harder than the overlying ice – for example, consolidated sediment or bedrock. This insight is supported with constraint from passive seismic analysis of seismic velocities beneath the ice, something not typically possible with active-source data alone. These AVA responses are observed inside and outside of the surface anomaly. Seismic analysis therefore suggests no evidence of significant water saturation or, indeed, a lake beneath the glacier bed. However, with vertical resolution no better than ~10-15 m, we cannot exclude the possibility of water films or small sediment pockets immediately beneath the glacier bed.
We conclude that any water feature previously inferred from ArcticDEM data was not present during seismic surveying and is therefore transient. Beyond IS, this work questions the degree to which examples in the wider archive of active subglacial lakes may be similarly transient, and expands the range of subglacial settings in which water can accumulate.
How to cite: Booth, A., Killingbeck, S., Paris, N., Gimbert, F., Hawkins, J., Doyle, S., Ross, N., Peacey, M., Livingstone, S., Ing, R., Veness, R., Craw, L., Thorpe, S., Sole, A., Moffatt, A., and Kulessa, B. and the Additional members of the SLIDE Project team: Seismic observations reveal a hard-bedded drained subglacial lake basin beneath Isunnguata Sermia, West Greenland , EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-1960, https://doi.org/10.5194/egusphere-egu26-1960, 2026.
Surface meltwater influences Antarctic ice-shelf stability by enhancing melt, lowering albedo, and promoting hydrofracture. Although supraglacial hydrology is now recognised as widespread across Antarctica, existing meltwater feature records vary in spatial resolution and temporal sampling. Recent work tends to only classify supraglacial lakes while inconsistently representing more transient features such as slush, underestimating the full extent of surface meltwater. Furthermore, traditional threshold-based supraglacial meltwater mapping approaches require extensive manual post-processing and are difficult to scale; machine learning offers a promising alternative but requires systematic evaluation to ensure classifiers generalise reliably across time and space.
Here, as part of the ESA 5D Antarctica project, we present a new continent-wide, high-resolution record of supraglacial hydrology across all Antarctic ice shelves from 2016 to 2026. The dataset is derived from ~135,100 Sentinel-2 images using supervised machine learning implemented in Google Earth Engine. To systematically evaluate machine-learning approaches, which remain underexplored in glaciological applications and are often applied without rigorous validation, we compared five algorithms: Random Forest, Gradient Boosting Decision Trees, Classification and Regression Trees, Support Vector Machines, and k-Nearest Neighbours. Each was assessed using five complementary validation experiments: repeated cross-validation to assess internal consistency, independent validation against expert-labelled data to test external accuracy, leave-one-year-out cross-validation to evaluate temporal transferability, leave-one-region-out testing to assess spatial transferability, and controlled label corruption to quantify sensitivity to annotation error. Random Forest achieved the most consistent performance and ranked first overall with a mean macro-F1 score of 0.992 and was selected for continent-wide deployment.
The resulting dataset provides monthly classifications of supraglacial hydrology, distinguishing open meltwater features, including lakes, channels, and water-filled crevasses, from non-open meltwater features such as saturated firn and slush. The dataset is delivered alongside an interactive cloud-based application that enables users to visualise, classify, and export products on demand. By resolving Antarctic surface hydrology at unprecedented spatial and temporal scales and enabling on-demand delineation of meltwater features through a publicly available application, this work supports the assessment of processes relevant to ice-shelf stability and provides constraints for climate and ice-sheet modelling. This capability is increasingly important for understanding the role of surface meltwater in Antarctic ice-shelf systems under future warming.
How to cite: Glen, E., Leeson, A., Donachie, F., McMillan, M., and Phillips, J.: Continental-scale mapping of Antarctic supraglacial hydrology using machine learning, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-11784, https://doi.org/10.5194/egusphere-egu26-11784, 2026.
Increasing atmospheric warming is enhancing surface melt across Antarctica, leading to the widespread formation of supraglacial lakes (SGLs) on ice shelves. These lakes play a critical role in ice-shelf stability by promoting hydrofracturing, increasing flexural stresses, and potentially triggering ice-shelf collapse, thereby accelerating grounded ice discharge and sea-level rise. Accurate estimation of SGL depth and volume is therefore essential for understanding Antarctic ice dynamics.
