Content uploaded by Sarah Fiona Barker
Author content
All content in this area was uploaded by Sarah Fiona Barker on Jan 16, 2018
Content may be subject to copyright.
This article appeared in a journal published by Elsevier. The attached
copy is furnished to the author for internal non-commercial research
and education use, including for instruction at the authors institution
and sharing with colleagues.
Other uses, including reproduction and distribution, or selling or
licensing copies, or posting to personal, institutional or third party
websites are prohibited.
In most cases authors are permitted to post their version of the
article (e.g. in Word or Tex form) to their personal website or
institutional repository. Authors requiring further information
regarding Elsevier’s archiving and manuscript policies are
encouraged to visit:
http://www.elsevier.com/authorsrights
Author's personal copy
Pathogen reduction requirements for direct potable reuse in Antarctica:
Evaluating human health risks in small communities
S. Fiona Barker
a,b,
⁎, Michael Packer
c
, Peter J. Scales
d
, Stephen Gray
e
, Ian Snape
c
, Andrew J. Hamilton
f
a
Department of Resource Management and Geography, The University of Melbourne, Parkville, VIC, 3010 Australia
b
Department of Primary Industries Victoria, Parkville, VIC, 3052 Australia
c
Department of Sustainability, Environment, Water, Population and Communities, Australian Antarctic Division, 203 Channel Highway, Kingston, Tasmania 7050 Australia
d
Particulate Fluids Processing Centre, Department of Chemical and Biomolecular Engineering, The University of Melbourne, Parkville, VIC 3010 Australia
e
Institute for Sustainability Innovation, Victoria University, PO Box 14428, Melbourne, VIC, 8001 Australia
f
Department of Agriculture and Food Systems, The University of Melbourne, Dookie College, VIC, 3647 Australia
HIGHLIGHTS
•Direct potable reuse (DPR) projects
should consider population size.
•Small community pathogen load in out-
break sewage is higher (pb0.001) than
municipal.
•LRVs for municipal sewage: 6.9 (norovirus),
8.0 (giardia), 7.4 (Campylobacter).
•LRVs for small community: 12.1 (norovirus),
10.4 (giardia), 12.3 (Campylobacter).
•Additional treatment barriers required for
small community DPR to meet 10−6
DALYs.
GRAPHICAL ABSTRACT
abstractarticle info
Article history:
Received 15 April 2013
Received in revised form 20 May 2013
Accepted 20 May 2013
Available online xxxx
Editor: D. Barcelo
Keywords:
Campylobacter
Drinking water
Giardia
Small, remote communities often have limited access to energy and water. Direct potable reuse of treated
wastewater has recently gained attention as a potential solution for water-stressed regions, but requires fur-
ther evaluation specific to small communities. The required pathogen reduction needed for safe implemen-
tation of direct potable reuse of treated sewage is an important consideration but these are typically
quantified for larger communities and cities. A quantitative microbial risk assessment (QMRA) was
conducted, using norovirus, giardia and Campylobacter as reference pathogens, to determine the level of
treatment required to meet the tolerable annual disease burden of 10
−6
DALYs per person per year, using
Davis Station in Antarctica as an example of a small remote community. Two scenarios were compared: pub-
lished municipal sewage pathogen loads and estimated pathogen loads during a gastroenteritis outbreak. For
the municipal sewage scenario, estimated required log
10
reductions were 6.9, 8.0 and 7.4 for norovirus, giar-
dia and Campylobacter respectively, while for the outbreak scenario the values were 12.1, 10.4 and 12.3 (95th
percentiles). Pathogen concentrations are higher under outbreak conditions as a function of the relatively
Science of the Total Environment 461–462 (2013) 723–733
Abbreviations: DALYs, disability adjusted life years; DPR, direct potable reuse; IPR, indirect potable reuse; LRV, log
10
reduction values; QMRA, quantitative microbial risk assessment.
⁎Corresponding author at: Department of Resource Management and Geography, The University of Melbourne, Parkville, VIC 3010, Australia. Tel.: +61 3 8341 2413.
E-mail addresses: fionabr@unimelb.edu.au (S.F. Barker), Michael.Packer@aad.gov.au (M. Packer), peterjs@unimelb.edu.au (P.J. Scales), Stephen.Gray@vu.edu.au (S. Gray),
Ian.Snape@aad.gov.au (I. Snape), andrewjh@unimelb.edu.au (A.J. Hamilton).
0048-9697/$ –see front matter © 2013 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.scitotenv.2013.05.059
Contents lists available at SciVerse ScienceDirect
Science of the Total Environment
journal homepage: www.elsevier.com/locate/scitotenv
Author's personal copy
greater degree of contact between community members in a small population, compared with interactions in
a large city, resulting in a higher proportion of the population being at risk of infection and illness. While the
estimates of outbreak conditions may overestimate sewage concentration to some degree, the results suggest
that additional treatment barriers would be required to achieve regulatory compliance for safe drinking
water in small communities.
© 2013 Elsevier B.V. All rights reserved.
1. Introduction
Small remote communities sometimes struggle to adequately
meet basic services such as power and water. In Australia, for exam-
ple, there are many small remote communities. This is exemplified
by the many remote indigenous communities, with nearly 13% of
people living in the 838 communities with a population of less than
50 people and a significant number in communities with between
50 and 199 residents (ABS, 2008). More than half of the people living
in remote indigenous communities rely on bore water as their main
water source, 62% rely on community generators for electricity, only
30% are connected to a town sewerage system while 28% and 3.2%
use septic tanks or pit toilets, respectively and high proportions of
people experience interruptions in supply of services (ABS, 2008).
In some of these communities, where water scarcity is an issue of
concern, alternative sources of water may be needed. While recent
droughts in Australia were accompanied by a drastic rise in the do-
mestic use of grey water (ABS, 2007a, 2010a, 2010b), alternative
sources of potable water have received less attention.
Indirect potable reuse schemes for the recycling of wastewater
(IPR is the discharge of treated water into a receiving body prior
to extraction and re-treatment for potable use) can be found in
many countries; however, direct potable reuse (DPR is reuse with-
out environmental mixing) is rare. There are currently only three
DPR schemes in the world: Windhoek in Namibia (Lahnsteiner
and Lempert, 2007), Cloudcroft in New Mexico and Big Springs in
Texas (Tchobanoglous et al., 2011). While the more immediate
driver of DPR is extreme water scarcity, various other factors also
favor DPR systems, including whole-of-system life-cycle costs, reli-
ability of water supply and quality and the exhaustion of econom-
ically feasible non-potable reuse options (Leverenz et al., 2011). An
important consideration for system design and operation is the im-
pact of population size on disease outbreaks, sewage quality and
ultimately the required level of treatment. A greater understanding
of these impacts is needed before the technology is implemented
broadly.
Quantitative microbial risk assessment (QMRA) is a useful tool to
assess pathogen reduction requirements for wastewater recycling
and has been used to inform the regulatory environment relevant to
wastewater schemes for non-potable reuse, IPR and DPR scenarios
(NRMMC et al., 2006b; NRMMC et al., 2008; NRMMC et al., 2009;
WHO, 2006). Reuse guidelines are usually based on water quality
characteristics of municipal sewage from large cities and, using a tol-
erable annual disease burden of ≤10
−6
disability adjusted life years
(DALYs) per person per year, QMRA has been used to inform guide-
lines where recommended pathogen log
10
reduction values (LRV)
are presented (NRMMC et al., 2008). Municipal sewage is typically
of consistent or relatively stable quality, as a function of the dilution
effect from a large population base (NRMMC et al., 2008), although
differences between peak and non-peak seasons may be detectable;
for example norovirus concentrations in sewage may be up to 1 or 2
logs units higher during peak season (Katayama et al., 2008;
Nordgren et al., 2009; Victoria et al., 2010). Localized disease out-
breaks and changes in population size may significantly alter sewage
microbial quality from a small population, potentially affecting treat-
ment requirements.
The objective of this study was to determine the required LRVs for
DPR in small communities as this has not been specifically considered
in reuse guidelines. While any of a number of small remote communi-
ties could have been chosen as a representative population for the
model, Davis Station, the largest of three permanent Australian research
stations in Antarctica, was selected for this exercise as there is current
interest in DPR. The Australian Antarctic Division is undertaking a pro-
ject to reduce the environmental impact of sewage treatment and
disposal at Davis Station. As part of this project, research is being
conducted into the potential implementation of DPR which, in addition
to providing a reliable potable water supply, could provide considerable
energy savings as compared with the existing water system. While
Davis Station may not be a typical small community, only minor modi-
fications (volume of drinking water and days of exposure) would be re-
quired to adequately reflect other populations. Regardless, the results of
this assessment were considered generalizable to a range of other small
communities, of which there are many in Australia and around the
world.
