Sea ice changes are closely related to global and regional climate. Sea ice thickness is one of the important parameters for sea ice change research. Aerial remote sensing is the most effective method for comprehensive and high-precision observation of sea ice thickness.And IceBridge is currently the largest aerial remote sensing project in the Arctic and Antarctic. Sea ice thickness of Bellingshausen Sea in Antarctic was used to study from IceBridge ATM elevation data and DMS optical images in 2009 - 2014. Furthermore， combined with meteorological data such as precipitation， the reasons for the variations of sea ice thickness in this area are discussed. The study found that the sea ice thickness in the sea area showed an overall slight increasing trend （0.07 m·a-1）， but it is insignificant at the 95% confidence level. In 2009 - 2011， there was a sharp change that first increased and then decreased. In 2010， it reached a maximum value of 2.42 m， and then it began to increase slowly. The interannual variation of sea ice thickness is related to meteorological factors such as precipitation and near-surface temperature. In contrast， precipitation is the main influencing factor.
The inaccessibility of many glaciers makes remote sensing data crucial at glacier change monitoring. Glacier outlines were digitized on-screen manually from topographic maps and Landsat images. Glacier change characteristics were analyzed at Zangser Kangri on the Qiangtang Plateau from 1971 to 2015. The results reveal there are 84 glaciers in 2015， covering an area of （297.65±4.29） km2， while 56 of them are smaller than 1 km2 and there are 44 and 16 glaciers facing north and south respectively. Glacier area decreased by （19.32±24.31） km2 from 1971 to 2015， at the annual shrinkage rate of （0.14±0.17）%， which is relatively slow. The annual shrinkage rates are （0.12±1.46）% for 1971 - 1977， （0.20±0.32）% for 1977 - 1993， （0.12±0.50）% for 1993 - 2000， （0.01±0.57）% for 2000 - 2006 and （0.16±0.31）% for 2006 - 2015， respectively. Based on the meteorological data， ERA-Interim reanalysis and ice cores records， the increased ablation season （May to September） temperature is the main driving force accounting for the shrinkage. Small glaciers （<0.5 km2） retreated faster at the rate of 14.00% than big glaciers （>2 km2） at the rate of 5.58%. Northward glaciers shrank faster at the rate of 8.06% than southward glaciers at the rate of 4.16%. The increased number （from 78 to 84） of glaciers indicates big glaciers subdivided to small glaciers while retreating quickly. Two glaciers advanced during our study period.
The glacier is sensitive to climate change， thus is widely used to understand the relationship between glacier and climate. In this paper， the mass balance of Qingbingtan Glacier No.72 was calculated based on the in situ data of stakes and pits from 2008 to 2014， combining with Landsat TM/ETM+/OLI images. Meanwhile， the method of Zero Equilibrium Line was used to calculate mass balance for entire glacier. The results indicated that mean gradient of mass balance with elevation was （0.86 ± 0.19） m w.e.·（100m）-1 during the study period. The mean equilibrium line altitude （ELA） was at （4 167.5 ± 33.2） m a.s.l.， which ranged from （4 317 ± 92） m a.s.l. to （4 109 ± 23） m a.s.l. during the period of 2008 - 2014. The mean annual mass balance was -0.38 m w.e， ranging from -1.23 m w.e. to +0.31 m w.e. During the study period， accumulative mass balance was -2.27 m w.e. from 2008 to 2014. In addition， based on the comparison glacier mass balances among Qingbingtan Glacier No.72， Tuyuksuyskiy Glacier and Urumqi Glacier No.1， we found that the changes in ELA and mass balance of Qingbingtan Glacier No.72 were similar to that of Tuyuksuyskiy Glacier， respectively. However， there was significant discrepancy in terms of the changes in ELA and mass balance between Qingbingtan Glacier No.72 and Urumqi Glacier No.1.
The connection between the glacier and the atmosphere was established by the energy balance model. Based on the meteorological data from August 24 to September 6， 2011， at the elevation of 4 550 m of the Laohugou Glacier No.12 in the Qilian Mountains， the characteristics of energy budget under different weather conditions were analyzed by the energy balance model. It is found that， the downward short wave radiation on fine days （318.3 W·m-2） is about 1.5 times as much as that on cloudy days and three times as much as that on overcast days. With the variation of weather conditions from fine to cloudy， the downward long wave radiation shows an increasing trend. Downward long wave radiation in fine days （215.4 W·m-2） < downward long wave radiation in cloudy days （267.4 W·m-2） < downward long wave radiation in overcast days （291.6 W·m-2）. Due to the influence of solid precipitation， the glacier albedo in overcast days （0.50） is more than twice as large as in fine days. Maximum ablation heat consumption in fine days （739.6 W·m-2） > maximum ablation heat consumption in cloudy days （582.8 W·m-2） > maximum ablation heat consumption in overcast days （324.5 W·m-2）. Net short wave radiation （98%） is the main source of energy income. However， there is significant differences in the proportion of energy output due to weather conditions. Under the three weather conditions， the proportion of net long wave radiation is 35%， 31% and 23%， respectively. The proportion of ablation heat consumption reached 62%， 64% and 75%， respectively. The proportion of latent heat flux varied little.
Glacier meltwater from Qilian Mountains plays an important role in maintaining the ecological balance in northwest China. In order to cognize the glacier change caused by climate warming in Qilian Mountains， we adopted Landsat images （TM， ETM+ and OLI） and band ratio threshold method to extract seven terms of glacier boundaries between 1987 and 2018， and interpreted the time series of glacier boundaries combining with topographic and climatic data. The results show： （1） During recent 31 years， the area of glacier in Qilian Mountains has shrank from 2 080.39 km2 to 1 442.09 km2， corresponding to an annual rate of 0.99%， which is significantly higher than the retreat rate （0.58%） from 1956 to 1990； （2） The equilibrium-line altitude has increased steadily during recent 31 years； （3） The glaciers are mainly distributed between 4 700 and 5 100 m a.s.l.， and the intensity of glacier shrinkage increases as the altitude decreases； （4） About 93% glaciers have an area below 2.0 km2； the total number and area of glaciers small than 0.1 km2 are increasing； （5） The glaciers between 0.5 km2 and 1.0 km2 experienced the fastest shrinkage， with an average annual retreat rate of 1.53%， while that larger than 10.0 km2 experienced the slowest shrinkage， with an average annual retreat rate of 0.59%； （6） The glacier shrinkage in Qilian Mountains is mainly caused by the rise of summer temperature， and more importantly， the glacier shrinkage has intensified during the last decade.
