Definition of the Subject and Its Importance

Pathogens can enter recreational waterbodies (lakes, rivers, ocean beaches) from a number of different sources. The illness acquired by swimmers after exposure to pathogen-polluted recreational waters is termed recreational waterborne illness (RWI). Since many RWI go unreported to health care agencies, the true number of RWI each year can only be estimated; however, numbers are believed to be high – over one million in southern California alone. The risk of RWI from exposure to recreational waters can be measured using epidemiological studies and estimated using quantitative microbial risk assessments.

Introduction

Exposure to microbially polluted recreational waters can cause a variety of adverse health effects in humans including neurological infections, skin infections, earaches, eye infections, gastrointestinal illnesses, and respiratory infections [1]. Microbial pollution refers to the presence of organisms that cause illness in humans either through the production of toxins or their colonization of the human body.

It is estimated that globally, exposure to coastal waters polluted with wastewater results in an excess 120 million gastrointestinal and 50 million severe respiratory illnesses per year [2], including illnesses acquired through consumption of contaminated shellfish. In southern California, there are an estimated 1.5 million cases of gastrointestinal illnesses each year due to recreational exposure to polluted waters [3]. Moreover, there were 259 recreational water outbreaks that occurred in the USA between 1970 and 2000 [4].

Estimating the number of individuals acquiring illness through exposure to recreational waters is challenging. Typically, individuals who acquire recreational waterborne illness (RWI) do not seek medical attention because most of the illnesses tend to be mild and self-limiting. In addition, most RWIs are not reportable, so incidence levels are highly uncertain. The estimates mentioned previously were obtained using a variety of assumptions about the contamination of water and exposures. Epidemiology studies and quantitative microbial risk assessments (QMRA) are two scientifically rigorous methods that are used to estimate rates of RWI as a function of water quality.

This chapter provides a brief overview of the pathogens present in coastal waters. Readers are referred to other references for more information on these topics [5–7]. The two most common methods of assessing risk of exposure to pathogens in coastal waters, epidemiology studies, and QMRA, are also described. Example applications of these methods to assess risk of illness from exposure to pathogen-polluted coastal waters are presented. Critical research gaps are identified and summarized.

Pathogens in Coastal Waters

Pathogens present in coastal waters can be characterized into two broad groups. The first group consists of autochthonous pathogens that have an ecological niche in recreational waters. The other group is composed of allochthonous pathogens that come from human and animal wastes that have been discharged into these waterbodies.

Autochthonous pathogens include harmful algae, organisms in the genus Vibrio, and some protozoa such as Naeglaria fowleri. Harmful diatoms and dinoflagellates can cause a variety of ailments in humans from their production of toxins [8, 9]. Vibrio vulnificus and V. parahaemolyticus can infect open wounds and cause gastroenteritis if ingested from seawater. Other Vibrio species (such as V. cholerae and V. mimicus) can be pathogenic and cause similar ailments. N. fowleri is primarily found in freshwater, and when it enters the human body, can cause a rare but serious brain infection. While these organisms are extremely important from a human health perspective, this chapter will not discuss risks associated with autochthonous pathogens.

The focus of this chapter is allochthonous pathogens

Table 16.1 Allothchonous human pathogens detected in coastal waters
Full size table

(Table 16.1) including the eight pathogens that cause a large proportion (>95%) of all non-foodborne illnesses in the USA [10]: the viruses norovirus, rotavirus, and adenovirus; the bacteria Campylobacter, Salmonella and pathogenic Escherichia coli; and the protozoans Cryptosporidium spp. and Giardia lamblia. Other pathogens that are important etiologies of RWI include enteroviruses, Staphylococcus aureus and methicillin-resistant S. aureus (MRSA). All these pathogens, with the exception of S. aureus and MRSA, cause gastrointestinal illness. Staphylococcus causes skin infections. Adenoviruses can cause gastrointestinal illness, as well as eye and respiratory infections.

The detection of pathogens in environmental matrices is methodologically challenging. Allochthonous pathogens are rare microbes in the environment. In seawater there are on the order of one million autochthonous bacteria and 10 million autochthonous viruses in a milliliter. Whereas allochthonous pathogens may be present at levels of 1 per 10 liters l or lower. Thus, enumerating the allochthonous pathogens is particularly difficult because the presence of all the other organisms in the sample can interfere with the detection of the rare target. Because the field of pathogen detection is still evolving and because allochthonous pathogens densities are typically low, there are limited data on pathogen concentrations and occurrence in recreational waters. In many cases, authors only provide data on the presence or absence of pathogens in recreational waters and do not provide concentrations. Some example concentrations are provided in Table 16.1.

