Event-aware SWAT+ calibration for stormflows with monthly nutrient data in the Qingjiang River Basin

Area of Science:

• Hydrology and Water Resources Management
• Environmental Engineering
• Geospatial Analysis

Background:

• Nonpoint-source (NPS) pollution export in steep mountain basins is sensitive to storm runoff peaks.
• Traditional monthly nutrient monitoring limits event-scale hydrological model calibration.
• Accurate simulation of storm hydrograph dynamics is crucial for managing water quality in these regions.

Purpose of the Study:

• To develop an event-aware Soil and Water Assessment Tool Plus (SWAT+) calibration framework.
• To improve the accuracy of discharge and nutrient load simulations during storm events.
• To enhance uncertainty quantification in hydrological modeling for mountainous basins.

Main Methods:

• Coupling Hippopotamus Optimization (HO) with SUFI-2 uncertainty fitting for SWAT+ calibration.
• Incorporating an event regularization term based on daily discharge data (peak magnitude, time-to-peak, recession slope).
• Utilizing split-sample calibration/validation (2008-2018) and out-of-sample testing (2020-2024).

Main Results:

• Improved monthly discharge Nash-Sutcliffe efficiency (NSE) from 0.72 to 0.79 (calibration) and 0.70 to 0.73 (validation).
• Reduced median relative peak-magnitude error (0.104 to 0.069) and tightened peak-timing dispersion.
• Maintained or improved monthly nutrient (TP) NSE (0.70 to 0.73) without degrading performance.

Conclusions:

• The HO-SUFI-2 framework effectively constrains storm hydrograph dynamics under data limitations.
• The approach provides practical, uncertainty-aware diagnostics for storm-period risk screening in mountainous basins.
• Out-of-sample performance indicates model robustness, though biases in 2020-2024 may relate to unrepresented management changes.