The reuse of graywater and stormwater has the potential to significantly impact the use of the potable water supply. However, the potential presence of a variety of contaminants in these water sources raises concerns about their uses. These concerns may be exacerbated by the lack of federal guidelines for the use of these types of waters and inconsistencies among limited existing state and local regulations or guidelines for the beneficial use of graywater and stormwater (see Chapter 8).
The presence of a contaminant does not automatically equate to significant risk. The situation in which the water is used has a significant impact on the potential risk. When there is little potential for human exposure to the water (such as subsurface irrigation with graywater), the risk will be lower than when the same quality of water is used in higher exposure environments (such as spray irrigation with graywater, where the potential for ingestion exists).
In a literature review, the committee could not find any documented reports of adverse health effects from the use of stormwater or graywater. Sharvelle et al. (2013) surveyed health departments from 15 states that allow graywater reuse, who reported no sicknesses resulting from graywater reuse, including Arizona, which has promoted graywater systems for more than 10 years. Although these results help bound the extent of risk—that is, the risk is unlikely to be large—waterborne infectious diseases tend to be underreported (Yang et al., 2012). Given the many potential exposure routes, linking illness with water supply contamination is challenging, particularly for distributed on-site sources such as graywater or stormwater. Therefore, in this chapter, the committee relies on risk assessment strategies to assess the risks of stormwater and graywater use.
When assessing the potential risk posed by graywater and/or stormwater reuse, a number of factors should be considered. These include the fact that potential risks may vary considerably, based on the specific nature of the water, how the water will be used, the potential exposure to the water, and the characteristics of the environment in which the water will be used. It is also critical to understand that it is not possible to reduce to zero the risks associated with the use of these waters, just as the risks associated with the use and consumption of conventional drinking water are not zero. However, an understanding of the risks associated with graywater and stormwater use can inform responsible decision making regarding the integrated use of all available water resources and facilitate communication with stakeholders.
This chapter provides an overview of the most common methods that can be used to assess human health risk associated with graywater or stormwater use. The chapter also summarizes what is known about human health and environmental risks.
Different methods can be used to calculate the risk associated with a particular activity or contaminant. In terms of potential impacts on human health, the National Research Council (NRC, 1983) risk assessment method is the most commonly used. The following sections describe the major elements of that process—hazard assessment, exposure assessment, dose-response assessment, and risk characterization—although readers seeking a complete description should consult NRC (1983, 2009c). Once the risk assessment is completed, the risk management phase considers the overall benefits and costs of various risk management approaches as well as social justice and legal issues (Figure 5-1). In its 2009 review of the U.S. Environmental Protection Agency’s (EPA) risk assessment process, the NRC reaffirmed the process but stated that more attention should be focused on the design of the risk assessment, especially in the beginning stages of the process. Specifically, there is a
need to bring all stakeholders into the process to determine the major factors to be considered, to define the decision-making context, and to agree upon the timeline and depth needed to ensure that the right questions are being asked in the context of the assessment. This will increase the likelihood that the outcome of the risk assessment will be more useful and better accepted by decision makers and the regulated community (NRC, 2009c).
The first step in the risk assessment process is hazard assessment, in which the contaminants of concern are identified. Such contaminants will vary, depending on the specific situation, including the source water quality and its end use. In the case of graywater, because it is derived from domestic wastes, the contaminants of most concern include human pathogenic microorganisms and inorganic and organic chemicals found in wash waters. In the case of stormwater, the contaminants of concern include metals, pesticides, other organic contaminants, and pathogens, derived from runoff from streets, roofs, lawns, and industrial and commercial areas. Chapter 4 discusses contaminants of concern in stormwater and graywater.
Under sufficient exposures, chemical and microbial contaminants can cause a range of adverse health effects, either acute or chronic. Acute illnesses occur suddenly and severely after only one or a few exposures to a contaminant and are common after exposure to human pathogenic microorganisms. Acute illnesses associated with exposures to waterborne pathogens include gastroenteritis (stomach flu), skin and respiratory infections, and conjunctivitis (pinkeye). Chronic health effects are long-lasting and typically occur after repeated, long-term exposures, most commonly by ingestion of contaminated water. Examples of possible chronic health effects include various cancers and adverse reproductive outcomes.
The next step in the process involves determination of the nature of an individual’s exposure to the hazard(s). Exposure assessment requires knowledge of the amount of contaminant to which the individual is exposed (e.g., the contaminant concentration multiplied by the volume of water ingested), as well as the route(s) and frequency of exposure. This can be one of the most challenging steps in the process, because exposures can vary from day to day, location to location, and person to person.
As discussed in Chapter 4, different source areas can generate different qualities of stormwater, and source control practices substantially impact the quality of graywater or stormwater. For graywater, factors affecting contaminant concentrations include whether best management practices are followed, such as bypassing laundry water when washing diapers or other soiled clothing, not disposing of hazardous chemicals down the sink, and avoiding storage of untreated water. For stormwater, rooftop rainwater capture tends to result in the highest water quality, but some roof and storage tank materials can leach high levels of metals such as copper, zinc, and lead. Runoff from parks and lawns may contain pesticides and fecal microorganisms. Larger source areas, including more roadways and parking areas, can result in higher concentrations of organic contaminants (see Chapter 4).
There may be hundreds of different types of chemicals or microorganisms in graywater or stormwater. Therefore, it is common to analyze the water for indicator or surrogate contaminants, rather than for each of the contaminants themselves. For example, rather than monitoring for human pathogenic microorganisms, which can be extremely costly and time-consuming, indicator microorganisms, such as fecal coliform bacteria, are commonly used. Although this simplifies the monitoring process, the use of surrogates to determine the risk of a water source can be problematic for a number of reasons.
