Bacterial Source Tracking in an Eastern Tennessee Stream using Bacteroides Host Associated Real-Time Polymerase Chain Reaction Assays

A bacterial source tracking study using Bacteroides host associated real-time PCR assays was performed to determine the sources of fecal contamination in Pond Creek (HUC 06010201013). Pond Creek, located in the Ridge and Valley physiographic region in Eastern Tennessee, is a 303(d) listed stream that fails to meet water quality standards for pathogens. Water samples and discharge were measured monthly at eight locations from November 2005 to November 2006. Grab samples were analyzed for several chemical parameters and for microbial fecal indicator organisms, namely Bacteroides spp., Escherichia coli, and Enterococcus. The objectives of the study were to quantify total, human, and bovine associated Bacteroides. Additionally, we investigated spatial and temporal variation of fecal indicator organisms and created load duration curves for each sampling site.

The results showed that Escherichia coli concentrations regularly exceeded water quality standards. Bacteroides host associated real-time PCR assays indicated that cattle were the dominant source of fecal pollution (99 percent of total Bacteroides). Although human-associated Bacteroides were detected, their concentrations remained relatively low across the watershed. Load data show that fecal contamination from bovine sources occurs at elevated levels throughout the watershed; no statistical differences between sites were observed for bovine associated Bacteroides (BoBAC) and Bacteroides spp. belonging to the Bacteroides genus (AllBAC) loads. Additionally, bovine-associated Bacteroides concentrations were very highly correlated (r² = .903) with the total Bacteroides concentrations. Load Duration Curves (LDCs) also indicated extensive bovine fecal pollution. Load data separated into human and bovine LDCs showed that Escherichia coli loads from bovine sources were mostly flow dependent whereas human associated Escherichia coli loads were generally flow independent.

Temporal variations followed seasonal weather patterns; mean loads of all fecal indicators, except Enterococcus, were greatest during the months of highest precipitation and lowest in the drier months. No temporal patterns were established for concentrations of fecal indicator organisms. This suggests that runoff transported the majority of fecal inputs to Pond Creek. Best management practices (BMPs) such as improving pastures, nutrient management, proper manure storage, controlling livestock stocking densities, vegetative filter strips, and riparian fencing with careful riparian grazing, should be implemented to reduce fecal inputs from cattle and help Pond Creek meet TMDL guidelines.