Application of curve-fitting techniques to develop numerical calibration procedures for a river water quality model

Area of Science:

• Environmental Science
• Water Resource Management
• Computational Hydrology

Background:

• River water quality models often face challenges due to complex structures and parameter estimation difficulties.
• Accurate parameterization is crucial for reliable water quality predictions.

Purpose of the Study:

• To introduce a novel and simplified calibration strategy for river water quality models.
• To enhance the efficiency and robustness of parameter estimation using limited data.

Main Methods:

• Development of a simple river water quality model (RWQM) structure.
• Application of cubic smoothing spline, polynomial curve-fitting, and nonlinear least squares for parameter estimation.
• Implementation of a sensitivity ranking system to assess parameter quality.

Main Results:

• Calibrated model parameters showed good agreement with river water quality data.
• Sensitivity analysis indicated minimal output variation for indices below 10% and significant variation above 50%.
• Statistical analysis (ANOVA) confirmed no significant difference between model and measured data means (p=0.8431).

Conclusions:

• The proposed calibration strategy is effective for river water quality modeling, even with small datasets.
• A MATLAB GUI tool was developed, offering a user-friendly platform for model calibration and data analysis.
• The approach provides a robust and simplified method to support water quality model applications.