Treated and untreated wastewater, human and other animal feces, stormwater and urban runoff, and agricultural runoff can all contain microbes that are pathogenic to humans [11, 12]. When discharged to coastal waters, concentrations of pathogens may be high and can pose high levels of human health risks.

Pathogens are also found in environmental reservoirs that in some cases may serve as a source of pathogens to recreational waters. Beach sands can harbor Campylobacter, Salmonella, and S. aureus [13–16]. Aquatic sediments can accumulate bacteria, protozoa, and viruses [17]. Marine and lacustrine kelp species may also harbor bacterial pathogens including pathogenic E. coli, Salmonella, and Campylobacter [18, 19].

Assessing Risk: Epidemiology and Indicator Organisms

The monitoring of recreational waters for all RWI pathogens to assess the safety of swimming is not scientifically or economically feasible [20]. Microbial indicator organisms have been used for centuries as indicators of the presence of human pathogens. Internationally, many countries use fecal and total coliforms as a basis for their recreational water quality criteria, standards, or guidelines [21]. Other countries rely on measurements of enterococci (or fecal streptococci), E. coli, or both for their recreational waters, most based on guidelines provided by World Health Organization [22] and/or the US Environmental Protection Agency (EPA). These organisms were chosen as indicators because their concentrations are high in human wastewater and feces, they are relatively simple to measure, and their presence in coastal waters is correlated to adverse health outcomes in swimmers through epidemiology studies conducted in wastewater-impacted waters [23–26]. The epidemiology studies that correlate indicator concentration to adverse health outcomes are key to the use of indicators to assess risk.

Epidemiology studies evaluate illness resulting from exposure to a particular contaminant or activity. When applied to RWI, epidemiology studies evaluate the illness rates in swimmers versus non-swimmers, and characterize illness rates as a function of indicator organism concentration. The studies involve the collection of health and behavior data contemporaneously with concentrations of indicator organisms. RWI epidemiology studies are either case-control randomized trial or prospective cohort designs. In case-control studies, swimmer and non-swimmer activities are prescribed by randomization at the onset of the study. In these studies, exposures are well controlled. Subject recruitment is done in advance of the study. In prospective cohort studies, subjects are recruited at the study site and are enrolled when they arrive at the shoreline. Exposures are self-prescribed by subjects and behavioral data on exposure is collected using self-reports at the end of the day. In the prospective cohort design, there is less control over exposure, but the exposures are more realistic as they are not prescribed by the study design. Additional types of studies that have been employed to study RWI include cross-sectional studies and event studies. The former is similar in design to a cohort study; the latter takes advantage of a sporting event, for example, for data collection.

Since the 1950s, numerous epidemiological studies have been conducted throughout the world to evaluate the association between recreational water quality and RWI (including GI symptoms; eye infections; skin complaints; ear, nose, and throat infections; and respiratory illness) [23–26]. Most of these studies investigated wastewater effluent-impacted marine and estuarine waters alone or in combination with freshwater. Several investigated freshwater recreational environments or non-wastewater effluent-impacted waters. These studies indicate that the rates of some adverse health outcomes are higher in swimmers compared with non-swimmers [23].

Taken as a whole, the weight of evidence from these studies indicates that fecal indicator bacteria (fecal streptococcus/Enterococcus, in particular) are able to predict GI and in some cases, respiratory illnesses from exposure to recreational waters [23, 25, 26]. This broad base of information stems from studies conducted throughout much of the developed world

Table 16.2 Recreational water epidemiology studies included in reviews by Prüss [23], Wade et al. [25], and Zmirou et al. [26]. Location refers to geographic location of study. Water Type refers to whether the study was conducted at a marine or fresh water. Study design denotes whether to study was a cohort, randomized trial, cross section, or event study
Full size table

(Table 16.2, Adapted from [27]).

Several meta-analyses and/or systematic reviews have summarized the available recreational water epidemiology studies [23, 25, 26]. Pruss et al. [23] conducted a systematic review to initiate development of new WHO guidelines for recreational use of the water environment. The comprehensive review of 22 published studies on sewage pollution of recreational water and health outcomes concluded that the epidemiological basis had been laid to develop WHO guidelines on fecal pollution based on a causal association between GI illness symptoms and increased concentrations of bacterial indicators (i.e., enterococci for marine, enterococci and E. coli for fresh) in recreational waters.

