Abstract

A quantitative microbial risk assessment (QMRA) was conducted to support renewal of the City of Vacaville wastewater discharge permit and seasonal (summer) filtration requirements. Influent and final disinfected effluent from the city's wastewater treatment plant, as well as 11 receiving water stations, were monitored for indicator organisms (i.e. total and fecal coliforms, Escherichia coli, Enterococcus, male-specific bacteriophage (MS2), and the Bacteroidales) and several pathogens (i.e. Giardia cysts, Cryptosporidium oocysts, infectious Cryptosporidium, and Norovirus GI and GII). QMRA annualized risks of infection for selected pathogens enteric viruses, Giardia and Cryptosporidium. Estimated median annualized risk for recreational exposure in either disinfected secondary and/or filtered disinfected secondary effluent is on the order of 1.1 × 10−3 per person per year (pppy) for enteric viruses and would be roughly one order of magnitude lower if local receiving water dilution of the treatment plant effluent was taken into account. Estimated median annual risk for recreation exposure in disinfected secondary effluent is 1.8 × 10−3 pppy for Cryptosporidium and 1 log10 less with filtration during the summer months. The estimated median annual risk for landscape exposure (e.g. golfing) to secondary disinfected effluent is 7.6 × 10−7 pppy for enteric viruses. Estimated median annualized risk is 1.7 × 10−7 pppy for enteric viruses and 3.0 × 10−5 to 3.6 × 10−6 pppy for parasites for use of secondary disinfected effluent with irrigated agriculture. Estimated annualized risks for recreational exposure to the local receiving waters were approximately 10 to 1,000 times greater than direct recreational exposure to the final filtered and disinfected effluent. All risk estimates associated with exposure to final treated plant effluent (i.e. secondary filtered and disinfected) were close to or lower than the California level of acceptable annual risk of infection of 10−4 pppy for recreational exposure. Risk estimates provide further evidence to support the use of seasonal treatment limits requiring summer filtration for public health protection.

INTRODUCTION

Quantitative microbial risk assessment (QMRA) involves evaluating the likelihood that an adverse health effect may result from human exposure to one or more pathogens. The risk assessment process requires the selection of appropriate microorganisms, routes of human exposure (e.g. receiving water recreation, and water reuse for food crops, recreation and landscape irrigation), and models to calculate risks to an individual or population. Two fundamental QMRA approaches are pervasive in the literature. They may be categorized as static (i.e. individual–based risk assessment) or dynamic (i.e. population-based risk assessments that consider person-to-person transmission, and/or the establishment of immunity within the population). The static model (National Research Council 1983) is commonly used as a generic framework for conducting microbial risk assessments of water- and food-borne pathogens (Haas et al. 2014). Assessments using a static model typically focus on estimating the probability of infection or disease for an individual as a result of a single exposure event. These assessments generally assume that multiple or recurring exposures constitute independent events with identical distributions of microbial contamination (Regli et al. 1991).

While QMRAs are now commonly used to inform many aspects of water planning and policy, an increasing number of studies are beginning to combine the QMRA approach with novel source tracking methods that use microbial genetics (Ashbolt et al. 2010). The Bacteroidales bacteria group in particular, is a useful indicator of contamination due to human versus other animal sources (EPA 2010). Genotyping of Cryptosporidium can also indicate likely origin (DiGiovanni et al. 2010). Evaluating the relative contribution of different types of Bacteroidales or Cryptosporidium from the wastewater treatment plant (WWTP) and at different locales within the discharge watershed can potentially provide greater insight into the proportion of total health risks from exposure to water that are attributed to wastewater discharge.

This paper describes the results of a study (Olivieri et al. 2012) in which the overall objective was to collect sufficient final disinfected effluent and various receiving water samples, upstream and downstream of a WWTP in California, USA, over the course of a year to better characterize microbial agents and update a previously-conducted QMRA (Olivieri & Soller 2002) to address various pathogenic agents of infection and/or disease through a number of exposure routes including: (1) recreation in the receiving waters, (2) using the recycled water for agricultural irrigation of food crops, and (3) non-potable reuse on parks, playgrounds, and golf courses. The results of the QMRA were then utilized to support the renewal City of Vacaville wastewater discharge permit and seasonal (summer only filtration) discharge limits.

METHODS

Study area

The study area (approximately 156,000 acres) lies in the northeastern portion of Solano County in the highly agricultural Central Valley of California. The main natural waterway in the area is Ulatis Creek, although agriculture canals and tidal channels are present throughout the area (Figure 1). All waterways inevitably drain to the Sacramento Delta and eventually to San Francisco Bay. The study area is comprised of both urban and non-urban land uses. The majority of the population in the area lives in the City of Vacaville. Although the study area includes urban land area, agriculture is the dominant land use (roughly 44% of area). Based on the most recent land use data, livestock-related land uses (i.e. pasture land, feedlots and dairies) make up roughly 11% of Solano County and a quarter of the study area (Table 1).

Table 1

Urban and non-urban land uses within Solano County and the study area

Land use classification% of study area% of Solano County
Urban 13% 11% 
Agricultural 
 Livestock (pasture, feedlots and dairies) 26% 11% 
 Row crops 18% 16% 
 Native vegetation, vacant and water 43% 62% 
TOTAL 100% 100% 
Land use classification% of study area% of Solano County
Urban 13% 11% 
Agricultural 
 Livestock (pasture, feedlots and dairies) 26% 11% 
 Row crops 18% 16% 
 Native vegetation, vacant and water 43% 62% 
TOTAL 100% 100% 
Figure 1

Summary of sampling locations and results.

Figure 1

Summary of sampling locations and results.

According to the County of Solano Department of Agriculture (2011), livestock production in the county consists of roughly 30,640 cows and 32,696 sheep. Assuming that livestock populations are associated with pasture, feedlot and dairy land uses, the density of livestock is roughly 0.48 cows and 0.52 sheep per acre of livestock-related land area in Solano County. Using these densities of livestock, an estimated 19,845 cows and 21,177 sheep were present within the study area at any point in time.

Vacaville's Easterly Wastewater Treatment Plant (EWWTP) discharges disinfected and dechlorinated secondary effluent (‘final disinfected effluent’) into the receiving waters of the watershed of the Sacramento–San Joaquin Delta. The EWWTP consists of two parallel plants, the older North Plant and the newer South Plant that went into service in 2004. The treatment system includes influent headworks, primary sedimentation basins, aeration basins, secondary clarifiers, and chlorination and dechlorination facilities. The nominal average dry weather flow of wastewater that can be treated by the facility is 15 MGD. After chlorination, the effluent is de-chlorinated using sodium bisulfite prior to discharge to Old Alamo Creek. At the time of the investigation improvements to the EWWTP under design/construction primarily consisted of facilities required to (1) bypass secondary treatment during wet season high influent plant flows and (2) provide a level of treatment equivalent to disinfected tertiary effluent as specified in the California Water Recycling Criteria for all flow prior to discharge during the dry season.

Pathogen sampling and analysis

Various microbial indicators and pathogens were sampled over the course of approximately one year (January 1, 2011 through January 31, 2012) to characterize both influent and effluent concentrations for use in the QMRA. Eleven receiving water sampling locations are shown on Figure 1 and were selected based upon the upstream and downstream position relative to the EWWTP, suspected agricultural influences, flow characteristics within the watershed, and proximity to a drinking water intake. All samples were collected during out-flow tidal events in order to measure the maximum influence of the EWWTP effluent and agricultural drains into the watershed.

In developing our pathogen sampling and QMRA approach, the following factors were considered: (1) for waterborne illness or disease to occur an agent of disease (pathogen) must be present, (2) the agent must be present in sufficient concentration to produce a probability of infection or disease in an unacceptable fraction of the population, and (3) a susceptible host must come into contact with the dose in a manner that results in infection or disease (Cooper 1991). From the long list of possible pathogens, those known to be present in wastewater, and most responsible for disease burden in the United States (Mead et al. 1999; Scallan et al. 2011) resulted in the following list of ‘pathogens of public health concern’: (Plant Influent and Final Disinfected Effluent) Noroviruses (NoV genogroup GI and GII), Cryptosporidium spp. and Giardia spp., and (Receiving Waters) Norovirus (GI and GII), Cryptosporidium spp. and Giardia spp. Norovirus (NoV) was added to the list since it is the leading cause of gastroenteritis worldwide (Atmar & Estes 2006). In addition, indicator organisms Enterococcus spp., male specific phage (MS2 Escherichia coli host 15597) the Bacteroidales group, total and fecal coliforms, and E. coli were included in the study. A probit analysis of the indicator data was conducted allowing for transforming sigmoid type curves to a straight line that can then be analyzed further through regression analysis (Ott 1995).