Based on the observation data of daily temperature of 269 sites and atmospheric circulation index in northern China from 1960 to 2017， linear regression， correlation analysis， inverse distance interpolation （IDW） and entropy weight method are used to， the temporal and spatial variation characteristics of cold wave frequency in northern China in the last 58 years and its causes were discussed， and then using the socio-economic statistics such as cultivated land area， per capita GDP and sown area of crops in the study area， the risk regionalization of cold wave in northern China was evaluated and analyzed from four aspects， including risk factors， environmental sensitivity， carrier vulnerability， disaster prevention and reduction ability. The results show that in the last 58 years， the frequency of cold wave in northern China decreased 0.20 times per 10 years， a total of 159.7 cold tides occurred in the last 58 years， of which the frequency of cold wave in winter and November was the most， and the maximum frequency appears in Jingyu County， Antu County of Jilin Province， Aershan City， Sonid Left Banner， Sonid Right Banner of Inner Mongolia Autonomous region， frequency up to 508 times， an average of 8.8 times a year， the minimum frequency of cold waves occurred in the south of north China， Kashi and Akesu City of Xinjiang Uygur Autonomous region. There is a big difference in the frequency of cold wave in different grades， the general cold wave frequency is the most， total 102.9 times， the general cold wave， the stronger cold wave， the strong cold wave and extra-heavy cold wave is similar to the cold wave frequency in spatial distribution， and generally shows the distribution law of increasing from south to north. The northeast region is the active region of cold wave， the northwest active region of cold wave is mainly in Aletai of Xinjiang Uygur Autonomous region， the North China active region is in central Inner Mongolia Autonomous region. On annual scale， the cold wave in northern China was dominated by CA， APVII， AO and APVAI， while the influence factors among seasons were detected， spring cold wave is mainly affected by CQ and CA， and the autumn cold wave was significantly correlated with APVII， CA and AO， while winter cold wave was significantly correlated with APVII， CA， AO and SHI. The comprehensive risk regionalization shows that the Aletai region of Xinjiang Uygur Autonomous region， the Daxinganling region of Heilongjiang Province， the southeast of Jilin Province and Shandong Province are the high risk areas of cold wave occurrence， and the low risk areas are mainly distributed in Akesu and Kashi City of Xinjiang， it is consistent with the spatial distribution of cold wave frequency.
The spatiotemporal heterogeneity of freezing and thawing processes within the active layer， especially on the ground surface， is the key to study the energy and water exchange between permafrost and atmosphere. Therefore， this study used the daily data of ground temperature and moisture from the Tanggula and Tongtianhe active layer observation sites in 2013 to compare variation of the daily freezing-thawing cycles， as well as the soil thermal diffusivity characteristics. The results showed that the seasonal freezing-thawing processes could be divided into four stages： unstable thawing stage， stable thawing stage， unstable freezing stage and stable freezing stage. The duration of unstable thawing stage was significantly longer than that of unstable freezing stage. The duration and depth of daily freezing-thawing cycles in alpine steppe were significantly bigger than those in alpine meadow. Most of the daily ground temperature dynamics showed obvious sinusoidal fluctuations， and the amplitudes of all freezing-thawing stages were quite different. The amplitude of the daily ground temperature changes in the unstable thawing stage was the smallest. Moreover， the amplitude of daily ground temperature in alpine meadow was significantly lower than that in alpine steppe， which indicated that vegetation had a significant impact on daily ground temperature dynamics. As a thermal insulation layer， vegetation cover can weaken the influence of atmosphere temperature fluctuation on ground temperature. The thermal diffusivity within the 5 - 10 cm soil layer was significantly higher than that within the 10 - 20 cm soil layer， and the thermal diffusivity in the warm season from May to October was significantly higher than that in the freezing season. The conduction-convection algorithm can reflect the moisture transfer trend of alpine grassland in permafrost regions.
The Tibetan Plateau is known as the “Chinese water tower”， where widely distributed permafrost and seasonally frozen soil， which playing an important role in ensuring the safety of water resources in China. Here， based on seasonally frozen soil and hydro-thermal data sets （the soil water content is unfrozen water content） from July 2015 to June 2016 in Haibei Station， Qinghai， we analyzed the seasonal variation of frozen depth and the characteristics of hydro-thermal coupling during freezing-thawing process. The results show that： Both soil temperature and soil water content profiles display a U-shape， indicating that there is a consistency between soil temperature and soil water content. The soil temperature profile shows a similar changing trend with mean air temperature， but the variation in mean air temperature is greater than the variation in soil temperature. Moreover， the change of soil temperature has lagged behind the change of mean air temperature， and the lag time depends on the depth of soil. The seasonal freezing-thawing process can be divided into three periods， i.e.， initial freezing period， stable freezing period and thawing period； and the soil freezing-thawing process is characterized by unidirectional freezing and bidirectional thawing. Moreover， the thawing rate is greater than the freezing rate， and in thawing process the thawing from surface downwards is the main. During freezing process of soil， the unfrozen soil water migrates to the frozen front， leading to a decline for soil water content across different soil layers. However， the soil water content in shallow soil decreases more than that in deep soil. During thawing process of the frozen soil， the unfrozen soil water content across different soil layers increases successively. In addition， a high water content region is observed in shallow soil layers. There is a good correlativity between unfrozen water content and soil temperature across different soil layers during freezing process， and the correlation is better in shallow soil than in deep soil. This study would be useful for revealing the key hydrological process of plateau and constructing the hydro-thermal coupling model in cold regions.
Soil temperature is an essential index for energy and material exchange between the earth-atmosphere system during land surface processes. Its dynamic changes and its response to climate change are also one of the key issues in the study of land surface processes. Under the background of global warming， the dynamic changes of soil thermal conditions in the permafrost active layer on the Tibetan Plateau are of great significance for understanding the variation of active layer thickness and the thermal effects of the underlying surface. In this paper， the BP neural network model is used to simulate the surface temperature of the Fenghuoshan area on the plateau. Using the output surface temperature to drive the FEFLOW model， the soil temperature at different depths in the active layer was simulated. Comparisons between the simulated and observed data in situ showed that： the root mean square error ranged from 0.09 to 1.78 ℃， and the Nash-Sutcliffe efficiency coefficient ranged from 0.86 to 0.98. Combined with the BP neural network model and the FEFLOW model， the dynamic changes of the active layer thermal conditions in the study area in the next 50 years are predicted. The results show that under the three warming scenarios of 0.02 ℃·a-1， 0.048 ℃·a-1， 0.07 ℃·a-1， the thickness of the active layer in the study area will increase by 19.4， 51.8， 64.7 cm 50 years later， respectively. Soil temperature at different depths showed a significant increase with the increase of air temperature， and the influence of air temperature on soil temperature decreases with depth. At the same time， it is found that under different warming scenarios， the time for the soil to start thawing is continuously advanced， and the time to start freezing is continuously delayed， and this law also decreases with the soil depth， but the difference of response of soil freezing and thawing process to temperature rise at different depths decreases with increasing of the warming rate.
