The scientific community is investing considerable effort in organising our current knowledge of the physical and biogeochemical properties of sea ice, exploring poorly understood sea ice processes, and forecasting future changes of the sea ice cover, such as in CMIP6.
In this session, we invite contributions regarding all aspects of sea ice science and sea ice-climate interactions in both the Arctic and Southern Ocean, including snow and sea ice thermodynamics and dynamics, sea ice-atmosphere and sea ice-ocean interactions, sea ice biological and chemical processes, sea ice observational and field studies and models. A focus on emerging processes and implications is particularly welcome.
Orals: Thu, 7 May, 10:45–12:30 | Room 1.34
Sea-ice ridges (or more precisely, deformed ice) constitute a large fraction of the Arctic ice pack, however, estimates range broadly from 30 to 70%. Yet, we know disproportionally little about their role in the Arctic sea-ice system, a system that is in rapid change. In situ studies of ridges are logistically challenging, and as a result, most research has focused on level sea ice, despite the significant proportion of sea ice residing in ridges.
Ridges evolve over time and provide an environment for ice growth very different than the typical level ice. Observations during the year-long MOSAiC drift expedition revealed several processes that could contribute to ice growth, even in summer, when most of the ice pack is melting. Indirectly ridges also affect, for example, melt pond formation, and also under-ice spreading of meltwater (under-ice ponds), that in turn affect the melt rates of level ice. However, these indirect effects of ridges have virtually never been quantified.
Rare observations from the MOSAiC expedition explored the unique habitats within sea-ice ridges, including surfaces of ice blocks in association with water-filled voids, showed distinct differences in biological properties compared to level ice bottom and pelagic biota. These ridge-specific habitats had unique protist and bacterial assemblages contributing to the high diversity and richness found in Arctic sea ice. In summer, ridges can be hotspots of protist biomass and can contain the majority (up to 80%) of sea ice algal biomass - an element not yet considered in Arctic assessments of sea-ice biomass.
We argue that sea-ice ridges should be included in assessments of Arctic Ocean biodiversity and biogeochemistry to fully understand the Arctic sea-ice ecosystem and its response to ongoing changes. Future efforts should not only investigate the complex physical and biological processes within sea-ice ridges but also integrate these processes into models. Only then can we predict how the changes in the Arctic icescape will affect atmosphere-ice-ocean-ecosystem interactions and how the ongoing changes of the ice pack will affect ridging in the future.
How to cite: Granskog, M. A. and the HAVOC (MOSAiC) team: Sea-ice ridges - an understudied yet key component of the Arctic sea-ice system, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-208, https://doi.org/10.5194/egusphere-egu26-208, 2026.
Over the last decade, Antarctic sea ice has experienced an abrupt and pronounced decline, the underlying causes of which remain incompletely understood. Using the eddy-permitting configuration of the coupled climate model AWI-CM3, we show that such an abrupt transition can result from wind-driven ocean–ice interactions. In the simulation, a sharp sea-ice collapse occurs in the late 2020s, after a prolonged period of relative stability. The decline begins in the Indian Ocean sector, where subsurface heat gradually accumulates beneath a persistently stratified surface layer. Episodic intensifications of the Southern Hemisphere westerlies, associated with positive phases of the Southern Annular Mode, can locally enhance wind-driven upwelling and weaken the upper-ocean buoyancy barrier. A pronounced event of this type, simulated in the late 2020s, is sufficiently strong to abruptly ventilate the accumulated Circumpolar Deep Water heat toward the surface triggering a rapid, but sustained, sea-ice retreat. In contrast, under the same forcing, the low-resolution configuration of the model exhibits a more surface-controlled warming regime with stronger upper-ocean stratification, limiting the ability of wind anomalies to induce a threshold-like transition and resulting in a more gradual, nearly monotonic sea-ice decline. These results highlight the key role of ocean stratification and wind–ocean coupling in enabling abrupt Antarctic sea-ice change.
How to cite: Zapponini, M., Kim, D.-W., Park, W., Rackow, T., Roach, L., Semmler, T., and Jung, T.: Delayed yet abrupt Antarctic sea-ice loss from wind-induced upwelling of ocean subsurface heat, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-12828, https://doi.org/10.5194/egusphere-egu26-12828, 2026.
Antarctic sea ice plays a critical role in regulating heat and gas exchanges between the atmosphere and the Southern Ocean by insulating the upper-ocean mixed layer. During winter, sea ice formation drives vertical mixing, deepening the mixed layer and entraining subsurface heat that feeds back on subsequent ice growth. Recent years have seen an alarming decline and increased variability in Antarctic sea ice. This trend highlights the urgent need to better understand and represent the processes controlling ice growth in predictive models. Despite its importance for climate projections, this coupled ice–ocean feedback remains poorly constrained because wintertime observations are sparse. Here, we use high-resolution, state-of-the-art large-eddy simulations to identify a previously unrecognized process in the Southern Ocean: overshoot convection. We show that sea ice formation generates energetic, meter-scale saline plumes that penetrate beyond the mixed layer and overshoot into the pycnocline. These plumes rebound upward, entraining warmer subsurface waters associated with Circumpolar Deep Water back into the mixed layer and enhancing upward heat fluxes that moderate further ice growth. Recent years have seen an alarming decline and increased variability in Antarctic sea ice, underscoring the urgency of improving the representation of such processes in predictive models. We develop a theoretical framework to quantify plume-driven heat fluxes and apply it to an ice–ocean multi-year reanalysis dataset over the recent decade, demonstrating the large-scale relevance of overshoot convection. Our results indicate that this process limits Antarctic sea ice growth and spatial expansion, providing a physical explanation for observed regional variations in ice thickness. Crucially, overshoot convection is absent from current climate-scale models, highlighting a key missing process in projections of future Antarctic sea ice evolution and climate change.