Noroviruses are members of the Caliciviridae family and are subdivided into five genogroups (da Silva et al. 2007). NoV GII has been shown to account for approximately 92% of the gastroenteritis cases, and NoV GI for a major percentage of the remaining cases (Atmar & Estes 2006). Unfortunately, molecular detection is currently the only method of detection and as such, accurate and precise quantification in wastewater and environmental samples is hampered by limitation in extraction efficiencies, the presence of inhibitory compounds, as well as low levels of viral pollution (da Silva et al. 2007; Hellmer et al. 2014). Further, as discussed under the section below covering dose response assumptions, while genome-based methods are more sensitive at detecting the presence of copies of the genome of a virus, these methods do not provide information on viral infectivity. The US EPA Method 600/R-95/178 modified with molecular detection by Ceeram, SAS PCR kit was used for the detection of Noroviruses.

Table 2 contains a summary of the methods and detection limits for microbial constituents monitored. Positive Cryptosporidium samples were analyzed by a technique known as the ‘Foci Detection Method and Most Probable Number Method’, or FDM-MPN (Slifko et al. 1999) to determine viability (i.e. infectivity). Positive samples for Giardia were examined microscopically to determine integrity of the cyst. The presence of intact internal structure was taken to indicate infectivity.

Table 2

Sample constituent, method and detection limit

ConstituentMethodDetection Limit & Units
Total, fecal and E. coli SM 9221 (MPN) 2 MPN/100 mL 
Enterococcus sp. Enterolet, SM 9230 D (MPN) 1 MPN/100 mL 
Human Bacteroidales EPA Method B, March 2010
(Lamendella et al. 2009)  
1/100 mL 
Cryptosporidium & Giardia (Total (oo)cysts) EPA Method 1623 0.1 cyst or oocyst/L
Assumes 10 L grab 
Cryptosporidium infectivity EPA Method 1623 (2005),
Assay by Slifko et al. (1999)  
0.1 Fluorescent foci (FF)/L 
Cryptosporidium genotyping Isolation by US EPA 1623
Assay by DiGiovanni et al. (2010)  
NA 
MS2 Bacteriophage Double-layer (Adams 1959); EPA Method 1602 (2001)  1 PFU/vol analyzed 
Norovirus genogroup I & II US EPA 600/R-95/178 Modified with molecular detection by Ceeram, S.A.S. PCR kit 2 Gene copies/L per EPA method 
Bio-banking DiGiovanni et al. (2010) (TAMU) NA 
ConstituentMethodDetection Limit & Units
Total, fecal and E. coli SM 9221 (MPN) 2 MPN/100 mL 
Enterococcus sp. Enterolet, SM 9230 D (MPN) 1 MPN/100 mL 
Human Bacteroidales EPA Method B, March 2010
(Lamendella et al. 2009)  
1/100 mL 
Cryptosporidium & Giardia (Total (oo)cysts) EPA Method 1623 0.1 cyst or oocyst/L
Assumes 10 L grab 
Cryptosporidium infectivity EPA Method 1623 (2005),
Assay by Slifko et al. (1999)  
0.1 Fluorescent foci (FF)/L 
Cryptosporidium genotyping Isolation by US EPA 1623
Assay by DiGiovanni et al. (2010)  
NA 
MS2 Bacteriophage Double-layer (Adams 1959); EPA Method 1602 (2001)  1 PFU/vol analyzed 
Norovirus genogroup I & II US EPA 600/R-95/178 Modified with molecular detection by Ceeram, S.A.S. PCR kit 2 Gene copies/L per EPA method 
Bio-banking DiGiovanni et al. (2010) (TAMU) NA 

Source tracking

Microbial source tracking employed during this investigation followed the Library Independent (LI) method, with the goal of determining whether the source was, or was not, human. The Bacteroidales bacteria group was used (EPA 2010). Initially, the general Bacteroidales marker was run to develop baseline data (Haugland et al. 2005; Siefring et al. 2008). Then the human Bacteroidales marker was run on all positive samples (Benhard & Field 2000). Also, molecular techniques were used to determine if positive samples for Cryptosporidium were the type(s) that are infectious to humans (DiGiovanni et al. 2010).

QMRA approach

A static QMRA approach was selected because information necessary for use of dynamic models (e.g. total number of diseased individuals, host immunity, etc.) is limited. For a static model, the population is assumed to be categorized into two epidemiological states: a susceptible state and an infected or diseased state. Susceptible individuals are exposed to the pathogen of interest and move into the infected/diseased state with a probability that is governed by a dose-response relationship (Figure 2). The risk computed from the dose-response is expressed as ‘per day’, and converted to an annualized risk (Haas et al. 2014):  
formula
where P is the probability of being infected at least once during the year, Probinf(d) is the probability of being infected for a given daily dose d, and the number of days of exposure is n. Note that the distribution of annualized risk estimates are based on simulated distribution of daily doses generated through an exposure model.
Figure 2

Static risk assessment conceptual model.

Figure 2

Static risk assessment conceptual model.

Pathogen concentrations and treatment reductions used in the QMRA

Data for each of the above pathogens and indicators in influent and secondary effluent as well as local receiving waters were obtained from plant and field monitoring program. To account for the variability observed in pathogen concentrations, lognormal distributions were fit to pathogen concentration data using Maximum Likelihood Estimation (MLE) (EPA 1991; Ott 1995).

EWWTP log10 removals for wastewater treatment processes were derived from the pathogen and indicator monitoring data (see Table 3). The log10 reductions are of the same order of magnitude to those identified in the WateReuse literature review (Olivieri et al. 2007) and the results of Rose et al. (2004). Further, review of the data provided in the National Water Resource Institute (NWRI) report (Olivieri et al. 2014) indicates that additional parasite removals by filtration during dry season operation would conservatively range from 0.5 to 1 log10 for both Giardia spp. and Cryptosporidium spp., and no additional virus removals. The EWWTP measured performance along with the estimated parasite reductions for summer filtration were then used for the QMRA simulations.

Table 3

EWWTP monitoring results for phage (MS2), Cryptosporidium, Giardia and Noroviruses (NoV GI and GII)

OrganismInfluent (raw sewage)
Final effluent (secondary disinfected)
MinMaxMedianSamples positive/totalMinMaxMedianSamples positive/total
Phage (MS2) 
 PFU/100 mL 6.00 × 103 1.30 × 106 2.60 × 105 32/32 <1 <1 2/32 
Norovirusa 
 GC/L 
  NoV GI 5.80 × 102 3.00 × 105 5.30 × 103 9/11 <DL <DL <DL 0/11 
  NoV GII 4.60 × 102 5.00 × 107 6.00 × 103 11/11 <DL <DL <DL 0/11 
Cryptosporidium 
 oocysts/L <1 10 5/11 <0.1 1.8 5/11 
 Adjusted for recoveryb 7.7 77 15 NA 0.2 3.4 1.9 NA 
Giardia 
 cysts/L 50 16.5 × 103 8.40 × 102 11/11 <0.1 0.4 0.1 2/11 
 Adjusted for recoveryb 3.85 × 102 1.27 × 105 6.47 × 103 NA 0.2 0.6 0.2 NA 
OrganismInfluent (raw sewage)
Final effluent (secondary disinfected)
MinMaxMedianSamples positive/totalMinMaxMedianSamples positive/total
Phage (MS2) 
 PFU/100 mL 6.00 × 103 1.30 × 106 2.60 × 105 32/32 <1 <1 2/32 
Norovirusa 
 GC/L 
  NoV GI 5.80 × 102 3.00 × 105 5.30 × 103 9/11 <DL <DL <DL 0/11 
  NoV GII 4.60 × 102 5.00 × 107 6.00 × 103 11/11 <DL <DL <DL 0/11 
Cryptosporidium 
 oocysts/L <1 10 5/11 <0.1 1.8 5/11 
 Adjusted for recoveryb 7.7 77 15 NA 0.2 3.4 1.9 NA 
Giardia 
 cysts/L 50 16.5 × 103 8.40 × 102 11/11 <0.1 0.4 0.1 2/11 
 Adjusted for recoveryb 3.85 × 102 1.27 × 105 6.47 × 103 NA 0.2 0.6 0.2 NA 

Note: <DL = less than detection limit of 2 Gene Copies/L, ND = non detected, GC = gene copy, NA = not applicable, PFU = plaque-forming unit (EPA Method 1602 (2001)).

aThe US EPA Method 600/R-95/178 modified with molecular detection by Ceeram, SAS PCR kits was used for the determination of presence and concentration of Noroviruses. The Ceeram reverse transcriptase real-time PCR kits could distinguish between Norovirus GI and GII.

bThe laboratory data were adjusted for percent recovery based on average of matrix spikes: Influent Giardia - 13% (increase by 7.7), Influent Cryptosporidium - 13% (increase 7.7), Effluent Giardia – 63% (increase 1.6) and Effluent Cryptosporidium- 54% (increase by 1.9) (IEH-Biovir Laboratories in Olivieri et al. 2012).

Exposure scenarios for QMRA

A total of three scenarios were explored and scenario assumptions are provided in Table 4.

