Monitoring and tracking contaminant sources in catchments and estuaries

Bivalve shellfish which include all species of clams, cockles, oysters, mussels and scallops feed by filtering food particles from the surrounding water. When these waters are contaminated, filter feeding shellfish accumulate the contaminants to many times the concentration of the surrounding water. Shellfish harvested from water contaminated by human sewage waste pose a significant threat to public health (Fleet 1978; Dore & Lees 1995) and, in recent years in Australia and worldwide, there have been a number of highly publicized cases of contaminated estuarine waters used for oyster growing (Geary & Whitehead 2011; Conaty et al. 2000; EFSA 2012). In each incident involving contaminated aquaculture products, agricultural waste management practices, runoff from urbanized areas and various human-derived faecal sources are usually found to be responsible. Faecal contamination from human sources may be a result of licenced wastewater treatment plant discharges to receiving waters, illegal practices, or the more diffuse contributions made by failing on-site wastewater management systems (OWMSs) from unsewered premises. To better manage water quality, it is important to be able to distinguish between the various sources of faecal contamination contributing to declining water quality currently being observed in many catchments and estuaries, particularly following heavy rainfall.

Of the tracking methods used to distinguish wastewater contaminant sources, a substantial amount of work has been undertaken using faecal sterol analysis (FSA) to distinguish sources of contamination in waters and sediments. All faecal material contains sterol compounds, and their breakdown products, stanols. The distribution of sterols found in faeces, and hence their usefulness in source tracking, arises from a combination of diet, an animal's ability to synthesize its own sterols, and the conversion of sterols by intestinal microbiota in the digestive tract. Coprostanol constitutes about 60% of the total sterol content of human faeces and is produced by biohydrogenation of cholesterol by anaerobic bacteria in the intestines of humans and higher mammals. It is unaffected by physical factors such as temperature and salinity. In the excreta of herbivores, 24-ethylcoprostanol has been found to be the principal faecal biomarker, whereas other animals which are ubiquitous in urban areas, such as dogs and birds, either do not have coprostanol in their faeces, or it is present in trace and/or smaller amounts, thus providing a diagnostic dichotomy of presence/absence. Distinguishable sterol profiles for humans, herbivores and birds have been found to be sufficiently distinctive to be of diagnostic value in determining whether faecal pollution in water is of human or animal origin (Leeming et al. 1998). As a consequence FSA has been used to successfully trace faecal pollution in marine, estuarine and freshwater environments.

Another tracking method which has been used is to examine the fate of human-derived effluent using the fluorescing property of optical brighteners. Fluorescent whitening compounds (FWCs) are added to washing powders to adsorb to fabrics and brighten clothing by fluorescing when exposed to ultraviolet light. In Australia there are two primary types of FWCs that are added to laundry powders: disodium distyrylbiphenyl disulfonate (DSBP) and disodium anilinomorpholinotriazinyl aminostilbene sulfonate (DAS1) (Figure 1). These optical brighteners emit light in the blue range (415–445 nm) to compensate for undesirable yellowing in clothes (Hartel et al. 2007). Laundry detergents contain approximately 0.10–0.20% (w/w) of FWCs (with DAS1 used in greater weight proportions), of which between 20 and 95% binds to the fabrics during washing, with the remainder being discharged (Devane et al. 2006; Stoll et al. 1997). There is no regulatory limitation on the dosing of FWCs in laundry applications so the dosage essentially comes down to cost versus performance. There is a peak (or plateau) in performance when using these compounds, thus rendering an overdose both expensive in terms of formulation cost, and useless in relation to performance enhancement.

FWCs as indicators of human sourced contamination have been used in several countries to identify leaking sewer mains (Chandler & Lerner 2015) and in sanitary surveys to identify wastewater contamination from failing OWMS. Devane et al. (2006), Gilpin et al. (2002) and Hartel et al. (2007) have reported the usefulness of this technique when combined with other indicators including human faecal sterols, although natural organic matter can also contribute to some of the observed fluorescence in water samples. FWCs are therefore considered good markers or indicators of a human wastewater source, and where present with faecal bacteria or human sterol compounds, may suggest that human sourced contaminants are present in stormwater runoff and receiving waters (Cao et al. 2009; Hartel et al. 2007). Table 1 provides a simple analysis of the likely cause of faecal contamination using indicator faecal bacterial numbers and FWC concentrations.

In this study the faecal sterol and FWC source identification methods described above have been applied to waters associated with the Bundabah Oyster Harvest area on the northern shore of Port Stephens, near Newcastle in NSW, Australia (Figure 2) to assess the presence of human-derived contamination in stormwater runoff entering the estuary and within the estuary itself. The source tracking results have been combined with traditional bacterial indicator counts (faecal coliforms (FCs)) for two rainfall events. Following significant rainfall and when specified threshold levels are exceeded, commercial shellfish producers in New South Wales are required to cease harvesting due to high faecal bacterial loads that may enter estuaries in stormwater runoff. The issue at this location is whether the apparent faecal contamination in the estuary detected using traditional indicators, is in part, derived from human sources such as failing OWMS.

