Li et al. (2025) Joint calibration of Manning’s roughness and seepage in canals using NSGA-II for precision hydrodynamic modeling

Identification

• Journal: Frontiers in Water
• Year: 2025
• Date: 2025-12-12
• Authors: Li Li, Dan Bai, Yibo Li, Meng Li, Jianhu Lei, Fangyong Zhen
• DOI: 10.3389/frwa.2025.1709125

Research Groups

• State Key Laboratory of Eco-Hydraulics in Northwest Arid Region, Xi'an University of Technology, Xi'an, China
• Xi'an Summit Intellilink Technologies Co., Ltd, Xi'an, China

Short Summary

This study introduces a novel framework using the NSGA-II algorithm to jointly calibrate Manning’s roughness and seepage parameters in irrigation canals, significantly improving hydrodynamic modeling precision and providing a quantitative basis for targeted maintenance and enhanced water resource utilization.

Objective

• To establish a novel calibration framework using the NSGA-II optimization algorithm to simultaneously invert Manning’s roughness coefficient and seepage parameters in irrigation canals.
• To precisely simulate canal flow processes by integrating accurate parameter identification with Adaptive Mesh Refinement (AMR) technology, enabling timely irrigation schedule adjustments.
• To scientifically assess canal section deterioration and prioritize anti-seepage measures to enhance water resource utilization efficiency, boost agricultural productivity, and advance sustainable development in irrigation districts.

Study Configuration

• Spatial Scale: A 2,904 m segment of the Han-Yan Canal in the Ningxia Yellow River Irrigation District, China, divided into three reaches (S1, S2, S3) with four measurement sections (D1, D2, D3, D4).
• Temporal Scale: Data collected in May 2020 and May 2021. Calibration dataset: May 4–5, 2021. Test Set 1: May 6, 2021. Test Set 2: May 9, 2020. Simulation time step: 60 seconds.

Methodology and Data

• Models used:
  • One-dimensional Saint-Venant equations (coupled with seepage calculation).
  • Davidson-Wilson semi-empirical model for channel seepage.
  • NSGA-II (Non-dominated Sorting Genetic Algorithm II) for multi-objective optimization.
  • Adaptive Mesh Refinement (AMR) technology.
  • Genetic Algorithm (GA) for comparative single-objective stepwise calibration.
• Data sources:
  • Field investigations in multiple irrigation districts.
  • 864 sets of water level and flow observation data from the Han-Yan Canal in Ningxia.
  • Flow measurement data from the Fourth Management Station of Hanyan Canal (Section D1).
  • Water level and flow data recorded by integrated measurement and control gates at 5-minute intervals (Sections D2, D3, D4).
  • Canal engineering conditions (cross-section data).
  • Crop planting areas and total water diversion data (2016–2018 average) from the Fourth Management Station of the Hanyan Canal.
  • Water Use Quota Standard of Ningxia Hui Autonomous Region (DB64/T 971-2020).

Main Results

• Manning’s roughness coefficients exhibit significant spatial heterogeneity in canal sections constructed with the same technique after several years of operation.
• The NSGA-II algorithm successfully calibrated Manning’s roughness coefficients (n₀) and seepage parameters (C₀) for the three canal segments (e.g., S1: n₀=0.0179, C₀=1.149; S2: n₀=0.0192, C₀=2.139; S3: n₀=0.0185, C₀=1.190).
• Simulated water levels at three control sections generally align with measured trends, with absolute water level errors within 0.03 m (RMSE ranging from 0.005 m to 0.025 m).
• The multi-objective joint calibration method (NSGA-II) avoids physically unrealistic parameter values (e.g., negative seepage coefficients) that can arise from single-objective stepwise calibration.
• Adaptive Mesh Refinement (AMR) technology significantly improved simulation accuracy, reducing RMSE by approximately 50% (from 0.0099 m to 0.0045 m) and MAPE by approximately 50% (from 0.6413 to 0.3804) at Section D1.
• Seepage rates vary spatiotemporally, with higher rates observed during peak water depths and in specific canal sections (e.g., S2).
• Implementing anti-seepage renovations is projected to expand corn irrigation area by over 3.2 hectares, wheat by over 2.65 hectares, and rice by over 1.24 hectares for the management station.
• Sensitivity analysis revealed that RMSE is highly sensitive to Manning’s roughness coefficient n₀ (a ±20% change in n₀ causes over 800% increase in RMSE), while the seepage parameter C₀ has a minor impact on RMSE (approximately 2% fluctuation for ±10% change).
• The interaction between n₀ and C₀ can cause the seepage evaluation indicator (Obj2) to vary by up to 80.85%, indicating that both parameters must be considered simultaneously in channel seepage analysis.

Contributions

• Proposes a novel multi-objective joint calibration framework based on the NSGA-II algorithm for simultaneous identification of channel roughness and seepage parameters, advancing from single-parameter to multi-objective collaborative inversion with parameter sensitivity analysis.
• Integrates optimized calibration results with Adaptive Mesh Refinement (AMR) technology to significantly enhance the accuracy of channel flow simulation, providing a reliable basis for irrigation scheduling.
• Develops a dual-factor evaluation system combining parameter calibration outcomes with field investigation results to scientifically assess canal segment deterioration, supporting targeted rehabilitation and optimized fund allocation for water conservation and increased agricultural productivity.

Funding

• National Natural Science Foundation of China (Nos. 41571222, 51909208)
• Shaanxi Provincial Department of Science and Technology Qinchuangyuan “Scientist+Engineer” Team Development Project (2024QCY-KXJ-100)

Citation

@article{Li2025Joint,
  author = {Li, Li and Bai, Dan and Li, Yibo and Li, Meng and Lei, Jianhu and Zhen, Fangyong},
  title = {Joint calibration of Manning’s roughness and seepage in canals using NSGA-II for precision hydrodynamic modeling},
  journal = {Frontiers in Water},
  year = {2025},
  doi = {10.3389/frwa.2025.1709125},
  url = {https://doi.org/10.3389/frwa.2025.1709125}
}