Abstract

The aim of this study was to propose a risk assessment method for infectious diseases, using reclaimed water as a case study. To evaluate the infectious risk of norovirus (NoV) in various applications of the reclaimed water, five types of ultrafiltration (UF) membrane treatments were considered: (1) coagulation with low pH + UF membrane process, (2) UF membrane process alone, (3) UF + ultraviolet (UV) irradiation process, (4) UF + nanofiltration (NF) membrane process, and (5) UF + reverse osmosis (RO) membrane process. These treatments were used in a pilot plant and the NoV concentration after each treatment process was studied over the long term (2010–2014). Infectious risk was described using disability-adjusted life-year (DALY) when the reclaimed water was applied for agricultural irrigation, landscape irrigation, recreational enhancement, and toilet and urinal flushing. The results show that reclaimed water could be considered acceptable for recreational enhancement by adding a UV, an NF membrane, or an RO membrane treatment to the UF membrane treatment process.

INTRODUCTION

As global water consumption increases in coming years, the reuse of wastewater as reclaimed water is expected to attract growing attention. However, before reclaimed water can be used for agricultural and landscape irrigation, it is important to reduce its levels of waterborne pathogens. An ultrafiltration (UF) membrane process could be selected as a reclamation treatment process, because high-quality water can be created at a relatively low cost (Huang et al. 2009) and it has high potential for removing Cryptosporidium, Giardia, and bacteriophage MS2 (Hirata & Hashimoto 1998; Fiksdal & Leiknes 2006; Langlet et al. 2009; Matsushita et al. 2013).

Enteric viruses, including adenovirus, rotavirus, and norovirus (NoV), are found at relatively high levels in sewage and treated sewage and should be removed before reclaimed water is put to various uses (Irving & Smith 1981; Lodder & Husman 2005). These viruses can cause gastroenteritis in humans, and many cases of infectious gastroenteritis have been attributed to NoV globally (Lopman et al. 2004; Marshall & Bruggink 2011). If NoV is not appropriately treated in the reclamation process, water users may become infected with the virus because of its high illness–infection ratio of approximately 80% (Moe 2009).

To use reclaimed water safely, an appropriate evaluation of human health risk should be conducted. For reclaimed water intended for use in agricultural irrigation, the World Health Organization (WHO) recommends conducting risk assessment via the quantitative microbial risk assessment (QMRA) method, using disability-adjusted life-year (DALY) as an indicator of human health risk (World Health Organization 2006). Therefore, an evaluation scheme of NoV infection risk from reclaimed water treated with the UF membrane process based on the QMRA method and monitoring data is proposed in this study.

MATERIALS AND METHODS

Monitoring of NoV concentrations in the pilot plant

Figure 1 outlines the UF membrane treatment process in the pilot plant. The reclaimed water was sourced from the secondary effluent from a wastewater treatment plant in Naha-shi, Japan and pretreated following the conventional activated sludge process. Five types of reclamation processes were conducted in the pilot plant: RUN 1, the coagulation-sedimentation + UF membrane process; and RUN 2, the UF membrane process alone. Three additional treatment processes were also performed: RUN 3, ultraviolet (UV) treatment; RUN 4, the nanofiltration (NF) membrane process; and RUN 5, the reverse osmosis (RO) membrane process. See Figure 1 for details of the treatment processes.

Figure 1

Schematic of the pilot plant for reclamation processing of the secondary effluent.

Figure 1

Schematic of the pilot plant for reclamation processing of the secondary effluent.

In the coagulation–sedimentation + UF membrane process, the coagulation and sedimentation tank was installed before the UF membrane process in order to improve NoV removal and the stability of the membrane process. Polyaluminum chloride (PACl) was used as the flocculant, at a dosage of 50 mg PACl/L (2.65 mg Al/L). After conducting the coagulation and sedimentation processes as pretreatment, accomplished by adjusting the pH to 5.5 using H2SO4 and adding PACl to the secondary effluent, the supernatant was supplied to the UF membrane at a flux rate of 2 m3/(m2·day). The UF membrane was made from polyvinylidene fluoride (PVDF) and the molecular weight cut off (MWCO) was 150 kDa. The process conditions were determined according to the results from preliminary experiments and previous publications (Lee et al. 2013). The second treatment process involved the application of the secondary effluent directly to a UF membrane (PVDF, 150 kDa) at a flux rate of 2 m3/(m2·day).