Bundabah Harvest Area is one of 12 harvest areas in Port Stephens. It has been classified as Restricted since 2006 which means that the growing area classified by the NSW Food Authority is an area from which shellfish may be harvested only with their approval and is then subjected to an effective purification process such as relaying or depuration. Owing to recent exceedances with respect to FC numbers in this harvest area, a revision to the current management plan for Bundabah is underway and the results of this work are planned to inform that review.

Adjacent to the Bundabah estuary there are two small unsewered communities on rocky and steep foreshores. Many of the OWMS in these communities are close to the water, visible from a survey boat and directly adjacent to the harvest area and leases. Inspections of the OWMS by Great Lakes Council which are conducted at intervals of 12 or 24 months have been primarily aimed at identifying system failure and/or illegal greywater disposal. In North Arm Cove, which is the larger of the two communities, the number of OWMS totalled 307 with the predominant systems being pump-out (101) involving a septic tank and collection tank, aerated treatment systems (94) and soil-based systems utilizing absorption and evapotranspiration (88). The remaining 24 systems comprised biological, chemical, composting, soil mound, sand and wet composting systems. Some unregulated greywater reuse is also practiced within the community. All pump-out systems with respect to pump truck volumes and dates are monitored, and records entered into a database to assess whether effluent is being recovered from the holding tanks at regular intervals. However there is concern on occasions that there may be illegal discharge of untreated effluent to the surface drainage system.

Water samples were collected from nine locations within this harvest area on the ebb tide following two rainfall events (24 November 2011 and 12 November 2013) and then analysed for salinity (electrical conductivity), FCs (membrane filtration method), FWCs (Turner Designs Trilogy Fluorimeter) and various faecal sterol compounds (gas chromatography-mass spectrometry). Apart from one control sample (Site 11) located in an open body of water in Port Stephens, two samples were collected from the main stormwater drains in the unsewered communities of North Arm Cove and Bundabah (Sites 9 and 10, respectively) prior to entry into the estuary, while two other samples were collected from streams entering the estuary (Site 7 Bundabah Creek and Site 8 Bulga Creek), with the remainder of sites located adjacent to the oyster leases within the estuary (Sites 1–4). Of these, Site 2 was at the head of the estuary where the tributary streams (Bundabah and Bulga) enter, while Site 4 was at the mouth of the estuary adjacent to Port Stephens waters.

The rainfall between 22 and 24 November 2011 (when sampling occurred) was 53.5 mm according to records obtained from near the estuary. This amount of rainfall was sufficient to result in closure of the oyster harvest area to commercial oyster harvesting (where the rainfall event exceeded 40 mm in 48 hours). For the second sampling event, 37 mm was recorded in the period 24–48 hours prior to sampling on 12 November 2013. Much of this part of eastern Australia experienced very dry conditions over this period and there were very few events which resulted in an exceedance of the rainfall intensity threshold which had been set by the NSW Food Authority. As a consequence there was little opportunity for any additional sampling between these events.

The sample results for salinity (electrical conductivity), FC bacteria, FWCs and coprostanol concentrations for each round of sampling are listed below in Tables 2 and 3.

With regard to the first round of water sampling, the FC bacterial analyses revealed high counts (>85 cfu/100 mL) at four sites; Sites 9 and 10, the North Arm Cove and Bundabah stormwater drains, respectively; Site 7 Bundabah Creek entrance and Site 2 at the head of the estuary (Table 2). These counts exceeded the NSW Food Authority guidelines for oyster growing waters in Restricted harvest areas (Safefoods NSW 2001) warranting the closure of the estuary to oyster harvesting following heavy rainfall. The four water samples with the highest FC counts also returned the highest FWC concentrations. While there are no standards for FWC compounds in environmental waters and little data available locally, these concentrations are elevated relative to the control site (Site 11 Port Stephens main bay) and Sites 1, 3 and 4 in the Bundabah estuary. Studies which have been undertaken overseas using different analytical instrumentation High Performance Liquid Chromatography (HPLC) to that used here for the compounds shown in Figure 1 have suggested that FWC concentrations similar to those at Sites 2, 7, 9 and 10 are also indicative of the presence of human sourced contamination (Devane et al. 2006, Gilpin et al. 2002).

