Climate warming and wetting accelerate permafrost degradation in the Xiao Xing’anling Mountains, Northeast China at the southern limit of Eurasian latitudinal permafrost. Conventional interferometric synthetic aperture radar (InSAR) retrievals of active-layer thickness (ALT) assume vertically uniform soil properties, potentially biasing estimates in layered soils. We developed a depth-stratified InSAR framework combining the amplitude of seasonal deformation (ASD) with layer-specific ERA5 soil moisture and USDA soil texture. We applied it to 149 Sentinel-1B acquisitions from October 2016 to December 2021. Long-term linear deformation rates ranged from -47 to +53 mm/year, ASD reached 140 mm, and retrieved thaw depths ranged from 0.4 to 6.5 m. Against borehole observations, soil stratification reduced mean absolute error from 1.023 to 0.853 m (16.6%), with the largest improvement where thin active layer overlies warm permafrost. Deep thaw occurred preferentially at elevations of 250–400 m a.s.l., on slopes (<5°), and on shaded aspects. ASD varied linearly with the E-factor, an edaphic index integrating soil moisture and texture, consistent with their coupled control on seasonal deformation. The sensitivity resembled that of poorly drained, near-saturated Arctic lowlands but exceeded that of the drier, coarser substrates of the Qinghai–Tibet Plateau. In this climatic transition zone, extreme retrieved depths likely represent the depth of the permafrost table rather than ALT, indicating active-layer–permafrost decoupling/detachment associated with supra-permafrost talik development. Low ASD combined with persistent subsidence may therefore indicate talik initiation. The framework improves thaw-depth retrieval in heterogeneous patchy permafrost regions and enables scalable monitoring of permafrost degradation and infrastructure risk.

Aufeis is a seasonal ice accumulation formed by successive freezing of groundwater or surface-water overflow on land surfaces, river ice or lake ice during winter. It is widespread across cold regions and plays important roles in hydrology, geomorphology, ecosystems, and infrastructure stability, yet remains underrepresented in regional and global cryospheric assessments. This review synthesizes recent advances in aufeis research, focusing on spatiotemporal distribution, formation mechanisms, seasonal dynamics, environmental functions, engineering impacts, and mitigation strategies, while identifying major knowledge gaps and future research priorities. Current evidence suggests that aufeis extent has generally declined in many cold-region landscapes over recent decades, although large uncertainties persist in forested, mountainous, and data-sparse regions, particularly across Asia. Under continued climate warming and hydroclimatic intensification, aufeis is likely to become smaller, more fragmented, and less spatially continuous. However, its response is not uniformly negative: permafrost degradation may locally enhance groundwater recharge, talik development, and winter baseflow, thereby promoting aufeis formation in some settings. This dual response highlights the strong dependence of aufeis dynamics on local hydrogeological structure, permafrost conditions, snow regime, and surface-subsurface connectivity. Beyond its geomorphic expression, aufeis can function as an important seasonal water reservoir, contributing substantially to spring and early-summer runoff and, in some basins, rivaling or exceeding the hydrological contribution of nearby small glaciers. These hydrological functions are critical for sustaining cold-region ecosystems, regulating streamflow seasonality, and buffering water scarcity during early thaw periods. At the same time, aufeis poses persistent and in some cases growing risks to roads, railways, pipelines, and other linear infrastructure, prompting a shift from conventional passive control measures toward integrated, process-informed, resilience-based mitigation. Aufeis-related landforms, sedimentary records, and geochemical signatures also provide valuable archives for reconstructing Quaternary hydrogeological and periglacial environments. A process-based, interdisciplinary understanding of aufeis is therefore essential for predicting its response to climate change, improving cold-region water resource assessment, interpreting paleoenvironmental records, and supporting the design, maintenance, and adaptation of resilient infrastructure in a warming cryosphere.

Integrated assessments of vegetation dynamics and ground thermal responses remain limited along warm discontinuous permafrost corridors on northeastern Qinghai–Tibet Plateau. This knowledge gap is particularly important for the compound G214-G0613 engineering corridor across the Bayan Har Mountains, where national highway and expressway construction jointly intensify anthropogenic disturbance to permafrost–ecosystem interactions. Here, we assessed vegetation and permafrost responses within a 10-km corridor buffer using Landsat-derived growing-season normalized difference vegetation index (NDVIgs; 2000–2020), mean annual ground temperature at the depth of zero annual amplitude from 16 monitoring sites (MAGTDZAA; 2010–2019), and modeled MAGT at 10 m depth (MAGT10m). Trend-free pre-whitening Mann-Kendall tests, Hurst analysis, segmented regression, and GeoDetector were used to quantify vegetation trends, disturbance thresholds, and dominant environmental controls. NDVIgs increased significantly across the corridor at 0.0045 a−1, while MAGTDZAA warmed at 0.012 °C a−1. Despite this regional greening, near-road vegetation degradation was concentrated within an acute disturbance zone extending 0.91 km from the corridor, with a broader transition zone reaching 2.13 km. GeoDetector results showed that annual precipitation and elevation were the dominant controls on NDVIgs variability, whereas elevation and soil type primarily governed MAGT10m patterns. Areas with high human footprint were the only class with a positive median MAGT10m (0.33 °C), indicating absent or actively degrading permafrost in warmer, lower-elevation terrain where engineering activity is concentrated. These results reveal a scale-dependent duality between corridor-wide greening and localized near-road degradation, a pattern that regional trend analyses may obscure. The study provides a quantitative basis for ecological buffer-zone design and permafrost-sensitive planning along compound engineering corridors in warm discontinuous permafrost regions.

Snow cover critically regulates the thermal stability of pipeline foundation soils in permafrost regions, yet its interaction with buried warm pipelines, particularly the lateral expansion of supra-permafrost subaerial taliks (SST), remains poorly quantified. Here, we conducted 30-year numerical model simulations for the China-Russia Crude Oil Pipeline I, incorporating dynamic snow boundary conditions and a linear climate warming rate (0.05 degrees C yr- 1) across three permafrost zones in the northern Da Xing'anling Mountains, Northeast China (mean annual ground temperature at -1.8 degrees C at Xing'an, -1.3 degrees C at Xinlin, and - 0.7 degrees C at Jagdaqi). Snow insulation markedly accelerates SST initiation, with taliks forming in years 3-8 compared with years 8-19 under snow-free conditions. By suppressing winter heat loss, snow drives cumulative subsurface heat storage. As a result, seasonal freezing weakens: frost penetration decreases and the seasonal frost depth decreases, while the total thaw depth progressively deepens under sustained pipeline heating and climate warming, indicating a decoupling between seasonal frost action and long-term thaw. Snow cover strongly enhances lateral SST expansion. In the isolated permafrost zone (Jagdaqi), SSTs transit from a localized thaw bulb to a laterally continuous thawed corridor. By year 30, SST areal extent per unit pipeline length increases by about one order of magnitude (e.g., 7.2-67.1 m2 at Xinlin), and total thaw depth reaches 18.5 m versus 10.9 m without snow. Overall, neglecting nival effects can substantially underestimate lateral foundation instability, highlighting the need for mitigation that couples snow management with thermal insulation design.

Aufeis is a sheetlike or layered accumulation of ice that forms on the ground surface or on top of river and lake ice when groundwater or surface water repeatedly discharges and freezes during the cold season. In the northern Da Xing'anling Mountains of Northeast China, the occurrence and distribution of aufeis are regulated by the coupled influences of hydroclimate, ground thermal conditions, topography, geomorphology, and hydrology. However, the interactions among these factors are highly complex, making it difficult to clearly resolve the spatial patterns and driving mechanisms of aufeis development. Focusing on the G111 National Highway corridor between Jagdaqi and Mo ' he, this study integrates spatial autocorrelation analysis with the Geodetector model to identify clustering characteristics, quantify dominant controls, and evaluate the synergistic effects of multiple environmental factors on aufeis formation and distribution. Field investigations identified 61 individual aufeis features, which exhibit significant spatial clustering. Geodetector analyses of 15 factors show that precipitation (mean q = 0.216), snow depth (q = 0.205), soil moisture (q = 0.183), and air temperature (q = 0.144) are the dominant controls at the regional scale. Among factor interactions, precipitation and slope aspect jointly provide the strongest explanatory power (q = 0.650). At the local scale, snow depth (q = 0.482), slope angle (0.438), and proximity to rivers (0.405) exert the strongest influences. Additional variables, including distance to faults, slope aspect, and slope angle, act as important modulators that selectively enhance spatial differentiation. Moreover, nonlinear interactions among factors substantially strengthen their explanatory power for aufeis distribution. Two years of field observations further indicate that aufeis exhibits marked spatial mobility and temporal periodicity. These findings highlight the coupled roles of climate, topography, hydrology, and permafrost-related environmental conditions in shaping aufeis formation and distribution and provide a scientific basis for infrastructure planning, ecological conservation, and water resource management in cold-region environments.

