The data set contains data of three stations in the middle reaches: (1) the eddy related flux observation data of point 4 in the flux observation matrix from May 31 to September 17, 2012. The station is located in the Yingke irrigation area of Zhangye City, Gansu Province, and the underlying surface is the village. The longitude and latitude of the observation point are 100.35753e, 38.87752n and 1561.87m above sea level. The height of the eddy correlator is 4.2m (after August 19, the height of the eddy correlator is adjusted to 6.2m), the sampling frequency is 10Hz, the ultrasonic direction is due north, and the distance between the ultrasonic anemometer and the CO2 / H2O analyzer is 17cm. (2) Eddy related flux data of point 12 in the flux observation matrix from May 28 to September 21, 2012. The site is located in the farmland of Daman irrigation area, Zhangye City, Gansu Province, with corn as the underlying surface. The longitude and latitude of the observation point are 100.36631e, 38.86515n and 1559.25m above sea level. The height of the eddy correlator is 3.5m, the sampling frequency is 10Hz, the ultrasonic direction is north, and the distance between the ultrasonic anemometer and the CO2 / H2O analyzer is 15cm. (3) Eddy related flux data of point 14 in the flux observation matrix from May 30 to September 21, 2012. The site is located in the farmland of Yingke Irrigation District, Zhangye City, Gansu Province, with corn as the underlying surface. The longitude and latitude of the observation point are 100.35310e, 38.85867n and 1570.23m above sea level. The height of the eddy correlator is 4.6m, the sampling frequency is 10Hz, the ultrasonic direction is north, and the distance between the ultrasonic anemometer and the CO2 / H2O analyzer is 15cm. The original observation data of the eddy correlator is 10Hz. The published data is the 30 minute data processed by the edire software. The main processing steps include: outliers elimination, delay time correction, coordinate rotation (secondary coordinate rotation), frequency response correction, ultrasonic virtual temperature correction and density (WPL) correction, etc. At the same time, the quality evaluation of each flux value is mainly the test of atmospheric stability (Δ st) and turbulence similarity (ITC). The 30min flux value output by edire software was also screened: (1) data in case of instrument error; (2) data in 1H before and after precipitation; (3) data with loss rate greater than 3% in every 30min of 10Hz original data; (4) observation data with weak turbulence at night (U * less than 0.1M / s). The average period of observation data is 30 minutes, 48 data in a day, and the missing data is marked as - 6999. Suspicious data caused by instrument drift and other reasons shall be identified with red font. The published observation data include: date / time, wind direction WDIR (?), horizontal wind speed wnd (M / s), standard deviation of lateral wind speed STD uuy (M / s), ultrasonic virtual temperature TV (℃), water vapor density H2O (g / m3), carbon dioxide concentration CO2 (mg / m3), friction velocity ustar (M / s), stability Z / L (dimensionless), sensible heat flux HS (w / m2), latent heat flux Le (w / m2), two Carbon dioxide flux FC (mg / (M2S)), quality mark of sensible heat flux QA ﹤ HS, quality mark of latent heat flux QA ﹐ Le, quality mark of carbon dioxide flux QA ﹐ FC. The quality identification of sensible heat, latent heat and carbon dioxide flux is divided into three levels (quality identification 0: (Δ st < 30, ITC < 30); 1: (Δ st < 100, ITC < 100); the rest is 2). The meaning of data time, for example, 0:30 represents the average of 0:00-0:30; data is stored in *. XLS format. For station information, please refer to Liu et al. (2015), and for observation data processing, please refer to Liu et al. (2011) and Xu et al. (2013).

The data is clipped from "1: 1 million wetland data of China". "1: 1 million wetland data of China" mainly reflects the national marsh wetland information in the 2000s. It is expressed in geographic coordinates using the decimal degree. The main contents include: marsh wetland types, wetland water supply types, soil types, main vegetation types, geographical area, etc. Implemented the "Standard for Information Classification and Coding of Sustainable Development Information Sharing System of China". Data source of this database: 1:20 swamp map (internal version), Tibetan Plateau 1: 500,000 swamp map (internal version), swamp survey data 1: 1 million and national 1: 4 million swamp map; processing steps are: data source selection, preprocessing, digitization and encoding of marsh wetland elements, data editing processing, establishing topological relationships, edge processing, projection conversion, linking with attribute databases such as place names and obtaining attribute data.

