High-resolution glacier monitoring data are essential for glaciological research， and unmanned aerial vehicle （UAV） data offer an important opportunity to resolve the systematic discrepancies among glacier change results obtained from different methods and different data sources. UAV-based monitoring can provide reliable validation benchmark for satellite remote sensing data and can compensate for the insufficient spatial coverage of in-situ measurements， enabling full-coverage and refined extraction of glacier surface information and thereby playing a crucial bridging role between satellite remote sensing and in-situ measurement. In this study， Qiyi Glacier in the Qilian Mountains， where long-term observations have been conducted， was selected as the study area. The 0.05 m-resolution UAV digital elevation model （UAV-DEM） and UAV digital orthophoto map （UAV-DOM）， generated from data acquired between July 28 and August 3， 2024 using a DJI M300 RTK UAV equipped with a Zenmuse L1 sensor， were used as the data foundation. By calculating glacier surface slope gradients derived from DEMs of different resolutions， the spatial resolution scale effect of DEMs in representing glacier terrain was analyzed， and the influence of DEM resolution on solar radiation received by the glacier surface was further quantified using a distributed solar radiation model. Additionally， a support vector machine （SVM） method was applied to classify the glacier surface according to the surface dirtiness level of Qiyi Glacier， and supraglacial channels and supraglacial lakes were identified through visual interpretation， so as to systematically summarize the refined glacier surface information extractable from UAV-DOM and provide data support for refined monitoring of Qiyi Glacier. Meanwhile， Sentinel-2A and Landsat-9 imagery， which are widely used in glacier change studies， were introduced for comparative analysis.The results showed that　（1） DEMs with different spatial resolutions exhibited significant differences in representing the topographically complex gully region of glacier terminus， while they had a relatively minor influence on the topographically flat central region of glacier tongue and the upper-middle region of the glacier. The analysis of the gully region of glacier terminus revealed that resolutions finer than 0.4 m overrepresented glacier topography by capturing microtopographic features， whereas resolutions coarser than 20 m caused terrain undulations to essentially disappear. An optimal DEM resolution of 4 m for representing the gully region of glacier terminus was identified based on the mean local variance of slope. （2） DEM spatial resolution also had a significant effect on solar radiation estimates in the gully region of glacier terminus. Within the resolution range of 4~20 m， each 1 m increase in resolution resulted in increases of 1.5 W·m^-2 in mean solar irradiance and 24.5 h in mean sunshine duration during the 2024/2025 mass balance year， whereas the influence became weak at resolutions coarser than 20 m. （3） Compared with the commonly used Sentinel-2A and Landsat-9 imagery， the overall accuracy of glacier surface dirtiness classification based on UAV-DOM （85%） was improved by 25% and 35%， respectively， providing validation data for glacier surface albedo inversion derived from satellite remote sensing. Furthermore， UAV-DOM enabled the extraction of complete supraglacial channel information， providing geometry parameters used to estimate glacier runoff. During the ablation season in 2024， UAV-DOM detected a newly developed supraglacial lake at an elevation of approximately 4 995 m near the summit of Qiyi Glacier， indicating that this glacier may undergo substantial ablation and even disappear within a short period. In summary， high-resolution UAV data can play an important bridging role between satellite remote sensing and in-situ measurement， and are of significant importance for improving the accuracy of refined mountain glacier monitoring. UAV technology has greatly advanced refined glacier monitoring research， particularly in acquiring high-precision topographic data and glacier surface information， thereby supporting refined monitoring of changes in Qiyi Glacier and its response to climate change.

Mountain glaciers， as sensitive indicators and regulators of climate change， are experiencing accelerated melting and mass loss under global warming. Despite extensive research， large uncertainties remain in predicting future glacier evolution. Glaciers possess unique thermodynamic characteristics—low temperature， high albedo， and low thermal conductivity—that profoundly affect local temperature， precipitation， and wind fields. Their retreat alters surface energy balance and regional climate feedbacks， thereby reshaping atmospheric circulation patterns and water resource distribution in mountainous regions. Understanding the two-way coupling between glaciers and the atmosphere is thus critical for improving regional climate prediction and high-mountain water resource management.Traditional climate models typically simplify or neglect this two-way interaction. The WRF （weather research and forecasting） model， featuring multi-scale nesting capability， sophisticated physical parameterization， and adaptability to complex terrain， provides an effective framework for simulating glacier-atmosphere interactions at high spatial resolution. This review systematically summarizes recent advances in the application of the WRF model in studies of glacier feedbacks on regional climate. It focuses on the physical mechanisms underlying glacier-atmosphere interactions， the development of high-resolution modeling， and the model’s critical role in elucidating energy and mass exchange processes.The results show that （1） Glacier-atmosphere interactions exhibit strong multi-scale coupling characteristics. Glaciers modulate surface energy balance and local circulation， significantly influencing regional climate. The coupling between glacier winds， valley winds， foehn flows， and monsoonal systems regulates local wind fields and precipitation distribution. Moreover， the glacier’s cold-source effect modifies boundary-layer stability and temperature gradients， thereby impacting cloud formation and precipitation phase transitions. （2） Glacier changes trigger pronounced climate feedbacks. Glacier retreat weakens katabatic flows and local cooling effects， shifts moisture transport pathways upward， and enhances high-altitude precipitation and snow accumulation， forming a nonlinear “glacier-precipitation-feedback” cycle. The dynamic equilibrium among glacier wind intensity， topographic slope， and energy flux exchange determines the spatiotemporal characteristics of local thermal anomalies. （3） The WRF model， by introducing glacier process parameterization and two-way coupling modules， successfully reproduces surface energy and mass balance， capturing key mesoscale processes such as glacier winds， foehn events， and mountain convection. High-resolution simulations （≤1 km） markedly improve the representation of wind fields， precipitation， and energy fluxes in complex terrain. （4） Recent progress in large-eddy simulation （WRF-LES） enables partial resolution of turbulent structures and energy transfer within glacier boundary layers at scales of tens to hundreds of meters， providing new insights into local energy balance mechanisms， though numerical stability and computational cost remain challenges.Improving simulation accuracy depends critically on enhanced observational systems. Future research should strengthen synergy among multi-platform and multi-scale observations. Ground-based and near-surface instruments—automatic weather stations， acoustic snow sensors， and cosmic-ray neutron probes—can provide high-frequency constraints on energy and mass fluxes. Drone and terrestrial LiDAR （or TLS） surveys can provide meter-scale monitoring of glacier surface geometry and roughness， supporting parameter calibration and turbulent flux estimation. Additionally， boundary-layer remote sensing （LiDAR， RASS） can capture turbulence and wind evolution above glaciers， compensating for the limitations of traditional meteorological stations. Satellite observations （e.g.， Sentinel， ICESat-2， CryoSat-2） continuously monitor glacier elevation， albedo， snow extent， and freeze-thaw dynamics.Integrating these multi-source observations with high-resolution modeling allows data assimilation and cross-validation from regional to local scales， significantly improving the representation of boundary-layer structures， energy fluxes， and precipitation phase transitions. Future research should promote an integrated “observation-assimilation-simulation-validation” framework， establish standardized glacier observation networks， and enhance data sharing and international collaboration. As observation technology and modeling capacity advance， understanding of nonlinear glacier-climate feedbacks will deepen， providing stronger scientific support for regional climate prediction and sustainable high-mountain water resource management.

Permafrost peatlands represent critical carbon reservoirs within terrestrial ecosystems， playing a significant role in the global carbon cycle. However， under ongoing climate warming， the accelerated loss of soil carbon from these ecosystems intensifies greenhouse gas emissions， forming a crucial climate feedback loop. The Greater Hinggan Mountains， as a typical distribution area of permafrost peatlands， are highly sensitive to climate change and provide an ideal natural laboratory for investigating the mechanisms and dynamics of soil carbon loss. Through a systematic literature review， this study summarizes current knowledge to elucidate the characteristics and key drivers of soil carbon loss in this vulnerable region. The main results are as follows. （1） Seasonal dynamics significantly influence carbon emissions and dissolved organic carbon （DOC） concentrations. During the growing season， carbon dioxide （CO₂） fluxes， methane （CH₄） fluxes， and DOC concentrations in pore water exhibited pronounced variability， with values ranging widely （CO₂： 5.03 to 1 384.59 mg·m^-2·h^-1； CH₄： -0.248 to 11.459 mg·m^-2·h^-1； DOC： 22.08 to 188.10 mg·L^-1）. This high variability underscores the complex interactions among biological activity， hydrological conditions， and temperature. （2） Warming enhances CO₂ and CH₄ emissions primarily by stimulating microbial activity and increasing soil DOC availability. （3） Soil moisture and groundwater levels exert differential effects on CO₂ and CH₄ emissions. CO₂ emissions display an inverted U-shaped relationship with soil moisture content and are negatively correlated with groundwater table depth. In contrast， CH₄ emissions show a positive correlation with both soil moisture and groundwater level， reflecting the anaerobic conditions required for methanogenesis. （4） Freeze-thaw cycles （FTCs） significantly suppress CO₂ emissions， with the suppression effect in deep soil layers approximately three times greater than that in surface layers. Conversely， FTCs enhance CH₄ emissions （by an average of ~308%） and increase DOC concentrations （1.38~1.62 times higher than the control）. Under FTC conditions， DOC content increases with rising soil moisture， indicating synergistic effects between hydrological and thermal regimes. （5） Combined warming and waterlogging treatments reduce CO₂ emissions （by 36.15%~41.15%） but substantially increase CH₄ emissions （increased by a factor of 17~32）. Furthermore， FTCs with larger temperature amplitudes （-15 to 5 ℃） result in significantly higher CO₂ and CH₄ emission rates—2.26 and 1.74 times greater， respectively—compared to those with smaller temperature variations （-5 to 5 ℃）. This highlights the importance of temperature oscillation magnitude in regulating carbon release during freeze-thaw events. Collectively， these findings reveal the complex response of the carbon cycle to climate warming in the permafrost peatlands of the Greater Hinggan Mountains and demonstrate that soil carbon loss is governed by multiple interacting factors. Despite these insights， current research still contains uncertainties and limitations. To address these gaps， future studies should focus on the following priorities. （1） Greenhouse gas flux observations should be conducted across permafrost peatlands with varying degrees of degradation. This requires establishing experimental setups with a gradient of warming intensities and extending the duration of field monitoring to capture long-term trends and interannual variability. （2） Multiple measurement techniques—such as static chamber-gas chromatography and eddy covariance analysis—should be employed to conduct multi-angle observations of soil carbon fluxes during both growing and non-growing seasons. Extending the temporal scope and frequency of monitoring is essential. Moreover， integrating field-measured environmental data into controlled laboratory incubation experiments will improve data accuracy and enhance mechanistic understanding. （3） Monitoring studies that trace the full pathway of DOC—from its production and mobilization to its ultimate mineralization into CO₂ and CH₄—should be implemented， and the underlying mechanisms， as well as the effects of multiple environmental factors and their interactions on soil carbon dynamics， should be elucidated. Addressing these research priorities will provide a more robust theoretical foundation for evaluating regional carbon cycle feedbacks， understanding the mechanisms of peatland soil carbon loss， and predicting future climate scenarios. Furthermore， the results will offer practical insights for supporting China’s carbon peaking and carbon neutrality strategy， informing ecological protection redline policies， and guiding the construction of ecological security barriers in the permafrost peatlands of the Greater Hinggan Mountains， Northeast China.