2. Methods
The focus of this model was human health risks from waterborne
pathogens, in particular diarrheal diseases, from ingestion of treated
drinking water. Two complementary approaches were employed to
estimate sewage pathogen concentrations: published values from
municipal sewage treatment plants and estimated gastroenteritis
outbreak conditions. Further detail is provided in supplementary
materials.
2.1. QMRA
The QMRA method was used to determine required LRVs for direct
potable reuse of wastewater starting from a health target—atolerable
annual burden of disease (DB)of≤10
−6
DALYs person
−1
year
−1
—that
has been widely adopted for both drinking water and non-potable
reuse (NRMMC et al., 2006b; WHO, 2006; WHO, 2011). All model
input parameters are listed in Table 1. Using the annual burden of
disease calculation
DB ¼PillBSf;ð1Þ
the tolerable annual probability of illness (P
ill
) was determined,
where Bis the disease burden (DALYs per case of illness) and S
f
is
the proportion of the population susceptible to the disease.
While country-specific estimates of disease burden (B)arepre-
ferred, they are often non-existent. In this model, published values
from a range of countries were used. For norovirus, a Uniform distribu-
tion (Cressey and Lake, 2009; Haagsma et al., 2008; Kemmeren et al.,
2006; Lake et al., 2010; Masago et al., 2006) was used to represent the
range of available values and similarly using Dutch data for giardia
(Havelaar, 2012; Vijgen et al., 2007)andCampylobacter (Havelaar,
2012; Havelaar and Melse, 2003).
Disease susceptibility (S
f
) is used to exclude the proportion of the
population shown to be resistant to infection. There is evidence of re-
sistance to norovirus infection (Johnson et al., 1990; Lindesmith et al.,
2003; Teunis et al., 2008) related to both histo-blood group antigens
and secretor status (Le Pendu, 2006) although it has been suggested
Norovirus
Quantitative Microbial Risk Assessment
(QMRA)
Sewage
724 S.F. Barker et al. / Science of the Total Environment 461–462 (2013) 723–733
Author's personal copy
Table 1
Model input parameters.
Parameter Units Distribution or point estimates
a
, [mean
b
] References and justification
Disease burden (B) DALYs case of illness
−1
Norovirus Uniform (3.71 × 10
−4
, 6.23 × 10
−3
), [3.30 × 10
−3
](Cressey and Lake, 2009; Haagsma et al., 2008;
Kemmeren et al., 2006; Lake et al., 2010; Masago et al., 2006)
Giardia Uniform (2.10 × 10
-3
, 2.68 × 10
−3
), [2.39 × 10
−3
](Havelaar, 2012; Vijgen et al., 2007)
Campylobacter Uniform (4.60 × 10
−3
, 4.10 × 10
−2
), [2.28 × 10
−2
](Havelaar, 2012; Havelaar and Melse, 2003)
Susceptibility fraction (S
f
) proportion
Norovirus Uniform (0.8, 1.0), [0.9] (Atmar, 2010; Denborough and Downing, 1968;
Soller et al., 2010; Thorven et al., 2005)
Giardia, Campylobacter 1
Exposure events (n) days year
−1
Uniform (62, 121), [91.5] Total number of days for months with population
≥30 (AAD, 2011) between 2005 and 2010
Dose–response models
Norovirus (a + b inoculum) Full beta-Poisson: α
NV
= 0.04, β
NV
= 0.055,
ŋ
NV
= 0.00255, r_NV = 0.086, a
NV
= 0.9997 (Teunis et al., 2008)
Giardia Exponential: r_G = Triangular(0.0044, 0.0566, 0.0199), [0.027] (Teunis et al., 1996); min/max are 95th confidence intervals
Campylobacter Full Beta-Poisson: α= 0.024, β= 0.011,
ŋ
C
= 3.63 × 10
−9
,r_C = 2.44 × 10
8
(Teunis et al., 2005)
Giardia infection:illness
(inf:ill)
proportion Uniform (0.24, 0.93), [0.58] (Birkhead and Vogt, 1989; Hoque et al., 2002;
Lopez et al., 1980; Yakoob et al., 2010)
Daily water consumption (V) L person
−1
Lognormal (3, 1) —truncated at 2 and 6; μ= 1.05, δ= 0.32 (Hunter et al., 2011; Roche et al., 2012;
Schijven et al., 2011; USEPA, 2004; USEPA, 2006)
Sewage concentration —
municipal sewage (c
sewage
)
Norovirus PCR units L
−1
Mixture (A, B), [3.12 × 10
6
];
A = Lognormal(2.19 × 10
6
, 2.60 × 10
6
); μ= 14.2, δ= 0.94,
B = Lognormal(4.06 × 10
6
, 6.27 × 10
6
); μ= 14.6, δ= 1.11
11.1% recovery efficiency (Katayama et al., 2008) applied
to A & B (Katayama et al., 2008)(Haramoto et al., 2006)
Giardia cysts L
−1
Mixture (G1, G2, G3), [2.51 × 10
3
];
G1 = 10^Normal(2.90, 0.56),
G2 = 10^Normal(2.94, 0.77),
G3 = 10^Normal(2.57, 0.72)
(Van Den Akker et al., 2011)
recovery included in values (32-47%)
Campylobacter cfu L
−1
Lognormal(1.90 × 10
3
, 5.00 × 10
3
); μ= 6.51, δ= 1.44 (NRMMC et al., 2006a)
Station population (P) # people Discrete distribution (min = 51, max = 106), [72] Daily station population in months with population ≥30;
data from 2005-2011, n = 601 (AAD, 2011)
Secondary attack rate (A
r
) proportion
Norovirus Uniform (0.14, 0.22), [0.18] (Alfano-Sobsey et al., 2012; Baron et al., 1982; Götz et al., 2002;
Johansson et al., 2002; ter Waarbeek et al., 2010)
Giardia Uniform (0.17, 0.18), [0.175] (Katz et al., 2006; Pickering et al., 1981)
Campylobacter Uniform (0, 0.15), [0.075] (Evans, 1996; Norkrans and Svedhem, 1982; Porter and Reid, 1980)
Peak shedding rate
Norovirus (S
NV
) copies g-feces
−1
Uniform (2.9 × 10
10
, 1.6 × 10
12
), [8.2 × 10
11
](Atmar et al., 2008; Chan et al., 2006; Lee et al., 2007)
Giardia (S
G
) cysts person
−1
day
−1
Uniform (6.42 × 10
8
, 7.05 × 10
8
), [6.73 × 10
8
](Tsuchiya, 1931)
Campylobacter (S
C
) cfu g-feces
−1
Uniform (10
4
,10
9
), [5 × 10
8
](Feachem et al., 1983; Lin et al., 2008)
Daily diarrheal fecal weight (F) g-feces person
−1
Uniform (200, 750), [475] (Rao, 2006)
Daily water use (W) L person
−1
day
−1
Uniform (90, 174), [132] Davis Station between 2010 and 2011 (AAD, 2011; AAD, 2012)
a
Distributions: Lognormal(mean, sd), values from 1,000,000 iterations, population parameters μand δcalculated as follows: μ= ln(x)−0.5ln(1 + s
2
/x
2
), δ= [ln(1 + (s
2
/x
2
))]
1/2
, where xis the sample mean and s
2
the sample standard
deviation; mixture is a set of random values drawn from each distribution with equal weighting; normal (mean, sd); triangular (min, max, mode/most likely); uniform (min, max).
b
Mean of 1,000,000 iterations (for information purposes only).
725S.F. Barker et al. / Science of the Total Environment 461–462 (2013) 723–733
Author's personal copy
that, due to the variation between norovirus genotypes, every person
may be genetically susceptible to at least one norovirus genotype
(Atmar, 2010). Since susceptibility to norovirus is uncertain, S
f
was
represented by a Uniform distribution accounting for a range from
secretor-positive individuals (0.8; Denborough and Downing, 1968;
Thorven et al., 2005) through to all individuals (1.0). Despite many
years of research, there remain many questions about the mecha-
nisms of pathogenicity, host responses to infection and immunity to
giardia infections (Roxström-Lindquist et al., 2006); therefore, in
this work, all individuals were assumed susceptible (S
f
= 1). No in-
formation on susceptibility to Campylobacter was found so the same
assumption was made.