In the northern Da Xing'anling Mountains, climate warming and intensive land use are accelerating thaw-related hazards that compromise transportation infrastructure. During field campaigns in August–September 2023, we mapped thaw-hazard distribution and acquired ground-temperature, electrical resistivity, and topographic data along major roads, railways, and the China–Russia Crude Oil Pipelines (CRCOPs). We identified 290 hazard sites: 85 thaw-settlement depressions with substantial pavement subsidence, 49 sites with extensive longitudinal cracking, and 156 sections of undulating (“wave”) roadway or pipeline right-of-way. Ground temperatures were generally lower on the western flank where permafrost is better preserved. At depths of 0–1 m beneath representative surfaces, ground temperatures were highest under asphalt (5.1 °C), followed by concrete (4.3 °C), bare ground (3.8 °C), and an experimental concrete-brick pavement (2.5 °C). The CRCOPs produced a stronger lateral thermal footprint than road and railway embankments, extending ≈16 m from the pipeline axis versus ≈10 m for other linear infrastructure. Hazards clustered on the west- to northwest-facing slopes with mean annual ground temperatures <+1 °C, predominantly at 400–600 m a. s. l., slope angles <8°, and latitudes 51°-53°N. Principal drivers include regional warming, enhanced ground-surface heat absorption, and water pooling that augments heat advection while reducing surface albedo. These findings provide actionable constraints for siting, design, and maintenance to improve the resilience of transportation corridors in permafrost-affected terrain.

In the northern Da Xing'anling Mountains of Northeast China, thermokarst lakes have undergone significant changes. These changes are particularly evident along the Highway X302, which stretches from Yakeshi in the south to Yitulihe in the north, across zones of discontinuous to patchy Xing'an permafrost (XAP). Thermokarst lakes are highly sensitive to both climate warming and human activities. However, changes in their areal extent and number over the past two decades remain unclear. To identify the drivers of thermokarst lake changes in the XAP region, this study analyzes the relationship among lake dynamics, climate change, and human activities. Rich data on lake changes were extracted using the modified normalized difference water index (MNDWI) from the Google Earth Engine (GEE) and verified against the Joint Research Centre (JRC) global surface water datasets. The results show a net increase in both the number and areal extent of thermokarst lakes along the Highway X302 from 2000 to 2020. Larger lakes (greater than 0.01 km2) expanded in surface area. Meanwhile, smaller lakes (less than 0.01 km2) increased in number. A bell-shaped trend was observed, with an initial increase in lake area followed by a decline around 2013. Before 2013, climate factors such as precipitation, air temperature, and potential evapotranspiration (PET) strongly influenced lake changes. After 2013, annual precipitation became the dominant driver of lake expansion. Human activities also contributed to changes in permafrost thermal regimes, further influencing lake dynamics. These findings provide valuable insights into the management of thermokarst lakes and wetlands in the region, and offer a reference for mitigating frost-related hazards along the Highway X302.

Abstract Northern high-latitude permafrost is facing unprecedented wildfire disturbances, driving an anomalous regional increase in annual carbon emissions (8.1 ± 2.9 TgC yr⁻¹ from 1997 to 2023) against a backdrop of declining global wildfire emissions. To elucidate these complex dynamics, this review conceptualizes the “Permafrost Critical Zone” (PCZ) and adopts a holistic Earth-system perspective to evaluate the cascading impacts of wildfires on vulnerable cryospheric landscapes. We synthesize how fire-induced organic layer combustion and surface albedo reduction destabilize the PCZ, drastically elevating ground surface temperatures by up to 7 °C and deepening the active layer by up to six times. These severe thermal shocks fundamentally rewire hydrological pathways, accelerating ground ice melt, altering supra-permafrost water storage, and amplifying surface runoff. Concurrently, wildfires abruptly reduce microbial diversity and restructure cold-adapted biological communities, initiating divergent post-fire vegetation succession trajectories. While ecological and hydrothermal recovery is essential for restoring carbon and water fluxes, the compounding effects of repeated fires under a warming climate threaten to irreversibly degrade these environments. We conclude by highlighting critical knowledge gaps and emphasizing the necessity of integrating PCZ dynamics into global models to predict impending climate tipping points and to inform long-term sustainable development strategies.

