The above examples do not constitute a complete list and we welcome all abstracts focusing on any physical, biophysical and biochemical permafrost-climate feedback, explicitly including submissions proposing novel mechanisms. We encourage contributions that aim at a quantification of the feedback strengths, at the regional to global scale, and those which improve our understanding of mechanisms, at the process-level. Likewise, we welcome abstracts targeting feedback mechanisms under scenario-based future projections as well during the historical period and in the deeper past. Contributions relying on modelling tools and observational data are equally welcome and so are submissions conceptually describing feedback-chains that have been overlooked by the scientific community.
Orals: Tue, 5 May, 08:30–10:10 | Room 1.34
Permafrost is experiencing rapid and widespread degradation in response to atmospheric warming, with profound implications for landscape stability, infrastructure planning, hazard assessment, and carbon-climate feedback. While probabilistic approaches to global permafrost modelling exist (Gruber, 2012), no framework currently spans the 1960–2020 period, a critical window encompassing the acceleration of anthropogenic climate change and the most pronounced observed warming in permafrost regions. Here, we present GRAPE-60 (Global ReAnalysis-driven Permafrost Extent estimate over a 60-year period), a high-resolution equilibrium-based permafrost model that integrates modern reanalysis datasets (ERA5 and JRA-55) with digital elevation models (ASTER and SRTM) to quantify global permafrost evolution from 1960 to 2020 and extend this framework to project future change under differing climate scenarios. Through systematic evaluation of multiple model configurations against borehole observations, we identify ERA5-ASTERDEM as the optimal combination, achieving an area under the receiver operating characteristic curve of 0.96 for historical periods. Independent comparison of the downscaled air temperatures product against ESA CCI satellite-derived land surface temperature (1996–2020) yielded a mean bias of +0.79°C (σ = 1.64°C), demonstrating agreement within the uncertainty bounds of the reference dataset. Our historical results reveal a net global permafrost loss of 2.49 × 10⁶ km² (8.9% of the exposed land area) between 1960 and 2020, with marked spatial patterns: continuous permafrost zones contracted by 21% while isolated patches expanded by 12%, indicating widespread degradation and fragmentation. To project future permafrost trajectories, we utilise GRAPE-60 as a spin-up and bias-correct temperature forcings from selected CMIP6 Earth System Models under SSP1-2.6, 2-4.5, 3-7.0, and 5-8.5 scenarios. Preliminary results reveal scenario-dependent pathways ranging from near-stabilisation under aggressive mitigation to substantial additional losses under high-emission futures. This work provides the first high-resolution, multi-decadal reconstruction of global permafrost evolution 1960-2020 and establishes a methodological framework for tracking and projecting future changes in this critical and rapidly transforming component of the cryosphere.
How to cite: McCourt, H. R., Westoby, M. J., Dunning, S. A., Buzzard, S., and Lim, M.: Global permafrost evolution 1960–2100: A new high-resolution model assessing past change and projecting scenario-based futures., EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-13457, https://doi.org/10.5194/egusphere-egu26-13457, 2026.
The Arctic is undergoing rapid warming, leading to significant changes in permafrost stability and ecosystem carbon exchange. Arctic tundra ecosystems store nearly half of the world's soil carbon, yet permafrost thaw may enhance carbon dioxide emissions, amplifying climate feedbacks. Despite their importance, large uncertainties remain regarding the role of high latitude permafrost regions in the Arctic carbon cycle. The Canadian Arctic covers extensive and heterogeneous landscapes, emphasizing the need for long-term site-based observations of net ecosystem exchange (NEE). In this study, we quantify and characterize carbon dioxide fluxes during the snow-free period using the eddy covariance method at a dry tundra site in Cambridge Bay, Nunavut, Canada. The analysis is based on nine years of eddy covariance observations collected between 2012 and 2022, excluding 2020 and 2021. Carbon dioxide exchange was examined during the snow-free period from mid-June to early September, when vegetation activity plays a dominant role in ecosystem carbon dynamics. The climate at Cambridge Bay is characterized by cold and dry conditions, with a 30-year mean annual temperature of approximately −13 °C and low annual precipitation (~145 mm), typical of a dry tundra ecosystem. The site has experienced a long-term warming trend of approximately +0.0415 °C per year over the period 1953–2022. Snow cover generally persists from September to May, and snowmelt typically begins in late May to early June, marking the onset of the snow-free and biologically active period. Results show that carbon dioxide fluxes during the snow-free period are strongly influenced by vegetation growth, with peak plant productivity and maximum carbon uptake occurring during the growing season between mid-July and early August. During this period, gross primary production exceeds ecosystem respiration, resulting in net carbon uptake. On average, the growing season NEE was −40.5 g C m⁻², with corresponding gross primary production (GPP) and ecosystem respiration (Reco) of 94.2 g C m⁻² and 53.6 g C m⁻², respectively. Notably, 2017 exhibited an unusually early onset of vegetation activity, leading to enhanced carbon uptake during the snow-free period. These flux estimates highlight the important role of the dry tundra ecosystem at Cambridge Bay in the pan-Arctic carbon dioxide budget. These findings provide robust observational evidence of snow-free period carbon dynamics in a high latitude Arctic dry tundra ecosystem and improve our understanding of how ongoing climate warming may influence carbon fluxes in permafrost regions. This work was supported by the National Research Foundation of Korea (NRF) grant funded by the Korea government (MSIT) (RS-2025-24683148)
How to cite: Hwang, H., Chae, N., Lee, B. Y., Jung, J. Y., and Choi, T.: Long-term Observations of CO₂ Fluxes during the Snow-free Period in a High Arctic Dry Tundra, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-15197, https://doi.org/10.5194/egusphere-egu26-15197, 2026.
