Permafrost regions in subarctic and arctic areas harbor substantial carbon reserves, which are becoming increasingly vulnerable to microbial decomposition as soils warm. As the seasonally thawed active layer deepens and anthropogenic disturbances escalate, accurately predicting carbon fluxes from disturbed environments underlain by permafrost requires a comprehensive understanding of soil respiration dynamics. This study investigated the impact of surface disturbance on seasonal soil biological properties in a boreal forest ecosystem near Fairbanks, Alaska. Further, we sought to identify the key environmental and geochemical factors influencing soil biology in the undisturbed and disturbed soils. Our results revealed a substantial rise in soil respiration at the disturbed boreal forest site, which exhibited a 14.4 % overall increase in CO2CO20.60±0.16 °C, in stark contrast to the sub-zero temperatures of
The second Ice Shelf-Ocean Model Intercomparison Project, ISOMIP+, compares 12 ice shelf-ocean models with a common, idealised, static configuration, aiming to assess inter-model variability. Models show similar basal melt rate patterns, ocean profiles and circulation but differ in ice-ocean boundary layer properties. Ice-ocean boundary layer representation is a key area for future work, as are realistic-domain ice sheet-ocean model intercomparisons.
Ocean-driven basal melting of Antarctic ice shelves plays an important role in the mass loss of the Antarctic Ice Sheet. Ice shelf cavity-resolving ocean models are a valuable tool for understanding ice shelf-ocean interactions and for simulating projections of ice shelf and ocean states under future climate. Designed to assess the current state of ice shelf–ocean modelling, the second Ice Shelf–Ocean Model Intercomparison Project, ISOMIP+, consists of 12 ocean model configurations submitted with a common, idealised experimental setup. Here, we focus on the experiments Ocean0–2 (Asay-Davis et al., 2016), which are ocean models with idealised, static ice shelf geometries, but where the ocean reaches a balance with prescribed far-field ocean conditions. Different thermal transfer coefficient values (ranging from 0.011 to 0.2) are used for each model in the melting parameterisation to achieve a common, tuned melt rate since the models cover a range of types of vertical coordinates, ice–ocean boundary layer treatments, and numerical schemes. These model differences lead to spread in the resultant ocean properties, circulation, boundary-layer structure and spatial distribution of melting. We also highlight similarities between models, such as a shared linear relationship across most models between melt rate and overturning and barotropic streamfunctions during the spin-up and spin-down, demonstrating a robust relationship between melt and circulation across models and forcing conditions. The ISOMIP+ results provide a systematic comparison of ice shelf cavity-capable ocean models. However, we also demonstrate the need for realistic ice shelf–ocean model intercomparison projects (some already underway) to assess model biases and inter-model variation against sparse observations. Further research is needed to understand the differences between models and further improve our modelled representations of the ice–ocean boundary layer and ice shelf cavity circulation.
Making accurate measurements and predictions of the West Antarctic Ice Sheet's (WAIS) contribution to present and future sea-level rise fundamentally depends on knowing its trajectory over the last few thousand years. We present new in situ 14Cat least
Thermokarst lakes are among the most common and dynamic landscape features in ice-rich lowland permafrost regions. They influence carbon, water and energy fluxes between atmosphere and land surface, and are an important component of Arctic lowland hydrology. Despite their significant role in the climate system, thermokarst lakes are only rudimentarily or not at all represented in Earth system models (ESMs). Attempts at stand-alone modelling of their dynamics have mostly been limited to the scale of individual lakes. Because lake formation, expansion, and drainage depend on small-scale surface and sub-surface heterogeneities that are difficult to measure, a deterministic modelling-approach would be a challenge at the regional or pan-Arctic scale. We therefore treat these processes as probabilistic across a landscape and create a model of thermokarst lake dynamics using stochastic approaches. With the inclusion of stochasticity and volatility, our method allows us to account for the diversity of individual lake behaviour that results from the small-scale differences in environmental conditions. We present idealized simulations and, additionally, test novel high-resolution remote sensing data products that capture annual lake areas for model initialization and the calibration of inherent or climate-induced lake dynamics. Our model is able to capture three plausible regimes by incorporating the main processes behind thermokarst lake dynamics and represents a new step towards stochastic representation of permafrost landscapes in ESMs. Furthermore, our findings emphasize the importance of continuing the retrieval of remote sensing data for model parameterization and acquiring additional data products containing information on past thermokarst lake behaviour.
