A distributed non-point source pollution model: calibration and validation in the Yellow River Basin

Area of Science:

• Environmental Hydrology
• Water Resource Management
• Non-point Source Pollution Modeling

Background:

• Assessing non-point source pollution in large river basins is crucial for water resource management.
• The Yellow River basin faces challenges with high sediment runoff modulus (770 t/km2).
• The applicability of the Soil and Water Assessment Tools (SWAT) model in such environments requires examination.

Purpose of the Study:

• To evaluate the applicability of the SWAT model for simulating flow and sediment in a large river basin.
• To calibrate and validate the SWAT model using observed hydrological and sediment data from the Yellow River basin.

Main Methods:

• Established a comprehensive database using GIS, including DEM, soil, land use, weather, and land management data.
• Employed a two-stage "Brute Force" optimization method for parameter calibration (1992-1997), separating direct and base flow using a digital filter technique.
• Validated the calibrated model using observed data from 1998-1999.

Main Results:

• The SWAT model successfully simulated monthly flow and sediment runoff in the Yellow River basin.
• Calibration and validation achieved high accuracy: relative error within 20%, coefficient of determination and Nash-Suttcliffe efficiency ≥ 0.70.
• The model demonstrated good performance in matching both flow and sediment yield.

Conclusions:

• The SWAT model is a suitable tool for non-point source pollution assessment in large river basins with high sediment loads.
• The established methodology provides a reliable framework for applying SWAT in similar complex hydrological systems.
• Accurate simulation of water and sediment dynamics is achievable with proper calibration and validation.