In the case of indicator bacteria, such as E. coli, these microorganisms are present in the intestines of numerous animals as well as humans. Therefore, they are always present in domestic wastewater and are often detected in stormwater (see Chapter 4, Table 4-3). However, pathogenic microorganisms are only present in the intestines when an individual is infected. Additionally, E. coli and enterococci have been detected in a variety of environmental reservoirs (Clary et al., 2014; Byappanahalli et al., 2006; Yamahara et al., 2007). Therefore, the detection of indicator bacteria in water does not necessarily mean that the water contains disease-causing microorganisms. Thus, the reliance on indicators alone limits the capacity to accurately assess human health hazards in a water source.
Additionally, wide variation in concentrations of specific pathogens in graywater exists between individual households, and therefore the use of indicator microorganisms (especially from individual households) to predict pathogen concentrations and associated risks can be problematic (O’Toole et al., 2014). The wide variation in pathogens exists for several reasons, including the numbers, ages, and health of the inhabitants; the fact that fecal shedding intensity and duration varies considerably from person to person; and the source of graywater used (e.g., shower, laundry), which can have a significant impact on the concentrations and types of microorganisms present in the graywater.
There are also situations in which pathogens are present in the absence of the indicator microorganisms, particularly for more environmentally stable pathogens such as protozoan parasites. Many waterborne disease outbreaks have occurred in which indicator bacteria were absent but pathogenic microorganisms were present in sufficient numbers to cause illness (e.g., MacKenzie et al., 1994). Overall, the reliance on indicator microorganisms, although practical from an economic and analytical perspective, may not provide accurate information regarding the water’s safety for its intended end use. Instead, indicator organisms serve as an imperfect but low-cost screening tool that can indicate possible concerns but cannot be used to prove safety.
Water quality treatment (see Chapter 6) may be applied to graywater and stormwater to reduce the concentrations of contaminants present for end uses with a higher degree of human exposure. Water quality may also degrade over time with extended storage, either from contact with tank materials (e.g., zinc [Hart and White, 2006]) or from the growth of microorganisms (Pitt and Talebi, 2012a).
Route(s) and Volumes of Exposure
Individuals can be exposed to contaminants through many different routes, including ingestion, inhalation, and dermal exposure. Some contaminants pose a potential hazard through many exposure routes, while others are harmful only if exposure is through a specific route. Some microorganisms, such as Legionella spp., are transmitted through inhalation. Others can cause harm through dermal exposure (e.g., Pseudomonas, Schistosoma) or exposure of the mucous membranes (e.g., adenovirus), resulting in eye and ear infections. For many microorganisms that are transmitted through water (e.g., Salmonella, Cryptosporidium, noroviruses), ingestion is the primary exposure route of concern. In addition to exposure from the direct ingestion of water, exposure can also result from consumption of food crops that
have been spray irrigated with graywater or stormwater that contains pathogens or chemicals. In determining the dose of the hazard to which an individual is exposed, it is critical that all potential exposure routes be considered so that an accurate risk assessment can be made.
The volume of water to which the individual is exposed is often determined by the specific exposure scenario. There are no standard exposure volumes for the different types of possible exposure. The District of Columbia’s Department of the Environment (DDOE, 2013) developed exposure volumes based on stormwater use and exposure conditions that vary from 0.01 mL from ingestion of aerosol spray from toilet flushing to 200 mL from ingestion of water in a swimming pool (Table 5-1). Other organizations may assume different exposure volumes.
In addition to the volume of water to which an individual is exposed, certain activities may affect the amount of contaminants present at the point of exposure. For example, foods that are irrigated with stormwater and then consumed uncooked will expose the consumer to a higher number of pathogens compared to the same food if it is consumed after cooking. Likewise, foods that are spray irrigated with graywater or stormwater have a higher probability of containing pathogenic microorganisms than food crops that are watered via subsurface irrigation (with the exception of root crops).
Frequency of Exposure
In the exposure assessment, the frequency with which the individual will be exposed to the hazard is another critical piece of information for consideration. In some situations, there may be only a single exposure; for example, an individual who inadvertently swallows some water while playing in a fountain on a hot day. In other cases, exposure might be on a daily basis, such as in the case of an individual who is exposed to aerosol spray from graywater used for toilet flushing. Frequency of exposure will also vary considerably depending on the water use and exposure conditions. For example, DDOE (2013) assumed that an individual would be exposed to aerosols via toilet flushing more than 1,000 times/yr but would accidentally ingest 100 mL of stormwater as a result of using stormwater for home lawn or garden spray irrigation only a single time per year (Table 5-1). Where data are lacking on the frequency of exposure, conservative assumptions are often made.
Other considerations related to the exposure of the individual to the hazard include those of scale—both temporal and spatial. The time that elapses from the release of the con-
TABLE 5-1 Exposure Assumptions Based on Stormwater Use and Exposure Conditions Developed by the District of Columbia’s Department of the Environment
|Stormwater Use||Route of Exposure, Conditions||Exposure Assumptions|
|Volume ingested (mL) in a single exposure||Events per year|
|Home lawn or garden spray irrigation||Ingestion of aerosol spray from typical watering||0.1||90|
|Ingestion after contact with plants/grass||1||90|
|Infrequent inadvertent ingestion of stormwater||100||1|
|Open space or municipal park drip or spray irrigation||Ingestion via casual contact with irrigated grass (picnic, walking pet)||0.1||32|
|Ingestion via low-intensity sports on irrigated field (golf, Frisbee)||1||32|
|Ingestion via high-intensity sports on irrigated field (baseball, soccer)||2.5||16|
|Ingestion on playground by child (frequent hand-to-mouth activity)||4||130|
|Indirect ingestion of spray from public fountain with spray element||0.1||130|
|Infrequent ingestion of public fountain water from standing pool on hot days||4||130|
|Home garden drip or spray irrigation||Ingestion of irrigated vegetables and fruit||7||50|
|Commercial farm produce drip or spray irrigation||Ingestion of irrigated vegetables and fruit||10||140|
|Home car wash spray application||Ingestion of water and spray||5||24|
|Commercial car wash spray||Ingestion of water and spray by car wash operator||3||250|
|Toilet||Ingestion of aerosol spray||0.01||1100|
|Washing machine use||Ingestion of sprays||0.01||365|
|Fire fighting||Ingestion of water and spray||20||50|
NOTE: In a correctly designed subsurface irrigation system (with no surface ponding and no application to food crops) no water would be ingested.
