xu et al. (2026) Enhanced early warning of extreme river discharge events in the Yangtze River Basin using atmospheric circulation signals

Identification

• Journal: International Journal of Applied Earth Observation and Geoinformation
• Year: 2026
• Date: 2026-09-22
• Authors: xiaoke xu, Yong Wang, Anning Huang, Yinghong Jing, Chunlei Gu, Xiaojun She, Lifu Zhang, Yao Li
• DOI: 10.1016/j.jag.2026.105590

Research Groups

• School of Computer, Chengdu University of Information Technology
• School of Geographical Sciences, Southwest University
• School of Atmospheric Sciences, Nanjing University
• Plateau Atmosphere and Environment Key Laboratory of Sichuan Province
• State Key Laboratory of Remote Sensing Science and Digital Earth, Aerospace Information Research Institute, Chinese Academy of Sciences

Short Summary

This study proposes a novel strategy to enhance early warning of extreme river discharge events (ERDEs) in the Yangtze River Basin by detecting associated atmospheric circulation signals. A machine learning model called DetRF is developed to integrate anomaly detection with random forest classification.

Objective

• Develop a physically guided machine learning framework that identifies large-scale circulation anomalies linked to ERDE-favorable conditions.
• Enhance early warning of ERDEs in the Yangtze River Basin by detecting atmospheric circulation precursor signals associated with ERDEs.

Study Configuration

• Spatial Scale: The study focuses on the Yangtze River Basin, which spans approximately 1.8 million km2 across southern China.
• Temporal Scale: The study period covers March to November from 2000 to 2024.

Methodology and Data

• Models used: DetRF (a machine learning model that integrates anomaly detection with random forest classification)
• Data sources:
  • ERA5 reanalysis data
  • Precipitation data from the half-hourly final run of GPM-IMERG V06
  • High-resolution daily river discharge data for China from the Global Flood Awareness System (GloFAS) reanalysis

Main Results

• The DetRF model demonstrated robust and consistent performance in detecting atmospheric circulation signals associated with ERDEs across all YRB subregions.
• The model exhibited strong detection capability, with recall values exceeding 0.815 across all regions.
• The detected signals accurately captured key synoptic features associated with hydrological extremes, including the positions and intensities of troughs, vortices, and cyclones.

Contributions

• This study provides a novel strategy to enhance early warning of ERDEs by detecting atmospheric circulation precursor signals associated with ERDEs.
• The DetRF model offers a physically guided machine learning framework that identifies large-scale circulation anomalies linked to ERDE-favorable conditions.

Funding

• This research was funded by the National Natural Science Foundation of China (Grant No. 91637206) and the Key Research Program of the Chinese Academy of Sciences (Grant No. XDA19040201).

Citation

@article{xu2026Enhanced,
  author = {xu, xiaoke and Wang, Yong and Huang, Anning and Jing, Yinghong and Gu, Chunlei and She, Xiaojun and Zhang, Lifu and Li, Yao},
  title = {Enhanced early warning of extreme river discharge events in the Yangtze River Basin using atmospheric circulation signals},
  journal = {International Journal of Applied Earth Observation and Geoinformation},
  year = {2026},
  doi = {10.1016/j.jag.2026.105590},
  url = {https://doi.org/10.1016/j.jag.2026.105590}
}