Reconceptualizing HRU Threshold Definition in the Soil and Water Assessment Tool

The Soil and Water Assessment Tool (SWAT) model simulates a watershed by dividing it into subbasins which are further divided into hydrologic response units (HRUs). User-defined area thresholds for land use, soil, and slope are often used when defining HRUs during model setup to improve computational efficiency by reducing the number of HRUs. This, however, results in loss of watershed biophysical information due to the reapportionment of HRUs that fail to exceed the threshold to other dominant HRUs. This study presents an improved HRU definition method that minimizes the loss of watershed biophysical information without considerably increasing the number of HRUs and lowering the user-defined HRU thresholds. Comparison of land use and soil distribution showed that the new HRU model closely matched the no-threshold full HRU model, unlike the default threshold SWAT model in which the landscape distribution characteristics were inadequately represented. Simulated hydrological and water quality variables, as well as model parameterization, were also better characterized when using the proposed HRU definition method when compared with the default SWAT model. The number of HRUs with the new method (934) was only slightly higher than the default threshold model (589) but considerably lower than the full HRU model (29,288). This new HRU definition method can help modelers perform computationally efficient modeling without compromising the accuracy of biophysical inputs to the model.

Automated summary

This study introduces an improved HRU definition method that minimizes the loss of watershed biophysical information without significantly increasing the number of HRUs or lowering user-defined thresholds. The new HRU model better represents land use and soil distribution characteristics and improves the simulation of hydrological and water quality variables compared to the default SWAT model.