Zmirou et al. [26] examined 18 published studies to provide a scientific basis for establishing new standards for the microbial quality of marine and fresh recreational waters to replace the 30 year-old European Union bathing water quality guidelines [28]. The researchers provided four major results: (1) increased concentrations of fecal coliforms or E. coli and enterococci in both fresh and marine recreational waters are associated with increased risks of acute GI illness, with enterococci eliciting four to eight times greater excess risks than fecal coliforms or E. coli at the same concentrations; (2) GI illness risks associated with enterococci occur at lower indicator concentrations in marine versus fresh recreational waters; (3) increased concentrations of total coliforms have little or no association with GI illness risk; and (4) no evidence exists of a concentration threshold of indicator microorganisms below which there would be no GI illness risk to bathers.

Wade et al. [25] conducted a systematic review and meta-analysis of 27 published studies to evaluate the evidence linking specific microbial indicators of recreational water quality to specific health outcomes under non-outbreak (endemic) conditions. The study was conducted at the request of the United States National Academy of Sciences. Secondary goals included identifying and describing critical study design issues and evaluating the potential for health effects at or below the current regulatory criteria [29]. They concluded that (1) enterococci and, to a lesser extent, E. coliare adequate indicators (predictors) of GI illness in marine recreational waters, but fecal coliforms are not; (2) the risk of GI illness is considerably lower in studies with enterococci and E. coli densities below those established by EPA [29], thus providing support for their regulatory use; (3) E. coli is a more reliable and consistent predictor of GI illness than enterococci or other indicators in fresh recreational waters; and (4) studies that used a non-swimming control group and that focused on children found elevated GI illness risks.

Based on these meta-analyses, the weight of evidence indicates that there is a relationship between levels of specific indicator bacteria and RWI in coastal waters impacted by wastewater. However, as discussed earlier there are many sources of indicator bacteria to coastal waters, and many of these sources contain different pathogens with diverse health risks. Along coastlines with good sewage infrastructure and regulated anthropogenic discharges, wastewater is unlikely to be contributing substantial amounts of indicator organisms to the swimming areas on a regular basis [27]. Important sources of indicator organisms and pathogens are probably nonpoint in nature, emanating from soils, animal feces, urban runoff, stormwater runoff, or agricultural runoff.

There are only a few epidemiology studies that examine the link between RWI and fecal indicator organisms in recreational waters polluted with sources other than wastewater. Review of these studies suggests the relationship between indicator concentration and RWI risks are equivocal. On one hand, Colford et al. [30] found that the incidence of swimmer illness was greater than the incidence of non-swimmer illness, but swimmer illness was not associated with any of the bacterial indicator organisms at a marine beach where bacterial contamination was not attributable to wastewater discharges. Similarly, Calderon et al. [31] found no statistically significant association between swimmers’ illness risk and indicator concentrations in a freshwater pond where agricultural runoff was the source of contamination. McBride [32] suggested that if more swimmers had been included in the Calderon et al. [31] study, achieving statistically significant results would have been possible, however. At a marine bathing study in New Zealand, McBride et al. [33] indicated that RWI risks posed by animal versus human fecal material were not substantially different; however, the study’s limited range of indicator organisms’ concentrations precluded the development of a detailed statistical model of health risks versus indicator density.

In the first study to be conducted in waters directly impacted by urban runoff, Haile et al. [34] reported associations between swimmer health and indicator densities. However, this nonpoint runoff source was known to contain human sources of fecal contamination, based on the presence of human enteric viruses. Dwight and colleagues [35] found that surfers exposed to Southern California urban runoff had higher illness rates than surfers exposed to Northern California rural runoff. The results from a Hong Kong marine water study [36] and a German freshwater study [37] are more difficult to interpret regarding risks from human versus nonhuman sources because in both studies, the analyses combined the results from sites with different predominant contamination sources.

An additional meta-analysis examined the differential risks associated with exposure to human and nonhuman animal fecal material [24]. Illness risk associated with bathing in water polluted primarily with human fecal material was reviewed based on studies from the USA [38, 39], Canada [40, 41], Israel [42, 43], Egypt [44, 45], Spain [46], France [47, 48], the UK [49, 50], Hong Kong [51], and Australia [52, 53]. Most of these studies showed a positive correlation between GI illness and fecal indicator density; there was little equivalent evidence from waters polluted primarily with animal feces. The only study specifically designed to address swimming-associated illness in animal-impacted waters was that of Calderon et al. [31] who found no statistically significant association between GI illness and fecal indicator bacteria densities. Based on this observation, Sinton and colleagues [24] concluded that reliable epidemiological evidence was lacking, and that other sources of information were needed to identify and apportion human and nonhuman fecal inputs to natural waters.