How to cite: Gayen, B., Bhadouriya, A., Naveira Garabato, A., and Silvano, A.: Overshooting convective plumes and their role in winter Antarctic sea-ice growth, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-15536, https://doi.org/10.5194/egusphere-egu26-15536, 2026.
A consistent monitoring of the sea ice freeboard is crucial for observing changes in sea ice thickness and for improving sea ice forecasts. Mainly, area-wide freeboard heights are determined by using surface elevations from satellite altimetry missions like ESA's Earth Explorer Cryosat-2 and NASA's ICESat-2. A sudden failure or interruption of these altimeter missions would lead to significant data gaps in the central Arctic Ocean. Therefore, additional ways to determine sea level and freeboard based on remote sensing techniques are of great importance.
One of these additional techniques is grazing angle GNSS reflectometry (GA GNSS-R), which uses surface reflections from Global Navigation Satellite System (GNSS) signals to capture elevation information and characterize surface roughness. Here, we utilize reflections collected by up to 15 Spire Global nanosatellites. These nanosatellites record both the directly transmitted signal (line of sight) and the signal reflected from the surface , The signals are used to determine the delays in the measured phase compared to a modelled phase, from which the elevation of a reflecting surface is then derived. Accurate height estimation is feasible when the reflected signal remains phase‑coherent and retains right‑hand circular polarization at grazing incidence angles between approximately 5° and 30°. Under these conditions, smooth surfaces such as sea ice or calm water produce strong coherent reflections, whereas rougher ocean surfaces induce decorrelation and significantly weaker reflected signatures.
As part of the ESA Earthnet Data Assessment Project (EDAP+), a new approach is developed to derive sea ice freeboard from GA GNSS-R observations providing complementary information to satellite altimetry. One key advantage of Spire’ GA GNSS-R constellation is the high geographic density over the poles and thus continuous coverage of high latitudes without a systematic polar observation gap, as is the case with altimetry missions. Geolocated surface reflections and elevation data provided by Spire Global are used to classify surfaces, to detect water openings within the sea-ice cover (e.g., leads), to retrieve local sea-surface elevations, and subsequently to derive along-track sectional freeboard heights from GA GNSS-R observations.
This study presents first monthly Arctic-wide freeboard maps for a complete winter season (2023/2024) and comparisons with upward-looking sonar measurements and CryoSat-2 freeboard products. Initial comparisons demonstrate that GNSS-R-derived freeboard provides valuable complementary information, particularly in regions with a dense data coverage, and achieves accuracy comparable to altimetry during autumn and early winter.
How to cite: Müller, F. L., Ricker, R., Hansen, E., Halsall, K., Talpe, M., Cartwright, J., Dettmering, D., Seitz, F., De Laurentiis, L., Di Bella, A., and Bouffard, J.: Observing Arctic sea ice freeboard with high-resolution spaceborne grazing angle GNSS-Reflectometry, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-9934, https://doi.org/10.5194/egusphere-egu26-9934, 2026.
The ICESat-2 laser altimeter was launched in 2018 and has been measuring surface elevation used to determine sea ice thickness. The sea ice freeboards during winter (November-April) show strong agreement with in-situ observations. However, meltwater ponds accumulating on the sea ice over summer have prevented generating valid sea ice thickness observations from ICESat-2 in the summer months. Melt ponds hinder the classification of lead surfaces, leading to errors in the sea surface reference.
Utilizing more than 80 Sentinel-2 optical images coinciding within 5 minutes with ICESat-2 tracks, we have trained a Gaussian Mixture clustering algorithm that correctly classifies laser returns throughout the melt season. On testing images, the newly developed classification performs considerably better than the ATL07 classification, which was designed to be applied over winter conditions. The new classification was used to create an ICESat-2 summer laser freeboard dataset. The freeboards were validated with NASA's summer airborne lidar data acquisition campaign over July 2022.
The new summer freeboard dataset captures the melt season well, showing a quick decrease in freeboard during the melt of the snowpack in May and June and a slight increase at the start of autumn in September.
The summer sea ice freeboard dataset from ICESat-2 provides promising opportunities to calculate year-round sea ice thickness and volume, improve seasonal predictability, validate and improve the representation of sea ice in coupled climate models, and improve shipping risk assessment.
How to cite: Glissenaar, I., Landy, J., Liu, W., and Buckley, E.: Advancing Arctic summer sea ice freeboard from ICESat-2, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-17515, https://doi.org/10.5194/egusphere-egu26-17515, 2026.
Ocean waves can trigger sea ice fractures. This process plays an important role on the evolution of the floe size distributions in the marginal ice zone (MIZ). However, the fracture of sea ice by waves is still poorly constrainted by observations, as it is difficult to precisely forecast, it mostly occurs on short time (minutes) and spatial scales (meters), and are often happening during extreme weather conditions. In order to better understand the fracture of ice by waves, we built on the work of Dumas-Lefebvre et al. [1] and designed an experiment where waves are generated by an icebreaker nearby a continous ice sheet. This experiment was realized in the Saguenay Fjord, Québec, Canada, in February 2024, in the context of the Transforming Climate Action (TCA) program. The continuous 12-cm thick layer of ice that recently formed was fully characterised before the CCGS Amundsen generated wave by sailing at a speed up to 15.8 knots. Wave propagation and ice break-up were recorded by wave buoys placed on the ice surface and by three unmanned aerial vehicles (UAV) in stationary flights overseeing the ice from different angles.
Using digital image correlation (DIC), geometric projections and rectifications, we developed a method to recover the three components of the waves velocity field from UAV observations. The accuracy of the method has been tested using the buoy signals as a reference, showing a quantitative agreement with a relative error of about 5%. Thanks to this stereo-DIC method, we obtained the full wave velocity field over a grid of 106x75 meters with a spatial resolution of 0.8 pix/m and a sampling frequency of 30 Hz. This reconstruction method offers a precise, high spatio-temporal sampling needed for future characterization of sea ice fracture, or any other dynamics of textured surfaces.
