Presumptive Sources of Fecal Contamination in Four Tributaries to the New River Gorge National River, West Virginia, 2004

Several methods were used to determine the sources of fecal contamination in water samples collected during September and October 2004 from four tributaries to the New River Gorge National River -- Arbuckle Creek, Dunloup Creek, Keeney Creek, and Wolf Creek. All four tributaries historically have had elevated levels of fecal coliform bacteria. The source-tracking methods used yielded various results, possibly because one or more methods failed. Sourcing methods used in this study included the detection of several human-specific and animal-specific biological or molecular markers, and library-dependent pulsed-field gel electrophoresis analysis that attempted to associate Escherichia coli bacteria obtained from water samples with animal sources by matching DNA-fragment banding patterns. Evaluation of the results of quality-control analysis indicated that pulsed-field gel electrophoresis analysis was unable to identify known-source bacteria isolates. Increasing the size of the known-source library did not improve the results for quality-control samples. A number of emerging methods, using markers in Enterococcus, human urine, Bacteroidetes, and host mitochondrial DNA, demonstrated some potential in associating fecal contamination with human or animal sources in a limited analysis of quality-control samples. All four of the human-specific markers were detected in water samples from Keeney Creek, a watershed with no centralized municipal wastewater-treatment facilities, thus indicating human sources of fecal contamination. The human-specific Bacteroidetes and host mitochondrial DNA markers were detected in water samples from Dunloup Creek, Wolf Creek, and to a lesser degree Arbuckle Creek. Results of analysis for wastewater compounds indicate that the September 27 sample from Arbuckle Creek contained numerous human tracer compounds likely from sewage. Dog, horse, chicken, and pig host mitochondrial DNA were detected in some of the water samples with the exception of the October 5 sample from Dunloup Creek. Cow, white-tailed deer, and Canada goose DNA were not detected in any of the samples collected from the four tributaries, despite the presence of these animals in the watersheds. Future studies with more rigorous quality-control analyses are needed to investigate the potential applicability and use of these emerging methods. Because many of the detections for the various methods could vary over time and with flow conditions, repeated sampling during both base flow and storm events would be necessary to more definitively determine the sources of fecal contamination for each watershed.