At the interface between the ocean and the ice shelf base, in the framework of the shear-controlled melt parameterisation, the ice melts due to combined actions of temperature, salinity and friction velocity. In the NEMO ocean model, the friction velocity is usually computed based on a constant drag coefficient and an ocean velocity averaged vertically within a distance from the ice, which is often referred to as the Losch layer. Instead, in this study, we use a logarithmic approach, where a constant hydrographic roughness length detetermines the drag coefficient through the law of the wall and the horizontal current speed is sampled in the first wet cell. The aim is to reduce the vertical resolution dependency, to homogeneise the sampling of horizontal current speed between the thermodynamic and the dynamic drag equation and to enable the use of a variable drag coefficient based on the subgrid-scale (or unresolved) ice shelf basal topography. The motivation behind a variable drag based on the topography comes from observations showing that regions with rough topographic features such as crevasses or basal melt channels experience more melts than flat ones. We compare different experiments in a configuration of the Amundsen Sea at
Sub-ice shelf melting is critical for the stability of the Antarctic ice sheet, as it influences ice-shelf buttressing that reduces grounded ice flow. Previous studies have emphasized that uncertainties in the state of sub-ice shelf melting contribute to uncertainties in future sea-level projections. To better understand how sub-ice shelf melt rates affect model initialization and predictions, we adopt a single ice-sheet model (PISM) and investigate two different sub-ice shelf melt rate schemes during model spin-ups. We then drive the Antarctic ice sheet into the future using identical environmental forcings. We find that, despite closely matched steady-state geometries achieved through the spin-up process with different sub-ice shelf melt rates, the prognostic simulations reveal significantly divergent ice mass changes, particularly in marine ice sheet regions. By 2100, the difference in global sea-level contributions from the Antarctic ice sheet can be as large as ∼ 57 %, primarily from West Antarctica. This discrepancy arises because the spin-up initialization method alters the ice sheet's dynamic state, such as basal friction and thermal regimes, leading to differences in ice-sheet mass changes over time. Therefore, this study underscores the importance of sub-ice shelf melting and ice-sheet model initialization methods in reducing uncertainties in predicting the Antarctic ice sheet's future.
Arctic supraglacial lakes volume changes serve as critical indicators of global temperature fluctuations. Accurate lake depth measurements are essential for reliable volume estimation, yet traditional bathymetry methods (e.g., airborne LiDAR and shipborne sonar) face significant challenges and high costs in the harsh Arctic environment, and are also inadequate for capturing the rapid temporal variability in lake depth. To address this, we utilized satellite-derived bathymetry (SDB) to achieve supraglacial lake water depth inversion. Specifically, three approaches were tested: (i) applying the radiative transfer equation (RTE) model (Philpot RTE model) using only Sentinel-2 multispectral optical imagery; (ii) combining ICESat-2 (Ice, Cloud, and Land Elevation Satellite-2) single photon-counting laser altimetry data with Sentinel-2 imagery through a semi-empirical log-transformed linear regression model (Lyzenga model); and (iii) a proposed novel approach that optimizes the model by considering the varying reflectance characteristics across different spectral bands in the water column. For the proposed method, we propose an SDB approach based on spectral stratification using the Otsu algorithm (maximum between-class variance method), and further integrate the spectral stratification with the Lyzenga model. To validate the effectiveness of the proposed method, we applied it to four relatively large and morphologically intact lakes on the southwest Greenland Ice Sheet (GrIS), using time-stamped ArcticDEM (Arctic Digital Elevation Model) strips as reference data. Compared with the Philpot RTE model, the root mean square error (RMSE) and mean absolute error (MAE) were reduced by up to 86.6 % and 89.0 %, respectively; compared with the traditional Lyzenga model, the RMSE and MAE decreased by up to 13.0 % and 14.0 %, respectively. The enhanced accuracy of our approach improves the ability to monitor volume changes in GrIS supraglacial lakes, providing valuable insights into their response to environmental changes.