References:
1. Dumas-Lefebvre, Elie and Dumont, Dany Aerial observations of sea ice breakup by ship waves TheCryosphere (2023)
How to cite: Kuchly, S., Auvity, B., Eddi, A., Dumont, D., and Perrard, S.: Reconstruction of a wave-induced ice break-up using unmanned aerial vehicles and stereo-DIC methods, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-13285, https://doi.org/10.5194/egusphere-egu26-13285, 2026.
The melt onset (MO) of sea ice plays a crucial role in the polar climate system and is detectable by satellite remote sensing. This study used scatterometer data to identify the melt onset dates of Arctic and Antarctic sea ice. The MO identified by this method primarily reflects early changes in the physical properties of the snow layer and is therefore usually earlier than the melt signals detected based on sea ice concentration data. Results show that the melt onset dates in both polar regions exhibit a clear latitudinal dependence. During the study period (2007-2022), no statistically significant long-term trends were found in the average MO for the entire Arctic or for any of its sub-regions. Instead, interannual variability dominates the temporal evolution of MO in the Arctic. Similarly, most Antarctic sub-regions show no significant trends, with the notable exception of the Ross Sea, where MO has advanced significantly. This study further analyzed the trends in the Arctic surface energy budget based on multiple time windows. The net surface energy flux shows a consistent and statistically significant increase only in the Bering Sea across all considered time windows, which supports the earlier MO observed in this region. Although the advance in MO in the Bering Sea does not reach statistical significance, its magnitude (-3 days/decade) is substantially larger than that in most other sub-regions. Following the pronounced decline in Arctic sea ice extent that began in 2007, the trend of advancing MO has not been statistically significant. The underlying mechanisms driving this recent change warrant further investigation.
How to cite: Xu, R., Shi, Q., Yang, Q., Liu, J., Su, J., and Leppäranta, M.: Melt Onset of Arctic and Antarctic Sea Ice from Scatterometer Data, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-11998, https://doi.org/10.5194/egusphere-egu26-11998, 2026.
We report on a novel data set of the melt-pond fraction on Arctic sea ice. Melt ponds on Arctic sea ice are an important phenomenon of the summer-melt process. They reduce the surface albedo of sea ice substantially, by that influence the net shortwave radiation balance, and with that the amount of solar radiation energy that is received by the sea ice-ocean system in the Arctic during summer. This has also implications for under-ice biogeochemical processes and ice mechanics. Melt ponds increase the uncertainty in the sea-ice concentration retrieved from satellite microwave radiometer measurements. Melt ponds have been observed by a number of satellite sensors, mostly in the optical and near-infrared wavelength range. Here we present an updated version of a spectral un-mixing approach published earlier that led to a data set of melt-pond fraction on Arctic sea ice with 8-daily sampling for months May through August from 2000 through 2011. The approach is based on reflectance measurements of channels 1, 3 and 4 of the Moderate Resolution Imaging Spectroradiometer (MODIS) on board the Earth Observation Satellite (EOS) Terra satellite. We modified the approach and derived the daily melt-pond fraction on Arctic sea ice at 500 m and at 12.5 km grid resolution for months June through August from 2000 through 2024 from MODIS v6.1 observations. In addition, we provide the net ice surface fraction – aka the fraction of sea ice without melt ponds – and the fraction of open water between the ice floes. We performed an evaluation of our data set against various independent observations. Our MODIS melt-pond fraction agrees within -3 % to +4 % with independent estimates of the melt-pond fraction from very-high resolution optical satellite imagery and from Operation Ice Bridge Digital Camera System imagery. The MODIS open-water fraction is smaller by 2 % to 6 % than that from independent estimates. The net ice-surface fraction tends to be larger by 2 % to 9 % than that from independent estimates. The 12.5 km gridded product shows a slightly worse (by 1 %) agreement. While our 12.5 km gridded MODIS product under-estimates the melt-pond fraction from very-high resolution optical satellite imagery by about 2 % in the mean (median: 3 %), the Medium Resolution Imaging Spectrometer (MERIS) product over-estimates these independent estimates by about 8 % (median: 9 %). Our MODIS melt-pond fraction data set is available from Sadikni and Kern (2025): https://doi.org/10.25592/uhhfdm.18069.
How to cite: Kern, S. and Sadikni, R.: 25-years of summer-time daily melt-pond fraction on Arctic sea ice derived from TERRA-MODIS sensor measurements, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-17914, https://doi.org/10.5194/egusphere-egu26-17914, 2026.
The formation of superimposed ice on Arctic sea ice during the melt season is studied using in situ observations from ice mass balance instruments (SIMBA/IAOOS) and a one-dimensional thermodynamic model of snow and sea ice (LIM1D). Ice mass balance instruments measure vertical profiles of temperature and a proxy for thermal conductivity over a vertical extent of 5 m through the atmosphere, snow, ice and ocean at regular time intervals.
In this study, we analyse four time series from instruments deployed near the North Pole in April and drifting towards Svalbard, all of which clearly show periods of superimposed ice formation. The dataset covers three melt seasons (2013, 2015, and 2017). Air–snow, snow–ice, and ice–ocean interfaces are retrieved from the temperature and conductivity profiles with an accuracy of about 2 cm, but this uncertainty can become substantially larger during the melt period. Snowmelt periods are identified, and the subsequent formation of superimposed ice at the sea ice surface is estimated.