To date, most large-scale SGL depth estimates have relied on optical remote sensing and radiative transfer methods (RTMs), which infer lake depth from spectral attenuation in satellite imagery. While effective, RTMs are sensitive to surface conditions, cloud cover, and water optical properties. The launch of ICESat-2 in 2018 provides a complementary approach, as its photon-counting lidar can detect returns from both lake surfaces and beds, enabling direct depth estimation. Several algorithms have been developed to extract SGL depths from ICESat-2 data, but applications remain spatially limited and none have yet focused on the George VI Ice Shelf (GVIIS).
Here, we present the first retrieval of supraglacial lake depths on the northern GVIIS using ICESat-2 ATL03 and ATL06 data. We apply the Watta algorithm (Datta and Wouters, 2021) to derive along-track SGL depth profiles and compare these results with independent depth estimates obtained using an RTM applied to Sentinel-2 and Landsat 8 imagery. This comparison is used to evaluate the strengths and limitations of ICESat-2–based depth retrievals relative to established optical methods.
Our study provides new constraints on supraglacial lake characteristics on the northern GVIIS and demonstrates the value of integrating active and passive remote sensing approaches to improve assessments of meltwater processes that influence Antarctic ice-shelf stability.
How to cite: Willis, I., Cheng, C., Metcalfe, C., Dell, R., and Banwell, A.: Retrieving Supraglacial Lake Depths on George VI Ice Shelf using ICESat-2, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-13736, https://doi.org/10.5194/egusphere-egu26-13736, 2026.
Since the 1950’s ice-shelf surface meltwater has been implicated as a driver of partial and complete collapse events across several ice shelves on the Antarctic Peninsula. Many of these events occurred alongside marked increases in 2m air temperatures in the later half of the 20th century, which drove increased rates of surface melting. Whilst this warming was briefly punctuated by a partial cooling across the Antarctic Peninsula from 1999 to 2014, it has since resumed. Given projections of non-linear increases in surface melting under future climate scenarios, surface meltwater-driven instabilities become increasingly important. Whilst most studies to date have focussed on the potential for hydrofracture, it is important that we begin to consider the potential for ice-shelf run-off and lateral meltwater export as Antarctica progresses towards Greenlandification.
Here we present remotely sensed evidence for the re-initiation of surface meltwater export from Bach Ice Shelf, following a 9-year hiatus. We combine optical remote sensing (Landsat 7, 8, and 9) with 43 years of regional climate model outputs (RACMO2.3p2 and ERA5) to evidence this change and consider the climate conditions that may have driven it. Variables considered include modelled surface mass balance, melt, and 2 m air temperatures. Melt-to-SMB ratios were calculated from modelled surface mass balance and melt.
In the austral summer of 2022/23, lateral meltwater export resumed on Bach Ice Shelf, ending the observed 9-year hiatus. This hiatus may have been driven by high surface mass balance and low snowmelt values, which resulted in low melt-to-SMB ratios. Such low melt-to-SMB ratios likely increased ice-shelf firn air content, which then took several years to overcome. The return of lateral meltwater export from Bach Ice Shelf in 2022/23 coincided with the highest modelled mean annual 2 m air temperatures. With continued atmospheric warming, we anticipate that Bach’s meltwater regime will continue to exhibit lateral meltwater export in the future, demonstrating the renewed importance of surface run-off for ice shelves on the Antarctic Peninsula.
How to cite: Dell, R., Maclennan, M., Trusel, L., and Bahrami, M.: Reactivation of lateral meltwater export from Bach Ice Shelf following a 9-year hiatus, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-19022, https://doi.org/10.5194/egusphere-egu26-19022, 2026.
Supraglacial channels are a primary pathway for transporting surface meltwater across ice sheets and ice shelves, and their geomorphology plays a key role in controlling meltwater routing. Meltwater commonly drains into supra- and subglacial features, where it can influence ice dynamics and mass loss. As surface melt rates increase under a warming climate, accurately constraining meltwater inputs to these systems becomes increasingly important.
While previous studies have modelled supraglacial channel evolution, few have explicitly accounted for the effects of solar radiation and spatially variable shear stresses on channel geometry. Here, we present a numerical model that simulates the temporal evolution of a supraglacial channel cross-section, incorporating atmospheric and radiative forcing, as well as hydraulic processes. Sensitivity analyses reveal that water temperature is a dominant control on channel incision, even for small variations in temperature. Additional simulations explore how water temperature evolves along the length of a channel and its implications for melt-driven erosion.