Given these studies, there are not sufficient epidemiological data to conclude that concentration of indicator organisms in coastal waters not impacted by wastewater are predictive of health risks in all cases. More epidemiology data may help address the lack of information. However, fecal contamination in coastal waters not impacted by wastewater is likely to be highly variable and emanate from a complex mix of sources. Other approaches may be more useful for addressing the relationship between indicators and health risk along these shorelines. Despite the lack of epidemiological evidence for a relationship between indicator organism concentration and health risk, it is well established in the outbreak literature that water contaminated with animal feces can cause illness in humans [4].

Assessing Risk: Quantitative Microbial Risk Assessment (QMRA) Modeling

QMRA is a health risk modeling approach that translates microbial exposures into infection or illness risk estimates. For RWI, the dose received by individuals is derived from estimates of the volume of water ingested during an exposure event and concentration of pathogen(s) in that volume of water. Once a dose (number of pathogens per exposure) is determined then a risk of infection or illness is derived from applying published dose-response models for specific pathogens [54], which are derived from human feeding trials or outbreak data [55, 56].

Estimations of key model parameters (such as ingested volume of water and pathogen concentration) are generally described as probability density functions (PDFs) (i.e., distributions to account for the stochastic nature of the parameter) to help account for inherent variability as well as various methods and model uncertainties. The characterized risk is best described as a distribution as well, to capture variability and uncertainty as best as possible. This characterized risk can then be compared to “tolerable risk” or a site benchmark for the recreational water (e.g., 8 or 19 gastrointestinal illnesses per 1,000 bathers as used in the 1986 EPA criteria for freshwater and marine recreation, respectively).

An initial screening-level risk assessment for a site may start by only using point estimates to describe model parameters (e.g., WHO, 2004). Then, to reduce uncertainties in the risk estimate, more complexity is built into a QMRA model using PDFs to better represent a specific site, in what is called a “static” model. An alternative approach known as “dynamic QMRA” takes into account secondary infections to the broader community, as well as addressing the susceptibility versus nonsusceptibility of individuals to infection and illness [57–59].

QMRA can be useful for a number of purposes. First, it can be used to look at hypothetical risks under different scenarios of pathogen sources and/or recreational activities and exposure routes [60, 61]. This can provide a human health interpretation of environmental pathogen concentrations, or it can provide guidance for decision-making with respect to alternative management options. Second, QMRA can be used to augment the understanding of recreational water epidemiology studies [62]. A number of studies have evaluated pathogen risks in recreational waters using QMRA [60, 61, 63–70]. A handful of these are reviewed below.

Diallo et al. [69] examined the risk associated with recreational exposure to canals in Thailand containing Giardia, Cryptosporidum, and diarrhegenic E. coli. They found the predicted risk of illness from the protozoa was two orders of magnitude higher than the actual protozoan infection rate in the region of Thailand, while the rate of gastrointestinal illness predicted from diarrhegenic E. coli matched actual observed rates of the disease in the region. The authors suggest that the illness rates predicted for the protozoa are much higher due to immunity in the community, which was not considered in the QMRA modeling. Another possibility is massive under-reporting of protozoa illness rates, which is a documented issue for diarrheal diseases.

Ashbolt and Bruno [65] used QMRA in conjunction with the published relationship between enterococci concentrations and probability of gastrointestinal illness and acute respiratory illness observed during a UK beach epidemiology study [71]. The authors showed that by assuming that the etiology of illnesses was viral, a fixed ratio between enterococci and viruses of 1–175, a volume of water ingested of 50 mL, and 50% illness rate of those infected with virus, they were able to model the observed illness rates (at exposures greater than 60 CFU/100 mL enterococci) assuming viruses had an exponential dose-response curve. Further, they were able to model the observed acute respiratory illness using the dose-response curve of adenovirus.

Gerba et al. [70] estimated risk of exposure to rotavirus in bathing waters using the few previously reported data available at the time on rotavirus concentrations in bathing waters. They found the risk of illness to be 1/10–1/100 with a one time bathing event. These authors chose to use rotavirus as a model pathogen because it is one of the most infectious viruses for which a dose-response model is available. A major limitation to their study is the lack of data on concentrations of rotavirus in bathing waters. Surface water concentrations were estimated to be between 0.24/L and 29/L. Ingested volumes used were 100 mL for recreational exposure.

