Khan et al. (2026) Comparative Evaluation of Empirical and Data‐Driven Techniques for Estimating Seepage Losses in Irrigation Canals

⚠️ Warning: This summary was generated from the abstract only, as the full text was not available.

Identification

• Journal: Irrigation and Drainage
• Year: 2026
• Date: 2026-09-14
• Authors: Faridullah Khan, Hamza Farooq Gabriel, Hassan Akhtar
• DOI: 10.1002/ird.70224

Research Groups

• Department of Water Resources Engineering, University of Engineering and Technology, Lahore, Pakistan
• Centre for Water Informatics and Engineering (CWIE), University of Queensland, Australia

Short Summary

This study evaluates seepage estimation methods in irrigation canal systems using 6 years of data from the Thal Canal System in Pakistan, comparing empirical equations with machine learning models.

Objective

• Investigate the accuracy of various seepage estimation methods for different types of canal linings (plain cement concrete, brick-lined, and unlined).

Study Configuration

• Spatial Scale: Local scale (Thal Canal System in Pakistan)
• Temporal Scale: 6 years

Methodology and Data

• Models used:
  • CatBoost model (metaheuristic-optimised)
  • Symbolic regression (PySR) model
• Data sources: Inflow-outflow data from the Thal Canal System

Main Results

• Seepage losses were estimated to be 5% for PCC-lined canals, 13% for brick-lined canals, and 22% for unlined canals.
• The CatBoost model showed the highest predictive performance (R² = 0.97, RMSE = 0.80 m³/s).
• PySR also performed well (R² = 0.94, RMSE = 1.22 m³/s) and produced a single closed-form equation.

Contributions

• This study provides insights into the accuracy of seepage estimation methods for different types of canal linings.
• The results highlight the potential of machine learning models (CatBoost and PySR) in predicting seepage losses, offering a balance between accuracy and mathematical transparency.

Funding

• This research was funded by the Higher Education Commission of Pakistan (HEC) under project # 2558.

Citation

@article{Khan2026Comparative,
  author = {Khan, Faridullah and Gabriel, Hamza Farooq and Akhtar, Hassan},
  title = {Comparative Evaluation of Empirical and Data‐Driven Techniques for Estimating Seepage Losses in Irrigation Canals},
  journal = {Irrigation and Drainage},
  year = {2026},
  doi = {10.1002/ird.70224},
  url = {https://doi.org/10.1002/ird.70224}
}