Wang et al. (2025) Method for selecting typical floods based on an unfavorable indicator and flood classification

Identification

• Journal: Environmental Fluid Mechanics
• Year: 2025
• Date: 2025-12-01
• Authors: Jiangyu Wang, Zhaohui Jia, Wei Zhang, Xin Li
• DOI: 10.1007/s10652-025-10062-0

Research Groups

• State Key Laboratory of Water Resources Engineering and Management, Wuhan University, Wuhan, China
• China Three Gorges Construction Engineering Corporation, Chengdu, China

Short Summary

This study proposes a multi-indicator method integrating an entropy-weighted unfavorable indicator with a two-dimensional return-period classification to scientifically identify and select typical unfavorable flood events. Applied to the Tongguan Station, the method enhances the representativeness and accuracy of flood event selection for disaster response planning.

Objective

• To enhance the scientific accuracy and representativeness of typical flood event selection by proposing a multi-indicator method that integrates an entropy-weighted unfavorable indicator with a two-dimensional return-period classification.

Study Configuration

• Spatial Scale: Tongguan Hydrological Station (34° 48′ N, 110° 00′ E) in the middle reaches of the Yellow River Basin, China, controlling a catchment area of approximately 682,000 km².
• Temporal Scale: Flood season (June to October) from 2002 to 2023, analyzing 180 independent flood events.

Methodology and Data

• Models used: Entropy-weighting method, two-dimensional return-period classification (4x4 matrix), K-means cluster analysis (for comparison), Pearson Type III distribution, Generalized Extreme Value (GEV) distribution, min-max normalization, Pearson correlation coefficient, Kruskal–Wallis test, bootstrap resampling.
• Data sources: Daily average discharge data observed at the Tongguan Hydrological Station.

Main Results

• Six key flood characteristic indicators were selected: peak discharge, flood volume, rising rate, falling rate, skewness coefficient, and proportion of high-pulse duration.
• The entropy-weighting method assigned the highest weight to the proportion of high-pulse duration (0.238) and the lowest to the skewness coefficient (0.022).
• The proposed two-dimensional return-period classification categorized 133 small, 38 medium, 6 large, and 3 extreme floods, consistent with actual flood occurrence patterns.
• Both the proposed return-period classification method and K-means cluster analysis passed significance tests (p < 0.05).
• The proposed method demonstrated a 67.78% consistency with K-means cluster analysis and over 94% consistency with traditional single-indicator classification methods.
• The method effectively identifies flood events with larger unfavorable indicators within the same flood type, particularly when peak discharge and total flood volume show significant discrepancies (e.g., "Sharp–Thin" vs. "Long-Duration" floods).
• Calculated peak discharges for 5, 10, 20, and 50-year return periods were 2637.81 m³/s, 3419.16 m³/s, 4185.76 m³/s, and 5185.18 m³/s, respectively.
• Calculated total flood volumes for 5, 10, 20, and 50-year return periods were 1.76 × 10⁹ m³, 2.55 × 10⁹ m³, 3.57 × 10⁹ m³, and 5.39 × 10⁹ m³, respectively.

Contributions

• Introduces a novel multi-indicator method for typical unfavorable flood event selection, integrating an entropy-weighted unfavorable indicator with a two-dimensional return-period classification.
• Enhances the scientific rigor, representativeness, and risk coverage of flood event selection by comprehensively considering flood magnitude, dynamic processes, and morphology.
• Overcomes the limitations of traditional single-indicator methods, which often underestimate risks or exhibit a preference for extreme values, and reduces dependence on input flood characteristics compared to K-means clustering.
• Provides more challenging and representative hydrological scenarios crucial for flood feature analysis, flood-management simulations, risk assessment, and the development of soil- and water-conservation strategies.

Funding

• National Key Research and Development Program of China (2023YFC3206204)
• National Natural Science Foundation of China (No. U2040218)

Citation

@article{Wang2025Method,
  author = {Wang, Jiangyu and Jia, Zhaohui and Zhang, Wei and Li, Xin},
  title = {Method for selecting typical floods based on an unfavorable indicator and flood classification},
  journal = {Environmental Fluid Mechanics},
  year = {2025},
  doi = {10.1007/s10652-025-10062-0},
  url = {https://doi.org/10.1007/s10652-025-10062-0}
}