Schoen and Ashbolt [67] explored the relative risk associated with exposure to seagull feces, poorly treated sewage, and a mixture of both sources in seawater with enterococci at levels of 35 CFU/100 mL (a USEPA standard). Authors assumed that seagull feces contained Campylobacter and Salmonella, while sewage contained norovirus, G. intestinalis, Cryptosporidium spp., and Salmonella. A distribution of ratios of enterococci to pathogen concentrations were considered to reflect the uncertainty in this parameter. The authors showed that when enterococci came exclusively from seagull feces, the risk of illness is less than the benchmark of 0.01 on which US standards are based.

Soller et al. [61] examined the risk of viral gastroenteritis associated with recreational and non-recreational use of a river downstream of a wastewater treatment plant discharge. Two wastewater treatment scenarios were compared with the goal of evaluating the public health benefit of increased treatment of the effluent. The authors employed a health protective approach by assuming that the etiology of illness would be a virus with clinical features identical to those of rotavirus. They also assumed that removal of the virus in treatment facilities would mirror that of coliphage and that coliphage and the virus occurred in a ratio that varied from 0.001 to 1. Exposures were informed by a hydrologic model of the area and observations of swimmer behavior. Unlike the other models discussed above, this model incorporated secondary transmission by allowing illness to be passed from person to person.

Steyn et al. [68] compared the risks in recreational surfaces waters in South Africa expected from measured E. coli concentrations (applying epidemiology study-derived dose-response curves) and measured Salmonella concentrations (applying the QMRA method and the Salmonella dose-response curve assuming ingestion of 100 mL of water for exposure). The researchers found that risks derived from the Salmonella QMRA model were higher than those derived from the E. coli concentrations. Their results suggest using E. coli to assess risk from exposure to Salmonella may be inadequate.

Wong et al. [63] estimated risk from exposure to adenoviruses during recreation contact with water at Great Lakes beaches impacted by point sources of treated wastewater. The authors measured adenoviruses using an MPN culture-dependent assay at their study site, assumed an adenovirus dose-response model, and an ingestion rate of 100 mL of water. The authors found that 0.24–2.4 illnesses per 1,000 swimmers were likely to have occurred from adenoviruses, a range of frequencies below the EPA guideline of 8 illnesses per 1,000 swimmers. Although an epidemiology study was done at the same time as their study, the epidemiology data were not available for comparison of actual illness rates.

Soller et al. [62] used QMRA to understand more fully the reported epidemiologic results from studies conducted on the Great Lakes in the US during 2003 and 2004 by identifying pathogens that could have caused the observed illnesses in those studies. The reference pathogens used for this analysis were Norovirus, rotavirus, adenovirus, Cryptosporidium spp., G. lamblia, Campylobacter jejuni, Salmonella enterica, and E. coli O157:H7. Two QMRA-based approaches were used to estimate the pathogen combinations that would be consistent with observed illness rates: in the first, swimming-associated gastrointestinal (GI) illnesses were assumed to occur in the same proportion as known illnesses in the US due to all non-foodborne sources, and in the second, pathogens were assumed to occur in the recreational waters in the same proportion as they occur in disinfected secondary effluent. The results indicated that human enteric viruses and in particular, norovirus could have caused the vast majority of the observed swimming-associated GI illnesses during the 2003/2004 water epidemiology studies [72, 73]. Evaluation of the time to onset of illness strongly supports the principal finding, and sensitivity analyses support the overall trends of the analyses even given their substantial uncertainties. These results are notable because little is known about the specific microbial agents that are responsible for the observed illnesses in swimmers. While several studies have attempted to collect biological specimens (blood or stool) as part of epidemiologic research at beach sites, these efforts have to date been largely unsuccessful in identifying the agents responsible for the observed increase in GI symptoms among swimmers [50].

Soller et al. [74] conducted a QMRA investigation to determine whether estimated risks following exposure to recreational waters impacted by gull, poultry, pig, or cattle fecal contamination are substantially different from those associated with waters impacted by human sources such as treated wastewater. Previously published QMRA methods were employed and extended to meet these objectives [67]. Health outcomes used in the analyses were infection via reference pathogens by water ingestion during recreation and subsequent GI illness. Illness risks from the reference pathogens were calculated for exposure to contaminated recreational water with fecal indicator bacterial densities at the U.S. regulatory limits: 35 CFU/100 mL enterococci and 126 CFU/100 mL E. coli. The probabilities of GI illness from reference pathogens were calculated using dose-response relationships from the literature and Monte Carlo simulations. The primary findings from the analysis were that: (1) GI illness risks associated with human exposure to recreational waters directly impacted by fresh cattle feces are not substantially different from those impacted by human sources; and (2) the risks associated with human exposure to recreational waters impacted by fresh gull, poultry, or pig feces are substantially lower than those impacted by human sources.

