As the main parameter in the land surface energy balance, surface temperature indicates the degree of land-atmosphere energy and water transfer and is widely used in research on climatology, hydrology and ecology. In the study of frozen soil, climate is one of the decisive factors for the existence and development of frozen soil. The surface temperature is the main climatic factor affecting the distribution of frozen soil and affects the occurrence, development and distribution of frozen soil. It is the upper boundary condition for modelling frozen soil and is significant to the study of hydrological processes in cold regions. The data set was based on the DEM and observation station data of the Tibetan Plateau Engineering Corridor and analysed the changing trend of surface temperature on the Tibetan Plateau from 2000 to 2014. Using the surface temperature data products MOD11A1/A2 and MYD11A1/A2 of MODIS aboard Terra and Aqua, the surface temperature information under cloud cover was reconstructed based on the spatio-temporal information of the images. The reconstruction information and surface temperature representativeness problems were analysed using information obtained from 8 sites, including the Kunlun Mountains (wetland, grassland), Beiluhe (grassland, meadow), Kaixinling (meadow, grassland), and Tanggula Mountain (meadow, wetland). According to the correlation coefficient (R2), root-mean-square error (RMSE), mean absolute error (MAE) and mean deviation (MBE), the following results were obtained: (1) the reconstruction accuracy of MODIS surface temperature under cloud cover is higher when it is based on spatio-temporal information; (2) the weighted average representation is the best when generalizing four observations of Terra and Aqua. By analysing the reconstruction of MODIS surface temperature information and representativeness problems, the average annual MODIS surface temperature data of the Tibetan Plateau and the engineering corridor from 2000 to 2010 were obtained. According to the data set, the surface temperature from 2000 to 2010 also experienced volatile rising trends from 2000 to 2010, which is basically consistent with the changing trend of the climate change in the permafrost regions of the Tibetan Plateau and the Qinghai-Tibet Engineering Corridor.

The field observation platform of the Tibetan Plateau is the forefront of scientific observation and research on the Tibetan Plateau. The land surface processes and environmental changes based comprehensive observation of the land-boundary layer in the Tibetan Plateau provides valuable data for the study of the mechanism of the land-atmosphere interaction on the Tibetan Plateau and its effects. This dataset integrates the 2005-2016 hourly atmospheric, soil hydrothermal and turbulent fluxes observations of Qomolangma Atmospheric and Environmental Observation and Research Station, Chinese Academy of Sciences (QOMS/CAS), Southeast Tibet Observation and Research Station for the Alpine Environment, CAS (SETORS), the BJ site of Nagqu Station of Plateau Climate and Environment, CAS (NPCE-BJ), Nam Co Monitoring and Research Station for Multisphere Interactions, CAS (NAMORS), Ngari Desert Observation and Research Station, CAS (NADORS), Muztagh Ata Westerly Observation and Research Station, CAS (MAWORS). It contains gradient observation data composed of multi-layer wind speed and direction, temperature, humidity, air pressure and precipitation data, four-component radiation data, multi-layer soil temperature and humidity and soil heat flux data, and turbulence data composed of sensible heat flux, latent heat flux and carbon dioxide flux. These data can be widely used in the analysis of the characteristics of meteorological elements on the Tibetan Plaetau, the evaluation of remote sensing products and development of the remote sensing retrieval algorithms, and the evaluation and development of numerical models.

Relationship between modern pollen and climate, and its representative to vegetation are the important references in explaining and reconstructing past climate and vegetation qualitatively or quantitatively. To extrct past climate and vegetation signals from fossil pollen spectrum of a lacustrine sediment, a corresponding modern pollen dataset collected from lake-sediment surface is necessary. At present, there are a few modern pollen datasets extracted from lake sediment-surface established on the Tibetan Plateau, however, the geographic gaps (e.g. the central and east Tibetan Plateau) of available sampled lakes influence the correct understanding. To ensure the even distribution of the representative lakes, we collected lake sediment-surface samples (n=117) covering the alpine meadow evenly on the east and central Tibetan Plateau, in July and August 2018. For pollen extraction, approximately 10 g (wet original sediment) per sample were sub-sampled. Pollen sample was processed by the standard acid-alkali-acid procedures followed by 7-μm-mesh sieving. More than 500 terrestrial pollen grains were counted for each sample. Pollen assemblages of the dataset from alpine meadow are dominated by Cyperaceae (mean is 68.4%, maximum is 95.9%), with other herbaceous pollen taxa as commen taxa including Poaceae (mean is 10.3%, maximum is 87.7%), Ranunculaceae (mean is 4.8%, maximum is 33.6%), Artemisia (mean is 3.7%, maximum is 24.5%), Asteraceae (mean is 2.1%, maximum is 33.6%), etc. Salix (mean is 0.4%, maximum is 5.3%) is the major shrub taxon in these pollen assemblages, while arboreal taxa occur with low percentages generally (mean of total arboreal percentages is 0.9% (maximum is 5.8%), including mainly Pinus (mean is 0.3%, maximum is 1.8%), Betula (mean is 0.1%, maximum is 0.9%) and Alnus (mean is 0.1%, maximum is 0.7%). These pollen assemblages represent the plant components well in the alpine meadow communities, although they are influenced slightly by long-distance pollen grain transported by wind or river (such as these arboreal pollen taxa). Together with pollen counts and percentages, we also provided the modern climatic data for the sampled lakes. The China Meteorological Forcing Dataset (CMFD; gridded near-surface meteorological dataset) with a temporal resolution of three hours and a spatial resolution of 0.1° was employed, and the climatic data of the nearest pixel of one sampled lake was defined to represent climatic conditions of the lake. Finally, the mean annual precipitation (Pann), mean annual temperature (Tann) and mean temperature of the coldest month (Mtco) and warmest month (Mtwa) are calculated for each sampled lake.