By explicitly resolving water temperature and energy exchange, this work provides a more complete description of supraglacial channel geomorphology. These results can be combined with field observations to improve estimates of meltwater routing and drainage volumes, with implications for surface hydrology modelling and ice-sheet mass-loss projections.
How to cite: Bianchi, G., Buzzard, S., Hawkins, J., and Prior-Jones, M.: Modelling the Geomorphology and Hydrology of Supraglacial Meltwater Channels, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-3544, https://doi.org/10.5194/egusphere-egu26-3544, 2026.
Moulins serve as critical hydrological conduits on the Greenland Ice Sheet (GrIS), facilitating the transfer of surface meltwater to the subglacial bed and directly modulating basal lubrication and ice velocity. Despite their significance, automated detection remains difficult; moulins are often sub-pixel scale in standard satellite imagery and are frequently misidentified as crevasses or inactive stream segments due to spectral overlap. This study introduces GrIS-MDM (GrIS Moulin Detection Model), a novel hydrology-informed framework designed to automate moulin extraction using ultra-high-resolution (0.06 m) unmanned aerial vehicle (UAV) imagery.
The GrIS-MDM framework synergises topographic data from Digital Elevation Models (DEMs) with spectral information from Digital Orthophotography Maps (DOMs) through a sequential tripartite workflow. The process begins by identifying tubular depressions using a contour-derived K-index to effectively eliminate shallow noise and surface artifacts. Subsequently, a multistage attention ResU-Net (MAResU-Net) is implemented to segment supraglacial river networks, utilising an automated sample collection protocol that substantially reduces manual labelling requirements. Finally, topological constraints are applied to isolate true moulins at river termini, distinguishing them from spurious depressions within the river interiors.
Validation conducted in the Sermeq Avannarleq region yielded a recall of 0.795 and a precision of 0.729. Experimental results demonstrate that GrIS-MDM achieves a 20.4% improvement in F1-score over traditional depth-based sink-filling methods. Integrating these detected moulins into hydrological models increased the spatial consistency of reconstructed stream networks by 5.8%. Furthermore, drainage analysis confirmed the model’s accuracy, with simulated water capture (90.8%) closely aligning with ground-truth observations (92.3%). Sensitivity tests indicate the framework remains effective at 2-m resolution, suggesting strong potential for deployment with high-resolution satellite platforms such as WorldView or ArcticDEM. This research offers a robust tool for enhancing high-precision supraglacial hydrological modelling and refining GrIS mass balance assessments.
How to cite: Chen, P., Chen, R., Cheng, X., and Chen, Z.: GrIS-MDM: A Hydrology Knowledge-Based Framework Combining Deep Learning Network for Moulin Detection Using Ultrahigh-Resolution UAV Imagery, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-4728, https://doi.org/10.5194/egusphere-egu26-4728, 2026.
Surface meltwater on ice shelves, together with runoff from upstream grounded ice, can pond and drain into crevasses, raising water pressure and promoting hydrofracturing. However, the timing of meltwater storage and release in coastal ablation and blue ice areas remains poorly constrained because continuous subsurface temperature records are scarce. Here we present one year of subsurface ice-temperature measurements from the blue ice area on Langhovde Glacier in coastal East Antarctica from January 2024 to January 2025, revealing persistent internal warming and a vertical temperature structure consistent with absorption of shortwave below the surface. We combine our observations with a coupled surface energy balance and firn model to isolate the effect of shortwave penetration into the surface. Simulations without subsurface shortwave absorption fail to reproduce the observed warming and its seasonal persistence, whereas including shortwave penetration substantially improves the simulated vertical temperature profile and supports an interpretation involving subsurface melt and refreezing. Observed near-surface ice temperature was ~3 °C higher than that model output. At 10 m depth, ice temperature was ~−6 °C, about 4 °C warmer than the site’s mean air temperature. These results indicate that shortwave driven subsurface heat storage is likely a key control on near surface thermal conditions in Antarctic blue ice area and may influence the seasonal opening and closure of meltwater pathways, thereby affecting the timing of runoff discharge to downstream ice shelves.