For these processes, water samples were collected during pilot plant operation in August 2012, February 2013, August 2013, and February 2014. Samples were obtained from the sampling ports (Figure 1) for 1 day in each of the 4 months at 4 h intervals to gauge the daily fluctuation of water quality treated with the UF membrane process.

Additional UV, NF membrane, and RO membrane treatments were performed after the UF membrane treatment process to obtain higher quality reclaimed water. Treated water was collected four times during the monitoring period. A 40-W low-pressure mercury vapor lamp was used for the UV treatment. The effective irradiation volume of the UV reactor was 0.72 L, and the UV dose was adjusted to 100 mJ/cm2 using bacteriophage MS2 as the biodosimetry. The flux of the NF and RO membrane processes was set at 11–13 L/m2/h.

Quantification of NoV genotypes GI and GII

The water samples were concentrated according to the polyethylene glycol sedimentation (PEG) method. The genomes of NoV genotypes GI and GII were quantified with a real-time PCR detector (LightCycler DX400, Roche Diagnostics). The specific primer and TaqMan probes for NoV GI and GII used in the real-time polymerase chain reaction (PCR) were described in a previous study (Kageyama et al. 2003). The standard curve was generated using 10-fold dilutions of plasmid DNA. For initial denaturation, the conditions of the real-time PCR were a temperature of 95 °C for 15 min, followed by 45 cycles of 95 °C for 30 s and 56 °C for 60 s. The details of the water sample concentration, RNA extraction, and related procedures were described in a previous study (Yasui et al. 2016). In this study, the standard curve was created for each measurement of NoV and the copy number was calculated. At that time, it was confirmed that the amplification efficiency was 95% or more and the R2 value was 0.9 or higher. Moreover, it was confirmed that the influence of inhibitors can be reduced in the method used.

Quantitative microbial risk assessment

The risk assessment of reclaimed water was performed following the procedures outlined in Figure 2.

  • 1:

    Selection of target pathogenic microorganism

Figure 2

Procedure for the risk assessment of NoV in the reclaimed water.

Figure 2

Procedure for the risk assessment of NoV in the reclaimed water.

Originally, all pathogenic microorganisms affecting human health should be targeted to evaluate health risk, but it is difficult to obtain and evaluate the substantial volume of data for all pathogenic microorganisms. Therefore, in this study, although there are various pathogenic microorganisms, NoV was targeted as the pathogenic microorganism because of its relatively high frequency in sewage and treated sewage.

  • 2-1:

    Monitoring the concentration of NoV in source water

Once the risk evaluation has been performed, investigation of the NoV concentration in the source water is required. After NoV concentration data was obtained for the source water, the distribution of NoV concentration was calculated following the procedure below.

  • 1.

    Arrange the obtained concentration data in order of concentration from low to high.

  • 2.
    Cumulative probability is calculated using Equation (1).  
    formula
    (1)
    where i is the rank number of the data point, and n is the total number of data points.
  • 3.

    Relationship between the calculated cumulative probability and NoV concentration is plotted on normal probability paper.

  • 4.

    From this plot, if the regression line is linear, the NoV concentration is considered to have normal distribution. This significance was tested using the Shapiro-Wilk test.

  • 2

    -2: Evaluation of removal characteristics for the treatment processes

To evaluate each treatment process shown in Figure 1, the NoV concentration was monitored for the selected treatment process (low pH + UF, UF only, UF + UV, UF + NF, or UF + RO). The removal ratio was calculated using the concentrations of source water and each treated water, and the removal fluctuation was estimated per the procedure shown in 2-1. For cases in which NoV was not detected in the treated water, to be cautious, the removal rate was estimated assuming that 1 copy/tube was detected by qPCR. Moreover, because NoV was not detected in the treated water of UF + UV, UF + NF, and UF + RO (n = 6), the effluent after the UF membrane process was spiked with coliphage MS2, and the removal ratio was estimated using the substitute coliphage MS2 removal rate fluctuation.

