Analysis of three stream samples and a duck pond.

These four sites provide a range of potential conclusions from application of faecal source tracking tools in an urban environment. The analytical results from analysis of samples from these sites are presented in the table below, with text below explaining how these results were interpreted.

The duck pond and stream A

The duck pond was adjacent to stream site A and functions as a useful comparison point. No visible exchange of water occurred between the pond and the stream. The visible presence of ducks, and visual colouration of the duck pond suggests duck faecal material should be present, something confirmed by elevated E. coli levels of 2,200 MPN/100ml. While high, these were considerably lower than the adjacent stream samples. The E. coli levels were also only 20% of the total coliforms compared with 60-80% in the stream samples.

Human indicative FWAs and human specific PCR markers were not detected in either the duck pond or the stream A samples.

High levels of sterols were present in the duck pond, but interpretation suggests that the relative levels do not support the presence of human effluent. In the first instance the low ratio of coprostanol:cholestanol (<0.5) is inconsistent with human or animal derived sterols. The ratio of coprostanol:24-ethylcoprostanol is also less than 1, suggesting that it is not a human source.

Very low levels of sterols were present in the stream sample. In itself this suggests source of E. coli is not human, not duck, nor farmed animals. Low levels negate the use of ratios.

There is no evidence of human effluent in the samples analysed. The absence of FWAs, human specific molecular markers, and faecal sterols, all suggest a non-human source of the E. coli present. Further investigation was recommended to identify the source of the E. coli.

Stream B

This stream contains strong evidence of human source of faecal contamination. Elevated FWAs, human specific PCR markers, and sterol ratios all indicate a human source of faecal contamination.

Further site evaluation suggested that recent work installing a retaining wall may have damaged sewer pipes. Sampling around this point identified that upstream E. coli levels were <100 MPN/100ml, while 20 meters downstream levels of 15,000 MPN/100ml were detected. Lower, but still very high levels were detected further downstream.

Stream C

Despite low levels of E. coli in this sample compared to the other sites, faecal source tools suggested human pollution to this stream. FWA levels were low, at a level we would normally include in background levels. PCR markers were however positive for human markers, and faecal sterol analysis was also strongly indicative of human pollution.

The reason this site was studied, was because of the higher levels of E. coli normally observed. The reasons for the seeming lack of correlation between the E. coli levels and sterols in particular are not known. At this site input may be from a particular household, and seeming lack of correlation of individual markers with E. coli levels may reflect the variability that can be associated with sampling inputs from an individual, rather than larger population. In this scenario FWA levels may depend upon when clothes are washed, E. coli levels on the natural fluctuations in microbial gut populations. Differential removal of indicators may also occur depending on local conditions.

The key lesson here, is the benefit of a microbial toolbox approach, and the need for a sampling programme covering a range of conditions and times.