Effects of Wastewater Disinfection on Waterborne Bacteria and Viruses

By Blatchley, Ernest R III; Gong, Woei-Long; Alleman, James E; Rose, Joan B; Et al

ABSTRACT:

Wastewater disinfection is practiced with the goal of reducing risks of human exposure to pathogenic microorganisms. In most circumstances, the efficacy of a wastewater disinfection process is regulated and monitored based on measurements of the responses of indicator bacteria. However, inactivation of indicator bacteria does not guarantee an acceptable degree of inactivation among other waterborne microorganisms (e.g., microbial pathogens).

Undisinfected effluent samples from several municipal wastewater treatment facilities were collected for analysis. Facilities were selected to provide a broad spectrum of effluent quality, particularly as related to nitrogenous compounds. Samples were subjected to bench-scale chlorination and dechlorination and UV irradiation under conditions that allowed compliance with relevant discharge regulations and such that disinfectant exposures could be accurately quantified. Disinfected samples were subjected to a battery of assays to assess the immediate and long-term effects of wastewater disinfection on waterborne bacteria and viruses.

In general, (viable) bacterial populations showed an immediate decline as a result of disinfectant exposure; however, incubation of disinfected samples under conditions that were designed to mimic the conditions in a receiving stream resulted in substantial recovery of the total bacterial community. The bacterial groups that are commonly used as indicators do not provide an accurate representation of the response of the bacterial community to disinfectant exposure and subsequent recovery in the environment. UV irradiation and chlorination/dechlorination both accomplished measurable inactivation of indigenous phage; however, the extent of inactivation was fairly modest under the conditions of disinfection used in this study. UV irradiation was consistently more effective as a virucide than chlorination/dechlorination under the conditions of application, based on measurements of virus (phage) diversity and concentration.

Taken together, and when considered in conjunction with previously published research, the results of these experiments illustrate several important limitations of common disinfection processes as applied in the treatment of municipal wastewaters. In general, it is not clear that conventional disinfection processes, as commonly implemented, are effective for control of the risks of disease transmission, particularly those associated with viral pathogens. Microbial quality in receiving streams may not be substantially improved by the application of these disinfection processes; under some circumstances, an argument can be made that disinfection may actually yield a decrease in effluent and receiving water quality. Decisions regarding the need for effluent disinfection must account for site-specific characteristics, but it is not clear that disinfection of municipal wastewater effluents is necessary or beneficial for all facilities. When direct human contact or ingestion of municipal wastewater effluents is likely, disinfection may be necessary. Under these circumstances, UV irradiation appears to be superior to chlorination in terms of microbial quality and chemistry and toxicology. This advantage is particularly evident in effluents that contain appreciable quantities of ammonia-nitrogen or organic nitrogen. Water Environ. Res., 79, 81 (2007).

KEYWORDS: disinfection, bacteria, virus, chlorine, UV, wastewater.

doi:10.2175/106143006X102024

Introduction

Wastewater disinfection has been practiced in the United States for approximately 100 years with the goal of providing protection to human populations from exposure to pathogenic waterborne microorganisms. Undisinfected wastewater effluents represent a potentially important source of pathogenic microorganisms in the environment and a possible vector for transmission of disease among human populations. Although bacterial pathogens are present in wastewater effluents, it is generally believed that enteric viruses represent the greatest risk to human health of all waterborne pathogens. The (oo)cysts of waterborne protozoan pathogens can also represent a substantial risk to humans. The risks of disease associated with waterborne pathogens are generally acute in nature.

Human contact with municipal wastewater effluents can occur through ingestion, swimming, direct or indirect contact with water from water reuse applications, or ingestion of seafood. Beyond these circumstances of direct contact, it is important to consider also that municipal wastewater effluents become a part of the hydrologie cycle. As such, wastewater effluents represent potentially important sources of biological and chemical constituents in water supplies; in a very real sense, “we all live downstream”.

Historically, most wastewater disinfection operations have used chlorine as the disinfectant. Chlorine is known to be effective for inactivation of common bacterial indicator organisms; however, several important drawbacks to chlorine-based disinfection have been identified, including its relative ineffectiveness against some microbial pathogens (e.g., bacterial spores, enteric viruses, and protozoan [oojcysts) and repair or regrowth of pathogens postdisinfection. Therefore, conditions that accomplish acceptable inactivation of indicator bacteria by chlorine-based disinfection strategies do not necessarily guarantee safety of a treated water supply.

UV irradiation is widely recognized as an alternative to chlorination/dechlorination for disinfection of municipal wastewaters. UV radiation is a broad-spectrum antimicrobial agent. However, some viral pathogens are known to be resistant to UV radiation, and the issue of repair and recovery of the exposed microbial community is also a potential drawback of UV-based disinfection processes. Therefore, as in the case of chlorination, conditions of UV irradiation that prove to be sufficient for inactivation of indicator bacteria do not guarantee an acceptable level of treatment for all microbial pathogens.