  • 3:

    Estimation of NoV concentration in reclaimed water

The NoV concentration distribution of reclaimed water was estimated by using the NoV concentration distribution in the source water and the distribution of removal rate calculated in sections 2-1 and 2-2. Thus, the NoV concentration distribution of the reclaimed water was predicted from the product of the distribution of concentration and the removal rate. Originally, it was only necessary to deal with the data by directly using the concentration of reclaimed water. However, then it was predicted that the NoV concentration in the reclaimed water will be low or undetectable, thus it will be calculated of a clear distribution of NoV concentration in the reclaimed water becomes difficult.

  • 4:

    Decision of applications

Although various applications of reclaimed water exist, toilet flushing, sprinkler irrigation, bathing, and the recreational soaking of hands and feet were selected in this study.

  • 5:

    Risk scenario

In the current case study, it was assumed that the reclaimed water provided by the Naha wastewater pilot plant is used as municipal water in central Naha City. Table 1 shows the applications of the reclaimed water, as well as the exposure frequencies of each application. The scenario of water supply to Naha City from the pilot plant follows.

  • The water supply was set at 4,200 m3/day based on the actual data for Naha City.

  • The distance of water supply from the pilot plant was 5 km, and the hydraulic retention time (HRT) of reclaimed water in the pipe-line was 1.5 h, owing to the diameter of the water pipe that was set at 250 mm.

  • Reclaimed water was stored in two tanks, each with a volume of 266 m3.

  • The HRT of the reclaimed water in the storage tanks was assumed as 3 h, considering the chlorination effect.

  • Chlorine was added to the water to prevent regrowth or scale in the water pipe during the delivery process.

  • The added chlorine was converted to chloramine, and the residual chlorine was set to 0.1 mg/L.

  • NoV may be reduced by Log 0.1 by adding the chlorine at a CT (residual chlorine concentration × contact time) value of 27 mg min/L. (This value was selected from our pre-experimental results which showed that NoV was reduced by Log 1 at a CT value of 270 mg min/L using real-time PCR using chloramine).

  • The total HRT of the reclaimed water supply was 4.5 h (1.5 + 3 h = 270 min). Because the residual chloramine concentration was set at 0.1 mg/L, the CT value became 27 mg min/L (0.1 mg/L × 270 min = 27 mg min/L).

Table 1

Applications of reclaimed water and exposure characteristics from such

ApplicationExposure frequencyTargetExposure type
Toilet flushing 0.02 mL/once, 3 times/year User Ingestion 
Sprinkler irrigation 0.1 mL/once, 20 times/year 
Recreation (Shallow stream or ponda0.3 mL/once, 20 times/year 
Recreation (Bathing) 30 mL/once, 8 times/year 
ApplicationExposure frequencyTargetExposure type
Toilet flushing 0.02 mL/once, 3 times/year User Ingestion 
Sprinkler irrigation 0.1 mL/once, 20 times/year 
Recreation (Shallow stream or ponda0.3 mL/once, 20 times/year 
Recreation (Bathing) 30 mL/once, 8 times/year 

aSoaking hands and feet.

  • 6:

    QMRA based on DALY

An evaluation of the applicability of reclaimed water from each UF membrane treatment technology via the QMRA method was used as an indicator of NoV infectious risk, based on DALY. DALY is a measure of overall disease burden (DB), expressed as the number of years of life lost (YLL) because of poor health, disability, or early death. The DALY was calculated as the DB per person year (DALYpppy) according to Equation (1):  
formula
(2)
where Pinf(D) is the probability of infection, D is the NoV dosage per day, Rinf is the illness–infectious probability, DB is the disease burden (year), and n is exposure time. DALYpppy was calculated for the current risk scenario using a Monte Carlo simulation (n = 100,000). Monte Carlo simulation was performed using Oracle Crystal Ball.

