Monthly Experimental Dataset of 0.1° Groundwater Storage Variations in the Tarim River Basin and Adjacent Areas (2002-2022)

XUE Dongping1,2,3GUI Dongwei1,3,4LIN Jingwu1,2,3LIU Qi1,2LIU Yunfei1,2JIN Qian1,2

1 Xinjiang Institute of Ecology and Geography，Chinese Academy of Sciences，Urumqi 830011，China2 Cele National Station of Observation & Research for Desert Grassland Ecosystem in Xinjiang，Cele，8483003 University of Chinese Academy of Sciences，Beijing 100049，China4 Xinjiang Institute of Physics and Chemistry，Chinese Academy of Sciences，Urumqi 830011，China

DOI:10.3974/geodb.2025.03.01.V1

Published:Mar. 2025

Visitors：606       Data Files Downloaded：4      
Data Downloaded：200.46 MB      Citations：

Key Words：

Tarim River Basin,Groundwater Storage Anomaly(GRACE),downscaling,machine learning

Abstract：

The Monthly Experimental Dataset of 0.1° Groundwater Storage Variations in the Tarim River Basin and Adjacent Areas was developed based on two downscaling schemes: a Semi-supervised Variational Autoencoder Algorithm (SSVAER) based on temporal continuity of pixels, and a Multiscale Geographically Weighted Regression (MGWR) model utilizing spatial correlations of pixels. Taking the Tarim River Basin and adjacent areas as the study area, we employed the mean values of GRACE Mascon products from two authoritative institutions - the Center for Space Research at University of Texas at Austin (CSR-RL06M) and NASA's Jet Propulsion Laboratory (JPL-RL06M) as primary data sources, covering the temporal span from April 2002 to June 2022 (210 monthly observations). Predictive variables included precipitation, snow water equivalent, land surface temperature, evaporation, runoff, and NDVI. Both schemes successfully enhanced the spatial resolution of GRACE-derived Groundwater Storage Anomalies (GWSA) from 0.5° to 0.1°. Results show that: (1) significant correlation preservation with coefficients ranging 0.94-0.98 between original and downscaled GWSA across >80% of the study area; (2) substantial improvement in groundwater characterization, with temporal continuity and spatial correlation approaches elevating correlation coefficients between GWSA and monitoring well data from 0.27 to 0.59 and 0.52 respectively; (3) robust model performance evidenced by root mean square errors (RMSE) consistently below 8. The dataset includes: (1) SSVAER-based downscaling results; (2) MGWR-derived downscaling outputs. The dataset is archived in .tif format with 0.1° spatial resolution. The dataset contains 420 data files with data size of 62.2 MB (compressed to one file with 50.1 MB). The analysis paper based on the dataset has been published at Science of the Total Environment, Volume 907, 2024.

Foundation Item：

Xinjiang Uygur Autonomous Region (2023TSYCLJ0049)；

Data Citation：

XUE Dongping, GUI Dongwei, LIN Jingwu, LIU Qi, LIU Yunfei, JIN Qian. Monthly Experimental Dataset of 0.1° Groundwater Storage Variations in the Tarim River Basin and Adjacent Areas (2002-2022)[J/DB/OL]. Digital Journal of Global Change Data Repository, 2025. https://doi.org/10.3974/geodb.2025.03.01.V1.

Data Product: