Fecal Bacterial Contamination of a Karst Watershed in Central Missouri.
Robert Lerch, USDA-ARS, Columbia, MO 65211-0001 and Robert Kremer, Univ. of Missouri, 302 ABNR Bldg., Columbia, MO 65211-7250.
The Bonne Femme watershed of Boone County, Missouri has a varied surface geology that includes karst topography with losing streams that are an especially vulnerable setting for ground water contamination. The study objective was to compare fecal contamination and detection of specific pathogenic water-borne bacteria within the major sub-watersheds and relate this contamination to land-use and hydrology. Ten sub-watersheds were sampled weekly, for one month per quarter-calendar year since 2003 for fecal coliforms, E. coli, enterohemorrhagic E. coli O157:H7, Salmonella, and Shigella. Fecal coliform and E. coli enumeration were done by the membrane filtration techniques, and pathogen specific analyses were performed through culture enrichment of water samples followed by DNA extraction of bacterial growth and PCR using pathogen-specific primers. Under low-flow conditions, fecal coliforms and E. Coli levels were typically less than 1000 cfu/100 mL, but many sites exceeded state and federal whole body contact limits in the 2nd and 3rd quarters of the year. Under high flow conditions, most sites exceed 10,000 cfu/100 mL, and whole body contact limits were always exceeded. Salmonella and Shigella were detected in at least two streams in each quarter of 2005 and 2006; E. coli O157:H7 was detected in at least one stream since 2nd quarter 2005. Frequency of detection varied with stream flow conditions; for example in 2nd quarter 2005, all pathogens were detected at half the sites under high flow conditions compared to detections in only 2 streams earlier in the quarter under baseflow conditions. During this same quarter, relative intensities (cell densities) of Salmonella varied inversely with stream flow, but E. coli O157:H7 detections were most intense under high flow conditions. Detection of pathogenic species showed that PCR-based detection can be effectively used to demonstrate the presence of live bacterial pathogens in different streams within a watershed.