Numerical simulations with the LIM1D model are performed to complement the analysis of the observed time series. The sensitivity of the results to the atmospheric forcing is also analysed. Summer temperature profiles strongly depend on the longwave radiative flux. Simulations forced with ERA5 exhibit an earlier onset of snowmelt than observed, whereas experiments using ERA-Interim radiation lead to a substantially improved agreement with the observations. In contrast, variations in snow albedo and snow density have a limited impact during the pre-melt period.
The formation of superimposed ice is successfully reproduced by the LIM1D model. The role of precipitation and surface runoff in the days preceding superimposed ice formation, as well as the changes in sea ice suggested by the observations, are confirmed. Lastly, the deployment of the instruments modifies the surrounding environment and may influence local observations. This should be taken into account when interpreting ice mass balance measurements.
How to cite: Kilian, L., Sirven, J., and Sennéchael, N.: Formation of superimposed ice in Arctic : study from observations and modeling, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-7024, https://doi.org/10.5194/egusphere-egu26-7024, 2026.
MOSAiC expedition in 2019/2020 collected an unprecedented volume of sea ice and snow data. For example, over 50 autonomous drifters, more than 160 km of sea ice thickness transect lines, nearly 250 snow pits, hundreds of satellite images, and thousands of ship radar images. Studies based on these data have yielded, among others, new data-based estimates of snow thermal conductivity, discovered new freezing mechanisms in the ridges and pinned the limits of sea ice deformation scale invariance. MOSAiC data analyses show that the snow depth and ice thickness was largely ice-age independent and that most of the winter’s snow was trapped in the deformed ice and that the loss of snow to leads was small.
Numerous sea ice deformation events formed the rough topography that trapped the snow throughout the winter. In order to better understand these processes and interactions, drifter, satellite remote sensing, and ship radar data can be used to analyse the deformation processes themselves, including reactivation of the leads and pressure ridges. These dynamics can not be represented in continuum models due to the spatial scale discrepancy. Here we will present how the sea ice deformation data from remote sensing can be used to create ice topography to constrain the snow distribution processes and, with it, the ice growth. We will use the recently developed methodology of data-model fusion by Itkin and Liston (2025, The Cryophere) to quantify the local snow-ice processes on spatial extents relevant also to the continuum models.
How to cite: Itkin, P., Haapala, J., and Liston, G. E.: Sea ice and snow six years after MOSAiC: bringing data to numerical models, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-18290, https://doi.org/10.5194/egusphere-egu26-18290, 2026.
The decline in Arctic sea ice thickness (SIT) is a key indicator response of the cryosphere to anthropogenic climate change. While the drivers of this decline are debated, they involve a competition between the ice-albedo positive feedback that enhances ocean heat content and the thin-ice negative feedback enhancing ice speed and alters dynamic-thermodynamic interactions. To date, pan-Arctic assessments of the dynamic and thermodynamic responses have been limited to the CryoSat-2 era (2010–Present Day).
Here we apply recently developed thickness products to produce novel 30-year records of Arctic sea ice volume (SIV) budget, resolving dynamic and thermodynamic terms. Our residual growth is validated on IMB buoy data and has a slight positive bias of +0.14 m-1 and a strong seasonal cycle. We identify a step change increase of 1,300 km-3 in thermodynamic growth, coincident with the pause in volume decline of the winter of 2007/08, representing a regime change in system behaviour where 76\% of seasonal growth is explained by the mean thickness of the surviving sea ice (p<0.001). Furthermore, enhancement in growth is regionalised and strongly correlates with sea ice drift modes. Specifically, we find that cyclonic atmospheric circulation modes associated with a negative Arctic Oscillation promote a dynamically coupled thin-ice feedback by enhancing thermodynamic growth through divergence and advection of SIV. This effect is observed throughout the periphery seas and is variable on inter-annual and decadal timescales. Finally, we investigate the contribution of thickness, concentration and drift to budget term variance. Finding that 87% of volume flux variance is accounted for by SIT anomaly. We highlight the utility of volume budgeting for identifying unphysical spatio-temporal patterns in SIT datasets and provide a new long-term benchmark for constraining sea ice growth in climate model assessments.
How to cite: Mellor, B., Tsamados, M., and Heorton, H.: A regime change in Arctic sea ice growth, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-5123, https://doi.org/10.5194/egusphere-egu26-5123, 2026.
The Canadian Arctic Archipelago is undergoing rapid climate-driven changes, including prolonged melt seasons and delayed freeze-up that accelerate the loss of thick multi-year ice. Existing freeze-onset datasets rely on localized in situ observations or coarse-resolution NASA passive microwave (PMW) products at km scale. However, in situ measurements lack systematic spatial coverage, while PMW data suffer from severe land contamination in the narrow channels and inlets of the Archipelago, limiting their reliability in this complex coastal environment. Detecting freeze-onset using high-resolution synthetic aperture radar (SAR) have historically been challenging due to sea ice motion that introduce uncertainty by disrupting pixel-level correspondence and backscatter continuity across image sequences. This study introduces a novel approach that integrates temporal backscatter slope analysis with environmental constraints, such as surface air temperature thresholds to overcome motion-related challenges. The method enables robust detection of freeze-onset events across both multi-year ice and open-water areas at 40 m pixel resolution, producing spatially comprehensive maps of freeze-up timing across the Canadian Arctic. Analysis during 2015–2025 will reveal fine-scale spatial and temporal variability in freeze-up patterns within the Last Ice Area. This work will deliver the first high-resolution freeze-up dataset for the Canadian Arctic, providing a scalable foundation for pan-Arctic freeze-onset estimation.
How to cite: Mahmud, M. and Smith, B.: Detecting sea ice freeze-onset from Sentinel-1 synthetic aperture radar, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-5640, https://doi.org/10.5194/egusphere-egu26-5640, 2026.