The data is clipped from "1: 1 million wetland data of China". "1: 1 million wetland data of China" mainly reflects the national marsh wetland information in the 2000s. It is expressed in geographic coordinates using the decimal degree. The main contents include: marsh wetland types, wetland water supply types, soil types, main vegetation types, geographical area, etc. Implemented the "Standard for Information Classification and Coding of Sustainable Development Information Sharing System of China". Data source of this database: 1:20 swamp map (internal version), Tibetan Plateau 1: 500,000 swamp map (internal version), swamp survey data 1: 1 million and national 1: 4 million swamp map; processing steps are: data source selection, preprocessing, digitization and encoding of marsh wetland elements, data editing processing, establishing topological relationships, edge processing, projection conversion, linking with attribute databases such as place names and obtaining attribute data.

This data is the NCEP/DOE reanalysis data of 6h interval nested downscaling by WRF model in northwest China to a horizontal resolution of 12km, 364 grid points in the east-west direction, 251 grid points in the north-south direction and 31 layers in the vertical direction. The simulation time starts from 1979-01-01,06:00:00 and ends at 2013-12-31,23:00:00. The parameterization schemes of the model are as follows: Kain Frisch cumulus convection scheme, WSM3 cloud microphysics scheme, RRTM long wave scheme, Dudhia short wave scheme, Noah land surface model, YSU planetary boundary layer scheme. The file naming rules in the data set are: wrf_t2_YYYY.nc and wrf_rain_YYYY.nc, where YYYY is the annual abbreviation, t2 is the 2m temperature (unit ℃), and rain is the total surface precipitation (unit mm).

Regional climate mode RegCM4.3 parameter settings are as follows: horizontal resolution is 40 km, east and west 112 divisions Point, 84 grid points in north-south direction, grid center is 74.21 ° E, 44.76 ° N, projection is Lambert projection, vertical stratification It is divided into 23 layers, and the top pressure is 50 hPa; the initial boundary field uses ERA40 with a horizontal resolution of 2.5 °. NCEP / NCAR then analyzes the data, the time series takes the pattern data and the site interpolation data is the common part of the time series. SST data selection GISST (1948-2002), the planet boundary layer in the model is selected as Holtslag format, cumulus convection The scheme is Emanuel MIT format, the side boundary conditions are exponential relaxation conditions, and the land surface process is described as BATs. In mode Terrain data selection USGS GTOPO 30 30 ″ DEM digital elevation model, surface vegetation data selection USGS points GLCC (global land cover characterization) with a resolution of 30 ″. The data set is daily data, including surface pressure (hpa), 2m maximum temperature (K), 2m minimum temperature (K), 2m average temperature (K), and average surface precipitation (kg · m-2 · s-1) Naming rules, xx_xxxx_YYYY.nc. Where YYYY is the year, xx represents Central Asia, and xxxx is the drive data name (ERA40, NCEP1 respectively).

The GCMs dataset used in this dataset is CMIP3 comparison plan data (A1B (Medium Carbon Emissions, Global Common Development Scenarios that Focus on Economic Growth), A2 (High Carbon Emissions, Focus on Regional development scenarios for economic growth) and B1 (low carbon emissions, global common development scenarios that emphasize environmentally sustainable development) from the 24 GCM outputs in IPCC AR4 provided by PCMDI. This dataset uses the Delta method for downscaling, uses the 20C3M dataset from 1961 to 1990 as a reference, and uses the SRES dataset from 2010 to 2099 as the future scenario.

This dataset mainly includes the passive microwave brightness temperature obtained from the Scanning Multichannel Microwave Radiometer (SMMR) carried by the Nimbus-7 satellite, including 06H, 06V, 10H, 10V, 18H, 18V, 21H, 21V, 37H, 37V, a total of ten microwave channels with two transits (ascending & descending) brightness temperature per day from October 25, 1978 to August 20, 1987, where H represents horizontal polarization and V represents vertical polarization. Nimbus-7, launched in October 1978, is a solar-synchronous polar-orbiting satellite. The microwave sensor SMMR is a dual-polarization microwave radiometer that measures the brightness temperature of five frequencies (6.6GHz, 10.69GHz, 18.0GHz, 21.0GHz, 37.0GHz) on the surface. It scans the surface at a fixed incident angle of about 50.3 °, with a width of 780 km, and passes through the equator at noon 12:00 (ascending orbit) and 24:00 (descending orbit). The time resolution of SMMR is daily, but due to the wide distance between swaths, the same surface will be revisited every 5-6 days. 1. File format and naming: Each set of data is composed of remote sensing data files. The name and naming rules of each group of data files in the SMMR_Grid_China directory are as follows: SMMR-MLyyyydddA / D.subset.ccH / V (remote sensing data) Among them: SMMR stands for SMMR sensor; ML stands for multi-channel low resolution; yyyy stands for year; ddd stands for Julian Day of the year (1-365 / 366); A / D stands for ascending (A) and derailing (D ); subset represents the brightness temperature data in China; cc represents the frequency (6.6GHz, 10.69GHz, 18.0GHz, 21.0GHz, 37.0GHz); H / V represents horizontal polarization (H) and vertical polarization (V). 2. Coordinate system and projection: The projection method is an equal area secant cylindrical projection, and the double standard parallels are 30 degrees north and south. For more information about EASE-GRID, please refer to http://www.ncgia.ucsb.edu/globalgrids-book/ease_grid/. If you need to convert the EASE-Grid projection to Geographic projection, please refer to the ease2geo.prj file, the content is as follows: Input projection cylindrical units meters parameters 6371228 6371228 1 / * Enter projection type (1, 2, or 3) 0 00 00 / * Longitude of central meridian 30 00 00 / * Latitude of standard parallel Output Projection GEOGRAPHIC Spheroid KRASovsky Units dd parameters end 3. Data format: Stored as integer binary, each data occupies 2 bytes. The actual data stored in this dataset is the brightness temperature * 10. After reading the data, you need to divide by 10 to get the real brightness temperature. Spatial resolution: 25km; Time resolution: daily, from 1978 to 1987. 4. Spatial range: Longitude: 60.1 ° -140.0 ° East longitude; Latitude: 14.9 ° -55.0 ° north latitude. 5. Data reading Remote sensing image data files for each set of data can be opened in ENVI and ERDAS software.

Microwave emissivity of the surface characterization of the object to launch the ability of microwave radiation, spaceborne passive microwave emissivity can on macro, large scale integral expression of epicontinental microwave radiation is a passive microwave surface parameters in quantitative inversion experience for one of the important basic data, is also on the large scale understand epicontinental microwave radiation in a way.This data set is considered to carry on the Aqua satellite advanced microwave scanning radiometer (amsr-e) and moderate resolution imaging spectroradiometer (MODIS) synchronous observation characteristics, using the MODIS land surface temperature and atmospheric water vapor data as input, by considering the effects of atmospheric emissivity estimation model, produced a global sky conditions during the running of amsr-e sensor (June 2002 ~ October 2011) of the epicontinental multichannel bipolar microwave instantaneous emission rate.Through product low-frequency radio signal, data alignment, statistic analysis, the different emissivity characteristics of surface coverage condition, frequency dependence and correlation studies conducted confirmatory analysis, the results show that the instantaneous dynamic details of emissivity is rich, standard deviation within 0.02 month daily variation, the change of time and space, frequency dependent on and related to the understanding of the natural physical process. This data set includes amsr-e global land surface daily, daily, daily, monthly and monthly products in the whole life cycle, which can be used to carry out satellite based passive microwave remote sensing simulation, land surface model, and inversion research of land surface temperature, snow cover, atmospheric precipitation/moisture/precipitation.The projection coordinates of the data adopt the standard EASE-GRID projection, and the data storage method is binary floating point lattice (the size of the matrix is 1383*586). After the data is obtained, ENVI/IDL and other software or the corresponding program code can be read in the form of binary files. All land surface emissivity data produced are named according to the following rules: RADI_AMSRE_EM # # # # _yyymmdd_EG_V. Bin For example, file name: RADI_AMSRE_EM01_20060101_EG_V# EM##: 01 means daily, 05 means 5 days, 10 means ten days, HM means half a month, MO means a month Yyyymmdd: yyyy means year, mm means month, and dd means date V##: version number, such as 0.1, 1.0, etc., the units digit is the official version RADI: institute of remote sensing and digital earth, Chinese academy of sciences AMSRE: advanced microwave scanning radiometer

