The SWAT model (https://swat.tamu.edu/) is one of the most widely used ecohydrological models worldwide and has been tested across an extensive range of watershed scales and environmental conditions (https://www.card.iastate.edu/swat_articles/). However, a considerable need remains for further testing of SWAT, especially to support applications for large river basin systems. This was a fundamental goal of the UMRB SWAT model described in this research, which was developed using the Hydrologic and Water Quality System (HAWQS) on-line platform (https://hawqs.tamu.edu/) with pre-loaded soil, land use, climate, topographic and other data layers. The testing was performed in support of DOE-funded research focused on evaluating climate model metrics and specifically included evaluation of (1) Hargreaves (HG) and Penman-Monteith (PM) PET methods, (2) National Climatic Data Center (NCDC) and Parameter-elevation Regressions on Independent Slopes Model (PRISM) climate datasets, available in HAWQS, and (3) testing of Livneh climate data obtained external to HAWQS.
The effects of the two different PET methods and the three climate data sets evaluated here have not been reported in previous applications of SWAT for the UMRB. The SWAT simulations showed that the Livneh-PM combination resulted in the highest average annual water yield of 380.6 mm versus the lowest estimated water yield of 193.9 mm for the Livneh-HG combination, in response to 23-year (1983 to 2005) uncalibrated simulations. The HG method was also found to result in consistently higher ET/PET estimates on both an annual and growing season basis. The research conducted in this study further revealed that the strongest replication of UMRB streamflows at multiple gauge sites during the 23-year SWAT simulations, based on a spatial validation testing approach, were obtained using the HG PET method and PRISM climate data set. Relatively accurate streamflow estimates were also found based on a combination of the HG PET method and Livneh climate data set.
Overall satisfactory results were found for the calibration and validation gauge sites, with the majority of R2 values ranging between 0.61 and 0.82, Nash-Sutcliffe modeling efficiency (NSE) values ranging between 0.50 and 0.79, and Kling-Gupta efficiency (KGE) values ranging between 0.61 and 0.84. The combination of PRISM climate data and HG PET method was determined to be the best option for simulation UMRB hydrologic dynamics although the Livneh climate data set and HG PET method combination was also acceptable. In contrast, the NCDC climate data provided via HAWQS resulted in distinctly weaker streamflow results and thus is not recommended for further analyses. The results obtained in this study are providing a foundation for the next phase of research focused on evaluating contemporary climate data provided by selected regional climate models (RCMs). The approach will be extrapolated to other major river basins within the overall Mississippi River Basin (MRB) systems to support broader research initiatives within the DOE-funded research initiative.