Spatially explicit pollutant load-integrated in-stream E. coli concentration modeling in a mixed land-use catchment. (1st November 2018)

• Record Type: Journal Article
• Authors: Thilakarathne, Madusanka; Sridhar, Venkataramana; Karthikeyan, Raghupathy
• Is Part Of: Water research. Volume 144(2018)
• Page Start: 87
• Page End: 103
• Publication Date: 2018-11-01
• Subjects: SWAT -- Bacteria -- E.coli -- Source characterization -- Resuspension; Water -- Pollution -- Research -- Periodicals; 363.7394
• DOI: 10.1016/j.watres.2018.07.021
• Languages: English
• ISSNs: 0043-1354

Abstract

Abstract: Pathogen contamination is a major cause of surface water impairment in the United States, and fecal bacteria levels are typically used to evaluate microbial loading in bodies of water. Environmental models are considered a useful tool for evaluating watershed management practices. In this study, we assessed E. coli contamination of the Upper Stroubles Creek, Virginia, USA using the Soil and Water Assessment Tool (SWAT) model. The study area has been declared an impaired body of water due to recent bacterial contamination. Bacterial source characterizations play a critical role in such modeling exercises and especially in the case of non-point sources. As the SWAT model involves bacteria load estimation at a Hydrological Response Unit (HRU) level, we use the Spatially Explicit Load Enrichment Calculation Tool (SELECT) for our E. coli load estimations. We also evaluate current approaches to the measurement of bacterial interactions of the sediment-water interface using SWAT and the frequent measurements of streambed E. coli concentrations. For the simulation of in-stream E. coli concentrations using estimates drawn from SELECT without (with) sediment bacteria resuspension-deposition, Nash–Sutcliffe Efficiency (NSE) values of −0.41 to 0.34 (−0.19 to 0.36) are found. Moreover, in-stream E. coli concentrations measured at flow duration intervals show that the model frequently overestimates mid-range flows while underestimating low-range flows even with model improvements. The use of high-resolution E. coli loads and the consideration of sediment bacteria resuspension-deposition processes, generated higher E. coli concentrations for forested areas compared to those of urban and pasture lands, suggesting the importance of using detailed bacteria load estimations and land use information when assessing E. coli distribution in the environment. Highlights: A fine-scale E. coli load estimation based on the SELECT approach was employed. We evaluated an algorithm that uses sediment bacteria resuspension measures. SELECT source characterization is employed for non-point bacteria sources. Sediment bacteria resuspension is unique for medium flow and highly permeable conditions. The newly developed framework is key for watershed-scale analyses of E. coli levels.