These four samples were also the only samples in which coprostanol, the key sterol indicator of potential human faecal contamination, was detected as shown in Table 2. Three of the four samples had elevated levels of coprostanol which as mentioned previously is a key human source marker. While Table 2 includes these concentrations, the various other faecal sterol compounds analysed and their ratios are not included here for brevity. With regard to the FSA undertaken, a ratio analysis is typically used to determine the likelihood of human and or animal faecal sources (Shah et al. 2007). Contamination of human and/or bovine origin is generally indicated by a coprostanol/chloestanol ratio ≥0.4 which was observed only in the case of Sites 2 (0.46) and 7 (0.42), although a value close to this threshold (0.36) was also recorded for Site 10. In these cases the low epicoprostanol/coprostanol ratios indicated that the contamination involved relatively fresh faecal material, while the non-detection of 24-ethylcoprostanol suggested that the contamination was more likely human than bovine in origin.

While the ratio values calculated for these samples are indicative of likely human faecal contamination, it is recognized that the interpretive reliability of the ratio analysis is far less robust at relatively low coprostanol concentrations. Notwithstanding this, the detection of coprostanol and high FC counts in these samples, in conjunction with the highest FWC concentrations (refer to interpretative Table 1), indicates a likelihood that these sites had been contaminated with a detectable level of human sourced wastewater.

The bacterial analyses in the second round of sampling again revealed high FC counts (>85 cfu/mL) at three sites where runoff enters the estuary – Sites 9 and 10, the North Arm Cove and Bundabah stormwater drains respectively, and Site 7 in Bundabah Creek (Table 3). The next highest FC count was Site 8 in Bulga Creek (70 cfu/100 mL) which was not sampled in Round 1. As mentioned previously, the NSW Food Authority faecal bacteria threshold was exceeded again for three of the sites (similar to the first round of sampling) even though the rainfall recorded was less than the designated threshold for closing the estuary to commercial oyster harvesting.

The stormwater samples at Sites 9 and 10 with the highest FC counts also returned the highest concentrations of FWCs. Again these concentrations are elevated relative to the control site (Site 11 Port Stephens main bay) and Sites 1, 3 and 4 which are in the estuary adjacent to oyster leases. Using the interpretative Table 1, it would appear that the high FC counts at Sites 9 and 10 coupled with their elevated FWC concentrations are indicative of the presence of human sourced contamination. With regard to FSA, none of the water samples collected in this round contained detectable levels of coprostanol, a key sterol indicator of potential human faecal contamination (refer Table 3). Other markers such as epicoprostanol and 24-ethylcoprostanol were not detected in the water samples, although several plant sterols were present in low concentrations. Given the non-detection of coprostanol, further calculation of sterol ratios was considered redundant and therefore not conducted in this case. Even given the non-detection of coprostanol, the elevated FWC signals observed at the stormwater drains (Sites 9 and 10), coupled with the elevated FC counts (particularly at Site 10) indicate that there is a human sourced contribution to stormwater runoff entering the estuary. In addition to the significant proportion of OWMS on pump-out and the possibility of illegal discharge, it is known that greywater (which is separated from black water and directly land applied in both communities) may enter the surface drains following land application and heavy rainfall. Using the interpretative Table 1 again, it is likely that there may also have been human sourced contributions to the stormwater runoff on this occasion.

Further sampling and analysis at this location has so far been hampered by the lack of rainfall which meets or exceeds the industry rainfall intensity threshold, although the poorer quality of the “fresh” stormwater runoff entering the estuary following rainfall is apparent. Faecal bacterial monitoring of oyster growing waters within the estuary indicates that there are many more exceedances occurring with respect to the food safety standards in place for oyster growing waters. It is very likely that the poorer quality stormwater runoff entering the estuary is contributing to this decline in estuary health.

In managing water quality at the catchment scale, it is difficult to demonstrate direct linkages between community wastewater management practices, stormwater runoff and estuarine water quality using standard monitoring techniques and microbiological indicators. In addition, to determine the fate of all contaminants generated from land-based practices is difficult and often expensive. While traditional faecal indicator bacteria are used in water quality monitoring for a variety of reasons (ease of analysis, cheaper cost and easier interpretation), chemical source tracking techniques appear to be useful in determining whether the contaminant sources in catchments and estuaries are likely to be human or of other animal origin. They assist in providing multiple lines of evidence with respect to the regular bacteriological monitoring which is undertaken in estuaries where aquaculture is practiced. This is particularly important from a public health point of view as any source contribution due to human activities which is identified can then be managed from a regulatory perspective.

On the basis of this approach and the methodology outlined, the use of chemical indicators to assist in tracking contaminant sources in estuaries which are used for oyster growing has now increased in the surveillance work undertaken by NSW Food Authority. The FWC test does have the potential to be used as a rapid initial field screening method to identify human-derived contamination in shoreline sanitary surveys, and the more expensive FSA assay has also been useful in confirming the presence of human sourced contamination in stormwater runoff. Where there are multiple lines of evidence, it can be assumed that where the traditional faecal microbiological results are high, the waters are likely to contain a contribution of human sourced faecal material.

The assistance of the NSW Food Authority and Great Lakes Council in this project work is gratefully acknowledged.