With global warming and the accelerated melting of sea ice， the navigability potential of the Arctic route continues to increase. Although the Arctic route offers significant economic benefits， it also poses serious threats to the local environment and climate. This review systematically reviews the variation characteristics of Arctic route navigability across historical， short-term， and future timescales， the methods for assessing Arctic navigation risks， and its economic， environmental， and climatic impacts. Currently， research on navigability mainly focuses on two aspects： first， using key quantitative indicators， such as navigable days， navigable windows， and route deviation， to evaluate the spatiotemporal evolution of Arctic route navigability under different scenarios from historical to future conditions； second， improving the accuracy of short-term navigability forecast to support operational navigation decisions. Research on navigation risk assessment focuses on developing risk identification frameworks for different accident scenarios， conducting navigation risk assessment and route planning across strategic， operational， and real-time navigation decision levels， and forming a technical logic chain of “risk identification-quantification-decision-making”. The focus of impact assessment is on the economic competitiveness of the Arctic route compared to traditional routes and their potential to reshape the global shipping patterns. It also explores the impact of shipping emissions on regional environment and climate， as well as the integrated emission reduction technology pathways of “fuel substitution， operational optimization， end-of-pipe treatment， and differentiated regulation.” Overall， this review clarifies current research progress on the Arctic route， identifies remaining knowledge gaps， and provides a scientific basis for climate-adaptive route planning， safe operational management， and the coordination of economic efficiency and environmental sustainability for the Arctic route.

With the intensification of global warming， glacier change has become a major issue concerning global water resource security and the development of human society. Located in the core area of the “Three Parallel Rivers”， Meili Snow Mountain is the mountain section with the most extensive distribution of low-latitude maritime glaciers in the Hengduan Mountains. Glaciers in this region are characterized by frequent avalanches， rapid movement， and high sensitivity to climate change. It is of great significance to study the response characteristics of these glaciers to climate change at a regional scale. Therefore， this study selected the modern glaciers in the Meili Snow Mountain region from 1990 to 2025 as the research object. Based on the GEE platform， eight periods of remote sensing images from May to October （the ablation season） were downloaded and synthesized. The ratio-threshold method and visual interpretation were employed to extract the glacier boundaries for each period， and the VOLTA model was used to simulate ice volume. Based on these data， the changes in glacier area and ice volume， as well as the distribution and variation characteristics of glaciers by scale， slope aspect， and altitude in the study area over the past 35 years were analyzed. Finally， combined with high-precision meteorological grid data， the response of glaciers to climate change in the Meili Snow Mountain region was explored. The results showed that： （1） over the past 35 years， a total of 57 glaciers were identified in the Meili Snow Mountain region. Both glacier area and ice volume exhibited a continuous retreating trend. The glacier area retreated from （131.51±6.14） km² （1990） to （113.67±6.13） km² （2025）， with an average annual change rate of -0.39%·a^-1. From 1995 to 2025， the ice volume gradually decreased from （12.74±0.61） km³ to （10.74±0.58） km³， with an average annual change rate of -0.52%·a^-1. Further statistical analysis of glacier area over the past 35 years by scale grade， slope aspect， and altitude revealed that the area of each individual glacier in the study area was less than 17 km²， and the number of glaciers with an area <2 km² was the largest. Glaciers on north-facing and northeast-facing aspects had the largest total area， while those on west-facing aspects had the smallest total area. The southeast-facing and south-facing aspects experienced the greatest retreat. The glacier area in the study area first increased and then decreased with increasing altitude， with glaciers mainly distributed in the altitude range of 4 400~5 800 m. （2） From 1980 to 2025， the warm-season average temperature in the study area showed a fluctuating upward trend， while warm-season precipitation exhibited a slight fluctuating decreasing trend. Considering the lag effect of glacier responses to climate change， correlation analysis of glacier area， ice volume， and meteorological factors indicated that glacier changes from 1990 to 2025 were primarily driven by rising temperatures， with precipitation playing a moderating role. It is predicted that glaciers in the study area will continue to accelerate their retreat over the next decade.

This study analyzed the frequency and intensity variation characteristics of cold wave processes in Altay， the “Snow Capital of China”， over the past 70 years （1954—2023）. Due to its unique geographical location and climatic conditions， Altay has become one of the regions in China with the most frequent occurrences of cold waves， which has a serious impact on the high-quality development of the local ice and snow economy. Based on the daily minimum temperature data from the Altay National Benchmark Climate Station， by defining the cold wave process， classifying its grades， and establishing statistical criteria， the long-term variation trends， seasonal and monthly distribution characteristics， intensity changes， and its relationships with other meteorological elements of cold waves were analyzed. The results showed that： （1） a total of 902 cold wave processes occurred in Altay from 1954 to 2023， with an average of 12.9 times per year. Winter was the main period for cold wave occurrences， accounting for as high as 68.0%. （2） The frequency of cold wave processes showed a significant decreasing trend， with a decreasing rate of -0.4 times·（10a）^-1 （P<0.05）， and the frequency of cold wave processes in autumn decreased most significantly. （3） In terms of cold wave grades， the occurrence frequencies of general， strong， and extremely strong cold waves accounted for 43.3%， 26.3%， and 30.4% respectively. Winter was dominated by extremely strong and strong cold wave processes， while spring and autumn were mainly dominated by general cold wave processes. （4） In terms of intensity changes， the average maximum 24-hour temperature drops for general， strong， and extremely strong cold wave processes were -7.7 ℃， -9.0 ℃， and -12.1 ℃， respectively， and the average cumulative temperature drop amplitudes were -11.2 ℃， -13.4 ℃， and -18.6 ℃ respectively. The maximum 24-hour temperature drop and cumulative temperature drop amplitudes of extremely strong cold wave processes were significantly stronger in winter than in other seasons. The comprehensive intensity of cold wave processes in winter was significantly higher than that in spring and autumn， being the strongest in December and the weakest in April on the monthly scale. （5） In addition， the comprehensive intensity of cold wave processes in spring and autumn showed continuous increasing and decreasing trends， respectively， while the comprehensive intensity of cold wave processes in winter has been showing a continuous upward trend since 1976. （6） The study also found that there was a good synchronization between higher winter precipitation， higher frequency of cold wave processes， and stronger comprehensive intensities of cold wave processes， with similarity rates as high as 84.3% and 77.1%， indicating a certain correlation between cold wave processes and precipitation. These findings provide a scientific basis for the monitoring， forecasting， disaster prevention， and mitigation of cold wave weather in the Altay region and contribute to a deeper understanding of the variation patterns of extreme weather events under the background of climate change.

As a nationally critical transportation hub， Henan Province is highly vulnerable to persistent low-temperature snowfall events. However， existing research on regional persistent low-temperature snowfall events in Henan has predominantly focused on case studies， while systematic investigations into the long-term variability of such events over extended time scales remain relatively scarce. Based on daily precipitation， average temperature， and minimum temperature data from meteorological stations in Henan Province during the winter half-years from 1961 to 2024， along with NCEP/NCAR reanalysis datasets， this study systematically analyzed the evolutionary characteristics of regional persistent low-temperature snowfall events in Henan. The K-means clustering method was employed to classify the 500 hPa geopotential height anomaly fields， enabling a refined classification of different circulation patterns along with their associated temperature and precipitation characteristics. Furthermore， the study revealed the upper- and lower-level circulation configurations for each pattern and explored the possible mechanisms driving the occurrence of such events. The results indicated that these events exhibited substantial interannual variability and pronounced interdecadal characteristics， with a notable increase in frequency and intensity after 2021. The events tended to occur in clusters， with regions south of the Yellow River identified as the most susceptible areas. The dominant circulation patterns were identified as the Ural Mountains-Okhotsk Sea double-blocking pattern， the Ural Mountains single-blocking pattern， and the Lake Baikal single-blocking pattern. Among these， the second pattern accounted for the highest proportion of events and has shown a significant increase in both frequency and duration since 1990， occurring predominantly in the late winter and featuring a broader core area compared to the other two patterns. The circulation configurations of the first two patterns were characterized by a positive phase of the Arctic Oscillation （AO）， a weakened mid- to high-latitude westerly jet， an intensified subtropical jet， and an anomalously strong Siberian High. In contrast， the third pattern exhibited a negative AO phase， with a stronger and eastward-extended westerly jet， and a northward-shifted subtropical jet that merged with the westerly jet. The Siberian High in this pattern was weaker than in the first two patterns but remained stronger than the climatological mean. Cyclonic circulation anomalies prevailed from the southern side of the Qinghai-Xizang Plateau to the Indochina Peninsula. Southerly anomalies at 850 hPa， combined with easterly anomalies from the Bohai Bay， provided abundant moisture conditions for these events. At the surface， the region was dominated by northerly winds， contributing to the formation of a temperature inversion structure， thereby providing the necessary thermal conditions for the development and persistence of low-temperature snowfall events. This study ultimately identifies the key circulation systems influencing regional persistent low-temperature snowfall events in Henan Province and elucidates the synergistic effects among these key circulation factors that collectively lead to the occurrence of such events. The diagnostic method established in this study can provide a scientific basis for the monitoring of regional persistent low-temperature snowfall events in Henan. Moreover， the primary circulation patterns derived from the objective classification offer significant reference for the accurate and advanced prediction of high-impact hazardous weather during the winter half-year in Henan Province. By addressing the gap in long-term systematic analysis and moving beyond individual case studies， this study contributes to a deeper understanding of the mechanisms underlying these extreme events and supports the enhancement of operational predictive capabilities for mitigating related disaster risks. Future efforts will focus on investigating the mechanisms underlying events occurring in the early winter versus the late winter， examining how the “Warm Arctic-Cold Eurasia” pattern influences the occurrence of such events， and exploring how different phases of the Arctic Oscillation contribute to the development of these events. These subsequent investigations aim to further advance the understanding of the physical processes governing persistent low-temperature snowfall events in Henan Province and to refine the predictive capabilities for these high-impact weather phenomena.