To estimate the tolerable daily probability of illness (p
ill
), the orig-
inal equation for annual probability of illness (WHO, 2006) was used
such that
Pill ¼1−1−pill
ðÞ
n
;ð2Þ
for nexposure events (days year
−1
). In the model, the summer peri-
od (months where population >30) was assumed to be the period of
exposure (due to the movement of people to and from the station)
and was represented by a Uniform distribution determined from
Davis Station records between 2005 and 2010. The tolerable daily
probability of infection (p
inf
) was determined using published dose–
response models for norovirus (Teunis et al., 2008), giardia (Teunis
et al., 1996) and campylobacter (Teunis et al., 2005). Full details of
dose–response models and determination of tolerable dose are pro-
vided in supplementary materials.
The tolerable pathogen concentration in treated drinking
water (c
tolerable
; organisms L
−1
) was estimated from the exposure
model,
λ¼ctolerableV;ð3Þ
using the estimated tolerable dose (λ) and the daily per capita
water consumption (V;Lperson
−1
day
−1
). Per capita water con-
sumption at Davis Station is much higher than that of the general
population (typically assumed to be 2 L day
−1
) as humidity is
very low in Antarctica. Some community members have indicated
they drink much more than the recommended 4 L, with consump-
tion of up to 6 L per day considered quite reasonable. Variability in
drinking water consumption was represented using a lognormal
distribution (Ãstrom et al., 2007; Pintar et al., 2012; Schijven et
al., 2011) with a mean daily drinking water consumption of 3 L.
In studies with mean daily drinking water consumption greater
than 1 L (Table S.1), standard deviations ranged from 0.8 to 1.2;
therefore, the middle value (1.0) was chosen to represent variation
and the distribution was truncated at the likely minimum and
maximum values (2 and 6 L).
Finally, the required log
10
reduction value (LRV) in sewage, neces-
sary to meet tolerable drinking water quality, was calculated as
LRV ¼log10 cðÞ−log10 ctolerable
ðÞ;ð4Þ
where the pathogen concentrations in sewage (c) were estimated
using two different methods: 1) published values of pathogen con-
centrations in municipal wastewater and 2) estimates of sewage
pathogen concentrations during a gastroenteritis outbreak at
Davis Station. There was no available information on concentra-
tions of pathogens or indicator organisms in raw sewage at Davis
Station.
Norovirus, giardia and Campylobacter concentrations in municipal
wastewater (c
sewage
;#L
−1
) were assumed to follow a Lognormal distri-
bution, with values drawn from published literature (refer to supple-
mentary materials). An estimate of outbreak conditions at Davis
Station was developed, with an outbreak defined as the arrival of one
infected person. Outbreak sewage pathogen concentrations (c
o
;#L
−1
)
were estimated using the following equations
co¼1þPAr
ðÞS
WP ;ð5Þ
S¼SNVFor S¼SCF;ð6Þ
where Pisthepopulationonagivensummerday,A
r
is the sec-
ondary attack rate (proportion), Sis the peak daily pathogen
shedding rate ( person
−1
day
−1
), Wis the per capita water use
(L person
−1
day
−1
), S
NV
and S
C
are the norovirus and Campylobacter
shedding concentrations (# g feces
−1
)andFis the daily diarrheal ex-
cretion rate (g feces person
−1
day
−1
).
To represent the summer population (P), months were selected
where the minimum number of people on station was >30, and
daily population values (n = 601) were used as a discrete distribu-
tion, using data from 2005 to 2011. Daily per capita water use (W)
was determined from monthly average population and monthly
total station water use during summer months (2010–2011; AAD,
2012), with the variation represented by a Uniform distribution.
The secondary attack rate (A
r
) is the proportion of people who,
after contact with the original infected person, become ill (typically
measured as the number of symptomatic cases). A
r
was used to esti-
mate the maximum number of people who might be ill at one time
(post-arrival of the one infected person), making the unrealistic
(highly conservative) assumption that all infections occurred instan-
taneously (rather than over a period of days or weeks). Uniform dis-
tributions were used to represent the range of published values for
secondary attack rate. Various studies have reported secondary
norovirus attack rates between 0.14 and 0.22 over periods of up to
14 days after the first reported case (Alfano-Sobsey et al., 2012;
Baron et al., 1982; Götz et al., 2002; Johansson et al., 2002; ter
Waarbeek et al., 2010). Two studies reported very similar secondary
attack rates for giardia (Katz et al., 2006; Pickering et al., 1981)
while a wide range (0 to 0.15) was reported for Campylobacter
(Evans, 1996; Norkrans and Svedhem, 1982; Porter and Reid, 1980).
Shedding rates (S) were also represented by Uniform distribu-
tions. The only known study of giardia shedding rates (S
G
; cysts per-
son
−1
day
−1
) was conducted with two infected individuals over a
period of 7 weeks (Tsuchiya, 1931) and the maximum shedding
rate from each participant was used to define the range of peak
daily shedding rates. Lin et al. (2008) reported viable Campylobacter
counts in feces (CFU g
−1
) from 10 samples while Feachem et al.
(1983) reported counts as high as 10
9
per g feces (minimum and
maximum values used to define the distribution). Three studies
(Atmar et al., 2008; Chan et al., 2006; Lee et al., 2007) reported a
range of norovirus shedding concentrations (S
NV
; copies g-feces
−1
)
and the maximum value from each of the four sets of data was used
to define the distribution. For both norovirus and Campylobacter,
a uniform distribution (# g-feces
−1
) was converted to shedding
rate using an estimate of daily diarrheal fecal weight (F; g person
−1
day
−1
). Individuals suffering from diarrhea are typically defined as
having a daily stool weight in excess of 200 g and a recent study
reported mean stool weights of 750 g in persons with diarrhea
(Rao, 2006); a uniform distribution was used to represent fecal
weights for ill individuals, making the assumption that all infected in-
dividuals have diarrhea (secondary attack rate counts only people
who are symptomatic).
2.2. Population size
The premise of this model is that small communities need to be
considered differently to large cities, with the assumption that out-
break conditions will be significantly different to those in a large
city as a function of the relatively greater degree of contact between
726 S.F. Barker et al. / Science of the Total Environment 461–462 (2013) 723–733
Author's personal copy
community members in a small population and the greater level of
dilution in a municipal sewage treatment plant due to the large pop-
ulation served (NRMMC et al., 2008). The estimate of municipal sew-
age concentrations reflects “average”conditions in a large city while
outbreak sewage concentration was estimated assuming that the
community at Davis Station operates in a similar fashion to the con-
fined populations assessed to determine secondary attack rates
(assuming a high degree of contact between all community members).
The difference in sewage concentrations during outbreaks in small or
large communities is a function of the proportion of the population
infected. To evaluate the impact of population size, a method was devel-
oped to estimate the likely sewage concentration, and therefore required
log
10
reduction, during a norovirus outbreak in a large city. Norovirus
was selected as the reference pathogen as the required epidemiological
data were available.
In Australia, there are 0.92 cases of gastroenteritis per person per
year (Hall et al., 2006) of which 10.7% are caused by norovirus
(Sinclair et al., 2005). If all norovirus infections occurred simulta-
neously (which is highly improbable), then 9.8% of the population
would be infected (~0.098 cases of novovirus infection per person
per year). A more realistic scenario can be developed using the results
of a Melbourne study of 600 households that reported a maximum of
2.5% of households with at least one case of norovirus per month
(Sinclair pers. comm.; Sinclair et al., 2005), assuming that monthly
incidence rates equate to outbreaks. Assuming four people per
household, 1.4 people would be infected per household event
(average value reported by Sinclair et al., 2005), and using the current
Melbourne population of 4,137,432 (ABS, 2012), an estimated 36,203
people would be infected during an outbreak, or ~0.88% of the popu-
lation. Applying this monthly infection rate across a whole year, there
would be 0.105 cases of norovirus per person per year which is
consistent with the estimated value above (0.098) and therefore a
norovirus outbreak in Melbourne was conservatively assumed to in-
fect 1% of the population. The following scenarios were compared to
evaluate the magnitude of the effect of population size: municipal
sewage (“average”city conditions), outbreak conditions in a large
city (population >1 million) and outbreak conditions at Davis
Station.
2.3. Method comparisons
The model presented herein uses a different approach to that taken
by regulatory bodies. For example, a stochastic approach was used here
to account for variability and uncertainty in the model while the Austra-
lian Guidelines for Water Recycling —Augmentation of Drinking Water
Supplies (NRMMC et al., 2008) use a deterministic approach, conceding
that stochastic analyses may provide a better understanding of uncer-
tainty and variability where sufficient data is available. In our model,
norovirus was chosen as the reference pathogen for viruses, giardia
for protozoans and Campylobacter for bacteria, while the Guidelines
use adenovirus measurements with the rotavirus dose–response
model for viruses, cryptosporidium for protozoans and Campylobacter
for bacteria. In addition, daily per capita drinking water consumption
was much higher to reflect conditions at Davis Station. The differences
in methods between the model used herein and that described in the
Guidelines are outlined in Table S.2. The Guideline method and input
parameters were used and then individual parameters were changed
sequentially (detailed in Table 2, Table S.4 and Table S.5) to evaluate
the impact of each change on the model output (required LRVs).
