Contamination of water resources by pathogenic bacteria.

Water-borne pathogen contamination in water resources and related diseases are a major water quality concern throughout the world. Increasing interest in controlling water-borne pathogens in water resources evidenced by a large number of recent publications clearly attests to the need for studies that synthesize knowledge from multiple fields covering comparative aspects of pathogen contamination, and unify them in a single place in order to present and address the problem as a whole. Providing a broader perceptive of pathogen contamination in freshwater (rivers, lakes, reservoirs, groundwater) and saline water (estuaries and coastal waters) resources, this review paper attempts to develop the first comprehensive single source of existing information on pathogen contamination in multiple types of water resources. In addition, a comprehensive discussion describes the challenges associated with using indicator organisms. Potential impacts of water resources development on pathogen contamination as well as challenges that lie ahead for addressing pathogen contamination are also discussed.

Introduction

Water-borne pathogen contamination in ambient water bodies and related diseases are a major water quality concern throughout the world. Pathogen contamination is a serious issue for almost all types of ambient water bodies, making its recognition and understanding essential (U.S. EPA [2012a]). The United Nations identified improving water quality as one of the eight Millennium Development Goals (MDGs). Its target is to reduce the number of people without access to safe water by 50% by 2015 (WHO [2011]). Because of the overwhelming scientific evidence for climate change (IPCC [2007]), it is also important to understand how perturbations in weather patterns can potentially impact pathogen levels in water resources. To meet future demands of water for food, energy, and ecosystems, increasing water storage structures (i.e., dams) must be a component of long-range planning (World Bank [2010]). However, such new structures can potentially degrade water quality and exacerbate public health risk.

While several review papers are currently available (Bradford et al. [2013]; Pachepsky and Shelton [2011]; Pang [2009]; Jin and Flury [2002]; John and Rose [2005]; Jamieson et al. [2004]; Jamieson et al. [2002]; Arnone and Walling [2007]; Kay et al. [2007]), there is a manifest need for additional transdisciplinary studies that assimilate knowledge gained from multi-research endeavors studying pathogen contamination, and provide a comprehensive synopsis in order to comprehend the entirety of the problem. Therefore, the goal of this review is to present a broad research scope assessment of pathogen contamination of water resources and the associated challenges it presents. We synthesize the potential health risks imposed by pathogens in water resources by providing existing knowledge that covers surface water, groundwater, fresh water, and saline water. Further, the impact of water resources development on pathogen contamination, future challenges, and recommendations are summarized. In addition, we provide a brief discussion describing water-borne pathogen footprints and potential challenges associated with the use of indicator organisms for assessing water quality.

Health risk

Water-borne diseases (i.e., diarrhea, gastrointestinal illness) caused by various bacteria, viruses, and protozoa have been the causes of many outbreaks (Craun et al. [2006]). In developing countries, such as those in Africa, water-borne diseases infect millions (Fenwick [2006]). According to World Health Organization (WHO), each year 3.4 million people, mostly children, die from water-related diseases (WHO [2014]). According to United Nations Children’s Fund (UNICEF) assessment, 4000 children die each day as a result of contaminated water (UNICEF [2014]). WHO ([2010]) reports that over 2.6 billion people lack access to clean water, which is responsible for about 2.2 million deaths annually, of which 1.4 million are in children. Improving water quality can reduce the global disease burden by approximately 4% (WHO [2010]).

Although water-associated diseases in developing countries are prevalent, they are also a serious challenge in developed countries. A study by Arnone and Walling ([2007]), who compiled data of outbreaks in the U.S. (1986 – 2000), reported 5,905 cases and 95 outbreaks associated with recreational water. Gastrointestinal Illness (GI) caused by variety of different microbes and germs, which causes symptoms, such as diarrhea, nausea, vomiting, fever, abdominal pain, was responsible for about 29.53% cases. More than 27% of cases were caused by Shigella spp. In addition, 10.99%, 10.08%, and 6.59% of the cases were caused by Cryptosporidium parvum, Adenovirus 3, and Leptospira, respectively. Nearly 23% and 21% of the outbreaks were caused by GI and Shigella spp, respectively. In addition, 16.84%, 12.63%, and 7.37% of the outbreaks were caused by Naegleria fowleri, E. coli 0157:H7, and Schistosoma spp., respectively. Besides acute gastroenteritis, major etiological agents such as Giarida, Cryptosporadium, E. coli 0157:H7, V. cholera, and Salmonella were the agents responsible for many outbreaks (Craun et al. [2006]). During the same period 437,082 cases and 48 outbreaks were caused by contaminated drinking water, of which about 95.89% of the cases were caused by Cryptosporidium parvum. Nearly 42% and 31% of the outbreaks were caused by Giardia lamblia and GI, respectively. Reporting statistics on water-borne outbreaks in the U.S., Craun et al. ([2006]) found that at least 1870 outbreaks (23 per year) occurred between 1920 and 2002. These reported outbreaks and their reported incidence of illnesses are likely to be an underestimation of actual numbers because of nonreported cases and missing exposure information. To protect public health, the U.S. EPA’s National Primary Drinking Water Regulations (NPDWRs) contain standards describing the Maximum Contaminant Level (MCL) – the highest level of a contaminate allowable in drinking water. The U.S. EPA has defined the MCL of various microorganisms, such as Cryptosporidium, Giardia lamblia, Legionella, and Total Coliforms (including fecal coliform and E. coli), and viruses (U.S. EPA [2012b]). The Maximum Contaminant Level Goal (MCLG) – the level of a contaminant in drinking water below which there is no known risk to public health, has also been proposed by the U.S. EPA. The MCLC levels for Cryptosporidium, Giardia lamblia, Legionella, and Total Coliforms are zero. The EPA requires 99% removal of Cryptosporidium in drinking water, and the removal percentages of Giardia lamblia and viruses are 99.9 and 99.99%, respectively. Although there is no limit for Legionella, EPA believes that if Giardia lamblia and viruses are removed/inactivated, then drinking water likely to be free of Legionella. The U.S. EPA requires routine sampling of drinking water for testing total coliform and E. coli, and if a routine sample is positive, then repeat samples are required. If, in any repeat sample, total coliform or E. coli is detected then the drinking water has an acute MCL violation. For a drinking water system that collects fewer than 40 routine samples per month, no more than one sample can be total coliform-positive per month. For a system that collects more than 40 routine samples, no more than 5% of samples total coliform-positive in a month is allowed (U.S. EPA [2012b]).

Each year approximately 42,000 cases of salmonellosis are reported in the U.S. (CDC [2014]). Schistosomiasis is not reported in the U.S. because it is not endemic; however, 200 million people are infected worldwide. In 2011, about 1,060 cases of Guinea worm disease, caused by the parasite Dracunculus medinensis, were reported in many remote parts of Africa that do not have safe drinking water. Malaria, a protozoal disease of the Genus Plasmodium transmitted by mosquitos breeding in contaminated water, affects 300–500 million people, and causes over one million deaths each year (more than 90% of deaths in Africa). Overall the morbidity and mortality caused by contaminated water are enormous and need to be controlled by improving the security of safe water (i.e., recreational as well as drinking water) in both developing and developed countries.

Historical perspective of water-borne diseases

Water contamination has a long presence in human history, with descriptions in the Sushruta Samshita about water-borne diseases resembling cholera in an Indian text written in Sanskrit as early as 500–400 B.C. (Colwell [1996]). Although cholera infections have not been reported in recent years in developed countries mainly due to improved sanitation, millions of people each year continue to get infected by Vibrio cholera in developing countries (Nelson et al. [2009]). The World Health Organization reports about 3–5 million cholera cases and 10,000 – 120,000 deaths, mainly in developing countries, due to cholera every year. Over time, cholera has caused millions of deaths in developing as well as developed countries (Colwell [1996]; Okun [1996]). For instance, a major outbreak of cholera was reported in London in 1849. Dr. John Snow, a physician to Queen Victoria, showed a relationship between people infected by cholera and contaminated water (Snow [1854]; Colwell [1996]). Jordan et al. ([1904]), Ruediger ([1911]), Simons et al. ([1922]), and Rudolfs et al. ([1950]) provide excellent reviews on incidents during the early 19th century. Colwell ([1996]) reported that in the mid and late 18th century, cholera infected millions of people all over the world. The worst outbreak in recent memory occurred in Haiti following the devastating earthquake affecting the capital and surrounding regions, with almost a half a million cases, killing thousands of people (CDC [2011]).

Water-borne pathogen footprints and challenges

Indicator organisms are commonly used to assess the levels of pathogens in water resources; i.e., water-borne pathogen footprints of water resources. Monitoring the levels of indicator organisms (such as fecal coliforms, E. coli) (Figure 1) is a common approach for quantifying the potential pathogen loads in ambient water bodies. For decades, public health officials/scientists have evaluated water quality by enumerating fecal coliforms and E. coli levels in rivers, lakes, estuaries, and coastal waters (Malakoff [2002]; Pandey et al. [2012a]; Pandey et al. [2012b]; Pandey and Soupir [2013]). There is, however, much debate regarding current indicator organisms and their ability to represent the potential presence of pathogenic bacteria. In addition, identifying the source of pathogens (e.g., human waste, animal waste, wildlife excreta, and waterfowl droppings) (Figure 2) is challenging (Malakoff [2002]; Dickerson et al. [2007]). There is potential to use a relatively new approach such as microbial source tracking (MST) to trace the origin of fecal coliform (Scott et al. [2002]; Grave et al. [2007]; Dickerson et al. [2007]; Ibekwe et al. [2011]; Ma et al. [2014]). In the past, the MST method was exploited by antibiotic resistance analysis to assess the impact of cattle on water quality on a watershed scale (Grave et al. [2007]). The authors suggested that host-origin libraries, based on a phenotypic method, are useful for tracking the pathogen sources. Many MST methods, however, rely on the assumption that some strains of bacteria are found only within a single kind or group of animals. This assumption can be debatable when it comes to the common fecal bacteria E. coli (Malakoff [2002]). Therefore, caution is needed while using E. coli for source tracking (Gordon [2001]). Further, the cost to develop libraries, implement extensive sampling programs needed for verifying the MST method, and calculate uncertainties associated with the method are legitimate issues, which requires attention before exploiting the MST method at watershed scale.

Figure 1

Transmission electron micrograph of (0157:H7; ATCC: 35150).

Figure 2

Challenges in identifying enteric pathogen sources (source: Malakoff,[2002]).

