Wang et al. (2026) Terrestrial water storage variations and drought characteristics in the upper yellow river basin revealed by joint GNSS-GRACE analysis

Identification

• Journal: Journal of Hydrology Regional Studies
• Year: 2026
• Date: 2026-02-17
• Authors: Zi Wang, Hai Zhu, Kejie Chen, Mingjia Li, Shunqiang Hu, Junguo Liu, Qingfeng Hu, Liang Sun, Ge Gu, He Li
• DOI: 10.1016/j.ejrh.2026.103255

Research Groups

• Department of Earth and Space Sciences, Southern University of Science and Technology, Shenzhen, China
• Department of Land Surveying and Geo-Informatics, The Hong Kong Polytechnic University, Hong Kong
• Guangdong Provincial Key Laboratory of Geophysical High-resolution Imaging Technology, Southern University of Science and Technology, Shenzhen, China
• Key Laboratory of Poyang Lake Wetland and Watershed Research, Ministry of Education, Jiangxi Normal University, Nanchang, China
• Henan Provincial Key Lab of Hydrosphere and Watershed Water Security, North China University of Water Resources and Electric Power, Zhengzhou, Henan, China
• College of Surveying and Geo-informatics, North China University of Water Resources and Electric Power, Zhengzhou, Henan, China
• Information Center (Hydrology Monitor and Forecast Center), Ministry of Water Resources, Beijing, China

Short Summary

This study integrates GNSS and GRACE observations to jointly invert terrestrial water storage (TWS) changes in the Upper Yellow River Basin (UYRB) from 2011 to 2023, revealing distinct spatial heterogeneity in TWS dynamics and drought drivers, with the south primarily climate-controlled and the north dominated by human activities.

Objective

• To jointly invert terrestrial water storage (TWS) changes in the Upper Yellow River Basin (UYRB) by integrating Global Navigation Satellite System (GNSS) and Gravity Recovery and Climate Experiment (GRACE) observations from 2011 to 2023.
• To characterize the spatiotemporal evolution of hydrological droughts in the UYRB and identify their dominant drivers through correlation analysis and a machine-learning framework with SHAP interpretation, using multi-source meteorological and hydrological datasets.

Study Configuration

• Spatial Scale: Upper Yellow River Basin (UYRB), divided into Region I (southern UYRB, Tibetan Plateau) and Region II (northern UYRB, Loess Plateau and Hetao Plain). Inversion grid cells of 0.5° × 0.5°.
• Temporal Scale: January 2011 to January 2023 (12 years), with monthly data.

Methodology and Data

• Models used:
  • GNSS-GRACE joint inversion framework (mascon-constrained, Green’s function-based)
  • Principal Component Analysis (PCA) for GNSS noise reduction
  • Light Gradient Boosting Machine (LightGBM) for TWS prediction
  • SHapley Additive exPlanations (SHAP) for model interpretation
  • Drought Severity Index (DSI) calculation
  • Standardized Precipitation Index (SPI)
• Data sources:
  • GNSS vertical displacement time series from 39 stations (Crustal Movement Observation Network of China - CMONOC).
  • GRACE RL06.3_v04 Mascon product (JPL, 0.5° grid).
  • Global Land Data Assimilation System Noah Land Surface Model version 2.1 (GLDAS-Noah v2.1) for root-zone soil moisture, canopy water content, and snow water equivalent.
  • ERA5-Land dataset for precipitation (P), evapotranspiration (ET), and runoff (R).
  • MODIS Leaf Area Index (LAI) product (MCD15A3H).
  • Groundwater level dataset (Wang et al., 2025) derived from the Annual Report of Groundwater Level Monitoring in China (2005–2022), gridded at 1 km and resampled to 0.5°.

Main Results

• Terrestrial water storage (TWS) seasonal variability decreases from southwest to northeast across the UYRB.
• The southern UYRB (Region I) shows overall TWS increases, while the northern UYRB (Region II) exhibits gradual TWS declines.
• The joint GNSS-GRACE inversion framework effectively integrates complementary observations, providing spatially balanced and physically realistic TWS estimates that outperform single-source solutions.
• Region I experienced increasing TWS trends, primarily linked to elevated precipitation, with severe anomalies peaking at -78.15 mm during 2015–2018.
• Region II showed persistent TWS depletion, attributed to limited precipitation and extensive human water use (agricultural irrigation, groundwater extraction), with peak deficits of -51.27 mm (2015–2018) and -78.33 mm (2019–2023).
• Hydrological droughts in Region I are strongly coupled with natural climate variability, showing a maximum correlation (r=0.91) between Joint-DSI and SPI at a 37-month accumulation period.
• Hydrological droughts in Region II are largely decoupled from precipitation variability, primarily driven by groundwater extraction and evapotranspiration, with a weak maximum correlation (r=0.31) between Joint-DSI and SPI at a 19-month timescale.
• Nine hydrological drought events (2–12 months duration) were identified in northern sub-regions, and six events (2–24 months duration) in the south.
• A long-term drought propagation timescale (e.g., 37 months in the south) governs TWS-based drought evolution, reflecting the integrated response of the entire hydrological column, including deep groundwater recharge.

Contributions

• Developed and applied a refined mascon-constrained GNSS-GRACE joint inversion framework, significantly enhancing the spatial resolution and robustness of TWS estimates, particularly in areas with limited GNSS coverage.
• Provided a comprehensive, high-resolution characterization of spatiotemporal TWS variations and hydrological drought events in the UYRB, overcoming the limitations of coarse-resolution GRACE and sparse GNSS data.
• Quantitatively identified and attributed the dominant drivers of hydrological drought in distinct sub-regions of the UYRB using an interpretable machine-learning framework (LightGBM with SHAP analysis), highlighting the contrasting roles of climate variability and human activities.
• Revealed long-term drought propagation timescales for TWS-based drought evolution, distinguishing these integrated responses from rapid surface-water propagation.
• Demonstrated the potential of GNSS–GRACE integration for fine-scale drought monitoring and informing sustainable water-resource management strategies in climate-sensitive basins.

Funding

• National Key Research and Development Program of China (No. 2024YFC3212200)
• 111 Project (Grant No. D25014)
• National Foreign Experts Program (Category S) (Grant No. S20240116)
• Henan Provincial Key Laboratory of Hydrosphere and Watershed Water Security
• Henan Province Foreign Scientist Studio (Grant No. GZS2024013)
• Research Projects (254000510004)

Citation

@article{Wang2026Terrestrial,
  author = {Wang, Zi and Zhu, Hai and Chen, Kejie and Li, Mingjia and Hu, Shunqiang and Liu, Junguo and Hu, Qingfeng and Sun, Liang and Gu, Ge and Li, He},
  title = {Terrestrial water storage variations and drought characteristics in the upper yellow river basin revealed by joint GNSS-GRACE analysis},
  journal = {Journal of Hydrology Regional Studies},
  year = {2026},
  doi = {10.1016/j.ejrh.2026.103255},
  url = {https://doi.org/10.1016/j.ejrh.2026.103255}
}