This data set comprises the plateau soil moisture and soil temperature observational data based on the Tibetan Plateau, and it is used to quantify the uncertainty of model products of coarse-resolution satellites, soil moisture and soil temperature. The observation data of soil temperature and moisture on the Tibetan Plateau (Tibet-Obs) are from in situ reference networks at four regional scales, which are the Nagqu network of cold and semiarid climate, the Maqu network of cold and humid climate, and the Ali network of cold and arid climate，and Pali network. These networks provided representative coverage of different climates and surface hydrometeorological conditions on the Tibetan Plateau. - Temporal resolution: 1hour - Spatial resolution: point measurement - Measurement accuracy: soil moisture, 0.00001; soil temperature, 0.1 °C; data set size: soil moisture and temperature measurements at nominal depths of 5, 10, 20, 40 - Unit: soil moisture, cm ^ 3 cm ^ -3; soil temperature, °C

The dataset includes soil physical and chemical attributes: pH value, organic matter fraction, cation exchange capacity, root abundance, total nitrogen (N), total phosphorus (P), total potassium (K), alkali-hydrolysable N, available P, available K, exchangeable H+, Al3+, Ca2+, Mg2+, K+ , Na+, horizon thickness, soil profile depth, sand, silt and clay fractions, rock fragment, bulk density, porosity, structure, consistency and soil color. Quality control information (QC) was provided. The resolution is 30 arc-seconds (about 1 km at the equator). The vertical variation of soil property was captured by eight layers to the depth of 2.3 m (i.e. 0- 0.045, 0.045- 0.091, 0.091- 0.166, 0.166- 0.289, 0.289- 0.493, 0.493- 0.829, 0.829- 1.383 and 1.383- 2.296 m) for convenience of use in the Common Land Model and the Community Land Model (CLM). 1.THSCH.nc: Saturated water content of FCH 2.PSI_S.nc: Saturated capillary potential of FCH 3.LAMBDA.nc: Pore size distribution index of FCH 4.K_SCH.nc: Saturate hydraulic conductivity of FCH 5.THR.nc: Residual moisture content of FGM 6.THSGM.nc: Saturated water content of FGM 7.ALPHA.nc: The inverse of the air-entry value of FGM 8.N.nc: The shape parameter of FGM 9.L.nc: The pore-connectivity parameter of FGM 10.K_SVG.nc: Saturated hydraulic conductivity of FGM 11.TH33.nc: Water content at -33 kPa of suction pressure, or field capacity 12.TH1500.nc: Water content at -1500 kPa of suction pressure, or permanent wilting point

The past frozen soil map of the Tibetan Plateau was based on a small number of temperature station observations and used a classification system based on continuity. This data set used the geographically weighted regression model (GWR) to synthesize MODIS surface temperature, leaf area index, snow cover ratio and multimodel soil moisture forecast products of the National Meteorological Information Center through spatiotemporal reconstruction. In addition, precipitation observations of more than 40 meteorological stations, the precipitation products of FY2 satellite observations and the multiyear average temperature observation data of 152 meteorological stations from 2000 to 2010 were integrated to simulate the average temperature data of the Tibetan Plateau, and the permafrost thermal condition classification system was used to classify permafrost into several types: Very cold, Cold, Cool, Warm, Very warm, and Likely thawing. The map shows that, after deducting lakes and glaciers, the total area of permafrost on the Tibetan Plateau is approximately 1,071,900 square kilometers. Verification shows that this map has higher accuracy. It can provide support for future planning and design of frozen soil projects and environmental management.