In a warming world of glacier changes, the scientific community has dedicated increasing attention to debris-covered glaciers and their response to climate. A variety of models with distinct complexity and data requirements have been developed and widely used to simulate melt under debris at different sites and scales, but their skills have never been compared. As part of the activities of the International Association of Cryospheric Sciences (IACS) Debris Covered Glacier Working Group, we present an intercomparison exercise aimed at advancing our understanding of model skills in simulating ice melt under a debris layer. We compare 15 models with different complexity at nine sites in the European Alps, Caucasus, Chilean Andes, Nepalese Himalaya and the Southern Alps of New Zealand, over one melt season. We run the models with measured meteorological data from automatic weather stations and estimated or measured debris properties. We consider four main model categories: (i) energy balance models that calculate melt by solving the physics of heat transfer to the debris layer, but require a high amount of input data; (ii) a simplified energy balance model; (iii) enhanced temperature-index models; and (iv) simple empirical temperature-index models that have been extensively used given their low data requirement but require calibration of their empirical parameters. Model performance is evaluated using on-site measurements of sub-debris melt (for all models) and surface temperature (for models based on the surface energy balance). Our results show that physically-based energy balance models and empirical temperature-index models perform in a distinct manner. At one end of the spectrum, simple temperature-index models are accurate when recalibrated or when using site-specific literature parameters, and show poor results when parameters are uncalibrated. At the other end, energy balance models show a range of performance: the most accurate energy balance models are those with the highest degree of complexity at the atmosphere-debris interface. An important data gap emerged from our experiment: the poor performance of all models at three sites was related to the poor knowledge of debris properties, and specifically of thermal conductivity. Future work should focus on both: (i) consistent data acquisition to evaluate existing models and support new model developments; (ii) advancing models by accounting for processes such as debris-snow interactions, moisture in the debris and refreezing. We suggest that a systematic effort of model development using a common model framework could be carried out in phase II of the Working Group.
Snow-avalanche hazard in mountainous areas may change in frequency and severity due to climatic change, especially in Arctic regions such as northern Norway where temperature change is amplified. Expanding earlier work, we train machine-learning (ML) models on dynamically downscaled reanalysis data including snow-cover simulations to predict avalanche danger for the Troms county in northern Norway. In contrast to earlier work, the trained ML models can distinguish between avalanche types, in particular those of dry and wet snow. Due to insufficient avalanche observations, we construct a binary metric (avalanche day/non-avalanche day) based on the avalanche danger warnings published in the Norwegian avalanche bulletin. The ML models provide a hindcast of the frequency of avalanche days for the period 1970–2024 (based on reanalysis) and a projection into the future for the 21st century (based on climate model simulations). Over the historical period the results confirm earlier studies showing that while multi-decadal linear trends are marginal, the interannual variability of the avalanche-day frequency is linked to the Arctic Oscillation. The projected future changes indicate a general decrease of avalanche danger, especially for dry-snow avalanches. In contrast, wet-snow avalanche danger exhibits changes dependent on elevation, increasing at all elevations until mid-century, but thereafter continuing the increase only at higher elevation, while at lower elevation reversing to a decrease. Our results are in line with an emerging consensus of an overall decline of avalanche danger in the 21st century and a shift in avalanche characteristics towards fewer dry and more wet-snow avalanches. Such a shift can be challenging for avalanche-prone populations as the current knowledge of the local avalanche conditions may become less relevant and increasingly fail to provide protection from the avalanche hazard.