Accelerating permafrost thaw may mobilize vast stores of deep and frozen soil carbon (>3 m), releasing CO2 into the atmosphere. Yet, the magnitude of this release remains uncertain due to the absent deep carbon processes in current Earth system models (ESMs). Here, we use an updated ORCHIDEE-MICT model that explicitly simulates Yedoma formation during the Pleistocene and the transient development of northern peatlands during the Holocene to project northern (>30°N) carbon responses under climate change. Incorporating these deep, frozen carbon pools improves agreement with carbon cycle observations and reduces previously projected net CO2 uptake by 47–74 Pg C between 1900 and 2100 across three future scenarios. Under high-emission pathways, the northern soil carbon balance shifts from a net sink to a net source of up to 32 Pg C, advancing the reversal predicted by the original model earlier in the 21st century. This earlier reversal is primarily driven by accelerated deepening of the active layer after mid-century, exposing more previously frozen carbon, particularly from Yedoma. Consistent with field data, our model shows that colder soils retain more labile carbon—contrary to assumptions in many IPCC models, which helps explain their prediction of a continuous carbon sink. Our results highlight the need to represent both the quantity and quality of permafrost carbon in ESMs to improve projections of permafrost–climate feedbacks.
How to cite: Xi, Y., Ciais, P., Zhu, D., Qiu, C., Zhang, Y., Peng, S., Bowring, S. P. K., Goll, D. S., Friedlingstein, P., and Hugelius, G.: Net release of CO2 from thawing permafrost soil carbon predicted to occur earlier in this century, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-4085, https://doi.org/10.5194/egusphere-egu26-4085, 2026.
Cryospheric retreat associated with deglaciation and permafrost thaw exposes previously stable carbon reservoirs to active biogeochemical cycling. While substantial effort has focused on constraining greenhouse gas emissions from thawing soil organic matter, many cryospheric environments also host large carbon stocks within bedrock and regolith, whose contribution to carbon-climate feedbacks remains poorly quantified. Mobilisation of these carbon stocks can enhance CO2 release to the atmosphere via an acceleration of sulphide mineral and rock-derived organic carbon oxidation. Multi-decadal increases in riverine sulphate concentrations across the Arctic, together with growing reports of “rusting rivers”, provide compelling evidence for a direct link between permafrost degradation and enhanced oxidative weathering. However, the magnitude, spatial distribution, and climatic significance of this feedback remain poorly constrained.
Here, we quantify the impact of cryospheric retreat on oxidative weathering through a field investigation spanning five valleys in Svalbard underlain by contrasting lithologies. Surface water geochemistry and discharge measurements collected from glacier termini to valley bottoms reveal that CO2 release from geological carbon stocks is sustained—and in some cases amplified—downstream of glacier margins. These observations indicate that enhanced oxidative weathering is not confined to subglacial environments but persists across regions subject to paraglacial processes. This improved spatial understanding of oxidative weathering is used to inform a forward model of carbon release from geological reservoirs resulting from permafrost thaw and and ice-sheet retreat, focused on the last deglaciation. Together, our results demonstrate that cryospheric retreat can enhance carbon release from geological reservoirs, representing a previously under-constrained positive feedback on the Earth system with implications for both past climate transitions and ongoing cryospheric retreat.
How to cite: Knight, A., Stokes, C., Stevens, L., Cosmidis, J., Wadham, J., Tipper, E., Wright, L., and Hilton, R.: Periglacial Regions as Hotspots of Oxidative Weathering that Drive Deglacial Acceleration of Rock Carbon Release, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-12991, https://doi.org/10.5194/egusphere-egu26-12991, 2026.
Acidification and metal mobilization linked to sulfide mineral oxidation pose an urgent risk to water quality and ecosystem health in thawing permafrost regions. Over the past six years, we have observed large increases in sulfate and metal concentrations and fluxes coupled with notable pH decreases in several headwater streams in the Tombstone Waters Observatory, Yukon, Canada. Field observations and satellite imagery reveal the emergence of acidic seepage zones characterized by extensive vegetation dieback and ochreous mineral precipitation. This seepage can exhibit pH < 3 and sulfate concentrations up to 5000 mg L-1, with metal (e.g., Fe, Al, Mn, Ni, Zn) concentrations ranging from 10s to 100s of mg L-1 and often exceeding water quality criteria. Subsequent pH buffering along groundwater discharge and stream mixing zones drives extensive precipitation of Al and Fe (oxyhydr)oxide and (hydroxy)sulfate phases, which influence metal transport and are visible many kilometers downstream. Notable increases in sulfate concentrations for major downstream rivers (e.g., Klondike, Ogilvie, Peel) show that coupled biogeochemical and hydrological processes in headwater catchments can have widespread impacts. Moreover, the growing occurrence of acidic seepage zones and ochreous mineral precipitates suggests impacts of permafrost thaw and sulfide mineral oxidation represent a substantial long-term risk to subarctic stream chemistry. Our ongoing research at the Tombstone Waters Observatory aims to advance understanding of the complex coupled processes influencing stream water chemistry in thawing permafrost regions.
How to cite: Lindsay, M., Skierszkan, E., Szeitz, A., and Carey, S.: Abrupt changes in subarctic stream chemistry linked to permafrost thaw and sulfide mineral oxidation, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-8385, https://doi.org/10.5194/egusphere-egu26-8385, 2026.
