Current recreational water quality science in the Great Lakes relies on measuring E. coli concentrations via cultivation techniques to estimate the risk associated with swimming in a particular waterbody. However, this dissertation showed such approaches inaccurately describe water quality across the entire beach and fail to represent underlining water quality issues. Water, sediment, and algae samples were collected in creeks, rivers, and beaches from multiple watersheds around Michigan. Samples were analyzed for cultivation based E. coli, enterococci, Clostridium perfringens, and coliphage as well as molecular markers for E. coli, Enterococcus spp., enterococci surface protein gene (human), Bacteroides (human and bovine) and Bacteroides thetaiotaomicron mannanase gene (human).In the Saginaw Bay, microbial indicators at four beaches were generally highest to lowest in stranded algae mats, sediment, shallow water, and deep water, respectively. Contamination in algae mats and sediment was identified in part as human specific using the enterococci surface protein gene. Higher concentrations of E. coli and enterococci in algae mats and sediment, compared to shallow and deep waters, were attributed largely to sediment bound bacteria and bacterial regrowth or persistence. Results demonstrated the potential for sediment and algae mats to act as non-point sources of pollution in the nearshore zone.Water and sediment samples collected from Mitchell Creek and Traverse City State Park beach quantified fecal indicator bacteria across space and time. Fecal indicator bacteriaconcentrations represented widespread, long-term, and recent fecal contamination in the Mitchell Creek. Despite the close proximity of the Mitchell Creek discharge to the Traverse City State Park beach, microbial concentrations were significantly lower (p < 0.01) in beach water which was partially impacted by creek discharge. Assessment of land use type at the watershed scale failed to identify consistent correlations with fecal indicator bacteria. However, Bacteroides thetaiotaomicron detections in both waterbodies indicated fecal contamination was partially human. Additional analysis on a subset of data identified significant disconnect between molecular and cultivation based results in creek and beach water. However, across all waterbodies, cultivated enterococci would have resulted in the greatest number of regulatory actions compared to cultivated E. coli and molecular based Enterococcus spp.A snapshot survey of 64 rivers discharging to the Great Lakes quantified E. coli and Bacteroides thetaiotaomicron under baseflow conditions. Bacteroides thetaiotaomicron was detected in all samples (X = 5.1 log10 Cell Equivalents 100 ml-1). The E. coli geometric mean across all rivers (1.4 log10 MPN 100 ml-1) suggests a potential regional reference condition. Classification And Regression Tree analysis indicated the total number of septic system in a watershed significantly impacted Bacteroides thetaiotaomicron concentrations under baseflow conditions. Land use characteristics better predicted microbial water quality than land use type.This work coupled molecular tools and novel monitoring strategies of unique environments (algae mats, sediments, beaches, small creek systems, and large river watersheds) to better understand the impact of human activities on Great Lakes water quality. Land use characteristics, not land use type, were related to qPCR markers in rivers which are transported to nearby beaches. Finally, septic systems, algae mats, and sediments were identified as non-point sources of pollution in Michigan surface water systems.
Read