Transferable, quantitative frameworks for assessing freeze–thaw hazards (FTHs) along engineering corridors remain limited in warming marginal permafrost regions, particularly those integrating hydroclimatic forcing, terrain controls, and anthropogenic disturbance. Here, we developed and validated an integrated machine-learning (ML)–GeoDetector framework, supported by field surveys and electrical resistivity tomography (ERT), for the Jagdaqi–Mo’he section of National Highway G111 in the Xing’an Permafrost (XAP) region of Northeast China, a representative thermally transitional discontinuous permafrost environment. Field investigations identified six major FTH types, among which thermokarst lakes and uneven pavement settlement were the most widespread, with hazard clusters showing strong spatial heterogeneity concentrated where permafrost degradation and engineering disturbance converge. Among five ML classifiers, the random forest (RF) model achieved the highest area under the curve (AUC = 0.93) and overall accuracy (OA = 0.86). Risk mapping shows that 23.7% of predicted FTH pixels fall within high-risk zones, primarily where uneven pavement settlement coincides with thermokarst lake development. Annual precipitation was identified as the dominant driver by both RF importance ranking (17.3%) and GeoDetector analysis (q = 0.54), and its interaction with slope angle yielded the highest explanatory power (q = 0.63), indicating that FTH hotspots are driven by hydroclimatic–topographic coupling rather than single-factor extremes. ERT profiles further show that similar near-surface thermal conditions can conceal markedly different subsurface permafrost structures, emphasizing the need for geophysical constraints in risk interpretation. The proposed RF−GeoDetector framework provides a transferable approach for corridor-scale FTH risk mapping and adaptive infrastructure management in warming marginal permafrost regions globally.

This study assesses the stability of the Bei'an-Hei'he Highway (BHH), located near the southern limit of latitudinal permafrost in the Xiao Xing'anling Mountains, Northeast China, where permafrost degradation is intensifying under combined climatic and anthropogenic influences. Freeze-thaw-induced ground deformation and related periglacial hazards remain poorly quantified, limiting regional infrastructure resilience. We developed an integrated framework that fuses multi-source InSAR (ALOS, Sentinel-1, ALOS-2), unmanned aerial vehicle (UAV) photogrammetry, electrical resistivity tomography (ERT), and theoretical modeling to characterize cumulative deformation, evaluate present stability, and project future dynamics. Results reveal long-term deformation rates from -35 to +40 mm/yr within a 1-km buffer on each side of the BHH, with seasonal amplitudes up to 11 mm. Sentinel-1, with its 12-day revisit cycle, demonstrated superior capability for monitoring the Xing'an permafrost. Deformation patterns were primarily controlled by air temperature, while precipitation and the topographic wetness index enhanced spatial heterogeneity through thermo-hydrological coupling. Wavelet analysis identified a 334-day deformation cycle, lagging climate forcing by similar to 107 days due to the insulating effects of peat. Early-warning analysis classified 4.99 % of the highway length as high-risk (subsidence <-18.18 mm/yr or frost heave >10.91 mm/yr). The InSAR-based landslide prediction model achieved high accuracy (Area Under the Receiver Operating Characteristic (ROC) Curve, or AUC = 0.9486), validated through field surveys of subsidence, cracking, and slow-moving failures. The proposed 'past-present-future' framework demonstrates the potential of multi-sensor integration for permafrost monitoring and provides a transferable approach for assessing infrastructure stability in cold regions.

Rising crude-oil flow temperatures have been observed at pump stations along the China-Russia Crude Oil Pipelines (CRCOPs) I and II, which traverse discontinuous, sporadic, and patchy permafrost zones and have been in operation since 2011 and 2018, respectively. However, the long-term thermal impacts of elevated oil temperatures for the stability of foundation soils around buried pipelines remain poorly constrained. A two-dimensional conductive heat transfer model incorporating ice-water phase change was developed to quantify vertical and lateral heat fluxes and permafrost degradation beneath buried pipelines under high (observed) and low (designed) oil-flow temperatures in a warming climate. The mitigation performance of insulation layers with thicknesses of 0, 3, 5, 8, 10, 12, and 15 cm is further evaluated. Results indicate that elevated oil temperature substantially intensifies heat transfer into the surrounding ground, causing the design-temperature condition to markedly underestimate long-term soil warming. The maximum difference in mean annual heat flux between high- and low-temperature scenarios reached 7.01 W & sdot;m-2, approximately 3.6 times higher than under designed temperatures. At 6 m depth, the ground warming rate reached 1.84 degrees C per decade under the high-temperature scenario, compared with 0.58 degrees C per decade under the low-temperature scenario. Insulation layers of 10 and 15 cm reduce heat loss by 76.8 and 85.2%, respectively. These findings quantify the long-term thermal impacts of oil temperature on buried pipelines in permafrost regions and provide a basis for optimizing thermal design and maintenance strategies.