Table 4

Exposure and treatment assumptions for recreational exposure, agricultural irrigation, and landscape irrigation

SExposures ScenariosTreatment AssumptionsExposure AssumptionsPathogens
One (I) Recreation exposure
  • (a)

    in final effluent (no dilution)

  • (b)

    in Cache Slough (main water mass background station located 18.5 miles from discharge)

 
Log10 reductions based on sampling a lognormal distribution described by data in Table 6 
Assumed uniform distribution of 0.5 to 1 log10 reduction for filtration of Giardia and Cryptosporidium during dry seasona 
Median exposure of 19 mL/event (Dufour et al. 2006 and Olivieri et al. 2007)
Assume 10 (Cooper et al. 2012) for base case analysis
Considered 40 events (Tanaka et al. 1998) per year summer season only 
Giardia spp., Cryptosporidium spp., Enteric virusesb 
Two (II) Landscape irrigation (golf course, parks) Same as above Uniform distribution with median exposure of 6 mL/event and upper and lower bound of 12 to 0.12 mL/event (Olivieri et al. 2007)
Assume 25 events per year (Tanaka et al. 1998
Giardia spp., Cryptosporidium spp., Enteric virusesb 
Three (III) Food crops (edible portion of crop in contact with water) Same as above Assume 40 events per year (Cooper et al. 2012) of consumption of lettuce per body weight 0.205 g/kg-day (EPA 2003).
Body weight: lognormal distribution with mean 61.4 and SD 13.4 kg (EPA 1997).
Volume of water on lettuce: zero-truncated normal distribution with mean 0.108 and SD 0.02 mL/g (Hamilton et al. 2006)
Assume 7 day environmental decay (Cooper et al. 2012)c 
Giardia spp., Cryptosporidium spp., Enteric virusesb 
SExposures ScenariosTreatment AssumptionsExposure AssumptionsPathogens
One (I) Recreation exposure
  • (a)

    in final effluent (no dilution)

  • (b)

    in Cache Slough (main water mass background station located 18.5 miles from discharge)

 
Log10 reductions based on sampling a lognormal distribution described by data in Table 6 
Assumed uniform distribution of 0.5 to 1 log10 reduction for filtration of Giardia and Cryptosporidium during dry seasona 
Median exposure of 19 mL/event (Dufour et al. 2006 and Olivieri et al. 2007)
Assume 10 (Cooper et al. 2012) for base case analysis
Considered 40 events (Tanaka et al. 1998) per year summer season only 
Giardia spp., Cryptosporidium spp., Enteric virusesb 
Two (II) Landscape irrigation (golf course, parks) Same as above Uniform distribution with median exposure of 6 mL/event and upper and lower bound of 12 to 0.12 mL/event (Olivieri et al. 2007)
Assume 25 events per year (Tanaka et al. 1998
Giardia spp., Cryptosporidium spp., Enteric virusesb 
Three (III) Food crops (edible portion of crop in contact with water) Same as above Assume 40 events per year (Cooper et al. 2012) of consumption of lettuce per body weight 0.205 g/kg-day (EPA 2003).
Body weight: lognormal distribution with mean 61.4 and SD 13.4 kg (EPA 1997).
Volume of water on lettuce: zero-truncated normal distribution with mean 0.108 and SD 0.02 mL/g (Hamilton et al. 2006)
Assume 7 day environmental decay (Cooper et al. 2012)c 
Giardia spp., Cryptosporidium spp., Enteric virusesb 

aBased on data and analysis contained in Soller et al. (2007) and Olivieri et al. (2014).

bEnteric viruses (EV) are based on the MS2 indicator data. The indicator data needs to be converted to EV concentration data in order to enable its use later in the dose-response equation. The following equation converts F + RNA coliphage concentration to enteric virus concentration: Log(y) = 0.17 + 0.98*Log(x), where x is the concentration of coliphage (per mL) and y is the concentration of enteric viruses (per L). The relationship is based on a study by Havelaar et al. (1993). In that study, a high correlation between these two organisms was found for several types of undisinfected environmental samples including river water. The relationship above may vary from site to site, but is currently the best available relationship. Given the assumptions implied in the risk calculations and that the ratio of phage to EV was reported to be relatively constant regardless of concentration and type of water investigated, it is anticipated that the risk estimates derived using coliphage for EWWTP final disinfected effluent are conservative upper bound estimates.

cOver 7 day decay period assume a mean 3.3 log10 reduction for enteric virus, and a 2 log10 reduction for Giardia and Cryptosporidium (Olivieri et al. 2014).

Scenario I addresses recreational exposure and associated ingestion of water during swimming in either of two concentrations: (a) the final effluent without dilution (a worst-case recreational scenario), or (b) in Cache Slough (a local receiving water).

Scenario II addresses ingestion of water that may occur from those that interact with landscape, golf courses, or parks that are irrigated with recycled water.

Scenario III addresses the consumption of food crops and the associated ingestion of water that may be on the surface of the edible portion of crops irrigated with recycled water. Human exposure through irrigation of food crops is based on Hamilton et al. (2006) and is consistent with earlier work (van Ginneken & Oron 2000; Petterson et al. 2001). The approach assumes that the ingestion of water is the product of three distributions: the rate of consumption of crops irrigated with recycled water (g/kg-day), body mass (kg), and volume caught on the plant surface (mL/g). Lettuce was used as the model crop for consumption because it is consumed at a relatively high rate compared to other vegetables (EPA 2003), and hence results in a conservative estimate of risk. The consumption value for lettuce is a point estimate of 0.205 g/kg-day (EPA 2003). Body mass is estimated by a normal distribution with mean of 61.429 and standard deviation of 13.362 kg (EPA 1997). The volume of irrigation water captured by the lettuce is estimated as a normal distribution with mean 0.108 and standard deviation of 0.02 mL/g (Hamilton et al. 2006). The resultant distribution of ingestion volume, that is, the amount of irrigation water ingested via lettuce, has a median value of approximately 1.3 mL/day with 5th and 95th percentiles of 0.018 mL/day and 97 mL/day.

Environmental decay assumptions were incorporated into Scenario III and are pathogen specific. It was assumed that virus concentrations in the environment decayed exponentially with time after irrigation (i.e. decay factor = ekt) based on findings from Petterson et al. (2001, 2002) and the approach of Hamilton et al. (2006). Based on Petterson's in-field study of decay, the decay constant k was assumed to be normally distributed with a mean of 1.07 and standard deviation of 0.07 (zero-truncated). This k is conservative due to Petterson's use of B. fragilis phage, a relatively hardy organism. Based on standard agricultural practices employed in California, 7 days (i.e. last day of irrigation prior to consumption) of environmental decay was assumed (i.e. mean of 3.3 log10 removal due to environmental decay) (Olivieri et al. 2014). Based on Mara et al. (2007), it was assumed that E. coli decayed at 3 log10 removal over 7 days. For Giardia and Cryptosporidium, a 2 log10 reduction from environmental decay over 7 days was assumed (Olivieri et al. 2014).

Dose-response assumptions for QMRA

For each of the pathogens investigated (i.e. enteric viruses, Cryptosporidium and Giardia), a summary of the functional forms, distributions used to describe the dose-response parameters, and the dose-response parameters along with corresponding references to support those data is presented in Table 5. To estimate enteric virus risk, MS2 indicator data, determined based on EPA method 1602 (EPA 2001), were used as an indicator of concentration, and a rotavirus dose-response was used to estimate the risk of infection. The dose-response relations for rotavirus, Cryptosporidium, and Giardia are relatively straightforward and commonly used in the field of QMRA. Norovirus data were used to estimate treatment plant enteric virus removals and were not explicitly analyzed as part of the risk assessment for the following reasons: (1) there remains significant uncertainty associated with the selection of a dose-response model (Van Abel et al. 2016); (2) no dose-response model has been accepted from a regulatory perspective; (3) norovirus has not been cultivated successfully using conventional tissue culture methods (consequently, no work is available to establish the ratio between genome density and infectious unit density in the water environment); and (4) while genome-based methods are more sensitive at detecting the presence of copies of the genome of a virus, these methods do not provide information on viral infectivity (NRC 2012).

Table 5

Summary of pathogen dose-response assumptions

PathogenDose-response form (Endpoint)Parameter distributionValue(s)References
Rotavirus Hypergeometric (InfectionPoint estimates a = 0.167
b = 0.191 
Teunis & Havelaar (2000) a 
Cryptosporidium spp. Exponential (InfectionUniform rlower = 0.04
rupper = 0.16 
EPA (2006)  
Giardia spp. Exponential (Infection) Point estimate r = 0.0199 Rose & Gerba (1991); Teunis et al. (1996)  
PathogenDose-response form (Endpoint)Parameter distributionValue(s)References
Rotavirus Hypergeometric (InfectionPoint estimates a = 0.167
b = 0.191 
Teunis & Havelaar (2000) a 
Cryptosporidium spp. Exponential (InfectionUniform rlower = 0.04
rupper = 0.16 
EPA (2006)  
Giardia spp. Exponential (Infection) Point estimate r = 0.0199 Rose & Gerba (1991); Teunis et al. (1996)  

aThe dose-response relationship utilizes dose units in focal forming units. The conversion between MPN and FFU was based on Havelaar et al. (1993). See footnote c in Table 4.

