Liu et al. (2026) Global estimation of bankfull river discharge reveals distinct flood recurrences across climate zones

Identification

• Journal: Nature Communications
• Year: 2026
• Date: 2026-08-18
• Authors: Yinxue Liu, Michel Wortmann, Laurence Paul Hawker, Jeffrey Charles Neal, Jiabo Yin, Marcus Suassuna Santos, Bailey J. Anderson, Richard James Boothroyd, Andrew Nicholas, Gregory H. Sambrook Smith, Philip J. Ashworth, Hannah Cloke, Solomon H. Gebrechorkos, Julian Leyland, Boen Zhang, Ellie Vahidi, Helen Griffith, Pauline Delorme, Stuart McLelland, Daniel R. Parsons, Stephen E. Darby, Louise J. Slater
• DOI: 10.1038/s41467-026-76433-3

Research Groups

• Department of Geography, University of Colorado Boulder
• Department of Civil and Environmental Engineering, University of California Berkeley
• NASA Goddard Space Flight Center
• National Centre for Earth Observation (NCEO)
• University College London (UCL)

Short Summary

This study estimates global bankfull river discharge at ~1 km resolution using machine learning approaches and a recently developed hydrography. The results reveal distinct flood recurrences across climate zones, with median return periods ranging from 1.5 years in tropical regions to 4.3 years in arid regions.

Objective

• Estimate the spatial distribution of bankfull river discharge globally
• Investigate the relationship between bankfull discharge and climate region
• Assess the bias of the commonly used 2-year return period assumption for bankfull discharge

Study Configuration

• Spatial Scale: Global, with a focus on ~1 km resolution along the GRIT hydrography
• Temporal Scale: Daily to annual time scales for flood frequency analysis

Methodology and Data

• Models used: Random Forest (RF) model for estimating bankfull discharge
• Data sources:
  • Observed bankfull discharge dataset from nine sources
  • Global River Topology (GRIT) hydrography
  • Daily discharge records at 8519 gauging stations

Main Results

• Estimated bankfull discharge return periods vary systematically across climate regions, with median values ranging from 1.5 years in tropical regions to 4.3 years in arid regions.
• The commonly used 2-year return period assumption for bankfull discharge is biased, overestimating bankfull discharge in tropical and temperate regions and underestimating it in arid regions.

Contributions

• This study provides a globally consistent benchmark for advancing riverine and water resources research across hydrology, geomorphology, and ecology.
• The results highlight the importance of incorporating climatic variation when estimating channel capacity at regional to continental scales.

Funding

• National Science Foundation (NSF)
• NASA's Earth Science Division
• European Research Council (ERC)

Citation

@article{Liu2026Global,
  author = {Liu, Yinxue and Wortmann, Michel and Hawker, Laurence Paul and Neal, Jeffrey Charles and Yin, Jiabo and Santos, Marcus Suassuna and Anderson, Bailey J. and Boothroyd, Richard  James and Nicholas, Andrew and Smith, Gregory H. Sambrook and Ashworth, Philip J. and Cloke, Hannah and Gebrechorkos, Solomon H. and Leyland, Julian and Zhang, Boen and Vahidi, Ellie and Griffith, Helen and Delorme, Pauline and McLelland, Stuart and Parsons, Daniel R. and Darby, Stephen E. and Slater, Louise J.},
  title = {Global estimation of bankfull river discharge reveals distinct flood recurrences across climate zones},
  journal = {Nature Communications},
  year = {2026},
  doi = {10.1038/s41467-026-76433-3},
  url = {https://doi.org/10.1038/s41467-026-76433-3}
}