Permafrost regions store vast amounts of soil carbon, and warming-driven thaw can enhance microbial decomposition and carbon dioxide release, strengthening the permafrost carbon feedback. Although this feedback has been widely studied under continued warming, the processes governing its magnitude and persistence remain highly uncertain. Here we use the Community Earth System Model version 2 (CESM2) to examine permafrost carbon-cycle dynamics under both warming and mitigation, and apply multiple regression to quantify the controls on net biome production. We identify soil moisture as a key regulator of carbon cycle dynamics. Thaw-driven increases in liquid soil water enhance plant photosynthesis, but more strongly accelerate microbial decomposition, shifting ecosystems toward net carbon release and reducing net biome production. Sensitivity to soil moisture is strongly heterogeneous, with the largest response in central Siberia where high litter carbon coincides with relatively low climatological soil moisture, whereas North America shows weaker sensitivity under a wetter background state. Soil moisture also delays recovery after mitigation by slowing permafrost refreezing and sustaining anomalously wet soils during the net-zero period. This maintains elevated heterotrophic respiration and prolongs negative net biome production even after mitigation. In conclusion, our results show that soil moisture can amplify and sustain permafrost carbon losses along mitigation pathways, highlighting improved representation of coupled soil hydrology and permafrost processes as a priority for reducing uncertainty in future carbon budget assessments.
How to cite: Mun, J.-H., Lee, H., Lee, D., Shin, Y., and Kug, J.-S.: Impact of Soil Moisture on Carbon Cycle Changes in Permafrost Regions in CESM2, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-9260, https://doi.org/10.5194/egusphere-egu26-9260, 2026.
Under changing climatic conditions, the Arctic is undergoing massive changes. Due to Arctic amplification, the Arctic is warming faster than any other region on Earth. The combination of warming and CO2 fertilization leads to increases in productivity and changes in vegetation composition, with shrubs invading the Tundra, and trees also shifting northwards. At the same time, permafrost thaws, adding previously frozen carbon deposits to the active carbon cycle. The net carbon balance resulting from all of these coupled processes is less clear than one might think and requires an integrated modelling approach.
We use ICON-Land, the land surface model of the ICON Earth System Model, to investigate changes in the carbon cycle of the permafrost region. We have extended the soil carbon model YASSO by introducing a vertical dimension in order to consider carbon storages in deeper frozen soil layers. Furthermore, we are considering Arctic-specific shrub PFTs in our dynamic vegetation scheme in order to represent the changes in vegetation composition expected in a changing climate, thus allowing a complete assessment of carbon cycle changes.
We initialise the soil C pools for the preindustrial climate state from the Northern Circumpolar Soil Carbon Database to insure initial C pool sizes close to measurements. We then determine changes in vegetation composition and soil C storage in transient model experiments following historical and future climate changes under RCPs 2.6, 4.5, and 8.5. Based on these experiments, we quantify the greenhouse gas balance under future climatic conditions. While the permafrost soils lose carbon in all scenarios, productivity increases, especially if the vegetation can adapt to the changed climatic conditions, leading to lower carbon release.
How to cite: Kleinen, T., de Vrese, P., Bergstedt, H., and Brovkin, V.: Compound effects of permafrost thaw, vegetation dynamics, and increasing CO2 alter carbon budget of the permafrost zone, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-9850, https://doi.org/10.5194/egusphere-egu26-9850, 2026.
The permafrost region holds vast amounts of carbon which, upon thaw, may be released to the atmosphere as CO2 through enhanced decomposition. While links between soil carbon content and respiration have been shown by, for example, numerous incubation studies, it remains challenging to establish similar relationships from in-situ data collected in the field – especially at the large landscape-scale of eddy covariance towers. Part of the reason is the high heterogeneity of Arctic landscapes, combined with frequently shifting footprint distributions, and general lack of detailed soil carbon data that make it difficult to separate the signal from the noise. In this study, therefore, we combine detailed surveys of soil carbon content with high resolution footprint analyses to explore whether inter-site differences in carbon fluxes across ten sites spanning the Arctic can be explained by soil carbon content. Soil carbon data at each site was collected across dominant landforms and analyzed with depth. At 5 sites, these data were further developed into high resolution soil carbon maps. At the remaining sites, spatially weighted estimates of soil carbon content were determined proportionally to dominant landforms. Net CO2 fluxes collected by the towers were processed according to the same pipeline and partitioned into GPP and ecosystem respiration. These fluxes were related to the amount of soil carbon in the active layer at dominant landforms through a spatially explicit footprint analysis. Our preliminary results suggest that active layer soil carbon storage is a strong predictor of inter-site differences in ecosystem respiration during the growing season. This study will further explore the robustness of this relationship across the Arctic and throughout the year.
How to cite: Parmentier, F.-J. W., Belelli Marchesini, L., Wille, C., Hugelius, G., Siewert, M. B., Aurela, M., Boike, J., Christensen, T. R., Dolman, H., Friborg, T., Goeckede, M., Hensgens, G., Holl, D., van Huissteden, K., Kutzbach, L., Pirk, N., Sachs, T., Shurpali, N., Zhao, Y., and Kuhry, P.: Can soil carbon content explain inter-site differences in carbon flux magnitude across the permafrost region?, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-14668, https://doi.org/10.5194/egusphere-egu26-14668, 2026.
Permafrost degradation across the Northern Hemisphere is projected to continue and intensify under ongoing climate warming, with important implications for the global carbon cycle. Thawing permafrost exposes previously frozen soil organic carbon (SOC) to microbial decomposition, resulting in the release of carbon dioxide and methane, thus potentially amplifying climate change through positive feedbacks. Robust projections of permafrost thaw are therefore essential for improving estimates of future carbon emissions and the global carbon budget.
Using output from the latest generation of global climate models participating in CMIP6, we assess changes in the annual active layer thickness (ALT), defined as the maximum seasonal thaw depth of permafrost, under four Shared Socioeconomic Pathway (SSP) scenarios (SSP1-2.6, SSP2-4.5, SSP3-7.0, and SSP5-8.5). We show that ALT estimates derived directly from CMIP6 soil temperature fields exhibit substantial deviations from observed ALT values, which can lead to inconsistent estimates of permafrost carbon release.