Modeling methods

A Monte Carlo simulation approach was used to address variability and uncertainty in the computed estimates of risk. For each pathogen/treatment process/route of exposure combination of interest, the following process was used (Figure 3). First, representative values from pathogen/indicator were used to characterize the pathogen concentration in influent (1) and fit to a lognormal distribution (2). A sample of 10,000 estimates from this distribution were generated. For each sample, reductions in the concentrations of the pathogens of interest that are expected to occur through wastewater treatment were estimated based on EWWTP data and literature data for filtration (3) and applied (4) to estimate effluent concentrations (5). Based on the exposure route of interest, ingestion rates were estimated (6). By combining the ingestion rate (7) with the effluent concentration, the dose of pathogen ingested per exposure event was estimated (8). A dose-response relationship, derived from the literature (9), was then used to estimate the risk associated with the dose for the exposure event and to estimate an annualized risk if the dose occurs throughout the year (10). Because all the dose-response functions for the pathogens considered have infection as an endpoint, the risk was expressed in terms of risk of infection. It is important to note that the estimated risks are for infection rather than disease, and that not all infections result in clinical disease. The QMRA simulations were conducted using the R language (version 3.1.0).

Figure 3

Flow diagram for the conducting QMRAs.

Figure 3

Flow diagram for the conducting QMRAs.

Acceptable risk assumption

California Department of Public Health (CDPH) implementation of the Water Recycling Criteria is based on a goal that the treatment-based standards provide sufficient overall plant reliability to achieve the US EPA drinking water regulation acceptable risk of one infection per 10,000 people per year (10−4 pppy), as a mean goal, for enteric viruses (Tanaka et al. 1998). This acceptable risk goal is considered conservative for non-voluntary exposures relevant to non-potable reuse. For voluntary recreation water exposures the 10−2 US EPA basis for criteria (EPA 2012) might be more appropriate. However, because the EPA acceptable risk is based on an indicator-illness (disease) relationship in receiving waters for single exposures they are problematic to compare to annualized risk estimates for pathogens based on an infection end point and are not used in this QMRA.

RESULTS

EWWTP performance

The plant influent and effluent water-monitoring results for the 13-month field study are summarized in Table 3, and plant performance (i.e. estimated log10 removals) described by the data shown in Table 6 were used for the QMRA simulations.

Table 6

EWWTP phage (MS2), Cryptosporidium, Giardia and Noroviruses estimated Log10 removal (adjusted for matrix spike recoverya) for secondary disinfected treatment

 Phage (MS2)CryptosporidiumaGiardiaaNorovirusbNorovirus
NoV GINoV GII
N (#samples) 32 11 11 11 11 
Log mean 5.4 1.4 4.5 3.9 4.1 
Log SD 0.5 0.7 0.7 0.8 1.7 
Maximum 6.1 2.6 5.9 5.5 7.7 
Minimum 3.0 0.4 3.4 2.8 2.7 
 Phage (MS2)CryptosporidiumaGiardiaaNorovirusbNorovirus
NoV GINoV GII
N (#samples) 32 11 11 11 11 
Log mean 5.4 1.4 4.5 3.9 4.1 
Log SD 0.5 0.7 0.7 0.8 1.7 
Maximum 6.1 2.6 5.9 5.5 7.7 
Minimum 3.0 0.4 3.4 2.8 2.7 

For purposes of calculating log10 removal, a value less than the method detection limit was considered to be at the detection limit. Normal distributions were zero truncated so that negative values were not sampled.

aThe laboratory data were adjusted for percent recovery based on average of matrix spikes: Influent Giardia - 13% (increase by 7.7), Influent Cryptosporidium - 13% (increase 7.7), Effluent Giardia – 63% (increase 1.6) and Effluent Cryptosporidium - 54% (increase by 1.9) (BIOVIR, April 30, 2012 contained in Olivieri et al. 2012).

bThe US EPA Method 600/R-95/178 modified with molecular detection by Ceeram, S.A.S PCR kits was used for the determination of presence and concentration of Noroviruses. The Ceeram reverse transcriptase real-time PCR kits could distinguish between Norovirus GI and GII.

For the indicator, total coliforms, the final effluent was found to be in compliance with the total coliform 23 MPN/100 mL single sample limit 100% of the time in both wet (November through April) and dry (May through October) seasons contained in the California wastewater treatment discharge permit (Olivieri et al. 2012). A probit analysis indicates compliance can be expected approximately 99.5% of the time during the wet season and >99.8% of the time during the dry season. The final effluent is in compliance with the final 2015 National Pollutant Discharge Elimination System (NPDES) limit for the total coliform 7-day median limit of 2.2 MPN/100 mL 99.5% of the time during the wet season (one value out of 389 exceeded the limit), and 100% of the time during the dry season.

The estimated treatment performance for enteric viruses was based on the male specific coliphage (MS2 provides a conservative estimate) data, and screening for Noroviruses (NoV). The log10 removal of viruses (based on MS2) ranged from 3.0 to 6.1 log10 with a mean of 5.4 log10. The US EPA Method 600/R-95/178 modified with molecular detection by Ceeram, SAS PCR kits was used for the determination of presence and concentration of Noroviruses. The Ceeram reverse transcriptase real-time PCR kits could distinguish between Norovirus GI and GII. Norovirus were positive in nine of the 11 EWWTP influent samples (raw wastewater) for NoV GI and 11 out of 11 samples for NoV GII. None of the 11 final disinfected and dechlorinated effluent samples were positive for NoV GI and GII. Based on the monitoring results, the mean log removal of Noroviruses from the plant appear to be on the order of at least 3.9 to 4.1 log10 and more likely 5 to 8 log10 removals based on the upper end of the range of observed concentrations (unadjusted for recovery). As noted previously, Norovirus data were used only to estimate plant removals and not used to conduct the QMRA.

Cryptosporidium concentrations in the influent ranged from <1 (detection limit) to 10 oocysts/L with a median of 1 oocyst/L (the estimated range and median are, adjusted for recovery, 7.7 to 77 oocysts/L with a median of 15 oocysts/L). Five of the 11 influent samples were positive for Cryptosporidium and none were infective. Four were found to be human-pathogenic by genetic markers. Cryptosporidium concentrations in the final disinfected effluent ranged from <0.1 (detection limit) to 1.8 oocysts/L with a median of 1 oocyst/L (the estimated range and median values, adjusted for recovery efficiency, were 0.2 to 1.8 oocysts/L with a median of 0.2 oocysts/L). The log10 removal of Cryptosporidium over a 13-month period of plant operation (assuming all non-detectable (ND) values are set at the detection limits) ranged from 0.4 to 2.6 log10 with a mean of 1.4 log10 removal. Five of the 11 final effluent samples were positive for Cryptosporidium, and none were infective. Genotyping (Olivieri et al. 2012) indicated that five samples were considered human-pathogenic by genetic markers.

Giardia concentrations in the influent ranged from 50 cysts/L to 16,500 cysts/L with a median of 840 cysts/L (the estimated range and median, adjusted for recovery efficiency, was 385 to 127,050 cysts/L with a median of 6,468 cysts/L). All 11 influent samples were positive for Giardia. Microscopic examination of the positive samples indicated that the percentage positive samples with identifiable internal structure ranged from 20% to 70% with a median value of 35%.

Giardia concentrations in the final effluent ranged from <0.1 (detection limit) to 0.4 cysts/L with a median of <0.1 cysts/L (the estimated range and median values, adjusted for recovery efficiency, was 0.2 to 0.6 cysts/L with a median of 0.2 cysts/L). Two of the 11 samples were positive for Giardia. A 60% median value was found for positive samples reviewed for final effluent and receiving water. The log10 removal of Giardia (assuming all non-detect (ND) values are set at the detection limits) ranged from 3.4 to 5.9 log10 with a mean of 4.5 log10. Please note that for the purpose of this analysis we conservatively assumed that a Giardia cyst with an identifiable internal structure was infective.

Receiving water quality

The results of the receiving water monitoring are graphically illustrated in Figure 1. Bacterial indicators (i.e. total and fecal coliforms, Enterococcus, and E. coli) were generally found at all receiving water stations at a mean concentration approximately 2 log10 and virus indicator (MS2 phage) concentration was found at a number of the stations on the order of 2 log10. The NPDES permit for the plant contains a fecal coliforms receiving water limitation for the immediate receiving waters (Old Alamo and New Alamo Creeks) of a geometric mean <200 MPN/100 mL in five samples over 30 day period and no more than 10% of the samples in a 30 day period greater than 400 MPN/100 mL). While the monitoring program was not designed for direct comparison against this limitation, data collected for Old Alamo Creek (assumes all non-detects are set at the detection limit of 2 MPN/100 mL) indicates that the geometric mean for the Old Alamo station upstream of the EWWTP is 318 MPN/100 mL with 20% exceeding the maximum 400 MPN/100 mL value, and data collected for Old and New Alamo Creek data downstream of plant indicates that the geometric mean is roughly 180 MPN/100 mL with less than 10% exceeding the maximum limitation (one sample out of 11). Norovirus were negative in 10 out of 12 receiving water samples (two samples had matrix interference). Parasites (i.e. Giardia spp. and Cryptosporidium spp.) were found at four of five receiving water stations (Old Alamo Upstream of the EWWTP, Old Alamo Downstream of the EWWTP, and Cache Slough – Main Water Mass), at generally low concentrations (i.e. <0.1 to 3 oocysts/L for Cryptosporidium and <0.1 to <1.7 cysts/L for Giardia). All were found to be non-infective for Cryptosporidium, and three of the four positive samples for Giardia had identifiable internal structure (assumed to be infective). Cryptosporidium was found in all five samples from Cache Slough ranging from <0.1 to 3 oocysts/L with a median of 0.6 oocysts/L. With one exception, the animal-associated Cryptosporidia (as determined by genetic markers) predominated at the receiving water stations.