The uniaxial mechanical behavior of frozen soil is influenced by the complex coupling between its intrinsic physical properties and external environmental factors， resulting in a highly nonlinear and strongly path-dependent stress-strain relationship that poses significant theoretical challenges for accurate characterization and prediction. Constitutive models based on physical principles are limited by their underlying assumptions and often fail to effectively handle working conditions beyond their predefined theoretical frameworks. In contrast， data-driven methods， such as neural networks， have strong nonlinear fitting capabilities and can extract complex patterns from data. However， these methods often produce predictions that violate fundamental physical laws due to the lack of physical constraints and rely heavily on large amounts of high-quality labeled data， leading to weak generalization and poor interpretability. To address these challenges， particularly in the context of generally scarce experimental data on frozen soil， this study proposed a novel adaptive data-physics-driven neural network model. The core idea of this method was to systematically embed prior physical knowledge into the data-driven learning architecture， thereby enhancing the model’s expressive power while ensuring physical rationality. Specifically， normalized deformation modes were introduced as foundational physical constraints to guide the network toward a physically consistent solution space from the initial stage. A key innovation of this study was the proposal of a triple joint loss function to balance the trade-off between data-driven objectives and physical constraints. This composite loss function consisted of three independent components： （1） data-driven loss term （DE）： minimize the error between model predictions and experimental observations， ensuring fitting accuracy on known data； （2） data-physics guidance loss term （DP）： serve as a bridge between data and physical mechanisms， guiding the network to learn physically meaningful latent representations from limited samples and enhancing the model’s generalization ability and structural rationality； （3） physics-driven loss term （PE）： penalize predictions that deviated from established physical laws by introducing physical constraints， thereby strengthening the model’s physical consistency. Given that the optimal weights among these objectives were highly problem-dependent and difficult to preset， an adaptive weight optimization strategy was further adopted， employing the tree-structured Parzen estimator （TPE） to efficiently and automatically optimize the weights of the three loss components. This mechanism could dynamically adjust the influence of each constraint according to specific problem characteristics， thereby maintaining robust model performance under diverse loading conditions and material states. Furthermore， the proposed model exhibited good theoretical flexibility. By adjusting the loss weight configuration， it could degenerate into a purely data-driven neural network or evolve into a physics-based parameter identification method， covering a wide spectrum of modeling paradigms. Systematic validation based on uniaxial compression test data of frozen soil demonstrated that， compared with traditional constitutive models and standard neural network methods， the proposed adaptive data-physics fusion model achieved significant improvements in prediction accuracy. It not only generated predictions in close agreement with experimental results， but also maintained strong physical consistency and extrapolation capability， effectively mitigating the performance degradation of traditional data-driven models under unseen working conditions. This study provides a new modeling approach for the mechanical behavior of frozen soil and other complex geotechnical materials under extreme conditions. The proposed method integrates physical interpretability， data efficiency， and superior generalization ability.

The northeastern transitional zone of the Qinghai-Xizang Plateau， located on the first topographic step of China， is a key area where climate sensitivity and geological instability intertwine. This region is extensively covered by strongly weathered mudstone. The strongly weathered mudstone in this area is subject to wet-dry and freeze-thaw cycles driven by glacial meltwater， concentrated rainfall， and significant seasonal temperature differences. Under the long-term and repeated action of multiple physical weathering forces， the rock mass exhibits pronounced weathering characteristics. This process also leads to the continuous accumulation of a large amount of fine-grained weathering products， significantly altering the physical composition and mechanical structure of slopes， thereby inducing and intensifying frequent geological disasters. To reveal the intrinsic mechanisms of this geological disaster process， wet-dry cycles were conducted using a soaking-drying method. In each cycle， the sample was placed in a saturator and completely immersed in water at room temperature （around 20 ℃） for 12 hours to ensure full moisture penetration into its internal voids. To ensure that the samples reached a fully saturated state， their saturation was measured using the mass method after the soaking stage. The saturated samples were then placed in an oven at a constant temperature of 50 ℃ for 12 hours for drying. This process constituted one complete wet-dry cycle. The freeze-thaw cycle tests employed a “low-temperature freezing-room-temperature thawing” method. Each cycle consisted of a 12-hour freezing stage and a 12-hour thawing stage. During the freezing stage， the samples were placed in a low-temperature refrigerator with the temperature stabilized at -20 ℃ to ensure complete freezing of their internal pores. During the thawing stage， the samples were transferred to a constant-temperature laboratory environment maintained at （18 ± 2） ℃ for thawing. To prevent moisture loss during the cycles， the samples were wrapped with cling film before testing. Laboratory triaxial shear tests were combined with scanning electron microscopy （SEM） to quantitatively extract the key microstructural parameter of pore area ratio. Based on this， a soil area damage coefficient （D_i ） was established and further analyzed in combination with the pore area ratio， probability entropy， and fractal dimension. This approach revealed the deterioration pattern of the mechanical properties of strongly weathered mudstone under cyclic loading. The results showed that the cyclic tests caused significant damage to the mechanical properties of the mudstone. The peak strength of the weathered mudstone decreased with an increasing number of cycles. After undergoing 3， 5， 7， and 10 wet-dry cycles， the cohesion of the mudstone decreased by 14.91%， 34.74%， 49.36%， and 54.78%， respectively， compared to the initial samples. The corresponding pore area ratios were 47.51%， 49.36%， 53.14%， and 58.88%， and the D_i values were 0.909， 0.866， 0.779， and 0.647. For the freeze-thaw cycles， after the mudstone underwent the same number of cycles （3， 5， 7， and 10）， its cohesion decreased by 26.32%， 28.27%， 31.95%， and 32.07% respectively， compared to the initial samples. The corresponding pore area ratios were 47.49%， 52.10%， 54.65%， and 56.89%. The D_i values were 0.909， 0.903， 0.707， and 0.693. Meanwhile， a sudden drop in strength began to occur in the strongly weathered mudstone when the confining pressure reached 100 kPa. Finally， the influence of wet-dry and freeze-thaw cycles on the mudstone was further revealed through the quantitative microscopic data， including pore area ratio， pore probability entropy， and pore fractal dimension. A significant correlation was observed between the attenuation of the mechanical properties of the mudstone and the increase in the number of wet-dry or freeze-thaw cycles. The research findings not only deepen the theoretical understanding of the disaster mechanisms of rock and soil masses in the marginal zone of the Qinghai-Xizang Plateau under the influence of cold and drought environments， but also provide important quantitative evaluation indicators and a theoretical basis for the prevention and control of related geological disasters. The proposed D_i can serve as an indicator to evaluate the weathering damage state of rock and soil masses. When combined with the identified critical confining pressure， it can provide reference data for slope stability evaluation in this area and offer important theoretical support for the prevention， monitoring， and early warning of geological disasters.

This study aims to explore the dynamic mechanical properties and damage constitutive model of fine sandstone under low-temperature conditions， address the current research gap regarding the dynamic mechanical properties of Jurassic-Cretaceous strata under different negative temperatures， clarify the influence of freezing temperature differences on rock mechanical properties， and provide theoretical support and experimental evidence for engineering construction in cold regions. Fine sandstone from the Jurassic-Cretaceous strata was selected as the research object. Dynamic impact compression tests were conducted on fine sandstone specimens at different negative temperatures （-5 ℃， -10 ℃， -15 ℃， -20 ℃， -25 ℃） using a split Hopkinson pressure bar （SHPB） test system. During the tests， the full stress-strain curves of fine sandstone under different temperature conditions were accurately obtained. The relationship between the dynamic peak stress of the specimens and temperature， as well as the energy dissipation patterns， were systematically analyzed. By introducing the fractal dimension D， the quantitative relationship between the fragment size fractal dimension and temperature was explored. Based on the assumption that the micro-element strength of fine sandstone followed the Weibull statistical distribution， combined with the Drucker-Prager （D-P） failure criterion， the key parameters F₀ and m were modified. Consequently， a dynamic strength-based statistical damage constitutive model of fine sandstone under different temperatures was established. The rationality of the model was verified by comparing the experimental curves with the theoretical curves.The research results showed that under impact loading， the stress-strain curves of fine sandstone at different temperatures could be divided into four stages： compaction， linear deformation， plastic deformation， and rapid unloading. The characteristics of each stage showed regular changes with decreasing temperature. Under impact loading， the overall dynamic peak stress of fine sandstone increased with decreasing temperature， demonstrating a clear temperature dependence. The growth trend of peak stress exhibited a phased pattern， characterized by a rapid initial increase followed by a deceleration. Compared with the value at -5 ℃， the peak stress at -10 ℃， -15 ℃， -20 ℃， and -25 ℃ increased by 18.1%， 55.2%， 71.7%， and 74.8%， respectively. In terms of energy evolution， with the decrease of temperature， the transmitted energy showed a trend opposite to that of the dissipated energy and reflected energy. Specifically， the dissipated energy and reflected energy increased monotonically， while the transmitted energy decreased monotonically. The fractal dimension of the fragmented fine sandstone under impact loading gradually increased with the decrease of temperature. The correlation coefficient between the fractal dimension of the crushed specimens and temperature reached 0.936， indicating a good exponential relationship between them. The dynamic strength-based statistical damage constitutive model of fine sandstone， modified based on the relationship between Weibull distribution parameters （F ₀， m） and temperature， had a high degree of agreement between theoretical and experimental curves. This model could accurately describe the stress-strain relationship and damage evolution process of fine sandstone under impact loading at different negative temperatures， thereby verifying its rationality.The findings of this study effectively improve the theoretical framework of low-temperature rock dynamics. They provide new experimental data and theoretical references for a deeper understanding of the dynamic failure mechanism of frozen rocks and for clarifying the influence of low temperature on the mechanical properties of fine sandstone. Furthermore， they provide important technical support for rock dynamic stability evaluation， blast impact design， and surrounding rock control in applications such as cold-region engineering， artificial frozen shafts， and deep low-temperature rock mass engineering. Therefore， these findings hold important theoretical significance and practical engineering value for improving the safety， reliability， and high-quality development of cold-region engineering construction.

