A3.6 Level of reference pathogen contamination in Australian source waters

Application of the QMRA framework for determining LRVs requires that the concentration of reference pathogens in source water is quantified or estimated.

Table A3.2 includes a summary of the prevalence (percentage of positive samples in any given study) and the range of reported concentrations among studies included from international reviews. From these reported studies, it is clear that pathogen concentrations are highly variable, spanning several orders of magnitude. This variability is not only between studies, but also within any given study over time. Estimation of representative pathogen concentrations for the source water categorisation is a challenge both because it is difficult to generalise between systems and because data from representative systems are limited.

Table A3.2 Summary of the range of reported concentrations of reference pathogens in surface water from international reviews⁽¹⁾

Pathogen	Prevalence (Percentage of samples positive)	Range of reported concentrations (Values reported as cited in publications and were calculated based on methods presented in original scientific papers)	Range of reported concentrations (Values reported as cited in publications and were calculated based on methods presented in original scientific papers)	Number of studies	Reviewed by
Bacteria		min	max
Campylobacter	0 - 86%	nr⁽²⁾	4.6 × 10⁶ CFU/L	19	Pitkanen and Hanninen 2017
Salmonella	3 - 100%	0.002 MPN/mL	1 ×10⁶ CFU/100 mL	21	Levantesi et al. 2012
Viruses
Adenovirus⁽³⁾	24.1 - 96%	1	1.38 ×10⁵ GC/L	12	Allard and Vantarakis 2017
Noroviruses	1 - 100%	9.4	2.70 ×104⁴ GC/L	11	Katayama and Vinje 2017
Rotavirus	13.9 - 100%	0.19	2.5 ×10⁷ GC/L	17	da Silva et al. 2016
Protozoa
Giardia	4 - 98.3 %	0.001	3.4 ×10⁴ cysts/L	43	Boarato et al. 2017
Cryptosporidium	5 - 95%	nr⁽⁴⁾	>1000ᵈ oocysts/L	38	Medema et al. 2009

Notes:

1. significant figures and units as reported in the review.

2. not reported in review.

3. by molecular methods.

4. value not reported, shown in a graph.

Abbreviations: GC Genome copies; CFU colony forming unit; MPN most probable number.

These results show the importance of having data of specific relevance to Australian conditions. Direct enumeration of pathogens in surface water is technically challenging and costly, and therefore few Australian datasets exist. Cryptosporidium data from Australian surface waters have been collated (Petterson and Deere 2013; Petterson et al. 2015). These data showed that average total immunofluoresence antibody positive Cryptosporidium concentration varied between systems from 2 $10^{-4}$ oocycst/L to nearly 10 oocycst/L. The higher concentrations were associated with smaller, unprotected systems and the observed E. coli concentrations provided some indication of Cryptosporidium concentrations. Unfortunately, no published data on concentrations of bacterial or viral agents in Australian surface waters have been identified for inclusion in this analysis. Water utilities may hold data, but this is not generally available for analysis by third parties. By publishing existing data, the accuracy of QMRA for drinking water in future revisions could be improved.

Other than for Cryptosporidium it is very difficult to make inferences about the pathogen concentration in source waters without local environmental Australian datasets. WHO has recommended assuming that the source water is contaminated with a certain percentage of wastewater as a starting point for estimating the LRV needed to address pathogen risks from a catchment (WHO 2011). The adenovirus (by culture) and Campylobacter concentrations expected for each catchment category were estimated by applying this principle and using E. coli as a surrogate to quantify the amount of raw sewage present in the source water. Point estimates (to one significant figure) of E. coli, Adenovirus (by culture) and Campylobacter in untreated sewage were calculated from reported concentrations in confidential Australian datasets reviewed by the Australian Guidelines for Water Recycling working committee while reviewing the 2006 Australian Guidelines for Water Recycling (NRMMC, EHPC and AHPC 2006). These were used as default 95ᵗʰ percentiles in those guidelines (summarised in Table A3.3). Resulting concentrations by category are summarised in Table A3.4.

To arrive at the concentrations in Table A3.4, order of magnitude estimates have been made relying on the assumptions outlined above. The only real alternative to these assumptions is the gathering of Australian data relevant to the catchment categories.

Table A3.3 Point estimates of E. coli, Adenovirus (by culture) and Campylobacter in untreated sewage

	Concentration Point estimate	Notes	Reference
E. coli (CFU/L)	1 × 10⁸	Estimated from range reported from Australian raw sewage samples	van den Akker et al. 2012; Sidhu et al. 2015
Cryptosporidium (infectious oocysts/L)	2 000	95ᵗʰ percentile based on Australian raw sewage data. Values selected for the upcoming revision of the Australian Guidelines for Water Recycling (AGWR)	Deere and Khan, 2016
Adenovirus (MPNIU/L)	8 000	95ᵗʰ percentile based on Australian raw sewage data. Values selected for the upcoming revision of the AGWR	Deere and Khan, 2016
Campylobacter (MPN/L)	7 000	95ᵗʰ percentile based on Australian raw sewage data. Values selected for the upcoming revision of the AGWR	Deere and Khan, 2016

Abbreviations: MPN most probable number; MPNIU most probable number infectious units.

Table A3.4 Assumed reference pathogen concentration by category for calculation of LRVs

Source water category	Maximum or 95ᵗʰ percentile⁽¹⁾ E. coli results from raw water monitoring (number/100 mL)	Indicative average Cryptosporidium concentration⁽¹⁾ (oocysts/L)	Adenovirus concentration⁽²⁾ (MPNIU/L)	Campylobacter concentration⁽²⁾ (MPN/L)
Category 1	<20 (WHO E. coli Band 1)	0.0	<0.016	< 0.014
Category 2	20 to 2000 (WHO E. coli Band 2)	0.01 – 0.1	0.016 – 0.16	0.014 – 0.14
Category 3	20 to 2000 (WHO E. coli Band 2)	0.1 – 1	0.16 – 1.6	0.14 – 1.4
Category 4	>2000 to 20,000 ⁽³⁾ (WHO E. coli Band 3)	1 - 10	1.6 - 16	1.4 - 14

Notes:

1. Maximum E. coli results from raw water representative of inlet to treatment plant should be used unless data set is robust enough to use 95ᵗʰ percentile.

2. Estimated from Australian monitoring data (Petterson and Deere 2013; Petterson et al. 2015).

3. Estimated using E. coli band and assuming E. coli sourced from raw sewage with concentrations as given in Table A3.3, MPN most probable number, MPNIU most probable number infectious units.

When allocating a microbial band for a water source the maximum E. coli result should be used (Walker et al. 2015) unless the data set is robust enough to use the 95ᵗʰ percentile. The selection of the most appropriate E. coli value for band allocation should be set in discussion with the relevant health authority or drinking water regulator. Further information on statistics for assessing performance can be found in Information Sheet 3.2 in Part IV of the Guidelines.

The absence of site-specific monitoring data for reference pathogens should result in a precautionary approach when developing LRVs. The upper concentration within each source water category should be used as a default until sufficient site-specific monitoring data can be generated. For example, if a catchment has been determined as Class 2 through the vulnerability classification, then the typical adenovirus concentration should be assumed to be 0.16 MPNIU/L and the typical Campylobacter concentration should be assumed to be 0.14 MPN/L. Variations to this approach should be discussed with the relevant health authority or drinking water regulator.