It is clear that the proper application of disinfectants can lead to removal or inactivation of microbial pathogens; however, given the drawbacks listed above, there is some question as to whether disinfection of municipal wastewater effluents should be applied in all cases. In many developed countries outside North America, wastewater disinfection is practiced only in situations where a direct, clear threat to human health is evident, such as discharges to bathing areas or shellfish breeding grounds. The occurrence of waterbome diseases in these areas is not substantially different from that of North America, where wastewater disinfection is required in the vast majority of cases. However, even within the United States, many states have chosen to require disinfection only on a seasonal basis. These attributes have prompted a reevaluation of the assignment of disinfection as a default application.

A broad range of possible options exist for disinfection of municipal wastewaters, from no disinfection at all to aggressive operations that accomplish extensive inactivation of recalcitrant waterbome microorganisms. The following sections present summaries of investigations that were performed in the mid 1970s that have become classics in the disinfection field. These summaries provide useful illustrations of the range of treatment objectives and wastewater disinfection operations in place in the United States today.

Sanitation Districts of Los Angeles County, California (Pomona Virus Study [Parkhurst, 1977]). The Los Angeles County Sanitation District, in conjunction with the U.S. Environmental Protection Agency (U.S. EPA) (Washington, D.C.) and the California State Water Resources Control Board (Sacramento, California), conducted a study with the objective of providing data regarding alternative tertiary treatment approaches for water reuse applications that could allow compliance with Title 22 of the California Administrative Code. Title 22 represents the California State Health Department’s Wastewater Reclamation Criteria. The study involved operation of four pilot-scale tertiary treatment processes. Disinfectants included in these four systems were inorganic combined chlorine, free chlorine, and ozone. At the time of this investigation, UV irradiation was not viewed as a viable alternative to chlorine.

While this landmark study was relevant for a number of reasons, perhaps the most tangible outcome of this study with regard to the issue of the need for municipal wastewater disinfection was the definition of the conditions of chlorine-based disinfection that are required to achieve acceptable treatment in a reuse setting. Specifically, the conditions of chlorination required to accomplish reliable compliance with Title 22 were as follows:

* Combined chlorine residual of at least 10 mg/L (as chlorine [Cl^sub 2^]) for a contact time of at least 2 hours, or

* Free chlorine residual of at least 4 mg/L (as Cl^sub 2^) for a contact time of at least 2 hours.

These disinfection conditions accomplished roughly 4 log^sub 10^ inactivation of seeded virus and were consistently in compliance with the coliform regulation. When used in conjunction with other appropr\iate physicochemical processes, treatment consistently accomplished overall virus inactivation (or removal) of roughly 5 log^sub 10^ units.

Metropolitan Water Reclamation District of Greater Chicago, Illinois. Leadership within the Metropolitan Water Reclamation District of Greater Chicago, Illinois (MWRDGC), has often challenged conventional thinking on topics relating to municipal wastewater treatment; in several cases, the approaches taken by the MWRDGC to solve water treatment and water quality problems have resulted in important innovations that have subsequently been adopted by other municipalities. An example is the MWRDGC’s approach to disinfection, which is described in great detail in a series MWRDGC publications (Lue-Hing et al., 1976; Sedita, Lue-hing, and Haas, 1987; Sedita, Zenz, Lue-Hing, and O’Brien, 1987).

Beginning in July 1972, the MWRDGC implemented continuous chlorination of the effluents from all of its facilities. Soon thereafter, the district initiated testimony before the Illinois Pollution Control Board questioning the wisdom of chlorine-based disinfection. To support this effort, the MWRDGC began an extensive investigation of the advantages and disadvantages of chlorination. In general terms, the MWRDGC demonstrated that water quality in the receiving waters downstream of their facilities, which at times contained as much as 90% effluent (therefore, very little dilution), was the same or better when chlorine-based disinfection was terminated than when chlorination was practiced. Chlorination was observed to result in an improvement in bacterial quality in their effluent only within a short reach of receiving stream (roughly 16 to 22 km [10 to 15 river miles] downstream of the outfall). Viable enteric viruses in receiving streams were not significantly affected by chlorination, as practiced at the district’s facilities. It is important to note that this conclusion was reached for a system that was based on a nitrified effluent. Therefore, the chlorine residual would probably have been present in the form of free chlorine, which is generally regarded as the most effective form of chlorine for inactivation of planktonic microorganisms. Chlorination also resulted in a substantial reduction in fish populations in receiving streams. In addition, the absence of fish allowed for increased populations of nuisance insects in and around Chicago waterways.

Based on their findings, MWRDGC presented a compelling argument that chlorination did not yield the benefits that were typically ascribed to disinfection practices; the risks of microbial exposure were not substantially affected by chlorination. Moreover, MWRDGC argued that the practice of chlorination was actually detrimental to water quality, based on measurements of aquatic life and other water quality parameters. It is noteworthy that these observations were made in receiving streams that, at times, were essentially undiluted municipal wastewater effluent from wastewater treatment facilities that used fairly conventional, though effective, treatment operations.

Based on the evidence presented to them, the Illinois EPA granted MWRDGC’s request to discontinue disinfection at its largest facilities (Stickney, Northside, Calumet, and Lemont water reclamation plants [WRPs]). Subsequent monitoring of effluents and receiving streams demonstrated improvements in water quality. Disinfection by chlorination/dechlorination continues at the three other district facilities (Egan, Kirie, and Hanover Park WRPs) on a seasonal basis.