DB is the specific indicator calculating by economic cost, mortality, morbidity rate due to go to hospital or daily life becoming difficult by infecting NoV. DB is calculated by considering each weight from the number of deaths per year, the number of hospitalized patients, etc., due to being infected with NoV. In this study, DB was set at 9.0 × 10−4 (year) based on the previous report from Kemmeren et al. (2006).

The illness–infectious probability (Rinf) was set to 80% in this study, since according to Moe (2009), it was reported that about 20% of people infected with NoV did not develop symptoms.

DALY was calculated as the YLL due to premature mortality in the population and the years lost due to disability (YLD) for people living with the health condition. The dose-response model of NoV reported by Masago et al. (2006) was calculated according to Equation (2):  
formula
(3)
where ID50 is the dosage of NoV for which there is a 50% probability of developing an illness, and D is the dosage of NoV (copy).

The dosage of NoV (D) was calculated from the estimation result of the NoV concentration in the reclaimed water and the amount of exposure for each application.

Although Teunis et al. (2008) proposed the dose-response model for NoV considering the state of aggregation and dispersion for the NoV virus particles, the model is complex and the value of ID50 is indicated as a wide range of 18 to 1,015 copies. Meanwhile, Masago et al. (2006) reported that the value of ID50 at 10 to 100 virus particles and its value range was estimated at safe side. In this study, the results of Masago et al. (2006) were selected as the value of ID50. However, because ID50 had a range of 10 to 100, we assumed the uniform distribution of 10 to 100 in consideration of the fluctuation of these values. The quantitative results via qPCR (copy number) were assumed as the number of infectious virus particles.

The assumption distributions and parameters used to calculate DALYpppy are summarized in Tables 2 and 3.

  • 7:

    Evaluation decision

The acceptable criterion for the reclaimed water was determined as 10−6 DALYpppy, in accordance with the recommendation of the WHO, and adopted for agricultural irrigation. It is necessary to understand that if the qPCR data were used to estimate the risk assessment instead of the infectivity data, the risk would be overestimated.

Table 2

Assumption distributions and parameters used to calculate DALYpppy

 Usage dataDistribution form
NoV concentration in the source water Monitoring data
μ = 1.2 × 106 copies/L
σ = 6.1 × 103 copies/L 
Lognormal distribution 
Reduction rate of each treatment process Monitoring data
μ = 3.9–7.7 Log
σ = 0.3–0.9 Log 
Lognormal distribution 
Concentration of reclaimed water Estimation (Concentration distribution in the source water × distribution of reduction rate) Lognormal distribution 
 Usage dataDistribution form
NoV concentration in the source water Monitoring data
μ = 1.2 × 106 copies/L
σ = 6.1 × 103 copies/L 
Lognormal distribution 
Reduction rate of each treatment process Monitoring data
μ = 3.9–7.7 Log
σ = 0.3–0.9 Log 
Lognormal distribution 
Concentration of reclaimed water Estimation (Concentration distribution in the source water × distribution of reduction rate) Lognormal distribution 
Table 3

Parameters used to calculate DALYpppy

ParameterDescriptionUsage data
 Probability of infection (Dose-response model)
 
References (Masago et al. 2006
 Illness–infectious probability (80%) References (Moe 2009
DB Disease burden (9.0 × 10−4 year) Estimated from references (Kemmeren et al. 2006
Dosage of NoV (copies) Calculation from estimated concentration of reclaimed water and exposure dosage 
 Amount of NoV with a 50% probability of developing as an illness From the reference (Masago et al. 2006), it was assumed that the value of ID50 was set at 10 to 100 with uniform distribution 
ParameterDescriptionUsage data
 Probability of infection (Dose-response model)
 