The change of ground temperature plays an important role in the climate feedback effect， and understanding the spatio-temporal relationship between ground temperature and its influencing factors is crucial to the prediction of global temperature change. The summer land surface temperature （LST） was constructed by wavelet analysis and BP neural network based on the daily observed data of meteorological stations in the Shiyang River basin from 1998 to 2017. The prediction results and the accuracy are tested. The results show that： （1） The prediction effect of daily average land surface temperature is the best at different stations， and the correlation coefficients between the predicted values and the observed values are both greater than 0.87， and the prediction probability within 3 ℃ is higher than 84%. Among them， Minqin has the best prediction results. The correlation coefficient between predicted and observed values reaches 0.91， and the prediction probability within 3 ℃ is 86%. （2） The prediction results of daily maximum LST in the Shiyang River basin can reflect its variation trend， and the correlation coefficient between the predicted value and the observed value is higher than 0.8. Among them， the simulation effect in Yongchang is the best， and the prediction probability within 3 ℃ is 83%. （3） For the daily minimum LST， the average correlation coefficient between the observed and simulated values is higher than 0.66， but it is little underestimated. The prediction probability of daily lowest LST in the Shiyang River basin at different stations within 3 ℃ is all higher than 83%. Among them， Wuwei has the best forecasting effect， the correlation coefficient between forecasting value and observation value is 0.72， and the forecasting probability within 3 ℃ reaches 94%. This study can provide some references for making up for the lack of ground temperature observation in arid and semi-arid areas and discussing its relationship with local climate.
Based on the daily minimum temperature observation data from 43 stations in Qinghai Province from 1961 to 2017， 61 typical regional persistence low temperature events were extracted according to the judgment criteria of regional persistence low temperature events， and the corresponding atmospheric circulation were analyzed. The results show that there were obvious precursory signals 15 days before a event， especially when there were anomalies in the Ural Mountains， Lake Baikal and Okhotsk Sea， the probability of regional sustained low temperature events in Qinghai was high. While events occurred in winter in Qinghai Province， the probability of positive anomaly at 500 hPa height field in Ural Mountains was the highest， followed by that in the Okhotsk Sea and the Lake Baikal. Meanwhile， the longer the events duration， the better the corresponding relation between events and the height field positive anomaly in the key area. Therefore， it is further revealed that the combination and configuration of height field anomalies in the two key areas （Ural Mountains and the Okhotsk Sea） can well explain the typical events in Qinghai in winter； all of which are useful for further improving the multi-scale climate prediction in winter in this area.
Based on the meteorological data from 46 meteorological stations in Jilin Province from 1961 to 2015， the spatiotemporal changes of annual freezing/thawing index and its relationship with longitude， latitude， altitude were studied by the method of climatic diagnosis analysis in Jilin Province. The results showed that the freezing index had decreased gradually from north to south， and the thawing index had decreased from west to east in the province. The freezing index showed a significant downward trend， with the climatic tendency rates of AFI （air freezing index） of -48.7 ℃·d·（10a）-1 and SFI （surface freezing index） of -166.8 ℃·d·（10a）-1. The ATI （air thawing index） and STI （surface thawing index） increased significantly at 57 ℃·d·（10a）-1 and 93.7 ℃·d·（10a）-1， respectively. The SFI， ATI， and STI had mutated in 2001， 1994， and 1997， respectively. In the 1960s and 1970s， the freezing index was very high and the thawing index was extremely low. The trend of freezing thawing index in the future was consistent with the past， when the freezing index had an downward trend and the thawing index had an upward trend. The freezing index was mainly affected by latitude and increased with the rise of latitude. The thawing index was mainly affected by altitude and decreased significantly with the rise of altitude. The climatic tendency of freezing index increased with the rise of altitude， and the climate tendency of the thawing index increased with the rise of latitude.
Using the daily and monthly average data of NCAR/NCEP from 1949 to 2015， the characteristics of the snowstorm water vapor transport over Liaoning Province in winter were studied. The results show that： under the influence of winter monsoon， the characteristics of the average winter water vapor transport over the province was that the water vapor mainly came from the middle latitude of the western boundary and there was no water vapor input in the meridional direction of the westerly， however， the occurrence of regional snowstorm was the result of the abnormal transport of water vapor in the meridional direction. There were two main water vapor transport belts. One was the mid-high latitude westerly belt airflow， and the other one was the northern branch of the northeast flow of the anticyclonic circulation over the south side of the Japan Sea. The two water vapor belts converged in Liaoning. Snowstorm water vapor sources mainly included the Western Pacific， the Japan Sea， the East China Sea and the Yellow Sea. The East China Sea and the Yellow Sea were the main and direct sources of water vapor. The Japan Sea high pressure was the key system for the transportation of snowstorm water vapor in Liaoning Province. There was Japan Sea high pressure in sea level pressure field during 82.4% of the snowstorm process in Liaoning Province. According to the location and intensity of the Japan Sea high pressure， it would be divided into northerly， southerly and high-pressure ridge types. Different types of high-pressure water vapor transmission intensity were different， corresponding to the intensity and distribution range of snowstorm in Liaoning Province. The north-north type （high-pressure center was north of 35° N， near or east of 140° E） had the strongest water vapor transmission， and the distribution range of the snowstorm was the largest. The southerly type （high-pressure center was near or east of 30° N， 140° E） had the second strongest water vapor transmission， and the distribution range of the snowstorm was small. High-pressure ridge type （with high-pressure ridges extending southeast or south to the Japan Sea） had the weakest water vapor transmission， and the distribution range of the snowstorm was the smallest. During the occurrence of snowstorm in Liaoning， the East China Sea， as well as the Yellow Sea， was a large area of positive humidity advection， which was the source of water vapor， and Liaoning Province was within the convergence of the wind field. Among them， the northerly area snowstorm was the largest and strongest in the water vapor source area in the north and south of the Yellow Sea. The corresponding wind field convergence area was also largest in range and strongest in strength. The advection of the water vapor in the East China Sea and the Yellow Sea and the convergence of the wind field were the contributing factors for the snowstorm in Liaoning Province.
The Tibetan Plateau is getting warmer and wetter. Landslides， such as earthflow， gelifluction （solifluction） and thaw slump occur frequently， seriously affecting engineering construction， ecological environment and human activities. However， there are only few studies on earthflow hazards on the Tibetan Plateau. And， it is difficult to distinguish earthflow， gelifluction （solifluction） and thaw slump due to some similar characteristics. This paper takes Zhimei landslide （Sept.7， 2017） in Yushu， Qinghai Province as an example， and uses multi-source data， such as measured data， multi-temporal remote sensing images and unmanned aerial vehicle data， and radar technology to investigate and analyze the landslide. It is found that the landslide occurs at the accumulation fan of the slope， of which the average thickness was about 5 m， the volume was about 2.4×104 m3， and the sliding direction of the slide body was consistent with the process of gravity. It is concluded that the landslide is a small traction landslide in the shallow layer of accumulated soil. Then the characteristics of earthflow， thaw slump and gelifluction （solifluction） have been summarized based on the recent research， and it is considered that Yushu landslide is an earthflow in seasonally frozen soil region， which relates to local geological conditions， continuous precipitation and freeze-thaw cycle. This study improves scientific understanding of earthflow and disasters on slope relating to Tibetan Plateau.
In order to analyze the effects of freezing and thawing process， ballast cover and rainfall on thermal conductivity， the in-situ monitoring of thermal conductivity， temperature and moisture of subgrade and natural soil in the test section of permafrost railway roadbed was carried out in the Qinghai-Tibet Plateau. The results show that the fluctuation of thermal conductivity during the melting period is more obvious than the freezing period. The thermal conductivity during the thawing period is larger than the freezing period in natural site， but the thermal conductivities of gravel soil subgrade and ballast covered subgrade during the freezing period are smaller than the thawing period. The monitoring results of subgrade soil are contrary to the general cognition and simulation value of temperature field. Moisture content and thermal conductivity of gravel soil subgrade are higher than ballasted subgrade soil due to the thermal insulation and water blocking effects of ballast layer. Meanwhile， the thermal conductivity of subgrade soil during freezing period tends to decrease， especially for the ballast subgrade soil. The rainfall infiltration increases the thermal conductivity of the soil. And the thermal conductivity of the low water content ballast subgrade has a considerable response to rainfall. In the numerical simulation of railway subgrade engineering in cold area， the influence of hydrothermal change on thermal conductivity should be considered. It is not appropriate to use the piecewise function or step function of the fixed phase change interval to estimate the thermal conductivity.
In order to study the mechanisms of pipeline stress in the process of soil freezing in soft clay area， in this paper， the freezing model test of buried pipeline in saturated soft clay was carried out， and the temperature field， water field and displacement field of soil during artificial freezing， as well as the mechanical properties of the pipe， were studied. The results show that： （1） During the freezing process， the change of soil temperature field directly affects the distribution of soil water in the soil； （2） There is a strong water migration phenomenon in front of the freezing front， and a large amount of water migrates to the frozen front， causing linear frost heaving of the soil； （3） The rate of development of frost heaving is directly affected by external loads； （4） When the freezing develops to the position of the pipe body， the pipe body at the position of the frost heaving and unfrozen transition section exhibits a maximum stress. The results of the study are useful for the safety assessment of pipelines in normal frozen soil.
As the key project of the Zhuhai connecting line of the Hong Kong-Zhuhai-Macao Bridge， the Gongbei Tunnel successfully used the Freeze-Sealing Pipe Roof Method （FSPR） for the first time at home and abroad. Based on this background， a model test of FSPR under different pipe arrangements have carried out and the temperature field variation have been studied. The results show that during the active freezing period， the temperature of each measurement point decreased sharply in the first four hours， then slowed down gradually， and finally tended to be stable after falling to the sand freezing point. The three configurations of pipe arrangements all met the freezing design requirements. The freezing tube inside the hollow pipe first cooled the pipe wall and then reduced the temperature of the surrounding soil via the surface area of the pipe. After 21 hours of active freezing， the vertical range of the frozen wall in Configuration C was the largest， and the average thickness of the frozen wall directly above the hollow pipe wall was about 105 mm. On the premise of meeting the design requirements of the rigidity of the pipe roof， this configuration could be adopted to achieve the purpose of forming a frozen curtain quickly. The temperature curve of the measurement point within the vertical distance of 100 mm at the concrete pipe middle line had a distinct rising stage in 4 hours after the opening of the limiting-tube， but it still remained below the freezing point and the frost wall between the pipes was stable. The existence of the frozen wall was stable， indicating that the limiting tube can limit the frost heave of the formation within a certain range under the condition that the water was effectively sealed between the pipes. The proposed optimized double circular freezing-tube can meet the requirements of freezing design， at the same time， it was more convenient to install， economical and environmentally friendly.
The pile foundation freezing uplift in cold region engineering is closely related to the shear behaviors of interface between the frozen soil and pile foundation. With the help of a self-design apparatus， series of the shear tests for concrete piles， steel piles and wood piles embed in gen-he silty clay at different freezing temperatures were carried out by using the static pressed pile method. Some conclusions can be obtained. At negative temperature， with the increased of shear displacement， the shear force experienced three stages： linear increased stage， brittle failure with an abrupt drop stage， and a stability stage. The lower the temperature， the greater the cementation force of ice between the pile and frozen soil. As a result， the freezing strength， the residual strength， and the allowable displacement of failure was higher in the lower negative temperature. It was worth mentioning that the freezing strength between the wooden pile and frozen soil interface was the highest， followed by that between the concrete pile and frozen soil interface， all big than that between the steel pile and frozen soil interface. The freezing strength and residual strength of the concrete pile and the steel pile had a linear relationship with the frozen temperature. However， the freezing strength and residual strength of the wooden pile has a quadratic function relationship. The allowable displacement of failure of the three kinds of piles had linear correlation with frozen temperature. This study is expected to provide a reference for the research on the shear characteristics between the frozen soil and the pile interface， and also supposed to help solving the pile foundation freezing uplift problem in cold region engineering.
