Rapid increase in West Siberia’s retrogressive thaw slumps since 1964 associated with Arctic winter warming

Rapid increase in West Siberia’s retrogressive thaw slumps since 1964 associated with Arctic winter warming | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Article Rapid increase in West Siberia’s retrogressive thaw slumps since 1964 associated with Arctic winter warming Nina Nesterova, Marina Leibman, Carl Stadie, Tobias Hölzer, Ingmar Nitze, and 5 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-7697239/v1 This work is licensed under a CC BY 4.0 License Status: Under Review Version 1 posted 10 You are reading this latest preprint version Abstract Retrogressive thaw slumps (RTSs) are key indicators of permafrost thaw in the West Siberian Arctic. Based on a new high-resolution remote sensing-derived, field-verified inventory of 6168 RTSs for the Yamal and Gydan peninsulas, we provide the first large-scale spatio-temporal and climate sensitivity analysis since 1964. In Gydan, RTS clusters are located at higher elevations, on rougher terrain, and in lake-rich areas, whereas such associations are weaker in Yamal. Temporal analysis of RTS using historical and modern satellite imagery for key sites (~ 6,103 km²) indicates that RTS numbers increased 23-fold since 1964, and initiation rates raised 26-fold. A discrete-time Bayesian hazards model identified summer maximum precipitation as the strongest short-term factor. However, winter warming was the dominant long-term driver, with an increase of 1.62 σ resulting in a 37-pp increase in annual initiation probability. RTS development in the region poses high risks to gas and transport infrastructure. Earth and environmental sciences/Climate sciences Earth and environmental sciences/Environmental sciences Earth and environmental sciences/Natural hazards Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Introduction In recent years, evidence is mounting that Arctic winter warming is profoundly changing the terrestrial Arctic, including permafrost warming (Smith et al., 2022), widespread talik formation (Farquharson et al., 2022 ), and accumulating ecological impacts of episodic but strong winter warming events (Pascual & Johanssen 2022; Bradley et al., 2025 ). Permafrost, defined as ground that remains below 0°C for at least two consecutive years (Harris et al., 1988), underlies about 15% of the Northern Hemisphere landmass (Obu et al., 2021). Arctic permafrost experiences thawing and a reduction in extent in response to climate warming (Biskaborn et al., 2019 ; Smith et al., 2022). Permafrost not only affects high northern latitude landscapes and ecosystems but also stores large amounts of carbon and nutrients, with carbon loss upon thaw contributing to global climate feedback (Schuur et al., 2015 ). What drives permafrost thaw and over which time scales is therefore important to understand. Thaw of ice-rich permafrost often occurs abruptly, leading to the formation of distinctive landforms (Webb et al., 2025 ). One of the most widespread and dynamic indicators of abrupt thaw is a retrogressive thaw slump (RTS) (Nesterova et al., 2024 ). An RTS is a slope failure initiated by the exposure and thaw of massive ground ice (Mackay, 1966 ; Nesterova et al., 2024 ) (Fig. 1 a). RTSs are reported throughout permafrost regions in the Arctic and high-mountain permafrost regions, where they significantly alter topography, hydrology, soil, and vegetation, while delivering large fluxes of sediments, carbon, and nutrients to downstream environments (Kokelj and Lewkowicz, 1999 ; Kokelj et al., 2005 ; Mesquita et al., 2010 ). These landforms evolve in a polycyclic fashion through phases of active ice ablation and mudflow, followed by stabilization and vegetation colonization and further possible reactivation (Mackay, 1966 ; Leibman and Kizyakov, 2007 ; Lantuit & Pollard, 2008 ). The initiation of RTSs was reported to be linked to climate drivers such as rising summer air temperatures and increased summer precipitation (Lantz and Kokelj, 2008; Kokelj et al., 2015 ; Leibman et al., 2021 ). However, predicting RTS initiation remains complicated due to the uncertainty of heterogeneous ground ice distribution (Pollard and French, 1980 ; Makopoulou et al., 2024 ), limited field data (Ward Jones et al., 2019 ), and the lack of suitable models (O’Neill et al., 2016). The West Siberian Arctic is mostly characterized by continuous permafrost (Obu et al., 2019 ; Fig. 1 b) and massive ground ice occurring close to the surface (Baulin et al., 1967 ; Streletskaya et al., 2013 ; Badu, 2015 ). These two factors define the noteworthy abundance of RTSs in this region (Leibman and Kizyakov, 2007 ; Khomutov et al., 2017 ). The observed regional warming and deepening of the active layer thickness catalyze further RTS initiation and expansion (Babkina et al., 2019 ; Biskaborn et al., 2019 ; Vasiliev et al., 2020 ). Yet, a lot of studies of this region have relied on local field sites (Leibman and Kizyakov, 2007 ; Khomutov et al., 2017 ; Novikova et al., 2018 ; Khomutov et al., 2017 ; Babkina et al., 2019 ) or have been parts of large-scale automatic mapping with clear limitations and uncertainties due to a lack of field verification (Nitze et al., 2018 ; Bernhard et al., 2022 a; Nitze et al., 2025 , Yang et al, 2023 ; Maier et al., 2025 ). A recent high-resolution field-verified, manually-mapped, and classified RTS inventory dataset for the West Siberian Arctic now, for the first time, enables a large-scale robust analysis without compromising accuracy (Nesterova et al., 2025 ). Here we present the first large-scale analysis of spatial and temporal patterns of RTS distribution in the West Siberian Arctic, particularly for the Yamal and Gydan peninsulas (Fig. 1 ), and an analysis of factors and drivers contributing to their accelerating development. Our study aims to answer the following research questions: (1) What are the main spatial patterns of RTS distribution in the West Siberian Arctic? (2) What are the temporal trends of RTS dynamics in the West Siberian Arctic? (3) What environmental factors are associated with RTS development trends? and (4) What are possible areas of future RTS development? What are the current and potential risks for the infrastructure? To answer the first research question, we analyzed RTS spatial clustering and its relation to elevation, water bodies, and landcover. To study temporal RTS dynamics from the 1960s, we analyzed historical as well as modern high-resolution satellite imagery for five key sites (Fig. 1 ). To estimate climate influence on RTS dynamics, we compared RTS temporal trends with different climatic variables. To assess the risks of RTS development, we analyzed the current and possibly impacted infrastructure. Results and discussion Spatial distribution To explore the spatial distribution of RTS clusters in the West Siberian Arctic, we aggregated RTS point locations from the inventory into hexagonal grid cells (Uber H3 grid, resolution 6; ~35 km²; Uber Technologies Inc., 2018). All further calculations and statistics were performed per hexagonal grid cell. Previously reported spatial clustering of RTSs in the region (Nesterova et al., 2025 ) was confirmed by our results, with multiple statistically significant clusters revealed using Anselin Local Moran’s I (p ⋜ 0.005) (black hatched areas in Fig. 2 a). These clusters represent 21,179 km², accounting for 6.9% of the total study area for both peninsulas (7,160 km² − 5.4% for Yamal; 14,019 km² − 8% for Gydan). This is consistent with a general tendency of spatial clustering of RTSs across the Arctic (Nitze et al., 2018 ; Lewkowicz, 2024 ). After initial pre-processing and data cleaning (Extended Methods 1.1 and 1.2 in Supplementary Information 2 (SI 2- 1.1 and 1.2)), the analysis of environmental characteristics of RTS clusters revealed differences between the Yamal and Gydan peninsulas. Landcover analysis of RTS clusters in both peninsulas revealed a dominant association with the moist tundra class, with slight differences between peninsulas (Fig. 2 a and b). RTS classes in Yamal are associated with moist tundra , abundant moss, prostrate to low shrubs/dwarf and low shrubs/dense dwarf and low shrubs (class IDs 10, 11, 12; IDs here and subsequently refer to the original landcover class ID developed by Bartsch et al., 2024 ), moist to wet tundra , dense dwarf and low shrubs (13) and mostly by moist tundra , low shrubs (14) landcover classes, altogether representing ~ 60% of landcover class proportion of RTS clusters in Yamal. Meanwhile, the class of dry tundra , characterized by abundant lichen and prostrate shrubs (7), which dominates in the north of Yamal, is the least represented (Fig. 2 a and b). The landcover class most commonly associated with RTS clusters in Gydan is moist tundra with abundant moss and prostrate to low shrubs (10), with low shrubs (14), as well as dwarf and low shrubs (11), and moist to wet tundra with dense dwarf and low shrubs (13). Also, the water (1) landcover class was among the dominant for the Gydan RTS clusters. Altogether, these classes represent ~ 70% of the landcover class proportion of RTS clusters in Gydan. The most underrepresented land cover class compared to the rest of the Gydan Peninsula is also moist tundra , but with abundant forbs and dwarf to tall shrubs (16). An elevation analysis of RTS clusters using ArcticDEM mosaics (Porter et al., 2023 ) also revealed some differences between the two peninsulas. All elevation parameter distributions were non-normally distributed, thus non-parametric tests were applied. The distribution of mean elevations (95th percentile) of RTS clusters in Yamal showed a statistically significant difference compared to the mean elevations (95th percentile) of the rest (without RTS clusters) of the Yamal peninsula based on the Mann-Whitney U-Test (U = 341,588, p < 0.005 after Bonferroni correction), as well as based on Kolmogorov-Smirnov Test (D = 0.198, p < 0.001 after Bonferroni correction). This reflects the differences, which are depicted in the Empirical Cumulative Density Function (CDF) in Fig. 2 c. The 95th percentile of mean elevation values for RTS clusters in Yamal ranged from 4.79 to 26.75 m a.s.l. The distribution of mean elevations of RTS clusters in Gydan demonstrated a strong statistically significant difference compared to the rest (without RTS clusters) of the Gydan peninsula based on both tests (U = 809,112, p < 0.001 after Bonferroni correction; D = 0.167, p < 0.001 after Bonferroni correction). RTS clusters in Gydan were significantly higher than the rest of the peninsula (median + 3m). The empirical CDF of Gydan RTS clusters compared to the rest of the peninsula is depicted in Fig. 2 d. The 95th percentile of mean elevation values for RTS clusters in Gydan ranged from 6.91 to 42.2 m a.s.l. The distribution of values of the vertical ruggedness measure (VRM, a characteristic of terrain roughness derived from the ArcticDEM mosaics) for RTS clusters did not show any statistical difference in Yamal (both tests); however, it showed a strong statistically significant difference compared to the rest of the peninsula in Gydan: U = 945,361, p < 0.001 after Bonferroni correction; D = 0.199, p < 0.001 after Bonferroni correction. RTS clusters in Gydan were rougher and characterized by higher VRM values (empirical CDF depicted in Fig. 2 e). The 95th percentile of VRM values for RTS clusters in Gydan ranged from 0.95e-5 to 3.1e-5. In general, these values represent rather smooth terrain. An analysis of the lake density revealed no significant difference in lake density between RTS clusters and the rest of the peninsula in Yamal (both tests), but a strongly significant difference in Gydan (U = 1,036,752, p < 0.001 after Bonferroni correction; D = 0.144, p < 0.001 after Bonferroni correction). RTS clusters in Gydan occur in significantly higher lake densities. This finding is shown in the empirical CDF (Fig. 2 f) and is supported by the overrepresentation of the water landcover class (Fig. 2 b). The 95th percentile values of lake density for clusters occur between 0.002 and 0.373. Our findings of RTS clusters tending to occur on higher elevations and rougher terrain are generally consistent with the literature review findings of inland RTSs occurring at the edges of terrains: river banks, lake shores, slopes of streams, and valleys (Nesterova et al., 2024 ). Moreover, most of the RTS clusters in Gydan are located in the hilly ridge area of Gydanskaya Gryada. The generally flat terrain found in the north of Yamal has almost a full absence of RTSs (Fig. 1 ). This might be due to the prevalence of very low elevations (< 20 m a.s.l.., see SI 1–2), low VRM (1.65e-5, see SI 1–2), and the dominance of the Dry Tundra landcover class. The latter is characterized by lower soil moisture and higher mineral component as well as less pronounced subsidence, suggesting limited ground ice melt and lower soil moisture dynamics (Bartsch et al., 2024 ). The following parameters did not show any influence on RTS spatial distribution: 1) lake parameters such as area, perimeter, eccentricity, orientation, solidity; 2) the bearing of RTS to the nearest lake; 3) the bearing of RTS to the nearest river; 4) Quaternary geology composition based on the State Geological Map of the Russian Federation; 5) ground ice characteristics based on historical Soviet permafrost maps. We describe in detail the datasets, their limitations, and the conducted tests in SI 1–1. RTS dynamics and association with climate variables To study the dynamics of RTSs from the 1960s to 2024, we analyzed the historical satellite imagery of CORONA (Grosse et al., 2005 ) and HEXAGON (Hammer et al., 2022 ) as well as modern satellite imagery of PlanetScope 2024 (Planet Team, 2018 ). We mapped and classified not only RTSs from the inventory (Nesterova et al., 2025 ), but also identified additional RTSs (53 RTSs on historical imagery and 78 new RTSs in 2024) that were missed. RTSs as features were classified as “RTS” (visually clearly identified RTS with no doubts), “disturbance feature” (uncertain on the type of permafrost disturbance), “undisturbed tundra”, “stabilized disturbance” (in historical imagery visible as previously tundra disturbance or RTS, currently no visible disturbance), and, in PlanetScope 2024 imagery, as “active” or “stabilized”. For more details on mapping and classification methodology, please see SI 2–1.6-1.8. RTSs were mapped and classified for available years (1964, 1969, 1972, 1977, 1982, 1984) at 4 key sites: 3 in Gydan (G1, G2, G3) and 1 (Y1) in Yamal (Fig. 1 ). Key sites in Gydan covered ~ 4,489 km 2 (~ 18% of RTS clusters in Gydan) and the key site in Yamal covered ~ 1,614 km 2 (~ 10% of RTS clusters in Yamal). The key sites were selected based on the availability of historical imagery for the 1960s and 1970s. For an additional key site (Y2), historical imagery was only available for one time period. Hence, we include the results of its RTS dynamics only in the Extended Results in SI 1–4. RTS mapping and classification approaches are described in detail in Extended Methods in SI 2–1.6-1.8. We used manual mapping to resolve RTS occurrence and dynamics. We estimated the uncertainty range associated with the operator subjectivity at 4.1%, based on previous work with the West Siberian RTS inventory (Nesterova et al., 2025 ). We found a general increase in the number of clearly detected RTS by 23-fold (2,338%) in total for all key sites (~ 6103 km 2 ) from the time period of 1964–1972 to 2024 (Figs. 3 and 4 ). Most of the increase in RTS number occurred after 1984, contributing 86% to the total increase (Fig. 3 ). Initiation rates of RTS increased over time, with the first large increase taking place as early as the 1980s and the largest increase in recent years reaching up to ~ 78 RTSs yr − 1 (26-fold increase). We have not normalized initiation rates per area due to the arbitrary key site boundaries that were selected based on data availability. Thus, key sites may include areas with no RTSs at all. Then, normalization by their areas can be misleading. Our finding of RTS initiation rates of ⋜3 RTSs yr − 1 in the 1960s and ~ 8.8 RTSs yr − 1 for 1972–1982 suggests that the initiation rates were comparable with the ones reported for Canada. The reported RTS initiation rate at the Aklavik Plateau (360 km 2 , a > 16 times smaller study area) was 0.35 RTS yr − 1 for the 1954–1971 time period and 0.68 RTSs yr − 1 for 1985–2004 (Lacelle et al., 2010 ). Meanwhile, modern rates of RTS initiation of ~ 78 RTSs yr − 1 for our area are comparable with those reported for current RTS hotspots. This includes, for instance, the neighboring Taymyr Peninsula, where one study in North Taymyr (68,000 km 2 , a > 10 times larger study area) reported ~ 120 RTS yr − 1 for 2010–2021 (Bernhard et al., 2022 b). Another study focused only on the extremely active area of East Taymyr (91 km 2 ) and identified rates of ~ 23 RTS yr − 1 for the 2011–2020 period (Barth et al., 2025 ). Comparable high initiation rates were reported for Banks Island (~ 70,000 km 2 , a > 10 times larger study area) in Canada from 1984 to 2013, with values of 136 RTS yr − 1 (Lewkowicz and Way, 2019 ). We found that headwall retreat rates did not change much over time. The median headwall retreat rates were 1.3 m yr − 1 for 1964–1982/84 and 1.6 m yr − 1 for 1982/84-2024, and the highest headwall retreat rate was 11.4 m yr − 1 during the 1964–1982/1984 time period (Fig. 3 ). These headwall retreat rates are similar to those reported for Tibetan RTSs (0.05-5 m yr − 1 ; Sun et al., 2017 ; Fan et al., 2025). West Siberian RTS headwall retreat rates are rather stable and slow, especially if compared to the 6.2 m yr − 1 estimated in the Canadian High Arctic (Ward Jones et al., 2019 ) or in the neighboring Yugorsky peninsula with 66 m yr − 1 (Leibman et al., 2021 ). The maximum headwall retreats were 256 m (1972–2024) in Y2 (SI 1–4) and 230 m (1984–2024) in key site G1 (SI 1–4). The analysis of the RTS development trajectory (Fig. 3 a) aimed to define whether RTSs found in recent imagery are located in previously disturbed areas in the 1960s and 1980s. About 42% of all RTSs were found in areas of undisturbed tundra both in 1962–1972 and 1982–1984 time periods, while 33% were found as disturbance features in the 1980s, and 23% were disturbance features in the 1960s (Fig. 3 ). We show that a substantial part of RTSs occurred in pre-disturbed conditions back in the 1960s, which is similar to findings from the Canadian Arctic (Aklavik Plateau and Yukon Coast), where RTS locations were already observed in the 1950s (Lacelle et al., 2010 ; Ramage et al., 2018). Figure 4 shows the increase in the number of mapped RTSs for the time periods of 1964–1972, 1982–1984, and 2019–2024. The climate data suggests an increase both in winter (October-April) and summer (May-September) air temperatures, with the winter air temperature warming at a much faster rate. There is also a small yet significant increase in winter precipitation. To study the sensitivity of RTS initiation to climate variables, we employed a discrete-time proportional-hazards model (Suresh et al., 2023). The model assesses the risk of RTS initiation, assumed to come only from climate covariates. The results for the first time period (1964–1984) and the second time period (2019–2024) suggest that summer maximum precipitation and winter mean air temperature are the most important factors for RTS initiation. An increase in summer maximum precipitation by one standard deviation raised the annual RTS initiation probability by 27.62 percentage points (pp) (95% Bayesian uncertainty interval 25.93–28.41) with a hazard ratio (HR) of 3.9 (3.38–4.22), making it the strongest short-term influencing factor. Summer rainfalls are reported to associate with thaw-related mass movements not only for West Siberia (Leibman et al., 2003 ) but also for the Canadian Arctic (Kokelj et al., 2015 ; Lewkowicz and Way, 2019 ) and Alaska (Balser et al., 2014 ). Table 1 Effects on annual initiation for a + 1 standard deviation increase (risk-set AMEs* in pp, with 95% HDIs*; HRs* as medians with 95% intervals). Bayes factors are Savage–Dickey BF 10 *. For two time periods as well as for long-term analysis. 