By comparison with most conventional disinfection practices in use today (for purposes of this manuscript, the term conventional disinfection will refer to disinfection operations that do not represent opportunities for water reuse, and therefore are not subject to reuse criteria), such as those that were in place in MWRDGC facilities at the time of their studies, the conditions of chlorination defined by the Pomona Virus Study can be characterized as extreme. The regulatory constraints that are imposed for most scenarios involving conventional disinfection are substantially less severe than those imposed by Title 22. As an example, Table 1 lists typical National Pollutant Discharge Elimination System (NPDES) permit limitations issued by the Indiana Department of Environmental Management (Indianapolis) that are relevant to disinfection operations. Similar discharge limitations are imposed by regulatory agencies in other states.

In practical terms, these constraints are met in well-run municipal wastewater treatment facilities by maintaining a chlorine residual of 1 to 2 mg/L (as Cl^sub 2^) for a detention time of 20 to 40 minutes, followed by tetravalent sulfur [S(IV)]-based dechlorination. When relevant, ammonia-nitrogen (NH^sub 3^-N) limitations are generally met by biochemical nitrification.

To put these treatment conditions into perspective, it is useful to characterize the conditions of chlorination used in each system. A “conventional disinfection” operation may accomplish disinfection based on a CT value (defined as the product of residual chlorine concentration and mean hydraulic detention time) of 40 to 80 mg . min/L, whereas a disinfection system that is implemented to satisfy the constraints of Title 22 may require chlorine exposure of more than 1000 mg . min/L.

Clearly, this range of possible chlorination conditions will yield a corresponding range of antimicrobial (and other) effects. In the time since the completion of the Pomona Virus Study and MWRDGC’s research on the subject of chlorine-based wastewater disinfection, UV irradiation practices have been adopted to meet the constraints imposed by wastewater treatment objectives. The conditions of UV irradiation required to satisfy Title 22 constraints (and similar reuse constraints in other areas) are substantially more severe than those required to meet the constraints of “conventional disinfection”.

Project Objectives

The examples described above illustrate a broad range of disinfection practices that are being applied across the United States. This range of practices brings with it a range of water quality issues that are relevant to human health. To address this range of issues, a research project was initiated to address the following two basic questions:

(1) Should wastewater disinfection be practiced?

(2) Under circumstances where the answer to question (1) is yes, how should disinfection be accomplished?

Because they represent the disinfectants of choice for municipal wastewaters in the vast majority of circumstances today, chlorination/dechlorination and UV irradiation were chosen for investigation in this research. Both disinfectants were applied in bench-scale systems to effluent samples collected from several municipal wastewater treatment facilities. Facilities were selected to provide a spectrum of effluent quality, particularly as related to effluent ammonia-nitrogen and organic nitrogen. Reduced nitrogen (in the forms of ammonia-nitrogen and organic nitrogen) was viewed as a critical factor relative to chlorine-based disinfection because of the formation of inorganic and organic chloramines, respectively. In general, inorganic chloramines are less effective than free chlorine for inactivation of planktonic microorganisms, while organic chloramines have little or no antimicrobial character (Donnermair, and Blatchley, 2003), but may represent sources of effluent toxicity (Gong et al., 2004).

Disinfectants were applied under conditions that were sufficient to accomplish compliance with relevant effluent discharge regulations based on microbial quality and chemistry. Bacterial and viral responses to disinfectants were then characterized. All compliance limits used as targets in this research corresponded with conventional disinfection, as defined above.

Materials and Methods

A complete description of methods and the results and conclusions of this research can be found in the final report for Water Environment Research Foundation (Alexandria, Virginia) project 99- HHE-l. Undisinfected effluent samples were collected atv five municipal wastewater treatment facilities, designated herein as facilities A to E for purposes of providing anonymity to the facilities. After collection, samples were packed in ice and shipped by express courier to participating laboratories for experimentation and analysis.

Facility A uses conventional primary clarification and activated sludge (with nitrification); effluent from this facility consistently displays high quality in terms of conventional bulk parameters such as five-day biochemical oxygen demand (BOD^sub 5^), total suspended solids (TSS), and ammonia-nitrogen. At the time of this research, facility B used conventional primary settling and activated sludge treatment without nitrification. Effluent quality for samples collected during this research was typically poorer than from the other facilities in terms of BOD^sub 5^, TSS, and ammonia- nitrogen. Facility C was somewhat unusual in that effluent was subjected to nitrification and denitrification before discharge. Facility D, which subjects the effluent to conventional secondary treatment (without nitrification) and filtration, produces a high quality effluent in terms of BOD^sub 5^ and TSS, but with a relatively high concentration of ammonia-nitrogen. Facility E uses conventional primary settling and activated sludge with nitrification.

Range-finding experiments were conducted with samples from each facility to determine bench-scale conditions of disinfection that would allow compliance with relevant discharge regulations, based on concentrations of viable indicator bacteria and residual chlorine. Based on these experiments, the conditions of disinfection required for each disinfectant were defined.

