The No. 8 hydrological section is located at Gaotai Heihe River Bridge (39 ° 23′22 .93 ″ N, 99 ° 49′37 .29″ E, 1347 m a.s.l.) in the middle reaches of the Heihe River Basin, Zhangye, Gansu Province. The dataset contains observations from the No.8 hydrological section from 17 June, 2012, to 24 November, 2012. The width of this section is 130 meters. The water level was measured using SR50 ultrasonic range and the discharge was measured using cross-section reconnaissance by the StreamPro ADCP. The dataset includes the following sections: Water level (recorded every 30 minutes) and Discharge. The data processing and quality control steps were as follows: 1) The water level data which collected from the hydrological station were averaged over intervals of 10 min for a total of 144 records per day. The missing data were denoted by -6999. 2) Data out the normal range records were rejected. 3) Unphysical data were rejected. For more information, please refer to Liu et al. (2016) (for multi-scale observation experiment or sites information), He et al. (2016) (for data processing) in the Citation section.

The dataset generated from the radiosonde observations in middle basin of Heihe River during 2012. The instrument type are RS92-SGP (Vaisala inc., Finland) or CF-06-A (Changfeng Micro-Electroinics, CHINA). Radiosondes were released during aerospace experiment, such as CASI/SAI, TASI, WIDAS sensors. Atmospheric parameters: pressure, temperature, relative humidity, wind speed and wind direction are measured or calculated at different altitude. This atmospheric parameter profiles can back up atmospheric correction in remote sensing. It can support meteorology research. Observation Site: 1. Wuxing Village: Latitude: 38°51′11.9″N，Longitude: 100°21′48.8″E，Altitude: 1563 m 2. Gaoya Hydrological Station Latitude: 39°8′7.2″N，Longitude: 100°23′59.0″E，Altitude: 1418 m 3. A’Rou Super Station Latitude: 38°03′17.9″N，Longitude: 100°27′28.1″E，Altitude: 2991 m Observation Instrument Type: RS92-SGP manufacture by Vaisala inc., Finland CF-06-A manufacture by Beijing Changfeng Micro-Electronics Technology Co., LTD, CHINA. Observation Time: Simultaneous observation time from 29 June, 2012 to 29 July, 2012 (UTC+8). Accessory data: Pressure, temperature, relative humidity, wind speed and wind direction profiles data.

The soil texture dataset of the Heihe River Basin (2011) is compiled by LIU Chao et al. (2011) by using the SOLIM model. Based on the famous Jenny equation of soil science, and according to the environmental factors such as climate, biology, topography and parent material, knowledge mining and fuzzy logic are combined on the basis of existing soil texture maps and soil profiles in Heihe River Basin. It is produced and integrated with thematic maps of glaciers and lakes. According to the different characteristics of the six ecological zones in Heihe River Basin, different mapping methods are used in the upper, middle and lower reaches. According to the different characteristics of six ecological zones in Heihe River Basin, different mapping methods are used in the upper, middle and lower reaches. The data is in grid format with 1KM spatial resolution and WGS-84 projection. Soil texture attributes and categories represent 0-30 cm topsoil texture attributes, derived from depth-weighted averages. The texname in the attribute table indicates the soil texture type name. Sandrange, siltrange, and clayrange respectively represent the sand, powder, and clay content ranges in the USDA soil triangle. Sandaverage, siltaverage and clayaverage are taken from the measured soil profiles, the average content of sand, silt and clay particles as the sand, silt and clay content of the soil type. (Note: The soil particle content of clay loam is derived from the soil quality map of Beijing Normal University). The soil texture classification standard is USDA, the sand grain size is defined as (2~0.05mm), the silt particle size is (0.05~0.002mm) and the clay size is defined as (<0.002mm).

The site No. 1 EC towers were used for the intercomparison field in the Yingke irrigation district (1552.75 m, 38°59′51.71″ N, 100°24′38.76″ E). The land surface is homogeneous and dominated by vegetables in the middle reaches of the Heihe River Basin. The precipitation comparison dataset was collected between 12 June, 2012, and 22 November, 2012. The dataset includes data for five different rain gauge types, i.e., pit gauge, Chinese standard manual precipitation gauge, siphon rain gauge, tipping bucket gauge, and weighting gauge. The mountain heights for these gauges were 0.0, 0.7, 1.2, 1.5, and 1.5 m, respectively. The data were recorded every 1 hour, 1 day, 10 minutes, 10 minutes, and 10 minutes, respectively. The main objective of the data collection was to perform an intercomparison of in situ rainfall measurements. The data processing and quality control steps were as follows: 1) The water level data which collected from the hydrological station were averaged over intervals of 10 min for a total of 144 records per day. The missing data were denoted by -6999. 2) Data out the normal range records were rejected. 3) Unphysical data were rejected. For more information, please refer to Liu et al. (2016) (for multi-scale observation experiment or sites information), He et al. (2016) (for data processing) in the Citation section.