Antarctic coastal polynyas are key components of Antarctic marine ecosystems, influencing light and nutrient availability and open water access for marine predators. Thus, changes in the physical characteristics of polynyas can influence how these ecosystems respond to a changing climate. Here, we explore challenges inherent in identifying climatologically and biologically relevant Antarctic coastal polynyas on gridded data in both satellite and Earth System Model data. We find that it is critical to consider grid type and resolution, season, metric and threshold when defining polynyas. Regridding data, both spatially and temporally, can have significant impacts on identified polynya statistics. The spatial distributions of Antarctic coastal polynyas are significantly correlated between the two observational products as well as between the observational products and the model data. The Earth System Model we use here captures coastal polynya-like features that occupy ∼ 3 % of the area of the winter sea ice zone and contribute ∼ 17 %–21 % of the total sea ice zone marine net primary productivity (NPP). Temporal self- and cross correlations of integrated polynya areas and numbers are inconsistent across observational and modelling products. Trends in polynya areas are not robust, changing significance, magnitude and sign across threshold, grid and product.
Accurate, high-resolution projections of the Greenland ice sheet’s surface mass balance (SMB) and surface temperature are essential for understanding future sea-level rise, yet current approaches are either computationally demanding or limited to coarse spatial scales. Here, we introduce a novel physics-constrained generative modeling framework based on a consistency model (CM) to downscale low-resolution SMB and surface temperature fields by a factor of up to 32 (from 160 to 5 km
The Greenland Ice Sheet (GrIS) and its peripheral glaciers and ice caps (PGIC) have received a lot of attention with respect to its marine-terminating margin, where ice discharge contributes a significant amount to ice mass loss. However, a similar fraction of the mass loss is caused by surface melt, leaving the ice predominately via the less studied terrestrial margins. Using existing ice masks and a lake dataset we extract the actual land-terminating sections, making up 93.1 % of the total GrIS (70 900 km) and PGIC (170 600 km) margin. The study provides evidence for the ability of the ice mask and ArcticDEM to capture margin morphologies across approximately 84 % of the land-terminating margin correctly, even able to identify very steep margin morphologies at a regional scale. We identify 28.4 % as near-vertical features over shallow terrain and a further 13.4 % as steep (∼ 20–45°), which roughly corresponds to an earlier estimate of 45 % for ice cliffs. 17.3 % of the land-terminating margin are identified as shallow ramps (< 20°). With geolocated morphologies, future studies will be able to identify to what degree past presence of lakes, bed topography or ice dynamics and past climate at and upstream of the margin can explain under which conditions the ice margin terminates in steep or shallow slopes.
Metrics such as the mean annual ground temperature (MAGT) and active layer thickness (ALT) are used to monitor and quantify permafrost change. However, these have limitations including those arising from the effects of latent heat, which reduce their sensitivity. We investigated the behaviour of existing and novel metrics derived from temperature observations (TSP metrics) using an ensemble of more than seventy 120-year simulations. We evaluated which TSP metrics provide new insight into permafrost change and evaluated how reliably each one indicates changes in sensible, latent, and total heat contents for different levels of sensor quality. We also quantified the effect of sensor placement on the magnitude of observed MAGT trends.We observed depth-related differences in decadal MAGT warming rates of more than 0.23 °C decade−1decade−1
Since 2016, the Sentinel-3 satellites have provided a continuous record of ice sheet elevation and elevation change. Given the unique, operational nature of the mission, and the planned launch of two additional satellites before the end of this decade, it is important to determine the performance of the altimeter across a range of ice sheet topographic surfaces. Whilst previous studies have assessed elevation accuracy, more detailed investigations of the underlying instrument and processor performance are lacking. This study therefore examines the performance of the Sentinel-3 Synthetic Aperture Radar (SAR) altimeter over the Antarctic Ice Sheet (AIS), utilising new detailed topographic information from the Reference Elevation Model of Antarctica (REMA). Applying Singular Value Decomposition to REMA, we firstly develop new self-consistent Antarctic surface slope and roughness datasets. We then use these datasets to assess altimeter performance across different topographic regimes, targeting a number of key steps in the altimeter processing chain. We also evaluate the impact of topography upon waveform decorrelation. For the new Sentinel-3 Thematic Product, we find that, for 94.1 % of acquisitions, the point of closest approach to the satellite is successfully captured within the range window – an improvement of ∼ 5 % compared to the previous non-thematic (BC-004) product. For both products, performance declines with increasing topographic complexity, which also limits the ability to record all backscattered energy within the beam footprint. We estimate that 57.4 % of the ice sheet exhibits greater topographic variance within the footprint than can be captured by the range window, and that the current window placement captures a median of 89.2 % of the total possible topography that could be recorded. These findings provide a better understanding of the performance of the Sentinel-3 altimeters over ice sheets, and can guide the design and optimisation of future satellite missions such as the Copernicus Polar Ice and Snow Topography Altimeter (CRISTAL) and the Sentinel-3 Next Generation Topography mission.