In order to explore the influencing factors and changing laws of the frozen tensile strength of soil-structure interface under freezing action， orthogonal test study on frozen tensile strength of silt-concrete interface under various moisture， temperature and dry density conditions was carried out by a self-developed test device. The test results show that under frost action， the tensile strength of silt-concrete interface is about 1/10 of the saturated loess， and 1/10 - 1/7 of the silty clay； the moisture has the greatest influence on the frozen interface tensile strength， followed by temperature and dry density. In the test moisture range， the interface tensile strength is positively correlated with the moisture. When temperature decreases， the interface frozen tensile strength increases rapidly between -2 ℃ and -6 ℃， and the growth rate decreases significantly after -6 ℃. The results show that the interface frozen tensile strength is optimal under low temperature， high moisture and low dry density； the test factors have different influence degree on the frozen tensile strength， and the water content and temperature have significant influence while the effect of dry density is not significant； in the test moisture range， the interface frozen tensile strength is positively correlated with the increase of moisture， and there is interaction between moisture and temperature； with the decrease of temperature， there is a critical negative temperature where interface frozen tensile strength changed from rapid growth to slow growth. The "critical negative temperature" phenomenon in the process that frozen interface tensile strength changes with temperature was discussed from the perspectives of unfrozen water content and moisture migration.
The mechanical properties of the interface between the frozen soil and various structures play an important role in most engineering in the cold region， which are affected by many factors. In order to study the influence of the freezing strength of the interface between the silty clay and the concrete lining， the orthogonal direct shear test considering the water content， freezing temperature and freezing time was carried out. Test results reveal that increasing water content， lowering freezing temperature and prolonging freezing time can improve the early freezing strength of the interface， and this promotion is linear at the beginning of the freezing process. Based on the theory of significance analysis， it is found that water content， freezing temperature and freezing time have significant influences on the early freezing strength of interface， while the interaction between the three factors has little effect on the freezing strength. Among the three factors， the significance queue is： freezing temperature > freezing time > water content. The second-order interaction between factors was ignored in the multivariate linear regression analysis. The prediction model of the freezing strength of the interface between silty clay and concrete is established， which is affected by the water content， freezing temperature and freezing time.
In order to study the unconfined compressive strength and failure characteristics of soil-cement under the coupling of chloride salt erosion and freeze-thaw cycles， the freeze-thaw cycle test of soil-cement with different concentrations of sodium chloride solution was carried out， and the freeze-thaw corrosion factor， volume change rate and modulus of deformation before and after freeze-thaw were obtained. The damage amount and microstructure characteristics were analyzed. It is found that the unconfined compressive strength and freeze-thaw corrosion factors of soil-cement under the coupling of chloride salt and freeze-thaw cycles decrease with increasing freeze-thaw cycles， and as the concentration of chloride salt increases gradually， the unconfined compressive strength and freeze-thaw corrosion factor of soil-cement decrease rapidly. In a same concentration solution， the volume change rate of soil-cement increases with increasing the number of freeze-thaw cycles， but the deformation modulus decreases with increasing the number of freeze-thaw cycles. As the concentration of chloride solution increases， the volume expansion of soil-cement increases， which results in a weaker internal structure and a weaker resistance to deformation. Under the same number of freeze-thaw cycles， the damage caused by sodium chloride solution is greater than that in specimen from clear water. As the concentration of sodium chloride solution increases， the damage of microstructure inside soil-cement becomes more serious.
To investigate the influence of confining pressure and temperature on strength of rock， a series of uniaxial and triaxial compression tests were conducted on weathered granite at different temperatures （15， -5， -15 ℃） and confining pressures （0， 4， 7， 10 MPa）. Results show that： （1） When the confining pressures range from 0 to 10 MPa， the triaxial compressive strength of weathered granite increased linearly under the same temperature. Under the same confining pressure， the compressive strength of weathered granite increased obviously as the temperature decreases. Meanwhile， the cohesion increased with the temperature decreasing， the internal friction angle also increased with the temperature decreasing. （2） The modulus of elasticity increased with the increase of confining pressure. However， with the decrease of temperature， the increasing range decreases gradually. Poisson’s ratio increased with the decrease of temperature and the increase of confining pressure. （3） Temperature and confining pressure are important factors affecting failure modes of weathered granite samples. The shrinkage of microcracks and interstices in mineral particles leads to the increase of cementation strength when the temperature decreases. With the increase of confining pressure and loading， micro-cracks developed and penetrated gradually. At the same time， pore ice fractured cracks were fully contacted， the friction force was further increased and the rock strength was improved.
In order to study the frost heaving pressure in fractured rock in cold regions， based on the elastic mechanics and phase transition theory， an analytical model for frost heaving pressure in a single fracture is established considering water migration and the content of unfrozen water； the evolution of frost heaving pressure in a single fracture is obtained by using the extended finite element method （XFEM） based on MATLAB program. The frost heaving pressure is also calculated by the method of effective coefficient of volume expansion and the result is in good agreement with the analytical solution and the other experimental results. Then the crack-tip stress is analyzed and the results are also in good agreement with the theoretical values， which validates the reliability of this numerical method. Finally， it found that the elastic module and the shape characteristics of fracture are two crucial parameters to analyze the main influencing factors affecting the development of frost heaving pressure.
Considering the engineering design demand， a simple analytical method for calculating soil frost heave deformation of horizontal field under steady-state condition is established based on the nonlinear elastic mechanical model of soil frost heave. Mathematical expression to determine the frost heave of soil element in semi - infinite space body under dead weight is deduced. Also， the prediction methods of soil lifting displacement， soil stress and strain field under different overburden and displacement conditions of horizontal site are proposed. Furthermore， the step-freezing test of Harbin silty clay in open system is carried out， and the reliability of the analytical method for predicting the surface lifting deformation caused by frost heave is verified. Besides， the analytical method is further applied to the evaluation of the frozen site. It can be concluded that the frost heave amount can be used as a main quantitative index to evaluate the frost heave of the site. However， it is still necessary to consider the frost heave sensitivity， the thickness of the frozen soil layer， the initial overburden pressure and the surface temperature.