Chlorination/dechlorination was conducted in well-mixed batch reactors using an initial chlorine concentration of 2.0 mg/L (asCl^sub 2^) and a contact time of 40 to 60 minutes. The forms of residual chlorine that were generated in solution were defined by N,N-diphenyl-p-phenylene-diamine/ferrouos ammonium sulfate (DPD/ FAS) titration (APHA et al., 1998) and by membrane introduction mass spectrometry (Shang and Blatchley, 1999). For effluent samples from the nitrifying facilities (A, C, and E), residual chlorine existed as free chlorine; for effluent samples from non-nitrifying facilities (B and D), residual chlorine was present in the form of combined chlorine, with the vast majority of the residual being present as monochloramine (NH^sub 2^Cl). Samples were dechlorinated by addition of sodium thiosulfate in slight stoichiometric excess of the chlorine residual present at the end of the exposure period.

For most experiments involving UV radiation, disinfectant exposure was accomplished in small, well-mixed, batch reactors under a monochromatic (λ = 254 nm) flat-plate collimated beam (Blatchley, 1997). For experiments requiring relatively large sample sizes (i.e., more than 1000 mL), irradiation was accomplished using a capillary-flow reactor (Gong and Blatchley, 2002). The UV dose delivered to samples examined in this research ranged from 0-20 mJ/ cm^sup 2^. Disinfectant exposures were conducted at each participating laboratory immediately before initiation of experiments.

Long-Term Respirometry: Repair and Regrowth of Bacterial Communities Postdisinfection. Disinfected samples were examined by long-term respirometry for purposes of characterizing recovery of the bacterial community postdisinfection. After treatment (i.e., disinfection or control), samples were incubated at 25C under dark conditions for 6 days in a respirometer (OO-104 system, N-CON Systems Co., Inc., Crawford, Georgia) to study the long-term respirometric behavior of the treated samples. Acetic acid was added to the disinfected samples and controls as an artificial substrate at a concentration of 14.1 mg/L (approximately 15 mg/L biochemical oxygen demand [BOD]). Artificial substrate was added at this concentration to mimic the BOD concentration of typical receiving waters. Acetic acid was selected as the substrate because acetate has characteristics that are representative of substrates that can be expected in natural receiving waters (Shuler and Kargi, 1992).

Concentrations of viable fecal coliform bacteria and total bacterial concentration (TBC) were monitored over the course of respirometry experiments as measures of the responses of the bacterial community to disinfectant exposure. These measurements were conducted daily by membrane filtration and acridine orange staining, respectively (APHA et al., 1998). The respirometer was used to monitor oxygen uptake throughout the 6-day period of each experiment.

For each treatment facility, effluent samples were collected on four different dates and subjected to the same treatments. For each sample, four treatments (original sample without substrate, original sample with substrate, UV-irradiated sample with substrate, and chlorinated/dechlorinated sample with substrate) were applied. The output variables in this study included total oxygen uptake, viable fecal coliform concentration, total bacteria concentration, and the ratio of viable fecal coliform to total bacteria concentration. Because there were orders of magnitude differences in bacterial concentrations between disinfected and undisinfected samples, viable fecal coliform concentrations were log]0 transformed.

Other Bacterial Assays. Disinfected samples were also analyzed for viable fecal coliforms, enterococci, and total culturable bacteria. Fecal coliforms were assayed using the membrane filtration method described above. Samples from facilities C and D, both of which discharge through marine outfalls, were assayed for the presence of viable enterococci using a two-step membrane filtration method that uses the selective medium mE and EIA agars (APHA et al., 1998). Total culturable bacteria (TCB) counts were obtained by plating samples onto R2A agar (Reasoner and Geldreich, 1985).

Responses of Indigenous Bacteriophage to Conventional Disinfection. Coliphage analysis was performed by the double agar overlay method (Adams, 1959). Total coliphages (somatic and F- specific) were assayed on tryptic soy agar using a log-phase host culture of Escherichia coli (E. coli) C-3000 (ATCC #15597). The F- specific coliphages (F+ phages) were assayed using a log-phase culture of E. coli F^sub amp^-HS (pFamp)R (ATCC #700891) on tryptic soy agar augmented with streptomycin/ampicillin, as specified by U.S. EPA (2001).

Individual phage plaques surviving bench chlorination or UV irradiation were harvested and stored in 0.5-mL aliquots of phosphate buffered saline at 4C, then regrown to high liter. Bacteriophages isolated in this manner were further characterized by their nucleic acid content. Bacteriophages containing RNA were differentiated from those containing DNA by suppressing growth of isolates in the presence of 100 g/mL of RNAse (bovine pancreas, type I-A, Sigma, St. Louis, Missouri). Isolates that did not produce plaques in the presence of RNAse were classified as RNA phage.

Bacteriophages that were able to form plaques in the presence of RNAse were classified as DNA phage.