Under the combined effects of climate change and human activities, permafrost is undergoing accelerated degradation, significantly affecting the boreal ecological, hydrological processes, climate systems, and engineering infrastructure. Northeast China, on the southern margin of permafrost regions on the East Asian continent, has experienced the formation, expansion, and interconnection of taliks (thawed or unfrozen parts in permafrost zones) due to human activities (e.g., mining and urbanization) and natural disturbances (e.g., wildfires), resulting in abrupt and accelerated permafrost degradation. This study leverages extensive remote sensing data and ground observations to explore the impacts of engineering activities on the permafrost environment in the Hola Basin in the northern Da Xing'anling Mountains in Northeast China. The results showed that engineering activities substantially altered the surface landscapes of the permafrost region. From 1969 to 2025, engineering disturbed areas expanded at a rate of 0.31 km2/yr, resulting in the loss of 14.27 km2 of natural surface landscapes, equivalent to 6.42% of the total basin area. Engineering activities also intensified vegetation degradation within and around the disturbed areas. The annual maximum normalized difference vegetation index (NDVImax) in engineering disturbed areas decreased significantly at a rate of −0.010 yr−1 (p < 0.001), and the primary spatial extent of their influence on surrounding vegetation was approximately 360 m, with a bootstrap 95% confidence interval of 296–426 m. Land surface temperatures (LST) were generally higher in engineering disturbed areas than in natural surface areas, with a mean difference of 2.12 ± 2.78 °C and pronounced seasonal variability. Borehole ground temperature observations further indicated that intense engineering disturbance increased the maximum thaw depth or active layer thickness (ALT) and substantially altered ground freeze-thaw processes. In areas of intense engineering disturbance, the ground thawing duration is prolonged, the total duration of ground freezing in the active layer is shortened, and taliks developed locally. These changes can further accelerate permafrost degradation, increase freeze-thaw hazards, and adversely affect the geological environment and engineering safety. These findings provide important empirical and a scientific basis for further elucidating the interactions among climate change, engineering activities, the active layer, and permafrost, while also supporting the sustainable management of northern forests ecosystems and the planning and maintenance of engineering infrastructure in cold regions.

The limited regional adaptability of soil moisture prediction models constrains their application under complex climatic conditions. Enhancing modeling accuracy and predictive capability is crucial for improving the precision of climate simulations and the effectiveness of extreme weather early warnings. This study proposes an interpretable and generalizable soil moisture prediction approach that employs the SABO algorithm to optimize CONV1D-BiLSTM model. The model's performance was evaluated and validated at Linzhi, Dangxiong, Mozhugongka, and Xietongmen in the southwestern Tibetan Plateau. Results indicate that, across four distinct environmental settings, the proposed model achieved an average root mean square error (RMSE) of 1.859 and an average coefficient of determination (R2) of 0.929. Furthermore, using the SHapley Additive exPlanations (SHAP) method, soil temperature and relative humidity were identified as key features across multiple stations. This method enhances the understanding of soil moisture dynamics in the context of climate change and provides a powerful tool for climate risk assessment and early warning on the Tibetan Plateau region.