Relaxation modulus of frozen soil is a most important factor in creep analysis. However， there is still no a more perfected formulation of relaxation modulus according to the spherical template test theoretically. Based on half spacious visco-elastic theory and fractional order thesis， this paper had found formulations of relaxation modulus and analyzed the impact of parameters included in formulations according to the nonlinear Kelvin model. We propose a universal method for acquiring relaxation modulus based on dead load spherical template indenter tests and uniaxial creep test. This means relaxation modulus analysis results based on uniaxial creep tests can be applied for evaluating the results based on spherical template indenter tests. Then， the model is used for predicting the spherical template indenter tests curves of frozen fine sand and uniaxial creep curves of frozen sand and frozen sandy clay， as well as relaxation modulus of them， respectively. Besides， other experimental data are provided to evaluate the calculation results. Analytical results show that relaxation modulus characterization function has the least material parameters and good monotony versus material parameters. Nonlinear Kelvin model can predict the reasonable and reliable value of relaxation modulus of frozen fine sand on the basis of spherical template indenter test. The nonlinear Kelvin model has a good description of uniaxial creep curves of frozen sand and frozen sandy clay. The changing regularity of relaxation modulus of frozen sand based on uniaxial creep tests agrees with that based on spherical template indenter test. The changing regularity of frozen sandy clay relaxation modulus with temperature and stress is consistent with existed conclusions.
Glacier meltwater runoff is an important channel in the migration of chemicals. Study on the hydrochemistry of glacier meltwater is helpful to reveal the biogeochemical cycle of the chemicals in the glacierized zones and serve as basis for understanding and evaluating the impact of glacier melting on the environment and human life. The Qinghai-Tibet Plateau and its surrounding areas hold the largest amount of glaciers outside the polar regions， and glaciers on the plateau have been retreating at an accelerated rate under the context of warming climate. Studies on the changes of various inorganic chemicals in glacial meltwater runoff and its climatic and environmental effects have been increasingly recognized as a hot topic. In this review， we reviewed studies on the concentrations and spatio-temporal variations of the inorganic chemical species in glacial meltwater runoff in the plateau. In addition， we summarized the main sources of the inorganic chemicals and the quantitative and qualitative methods of provenance determination. It is suggested that the contents and spatio-temporal variation of inorganic chemical species in glacial meltwater were influenced by the melting of glaciers， the properties of bedrocks， the hydrological characteristics of the meltwater runoff and some hydro-physiochemical processes. We summarized the deficiencies of the current research on the inorganic hydrochemistry of meltwater runoff in the Qinghai-Tibet Plateau and its surrounding areas， and proposed the future research needs. It is necessary to strengthen data accumulation and the spatial -time scale of field observation， to clarify the transport mechanism and rule of inorganic chemicals， to reveal the synergistic and antagonistic effects of multiple factors on the hydrochemistry， and to evaluate the effects of glacial meltwater hydrochemistry on the climate and environment. All these research will provide scientific guidance for dealing with the environmental impact of glacier ablation over the Qinghai-Tibet Plateau.
In this study， we have evaluated three satellite rainfall products （IMERG-E/IMERG-L/IMERG-F） from Global Precipitation Measurement （GPM） and two satellite rainfall products （TMPA 3B24 and 3B42RT） from Tropical Rainfall Measuring Mission （TRMM） in the source regions of the Yellow River in daily scale， with special focus on the dependence of performance on spatial distribution and rainfall intensity. Daily rainfall data series from 38 gauge stations has been used. Results indicate that： （1） GPM products outperform TRMM products in both consistency and deviation with gauge observation data， while 3B42 is better than 3B42RT. The correlation coefficients between five products and gauge observation show significant decline from southeast to northwest， while the RMSE in northern regions are larger than in southern regions generally. （2） In terms of the ability of rainfall detection， probability of detection and critical success index of IMERG products are both larger than those of TMPA products， while false alarming ratio of IMERG products are larger than that of TMPA products. High false alarm rates are occurred in every product at individual stations. These stations are almost located in the northwestern regions. （3） The performance of IMERG products is negatively correlated with elevation， while that of TMPA products has not obvious correlation with elevation. （4） The deviation of all products is positively correlated with the rainfall intensity except for 3B42RT. In terms of probability of detection， IMERG products are much better than TMPA products whether in light rainfall events， moderate rainfall events or heavy rainfall events.
Baseflow is the vital supply sources of the runoff in the Yellow River， which plays an important role in water resource planning and ecological environment construction in the source regions. In this study， the Baihe River basin was selected as research area in the source area of the Yellow River. Based on the daily runoff data observed at Tangke Hydrological Station from 1981 to 2015， 12 kinds of five types commonly and widely used numerical simulation methods were used to separate the baseflow and compare and analyze the applicability in the Baihe River basin. The results showed that there were significantly difference among the annual baseflow indexes estimated by the 12 methods； the baseflow index from the Lyne-Hollick digital filter method was the maximum， 0.91， and that from the Kalinin method was the minimum， 0.39； the standard deviation and the extreme value ratio from the Eckhard filter method were the worst， with the best stability； the standard deviation and the extreme value ratio from the Kalinin method were the biggest with the worst stability. The daily runoff process in 1992 （a wet year） was separated， of which the results showed that the baseflow hydrographs from Chapman digital filter method， Chapman-Maxwell digital filter method and Kalinin method were much more smoother and slower， which were more accordant with practical circumstances. The baseflow simulation results of Boughton-Chapman digital filter method， BFI and HYSEP methods were better， their Nash-Sutcliffe efficiency coefficients were higher than 0.87， their decisive coefficients were close to 0.9 and their average relative errors were less than 10%. It was concluded that the Boughton-Chapman digital filter method and BFI method were more suitable for the baseflow separation in the Baihe River basin.
Based on the summer daily precipitation data from 47 observational stations in northern Xinjiang from 1961 to 2017， the extreme summer precipitation thresholds at different observational stations were defined according to the percentile method. The temporal and spatial distribution of the extreme summer precipitation events and the maximum daily precipitation in northern Xinjiang and their contribution， relationship with altitude were analyzed. The results indicate that： Both extreme summer precipitation events and the maximum daily precipitation in the northern Xinjiang had presented obviously temporal and spatial differences. The extreme summer precipitation and the maximum daily precipitation were high in the mountains and low in the basin. There was a maximum precipitation band occurred at approximately 2 000 m above sea level. The extreme summer precipitation events and the maximum daily precipitation in northern Xinjiang had been significantly increased since the 1990s. The extreme precipitation events in summer were mainly on a single day. The contribution of the extreme summer precipitation had been slowly increased. But the contribution of the precipitation of the extreme precipitation process and the maximum daily precipitation in summer had showed a downward trend over time.
