Qi et al. (2026) Waterlogging Simulation Model Based on Bidirectional Coupling Between Runoff Production and Confluence in Plain Areas

Identification

• Journal: Water Resources Management
• Year: 2026
• Date: 2026-02-23
• Authors: Zhen Qi, Huaqing Zhao, Ranhang Zhao, Xingju Wang, Yuehua Xu
• DOI: 10.1007/s11269-026-04512-7

Research Groups

• School of Civil Engineering, Shandong University, Jinan, Shandong, China

Short Summary

This paper develops a raster-based waterlogging simulation model for plain areas, introducing a novel bidirectional coupling mechanism between runoff production and confluence to overcome the limitations of traditional unidirectional models. The model significantly improves the accuracy of plain waterlogging simulation, with peak flow errors within ±15% and Nash-Sutcliffe Efficiency (NSE) values over 0.84, providing enhanced tools for flood warning and water resource management.

Objective

• To establish a targeted plain waterlogging simulation model with a bidirectional coupling mechanism between runoff production and confluence.
• To improve the simulation accuracy of inundation characteristics (extent, water depth, duration) in plain areas.
• To provide reliable technical support for flood warning and water resource management in floodplains.

Study Configuration

• Spatial Scale: Tuhaihe River Catchment, upper reaches, Shandong Province, China, covering 7,685.61 square kilometers. The model uses a raster resolution of 500 meters, treating each raster as a hydrological response unit (HRU).
• Temporal Scale: Hydrographic data (hourly rainfall and flow) from 1990 to 2024 were used for calibration and validation. Simulation examples for a 50-year design rainstorm show flood events lasting tens of days, with results depicted at 60, 120, 180, 240, 300, and 360 hours.

Methodology and Data

• Models used:
  • Overall Model: Raster-based waterlogging simulation model with bidirectional coupling of runoff production and confluence.
  • Runoff Production: Adaptive Soil Conservation Service Curve Number (SCS-CN) method, modified to account for upstream water recharge and dynamic infiltration capacity.
  • Slope Confluence: Combined water balance equation and Manning’s equation, incorporating bidirectional coupling.
  • River Confluence: Muskingum method.
  • DEM Processing: MATLAB’s “imfill” function for depression storage calculation; empirical coefficients (ep1, ep2) introduced for correcting DEM gradient errors and depression filling deviations.
  • Flow Direction: Steepest slope method.
• Data sources:
  • Underlying Surface Data: Catchment boundaries, river locations, Digital Elevation Model (DEM), land use types, soil types, and river sections.
  • Hydrographic Data: Hourly rainfall data from ten rainfall stations and hourly flow data from two hydrographic stations (1990-2024).
  • Calibration/Validation Data: Five historical flood events (July 1990, August 1990, September 1991, July 2012, July 2013).

Main Results

• The proposed model achieved high simulation accuracy, with peak flow errors within ±15% and Nash-Sutcliffe Efficiency (NSE) values consistently over 0.84 for both calibration and validation events.
• Simulated and observed flow processes showed strong consistency in trends, peak flows, peak times, and rates of flow change, demonstrating good model stability.
• Waterlogging simulations for a 50-year design rainstorm (207 mm rainfall over 168 hours) showed inundation peaking around the 120th hour and water ceasing to flow, accumulating in depressions, by approximately the 360th hour.
• Inundation depths were consistent with depression storage (demf0), with larger depths in areas of greater demf0, and water concentrated in these areas after the flood receded.
• The calibrated empirical coefficients ep1 and ep2 were 0.8 and 0.5, respectively, indicating the objective existence and significant impact of DEM gradient and depression filling errors on simulation results.

Contributions

• Novel Bidirectional Coupling: Breaks from traditional unidirectional coupling in distributed hydrological models by dynamically integrating the mutual influence of runoff production and confluence processes, particularly crucial for flat plain areas.
• Adaptive Runoff Production Model: Enhances the SCS-CN method by adaptively adjusting infiltration capacity based on total HRU water volume (rainfall and upstream inflow), reflecting the dynamic interaction in the hydrological cycle.
• Refined Underlying Surface Treatment: Incorporates detailed DEM processing, including artificial raising of embankment elevations and lowering of river/outlet elevations, along with depression filling calculations, to create a more realistic flood flow environment.
• Error Correction with Empirical Coefficients: Introduces two empirical parameters (ep1 for gradient deviation, ep2 for depression storage deviation) to effectively correct inherent errors in DEM data, significantly improving model accuracy and terrain adaptability.
• Confluence Hierarchy Concept: Proposes a novel method for determining HRU flow convergence relationships, enhancing simulation accuracy and computational efficiency by fixing and calculating confluence layers once.
• Runoff Segmentation: Divides runoff into static (depression filling) and dynamic components, accurately describing flood flow states under different terrains and improving confluence calculation accuracy.

Funding

• Water Resources Department of Shandong Province, China [No. SDSLKY201902]

Citation

@article{Qi2026Waterlogging,
  author = {Qi, Zhen and Zhao, Huaqing and Zhao, Ranhang and Wang, Xingju and Xu, Yuehua},
  title = {Waterlogging Simulation Model Based on Bidirectional Coupling Between Runoff Production and Confluence in Plain Areas},
  journal = {Water Resources Management},
  year = {2026},
  doi = {10.1007/s11269-026-04512-7},
  url = {https://doi.org/10.1007/s11269-026-04512-7}
}