2.4. Sensitivity analysis
A sensitivity analysis, using Spearman rank order correlation coef-
ficients, was conducted using values from the first 1000 random
draws of each input distribution to identify those input parameters
that had the greatest influence on the uncertainty of the model out-
put. Input distributions were assessed to ensure there was no correla-
tion between unrelated variables and then relevant input parameters
were tested against the final model output (LRV). To further evaluate
the impact of variation of input parameters on the magnitude of re-
quired LRVs, the model was run with key inputs set at discrete per-
centile values (5th, 50th and 95th), with no other alteration to the
model; median required LRVs were reported.
2.5. Model structure and implementation
For all input parameters, a set of random values (n = 1,000,000)
was drawn from the distribution and used for all model calculations.
For all model outputs, the median and 90% confidence intervals were
reported. Confidence intervals were estimated using the percentile
method (Buckland, 1984) and values are reported as follows: 50th
[5th, 95th]; single values are 95th percentile values unless otherwise
indicated. Statistical differences were determined from the first
10,000 random draws from each output distribution using analysis
of variance (ANOVA) and comparison of means using Tukey's HSD
(Honestly Significant Difference) test. Differences were considered
significant at p ≤0.05. All modeling and analyses were performed
in ‘R’version 2.12.2 (The R Foundation for Statistical Computing,
2011) and some distribution fitting was conducted in @Risk (version
5.7).
3. Results
Estimates of norovirus, giardia and Campylobacter concentrations
in municipal sewage (1 × 10
7
,9×10
3
and 7.2 × 10
3
#L
−1
, respec-
tively) were significantly lower (p b0.0001) than those determined
for Davis Station outbreak conditions, 1.4 × 10
12
, 1.4 × 10
6
and
4.9 × 10
8
(Fig. 1), which had a direct effect on the required LRVs.
Table 2
Estimated required enteric virus log
10
reduction values (LRVs) for stepwise methodological changes from the Guideline method (NRMMC et al., 2008) to a deterministic approx-
imation of the model using municipal sewage concentrations.
Step LRV Model input parameters
a
Vc B S
f
inf:ill d–rn
1. 9.4 2 8000 1.3 × 10
−2
(RV) 0.06 (RV) 0.88 RV
b
365
2. 12.5 2 1.02 × 10
7
(95th NV) 1.3 × 10
−2
(RV) 0.06 (RV) 0.88 RV
b
365
3. 9.8 4.8 (95th AAD) 8000 1.3 × 10
−2
(RV) 0.06 (RV) 0.88 RV
b
365
4. 7.2 2 1.02 × 10
7
(95th NV) 5.94 × 10
3
(95th NV) 0.99 (95th NV) NV NV
c
365
5. 7.6 4.8 (95th AAD) 1.02 × 10
7
(95th NV) 5.94 × 10
3
(95th NV) 0.99 (95th NV) NV NV
c
365
6. 6.9 2 1.02 × 10
7
(95th NV) 5.94 × 10
3
(95th NV) 0.99 (95th NV) NV NV
c
118 (95th AAD)
7. 7.3 4.8 (95th AAD) 1.02 × 10
7
(95th NV) 5.94 × 10
3
(95th NV) 0.99 (95th NV) NV NV
c
118 (95th AAD)
a
Model input parameters: V= daily water consumption (L person
−1
), c= sewage pathogen concentration (# L
−1
), B= disease burden (DALYs case
−1
), S
f
= susceptibility
fraction, inf:ill = ratio of infection to illness, d–r = dose–response model, n= days of exposure per year. 95th refers to 95th percentile of the input distribution. AAD = Davis Sta-
tion data. NV = norovirus, RV = rotavirus.
b
Simplified approximate beta-Poisson.
c
Full beta-Poisson.
727S.F. Barker et al. / Science of the Total Environment 461–462 (2013) 723–733
Author's personal copy
The required LRVs to meet the ≤10
−6
DALYs person
−1
year
−1
health
target, for potable reuse of treated wastewater at Davis Station, were
6.9 for norovirus, 8.0 for giardia and 7.4 for Campylobacter using esti-
mates of municipal sewage, while for the Davis Station outbreak sce-
nario they were 12.1, 10.4 and 12.3 respectively (Fig. 2).
The estimate of norovirus concentration in municipal sewage
(1 × 10
7
L
−1
) was similar (within 1 order of magnitude) to many pre-
viously reported maximum sewage concentrations in Japan, UK, Italy,
Finland, Germany, Sweden, Singapore and the Netherlands (Aw and
Gin, 2010; Haramoto et al., 2006; Katayama et al., 2008; La Rosa et al.,
2010; Laverick et al., 2004; Nordgren et al., 2009; Pusch et al., 2005;
Van Den Berg et al., 2005; VonBonsdorff et al., 2002), while the estimate
of outbreak concentration (1 × 10
12
L
−1
) was 5 orders of magnitude
higher. Similarly, the estimate of giardia concentration in municipal
sewage (9 × 10
3
L
−1
) was within 1 order of magnitude of most of the
previously reported maximum sewage concentrations in Japan, the
Netherlands, Spain, Sweden and the USA (Castro-Hermida et al., 2008;
Castro-Hermida et al., 2010; Gassmann and Schwartzbrod, 1991;
Medema and Schijven, 2001; Oda et al., 2005; Ottoson et al., 2006a;
Ottoson et al., 2006b; Sykora et al., 1991) while the estimate of giardia
outbreak concentration (1.4 × 10
6
L
−1
) was 3 orders of magnitude
higher. The estimate of Campylobacter concentration in municipal
sewage (7.2 × 10
3
cfu L
−1
) was similar to published values from Italy
and Spain (Rodríguez and Araujo, 2010; Stellacci et al., 2010), but
lower (by as much as 2 orders of magnitude) than published concentra-
tions in Germany and the Baltic Sea region (Holler, 1988; Rechenburg
and Kistemann, 2009). The estimate of outbreak concentration
(4.9 × 10
8
) was up to 5 orders of magnitude higher than municipal
sewage estimates.
The situation considered here is a worst case scenario where raw
wastewater is not diluted with other wastewater sources (stormwater,
rainwater, etc.). Each of the different scenarios and estimation methods
had a significant effect (p b0.001) on the estimated sewage pathogen
concentrations and subsequently the required LRVs. To evaluate the im-
pact of population size on required LRVs, an epidemiological method
was developed to estimate norovirus concentrations in Melbourne sew-
age during an outbreak. Melbourne outbreak sewage concentration
(7.2 × 10
10
#L
−1
) was nearly 4 orders of magnitude greater than mu-
nicipal sewage (1.0 × 10
7
#L
−1
) and ~1 order of magnitude less than
Davis Station outbreak concentration (1.4 × 10
12
#L
−1
), requiring
10.8 compared with 12.1 LRVs for Davis Station (Fig. 3).
The Guidelines recommend a minimum enteric virus LRV of 9.5 for
the production of drinking water from sewage while the model, using
municipal sewage pathogen concentrations, determined a LRV of 6.9
Fig. 1. Cumulative probability distributions of estimated sewage concentrations under two scenarios: municipal sewage and estimated Davis Station outbreak sewage. The circles
are the 50th percentiles and dashed vertical lines are the 95th percentiles.
Fig. 2. Cumulative probability distributions of required log
10
reductions (LRVs) for direct potable reuse of wastewater under two scenarios: municipal sewage and estimated Davis
Station outbreak sewage. The circles are the 50th percentiles and dashed vertical lines are the 95th percentiles.
728 S.F. Barker et al. / Science of the Total Environment 461–462 (2013) 723–733
Author's personal copy
for norovirus. To compare these two methods, sequential steps from
the Guideline method to a deterministic approximation of the model
are reported in Table 2. The difference in LRVs between steps 2 and 4
shows that the full norovirus dose–response model reduces the re-
quired LRV from 12.5 with the rotavirus dose–response model to 7.2;
this is likely the primary contributing factor to the difference between
the Guideline value and the model value, although the higher virus con-
centration was also important (increased the LRV from 9.4 to 12.5). The
difference between steps 4 and 5 shows the impact of using the higher
drinking water volume (7.2 to 7.6) and the difference between steps 5
and 7 shows the impact of a shorter exposure period (7.6 to 7.3);
none of these changes greatly altered the final model output. Compar-
ing the 95th percentile of the full stochastic model (6.9) with a deter-
ministic approximation of the model (step 7; 7.3), the difference is
small, demonstrating that the understanding gained from the stepwise
evaluation of parameter changes can be applied to the full model. A sim-
ilar step-wise process wasconducted for the other reference pathogens
and results are presented in supplementary materials (Tables S.4 and
S.5). The impact of the full stochastic model had much less impact on
the LRVs for giardia and Campylobacter.