Based on the hourly precipitation data from ten stations over Ili River valley during 2012 - 2017， the daily variations of winter precipitation were analyzed. The results showed that the diurnal variation of the winter precipitation in the west valley presented nearly a single peak and a single trough， with the maximum at morning （09：00 - 12：00， Beijing time）. The daily variation of winter precipitation in other areas of the valley presented multiple peaks with more dispersed small peaks. The occurrence time of maximum average precipitation intensity and maximum hourly accumulative precipitation was more consistent with maximum hourly accumulative precipitation frequency. The relationship between precipitation and precipitation frequency in most areas of the valley was closer. The winter precipitation in the valley was mainly short-duration precipitation within 6 hours， and its contribution rate to the total winter precipitation was less than 30%. There were less precipitation events lasting longer than 12 hours， less than 30% in most parts of the valley， but their contributed more to the valley. Short duration and continuous precipitation events were important contributors to the diurnal variation of precipitation in the western part of the valley. Continuous and long precipitation events were important contributors to the diurnal variation peaks of precipitation in the center and north of the valley. The relationship between the daily cycle of winter precipitation and precipitation continuity was not significant in the southwest of the valley.
During icing period， with the exception to freezing， large amount of water from the section of the Yellow River in Inner Mongolia is converted into channel storage. This will lead to a long low discharge process at Toudaoguai Hydrology Station. Longer low discharge process will lead to ice flood during the break-up period. The critical low discharge value， 330 m3·s-1， was defined after analyzing the changing process and characteristics of Toudaoguai Hydrology Station discharge from 1998 to 2016. Based on this value， analysis of low discharge process was conducted using R/S range analysis method， Fourier transformation method and multiple linear regression analysis. The Logit and Probit nonlinear probabilistic models were used to discuss factors of low discharge process at the same time. The result showed that the number of low discharge days was decreasing， and there was a significant impact relation between low discharge process and change in upstream relative discharge. The low discharge process synchronized with change in upstream relative discharge， but lagged behind the change in channel storage. By using Logit and Probit models to analyze the response probability of duration of low discharge changing with each influence factors， ice discharge was a primary influence factor of low discharge process. The temperature and condition of the freeze-up site， as well as the relative coming discharge， were the main factors also.
The stable oxygen isotope composition （δ18O） was measured in different water bodies such as soil water， plant water and atmospheric water vapor of Picea crassifolia forest ecosystem in the Pailugou catchment of upper reaches of the Heihe River at 2 700 m and 2 900 m altitude， respectively. The Craig-Gordon model， isotope steady state hypothesis and Keeling Plot model were used to obtain the δ18O of evaporation， transpiration and evapotranspiration， respectively. Combined the multiple linear mixed model to partition evapotranspiration into evaporation and transpiration. The results showed that： （1）The δ18OE of evaporation， the δ18OT of transpiration andthe δ18OET of evapotranspiration was between -35.9‰ and -25.2‰， -9.0‰ and -4.2‰， -18.5‰ and -10.2‰， respectively. The order of them was δ18OT > δ18OET > δ18OE，which according with the assumption of isotope steady state. （2） Contribution rate of evaporation to evapotranspiration （fE） varied from 11.6% to 47.8% and contribution rate of transpiration to evapotranspiration （fT） varied from 52.3% to 88.4%. fT was much larger than fE， indicating that most evapotranspiration came from transpiration. Transpiration was a vital part of evapotranspiration in the Picea crassifolia forest ecosystem. （3） The fT was negatively correlated with air temperature but positively correlated with relative humidity， indicating air temperature restrained fT and relative humidity promoted fT although the correlation was not significant. Addition， fT was also influenced by various environmental and biological factors except air temperature and relative humidity in natural environment. Temperature and relative humidity are not the only control factors of fT. The results of this study can provide a theoretical basis for further studies on intra-regional circulation and watershed scale water circulation research in Heihe River basin.
The vegetation ecosystem on the Tibetan Plateau is vulnerable under global and regional climate changes. It is an ideal place to study the response of terrestrial vegetation ecosystems to global climate change. Based on GIMMS NDVI， temperature and precipitation， and vegetation type data， the linear regression equation， correlation coefficient， partial correlation coefficient and t-test method were used to analyze the temporal and spatial variations of NDVI and their responses to the variations of temperature and precipitation on the plateau from 1982 to 2015. The result showed that： （1） The temporal change process of NDVI in the plateau from 1982 to 2015 showed an insignificant increase process， while the spatial change process was increased significantly， accounting for 63.26% of the total area， mainly in the north， west and south of the plateau， such as the west and north of Qinghai Province， southern Xinjiang Region， western and southern Tibet. A significant reduction was concentrated in the southeast， such as the eastern part of Tibet， the southern part of Qinghai Province， the western part of Sichuan Province and the border areas of the three provinces （region）， accounting for only 3.45% of the total area； （2） The average NDVI of the main vegetation types on the plateau were broad-leaved forest > coniferous forest > shrub > meadow > alpine vegetation > grassland >semi-desert， in which NDVI of grassland， alpine vegetation and semi-desert vegetation having a significant linear increase， while the NDVI of shrub， coniferous and broad-leaved forest vegetation having less significant reduction； （3） The spatial distribution of NDVI and temperature correlation coefficient on the plateau was north-south， with the characteristics of latitude and zonal. The significant positive correlation was distributed in the central and northern part of the plateau， while the significant negative correlation was distributed in the central and southern part of the plateau. The correlation coefficients between NDVI and precipitation distributed in an east-west direction with dry zonal characteristics， and a significant positive correlation distributed in the middle of the plateau， while the significant positive correlation distributed in the east and west sides of the plateau. The following conclusion can be drawn that the NDVI in poor hydro-thermal condition in the northern part of the plateau had increased significantly from 1982 to 2015， but had a significant reduction in the areas with sufficient hydro-thermal conditions. Further research on the NDVI and their response to climate change in different vegetation regions will help to understand the regional differences in global climate change and scientifically formulate vegetation ecological protection policies.
Based on the TRMM satellite precipitation data and the observed temperature data， using the Thornthwaite Memorial model， the net primary production （NPP） variation in the alpine pastoral areas of northern Tibetan Plateau from 1998 to 2016 and the NPP variation tendency under the future climate change were estimated. The results show that the area with rising tendency of NPP in the study area， accounting for 71.9% of the total area， was greater than declining tendency. The average annual increase rate of NPP was 0.54%， and the temperature and precipitation had increased in the same period. NPP and temperature increased significantly around 2007. By analyzing the influence of climate factors on NPP， it was found that precipitation was the dominant factor in the region， and with the increase of latitude， the influence of temperature was more significant only in a small eastern region. It was estimated that the NPP changing tendency under climate change showed that the average state of NPP in the study area has almost no change under the three emission scenarios （RCP2.6， RCP4.5， RCP8.5）. The effect was only limited to the higher NPP areas in the southeast of the study region， where there was a slight improvement effect； the improvement effect was RCP8.5 ≈ RCP4.5 > RCP2.6， indicating that climate warming had limited effect on NPP improvement in the study area.
