Bäthge et al. (2026) A Global-Scale Time Series Dataset for Groundwater Studies within the Earth System

Identification

• Journal: Scientific Data
• Year: 2026
• Date: 2026-03-09
• Authors: Annemarie Bäthge, Claudia Ruz Vargas, Gunnar Lischeid, Raoul Collenteur, Mark O Cuthbert, Jan H. Fleckenstein, Martina Flörke, Inge de Graaf, Sebastian Gnann, Andreas Hartmann, Xander Huggins, Nils Moosdorf, Yoshihide Wada, Thorsten Wagener, Robert Reinecke
• DOI: 10.1038/s41597-026-06966-1

Research Groups

• Institute of Geography, Johannes Gutenberg-University, Mainz, Germany
• International Groundwater Resources Assessment Centre (IGRAC), Delft, The Netherlands
• Research Area 4: ‘Simulation & Data Science’, Leibniz Centre for Agricultural Landscape Research, Müncheberg, Germany
• Institute of Environmental Science and Geography, University of Potsdam, Germany
• Department Water Resources and Drinking Water, Eawag, Dübendorf, Zürich, Switzerland
• School of Earth and Environmental Sciences, Cardiff University, UK
• Department of Hydrogeology, Helmholtz Center for Environmental Research, Leipzig, Germany
• Institute of Engineering Hydrology and Water Resources Management, Ruhr University Bochum, Bochum, Germany
• Earth Systems and Global Change group, Wageningen University and Research, Wageningen, the Netherlands
• Chair of Hydrology, University of Freiburg, Freiburg, Germany
• Institute of Groundwater Management, Dresden University of Technology
• Institute for Resources, Environment, and Sustainability, University of British Columbia, Vancouver, BC, Canada
• High Meadows Environmental Institute, Princeton University, Princeton, NJ, United States of America
• Stockholm Resilience Centre, Stockholm University, Stockholm, Sweden
• Leibniz Centre for Tropical Marine Research (ZMT), Bremen, Germany
• Kiel University, Kiel, Germany
• Biological and Environmental Science and Engineering Division, King Abdullah University of Science and Technology, Thuwal, Saudi Arabia

Short Summary

This paper introduces GROW, a global-scale, quality-controlled dataset integrating over 200,000 groundwater depth and level time series with 36 associated Earth system variables to facilitate understanding of groundwater dynamics and model evaluation. It provides an analysis-ready foundation for studying large-scale groundwater processes in space and time within the Earth system.

Objective

• To develop and present GROW, a global-scale, analysis-ready, quality-controlled dataset that integrates groundwater depth and level time series with a comprehensive set of Earth system variables to enable a better understanding of groundwater dynamics and their interactions within the Earth system, and to support the calibration and evaluation of groundwater models.

Study Configuration

• Spatial Scale: Global-scale, covering 55 countries. Earth system variables are provided at spatial resolutions ranging from 3 arc seconds to 0.5°.
• Temporal Scale: Time series with daily, monthly, or yearly resolutions. Data spans from 1891 to near present, with individual time series lengths up to 135 years.

Methodology and Data

• Models used:
  • DBSCAN algorithm (for outlier and change point detection)
  • Mann-Kendall trend test (original and Hamed and Rao Modified for autocorrelated data)
  • Sen’s slope (for trend magnitude)
  • Penman’s equation (for GLEAM potential evapotranspiration)
• Data sources:
  • Groundwater time series: Global Groundwater Monitoring Network (GGMN) and Groundwater Observations Repository from the International Groundwater Resources Assessment Centre (IGRAC).
  • Earth system variables: Derived from various observation-based, reanalysis, and modelled global data products, including:
    • Atmosphere: MSWEP V2, GPCC, ERA5-Land, GLEAM4, CHELSA v2.1, Meybeck et al. 2013.
    • Geosphere: MERIT DEM v1.0.3, Geomorpho90m, GLiM v1, WHYMAP WOKAM, GLHYMPS2.0, HiHydroSoil v2.0.
    • Hydrosphere: Cuthbert, Gleeson et al., HydroRivers, BasinATLAS Level 9.
    • Cryosphere: BasinATLAS Level 9, ERA5-Land.
    • Biosphere: AVHRR NDVI, VIIRS NDVI, ERA5-Land, Huggins et al.
    • Anthroposphere: Wada et al., Volkholz & Ostberg 2024, Huggins et al.
  • Data processing: Included temporal harmonization (daily, monthly, yearly means), gap capping (maximum 10% total gap fraction), linear gap filling, flagging of negative depth values, autocorrelation, outliers/change points (using DBSCAN), plateaus, trend analysis (Mann-Kendall, Sen's slope), removal of duplicates, and checks for erroneous coordinates.

Main Results

• The GROW dataset comprises 204,292 quality-controlled groundwater time series.
• Of these, 85% have a yearly resolution, 9% a monthly resolution, and 6% a daily resolution.
• 51% of the time series are at least 10 years long, with a maximum length of 135 years.
• The dataset integrates 36 groundwater-associated Earth system variables (time series and static attributes), categorized into atmosphere (6), geosphere (10), hydrosphere (2), cryosphere (4), biosphere (5), and anthroposphere (9).
• 34 data flags are provided for well features and time series characteristics to enable targeted data filtering.
• The dataset exhibits a spatial bias, with 91% of wells located in North America (51%), India (17%), Europe (13%), and Australia (10%).
• There is also a bias towards arid (26%) and temperate (58%) climates, low elevations (62% of wells below 200 m altitude), and anthropogenically used land (67%).
• The median depth to groundwater in the dataset is 8 m, indicating a representation of shallow groundwater.
• Trend analysis revealed that 22% of time series show a decreasing trend, 12% an increasing trend, and 66% no significant trend.

Contributions

• Presents the first global-scale, analysis-ready, and quality-controlled dataset that comprehensively integrates groundwater depth and level time series with a wide array of 36 associated Earth system variables.
• Addresses a critical gap in the literature by providing a standardized, harmonized, and freely available dataset (adhering to FAIR principles) with extensive metadata and data flags, significantly reducing user preprocessing time.
• Enables large-sample spatiotemporal groundwater analysis, facilitating the investigation of cumulative effects of multiple controls on groundwater dynamics across diverse environmental settings.
• Offers a robust foundation for calibrating and evaluating global hydrological, land-surface, and climate models, enhancing the integrated process understanding of groundwater within the Earth system.
• Facilitates the transfer of conceptualized groundwater processes from data-rich to data-scarce regions with similar environmental conditions.

Funding

• European Research Council (ERC) Starting Grant (GROW: No. 10104110)
• Alexander von Humboldt Foundation (Alexander von Humboldt Professorship endowed by the German Federal Ministry of Education and Research (BMBF))
• Copernicus Climate Change Service (C3S)
• ISIMIP3a project

Citation

@article{Bäthge2026GlobalScale,
  author = {Bäthge, Annemarie and Vargas, Claudia Ruz and Lischeid, Gunnar and Collenteur, Raoul and Cuthbert, Mark O and Fleckenstein, Jan H. and Flörke, Martina and Graaf, Inge de and Gnann, Sebastian and Hartmann, Andreas and Huggins, Xander and Moosdorf, Nils and Wada, Yoshihide and Wagener, Thorsten and Reinecke, Robert},
  title = {A Global-Scale Time Series Dataset for Groundwater Studies within the Earth System},
  journal = {Scientific Data},
  year = {2026},
  doi = {10.1038/s41597-026-06966-1},
  url = {https://doi.org/10.1038/s41597-026-06966-1}
}