The Chinese regional surface meteorological element data set is a set of near-surface meteorological and environmental element reanalysis data set developed by the Qinghai-Tibet Plateau Research Institute of the Chinese Academy of Sciences. The data set is based on the existing Princeton reanalysis data, GLDAS data, GEWEX-SRB radiation data and TRMM precipitation data in the world, and is made by combining the conventional meteorological observation data of China Meteorological Administration. The temporal resolution is 3 hours and the horizontal spatial resolution is 0.1, including 7 factors (variables) including near-surface air temperature, near-surface air pressure, near-surface air specific humidity, near-surface full wind speed, ground downward short wave radiation, ground downward long wave radiation and ground precipitation rate. The physical meaning of each variable: | Meteorological Element || Variable Name || Unit || Physical Meaning | near-surface temperature ||temp|| K || instantaneous near-surface (2m) temperature | surface pressure || pres|| Pa || instantaneous surface pressure | specific humidity of near-surface air || shum || kg/ kg || instantaneous specific humidity of near-surface air | near ground full wind speed || wind || m /s || instantaneous near ground (anemometer height) full wind speed | downward short wave radiation || srad || W/m2 || 3-hour average (-1.5 HR ~+1.5 HR) downward short wave radiation | Downward Long Wave Radiation ||lrad ||W/m2 ||3-hour Average (-1.5 hr ~+1.5 hr) Downward Long Wave Radiation | precipitation rate ||prec||mm/hr ||3-hour average (-3.0 HR ~ 0.0 HR) precipitation rate For more information, please refer to the "User's Guide for China Meteorological Al Forcing Dataset" published with the data. The main changes in the latest version (01.06.0014) are: 1. Extend the data to December 2015 (except for short-wave and long-wave data, only until October 2015; the data from November to December 2015 are interpolated based on GLDAS data, and the error may be too large); 2. Set the minimum wind speed at 0.05 m/s; 3. Fixed a bug in the previous radiation algorithm to make our short wave and long wave data more reasonable in the morning and evening periods. 4. bug of precipitation data has been corrected, and the period involved in the change is 2011-2015.

Based on the data information provided by the data management center of Heihe project, the daily humidity data of 21 regular meteorological observation stations in Heihe River Basin and its surrounding areas and 13 national reference stations around Heihe River were collected and calculated. The spatial stability analysis is carried out to calculate the coefficient of variation. If the coefficient of variation is greater than 100%, the geographical weighted regression is used to calculate the relationship between the station and the geographical terrain factors, and the monthly humidity distribution trend is obtained; if the coefficient of variation is less than or equal to 100%, the common least square regression is used to calculate the relationship between the station humidity value and the geographical terrain factors (latitude, longitude, elevation, slope, aspect, etc.) The residual after removing the trend was fitted and corrected by HASM (high accuracy surface modeling method). Finally, the monthly average humidity distribution of the Heihe River Basin in 1961-2010 is obtained by adding the trend surface results and the residual correction results. Time resolution: monthly average humidity for many years from 1961 to 2010. Spatial resolution: 500M.

The station data information of 21 regular meteorological observation stations in Heihe River Basin and surrounding areas and 13 national benchmark stations around Heihe River provided by Heihe plan data management center are used to make statistics and collation of daily wind speed and calculate the monthly wind speed data of 1961-2010 for many years. The spatial stability analysis is carried out to calculate the variation coefficient. If the variation coefficient is greater than 100%, the geographical weighted regression is used to calculate the relationship between the station and the geographical terrain factors, and the monthly wind speed distribution trend is obtained; if the variation coefficient is less than or equal to 100%, the common least square regression is used to calculate the relationship between the station wind speed value and the geographical terrain factors (longitude and latitude, elevation, slope, aspect, etc.) The trend of monthly wind speed distribution is obtained, and the residual after removing the trend is fitted and corrected by HASM (high accuracy surface modeling method). Finally, the monthly average wind speed distribution of the Heihe River Basin in 1961-2010 is obtained by adding the trend surface results and the residual correction results. Time resolution: monthly average wind speed for many years from 1961 to 2010. Spatial resolution: 500M.

Based on the data of 21 regular meteorological observation stations in Heihe River Basin and its surrounding areas and 13 national benchmark stations around Heihe River provided by the data management center of Heihe plan, the daily sunshine hours are statistically sorted out and the monthly sunshine hours data of 1961-2010 for many years are calculated. The spatial stability analysis is carried out to calculate the variation coefficient. If the variation coefficient is greater than 100%, the geographical weighted regression is used to calculate the relationship between the station and the geographical terrain factors, and the monthly sunshine hours distribution trend is obtained; if the variation coefficient is less than or equal to 100%, the ordinary least square regression is used to calculate the sunshine hours and the geographical terrain factors (longitude, latitude, elevation, slope, aspect, etc.) of the station ）The distribution trend of sunshine hours per month is obtained, and the residuals after removing the trend are fitted and corrected by HASM (high accuracy surface modeling method). Finally, the monthly average sunshine hours distribution of the Heihe River Basin in 1961-2010 is obtained by adding the trend surface results and the residual correction results. Time resolution: monthly average sunshine hours for many years from 1961 to 2010. Spatial resolution: 500M.