Dou et al. (2025) Comparative Analysis of Rainfall-Based and Discharge-Based Early Warning Methods for Flash Floods

Identification

• Journal: Water
• Year: 2025
• Date: 2025-12-25
• Authors: Yanhong Dou, Junyao Wen, Xiangning Liu, Ronghua Liu, Jichao Sun
• DOI: 10.3390/w18010064

Research Groups

• State Key Laboratory of Water Cycle and Water Security, China Institute of Water Resources and Hydro-Power Research, Beijing, China.
• School of Water Resources and Environment, China University of Geosciences Beijing, Beijing, China.

Short Summary

This study evaluates the comparative performance of rainfall-based (RW) and discharge-based (DW) early warning methods for flash floods using historical case studies and hydrological simulations. The findings identify specific environmental and infrastructural scenarios where each method excels, advocating for an integrated, dynamically weighted warning framework.

Objective

• To identify the relative strengths and limitations of rainfall-based versus discharge-based warning methods under varying conditions of rainfall intensity, spatial distribution, antecedent soil moisture, and watershed characteristics.

Study Configuration

• Spatial Scale: Six small-to-medium mountainous catchments in Fujian Province, China (drainage areas ranging from 18.54 $km^2$ to 468.52 $km^2$) and the Xiao’anxia watershed.
• Temporal Scale: Analysis of historical flash flood events occurring between June and August 2024, supplemented by simulation scenarios with 1 h, 3 h, and 6 h durations.

Methodology and Data

• Models used: Distributed Xin’anjiang rainfall-runoff model (integrated into the Fujian Provincial Flash Flood Early Warning Platform).
• Data sources: National Flash Flood Survey and Assessment database (warning thresholds and watershed characteristics); Water Yearbook of Fujian Province (rainfall and runoff data); QGIS 3.34.6 for spatial processing and watershed delineation.

Main Results

• DW Superiority: Discharge-based warnings outperform RW during persistent moderate-intensity rainfall when antecedent soil moisture is moderate to high (triggering alerts up to 2 h 18 min earlier). DW also avoids false alarms caused by spatial rainfall heterogeneity (e.g., heavy rain occurring just outside the catchment boundary).
• RW Superiority: Rainfall-based warnings are more effective for localized, short-duration heavy rainfall events when antecedent soil moisture is low. RW is more reliable in watersheds containing numerous small-to-medium reservoirs lacking operational data (where DW produces false alarms) and in regions with sparse or unevenly distributed rain gauges.
• Quantitative Sensitivity: Simulations show that DW trigger timing is highly sensitive to initial soil moisture but relatively insensitive to the temporal distribution of rainfall (decreasing, uniform, or increasing patterns).

Contributions

• Establishes a clear classification of "advantageous scenarios" for both RW and DW methods, filling a gap in existing literature regarding the operational selection of warning criteria.
• Quantifies the impact of unmonitored hydraulic structures (reservoirs) and rain gauge density on the reliability of hydrological model-based warnings.
• Proposes a dynamically weighted integrated warning framework that adapts to real-time hydro-meteorological conditions to reduce false and missed alarms.

Funding

• Ningbo Water Resources Science and Technology Program Project (Grant No. NSKA202507).
• National Key R&D Program of China (Grant No. 2023YFC3006700).
• State Key Laboratory of Water Cycle and Water Security (Project No. SKL2025TDGG05).

Citation

@article{Dou2025Comparative,
  author = {Dou, Yanhong and Wen, Junyao and Liu, Xiangning and Liu, Ronghua and Sun, Jichao},
  title = {Comparative Analysis of Rainfall-Based and Discharge-Based Early Warning Methods for Flash Floods},
  journal = {Water},
  year = {2025},
  doi = {10.3390/w18010064},
  url = {https://doi.org/10.3390/w18010064}
}

Generated by BiblioAssistant using gemini-3-flash-preview (Google API)