SOURCE: DDOE (2013).
taminant to the exposure may have a significant impact on the concentration or nature of the contaminant. For example, human pathogenic microorganisms have a finite lifetime in the environment. This may range from less than a day to several months or years, depending on the specific microorganism and the environmental conditions. Chemical contaminants may degrade over time, and depending on the specific chemical, the transformation products could be more or less harmful than the parent compound. Modified risk assessment models have been developed to include dynamic modeling (see, e.g., Eisenberg et al., 2004), as well as to allow for the effects of environmental factors on the concentrations of the contaminants (see, e.g., Whelan et al., 2014)
Spatial scale is also important. If the graywater or stormwater is collected over a large area, then a large number of individuals/households/businesses will contribute to the water’s composition. This may result in the water containing microorganisms or chemicals that it would not have contained had the contributing area been smaller. The project’s scale will also affect the number of people who may be exposed to potential hazards in the water. At the household scale, graywater use does not significantly increase risk of illness from pathogens because there are many other pathways for spreading communicable illnesses among members of the same household (Maimon et al., 2010). However, untreated graywater used in larger scale (e.g., multi-residential) projects could substantially increase risk, because graywater use creates exposure pathways between infected and uninfected individuals that otherwise did not exist. Therefore, larger-scale projects will be more likely to involve some type of treatment than will projects that occur on an individual homeowner’s property (see Chapter 6). A project that occurs over a large spatial scale may also, by necessity, involve a longer time scale, so attenuation of contaminants may occur as a result.
After the amount of the contaminant to which the individual will be exposed is known, it is necessary to understand the effect that that amount of the contaminant will have on the exposed individual—in other words, what response will a specific dose produce? Typically, information on the dose-response relationship of a particular contaminant is obtained from published literature values, rather than from conducting a dose-response study for each contaminant in each specific situation. Box 5-1 describes the approaches used to determine dose-response relationships for individual microbial or chemical contaminants.
For both chemicals and microorganisms, however, exposure is commonly to mixtures of the contaminants, rather than to single contaminants. Unfortunately, little is known about the effects on the dose-response relationship of contaminants when an individual is exposed to more than one contaminant at a time. Exposure to a mixture of contaminants can cause effects that are equal to, less than, or greater than that of the individual components, and understanding these effects is the focus of ongoing research (see, e.g., Backhaus, 2014; Jarvis et al., 2014).
The next step in the process is to calculate the risk, considering the contaminant of interest, the level and frequency of exposure, and the dose-response relationship. The risk may be expressed in several different ways. For microorganisms, the acute risk of infection, illness, or mortality can be presented. In the case of many chemicals, the risk of cancer over the course of a lifetime is the endpoint of interest.
The different outcomes associated with the contaminants of interest can make it difficult to compare risks and to make decisions regarding possible risk-risk tradeoffs. For example, how does one compare the risk of getting cancer to the risk of getting a norovirus infection? Cancer is typically acquired after long-term exposure to a relatively low dose of a chemical, while hepatitis can be acquired from a single exposure to contaminated food or water. Infection caused by norovirus is typically self-limiting, while cancer can be a short-term or long-term illness.
One way to more easily compare the health effects caused by exposure to different types of contaminants with different health effects is to quantify the burden of disease morbidity and mortality through the use of disability-adjusted life years (DALYs). Per the World Health Organization, “One DALY can be thought of as one lost year of ‘healthy’ life. The sum of these DALYs across the population, or the burden of disease, can be thought of as a measurement of the gap between current health status and an ideal health situation where the entire population lives to an advanced age, free of disease and disability.”1 DALYs are calculated by adding the number of years of life lost due to disability (for people who are living with the adverse health effect) to the number of years of life lost due to premature deaths across the exposed population. Results for specific scenarios can then be evaluated in the context of acceptable risk targets. For example, Australia developed guidelines for potable reuse (NRMMC et al., 2006) that set a tolerable microbial risk of 10-6 DALYs per person per year (approximately 1 diarrheal illness per 1,000 people per year).
The process of calculating the risk from exposure to a contaminant is a scientific process. Once the risk is calculated, the determination of whether that risk is acceptable involves not only science but also technological feasibility, economics, politics, and societal factors. Interested readers should consult NRC (2009c) for a detailed discussion of risk management.
Several cities have taken a tiered approach for managing the risks of harvested stormwater use, with increasing levels of regulation or treatment required with increasing exposures (see Los Angeles County Department of Public Health tiered framework in Table 8-2). The District of Columbia recently developed a quantitative process for tiered risk assessment and management for nonpotable uses of harvested stormwater (DDOE, 2013). The process (shown in Figure 5-2) uses tiered risk-based screening levels for
individual chemical and microbial contaminants and four exposure classes (low to severe; see Tables 5-2 and 5-3). The process allows planners to compare typical stormwater concentrations against tiered risk-based levels, with higher contaminant concentrations allowed for activities with low exposure. For cases when stormwater concentrations exceed these screening levels (i.e., the risk is considered unacceptable), the risk management process (Figure 5-2) requires treatment or additional justification for why treatment is not needed. Note that the examples in Table 5-2 are based only on direct human exposure—not ecological risk or indirect human exposure. Consideration of other pathways, including the use of stormwater to recharge an aquifer used for water supply, would include different exposures and could result in different chemical concentration limits.