In the construction of transportation engineering infrastructure in the seasonally frozen soil regions of Northwest China， a certain proportion of loess is usually incorporated to improve the poor engineering geological properties of red-bed mudstone， thereby optimizing and enhancing the long-term service performance of foundation fill materials in this special engineering area. The evolutionary patterns of the mechanical properties and deformation resistance of red-bed mudstone-loess mixtures under the continuous action of freeze-thaw cycles have long been a core research focus in cold-region geotechnical engineering， as they are closely related to the structural stability and long-term safety of subgrade and foundation engineering in frozen soil regions. However， systematic and in-depth research on this scientific issue remains relatively scarce in the academic community. To comprehensively clarify the specific influence patterns of freeze-thaw cycles on the mechanical behavior of red-bed mudstone-loess mixtures with different mixing ratios， this study designed a systematic phased experimental scheme and carried out a series of indoor geotechnical tests. First， standard compaction tests were conducted in strict accordance with geotechnical engineering testing specifications， with loess content as the independent variable， to determine the influence of different loess contents on two key compaction indicators of the mixture—maximum dry density and optimal water content. Based on the results of the compaction tests， the triaxial compression test method， which was widely used in geotechnical research， was adopted to investigate the evolution patterns of the compressive strength and creep characteristics of pure loess and pure red-bed mudstone samples prepared under the conditions of maximum dry density and optimal water content under the action of freeze-thaw cycles. Furthermore， on the basis of single-soil sample tests， this study conducted repeated freeze-thaw cycle tests on red-bed mudstone-loess mixture samples with different mixing ratios under the conditions of maximum dry density and optimal water content， and then carried out systematic triaxial compressive strength tests and triaxial creep tests on all mixture samples. Through comprehensive analysis and comparison of the large amount of data generated from all the above tests， the results showed that loess content had a significant regulatory effect on the compaction performance of red-bed mudstone-loess mixtures. With increasing loess content， the maximum dry density of the mixture first increased and then decreased， reaching its peak at a loess content of 50%， while the optimal water content of the mixture changed only slightly and gradually with loess content. Freeze-thaw cycles had a significant differential impact on the triaxial strength of loess and red-bed mudstone. The influence of freeze-thaw cycles on the triaxial compressive strength of loess was slight， but it significantly reduced the triaxial compressive strength of red-bed mudstone， and the degree of strength attenuation of red-bed mudstone increased sharply with the number of freeze-thaw cycles. In terms of creep deformation characteristics， the axial creep deformation of both loess and red-bed mudstone generally increased with the number of freeze-thaw cycles， but the magnitudes of the increase differed significantly， with the deformation increment of red-bed mudstone being far more significant than that of loess. For red-bed mudstone-loess mixtures， when the loess content was 20%， the short-term compressive strength and long-term deformation resistance of the mixture were both significantly improved compared with pure red-bed mudstone. If the loess content was further increased beyond 20%， the mechanical properties of the mixture continued to improve， but the improvement magnitude and growth rate slowed down significantly. Therefore， in the practical engineering of improving the bearing performance of red-bed mudstone foundations with loess in the seasonally frozen soil regions of Northwest China， it is suggested that the loess content should not be less than 20%. If local economic conditions permit， construction costs are controllable， and the engineering project has higher requirements for the mechanical reliability of fill materials， the loess content can be further increased above 20% to achieve more excellent and stable engineering performance under the action of freeze-thaw cycles.

This study explores the feasibility of applying microbially induced carbonate precipitation （MICP） technology to the remediation of heavy metal pollution in coal gangue from cold-region mining areas. Coal gangue in such areas is frequently subjected to repeated freeze-thaw cycles， which gradually weaken its structure and increase the mobility of heavy metals. These environmental conditions pose significant risks， particularly in mining areas where large volumes of gangue are stored. Therefore， it is necessary to verify whether MICP can maintain stable solidification performance under freeze-thaw conditions. The purpose of this study is to provide technical guidance and a theoretical basis for the application of MICP in heavy metal pollution control in cold-region mining areas.In the experimental program， MICP was employed to solidify heavy metal ions in coal gangue， with Pb²⁺ concentration selected as the primary variable for specimen preparation. Specimens containing different Pb²⁺ concentrations were prepared， and the calcium carbonate content was measured to evaluate the degree of microbial mineralization. This measurement was critical because the amount of carbonate formed directly reflected the effectiveness of the microbial process under varying chemical conditions. To investigate the influence of freeze-thaw cycles， the specimens were subjected to different numbers of cycles. After each designated cycle， penetration strength， mass loss rate， and Pb²⁺ leaching concentration were measured. These indicators were selected because they represented different aspects of material performance. Penetration strength reflected mechanical integrity， mass loss rate indicated structural degradation， and leaching tests revealed the stability of solidified heavy metals.X-ray diffraction （XRD） and scanning electron microscopy （SEM） were used to analyze the mineral phases formed during MICP and to observe the morphology and distribution of precipitates on the surfaces of coal gangue particles. These analyses were intended to determine whether freeze-thaw cycles altered the mineral composition of MICP products or primarily affected their physical structure. SEM observations provided detailed information about how carbonate precipitates filled pores or coated particle surfaces， and how these microstructures changed after repeated freezing and thawing. Such microstructural insights were essential for understanding the mechanisms behind strength loss and recovery during freeze-thaw cycles.The results showed that lower Pb²⁺ concentrations promoted MICP mineralization， leading to increased calcium carbonate production. This suggested that small amounts of Pb²⁺ did not inhibit microbial activity and may even have facilitated nucleation. However， as Pb²⁺ concentration continued to rise， carbonate formation decreased， indicating that excessive Pb²⁺ interfered with microbial processes or crystal growth. Under freeze-thaw conditions， the mass loss rate reached its maximum after seven cycles， while penetration strength dropped to its lowest point after nine cycles. Beyond these points， both indicators exhibited partial recovery. This opposite trend indicated that freeze-thaw damage and subsequent densification occurred simultaneously. Early cycles primarily caused structural deterioration， whereas later cycles may have led to particle rearrangement or secondary precipitation， which slightly improved compactness.The solidification efficiency of Pb²⁺ remained above 85% after MICP treatment， demonstrating that the technique performed well under low to moderate Pb²⁺ concentrations. When the Pb²⁺ concentration was below 300 mg·kg^-1 and the number of freeze-thaw cycles did not exceed nine， the solidified specimens maintained good structural stability and solidification performance. Although freeze-thaw cycles did not alter the mineral phases of the MICP products， repeated cycles damaged the carbonate framework and the bacterial coating formed during the process. This structural weakening increased the likelihood of Pb²⁺ leaching and reduced the mechanical strength of the solidified material. These findings highlight the importance of considering long-term freeze-thaw durability when applying MICP in cold-region mining areas.Overall， the study demonstrates that MICP has considerable potential for stabilizing heavy-metal-polluted coal gangue in cold-region mining areas. This technology can achieve high solidification efficiency and maintain structural stability under certain conditions. However， its performance is sensitive to Pb²⁺ concentration and the number of freeze-thaw cycles. The dual effects of freeze-thaw damage and densification must be carefully considered， as they influence both mechanical behavior and leaching characteristics. Future efforts should focus on long-term durability assessments， optimization of microbial activity under varying chemical environments， and scaling up laboratory findings to field applications. By addressing these issues， MICP could become a reliable and sustainable method for heavy metal pollution control in cold-region mining areas.

In artificial ground freezing， the frozen wall serves as a key structure for maintaining underground excavation stability and water-sealing performance. The rational control of its thickness is of great importance for ensuring excavation safety， reducing the risk of surrounding ground deformation， and mitigating frost heave and thaw settlement. An insufficient frozen wall thickness may result in inadequate load-bearing capacity and increased leakage risk. In contrast， an excessively thick frozen wall can lead to cooling energy waste and may aggravate frost heave and thaw settlement. However， systematic studies on the heat exchange mechanism between the outer boundary of the frozen wall and the surrounding unfrozen ground remain relatively limited. In particular， under no-groundwater-seepage conditions， the heat exchange patterns during the stable formation stage of frozen wall thickness have not been clearly revealed， thereby limiting the theoretical and quantitative development of frozen wall thickness control methods. To address these issues， this study established a numerical simulation method based on COMSOL Multiphysics to describe the water-heat phase change behavior during ground freezing， and to systematically simulate the formation and evolution process of the frozen wall under artificial freezing conditions. Under no-seepage conditions， the heat exchange characteristics between the outer boundary of the frozen wall and the surrounding ground during freezing were analyzed. The results indicated that when the ground temperature field gradually developed and eventually reached a steady state， the radial heat flux in the soil around the frozen wall exhibited pronounced spatial consistency. Based on this stable heat transfer characteristic， the concept of “cooling energy demand of the frozen wall” was proposed and defined as the steady-state heat flux between the outer boundary of the frozen wall and the surrounding unfrozen soil when the ground temperature field was in steady state. On this basis， four main influencing factors—initial ground temperature， soil particle thermal conductivity， target frozen wall radius， and soil moisture content—were selected to conduct an orthogonal numerical experimental study on the cooling energy demand of the frozen wall. The influence patterns of these factors on the cooling energy demand were systematically analyzed， and a corresponding prediction model was established. The results showed that the established water-heat phase change numerical simulation method effectively reflected the heat exchange process between the outer boundary of the frozen wall and the surrounding unfrozen soil， meeting the requirements for calculating and analyzing the cooling energy demand of the frozen wall. When the frozen wall thickness approached a stable state， the heat flux at any radial cross-section in the ground remained nearly constant. This value was defined as the cooling energy demand of the frozen wall， namely， the heat flux between the outer boundary of the frozen wall and the surrounding unfrozen soil under stable frozen wall thickness conditions. The degrees of influence of different factors on the cooling energy demand of the frozen wall， from large to small， were： initial ground temperature， soil particle thermal conductivity， target frozen wall radius， and moisture content. Among them， initial ground temperature and soil particle thermal conductivity were the dominant controlling factors， with contribution proportions of approximately 70% and 24%， respectively. The cooling energy demand of the frozen wall exhibited a significant linear increasing trend with each dominant factor. Based on the above patterns， a prediction model for the cooling energy demand of the frozen wall considering the influence of the dominant factors was established. This model can effectively characterize the variation characteristics of cooling energy demand of the frozen wall under no-seepage conditions， and can provide a reliable theoretical basis and technical support for construction process design， freezing parameter selection， and energy demand assessment in artificial ground freezing under no-seepage conditions.

In freezing method construction， the thickness of the freezing wall serves as the critical factor determining structural strength and soil deformation. Coastal strata exhibit asymmetric freezing wall morphology due to seepage effects， necessitating precise design of freezing wall thickness. Focusing on freezing wall thickness design under different seepage conditions， this study， based on the Fuzhou Metro connecting passage project， and established the ultimate bearing capacity of the freezing wall and the threshold of surface frost heave deformation as criteria for preliminary and optimal selection of freezing wall thickness， respectively. Numerical simulations were employed to identify design intervals for freezing wall thickness applicable to seepage-affected strata. For high seepage conditions， an optimized freezing pipe layout method was developed to ensure that the designed thickness simultaneously satisfies structural strength and surface deformation control requirements. The results demonstrated that surface frost heave deformation exhibited positive correlation with freezing wall thickness. Seepage prolonged freezing wall formation by impeding cold energy transfer. Conventional designs fail to meet dual requirements of strength and deformation control when seepage velocity （v） ≥0.9 m·d^-1. By using the method of arched arrangement of freezing pipes at the top of the connecting passage and local densification of freezing pipes on the upstream side， construction risks caused by enhanced seepage-induced surface frost heave deformation can be effectively mitigated. After optimization， the freezing wall thickness can simultaneously satisfy strength and deformation control requirements even at v=1.0 m·d^-1.

