Enhanced SWAT calibration through intelligent range-based parameter optimization.

Area of Science:

• Environmental Science
• Hydrology
• Water Resource Management

Background:

• Hydrological models are essential for environmental management, but parameter uncertainties affect output reliability.
• Optimized parameters are a major source of uncertainty in streamflow simulations.
• Improving model calibration is crucial for reducing prediction uncertainty and enhancing robustness.

Purpose of the Study:

• To introduce and evaluate a novel parameter calibration strategy using the Self-Organizing Map (SM) for the Soil and Water Assessment Tool (SWAT) model.
• To compare the performance of the SM strategy against traditional methods like Sequential Uncertainty Fitting version 2 (SUFI-2) for monthly streamflow simulation.
• To assess the impact of the SM strategy on simulation accuracy and prediction uncertainty in hydrological modeling.

Main Methods:

• A clustering technique, the Self-Organizing Map (SM), was applied to optimize parameter ranges in the SWAT model.
• Three calibration schemes were developed: SUFI-2, SUFI-2-Narrowing Down (SUFI-2-ND), and SM.
• Multiple diagnostic error metrics were used to evaluate simulation accuracy and prediction uncertainty for monthly streamflow in the Baishan Basin, China.

Main Results:

• The SM strategy demonstrated superior performance in describing watershed streamflow compared to SUFI-2 and SUFI-2-ND.
• The SM method particularly excelled in simulating spring snowmelt runoff (baseflow period).
• The SM approach effectively controlled prediction uncertainty, showing adaptability and reliability in interval optimization.

Conclusions:

• The Self-Organizing Map (SM) offers a valuable and reliable calibration tool for hydrological modeling.
• This strategy enhances the robustness and credibility of SWAT model predictions for environmental management.
• The SM method provides a promising approach to reduce streamflow prediction uncertainty and improve model performance.