Australia also adopted risk-based guidelines for managing the beneficial use of stormwater and graywater (NRMMC et al., 2006, 2009). The guidelines include treatment recommendations for different applications based on 95th percentile concentrations from existing pathogen data or conservative estimates where data were lacking. In addition, the risk management framework includes a commitment to responsible use and recommendations for preventive measures, management of system failures, employee and community awareness, evaluation, review, and continual improvement.
Several quantitative human health risk assessments have been published on graywater and stormwater. The most relevant to the uses considered in this report are presented here.
In a review of onsite reuse of graywater for irrigation, Maimon et al. (2010) attempted to summarize the potential risks from exposures to human pathogenic microorganisms, using rotavirus as the example. Their scenario included exposures from the accidental ingestion of untreated graywater (100 mL; once/yr), the routine ingestion of the graywater from touching irrigated plants (1 mL; 90 times/yr), and in-
TABLE 5-2 Examples from the DDOE Chemical Risk-based Levels for Stormwater Use Based on Human Exposure Category and Comparison with DDOE’s Drinking Water Standards
|Contaminant (μg/L)||Drinking Water Standard||Direct Human Exposure Category|
aHigh exposure includes applications such as commercial farm produce drip or spray irrigation, firefighting, and commercial car washes.
bMedium exposure includes public fountains, spray irrigation of playgrounds, home garden spray irrigation, home drip irrigation of fruits and vegetables, and home car washing.
cLow exposure includes toilet flushing, washing machine use, and open space spray irrigation of parks (non-playgrounds).
NOTES: For each of the exposure classes (grouped by the amount of water ingested), the risk-based levels represent the contaminant concentrations corresponding to a cancer risk of 10-6 or a non-cancer hazard index of 1.0. DDOE states that although EPA suggests a discretionary cancer risk level between 10-4 and 10-6, it selected this cancer risk level to account for the presence of multiple contaminants.
SOURCE: DDOE (2013).
TABLE 5-3 Examples from the DDOE Microbial Risk-based Levels for Stormwater Use Based on Human Exposure Category
|Contaminant (μg/L)||Swimming||Direct Human Exposure Category|
|E. coli (CFU/100 mL)||126a||1,714||4,615||50,000|
aRSLs correspond to a risk level of 8 in 1,000 of developing a gastrointestinal disease.
bRSLs correspond to a 10-6 risk level of developing a gastrointestinal disease.
NOTE: See Table 3-2 for examples of exposure categories.
SOURCE: DDOE (2013).
gestion of the graywater spray (0.1 mL; 90 times/yr). Using the assumption that the acceptable annual risk of infection was 1.4 x 10-3 infections per person per year, they calculated the acceptable safe dose of rotavirus as 0.0024 viruses per exposure if the exposure occurs from the accidental ingestion of 100 mL of untreated graywater once per year or 0.00014 viruses per exposure if the exposure is to 1 mL and occurs 90 times per year (e.g., from the routine ingestion of the graywater from touching irrigated plants). Using this acceptable dose and the estimated concentrations of rotavirus in graywater (based on three different assumed relationships between rotavirus and measured E. coli concentrations), the maximum volume of graywater that can be “safely” ingested in a single exposure occurring once per year was calculated to range from 0.003 mL (assuming 0.8 rotaviruses/mL) to 0.24 mL (assuming a rotavirus concentration of 0.01 rotaviruses/mL). Maimon et al. (2010) concluded that most of the exposure scenarios examined would result in exposure that would exceed the acceptable safe dose and that using graywater for spray irrigation or food crop irrigation would necessitate disinfection to protect against rotavirus.
The committee could not find any risk assessments of exposure from surface drip irrigation (no landscape cover) with graywater. Such exposures would presumably be higher than those of subsurface irrigation, but exposure estimates and the increase in risk at the household scale, where many vehicles of disease transmission already exist, are needed to inform safe design practices.
Using a different approach, Ottoson and Stenstrom (2003) calculated the risks associated with exposure to human pathogens in treated (but not disinfected) graywater.2 Because fecal indicator bacteria can overestimate the amount of fecal material in the graywater, the authors used measured concentrations of coprostanol (a fecal sterol) at the site as an estimator of the amount of fecal material in the graywater, and estimated pathogen concentrations based on the prevalence of pathogens in the general population. Risks were modeled in three scenarios—direct contact, spray irrigation of sports fields, and daily consumption of groundwater recharged
2 The treatment system included settling tanks, activated sludge, a biofilter, and surface storage in ponds.
with treated graywater. In the case of groundwater recharge, it was assumed that the microorganisms traveled through a 3-m-thick vadose zone and that reductions in the numbers of microorganisms occurred during transport as well as in the groundwater during retention. The study considered several pathogens, including Campylobacter jejuni, Cryptosporidium parvum, Giardia lamblia, rotavirus, and Salmonella, and modeled rates of natural attenuation in the subsurface. These organisms were chosen as conservative representatives for the behavior of fecal bacteria, viruses, and protozoan parasites. Lack of disinfection resulted in sizable risk across a range of exposures, although the risks ranged considerably by pathogen and exposure scenario. The highest risks across all scenarios were associated with rotavirus, ranging from 0.25 probability of infection from a single event with direct contact to 0.63 annual probabilities for both spray exposure and groundwater consumption after 1 month retention (see Table 5-4). Risks from groundwater consumption decreased notably over time and were negligible (less than 10-11) for all organisms after 6 months retention. In addition, the effects of removal during transport through the soil are clearly seen in the much lower risks for the larger protozoan parasites (Cryptosporidium and Giardia) compared to those for the smaller, more easily transported bacteria and viruses.
