Inter seasonal precipitation patterns impact the occurrence of waterborne pathogens in an agricultural watershed

Background/Question/Methods: Runoff from agricultural fields undergoing manure applications or housing livestock operations may carry a variety of chemical and microbial contaminants that compromise water quality and increase the possibility of human exposure to pathogenic microorganisms. Fecal indicator bacteria (FIB) such as Escherichia coli or Enterococcus spp. are commonly used to demonstrate fecal contamination in water. Because the prevalence of FIB may be used to suggest the potential presence of waterborne pathogens, understanding the behavior of waterborne pathogens in reference to that of the traditional indicators is an important step for development of Risk Assessment models. The purpose of this study was to assess the environmental factors affecting occurrence of waterborne pathogens and FIB in a watershed impacted by mixed land uses. Synoptic sampling was conducted under baseflow conditions from November 2012 to December 2013. Escherichia coli, Enterococcus sp., Salmonella (invA), and Shiga toxin producing pathogens (Stx2) were measured at 15 locations in the South Fork Broad River (SFBR) watershed in N.E. Georgia, USA. Decision trees (i.e. classification trees), a nonparametric data mining method, were used to determine independent variables describing the occurrence of waterborne pathogens. Environmental parameters used to predict occurrence of individual pathogens included: E. coli and enterococci concentrations in water and sediment, pH, water temperature, conductivity, total suspended solids (TSS), dissolved oxygen (DO), percent sand in the stream bed, percent forest cover, 24- and 48-h rainfall prior to the sampling day, and concentrations of microbial source tracking (MST) markers (in water and sediment) for cattle, ruminants, and humans. GIS analysis was performed to describe the distribution of pathogens across the watershed and identify seasonal and land use effects.

Results/Conclusions: Salmonella and Stx2 were detected in 76.2%, and 61.1% of the water samples, respectively, while they were detected in 51.1%, and 43.3% of the sediment samples, respectively. Water E. coli concentration, temperature and ruminant marker concentration in sediment predicted occurrence of Stx2 in water, while DO, pH and water E. coli concentrations predicted presence of Salmonella in water. Salmonella was identified at a higher frequency (>75%) during the wet season while the Stx 2 virulence gene was more frequently identified (>75%) during the dry season. Results indicate a strong relationship between pathogen frequencies, dry/wet seasonal patterns, and FIB and MST marker measurements. More importantly, this study shows the utility of integrating microbial water quality measurements with agricultural management practices to predict pathogen occurrences in headwater streams in mixed used watersheds.