Source tracking

The human Bacteroidales marker was then run on all the positive samples. The results for a total of 58 samples (28 influent wastewater and 30 final disinfected effluent) indicated that Bacteroidales copies were found in the influent at a median level of 3 × 106 copies/mL and in the final disinfected effluent at a median level of 1 × 102 copies/mL. The results from 49 receiving water samples collected at 11 stations indicated that Bacteroidales were found at median level of 1.4 × 103 copies/mL, and range from 1.8 × 102 to 5.6 × 104 copies/mL. The EWWTP influent tested positive for the human marker while the receiving water stations and final disinfected effluent were negative for the human marker clearly indicating that the predominate Bacteroidales were from non-human sources.

QMRA findings

The wastewater plant influent data employed as input for the QMRA simulations are included in Table 7, and annualized risk results for the water recycling exposure scenarios and the one receiving water exposure scenario are shown in Tables 811. Estimated ‘per event’ risks are generally about two orders of magnitude less than annualized estimates (Olivieri et al. 2012). Seasonality of the data was not considered for this analysis.

Table 7

Summary of wastewater influent pathogen concentration distributions used for modeling

Pathogen (units)Distributiona
Enteric virus (MPN/L)b Lognormal (log mean 3.51, log SD 0.45) 
Giardia lamblia (cysts/L)c Lognormal (log mean 3.82, log SD 0.61) 
Cryptosporidium parvum (oocysts/L)c Lognormal (log mean 1.30, log SD 0.39) 
Pathogen (units)Distributiona
Enteric virus (MPN/L)b Lognormal (log mean 3.51, log SD 0.45) 
Giardia lamblia (cysts/L)c Lognormal (log mean 3.82, log SD 0.61) 
Cryptosporidium parvum (oocysts/L)c Lognormal (log mean 1.30, log SD 0.39) 

aBase 10 log mean and SD are provided here for clarity, however, the Monte Carlo risk simulations in R are based on the natural logarithm.

bBased on phage (MS2), and the Havelaar et al. (1993) coliphage to enteric virus relationship.

cBased on adjusting for recovery see Table 6.

Table 8

Scenario I (a) secondary disinfected (chlorine) effluent (no filtration)

StatisticaEnteric virusGiardiabCryptosporidiumc
Min 4.91 × 10−7 2.58 × 10−8 3.11 × 10−7 
25th 2.70 × 10−4 4.11 × 10−5 3.88 × 10−4 
Median 1.10 × 10−3 2.29 × 10−4 1.80 × 10−3 
75th 4.43 × 10−3 1.25 × 10−3 8.41 × 10−3 
90th 1.54 × 10−2 5.73 × 10−3 3.32 × 10−2 
95th 3.30 × 10−2 1.48 × 10−2 7.62 × 10−2 
Max 9.88 × 10−1 9.96 × 10−1 1.00 × 101 
Mean 8.29 × 10−3 4.58 × 10−3 1.83 × 10−2 
SD 3.51 × 10−2 3.12 × 10−2 6.87 × 10−2 
StatisticaEnteric virusGiardiabCryptosporidiumc
Min 4.91 × 10−7 2.58 × 10−8 3.11 × 10−7 
25th 2.70 × 10−4 4.11 × 10−5 3.88 × 10−4 
Median 1.10 × 10−3 2.29 × 10−4 1.80 × 10−3 
75th 4.43 × 10−3 1.25 × 10−3 8.41 × 10−3 
90th 1.54 × 10−2 5.73 × 10−3 3.32 × 10−2 
95th 3.30 × 10−2 1.48 × 10−2 7.62 × 10−2 
Max 9.88 × 10−1 9.96 × 10−1 1.00 × 101 
Mean 8.29 × 10−3 4.58 × 10−3 1.83 × 10−2 
SD 3.51 × 10−2 3.12 × 10−2 6.87 × 10−2 

Summary of annualized risks of infection assuming 10 exposure events per year of recreation in the final effluent.

aValues are percentiles.

bMedian estimated risk assumes 35% with complete structure (infective). Adjusting infectivity from 35% to 60% yields an estimated median annualized risk of 3.73 × 10−04.

cEstimated risks assume 24% infective (Note: none were infective).

Table 9

Scenario I (b) summary of annualized risks of infection assuming 10 exposure events per year of recreation in the Cache Slough receiving water

StatisticaEnteric virusGiardiabCryptosporidiumc
Min 1.35 × 10−4 6.95 × 10−9 3.36 × 10−6 
25th 2.03 × 10−1 7.83 × 10−5 1.82 × 10−3 
Median 6.31 × 10−1 6.23 × 10−4 7.80 × 10−3 
75th 9.58 × 10−1 4.69 × 10−3 3.36 × 10−2 
90th 9.88 × 10−1 2.83 × 10−2 1.19 × 10−1 
95th 1.00 × 10−0 8.43 × 10−2 2.47 × 10−1 
Max 1.00 × 100 1.00 × 100 1.00 × 100 
Mean 5.75 × 10−1 2.05 × 10−2 4.84 × 10−2 
SD 3.65 × 10−1 8.94 × 10−2 1.20 × 10−1 
StatisticaEnteric virusGiardiabCryptosporidiumc
Min 1.35 × 10−4 6.95 × 10−9 3.36 × 10−6 
25th 2.03 × 10−1 7.83 × 10−5 1.82 × 10−3 
Median 6.31 × 10−1 6.23 × 10−4 7.80 × 10−3 
75th 9.58 × 10−1 4.69 × 10−3 3.36 × 10−2 
90th 9.88 × 10−1 2.83 × 10−2 1.19 × 10−1 
95th 1.00 × 10−0 8.43 × 10−2 2.47 × 10−1 
Max 1.00 × 100 1.00 × 100 1.00 × 100 
Mean 5.75 × 10−1 2.05 × 10−2 4.84 × 10−2 
SD 3.65 × 10−1 8.94 × 10−2 1.20 × 10−1 

aValues are percentiles.

bMedian estimated risk assumes 35% with complete structure (infective). Adjusting infectivity from 35% to 60% yields an estimated median annualized risk of 1.01 × 10−3.

cEstimated risks assume 24% infective (Note: none were infective).

Table 10

Scenario II secondary disinfected (chlorine) effluent (no filtration)

StatisticaEnteric virusGiardiabCryptosporidiumc
Min 8.06 × 10−10 2.99 × 10−11 4.05 × 10−10 
25th 2.29 × 10−7 3.33 × 10−8 3.14 × 10−7 
Median 7.55 × 10−7 1.54 × 10−7 1.21 × 10−6 
75th 2.22 × 10−6 7.09 × 10−7 4.65 × 10−6 
90th 6.58 × 10−6 2.70 × 10−6 1.57 × 10−5 
95th 1.19 × 10−5 6.00 × 10−6 3.16 × 10−5 
Max 4.39 × 10−4 7.91 × 10−4 2.85 × 10−3 
Mean 3.02 × 10−6 1.76 × 10−6 8.27 × 10−6 
SD 1.00 × 10−5 1.27 × 10−5 4.21 × 10−5 
StatisticaEnteric virusGiardiabCryptosporidiumc
Min 8.06 × 10−10 2.99 × 10−11 4.05 × 10−10 
25th 2.29 × 10−7 3.33 × 10−8 3.14 × 10−7 
Median 7.55 × 10−7 1.54 × 10−7 1.21 × 10−6 
75th 2.22 × 10−6 7.09 × 10−7 4.65 × 10−6 
90th 6.58 × 10−6 2.70 × 10−6 1.57 × 10−5 
95th 1.19 × 10−5 6.00 × 10−6 3.16 × 10−5 
Max 4.39 × 10−4 7.91 × 10−4 2.85 × 10−3 
Mean 3.02 × 10−6 1.76 × 10−6 8.27 × 10−6 
SD 1.00 × 10−5 1.27 × 10−5 4.21 × 10−5 

Summary of annualized risks of infection assuming 25 exposure events per year from landscape (golf course) irrigation.

aValues are percentiles.

bMedian estimated risk assumes 35% with complete structure (infective). Adjusting infectivity from 35% to 60% yields an estimated median annualized risk of 2.51 × 10−7.

cEstimated risks assume 24% infective (Note: none were infective).