This dataset uses daily temperature data from SMMR (1978-1987), SSM/I (1987-2009) and SSMIS (2009-2015). It is generated by the dual-index (TB, 37v, SG) freeze-thaw discrimination algorithm. The classification results include the frozen surface, the thawed surface, the deserts and water bodies. The data coverage is the main part of China’s mainland, with a spatial resolution of 25.067525 km via the EASE-Grid projection method, and it is stored in ASCIIGRID format. All the ASCII files in this data set can be opened directly with a text program such as Notepad. Except for the head file, the body content is numerically characterized by the freeze/thaw status of the surface soil: 1 for frozen, 2 for thawed, 3 for desert, and 4 for precipitation. If you want to use the icon for display, we recommend using the ArcView + 3D or Spatial Analyst extension module for reading; in the process of reading, a grid format file will be generated, and the displayed grid file is the graphical expression of the ASCII file. The read method comprises the following. [1] Add the 3D or Spatial Analyst extension module to the ArcView software and then create a new View. [2] Activate View, click File menu, and select the Import Data Source option. When the Import Data Source selection box pops up, select ASCII Raster in the Select import file type box. When the dialog box for selecting the source ASCII file automatically pops up, click to find any ASCII file in the data set, and then press OK. [3] Type the name of the Grid file in the Output Grid dialog box (it is recommended that a meaningful file name is used for later viewing) and click the path to store the Grid file, press OK again, and then press Yes (to select integer data) and Yes (to put the generated grid file into the current view). The generated files can be edited according to the Grid file standard. This completes the process of displaying an ASCII file into a Grid file. [4] In the batch processing, the ASCIGRID command of ARCINFO can be used to write AML files, and then use the Run command to complete the process in the Grid module: Usage: ASCIIGRID <in_ascii_file> <out_grid> {INT | FLOAT}. The production of this data is supported by the following Natural Science Foundation Projects: Environmental and Ecological Science Data Center of West China (90502010), Land Data Assimilation System of West China (90202014) and Active and Passive Microwave Radiation Transmission Simulation and Radiation Scattering Characteristics of the Frozen Soil (41071226).

A comprehensive understanding of the permafrost changes in the Qinghai Tibet Plateau, including the changes of annual mean ground temperature (Magt) and active layer thickness (ALT), is of great significance to the implementation of the permafrost change project caused by climate change. Based on the CMFD reanalysis data from 2000 to 2015, meteorological observation data of China Meteorological Administration, 1 km digital elevation model, geo spatial environment prediction factors, glacier and ice lake data, drilling data and so on, this paper uses statistics and machine learning (ML) method to simulate the current changes of permafrost flux and magnetic flux in Qinghai Tibet Plateau The range data of mean ground temperature (Magt) and active layer thickness (ALT) from 2000 to 2015 and 2061 to 2080 under rcp2.6, rcp4.5 and rcp8.5 concentration scenarios were obtained, with the resolution of 0.1 * 0.1 degree. The simulation results show that the combination of statistics and ML method needs less parameters and input variables to simulate the thermal state of frozen soil, which can effectively understand the response of frozen soil on the Qinghai Tibet Plateau to climate change.

A multi-layer soil particle-size distribution dataset (sand, silt and clay content), based on USDA (United States Department of Agriculture) standard for regional land and climate modelling in China. was developed The 1:1,000,000 scale soil map of China and 8595 soil profiles from the Second National Soil Survey served as the starting point for this work. We reclassified the inconsistent soil profiles into the proper soil type of the map as much as possible because the soil classification names of the map units and profiles were not quite the same. The sand, silt and clay maps were derived using the polygon linkage method, which linked soil profiles and map polygons considering the distance between them, the sample sizes of the profiles, and soil classification information. For comparison, a soil type linkage was also generated by linking the map units and soil profiles with the same soil type. The quality of the derived soil fractions was reliable. Overall, the map polygon linkage offered better results than the soil type linkage or the Harmonized World Soil Database. The dataset, with a 1-km resolution, can be applied to land and climate modelling at a regional scale. Data characteristics： projection：projection Coverage: China Resolution: 0.00833 (about 1 km) Data format: FLT, TIFF Value range: 0%-100% Document describing： Floating point raster files include: Sand1. FLT, clay1. FLT -- surface (0-30cm) sand, clay content. Sand2. FLT, clay2. FLT -- content of sand and clay in the bottom layer (30-100cm). PSD. HDR -- header file: Ncols - the number of columns Nrows- rows Xllcorner - latitude in the lower left corner Yllcorner - longitude of the lower left corner Cellsize - cellsize NODATA_value - a null value byteorder - LSBFIRST, Least Significant Bit First TIFF raster files include: Sand1. Tif, clay1. Tif - surface (0-30cm) sand, clay content. Sand2. Tif, clay2. Tif - bottom layer (30-100cm) sand, clay content.

The distribution of lakes in space and its change over time are closely related to agricultural, environmental and ecological issues, and are critical factors for human socio-economic development. In the past decades, satellite based remote sensing has been developed rapidly to provide essential data sources for monitoring temporal lakes dynamics with its advantage of rapidness, wide coverage, and lower cost. This dataset was produced from Landsat images using the automated water detection method (Feng et al, 2015). We collected 96,278 Landsat images (about 25 terabytes) that acquired since 2000 with less than 80% cloud contamination in the arid region of central Asia and Tibetan Plateau. Water is detected in each of the image and then aggregated to monthly temporal resolution by taking advantage of the high-performance processing capability and large data storage provided by Global Land Cover Facility (GLCF) at University of Maryland. The results are validated systematically and quantitatively using manually interpreted dataset, which consists of a set of locations collected by a stratified random sampling strategy to effectively represent different spatial-temporal distributions in the region. The validation suggests high accuracy of the results (overall accuracy: 99.45(±0.59); user accuracy: 85.37%±(3.74); produce accuracy: 98.17(±1.05)).