Calculating risk based on indicator concentrations presents many challenges. Ideally, risk assessment would be based on direct pathogens data. This effort, however, has been limited by the relative lack of quantitative data on the concentrations of pathogenic microorganisms in graywater and stormwater. One report on the concentrations of pathogens in untreated graywater was located (see Table 4-2; Birks and Hills, 2007), and these data were used to perform example calculations of infection risk infection for two different scenarios in which individuals could be exposed to graywater (see Box 5-2), although these limited data should not be assumed to describe the risk of graywater use generally.
The committee could not find any published chemical risks assessments for graywater. Debroux et al. (2012) recently reviewed the potential human health effects associated with exposure to nonregulated trace organic compounds (including pharmaceuticals and personal care products) in recycled municipal wastewater from potable and nonpotable reuse. They concluded that none of the risk assessments conducted over the past 10 years found any adverse human health effects or significant risks. The risk of ingesting trace organic compounds after uptake into food crops is not well understood and has been identified as a research need for nonpotable reclaimed water (NRC, 2012a). Concentrations of trace organic chemicals in graywater may differ from those in recycled municipal wastewater. Compared to nonpotable recycled water, graywater would contain higher concentrations of personal care products (which are primarily derived from sink and shower water, and therefore would be diluted by other water sources in wastewater) and lower concentrations of pharmaceuticals, which are primarily excreted in urine and occasionally flushed directly down the toilet. In addition, the treatment levels for potable and nonpotable reuse tend to be much greater than those for typical graywater, if treatment is even applied, although the exposure levels for typical potable and nonpotable reuse projects are also much greater. Nevertheless, this analysis provides potentially useful reference information. The risks of long-term, low-level exposures to mixtures of trace organic compounds remain unclear for conventional drinking water sources, and typical nonpotable graywater exposures under best management practices would be lower.
Very limited information is available in the refereed literature on the risks of various uses of stormwater. A study of the risks associated with chemicals and pathogens in stormwater that was treated in a reed bed and then recharged into
TABLE 5-4 Comparative Risks of Infection across Pathogens from Treated, Non-disinfected Graywater
|Pathogen||Risk from Single Exposure||Annual Risk|
|Direct contact (1 mL/event)||Spray irrigation (1 mL exposure, 26 times/yr)||Groundwater recharge (2 liter/day consumption)|
|1 mo. retention||3 mo. retention||6 mo. retention|
|Campylobacter jejuni||0.00158||0.00316||0.00316||3.2 x 10-9||<10-11|
|Cryptosporidium parvum||0.0000251||0.0002||7.9 x 10-8||<10-11||<10-11|
|Giardia lamblia||0.00000316||0.0000316||1.26 x 10-8||<10-11||<10-11|
SOURCE: Based on data in Ottoson and Stenstrom (2003).
an aquifer prior to recovery and use was performed by a group of scientists at CSIRO (Page et al., 2013). Risks from three pesticides (i.e., diuron, simazine, and chlorpyrifos) and three human pathogens (i.e., rotavirus, Campylobacter, and Cryptosporidium) were calculated. Box 5-3 provides the details of this quantitative risk assessment.
Albrechtsen (2002) published data on Cryptosporidium in roof runoff used for toilet flushing. These data are used in example calculations of risk for toilet flushing with captured rainwater (see Box 5-4). The results suggest a 1 in 2,170 annual chance of Cryptosporidium infection using the data reported and exposure assumptions from DDOE (2013).
These very limited data suggest that the risks of stormwater use should not be taken lightly in applications where human exposures are likely. They represent only two case studies, but they show how a risk framework can be useful for informing project treatment needs.
The risk assessment calculations discussed in the previous section represent risks incurred under routine exposures, but the risks of treatment failure, cross-connection between potable and nonpotable water lines, or inadvertent groundwater contamination also need to be considered as part of an overall understanding of risk.
To date, there have been no reported adverse health effects resulting from failures of graywater reuse systems (Sharvelle et al., 2013). However, failures in highly engineered systems will eventually occur, and the potential impacts of such events must be understood. In systems with substantial human exposures, real-time monitoring or redundant treatment systems can be developed to reduce these risks (see Chapter 6).
No graywater or stormwater cross-connections have been reported, but accidental or intentional cross-connections between nonpotable and potable water supplies are a public health concern, particularly for non-disinfected water supplies. There have been several reports in which cross-connections have occurred between potable water systems and nonpotable reclaimed water, some of which resulted in illnesses. For example, in 2007 it was reported that 12 individuals who worked at Melbourne Water became ill after consuming water in an administration building from a tap
that had been mistakenly connected to a reclaimed water pipe (Herald Sun, 2007). Another incident in Queensland, Australia, involved 630 homes in a housing development with dual plumbing for toilet flushing and outdoor use (WQRA, 2010). Within 2 days of the nonpotable reclaimed water being delivered to the homes, complaints of foul taste and odor were received from residents and a cross-connection was discovered. Hambly et al. (2012) described several cases in Australia in which cross-connections between potable water and reclaimed water occurred. Guidance and testing programs have been developed to minimize accidental cross-connections in water systems (AWWA, 2009; EPA, 2003) that could be applied to dual-plumbed graywater or stormwater use systems, as appropriate. The California Plumbing Code recommends annual cross-connection inspections and testing for permitted graywater and stormwater use systems.
Another potential failure that could affect human health is the unplanned recharge of stormwater or graywater into an aquifer used for a drinking water supply. Or planned aquifer recharge may not adequately remove contaminants before they reach a potable aquifer. Box 5-3 describes one published risk assessment for groundwater impacted by stormwater-derived contaminants. Because of the magnitude
of the potential health risks associated with long-term ingestion of contaminated water, the committee describes what is known about the potential for groundwater contamination from graywater and stormwater in this section. Chapter 4 discusses contaminants of greatest concern to groundwater infiltration projects.