Thermokarst lake dynamics crucially affect permafrost degradation, greenhouse gas emissions, hydrological and geomorphic processes and ecological responses in cold-region environments. However, the spatiotemporal evolution and driving mechanisms of thermokarst lakes in mid- to high-latitude regions characterised by discontinuous, sporadic and isolated patches of permafrost, such as Northeast China, remain insufficiently investigated, particularly in the context of their long-term interactions with climatic and anthropogenic factors. This study integrates machine learning techniques with Landsat imagery to investigate changes in the number and area of thermokarst lakes along the Mo’he county section of National Highway G111 in the northern Da Xing’anling Mountains from 1989 to 2020. Results indicate a 243.7% increase in the number of lakes larger than 0.1 hm2 (from 355 to 865) and a 345.3% expansion in total lake area (from 233.6 to 806.6 hm2), with the most significant changes (p ＜ 0.05) occurring in riverine and anthropogenically disturbed areas. Lake expansion is significantly correlated with land use indicators and climatic factors, including cropland area (r = 0.96), built-up area (r = 0.97), and potential evapotranspiration (r = 0.94). These findings suggest that anthropogenic disturbances, such as road construction and land use change, have intensified permafrost thaw by increasing ground heat flux and altering surface hydrology. Moreover, the interplay between climate warming and human activities has accelerated thermokarst lake expansion. This study underscores the need for land-use planning and highlights the importance of identifying disturbance-prone areas to support sustainable development and ecosystem management in permafrost regions.

Engineering disturbances are increasing in permafrost regions of northeastern China, where soil microorganisms play essential roles in biogeochemical cycling and are highly sensitive to linear infrastructure disturbances. However, limited research has addressed how microbial communities respond to different post-engineering-disturbance recovery stages. This study investigated the impacts of the China–Russia Crude Oil Pipelines (CRCOPs) on soil microbial communities in a typical boreal forest permafrost zone of the Da Xing’anling Mountains. Soil samples were collected from undisturbed forest (the control, CK); short-term disturbed sites associated with Pipeline II, which was constructed in 2018 (SD); and long-term disturbed sites associated with Pipeline I, which was constructed in 2011 (LD). Pipeline engineering disturbances significantly increased soil clay content and pH while reducing soil water content (SWC), soil organic carbon (SOC), total nitrogen (TN), and total phosphorus (TP) (p < 0.05). No significant differences in these soil properties were observed between SD and LD. Bacterial diversity increased significantly, whereas fungal diversity significantly decreased following pipeline disturbances (p < 0.05). The beta diversity of both bacterial and fungal communities differed significantly among the three disturbance types. At the phylum level, pipeline disturbance increased the relative abundances of Proteobacteria, Acidobacteriota, Actinobacteriota, Ascomycota, and Mortierellomycota while reducing those of Bacteroidota and Basidiomycota. These shifts were associated with disturbance-induced changes in soil properties. Microbial co-occurrence networks in SD exhibited greater complexity and connectivity than those in CK and LD, suggesting intensified biotic interactions and active ecological reassembly during the early recovery phase. These findings suggest that pipeline disturbance could drive soil microbial systems into a new stable state that is difficult to restore over the long term, highlighting the profound impacts of linear infrastructure on microbial ecological functions in cold regions. This study provides a scientific basis for ecological restoration and biodiversity conservation in permafrost-affected areas.

In boreal permafrost regions, forests are among the primary ecotypes, where fire severity, extent and frequency have been rising due to a warming climate. Following a fire, the composition (species types and diversity), structure (arrangement and layering of trees and undergrowth), and successional trajectory (changes in the ecosystem over time) of forest vegetation are modified. These changes have significant impacts on ecosystem carbon balance, ecosystem services, and forest management, influencing carbon release and storage, biodiversity, hydrological cycles, as well as fire prevention measures and risk assessment. However, at the southern edge of the boreal forest, where permafrost ecosystems are particularly vulnerable to climate change and wildfire disturbances, few studies have focused on post-fire understory vegetation renewal and succession. In this study, we selected six burned areas in Northeast China, including four larch forests and two shrub wetlands, to examine vegetation succession pathways under varying fire severities from 4 to 32 years after wildfires. The results showed that in larch forests, herbaceous vegetation, particularly graminoids, benefited from fire disturbances in the early post-fire period, showing significant increases in coverage, biomass, height, and species diversity. In the mid- post-fire period (four to ten years after the fire), shrubs, especially tall shrubs, began to benefit from forest fires, showing significant increases in cover, biomass and height. Tall shrub species increased during the early post-fire period (first four years), then declined, gradually recovering in the middle and late post-fire period (10 to 32 years). In larch forests, fire severity and post-fire increase in soil temperature and moisture contents constrained shrub growth while promoting herbaceous vegetation. Similarly, in shrub wetlands, shrub coverage and biomass remained higher at burned sites than at unburned ones, while the opposite trend was observed for herbaceous vegetation. In contrast to larch forest, shrub coverage significantly increased in shrub wetlands following a fire. During the period of 17 to 32 years after the fire, the height and species diversity of herbs and shrubs were lower at the burned sites compared to unburned sites, leading to the transformation of herb- dominated wetlands into shrub-dominated ones. Fire severity and permafrost degradation significantly inhibited herb growth while promoting shrub growth in shrub wetlands. These fire-induced effects intensified with rising fire severity, and the legacy effects of forest fires on herbs and shrubs persisted for up to 32 years after the event. Additionally, forest fires increased the species diversity of flowering plants. Therefore, this study provides valuable data to support the restoration and management of vegetation following forest fires in boreal forest regions.

