Wolkeba et al. (2026) Dynamic integration of GRACE-based water storage into global hydrological models using machine learning

⚠️ Warning: This summary was generated from the abstract only, as the full text was not available.

Identification

• Journal: Environmental Research Water
• Year: 2026
• Date: 2026-09-11
• Authors: Fitsume T. Wolkeba, Mesfin Mergia Mekonnen, Hamid Moradkhani, Mukesh Kumar
• DOI: 10.1088/3033-4942/aea63c

Research Groups

• Department of Hydrology and Water Resources, University of Arizona
• NASA's Jet Propulsion Laboratory (JPL)
• US Geological Survey (USGS)

Short Summary

This study explores the enhancement of grid-level water availability estimates by dynamically coupling machine learning models with global hydrological models using GRACE-based Total Water Storage data. The results show significant improvements in daily and monthly streamflow estimates.

Objective

• Investigate the impact of integrating machine learning models with global hydrological models on groundwater storage dynamics and subsequent effects on water availability estimates.

Study Configuration

• Spatial Scale: Continental United States (CONUS)
• Temporal Scale: Daily to monthly time steps

Methodology and Data

• Models used: Long Short-Term Memory (LSTM) models, CWatM global hydrological model
• Data sources: GRACE-based Total Water Storage data, USGS observation data, reanalysis datasets

Main Results

• Average 16% improvement in daily streamflow estimates
• KGE scores increased at 61% of USGS stations (64 stations)
• 10% more stations reported KGE values above 0 for monthly streamflow estimates
• 58% of stations (61 stations) showed improvements, with an average monthly KGE increase of 13%

Contributions

• Original value lies in the development and application of machine learning-based data integration approaches to improve groundwater storage dynamics representation in global hydrological models.

Funding

• NASA's Earth Science Division (Grant Number: NNX17AQ49G)
• National Science Foundation (Grant Number: EAR-1844467)

Citation

@article{Wolkeba2026Dynamic,
  author = {Wolkeba, Fitsume T. and Mekonnen, Mesfin Mergia and Moradkhani, Hamid and Kumar, Mukesh},
  title = {Dynamic integration of GRACE-based water storage into global hydrological models using machine learning},
  journal = {Environmental Research Water},
  year = {2026},
  doi = {10.1088/3033-4942/aea63c},
  url = {https://doi.org/10.1088/3033-4942/aea63c}
}