Currently, public health officials/scientists rely on exposure limits for assessing pathogen levels in water resources, which have been established to protect human health. The EPA defines acceptable recreational limits as those that will result in eight or fewer swimming-related gastrointestinal (GI) illnesses out of every 1,000 swimmers (U.S. EPA [1986]). The current U.S. EPA fresh water quality criteria for E. coli is a geometric mean not exceeding 126 CFU/100 ml, or no samples exceeding a single sample maximum of 235 CFU/100 ml (U.S. EPA [2001]). Criteria were developed based on the U.S. EPA measurements of total and Highly Credible Gastrointestinal Illnesses (HCGI), which correlated with E. coli densities (r = 0.804) in fresh recreational waters (Dufour [1984]). Multiple studies have identified trends between indicator organisms in water and GI illness in humans, including vomiting, diarrhea, and fever (Cabelli [1983]; Wade et al. [2006]). Recent work by Edge et al. ([2010]) detected water-borne E. coli in 80% of water samples with E. coli levels of less than 100 CFU/100 ml. Another study by Wade et al. ([2006]) reported significant positive trends between increased GI illness and indicator organisms at the Lake Michigan beach, and a positive trend with indicators such as E. coli at a Lake Erie beach. Recently, the use of indicator organisms (e.g., fecal coliforms, E. coli) for assessing pathogen levels has been debated more often than ever; however, the use of indicator organisms is likely to continue for assessing pathogen levels in water resources potentially because of the lack of an alternative reliable solution.

Pathogen contamination in water resources

The U.S. EPA, which monitors water quality of various ambient water bodies, estimated that pathogens impair more than 480,000 km of rivers and shorelines and 2 million ha of lakes in the U.S. (U.S. Environmental Protection Agency [2010a]). According to EPA estimates, pathogens are the leading cause of impairment for 303 (d) listed waters (i.e., list of impaired and threatened waters that the Clean Water Act requires all states to submit for EPA approval) (Figure 3) (U.S. EPA [2014a], [2014b], [2014c]). A total of 71,917 causes of impairment have been reported, and the top five causes of impairment are shown in Figure 3. Pathogen contamination clearly dominates the causes of impairment (U.S. EPA [2014a], [2014b], [2014c]).

Figure 3

Causes of impairment in the U.S. (data source: U.S. EPA ([2014a],[2014b],[2014c])).

Studies by Diffey ([1991]), Brookes et al. ([2004]), Jamieson et al. ([2004]), Gerba and Smith ([2005]), Gerba and McLeod ([1976]), Hipsey et al. ([2008]), Pachepsky and Shelton ([2011]) reviewed the current studies of water-borne pathogen transport, with particular reference to freshwater and estuarine sediments. In addition, many current reviews focus on specific aspects of water resources, for instance, John and Rose ([2005]) focused on groundwater, Brookes et al. ([2004]) focused on reservoirs and lakes, Jamieson et al. ([2004]) focused on agricultural watersheds, and Kay et al. ([2007]) reviewed catchment microbial dynamics. The review study presented here uses a relatively broader approach for understanding how water-borne pathogens can potentially impact public health and various ambient water bodies. In addition, existing challenges, while assessing pathogen levels in water resources are discussed.

Coastal and estuarine environments

In the U.S., elevated pathogen levels are a leading cause of impairments of coastal environments (U.S. EPA [2014a], [2014b], [2014c]). Urban runoff and sewers have been identified as the primary source of coastal water impairments. Rippey ([1994]) reported about 400 outbreaks and 14,000 cases caused by pathogen contaminated coastal water since the late 1800s in the U.S. Impairments of coastal environments have major economic impacts on the U.S. For example, losses caused by pathogen contamination in Massachusetts are more than $75 million each year (Weiskel et al. [1996]). The studies, which elaborate various pathogens in coastal environment and their survival mechanism, are summarized in Table 1.

Table 1

Studies describing pathogen contamination in saline water (coastal and estuary environments)

Author(s) year	Organisms	Results	Study remarks
Coastal water
Gerba and McLeod ([1976])	E. coli, fecal coliform	Longer survival in the sediment presence in seawater	Sediment influence on pathogen survival
Goyal et al. ([1977])	Total coliform, fecal coliform	Sediments of shallow canal can act as a reservoir	Pathogens distribution in water and sediment
Kapuscinski and Mitchell ([1983])	E. coli, bacteriophages	E. coli survives longer than bacteriophages	Survival of pathogens in sunlight
Rao et al. ([1984])	Enteroviruses (Polio and Rota)	Abundance viruses were attached with sediment	Viruses distribution in water and sediment
Weiskel et al. ([1996])	Fecal coliform	Waterfowl major source of fecal coliform inputs	Pathogen source and transport pathways
Sinton et al. ([1999])	Bacteria and Fecal bacteriophages	Somatic coliphages shown prolonged survival	Sunlight influence on sewage-polluted seawater
Solo-Gabriele et al. ([2000])	E. coli	Riverbanks as the primary pathogen source	Sources of pathogens on subtropical environment
Nasser et al. ([2003])	Cryptosporidium, viruses, E. coli	E. coli die-off was faster than other pathogens	Comparative survival of various pathogens
Schriewer et al. ([2010])	Bacteroidales, fecal indicator, protozoa, bacteria	Bacteroidales have shown higher predictive skill than fecal indicators	Bacteroidales as a predictor of pathogens in coastal water
Estuaries
Ketchum et al. ([1952])	Coliform, zooplankton	Sactericidal and predation caused coliform die-off	Processes responsible for pathogens decrease
Smith et al. ([1978])	Echovirus 1, coxsackieviruses	Sediment prolonged viruses survival	Persistence of pathogen viruses
Rhodes and Kator ([1990])	E. coli	Mortality rises in sunlight	Indigenous microbiota and sunlight influence
White et al. ([1998])	Perkinsus marinus	Use of Kriging analysis for disease prevalence	Use of GIS in pathogen distribution analysis
Lipp et al. ([2001])	Vibrio vulnificus	Salinity controls the distribution of pathogens	Distribution of human pathogens
Desmarais et al. ([2002])	E. coli, C. perfringens	Pathogens re-grown with tides and sterile sediment	Influence of soil on fecal indicator in tidally influenced environment
Frias-Lopez et al. ([2002])	Clostridium, Campylobacter, Arobacter	Pathogen partitioning was found in surface and overlying water	Bacterial communities partitioning between sea water, dead coral surface
Chandran and Hatha ([2005])	E. coli, Salmonella	E. coli shows better survival capacity than Salmonella	Relative survival in microcosm studies
Baker-Austin et al. ([2009])	Vibrio vulnificus	Prevalence of antibiotic resistance in a human pathogen	Multi-site analysis shows widespread antibiotic resistance in pathogens

