Improving Hydrologic Model Performance by Using the UNET Model: A Case Study for the Illinois River Basin

Abstract

Using the U.S. Environmental Protection Agency’s (USEPA) Better Assessment Science Integrating Point and Nonpoint Sources (BASINS) system, the Illinois State Water Survey developed a Soil Water Assessment Tool (SWAT) model for the Illinois River basin. The SWAT model was calibrated and verified using observed flows at Marseilles, Kingston Mines, and Valley City along the Illinois River. Over the 1993-2000 simulation period, Nash-Sutcliffe efficiencies for daily, monthly, and annual flows exceeded 0.7 at the three gaging stations, except daily flows at Kingston Mines. However, comparisons between the simulated and observed daily flows indicated that simulated and observed peak flows deviated in magnitude and timing. The SWAT model uses Muskingum and variable storage-flow routing schemes, which are less capable of routing flows in the complex Illinois River that has locks and dams, backwater lakes, and gentle hydraulic gradients with backwater effects from the Mississippi River. The goal of this study was to improve the hydrologic modeling accuracy for daily flow routing on the Illinois River. A one-dimensional unsteady state hydraulic model (UNET) developed for the Illinois River was coupled with the calibrated SWAT model. This paper compares modeling accuracies of the coupled SWAT-UNET models and SWAT model results.