There is some concern about graywater constituents leaching to groundwater when graywater is applied for irrigation, particularly in cases where graywater is applied over large areas at rates greater than required based on evapotranspiration. When graywater is used for irrigation, constituents of primary concern for groundwater quality include nitrogen, salts (including sodium, chloride, and boron), pathogens, and organic contaminants from cleaning or personal care products. However, the actual human health risk would depend on many factors, including contaminant concentrations in graywater (see Chapter 4), irrigation rates, potential for contaminant sorption or biodegradation, soil and aquifer characteristics that affect contaminant transport, depth to the water table, and distance to the point of groundwater withdrawal. In a soil-column study, Negahban-Azar et al. (2013) showed that salts have potential to leach through graywater-irrigated soil. This would pose the most concern in soils with high sand content and/or where high infiltration rates are observed. Stevens et al. (2011) noted a risk for salt transport to groundwater, which would pose the most risk in arid climates where evapotranspiration is high. Although leaching of phosphorus is generally not a concern because of limited mobility in soil, leaching of inorganic nitrogen is possible. Surfactants leached from columns ranged from 0 to 20 percent of what was added, resulting in low concentrations in leachate, although surfactant concentration in leachate increased over the 17-month duration of experiments. To reduce these risks, at least two states (i.e., New Mexico, Arizona) recommend that household graywater not be applied in areas where the seasonally high groundwater table is less than 5 ft from the application point (Sharvelle et al., 2013). Source control (best management practices) can also be used to reduce the concentrations of contaminants in graywater used for irrigation (see Chapter 4).
As described in Chapters 1 and 2 of this report, there is increased interest in urban stormwater capture and enhanced infiltration through engineered structures as a means to manage urban stormwater and peak flows, reduce non-point pollution, and replenish groundwater supplies. Because urban runoff also contains pollutants, there exists the potential to contaminate groundwater during infiltration, especially in the future if large volumes of urban stormwater are captured for groundwater recharge, thereby increasing human health risks for current or future groundwater users. As outlined in Table 4-4, pollutants in urban stormwater include salts, suspended solids, nutrients (e.g., nitrogen and phosphorus), heavy metals (e.g., copper, lead, chromium, nickel, and zinc), organic compounds from automotive use and biocide applications, and pathogens. The likelihood for these pollutants to migrate through the soil and contaminate groundwater during stormwater infiltration depends on a number of factors including the infiltration rate, permeability and character of the soil or infiltration media, biological activity in the subsurface, depth to the water table, and the properties of the pollutants. With the growing practice of “enhanced infiltration” for groundwater replenishment, there is concern that these practices may put the groundwater at risk from chemical and microbial contaminants (Nieber et al., 2014).
Risk Factors for Chemical Contamination of Groundwater. Chemical pollutants in urban stormwater that are most likely to contaminate groundwater are those that are relatively non-volatile, hydrophilic (dissolve or mix easily in water), ionic, and non-sorbing. Soluble, non-sorbing salts, such as road deicing compounds, will flow with the infiltrating runoff and not be removed during infiltration (Bannerman et al., 2014; Mullaney et al., 2009). The existing literature on the fate of organic compounds in stormwater infiltration systems, however, is much less than for heavy metals (Weiss et al., 2008). Mikkelsen et al. (1996) studied metal movement in percolating stormwater at two infiltration systems in Switzerland, and metal concentrations in the water were found to decrease rapidly to background conditions within 1.5 m of depth. In Perth, Australia, Appleyard (1993) reported sediment concentrations of 3,500 ppm of lead in stormwater infiltration basins because of strong sorption to iron oxides. Nightingale (1987) studied water quality beneath five stormwater recharge basins in Fresno, California. The basins drained single-family residential neighborhoods and captured winter stormwater; sampling at depths up to 26 m showed no contamination except for trace levels of diazinon.
The potential for subsurface transport of metals and sorbing organic chemicals into groundwater depends on the character of the media and whether fine solids are retained by filtration. Stormwater contaminants are unlikely to be removed if stormwater infiltrates directly into coarse media or karst formations with extremely high percolation rates with little or no opportunity for attenuation or filtration (e.g., Stephensen et al., 1999). A study of 15 dry wells—precast concrete structures with open bottoms resting on and surrounded by crushed
stone for subsurface disposal of stormwater—found no subsurface changes in water quality for filtered forms of copper, lead, and zinc or for E. coli and enterococci, even after percolating through gravel and at least 4 feet of urban subsurface soils (Pitt and Talebi, 2012a). Similarly, groundwater is more at risk from stormwater contaminants in areas where the soil is sandy and the groundwater is shallow (Fischer et al., 2003). In a study of groundwater beneath 16 stormwater detention basins in New Jersey that had sandy and unconsolidated soils, the sampling showed elevated levels of petroleum hydrocarbons, as well as the herbicides metolachlor and prometon. Fischer et al. (2003) concluded that high recharge in urban stormwater basins may impact groundwater even when the constituent concentrations are low.
Risk Factors for Pathogen Contamination of Groundwater. Few studies have examined the efficacy of infiltration practices for removing pathogenic organisms (Weiss et al., 2008). Because pathogens are typically associated with particles, physical straining through the soil or engineered media may remove pathogens just as sand filters are used in water treatment, although the effectiveness will depend on the organism and the porous media properties. Straining is most effective for protozoan pathogens (greater than 3 microns), such as Giardia lamblia and Cryptosporidium parvum, and larger bacteria (approximately 1-2 microns) than viruses, which are too small (0.02 to 0.08 microns) to be effectively removed by filtration through porous media. However, virus removal during passage through porous media may also occur via attachment to soil particles or aquifer material, depending on the specific characteristics of the virus and the environment (see Schijven and Hassanizadeh, 2000). Clark and Pitt (2007) documented pathogen contamination of groundwater due to infiltration practices that included stormwater sand filters. High bacterial and virus concentrations were found in groundwater on Long Island where the groundwater table was close to the land surface. In contrast, the Orange County Water District performed extensive analysis of viruses and protozoa in groundwater in the Santa Ana River basin, where reclaimed wastewater and stormwater recharges groundwater, and concluded that the surface recharge is not a significant source of pathogens to groundwater (NWRI, 2004; OCWD, 2004).