Results and Discussion

Long-Term Respirometry: Repair and Regrowth of Bacterial Communities Postdisinfection. The goal of these experiments was to assess responses of bacterial communities in wastewater effluents to disinfectant exposure. For qualitative characterization of bacterial community responses, the following two basic factors were considered: (1) the behavior of indicator organisms (fecal coliforms, used as a surrogate for bacterial pathogens) and (2) the behavior of the total bacterial community. Based on this approach, nine possible outcomes are possible, as summarized in Figure 1. Although it is clear that this method of analysis ignores many potentially important aspects of microbial (bacterial) ecology, it does allow for simple screening of the effectiveness of disinfection processes.

From Figure 1, it is evident that not all disinfection scenarios should (necessarily) be considered effective in terms of reducing human exposure to (bacterial) pathogenic microorganisms. It should be emphasized that this illustration is aimed at a qualitative assessment of the responses of the bacterial community to disinfectant exposure. In examination of the responses of viral and protozoan organisms, the outcomes may be somewhat different in that inactivation, and repair and recovery responses are likely to be substantially different than with bacteria.

Figure 1 lists nine possible scenarios that could develop among wastewater bacteria following disinfection; the effectiveness of a disinfection process can be judged by variations in the total bacterial community and the pathogenic (indicator) fraction. For example, cases c, g, and i may be judged to represent a positive effect of disinfection because they imply a reduction in pathogenic (indicator) bacteria. On the other hand, cases a, b, d, and e have an adverse effect because pathogenic (indicator) bacteria concentrations are not reduced. It is also interesting to note that, in cases f and h, it is difficult to unambiguously judge disinfection efficacy based on these two criteria. For these two cases, judgment of antibacterial efficacy requires additional information, such as the absolute concentration of viable pathogenic (indicator) bacteria.

To answer the question “is a disinfection process effective?” from the standpoint of bacterial risk, it is necessary to consider both regrowth and the pathogen (indicator) ratio. To do this, it is necessary to investigate the effects of upstream treatment processes, disinfection, and receiving waters on regrowth and the pathogen ratio. Under conditions of abundant substrate supply, rapid- growing microorganisms generally dominate populations. This is true in municipal wastewater treatment facilities, where the abundance of available organic substrates favors the growth of rapidly dividing bacteria, such as coliforms and pseudomonads. These dominant microbial populations in wastewater, which gain a competitive advantage because of their high intrinsic growth rates, are rapidly displaced in competition with other microbial populations of receiving waters as the concentration of organic compounds diminishes, owing to decomposition and dilution; under lower nutrient conditions, a more diverse community of slowly growing bacteria is favored.

For interpretation of the results of these experiments, several assumptions have been made, including the following:

* Fecal coliform bacteria can be used as indicators for pathogenic bacteria in disinfected wastewater effluent. This implies that fecal coliforms can be characterized as having (a) similar susceptibility to disinfection processes, (b) similar intrinsic growth rates, and (c) similar requirements for nutrients as pathogenic bacteria.

* The substrate (acetic acid) used in this study can represent the substrate condition of receiving waters.

* The addition of substrate does not affect microbial ecology relationships during incubation.

It is clear that these assumptions are not entirely valid, because of the following:

* Susceptibility to disinfection processes, intrinsic growth rates, and nutrient requirements for fecal coliform and pathogenic bacteria will be different; they are different even between two pathogenic bacteria. Coliform bacteria are commonly used as an “indicator” of microbial quality. However, it is clear that no single species can truly represent the broad range of microbial pathogens that could be present in a municipal wastewater effluent.

* The nutrient conditions of receiving waters are site-specific, so it is impossible to find a substrate that can be representative of all situations.

* Acetic acid canbe biodegraded easily; this will benefit those bacteria with high intrinsic growth rates. The composition of biodegradable compounds in receiving waters will vary.

The conditions used in these experiments provided a common basis for examination of the behavior of municipal wastewater effluents from several different facilities. As such, it was possible to compare the long-term behavior of multiple samples collected from each of the four facilities. The conditions of these experiments were believed to be generally representative of actual conditions in receiving waters; however, it is not reasonable to expect direct, quantitative translation of these results to conditions in the respective receiving waters.

Assessments of disinfection efficacy have traditionally been based on the inactivation or removal of fecal indicators, such as total coliforms, fecal coliforms, and fecal streptococci. However, there is little information to allow correlation between these indicator organisms and real pathogens, particularly in terms of their long-term behavior. Although the assumptions listed above are not entirely justified, it is necessary to use this approach because relatively little information has been obtained regarding the ecological relationships between fecal coliform and pathogenic bacteria, many of which are not culturable. Fecal coliforms have been chosen as the target microorganism because they represent a common indicator microorganism for wastewater effluent regulation.

Responses of Fecal Coliforms (Indicator) and Total Bacterial Counts. As described previously, the effectiveness of disinfection treatment processes was assessed based on an index test in which the dynamic behavior of fecal coliforms and total bacterial counts were examined throughout the course of incubation used in the long-term respirometry assays. Based on the 16 experimental runs (four treatment facilities, four replicates per facility), four different treatments were applied; fecal coliform and total bacteria concentrations were recorded for each case from t = 0 hours (immediately after treatment) to t = 144 hours. Because there were four replicates involved in each treatment facility, an, average value of the four replicates was used for representation of the total bacteria concentration and fecal-to-total-bacteria ratio. Based on this information, classification of disinfection process proceeded using Figure 1. Table 2 provides a summary of the results of these measurements for all four facilities and all four exposure scenarios (treatments). Recall that effluent samples from facilities B and D were non-nitrified, whereas those from facilities A and C had been subjected to nitrification (and denitrification, in the case of facility C).