An assessment of all input parameters confirmed that there were no
unexpected relationships or correlations and variation in many of the
input parameters contributed significantly to the variation in the
model outputs (Table 3). Using the municipal sewage method, sewage
concentration had the largest impact on variation in the estimate of
required LRVs, while drinking water volume, disease burden and
exposure period contributed smaller amounts. Exposure period
did not affect Campylobacter, while for giardia the dose–response
parameter (r) and the infection to illness relationship also made
significant contributions to variation. The outbreak scenario meth-
od was similar for norovirus and Campylobacter, with the greatest
effect on variation in LRV due to variation in the estimate of sewage
concentration which was a function of the other input parameters.
Pathogen shedding rate contributed the most to the variation in
LRVs for norovirus and Campylobacter, followed by fecal weight,
disease burden, volume of drinking water and daily per capita
water use. Secondary attack rate was also a significant contributor
for Campylobacter. The variation in required LRVs for giardia was
somewhat different and largely influenced by the variation in the
dose–response parameter and illness to infection ratio, followed
by drinking water volume, exposure period and daily per capita
water use.
Similar trends were observed in the impact on LRVs when input
parameters were fixed at discrete percentile values (Fig. 4). For mu-
nicipal sewage scenarios, median LRVs were most affected by the
variation in the estimate of sewage concentration, with the spread
in estimated LRVs as high as 2.3 log
10
for giardia. For outbreak sew-
age scenarios, median LRVs were most affected by pathogen shed-
ding rate for norovirus and Campylobacter with a difference in
LRVs as large as 1.3 log
10
(Campylobacter). The effect of input pa-
rameter variation on LRVs for giardia was minimal for outbreak
conditions.
4. Discussion
While there have been recent arguments that the 10
−6
DALY
threshold is too conservative, even for developed countries with
lower background levels of water-borne disease (Mara, 2011; Mara
et al., 2010), the more cautious approach appears sensible in the con-
text of small communities where, as a result of isolation, the implica-
tions of illness may be much greater. Using the 10
−6
DALY health
target, required LRVs were calculated to be 6.9, 8.0 and 7.4 for
norovirus, giardia and Campylobacter using municipal sewage values
and 12.1, 10.4 and 12.3 for estimated Davis Station outbreak condi-
tions, compared with 9.5, 8.0 and 8.1 reported in the Guidelines
(NRMMC et al., 2008). Using municipal sewage concentrations, the
LRVs for giardia and Campylobacter were very similar to the Guideline
values while the LRV for norovirus was much lower, largely due to the
difference between the rotavirus and norovirus dose–response
models.
Under outbreak conditions, LRVs were much higher than Guideline
values as a direct result of the much higher sewage pathogen concentra-
tions (3–5 orders of magnitude greater) estimated for Davis Station
Fig. 3. Cumulative probability distributions of required log
10
reductions (LRVs) for direct
potable reuse of wastewater comparing estimate methods: 1) municipal sewage, 2) out-
break sewage—Davis Station and3) outbreak sewage —largecity. The circlesare the 50th
percentiles and dashed vertical lines are the 95th percentiles.
Table 3
Spearman's rank order correlation coefficients for required log
10
pathogen reductions.
Pathogen Method Model input parameters
a
c
sewage
VS
f
n B r inf:ill P A
r
SFW
Norovirus Municipal 0.90
b
0.22
b
−0.04 0.09
b
0.28
b
n/a n/a n/a n/a n/a n/a n/a
Outbreak 0.88
b,c
0.24
b
−0.02 0.15
b
0.32
b
n/a n/a −0.001 0.09
b
0.75
b
0.41
b
−0.21
b
Giardia Municipal 0.86
b
0.20
b
n/a
d
0.16
b
0.08
b
0.27
b
0.23
b
n/a n/a n/a n/a n/a
Outbreak 0.26
b,c
0.34
b
n/a 0.32
b
0.12
b
0.66
b
0.50
b
−0.02 0.07
b
0.05 n/a −0.25
b
Campylobacter Municipal 0.99
b
0.07
b
n/a 0.00 0.14
b
n/a n/a n/a n/a n/a n/a n/a
Outbreak 0.93
b,c
0.23
b
n/a 0.09
b
0.25
b
n/a n/a −0.07
b
0.45
b
0.66
b
0.33
b
−0.16
b
a
Model input parameters: c
sewage
= estimated sewage pathogen concentration (# L
−1
), V= daily water consumption (L person
−1
), S
f
= susceptibility fraction, n= exposure
period (days year
−1
), n= exposure period (days year
−1
), B= disease burden (DALYs case
−1
), r= dose–response parameter for giardia, inf:ill = ratio of infection to illness for
giardia, P= station population, A
r
= secondary attack rate, S= peak pathogen shedding, F= daily fecal weight (g-feces person
−1
), W= daily water use (L person
−1
day
−1
).
b
p≤0.05.
c
Outbreak sewage pathogen concentration was calculated from some or all of the following inputs: station population, secondary attack rate, shedding rate, fecal weight, daily
water use and dose–response fit parameters. Its inclusion in the sensitivity analysis reflects the sum of variation contributed by the other model input parameters.
d
n/a = not applicable.
729S.F. Barker et al. / Science of the Total Environment 461–462 (2013) 723–733
Author's personal copy
outbreak conditions. These values, particularly norovirus, were orders
of magnitude higher than other published values of municipal sewage
pathogen concentrations, reporting peaks of 10
3
–10
7
for norovirus,
10
2
–10
4
for giardia and 10
4
–10
7
for Campylobacter (Table S.6). There
is very little information available on sewage pathogen concentrations
during community gastroenteritis outbreaks, although the Guidelines
use 95th percentile values assumedly to represent peak pathogen
loads that might occur during an outbreak. To further evaluate the
outbreak method, norovirus concentrations at Davis Station were
compared with estimated concentrations during an outbreak in
Melbourne. The proportion of people that become infected during a
Melbourne norovirus outbreak (1%) was much less than the secondary
attack rate (14–18%) used for the Davis Station outbreak scenario;
therefore, Melbourne sewage was more dilute (i.e. lower pathogen
concentration) and required 10.8 compared with 12.1 LRVs for Davis.
Assuming that the 95th percentile of the municipal concentration esti-
mate represents outbreak conditions, the median Melbourne outbreak
concentration (2.6 × 10
10
#L
−1
) was nearly 3 orders of magnitude
higher and may represent an overestimation of outbreak concentra-
tions. There are various possible explanations for this disparity in
concentration estimates: 1) the estimate of municipal sewage,
based on data from Japan, does not reflect Melbourne conditions
(i.e. norovirus rates in Japan are lower than in Melbourne); 2) the es-
timate of municipal sewage, based on monthly measurements,
missed outbreak conditions; 3) the outbreak method does not ac-
count for pathogen decay through the distribution system; or 4)
the outbreak sewage estimation method is too conservative. The im-
pact of each of these potential contributors cannot be quantified but
importantly, even if the outbreak method overestimates sewage
concentration, the required LRVs are still higher than those in the
Guidelines suggesting that additional treatment will be required. A
greater understanding of sewage pathogen concentrations from
small communities is needed to reduce the uncertainty around the
estimated LRVs.
Fig. 4. Median required log
10
reductions when individual input parameters were held at discrete values. Boxes represent values for 5th, 50th and 95th percentile input values (bot-
tom, middle and upper lines, respectively). The left hand side depicts municipal sewage scenarios and the right hand side depicts the Davis Station outbreak scenarios. Input pa-
rameters are defined as follows: c= sewage concentration, V= daily water consumption, n= exposure period, B= disease burden (DALYs case
−1
), Sf = susceptibility fraction,
r= dose–response parameter, inf:ill = ratio of infection to illness, P= station population, W= daily water, Ar = secondary attack rate, inf = number of people ill, SR = shed-
ding rate, f= daily fecal weight.