How to cite: Saito, J. and Minowa, M.: Shortwave Penetration Drive Subsurface Warming and Melt on Langhovde Glacier, East Antarctica, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-11462, https://doi.org/10.5194/egusphere-egu26-11462, 2026.
Ponded crevasses on the Greenland Ice Sheet impact the retention and drainage of surface meltwater in the low elevation zone and enhance the melt of ice sheet. However, current understanding is limited to a few glaciers, leaving their inter- and intra-annual evolution largely unknown at an ice-sheet-wide scale. This study presents an automated methodology to map ponded crevasses from 10-m resolution Sentinel-2 satellite images to produce the first pan-Greenland dataset describing their intra-annual evolution during the warm 2019 melt season. The results indicate that: 1) the central-western and southwestern basins exhibit earlier drainage onsets, due to the high surface meltwater runoff in the early melt season; 2) local topographic depressions favor higher areal expansion and shrinking rates of ponded crevasses; 3) towards the end of the melt season, a considerable amount of ponded crevasse area remains (~10% relative to the peak), suggesting the possible retention. We will further apply the proposed method to the Sentinel-2 images from 2016 to 2025 to reveal how the intra-annual evolution of ponded crevasses responds to varying climatological settings.
How to cite: Zhang, W., Liu, L., van den Broeke, M., Wouters, B., Xu, X., Wang, Y., and Yang, K.: Intra-annual evolution of ponded crevasses across the Greenland Ice Sheet, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-14990, https://doi.org/10.5194/egusphere-egu26-14990, 2026.
Nearly all meltwater produced on the Greenland Ice Sheet surface is routed through the interior of the ice, and the spatial and temporal patterns – as well as the mode of delivery – of discharge to the bed can have significant consequences for processes including ice fracture, rheology, and basal sliding. Existing suggest that a majority of meltwater in Greenland is transferred to the bed via surface crevasse fields, rather than lakes or moulins. However, in contrast to well-observed phenomena such as supraglacial lake drainages, little work has been done to explore how this process should be parameterised in regional-scale models that route meltwater from the surface to the bed. Here, we explore: (i) how well observations of crevasse field filling and drainage support existing parameterisations based upon linear elastic fracture mechanics (LEFM); (ii) what modifications may need to be implemented to better represent crevasse field hydrology, including the choice of proxy for resistive stress (Rxx) and the inclusion of seasonally-varying stress; and (iii) the potential consequences for effective and sliding at the glacier bed, as represented through subglacial hydrological models.
How to cite: Chudley, T., Stokes, C., Lea, J., Law, R., Hepburn, A., and Clason, C.: Parameterising crevasse field drainage into meltwater routing models for the Greenland Ice Sheet, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-12454, https://doi.org/10.5194/egusphere-egu26-12454, 2026.
At present, there is minimal surface meltwater over Antarctica and Antarctic subglacial drainage systems are isolated from supraglacial water, differentiating them from the surface meltwater-fed hydrological networks of Greenland. However, projected increases in surface melt across grounded regions of the Antarctic Ice Sheet raise the possibility that surface-to-bed hydrological connections may begin to form via moulins, features known in Greenland to drive seasonal ice velocity variability by modulating subglacial water pressure. In Antarctica, additional seasonal water fluxes into the subglacial drainage system could amplify the effects that subglacial channels have on ice dynamics and sub-ice shelf melt rates, which could, in turn, impact grounding line positions and stability and alter the rate of sea level rise. Here, we investigate the response of the Amundsen Sea Embayment (ASE), where rapid, potentially irreversible changes are underway, to new subglacial meltwater forcing from moulin inputs. To accomplish this, we develop a moulin prediction algorithm that uses surface melt projections at 2100 and 2300 from UKESM and strain rates derived from the Ice Sheet and Sea Level Systems Model (ISSM). We then use these moulin locations and discharges as input to the Glacier Drainage System (GlaDS) subglacial hydrology model. We simulate five consecutive melt seasons followed by an extended recovery period to evaluate whether episodic meltwater inputs leave a long-term imprint on the ASE's drainage system. We provide a mosaic of possible trajectories for ASE's subglacial drainage system by varying GlaDS internal parameters and the strain-rate threshold, complemented by an additional set of experiments that use a randomly generated moulin distribution. This approach allows us to better gauge the sensitivity and responsiveness of the ASE's drainage systems to perturbations from surface meltwater inputs, with knock-on effects for glacier stability. Ultimately, our work provides an improved understanding of how surface-bed hydrological pathways may influence the future evolution of the Antarctic Ice Sheet in a warming climate.