In general, the treatments involving no disinfection (designated in Table 2 as “original without substrate addition” and “original with substrate addition”) resulted in an improvement in bacterial quality over the course of the 6-day incubation procedure. In contrast, overall bacterial quality remained essentially unchanged or degraded following the disinfection procedures. The decreases in bacterial quality were most evident in the application of chlorination/dechlorination to non-nitrified effluent samples, where +1-valent chlorine would have been present predominantly in the form NH^sub 2^Cl. While it is clear that chlorine- or UV-based disinfection will accomplish an immediate decrease in the concentrations of viable bacteria, it appears that the long-term effects of chlorination/dechlorination or UV irradiation may actually be detrimental to water quality, in terms of bacterial composition.

It is important to recognize that the changes among the bacterial populations were all normalized against the bacterial composition at t = 0, corresponding to the time at which disinfectant exposure was terminated. This method of normalization can provide a misleading representation of bacterial population dynamics, in that the basis of normalization was different for all samples. To address this issue, the data from the long-term respirometry experiments were presented in another form. Table 3 provides a summary of average total bacterial counts and average viable fecal coliform concentration for each facility and treatment. The non-normalized data in this table allow a direct comparison of bacterial changes among all of the treatments, over the period of incubation used in these experiments.

Several important trends are evident in the data presented in Table 3. First, the disinfection procedures generally accomplished effective inactivation of fecal coliform bacteria, although viable fecal coliform concentrations in some samples exceeded regulatory limits. The bacterial population that existed post-UV irradiation tended to decline over the period of incubation, as measured both by TBC and fecal coliform. In contrast, the bacterial population that existed postchlorination consistently increased, both in terms of TBC and fecal coliform. The TBC at the end of the incubation period was consistently higher in the samples that had been subjected to chlorination/dechlorination than any other treatment. In some cases, fecal coliform concentrations in the chlorinated/ dechlorinated samples were higher at the end of the period of incubation than the undisinfected samples. Disinfectant exposure appears to be effective for short-term control of bacterial populations; however, the data presented in Table 3 suggest that, from the standpoint of bacterial composition in the long-term, opting to skip disinfection may yield better water quality than application of disinfection. If disinfection needs to be implemented, these data indicate an advantage of UV irradiation relative to chlorination/ dechlorination.

Oxygen Uptake. Final oxygen consumption of each treatment was recorded after an incubation period of 144 hours. The resulting oxygen consumptions from each treatment were compared with initial ammonium concentrations based on 64 test samples, and the results are shown in Figure 2. The superimposed straight line represents the theoretical nitrogenous oxygen demand with a slope of 4.3 g (Vg [NH^sub 4^^sup +^]-N (Tchobanoglous and Burton, 1991). There was no clear trend between oxygen consumption and initial ammonium concentration when ammonium concentration was low (<1 mg/L), corresponding to effluent samples collected from nitrifying facilities. However, in samples collected from nonnitrifying facilities, where ammonium concentration was substantially higher, it was clear that the total oxygen consumption was strongly related to initial ammonium concentration, and higher than theoretical oxygen consumption based on ammonia oxidation was observed in most cases (except for UV irradiated samples). This suggests that most oxygen was consumed in the oxidation of ammonium (nitrification). For the range of ammonium-nitrogen concentration present in the samples that were tested in this research ([NH^sub 3^]^sub 0^ = 0 to 18 mg/L as nitrogen), the undisinfected samples, with or without substrate, typically yielded similar oxygen consumption, indicating that the artificial substrate did not cause a significant increase in overall oxygen uptake. Also, oxygen consumption in the undisinfected samples was generally higher than in the disinfected samples.

Oxygen consumption in UV-irradiated samples under high ammonium concentrations (14.35 and 18.33 mg/L as nitrogen) was substantially lower than theoretical oxygen consumption predictions based on biochemical oxidation of ammonia-nitrogen. The reason for this behavior is not clear, but it should be pointed out that these two samples were both from facility D, which uses filtration as an upstream treatment process. Guerrero and Jones (1996) indicated that exposure to visible light (400 to 475 nm) and near-UV irradiation (300 to 375 nm) will cause inhibition of nitrite oxidizers and ammonium oxidizers. Furthermore, they found that ammonium oxidizers are more sensitive to photoirradiation than nitrite oxidizers. Guerrero and Jones (1996) also investigated dark recovery of nitrifying bacteria after photoinhibition. They found that the recovery rates of Nitrosomonas cryotolerans and Nitrosococcus oceanus were slower when exposed to a short wavelength of radiation (300 nm), and the maximum recovery percentage appeared to be dependent on the wavelength of irradiation. However, no information regarding nitrifier response to UV^sub 254^ (or other germicidally active wavelengths) was discussed in their studies, nor was the relative susceptibility of nitrifying bacteria to UV irradiation included. Therefore, similarities in observed responses of nitrifying bacteria in this research can only be made by inference of a similar set of processes being induced by solar UV radiation and the UV radiation that characterizes low-pressure mercury lamps (λ = 254 nm), as used in this research.