1964–1984 and 2019–2024 time periods Covariate Δ𝛾* (pp) HR (median [95%]) BF 10 Winter mean air temperature + 21.25 [19.34, 23.18] 3.11 [2.82, 3.46] ≫ 10 12 Summer maximum precipitation + 27.17 [25.93, 28.41] 3.80 [3.38, 4.22] ≫ 10 12 Winter maximum precipitation + 0.76 [0.08, 1.63] 1.02 [1.00, 1.06] 0.07 Summer mean air temperature + 0.40 [0.00, 1.14] 1.02 [1.00, 1.06] 0.08 Memory: winter maximum precipitation (2 years) + 3.13 [0.92, 5.10] 1.23 [1.09, 1.37] 57.68 Memory: winter mean air temperature (2 years) + 0.52 [0.00, 1.43] 1.03 [1.00, 1.08] 0.17 Long-term period 1964–1984 → 2019–2024 Covariate 1964–1984 →2019–2024 change (natural; SD*) Implied Δ𝛾 (pp, linear) Winter mean air temperature + 4.82 K* (+ 1.62 SD) + 33.56 Summer maximum precipitation -0.05 K (-0.08 SD) -2.2 Winter maximum precipitation + 0.75 K (+ 1.58 SD) + 1.43 Summer mean air temperature + 3.17 K (+ 2.28 SD) + 0.9 *Abbreviations and definitions: AME - Average marginal effect, an absolute change in the annual initiation probability; HDI - Highest density interval, a Bayesian uncertainty interval; HR - hazard ratio, the ratio of the hazard rates; Savage–Dickey BF 10 - Bayes factor (BF) - evidence estimate, BF > 10 indicates a strong evidence; SD - standard deviation; pp - percentage points, an absolute change on the probability scale; Δ𝛾 - the change in RTS initiation probability; K - Kelvin. The increase in one standard deviation of winter mean air temperature raises the probability of annual RTS initiation by 21.25 pp (19.34–23.18) with an HR of 3.11 (2.82–3.46), indicating strong preconditioning. Winter maximum precipitation over the two preceding years (“memory” climate variable that stores the cumulative information of the last two winters) has a rather small but still strongly evident effect on RTS initiation probability by adding 3.13 pp (0.92–5.10) with one standard deviation increase. Other climatic variables have rather small effects or evidence. Our results suggest that winter mean air temperature is a dominant long-term driver of RTS initiation. The increase in winter mean air temperature by 1.62 standard deviations leads to a rise in annual RTS initiation probability over the long term by 37.4 pp. The other analysed climate variables were found to be of weak evidence (Table 1 ). Multiple studies reported warm summer air temperature being associated with RTS development (Lacelle et al., 2010 ; Balser et al., 2014 ; Swanson and Nolan, 2018 ; Lewkowicz and Way, 2019 ; Ward Jones et al., 2019 ). Our findings suggest that the mean winter air temperature is not only a significant factor influencing RTS initiation over short time periods, but also is a long-term driver. Winter air temperatures in the West Siberian Arctic increased much faster than in summer (Fig. 4 ), which is consistent with global Arctic winter warming that exceeds summer warming by at least four times (Bintanja and van der Linden, 2013 ). Warmer winters imply shallower freezing of the ground and thus, thickening of the active layer, which was reported to deepen in West Siberia over the last 50 years (Vasiliev et al., 2020 ). A slight but yet significant increase in mean winter precipitation (Fig. 4 ), as well as a small but still evident effect of higher maximum winter precipitation for the preceding two years (“memory” climate variable) on RTS initiation found in our model, suggest that winter precipitation in the form of snow may also contribute to the thermal insulation of permafrost from cold air temperatures in winter and thus its warming. Warming of permafrost as well as deepening of the active layer thickness were reported to activate thaw-related mass movements in Central Yamal (Babkina et al., 2019 ). Potential for future RTS development and infrastructure risks Based on our findings, we anticipate more RTSs to occur in the West Siberian Arctic with ongoing climate warming and deepening of the active layer in the future. To identify areas of possible future RTS formation in the region, we analyzed which areas share the same environmental parameters as current clusters. We then filtered only areas with at least one RTS already mapped in Nesterova et al. ( 2025 ), which indicates the presence of massive ground ice. Next, we applied filters of the 95th percentile of the value range for the parameters that were found to be significantly different in the clusters compared to the rest of the peninsula (see Results - Spatial distribution). For the Yamal Peninsula, the filtering was performed by mean elevation and proportion of landcover classes, while for the Gydan Peninsula, in addition to these two factors, VRM and lake density were also included. As a result, the potential future RTS cluster areas for Yamal comprised 3,092.4 km², or 2.34% of the entire Yamal Peninsula (Fig. 5 ), which would account for an increase of 43.19% of highly RTS-impacted areas. The potential future RTS cluster areas for Gydan comprised 4,508.3 km², or 2.59% of the entire Gydan Peninsula (Fig. 5 ), representing an increase of 32.16% of highly RTS-impacted areas. Thaw-related mass wasting processes across the Arctic were reported to pose a risk to the infrastructure stability (Luo et al., 2019 ; Hjort et al., 2022 ). To estimate the potential risk from RTS development to infrastructure, we analyzed the infrastructure that is currently impacted by RTS clusters, as well as where infrastructure overlaps with areas that may develop into highly RTS-impacted areas in the future. The infrastructure in Yamal was found to be at much higher risk compared to Gydan. The currently largest area-wise affected infrastructure in Yamal belongs to the gas and oil industry, comprising ~ 5.3 km² of affected area. This includes the Bovanenkovo gas field settlement, including the airport with a runway, as well as the Bovanenkovo-Ukhta pipeline (410 km, which is ~ 37% of total length). Around ~ 2.5 km² of transport infrastructure, such as roads, bridges, and tunnels, is also affected, mostly within the Bovanenkovo gas field settlement. Around 66% of all power & energy infrastructure (49 km of length) within Bovanenkovo was found at risk of damage, possibly posing substantial environmental hazards. The potential future RTS clustering could lead to further risk for the Obskaya-Bovanenkovo-Karskaya railway line, which could increase by more than twice from 19 km of the currently affected length to 52 km. Oil and gas, as well as transport infrastructure with potential RTS clusters, can experience an increase of 15 km in length. Unlike in Yamal, the infrastructure of the Gydan peninsula is currently not affected. Gydan’s largest Salmanovskoye gas and oil field, including the airport with a runway, is in the vicinity of the potential RTS clusters. Potential future RTS-impacted clusters could also develop in Tadebya-Yakha, a small settlement on the western coast of Gydan. The majority (> 75%) of all RTSs from the inventory are located on lake shores (Nesterova et al., 2025 ). We found that more than ~ 2400 lakes in the West Siberian Arctic are currently affected by RTSs. As RTS development affects not only hydrology, but also the water quality (Kokelj et al., 2005 ; Kokelj & Jorgenson, 2013 ), this correlation between RTS and lakes may pose substantial risks to aquatic habitats and local indigenous communities, which requires further detailed investigation in future research. Methods Spatial analysis To perform all geospatial operations as well as map layout, we used Python language v.3.11.10 in Visual Studio Code v. 1.103.2 and QGIS v. 3.36.2-Maidenhead. All geospatial operations and mapping were performed in WGS84 / UTM zone 43 projection (EPSG:32643). To investigate the spatial distribution of RTSs as well as environmental parameters (elevation, lake density, landcover) within the study area, we aggregated our data into Uber H3 hexagonal grid with a resolution of 6, which is ~ 35 km² (Uber Technologies Inc., 2018). This grid system was chosen since it is widely used (Aini et al., 2023 ; Nitze et al., 2025 ), well-integrated in Python, and fully reproducible due to its set boundaries. To estimate elevation, we used the ArcticDEM dataset (Porter et al., 2023 ). To assess lake density, we created a dataset based on OpenStreetMap dataset (©OpenStreetMap contributors, ODbL 1.0, website, accessed 08.08.2025). To analyze a landcover, we used the circumarctic land cover (Bartsch et al., 2024 ). The infrastructure dataset was also accessed via OSM. RTS polygons for the 1960s and 1980s were digitized based on georeferenced CORONA and HEXAGON imagery (Grosse et al., 2005 ; Hammer et al., 2022 ). RTS polygons for 2024 were created based on PlanetScope imagery (Planet Team, 2018 ). The detailed description of all datasets used in this study, the limitations, and all the pre-processing steps applied are described in detail in the SI 2–1.1-1.4. Statistical Analysis Clusters of RTS accumulation were identified using Anselin Local Moran’s I statistic (Moran, 1950 ), with Queen’s Case (Lloyd, 2010 ) employed as the conceptualization of spatial relationships to account for adjacency effects. To select a test of hypothesis testing we first estimated the normality of analyzed distributions of mean elevations, VRM and lake density using multiple tests such as Shapiro-Wilk test (Shapiro and Wilk, 1965 ), D’Agostino’s K² test combines skewness and kurtosis (D’Agostino et al., 1990), and two modifications of Kolmogorov-Smirnov test: Anderson–Darling test that puts more weight to tails (Anderson and Darling, 1954 ), Lilliefors test that accounts for the fact that the mean and variance are estimated from the sample (Lilliefors, 1967 ). All of the tests eventually showed that distributions of all parameters, both for all the peninsula and for clusters, were not normal; thus, for the hypothesis testing, we chose a non-parametric test. To get complementary evidence, we applied two non-parametric tests: the Mann-Whitney U-Test, which compares central tendencies through differences in medians (Mann and Whitney, 1947), and the Kolmogorov-Smirnov Test that compares overall distribution differences, including shape, spread, and medians of the data (Massey, 1951 ). Bayesian climate hazard modelling of retrogressive thaw slump initiation We analyzed the climate sensitivity of retrogressive thaw slump (RTS) initiation using a discrete-time proportional hazards model (Allison, 1982 ). Annual initiation hazards were linked to covariates through a complementary log–log function (Allison, 1982 ). The linear predictor included four standardized climate covariates: thawing degree days, summer maximum precipitation, winter mean temperature, and winter maximum precipitation. Coefficients expected to increase risk were constrained to be non-negative with HalfNormal priors (Gelman et al., 2008 ). Models were estimated in a Bayesian framework with Hamiltonian Monte Carlo (Betancourt, 2018 ). Prior predictive simulations confirmed plausibility, and posterior predictive checks compared yearly onset counts with observations (Auger-Méthé et al., 2019). Effect sizes are reported as hazard ratios (HR) and as average marginal effects (AME), expressed as percentage-point changes in annual initiation probability averaged across the pre-initiation risk set. For more detailed settings, see SI 2–2. Declarations Competing financial interest The authors declare no competing financial interests. Materials & Correspondence. Correspondence and material requests should be addressed to NN ( [email protected] ). Funding We acknowledge support for this study by DAAD to NN (“STIBET-I”). NN, GG, IN, and TH were supported by the BMWK project ML4EARTH and the google.org Impact Challenge on Climate Innovation to the Permafrost Discovery Gateway development team. IN and GG were additionally supported by the NSF Navigating the New Arctic Permafrost Discovery Gateway (#1927872 #2052107). ML was supported by the state assignment of the Ministry of Science and Higher Education of the Russian Federation (Project No. FWRZ-2021-0012). HL has received funding from the EU Horizon Europe, grant agreement No. 101133587 (ILLUQ). CS was supported by HEIBRiDS. PlanetScope data were provided through the NASA CSDA program (for NSF affiliated research on the Permafrost Discovery Gateway) and the Planet Research Program. Author Contribution Conceptualization - NN, ML, GG, IN; Methodology - NN, CS; Software -NN, TH; Validation - NN, MV; Formal analysis - NN, CS; Investigation - NN, CS; Data Curation - TH, MV, KM, IT; Writing - Original Draft - NN, CS; Writing - Review & Editing - all authors; Visualization - NN, CS; Supervision - ML, GG, IN, HL. Acknowledgement The first author would like to thank Rustam Khairullin (b-geos) for processing the missing part of the landcover, and Chiara Philips for the design recommendations. Data Availability The datasets that were generated during the study are currently available from the corresponding author upon request to the reviewers and the editors, but will be publicly available on PANGAEA upon publication of the study. References Aini, A. N., Dewantari, O. A., Mandala, D. P. A. & Bisri, M. B. F. An Enhanced Earthquake Risk Analysis using H3 Spatial Indexing. IOP Conference Series: Earth and Environmental Science. 1245(1):012014. (2023). https://doi.org/10.1088/1755-1315/1245/1/012014 Allison, P. D. Discrete-Time Methods for the Analysis of Event Histories. Sociol. Methodol. 13 , 61–98. https://doi.org/10.2307/270718 (1982). Anderson, T. W. & Darling, D. A. A Test of Goodness of Fit. J. Am. Stat. Assoc. 49 (268), 765–769. https://doi.org/10.2307/2281537 (1954). Auger-Méthé, M. et al. A guide to state–space modeling of ecological time series. Ecol. Monogr. 91 (4), e01470 (2021). Babkina, E. A. et al. Activation of Cryogenic Processes in Central Yamal as a Result of Regional and Local Change in Climate and Thermal State of Permafrost. Russ. Meteorol. Hydrol. 44 (4), 283–290. https://doi.org/10.3103/S1068373919040083 (2019). Badu, Y. B. Ice content of cryogenic strata (permafrost interval) in gas-bearing structures, Northern Yamal. Earth’s Cryosphere . 19 (3), 9–18 (2015). Balser, A. W., Jones, J. B. & Gens, R. Timing of retrogressive thaw slump initiation in the Noatak Basin, northwest Alaska, USA. J. Geophys. Research: Earth Surf. 119 (5), 1106–1120. https://doi.org/10.1002/2013JF002889 (2014). Barth, S. et al. Rapid Changes in Retrogressive Thaw Slump Dynamics in the Russian High Arctic Based on Very High-Resolution Remote Sensing. Geophys. Res. Lett. 52 (7). https://doi.org/10.1029/2024GL113022 (2025). e2024GL113022. Bartsch, A. et al. Circumarctic land cover diversity considering wetness gradients. Hydrol. Earth Syst. Sci. 28 (11), 2421–2481. https://doi.org/10.5194/hess-28-2421-2024 (2024). Baulin, V. V., Belopukhova, E. B., Dubikov, G. I. & Shmelev, L. M. Geocryological conditions of western Siberia Lowland (Nauka, 1967). Bernhard, P., Zwieback, S. & Hajnsek, I. Accelerated Mobilization of Organic Carbon from Retrogressive Thaw Slumps on the Northern Taymyr Peninsula. Cryosphere Discuss. 2022 , 1–24. https://doi.org/10.5194/tc-16-2819-2022 (2022). Bernhard Philipp, Zwieback, S., Bergner, N. & Hajnsek, I. Assessing volumetric change distributions and scaling relations of retrogressive thaw slumps across the Arctic. Cryosphere 16 (1), 1–15. https://doi.org/10.5194/tc-16-1-2022 (2022). Betancourt, M. A Conceptual Introduction to Hamiltonian Monte Carlo. [accessed 2025 Sept 9]. http://arxiv.org/abs/1701.02434 . (2018). https://doi.org/10.48550/arXiv.1701.02434 Bintanja, R. & van der Linden, E. C. The changing seasonal climate in the Arctic. Sci. Rep. 3 (1), 1556. https://doi.org/10.1038/srep01556 (2013). Biskaborn, B. K. et al. Permafrost is warming at a global scale. Nat. Commun. 10 (1), 264. https://doi.org/10.1038/s41467-018-08240-4 (2019). Bradley, J. A. et al. Svalbard winter warming is reaching melting point. Nat. Commun. 