Table 11

Scenario III secondary disinfected (chlorine) effluent applied directly to crops

StatisticaEnteric virusGiardiabCryptosporidiumc
Min 1.69 × 10−11 3.10 × 10−10 
25th 2.34 × 10−8 3.84 × 10−7 3.35 × 10−6 
Median 1.72 × 10−7 3.63 × 10−6 2.99 × 10−5 
75th 1.34 × 10−6 3.40 × 10−5 2.44 × 10−4 
90th 8.38 × 10−6 2.63 × 10−4 1.65 × 10−3 
95th 2.41 × 10−5 8.67 × 10−4 5.35 × 10−3 
Max 4.79 × 10−2 9.92 × 10−1 9.99 × 10−1 
Mean 1.81 × 10−5 7.37 × 10−4 2.62 × 10−3 
SD 5.17 × 10−4 1.49 × 10−2 2.49 × 10−2 
StatisticaEnteric virusGiardiabCryptosporidiumc
Min 1.69 × 10−11 3.10 × 10−10 
25th 2.34 × 10−8 3.84 × 10−7 3.35 × 10−6 
Median 1.72 × 10−7 3.63 × 10−6 2.99 × 10−5 
75th 1.34 × 10−6 3.40 × 10−5 2.44 × 10−4 
90th 8.38 × 10−6 2.63 × 10−4 1.65 × 10−3 
95th 2.41 × 10−5 8.67 × 10−4 5.35 × 10−3 
Max 4.79 × 10−2 9.92 × 10−1 9.99 × 10−1 
Mean 1.81 × 10−5 7.37 × 10−4 2.62 × 10−3 
SD 5.17 × 10−4 1.49 × 10−2 2.49 × 10−2 

Summary of annualized risks of infection assuming 40 exposure events per year to crops irrigated with recycled water (no filtration).

aValues are percentiles.

bMedian estimated risk assumes 35% with complete structure (infective). Adjusting infectivity from 35% to 60% yields an estimated median annualized risk of 5.91 × 10−6.

cEstimated risks assume 24% infective (Note: none were infective).

The estimated annualized risks for recreational exposure (Scenario I(b) – Cache Slough, see Table 11) in the local receiving waters were conducted using the monitoring results for the Cache Slough (the main water mass, strongly influenced by tides and where Cryptosporidium spp. were detected more frequently and at higher concentrations during the monitoring program). The estimated annualized median risks from recreational exposure to local receiving waters (assumed to not be influenced by wastewater effluent) were all equal to or greater than the acceptable California risk of infection of 10−4 pppy. Comparison of Cache Slough recreational risk estimates against the risk estimates for recreational exposure to non-filtered EWWTP effluent (Scenario I(a), see Table 8) indicate that the risk associated with recreation in the non-filtered effluent is on the order of 0.4 to 0.6 log10 less for the parasites, and on the order of 3 log10 less risk for enteric viruses. Risks associated with recreation in filtered effluent is estimated to be 1.2 log10 less for Giardia and 1.4 log10 less for Cryptosporidium than recreating in Cache Slough (Olivieri et al. 2012). Note that none of the Cryptosporidium in Cache Slough was infective, and thus the risk assessment relied on the same conservative assumption used for recreational exposure to the final EWWTP effluent. Further, the addition of filtration did not change estimated virus reductions and thus, did change the annualized risk estimates for all scenarios investigated. Further note that filtration only occurs during the dry season and is appropriate to consider for all exposure scenarios.

For the recreational exposure in Cache Slough (Scenario I (b)) 10 event per year were assumed for the base case. If during the summer season 40 events are assumed the estimated risk of infection roughly increases by 1 log10.

For Scenario II (see Table 10), the estimated median annualized risk was 7.6 × 10−7 for enteric viruses and 1.5 × 10−7 to 1.2 × 10−6 for parasites for use of secondary disinfected effluent for landscape irrigation. All estimates are lower than the one annual infection per 10−4 California assumed level of acceptable infection for recreational exposure (CDPH 2010). The addition of filtration results in lowering the annualized median risk estimates for the parasites by about 1 log10 resulting in risk estimates for parasites on the order of 10−8 to 10−7 pppy.

For Scenario III (see Table 11), direct contact of produce with agricultural irrigation using recycled water, estimated median annualized risk was 1.7 × 10−7 for enteric viruses and 3.0 × 10−5 to 3.6 × 10−6 for parasites for use of secondary disinfected effluent. All estimates are lower than the California assumed level of acceptable infection for recreational exposure. The addition of filtration results in lowering the annualized median risk estimates for the parasites by about 1 log10 resulting in risk estimates for parasites on the order of 10−6 to 10−7 pppy.

Generally for all the scenarios, sensitivity of risk results to dose-response parameters, which are also subject to uncertainty and population variability, was not explored. For example, the dose-response relationship utilized for Cryptosporidium (Table 5) relies on US EPA (EPA 2006) exponential equation and assumes a uniform distribution to incorporate uncertainty and variability with the r value. It is expected that incorporation of dose-response variability and uncertainty would most likely broaden the distribution of risk, however, relative to the variability associated with exposure assumptions it is anticipated that risk estimates would tend to be insensitive to varying r values.

The EWWTP is required by the State of California to filter the secondary effluent prior to disinfection in the dry season (i.e. the months of May through October) of each year. As previously discussed it is conservatively assumed that additional parasite removals through the addition of filtration are on the order of 0.5 to 1 log10 reductions (uniform distribution) for both Giardia spp. and for Cryptosporidium spp. (Olivieri et al. 2014). Based on the QMRA conducted as part of this investigation, adding filtration results in approximately a 1 to 1.5 log10 reduction of the annualized median risk estimate for exposure to pathogenic parasites. No additional removals through filtration were assumed for enteric viruses, thus the estimated risks presented above were not changed.

DISCUSSION

In summary, all median annualized risks, except for Cryptosporidium during recreation in the final effluent, based on the representative microbial concentrations and exposure scenarios (i.e. landscape irrigation, crop irrigation and consumption, and recreation in the final effluent) to secondary disinfected treatment effluent are below the 10−4 pppy risk of infection goal (results are based on adjusting for recovery and infectivity of parasites and assuming that 24% (Rose et al. 2004) of the Cryptosporidium were infective when all results indicate none were infective). The addition of filtration results in lowering the annualized median risk estimates for the parasites by about 1 log10 resulting in risk estimates for parasites on the order of 10−6 to 10−7. Estimated median annualized risk of infection for direct recreational exposure in either disinfected secondary and/or filtered disinfected secondary effluent is on the order of 1.1 × 10−3 pppy for enteric viruses, which is at approximately the level of acceptable risk of infection utilized in California, but would be roughly an order of magnitude lower if local receiving water dilution of the treatment plant effluent was taken into account.

Estimated median annualized risk is 1.7 × 10−7 pppy for enteric viruses and 3.0 × 10−5 to 3.6 × 10−6 pppy for parasites for use of secondary disinfected effluent with irrigated agriculture. All irrigated crop exposure estimates are lower than the California 10−4 pppy assumed level of acceptable infection for recreational exposure.

The 13-month EWWTP and receiving water monitoring for microbial indicators, including specific human pathogens, and the current QMRA results are consistent with and support the basis for the 2002 CDPH decision regarding seasonal treatment limits for the protection of public health from potential exposure associated with recreation in undiluted effluent, and from golf course and food crop irrigation with treated effluent. Specifically, the California Department of Public Health (CDPH) reviewed the 2001 QMRA and concluded that tertiary filtration is appropriate and necessary during the dry weather period (e.g. May through October) for public health protection and that, during winter months when the beneficial uses of agricultural irrigation and contact recreational activities do not occur, secondary treatment with disinfection to an MPN of 23 total coliforms/100 mL is necessary for public health protection (CDPH 2002). The EWWTP is in the process of providing additional treatment (e.g. sand filtration and additional disinfection flexibility) to address the Regional Water Board's (RWB) new seasonal effluent requirements based on the CDPH recommendations. No changes to the RWB seasonal limits are necessary as they are supported by the results of the microbial monitoring program and QMRA.

Utilizing the same conservative assumptions, the QMRA indicates that potential risks associated with exposure to receiving waters present annualized risk levels 10 to 1,000 times greater than direct exposure to final filtered and disinfected effluent for the various routes of exposure investigated. Further, the results of bacterial indicator monitoring and source tracking investigation, coupled with the heavy agriculture and farm land uses adjacent to the sloughs, clearly indicate that predominate sources for these microbial organisms in the receiving water are of non-human origin.

Investigation of the viability/non-viability of protozoan organisms was conducted where practicable. Since it is not unusual to observe protozoan parasites in the effluent from a WWTP, such observations can cause concern because the standard methodology for detecting protozoans does not determine the physiological state of the organism. We attempted to account for this for Cryptosporidium samples using the FDM-MPN (Slifko et al. 1999) method to determine infectivity. Unfortunately, there is not a convenient cell culture method for infectious Giardia. Thus, testing for infectivity would require the use of a whole animal assay, which was impractical and too expensive for this study. However, microscopic examination of cyst integrity for Giardia was conducted on positive samples. The presence of morphological features, or, the reaction to a stain, does not definitively demonstrate the physiological or infective nature of the organism. Nevertheless, for the purpose of the QMRA conducted for this investigation it was conservatively assumed that the presence of intact internal structure indicated infectivity. The conservative assumption regarding infectivity did not change the overall results of this paper.