Freeze-thaw cycles reduce the integrity and cause structural damage to the widely distributed clay on the Qinghai-Xizang Plateau， aggravate the generation and expansion of soil cracks， form preferential flow paths dominated by cracks， accelerate water infiltration， affect the surface energy balance. These may further increase the frequency and severity of phenomena such as alpine meadow degradation， subgrade slope erosion， failure of frozen soil slope retaining structures， and thermal thaw slumping， posing a serious threat to the infrastructure safety and ecological environment of the Qinghai-Xizang Plateau. Although current research has made some progress in characterizing the geometric morphology of cracks under freeze-thaw actions， the essential premise of soil cracking is drying shrinkage caused by evaporation. There is still a lack of systematic and qualitative research on the mechanism of how the crack network and water evaporation interact and promote each other. Therefore， this study aims to clarify， through systematic laboratory tests， the effects of freeze-thaw cycles on the evaporation process and cracking behavior of Qinghai-Xizang clay and the interaction mechanisms between them. Clay samples with three initial water contents （16%， 23%， 30%） and a dry density controlled at 1.53 g⋅cm^-3 were prepared. The samples were subjected to 0， 1， 5， and 10 freeze-thaw cycles using the full freezing method. The freeze-thaw cycle was set as freezing at -20 ℃ and thawing at 20 ℃ for 12 hours each to simulate the seasonal variation characteristics of the plateau area. After the freeze-thaw cycles， continuous evaporation tests were carried out to systematically reveal the influence of prior freeze-thaw cycles on the evaporation and cracking characteristics of Qinghai-Xizang clay. During evaporation， an electronic balance with an accuracy of 0.01 g was used to monitor the sample mass every 2 hours to obtain the dynamic change of water content， thereby dividing the evaporation stages and analyzing the evolution of duration and evaporation rate. At the same time， a high-definition digital camera was used to photograph the surface crack development morphology of the samples until evaporation stabilized. Based on crack image analysis software， crack parameters including surface crack ratio， total crack length， average crack width， and fractal dimension were quantitatively extracted to systematically explore the crack evolution behavior under freeze-thaw cycles. In addition， scanning electron microscopy （SEM） was used to observe the changes in soil microstructure， revealing the effects of freeze-thaw cycles on pore structure and micro-cracks. The results showed that the evaporation process of the soil exhibited distinct stage characteristics and could be divided into a constant-rate stage， a deceleration stage， and a residual stage. When the initial water content was low， the constant-rate stage disappeared， and the evaporation process showed only two distinct stages. The time required for the soil to reach the stable stage of evaporation increased with initial water content， while the initial evaporation rate increased with the number of freeze-thaw cycles. The surface crack morphology of the soil became more complex with increasing initial water content， forming a denser and more interconnected network crack structure. Freeze-thaw cycles significantly promoted the development of surface cracks on the soil， and each crack parameter was positively correlated with both the number of freeze-thaw cycles and the initial water content. When the number of freeze-thaw cycles reached 10 and the initial water content was 30%， the extent of crack development was the most significant， with surface crack ratio， total crack length， average crack width， and fractal dimension reaching 2.4%， 57.4 cm， 0.046 cm， and 1.52， respectively. Microstructural observations showed that freeze-thaw cycles led to weakened inter-particle bonding and pore reorganization， providing pathways for water migration and evaporation. The evaporation process further aggravated water loss and soil shrinkage， thus driving crack propagation， forming a feedback mechanism of “freeze-thaw-crack-evaporation” synergistic interaction among the three. This study reveals， from macro-to-micro multi-scales， the interaction among water migration， evaporation and drying， and crack development in Qinghai-Xizang clay under the influence of freeze-thaw cycles， providing experimental evidence and theoretical support for a deeper understanding of the water-heat-mechanical coupling process in cold-region soil. The relevant conclusions have important reference value for subgrade slope protection， geological hazard risk assessment， and ecological protection and restoration in frozen soil regions of the Qinghai-Xizang Plateau.

The coefficient of subgrade reaction is a critical parameter for calculating deformation and settlement of roads and bridges using the elastic foundation beam method. However， current codes regard this coefficient as a constant without considering the effects of freeze-thaw cycles， which is particularly problematic in seasonally frozen soil regions where loess is widely distributed. In these regions， drastic temperature changes cause repeated freeze-thaw of the soil， leading to significant changes in mechanical properties and large deformations that may induce longitudinal and transverse cracks in pavements， severely compromising pavement stability and reducing road service life and safety. To investigate the influence of freeze-thaw environments on the coefficient of subgrade reaction of loess， this study selected sandy loess from Guyuan， Ningxia as the research material. Four groups of remolded soil specimens with different frost heave ratios were prepared by controlling sand content and water content， named as soil specimen I， soil specimen II， soil specimen III， and soil specimen IV. Tests were conducted at temperatures of 10 ℃， 0 ℃， -10 ℃， and -20 ℃， with freeze-thaw cycles of 0， 1， and 2 times， a fixed confining pressure of 100 kPa， and a shear rate of 1 mm•min^-1. The tests were combined with finite element analysis to comprehensively investigate the mechanical behavior and deformation characteristics of loess specimens under various environmental conditions. The experimental results demonstrated that the peak strength of the stress-strain curve increased significantly with decreasing temperature but decreased with increasing number of freeze-thaw cycles. This phenomenon was attributed to the phase change of soil water. Liquid water transformed into ice crystals at lower temperatures， forming ice cementation that enhanced the bonding between soil particles and increased the bearing capacity. Conversely， freeze-thaw cycles caused irreversible structural damage and increased porosity， leading to a gradual reduction in strength. The failure modes of non-frost-heaving soil and frost-heaving soil exhibited significant differences due to fundamental differences in pore structure and particle cementation state. Non-frost-heaving soil exhibited brittle failure， characterized by a rapid stress drop after peak stress， no obvious plastic deformation phase， and the formation of shear bands with an inclination angle of approximately 45° and sharp fracture surfaces. In contrast， frost-heaving soil exhibited ductile failure with gradual stress reduction or stabilization， a prolonged plastic deformation phase， and diffuse shear bands with a width of approximately 5~8 mm. Soil specimens with lower frost heave ratios had higher coefficients of subgrade reaction but were prone to brittle failure， posing potential safety risks in engineering applications. The damage to soil specimens was most significant during the first five freeze-thaw cycles and gradually stabilized after 15 cycles， indicating that freeze-thaw damage had a cumulative effect and a diminishing marginal effect. To accurately characterize the freeze-thaw damage effect， an elastoplastic constitutive model incorporating a damage factor was developed based on the Mises yield criterion. The model incorporated the effects of the number of freeze-thaw cycles and water content on the elastic modulus and yield stress through the damage factor， and the damage factor had a logarithmic relationship with the number of freeze-thaw cycles. The current elastic modulus was modified by both the water content influence coefficient and the freeze-thaw damage factor， while the yield stress was modified by the confining pressure， water content， and freeze-thaw damage. Finite element analysis was performed using ABAQUS， and the stress-strain behavior was simulated using a user-defined material subroutine UMAT to calculate the coefficient of subgrade reaction for up to 20 freeze-thaw cycles. The three-dimensional finite element model consisted of 7 950 elements with dimensions of 50 mm×100 mm， representing the triaxial compression test specimen. The comparison between experimental and simulation results showed that the error was within 15%， validating the reliability and accuracy of the finite element model in predicting soil behavior under freeze-thaw conditions. Based on experimental data and finite element analysis， this study proposed freeze-thaw deterioration coefficients for the first time. For non-frost-heaving soil， the deterioration coefficient was 0.6， and for frost-heaving soil， it was 0.7. These coefficients represented the ratio of the coefficient of subgrade reaction after 20 freeze-thaw cycles to that in the unfrozen state. Considering the freeze-thaw deterioration effects， the recommended values of the coefficient of subgrade reaction in seasonally frozen soil regions were 32~39 MPa⋅m^-1 for non-frost-heaving soil and 6~7 MPa⋅m^-1 for frost-heaving soil. These recommended values were significantly lower than those in current standards and relevant literature， which did not consider the freeze-thaw effects. The findings of this study provide important technical support for determining appropriate coefficients of subgrade reaction in seasonally frozen soil regions and highlight the necessity of considering freeze-thaw damage in engineering practice to ensure construction safety and extend the service life of infrastructure in freeze-thaw environments.

The implementation of carbon sequestration in permafrost areas using the hydrate method is a stable and high-capacity new carbon sequestration technology. The exploration of the mechanism of the hydrate formation process is the cornerstone that affects the implementation efficiency and capacity of the sequestration project. Clay， as a key component in permafrost strata， has no unified conclusion on its influence mechanism on the formation characteristics of hydrates. This study takes the influence of series concentration （1 wt%~5 wt%） illite on the formation kinetics of hydrates as the research topic， and combines two experimental systems： clay suspension droplets and clay-containing porous media. The kinetic and morphological studies on the nucleation and growth process of hydrates were carried out. The experimental results show that the multiple nucleation sites， lamellar structure and surface charging characteristics of clay accelerate the nucleation rate of hydrates， reducing the induction time in the droplet experimental group by up to 365.5 minutes. However， its high viscosity characteristics have a negative impact on the formation of hydrates and the final conversion rate. From a morphological perspective， it is found that clay hydrates are more compact and structurally regular， and their inhibitory effect on gas-phase diffusion significantly weakens the conversion rate of hydrates. Under the promoting effect of illite on the initial growth of hydrates， throat blockage is prone to occur in porous medium reservoirs with narrow pores， which hinders the further formation of hydrates. This study provides new insights into the mechanism of hydrate formation under the influence of illite and offers reference suggestions for the engineering design of carbon sequestration injection.

The evolution of unfrozen water content in red-bed weakly expansive soil during freeze-thaw cycles is one of the key factors affecting the heave/subsidence of pile-supported high-speed railway cuttings. In this study， nuclear magnetic resonance （NMR） was employed to conduct experiments on the unfrozen water content of weakly expansive soil during the freeze-thaw process. The evolution characteristics of the total unfrozen water content and the unfrozen water content in micropores， mesopores， and macropores were analyzed with respect to temperature， swelling value （quantified swelling potential）， and initial water content. This study aims to reveal the variation patterns of water retention in pores of different sizes of weakly expansive soil after the freeze-thaw process， to investigate the hysteresis effect of unfrozen water during the freeze-thaw cycles， and to establish a predictive model for unfrozen water content under freeze-thaw conditions. The results showed that during the freeze-thaw cycles， the unfrozen water content of weakly expansive soil underwent an intense phase change within the temperature range of -2 °C to -4 °C. Specifically， during freezing， the unfrozen water content transitioned through three stages with decreasing temperature： unfrozen， rapid freezing， and stable freezing. During thawing， it progressed through three stages with increasing temperature： slow thawing， rapid thawing， and stable thawing， presenting the characteristic that “water in macropores froze first， while water in micropores thawed first”. The greater the swelling value of the sample， the lower its freezing and thawing points. Moreover， during the freeze-thaw process， the unfrozen water contents in micropores and mesopores， as well as the total unfrozen water contents， were higher. In contrast， the unfrozen water content in macropores was lower. A higher initial water content led to higher freezing and thawing points， as well as greater unfrozen water content in micropores， mesopores， and macropores. The pore water in micropores， mesopores， and macropores accounted for approximately 40%， 55%， and 5% of the total pore water， respectively. After the freeze-thaw process， the pore water content in micropores and macropores increased， while that in mesopores decreased. The increment in micropore water content reached its maximum at a swelling value （Z） of 40， accounting for 2.8% to 4.2% of the initial water content. This indicated that Z=40 represented a critical point： below this value， the clay mineral content was insufficient， hydrophilicity was weak， and few micropores were generated； above it， excessive clay minerals enhanced overall expansibility， resulting in dense particle packing， a reduction in macropores available for division after the freeze-thaw process， and consequently a limited increase in micropores. The variations in mesopore and macropore water content reached the maximum at an initial water content （w） of 15%， accounting for 2.8% to 7.1% of the initial water content. The swelling value and initial water content significantly influenced the hysteresis effect of unfrozen water in weakly expansive soil. A larger swelling value or a higher initial water content corresponded to a greater degree of hysteresis. Moreover， the hysteresis effect exhibited distinct temperature dependence， first increasing and then decreasing with rising temperature， with the maximum hysteresis degree typically occurring between -2 ℃ and -4 ℃. Finally， to accurately predict the variations in unfrozen water content， the correlations between three key factors （swelling value， initial water content， and temperature） and unfrozen water content were quantified based on Pearson correlation analysis， leading to the establishment of a predictive model incorporating these variables. For approximately 90% of the data， the errors between the theoretical values from the model and the experimental values were less than 15%. The model demonstrates good agreement between calculated and measured values， providing effective theoretical support and data reference for evaluating and analyzing the freeze-thaw characteristics of weakly expansive soil in engineering practice.