16 (1), 6409. https://doi.org/10.1038/s41467-025-60926-8 (2025). D’Agostino, R. B. & Belanger, A. A Suggestion for Using Powerful and Informative Tests of Normality. Am. Stat. 44 (4), 316–321. https://doi.org/10.2307/2684359 (1990). Farquharson, L. M., Romanovsky, V. E., Kholodov, A. & Nicolsky, D. Sub-aerial talik formation observed across the discontinuous permafrost zone of Alaska. Nat. Geosci. 15 (6), 475–481. https://doi.org/10.1038/s41561-022-00952-z (2022). Gelman, A., Jakulin, A., Pittau, M. G. & Su, Y-S. A weakly informative default prior distribution for logistic and other regression models. Annals Appl. Stat. 2 (4), 1360–1383. https://doi.org/10.1214/08-AOAS191 (2008). Grosse, G., Schirrmeister, L., Kunitsky, V. V. & Hubberten, H-W. The use of CORONA images in remote sensing of periglacial geomorphology: an illustration from the NE Siberian coast. Permafrost Periglac. Process. 16 (2), 163–172. https://doi.org/10.1002/ppp.509 (2005). Hammer, E., FitzPatrick, M. & Ur, J. Succeeding CORONA: declassified HEXAGON intelligence imagery for archaeological and historical research. Antiquity 96 (387), 679–695. https://doi.org/10.15184/aqy.2022.22 (2022). Hjort, J. et al. Impacts of permafrost degradation on infrastructure. Nat. Reviews Earth Environ. 3 (1), 24–38. https://doi.org/10.1038/s43017-021-00247-8 (2022). Khomutov, A. et al. Activation of Cryogenic Earth Flows and Formation of Thermocirques on Central Yamal as a Result of Climate Fluctuations. In: Mikoš K, Vilímek V, Yin Y, Sassa K., editors. Advancing Culture of Living with Landslides. Proceedings of World Landslide Forum 4, May 29 – June 2, 2017, Ljubljana, Slovenia. Vol.5, Landslides in Different Environments. Springer International Publishing AG; p 209–216 (2017). Kokelj, S. V. et al. Increased precipitation drives mega slump development and destabilization of ice-rich permafrost terrain, northwestern Canada. Glob. Planet Change . 129 , 56–68. https://doi.org/10.1016/j.gloplacha.2015.02.008 (2015). Kokelj, S. V. et al. The influence of thermokarst disturbance on the water quality of small upland lakes, Mackenzie Delta region, Northwest Territories, Canada. Permafrost Periglac. Process. 16 (4), 343–353. https://doi.org/10.1002/ppp.536 (2005). Kokelj, S. V. & Lewkowicz, A. G. Salinization of permafrost terrain due to natural geomorphic disturbance, Fosheim Peninsula, Ellesmere Island 372–385 (Arctic, 1999). Kokelj, S. V. & Jorgenson, M. T. Advances in thermokarst research. Permafrost Periglac. Process. 24 (2), 108–119. https://doi.org/10.1002/ppp.1779 (2013). Kokelj, S. V. et al. Climate-driven thaw of permafrost preserved glacial landscapes, northwestern Canada. Geology 45 (4), 371–374. https://doi.org/10.1130/G38626.1 (2017). Lacelle, D., Bjornson, J. & Lauriol, B. Climatic and geomorphic factors affecting contemporary (1950–2004) activity of retrogressive thaw slumps on the Aklavik Plateau, Richardson Mountains, NWT, Canada. Permafrost Periglac. Process. 21 (1), 1–15. https://doi.org/10.1002/ppp.666 (2010). Lantuit, H. & Pollard, W. H. Fifty years of coastal erosion and retrogressive thaw slump activity on Herschel Island, southern Beaufort Sea, Yukon Territory, Canada. Geomorphology 95 (1–2), 84–102. https://doi.org/10.1016/j.geomorph.2006.07.040 (2008). Leibman, M. O. & Kizyakov, A. I. Cryogenic landslides of the Yamal and Yugorsky Peninsulas (In Russian) Vol 206 (Earth Cryosphere Institute SB RAS, 2007). Leibman, M. O., Kizyakov, A. I., Sulerzhitsky, L. D. & Zaretskaia, N. E. Dynamics of landslide slopes and their development on Yamal Peninsula. In: Permafrost. Proceedings of the 8th international conference on permafrost. Swets and Zeitlinger, Lisse. p 651–656 (2003). Leibman, M. et al. Coastal Retreat Due to Thermodenudation on the Yugorsky Peninsula, Russia during the Last Decade, Update since 2001–2010. Remote Sens. 13 (20), 4042. https://doi.org/10.3390/rs13204042 (2021). Lewkowicz, A. G. & Way, R. G. Extremes of summer climate trigger thousands of thermokarst landslides in a High Arctic environment. Nat. Commun. 10 (1), 1329. https://doi.org/10.1038/s41467-019-09314-7 (2019). Lewkowicz, A. G. Retrogressive thaw slump activity in the western Canadian Arctic (1984–2016). (2024). Lilliefors, H. W. On the Kolmogorov-Smirnov Test for Normality with Mean and Variance Unknown. J. Am. Stat. Assoc. 62 (318), 399–402. https://doi.org/10.2307/2283970 (1967). Lloyd, C. D. Spatial data analysis: an introduction for GIS users (Oxford University Press, 2010). Luo, J. et al. Recent acceleration of thaw slumping in permafrost terrain of Qinghai-Tibet Plateau: An example from the Beiluhe Region. Geomorphology 341 , 79–85. https://doi.org/10.1016/j.geomorph.2019.05.020 (2019). Mackay, J. R. Segregated epigenetic ice and slumps in permafrost, Mackenzie Delta area, NWT. Geographical Bull. 8 , 59–80 (1966). Maier, K. et al. Detecting mass wasting of Retrogressive Thaw Slumps in spaceborne elevation models using deep learning. Int. J. Appl. Earth Obs. Geoinf. 137 , 104419. https://doi.org/10.1016/j.jag.2025.104419 (2025). Makopoulou, E. et al. Retrogressive thaw slump susceptibility in the northern hemisphere permafrost region. Earth. Surf. Proc. Land. 49 (11), 3319–3331. https://doi.org/10.1002/esp.5890 (2024). Massey, F. J. The Kolmogorov-Smirnov Test for Goodness of Fit. J. Am. Stat. Assoc. 46 (253), 68–78. https://doi.org/10.2307/2280095 (1951). Mesquita, P. S., Wrona, F. J. & Prowse, T. D. Effects of retrogressive permafrost thaw slumping on sediment chemistry and submerged macrophytes in Arctic tundra lakes. Freshw. Biol. 55 (11), 2347–2358 (2010). Moran, P. A. P. Notes on Continuous Stochastic Phenomena. Biometrika 37 (1/2), 17–23. https://doi.org/10.2307/2332142 (1950). Nesterova, N. et al. Review article: Retrogressive thaw slump characteristics and terminology. Cryosphere 18 (10), 4787–4810. https://doi.org/10.5194/tc-18-4787-2024 (2024). Nesterova, N. et al. High-resolution inventory and classification of retrogressive thaw slumps in West Siberia. Earth Syst Sci Data Discuss. 2025:1–30. (2025). https://doi.org/10.5194/essd-2025-164 Nitze, I. et al. Remote sensing quantifies widespread abundance of permafrost region disturbances across the Arctic and Subarctic. Nat. Commun. 9 (1). https://doi.org/10.1038/s41467-018-07663-3 (2018). Nitze, I. et al. DARTS: Multi-year database of AI-detected retrogressive thaw slumps in the circum-arctic permafrost region. Sci. Data . 12 (1), 1512. https://doi.org/10.1038/s41597-025-05810-2 (2025). Novikova, A. et al. Dynamics of Permafrost Coasts of Baydaratskaya Bay (Kara Sea) Based on Multi-Temporal Remote Sensing Data. Remote Sens. 10 (9). https://doi.org/10.3390/rs10091481 (2018). O’Neill, H. B., Wolfe, S. A. & Duchesne, C. New ground ice maps for Canada using a paleogeographic modelling approach. Cryosphere 13 (3), 753–773. https://doi.org/10.5194/tc-13-753-2019 (2019). Obu, J. How Much of the Earth’s Surface is Underlain by Permafrost? J. Geophys. Research: Earth Surf. 126 (5). https://doi.org/10.1029/2021JF006123 (2021). e2021JF006123. Obu, J. et al. Northern Hemisphere permafrost map based on TTOP modelling for 2000–2016 at 1 km2 scale. Earth Sci. Rev. 193 , 299–316. https://doi.org/10.1016/j.earscirev.2019.04.023 (2019). Pascual, D. & Johansson, M. Increasing impacts of extreme winter warming events on permafrost. Weather Clim. Extremes . 36 , 100450. https://doi.org/10.1016/j.wace.2022.100450 (2022). Planet Team. Planet imagery product specifications. (2018). https://assets.planet.com/docs/combined-imagery-product-spec-april-2019.pdf Pollard, W. H. & French, H. M. A first approximation of the volume of ground ice, Richards Island, Pleistocene Mackenzie Delta, Northwest Territories, Canada. Can. Geotech. J. 17 (4), 509–516 (1980). Porter, C. et al. ArcticDEM - Mosaics, Version 4.1. /10.7910/DVN/3VDC4W. https://doi.org/10.7910/DVN/3VDC4W (2023). Schuur, E. A. G. et al. Climate change and the permafrost carbon feedback. Nature 520 (7546), 171–179. https://doi.org/10.1038/nature14338 (2015). Shapiro, S. S. & Wilk, M. B. An Analysis of Variance Test for Normality (Complete Samples). Biometrika 52 (3/4), 591–611. https://doi.org/10.2307/2333709 (1965). Smith, S. L. et al. The changing thermal state of permafrost. Nat. Reviews Earth Environ. 3 (1), 10–23. https://doi.org/10.1038/s43017-021-00240-1 (2022a). Smith, S. L. et al. The changing thermal state of permafrost. Nat. Reviews Earth Environ. 3 (1), 10–23. https://doi.org/10.1038/s43017-021-00240-1 (2022b). Streletskaya, I. et al. Pleistocene-Holocene palaeoenvironmental records from permafrost sequences at the Kara Sea coast (NW Siberia, Russia). Geography, environment, sustainability. 6(3):60–76 (2013). Sun, Z. et al. Creep characteristics and process analyses of a thaw slump in the permafrost region of the Qinghai-Tibet Plateau, China. Geomorphology 293 , 1–10. https://doi.org/10.1016/j.geomorph.2017.04.045 (2017). Suresh, K., Severn, C. & Ghosh, D. Survival prediction models: an introduction to discrete-time modeling. BMC Med. Res. Methodol. 22 (1), 207. https://doi.org/10.1186/s12874-022-01679-6 (2022). Swanson, D. K. & Nolan, M. Growth of retrogressive thaw slumps in the Noatak Valley, Alaska, 2010–2016, measured by airborne photogrammetry. Remote Sens. 10 (7). https://doi.org/10.3390/rs10070983 (2018). Uber Technologies Inc. H3: Uber’s hexagonal hierarchical spatial index. (2018). Vasiliev, A. A. et al. Permafrost degradation in the Western Russian Arctic. Environ. Res. Lett. 15 (4), 045001. https://doi.org/10.1088/1748-9326/ab6f12 (2020). Ward Jones, M. K., Pollard, W. H. & Jones, B. M. Rapid initialization of retrogressive thaw slumps in the Canadian high Arctic and their response to climate and terrain factors. Environ. Res. Lett. 14 (5). https://doi.org/10.1088/1748-9326/ab12fd (2019). Webb, H. et al. A Review of Abrupt Permafrost Thaw: Definitions, Usage, and a Proposed Conceptual Framework. Curr. Clim. Change Rep. 11 (1), 7. https://doi.org/10.1007/s40641-025-00204-3 (2025). Yang, Y. et al. Mapping retrogressive thaw slumps using deep neural networks. Remote Sens. Environ. 288 , 113495. https://doi.org/10.1016/j.rse.2023.113495 (2023). OpenStreetMap. OpenStreetMap; [accessed 2025 Aug 8]. https://www.openstreetmap.org/ Additional Declarations No competing interests reported. Supplementary Files SI1Nesterovaetal.docx SI2Nesterovaetal.docx Cite Share Download PDF Status: Under Review Version 1 posted Editorial decision: Revision requested 20 Nov, 2025 Reviews received at journal 12 Nov, 2025 Reviewers agreed at journal 10 Nov, 2025 Reviews received at journal 03 Nov, 2025 Reviewers agreed at journal 30 Oct, 2025 Reviewers agreed at journal 16 Oct, 2025 Reviewers invited by journal 14 Oct, 2025 Editor assigned by journal 26 Sep, 2025 Submission checks completed at journal 25 Sep, 2025 First submitted to journal 23 Sep, 2025 You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-7697239","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":529673862,"identity":"dd450abd-e495-4ce5-abbd-f486c2a99f4a","order_by":0,"name":"Nina Nesterova","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA5UlEQVRIiWNgGAWjYDACZlSuDVhEgigtPBAqjQgtMADVchhM4tVizs578APjDht7e/beh495as7LGxxnYLzxAY8Wy2a+ZAnGM2mJPTzHjY15jt023HCYgdlyBh4tBod5DCQY2w4n8EiksUnOYLvNuO0wA5s0D34txj8Y2/7b88g/A2r5d84erOUPfi1mQFsOMPZIsLFJfGw7kAjWgs/7IC0WiW3JiT1n0pgNPvYlJ+8/zNhs2YNPy/kzxjc+ttnZs7cfY3yQ8M3Odmb/4YM3fuCzBgQSULmMDYQ0jIJRMApGwSggAAChPEV5r69s5AAAAABJRU5ErkJggg==","orcid":"","institution":"Alfred Wegener Institute for Polar and Marine Research","correspondingAuthor":true,"prefix":"","firstName":"Nina","middleName":"","lastName":"Nesterova","suffix":""},{"id":529673863,"identity":"3841669c-b253-4848-a266-d53c7382abab","order_by":1,"name":"Marina Leibman","email":"","orcid":"","institution":"Tyumen Scientific Centre SB RAS","correspondingAuthor":false,"prefix":"","firstName":"Marina","middleName":"","lastName":"Leibman","suffix":""},{"id":529673864,"identity":"fea622da-53e2-4fb4-b9bd-c122a5fd03b6","order_by":2,"name":"Carl Stadie","email":"","orcid":"","institution":"Alfred Wegener Institute for Polar and Marine Research","correspondingAuthor":false,"prefix":"","firstName":"Carl","middleName":"","lastName":"Stadie","suffix":""},{"id":529673865,"identity":"70004d1d-e416-4b48-a0e6-53b216dfa914","order_by":3,"name":"Tobias Hölzer","email":"","orcid":"","institution":"Alfred Wegener Institute for Polar and Marine Research","correspondingAuthor":false,"prefix":"","firstName":"Tobias","middleName":"","lastName":"Hölzer","suffix":""},{"id":529673866,"identity":"804905ee-febd-47c7-bc86-5b9a3fe8044f","order_by":4,"name":"Ingmar Nitze","email":"","orcid":"","institution":"Alfred Wegener Institute for Polar and Marine Research","correspondingAuthor":false,"prefix":"","firstName":"Ingmar","middleName":"","lastName":"Nitze","suffix":""},{"id":529673867,"identity":"41914bc8-124c-436f-b07b-33f1e200e1f0","order_by":5,"name":"Ilia Tarasevich","email":"","orcid":"","institution":"University of Houston","correspondingAuthor":false,"prefix":"","firstName":"Ilia","middleName":"","lastName":"Tarasevich","suffix":""},{"id":529673868,"identity":"8f389e8c-054f-4b63-bb0c-dbc93d6d5b6c","order_by":6,"name":"Kathrin Maier","email":"","orcid":"","institution":"ETH Zurich","correspondingAuthor":false,"prefix":"","firstName":"Kathrin","middleName":"","lastName":"Maier","suffix":""},{"id":529673869,"identity":"660b50c0-66ec-43ca-a9a6-ed519766fb45","order_by":7,"name":"Maiia Vasileva","email":"","orcid":"","institution":"Alfred Wegener Institute for Polar and Marine Research","correspondingAuthor":false,"prefix":"","firstName":"Maiia","middleName":"","lastName":"Vasileva","suffix":""},{"id":529673870,"identity":"764086f8-4b6f-400c-a46f-7e2385c75b37","order_by":8,"name":"Hugues Lantuit","email":"","orcid":"","institution":"Alfred Wegener Institute for Polar and Marine Research","correspondingAuthor":false,"prefix":"","firstName":"Hugues","middleName":"","lastName":"Lantuit","suffix":""},{"id":529673871,"identity":"097acfd6-069e-4904-bb6e-aa902631554e","order_by":9,"name":"Guido Grosse","email":"","orcid":"","institution":"Alfred Wegener Institute for Polar and Marine Research","correspondingAuthor":false,"prefix":"","firstName":"Guido","middleName":"","lastName":"Grosse","suffix":""}],"badges":[],"createdAt":"2025-09-23 19:08:15","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-7697239/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-7697239/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":93556107,"identity":"ab8dd86d-e534-4444-8fde-c827fa077e86","added_by":"auto","created_at":"2025-10-15 06:41:15","extension":"docx","order_by":0,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":5806658,"visible":true,"origin":"","legend":"","description":"","filename":"Nesterovaetalmanuscript240925.docx","url":"https://assets-eu.researchsquare.com/files/rs-7697239/v1/f562c24eae72801367ed5d1d.docx"},{"id":93558476,"identity":"300920ae-a2ae-44f9-969f-a2efd9c30c75","added_by":"auto","created_at":"2025-10-15 07:05:15","extension":"json","order_by":1,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":10117,"visible":true,"origin":"","legend":"","description":"","filename":"f1b884eebd0d434a8a384a6a31b32621.json","url":"https://assets-eu.researchsquare.com/files/rs-7697239/v1/2d55c7a808118ae865125b48.json"},{"id":93556104,"identity":"97b1978b-7822-46b0-8f15-a82b9306fb4e","added_by":"auto","created_at":"2025-10-15 06:41:15","extension":"docx","order_by":2,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":1859262,"visible":true,"origin":"","legend":"","description":"","filename":"SI1Nesterovaetal.docx","url":"https://assets-eu.researchsquare.com/files/rs-7697239/v1/90ac77234c9759d859bfbcb4.docx"},{"id":93556111,"identity":"676390d7-2198-4f12-94f7-ebe2b0299992","added_by":"auto","created_at":"2025-10-15 06:41:15","extension":"docx","order_by":3,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":308999,"visible":true,"origin":"","legend":"","description":"","filename":"SI2Nesterovaetal.docx","url":"https://assets-eu.researchsquare.com/files/rs-7697239/v1/670366588615eb9c96f65e70.docx"},{"id":93556112,"identity":"68388b83-f35e-4ca0-a0ae-0785a3283393","added_by":"auto","created_at":"2025-10-15 06:41:15","extension":"xml","order_by":4,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":140871,"visible":true,"origin":"","legend":"","description":"","filename":"f1b884eebd0d434a8a384a6a31b326211enriched.xml","url":"https://assets-eu.researchsquare.com/files/rs-7697239/v1/79855c431ecfb4e720de3d40.xml"},{"id":93557283,"identity":"c86a7547-fc9c-4eb7-b957-994add3f680e","added_by":"auto","created_at":"2025-10-15 06:49:16","extension":"png","order_by":10,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":552243,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-7697239/v1/1d51f1f1cc4a9d9107b2c4d4.png"},{"id":93557280,"identity":"be9b97b4-51d7-4ff6-bef8-ecab32fd81d1","added_by":"auto","created_at":"2025-10-15 06:49:16","extension":"png","order_by":11,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":109936,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-7697239/v1/ed22e5ec11bfef2b2d70801b.png"},{"id":93558270,"identity":"6d147ba0-f0f3-4384-ac01-827cfdfc0dc1","added_by":"auto","created_at":"2025-10-15 06:57:15","extension":"png","order_by":12,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":38005,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage3.png","url":"https://assets-eu.researchsquare.com/files/rs-7697239/v1/c5ac1af1b35484042cdea69e.png"},{"id":93556108,"identity":"26710ce3-d562-4c02-b4b2-38c5840b5276","added_by":"auto","created_at":"2025-10-15 06:41:15","extension":"png","order_by":13,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":58632,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage4.png","url":"https://assets-eu.researchsquare.com/files/rs-7697239/v1/6a704f26da6fcf343999376e.png"},{"id":93556115,"identity":"6f8672cb-4d4f-4387-b2a3-6dab25add75d","added_by":"auto","created_at":"2025-10-15 06:41:15","extension":"png","order_by":14,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":224148,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage5.png","url":"https://assets-eu.researchsquare.com/files/rs-7697239/v1/afbecb3606385252815ad3e3.png"},{"id":93557282,"identity":"ffec8c9e-9b76-4060-b6a0-7c44bb4c9d7e","added_by":"auto","created_at":"2025-10-15 06:49:16","extension":"xml","order_by":15,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":141742,"visible":true,"origin":"","legend":"","description":"","filename":"f1b884eebd0d434a8a384a6a31b326211structuring.xml","url":"https://assets-eu.researchsquare.com/files/rs-7697239/v1/b8c1e3e5dd53fac4e1b6287a.xml"},{"id":93556118,"identity":"e46cf018-514f-4be6-b894-3ed232cdad8e","added_by":"auto","created_at":"2025-10-15 06:41:16","extension":"html","order_by":16,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":153943,"visible":true,"origin":"","legend":"","description":"","filename":"earlyproof.html","url":"https://assets-eu.researchsquare.com/files/rs-7697239/v1/056d84105ee506eb3a56c2e0.html"},{"id":93557276,"identity":"2b23e5ca-4ba9-453b-8532-2b33942f58af","added_by":"auto","created_at":"2025-10-15 06:49:15","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":2963601,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cem\u003e\u003cstrong\u003eRTSs in West Siberia. (a) \u003c/strong\u003e\u003c/em\u003e\u003cem\u003eRetrogressive thaw slump in Central Yamal (photo by Nina Nesterova, July 2019). \u003c/em\u003e\u003cem\u003e\u003cstrong\u003e(b) \u003c/strong\u003e\u003c/em\u003e\u003cem\u003eStudy area in the West Siberian Arctic: Yamal and Gydan peninsulas; RTSs mapped in the inventory by Nesterova et al., 2025; Dashed white and yellow polygons show the location of five key sites for time series analysis with high-resolution historical and recent imagery. Permafrost distribution is based on Obu et al. (2019). Basemap of panel (b): AWI Basemap ©2013-2025 Alfred-Wegener-Institut Helmholtz-Zentrum für Polar- und Meeresforschung.