Impacts of Reservoir Operation in the SWAT Model Calibration

Abstract

Listen

Translate article

The Soil and Water Assessment Tool (SWAT) model has widely been used as an effective tool for assessing water resources. However, an evaluation of the reservoir component has rarely been addressed despite the accuracy of simulated streamflow, which can be reduced by not considering a reservoir in the model. The purpose of this paper is to evaluate the impacts of reservoir outflow estimation methods in calibrating the SWAT model. This study was conducted within the Grenada Lake Reservoir watershed (4,500 km2) within the Yazoo River Basin. The SWAT model was calibrated and validated using long-term measured streamflow data from the U.S. Geological Survey gage stations (07282000, 07283000, and 07285500) and reservoir outflow data from U.S. Army Corps of Engineers. Two upper sub-watersheds contributing to inflow of reservoir were first calibrated and validated for streamflow. Comparing with the measured streamflow data, the model performance showed satisfactory during both calibration and validation periods. Based on the calibrated parameters, the impacts of the reservoir outflow estimation methods were evaluated at the watershed outlet (07285500) of the study watershed, which is located below the Grenada reservoir. According to the statistical analysis, the measured outflow options showed satisfactory results for daily stream flow data, whereas the target release options showed unsatisfactory results. However, all four reservoir outflow options showed satisfactory results for monthly streamflow data. These results indicates that the target release approach using non-specific or specific target release approach in the SWAT model can effectively use to simulate the monthly streamflow with the limited information of reservoir. The results from our study can provide the useful information in applying the SWAT model for the watershed including the flood control reservoirs.