References (Masago et al. 2006
 Illness–infectious probability (80%) References (Moe 2009
DB Disease burden (9.0 × 10−4 year) Estimated from references (Kemmeren et al. 2006
Dosage of NoV (copies) Calculation from estimated concentration of reclaimed water and exposure dosage 
 Amount of NoV with a 50% probability of developing as an illness From the reference (Masago et al. 2006), it was assumed that the value of ID50 was set at 10 to 100 with uniform distribution 

RESULTS AND DISCUSSION

NoV concentration in the pilot plant

Because no significant difference in the concentration of the influent at the two influent sampling ports was observed on the same day, the data for these two ports were combined. There was an increasing trend in both NoV GI and GII in winter. In summer (August 2012–2013), the average concentration of GI was 1.1 × 105 copies/L and that of GII was 6.7 × 104 copies/L . In winter, the average concentration of GI was 1.3 × 106 copies/L and GII was 7.3 × 106 copies/L .

Figure 3 shows the NoV concentration of the treated effluents for PAC + UF and UF only. NoV was under the detection limit for all of the measurements after the UV, NF membrane, and RO membrane treatment processes (data not shown).

Figure 3

(a) NoV genotype GI concentration in the treated effluents of PAC + UF and UF; (b) NoV genotype GII concentration in the treated effluents of PAC + UF and UF.

Figure 3

(a) NoV genotype GI concentration in the treated effluents of PAC + UF and UF; (b) NoV genotype GII concentration in the treated effluents of PAC + UF and UF.

Estimation of NoV concentration in each reclaimed water

The normality of the distribution of NoV concentration in the influent was determined using the Shapiro-Wilk test. The distribution of NoV concentration in the influent water was calculated by summing the concentrations of genotypes GI and GII, because both genotypes of NoV have the ability to infect humans.

From the calculated result, the concentration of NoV was determined according to the logarithmic normal distribution, with a 5% significance level. The average concentration of NoV GI + GII was 1.2 × 106 copies/L (Log 6.0), with a standard deviation of 6.1 × 103 copies/L (Log 3.8). Values that were under the detection limit in the treated water were assumed to be 1 copy/tube, which we considered a conservative estimate for risk calculation.

After treatment by the UV, NF membrane, and RO membrane processes, NoV was not detected (i.e. was under the detection limit) in each sample of treated water. Therefore, the NoV removal ratio for the NF and RO membrane processes were estimated using the removal ratio of coliphage MS2. A stock solution of coliphage MS2, with a concentration of 107–108 PFU/mL and without medium components, was added to the treated water after the UF membrane process. The concentration of NoV in the UF + UV treatment process was estimated by referring to the pre-experimental result using the low-pressure UV lamp. It was assumed that the average removal rate for the UF + UV treatment was Log 5.0 (σ = Log 0.3), for the UF + NF treatment Log 6.7 (σ = Log 0.5), and for the UF + RO treatment Log 7.7 (σ = Log 0.5).

Results of the trial calculation of NoV infectious risk in municipal water applications

Figure 4 shows the distribution of DALYpppy in the treated water of each reclamation treatment process per application. When the calculated value of DALYpppy matched the associated criteria, the usability of the reclaimed water was classified into one of three categories, which are described below.

  • Acceptable: 95% less than 10−6 (DALY/person year) of the distribution DALYpppy

  • Inadequate: 5–68% with risks greater than 10−6 (DALY/person year)

  • Unacceptable: 68% more than 10−6

Figure 4

Distribution of DALYpppy for each application after several reclamation treatment processes.

Figure 4

Distribution of DALYpppy for each application after several reclamation treatment processes.

The usability of reclaimed water for each application shown in Figure 4 is summarized in Table 4. If the treated sewage was used as reclaimed water, it could not be applicable to all applications from the viewpoint of NoV infectious risk to the water user. Reclaimed water could be made available for recreational use in shallow streams or ponds by performing UV, NF membrane, and RO membrane treatments after the UF membrane process.