Other Bacterial Assays. The concentration of indicator organisms present in a water sample is often used to represent microbial quality as an index parameter. Coliform bacteria are commonly used for this purpose, but many alternative indicator organisms have been proposed and are in use today. For example, enterococci are commonly used with facilities that discharge to a marine environment. Experiments were conducted to characterize the responses of fecal coliforms and enterococci to the bench-scale disinfection procedures used in this research. Figure 3 provides summaries of the results of these experiments for samples collected from facilities D and C.

While some variation was evident in the log^sub 10^ inactivation responses of fecal coliform bacteria among the various samples collected, inactivation responses measured on any individual sample by the two disinfectants were quite similar. Similar conclusions can be reached with regard to the enterococcus data. Collectively, t\hese data indicate that the conditions of disinfection by UV irradiation and chlorination/dechlorination were comparable, both in terms of coliform inactivation and enterococcus inactivation. Moreover, it appears that fecal coliforms and enterococcus are similar in terms of their behavior as indicator organisms.

These conclusions are based on effluent samples from facilities that produce effluents with substantially different residual nitrogen composition. Facility D yields an effluent that typically contains a relatively high concentration of ammonia-nitrogen, while effluent from facility C has been subjected to nitrification and denitrification. Therefore, residual chlorine composition in effluent samples from facility D was dominated by NP^sub 2^Cl, whereas the chlorinated samples of effluent from facility C were dominated by free chlorine.

The responses of TCB were also examined in disinfected samples with regard to their ability to function as an indicator group and because it represents an index of the total microbial burden to be imposed on a disinfection system. The TCB are those bacteria that can grow on laboratory media at a specific temperature during a given period of time. The assay used in this study is commonly referred to as the heterotrophic plate count assay (APHA et al., 1998). This method does not recover strict anaerobes.

The TCB made up from 4.70% (facility E) to 107.3% (facility D) of the TDC in the samples analyzed during this study. For comparison, fecal coliform comprised between 0.05% (facility C) and 39.4% (facility B) of the TCB, while enterococcus comprised between 0.01% (facility C) and 16.2% (facility D) of the TCB. These data clearly demonstrate that the conventional indicator groups of bacteria comprised only a small percentage of the total bacterial population that were present in the wastewater samples .. analyzed in this work. Even the most general and efficient method of culturing the total bacterial populations from these samples performed poorly. It is an established concept in microbial ecology, that culture-based methods will not recover all of the bacteria from any type of sample (Amann et al, 1995; Brock, 1987).

Pooled data from all facilities were used to assess the efficacy of each disinfectant. There were no significant differences (P = 0.052) between TCB abundances following disinfection with chlorine and UV radiation, but both treatments were significantly different (P < 0.001) from TCB abundances in the untreated samples. The TCB abundances in untreated, chlorinated, and UV-irradiated samples did not correlate with the fecal coliform or enterococcus abundances, in these same samples from the respective facilities. Additionally, there were significant differences (P < 0.001) between the TCB and the fecal coliform and enterococcus abundances in untreated, chlorinated, and UV-irradiated samples. These data indicate that the occurrence of and changes in the abundances of TCB is a dynamic process, as observed with fecal coliform and enterococcus. However, the TCB populations in the samples assayed in this study were ; not affected to the same extent by the actions of the disinfectants, as were the fecal coliform and enterococcus. This difference was most likely a result of significantly greater numbers of TCB than fecal ' coliform and enterococcus in each sample and the fact that the indigenous bacteria were inherently more resistant, based on culturability, to chlorine and UV disinfection than fecal coliform and enterococcus (Belkin et al., 1999; Matin and Harakeh, 1990; Olson and Stewart, 1987; Russell et al., 1997). Moreover, the TCB assay responds to a broader spectrum of bacteria than the fecal coliform or enterococcus assays. Consequently, it is reasonable to expect that the TCB population will display a broad range of susceptibility to externally applied stresses. The TCB assay does not differentiate among bacteria based on their sensitivity to disinfectants. Therefore, the observed behavior of the population of bacteria that respond positively to the TCB assay will be heavily influenced by bacteria that display natural resistance to a form of external stress, such as a disinfectant.

Current regulations that reference the use of this method (i.e., U.S. EPA’s Surface Water Treatment Rule, 40 CFR 141.74; U.S. EPA [2005]) recommend using a pour plate method and incubating at 35C for 2 days. Additional guidelines for the use of this method suggest that increasing the incubation period to 5 to 7 days and lowering the incubation temperature to between 20 and 28C will provide conditions for obtaining “the highest counts” (APHA et al., 1998). Previous work by Lisle et al. (1998 and 1999) has shown that bacterial growth rates on culture media are significantly reduced following exposure to disinfectants and that prolonged incubation at room temperatures significantly increased recovery efficiencies. It is worth noting that incubation periods for fecal coliform and enterococcus cannot be extended, as both media used in these assays would be overgrown with non-fecal coliform and non-enterococcus bacteria within 5 days. Additionally, the respective methods have been standardized for regulatory applications, and alterations to these methods invalidate any resulting data.