Nitrous oxide （N₂O） is a potent long-lived greenhouse gas with an atmospheric lifetime of （116±9） years and a global warming potential （GWP） approximately 310 times that of carbon dioxide （CO₂） over a 100-year timescale. While rivers are recognized as significant N₂O sources， current emission estimates remain highly uncertain， particularly in understudied high-altitude glacier-fed river regions. These sensitive systems are vital lifelines in arid regions and possess unique environmental conditions （e.g.， low temperatures， high turbulence， oligotrophic status） that may drive the microbial pathways of nitrogen cycling （e.g.， nitrification and denitrification） in a manner distinctly different from other systems. Accelerated glacier retreat under global warming further alters their hydrology， sediment load， and biogeochemical processes， yet the specific impacts of these changes on N₂O dynamics remain unclear. This knowledge gap limits the ability to accurately assess climate feedback effects associated with cryospheric degradation. This study presented a systematic investigation of N₂O dynamics in the Urumqi River， a typical glacier-fed river in the Tianshan Mountains of western China. Systematic field monitoring was conducted during the 2021 ablation season （June 28 to October 4） at two key sites： the terminus of Urumqi Glacier No. 1 （the source， sampled every 1~2 weeks） and the main stem of the Urumqi River （high-frequency daily sampling）. Dissolved N₂O concentrations were determined using the headspace equilibrium method and analyzed by gas chromatography equipped with an electron capture detector （ECD）. Water-atmosphere N₂O diffusive fluxes were calculated using the thin boundary layer model， which was parameterized based on in-situ wind speed and water temperature data to determine the gas exchange velocity （k）. A comprehensive set of hydrochemical parameters （including NO₃ ^-， TDS， pH， major ions such as Na⁺， Cl^-， Ca²⁺， and SO₄ ^2-， and total nitrogen） was measured concurrently. To identify key driving factors and manage multicollinearity， a robust statistical approach was employed， namely stepwise multiple linear regression （SMLR） based on the Akaike information criterion （AIC） and relative importance analysis （LMG method）. Regional-scale estimation was achieved using river surface area extracted via the Google Earth Engine （GEE） and the JRC Global Surface Water （GSW） dataset. The results demonstrated that the Urumqi River system was a persistent source of atmospheric N₂O throughout the entire observation period. The glacier terminus was identified as a significant “hotspot” of N₂O production with a mean N₂O concentration ［（18.1±1.24） nmol⋅L^-1］ significantly higher than that in the downstream main stem ［（14.8±0.41） nmol⋅L^-1； P < 0.05］. This difference was even more pronounced in flux. The mean flux at the glacier terminus ［（182.1±13.31） μmol⋅m^-2⋅d^-1］ was much higher than that at the downstream reach ［（112.08±4.59） μmol⋅m^-2⋅d^-1； P < 0.001］. This “source higher than downstream” spatial pattern suggested that subglacial and proglacial zones were critical sites of N₂O production. Elevated N₂O concentrations may originate from microbial processes （e.g.， denitrification） in anoxic subglacial environments， as well as from biogeochemical activity in newly exposed， nutrient-rich proglacial sediments flushed by meltwater. The SMLR and LMG analyses revealed that N₂O dynamics were primarily controlled by nitrate （NO₃ ^-）， representing substrate availability （explaining 36.5% of the flux variance）， and total dissolved solids （TDS）， reflecting catchment-scale water-rock interaction intensity （explaining 32.7% of the variance）. The high correlation with TDS suggested that dissolved ions from chemical weathering may also support microbial metabolism by providing essential nutrients or maintaining osmotic balance. The weakly alkaline water （mean pH 7.95） also contributed to the model （about 10% contribution）， potentially enhancing the metabolism of nitrifying bacteria that thrived under neutral to alkaline conditions. The estimated total annual N₂O emission from the upper Urumqi River Basin was 2.4 tons， with 54.0% occurring during the emission window in late summer and early autumn （August-October）. Compared with other natural high-altitude river systems worldwide， the N₂O flux in this region was at a relatively high level， indicating that proglacial environments may be globally under-recognized but important N₂O production areas. These findings provide a critical observational basis for understanding nitrogen cycling in alpine glacial meltwater systems and for more accurately assessing N₂O budgets in mountainous regions under rapid climate change.

Surface water and groundwater， as a hydrological continuum， are characterized by frequent transformation， which constitutes the core of river basin water balance and precise water resources management. In terms of simulation methods， compared with distributed physical models featuring extremely high computational costs， semi-distributed coupled models represented by SWAT+ strike a balance between physical accuracy and computational efficiency， making them more suitable for large-scale and long-term scenario analysis. At present， although the macroscopic water cycle in the arid regions of Northwest China has been clarified， research on the quantitative characterization of surface water-groundwater interaction at the grid scale and across the “mountain-plain” system remains scarce. Taking the Aksu River Basin as a case study， cropland expansion and excessive groundwater exploitation driven by human activities have significantly altered regional hydraulic connectivity. Therefore， it is urgent to carry out coupled simulation research to clarify its spatiotemporal evolution mechanism， thereby providing theoretical support for the coordinated allocation of water resources. In this study， a SWAT+gwflow coupled model was constructed for the middle reaches of the Aksu River. The model was calibrated and validated using runoff data and groundwater level observations. The results demonstrated that the coupled model could well simulate the hydrological regime of surface water and groundwater in the river basin， with high accuracy in calculating the exchange volume between surface water and groundwater. The results at the grid-scale spatial unit could achieve mutual verification and complementarity with relevant studies， accurately depicting the surface water-groundwater exchange patterns in the river basin. Through the quantitative analysis of multiple hydrological element characteristics in the basin， the spatial characteristics of surface water-groundwater interactions were obtained. The main results were as follows： （1） The study area featured a wide scope and high frequency of surface water-groundwater exchange. The interaction between surface water and groundwater in mountain tributaries was complex. The exchange showed a block distribution pattern from the alluvial-proluvial fans to the floodplain areas on both sides of the mainstream. Surface water and groundwater underwent multiple exchanges through river seepage and discharge， groundwater rise to the soil profile， and surface overflow. （2） Groundwater depletion mainly occurred in areas with elevations below 1 500 m and slope gradients of 0° to 2°， while the minimum depletion was observed in areas with elevations between 2 000 m and 2 500 m and slope gradients ranging from 2° to 6°. Water exchange was concentrated in regions with slope gradients less than 15° and elevations below 2 000 m， accounting for over 75% of the total exchange volume of the river basin. Overall， the exchange volume decreased with increasing elevation and slope gradient across different elevation zones and slope gradient grades， and each hydrological element exhibited extreme values under specific elevation and slope gradient grades. （3） From 2005 to 2019， groundwater exploitation was the key factor leading to groundwater storage depletion in the middle reaches of the Aksu River， with an average annual decline of 17.9 mm. The drop in groundwater levels weakened the hydraulic connection between river and groundwater， resulting in a subsequent decrease in groundwater recharge.

In the context of climate change， rapid glacier retreat is intensifying the release of glacial meltwater， which is projected to become a major source of dissolved organic matter （DOM） for downstream ecosystems. Glacier surface environments， particularly ablation zones， host complex biogeochemical processes where solar radiation and microbial communities continuously transform DOM. Although the composition of DOM in glacial snow and ice is known to be influenced by light exposure and microbial activity， the dynamics and controlling factors of its concentration and molecular composition remain poorly understood. In particular， how photochemical and microbial processes individually and interactively regulate the fate of DOM on glacier surfaces has not been systematically examined. To address this knowledge gap， this study carried out a 30 dayin-situ incubation experiment on snow and ice collected at the terminus of the ablation zone of Laohugou Glacier No.12， a typical continental glacier in the northern Qinghai-Xizang Plateau. The experiment included three parallel treatments： photochemical processing alone （exposed to natural sunlight under sterile conditions）， microbial processing alone （dark incubation with indigenous microbes）， and coupled photochemical microbial processing （natural sunlight with active microbial communities）. DOM was characterized using three dimensional fluorescence spectroscopy coupled with parallel factor analysis （EEM PARAFAC） and Fourier transform ion cyclotron resonance mass spectrometry （FT-ICR-MS）， enabling comprehensive assessment of changes in DOM concentration， optical properties， and molecular composition. Photochemical processes alone had a limited effect on dissolved organic carbon （DOC） concentration but significantly altered the composition and structure of DOM. With increasing light exposure， SUVA₂₅₄ continuously decreased， indicating reduced aromaticity of colored DOM. Fluorescence analysis revealed a decline in humic like components and a parallel accumulation of protein like components. At the molecular level， the relative abundances of labile compounds such as lipids， peptides， and unsaturated hydrocarbons showed a net decline， whereas more recalcitrant phenols and polycyclic aromatic hydrocarbons increased. Moreover， photochemical processes promoted the accumulation of heteroatom containing compounds （N， S， and P）， suggesting that photo-transformation either introduces heteroatoms into DOM molecules or selectively preserves them. Under coupled photochemical and microbial processes， both DOC concentration and SUVA₂₅₄ showed a decreasing trend， indicating combined effects on carbon removal and aromaticity reduction. The changes in fluorescent components were consistent with those induced by photochemistry alone （i.e.， a decrease in humic like substances and an increase in protein like components）. In terms of molecular composition， the coupled processes favored the net accumulation of lipids and peptide like compounds， contrasting with the depletion of these compound classes observed under the individual microbial treatment. Furthermore， the coupled interaction promoted the net enrichment of CHO compounds， implying that the coexistence of light and microbes creates unique transformation pathways not predictable from individual processes alone. In contrast， microbial processes alone primarily reduced DOC concentration， demonstrating that microbes are the main direct consumers of labile organic carbon on the glacier surface. SUVA₂₅₄ and fluorescent components showed no significant changes， suggesting that microbial activity alone does not substantially alter the bulk aromaticity or humic like fluorescence of DOM. At the molecular level， the relative abundances of lipids and peptides decreased， whereas unsaturated hydrocarbons increased. In contrast to the photochemical treatment， phenols and polycyclic aromatic hydrocarbons showed a net decrease under microbial processing alone. The individual microbial effect on the abundance of compounds with different elemental compositions （CHO， CHON， CHOS， etc.） was limited， highlighting that microbes preferentially utilize specific labile molecules without dramatically reorganizing the overall elemental fingerprint. In summary， variations in DOC concentration on the glacier surface were primarily regulated by microbial processes， whereas the aromaticity of colored DOM and the abundance of humic like fluorescent components were mainly controlled by photochemical processes. Overall， the molecular composition of DOM was jointly shaped by photochemical-microbial interactions， which further regulated its bioavailability and environmental fate. The significant net accumulation of lipids and peptides under the coupled treatment strongly supports the view that glacier DOM has a substantial microbial source and high biological reactivity.