To address this limitation, we propose a simplified, observation-constrained approach that focuses on projected changes in ALT rather than absolute model-derived values. These projected ALT changes, combined with present-day ALT observations, are used to estimate vulnerable carbon under future climate projections. We validate our ALT-based approach through comparison with simulations from the Lund–Potsdam–Jena managed Land (LPJmL) dynamic global vegetation model, which explicitly represents permafrost and soil thermal processes. This comparison shows consistent spatial patterns of active layer thickness, supporting the robustness of our simplified estimation framework.
CMIP6 models project ALT changes of 0.1–0.3 m per degree rise in local temperature, resulting in an average deepening of approximately 1.2–2.1 m in the northern high latitudes under different scenarios. With increasing temperatures, permafrost thawing initiates in southern Siberia, northern Canada, and Alaska, and progressively extends poleward, reaching much of the Arctic under high-emissions scenarios (SSP5-8.5) by the end of the century. Using projections of ALT changes and vertically resolved SOC data, we estimate ensemble-mean decomposable carbon stocks in thawed permafrost of approximately 115 GtC under SSP1-2.6, 180 GtC under SSP2-4.5, 260 GtC under SSP3-7.0, and 300 GtC under SSP5-8.5 by 2100.
How to cite: Nadeem, I., Nakicenovic, N., Yaqub, A., Sakschewski, B., Loriani, S., Bala, G., Tharammal, T., Zimm, C., and Aeman, H.: Estimating Vulnerable Carbon in Thawing Northern High-Latitude Permafrost Using CMIP6 Climate Projections, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-19398, https://doi.org/10.5194/egusphere-egu26-19398, 2026.
Within the Tipping Point Modeling Intercomparison Project (TIPMIP; www.tipmip.org), we outline a coordinated experimental protocol designed for standalone land surface models (LSMs) as well as coupled Earth System Models (ESMs). The protocol targets key questions on permafrost feedbacks, thresholds, timescales, abrupt versus gradual, and irreversible changes, and interactions with other tipping elements, aiming to quantify plausible landscape transformations, associated greenhouse gas emissions, and their impacts on global energy, carbon, and hydrological cycles, as well as large scale circulation.
Building on TIPMIP-ESM, the TIPMIP-permafrost Tier 1 experiments, defined as the core, priority experiments required for MIP participation, aim to use LSMs to quantify the extent to which permafrost area, ground ice volume, and soil carbon stocks exhibit path dependent behavior and reversibility when subjected to idealized warming, stabilization, and subsequent cooling phases. This includes evaluating whether the permafrost returns to its initial state as temperatures decline, or the extent to which certain changes, such as area loss, carbon redistribution, hydrological reorganization, or ground subsidence, persist despite cooling and thus remain effectively irreversible over time scales spanning a century to a several hundred years. In addition, during the stabilization phase we will examine whether permafrost degradation continues even after warming ceases. This involves assessing time lagged thaw response, identifying critical thresholds that may trigger rapid acceleration of degradation, and determining whether internal processes, such as shifts in soil moisture, ground subsidence and thermokarst lake formation, or carbon redistribution, amplify change through self reinforcing feedbacks. The emergence of such behaviors would indicate nonlinear system dynamics and a heightened susceptibility to tipping point transitions.
Additional sensitivity experiments (Tier 2), aimed at understanding and quantifying specific processes that could induce permafrost tipping, will apply idealized forcing. The focus is on altering hydrologic conditions and modifying surface properties (e.g., vegetation, albedo, thermokarst lake distribution) to explore how these factors influence the onset and dynamics of tipping behavior.
Tier 3 experiments, also sensitivity oriented, investigate the coupling between permafrost dynamics and other climate tipping elements using ESMs, for example, assessing the consequences of an AMOC collapse for high latitude permafrost stability, as well as the broader Earth system impacts of abrupt permafrost tipping.
How to cite: Georgievski, G., Brovkin, V., Burke, E., Nitzbon, J., Steinert, N., Tardif, D., Dennis, D., Loriani, S., Donges, J., and Winkelmann, R.: Permafrost Experimental Protocol within the Tipping Point Modelling Intercomparison Project (TIPMIP), EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-12875, https://doi.org/10.5194/egusphere-egu26-12875, 2026.