The sources of coastal water contamination are: point discharges of treated and untreated sewage from shoreline outfalls, and non-point discharges. The non-point sources, such as runoff from naturally vegetated areas, discharge pathogens into coastal waters. Besides runoff from vegetated areas, the storm water runoff from urban, commercial, and industrial lands also discharges pathogens into coastal waters. In addition, other sources, such as malfunctioning or poorly sited septic systems, can also introduce significant amounts of pathogens (Sayler et al. [1975]; Howe et al. [2002]). Weiskel et al. ([1996]) reported that direct deposition of waterfowl feces was a considerable source of pathogens. Fayer and Trout ([2005]) summarized the transport of various pathogens, such as Giardia, Toxoplasma, and Cryptosporidium (zoonotic parasites) in the coastal environment. Moreover, the presence of sediment in seawater can also increase the survival chance of fecal coliforms, such as E. coli (Gerba nd McLeod [1976]; Goyal et al. [1977]). Solo-Gabriele et al. ([2000]) showed that the location and timing of storms of the coastal area in tropical and subtropical environments are also important factors that can potentially influence coastal water quality.

Previous studies have shown that the direct discharge of storm water runoff into coastal waters through storm drain systems can cause pathogen contamination, even where separate storm and sanitary sewer systems are in place. For instance, Weiskel et al. ([1996]) found that about 16% of the total fecal coliform inputs were caused by storm water entering Buttermilk Bay in Massachusetts. In addition, coastal rivers draining largely undeveloped watersheds with extensive riparian wetlands can be a natural source of fecal pathogens to coastal waters (Viau et al. [2011]; Staley et al. [2014]; Roberts et al. [2013]; Liang et al. [2013]; Wilkes et al. [2014]). On-site septic systems can also contribute significant amounts of fecal pathogens to coastal waters in low-lying fine-grained geological settings where saturated soils enhance pathogen growth. Weiskel et al. ([1996]) reported that shoreline wrack deposits could act as a reservoir of fecal bacteria, and the removal of wrack deposits from inter-tidal zones can improve the water quality of adjacent coastal waters.

Similar to the coastal environment, increasing water-borne pathogen levels in estuaries are a serious threat to public health. Human activities can impact estuary pathogen levels when they are adjacent to populated areas, and often provide a means of transportation and substantial recreation (Schriewer et al. [2010]; Pachepsky and Shelton [2011]). The most common pathogens, previously identified in estuaries by Rhodes and Kator ([1990]), were Vibrio cholerae, Giardia, Cryptosporidium, Salmonella, and Campylobacter spp. As shown in Table 1, the presence of various pathogens (e.g., E. coli, C. perfringens, Clostridium, Salmonella) has been reported in many previous studies. Municipal point sources are the primary cause of pathogen contamination in estuaries. Urban water disposed through combined sewer outflows is the cause of approximately 12% of estuary impairments in the U.S. (Arnone and Walling [2007]). Pathogens, including Vibrio vulnificus which carries the highest fatality rate of any food-borne pathogen in the U.S., were detected in the Gulf of Mexico Estuary (Lipp et al. [2001]; Baker-Austin et al. [2009]). Several studies discovered that bed sediment plays a vital role (i.e., with the release of particle-attached pathogens from bed sediment to a water column through a resuspension process) for the persistence and transport of pathogens in the estuaries (Smith et al. [1978]; Desmarais et al. [2002]). Previous studies have shown that pathogen growth and decay are influenced by environmental conditions. For instance, a study by Chandran and Hatha ([2005]) revealed that sunlight is a major factor that influences survival of pathogens like E. coli and S. typhimurium in the estuarine water.

Groundwater

Groundwater is heavily used all over the world as the primary source of domestic drinking water supplies, and contaminated groundwater certainly enhances risk to public health. Nationally, 40% of the U.S. domestic water supply originates from groundwater, and over 40 million people use groundwater as their drinking water via private wells (Alley et al. [1999]). Groundwater pathogen contamination has led to numerous disease outbreaks in the U.S.; for example, at least 46 outbreaks of disease occurred between 1992 and 1999, resulting in 2,739 cases of illness and several deaths (John and Rose [2005]). These are reported cases; due to underdiagnosis and underreporting, the actual morbidity is almost certainly higher.