Glacier terminus retreat involves complex processes superimposed at the interface between the ice sheet, the ocean, and the subglacial substrate, posing challenges for accurate physical modeling of terminus change. To enhance our understanding of outlet glacier ablation, numerous studies have focused on investigating terminus position changes on a seasonal scale with no clear control on seasonal terminus change that has been identified across all glaciers. Here, we explore the potential of machine learning to analyze glaciological time series data to gain insight into the seasonal changes of outlet glacier termini. Using XGBoost machine learning models, we forecast seasonal changes in terminus positions for 46 outlet glaciers in Greenland. Through the SHapley Additive exPlanations (SHAP) feature importance analysis, we identify the dominant predictors of seasonal terminus position change for each glacier. We find that glacier geometry is important for accurate predictions of the magnitude of terminus seasonality and that environmental variables (mélange, ocean thermal forcing, runoff, and air temperature) are important for determining the onset of seasonal terminus change. Our work highlights the utility of machine learning in understanding and forecasting glacier behavior.
Subglacial lakes influence glacier hydrology, dynamics, and mass balance; however, they are poorly documented outside the polar ice sheets. Here we use high-resolution digital elevation models during 2011–2021 and regression analysis to characterize subglacial lakes. We identified 37 subglacial lakes across the Canadian Arctic, 35 of which are newly identified. These lakes have an area of 0.3–48.5 km23. We classify these subglacial lakes into three types: (1) classic subglacial lakes, (2) terminal subglacial lakes at places where two glacier termini converge and coalesce, and (3) partial subglacial lakes with an area of open water at the ice margin. Types 2 and 3 are newly introduced in this study, there are 11 and 15 lakes classified as these two types, respectively. Lake activities negatively correlate with regional mass balance (
Climate change (CC) is significantly impacting the snow cover of the European Alps, compromising hydrological cycles, water stock for agricultural and civil supply, winter tourism. This study investigates Snow Cover Changes (SCC) in the Western Italian Alps (Piemonte and Valle d'Aosta regions) from 2000 to 2023, using MODIS satellite data. In particular, MOD10A1 images were processed in Google Earth Engine to derive daily snow cover, integral snow cover area (iSCA), snow persistence (SP), and mean daily snowed area (MDSA). Ground data from 96 snowmeter stations were used to validate the satellite-derived SP. The analysis of SCC was performed by quantifying long-term trends of MDSA at the pixel level. The normalized trend (nT) index represents the percentage change rate in snow-covered area per yearly mean snow event. It was mapped showing different spatial patterns of SCC in the study area. Results reveal an altitudinal gradient in nT, with the higher snow cover reduction occurring in lowland and within main valley areas, reaching −5 % below 1000 m a.s.l. and −1.8 % between 1000–1500 m a.s.l. These findings highlight the vulnerability of snow resources due to CC, impacting water availability, winter sports, and regional economies. This study can support adaptation strategies and sustainable resource management in the Western Alps by mapping critical areas where CC effects on snow must be mitigated.