\u003c/em\u003e\u003c/p\u003e","description":"","filename":"floatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-7697239/v1/feb9eef45132cde76f27dde5.png"},{"id":93556099,"identity":"ece07b3e-8ba2-4032-86e1-21a87dbefb38","added_by":"auto","created_at":"2025-10-15 06:41:15","extension":"jpeg","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":185682,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cem\u003e\u003cstrong\u003eSpatial distribution of RTSs in West Siberia. (a) \u003c/strong\u003e\u003c/em\u003e\u003cem\u003eLandcover composition (Bartsch et al., 2024) for the study area and statistically significant RTS clusters (black shaded). (\u003c/em\u003e\u003cem\u003e\u003cstrong\u003eb) \u003c/strong\u003e\u003c/em\u003e\u003cem\u003eRelative difference in landcover class proportion for RTS clusters versus the rest of Yamal and Gydan peninsulas. (\u003c/em\u003e\u003cem\u003e\u003cstrong\u003ec) \u003c/strong\u003e\u003c/em\u003e\u003cem\u003eEmpirical CDF of mean elevations (m a.s.l.) excluding Ural Mountains (\u0026gt;200 m a.s.l.) for RTS clusters and the rest of the Yamal Peninsula. (\u003c/em\u003e\u003cem\u003e\u003cstrong\u003ed) \u003c/strong\u003e\u003c/em\u003e\u003cem\u003eEmpirical CDF of mean elevations (m a.s.l.) excluding Ural Mountains (\u0026lt;200 m a.s.l.) for RTS clusters and the rest of Gydan Peninsula. (\u003c/em\u003e\u003cem\u003e\u003cstrong\u003ee) \u003c/strong\u003e\u003c/em\u003e\u003cem\u003eEmpirical CDF of VRM values for RTS clusters and the rest of the Gydan Peninsula. (\u003c/em\u003e\u003cem\u003e\u003cstrong\u003ef) \u003c/strong\u003e\u003c/em\u003e\u003cem\u003eP Empirical CDF of lake density values for RTS clusters and the rest of the Gydan Peninsula. Basemap of panel (a): AWI Basemap ©2013-2025 Alfred-Wegener-Institut Helmholtz-Zentrum für Polar- und Meeresforschung.\u003c/em\u003e\u003c/p\u003e","description":"","filename":"floatimage2.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-7697239/v1/c166f0b9590caa7bdcd90eb0.jpeg"},{"id":93556101,"identity":"a5d295d4-ce47-4729-9e40-c56e239ea484","added_by":"auto","created_at":"2025-10-15 06:41:15","extension":"jpeg","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":124453,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cem\u003e\u003cstrong\u003eRTS dynamics in the West Siberian Arctic 1964-2024 (same x-axis of time for all plots a to d): (a) \u003c/strong\u003e\u003c/em\u003e\u003cem\u003eRTS development trajectory including categories of undisturbed tundra, disturbance feature, RTS, as well as stabilized disturbance for the periods 1964-1972 and 1982-1984. The period 2019-2023 represents the majority of the years of satellite images used in the ESRI basemap mosaic in the inventory by Nesterova et al. (2025); thus, the categories for RTSs are either included in the inventory or not. For 2024, PlanetLabs imagery was used, where only active or stabilized categories can be distinguished. (\u003c/em\u003e\u003cem\u003e\u003cstrong\u003eb) \u003c/strong\u003e\u003c/em\u003e\u003cem\u003eHeadwall retreat rate distribution over the periods of 1964-1982/84 and 1982/84-2024. (\u003c/em\u003e\u003cem\u003e\u003cstrong\u003ec) \u003c/strong\u003e\u003c/em\u003e\u003cem\u003eRTS initiation rates per year for the time periods of 1960-1972 (or earlier), 1972-1984, 1985-2023, and 2024. The total headwall retreat rates inherit the positional uncertainty of historical imagery georeferencing of ~4 m, which was recalculated for the displacement uncertainty over the years. (\u003c/em\u003e\u003cem\u003e\u003cstrong\u003ed) \u003c/strong\u003e\u003c/em\u003e\u003cem\u003eNumber of initiated RTSs for the time periods of 1960-1972 (or earlier), 1972-1984, 1985-2023, and 2024.\u003c/em\u003e\u003c/p\u003e","description":"","filename":"floatimage3.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-7697239/v1/8d46640b8713c4d3d9bf9170.jpeg"},{"id":93558268,"identity":"595f01dc-35ec-45af-97fb-a04e71ba7cee","added_by":"auto","created_at":"2025-10-15 06:57:15","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":233627,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cem\u003e\u003cstrong\u003eRTS occurrence and climate variables 1964-2024 (the plots share a common time line): \u003c/strong\u003e\u003c/em\u003e\u003cem\u003eMapped RTS count, summer (May-September) mean air temperature (°C), winter (October-April) mean air temperature(°C), total summer precipitation (m), total winter precipitation (m).\u003c/em\u003e\u003c/p\u003e","description":"","filename":"floatimage4.png","url":"https://assets-eu.researchsquare.com/files/rs-7697239/v1/233cbfb2eedbad6037ae9cfa.png"},{"id":93556105,"identity":"961edf44-ae03-468b-84df-73c00666c060","added_by":"auto","created_at":"2025-10-15 06:41:15","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":2227813,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cem\u003e\u003cstrong\u003eMap of current and potential future RTS cluster areas in relation to major infrastructure. \u003c/strong\u003e\u003c/em\u003e\u003cem\u003eBasemap: AWI Basemap ©2013-2025 Alfred-Wegener-Institut Helmholtz-Zentrum für Polar- und Meeresforschung.\u003c/em\u003e\u003c/p\u003e","description":"","filename":"floatimage5.png","url":"https://assets-eu.researchsquare.com/files/rs-7697239/v1/4e014a1519f8913473d16bd7.png"},{"id":93559159,"identity":"97f23c76-c39e-4624-a349-6c4d60e0948d","added_by":"auto","created_at":"2025-10-15 07:13:18","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":6049585,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7697239/v1/af41c290-170e-4dcd-964d-47616779e456.pdf"},{"id":93556106,"identity":"2924db56-ef09-4bcb-82fc-618a59e9de35","added_by":"auto","created_at":"2025-10-15 06:41:15","extension":"docx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":1859262,"visible":true,"origin":"","legend":"","description":"","filename":"SI1Nesterovaetal.docx","url":"https://assets-eu.researchsquare.com/files/rs-7697239/v1/4b8c175f5d1e83bb59085311.docx"},{"id":93557279,"identity":"f33a9b7f-db23-4232-8b67-1ac44c9a1acc","added_by":"auto","created_at":"2025-10-15 06:49:15","extension":"docx","order_by":2,"title":"","display":"","copyAsset":false,"role":"supplement","size":308999,"visible":true,"origin":"","legend":"","description":"","filename":"SI2Nesterovaetal.docx","url":"https://assets-eu.researchsquare.com/files/rs-7697239/v1/3094ff745257c93f469c740e.docx"}],"financialInterests":"No competing interests reported.","formattedTitle":"Rapid increase in West Siberia’s retrogressive thaw slumps since 1964 associated with Arctic winter warming","fulltext":[{"header":"Introduction","content":"\u003cp\u003eIn recent years, evidence is mounting that Arctic winter warming is profoundly changing the terrestrial Arctic, including permafrost warming (Smith et al., 2022), widespread talik formation (Farquharson et al., \u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e2022\u003c/span\u003e), and accumulating ecological impacts of episodic but strong winter warming events (Pascual \u0026amp; Johanssen 2022; Bradley et al., \u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e2025\u003c/span\u003e). Permafrost, defined as ground that remains below 0\u0026deg;C for at least two consecutive years (Harris et al., 1988), underlies about 15% of the Northern Hemisphere landmass (Obu et al., 2021). Arctic permafrost experiences thawing and a reduction in extent in response to climate warming (Biskaborn et al., \u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e2019\u003c/span\u003e; Smith et al., 2022). Permafrost not only affects high northern latitude landscapes and ecosystems but also stores large amounts of carbon and nutrients, with carbon loss upon thaw contributing to global climate feedback (Schuur et al., \u003cspan citationid=\"CR57\" class=\"CitationRef\"\u003e2015\u003c/span\u003e). What drives permafrost thaw and over which time scales is therefore important to understand.\u003c/p\u003e\u003cp\u003eThaw of ice-rich permafrost often occurs abruptly, leading to the formation of distinctive landforms (Webb et al., \u003cspan citationid=\"CR68\" class=\"CitationRef\"\u003e2025\u003c/span\u003e). One of the most widespread and dynamic indicators of abrupt thaw is a retrogressive thaw slump (RTS) (Nesterova et al., \u003cspan citationid=\"CR45\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). An RTS is a slope failure initiated by the exposure and thaw of massive ground ice (Mackay, \u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e1966\u003c/span\u003e; Nesterova et al., \u003cspan citationid=\"CR45\" class=\"CitationRef\"\u003e2024\u003c/span\u003e) (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003ea). RTSs are reported throughout permafrost regions in the Arctic and high-mountain permafrost regions, where they significantly alter topography, hydrology, soil, and vegetation, while delivering large fluxes of sediments, carbon, and nutrients to downstream environments (Kokelj and Lewkowicz, \u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e1999\u003c/span\u003e; Kokelj et al., \u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e2005\u003c/span\u003e; Mesquita et al., \u003cspan citationid=\"CR43\" class=\"CitationRef\"\u003e2010\u003c/span\u003e).\u003c/p\u003e\u003cp\u003eThese landforms evolve in a polycyclic fashion through phases of active ice ablation and mudflow, followed by stabilization and vegetation colonization and further possible reactivation (Mackay, \u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e1966\u003c/span\u003e; Leibman and Kizyakov, \u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e2007\u003c/span\u003e; Lantuit \u0026amp; Pollard, \u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e2008\u003c/span\u003e). The initiation of RTSs was reported to be linked to climate drivers such as rising summer air temperatures and increased summer precipitation (Lantz and Kokelj, 2008; Kokelj et al., \u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e2015\u003c/span\u003e; Leibman et al., \u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e2021\u003c/span\u003e). However, predicting RTS initiation remains complicated due to the uncertainty of heterogeneous ground ice distribution (Pollard and French, \u003cspan citationid=\"CR55\" class=\"CitationRef\"\u003e1980\u003c/span\u003e; Makopoulou et al., \u003cspan citationid=\"CR41\" class=\"CitationRef\"\u003e2024\u003c/span\u003e), limited field data (Ward Jones et al., \u003cspan citationid=\"CR67\" class=\"CitationRef\"\u003e2019\u003c/span\u003e), and the lack of suitable models (O\u0026rsquo;Neill et al., 2016).\u003c/p\u003e\u003cp\u003eThe West Siberian Arctic is mostly characterized by continuous permafrost (Obu et al., \u003cspan citationid=\"CR52\" class=\"CitationRef\"\u003e2019\u003c/span\u003e; Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eb) and massive ground ice occurring close to the surface (Baulin et al., \u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e1967\u003c/span\u003e; Streletskaya et al., \u003cspan citationid=\"CR61\" class=\"CitationRef\"\u003e2013\u003c/span\u003e; Badu, \u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e2015\u003c/span\u003e). These two factors define the noteworthy abundance of RTSs in this region (Leibman and Kizyakov, \u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e2007\u003c/span\u003e; Khomutov et al., \u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e2017\u003c/span\u003e). The observed regional warming and deepening of the active layer thickness catalyze further RTS initiation and expansion (Babkina et al., \u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e2019\u003c/span\u003e; Biskaborn et al., \u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e2019\u003c/span\u003e; Vasiliev et al., \u003cspan citationid=\"CR66\" class=\"CitationRef\"\u003e2020\u003c/span\u003e). Yet, a lot of studies of this region have relied on local field sites (Leibman and Kizyakov, \u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e2007\u003c/span\u003e; Khomutov et al., \u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e2017\u003c/span\u003e; Novikova et al., \u003cspan citationid=\"CR49\" class=\"CitationRef\"\u003e2018\u003c/span\u003e; Khomutov et al., \u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e2017\u003c/span\u003e; Babkina et al., \u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e2019\u003c/span\u003e) or have been parts of large-scale automatic mapping with clear limitations and uncertainties due to a lack of field verification (Nitze et al., \u003cspan citationid=\"CR47\" class=\"CitationRef\"\u003e2018\u003c/span\u003e; Bernhard et al., \u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e2022\u003c/span\u003ea; Nitze et al., \u003cspan citationid=\"CR48\" class=\"CitationRef\"\u003e2025\u003c/span\u003e, Yang et al, \u003cspan citationid=\"CR69\" class=\"CitationRef\"\u003e2023\u003c/span\u003e; Maier et al., \u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e2025\u003c/span\u003e).\u003c/p\u003e\u003cp\u003eA recent high-resolution field-verified, manually-mapped, and classified RTS inventory dataset for the West Siberian Arctic now, for the first time, enables a large-scale robust analysis without compromising accuracy (Nesterova et al., \u003cspan citationid=\"CR46\" class=\"CitationRef\"\u003e2025\u003c/span\u003e). Here we present the first large-scale analysis of spatial and temporal patterns of RTS distribution in the West Siberian Arctic, particularly for the Yamal and Gydan peninsulas (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e), and an analysis of factors and drivers contributing to their accelerating development.\u003c/p\u003e\u003cp\u003eOur study aims to answer the following research questions: (1) What are the main spatial patterns of RTS distribution in the West Siberian Arctic? (2) What are the temporal trends of RTS dynamics in the West Siberian Arctic? (3) What environmental factors are associated with RTS development trends? and (4) What are possible areas of future RTS development? What are the current and potential risks for the infrastructure?\u003c/p\u003e\u003cp\u003eTo answer the first research question, we analyzed RTS spatial clustering and its relation to elevation, water bodies, and landcover. To study temporal RTS dynamics from the 1960s, we analyzed historical as well as modern high-resolution satellite imagery for five key sites (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e). To estimate climate influence on RTS dynamics, we compared RTS temporal trends with different climatic variables. To assess the risks of RTS development, we analyzed the current and possibly impacted infrastructure.\u003c/p\u003e\u003cp\u003e\u003c/p\u003e"},{"header":"Results and discussion","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e\u003ch2\u003eSpatial distribution\u003c/h2\u003e\u003cp\u003eTo explore the spatial distribution of RTS clusters in the West Siberian Arctic, we aggregated RTS point locations from the inventory into hexagonal grid cells (Uber H3 grid, resolution 6; ~35 km\u0026sup2;; Uber Technologies Inc., 2018). All further calculations and statistics were performed per hexagonal grid cell.\u003c/p\u003e\u003cp\u003ePreviously reported spatial clustering of RTSs in the region (Nesterova et al., \u003cspan citationid=\"CR46\" class=\"CitationRef\"\u003e2025\u003c/span\u003e) was confirmed by our results, with multiple statistically significant clusters revealed using Anselin Local Moran\u0026rsquo;s I (p ⋜ 0.005) (black hatched areas in Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003ea). These clusters represent 21,179 km\u0026sup2;, accounting for 6.9% of the total study area for both peninsulas (7,160 km\u0026sup2; \u0026minus;\u0026thinsp;5.4% for Yamal; 14,019 km\u0026sup2; \u0026minus;\u0026thinsp;8% for Gydan). This is consistent with a general tendency of spatial clustering of RTSs across the Arctic (Nitze et al., \u003cspan citationid=\"CR47\" class=\"CitationRef\"\u003e2018\u003c/span\u003e; Lewkowicz, \u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e2024\u003c/span\u003e).