Zhao et al. (2025) Simulating Rainfall for Flood Forecasting in the Upper Minjiang River

Identification

• Journal: Water
• Year: 2025
• Date: 2025-12-19
• Authors: Wenjie Zhao, Yang Zhao, Qijia Zhao, Xingping Wang, Tiantian Su, Yuan Guo
• DOI: 10.3390/w18010004

Research Groups

• Sichuan Province Zipingpu Development Co., Ltd., Chengdu 610091, China
• College of Geosciences and Engineering, North China University of Water Resources and Electric Power, Zhengzhou 450001, China
• College of Water Conservancy and Transportation, Zhengzhou University, Zhengzhou 450001, China
• Key Laboratory of Urban Stormwater System and Water Environment (Beijing University of Civil Engineering and Architecture), Ministry of Education, Beijing 100044, China

Short Summary

This study integrates the WRF numerical weather prediction model with the InfoWorks ICM hydrodynamic model to improve flood forecasting in the Upper Minjiang River, demonstrating that optimized WRF configurations and updated land surface data enhance rainfall and flood hydrograph simulations, though peak discharge underestimation persists.

Objective

• To develop and evaluate a combined flood forecasting chain for the Upper Minjiang River by assessing WRF model sensitivity to initial conditions and land use, then using optimized WRF outputs to drive InfoWorks ICM for flood hydrograph and inundation pattern simulation.

Study Configuration

• Spatial Scale: Upper Minjiang River Basin (22,576 km²); WRF model with two nested layers (d01: 9 km resolution, d02: 3 km resolution); initial field data at 1° × 1° and 0.25° × 0.25° resolutions; DEM at 30 m resolution; land use data at 30 m and 500 m resolutions.
• Temporal Scale: Five heavy rainfall events from 2018 to 2022 (durations from 24 hours to 84 hours); initial field data temporal intervals of 1 hour, 3 hours, and 6 hours; WRF simulation integration step of 30 seconds; forecast lead times evaluated from 2 hours to 20 hours.

Methodology and Data

• Models used:
  • Weather Research and Forecasting Model (WRF)
  • InfoWorks Integrated Catchment Modelling (ICM)
• Data sources:
  • Initial meteorological field data: FNL (National Center for Weather and Environmental Prediction - NCEP) and ERA5 (European Center for Medium-Range Weather Forecasts - ECMWF).
  • Measured precipitation data: From 9 rainfall stations, interpolated using Inverse Distance Weighting (IDW).
  • Digital Elevation Model (DEM): 30 m resolution SRTM DEM, updated with ASTER elevation data (30 m resolution, 2013).
  • Soil data: Second National Land Survey and Harmonized World Soil Database (HWSD).
  • Land use data: 30 m resolution Chinese Academy of Sciences (CNLUCC) datasets, updated with MODIS land use data (500 m resolution, 2020).
  • Vegetation data: 1:1,000,000 vegetation dataset of China.
  • Post-processing: Python scripts for assimilating WRF grid precipitation into InfoWorks ICM.

Main Results

• The WRF model effectively simulates the spatial distribution and peak timing of precipitation in the Upper Minjiang River, but systematically underestimates peak rainfall intensity and cumulative precipitation.
• Initial field data with 0.25° spatial resolution and 3-hour temporal intervals provide the best performance for WRF rainfall simulations.
• A forecast lead time of 10–14 hours demonstrates superior predictive capability in numerical simulations.
• Updating elevation and land use conditions (ASTER DEM, 2020 MODIS land use) increases cumulative rainfall estimates and improves approximation of measured values, although simulated peaks remain lower than observed.
• Segmented integration of rainfall forecasts significantly enhances the representation of subsequent rainfall peaks and expands rainfall coverage.
• InfoWorks ICM, driven by observed precipitation, generally reproduces hydrograph patterns and peak timing but underestimates peak magnitude.
• WRF-driven simulations (ERA5 and FNL) significantly underestimate peak discharge and often have peak timing errors exceeding 10 hours, but show improved temporal consistency after land surface data updates.
• Numerical precipitation forecasts improve the correlation of flood hydrographs and extend early warning lead times, despite persistent systematic underestimation in peak discharge quantification.

Contributions

• Develops and evaluates a novel combined flood forecasting chain for the complex Upper Minjiang River Basin by integrating the WRF numerical weather prediction model with the InfoWorks ICM hydrodynamic model.
• Systematically assesses the sensitivity of WRF-simulated heavy rainfall to various initial field data (FNL, ERA5) across different spatial (1°, 0.25°) and temporal (1 hour, 3 hours, 6 hours) resolutions.
• Identifies and quantifies the optimal initial field configuration for the study area: ERA5 data with 0.25° spatial resolution, 3-hour temporal intervals, and a 10–14 hour forecast lead time.
• Quantifies the impact of updated high-precision topographic (ASTER DEM) and land use (MODIS 2020) data on WRF rainfall simulations and subsequent flood hydrographs, showing improvements in cumulative rainfall estimates.
• Demonstrates the practical value of numerical precipitation forecasts in improving flood hydrograph correlation and extending early warning lead times, even while acknowledging persistent challenges in accurate peak discharge prediction.

Funding

• National Natural Science Foundation of China (No: 52579026)
• Natural Science Foundation of Henan Province (232300421208, 242300421038)
• Key Research Project Program of Higher Education Institutions in Henan Province (23A170003)
• Open Research Fund Program of Key Laboratory of Urban Stormwater System and Water Environment (Beijing University of Civil Engineering and Architecture), Ministry of Education (USSWE2023KF01)

Citation

@article{Zhao2025Simulating,
  author = {Zhao, Wenjie and Zhao, Yang and Zhao, Qijia and Wang, Xingping and Su, Tiantian and Guo, Yuan},
  title = {Simulating Rainfall for Flood Forecasting in the Upper Minjiang River},
  journal = {Water},
  year = {2025},
  doi = {10.3390/w18010004},
  url = {https://doi.org/10.3390/w18010004}
}