DEM is the English abbreviation of Digital Elevation Model, which is the important original data of watershed topography and feature recognition.DEM is based on the principle that the watershed is divided into cells of m rows and n columns, the average elevation of each quadrilateral is calculated, and then the elevation is stored in a two-dimensional matrix.Since DEM data can reflect local topographic features with a certain resolution, a large amount of surface morphology information can be extracted through DEM, which includes slope, slope direction and relationship between cells of watershed grid cells, etc..At the same time, the surface flow path, river network and watershed boundary can be determined according to certain algorithm.Therefore, to extract watershed features from DEM, a good watershed structure pattern is the premise and key of the design algorithm. Elevation data map 1km data formed according to 1:250,000 contour lines and elevation points in China, including DEM, hillshade, Slope and Aspect maps. Data set projection: Two projection methods: Equal Area projection Albers Conical Equal Area (105, 25, 47) Geodetic coordinates WGS84 coordinate system

This data comes from the National Geographic Information Resources Catalogue Service System, which was provided free to the public by the National Basic Geographic Information Center in November 2017. We have spliced and cut the source of the three rivers as a whole, so as to facilitate the use of the study of the source area of the three rivers. The data trend is 2015. This data set includes 1:250,000 residential place names (AANP) in Sanjiangyuan area, including administrative place names at all levels and urban and rural residential place names. Names and Definitions of Attribute Items of Residential Place Name Data (AANP): Attribute Item Description Fill in Example NAME Name Quanqu Village PINYIN Chinese Pinyin Quanqucun CLASS Geographical Name Classification Code AK GNID Place Name Code 632524000000 XZNAME Township Name Ziketan Township

This dataset contains the flux measurements from the Daman superstation eddy covariance system (EC) at the highest layer in the flux observation matrix from 30 May to 15 September, 2012. The site (100.37223° E, 38.85551° N) was located in a cropland (maize surface) in Daman irrigation district, which is near Zhangye, Gansu Province. The elevation is 1556.06 m. The EC was installed at a height of 34 m; the sampling rate was 10 Hz. The sonic anemometer faced north, and the separation distance between the sonic anemometer and the CO2/H2O gas analyzer (CSAT3&Li7500A) was 0.17 m. Raw data acquired at 10 Hz were processed using the Edire post-processing software (University of Edinburgh, http://www.geos.ed.ac.uk/abs/research/micromet/EdiRe/), including spike detection, lag correction of H2O/CO2 relative to the vertical wind component, sonic virtual temperature correction, coordinate rotation (2-D rotation), corrections for density fluctuation (Webb-Pearman-Leuning correction), and frequency response correction. The EC data were subsequently averaged over 30 min periods. Moreover, the observation data quality was divided into three classes according to the quality assessment method of stationarity (Δst) and the integral turbulent characteristics test (ITC), which was proposed by Foken and Wichura [1996]: class 1 (level 0: Δst<30 and ITC<30), class 2 (level 1: Δst<100 and ITC<100), and class 3 (level 2: Δst>100 and ITC>100), representing high-, medium-, and low-quality data, respectively. In addition to the above processing steps, the half-hourly flux data were screened in a four-step procedure: (1) data from periods of sensor malfunction were rejected; (2) data before or after 1 h of precipitation were rejected; (3) incomplete 30 min data were rejected when the missing data constituted more than 3% of the 30 min raw record; and (4) data were rejected at night when the friction velocity (u*) was less than 0.1 m/s. There were 48 records per day; the missing data were replaced with -6999. Moreover, suspicious data were marked in red. The released data contained the following variables: data/time, wind direction (Wdir, °), wind speed (Wnd, m/s), the standard deviation of the lateral wind (Std_Uy, m/s), virtual temperature (Tv, ℃), H2O mass density (H2O, g/m^3), CO2 mass density (CO2, mg/m^3), friction velocity (ustar, m/s), stability (z/L), sensible heat flux (Hs, W/m^2), latent heat flux (LE, W/m^2), carbon dioxide flux (Fc, mg/ (m^2s)), quality assessment of the sensible heat flux (QA_Hs), quality assessment of the latent heat flux (QA_LE), and quality assessment of the carbon flux (QA_Fc). In this dataset, the time of 0:30 corresponds to the average data for the period between 0:00 and 0:30; the data were stored in *.xlsx format. For more information, please refer to Liu et al. (2016) (for multi-scale observation experiment or sites information), Xu et al. (2013) (for data processing) in the Citation section.

On 25 August 2012, a Leica ALS70 airborne laser scanner boarded on the Y-12 aircraft was used to obtain LiDAR DSM point cloud data. Leica ALS70 airborne laser scanner has unlimited numbers of returns intensities measurements including the first, second, third return intensities. The wavelength of laser light is 1064 nm. The absolute flight altitude is 5200 m with the point cloud density 1 point per square meter. Aerial LiDAR-DSM was obtained through parameter calibration, automatic classification of point cloud density and manual editing.