The alpine meadow at the southern foot of the Qilian Mountains， a key vegetation type on the Qinghai-Xizang Plateau， is highly sensitive to its carbon cycle processes to climate change， particularly temperature variations. Gross primary productivity （GPP） serves as a core indicator of ecosystem carbon uptake capacity， while growing degree days （GDD） represent a critical thermal factor driving plant phenology and photosynthesis. Against the backdrop of global warming， the Qinghai-Xizang Plateau acts as an “amplifier”， with a warming rate significantly exceeding the global average， profoundly impacting the structure and function of alpine ecosystems. However， previous research has predominantly focused on low-altitude ecosystems or relied on short-term observations， resulting in a lack of systematic understanding regarding the response patterns of alpine meadow carbon fluxes to heat accumulation across different temporal scales. This gap limits accurate prediction of regional carbon sink potential and its future dynamics. To address this， this study conducts long-term continuous in-situ observations of CO₂ fluxes in the alpine meadow ecosystem using the eddy covariance technique at the Haibei Alpine Meadow Ecosystem Research Station. Utilizing a decade （2007—2016） of continuous eddy covariance observations from an alpine meadow at the southern foot of the Qilian Mountains， this study aims to systematically quantify the dynamic characteristics of ecosystem GPP across daily， monthly， and growing season scales， and to elucidate its response patterns and scale dependence on GDD （and mean daily air temperature， T _a）. The objective is to clarify the driving mechanism of heat accumulation on carbon sink function and to identify potential ecological thresholds. The results demonstrated a significant scale effect in the response of GPP to heat accumulation. At the daily scale， GPP and T _a exhibited synchronous unimodal variation， peaking in July， with a significant positive correlation （P<0.001）， reflecting the immediate regulatory role of temperature on photosynthetic physiology. At the monthly scale， GPP and GDD also showed consistent unimodal patterns， both peaking in July， and showed a highly significant positive correlation overall （P<0.001）. However， this relationship exhibited distinct phenological phases. GPP and GDD were not significantly correlated in May and September， whereas a significant positive correlation was observed from June to August， strongest in July （R ²=0.66， P=0.004）. This revealed the optimal enhancement of vegetation photosynthetic capacity during midsummer due to concurrent water and heat availability. At the growing season scale， despite considerable interannual variability in both GPP and GDD， a significant positive correlation remained （P=0.023）， indicating a macro-level driving role of accumulated GDD on the annual carbon sink. However， its explanatory power was weaker than at the monthly scale， suggesting that interference from other interannual factors such as precipitation， radiation， and non-growing season climate events cannot be ignored. This study concludes that the response of GPP to GDD in this alpine meadow at the southern foot of the Qilian Mountains exhibits strong scale dependence and phenological window effects. In the short term， climate warming may enhance regional carbon sink function by increasing GDD. However， this positive feedback mechanism has clear limitations. Low temperatures and phenological constraints early in the growing season， combined with limited light resources and plant senescence late in the season， make GPP insensitive to heat during these periods. At interannual scales， water stress， extreme climate events， and potential vegetation community shifts （e.g.， shrub encroachment） induced by long-term warming could weaken or even reverse the positive correlation between GPP and GDD. The scientific significance of this study lies in systematically clarifying the non-linear response characteristics of alpine meadow GPP to heat accumulation and emphasizing the importance of scale analysis for accurately assessing ecosystem carbon cycles. The findings provide crucial theoretical and data support for predicting the evolution of carbon sink function on the Qinghai-Xizang Plateau under future climate scenarios， and hold important reference value for refining regional climate models and formulating scientific ecological management strategies.

Resource substrates， including litter， plant roots （e.g.， root types， root architecture， dead roots， and their products）， and root exudates， constitute the foundation for driving material cycling and energy flow in the soil food web of artificial grasslands. Their input and transformation drive the interactions between aboveground and belowground ecological processes. The quality， physical structure， and chemical composition of resource substrates derived from different plants determine the dynamic development of soil physicochemical properties， soil biological community structure， and plant productivity in artificial grasslands， playing a crucial role in the stability and sustainability of nutrient input and turnover processes in ecosystems. However， systematic descriptions of resource substrates have so far been addressed primarily in the soil micro-food web of natural grasslands， while their role in the establishment of artificial grasslands remains poorly understood. On this basis， this study systematically reviews the definition of resource substrates， the current research status of their main components， and their impacts on soil physicochemical properties， soil organisms （soil microorganisms， nematodes， and arthropods）， and grassland productivity， by integrating domestic and international studies on artificial grasslands， and summarizes and provides an outlook for future research. Future research on resource substrates should focus on the following aspects： （1） Synergistic research on multiple resource substrates and integrated studies of the entire food web in artificial grasslands； （2） Application of new technologies and interdisciplinary approaches in resource substrate studies； （3） Enhanced research on the regional-scale effects of grassland resource substrates and their responses to climate change. This will provide a theoretical basis for integrated studies on ecosystem functions and processes.

High-altitude alpine regions， characterized by extensive glaciers， seasonal snowpacks， and permafrost， constitute the “Third Pole” and serve as indispensable global ecological barriers and strategic reservoirs for water conservation. Their unique geo-climatic constraints and the inherent fragility of their thermal regimes make these ecosystems especially sensitive to global environmental changes. In recent years， the occurrence and accumulation of microplastics （MPs）—a prominent and persistent category of emerging anthropogenic contaminants—within these remote alpine environments have attracted extensive research attention. Recent empirical evidence underscores their extensive distribution across complex environmental media including the troposphere， glacial cryoconite， fluvial and lacustrine systems， permafrost-affected soils， and diverse biological organisms ranging from soil mesofauna and aquatic invertebrates to high-altitude vertebrates. Consequently， high-altitude alpine regions are increasingly recognized as critical sinks for the long-term storage of global atmospheric MP flux， as well as sensitive sources for their secondary release， particularly as storage and transport are intrinsically coupled with dynamic， phase-changing processes of the cryosphere. The transport dynamics within these high-altitude catchments are governed by complex processes， where MPs are primarily intercepted from the atmosphere and accumulate in glaciers through complex dry and wet deposition， potentially remaining preserved in ice. However， as contemporary global warming triggers unprecedented glacial ablation and permafrost degradation， these cryospheric reservoirs undergo a critical functional transition from stable sinks to active secondary sources， systematically discharging entrained plastic debris and associated adsorbed chemical additives into proglacial streams and downstream terrestrial habitats. Subsequent to their release， these contaminants are subjected to strong environmental factors unique to high-altitude settings， including high-intensity ultraviolet （UV） irradiation that accelerates photo-oxidative aging and surface functionalization， as well as extreme freeze-thaw cycles that facilitate the physical breakdown of brittle polymers into even smaller， more bioavailable MPs through ice-crystal expansion and mechanical stress. These particles then undergo multifaceted trophic transfer and bioaccumulation within specialized high-altitude alpine food webs， thereby exerting significant physiological and toxicological pressures on biodiversity， impairing fundamental ecological functions such as nutrient cycling and primary productivity， and compromising the overall structural resilience and stability of these vulnerable high-altitude systems. Furthermore， the perturbation of alpine hydrological and biogeochemical cycles by MPs may generate widespread ecological effects far beyond local watersheds， impacting global environmental health and the sustainability of downstream water supplies. This review provides a systematic synthesis of the current research progress regarding occurrence profiles and pivotal transport pathways of microplastics in typical environmental matrices of high-altitude alpine regions， highlighting that the mechanistic understanding of multi-phase transport and interfacial transformation governed by extreme physicochemical constraints remains limited and largely qualitative. To bridge these critical knowledge gaps and provide a robust scientific framework for future research， this review highlights several priority directions： firstly， elucidating the mechanisms of multi-interface transport and transformation processes， focusing on how extreme cold and high radiation jointly affect the environmental persistence and degradation kinetics of polymers； secondly， standardizing specialized analytical protocols for complex alpine matrices such as deep ice cores and cryoconite， optimizing extraction and identification techniques for low-concentration and highly weathered samples to avoid analytical bias； thirdly， establishing integrated， transboundary monitoring networks to capture long-range spatiotemporal patterns across the global water towers， facilitating a global mass-balance understanding of plastic flux； and fourthly， developing sophisticated ecological risk assessment models that integrate climate-plastic interactions and account for the unique characteristics and fragility of high-altitude alpine species. Such an integrated approach is essential for providing a rigorous scientific basis for formulating targeted ecological conservation strategies， implementing effective pollution mitigation measures， and developing international policies to ensure the ecological integrity， biodiversity， and water security of these sensitive and irreplaceable alpine headwater ecosystems.