Inactivation also plays a key role in determining whether pathogens present in the infiltrating water will survive to contaminate the underlying groundwater. A large body of literature describes the numerous factors that affect the length of time that microorganisms can survive in the subsurface (see, e.g., John and Rose, 2005). Typically, microorganisms survive longer at cooler temperatures and near-neutral pH conditions (John and Rose, 2005; Yates et al., 1988). Some bacteria are reported to survive longer in acidic soils and soils with a large amount of organic matter, as long as several months (Pitt et al., 1999). Viruses typically survive longer than bacteria (Yates et al., 1988; Sidhu et al., 2010). As a general rule, protozoan parasites and helminth ova survive longer in the environment than the other types of enteric pathogens because of their environmentally resistant non-metabolically active forms (e.g., Cryptosporidium parvum oocysts).
In addition to health effects on humans, the potential impacts of contaminants in graywater and stormwater on the environment should be considered when significant environmental exposures are likely. Assuming best practice are followed and graywater is not ponded or discharged directly to surface water during irrigation, aquatic organisms should not experience significant contaminant exposures. For stormwater capture projects, ecological exposure scenarios may not be common, but for projects with surface impoundments or wetland treatment cells (see Box 2-6), aquatic life may become an intentional or unintentional component of the project, where they could be impacted by trace metals or organic contaminants (Grebel et al., 2013). Risks could also include algal blooms and low dissolved oxygen associated with elevated nutrients, leading to fish die off in stormwater ponds. The committee found only a few analyses of ecological risks involving graywater (Gross et al., 2005; Maimon et al., 2010), although more work has been done on ecological risks associated with stormwater ponds (reviewed in Tixier et al., 2011).
Stormwater retention ponds and wetlands typically provide entirely new aquatic habitat, enhancing biodiversity in the urban environment (Brand and Snodgrass, 2010; Le Viol et al., 2009). However, in these settings stormwater-derived contaminants in water and sediment can exceed probable-effect levels for aquatic life (VanLoon et al., 2000; Wik et al., 2008) and cause adverse ecological impacts, such as lethal and sublethal effects on embryonic and larval amphibians (Bishop et al., 2000b; Snodgrass et al., 2008). Thus, stormwater ponds tend to be populated by more pollution-tolerant organisms (Wik et al., 2008) and have low species richness (Bishop et al., 2000a). A central issue when considering ecological risk, therefore, is to determine what level of impairment is acceptable in consideration of the new environmental benefits provided. Without stormwater capture features, urban streams are impacted by the same contaminants as well as extremes in flow associated with runoff from largely impervious surfaces. The EPA (1998) developed a framework for ecological risk assessment that can be used to evaluate the probability of adverse ecological effects, which includes comparison of field data to reference sites. Selection of an appropriate reference site in the urban envi-
ronment is critical because the degree of impairment in the reference site determines the magnitude of calculated risk and the habitat objectives of the stormwater project (Tixier et al., 2011). This challenge has led some researchers to call for new strategies to understand the ecological functioning of stormwater-based habitats to better develop ecological objectives and management measures (Lafont et al., 2007; Tixier et al., 2011).
In addition to assessing potential toxicological effects on aquatic life, graywater or stormwater irrigation projects should be aware of potential water quality impacts to plants and to soil properties. Constituents of greatest concern for stormwater include salinity, sodium, chloride, and metals (EPA, 2012a). For graywater, contaminants of concern include nitrogen, phosphorus, salinity, boron, sodium, chloride, and surfactants (Sharvelle et al., 2013). As discussed in Chapter 4, salinity is a key concern, because of potential negative effects on plant health and because high quantities of sodium relative to calcium and magnesium (measured as the sodium adsorption ratio [SAR]) can impact soil structure, making the soil less permeable and more erodible, particularly soils with a high clay content. In a study of seven U.S. households using graywater for irrigation, soil SAR levels were elevated in graywater-irrigated soil compared to freshwater-irrigated soil, and the soil SAR at a site with more than 30 years of graywater irrigation was 2-22 times greater than at the control site. However, at all sites, the soil SAR was less than 5, low enough to prevent any harmful effect for plants’ water uptake (Sharvelle et al., 2012). Salinity is likely to be a greater concern in arid climates with high evapotranspiration rates and fewer rainfall events to flush the soils of salt build-up (Stevens et al., 2011).
Excess concentrations of boron, metals, and surfactants can be toxic to plants, and surfactants in graywater can also cause the soil to become more hydrophobic, impacting plant health (Garland et al., 2000; Gross et al., 2005). A large percentage of the applied surfactants have been shown to biodegrade in the soil, and although surfactant concentrations are elevated in graywater-irrigated soils compared to control sites, surfactants do not accumulate in soils over time (Sharvelle et al., 2012). Boron has been found to accumulate in soils irrigated with graywater for more than 5 years (Negahban-Azar et al., 2012), which may be a concern. In addition, some plants may be more sensitive to contaminant effects than others. For example, Sharvelle et al. (2013) observed reduced growth or adverse plant health effects from graywater irrigation on only 3 species (i.e., avocado, lemon tree, and Scotch pine) out of 22 studied.