730 S.F. Barker et al. / Science of the Total Environment 461–462 (2013) 723–733
Author's personal copy
Various assumptions were made in the development of the model
that may be important constraints in the application of the model re-
sults. Secondary attack rate was used to estimate outbreak sewage
pathogen concentrations and is a measure of the spread of illness by di-
rect (person-to-person contact, inhalation of aerosols, etc.) and indirect
(transfer from contaminated surfaces, etc.) contact. Studies are typically
conducted in relatively confined populations such as households and
school camps. While there is evidence that pathogen shedding can
occur in the absence of symptoms (Atmar et al., 2008; Birkhead and
Vogt, 1989; Yakoob et al., 2010), the secondary attack rate accounts
for symptomatic cases only. Therefore, the model has not accounted
for asymptomatic infections that could contribute to the pathogen
load in sewage. This may be of limited concern, at least for norovirus,
as recent investigations have found that asymptomatic cases are unlike-
ly to cause transmission despite high shedding rates (Sukhrie, 2012).
We have also made highly conservative assumptions thatall individuals
became ill instantaneously and shed pathogens at the peak rate, and
that all infected or ill individuals had diarrhea. In an actual outbreak, it
is likely that the spread of infection would occur over a few weeks
(the time span of studies used to estimate secondary attack rate). At
the same time, pathogen shedding can occur for extended periods of
time –both prior to symptomatic illness and after apparent recovery
–and it would seem unlikely that peak shedding amongst all individ-
uals would occur simultaneously.
Careful consideration will be required to design a treatment plant
to meet safe drinking water requirements in the event of an outbreak
of gastroenteritis in a small community. The higher required LRVs for
norovirus, giardia and Campylobacter will demand a combination of
treatment systems. At Davis Station, a secondary treatment plant
will be installed to remove the majority of the wastewater contami-
nants, with additional tertiary and polishing treatment steps to
meet potable water quality requirements. The tertiary and polishing
processes of large scale indirect potable water systems generally con-
sist of ultrafiltration, reverse osmosis and advanced oxidation
followed by final disinfection. Such systems provide a multi-barrier
approach to ensure water quality and are required to achieve a
virus LRV of 9.5. Such processes can achieve higher LRVs (e.g. virus
LRV of 10 for Western Corridor in Brisbane, Australia), but neverthe-
less, the higher required LRVs for small scale treatment plants as sug-
gested by this model (e.g. an extra LRV of 2.6 for viruses) will
necessitate additional treatment units such as UV disinfection. The
higher protozoa and bacteria LRVs required for small systems also ne-
cessitate this extra treatment barrier.
In considering the higher required LRV requirements suggested by
this model, it is important to contextualize the risk of exposure to
treated wastewater relative to other forms of exposure. A small com-
munity such as Davis Station operates similar to a household in that
the level of contact between community members is quite high. The
potential exposure pathways include person-to-person contact, con-
tact with contaminated surfaces and inhalation/ingestion of aerosols.
The assumption of the model, that one infected person arrives at
Davis Station, would result in 18, 19 or 12 people sick with norovirus,
giardia or Campylobacter respectively, based on the secondary attack
rate (direct or indirect contact with the infected person). In contrast,
assuming all infected individuals are shedding pathogens at a peak
rate and that treatment of sewage conforms to the required LRVs
needed to meet the 10
−6
DALY health target, consumption of the
treated water would result in up to 17 cases of norovirus, 5 cases of
giardia or 2 cases of Campylobacter illness per 10,000 people or 0.18,
0.05 and 0.02 additional cases of norovirus, giardia and Campylobacter
per summer season (using 95th percentile station population).
While Davis Station may be considered an extreme example, a
similar approach could be applied to many small remote communi-
ties in Australia. In the Northern Territory alone, there are 41 pre-
dominantly indigenous communities (95% indigenous) that range in
size from 85 to 886 residents, with 13 of those communities having
a population under 200 (ABS, 2007b). Other reports have found that
of the 1,139 remote indigenous communities across Australia, more
than half (54%) reported less than 20 residents and 23% reported pop-
ulations of 20 to 49 (ABS, 2003). DPR may be an appropriate solution
in some of these communities and the results of this model demon-
strate the importance of consideration of small communities in deter-
mining appropriate treatment trains.
5. Conclusion
Direct potable reuse is a relatively new concept that has legiti-
mate potential to enhance water security in both small and large
communities. This analysis has highlighted the need to consider
population size and vulnerability when assessing treatment require-
ments, a conclusion based on a quantitative microbial risk assess-
ment (QMRA) that was conducted using norovirus, giardia and
Campylobacter as reference pathogens. Two scenarios were com-
pared, municipal sewage pathogen loads and potential pathogen
loads during a community gastroenteritis outbreak, and pathogen
concentrations were significantly higher (p b0.001) in the outbreak
scenario. For the municipal sewage scenario, required LRVs were 6.9,
8.0 and 7.4 for norovirus, giardia and Campylobacter respectively,
while for outbreak conditions, the values were 12.1, 10.4 and 12.3.
While the outbreak values could overestimate LRVs by as much as
3 (for norovirus), they still indicate a need for additional treatment
barriers for small communities in order to provide safe drinking
water in the event of an outbreak. This higher treatment require-
ment is predominately attributed to the significantly increased path-
ogen levels in outbreak sewage relative to municipal sewage from a
large city as a result of dilution and the relatively smaller proportion
of the population infected. The recommended pathogen LRVs clearly
represent a worst case scenario, assuming high pathogen concentra-
tions and close community contact (high secondary attack rate).
Generalization to other small communities is relevant nonetheless,
and the model results indicate that in the event of an outbreak addi-
tional treatment barriers will be necessary to achieve safe drinking
water in such communities.
Conflict of interest
There is no conflict of interest to report.
Acknowledgments
The authors would like to acknowledge the members of the AAD
Polar Medicine Unit for providing context and a clearer understanding
of station conditions. Helpful comments and suggestions were also re-
ceived from Dr. Martha Sinclair (Monash University). Funding for this
work through the Cooperative Research Network (CRN) program of
the Australian Government and the Australian Antarctic Division of the
Department of Sustainability, Environment, Water, Populations and
Communities of the Australian Government is gratefully acknowledged.
Appendix A. Supplementary data
Supplementary data to this article can be found online at http://
dx.doi.org/10.1016/j.scitotenv.2013.05.059.
References
AAD. Antarctic explorer —expeditioner's database. Kingston, Tasmania: Australian
Antarctic Division; 2011.
AAD. State of environment. Indicator 61 —total potable water consumption at Australian
Antarctic Stations. SIMR —(state ofenvironment). Systemfor indicator management
and reporting —an on-line state of environment system for the Antarctic. SIMR —
(state of environment). Australian Antarctic Division (AAD); 2012.
731S.F. Barker et al. / Science of the Total Environment 461–462 (2013) 723–733
Author's personal copy
ABS. 4102.0 Australian social trends 2003. Canberra, Australia: Australian Bureau of
Statistics; 2003.
ABS. 4602.0 —Environmental issues: people's views and practices, Mar 2007. Com-
monwealth of Australia, Canberra: Australian Bureau of Statistics; 2007a.
ABS. 4705.0 Population distribution, aboriginal and Torres Strait Islander Australians,
2006. Canberra, Australia: Australian Bureau of Statistics; 2007b.
ABS. 4102.0 Australian social trends 2008. Commonwealth of Australia, Canberra: Aus-
tralian Bureau of Statistics; 2008.
ABS. 4602.0.55.003 —environmental issues: water use and conservation. Common-
wealth of Australia, Canberra: Australian Bureau of Statistics; 2010a.
ABS. 4602.2 —household water, energy use and conservation, Victoria, Oct 2009. Com-
monwealth of Australia, Canberra: Australian Bureau of Statistics; 2010b.
ABS. 3101.0 —Australian demographic statistics, Mar 2012. Canberra: Australian Bu-
reau of Statistics; 2012.
Alfano-Sobsey E, Sweat D, Hall A, Breedlove F, Rodriguez R, Greene S, et al. Norovirus
outbreak associated with undercooked oysters and secondary household transmis-
sion. Epidemiol Infect 2012;140:276–82.
Ãstrom J, Petterson S, Bergstedt O, Pettersson TJR, Stenström TA. Evaluation of the mi-
crobial risk reduction due to selective closure of the raw water intake before drink-
ing water treatment. J Water Health 2007;5:81–97.
Atmar RL. Noroviruses: state of the art. Food Environ Virol 2010;2:117–26.
Atmar RL, Opekun AR, Gilger MA, Estes MK, Crawford SE, Neill FH, et al. Norwalk virus
shedding after experimental human infection. Emerg Infect Dis 2008;14:1553–7.
Aw TG, Gin KYH. Environmental surveillance and molecular characterization of human
enteric viruses in tropical urban wastewaters. J Appl Microbiol 2010;109:716–30.