The EWWTP data presented in Table 3 are generally of the same concentrations found in raw wastewater of Cryptosporidium parvum range of 1 to 70 oocysts/L, and Giardia lamblia range of 70 to 104 cysts/L (Olivieri et al. 2014). NoV concentrations found in the raw wastewater samples presented in Table 3 for NoV GI had a geometric mean of 7.3 × 103 GC per L (ranged from 580 to 3 × 105 genome copies (GC) per L) and for NoV GII had a geometric mean of 2.4 × 104 GC per L (ranged from 460 to 5 × 107 GC per L). NoV concentrations found in raw wastewater samples presented in Table 3 are about the same concentrations noted in other investigations that ranged from 101 to 109 GC per L for NoV GI, and from 104 to 107 GC per L for NoV GII (da Silva et al. 2007; Victoria et al. 2010; Flannery et al. 2012; Hellmer et al. 2014). We did not try to normalize the methods used or statistical compare the concentrations. Further, the general observation that the concentration of NoV GII being greater than NoV GI appears to be consistent with the results of these and the other noted investigations (da Silva et al. 2007; Victoria et al. 2010; Flannery et al. 2012; Hellmer et al. 2014). More importantly, for this limited investigation, none of the 11 secondary treated and disinfected/dechlorinated effluent samples were positive for NoV GI and GII. As previously noted, estimated enteric viruses were based on measured MS2 concentrations in untreated wastewater ranging from 60 to 13,000 PFU per mL and the Havelaar et al. (1993) relationship between FRNA and enteric viruses. Based on these assumptions, the range of estimated enteric virus in untreated wastewater is from 101 to 104 PFU/L which is similar to the reported range of enteric viruses of from 102 to 103 PFU/L (Havelaar et al. 1993).

Given the predominant agricultural land uses and potential runoff the level of indicator organisms found in the receiving waters are not surprising. Further, given the quality of the EWWTP secondary disinfected effluent (i.e. during the period January 1, 2011 through January 31, 2012, 50 of the 55 final effluent samples were <2 MPN/100 mL (detection limit) and five were at the limit of detection) it was apparent that the EWWTP was not the source of fecal coliform organisms identified within the receiving water.

This study's use of the static model employing Monte Carlo simulations in a comparative screening level risk characterization is consistent with the literature in the field describing conditions in which the use of the static model is appropriate (Haas 1983; Haas et al. 1999; Soller et al. 2004). Also, as part of the WERF 2004 development of QMRA tools, the question of convergence using a dynamic versus the static model was investigated. The analysis indicated that, generally, as acceptable risk levels approached less than one per 10,000 per person year for low doses, the static and dynamic model estimates were similar (Soller et al. 2004). Thus, the static model is appropriate for this assessment.

Overall, the results of this investigation indicate that the median risk estimates associated with exposure (i.e. landscape irrigation, crop irrigation and consumption, and recreation in the final effluent) to the secondary disinfected final effluent, except for Cryptosporidium during direct recreation in final effluent, are all close to or lower than the annual risk of infection of 10−4 pppy, the assumed level of acceptable infection for recreational exposure by the State of California. If dilution from receiving water is accounted for, the median risk estimates for enteric virus and recreational exposure are the same order of magnitude as the acceptable risk level. The addition of seasonal treatment with filtration lowered the annualized median risk estimates for the parasites by about 1 log10. The risk assessment also indicate that potential risks associated with exposure to the local receiving waters (i.e., Cache Slough) present median annualized risk level of infection of approximately 0.5 log10 greater for parasites and 3 log10 greater enteric viruses than direct recreational exposure to final disinfected secondary effluent. Further, the results of bacterial indicator monitoring and the source tracking investigation, coupled with the heavy agriculture and farmland uses adjacent to the sloughs, clearly indicates that sources for these microbial organisms in the receiving water are non-human. Finally, the risk estimates provide further evidence to support the 2002 California Department of Public Health decision (CDPH 2002) regarding seasonal treatment limits for the protection of public health from potential exposure associated with recreation in undiluted effluent, and from golf course and food crop irrigation with treated effluent.