Permafrost is extensively distributed across the Qinghai-Tibet Plateau (QTP) in China and is interspersed with numerous streams, rivers, and lakes, forming a highly sensitive cryo-hydrogeological environment. Beneath these aquatic systems, permanently unfrozen zones, known as taliks, commonly develop within the permafrost and play a critical role in regulating hydraulic connectivity and water exchange between groundwater and surface water. As such, river-talik systems exert a strong influence on regional water resources, river discharge regimes, and the stability of engineering infrastructure. Nevertheless, the functioning and dynamics of river-talik systems in discontinuous permafrost regions remain complex and insufficiently understood. During the summer season, when taliks remain hydraulically connected to the riverbed, groundwater flow through taliks can sustain river baseflow, thereby influencing seasonal water availability and downstream water resources. In contrast, during winter, progressive refreezing of the riverbed disrupts this hydraulic connection, resulting in groundwater pressure accumulation within aquifers at talik constrictions. Elevated groundwater pressure may fracture the overlying ice cover, allowing supra-permafrost groundwater to discharge onto river floodplains and form extensive icings. These processes not only alter winter runoff pathways but may also inundate adjacent land and infrastructure corridors. In addition, ongoing permafrost degradation beneath riverbeds can induce differential ground settlement, which poses significant risks to bridge foundations, embankments, and other critical infrastructure in cold regions. To investigate the cryo-hydrogeological characteristics and engineering implications of a complex river-talik system in a discontinuous permafrost region of the QTP, a comprehensive field-based approach was employed, integrating continuous monitoring of ground temperature, surface temperature, and hydraulic head with electrical resistivity tomography (ERT) surveys. The presence of flowing river water enhances subsurface heat storage beneath the riverbed, substantially delaying riverbed freezing during autumn and early winter. Moreover, the formation, thickening, and lateral expansion of icings during winter provide an additional thermal buffer. Together, these processes maintain ground temperatures above 0 °C at depth throughout the year, resulting in the development of an approximately 2 m thick talik beneath the riverbed. ERT inversion results further identify both supra-permafrost taliks and taliks penetrating through the permafrost. The river talik exhibits a complex three-dimensional geometry, with cross-sectional dimensions varying along the channel and extending locally above or intermittently through the permafrost. Furthermore, the warm permafrost underlying the talik retains partial permeability, allowing limited groundwater flow within the permafrost matrix. Collectively, these characteristics give rise to intricate local-scale cryo-hydrogeological processes that strongly affect groundwater-surface water interactions, winter river dynamics, and the long-term stability of infrastructure in permafrost regions.
How to cite: Liu, W.: River Talik Development and Its Implications for Cryo-Hydrogeological Processes and Infrastructure in Permafrost Regions on the Qinghai-Tibet Plateau, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-4328, https://doi.org/10.5194/egusphere-egu26-4328, 2026.
Thermokarst lakes are widespread and dynamic features of ice-rich permafrost landscapes. They accelerate permafrost thaw, enhance methane production, and alter soil hydrology as well as energy and water exchanges between land and atmosphere. These effects can alter local and global climate. But despite their important role in the climate sysem, thermokarst lakes are largely absent from Earth system models (ESMs), because deterministic and physics-based modelling approaches require extensive high-resolution ground-ice data that are not available. To close this gap, we develop a probabilistic modeling framework that represents lake dynamics as stochastic processes and can be parameterized using remote sensing data. The model has the potential to provide evolving Arctic water-area fractions and lake-size distributions that can be coupled to ESMs, improving the representation of permafrost dynamics and high-latitude carbon emissions under climate change.
How to cite: Reinken, C., Brovkin, V., de Vrese, P., Nitze, I., Bergstedt, H., and Grosse, G.: Stochastic Modelling of Thermokarst Lakes, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-10212, https://doi.org/10.5194/egusphere-egu26-10212, 2026.
Mean annual permafrost table temperature (MAPT) and active-layer thickness (ALT) are key variables for assessing the thermal regimes of permafrost and active layer and their responses to climate changes. Models for estimating MAPT and ALT have typically been forced by air or ground temperatures and some ground physical properties. While temperature measurements are relatively widely available and reliable, ground physical properties are frequently unavailable or unrepresentative and therefore need to be estimated, which introduces uncertainties into model outputs.
Hence, we devised two simple analytical−statistical models (ASMs) for estimating MAPT and ALT, which are driven solely by pairwise combinations of freezing and thawing indices in the active layer. We tested the models at the total of 55 sites in the Earth's major permafrost regions comprising five different ground surface covers and four permafrost zones where the models showed the total mean errors of less than 0.05 °C for MAPT and 9 % for ALT. Besides, the models can also be used to establish typical values of some ground physical parameters for MAPT and ALT estimates. We believe that ASMs can find useful applications in permafrost and active-layer modelling under a wide range of climates, ground surface covers, and ground physical conditions.
Acknowledgement
The research was funded by the Czech Science Foundation (project numbers GM22-28659M and GA25-18272S) and by the Ministry of Education, Youth and Sports (project number LL2505).
How to cite: Uxa, T., Hrbáček, F., and Kňažková, M.: Analytical–statistical models (ASMs) for mean annual permafrost table temperature and active-layer thickness, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-10730, https://doi.org/10.5194/egusphere-egu26-10730, 2026.
In the High Arctic, new channel networks are developing relatively rapidly within 10s of years, typically attributed to exteneded thaw seasons under climate change and late season permafrost loss. During the 2024 field season on Tallurutit (Devon Island), we visited a developing channel network where relatively flat pools are connected by steeper channelised segments which exhibit evidence of recent gravel and sand transport. However, despite visiting during a storm event, we did not observe active sediment transport and only limited surface-water runoff. This brings up the question: How does the hydrologic connectivity between channel segments and surface runoff fraction change across a thaw season?
We developed a new box model that couples surface and subsurface water flow between pools conserving enthalpy to explore the hydrologic response of an analogue landscape segment across a thaw season. Using this model, we first use synthetic rainevents to identify the key factors modulating the hydrologic response and then use new weather data from Tallurutit and historic climate data to explore when surface water erosion is likely to occur under typical climate conditions and for future climate scenarios.
Subjecting the landscape to the same magnitude-duration rainevent in the early versus late thaw season shows that surface water runoff fraction is greatest early in the thaw season as the shallow thaw front limits subsurface water storage. This results in successive overspill events that rapidly transport water across pools and promote erosion. In the late thaw season, subsurface water storage dominates and pools are connected via subsurface water flow. As a result surface water flow and resultant erosion is minimised. We summarise a key implication of our work through a reservoir response time that depends on the bed permeability, pool length, and thaw depth which modulates whether a pool will dampen a hydrologic signal or transmit it downstream.