Under multiple anthropogenic warming scenarios in CMIP6, Arctic sea ice is projected to undergo a rapid seasonal decline by the mid-21st century. Stratospheric Aerosol Injection (SAI) has been proposed as a potential intervention to mitigate Arctic warming, yet the sensitivity of Arctic sea-ice recovery to aerosol injection latitude, including its magnitude and governing processes, remains insufficiently quantified. Here, we investigate how SAI injection latitude influences Arctic sea ice recovery using simulations with CESM2-WACCM6, in which sulfate aerosols are injected at latitudes ranging from 45°S to 45°N under fixed injection rates.
Our results show that Arctic sea ice recovery exhibits a strong dependence on injection latitude, with injections closer to the North Pole producing a more rapid and robust recovery in both sea-ice extent and volume. This response is associated with coordinated changes in clear-sky and cloud radiative fluxes, as well as enhanced surface albedo, which together favor a surface energy balance conducive to ice growth. Notably, we find that Arctic sea ice recovery does not scale linearly with global mean surface temperature under SAI, highlighting the importance of injection latitude in shaping regional cryospheric responses.
How to cite: Kim, H., Kim, H., Visioni, D., and Bednarz, E.: Injection-Latitude Dependence of Arctic Sea Ice Recovery in Stratospheric Aerosol Injection Simulations, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-4698, https://doi.org/10.5194/egusphere-egu26-4698, 2026.
Various studies identified possible drivers of extreme Arctic sea ice reduction, such as observed in the summers of 2007 and 2012, including sea ice-ocean preconditioning, large-scale atmospheric circulation variability and synoptic-scale cyclones. However, a robust quantitative statistical analysis and a better understanding of the predictability of extreme sea ice lows are hindered by the small number of events that can be sampled in observations and numerical simulations. Recent studies tackled the problem of sampling climate extremes by using rare event algorithms, i.e., computational techniques developed in statistical physics to reduce the computational cost required to sample rare events in numerical simulations. Here we apply a rare event algorithm to ensemble simulations with the European community Earth-System-Model version 3 (EC-Earth3) to study extremes of intra seasonal pan-Arctic sea ice area reduction under present-day climate conditions. The rare event simulations produce sea ice area anomalies larger in magnitude than observed in 2012, and we compute statistically significant composite maps of dynamical quantities conditional on the occurrence of extremes with probabilities of less than 1%. We exploit the improved statistics of low sea ice states to study their drivers on synoptic to seasonal time scales, including a sea ice area and sea ice volume budget analysis to disentangle the roles of dynamic vs. thermodynamic forcing on the sea ice. On statistical average over the extremes, enhanced thermodynamic melting accounts for approximately 75% of enhanced sea ice area and volume loss and predominately occurs on the Pacific-North American side, while enhanced dynamic sea ice loss appears on the Eurasian side of the Arctic. Finally, we show that extreme sea ice lows are on average preceded by persistent cyclonic mean sea level pressure anomalies over the central to eastern Arctic during the spring-summer transition. These low-pressure systems promote sea ice loss thermodynamically due to enhanced moisture and heat flux convergence, cloudiness, surface downward longwave radiative fluxes and dynamically through the impact of anomalous winds on the transport of sea ice.
How to cite: Sauer, J., Zappa, G., Ragone, F., and Massonnet, F.: Insights on the statistics of extreme Arctic sea ice conditions from the EC-Earth3 climate model, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-12121, https://doi.org/10.5194/egusphere-egu26-12121, 2026.
Sea ice has exhibited a number of record lows in both hemispheres over the past two decades. While the causes of individual
sea ice lows have already been investigated, no systematic comparison across events and hemispheres has been conducted
in a consistent framework yet. Here, the global standalone ocean–sea ice model NEMO4.2.2-SI3 at 1/4° resolution is used
to decompose the sea ice mass budget. We separate the relative contributions of ice melt/growth and thermodynamic/dynamic
processes, both from a climatological perspective and for selected individual years. The seasonal cycles of Arctic and Antarctic
ice mass fluxes show similarities, such as the prevalence of basal growth and melt in the mass budget. The long-term evolution
of the mass budget terms reveals an increased importance of basal melt in both hemispheres, at the expense of surface and
lateral melt. Regarding sea ice lows, the model indicates that the Arctic summer 2007 anomaly was chiefly caused by dynamic
factors, while the Arctic summer 2012 event was rather explained by thermodynamic factors. The Antarctic summer 2022 event
was driven by dynamic processes transporting ice towards sectors where more melt than usual occurred. The Antarctic winter
2023 event was characterized by a lack of basal growth. This study emphasises the dominance of processes at the ice-ocean
interface in driving the ice mass evolution at all time scales considered here, and highlights the potential of the ice mass budget
decomposition to disentangle oceanic and atmospheric contributions in the evolution of the ice state in a changing climate.
How to cite: Richaud, B., Massonnet, F., Fichefet, T., Topál, D., Barthélemy, A., and Docquier, D.: Anatomy of Arctic and Antarctic sea ice lows in an ocean–sea ice model, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-9496, https://doi.org/10.5194/egusphere-egu26-9496, 2026.
The Atlantic meridional overturning circulation (AMOC) could substantially weaken over the next century due to climate change. The Southern Ocean (SO) is a key control of global ocean circulation and climate. Here, we use the latest generation of climate models to assess the impacts of this potential AMOC weakening on the SO and Antarctic sea ice, on timescales of less than a century. Following AMOC weakening, ocean transports move heat southwards into the SO, causing SO surface warming and sea-ice loss. We also identify a new atmospheric connection, from the tropics to Antarctica: this connection enhances warming and sea-ice loss in one SO region, but causes cooling and sea-ice growth in another. This shows that substantial AMOC weakening could impact the SO on multidecadal timescales. However, these SO changes resulting from AMOC collapse are much smaller than the projected direct impacts of greenhouse-gas-induced warming.