This dataset contains the spectra of white cloth and black cloth obtained in the simultaneous time during the airborn remote sensing which supports the airboren data preprocessing as CASI, SASI and TASI , and the spetra of the typical targets in the middle reaches of the Heihe River Basin. Instruments: SVC-HR1024 from IRSA, ASD Field Spec 3 from CEODE, Reference board Measurement method: the spectra radiance of the targets are vertically measured by the SVC or ASD; before and after the target, the spectra radiance of the reference board is measured as the reference. This dataset contains the spectra recorded by the SVC-HR1024 ( in the format of .sig which can be opened by the SVC-HR1024 software or by the notepad ) and the ASD (in the format of .asd), the observation log (in the format of word or excel), and the photos of the measured targets. Observation time: 15-6-2012, the spectra of typical targets in the EC matrix using SVC 16-6-2012, the spectra of typical targets in the wetland by SVC 29-6-2012, the spectra of typical vegetation and soil in Daman site and Gobi site by ASD 29-6-2012, the spectra of white cloth and black cloth by ASD which is simultaneous with the airborne CASI data 30-6-2012, the spectra of vegetation and soil in the desert by ASD 5-7-2012, the spectra of white cloth and black cloth by ASD which is simultaneous with the airborne CASI data 7-7-2012, the spectra of corn in the Daman site for the research of daily speral variation. 8-7-2012, the spectra of white cloth and black cloth by ASD which is simultaneous with the airborne CASI data 8-7-2012, the spectra of corn in the Daman site by ASD for the research of daily speral variation 9-7-2012, the spectra of corn in the Daman site by ASD for the research of daily speral variation 10-7-2012, the spectra of corn in the Daman site by ASD for the research of daily speral variation 11-7-2012, the spectra of corn in the Daman site by ASD for the research of daily speral variation. The time used in this dataset is in UTC+8 Time.

Taking 2005 as the base year, the future population scenario prediction adopted the Logistic model of population; not only is it better able to describe the change pattern of population and biomass, but it is also widely applied in the economic field. The urbanization rate was predicted using the urbanization Logistic model. Based on the existing urbanization horizontal sequence value, the prediction model was established by acquiring the parameters in the parametric equation applying nonlinear regression. The urban population was calculated by multiplying the predicted population by the urbanization rate. The Logistic model was used to predict the future gross national product of each county (or city), and then according to the economic development level of each county (or city) in each period (in terms of real GDP per capita), the corresponding industrial structure scenarios in each period were set, and the output value of each industry was predicted. The trend of changing industrial structure in China and the research area lagged behind the growth of GDP and was therefore adjusted according to the need of the future industrial structure scenarios of the research area.

The 1:1,000,000 road data set of the North Pole includes the Arctic_Major_Routes, the Arctic_Minor_Routes, the Arctic_railway vector space data and the related attribute data: road Name and Type. The data comes from the 1:100,000 ADC_WorldMap global data set，The data through topology, warehousing and other data quality inspection，It's most comprehensive, current and seamless geographic digital data for the whole earth. The world map coordinate system is latitude and longitude, WGS84 datum surface，Arctic specific projection parameters（North_Pole_Stereographic）.

This dataset contains the flux measurements from the large aperture scintillometer (LAS) at site No.3 in the flux observation matrix. There were two types of LASs at site No.3: German BLS900 and Netherland Kipp&zonen. The observation periods were from 6 June to 20 September, 2012, and 19 June to 20 September, 2012, for the BLS900 and the Kipp&zonen, respectively. The north tower is placed with the receiver of BLS900 and the transmitter of Kipp&zonen, and the south tower is placed with the transmitter of BLS900 and the receiver of Kipp&zonen. The site ( (north: 100.373° E, 38.883° N; south: 100.372° E, 38.856° N) was located in the Yingke irrigation district, which is near Zhangye, Gansu Province. The elevation is 1552.75 m. The underlying surface between the two towers contains corn, greenhouse, and village. The effective height of the LASs was 33.45 m; the path length was 3111 m. Data were sampled at 1 min intervals. Raw data acquired at 1 min intervals were processed and quality-controlled. The data were subsequently averaged over 30 min periods. The main quality control steps were as follows. (1) The data were rejected when Cn2 was beyond the saturated criterion (Cn2>3.36E-14). (2) Data were rejected when the demodulation signal was small (BLS900: Average X Intensity<1000; Kipp&zonen: Demod<-20 mv). (3) Data were rejected within 1 h of precipitation. (4) Data were rejected at night when weak turbulence occurred (u* was less than 0.1 m/s). The sensible heat flux was iteratively calculated by combining with meteorological data and based on Monin-Obukhov similarity theory. There were several instructions for the released data. (1) The data were primarily obtained from BLS900 measurements; missing flux measurements from the BLS900 were filled with measurements from the Kipp&zonen. Missing data were denoted by -6999. (2) The dataset contained the following variables: data/time (yyyy-mm-dd hh:mm:ss), the structural parameter of the air refractive index (Cn2, m-2/3), and the sensible heat flux (H_LAS, W/m^2). (3) In this dataset, the time of 0:30 corresponds to the average data for the period between 0:00 and 0:30; the data were stored in *.xlsx format. Moreover, suspicious data were marked in red. For more information, please refer to Liu et al. (2016) (for multi-scale observation experiment or sites information), Xu et al. (2013) (for data processing) in the Citation section.