Although indicator organisms are present in graywater, graywater-irrigated soil has not been found to consistently contain elevated concentrations of indicator organisms compared to potable water-irrigated soil (City of Los Angeles, 1992; Negahban-Azar et al, 2012; Sharvelle et al., 2012). In these studies, animals were known to contribute to indicator organisms in both graywater- and potable water-irrigated areas, and it was not possible to differentiate the contribution of indicator organisms from graywater from other natural contributions of indicator organisms. Antimicrobial chemicals have been found to accumulate in graywater-irrigated soil (Negahban-Azaer et al., 2012). The effects of antimicrobial chemicals on soils are not well understood, and there is increasing concern that they might contribute to the abundance and persistence of antibiotic resistance in soil microorganisms (Auerbach et al., 2007).
TABLE 5-5 Published Water Quality Guidelines for Irrigation
|Hazard||Australian and New Zealand Trigger Values||2012 EPA Water Reuse Guidelines|
|Salinity (µS/cm)||<950 to 12,200a||<700 to 3,000|
|Chloride||<175 to >700 mg/L|
|Sodium||<115 to >460 mg/L|
aDepending on crop sensitivity.
NOTE: Trigger values are established to minimize soil build-up of contaminants and prevent direct toxic effects to crops. Australian trigger values for metals reflect the long-term trigger values, assuming tolerance for 100 years.
Guidelines for irrigation quality to minimize adverse effects have been established by the EPA and in Australia and New Zealand (see Table 5-5). Most stormwater outfall samples fall within these guidelines (see Table 4-4), although household-scale projects capturing runoff from roofs with certain materials could exceed guidelines for metals (see Box 4-1). Source control strategies can be used to control boron and sodium in graywater and salts in stormwater (see Chapter 3).
Although no documented reports of adverse human health effects from the beneficial use of stormwater or graywater have been identified, additional examination of risk is necessary to support safe and appropriate design and implementation of stormwater and graywater use systems. This effort will be especially important as the use of graywater and stormwater becomes more widespread, particularly in water-scarce regions.
Risk assessment provides a means to determine “fit-for-purpose” water quality criteria or treatment needs based on human exposures. Risk from graywater or stormwater is a factor of chemical or microbial concentrations and exposure (typically, the amount or water ingested). Thus, unlike drinking water criteria, which are established based on 2 liters of water consumed per day, criteria for applications with minimal human exposures might allow for much higher concentrations of contaminants in graywater or stormwater and still result in acceptably low health risks. Risk assessment tools provide a ready means for developing such criteria for many chemicals and microbes for which drinking water criteria exist. As nonpotable on-site use of graywater and stormwater becomes more common, additional public health risk communication efforts would be beneficial to help the public understand risk-based treatment objectives and appropriate safeguards.
Considering the low exposures in most nonpotable graywater and stormwater applications, pathogens represent the most significant acute risks. Available risk assessments and the committee’s risk calculations using limited, observed pathogen data and various possible exposure scenarios suggest that disinfection is necessary for many uses of graywater, including spray irrigation, food crop irrigation, and toilet flushing, to protect human health. Subsurface landscape irrigation with graywater does not pose significant risk, if best practices are followed, because human exposure is minimized. These findings are consistent with most regulatory guidance (see Chapter 8), although the risk of surface drip irrigation (without landscape cover) at the household scale remains unresolved. Limited data on pathogens in roof runoff suggest that treatment may also be needed, even for low levels of human exposure, such as toilet flushing, although more research on pathogens in roof runoff is needed. Chemicals become of concern in groundwater infiltration projects, where drinking water supplies could be impacted.
Extremely limited data are available on the pathogen content in graywater and roof runoff, which precludes a full assessment of microbial risks. Most water quality monitoring assesses microbial indicator data, and microbial risk assessments are conducted using assumed relationships between the concentrations of indicator microorganisms and pathogenic microorganisms. Consistent relationships between surrogates and contaminants have not been established for graywater or stormwater. Such relationships would be extremely variable at smaller scales but, even at large scales, could differ substantially from traditional indicator-pathogen relationships derived for municipal wastewater. This is a particular concern for roof runoff, which may include microbial indicator organisms from the waste of animals that do not transmit human pathogens. Therefore, the actual concentration of human pathogens in the water—and the associated risk of exposure to that water—may be much higher or lower than that calculated using the concentration of indicator microorganisms.
Enhanced infiltration of stormwater for groundwater recharge poses risks of groundwater contamination and necessitates careful design to minimize those risks. The risk of groundwater contamination from stormwater recharge is related to the contaminants present, any pretreatment processes installed, the capacity for the subsurface soil and engineered media used in the infiltration basin to remove them, and the proximity to groundwater used as a drinking water supply. Dry wells, which directly inject water into the subsurface and surface infiltration through sandy soils do not effectively attenuate chemical contaminants, and treatment prior to injection might be needed to prevent groundwater contamination.
As the uses of graywater and stormwater become more common, care to prevent cross-connections needs to be taken. Cross-connections between potable and nonpotable water systems can expose residents to elevated risks. No reports of adverse health effects from cross-connections between graywater or stormwater systems and potable systems have been documented, but there have been reports of cross-connections between reclaimed water and potable systems that have resulted in illnesses. Regulatory guidance and inspection criteria can help reduce cross-connection risks.
Environmental impacts from the outdoor use of graywater and stormwater generally appear low, but risks depend upon several factors, including water quality, application rates, and plant or animal species exposed.
Effects of irrigation on plant and soil health can occur from salts, boron, and metals, but source control practices and appropriate irrigation rates can reduce these impacts. If not controlled at the source, then long-term build-up of boron or salt can pose risk to plant and soil health, depending on soil and climatic conditions. Constructed stormwater ponds and wetlands typically contain elevated contaminant levels sufficient to impair reproduction among some aquatic species, often leading to a habitat dominated by pollution-tolerant organisms. Such ecological affects may be acceptable, considering the overall environmental benefits provided by such features, including reduced pollution to other surface waters, but the ecological objectives of such projects are often unclear, hindering efforts to limit ecological risks through improved management and design.