How to cite: Diamond, R., Sime, L., Schroeder, D., Jackson, L., and Holland, P.: A Weakened AMOC Could Cause Southern Ocean Temperature and Sea-Ice Change on Multidecadal Timescales, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-13249, https://doi.org/10.5194/egusphere-egu26-13249, 2026.
Antarctic sea ice expanded during much of the late 20th and early 21st century, in contrast to the rapid loss of sea ice observed in the Arctic over the same period. However, since the mid-2010s, Antarctic sea ice has produced a series of record minima, with 2022 – 2025 collectively producing the lowest four extents on record. Climate models with adequate global warming have struggled to reproduce the general increase in sea ice coverage prior to 2016 and only very rarely simulate losses of the magnitude seen thereafter, raising concerns about their capability in simulating realistic variability in Antarctic sea ice state. Moreover, climate models have significant Southern Ocean biases, in particular an overestimation of Southern Ocean convection, which is a related source of low confidence in sea ice variability. Understanding the relationship between sea ice variability and the Southern Ocean state is particularly important because several recent studies have suggested that warming in the subsurface ocean is a significant driver of the repeated sea ice lows seen in the Antarctic since 2016.
In this study, we identify a series of substantial sea ice loss events within model output from simulations produced using a regional Southern Ocean circumpolar configuration of NEMO-SI3. We also characterise the modelled ocean salinity and temperature profiles leading up to each ice loss event. We then run a series of sensitivity studies where we either directly modify the ocean state or change parameters that impact the subsurface ocean. We will present results exploring how the magnitude of ice loss events is impacted by these changes in ocean state and discuss the implications of these results for the role of the ocean as a driver of Antarctic sea ice change and variability.
How to cite: Bateson, A., Feltham, D., Rogalla, B., Bilge, T., Naughten, K., Holland, P., Holmes, C., and Schröder, D.: Exploring the role of ocean preconditioning as a driver of Antarctic sea ice loss events, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-4061, https://doi.org/10.5194/egusphere-egu26-4061, 2026.
The formation of dense water in the Southern Ocean plays a key role in the global ocean overturning circulation, affecting the distribution of heat, carbon, oxygen and nutrients across the World Ocean. However, its representation in large-scale ocean–sea ice models used in climate studies remains biased. These models often produce dense water in the wrong locations and for incorrect reasons. We hypothesize that this partly stems from a poor representation of coastal polynyas and their drivers, particularly landfast ice. In a recent study, we introduced a velocity-restoring method to represent Antarctic landfast ice in the NEMO4-SI³ ocean–sea ice model and demonstrated its essential role in shaping coastal polynyas and controlling sea ice production. Here, we investigate the impact of this landfast ice representation on Antarctic shelf water properties and ice shelf melt. When the landfast ice scheme is activated, continental shelf waters densify in some coastal polynya areas, as expected. However, freshening is also observed beneath extensive landfast ice tongues, influencing salinity in several downstream polynyas. At a circumpolar scale, landfast ice improves the realism of bottom shelf water salinity and temperature. Notably, changes in mixed layer depth modulate the exchanges between the continental shelves and the open ocean, resulting in enhanced ice shelf melt. Overall, we show that representing landfast ice impacts the simulated ocean stratification, and formation and transformation of key Antarctic water masses. Our results thus further highlight the need for a physically-based representation of Antarctic landfast ice in Earth system models.
How to cite: Pirlet, N., Fichefet, T., Vancoppenolle, M., de Lavergne, C., and Jourdain, N. C.: Effects of a landfast ice representation on Antarctic shelf water properties and ice shelf melt simulated by NEMO4-SI³, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-5368, https://doi.org/10.5194/egusphere-egu26-5368, 2026.
The spread of climate model results in terms of quantifying the sea ice surface albedo feedback is partly caused by the sensitivity of the simulated sea ice surface albedo to surface warming. The accurate representation of the Arctic sea ice and its evolution throughout the year, particularly in the melting period, is crucial to obtain reliable climate model projections.
The rapidly warming Arctic leads to changes in sea ice properties. Arctic sea ice is becoming younger, transitioning from rough multi-year ice to flatter first-year ice. This influences the distribution and occurrence of melt ponds, which increase surface heterogeneity during the melting season. As a result, the surface albedo is altered and modelling it becomes more complex. However, many models are not able to simulate ice properties, like surface roughness, which would be beneficial for simulating melt pond fractions. A new melt pond fraction parametrization for the coupled Arctic climate model HIRHAM-NAOSIM has been developed using satellite data. This new parametrization includes a retarded response to temperature changes with different change rates for thin and thick ice. By taking only surface temperature and ice thickness as input variables, this parametrization can be applied to many climate models. An offline analysis with satellite data shows a higher correlation between the new parametrization and satellite observations (R=0.71) compared to the old parametrization (R=0.62). The offline analysis, first model results, and a comparison of model results with MOSAIC and satellite observations will be presented. The effects of the new parametrization on the radiative energy balance will be discussed.
How to cite: Schoch, P., Jäkel, E., Dorn, W., and Wendisch, M.: Recent improvements of the melt pond albedo parametrization in HIRHAM-NAOSIM, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-10675, https://doi.org/10.5194/egusphere-egu26-10675, 2026.
Melt ponds formed during summer play a crucial role in the evolution of Arctic sea ice. Observations show that the salinity in melt ponds ranges from 1 to 29 PSU, and saline ponds have different thermal properties from freshwater ponds. During the melt season, ponds with different salinities can exhibit distinct flow regimes and heat-transport efficiencies under the same radiative forcing, which can affect the relative fractions of absorbed heat that is emitted back to the atmosphere versus down into the ice. These feedbacks thus impact the evolution of pond depth. In the freezing season, the brine solution within a pond forms a porous mushy layer as it solidifies. If gravity drainage is triggered, the resulting plumes may induce complex circulation within the remnant unfrozen liquid beneath the ice lid and modify salinity transport within the underlying ice layer. These effects have not yet been fully quantified in existing models, despite their potential impact on the coupled pond–ice system.
