Assessing the use of alternative soil data in hydrological and water quality modeling with SWAT+: SSURGO and POLARIS at subbasin and field scales

Authors: NAO-YARASCA, EFRAIN - Blackland Research And Extension Center; OSORIO LEYTON, JAVIER - Blackland Research And Extension Center; White, Michael; GAO, JUNGANG - Blackland Research And Extension Center; Arnold, Jeffrey

Submitted to: Water Research
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 2/23/2025
Publication Date: 2/25/2025
Citation: Nao-Yarasca, E., Osorio Leyton, J., White, M.J., Gao, J., Arnold, J.G. 2025. Assessing the use of alternative soil data in hydrological and water quality modeling with SWAT+: SSURGO and POLARIS at subbasin and field scales. Water Research. https://doi.org/10.3390/w17050670.
DOI: https://doi.org/10.3390/w17050670

Interpretive Summary: Accurate soil data is vital for hydrological modeling, but few studies compare new datasets like POLARIS with traditional ones like SSURGO. This study used the SWAT model to evaluate streamflow and sediment yields in the Big Muddy Creek watershed using both datasets. Results showed similar performance between these datasets at the sub-basin level, but with more variability at the HRU scale. Streamflow predictions were expectedly more reliable than sediment yield, which is more sensitive to soil differences. Comparisons to observed data revealed varying accuracy, highlighting the need for better calibration and more data. POLARIS is a viable alternative to SSURGO, especially in diverse or data-limited areas for the purpouses of SWAT modeling.

Technical Abstract: The accuracy of soil databases is essential in hydrological modeling, yet limited studies have evaluated the implications of using emerging soil datasets like POLARIS compared to traditional ones such as SSURGO. This study investigates the performance of the Soil and Water Assessment Tool (SWAT) in simulating streamflow and sediment yield within the Big Muddy Creek watershed using both datasets. The objectives were to evaluate the consistency of model outputs at two spatial scales—sub-basin and field (HRU)—and to assess the predictive capabilities of the datasets under varying spatial resolutions. Using 81 sub-basins and selected HRUs representing croplands, simulations were analyzed for streamflow and sediment yield. Performance metrics, including Nash-Sutcliffe Efficiency (NSE), correlation coefficients (r), and Percent Bias (PBIAS), revealed comparable results at the sub-basin scale, with 88.9% of sub-basins achieving "very good" PBIAS thresholds. However, alignment between datasets diminished at the HRU scale, where only 42.9% and 33.3% of HRUs achieved similar thresholds in two analyzed sub-basins. Streamflow predictions demonstrated stronger consistency between datasets than sediment yield, which exhibited higher sensitivity to differences in soil parameterization. Simulated streamflow using both soil datasets compared to observed streamflow highlighted variability in model performance across six gauging sites, ranging from "very good" to "unsatisfactory," suggesting the need for more rigorous calibration and expanded observational data. This study highlights POLARIS as a robust alternative to SSURGO for hydrological modeling, particularly in heterogeneous or data-sparse regions. Future research should focus on the broader application of POLARIS across diverse hydrological processes, climatic conditions, and land uses, emphasizing its potential to advance water resource management and enhance model precision.