Several studies have shown that microbial pathogens, such as Salmonella, E. coli, S. faecalis, and enteroviruses are relatively stable in groundwater (Bittion et al. 1983; Schijven and Hassanizadeh [2000]; Pang et al. [2004]) (Table 2). Controlling groundwater pathogen contamination has recently been emphasized in many countries, as pathogens can survive up to 400 days depending on the soil temperature (Nevecherya et al. [2005]; Filip and Demnerova [2009]). For example, identifying sources of groundwater pathogen contamination has received significant attention in France (Grisey et al. [2010]). Many studies reported that health risks caused by chlorine-resistant protozoans, such as Cryptosporidium spp. (Ferguson et al. [2003]; Kay et al. [2007]; Kay et al. [2008]), are considerable. One of the major concerns is that wetlands without lining might cause pathogen contamination of groundwater (Kay et al. [2007]). Similar concerns have been expressed in the United Kingdom by water regulators. The European Union (EU) has also emphasized protecting groundwater from pathogen contamination. Pathogen-contaminated groundwater can cause pollution in coastal environments. For example, a study of Buttermilk Bay has shown that groundwater is capable of transporting a large quantity of pathogens from surface to sub-surface water either by direct discharge or by discharge to rivers flowing into the bay (Moog [1987]; Weiskel et al. [1996]). The risk of contaminating groundwater particularly increases in areas where shallow aquifers exist. In these situations it is more likely that contaminated surface water or water from septic tanks can reach groundwater (Weiskel et al. [1996]). Precipitation events are likely to increase groundwater pathogen contamination because of contaminated ground water recharge.

Table 2

Pathogen contamination in freshwater environment (ground water, rivers and lakes and reservoirs)

Author(s) year	Organisms	Results	Study remarks
Groundwater
Bitton et al. ([1983])	Salmonella, E. coli, S. faecalis, enteroviruses	S. faecalis decay rate was similar to viruses	Survival of pathogens in groundwater
Schijven and Hassanizadeh ([2000])	Viruses	Viruses attachment with soil was influenced by pH, and favorable sites	Soil passage impacts on virus removal at field-scale
Pang et al. ([2004])	E. coli and F-RNA phages	Pathogens were sorbed in aquifer material	Transport of E. coli and F-RNA phages
Nevecherya et al. ([2005])	Salmonella, viruses, E. coli, shigellos	Mathematical model was derived for temperature depends inactivation rate	Pathogenic bacteria and viruses survival in groundwater
Filip and Demnerova ([2009])	Bacillus megaterium and Staphylococcus	Pathogens survived 10 to 100 days	Pathogens survival in groundwater; FT-IR characterization
Grisey et al. ([2010])	Total coliforms, E. coli, Enterococci, Pseudomonas aeruginosa, Salmonella and Staphylococcus aureus	Bacterial density monitoring coupled with artificial trace experiments proved useful in locating pathogens sources	Pathogens survival in groundwater and landfill leachate
Rivers
Chin ([2010])	Fecal coliform	Summer and rainfall raises pathogens	Urban areas impacts on stream pathogens
Smith et al. ([1973])	Salmonella, fecal coliforms, streptococci	Salmonella decline was close to fecal coliform	Usefulness of indicators organisms
Burton et al. ([1987])	Pseudomonas aeruginosa, Salmonella newport, E. coli	Clay in sediments improves E. coli survival	Survival of pathogens in fresh water sediments
Smith et al. ([1987])	E. coli, Fecal bacteria	U.S. rivers shows decline in fecal indicator	Water quality assessment from 1974 to 1981
Terzieva and McFeters ([1991])	E. coli, Campylobacter jejuni, Yersinia enterocolitica	Temperate zone surface water serve as a rersisitent verhicle in transimsision of bacteria between animals and humans	Survival and injury study of bacteria in agricultural surface water
Terzieva and McFeters ([1991])	Yersinia enterocolitica and E. coli	Experimental design and environmental play major role in pathogens survival	Survival in stream water, comparison between field and lab studies
Fraser et al. ([1998])	Fecal coliform	Model predicted pathogens	Modelling non-point source pollution
Lakes and reservoirs
Beaver and Crisman ([1989])	Ciliates	Grazing habits of ciliates are discussed	Predators roles in fresh water
Rubentschik et al. ([1936])	Serratia marcescens	Adsorption of bacteria could be questionable	Adsorption of bacteria in salt lakes
Mac Kenzie et al. ([1994])	Cryptosporidium	C. oocysts study passes through the filtration system of water supply	Contaminated water from Milwaukee lake caused outbreak
Wcislo and Chrost ([2000])	E. coli	Predators controlled pathogen levels	Survival of pathogens in man-made reservoir
Kistemann et al. ([2002])	E. coli, coliform, fecal streptococcal, and Clostridium perfringens	Most of the pathogens increases during extreme runoff events	Microbial load in drinking water reservoir during rainfall events
Howe et al. ([2002])	Cryptosporidium oocysts	Animal feces was a major source of pathogens	Water supply’s oocysts caused outbreak in northern England
Ishii et al. ([2006])	E. coli	E. coli survived longer in soil	Presence and growth of E. coli in Lake superior watershed

Pathogen contamination (e.g., bacteria, protozoa, and viruses) poses a serious risk in water resources. The transport of pathogens from surface water to groundwater increases the vulnerability of groundwater (Jin and Flury [2002]). These authors reported that 70% of the water-borne microbial illness outbreaks in the United States have been associated with ground water. Pathogens such as viruses are much smaller than bacteria and protozoa, and many can potentially reach groundwater through porous soil matrices. Jin and Flury ([2002]) reviewed the fate and transport of viruses in porous media to understand mechanisms and modeling of virus sorption, and concluded that factors such as solution chemistry, virus properties, soil properties, temperature, and association with solid particle influences virus survival, transport, and sorption in porous media. Pang ([2009]) studied microbial removal rates in subsurface media, and reported that soil types considerably influence microbial removal rates. For instance, volcanic soils, pumice sand, fine sand, and highly weathered aquifer rocks showed high removal rates. The author found that microbial removal rates were inversely correlated with infiltration rates and transport velocity.