Combined with grain-scale laboratory experiments which show that bed erodibility is inversely proportional to thaw depth, this work suggests that the formation and evolution of channel networks in the High Arcitc is primarily driven by early season discharge events. Under climate change, the frequency and magnitude of early season rainstorms and heatwaves, resulting in rapid snowmelt, are becoming more common. This suggests that the observed rapid channelisation in the High Arctic is related to early season climate extremes, instead of long-term warming averages and late season permafrost loss.
How to cite: Eschenfelder, J., Chartrand, S. M., and Jellinek, A. M.: Seasonal variation of hydrologic connectivity for High Arctic stream segments, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-5967, https://doi.org/10.5194/egusphere-egu26-5967, 2026.
The melting of segregated ice in ice-rich frozen soils is a primary driver of soil structural instability and thaw settlement, yet the dynamic morphological evolution of segregated ice and its quantitative linkage to thaw settlement remain poorly understood. To address this issue, a series of physical model experiments were conducted using a self-developed visualized thaw settlement experimental platform for ice-rich frozen soils. Thus, the different surface thawing temperatures (5 °C, 7 °C, and 10 °C) , multiple particle-size gradation schemes (G-I to G-V), and different external load levels (0.5 kPa, 1.0 kPa, and 1.5 kPa) were introduced to investigate the influence of temperature, gradation, and pressure on segregated ice melting and thaw settlement behavior. The results indicate that the melting process of segregated ice can be divided into a rapid phase-change stage and a stable thaw settlement stage. An increase in surface thawing temperature significantly shortens the duration of the phase-change stage and enhances downward migration of liquid water toward the unfrozen zone, resulting in a pronounced increase in thaw settlement. Particle-size gradation regulates moisture accumulation and migration by modifying pore structure and capillary force intensity; segregated ice is more readily formed in fine-grained soils, which exhibit a higher risk of thaw settlement after melting. External loading has a limited influence on the phase-change pattern; however, by increasing pore water pressure, it significantly intensifies settlement deformation during the stable stage. Pore structure collapse induced by segregated ice melting dominates the rapid settlement stage, whereas pore water drainage and soil skeleton reorganization govern deformation during the stable thaw settlement stage.
How to cite: Wang, Z., Huo, Z., and Zhang, C.: Mechanism of thaw settlement induced by segregated ice melting based on a visualizing study, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-12648, https://doi.org/10.5194/egusphere-egu26-12648, 2026.
Saline permafrost exists beneath shallow shelf seas, coastal plains shaped by past marine transgressions, and post-glacially uplifted landscapes that were once submerged. Salinity influences the freezing point and mechanical strength of permafrost; it is, therefore, a critical parameter for assessing its stability. On Svalbard, the Kvadehuksletta region northwest of Ny-Ålesund features a diverse landscape comprising raised beach terraces, lagoons, paleo-lagoons (now lakes), and surface seeps. Our research aims to decipher how marine sediments transform after emergence. We hypothesize that ice formation during permafrost aggradation produces a porewater salinity gradient that triggers the downwards migration of salt in slowly uplifting sediments that are weakly susceptible to groundwater flushing. Sufficient salt build-up may lead to the formation of cryopegs. Cryopegs, a type of talik, are unfrozen layers or pockets within permafrost that persist at subzero temperatures due to their elevated salt content. In summer 2024 and 2025, we carried out several electrical resistivity tomography (ERT) profiles, including three profiles (ranging from 800 to 2300 m in length) perpendicular to the coastline. The westernmost profile (collected in 2025) intersected a dynamic lagoon that was connected to the sea in 2024 but became completely cut off in 2025 by storm-surge deposits. To help delineate frozen and unfrozen permafrost conditions, electrical resistivity-temperature analyses of field samples collected from shallow cores (down to 2.5 m) are currently underway. Laboratory tests indicate that the near-surface marine clays adjacent to the lagoon have low resistivities (< 10 Ωm) when thawed and freezing point temperatures down to -1.6 °C. The field samples are also being analysed for porewater chemistry (electrical conductivity, cations & anions, pH, stable water isotopes) and basic sedimentological properties like grain size. At two coring sites (1 paleo-lagoon, 1 beach setting), an annual ground temperature time series was also collected between field seasons. While the physical and electrical properties of the marine sediments are important to establish, so is their thickness. To potentially provide additional information on the depth to bedrock along selected ERT profile segments, we conducted multiple seismic refraction tomography (SRT) surveys (115 m length) in 2025 using a sledgehammer as an energy source. The synthesis of all datasets to describe uplifted permafrost is work in progress, but preliminary conclusions suggest that cryopeg occurrence is most likely in low-lying coastal areas characterized by warm permafrost, occasional seawater submergence, and saline marine clays with low hydraulic conductivity.
How to cite: Angelopoulos, M., Boie, K., Rau, M., Offer, M., Eppinger, S., Hauber, E., Zanetti, M., Sassenroth, C., Johnsson, A., Hiesinger, H., Schmedemann, N., Overduin, P. P., Boike, J., Westermann, S., Hallet, B., and Krautblatter, M.: Post-glacial emergent permafrost processes within coastal and paleo-lagoon settings on Svalbard, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-13227, https://doi.org/10.5194/egusphere-egu26-13227, 2026.
Land surface models (LSMs) still exhibit widespread deficiencies in frozen soil regions, particularly in overestimating soil temperature responses to air temperature under snow cover and in simulating soil moisture dynamics (see, e.g., Luo et al (2025)). Analysis of CMIP6 simulations reveals that, in frozen soil areas, the influence of LSMs on coupled climate model results is comparable in magnitude to that of atmospheric forcing. Moreover, compensating effects between land and atmosphere components often lead to apparently better performance in coupled simulations than in offline LSM experiments. This compensation poses a risk that structural deficiencies in LSMs may remain obscured when evaluating coupled model performance.