Baron RC, Murphy FD, Greenberg HB, Davis CE, Bregman DJ, Gary GW, et al. Norwalk
gastrointestinal illness: an outbreak associated with swimmin in a recreational
lake and secondary person-to-person transmission. Am J Epidemiol 1982;115:
163–72.
Birkhead G, Vogt RL. Epidemiologic surveillance for endemic Giardia lamblia infec-
tion in vermont —the roles of waterborne and person-to-person transmission.
Am J Epidemiol 1989;129:762–8.
Buckland ST. Monte carlo confidence intervals. Biometrics 1984;40:811–7.
Castro-Hermida JA, García-Presedo I, Almeida A, González-Warleta M, Da Costa JMC,
Mezo M. Contribution of treated wastewater to the contamination of recreational
river areas with Cryptosporidium spp. and Giardia duodenalis. Water Res 2008;42:
3528–38.
Castro-Hermida JA, García-Presedo I, González-Warleta M, Mezo M. Cryptosporidium
and Giardia detection in water bodies of Galicia, Spain. Water Res 2010;44:
5887–96.
Chan MCW, Sung JJY, Lam RKY, Chan PKS, Lee NLS, Lai RWM, et al. Fecal viral load and
norovirus-associated gastroenteritis. Emerg Infect Dis 2006;12:1278–80.
Cressey P, Lake R. Risk ranking: DALY estimates for selected foodborne diseases in New
Zealand using revised Dutch disability weights. Ilam, Christchurch: Institute of En-
vironmental Science & Research Limited; 2009.
Denborough MA, Downing HJ. Secretor status in asthma and hay fever. J Med Genet
1968;5:302–5.
Evans MR. A milk-borne campylobacter outbreak following an educational farm visit.
Epidemiol Infect 1996;117:457–62.
Feachem RG, Bradley DJ, Garelick H, Mara DD. Sanitation and disease. Health aspects
of excreta and wastewater management. Washington, D.C.: John Wiley & Sons;
1983
Gassmann L, Schwartzbrod J. Wastewater and giardia cysts. Water Sci Technol 1991;24:
183–6.
Götz H, De Jong B, Lindbäck J, Parment PA, Hedlund KO, Torvén M, et al. Epidemiolog-
ical investigation of a food-borne gastroenteritis outbreak caused by Norwalk-like
virus in 30 day-care centres. Scand J Infect Dis 2002;34:115–21.
Haagsma JA, Havelaar AH, Janssen BMF, Bonsel GJ. Disability adjusted life years and
minimal disease: application of a preference-based relevance criterion to rank en-
teric pathogens. Popul Health Metr 2008:6.
Hall GV, Kirk MD, Ashbolt R, Stafford R, Lalor K, Bell R, et al. Frequency of infectious gas-
trointestinal illness in Australia, 2002: regional, seasonal and demographic varia-
tion. Epidemiol Infect 2006;134:111–8.
Haramoto E, Katayama H, Oguma K, Yamashita H, Tajima A, Nakajima H, et al. Seasonal
profiles of human noroviruses and indicator bacteria in a wastewater treatment
plant in Tokyo, Japan. Water Sci Technol 2006;54:301–8.
Havelaar AH. Disease burden of foodborne pathogens in the Netherlands, 2009. Int J
Food Microbiol 2012;156:231–8.
Havelaar AH, Melse JM. Quantifying public health risk in the WHO Guidelines for drinking
water quality. A burden of disease approach. Bilth oven,BA: World Health Organization,
Netherlands Ministry of Housing, Physical Planning and the Environment, Directorate
General for Environmental Protection, Directorate for Soil, Water and Countryside;
2003.
Holler C. Long-term study of occurrence, distribution and reduction of Campylobacter
sp. in the sewage system and wastewater treatment plant of a big town. Water
Sci Technol 1988;20:529–31.
Hoque EM, HopeVT, Kjellström T, Scragg R,Lay-Yee R. Risk of giardiasisin Aucklanders: a
case—control study. Int J Infect Dis 2002;6:191–7.
Hunter PR, De Sylor MA, Risebro HL, Nichols GL, Kay D, Hartemann P. Quantitative mi-
crobial risk assessment of Cryptosporidiosis and Giardiasis from very small private
water supplies. Risk Anal 2011;31:228–36.
Johansson PJH, Torvén M, Hammarlund AC, Björne U, Hedlund KO, Svensson L.
Food-borne outbreak of gastroenteritis associated with genogroup I calicivirus. J
Clin Microbiol 2002;40:794–8.
Johnson PC, Mathewson JJ, DuPont HL, Greenberg HB. Multiple-challenge study of host
susceptibility to Norwalk gastroenteritis in US adults. J Infect Dis 1990;161:18–21.
Katayama H, Haramoto E, Oguma K, Yamashita H, Tajima A, Nakajima H, et al. One-year
monthly quantitative survey of noroviruses, enteroviruses, and adenoviruses in
wastewater collected from six plants in Japan. Water Res 2008;42:1441–8.
Katz DE, Heisey-Grove D, Beach M, Dicker RC, Matyas BT. Prolonged outbreak of giardi-
asis with two modes of transmission. Epidemiol Infect 2006;134:935–41.
Kemmeren JM, Mangen M-JJ, van Duynhoven YTHP, Havelaar AH. Priority setting of
foodborne pathogens: disease burden and costs of selected enteric pathogens
(RIVM report 330080001/2006). Bilthoven, The Netherlands: National Institute
for Public Health and the Environment (RIVM); 2006.
La Rosa G, Pourshaban M, Iaconelli M, Muscillo M. Quantitative real-time PCR of enteric
viruses in influent and effluent samples from wastewater treatment plants in Italy.
Ann Ist Super Sanita 2010;46:266–73.
Lahnsteiner J, Lempert G. Water management in Windhoek, Namibia. Water Sci Technol
2007;55:441–8.
Lake RJ, Cressey PJ, Campbell DM, Oakley E. Risk ranking for foodborne microbial haz-
ards in New Zealand: burden of disease estimates. Risk Anal 2010;30:743–52.
Laverick MA, Wyn-Jones AP, Carter MJ. Quantitative RT-PCR for the enumeration of
noroviruses (Norwalk-like viruses) in water and sewage. Lett Appl Microbiol 2004;39:
127–36.
Le Pendu J, Ruvoen-Clouet N, Kindberg E, Svensson L. Mendelian resistance to human
norovirus infections. Semin Immunol 2006;18:375–86.
Lee N, Chan MCW, Wong B, Choi KW, Sin W, Lui G, et al. Fecal viral concentration and
diarrhea in norovirus gastroenteritis. Emerg Infect Dis 2007;13:1399–401.
Leverenz HL, Tchobanoglous G, Asano T. Direct potable reuse: a future imperative. J
Water Reuse Desalination 2011;1:2–10.
Lin S, Wang X, Zheng H, Mao Z, Sun Y, Jiang B. Direct detection of Campylobacter jejuni
in human stool samples by real-time PCR. Can J Microbiol 2008;54:742–7.
Lindesmith L, Moe C, Marionneau S, Ruvoen N, Jiang X, Lindblad L, et al. Human suscep-
tibility and resistance to Norwalk virus infection. Nat Med 2003;9:548–53.
Lopez CE, Dykes AC, Juranek DD. Waterborne giardiasis: a communitywideoutbreak ofdis-
ease and a high rate of asymptomatic infection. Am J Epidemiol 1980;112:495–507.
Mara DD. Water- and wastewater-related disease and infection risks: what is an appro-
priate value for the maximum tolerable additional burden of disease? J Water
Health 2011;9:217–24.
Mara DD, Hamilton AJ, Sleigh A, Karavarsamis N. Discussion paper: options for updating
the 2006 WHO guidelines. WHO, FAO, IDRC, IWMI; 2010.
Masago Y, Katayama H, Watanabe T, Haramoto E, Hashimoto A, Omura T, et al. Quan-
titative risk assessment of noroviruses in drinking water based on qualitative
data in Japan. Environ Sci Technol 2006;40:7428–33.
Medema GJ, Schijven JF. Modelling the sewag e discharge and dispersion of Cryptosporidium
and giardia in surface water. Water Res 2001;35:4307–16.
Nordgren J, Matussek A, Mattsson A, Svensson L, Lindgren PE. Prevalence of norovirus
and factors influencing virus concentrations during one year in a full-scale waste-
water treatment plant. Water Res 2009;43:1117–25.
Norkrans G, Svedhem Å. Epidemiological aspects of Campylobacter jejuni enteritis. J Hyg
1982;89:163–70.
NRMMC, EPHC, AHMC. National guidelines for water recycling: managing health and en-
vironmental risks (phase 1). National water quality management strategy. Natural
Resource Management Ministerial Council, Environment Protection and Heritage
Council. Canberra: Australian Health Ministers' Conference; 2006a.