Considerable work recently has been done towards understanding pathogen transport in the vadose zone (Wang et al. [2014a]; [2014b]; Unc and Goss [2004]; Darnault et al. [2004]). Groundwater can be contaminated by seepage and percolation of contaminated water from the vadose zone (Darnault et al. [2004]). The macropores of agricultural land are also known to play a considerable role in polluting groundwater, particularly from fields where manure is applied (Jamieson et al. [2002]). Unc and Goss ([2004]) evaluated the influence of manure on the transport of bacteria from land receiving manure to water resources. These authors reported that manure application in the land influences pathogen transport in the vadose zone. The presence of straw and coarse organic matter influences the persistence of bacteria, and manure application changes the physical configuration of soil, the soil chemistry, and the properties of the microbial cells, which control the survival and persistence of bacteria in soils. Another recent study by Wang et al. ([2014]) assessed the transport of E. coli in soils with preferential flow. The authors reported that a decrease in macropore length resulted in a decreased apparent saturated hydraulic conductivity of the macropore and an increase in the mass transfer. Wang et al. ([2014a]; [2014b]) concluded that macropore length has a considerable influence on preferential transport of E. coli.

Reservoirs and lakes

Previous studies have shown the presence of many pathogens in lakes and reservoirs (Table 2), and that these pathogens can pose risks to human health. In many countries surface reservoirs serve as the main source of drinking water, and these surface water bodies are often vulnerable to pathogen contamination (Kistemann et al. [2002]). In the developed world, although there is increased awareness of water treatment for pathogen contamination and water quality, outbreaks of water-borne disease via public water supplies continue to be reported (Gibson et al. [1998]; Howe et al. [2002]; Brookes et al. [2004]).

In the past, more than 403,000 residents of the greater Milwaukee, Wisconsin area experienced gastrointestinal illnesses due to infection with the parasite Cryptosporidium parvum following contamination of the city’s water supply, which was associated with inadequate filtration of contaminated water from Lake Michigan (Mac Kenzie et al. [1994]; Cicirello et al. [1997]). In the 1990s, Cryptosporidiosis became the most common cause of outbreaks associated with public drinking water supplies in the United Kingdom (Howe et al. [2002]). In developing countries, diseases such as diarrhea and cholera are the leading cause of morbidity (Nelson et al. [2009]). Overall, diarrhea associated with drinking contaminated water is responsible for 2 to 2.5 million deaths annually (Fenwick [2006]). In lakes and reservoirs, increased pathogens are often associated with storm events, and the stream inflow is considered to be the major source of pathogens during storm events. Elevated flows in rivers most likely agitate bed sediment, which causes enhancement of pathogen levels of the water column (Jamieson et al. [2005a]; Jamieson et al. [2005b]; Pandey and Soupir [2013]; Bai and Lung [2005]). During the rainy season, the influx of contaminated water from rivers to lakes and reservoirs can substantially increase pathogen levels (Kistemann et al. [2002]). The quantity of pathogen influxes from tributaries of lakes and reservoirs during the rainy season is of particular importance in determining pathogen transport and distribution (Brookes et al. [2004]).

Rivers

Pathogen contamination is a major cause of stream impairments. The sources of impairment and health risks induced by water-borne pathogens are extensively reported (Table 2). In the U.S. pathogen contamination is the leading cause of stream water pollution. The EPA’s National Water Quality Inventory Report suggests that about 53% of the assessed rivers are impaired, and a majority of them are contaminated by pathogens (U.S. EPA [2012a]). The cost to implement the total maximum daily load (TMDL) plans to improve stream water is estimated as $0.9 to $4.3 billion per year (U.S. EPA [2010b]).

Pathogen influxes into rivers from agricultural lands (Figure 4) are the main cause of stream impairments (Chin [2010]; U.S. EPA [2012a]). A weak understanding of pathogen transport from agricultural lands to rivers is considered to be a major challenge in implementing and deriving suitable land management practices capable of improving stream water quality. For instance, despite common knowledge that agricultural land’s non-point source pollution is a leading cause of stream impairment, it is difficult to identify points of origin of pathogens and the pathways by which they enter streams. As an example, pathogens are likely to enter rivers from many potential sources, including lateral inputs from pastures and riparian zones, influx of pathogen-contaminated groundwater, direct deposit of fecal matter from livestock and wildlife, discharge of contaminated sanitary sewer flows, and wastewater treatment plant effluents. In rainy events, pathogens in rivers are influenced by fresh input from watersheds as well as sub-surface flow. In addition, the resuspension of legacy pathogens from bed sediments can considerably increase pathogen levels (Cho et al. [2010]; Droppo et al. [2009]; Jamieson et al. [2005b]; Kiefer et al. [2012]; Nagels et al. [2002]; Muirhead et al. [2004]; Kim et al. [2010]; Smith et al. [2008]).

Figure 4

Simplified path of animal waste pathogen transport from agricultural land to rivers.

Controlling pathogen contamination from livestock/wildlife to streams is challenging (Terzieva and McFeters [1991]). For example, it is doubtful that pathogen contamination can be prevented by fencing off riparian buffers, and even if buffers are useful in controlling stream water pathogens, it is not certain what their width must be (Nagels et al. [2002]). There are review studies that elaborate on stream water pathogen contamination (Fraser et al. [1998]; Jamieson et al. [2004]; Pachepsky et al. [2006]).