To identify specific weaknesses in current LSM formulations, we conduct offline simulations using three land surface models—CLM5, TERRA standalone, and JSBACH—at permafrost observation sites, including Bayelva (Svalbard) and Samoylov (Lena River Delta). The models are driven by meteorological forcing derived from the WATCH Forcing Data methodology applied to ERA5 (WFDE5), which is further bias-corrected to the sites using in situ observations.
We evaluate simulated snow water equivalent, soil temperature, soil moisture, and surface energy fluxes against observations, complemented by targeted sensitivity experiments. This approach aims to diagnose the key processes responsible for model biases in permafrost regions and to assess potential pathways for improving land surface model performance under cold-region conditions.
Zhicheng Luo, Risto, D., B. Ahrens (2025) Assessing Climate Modeling Uncertainties in the Siberian Frozen Soil Regions by Contrasting CMIP6 and LS3MIP. The Cryosphere, 19, 6547–6576. https://doi.org/10.5194/tc-19-6547-2025
How to cite: Ahrens, B., Luo, Z., Parmar, M., and Risto, D.: Improved Process Understanding Using Stand-alone Land Surface Models in Simulating Permafrost, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-13913, https://doi.org/10.5194/egusphere-egu26-13913, 2026.
Permafrost evolution and subsurface thermal dynamics play a key role in the climate system, yet their representation in Earth System Models (ESMs) remains constrained by shallow soil configurations. This study investigates millennial-scale land–climate interactions using simulations with the Land Surface Model CESM2/CLM5 employing a modified deep lower boundary extending to 500 m, alongside the standard 43 m configuration. Simulations span 500–2014 CE and are driven by boundary conditions obtained from two NorESM1-F PARCIM experiments over the same time interval. These experiments follow PMIP last-millennium protocols and include prescribed variations in solar irradiance, volcanic forcing, and greenhouse gas concentrations, while anthropogenic aerosol and land cover are held fixed at pre-industrial conditions. Two solar forcing reconstructions are used, representing low and high solar variability, enabling assessment of how land model depth and natural external forcing shape subsurface thermal states, permafrost extent, active-layer thickness, and soil carbon evolution on centennial to millennial timescales. Simulations using the deep land configuration exhibit reduced variability in the simulated permafrost area during the pre-industrial period (500–1850 CE) relative to the standard shallow configuration. Evaluation against observational and reanalysis-based datasets over 1997–2021 indicates closer agreement for the high solar variability forcing than for the low variability forcing with respect to permafrost extent and subsurface thermal conditions. High-latitude vegetation responses are further explored using modified CLM-FATES configurations. Finally, selected CAM–CLM coupled simulations are used to assess land–atmosphere feedback and to examine how differences between deep and shallow land states propagate into future projections under SSP scenarios from 2015 to 2100.
How to cite: Heidari, M., Beltrami, H., Pausata, F. S. R., Counillon, F., and Rouco, J. F. G.: Millennial-Scale Land–Climate Interactions Using a Deepened CLM5 Lower Boundary: Permafrost Evolution, Vegetation Dynamics, and Coupled Feedback in CESM2, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-14309, https://doi.org/10.5194/egusphere-egu26-14309, 2026.
Rising temperatures in the northern permafrost zone are profoundly altering key surface and subsurface processes, triggering important climate feedbacks. The most prominent of these is the accelerated decomposition of formerly frozen soil organic matter, leading to the release of carbon into the atmosphere. In addition, changes in surface and soil hydrology may influence regional and global climate through cloud radiative effects. The thawing of previously impervious soil layers increase hydrologic connectivity, thereby enhancing drainage and increasing landscape drying. This process limits evapotranspiration during the spring and summer months, reducing the formation of low-altitude clouds throughout the snow-free season. The decrease in summertime cloud cover, in turn, allows for greater shortwave absorption at the surface, amplifying temperatures and further accelerating permafrost degradation.
In this study, we investigate recent trends in cloud cover and soil moisture to identify potential signals of a permafrost-cloud feedback in observational and reanalysis datasets. While substantial disagreement exists between data sources regarding the signal strength and spatial patterns, we observe a consistent increase in the correlation between summertime cloud cover and soil moisture over recent decades, coupled with widespread drying in formerly permafrost-affected regions—supporting the hypothesis of a permafrost-cloud feedback. Additionally, we use multiple Earth-system model simulations to quantify the temperature contribution of this feedback under a high-warming scenario. Our results show a robust global temperature increase across all model setups, driven by reduced water availability due to permafrost degradation.
How to cite: de Vrese, P., Stacke, T., Gayler, V., and Brovkin, V.: Permafrost cloud feedback, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-15161, https://doi.org/10.5194/egusphere-egu26-15161, 2026.
The Arctic permafrost is a large carbon pool that is highly sensitive to climate change. Carbon fluxes were examined to understand the characteristics of the carbon cycle in tundra ecosystems adjacent to a pond. Freshwater lakes (large-scale) and ponds (small-scale) within the Arctic terrestrial ecosystem play important biogeochemical roles and, in some cases, are major sources of greenhouse gas emissions. Globally, lakes are primarily distributed in Canada and Russia, accounting for 42% and 49% of the total, respectively. The study site is located on dry tundra with a pond in the High Arctic near Cambridge Bay, Nunavut, Canada (69°7′47.7″N, 105°3′35.3″W). CO₂ and CH₄ fluxes were measured in the tundra ecosystem to evaluate the potential future sensitivity of the carbon cycle to climate change during the summers of 2019, 2022, and 2025. The vegetation cover around the site mainly consists of dwarf shrubs, graminoids, and lichens. Carbon fluxes were compared under different soil water content conditions. Based on the chamber measurements, the variability of net CO₂ exchange was more sensitive in grasses under wet conditions than in vegetation under dry conditions, and both the variability and magnitude of CH₄ emissions near ponds were greater than those under dry conditions. CO₂ and CH₄ fluxes showed a positive relationship in nearly bare soil under wet conditions and a negative relationship in areas with various vegetation under dry conditions. Net ecosystem exchange, ecosystem respiration, and gross primary production were estimated using two types of chambers to investigate the influence of carbon dynamics on the tundra carbon cycle. Carbon fluxes were compared across three years during the snow-free season, and the controlling factors of the carbon cycle were examined. This work was supported by the National Research Foundation of Korea (NRF) grant funded by the Korea government (MSIT) (RS-2025-24683148).