Table 4

Summary of evaluation results for the case study

Treatment processSecondary effluentPAC + UFUF + UVUF + NFUF + RO
Removal performance Avg. log reduction rate (Log) – 3.9 5.0 6.7 7.7 
Dispersion (Log) – ± 0.9 ± 0.3 ± 0.5 ± 0.5 
Municipal water uses Toilet flushing 0.02 mL/once
3 times/year 
Unacceptable Acceptable Acceptable Acceptable Acceptable 
Sprinkler irrigation 0.1 mL/once
20 times/year 
Unacceptable Inadequate Acceptable Acceptable Acceptable 
Recreation (Shallow stream/pond) 0.3 mL/once
20 times/year 
Unacceptable Inadequate Acceptable Acceptable Acceptable 
Recreation (Bathing) 30 mL/once
8 times/year 
Unacceptable Unacceptable Inadequate Acceptable Acceptable 
Treatment processSecondary effluentPAC + UFUF + UVUF + NFUF + RO
Removal performance Avg. log reduction rate (Log) – 3.9 5.0 6.7 7.7 
Dispersion (Log) – ± 0.9 ± 0.3 ± 0.5 ± 0.5 
Municipal water uses Toilet flushing 0.02 mL/once
3 times/year 
Unacceptable Acceptable Acceptable Acceptable Acceptable 
Sprinkler irrigation 0.1 mL/once
20 times/year 
Unacceptable Inadequate Acceptable Acceptable Acceptable 
Recreation (Shallow stream/pond) 0.3 mL/once
20 times/year 
Unacceptable Inadequate Acceptable Acceptable Acceptable 
Recreation (Bathing) 30 mL/once
8 times/year 
Unacceptable Unacceptable Inadequate Acceptable Acceptable 

CONCLUSIONS

The infectious risk of NoV to a user was evaluated based on QMRA when reclaimed water from the pilot plant was applied for municipal water use, using monitoring data of NoV concentration in the pilot plant. The following conclusions were reached after evaluating the results.

  • When the secondary effluent, the source of reclaimed water in the pilot plant, was used as municipal water without any reclamation treatment process, it could not be considered usable because of the risk of NoV infections to the water user.

  • Reclaimed water could be made available for recreational use as shallow streams or ponds by performing UV, NF membrane, and RO membrane treatments after the UF membrane process.

  • The dosage of NoV was calculated using the value from real-time PCR, if the ineffective virus particles were equal to the copy number obtained from real-time PCR. Therefore, the risk assessment in the trial was not reflected directly in the NoV virus titer, as the risk in each scenario was estimated using the value of real-time PCR.

The proposed risk assessment in this study can be applied in various situations, as long as the concentration distribution and removal characteristics of the treatment technology for the target facilities are clarified. Therefore, versatile evaluation is possible when the risk scenario is reviewed according to regional characteristics.

The concept of this study is risk assessment for NoV using a qPCR method if reclaimed water is used for municipal water applications. Because the output from the qPCR method is a genome number, it is not appreciated to use these values directly for risk assessment. However, a method for culturing NoV in either cell culture or test animals has not been established yet (Duizer et al. 2004). Therefore, the trial risk assessment of NoV in this study indicated risk for usage of municipal water applications as being on the safe side.