Several observations may be drawn from the QMRA studies summarized above. First, the assembled studies focused on a small subset of the pathogens potentially important in waterborne exposure during recreation. The pathogen analyzed most frequently is rotavirus, primarily due to its high infectivity and the availability of dose-response data. It is likely that future QMRAs will also focus on norovirus [62, 67, 74] now that a published dose-response relationship is available [75].

Second, modeling variability in pathogen source density appears to be hampered by scarcity of both data and analysis techniques. Two common methods for accounting for source variability among studies are (1) use of empirical distributions for pathogen density based on relatively short time series, and (2) assumption of log-normal distribution of pathogen densities. There are advantages and drawbacks to both of these approaches. Drawbacks to the use of empirical distributions are inconsistency in sampling strategies used to develop databases, frequent non-detects, constraint of pathogen densities to those observed in a limited number of samples, and lack of availability of pathogen concentration data. Use of distributions to describe pathogen density in sources overcomes some of the constraints associated with use of empirical distributions, but choosing a distributional family can be problematic. Among the studies reviewed, many studies employed point estimates for pathogen density. Among studies using distributions to describe pathogen variability, the following distributions were employed: normal, triangular, log-normal, negative binomial, uniform, and Poisson.

Third, most of the studies reviewed do not account for variability and uncertainty in dose-response model parameters. As with variability in exposure, this observation likely indicates that high quality and diverse dose-response model data are not available. The use of a small number of dose-response models may indicate that some QMRA modelers chose the pathogens to model based on the availability of dose-response models. Lack of dose-response models for many pathogens of concern and for differing routes of exposure (i.e., cutaneous exposure to S. aureus) is a major data gap. The need for dose-response models corresponding to different exposure routes (i.e., ingestion, inhalation, etc.) arises from the ability of some waterborne pathogens such as adenovirus to initiate infection via multiple routes.

Finally, secondary transmission and immunity are often neglected in risk estimation. Studies that have included secondary transmission and waterborne illnesses [57, 58, 60, 76–78] have demonstrated that consideration of secondary transmission and immunity can influence overall risk associated with exposure to pathogens significantly and in unintuitive ways.

Future Directions

Health data collected from recreational swimmers confirm measurable health effects from exposure to contaminated coastal waters. To adequately protect the health of swimmers and others who recreate in coastal waters or consume fish and shellfish from coastal waters, it is essential to understand the microbial hazards present and the risks they represent to human health. While there are data available on microbial hazards and risks, many aspects remain poorly understood and characterized.

Research is needed to further characterize microbial hazards in recreational waters. Pathogen detection techniques that allow detection of infectious pathogens rapidly in environmental waters are needed. There are a number of pathogen and indicator detection technologies that are in development or have been recently developed [79, 80]. However, many of these have not been applied to natural waters, or are just in the “proof of concept” stage. Work is needed to transit new detection technologies that work at the bench-scale to the field scale – to detect pathogens in environmental waters. Once this is accomplished, they should be applied to a wide array of waterbodies to fully understand pathogen and indicator occurrence, concentrations, fate, and transport.

A better understanding of the risks of exposure to pathogens from different sources is needed. Some QMRA studies have addressed this issue, but further research is needed to ground truth the QMRA results with epidemiology data, or data on infection rates and etiologies, as determined through analysis of bodily fluid from individuals with RWI. Gastroenteritis is the most well-studied RWI; more work is needed to understand the importance of other RWI including skin infections and respiratory ailments. More dose-response data for a wider array of pathogens are needed to provide more refined estimates of risk using QMRA. The importance of secondary infections and immunity for RWI should also be further characterized.

Finally, better surveillance systems are needed for RWI so that prediction systems can be developed. As global climate changes in the coming decades, the scientific community needs to be able to anticipate how this might change the burden of RWI. Microbial pollution of coastal waters is expected to change as temperatures, runoff frequency and volumes, and rainfall pattern change. A thorough understanding of the occurrence of RWI, and the distribution, fate, and transport of waterborne pathogens will enable us to better anticipate the effects of climate change.