Based on massive processing of heritage radar data from the satellite missions ERS-1/2, JERS-1, ENVISAT ASAR, ALOS PALSAR and Radarsat-2, and in combination with data from the current Sentinel-1 and ALOS-2 PALSAR-2 missions, we compiled a ∼ 30-year time series of radar backscatter over Svalbard. We exploited this data to detect glacier surges by using changes in backscatter as an indicator of increased or decreased surge-related crevassing. In this way, we reconstructed an as consistent as possible time series of surge activity on Svalbard for 1992 to 2025. We recorded 24 surge-type events during the pre Sentinel-1 period 1992–2014 (23 years) and 34 surge-type events during the post Sentinel-1 period 2015–2025 (11 years). This time series shows an approximately threefold increase in surges since 2015, from an average of about one surge per year to more than three surges per year. We show that this increase is unlikely to be explained alone by the better resolution, coverage and quality of the Sentinel-1 data compared to the data from the earlier SAR heritage missions. Simulation results indicate that the observed increase is extremely unlikely to be attributed to random perturbations in surge cyclicity, and instead suggest the influence of an external forcing mechanism. The number of surges during the recent decade seems high, but due to uncertainties in historical records, it remains unclear whether this frequency is exceptional or if earlier decades were unusually quiet. The cause of the observed threefold increase in surge activity also remains uncertain, given our incomplete understanding of surge initiation in relation to climate variability and non-climatic surge controls.
Crevasses play a crucial role in glacier-related hazards by facilitating water intrusion into the ice body and potentially triggering the collapse of large ice masses. However, the stress conditions governing their initiation and propagation remain uncertain. In particular, there is ongoing debate regarding the most relevant stress invariants to define fracture initiation (the failure criterion) and the corresponding failure strength, i.e. the stress threshold beyond which crevasses form. Laboratory estimates are hampered by the difficulty of reproducing natural glacier conditions, while in situ studies encounter uncertainties when converting strain or strain rate into stress estimates. This study investigates crevasse initiation processes by analyzing the artificial drainage of a water-filled cavity on Tête Rousse Glacier in 2010. Using the finite element code Elmer/Ice, we simulate the drainage and subsequent cavity refilling over three consecutive years. Given the well-constrained cavity geometry and water levels, stress fields are inferred directly from the force balance, removing the need to convert deformation data into stress estimates. Simulated stress distributions are compared with a pattern of circular crevasses mapped around the cavity after the first drainage event. Our results suggest that crevasses started forming in the delayed non-linear viscous regime governed by the Glen-Nye flow law, rather than as a result of the immediate elastic response to drainage. Additionally, by evaluating four failure criteria commonly used in glaciology, we show that the maximum principal stress criterion, with a stress threshold of 100 to 130 kPa, provides the best match to the observed crevasse field.
Seasonally frozen ground (SFG) is a critical component of the cryosphere, yet its freezing dynamics are often oversimplified in large-scale monitoring frameworks – particularly in remote sensing (RS) and land surface modeling – through the use of binary 0 °C thresholds. This approach overlooks the physically significant “transitional” state where liquid water and ice coexist, leading to systematic errors in quantifying the timing and duration of the frozen season. To address this, we recast the Soil Freezing Characteristic Curve (SFCC) framework directly into permittivity–temperature space. By operating in dielectric space, we bypass the high uncertainty associated with soil-specific liquid water content calibrations and enable a robust categorization of soil into unfrozen, transitional, and frozen states. We fitted this model to in-situ measurements from eight monitoring networks (87 sites) across Canadian boreal forest, prairie, and tundra ecozones. Using Bayesian hierarchical partial pooling, we derived stabilized estimates of the freezing onset (Tf) and transition sharpness (b). Network-level Tfb°C−1, reflecting distinct freezing regimes. We found that the transitional state is a dominant seasonal feature at these sites, challenging binary 0 °C assumptions used in RS evaluation. In high-moisture sites characterized by thick organic insulation (e.g., within the observed eastern boreal forest networks), this state persisted for over 100 d – effectively the entire winter – despite persistent subzero air temperatures. In contrast, sites in the western boreal and prairie networks, which generally lack thick surface organic layers and have lower soil moisture, exhibited shorter but still significant transitional periods (30 and 60 d, respectively). Even in the extreme cold of the tundra network sites, the transitional phase persisted for over 40 d. These results confirm that surface insulation and soil moisture, rather than air temperature alone, govern the SFG regime at the observed locations, providing a reproducible, physically-based reference framework for the next generation of freeze–thaw products.