How to cite: Chae, N., Hwang, H., Choi, T., and Lee, B. Y.: Characteristics of Carbon Fluxes for Tundra Ecosystems Adjacent to a Pond in Arctic Canada, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-15489, https://doi.org/10.5194/egusphere-egu26-15489, 2026.
Thermal properties of permafrost-affected soils play a key role in determining their response to ongoing climate warming. These properties influence the rate of active layer thickening and govern whether permafrost degradation is amplified or inhibited. Soil thermal characteristics are closely linked to other physical soil factors, with moisture often considered the most critical due to its potentially high interannual variability. In the eastern Antarctic Peninsula, projections indicate precipitation changes between −5 and +10%. However, it remains unclear whether these changes will have a noticeable effect on soil moisture, as the area is generally classified as polar-arid, with very low effective precipitation throughout the year. We therefore hypothesize that drying mechanisms will prevail, namely (a) increasing surface evapotranspiration driven by surface warming and (b) enhanced infiltration into the ground due to thickening of the active layer.
In this study, we present potential trajectories of active layer thickness (ALT) evolution on James Ross Island, Antarctic Peninsula, for the period 2010–2050 and different soil moisture scenarios. ALT was predicted using the Stefan model parameterized by: (a) climate outputs from the WRF model forced by SSP2-4.5 at a 2 km spatial resolution and (b) laboratory analysis of the relationship between soil thermal conductivity and soil moisture. Experiments were conducted using an ISOMET 2114 soil thermal properties analyzer on samples collected from three sites. Soil thermal properties were measured on samples with a fixed bulk density across several moisture states, ranging from fully saturated to completely dry.
The Stefan model was run for four soil moisture scenarios (5%, 15%, 25%, and 35%). Predicted ALT under the driest scenarios (5% and 15%) was approximately 40 cm deeper than under the wettest scenario (35%). Overall, the results indicate an increase in ALT under SSP2-4.5 conditions by 2050.
How to cite: Hrbáček, F., Kňažková, M., Kaplan Pastíriková, L., Matějka, M., Kohoutková, K., and Uxa, T.: Soil moisture as a key control on active layer thickness prediction on James Ross Island, Antarctica, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-17019, https://doi.org/10.5194/egusphere-egu26-17019, 2026.
Physically based frozen soil models are essential for understanding hydrological processes in cold regions, particularly snowmelt infiltration into seasonally frozen soils and permafrost thaw. While significant progress has been made in modeling coupled heat and mass transport in frozen soils, and several sophisticated physically based models exist, practical applications require robust coupling between snow and soil models. Although we have well-developed physically based snow models, state-of-the-art soil models are typically not integrated with them. Notable exceptions include the Cold Regions Hydrological Model (CRHM) and various land surface models such as CLASS, CLASSIC, SVS, and SUMMA. However, these models are often applied with coarse vertical resolution and, in some cases, rely on oversimplified process representations. The objective of this study is to develop a simple, point scale mass- and energy-conservative coupled snow-soil model that can be used to systematically evaluate the numerous implicit and explicit assumptions embedded in existing models. A particular focus is to evaluate various approaches for representing the upper boundary condition of the soil, which plays a critical role in governing heat and mass fluxes and, consequently, the thermal and hydrological behavior of the soil.
How to cite: Ireson, A., Muenchrath, A., and Spence, C.: Energy conservative solutions for coupled heat-mass transport in frozen soils and snow, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-22122, https://doi.org/10.5194/egusphere-egu26-22122, 2026.
During the Last Glacial Maximum (LGM, 21000 years ago), atmospheric CO₂ concentrations were approximately 100 ppm lower than during the pre-industrial period, yet the mechanisms responsible for this difference remain poorly understood. Several hypotheses have been proposed, involving ocean circulation, the marine biosphere, or continental carbon stocks, but none has been able to fully explain this difference. Carbon stored in permafrost soils may represent a significant and still poorly constrained component of this missing carbon reservoir.
This study investigates the role of permafrost as a long-term carbon reservoir from the LGM to the present using a modelling approach. A simple soil thermal model (FROG) is used to simulate permafrost extent, depth, active layer thickness and soil carbon content down to 1000 meters depth. The model is coupled to iLOVECLIM, an intermediate-complexity Earth System Model coupling atmosphere, ocean, and biosphere components. This approach allows transient simulations to be performed over several thousand years, which is necessary to better represent carbon cycle evolution during glacial–interglacial cycles.
Simulations are conducted for present-day and LGM climate conditions. The model produces spatially explicit estimates of permafrost extent, active layer thickness, and soil carbon distribution for both periods. Results indicate substantial differences in permafrost depth and spatial coverage between the LGM and present day. Model results are compared with available observational datasets for permafrost and soil carbon, as well as outputs from other climate models.
How to cite: Voisin, M., Quiquet, A., Bouttes, N., and Roche, D.: Modelling the carbon cycle in permafrost in a simplified thermal soil model coupled with iLOVECLIM , EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-18508, https://doi.org/10.5194/egusphere-egu26-18508, 2026.