We develop a one-dimensional pond-ice model based on an enthalpy method and a brine drainage model to explore how initial pond salinity influences the system over a melting–freezing cycle. We constrain the parameterised fluxes in the one-dimensional model using insight from a suite of two-dimensional high-resolution simulations, including double-diffusive convection and mushy-layer dynamics. Our two-dimensional simulations of double-diffusive convection during the melting stage show that salinity regulates the internal flow regime by controlling stratification, thus inhibiting turbulent convection at relatively high salinities. During the refreezing stage, two-dimensional simulations using the enthalpy method show that gravity drainage can occur across a wide salinity range, initiating turbulence even in the absence of external heat sources. This turbulence leads to highly efficient vertical salt transport. By varying the initial salinity in the one-dimensional model, we find that salinity can consequently influence both the maximum pond depth and the timescale of pond refreezing.
How to cite: Zhang, Y. and Wells, A.: A One-Dimensional Enthalpy Model for Melt and Refreezing of Saline Arctic Melt Ponds Constrained by Two-Dimensional Simulations, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-1093, https://doi.org/10.5194/egusphere-egu26-1093, 2026.
Most continuum models of sea ice follow the principles laid down by the Arctic Ice Dynamics Joint Experiment (AIDJEX), a US-Canadian initiative from the 1970s (McLaren, 1981). AIDJEX developed a model in which sea-ice floes break like a deformable plastic material at spatial scales of tens to hundreds of kilometres. This approach treats sea ice as a continuum, with the ice cover varying smoothly in space and time. Discrete element models (DEMs) model sea-ice floes as bonded groups of individual elements, explicitly capturing interactions with surrounding floes. Bond failure represents out-of-plane processes, such as ridging, or the formation of open water, such as leads.
We consider the results from simulating the dynamic and deformation properties of sea ice from a DEM and a continuum model using the same set of highly idealised scenarios. The results illustrate how the DEM can be used to provide an ensemble of possible outcomes around the mean behaviour provided by the continuum model, with the ensemble spread being linked to the spatial resolution being considered. The impacts of different stress confinement ratios are considered, illustrating a link between the DEM fracture patterns and fracture patterns that are observed in laboratory experiments on sea ice as well as in field observations.
Although the patterns of deformation from the DEM look more comparable to observed sea ice fracture patterns, the conclusions from this study support the use of continuum models for pack ice dynamics, especially at larger spatial scales.
How to cite: Coulthard, P., Feltham, D., Bateson, A., Schroeder, D., and Blockley, E.: Comparing stress and deformation characteristics of sea ice using continuum and discrete element models, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-7684, https://doi.org/10.5194/egusphere-egu26-7684, 2026.
As the summer Arctic sea ice extent has retreated, the marginal ice zone (MIZ) is becoming a larger fraction of the ice cover. The MIZ is defined as the region of the ice cover that is influenced by waves and for convenience here is defined as the region of the ice cover between sea ice concentrations (SIC) of 15 % to 80 %.
We use model simulations to analyse individual processes of ice volume gain and loss in the ice pack (SIC > 80 %) versus those in the MIZ. We use an atmosphere-forced, physics-rich, sea-ice-mixed layer model based on CICE, that includes a joint prognostic floe size and ice thickness distribution (FSTD) model including brittle fracture and form drag. Our model produces realistic simulations as compared with satellite observations of sea ice extent and PIOMAS (the Pan-Arctic Ice Ocean Modeling and Assimilation System) estimates of thickness.
We compare the ice cover and mass balance processes between the 1980s, 2010s and 2040s. The MIZ fraction of the July sea ice cover, when the MIZ is at its maximum extent, increases by a factor of 2 to 3, from 14 % (20 %) in the 1980s to 46 % (50 %) in the 2010s in NCEP (HadGEM2-ES) atmosphere-forced simulations. In a HadGEM2-ES forced projection, the July sea ice cover is almost entirely MIZ (93 %) in the 2040s.
Basal melting accounts for the largest proportion of melt in regions of pack ice and MIZ for all time periods. During the historical period, top melt is the next largest melt term in pack ice, but in the MIZ, top melt and lateral melt are comparable. This is due to a relative increase of lateral melting and a relative reduction of top melting by a factor of 2 in the MIZ compared to the pack ice. The volume fluxes due to dynamic processes decrease due to the reduction in ice volume in both the MIZ and pack ice.
For areas of sea ice that transition to being MIZ in summer, we find an earlier melt season: in the region that was pack ice in the 1980s and became MIZ in the 2010s, the peak in the total melt volume flux occurs 20(12) d earlier. This continues in the projection where melting in the region that becomes MIZ in the 2040s shifts 14 d earlier compared to the 2010s.
How to cite: Feltham, D., Bateson, A., Frew, R., and Schroeder, D.: Melting, freezing and dynamics of Arctic sea ice: pack ice versus marginal ice zone , EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-12450, https://doi.org/10.5194/egusphere-egu26-12450, 2026.
According to Copernicus data, the last three years (2023–2025) have been the hottest on record globally. In Northern Europe, however, the pace of warming has even exceeded the global average. This case study focuses mainly on the Estonian coastal area of the Baltic Sea, where air temperatures at coastal stations have risen by about 2.5°C between 1950 and 2025. As a result, the annual maximum Baltic Sea ice extent has declined (according to the endpoints of a linear trend) from 212 to 140 thousand km² (34%) over the centennial period between 1924/25 and 2024/25, and ice-cover duration in the Estonian coastal sea has decreased, depending on the station, by roughly 30–60% since 1950. This decline has likely contributed to the intensification of coastal erosion observed along several Estonian coastal stretches and more broadly along the southeastern Baltic Sea.