\u003c/p\u003e\u003cp\u003eAfter initial pre-processing and data cleaning (Extended Methods 1.1 and 1.2 in Supplementary Information 2 (SI 2- 1.1 and 1.2)), the analysis of environmental characteristics of RTS clusters revealed differences between the Yamal and Gydan peninsulas.\u003c/p\u003e\u003cp\u003eLandcover analysis of RTS clusters in both peninsulas revealed a dominant association with the \u003cem\u003emoist tundra\u003c/em\u003e class, with slight differences between peninsulas (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003ea and b). RTS classes in Yamal are associated with \u003cem\u003emoist tundra\u003c/em\u003e, abundant moss, prostrate to low shrubs/dwarf and low shrubs/dense dwarf and low shrubs (class IDs 10, 11, 12; IDs here and subsequently refer to the original landcover class ID developed by Bartsch et al., \u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e2024\u003c/span\u003e), \u003cem\u003emoist to wet tundra\u003c/em\u003e, dense dwarf and low shrubs (13) and mostly by \u003cem\u003emoist tundra\u003c/em\u003e, low shrubs (14) landcover classes, altogether representing\u0026thinsp;~\u0026thinsp;60% of landcover class proportion of RTS clusters in Yamal. Meanwhile, the class of \u003cem\u003edry tundra\u003c/em\u003e, characterized by abundant lichen and prostrate shrubs (7), which dominates in the north of Yamal, is the least represented (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003ea and b).\u003c/p\u003e\u003cp\u003eThe landcover class most commonly associated with RTS clusters in Gydan is \u003cem\u003emoist tundra\u003c/em\u003e with abundant moss and prostrate to low shrubs (10), with low shrubs (14), as well as dwarf and low shrubs (11), and \u003cem\u003emoist to wet tundra\u003c/em\u003e with dense dwarf and low shrubs (13). Also, the water (1) landcover class was among the dominant for the Gydan RTS clusters. Altogether, these classes represent\u0026thinsp;~\u0026thinsp;70% of the landcover class proportion of RTS clusters in Gydan. The most underrepresented land cover class compared to the rest of the Gydan Peninsula is also \u003cem\u003emoist tundra\u003c/em\u003e, but with abundant forbs and dwarf to tall shrubs (16).\u003c/p\u003e\u003cp\u003eAn elevation analysis of RTS clusters using ArcticDEM mosaics (Porter et al., \u003cspan citationid=\"CR56\" class=\"CitationRef\"\u003e2023\u003c/span\u003e) also revealed some differences between the two peninsulas. All elevation parameter distributions were non-normally distributed, thus non-parametric tests were applied. The distribution of mean elevations (95th percentile) of RTS clusters in Yamal showed a statistically significant difference compared to the mean elevations (95th percentile) of the rest (without RTS clusters) of the Yamal peninsula based on the Mann-Whitney U-Test (U\u0026thinsp;=\u0026thinsp;341,588, p\u0026thinsp;\u0026lt;\u0026thinsp;0.005 after Bonferroni correction), as well as based on Kolmogorov-Smirnov Test (D\u0026thinsp;=\u0026thinsp;0.198, p\u0026thinsp;\u0026lt;\u0026thinsp;0.001 after Bonferroni correction). This reflects the differences, which are depicted in the Empirical Cumulative Density Function (CDF) in Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003ec. The 95th percentile of mean elevation values for RTS clusters in Yamal ranged from 4.79 to 26.75 m a.s.l.\u003c/p\u003e\u003cp\u003eThe distribution of mean elevations of RTS clusters in Gydan demonstrated a strong statistically significant difference compared to the rest (without RTS clusters) of the Gydan peninsula based on both tests (U\u0026thinsp;=\u0026thinsp;809,112, p\u0026thinsp;\u0026lt;\u0026thinsp;0.001 after Bonferroni correction; D\u0026thinsp;=\u0026thinsp;0.167, p\u0026thinsp;\u0026lt;\u0026thinsp;0.001 after Bonferroni correction). RTS clusters in Gydan were significantly higher than the rest of the peninsula (median\u0026thinsp;+\u0026thinsp;3m). The empirical CDF of Gydan RTS clusters compared to the rest of the peninsula is depicted in Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003ed. The 95th percentile of mean elevation values for RTS clusters in Gydan ranged from 6.91 to 42.2 m a.s.l.\u003c/p\u003e\u003cp\u003eThe distribution of values of the vertical ruggedness measure (VRM, a characteristic of terrain roughness derived from the ArcticDEM mosaics) for RTS clusters did not show any statistical difference in Yamal (both tests); however, it showed a strong statistically significant difference compared to the rest of the peninsula in Gydan: U\u0026thinsp;=\u0026thinsp;945,361, p\u0026thinsp;\u0026lt;\u0026thinsp;0.001 after Bonferroni correction; D\u0026thinsp;=\u0026thinsp;0.199, p\u0026thinsp;\u0026lt;\u0026thinsp;0.001 after Bonferroni correction. RTS clusters in Gydan were rougher and characterized by higher VRM values (empirical CDF depicted in Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003ee). The 95th percentile of VRM values for RTS clusters in Gydan ranged from 0.95e-5 to 3.1e-5. In general, these values represent rather smooth terrain.\u003c/p\u003e\u003cp\u003eAn analysis of the lake density revealed no significant difference in lake density between RTS clusters and the rest of the peninsula in Yamal (both tests), but a strongly significant difference in Gydan (U\u0026thinsp;=\u0026thinsp;1,036,752, p\u0026thinsp;\u0026lt;\u0026thinsp;0.001 after Bonferroni correction; D\u0026thinsp;=\u0026thinsp;0.144, p\u0026thinsp;\u0026lt;\u0026thinsp;0.001 after Bonferroni correction). RTS clusters in Gydan occur in significantly higher lake densities. This finding is shown in the empirical CDF (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003ef) and is supported by the overrepresentation of the water landcover class (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eb). The 95th percentile values of lake density for clusters occur between 0.002 and 0.373.\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003eOur findings of RTS clusters tending to occur on higher elevations and rougher terrain are generally consistent with the literature review findings of inland RTSs occurring at the edges of terrains: river banks, lake shores, slopes of streams, and valleys (Nesterova et al., \u003cspan citationid=\"CR45\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). Moreover, most of the RTS clusters in Gydan are located in the hilly ridge area of Gydanskaya Gryada. The generally flat terrain found in the north of Yamal has almost a full absence of RTSs (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e). This might be due to the prevalence of very low elevations (\u0026lt;\u0026thinsp;20 m a.s.l.., see SI 1\u0026ndash;2), low VRM (1.65e-5, see SI 1\u0026ndash;2), and the dominance of the Dry Tundra landcover class. The latter is characterized by lower soil moisture and higher mineral component as well as less pronounced subsidence, suggesting limited ground ice melt and lower soil moisture dynamics (Bartsch et al., \u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e2024\u003c/span\u003e).\u003c/p\u003e\u003cp\u003eThe following parameters did not show any influence on RTS spatial distribution: 1) lake parameters such as area, perimeter, eccentricity, orientation, solidity; 2) the bearing of RTS to the nearest lake; 3) the bearing of RTS to the nearest river; 4) Quaternary geology composition based on the State Geological Map of the Russian Federation; 5) ground ice characteristics based on historical Soviet permafrost maps. We describe in detail the datasets, their limitations, and the conducted tests in SI 1\u0026ndash;1.\u003c/p\u003e\u003c/div\u003e\n\u003ch3\u003eRTS dynamics and association with climate variables\u003c/h3\u003e\n\u003cp\u003eTo study the dynamics of RTSs from the 1960s to 2024, we analyzed the historical satellite imagery of CORONA (Grosse et al., \u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e2005\u003c/span\u003e) and HEXAGON (Hammer et al., \u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e2022\u003c/span\u003e) as well as modern satellite imagery of PlanetScope 2024 (Planet Team, \u003cspan citationid=\"CR54\" class=\"CitationRef\"\u003e2018\u003c/span\u003e). We mapped and classified not only RTSs from the inventory (Nesterova et al., \u003cspan citationid=\"CR46\" class=\"CitationRef\"\u003e2025\u003c/span\u003e), but also identified additional RTSs (53 RTSs on historical imagery and 78 new RTSs in 2024) that were missed. RTSs as features were classified as \u0026ldquo;RTS\u0026rdquo; (visually clearly identified RTS with no doubts), \u0026ldquo;disturbance feature\u0026rdquo; (uncertain on the type of permafrost disturbance), \u0026ldquo;undisturbed tundra\u0026rdquo;, \u0026ldquo;stabilized disturbance\u0026rdquo; (in historical imagery visible as previously tundra disturbance or RTS, currently no visible disturbance), and, in PlanetScope 2024 imagery, as \u0026ldquo;active\u0026rdquo; or \u0026ldquo;stabilized\u0026rdquo;. For more details on mapping and classification methodology, please see SI 2\u0026ndash;1.6-1.8.\u003c/p\u003e\u003cp\u003eRTSs were mapped and classified for available years (1964, 1969, 1972, 1977, 1982, 1984) at 4 key sites: 3 in Gydan (G1, G2, G3) and 1 (Y1) in Yamal (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e). Key sites in Gydan covered\u0026thinsp;~\u0026thinsp;4,489 km\u003csup\u003e2\u003c/sup\u003e (~\u0026thinsp;18% of RTS clusters in Gydan) and the key site in Yamal covered\u0026thinsp;~\u0026thinsp;1,614 km\u003csup\u003e2\u003c/sup\u003e (~\u0026thinsp;10% of RTS clusters in Yamal). The key sites were selected based on the availability of historical imagery for the 1960s and 1970s. For an additional key site (Y2), historical imagery was only available for one time period. Hence, we include the results of its RTS dynamics only in the Extended Results in SI 1\u0026ndash;4. RTS mapping and classification approaches are described in detail in Extended Methods in SI 2\u0026ndash;1.6-1.8. We used manual mapping to resolve RTS occurrence and dynamics. We estimated the uncertainty range associated with the operator subjectivity at 4.1%, based on previous work with the West Siberian RTS inventory (Nesterova et al., \u003cspan citationid=\"CR46\" class=\"CitationRef\"\u003e2025\u003c/span\u003e).\u003c/p\u003e\u003cp\u003eWe found a general increase in the number of clearly detected RTS by 23-fold (2,338%) in total for all key sites (~\u0026thinsp;6103 km\u003csup\u003e2\u003c/sup\u003e) from the time period of 1964\u0026ndash;1972 to 2024 (Figs.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e and \u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003e). Most of the increase in RTS number occurred after 1984, contributing 86% to the total increase (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e).\u003c/p\u003e\u003cp\u003eInitiation rates of RTS increased over time, with the first large increase taking place as early as the 1980s and the largest increase in recent years reaching up to ~\u0026thinsp;78 RTSs yr\u003csup\u003e\u0026minus;\u0026thinsp;1\u003c/sup\u003e (26-fold increase). We have not normalized initiation rates per area due to the arbitrary key site boundaries that were selected based on data availability. Thus, key sites may include areas with no RTSs at all. Then, normalization by their areas can be misleading. Our finding of RTS initiation rates of ⋜3 RTSs yr\u003csup\u003e\u0026minus;\u0026thinsp;1\u003c/sup\u003e in the 1960s and ~\u0026thinsp;8.8 RTSs yr\u003csup\u003e\u0026minus;\u0026thinsp;1\u003c/sup\u003e for 1972\u0026ndash;1982 suggests that the initiation rates were comparable with the ones reported for Canada. The reported RTS initiation rate at the Aklavik Plateau (360 km\u003csup\u003e2\u003c/sup\u003e, a\u0026thinsp;\u0026gt;\u0026thinsp;16 times smaller study area) was 0.35 RTS yr\u003csup\u003e\u0026minus;\u0026thinsp;1\u003c/sup\u003e for the 1954\u0026ndash;1971 time period and 0.68 RTSs yr\u003csup\u003e\u0026minus;\u0026thinsp;1\u003c/sup\u003e for 1985\u0026ndash;2004 (Lacelle et al., \u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e2010\u003c/span\u003e). Meanwhile, modern rates of RTS initiation of ~\u0026thinsp;78 RTSs yr\u003csup\u003e\u0026minus;\u0026thinsp;1\u003c/sup\u003e for our area are comparable with those reported for current RTS hotspots. This includes, for instance, the neighboring Taymyr Peninsula, where one study in North Taymyr (68,000 km\u003csup\u003e2\u003c/sup\u003e, a\u0026thinsp;\u0026gt;\u0026thinsp;10 times larger study area) reported\u0026thinsp;~\u0026thinsp;120 RTS yr\u003csup\u003e\u0026minus;\u0026thinsp;1\u003c/sup\u003e for 2010\u0026ndash;2021 (Bernhard et al., \u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e2022\u003c/span\u003eb). Another study focused only on the extremely active area of East Taymyr (91 km\u003csup\u003e2\u003c/sup\u003e) and identified rates of ~\u0026thinsp;23 RTS yr\u003csup\u003e\u0026minus;\u0026thinsp;1\u003c/sup\u003e for the 2011\u0026ndash;2020 period (Barth et al., \u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e2025\u003c/span\u003e). Comparable high initiation rates were reported for Banks Island (~\u0026thinsp;70,000 km\u003csup\u003e2\u003c/sup\u003e, a\u0026thinsp;\u0026gt;\u0026thinsp;10 times larger study area) in Canada from 1984 to 2013, with values of 136 RTS yr\u003csup\u003e\u0026minus;\u0026thinsp;1\u003c/sup\u003e (Lewkowicz and Way, \u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e2019\u003c/span\u003e).\u003c/p\u003e\u003cp\u003eWe found that headwall retreat rates did not change much over time. The median headwall retreat rates were 1.3 m yr\u003csup\u003e\u0026minus;\u0026thinsp;1\u003c/sup\u003e for 1964\u0026ndash;1982/84 and 1.6 m yr\u003csup\u003e\u0026minus;\u0026thinsp;1\u003c/sup\u003e for 1982/84-2024, and the highest headwall retreat rate was 11.4 m yr\u003csup\u003e\u0026minus;\u0026thinsp;1\u003c/sup\u003e during the 1964\u0026ndash;1982/1984 time period (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e). These headwall retreat rates are similar to those reported for Tibetan RTSs (0.05-5 m yr\u003csup\u003e\u0026minus;\u0026thinsp;1\u003c/sup\u003e; Sun et al., \u003cspan citationid=\"CR62\" class=\"CitationRef\"\u003e2017\u003c/span\u003e; Fan et al., 2025). West Siberian RTS headwall retreat rates are rather stable and slow, especially if compared to the 6.2 m yr\u003csup\u003e\u0026minus;\u0026thinsp;1\u003c/sup\u003e estimated in the Canadian High Arctic (Ward Jones et al., \u003cspan citationid=\"CR67\" class=\"CitationRef\"\u003e2019\u003c/span\u003e) or in the neighboring Yugorsky peninsula with 66 m yr\u003csup\u003e\u0026minus;\u0026thinsp;1\u003c/sup\u003e (Leibman et al., \u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e2021\u003c/span\u003e). The maximum headwall retreats were 256 m (1972\u0026ndash;2024) in Y2 (SI 1\u0026ndash;4) and 230 m (1984\u0026ndash;2024) in key site G1 (SI 1\u0026ndash;4).\u003c/p\u003e\u003cp\u003eThe analysis of the RTS development trajectory (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003ea) aimed to define whether RTSs found in recent imagery are located in previously disturbed areas in the 1960s and 1980s. About 42% of all RTSs were found in areas of undisturbed tundra both in 1962\u0026ndash;1972 and 1982\u0026ndash;1984 time periods, while 33% were found as disturbance features in the 1980s, and 23% were disturbance features in the 1960s (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e). We show that a substantial part of RTSs occurred in pre-disturbed conditions back in the 1960s, which is similar to findings from the Canadian Arctic (Aklavik Plateau and Yukon Coast), where RTS locations were already observed in the 1950s (Lacelle et al., \u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e2010\u003c/span\u003e; Ramage et al., 2018).\u003c/p\u003e\u003cp\u003eFigure \u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003e shows the increase in the number of mapped RTSs for the time periods of 1964\u0026ndash;1972, 1982\u0026ndash;1984, and 2019\u0026ndash;2024. The climate data suggests an increase both in winter (October-April) and summer (May-September) air temperatures, with the winter air temperature warming at a much faster rate. There is also a small yet significant increase in winter precipitation.\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003eTo study the sensitivity of RTS initiation to climate variables, we employed a discrete-time proportional-hazards model (Suresh et al., 2023). The model assesses the risk of RTS initiation, assumed to come only from climate covariates. The results for the first time period (1964\u0026ndash;1984) and the second time period (2019\u0026ndash;2024) suggest that summer maximum precipitation and winter mean air temperature are the most important factors for RTS initiation. An increase in summer maximum precipitation by one standard deviation raised the annual RTS initiation probability by 27.62 percentage points (pp) (95% Bayesian uncertainty interval 25.93\u0026ndash;28.41) with a hazard ratio (HR) of 3.9 (3.38\u0026ndash;4.22), making it the strongest short-term influencing factor. Summer rainfalls are reported to associate with thaw-related mass movements not only for West Siberia (Leibman et al., \u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e2003\u003c/span\u003e) but also for the Canadian Arctic (Kokelj et al., \u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e2015\u003c/span\u003e; Lewkowicz and Way, \u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e2019\u003c/span\u003e) and Alaska (Balser et al., \u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e2014\u003c/span\u003e).\u003c/p\u003e\u003cp\u003e\u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab1\" border=\"1\"\u003e\u003ccaption language=\"En\"\u003e\u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e\u003cdiv class=\"CaptionContent\"\u003e\u003cp\u003e\u003cb\u003eEffects on annual initiation for a\u0026thinsp;+\u0026thinsp;1 standard deviation increase (risk-set AMEs* in pp, with 95% HDIs*; HRs* as medians with 95% intervals). Bayes factors are Savage\u0026ndash;Dickey BF\u003c/b\u003e\u003csub\u003e\u003cb\u003e10\u003c/b\u003e\u003c/sub\u003e\u003cb\u003e*. For two time periods as well as for long-term analysis.