The dataset contains all individual glacial storage (unit: km3) over the Qinghai-Tibetan Plateau in 1970s and 2000s. It is sourced from the resultant data of the paper entitled "Consolidating the Randolph Glacier Inventory and the Glacier Inventory of China over the Qinghai-Tibetan Plateau and Investigating Glacier Changes Since the mid-20th Century". The first draft of this paper has been completed and is planned to be submitted to Earth System Science Data journal. The baseline glacier inventories in 1970s and 2000s are the Randolph Glacier Inventory 4.0 dataset, and the Glacier Inventory of China, respectively. Based on the individual glacial boundaries extracted from the above-mentioned two datasets, the grid-based bedrock elevation dataset (https://www.ngdc.noaa.gov/mgg/global/global.html, DOI: 10.7289/v5c8276m), and the glacier surface elevation obtained by a slope-dependent method, the individual glacier volumes in 1970s and 2000s are then calculated. In addition, the calculated results of individual glacier volumes in this study have been compared and verified with the existent results of several glacier volumes, relevant remote sensing datasets, and the global glacier thickness dataset based on the average of multiple glacier model outputs (https://www.research-collection.ethz.ch/handle/20.500.11850/315707, doi:10.3929/ethz-b-000315707), and the errors in the calculations have also been quantified (the error rates of all mountains are within 10%). The established dataset in this study is expected to provide the data basis for the future regional water resources estimation and glacier ablation-involved researches. Moreover, the acquisition of the data also provides a new idea for the future glacier storage estimation.

This data set includes 2002/04-2019/12 Greenland ice sheet mass changes derived from satellite gravimetry measurements. The satellite gravimetry data come from the joint NASA/DLR Gravity Recovery And Climate Experiment mission twin satellites (GRACE, 2002/04 to 2017/06) and its successor, GRACE Follow-On (GRACE-FO, 2018/06 to present). In order to fill the data gap between GRACE and GRACE-FO, we further utilize gravity field solutions derived from high-low GNSS tracking data of ESA's Swarm 3-satellite constellation whose primary scientific objective is geomagnetic surveying. The data set is provided in Matlab data format, the ice sheet mass changes are transformed to equivalent water height in meters, expressed on 0.25°x0.25° grid with monthly temporal resolution. This data set can be used to study the characteristics of Greenland ice sheet mass changes in recent two decades and their relation with the global climate change.

Based on GRACE Level-1b satellite gravity data, a time series of mass change over Greenland for the period 2002 to 2016, with a spatial resolution of 1 degree × 1 degree and a time resolution of one month was developed by the satellite gravity team led by Professor Shen Yunzhong from Tongji University. The reference time of this time series is the mean time span between January 2004 and December 2009. During data processing, ICE5G model was used to reduce the effect of GIA, and the contribution of GAD was added back by using AOD1B RL06 from GFZ

From 1982 to 2015, the NDVI change data sets of different types of permafrost regions in the northern hemisphere have a temporal resolution of once every five years, covering the entire Arctic countries with a spatial resolution of 8km. Based on multi-source remote sensing, simulation, statistics and measured data, the regulation and service functions of Permafrost on Ecosystem in the northern hemisphere are quantified by using GIS and ecological methods, All the data are under quality control.

In recent years, the Antarctic Ice Sheet experiences substantial surface melt, and a large amount of meltwater formed on the ice surface. Observing the spatial distribution and temporal evolution of surface meltwater is a crucial task for understanding mass balance across the Antarctic Ice Sheet. This dataset provides a 30 m surface meltwater coverage, extracted from Landsat images, in the typical ablation zone of the ice sheet (Alexandria Island, Antarctic Peninsula) from 2000 to 2019. The projection of this dataset is South Polar Stereographic. The formats of the dataset are vector (.shp) and raster (.tif).

Data content: national economy_ Industrial value added (monthly) (2010-2019) Data source and processing method: the original industrial economic data of China (including the third pole) from the official website of the world bank and sina.com from 2010 to 2019 are obtained through data sorting, screening and cleaning. The data start time is from 2010 to 2019 in Microsoft Excel (xlsx) format.

Data content: Foreign Economic and trade_ Total import and export of goods (1952-2019) and foreign economic and trade_ Total import and export by trade (1981-2019) Data sources and processing methods: the original data of China's foreign trade and investment from 2015 to 2019 (including the third pole) were obtained from the official website of the world bank and sina.com, and the foreign trade and investment data set of China (including the third pole) from 1952 to 2019 was obtained through data sorting, screening and cleaning. The data start time is from 1952 to 2019 in Microsoft Excel (xlsx) format.