This paper investigates the spatiotemporal dynamics and their changes of the southern limit of latitudinal permafrost (SLLP) and the lower limit of mountain permafrost (LLMP) in Northeast China, emphasizing the roles of climate change and human activities. Permafrost in this region is primarily distributed in the northern parts of the Da and Xiao Xing'anling mountain ranges and in the upper parts of the Changbai Mountains and at the summits of the Huanggangliang Mountains in the southern part of the Da Xing'anling Mountain Range. Permafrost degradation, ongoing since at least the local Holocene Megathermal Period (8.5-6.0 ka BP), has intermittently reversed during cooler climatic intervals but continues to exert significant impacts on regional environments, infrastructure stability, and carbon storage. Notably, the northward retreats of the SLLP since the mid-19th century underscore the sustained nature of this degradation, especially in southern patchy permafrost zones increasingly sensitive to warming and anthropogenic influences. LLMP variability is similarly shaped by a combination of climatic, hydrometeorological, ecological, and topographic factors. The distributions of SLLP and LLMP are further complicated by the presence of relict and sporadic permafrost, as well as the hydrothermal effects of vegetation and snow cover. Addressing the challenges of mapping and modeling boreal permafrost in Northeast China requires comprehensive field investigations, long-term in situ monitoring via station networks, and advanced numerical modeling. Emerging technologies, including satellite and airborne remote sensing (RS), geographic information systems (GIS), unmanned aerial vehicles (UAVs), surface geophysical methods, and big data analytics, offer new possibilities for enhancing permafrost monitoring and mapping. Integrating these tools with conventional field studies can significantly improve our understanding of permafrost dynamics. Continued efforts in monitoring, technological innovation, multidisciplinary collaboration, and international cooperation are essential to meet the challenges posed by permafrost degradation in a changing climate.

Permafrost, a major component of the cryosphere, is undergoing rapid degradation due to climate change, human activities, and other external disturbances, profoundly impacting ecosystems, hydroclimate, engineering geological stability, and infrastructure. In Northeast China, the thermal dynamics of the Xing’an permafrost are particularly complex, complicating the accurate assessment of its spatial extent. Many earlier mapping efforts, despite significant progress, fall short in accounting for some key local geoenvironmental factors. Thus, this study introduces a new approach that corporates four key driving factors—biotic, climatic, physiographic, and anthropogenic—by integrating multi-source datasets and in-situ observations. Four machine learning (ML) models (Random Forest (RF), Support Vector Machine (SVM), Logistic Regression (LR), and XGBoost (XGB)) are applied to simulate permafrost distribution and probability, as well as to evaluate their performance. The results indicate that models’ accuracy, ranked from highest to lowest, is as follows: RF (Area Under the Curve (AUC)=0.88, and Accuracy=0.81), XGB (0.86 and 0.77), LR (0.81 and 0.73), and SVM (0.76 and 0.66), with RF emerging as the most effective model for permafrost mapping in Northeast China. Analysis of the relationships between predictors and permafrost occurrence probability (POP) indicates that vegetation and snow cover exert non-linear effects on permafrost, while human activities significantly reduce POP. Additionally, finer soil textures and higher soil organic matter content are positively correlated with increased POP. The modeling results, combined with field survey data, also show that permafrost is more prevalent in lowlands than in uplands, confirming the symbiotic relationship between permafrost and wetlands in Northeast China. This spatial variation is influenced by local microclimates, runoff patterns, and soil thermal properties. The primary sources of model error are uncertainties in the accuracy of multi-source datasets at different scales and the reliability of observational data. Overall, ML models demonstrate great potential for mapping permafrost in Northeast China.