Many studies have emphasized the use of mathematical models to understand pathogen transport from agricultural land to rivers (Kim et al. [2010]; Muirhead et al. [2004]; Jamieson et al. [2005a]; Jamieson et al. [2005b]). Previous studies (Gerba and Smith [2005]; Pandey et al. [2012b]; Pandey and Soupir [2013]; Pachepsky and Shelton [2011]; Martinez et al. [2014]) have emphasized the need to improve existing models for calculating the fate and transport of pathogens at the watershed scale. Currently, empirical as well as mechanistic models are being used to calculate microbial fate and transport (Muirhead and Monaghan [2012]). Numerous studies are available for exploiting watershed scale models such as WATFLOOD (Dorner et al. [2006]), the Soil and Water Assessment Tool (SWAT) (Neitsch et al. [2005]), the Spatially Explicit Deliver Model (SEDMOD) (Fraser [1999]), and KINEROS/STWIR (Guber et al. [2011]) for predicting pathogen transport. Though numerous watershed scale models are available, which can be exploited for calculating pathogen transport at watershed scale, considerable difficulties exist while using the models (Pandey et al. [2012b]; Pandey and Soupir [2013]). For instance, correctly identifying the model input parameter values is a daunting task. Recently, considerable emphasis has been given to understand the sensitivity of the input parameters to the model output (Martinez et al. [2014]; Parajuli et al. [2009]).

When implementing watershed scale models for predicting pathogen transport, difficulties related to the selection of model input parameters are common (Martinez et al. [2014]). A study by Parajuli et al. ([2009]) evaluated the sensitivity of fecal coliform bacteria loads modeled with SWAT, and the authors reported that many of the parameters (e.g., bacterial die-off rates, the temperature adjustment factor) were insensitive to the model output. Similarly, Coffey et al. ([2010]) reported that initial concentration of E. coli and the bacterial partition coefficient parameters of the SWAT model were overly sensitive in affecting the model’s output. Another recent study by Martinez et al. ([2014]) evaluated the KINEROS/STWIF model input parameter sensitivities to the output. These authors reported that environmental controls such as soil saturation, rainfall duration, and rainfall intensity had the most sensitivity, while parameters such as soil and manure properties were the least sensitive in affecting model output. Many previous studies (Pandey and Soupir [2012a], [2012b]; Kim et al. [2010]; Martinez et al. [2014]; Parajuli et al. [2009]) reported that identification and selection of model input parameters are major challenges when implementing the model for predicting fate and transport of pathogens at the watershed scale.

Blooming of cyanobacteria

While water-borne pathogens are a serious concern, excessive algal bloom (Figure 5) in water resources can potentially limit their uses for recreation activities as well as for drinking water. Cyanobacteria (blue-green algae) have unique roles in oxygenation of the atmosphere (Hofer [2013]); however, their excessive growth or dense algal bloom in water resources diminishes the quality and quantity of light in the water column ([U.S. EPA 2014b]). When blooms are excessive, the risk of toxin contamination (released by harmful algal bloom (HABs)) is likely to be elevated. The HAB includes many types of algal taxa such as dinoflagellates, diatoms, and cyanobacteria. Eutrophication caused by excessive algal bloom can cause fish kills and reduce the diversity of aquatic life ([U.S. EPA 2014b]). In hypoxic water, dissolved oxygen levels can be less than 2 – 3 ppm (U.S. EPA [2014c]). For example, in the hypoxic zone in the northern Gulf of Mexico, an area along the Louisiana-Texas coast, less than 2 ppm of dissolved oxygen concentration has been reported. This is believed to be caused by excess nutrients delivered from the Mississippi River, in combination with seasonal stratification of Gulf waters (U.S.G.S., [2014]). The largest U.S. hypoxic zone, which occurred in 2002 in the Gulf of Mexico, was about 13,518 square kilometers (U.S.G.S. [2012]).

Figure 5

Algal bloom in Squaw Creek, Iowa, U.S.

Freshwater cyanobacterial blooms produce highly potent cyanotoxins and cyanobacterial HABs, which can affect the liver, nervous system, and skin (U.S. EPA [2014b]). Cyanobacterial blooms can be potentially detrimental to human and animal health, aquatic habitats, and aquaculture industries (Kaloudis et al. [2013]; Carmichael [2001]; Falconer [2005]; Codd et al. [1999]). Previous studies (Mackintosh et al. [1990]; Yoshizawa et al. [1990]) have shown that microcystins (i.e., cynotoxins) are hepatotoxic and act as tumor promoters through the inhibition of protein phosphatases, which play a key role in cellular regulation. Eutrophic water conditions combined with warm surface water temperatures (15 – 30°C) can potentially enhance cyanobacterial blooms in water (Oikonomou et al. [2012]; Vareli et al. [2009]). Currently more than 400 hypoxic zones exist in the world affecting 245,000 square kilometers (Diaz and Rosenberg [2008]). These dead zones in coastal waters have spread exponentially since the 1960s and have considerably impacted ecosystems. The increase in these dead zones is likely to be influenced by climate change. Various factors such as how climate change affects water-column stratification and how nutrient runoff affects organic matter production will determine further expansion of the dead zones (Diaz and Rosenberg [2008]). For instance, climate predictions of the Mississippi River basin indicate a 20% increase in river discharge, which will enhance nutrient loading and result in expansion of the oxygen depleted area (IPCC [2007]). Considering the importance of fresh water, estuarine, and marine environment to public health and aquatic life, identifying approaches capable of controlling excessive algal blooms is required before these dead zones spread globally.

Impact of water resources development

Water resources development involves altering the natural flow path of rivers and lakes, as well as designing irrigation schemes and dams. These activities have been alleged to be responsible for causing new diseases and enhancing health risks (Fenwick [2006]; Steinmann et al. [2006]).