How to cite: Hayden, A.-M., Dow, C., Hill, T., Ehrenfeucht, S., and Pelle, T.: Modelling subglacial responses to future moulin inputs in the Amundsen Sea Embayment, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-14926, https://doi.org/10.5194/egusphere-egu26-14926, 2026.
Calving glaciers (lake-terminating or tidewater) represent a key uncertainty in future glacier projections, particularly for the ice sheets, where they are responsible, respectively, for around 40% of ice mass loss (Greenland), and nearly 100% (Antarctica). Accurately modelling these glaciers into the future is therefore crucial for being able to forecast future mass loss and the associated sea-level rise. Yet, long-term predictions of the evolution of these systems remain extremely challenging, as both calving and subglacial hydrology, to which these glaciers are highly sensitive, are difficult and/or expensive to model at longer timescales.
One possible solution to better model long-term calving is to represent it as a stochastic process based on the theory of self-organised criticality. Calving can be typified by two types of self-organisation around, respectively, ice cliffs (serac-style events) and ice tongues (full-thickness events). We show that both modes of calving can be influenced by subglacial hydrology.
Here, we therefore present recent work building on the calving implementation of the stochastic crevasse-depth calving function in the open-source ice-flow model, Elmer/Ice, allowing significant improvements in long-term predictions of calving rates and styles. We couple the new calving implementation with the version of the Glacier Drainage System (GlaDS) subglacial hydrology model available in the Elmer/Ice code and present results from a Greenlandic tidewater glacier, Sermeq Kujalleq (Store Glacier). This allows us to explore the interaction of calving, subglacial hydrology, and meltwater plumes, and also provides useful insight for similar future modelling efforts. Using the capabilities of the Elmer/Ice model to resolve the full 3D stress field and allow unconstrained terminus geometries gives us unparalleled insight into the melt-driven serac calving that results from undercutting of the terminus. Furthermore, the subglacial hydrology modelled by GlaDS can directly influence full-thickness calving through increased basal water pressure promoting fractures near the base of the glacier. Consequently, we show that coupled hydrology-calving modelling promotes increased calving in summer when water pressures are high compared to winter.
How to cite: Cook, S., Benn, D., and Wheel, I.: Water/Falls: Coupling stochastic calving and subglacial hydrology at Sermeq Kujalleq (Store Glacier), EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-3041, https://doi.org/10.5194/egusphere-egu26-3041, 2026.
Understanding the subglacial hydrology of glaciers is key in adequately modelling ice flow and future retreat of glaciers under climate change. Subglacial drainage systems are observed and modelled to change with the seasons, varying between an inefficient, distributed drainage system in winter and an efficient, channelised drainage system during the melt season. This directly influences a glacier’s basal flow velocity. Subglacial channels close at the end of the melt season by ice creep and the decrease of melt reducing subglacial discharge, and are typically supposed to disappear completely on many Alpine glaciers before the start of the next melt season. We modelled the evolution of the subglacial discharge system over the course of a year on Hintereisferner, Austrian Alps, using the Glacier Drainage System model (GlaDS). Additionally, we did three ground-penetrating radar (GPR) acquisitions of the glacier tongue over the course of a year (April 2025-March 2026). In both winter and summer acquisitions, en- and subglacial channels could be observed. Besides minor changes in the drainage system, a large subglacial channel was repeatedly detected, indicating possible long-term persistence of channelised drainage throughout the winter despite the expected shutdown of the efficient subglacial drainage system. We present our observations and compare them to the model results. Furthermore, the implications for ice flow and glacier evolution are discussed.
How to cite: Walker, C., Jovanovic, N., and Cook, S.: Interseasonal persistence of large subglacial channel on Hintereisferner, Austria., EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-5266, https://doi.org/10.5194/egusphere-egu26-5266, 2026.
