Abstract

The use of host-associated Bacteroidales-based 16S ribosomal ribonucleic acid genetic markers was investigated as a tool for providing information to managers on sources of bacterial impairment in Vermont streams. The study was conducted during 2009 in two watersheds on the U.S. Environmental Protection Agency’s 303(d) List of Impaired Waters, the Huntington and the Mettawee Rivers. Streamwater samples collected during high-flow and base-flow conditions were analyzed for concentrations of Escherichia coli (E. coli) and Bacteroidales genetic markers (General AllBac, Human qHF183 and BacHum, Ruminant BoBac, and Canid BacCan) to identify humans, ruminants, and canids as likely or unlikely major sources of fecal contamination. Fecal reference samples from each of the potential source groups, as well as from common species of wildlife, were collected during the same season and from the same watersheds as water samples. The results were combined with data from other states to assess marker cross reaction and to relate marker results to E. coli, the regulated water-quality parameter, with a higher degree of statistical significance.

Results from samples from the Huntington River collected under different flow conditions on three dates indicated that humans were unlikely to be a major source of fecal contamination, except for a single positive result at one station that indicated the potential for human sources. Ruminants (deer, moose, cow, or sheep) were potential sources of fecal contamination at all six stations on the Huntington River during one high-flow event and at all but two stations during the other high-flow event. Canids were potential sources of fecal contamination at some stations during two high-flow events, with genetic-marker concentrations in samples from two of the six stations showing consistent positive results for canids for both storm dates. A base-flow sample showed no evidence of major fecal contamination in the Huntington River from humans, ruminants, or canids.

Results from samples from the Mettawee River watershed collected during high-flow conditions (12 storm samples on 2 dates at 6 stations) indicated that there was no evidence of fecal contamination from humans in seven samples and possible evidence in five samples. Results for humans were positive for only one station during both storm events. For two of the five samples with evidence for human fecal contamination, results for two different human genetic markers agreed, but results from three samples were inconsistent. In samples from five of the six Mettawee stations, ruminants were a potential source of fecal contamination on at least one of the three sampled dates, including three positive results for the base-flow sample. Yet samples from all of the stations that showed positive results for ruminants did so for only one or two of the three sampled dates. Samples from only one of the six stations gave consistent results, which were negative for ruminants for all three dates. In the Mettawee River base-flow sample, humans were an unlikely source of major fecal contamination.

Factors that may influence results and conclusions include the timing of sample collection relative to the storm event; variability of E. coli and Bacteroidales concentrations in fecal reference samples and in water; sampling and analytical errors; the potential cross reactivity of host-associated genetic markers; and different persistence and survival rates of E. coli bacteria and Bacteroidales genetic markers on land, in water, and by season. These factors interfere with the ability to directly relate Bacteroidales concentrations to E. coli concentrations in river samples. It must be recognized that while use of Bacteroidales genetic markers as a source tracking tool coupled with the interpretive approach described in this report cannot be used quantitatively to pinpoint sources, it can be used to exclude potential sources as major contributors to fecal contamination.

First posted October 18, 2011