REFERENCES

REFERENCES
Adams
,
M. H.
1959
Bacteriophages
.
Interscience Publishers
,
New York, NY, USA
.
Ashbolt
,
N. J.
,
Schoen
,
M. E.
,
Soller
,
J. A.
&
Roser
,
D. J.
2010
Predicting pathogen risks to aid beach management: the real value of quantitative microbial risk assessment (QMRA)
.
Water Research
44
(
16
),
4692
4703
.
Atmar
,
R. T.
&
Estes
,
M. K.
2006
The epidemiology and clinical importance of norovirus infection
.
Gastroenterology Clinics of North America
35
,
275
290
.
California Department of Public Health (CDPH)
2002
North Coastal Region, Drinking Water Field Operations Branch, Division of Drinking Water and Environmental Management, in a letter dated 22 March 2002 (referenced in RWB Order No. R5-2008-0055, Fact Sheet, pages F-24-27)
.
California Department of Public Health (CDPH)
2010
Yamamoto, G.H. Letter to Assistant Executive Officer K.D. Landau
.
Central Valley Water Board
,
June 15, 2010
.
Cooper
,
R. C.
1991
Public health concerns in wastewater reuse
.
Water Science and Technology
24
(
9
),
55
65
.
Cooper
,
R. C.
,
Olivieri
,
A. W.
,
Cahn
,
M.
,
Colford
,
J.
,
Crook
,
J.
,
Debroux
,
J. F.
,
Mandrell
,
R.
,
Suslow
,
T.
&
Tchobanoglous
,
G.
2012
Review of California Water Recycling Criteria for agricultural irrigation
.
National Water Research Institute (NWRI) Independent Advisory Panel for California Department of Public Health
.
Fountain Valley, CA, USA
.
da Silva
,
A. K.
,
Le Saux
,
J. C.
,
Parnaudeau
,
S.
,
Pommepuy
,
M.
,
Elimelech
,
M.
&
Le Guyadar
,
F. S.
2007
Evaluation of removal of norovirus during wastewater treatment, using real-time reverse transcription-PCR: different behaviors of genogroups I and II
.
Applied and Environmental Microbiology
73
(
24
),
7891
7897
.
DiGiovanni
,
G. D.
,
Hoffman
,
R. M.
&
Sturbaum
,
G. D.
2010
Cryptosporidium Genotyping Method for Regulatory Slides
.
Water Res Foundation
.
Rpt 4099
.
Dufour
,
A.
,
Evans
,
O.
,
Behymer
,
T.
&
Cantu
,
R.
2006
Water ingestion during swimming activities in a pool: a pilot study
.
Journal of Water and Health
,
4
(
4
),
425
430
.
EPA
1991
Technical Support Document Water Quality-Based Toxics Control
.
Office of Water
,
Vol. EPA/505/2-90-001
.
EPA
1996
ICR Microbial Laboratory manual
.
600/R-95/178ICR
.
EPA
1997
Method 1600: Membrane Filter Test Method for Enterococci in Water
.
EPA
2001
Method 1602: Male Specific (F+) and Somatic Coliphage in Water by Single Agar Layer 9SAL) Procedure
.
EPA 821-R-01-029. April 2001
.
EPA
2003
CSFII Analysis of Food Intake Distributions
.
EPA-600-R-03-029
,
Washington, DC, USA
.
EPA
2005
Method 1623: Cryptosporidium and Giardia in Water by Filtration/IMS/FA
.
EPA 815-R-05-002. December 2005
.
EPA
2006
National Primary Drinking Water Regulations: Long Term 2 Enhanced Surface Water Treatment Rule (LT2ESWTR); Final Rule
.
Vol. 40CFR Parts 9, 141 and 142, volume 71, Number 654, January 5, 2006
.
EPA
2010
Method B: Bacteroidales in Water by TaqMan Quantitative Polymerase Chain Reaction (qPCR) Assay
.
March. Draft
.
EPA
2012
Recreational Water Quality Criteria
.
EPA 820-F-12-058
;
United States Environmental Protection Agency (USEPA) Office of Water
,
Washington, DC, USA
.
Flannery
,
J.
,
Keaveney
,
S.
,
Rajko-Nenow
,
P.
,
O'Flaherty
,
V.
&
Dore
,
W.
2012
Concentrations of norovirus during wastewater treatment and its impact on oyster contamination
.
Applied and Environmental Microbiology
78
(
9
),
3400
3406
.
Haas
,
C. N.
1983
Effect of effluent disinfection on risks of viral disease transmission via recreational exposure
.
Journal of the Water Pollution Control Federation
55
,
1111
1116
.
Haas
,
C. N.
,
Rose
,
J. B.
&
Gerba
,
C. P.
1999
Quantitative Microbial Risk Assessment
.
Wiley
,
New York, NY, USA
.
Haas
,
C. N.
,
Rose
,
J. B.
&
Gerba
,
C. P.
2014
Quantitative Microbial Risk Assessment
, 2nd edn.
Wiley
,
New York, NY, USA
.
Hamilton
,
A. J.
,
Stagnitti
,
F.
,
Premier
,
R.
,
Boland
,
A.
&
Hale
,
G.
2006
Quantitative microbial risk assessment models for consumption of raw vegetables irrigated with reclaimed water
.
Applied and Environmental Microbiology
72
(
5
),
3284
3290
.
Havelaar
,
A. H.
,
van Olphen
,
M.
&
Drost
,
Y. C.
1993
F-specific RNA bacteriophages are adequate model organisms for enteric viruses in fresh water
.
Applied and Environmental Microbiology
59
(
9
),
2956
2962
.
Hellmer
,
M.
,
Paxeus
,
N.
,
Enacha
,
L.
,
Arnholm
,
B.
,
Johansson
,
A.
,
Bergstrom
,
T.
&
Norder
,
H.
2014
Detection of pathogenic viruses in sewage provided early warnings of hepatitis A virus and norovirus outbreaks
.
Applied and Environmental Microbiology
80
(
21
),
6771
6781
.
Lamendella
,
R.
,
Santo-Domingo
,
J. W.
,
Yannarell
,
A. C.
,
Ghosh
,
S.
,
DiGionvanni
,
G.
,
Mackie
,
R. I.
&
Oerther
,
D. B.
2009
Evaluation of Swine specific PCR assays used for fecal source tracking and analysis of molecular diversity of swine specific bacteroidales populations
.
Applied and Environmental Microbiology
75
(
18
),
5787
5796
.
Mara
,
D. D.
,
Sleigh
,
P. A.
,
Blumenthal
,
U. J.
&
Carr
,
R. M.
2007
Health risks in wastewater irrigation: comparing estimates from quantitative microbial risk analyses and epidemiological studies
.
Journal of Water and Health
05
(
1
),
39
50
.
Mead
,
P. S.
,
Slutsker
,
L.
,
Dietz
,
V.
,
McCaig
,
L. F.
,
Bresee
,
J. S.
,
Shapiro
,
C.
,
Griffin
,
P. M.
&
Tauxe
,
R. V.
1999
Food related illness and death in the United States
.
Emerging Infectious Diseases
5
(
5
),
607
625
.
National Research Council (NRC)
1983
Risk Assessment in the Federal Government: Managing the Process
.
National Academy Press
,
Washington, DC, USA
.
National Research Council (NRC)
2012
Water Reuse – Potential for Expanding the Nation's Water Supply Through Reuse of Municipal Wastewater
.
The National Academies Press
,
Washington, DC, USA
.
Olivieri
,
A. W.
&
Soller
,
J. A.
2002
City of Vacaville – Evaluation of Public Health Risks Concerning Infectious Disease Agents Associated with Exposure to Treated Wastewater Discharged by the City of Vacaville EWWTP
.
EOA, Inc., Alexandria, VA, USA
.
Olivieri
,
A. W.
,
Seto
,
E.
&
Soller
,
J.
2007
Application of Microbial Risk Assessment Techniques to Estimate Risk due to Exposure to Reclaimed Waters
.
WateReuse, Final Report WRF-04-001
.
Olivieri
,
A. W.
,
Danielson
,
R. E.
&
Seto
,
E.
2012
City of Vacaville, Easterly Wastewater Treatment Plant Monitoring Microbial Pathogens and Quantitative Microbial Risk Assessment in Effluent and Receiving Water
.
EOA, Inc., Oakland, CA, USA
.
Olivieri
,
A.
,
Seto
,
E.
,
Cooper
,
R.
,
Cahn
,
M.
,
Colford
,
J.
,
Crook
,
J.
,
Debroux
,
J.
,
Mandrell
,
R.
,
Suslow
,
T.
,
Tchobanoglous
,
G.
,
Hultquist
,
R.
,
Spath
,
D.
&
Mosher
,
J.
2014
Risk-based review of California's water-recycling criteria for agricultural irrigation
.
Journal of Environmental Engineering
04014015
.
Ott
,
W. R.
1995
Environmental Statistics and Data Analysis
.
Lewis Publishers
,
Boca Raton, FL, USA
.
Petterson
,
S. R.
,
Ashbolt
,
N. J.
&
Sharma
,
A.
2001
Microbial risks from wastewater irrigation of salad crops: a screening-level risk assessment
.
Water Environment Research
73
(
6
),
667
672
.
Petterson
,
S. R.
,
Ashbolt
,
N. J.
&
Sharma
,
A.
2002
Microbial risks from wastewater irrigation of salad crops: a screening-level risk assessment
.
Water Environment Research
74
(
4
),
411
.
Regli
,
S.
,
Rose
,
J. B.
,
Haas
,
C. N.
&
Gerba
,
C. P.
1991
Modeling the risk from Giardia and viruses in drinking-water
.
Journal American Water Works Association
83
(
11
),
76
84
.
Rose
,
J. B.
&
Gerba
,
C. P.
1991
Assessing potential health risks from viruses and parasites in reclaimed water in Arizona and Florida, USA
.
Water Science and Technology
23
(
10/2
),
2091
2098
.
Rose
,
J. B.
,
Nowlin
,
H.
,
Farrah
,
S. R.
,
Harwood
,
V.
,
Levine
,
A.
,
Lukasik
,
J.
,
Menendez
,
P.
&
Scott
,
T. M.
2004
Reduction of Pathogens, Indicator Bacteria, and Alternative Indicators by Wastewater Treatment and Reclamation Processes
.
Water Environment Research Foundation
.
Report 00-PUM2T
.
Scallan
,
E.
,
Hoekstra
,
R. M.
,
Angulo
,
F. J.
,
Tauxe
,
R. V.
,
Widdowson
,
M. A.
,
Roy
,
S. L.
,
Jones
,
J. L.
&
Griffin
,
P. M.
2011
Foodborne illness acquired in the United States – major pathogens
.
Emerging Infectious Diseases
17
(
1
),
7
15
.
Siefring
,
S.
,
Varma
,
M.
,
Atikovic
,
E.
,
Wymer
,
L.
&
Haugland
,
R. A.
2008
Improved real-time PCR assays for the detection of fecal indicator bacteria in surface waters with different instrument and reagent systems
.
Journal of Water and Health
6
,
225
237
.
Slifko
,
T. R.
,
Friedman
,
D. E.
&
Rose
,
J. B.
1999
A most probable number assay for the enumeration of infectious Cryptosporidium parvum oocysts
.
Applied and Environmental Microbiology
65
(
9
),
3936
3941
.
Solano County Department of Agriculture
2011
Solano County 2010 Crop and Livestock Report
.
Agricultural Commissioner/Weights and Measures
. .
Soller
,
J.
,
Olivieri
,
A. W.
,
Eisenberg
,
J. N. S.
,
Sakaji
,
R.
&
Danielson
,
R.
2004
Evaluation of Microbial Risk Assessment Techniques and Applications
.
Water Environment Research Foundation
,
Project 00-PUM-3, Final Project Report
.
Soller
,
J.
,
Seto
,
E.
&
Olivieri
,
A. W.
2007
Microbial Risk Assessment Interface Tool: User Documentation
(WERF 04-HHE-3)
,
Water Environment Research Foundation
,
Alexandria, VA, USA
.
Standard Methods (SM 9221) for the Examination of Water and Wastewater. Method 9221. Multiple Tube Fermentation Techniques for Members of the Coliform Group. Online. Approved 2001. www.standardmethods.org (accessed 13 February 2017)
.
Standard Methods (SM 9230 D) for the Examination of Water and Wastewater. Method 9230B. Fecal Enterococcus/Streptococcus Groups. Approved 2007. www.standardmethods.org (accessed 13 February 2017)
.
Tanaka
,
H.
,
Asano
,
T.
,
Schroeder
,
E. D.
&
Tchobanoglous
,
G.
1998
Estimating the safety of wastewater reclamation and reuse using enteric virus monitoring data
.
Water Environment Research
70
(
1
),
39
51
.
Teunis
,
P. F.
&
Havelaar
,
A. H.
2000
The beta poisson dose-response model is not a single-hit model
.
Risk Analysis
20
(
4
),
513
521
.
Teunis
,
P. F. M.
,
van der Heijden
,
O. G.
,
van der Giessen
,
J. W. B.
&
Havelaar
,
A. H.
1996
The Dose-Response Relation in Human Volunteers for Gastro-Intestinal Pathogens
.
RIVM Report 1996, 284550002
.
Van Abel
,
N.
,
Schoen
,
M. E.
,
Kissel
,
J. C.
&
Meschke
,
J. S.
2016
Comparison of risk predicted by multiple norovirus dose-response models and implications for quantitative microbial risk assessment
.
Risk Analysis: An Official Publication of the Society for Risk Analysis
37
,
245
264
.
Victoria
,
M.
,
Guimaraes
,
F. R.
,
Fumian
,
T. M.
,
Ferreira
,
F. F. M.
,
Vieira
,
C. B.
,
Shubo
,
T.
,
Leite
,
J. P. G.
&
Miagostovich
,
M. P.
2010
One year monitoring of norovirus in a sewage treatment plant in Rio de Janeiro, Brazil
.
Journal of Water and Health
8
(
1
),
158
165
.