\u003c/b\u003e\u003c/p\u003e\u003c/div\u003e\u003c/caption\u003e\u003ccolgroup cols=\"6\"\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e\u003cthead\u003e\u003ctr\u003e\u003cth align=\"left\" colspan=\"6\" nameend=\"c6\" namest=\"c1\"\u003e\u003cp\u003e1964\u0026ndash;1984 and 2019\u0026ndash;2024 time periods\u003c/p\u003e\u003c/th\u003e\u003c/tr\u003e\u003c/thead\u003e\u003ctbody\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eCovariate\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c3\" namest=\"c2\"\u003e\u003cp\u003eΔ\u0026#120574;* (pp)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c5\" namest=\"c4\"\u003e\u003cp\u003eHR (median [95%])\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003eBF\u003csub\u003e10\u003c/sub\u003e\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eWinter mean air temperature\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c3\" namest=\"c2\"\u003e\u003cp\u003e+\u0026thinsp;21.25 [19.34, 23.18]\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c5\" namest=\"c4\"\u003e\u003cp\u003e3.11 [2.82, 3.46]\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e≫ 10\u003csup\u003e12\u003c/sup\u003e\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eSummer maximum precipitation\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c3\" namest=\"c2\"\u003e\u003cp\u003e+\u0026thinsp;27.17 [25.93, 28.41]\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c5\" namest=\"c4\"\u003e\u003cp\u003e3.80 [3.38, 4.22]\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e≫ 10\u003csup\u003e12\u003c/sup\u003e\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eWinter maximum precipitation\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c3\" namest=\"c2\"\u003e\u003cp\u003e+\u0026thinsp;0.76 [0.08, 1.63]\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c5\" namest=\"c4\"\u003e\u003cp\u003e1.02 [1.00, 1.06]\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e0.07\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eSummer mean air temperature\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c3\" namest=\"c2\"\u003e\u003cp\u003e+\u0026thinsp;0.40 [0.00, 1.14]\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c5\" namest=\"c4\"\u003e\u003cp\u003e1.02 [1.00, 1.06]\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e0.08\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eMemory: winter maximum precipitation (2 years)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c3\" namest=\"c2\"\u003e\u003cp\u003e+\u0026thinsp;3.13 [0.92, 5.10]\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c5\" namest=\"c4\"\u003e\u003cp\u003e1.23 [1.09, 1.37]\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e57.68\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eMemory: winter mean air temperature (2 years)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c3\" namest=\"c2\"\u003e\u003cp\u003e+\u0026thinsp;0.52 [0.00, 1.43]\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c5\" namest=\"c4\"\u003e\u003cp\u003e1.03 [1.00, 1.08]\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e0.17\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colspan=\"6\" nameend=\"c6\" namest=\"c1\"\u003e\u003cp\u003eLong-term period 1964\u0026ndash;1984 \u0026rarr; 2019\u0026ndash;2024\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e\u003cp\u003eCovariate\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c4\" namest=\"c3\"\u003e\u003cp\u003e1964\u0026ndash;1984 \u0026rarr;2019\u0026ndash;2024 change (natural; SD*)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c6\" namest=\"c5\"\u003e\u003cp\u003eImplied Δ\u0026#120574; (pp, linear)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e\u003cp\u003eWinter mean air temperature\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c4\" namest=\"c3\"\u003e\u003cp\u003e+\u0026thinsp;4.82 K* (+\u0026thinsp;1.62 SD)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c6\" namest=\"c5\"\u003e\u003cp\u003e+\u0026thinsp;33.56\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e\u003cp\u003eSummer maximum precipitation\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c4\" namest=\"c3\"\u003e\u003cp\u003e-0.05 K (-0.08 SD)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c6\" namest=\"c5\"\u003e\u003cp\u003e-2.2\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e\u003cp\u003eWinter maximum precipitation\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c4\" namest=\"c3\"\u003e\u003cp\u003e+\u0026thinsp;0.75 K (+\u0026thinsp;1.58 SD)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c6\" namest=\"c5\"\u003e\u003cp\u003e+\u0026thinsp;1.43\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e\u003cp\u003eSummer mean air temperature\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c4\" namest=\"c3\"\u003e\u003cp\u003e+\u0026thinsp;3.17 K (+\u0026thinsp;2.28 SD)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c6\" namest=\"c5\"\u003e\u003cp\u003e+\u0026thinsp;0.9\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003c/tbody\u003e\u003c/colgroup\u003e\u003c/table\u003e\u003c/div\u003e\u003c/p\u003e\u003cp\u003e\u003cem\u003e*Abbreviations and definitions: AME - Average marginal effect, an absolute change in the annual initiation probability; HDI - Highest density interval, a Bayesian uncertainty interval; HR - hazard ratio, the ratio of the hazard rates; Savage\u0026ndash;Dickey BF\u003c/em\u003e\u003csub\u003e\u003cem\u003e10\u003c/em\u003e\u003c/sub\u003e \u003cem\u003e- Bayes factor (BF) - evidence estimate, BF\u0026thinsp;\u0026gt;\u0026thinsp;10 indicates a strong evidence; SD - standard deviation; pp - percentage points, an absolute change on the probability scale; Δ\u0026#120574; - the change in RTS initiation probability; K - Kelvin.\u003c/em\u003e\u003c/p\u003e\u003cp\u003eThe increase in one standard deviation of winter mean air temperature raises the probability of annual RTS initiation by 21.25 pp (19.34\u0026ndash;23.18) with an HR of 3.11 (2.82\u0026ndash;3.46), indicating strong preconditioning. Winter maximum precipitation over the two preceding years (\u0026ldquo;memory\u0026rdquo; climate variable that stores the cumulative information of the last two winters) has a rather small but still strongly evident effect on RTS initiation probability by adding 3.13 pp (0.92\u0026ndash;5.10) with one standard deviation increase. Other climatic variables have rather small effects or evidence.\u003c/p\u003e\u003cp\u003eOur results suggest that winter mean air temperature is a dominant long-term driver of RTS initiation. The increase in winter mean air temperature by 1.62 standard deviations leads to a rise in annual RTS initiation probability over the long term by 37.4 pp. The other analysed climate variables were found to be of weak evidence (Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e).\u003c/p\u003e\u003cp\u003eMultiple studies reported warm summer air temperature being associated with RTS development (Lacelle et al., \u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e2010\u003c/span\u003e; Balser et al., \u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e2014\u003c/span\u003e; Swanson and Nolan, \u003cspan citationid=\"CR64\" class=\"CitationRef\"\u003e2018\u003c/span\u003e; Lewkowicz and Way, \u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e2019\u003c/span\u003e; Ward Jones et al., \u003cspan citationid=\"CR67\" class=\"CitationRef\"\u003e2019\u003c/span\u003e). Our findings suggest that the mean winter air temperature is not only a significant factor influencing RTS initiation over short time periods, but also is a long-term driver. Winter air temperatures in the West Siberian Arctic increased much faster than in summer (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003e), which is consistent with global Arctic winter warming that exceeds summer warming by at least four times (Bintanja and van der Linden, \u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e2013\u003c/span\u003e). Warmer winters imply shallower freezing of the ground and thus, thickening of the active layer, which was reported to deepen in West Siberia over the last 50 years (Vasiliev et al., \u003cspan citationid=\"CR66\" class=\"CitationRef\"\u003e2020\u003c/span\u003e). A slight but yet significant increase in mean winter precipitation (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003e), as well as a small but still evident effect of higher maximum winter precipitation for the preceding two years (\u0026ldquo;memory\u0026rdquo; climate variable) on RTS initiation found in our model, suggest that winter precipitation in the form of snow may also contribute to the thermal insulation of permafrost from cold air temperatures in winter and thus its warming. Warming of permafrost as well as deepening of the active layer thickness were reported to activate thaw-related mass movements in Central Yamal (Babkina et al., \u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e2019\u003c/span\u003e).\u003c/p\u003e\n\u003ch3\u003ePotential for future RTS development and infrastructure risks\u003c/h3\u003e\n\u003cp\u003eBased on our findings, we anticipate more RTSs to occur in the West Siberian Arctic with ongoing climate warming and deepening of the active layer in the future. To identify areas of possible future RTS formation in the region, we analyzed which areas share the same environmental parameters as current clusters. We then filtered only areas with at least one RTS already mapped in Nesterova et al. (\u003cspan citationid=\"CR46\" class=\"CitationRef\"\u003e2025\u003c/span\u003e), which indicates the presence of massive ground ice. Next, we applied filters of the 95th percentile of the value range for the parameters that were found to be significantly different in the clusters compared to the rest of the peninsula (see Results - Spatial distribution). For the Yamal Peninsula, the filtering was performed by mean elevation and proportion of landcover classes, while for the Gydan Peninsula, in addition to these two factors, VRM and lake density were also included.\u003c/p\u003e\u003cp\u003eAs a result, the potential future RTS cluster areas for Yamal comprised 3,092.4 km\u0026sup2;, or 2.34% of the entire Yamal Peninsula (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003e), which would account for an increase of 43.19% of highly RTS-impacted areas. The potential future RTS cluster areas for Gydan comprised 4,508.3 km\u0026sup2;, or 2.59% of the entire Gydan Peninsula (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003e), representing an increase of 32.16% of highly RTS-impacted areas.\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003eThaw-related mass wasting processes across the Arctic were reported to pose a risk to the infrastructure stability (Luo et al., \u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e2019\u003c/span\u003e; Hjort et al., \u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e2022\u003c/span\u003e). To estimate the potential risk from RTS development to infrastructure, we analyzed the infrastructure that is currently impacted by RTS clusters, as well as where infrastructure overlaps with areas that may develop into highly RTS-impacted areas in the future. The infrastructure in Yamal was found to be at much higher risk compared to Gydan. The currently largest area-wise affected infrastructure in Yamal belongs to the gas and oil industry, comprising\u0026thinsp;~\u0026thinsp;5.3 km\u0026sup2; of affected area. This includes the Bovanenkovo gas field settlement, including the airport with a runway, as well as the Bovanenkovo-Ukhta pipeline (410 km, which is ~\u0026thinsp;37% of total length). Around ~\u0026thinsp;2.5 km\u0026sup2; of transport infrastructure, such as roads, bridges, and tunnels, is also affected, mostly within the Bovanenkovo gas field settlement. Around 66% of all power \u0026amp; energy infrastructure (49 km of length) within Bovanenkovo was found at risk of damage, possibly posing substantial environmental hazards. The potential future RTS clustering could lead to further risk for the Obskaya-Bovanenkovo-Karskaya railway line, which could increase by more than twice from 19 km of the currently affected length to 52 km. Oil and gas, as well as transport infrastructure with potential RTS clusters, can experience an increase of 15 km in length. Unlike in Yamal, the infrastructure of the Gydan peninsula is currently not affected. Gydan\u0026rsquo;s largest Salmanovskoye gas and oil field, including the airport with a runway, is in the vicinity of the potential RTS clusters. Potential future RTS-impacted clusters could also develop in Tadebya-Yakha, a small settlement on the western coast of Gydan.\u003c/p\u003e\u003cp\u003eThe majority (\u0026gt;\u0026thinsp;75%) of all RTSs from the inventory are located on lake shores (Nesterova et al., \u003cspan citationid=\"CR46\" class=\"CitationRef\"\u003e2025\u003c/span\u003e). We found that more than ~\u0026thinsp;2400 lakes in the West Siberian Arctic are currently affected by RTSs. As RTS development affects not only hydrology, but also the water quality (Kokelj et al., \u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e2005\u003c/span\u003e; Kokelj \u0026amp; Jorgenson, \u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e2013\u003c/span\u003e), this correlation between RTS and lakes may pose substantial risks to aquatic habitats and local indigenous communities, which requires further detailed investigation in future research.\u003c/p\u003e"},{"header":"Methods","content":"\u003cdiv id=\"Sec7\" class=\"Section2\"\u003e\u003ch2\u003eSpatial analysis\u003c/h2\u003e\u003cp\u003eTo perform all geospatial operations as well as map layout, we used Python language v.3.11.10 in Visual Studio Code v. 1.103.2 and QGIS v. 3.36.2-Maidenhead. All geospatial operations and mapping were performed in WGS84 / UTM zone 43 projection (EPSG:32643). To investigate the spatial distribution of RTSs as well as environmental parameters (elevation, lake density, landcover) within the study area, we aggregated our data into Uber H3 hexagonal grid with a resolution of 6, which is ~\u0026thinsp;35 km\u0026sup2; (Uber Technologies Inc., 2018). This grid system was chosen since it is widely used (Aini et al., \u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e2023\u003c/span\u003e; Nitze et al., \u003cspan citationid=\"CR48\" class=\"CitationRef\"\u003e2025\u003c/span\u003e), well-integrated in Python, and fully reproducible due to its set boundaries.\u003c/p\u003e\u003cp\u003eTo estimate elevation, we used the ArcticDEM dataset (Porter et al., \u003cspan citationid=\"CR56\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). To assess lake density, we created a dataset based on OpenStreetMap dataset (\u0026copy;OpenStreetMap contributors, ODbL 1.0, website, accessed 08.08.2025). To analyze a landcover, we used the circumarctic land cover (Bartsch et al., \u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). The infrastructure dataset was also accessed via OSM. RTS polygons for the 1960s and 1980s were digitized based on georeferenced CORONA and HEXAGON imagery (Grosse et al., \u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e2005\u003c/span\u003e; Hammer et al., \u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e2022\u003c/span\u003e). RTS polygons for 2024 were created based on PlanetScope imagery (Planet Team, \u003cspan citationid=\"CR54\" class=\"CitationRef\"\u003e2018\u003c/span\u003e). The detailed description of all datasets used in this study, the limitations, and all the pre-processing steps applied are described in detail in the SI 2\u0026ndash;1.1-1.4.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec8\" class=\"Section2\"\u003e\u003ch2\u003eStatistical Analysis\u003c/h2\u003e\u003cp\u003eClusters of RTS accumulation were identified using Anselin Local Moran\u0026rsquo;s I statistic (Moran, \u003cspan citationid=\"CR44\" class=\"CitationRef\"\u003e1950\u003c/span\u003e), with Queen\u0026rsquo;s Case (Lloyd, \u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e2010\u003c/span\u003e) employed as the conceptualization of spatial relationships to account for adjacency effects. To select a test of hypothesis testing we first estimated the normality of analyzed distributions of mean elevations, VRM and lake density using multiple tests such as Shapiro-Wilk test (Shapiro and Wilk, \u003cspan citationid=\"CR58\" class=\"CitationRef\"\u003e1965\u003c/span\u003e), D\u0026rsquo;Agostino\u0026rsquo;s K\u0026sup2; test combines skewness and kurtosis (D\u0026rsquo;Agostino et al., 1990), and two modifications of Kolmogorov-Smirnov test: Anderson\u0026ndash;Darling test that puts more weight to tails (Anderson and Darling, \u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e1954\u003c/span\u003e), Lilliefors test that accounts for the fact that the mean and variance are estimated from the sample (Lilliefors, \u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e1967\u003c/span\u003e). All of the tests eventually showed that distributions of all parameters, both for all the peninsula and for clusters, were not normal; thus, for the hypothesis testing, we chose a non-parametric test. To get complementary evidence, we applied two non-parametric tests: the Mann-Whitney U-Test, which compares central tendencies through differences in medians (Mann and Whitney, 1947), and the Kolmogorov-Smirnov Test that compares overall distribution differences, including shape, spread, and medians of the data (Massey, \u003cspan citationid=\"CR42\" class=\"CitationRef\"\u003e1951\u003c/span\u003e).