REFERENCES

REFERENCES
Duizer
,
E.
,
Schwab
,
K. J.
,
Neill
,
F. H.
,
Atmar
,
R. L.
,
Koopmans
,
M. P. G.
&
Estes
,
M. K.
2004
Laboratory efforts to cultivate noroviruses
.
J. Gen. Virol.
85
,
79
87
.
Huang
,
H.
,
Schwab
,
K.
&
Jacangelo
,
J. G.
2009
Pretreatment for low pressure membranes in water treatment: a review
.
Environ. Sci. Technol.
43
(
9
),
3011
3019
.
Irving
,
L. G.
&
Smith
,
F. A.
1981
One-year survey of enteroviruses, adenoviruses, and reoviruses isolated from effluent at an activated-sludge purification plant
.
Appl. Environ. Microbiol.
41
(
1
),
51
59
.
Kageyama
,
T.
,
Kojima
,
S.
,
Shinohara
,
M.
,
Uchida
,
K.
,
Fukushi
,
S.
,
Hoshino
,
F.B.
,
Takeda
,
N.
&
Katayama
,
K.
2003
Broadly reactive and highly sensitive assay for Norwalk-like viruses based on real-time quantitative reverse transcription-PCR
.
J. Clin. Microbiol.
41
(
4
),
1548
1557
.
Kemmeren
,
J. M.
,
Mangen
,
M. J. J.
,
van Duynhoven
,
Y. T. H. P.
&
Havelaar
,
A. H.
2006
Priority Setting of Foodborne Pathogens
,
RIVM report 2006
, p.
68
.
Lee
,
S. T.
,
Yamashita
,
N.
,
Ihara
,
M.
,
Tanaka
,
H.
,
Kobayashi
,
K.
,
Tanaka
,
H.
,
Takabatake
,
H.
&
Tanaka
,
Y.
2013
Nexus of energy and virus reduction of UF membrane filtration process with coagulation/UV irradiation for wastewater reclamation
.
Proceedings of Water Environ. Fed. Energy
9
,
5
13
.
Lopman
,
B.
,
Vennema
,
H.
,
Kohli
,
E.
,
Pothier
,
P.
,
Sanchez
,
A.
,
Negredo
,
A.
,
Buesa
,
J.
,
Schreier
,
E.
,
Gray
,
J.
,
Gallimore
,
C.
,
Bottiger
,
B.
,
Hedlund
,
K.-O.
,
Torvén
,
M.
,
von Bonsdorff
,
C.-H.
,
Maunula
,
L.
,
Poljsak-Prijatelj
,
M.
,
Zimsek
,
J.
,
Reuter
,
G.
,
Szücs
,
G.
,
Melegh
,
B.
,
Svennson
,
L.
,
van Duijnhoven
,
Y.
,
Koopmans
,
M.
,
Reacher
,
M.
,
Brown
,
D.
&
Iturriza
,
M.
2004
Increase in viral gastroenteritis outbreaks in Europe and epidemic spread of new norovirus variant
.
Lancet
363
(
9410
),
682
688
.
Marshall
,
J. A.
&
Bruggink
,
J. L.
2011
The dynamics of norovirus outbreak epidemics: recent insights
.
Int. J. Environ. Res. Public Health
8
(
4
),
1141
1149
.
Masago
,
Y.
,
Katayama
,
H.
,
Watanabe
,
T.
,
Haramoto
,
E.
,
Hashimoto
,
A.
,
Omura
,
T.
,
Hirata
,
T.
&
Ohgaki
,
S.
2006
Quantitative risk assessment of Noroviruses in drinking water based on qualitative data in Japan
.
Environ. Sci. Tech.
40
(
23
),
7428
7433
.
Teunis
,
P. F. M.
,
Moe
,
C. L.
,
Liu
,
P.
,
Miller
,
S. E.
,
Lindesmith
,
L.
,
Baric
,
R. S.
,
Pendu
,
L. J.
&
Calderon
,
R. L.
2008
Norwalk Virus: How Infectious is It?
J. Med. Virol.
80
(
8
),
1468
1476
.
World Health Organization
2006
WHO Guidelines for the Safe use of Wastewater, Excreta and Greywater, Volume II; Wastewater Use in Agriculture
.
WHO
,
Geneva, Switzerland
, pp.
56
64
.
Yasui
,
N.
,
Suwa
,
M.
,
Sakurai
,
K.
,
Suzuki
,
Y.
,
Tsumori
,
J.
,
Kobayashi
,
K.
,
Takabatake
,
H.
,
Lee
,
S. T.
,
Yamashita
,
N.
&
Tanaka
,
H.
2016
Removal characteristics and fluctuation of norovirus in a pilot-plant by an ultrafiltration membrane for the reclamation of treated sewage
.
Environ. Technol.
37
(
12
),
2793
2801
.