Data content: annual GDP statistics (1990-2019), quarterly cumulative GDP statistics (1990-2019) and GDP (2010-2019) Data sources and processing methods: the original macroeconomic data of China (including the third pole) from the official website of the world bank and sina.com from 1990 to 2019 are obtained through data sorting, screening and cleaning. The data are stored in Microsoft Excel (xlsx) format.

The sand drift potential (DP, in vector units (VU)) is calculated by DPi=∑U^2*[U-Ut]*fu where i represents 16 directions: N, NNE, NE, NEE, E, EES, ES, ESS, S, SSW, WS, WWS, W, WWN, NW and NNW; U is the sand-moving wind speed at the standard height of 10 m; Ut is the threshold wind velocity, which is the minimum wind velocity at the 10 m height to cause sediment particles in saltation; and fu is the fraction of time when the wind speed is higher than Ut. The 2 m s-1 bin is adopted in each sand-moving wind direction, corresponding to the mean wind speeds of 7, 9, 11, 13, 15, 17, 19, 21, 23, 25, 27, 29, 31, 33 and 34 m s-1, and the sum of these bins is the final DP in the wind direction. Note that these mean wind speeds have been expressed in knots by the approximate conversion (1 knot = 0.5144 m s-1) to ensure the valid classification of wind energy (low energy, <200 VU; intermediate energy, ≥200 VU and <400 VU; high energy, ≥400 VU) developed by Fryberger. The divisor used in calculating the frequency of effective sand-moving winds from different directions is the total hour number of Julian years (8760 hours for common years or 8784 hours for leap years). The wind speed and wind direction data from 1950 to 2021 were hourly estimates of 10 m u-component of wind and 10 m v-component wind with a horizontal resolution of 0.1°×0.1° generated with the ERA5-Land dataset.

(1) This is a literature-based eddy covariance carbon exchange dataset on the Tibetan Plateau, including air temperature, soil temperature, precipitation, ecosystem productivity and other parameters. (2) The data set is based on the field measured data of vorticity, and adopts the internationally recognized standard processing method of vorticity related data. The basic process includes: outlier elimination coordinate rotation WPL correction storage item calculation precipitation synchronization data elimination threshold elimination outlier elimination U * correction missing data interpolation flux decomposition and statistics. This data set also contains the model simulation data calibrated based on the vorticity correlation data set. (3) the data set has been under data quality control, and the data missing rate is 37.3%, and the missing data has been supplemented by interpolation. (4) The data set has scientific value for understanding carbon sink function of alpine wetland, and can also be used for correction and verification of mechanism model.

Iceberg calving, one of the key process of Antarctic mass balance, has been regarded as an important variable in fine monitoring the changes of ice shelves. The authors used multi-source remote sensing data near early August of each year from 2005 to 2020, including ENVISAT ASAR (WSM) images from 2005 to 2011, Terra/Aqua MODIS 7-2-1 band composite images from 2012 to 2014, Landsat-8 OLI 4-3-2 band composite images from 2013 to 2020, and Sentinel-1 SAR (EW) images from 2015 to 2020, to generate annual circum-Antarctic image mosaics after pre-processing. Next, combining MEaSUREs ice velocity dataset, grounding line, ice thickness dataset (Bedmap 2 and Bedmachine), spatial calculation and map digitization techniques were applied to extract all annual calving events larger than 1 km² that occurred on the Antarctic ice shelves from August 2005 to August 2020. Also, their area, thickness, mass and calving recurrence cycle were calculated to derive the annual iceberg calving dataset of the Antarctic ice shelves (2005-2020). This dataset contains the distribution of 15-year annual calving events, along with the attributes of each individual calving event including calving year, length, area, average thickness, mass, and recurrence interval. This dataset can directly reflect the magnitude characteristics and distribution of Antarctic iceberg calving in different years, which fills the gap of fine monitoring dataset of iceberg calving and provides fundamental data for subsequent research on calving mechanism and mass balance of Antarctic ice shelf-ice sheet system.

The rainfall data set of the southern Qinghai Tibet Plateau is fused by the satellite and the ground station. The data is in ASCII format, with a temporal resolution of 1 day and a horizontal spatial resolution of 0.1 °， The time coverage is from June 10 to October 31 in 2014-2019, which can provide driving data for rainfall verification and hydrological simulation in the southern Tibetan Plateau. The data set is based on the rainfall data of China Meteorological Administration and Hydrological Bureau of the Ministry of water resources after strict quality control , which is the highest density ground station network in the region so far. Dynamic Bayesian Model Average method is used to merge satellite precipitation products, i.e., GPM-IMERG, GSMaP, and CMORPH, based on the likelihood measurements of a high-density rainfall gauge network. The statistical accuracy evaluation and hydrological simulation verification of the merged data preforms better than the source satellite data, and also better than the popular reanalysis data CHIRPS and MSWEP.