The influence of water resources development in spreading diseases, such as schistosomiasis, a parasitic disease which is ranked second only to malaria with regard to the number of people infected, has been reported extensively; one estimate says that about 103 million out of 779 million infected people live in close proximity to large reservoirs and irrigation schemes (Steinmann et al. [2006]).

Designing dams and irrigation schemes in tropical and subtropical climate zones has often resulted in disease outbreaks caused by water-borne pathogens. Consider, for example, the Sennar Dam on the Blue Nile River and Sudan’s Gezira Scheme, the world’s largest irrigation project. Because of the dam’s commercial success, irrigation in the region has doubled from the 1940s and 1950s. After the 1950s, infections from malaria and schistosomiasis increased significantly, becoming the subject of the first integrated disease-control program, the Blue Nile Health Project, implemented from 1978 to 1990. The project failed to have any impact in controlling the prevalence of schistosomiasis (Eltoum et al. [1993]; Fenwick [2006]; Steinmann et al. [2006]). Another example is China’s Three Gorges Dam, built across the Yangtze River and completed in 2009, which created a 50,700 km2 reservoir and submerged more than 220 counties. Hotez et al. ([1997]) reported that the reservoir would produce environmental changes that could lead to the transmission of schistosomiasis in the area served by the dam. A recent study by Schrader et al. ([2013]) found major high risk areas for schistosomiasis occurrence in the large lakes and flood plain regions of the Yangtze River. Another study by Gray et al. ([2012]) reported that the Three Gorges Dam will likely to impact the transmission of schistosomiasis in China.

In the U.S., because of increasing concern for produce safety, pathogen-free irrigation water is attracting considerable attention (Martinez et al. [2014]). Painter et al. ([2013]) reported that produce accounted for nearly half of food-borne illnesses in the U.S. between 1998 and 2008. Growing concern about the safety of food and water will likely help in developing improved strategies while planning and designing large dams for irrigation purposes.

Challenges and recommendations

Infectious diseases caused by pathogens are the third leading cause of death in the United States, and the leading cause in the world (Binder et al. [1999]). The past two decades have seen the emergence of many new pathogenic infectious diseases (Daszak et al. [2000]). Many of these are caused by anthropogenic changes, such as water resources development, climate warming, and interactions between humans and animals, both domestic and wild (Krause [1994]; Epstein [2001]; Woolhouse [2002]; Fenwick [2006]; Schriewer et al. [2010]).

Multidisciplinary knowledge about how ambient water bodies, wildlife, domestic animals, and human populations interact with and impact each other are crucial in dealing with future challenges. Generally, domestic animals, wildlife, and humans are considered to be major sources of water-borne pathogens; however, finding the specific culprit—the primary pathogen source—is challenging (Malakoff [2002]). A watershed, for example, can have many pathogen sources, such as agricultural land, riparian areas, agricultural feeding operations, livestock, wildlife, and humans.

Developing models that are reliable in predicting pathogen survival and transport at the watershed scale can be helpful in implementing/evaluating the strategies for mitigating ambient water body pathogen levels. Evaluating the impacts of various environmental factors on pathogen survival in water resources is crucial. Various publications (i.e., Dorner et al. [2006]; Kim et al. [2010]; Rehmann and Soupir [2009]; Droppo et al., [2011]; Cho et al. [2010]; Pandey et al. [2012b]; Pandey and Soupir [2013]) are available describing models capable of predicting pathogen contamination levels in ambient water bodies; however, further work is required to improve model predictions. The evaluation of predictions by existing models clearly demonstrates the need for improvement. Many pathogen transport models use only temperature-induced mortality and growth, and do not include interactions among other environmental factors (e.g., pH, nutrients, DO, solar radiation); future inclusion of these environmental factors will likely to improve the model predictions.

Typically most studies have relied on E. coli and other indicator bacteria to indicate pathogen levels in water. Although widely used in monitoring contamination levels, E. coli alone can lead to mercurial and misleading information (Gordon [2001]). Schriewer et al. ([2010]) suggested that with improved pathogen detection technology (i.e., PCR-based detection) an indicator organism, such as E. coli, can be sufficiently accurate in most cases. Overall, improving technology to identify causative agents more accurately, creating standard epidemiological data for diseased populations, and enhancing the knowledge of disease dynamics can improve the understanding of risks caused by interactions among various populations (Harvell et al. [1999]; Daszak et al. [2000]; Harvell et al. [2002]).

In the past, a number of studies on pathogen contamination have been conducted on a scale where the conditions of ambient water bodies were simulated in laboratories. These studies are helpful in understanding pathogen behavior only up to a point. For enhancing the understanding of pathogen interactions in the environment, more emphasis should be given to field-scale studies.

Conclusions

This review examines studies from various disciplines to understand pathogen contamination in ambient water bodies. The worldwide prevalence of pathogen contamination is a serious concern, and enhancing the understanding of major pathogen sources and their significant impacts on water resources is crucial. A considerable number of studies on pathogen contamination have been conducted on a laboratory-scale; more emphasis should be given to field-scale studies for enhancing the understanding of pathogen interactions in the environment. Developing new models, and improving existing modeling approaches commonly used for predicting water-borne pathogen levels will likely to help in assessing pathogen contamination at watershed-scale. Considering the limited ability of existing models to predict pathogen contamination, improvement and development of new models are needed so that pathogen levels can be predicted more accurately. Integrating knowledge from multiple fields (e.g., hydrology, microbiology, and ecology) would increase the understanding of pollution levels and potential causes of pollution, and can also help devise long-term strategies to improve water quality.

Competing interest

The authors declare that they have no competing interests.