The GLOBE (Greenland Subglacial Lake Observatory) project aims to create a comprehensive inventory of Greenland's active subglacial lakes through systematic analysis of high-resolution Digital Elevation Models (DEMs) combined with satellite altimetry data. In particular, the complete 2-meter resolution ArcticDEM database covering the entire ice sheet between 2008 and 2025 is utilised for this purpose. Since initial proof-of-concept studies demonstrated the viability of detecting subglacial lake dynamics using elevation variance, substantial progress has been made in extending this approach across larger spatial scales and more diverse glaciological conditions.
In this work, we present methodological refinements necessary for robust, systematic detection of subglacial lake signatures at the ice-sheet scale. We evaluate the impact of different coregistration methods (including no coregistration) on the reliability of the standard deviation of elevation maps used for lake identification. Additionally, we assess the effectiveness of detrending long-term ice-sheet elevation changes to isolate the shorter-term, localised elevation variations associated with subglacial lake drainage and filling events. As the project scales to cover all the ice sheet, understanding how these processing choices affect detection accuracy, precision, and the separation of signal from noise is critical for ensuring robust, reproducible results.
Preliminary results presented here offer insights into the relative importance of the different parameters and steps involved in the GLOBE pipeline and their contribution to lake identification. We demonstrate how these results inform best practices for systematic subglacial lake mapping across Greenland and the subsequent integration of subglacial hydrology into larger-scale ice sheet models, thereby improving predictions of ice sheet stability and mass loss.
How to cite: Moral Pombo, D., McMillan, M., Bowling, J., Hardy, D., Close, R., and Phillips, J.: Optimising detection of Greenland's active subglacial lakes with DEMs: evaluating coregistration and detrending strategies, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-7285, https://doi.org/10.5194/egusphere-egu26-7285, 2026.
We present observations documenting the drain–fill cycle of a large subglacial lake and the associated velocity response on the Totten Glacier, East Antarctica. The lake transitioned from a period of stability to rapid drawdown, dropping by more than 50 m in less than a year. Following drainage, glacier flow speed decelerated by ~100 m yr⁻¹ (~20%) over the lake and by ~30 m yr⁻¹ in a region immediately downstream. We hypothesize that the pronounced slowdown over the lake reflects complete drainage and increased basal traction associated with grounding, while the downstream deceleration results from disruption of the subglacial hydrologic system. Over the subsequent three years, as the lake refilled, glacier flow speeds recovered to pre-drainage levels.
How to cite: Winberry, J. P., Greene, C., McCormack, F., Cook, S., and Dow, C.: Temporary slow-down associated with drainage of a large subglacial lake, Totten Glacier, East Antarctica, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-8915, https://doi.org/10.5194/egusphere-egu26-8915, 2026.
Active subglacial lakes beneath the Antarctic Ice Sheet undergo repeated filling and drainage cycles that are now routinely identified using satellite altimetry. Recent studies have shown that such lakes are widespread and dynamically active, with more than few dozen active systems detected across Antarctica (Wilson et al., 2025). Even though these lakes are expected to influence basal water pressure, their impact on surface ice-stream velocity are inadequately constrained by observations.
In this study, we examine whether subglacial lake fill and drain events cause measurable, time-lagged changes in ice-stream surface velocity. The analysis focuses on a few well-documented active subglacial lakes located near major ice streams. Surface elevation time series from CryoSat-2 are used to identify lake filling and drainage phases and to quantify the timing and magnitude of individual events, also with ICESat-2 data used for validation where available. Ice velocity time series derived from Sentinel-1 are extracted at multiple locations upstream and downstream of each lake and are detrended to isolate velocity anomalies.
We analyse the relationship between lake elevation changes and velocity anomalies using correlation and lead lag methods, and assess how any velocity response varies with distance from the lake. Statistical significance is evaluated relative to background velocity variability and sensitivity to spatial averaging. The results provide observational constraints on the coupling between subglacial hydrology and ice dynamics, and help to assess whether subglacial lake activity can produce detectable surface velocity responses at satellite resolution.
Wilson, S.F., Hogg, A.E., Rigby, R. et al. Detection of 85 new active subglacial lakes in Antarctica from a decade of CryoSat-2 data. Nat Commun 16, 8311 (2025). https://doi.org/10.1038/s41467-025-63773-9
How to cite: Moqadam, H. and Sasgen, I.: Observations of ice-stream velocity variability associated with subglacial lake activity in Antarctica, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-15205, https://doi.org/10.5194/egusphere-egu26-15205, 2026.