Vegetation community， soil respiration characteristics and their relationships at different degradation stages were explored， and the temperature sensitivity of soil respiration was analyzed by monitoring the vegetation and soil respiration at different degradation stages of an alpine meadow in Maqen County， Sanjiangyuan region in 2017. The results show that： As the degree of degradation increase， the Gramineae decreases， but poisonous weed increases significantly （P<0.05）； vegetation coverage， species index， and diversity index decrease significantly （P<0.05）. The aboveground biomass under the condition of heavy degradation has decreased by 25.36% and 22.37% in comparison with light and moderate degradation （P<0.05）. Under the condition of moderate degradation， evenness index and underground biomass increased significantly （P<0.05）. Soil respiration at each degradation stage has a unimodal change during the year， showing high in growing season and low in non-growing season， especially in a highest peak growth period （July - August） of the vegetation and a significant exponential relationship with soil temperature at 5 cm depth （P<0.05）； total soil respiration emissions in 2017 under light， moderate and heavy degradation are 626.89 gC·m-2， 386.66 gC·m-2 and 393.81 gC·m-2， respectively. Soil respiration rate decreases with the degree of degradation. Temperature sensitivity coefficients （Q10） of soil respiration at light degradation， moderate degradation and heavy degradation stages are 2.82， 3.54， and 2.35， respectively， indicating that the temperature sensitivity is the highest under moderate degradation and the lowest under heavy degradation. These results would be useful for understanding the vegetation and soil respiration characteristics at different degradation stages of alpine meadows on the Tibetan Plateau.
Utilizing datasets of the Aral Sea area， global CRU meteorological data and the irrigation area and reservoir capacity， we quantitatively assessed the long-term area variation of the Aral Sea during the period of 1960 - 2018， and then investigated the dominant influence factors including climate change and human activities. It is revealed that area of the Aral Sea had shrunk dramatically from 6.85×104 km2 to （8.32±0.19）×103 km2 with a shrinkage of （6.02±0.02）×104 km2 （about 87.85%） from 1960 to 2018. The area of the Aral Sea had shrunk by （5.94±0.02）×104 km2 （about 86.77%） during the period of 1960 - 2009， while the shrinkage rate had slowed down obviously and the area had decreased by 740.04 km2 （about 8.17%） during the period of 2009 - 2018. The results of this statistical analysis show that the enhanced human activities since 1960， especially the increase of irrigation water consumption and reservoir capacity， is the dominant factor rendering the rapid shrinkage of the Aral Sea. Hence， water consumption， especially for irrigation should be adjusted to adapt on-going warming in the Aral Sea basin as soon as possible.
Logarithmic transformation is used to transform the gray values of two kinds of nighttime light data into the same range， which improves the timeliness of the study. Based on the nighttime light data of 1992， 1997， 2002， 2007， 2012 and 2017， the urban built-up areas in the three provinces of Northeast China are extracted. In this paper， the position of light center in the three provinces and the whole Northeast China has been counted， the changes of the total amount of light in various resource-based cities have been counted also， and the influence factors of light changing has been discussed. The results show that Shenyang， Dalian， Harbin and Changchun have been developing continuously in the past 25 years and have a pull effect on the light obviously and increasingly. The light center of each province and the whole Northeast China are close to a line of Harbin-Changchun-Shenyang. The development gap between non-resource-based cities and resource-based cities gradually increases， and the overall development of resource-based cities is relatively slow. The total amount of light in Shenyang， Dalian， Harbin and Changchun has always maintained rapid growth， and has always maintained absolute advantages with other cities within non-resource-based cities. The development of resource-renewable cities is different and the internal differences are not obvious. Individual differences between resource-mature cities are gradually increasing， while those between resource-declining cities are relatively small. On the whole， the polarization and regional imbalance of urban development in the three provinces of Northeast China are becoming more and more prominent.
Aiming at the dynamic changes of complex geological and ecological environment along the mountain highways， in this paper， the area along the Dujiangyan-Wenchuan Highway was taken as a research object， and a method was proposed of evaluating regional scientific ecological environmental bearing capacity based on comprehensive quality and comprehensive quality clustering relationship to achieve the scientific evaluation of the bearing capacity of mountainous highways with geological disasters. The natural-environment-humanity model was used to construct the regional carrying capacity assessment index system， with the help of the analytic hierarchy process. Technique for Order Preference by Similarity to an Ideal Solution algorithm was adopted for calculation and weighted superposition to obtain the comprehensive status of the bearing capacity in the study area and to make grade classification according to the result of the bearing capacity in the study area. The results show that the spatial distribution of geological environment， ecological environment and socio-economic carrying capacity of the study area is highly clustered， accounting for 31.12% of the area suitable for the construction area， and 31.98% of the reserve control area， 36.79% of the area not suitable for construction. The triangle area being consisted of Yingxiu Town， Xuankou Town and Dujiangyan in the study area presents a large high-value clustering area with good comprehensive bearing capacity and high spatial concentration， which is conducive to planning and construction. The research results provide new idea for the accurate evaluation of regional carrying capacity and spatial pattern research， and will be useful for disaster prevention and mitigation work.
The winter cattle camp is a place where cattle and people can rest together during the nomadic process. The article interpreted the Landsat images of the Tizinafu River basin in 2000， 2010 and 2018， extracted the information of the winter cattle camps in the corresponding period， and explore the spatiotemporal characteristics of the winter cattle camps in the past 18 years. The results showed that： （1） The number of winter cattle camps in the study area increased from 306 in 2000 to 712 in 2018， and the number and acreage had increased significantly. （2） The hot spots distributed among the camps spread from the middle to the east， and the cold spots spread along the northwest-southeast direction. （3） The camps were distributed in gentle slopes with good water and grass conditions at an altitude of 2 100 to 3 000 m， and the transition from the shady slope to the sunny slope to the residents Point-direction migration， snow and ice melting was the main source of water for cattle and herders， and the distance from the road was two-level differentiation. （4） Population， stocking capacity， NDVI and rural road density were the main driving force for the time and space changes of winter cattle camps. In view of the relatively poor production and living conditions of herders， it is recommended that camping infrastructure should be increased and improved according to local conditions to help overcome poverty.
