Zhang et al. (2026) Overestimation of past and future increases in global river flow by Earth system models

Identification

• Journal: Nature Geoscience
• Year: 2026
• Date: 2026-01-09
• Authors: Yongqiang Zhang, Günter Blöschl, Haoshan Wei, Dongdong Kong, Ning Ma, Thorsten Wagener, Jing Tian, Jun Xia, Congcong Li, Longhao Wang, Francis H. S. Chiew, L. Ruby Leung, X. Liu, Hongxing Zheng, Xuanze Zhang, Changming Liu
• DOI: 10.1038/s41561-025-01897-9

Research Groups

• Key Laboratory of Water Cycle and Related Land Surface Processes, Institute of Geographic Sciences and Natural Resources Research, Chinese Academy of Sciences, Beijing, China
• Institute of Hydraulic and Water Resources Engineering, Technische Universität Wien, Vienna, Austria
• University of Chinese Academy of Sciences, Beijing, China
• Department of Atmospheric Science, School of Environmental Studies, China University of Geosciences, Wuhan, China
• Institute of Environmental Science and Geography, University of Potsdam, Potsdam, Germany
• State Key Laboratory of Water Resources and Hydropower Engineering Sciences, Wuhan University, Wuhan, China
• CSIRO Environment, Canberra, Australian Capital Territory, Australia
• Pacific Northwest National Laboratory, Richland, WA, USA

Short Summary

This study refines global water partitioning estimates by combining Earth system model outputs with river flow observations using an emergent constraint approach. It reveals that Earth system models significantly overestimate past and future increases in global river flow, providing more accurate historical estimates and strengthening future projections with reduced uncertainty.

Objective

• To refine estimates of global water partitioning, including river flow and land evapotranspiration, and improve future projections by applying an emergent constraint approach that integrates outputs from multiple Earth system models with river flow observations from 50 large basins.

Study Configuration

• Spatial Scale: Global, with observational constraints derived from 50 large river basins.
• Temporal Scale: Historical period (1980–2014) and future climate change projections.

Methodology and Data

• Models used: Multiple Earth system models (ESMs), emergent constraint approach, WaterGAP model.
• Data sources: River flow observations from 50 large basins (e.g., Global Runoff Data Centre, Arctic Great Rivers Observatory, Mekong River Commission, India Water Resources Information System), outputs from multiple Earth system models (CMIP6), evapotranspiration datasets, MSWEP V2, GPCP, ERA5-Land reanalysis.

Main Results

• Between 1980 and 2014, global river flow was estimated at (39.1 ± 5.4) × 10^3 km^3 yr^−1, with a river flow-to-precipitation ratio of 0.35 ± 0.03, both lower than previous estimates.
• Land evapotranspiration during the same period reached (73.4 ± 6.2) × 10^3 km^3 yr^−1.
• Under climate change, global river flow is projected to rise by 7.8 ± 5.5 mm per year per degree of warming.
• This emergent constraint-refined projection is 9.3% lower than the ensemble mean of Earth system models and reduces inter-model uncertainty by 66%.

Contributions

• Provides more accurate historical estimates of global river flow and land evapotranspiration by integrating observational data.
• Strengthens future projections of global water-cycle components by reducing inter-model uncertainty.
• Highlights a significant overestimation of past and future increases in global river flow by current Earth system models.
• Demonstrates the effectiveness of the emergent constraint approach in refining global water-cycle quantification.

Funding

• National Natural Science Foundation of China (grants 42330506, 42361144709, 42430610)
• Talent Program of the Ministry of Science and Technology of China
• Alexander von Humboldt Foundation (Alexander von Humboldt Professorship endowed by the German Federal Ministry of Education and Research (BMBF))
• PIFI outstanding international team project by the Chinese Academy of Sciences

Citation

@article{Zhang2026Overestimation,
  author = {Zhang, Yongqiang and Blöschl, Günter and Wei, Haoshan and Kong, Dongdong and Ma, Ning and Wagener, Thorsten and Tian, Jing and Xia, Jun and Li, Congcong and Wang, Longhao and Chiew, Francis H. S. and Leung, L. Ruby and Liu, X. and Zheng, Hongxing and Zhang, Xuanze and Liu, Changming},
  title = {Overestimation of past and future increases in global river flow by Earth system models},
  journal = {Nature Geoscience},
  year = {2026},
  doi = {10.1038/s41561-025-01897-9},
  url = {https://doi.org/10.1038/s41561-025-01897-9}
}