Subglacial hydrology exerts an important control on ice flow and influences the evolution of downstream hydrology, as well as the occurrence of glacial lake outburst floods. However, large-scale modelling of subglacial hydrology remains computationally expensive due to the presence of nonlinear processes.
Within our DeLIGHT (Deep-Learning-Informed Glacio-Hydrological Threat) framework, we aim at enabling coupled simulations of ice flow, subglacial hydrology, and downstream hydrology, with the goal of improving predictions of ice flow evolution and the timing of peak runoff. For this purpose, we will leverage recent advances in deep learning. As a first step, this research focuses on the development of a subglacial hydrology emulator trained using output from the Glacier Drainage System (GlaDS) model implemented within Elmer/Ice, with the aim of applicability to mountain glacier catchments worldwide. The emulator is based on a class of deep learning architecture called neural operators, which allow for better generalisation compared to classical neural networks.
To generate the training set, GlaDS is forced using meltwater inputs derived from a calibrated degree-day model, which is driven by daily climate data spanning the 2000–2010 period. We select 70 glaciers spanning a wide range of physiographic characteristics across Svalbard, Scandinavia, the Alps, and Central and Southeast Asia to provide a representative range of mountain-glacier subglacial hydrological scenarios within the training set. We present results from the training simulations and initial directions for the development of the emulator.
How to cite: Jovanovic, N., Cook, S., Zwinger, T., Fürst, J., and Walker, C.: Emulating a Subglacial Hydrology Model with a Neural Operator, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-9501, https://doi.org/10.5194/egusphere-egu26-9501, 2026.
The evolution of the subglacial drainage system plays a key role in the seasonal behavior of ice surface velocities both on mountain glaciers and ice sheets. This relationship offers an opportunity to evaluate subglacial hydrology models against observations of large spatial and temporal coverage. Although surface velocities are only an indirect measure of the subglacial water pressure, and such a comparison therefore requires the coupling to an ice flow model through a sliding law, it presents a promising but currently still under-explored opportunity for the challenging task of constraining subglacial hydrology models.
In this study, we test a coupled ice flow and subglacial hydrology model in its ability to reproduce observed patterns of seasonal surface velocity on a group of land-terminating glaciers in southwest Greenland. These glaciers exhibit a variety of seasonal behaviors, including a spring acceleration with constant winter speed, a pronounced speed minimum in autumn, as well as a sustained speed-up throughout the winter. We use the Glacier drainage system model GlaDS, coupled to a higher-order ice flow model in a way that the whole system of equations is solved simultaneously, which allows an immediate two-way coupling. While the original GlaDS is restricted to constant winter water pressures (and velocities), we explore model extensions such as a laminar/turbulent transition that yield more complex seasonal behaviors. We investigate the interplay of the physical model choices, model parameters, properties of the bed topography, meltwater input and other factors to determine which conditions are necessary to produce certain types of seasonal behavior, and we evaluate if the observed patterns can be reproduced successfully in a realistic modeling framework.
How to cite: Pohle, A. and Aschwanden, A.: Reproducing seasonal patterns of surface velocity with a coupled ice flow and subglacial hydrology model, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-14994, https://doi.org/10.5194/egusphere-egu26-14994, 2026.
Subglacial drainage through ice and sediment incised channels, so-called canals, likely impacts the dynamics of soft bedded glaciers and ice streams as well as dictating sediment evacuation from ice sheet and glaciers. The original work of Walder and Fowler (1993) found that canals would form a distributed drainage system, i.e. that many small canals would be favoured over one large canal for a given discharge. We present a simple 1D numerical canal model which simulates both ice and sediment incision, water flow and sediment transport whilst assuming a fixed shape of the channel. With this model we investigate the distributed vs channelised behaviour of canals and extract which processes impact this dichotomy.
How to cite: Werder, M. and Utkin, I.: Canals: distributed versus channelised, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-20314, https://doi.org/10.5194/egusphere-egu26-20314, 2026.
Accelerated global warming is driving significant mass loss from the Greenland Ice Sheet (GrIS), leading to widespread ice-margin retreat and increased meltwater runoff. Both processes have resulted in a widespread increase in the number, area, and volume of ice-marginal lakes. These proglacial water bodies are known to possess significant potential to influence ice motion and, thus, ice sheet stability by altering thermal and mechanical boundary conditions at the terminus. Consequently, there is an imperative to understand how recent expansions in lake extent have impacted ice dynamics around the margins of the GrIS.