In the context of China’s “dual carbon” goals， investigating the impact of land use change on regional carbon storage patterns and their driving mechanisms is of great significance for optimizing land use structure and enhancing regional carbon sequestration capacity. Arid and semi-arid regions are particularly sensitive to both climate change and human activities due to their fragile ecological environments and limited water resources. However， the dynamics of carbon storage and their underlying driving mechanisms in these regions remain insufficiently understood， especially under the combined pressure of climate warming and rapid socioeconomic development. Gansu Province， located in Northwest China， is characterized by complex and diverse geomorphological features， including plateaus， mountains， basins， and desert landscapes. As a representative arid and semi-arid region， it exhibits pronounced spatial heterogeneity in climate， vegetation， and soil conditions. Under the dual influence of climate change and the implementation of “dual carbon” strategies， understanding the spatiotemporal dynamics of carbon storage and identifying its key driving factors in this region are essential for improving regional carbon sequestration capacity and supporting sustainable land use planning. Based on multi-temporal land use data of Gansu Province from 2000 to 2023， this study incorporated regional climatic heterogeneity to refine biomass and soil carbon density parameters， thereby improving the accuracy of carbon storage estimation. A coupled modeling framework integrating the PLUS-InVEST-Geodetector models was established to systematically characterize the spatiotemporal evolution of land use and carbon storage， quantitatively identify the driving mechanisms of carbon storage changes， and simulate future carbon storage dynamics under multiple development scenarios， including natural development， economic development， and ecological protection scenarios for 2035. The results indicated that： （1） from 2000 to 2023， grassland consistently remained the dominant land cover type in Gansu Province， with a relatively stable area， while cultivated land and unused land continuously decreased， and construction land expanded significantly. Overall， land use change was mainly characterized by the reduction of cultivated land and unused land， accompanied by a substantial increase in construction land. （2） Carbon density correction coefficients exhibited slight stage-wise fluctuations， with biomass carbon density correction coefficients consistently lower than those of soil carbon across all periods， indicating the dominant contribution of soil carbon to total carbon storage in arid and semi-arid ecosystems. （3） Total carbon storage showed an overall trend of first increasing and then decreasing， reaching a peak of 3 108.00 × 10⁶ t C in 2010. Spatially， carbon storage displayed pronounced heterogeneity， forming a gradient pattern of higher values in the southeast and lower values in the northwest， which was closely associated with variations in precipitation， vegetation cover， and topography. （4） The spatial differentiation of carbon storage was jointly influenced by natural environmental and socioeconomic factors， among which natural factors exerted stronger explanatory power. NDVI and annual precipitation were identified as the dominant driving factors， and factor interactions were mainly characterized by nonlinear or two-factor enhancement effects. （5） Scenario simulations for 2035 revealed significant differences in carbon storage responses under different development pathways. The ecological protection scenario exhibited the most pronounced increase in carbon storage， primarily driven by the conversion of cultivated land to forestland and grassland， thereby enhancing regional carbon sequestration capacity. In contrast， the economic development scenario showed relatively limited growth due to the continued expansion of construction land. Overall， this study establishes an integrated framework for carbon storage assessment and driving mechanism analysis in arid and semi-arid regions by coupling land use simulation， ecosystem service modeling， and spatial statistical analysis. The findings provide valuable scientific support for optimizing land use structure， enhancing carbon sequestration capacity， and promoting sustainable carbon management under the “dual carbon” goals.

Cities， as core carriers that gather resources and promote economic growth， are major sources of energy use and carbon emissions， and their low-carbon transition has become a strategic priority in the global response to climate change. Alpine cold regions mainly refer to the Qinghai-Xizang Plateau and its adjacent high-altitude severe-cold areas. In China’s physiographic regionalization， this area is classified as the Qinghai-Xizang Alpine Cold Region， one of China’s three major natural regions. In climatic regionalization， it is designated as an independent alpine cold zone， with an average altitude of more than 3 000~4 000 m. Its climatic characteristics include low annual average temperature， large diurnal temperature difference， sufficient sunshine， intense ultraviolet radiation， thin air， and widespread glaciers and frozen soil. This unique environment not only intensifies the sensitivity and vulnerability of ecosystems， but also reduces the durability of urban buildings and infrastructure. Meanwhile， due to the long-term reliance of regional industries on resource development， as well as their single structure and extensive mode， the low-carbon transition of urban economy is difficult， making the region a critical and challenging area of low-carbon research and practice. Although urban low-carbon development has received widespread attention in recent years， studies on alpine cold regions still lack systematic review. Using a structured literature review， this study synthesizes major progress and challenges in low-carbon energy， low-carbon industry， low-carbon spatial planning， and low-carbon building. Research on low-carbon urban development in alpine cold regions has gradually expanded from early energy assessment and low-carbon buildings to low-carbon economy， spatial processes， and institutional coordinated innovation. In alpine cold regions， new energy not only serves as a key force for replacing fossil fuels and promoting carbon peaking， but also the core support for constructing a long-term zero-carbon energy system and achieving carbon neutrality goals. This region also has substantial carbon sink potential including grasslands and forests， and can simultaneously play the dual roles of “clean energy export base” and “carbon sink functional zone” in the national “dual-carbon” strategy， thereby forming a source-sink coordinated pathway that links energy transition with ecological protection. Research on low-carbon industries in alpine cold regions has developed a relatively systematic theoretical and methodological framework and has been applied in agriculture and animal husbandry， tourism， and construction. Empirical studies reveal multiple pathways for green transformation， including system coupling and coordination， improvements in production factor efficiency， the low-carbon development of industrial parks and characteristic economic forms， and ecological value transformation models. Meanwhile， the innovation of relevant laws and policies continues to advance. Ecological protection laws， negative lists for industrial access， and carbon compensation mechanisms jointly establish the institutional foundation for regional characteristic ecological industries， offering policy guarantee and practical support for low-carbon development in alpine cold regions. Research on low-carbon spatial planning in alpine cold regions has formed multi-level explorations from macro to micro scale. At the regional scale， a “moderately compact + multi-center” spatial pattern and differentiated strategies for optimizing carbon budget are proposed. At the land use level， model simulation reveals that the transformation from grassland to forest enhances future carbon sink potential. Simultaneously， the planning practices of green transportation and low-carbon communities continue to expand， ranging from the technological innovation of plateau highway engineering to culturally adaptive community design for Xizang culture， both providing diverse pathways and empirical support for the green transformation of cities and towns in cold regions. Research on low-carbon buildings in alpine cold regions is forming a systematic exploration from design adaptability and renewable energy utilization to green building material innovation. Empirical cases show that optimizing building form and photovoltaic integrated design can markedly reduce energy consumption and carbon emissions. Technologies such as solar energy-heat pump exhibit high efficiency in heating and power generation， while the combination of green building materials and local resources effectively expands the carbon reduction potential. These results collectively demonstrate that the building sector is a crucial breakthrough for achieving “dual-carbon” goals and enhancing human settlement adaptability in alpine cold regions. However， deficiencies remain in developing urban low-carbon adaptation theory， advancing technology integration， and building innovation systems. Future research should focus on the following four priorities： （1） alpine adaptation and multi-energy coordinated optimization of low-carbon energy systems. Efforts should be made to overcome the issues of output fluctuation and efficiency degradation of renewable energy under alpine cold climate， and develop multi-energy complementary systems with strong climate adaptability and high resilience. （2） Refined assessment and climate resilience enhancement of low-carbon industrial systems. A long-term series and high-precision energy-resource-emission basic database should be systematically established， and life-cycle carbon footprint assessment tools tailored to the characteristics of alpine cold regions should be developed. （3） Paradigm innovation of low-carbon planning for ecological security and spatial coordination. Research on the interaction mechanism between ecological security and urban spatial structure under climate change should be conducted， and low-carbon wisdom in the spatial form of traditional settlements should be explored. （4） Integrated innovation and regional adaptation in low-carbon buildings and energy supply technologies. Research on the mechanisms of building thermal performance， material durability， and human thermal comfort under special plateau climate environment should be strengthened， and accurate assessment of building carbon life cycle and low-energy design should be promoted.

The basal thermal regime of ice sheets and ice caps exerts a fundamental control on ice dynamics， subglacial erosion， and the preservation or modification of glacial landforms. Whether the ice is cold-based or warm-based directly determines the efficiency of basal sliding， meltwater production， and erosion intensity， thereby profoundly influencing long-term geomorphic evolution. Therefore， determining the spatial structure of the basal thermal regime and its controlling factors is essential for understanding Quaternary glacial-geomorphic processes. However， although research on large continental ice sheets has advanced in recent years， understanding of the basal thermal regime of local ice caps， such as paleo-Daocheng Ice Cap （P-DIC） on the Qinghai-Xizang Plateau （QXP）， remains relatively limited. This study focused on the main coverage area （high planation surface） of the P-DIC in the Haizishan area of the southeastern QXP. In-situ cosmogenic ¹⁰Be and ²⁶Al surface exposure dating was applied to widely distributed bedrock surfaces and glacial erratics in the area. Along the main axis of the planation surface from west to east， nine ¹⁰Be exposure ages ［ranging from （15.1±1.0） ka to （148.3±9.3） ka］ and two paired ²⁶Al exposure ages ［（147.3±13.7） ka and （17.8±1.6） ka］ were obtained， and the paired ²⁶Al ages were consistent with their corresponding ¹⁰Be ages within 1σ error. Combined with previously published exposure ages that were uniformly recalculated， this study compiled an exposure age dataset covering the planation surface， consisting of 33 ¹⁰Be ages and 11 ²⁶Al ages. The results exhibited a pronounced spatial differentiation in exposure ages. Samples from the western part of the planation surface were mainly concentrated in Marine Isotope Stage （MIS）6. Samples from the central part spanned MIS5 to MIS2， whereas samples from the eastern part were predominantly concentrated in the late MIS2. The paired ¹⁰Be-²⁶Al results indicated that samples in the study area generally showed simple exposure histories， without significant later burial or complex exposure/burial processes. This suggested that the spatial differences in exposure ages were unlikely to result from differences in the exposure histories of the samples. Regional glacial-geomorphic and chronological studies consistently indicated that the Haizishan high planation surface was entirely covered by the P-DIC during MIS6 and the Last Glacial Maximum （LGM）. Therefore， during this period， there was no scenario of gradual west-to-east deglaciation of the planation surface. Meanwhile， the lithology of the study area was relatively uniform， dominated by biotite monzogranite， and lithological differences were also insufficient to cause the spatial partitioning of exposure ages. These constraints indicated that the spatial distribution of exposure ages was more likely to reflect spatial differences in the basal thermal regime at the base of the ice cap. On this basis， by integrating the regional glacial evolution context with existing chronological data， geomorphic features， and erosion rate results， this study identified the polythermal spatial structure of the basal regime of P-DIC in its main area during the LGM. From west to east， it was characterized by three types of basal thermal zones. The western part of the planation surface was characterized by old exposure ages concentrated in MIS6， well-preserved landforms， and low erosion intensity， indicating a stable cold-based preservation zone where ice was largely frozen to the bedrock. The central part exhibited ages spanning MIS5 to MIS2， together with preserved weathering crusts and localized abrasion landforms， with comparable erosion rates. This suggested a mixed cold-warm transitional zone dominated by warm-based conditions with locally residual cold-based patches. The eastern part of the planation surface was characterized by young exposure ages concentrated in the late MIS2， widely developed cirques and strongly abraded landforms， and relatively high erosion rates， indicating effective resetting of cosmogenic nuclides. This was consistent with a warm-based erosion zone during the LGM. Further analysis indicated that this spatial heterogeneity was primarily caused by the ice thickness distribution and heat accumulation capacity， which were controlled by the underlying topography. Combined with long-term differential erosion， these factors further shaped the topographic relief and reinforced the differentiation between the cold- and warm-based structures. This study is of great significance for a deeper understanding of the basal thermal structure， glacial dynamics， and geomorphic responses of local paleo-ice caps on the QXP， and provides key constraints for quantifying the spatial differentiation of paleo-glacial erosion in the region and for evaluating the basal boundary conditions in ice-cap models.