\u003c/p\u003e\u003c/div\u003e\n\u003ch3\u003eBayesian climate hazard modelling of retrogressive thaw slump initiation\u003c/h3\u003e\n\u003cp\u003eWe analyzed the climate sensitivity of retrogressive thaw slump (RTS) initiation using a discrete-time proportional hazards model (Allison, \u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e1982\u003c/span\u003e). Annual initiation hazards were linked to covariates through a complementary log\u0026ndash;log function (Allison, \u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e1982\u003c/span\u003e). The linear predictor included four standardized climate covariates: thawing degree days, summer maximum precipitation, winter mean temperature, and winter maximum precipitation. Coefficients expected to increase risk were constrained to be non-negative with HalfNormal priors (Gelman et al., \u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e2008\u003c/span\u003e). Models were estimated in a Bayesian framework with Hamiltonian Monte Carlo (Betancourt, \u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e2018\u003c/span\u003e). Prior predictive simulations confirmed plausibility, and posterior predictive checks compared yearly onset counts with observations (Auger-M\u0026eacute;th\u0026eacute; et al., 2019). Effect sizes are reported as hazard ratios (HR) and as average marginal effects (AME), expressed as percentage-point changes in annual initiation probability averaged across the pre-initiation risk set. For more detailed settings, see SI 2\u0026ndash;2.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003ch2\u003eCompeting financial interest\u003c/h2\u003e\u003cp\u003eThe authors declare no competing financial interests.\u003c/p\u003e\u003c/p\u003e\u003cp\u003e\u003ch2\u003eMaterials \u0026amp; Correspondence.\u003c/h2\u003e\u003cp\u003eCorrespondence and material requests should be addressed to NN ([email protected]).\u003c/p\u003e\u003c/p\u003e\u003ch2\u003eFunding\u003c/h2\u003e\u003cp\u003eWe acknowledge support for this study by DAAD to NN (\u0026ldquo;STIBET-I\u0026rdquo;). NN, GG, IN, and TH were supported by the BMWK project ML4EARTH and the google.org Impact Challenge on Climate Innovation to the Permafrost Discovery Gateway development team. IN and GG were additionally supported by the NSF Navigating the New Arctic Permafrost Discovery Gateway (#1927872 #2052107). ML was supported by the state assignment of the Ministry of Science and Higher Education of the Russian Federation (Project No. FWRZ-2021-0012). HL has received funding from the EU Horizon Europe, grant agreement No. 101133587 (ILLUQ). CS was supported by HEIBRiDS.\u003c/p\u003e\u003cp\u003ePlanetScope data were provided through the NASA CSDA program (for NSF affiliated research on the Permafrost Discovery Gateway) and the Planet Research Program.\u003c/p\u003e\u003ch2\u003eAuthor Contribution\u003c/h2\u003e\u003cp\u003eConceptualization - NN, ML, GG, IN; Methodology - NN, CS; Software -NN, TH; Validation - NN, MV; Formal analysis - NN, CS; Investigation - NN, CS; Data Curation - TH, MV, KM, IT; Writing - Original Draft - NN, CS; Writing - Review \u0026amp; Editing - all authors; Visualization - NN, CS; Supervision - ML, GG, IN, HL.\u003c/p\u003e\u003ch2\u003eAcknowledgement\u003c/h2\u003e\u003cp\u003eThe first author would like to thank Rustam Khairullin (b-geos) for processing the missing part of the landcover, and Chiara Philips for the design recommendations.\u003c/p\u003e\u003ch2\u003eData Availability\u003c/h2\u003e\u003cp\u003eThe datasets that were generated during the study are currently available from the corresponding author upon request to the reviewers and the editors, but will be publicly available on PANGAEA upon publication of the study.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eAini, A. N., Dewantari, O. A., Mandala, D. P. A. \u0026amp; Bisri, M. B. F. An Enhanced Earthquake Risk Analysis using H3 Spatial Indexing. IOP Conference Series: Earth and Environmental Science. 1245(1):012014. (2023). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1088/1755-1315/1245/1/012014\u003c/span\u003e\u003cspan address=\"10.1088/1755-1315/1245/1/012014\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eAllison, P. D. Discrete-Time Methods for the Analysis of Event Histories. \u003cem\u003eSociol. Methodol.\u003c/em\u003e \u003cb\u003e13\u003c/b\u003e, 61\u0026ndash;98. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.2307/270718\u003c/span\u003e\u003cspan address=\"10.2307/270718\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (1982).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eAnderson, T. W. \u0026amp; Darling, D. A. A Test of Goodness of Fit. \u003cem\u003eJ. Am. Stat. Assoc.\u003c/em\u003e \u003cb\u003e49\u003c/b\u003e (268), 765\u0026ndash;769. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.2307/2281537\u003c/span\u003e\u003cspan address=\"10.2307/2281537\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (1954).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eAuger-M\u0026eacute;th\u0026eacute;, M. et al. A guide to state\u0026ndash;space modeling of ecological time series. \u003cem\u003eEcol. Monogr.\u003c/em\u003e \u003cb\u003e91\u003c/b\u003e (4), e01470 (2021).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eBabkina, E. A. et al. Activation of Cryogenic Processes in Central Yamal as a Result of Regional and Local Change in Climate and Thermal State of Permafrost. \u003cem\u003eRuss. Meteorol. Hydrol.\u003c/em\u003e \u003cb\u003e44\u003c/b\u003e (4), 283\u0026ndash;290. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.3103/S1068373919040083\u003c/span\u003e\u003cspan address=\"10.3103/S1068373919040083\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2019).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eBadu, Y. B. Ice content of cryogenic strata (permafrost interval) in gas-bearing structures, Northern Yamal. \u003cem\u003eEarth\u0026rsquo;s Cryosphere\u003c/em\u003e. \u003cb\u003e19\u003c/b\u003e (3), 9\u0026ndash;18 (2015).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eBalser, A. W., Jones, J. B. \u0026amp; Gens, R. Timing of retrogressive thaw slump initiation in the Noatak Basin, northwest Alaska, USA. \u003cem\u003eJ. Geophys. Research: Earth Surf.\u003c/em\u003e \u003cb\u003e119\u003c/b\u003e (5), 1106\u0026ndash;1120. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1002/2013JF002889\u003c/span\u003e\u003cspan address=\"10.1002/2013JF002889\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2014).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eBarth, S. et al. Rapid Changes in Retrogressive Thaw Slump Dynamics in the Russian High Arctic Based on Very High-Resolution Remote Sensing. \u003cem\u003eGeophys. Res. Lett.\u003c/em\u003e \u003cb\u003e52\u003c/b\u003e (7). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1029/2024GL113022\u003c/span\u003e\u003cspan address=\"10.1029/2024GL113022\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2025). e2024GL113022.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eBartsch, A. et al. Circumarctic land cover diversity considering wetness gradients. \u003cem\u003eHydrol. Earth Syst. Sci.\u003c/em\u003e \u003cb\u003e28\u003c/b\u003e (11), 2421\u0026ndash;2481. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.5194/hess-28-2421-2024\u003c/span\u003e\u003cspan address=\"10.5194/hess-28-2421-2024\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2024).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eBaulin, V. V., Belopukhova, E. B., Dubikov, G. I. \u0026amp; Shmelev, L. M. \u003cem\u003eGeocryological conditions of western Siberia Lowland\u003c/em\u003e (Nauka, 1967).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eBernhard, P., Zwieback, S. \u0026amp; Hajnsek, I. Accelerated Mobilization of Organic Carbon from Retrogressive Thaw Slumps on the Northern Taymyr Peninsula. \u003cem\u003eCryosphere Discuss.\u003c/em\u003e \u003cb\u003e2022\u003c/b\u003e, 1\u0026ndash;24. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.5194/tc-16-2819-2022\u003c/span\u003e\u003cspan address=\"10.5194/tc-16-2819-2022\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2022).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eBernhard Philipp, Zwieback, S., Bergner, N. \u0026amp; Hajnsek, I. Assessing volumetric change distributions and scaling relations of retrogressive thaw slumps across the Arctic. \u003cem\u003eCryosphere\u003c/em\u003e \u003cb\u003e16\u003c/b\u003e (1), 1\u0026ndash;15. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.5194/tc-16-1-2022\u003c/span\u003e\u003cspan address=\"10.5194/tc-16-1-2022\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2022).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eBetancourt, M. A Conceptual Introduction to Hamiltonian Monte Carlo. [accessed 2025 Sept 9]. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttp://arxiv.org/abs/1701.02434\u003c/span\u003e\u003cspan address=\"http://arxiv.org/abs/1701.02434\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. (2018). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.48550/arXiv.1701.02434\u003c/span\u003e\u003cspan address=\"10.48550/arXiv.1701.02434\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eBintanja, R. \u0026amp; van der Linden, E. C. The changing seasonal climate in the Arctic. \u003cem\u003eSci. Rep.\u003c/em\u003e \u003cb\u003e3\u003c/b\u003e (1), 1556. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1038/srep01556\u003c/span\u003e\u003cspan address=\"10.1038/srep01556\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2013).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eBiskaborn, B. K. et al. Permafrost is warming at a global scale. \u003cem\u003eNat. Commun.\u003c/em\u003e \u003cb\u003e10\u003c/b\u003e (1), 264. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1038/s41467-018-08240-4\u003c/span\u003e\u003cspan address=\"10.1038/s41467-018-08240-4\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2019).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eBradley, J. A. et al. Svalbard winter warming is reaching melting point. \u003cem\u003eNat. Commun.\u003c/em\u003e \u003cb\u003e16\u003c/b\u003e (1), 6409. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1038/s41467-025-60926-8\u003c/span\u003e\u003cspan address=\"10.1038/s41467-025-60926-8\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2025).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eD\u0026rsquo;Agostino, R. B. \u0026amp; Belanger, A. A Suggestion for Using Powerful and Informative Tests of Normality. \u003cem\u003eAm. Stat.\u003c/em\u003e \u003cb\u003e44\u003c/b\u003e (4), 316\u0026ndash;321. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.2307/2684359\u003c/span\u003e\u003cspan address=\"10.2307/2684359\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (1990).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eFarquharson, L. M., Romanovsky, V. E., Kholodov, A. \u0026amp; Nicolsky, D. Sub-aerial talik formation observed across the discontinuous permafrost zone of Alaska. \u003cem\u003eNat. Geosci.\u003c/em\u003e \u003cb\u003e15\u003c/b\u003e (6), 475\u0026ndash;481. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1038/s41561-022-00952-z\u003c/span\u003e\u003cspan address=\"10.1038/s41561-022-00952-z\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2022).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eGelman, A., Jakulin, A., Pittau, M. G. \u0026amp; Su, Y-S. A weakly informative default prior distribution for logistic and other regression models. \u003cem\u003eAnnals Appl. Stat.\u003c/em\u003e \u003cb\u003e2\u003c/b\u003e (4), 1360\u0026ndash;1383. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1214/08-AOAS191\u003c/span\u003e\u003cspan address=\"10.1214/08-AOAS191\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2008).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eGrosse, G., Schirrmeister, L., Kunitsky, V. V. \u0026amp; Hubberten, H-W. The use of CORONA images in remote sensing of periglacial geomorphology: an illustration from the NE Siberian coast. \u003cem\u003ePermafrost Periglac. Process.\u003c/em\u003e \u003cb\u003e16\u003c/b\u003e (2), 163\u0026ndash;172. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1002/ppp.509\u003c/span\u003e\u003cspan address=\"10.1002/ppp.509\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2005).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eHammer, E., FitzPatrick, M. \u0026amp; Ur, J. Succeeding CORONA: declassified HEXAGON intelligence imagery for archaeological and historical research. \u003cem\u003eAntiquity\u003c/em\u003e \u003cb\u003e96\u003c/b\u003e (387), 679\u0026ndash;695. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.15184/aqy.2022.22\u003c/span\u003e\u003cspan address=\"10.15184/aqy.2022.22\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2022).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eHjort, J. et al. Impacts of permafrost degradation on infrastructure. \u003cem\u003eNat. Reviews Earth Environ.\u003c/em\u003e \u003cb\u003e3\u003c/b\u003e (1), 24\u0026ndash;38. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1038/s43017-021-00247-8\u003c/span\u003e\u003cspan address=\"10.1038/s43017-021-00247-8\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2022).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eKhomutov, A. et al. Activation of Cryogenic Earth Flows and Formation of Thermocirques on Central Yamal as a Result of Climate Fluctuations. In: Mikoš K, Vil\u0026iacute;mek V, Yin Y, Sassa K., editors. Advancing Culture of Living with Landslides. Proceedings of World Landslide Forum 4, May 29 \u0026ndash; June 2, 2017, Ljubljana, Slovenia. Vol.5, Landslides in Different Environments. Springer International Publishing AG; p 209\u0026ndash;216 (2017).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eKokelj, S. V. et al. Increased precipitation drives mega slump development and destabilization of ice-rich permafrost terrain, northwestern Canada. \u003cem\u003eGlob. Planet Change\u003c/em\u003e. \u003cb\u003e129\u003c/b\u003e, 56\u0026ndash;68. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.gloplacha.2015.02.008\u003c/span\u003e\u003cspan address=\"10.1016/j.gloplacha.2015.02.008\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2015).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eKokelj, S. V. et al. The influence of thermokarst disturbance on the water quality of small upland lakes, Mackenzie Delta region, Northwest Territories, Canada. \u003cem\u003ePermafrost Periglac. Process.\u003c/em\u003e \u003cb\u003e16\u003c/b\u003e (4), 343\u0026ndash;353. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1002/ppp.536\u003c/span\u003e\u003cspan address=\"10.1002/ppp.536\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2005).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eKokelj, S. V. \u0026amp; Lewkowicz, A. G. \u003cem\u003eSalinization of permafrost terrain due to natural geomorphic disturbance, Fosheim Peninsula, Ellesmere Island\u003c/em\u003e372\u0026ndash;385 (Arctic, 1999).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eKokelj, S. V. \u0026amp; Jorgenson, M. T. Advances in thermokarst research. \u003cem\u003ePermafrost Periglac. Process.\u003c/em\u003e \u003cb\u003e24\u003c/b\u003e (2), 108\u0026ndash;119. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1002/ppp.1779\u003c/span\u003e\u003cspan address=\"10.1002/ppp.1779\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2013).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eKokelj, S. V. et al. Climate-driven thaw of permafrost preserved glacial landscapes, northwestern Canada. \u003cem\u003eGeology\u003c/em\u003e \u003cb\u003e45\u003c/b\u003e (4), 371\u0026ndash;374. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1130/G38626.1\u003c/span\u003e\u003cspan address=\"10.1130/G38626.1\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2017).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eLacelle, D., Bjornson, J. \u0026amp; Lauriol, B. Climatic and geomorphic factors affecting contemporary (1950\u0026ndash;2004) activity of retrogressive thaw slumps on the Aklavik Plateau, Richardson Mountains, NWT, Canada. \u003cem\u003ePermafrost Periglac. Process.\u003c/em\u003e \u003cb\u003e21\u003c/b\u003e (1), 1\u0026ndash;15. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1002/ppp.666\u003c/span\u003e\u003cspan address=\"10.1002/ppp.666\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2010).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eLantuit, H. \u0026amp; Pollard, W. H. Fifty years of coastal erosion and retrogressive thaw slump activity on Herschel Island, southern Beaufort Sea, Yukon Territory, Canada. \u003cem\u003eGeomorphology\u003c/em\u003e \u003cb\u003e95\u003c/b\u003e (1\u0026ndash;2), 84\u0026ndash;102. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.geomorph.2006.07.040\u003c/span\u003e\u003cspan address=\"10.1016/j.geomorph.2006.07.040\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2008).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eLeibman, M. O. \u0026amp; Kizyakov, A. I. \u003cem\u003eCryogenic landslides of the Yamal and Yugorsky Peninsulas (In Russian)\u003c/em\u003eVol 206 (Earth Cryosphere Institute SB RAS, 2007).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eLeibman, M. O., Kizyakov, A. I., Sulerzhitsky, L. D. \u0026amp; Zaretskaia, N. E. Dynamics of landslide slopes and their development on Yamal Peninsula. In: Permafrost. Proceedings of the 8th international conference on permafrost. Swets and Zeitlinger, Lisse. p 651\u0026ndash;656 (2003).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eLeibman, M. et al. Coastal Retreat Due to Thermodenudation on the Yugorsky Peninsula, Russia during the Last Decade, Update since 2001\u0026ndash;2010. \u003cem\u003eRemote Sens.\u003c/em\u003e \u003cb\u003e13\u003c/b\u003e (20), 4042. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.3390/rs13204042\u003c/span\u003e\u003cspan address=\"10.3390/rs13204042\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2021).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eLewkowicz, A. G. \u0026amp; Way, R. G. Extremes of summer climate trigger thousands of thermokarst landslides in a High Arctic environment. \u003cem\u003eNat. Commun.\u003c/em\u003e \u003cb\u003e10\u003c/b\u003e (1), 1329. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1038/s41467-019-09314-7\u003c/span\u003e\u003cspan address=\"10.1038/s41467-019-09314-7\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2019).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eLewkowicz, A. G. Retrogressive thaw slump activity in the western Canadian Arctic (1984\u0026ndash;2016). (2024).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eLilliefors, H. W. On the Kolmogorov-Smirnov Test for Normality with Mean and Variance Unknown. \u003cem\u003eJ. Am. Stat. Assoc.\u003c/em\u003e \u003cb\u003e62\u003c/b\u003e (318), 399\u0026ndash;402. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.2307/2283970\u003c/span\u003e\u003cspan address=\"10.2307/2283970\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (1967).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eLloyd, C. D. \u003cem\u003eSpatial data analysis: an introduction for GIS users\u003c/em\u003e (Oxford University Press, 2010).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eLuo, J. et al. Recent acceleration of thaw slumping in permafrost terrain of Qinghai-Tibet Plateau: An example from the Beiluhe Region. \u003cem\u003eGeomorphology\u003c/em\u003e \u003cb\u003e341\u003c/b\u003e, 79\u0026ndash;85. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.geomorph.2019.05.020\u003c/span\u003e\u003cspan address=\"10.1016/j.geomorph.2019.05.020\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2019).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eMackay, J. R. Segregated epigenetic ice and slumps in permafrost, Mackenzie Delta area, NWT. \u003cem\u003eGeographical Bull.\u003c/em\u003e \u003cb\u003e8\u003c/b\u003e, 59\u0026ndash;80 (1966).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eMaier, K. et al. Detecting mass wasting of Retrogressive Thaw Slumps in spaceborne elevation models using deep learning. \u003cem\u003eInt. J. Appl. Earth Obs. Geoinf.\u003c/em\u003e \u003cb\u003e137\u003c/b\u003e, 104419. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.jag.2025.104419\u003c/span\u003e\u003cspan address=\"10.1016/j.jag.2025.104419\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2025).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eMakopoulou, E. et al. Retrogressive thaw slump susceptibility in the northern hemisphere permafrost region. \u003cem\u003eEarth. Surf. Proc. Land.\u003c/em\u003e \u003cb\u003e49\u003c/b\u003e (11), 3319\u0026ndash;3331. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1002/esp.5890\u003c/span\u003e\u003cspan address=\"10.1002/esp.5890\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2024).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eMassey, F. J. The Kolmogorov-Smirnov Test for Goodness of Fit. \u003cem\u003eJ. Am. Stat. Assoc.\u003c/em\u003e \u003cb\u003e46\u003c/b\u003e (253), 68\u0026ndash;78. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.2307/2280095\u003c/span\u003e\u003cspan address=\"10.2307/2280095\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (1951).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eMesquita, P. S., Wrona, F. J. \u0026amp; Prowse, T. D. Effects of retrogressive permafrost thaw slumping on sediment chemistry and submerged macrophytes in Arctic tundra lakes. \u003cem\u003eFreshw. Biol.\u003c/em\u003e \u003cb\u003e55\u003c/b\u003e (11), 2347\u0026ndash;2358 (2010).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eMoran, P. A. P. Notes on Continuous Stochastic Phenomena. \u003cem\u003eBiometrika\u003c/em\u003e \u003cb\u003e37\u003c/b\u003e (1/2), 17\u0026ndash;23. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.2307/2332142\u003c/span\u003e\u003cspan address=\"10.2307/2332142\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (1950).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eNesterova, N. et al. Review article: Retrogressive thaw slump characteristics and terminology. \u003cem\u003eCryosphere\u003c/em\u003e \u003cb\u003e18\u003c/b\u003e (10), 4787\u0026ndash;4810. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.5194/tc-18-4787-2024\u003c/span\u003e\u003cspan address=\"10.5194/tc-18-4787-2024\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2024).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eNesterova, N. et al. High-resolution inventory and classification of retrogressive thaw slumps in West Siberia. Earth Syst Sci Data Discuss. 2025:1\u0026ndash;30. (2025). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.5194/essd-2025-164\u003c/span\u003e\u003cspan address=\"10.5194/essd-2025-164\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eNitze, I. et al. Remote sensing quantifies widespread abundance of permafrost region disturbances across the Arctic and Subarctic. \u003cem\u003eNat. Commun.\u003c/em\u003e \u003cb\u003e9\u003c/b\u003e (1). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1038/s41467-018-07663-3\u003c/span\u003e\u003cspan address=\"10.1038/s41467-018-07663-3\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2018).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eNitze, I. et al. DARTS: Multi-year database of AI-detected retrogressive thaw slumps in the circum-arctic permafrost region. \u003cem\u003eSci. Data\u003c/em\u003e. \u003cb\u003e12\u003c/b\u003e (1), 1512. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1038/s41597-025-05810-2\u003c/span\u003e\u003cspan address=\"10.1038/s41597-025-05810-2\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2025).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eNovikova, A. et al. Dynamics of Permafrost Coasts of Baydaratskaya Bay (Kara Sea) Based on Multi-Temporal Remote Sensing Data. \u003cem\u003eRemote Sens.\u003c/em\u003e \u003cb\u003e10\u003c/b\u003e (9). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.3390/rs10091481\u003c/span\u003e\u003cspan address=\"10.3390/rs10091481\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2018).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eO\u0026rsquo;Neill, H. B., Wolfe, S. A. \u0026amp; Duchesne, C. New ground ice maps for Canada using a paleogeographic modelling approach. \u003cem\u003eCryosphere\u003c/em\u003e \u003cb\u003e13\u003c/b\u003e (3), 753\u0026ndash;773. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.5194/tc-13-753-2019\u003c/span\u003e\u003cspan address=\"10.5194/tc-13-753-2019\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2019).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eObu, J. How Much of the Earth\u0026rsquo;s Surface is Underlain by Permafrost? \u003cem\u003eJ. Geophys. Research: Earth Surf.\u003c/em\u003e \u003cb\u003e126\u003c/b\u003e (5). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1029/2021JF006123\u003c/span\u003e\u003cspan address=\"10.1029/2021JF006123\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2021). e2021JF006123.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eObu, J. et al. Northern Hemisphere permafrost map based on TTOP modelling for 2000\u0026ndash;2016 at 1 km2 scale. \u003cem\u003eEarth Sci. Rev.\u003c/em\u003e \u003cb\u003e193\u003c/b\u003e, 299\u0026ndash;316. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.earscirev.2019.04.023\u003c/span\u003e\u003cspan address=\"10.1016/j.earscirev.2019.04.023\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2019).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003ePascual, D. \u0026amp; Johansson, M. Increasing impacts of extreme winter warming events on permafrost. \u003cem\u003eWeather Clim. Extremes\u003c/em\u003e. \u003cb\u003e36\u003c/b\u003e, 100450. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.wace.2022.100450\u003c/span\u003e\u003cspan address=\"10.1016/j.wace.2022.100450\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2022).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003ePlanet Team. Planet imagery product specifications. (2018). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://assets.planet.com/docs/combined-imagery-product-spec-april-2019.pdf\u003c/span\u003e\u003cspan address=\"https://assets.planet.com/docs/combined-imagery-product-spec-april-2019.pdf\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003ePollard, W. H. \u0026amp; French, H. M. A first approximation of the volume of ground ice, Richards Island, Pleistocene Mackenzie Delta, Northwest Territories, Canada. \u003cem\u003eCan. Geotech. J.\u003c/em\u003e \u003cb\u003e17\u003c/b\u003e (4), 509\u0026ndash;516 (1980).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003ePorter, C. et al. ArcticDEM - Mosaics, Version 4.1. /10.7910/DVN/3VDC4W. https://doi.org/10.7910/DVN/3VDC4W (2023).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eSchuur, E. A. G. et al. Climate change and the permafrost carbon feedback. \u003cem\u003eNature\u003c/em\u003e \u003cb\u003e520\u003c/b\u003e (7546), 171\u0026ndash;179. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1038/nature14338\u003c/span\u003e\u003cspan address=\"10.1038/nature14338\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2015).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eShapiro, S. S. \u0026amp; Wilk, M. B. An Analysis of Variance Test for Normality (Complete Samples). \u003cem\u003eBiometrika\u003c/em\u003e \u003cb\u003e52\u003c/b\u003e (3/4), 591\u0026ndash;611. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.2307/2333709\u003c/span\u003e\u003cspan address=\"10.2307/2333709\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (1965).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eSmith, S. L. et al. The changing thermal state of permafrost. \u003cem\u003eNat. Reviews Earth Environ.\u003c/em\u003e \u003cb\u003e3\u003c/b\u003e (1), 10\u0026ndash;23. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1038/s43017-021-00240-1\u003c/span\u003e\u003cspan address=\"10.1038/s43017-021-00240-1\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2022a).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eSmith, S. L. et al. The changing thermal state of permafrost. \u003cem\u003eNat. Reviews Earth Environ.\u003c/em\u003e \u003cb\u003e3\u003c/b\u003e (1), 10\u0026ndash;23. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1038/s43017-021-00240-1\u003c/span\u003e\u003cspan address=\"10.1038/s43017-021-00240-1\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2022b).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eStreletskaya, I. et al. Pleistocene-Holocene palaeoenvironmental records from permafrost sequences at the Kara Sea coast (NW Siberia, Russia). Geography, environment, sustainability. 6(3):60\u0026ndash;76 (2013).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eSun, Z. et al. Creep characteristics and process analyses of a thaw slump in the permafrost region of the Qinghai-Tibet Plateau, China. \u003cem\u003eGeomorphology\u003c/em\u003e \u003cb\u003e293\u003c/b\u003e, 1\u0026ndash;10. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.geomorph.2017.04.045\u003c/span\u003e\u003cspan address=\"10.1016/j.geomorph.2017.04.045\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2017).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eSuresh, K., Severn, C. \u0026amp; Ghosh, D. Survival prediction models: an introduction to discrete-time modeling. \u003cem\u003eBMC Med. Res. Methodol.\u003c/em\u003e \u003cb\u003e22\u003c/b\u003e (1), 207. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1186/s12874-022-01679-6\u003c/span\u003e\u003cspan address=\"10.1186/s12874-022-01679-6\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2022).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eSwanson, D. K. \u0026amp; Nolan, M. Growth of retrogressive thaw slumps in the Noatak Valley, Alaska, 2010\u0026ndash;2016, measured by airborne photogrammetry. \u003cem\u003eRemote Sens.\u003c/em\u003e \u003cb\u003e10\u003c/b\u003e (7). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.3390/rs10070983\u003c/span\u003e\u003cspan address=\"10.3390/rs10070983\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2018).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eUber Technologies Inc. H3: Uber\u0026rsquo;s hexagonal hierarchical spatial index. (2018).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eVasiliev, A. A. et al. Permafrost degradation in the Western Russian Arctic. \u003cem\u003eEnviron. Res. Lett.\u003c/em\u003e \u003cb\u003e15\u003c/b\u003e (4), 045001. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1088/1748-9326/ab6f12\u003c/span\u003e\u003cspan address=\"10.1088/1748-9326/ab6f12\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2020).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eWard Jones, M. K., Pollard, W. H. \u0026amp; Jones, B. M. Rapid initialization of retrogressive thaw slumps in the Canadian high Arctic and their response to climate and terrain factors. \u003cem\u003eEnviron. Res. Lett.\u003c/em\u003e \u003cb\u003e14\u003c/b\u003e (5). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1088/1748-9326/ab12fd\u003c/span\u003e\u003cspan address=\"10.1088/1748-9326/ab12fd\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2019).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eWebb, H. et al. A Review of Abrupt Permafrost Thaw: Definitions, Usage, and a Proposed Conceptual Framework. \u003cem\u003eCurr. Clim. Change Rep.\u003c/em\u003e \u003cb\u003e11\u003c/b\u003e (1), 7. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1007/s40641-025-00204-3\u003c/span\u003e\u003cspan address=\"10.1007/s40641-025-00204-3\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2025).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eYang, Y. et al. Mapping retrogressive thaw slumps using deep neural networks. \u003cem\u003eRemote Sens. Environ.\u003c/em\u003e \u003cb\u003e288\u003c/b\u003e, 113495. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.rse.2023.113495\u003c/span\u003e\u003cspan address=\"10.1016/j.rse.2023.113495\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2023).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eOpenStreetMap. OpenStreetMap; [accessed 2025 Aug 8]. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.openstreetmap.org/\u003c/span\u003e\u003cspan address=\"https://www.openstreetmap.org/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":true,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"scientific-reports","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"scirep","sideBox":"Learn more about [Scientific Reports](http://www.nature.com/srep/)","snPcode":"","submissionUrl":"","title":"Scientific Reports","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"Scientific Reports","inReviewEnabled":true,"inReviewRevisionsEnabled":true},"keywords":"","lastPublishedDoi":"10.21203/rs.3.rs-7697239/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-7697239/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eRetrogressive thaw slumps (RTSs) are key indicators of permafrost thaw in the West Siberian Arctic. Based on a new high-resolution remote sensing-derived, field-verified inventory of 6168 RTSs for the Yamal and Gydan peninsulas, we provide the first large-scale spatio-temporal and climate sensitivity analysis since 1964. In Gydan, RTS clusters are located at higher elevations, on rougher terrain, and in lake-rich areas, whereas such associations are weaker in Yamal. Temporal analysis of RTS using historical and modern satellite imagery for key sites (~\u0026thinsp;6,103 km\u0026sup2;) indicates that RTS numbers increased 23-fold since 1964, and initiation rates raised 26-fold. A discrete-time Bayesian hazards model identified summer maximum precipitation as the strongest short-term factor. However, winter warming was the dominant long-term driver, with an increase of 1.62 σ resulting in a 37-pp increase in annual initiation probability. RTS development in the region poses high risks to gas and transport infrastructure.\u003c/p\u003e","manuscriptTitle":"Rapid increase in West Siberia’s retrogressive thaw slumps since 1964 associated with Arctic winter warming","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-10-15 06:41:11","doi":"10.21203/rs.3.rs-7697239/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"decision","content":"Revision requested","date":"2025-11-20T05:03:09+00:00","index":"","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2025-11-12T17:01:26+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"210756301706512629549088309932086940075","date":"2025-11-10T07:13:08+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2025-11-03T14:53:34+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"176036542308322366057085832878852590473","date":"2025-10-30T12:56:48+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"1827299174686331717850593487400029434","date":"2025-10-16T23:00:15+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2025-10-14T20:24:11+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2025-09-26T10:38:36+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2025-09-25T08:29:33+00:00","index":"","fulltext":""},{"type":"submitted","content":"Scientific Reports","date":"2025-09-23T18:53:54+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"scientific-reports","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"scirep","sideBox":"Learn more about [Scientific Reports](http://www.nature.com/srep/)","snPcode":"","submissionUrl":"","title":"Scientific Reports","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"Scientific Reports","inReviewEnabled":true,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"f4a8f427-76dc-4455-9b17-f8929a5f107d","owner":[],"postedDate":"October 15th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"under-review","subjectAreas":[{"id":56301664,"name":"Earth and environmental sciences/Climate sciences"},{"id":56301665,"name":"Earth and environmental sciences/Environmental sciences"},{"id":56301666,"name":"Earth and environmental sciences/Natural hazards"}],"tags":[],"updatedAt":"2026-05-13T20:11:27+00:00","versionOfRecord":[],"versionCreatedAt":"2025-10-15 06:41:11","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-7697239","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-7697239","identity":"rs-7697239","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}