Li et al. (2026) Quantitative Estimation of the Effects of Parameter Classification and Optimization Methods on Flood Simulations

Identification

• Journal: Water Resources Management
• Year: 2026
• Date: 2026-01-01
• Authors: Ji Li, Jiao Liu, Zhiqiang Xia, Chenrun Liu, Yuechen Li
• DOI: 10.1007/s11269-025-04415-z

Research Groups

• Chongqing Jinfo Mountain Karst Ecosystem National Observation and Research Station, Chongqing Key Laboratory of Karst Environment, Southwest University, Chongqing, China
• Chongqing Engineering Research Center for Remote Sensing Big Data Application, School of Geographical Sciences, Southwest University, Chongqing, China
• Chongqing Municipal Water Resources Bureau, Chongqing, China

Short Summary

This study quantitatively evaluates the effects of parameter classification and an improved chaotic particle swarm optimization (CPSO) algorithm on flood simulation performance using the Liuxihe Model, demonstrating that optimized parameters significantly enhance model accuracy and calibration efficiency for flood prediction.

Objective

• To systematically evaluate the impact of parameter types, physical properties, and optimization methods on the flood forecasting performance of hydrological models.
• To verify the effectiveness of parameter calibration for the distributed hydrological model by improving the chaotic particle swarm optimization algorithm and comparing its flood simulation effect with initial parameters.
• To evaluate the effects of different model parameters (classified into six categories) and key hydrogeological parameters on model uncertainty.

Study Configuration

• Spatial Scale: Tiantoushui watershed, a first-level tributary of the Wujiang River in Lechang city, Guangdong Province, China. The drainage area is 523 km². Spatial resolution of physical data (DEM, land use, soil type) is 1000 m × 1000 m.
• Temporal Scale: 22 sets of measured historical rainfall and flood process data collected from 2000–2022.

Methodology and Data

• Models used: Liuxihe Model (a new-generation distributed physical hydrological model for watershed flood prediction), improved chaotic particle swarm optimization (CPSO) algorithm for parameter optimization.
• Data sources:
  • Watershed digital topographic elevation model (DEM)
  • Land use type data
  • Soil type data
  • Measured historical rainfall data from 34 high-density rainfall stations.
  • Measured historical flood process data (22 sets).
  • Rainfall data spatially interpolated using the inverse distance weighted interpolation method.

Main Results

• The improved chaotic particle swarm optimization (CPSO) algorithm enhanced model calibration efficiency by up to 2 times, achieving convergence in approximately 19 iterations compared to 38 iterations for initial parameters.
• Flood simulations with optimized parameters were significantly better than those with initial parameters; the average Nash certainty coefficient increased by 26%, and the flood peak error decreased by 70%.
• Model parameters were classified into six categories (hydrological, geological, soil, vegetation, landform, meteorological) and their sensitivity ranked as: hydrological > geological > soil > vegetation > landform > meteorological. The infiltration coefficient was identified as the most sensitive parameter.
• The Liuxihe model demonstrated the best simulation effects for large floods, followed by moderate floods, and then minor floods. Simulation errors for medium and small floods were less than 20%, meeting local hydrological forecast error requirements.

Contributions

• Provides a systematic and quantitative evaluation of the impact of parameter classification and optimization methods on the flood simulation performance of distributed hydrological models.
• Introduces and validates an improved chaotic particle swarm optimization (CPSO) algorithm, demonstrating its superior efficiency and effectiveness in hydrological model calibration.
• Offers a detailed parameter sensitivity analysis for the Liuxihe Model, classifying parameters into physically meaningful categories and ranking their influence on flood prediction.
• Delivers crucial technical support and theoretical guidance for parameter optimization, sensitivity/uncertainty analysis, and practical application of distributed hydrological models in regional and watershed flood forecasting and disaster relief.

Funding

• State Key Program of the National Natural Science Foundation of China (U2244216)
• Special Fund for the Youth Team of Southwest University (SWU-XDJH202306)
• Fundamental Research Funds for the Central Universities (SWU-KQ22001)

Citation

@article{Li2026Quantitative,
  author = {Li, Ji and Liu, Jiao and Xia, Zhiqiang and Liu, Chenrun and Li, Yuechen},
  title = {Quantitative Estimation of the Effects of Parameter Classification and Optimization Methods on Flood Simulations},
  journal = {Water Resources Management},
  year = {2026},
  doi = {10.1007/s11269-025-04415-z},
  url = {https://doi.org/10.1007/s11269-025-04415-z}
}