NRMMC, EPHC, AHMC. Australian guidelines for water recycling: managing health and
environmental risks (phase 1). National waterquality management strategy. Natural
Resource Management Ministerial Council, Environment Protection and Heritage
Council, Australian Health Ministers' Conference, Canberra; 2006b.
NRMMC, EPHC, NHMRC. Australian guidelines for water recycling: managing health
and environmental risks (Phase 2). Augmentation of drinking water supplies. Na-
tional Water Quality Management Strategy. Canberra: Natural Resource Manage-
ment Ministerial Council, Environment Protection and Heritage Council, National
Health and Medical Research Council; 2008.
NRMMC, EPHC, NHMRC. Australian guidelines for water recycling: managing health and
environmental risks (phase 2). Managed aquifer recharge. National water quality
management strategy. Canberra: Natural Resource Management Ministerial Council,
Environment Protection and Heritage Council, National Health and Medical Research
Council; 2009.
Oda T, Kawabata M, Uga S. Detection of giardia cysts in sewage and estimations of giar-
diasis prevalence among inhabitants in Hyogo Prefecture, Japan. Trop Med Health
2005;33:1–5.
Ottoson J, Hansen A, Bjorlenius B, Norder H, Stenström TA. Removal of viruses, parasitic
protozoa and microbial indicators in conventional and membrane processes in a
wastewater pilot plant. Water Res 2006a;40:1449–57.
OttosonJ,HansenA,WestrellT,JohansenK,NorderH,StenströmTA.Removal of noro- and
enteroviruses, Giardia cysts, Cryptosporidium oocysts, and fecal indicators at four sec-
ondary wastewater treatment plants in Sweden. Water Environ Res 2006b;78:828–34.
Pickering LK, Evans DG, DuPont HL. Diarrhea caused by Shigella, rotavirus, and Giardia
in day-care centers: prospective study. J Pediatr 1981;99:51–6.
Pintar KDM, Fazil A, Pollari F, Waltner-Toews D, Charron DF, McEwen SA, et al. Consider-
ing the risk of infection by Cryptosporidium via consumption of municipally treated
drinking water from a surface water source in a southwestern Ontario community.
Risk Anal 2012;32:1122–38.
Porter IA, Reid TMS. A milk-borne outbreak of Campylobacter infection. J Hyg 1980;84:
415–9.
Pusch D, Oh DY, Wolf S, Dumke R, Schröter-Bobsin U, Höhne M, et al. Detection of en-
teric viruses and bacterial indicators in German environmental waters. Arch Virol
2005;150:929–47.
Rao SSC. Oral rehydration for viral gastroenteritis in adults: a randomized, controlled
trial of 3 solutions. JPEN J Parenter Enteral Nutr 2006;30:433–9.
732 S.F. Barker et al. / Science of the Total Environment 461–462 (2013) 723–733
Author's personal copy
Rechenburg A, Kistemann T. Sewage effluent as a source of Campylobacter sp. in a sur-
face water catchment. Int J Environ Health Res 2009;19:239–49.
Roche SM, Jones AQ, Majowicz SE, McEwen SA, Pintar KDM. Drinking water consumption
patterns in Canadian communities (2001–2007). J Water Health 2012;10:69–86.
Rodríguez S, Araujo R. Occurrence of thermotolerant Campylobacter species in surface
waters of a Mediterranean area and in its prevailing pollution sources. J Appl
Microbiol 2010;109:1027–34.
Roxström-Lindquist K, Palm D, Reiner D, Ringqvist E, Svärd SG. Giardia immunity —an
update. Trends Parasitol 2006;22:26–31.
SchijvenJF, Teunis PFM, Rutjes SA,Bouwknegt M, de Roda HusmanAM. QMRAspot: a tool
for quantitative microbial risk assessment from surface water to potable water.
Water Res 2011;45:5564–76.
Sinclair MI, Hellard ME, Wolfe R, Mitakakis TZ, Leder K, Fairley CK. Pathogens causing
community gastroenteritis in Australia. J Gastroenterol Hepatol 2005;20:1685–90.
Soller JA, Bartrand T, Ashbolt NJ, Ravenscroft J, Wade TJ. Estimating the primary etio-
logic agents in recreational freshwaters impacted by human sources of faecal con-
tamination. Water Res 2010;44:4736–47.
Stellacci P, Liberti L, Notarnicola M, Haas CN. Hygienic sustainability of site location of
wastewater treatment plants. A case study. II. Estimating airborne biological
hazard. Desalination 2010;253:106–11.
Sukhrie FHA. Nosocomial transmission of norovirus is mainly caused by symptomatic
cases. Clin Infect Dis 2012;54:931–7.
Sykora JL, Sorber CA, Jakubowski W, Casson LW, Gavaghan PD, Shapiro MA, et al. Dis-
tribution of Giardia cysts in wastewater. Water Sci Technol 1991;24:187–92.
Tchobanoglous G, Leverenz H, Nellor MH, Crook J. Direct potable reuse: a path forward.
Alexandria, VA: WateReuse Research Foundation; 2011.
ter Waarbeek HLG, Dukers-Muijrers NHTM, Vennema H, Hoebe CJPA. Waterborne gas-
troenteritis outbreak at a scouting camp caused by two norovirus genogroups: GI
and GII. J Clin Virol 2010;47:268–72.
Teunis P, van der Heijden O, van der Giessen J, Havelaar A. The dose–response relation in
human volunteers for gastro-intestinal pathogens. Report nr. 284550002. Bilthoven,
The Netherlands: National Institute of Public Health and the Environment (RIVM);
1996.
Teunis PFM, van den Brandhof W, Nauta M, Wagenaar J, van den Kerkhof H, van Pelt W.
A reconsideration of the Campylobacter dose–response relation. Epidemiol Infect
2005;133:583–92.
Teunis PFM, Moe CL, Liu P, Miller SE, Lindesmith L, Baric RS, et al. Norwalk virus: how
infectious is it? J Med Virol 2008;80:1468–76.
The R Foundationfor Statistical Computing. The R project for statistical computing; 2011.
Thorven M, Grahn A, Hedlund KO, Johansson H, Wahlfrid C, Larson G, et al. A homozy-
gous nonsense mutation (428G-A) in the human secretor (FUT2) gene provides re-
sistance to symptomatic norovirus (GGII) infections. J Virol 2005;79:15351–5.
Tsuchiya H. Astudy on variabilities in dimensions and numbers of discharged cysts
Giardia lamblia (stiles 1915) from day to day under normal conditions. Am J Hyg
1931;13:544–67.
USEPA. Estimated per capita water ingestion and body weight in the United States —an
update. Washington, D C: U.S. EPA, Office of Water, Office of Science and Technology;
2004.
USEPA. Economic analysis for the final ground water rule. United States Environmental
Protection Agency; 2006.
Van Den Akker B, Whiffin V, Cox P, Beatson P, Ashbolt NJ, Roser DJ. Estimating the risk
from sewage treatment plant effluent in the Sydney catchment area. Water Sci
Technol 2011;63:1707–15.
Van Den Berg H, Lodder W, Van Der Poel W, Vennema H, De Roda Husman AM. Genetic
diversity of noroviruses in raw and treated sewage water. Res Microbiol 2005;156:
532–40.
Victoria M, Guimarães FR, Fumian TM, Ferreira FFM, Vieira CB, Leite JPG, et al. One
year monitoring of norovirus in a sewage treatment plant in Rio de Janeiro, Brazil.
J Water Health 2010;8:158–65.
Vijgen SMC, Mangen MJM, Kortbeek LM, van Duijnhoven YTHP, Havelaar AH. Disease
burden and related costs of cryptosporidium and giardiasis in the Netherlands.
Bilthoven: RIVM; 2007.
Von Bonsdorff CH, Maunula L, Niemi RM, Rimhanen-Finne R, Hänninen ML, Lahti K. Hy-
gienic risk assessment by monitoring pathogens in municipal sewage. Water Sci
Technol 2002;2:23–8.
WHO. Guidelines for the safe use of wastewater, excreta and greywater. Geneva,
Switzerland: World Health Organisation; 2006.
WHO. Guidelines for drinking water quality. 4th ed Geneva: World Health Organization;
2011.
Yakoob J, Abbas Z, Beg MA, Naz S, Khan R, Islam M, et al. Prevalences of Giardia lamblia
and Cryptosporidium parvum infection in adults presenting with chronic diarrhoea.
Ann Trop Med Parasitol 2010;104:505–10.
733S.F. Barker et al. / Science of the Total Environment 461–462 (2013) 723–733