The study reviews shifts in ice-related processes with coastal geomorphic consequences, including changes in wave conditions due to longer ice-free seasons, reduced effective fetch, and shorter periods of frozen coastal sediments. Two less-studied effects are examined in detail: (1) using the ERA5-forced WRF–FVCOM modelling suite to quantify how the absence of ice affects sea-level patterns in the Gulf of Riga and the suppression of sea-level maxima during winter storms; and (2) analysing the occurrence and impact of winter (“warm”) upwelling on ice dynamics using ADCP measurements, meteorological–oceanographic observations, ice charts, and SST imagery.
The results show that reduced ice cover can help explain the higher storm surges observed in the Baltic Sea. In recent, more ice-free decades, winter storms have more freely produced surges, whereas before the 1980s they were often suppressed by sea ice. The combination of declining ice cover and the increasing probability of undamped storm surges likely contributes to steeper sea-level maxima and higher upper-quantile sea-level trends in long-term records. Secondly, although coastal upwelling in the Baltic Sea is usually considered a summertime process, similar forcing in winter can bring slightly warmer subsurface water (2–4°C) to the surface, contrasting with pre-freezingly cooled (0–1°C) surface water. The process is quite frequent along the straight North Estonian coast of the Gulf of Finland, when during sustained easterly winds this upwelled water creates a stark contrast with cold (–10…–20°C) weather conditions in the area. The phenomenon often delays coastal ice formation relative to the Finnish side and may help explain an anomaly in regional ice-pattern statistics in the mouth section of the Gulf of Finland.
Finally, as the transition zone between ice-free and seasonally frozen seas shifts northward under climate warming, the regions affected by these processes will also migrate. Winter upwelling has little effect in seas that are either fully ice-free or that freeze over rapidly. It is likely that about a century ago these processes were more pronounced south of Estonia (e.g., along the Latvian and Lithuanian coasts), and in the future they may shift farther north, such as into the Bothnian Sea.
How to cite: Suursaar, Ü., Mäll, M., Luik, K., and Tõnisson, H.: Effects on the transition zone of retreating seasonal sea ice in the Baltic Sea, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-2520, https://doi.org/10.5194/egusphere-egu26-2520, 2026.
To understand the response of sea ice to geophysical forcing, collecting field data remains essential. However, in situ monitoring of sea ice physical properties is challenging due to the complex logistics of the polar environment. Remote sensing methods such as Synthetic Aperture Radar (SAR) address some of these logistical constraints, but typically provide spatial information at regional scales and with temporal resolutions on the order of several days. While this is sufficient for large-scale monitoring, current sea ice models still struggle to accurately reproduce its decline, partly because small-scale processes are not adequately captured. Consequently, the next generation of models will require finer resolutions to describe local-scale ice–ocean interactions, especially in the context of the expanding Marginal Ice Zone (MIZ). This need is closely linked to our ability to observe small-scale sea ice properties so that breakup mechanisms associated with swell forcing can be better understood.
We introduce a methodology based on Distributed Acoustic Sensing (DAS) to recover small-scale variations in ice thickness and Young’s modulus. We test this approach on a dataset recorded in February 2025 on fast ice in the St. Lawrence Estuary (Canada). We demonstrate that sea ice properties can be estimated from both active and passive acquisitions on distances of the order of the km, with a spatial resolution of ~20 m.
How to cite: Moreau, L., Kuchly, S., Eddi, A., Perrard, S., and Dumont, D.: High resolution monitoring of sea ice properties with Distributed Acoustic Sensing, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-1462, https://doi.org/10.5194/egusphere-egu26-1462, 2026.
Taiwan’s Triton satellite carries a GNSS-Reflectometry (GNSS-R) payload designed to investigate bistatic L-band microwave scattering over the ocean and cryosphere. In this study, we present Arctic observations acquired during multiple overpasses from the late 2025 to early 2026, focusing on both the spatial coverage of specular points (SPs) and the physical interpretation of delay–Doppler map (DDM) signatures over sea-ice-covered regions.
Accumulated SP tracks over a three-months period demonstrate that Triton’s high-inclination orbit enables systematic sampling beyond 88°N, extending into the central Arctic Basin where conventional monostatic microwave sensors and existing GNSS-R missions, such as CYGNSS, do not provide coverage. In addition, Triton offers approximately four times higher resolution in both delay and Doppler dimensions compared to CYGNSS, enhancing sensitivity to subtle variations in surface scattering regimes.
Beyond spatial coverage, first-look analyses of high-resolution DDMs collocated with passive microwave sea ice concentration products reveal distinct scattering characteristics over marginal ice zone and partial ice cover conditions. Observed DDMs exhibit energy concentrated near the specular delay with pronounced elongation in the Doppler dimension, while remaining relatively confined in delay. This behavior is consistent with quasi-specular scattering from ice floes and reduced surface roughness.
The enhanced delay resolution further provides a framework to assess the potential contribution of subsurface or multi-layer reflections at L-band, which may become detectable under thicker or more consolidated ice conditions. These results indicate that high-resolution GNSS-R observations from Triton are sensitive not only to the presence of sea ice, but also to changes in scattering mechanisms related to ice structure and surface state. Ongoing work aims to systematically classify DDM observables across ice regimes and seasons, assessing the feasibility of GNSS-R as a complementary tool for Arctic sea ice characterization.
Figure. Spatial Cover over the Arctic region from 2025/10/02 to 2026/01/08 by TRITON satellite
How to cite: Duong, H., Chien, H., Chen, M.-Y., Lin, L.-C., and Yeh, W.-H.: High-Resolution GNSS-Reflectometry Observations of Arctic Sea Ice from Taiwan’s Triton Satellite, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-8600, https://doi.org/10.5194/egusphere-egu26-8600, 2026.