In this study, we investigate changes in near-terminus ice dynamics for the 76 glaciers draining into the 70 largest ice-marginal lakes around the GrIS over a 25-year period (1999–2024). We analyze velocity variation within the near-terminus region to isolate the direct influence of the ice-marginal lake. The results show that the studied glaciers have experienced an overall slowdown in ice motion of 19% between 1999 and 2024. While we observe regional variability across the ice sheet, no individual sector exhibits aggregate acceleration. Overall, only 32% of the studied glaciers accelerated in their near-terminus region. In these specific instances, speed-up is driven primarily by ice-marginal thinning and/or proglacial lake deepening, both of which increase flotation and consequently reduce basal drag.
These findings challenge the prevailing view that ice-marginal lake growth inevitably promotes dynamic instability. While ice-marginal lakes are increasing around the GrIS, we find no evidence over the last 25 years that these systems are making an increasing contribution to ice mass loss via ice dynamics. This suggests that many ice-marginal lakes lack the depth or overdeepened bed topography necessary to induce significant flotation-driven instability.
How to cite: Wang, Y., Nienow, P., Otero, J., and Goldberg, D.: Widespread lake-terminating glacier slowdown despite ice-marginal lake expansion in Greenland, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-20359, https://doi.org/10.5194/egusphere-egu26-20359, 2026.
Posters virtual: Tue, 5 May, 14:00–18:00 | vPoster spot 1a
EGU26-767 | ECS | Posters virtual | VPS20
Seasonal evolution of supraglacial lakes in Northeast GreenlandTue, 05 May, 15:12–15:15 (CEST) vPoster spot 1a
Supraglacial lakes form seasonally on the Greenland Ice Sheet (GIS) during the melt season as surface meltwater accumulates in topographic depressions. These lakes are dynamic, rapidly draining through supraglacial channels or discharging via hydrofractures, contributing to surface runoff and triggering cascading drainage of nearby lakes. Quantifying the spatial and temporal variability of their area, depth and drainage patterns is critical for understanding GIS hydrology and their role in modulating ice sheet behavior. Here we present a quantitative comparison of supraglacial lake evolution and rapid drainage cascade dynamics between contrasting melt years on Northeast Greenland Ice Stream. We analyzed Sentinel-2 observations from the 2019 and 2020 melt seasons using an automated Otsu thresholding approach combining dual water indices such as NDWIice and NDWIGN with topographic depressions from ArcticDEM to map the lakes. Lake depths and volumes were estimated using an empirical relationship between Sentinel-2 reflectance and lake depth calibrated with ICESat-2 ATL03 photon altimetry. We identified rapid drainage events and quantified their spatial and temporal clustering into cascade sequences.
The analysis revealed distinct interannual contrasts in the timing, persistence, and areal extent of supraglacial lakes, reflecting the influence of seasonal temperature variability. In 2019, warmer conditions favored more sustained lake development and prolonged persistence, whereas cooler conditions of 2020 year led to a more rapid rise-and-fall pattern with reduced total storage. Lake formation exhibited a clear elevation dependence, initiating earlier at lower elevations and progressing upward as the melt season advanced. Mid-elevation zones such as 800 to 1000m acted as key reservoirs storing 80% of the total lake volume, hosting the most persistent and voluminous lakes, suggesting their importance in surface-to-bed meltwater routing. Rapid drainage events were different between years despite similar lake inventories. A total of approximately 600 drainage events were identified across both years. Among these approximately 30% of drainage events participated in cascades. Rapid drainage events were concentrated at lower elevations typically below 800m, with a substantial proportion occurring as part of cascading drainage sequences.
Overall, our results demonstrate that variations in melt-season intensity could modulate supraglacial lake persistence, drainage behavior, and cascading dynamics. These findings emphasize the importance of mid-elevation lakes as critical nodes in meltwater transfer and provide new insights into understanding of surface lake water storage and surface-to-bed hydrological connectivity across the NEGIS sector.
How to cite: Das K, G., Vijay, S., and Singh, S. K.: Seasonal evolution of supraglacial lakes in Northeast Greenland, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-767, https://doi.org/10.5194/egusphere-egu26-767, 2026.