In this study, TCB incubation at 21 to 23C was extended to 14 days, with counts being conducted on days 2, 5, and 14. Figure 4 illustrates TCB recovery as a function of time for samples collected from facilities B and C. Similar data were collected for samples collected from facilities A, D, and E (data not shown). Several common trends were evident among these data sets. First, the TCB concentrations for the untreated samples were consistently greater than those for the chlorinated and UV-irradiated samples throughout the incubation period. second, there was a general increase in TCB values during the incubation period. In most, but not all cases, the samples that were subjected to UV irradiation yielded lower TCB counts than samples that had been subjected to chlorination/ dechlorination.

These data indicate that the TCB assay may represent a desirable alternative as an indicator test for microbial (bacterial) composition. An important advantage of this assay over conventional indicator testing based on coliform bacteria or enterococcus is that this assay represents a larger fraction of the bacterial population than either of the conventional indicator groups. This greater diversity in testing also yields a bacterial population that is likely to display greater apparent resistance to environmental stresses (e.g., disinfectant exposure) than testing based on conventional indicators because the bacterial population that yields a response to the assay has a greater range of sensitivities to disinfectant exposure than do coliforms or enterococci. In addition, TCB incubation can be conducted over a relatively long period of time, thereby allowing assessment of bacterial repair and recovery.

Viewed differently, diversity and incubation time can represent disadvantages of the TCB assay relative to conventional indicator assays. A drawback of diversity associated with this assay is that it provides less detailed information regarding the specific bacteria that respond to testing than do the more species-specific assays that are used for coliform or enterococcus testing. Furthermore, while extended incubation time does permit an assessment of repair and regrowth, it also represents a greater analytical burden than either of the conventional assays.

Responses of Indigenous Bacteriophage to Conventional Disinfection. Traditionally, assessments of antimicrobial efficacy in disinfection operations used for treatment of municipal wastewater have been based on measurements of the concentration of viable indicator bacteria. While these organisms satisfy some of the basic requirements of indicator organisms, several important shortcomings of their application for this purpose have been identified (see preceding discussion). Among the most important of these limitations are the relative ease with which most bacterial indicators are inactivated by common disinfectants and the fact that enteric viruses generally represent the most serious risk to human health among wastewater microorganisms.

Unfortunately, the assays used to assess viability (or infectivity) among human enteric viruses are time-consuming and expensive to conduct. In most situations, it is not practical to monitor for human enteric viruses. However, several indigenous phages have been identified that are structurally or otherwise similar to human viruses. Assays of phage viability (infectivity) are comparatively easy to conduct. Therefore, a series of experiments was conducted to assess the effects of common wastewater disinfectants on the concentrations of viable (infective) indigenous phages.

The concentration of indigenous bacteriophages in effluent samples from the five wastewater treatment facilities varied considerably, with the highest phage concentrations isolated from facility B. The phage population for this facility was comprised of both somatic and F-specific phages, with facility B displaying the highest concentration of F-specific phages of all the facilities examined. In decreasing order of initial phage concentration, the facilities were ranked as: B > A > D [asymptotically =] E > C. The samples containing the highest concentration of phages surviving either chlorine or UV disinfection generally reflected the ranking of the facilities with regard to initial phage concentration.

Although samples from all five facilities were analyzed for phage composition and dose-response behavior, the vast majority of useable data came from the analysis of samples collected from facilities A and B. Samples collected from facilities C and D had extremely low phage concentrations, such that it was difficult to asse\ss their dose-response behavior or nucleic acid content. The samples collected from facility E had quantifiable concentrations of viable phages; however, both disinfection schemes yielded samples in which F+ phage concentrations were below the limit of detection. One of the UV irradiated samples yielded a measureable concentration of somatic phage (see below).

Figure 5 illustrates representative examples of phage responses to exposure to UV and chlorine in samples collected from facilities A and B. The data in these figures illustrate several of the important trends that were observed in the data from the experiments focused on phage inactivation. First, the concentration of viable phages present in the samples was variable and low. Some evidence of seasonal effects was apparent in samples collected from winter and spring months at these two facilities, with summer phage concentrations being substantially higher than those observed in winter. second, the assay based on E. coli C-3000 consistently yielded higher concentrations of viable phages than the assay based on E. coli F^sub amp^.

Of particular importance in this work were the abilities of residual chlorine and UV radiation to accomplish inactivation of the indigenous phage. In the case of samples from facility A (a nitrifying facility), residual chlorine existed largely in the form of free chlorine. Exposure to chlorine under conditions that were shown to be capable of complying with discharge limitations generally yielded poor phage inactivation.

For samples that were subjected to UV irradiation from facilities A and B, phage inactivation was generally good. For the examples illustrated in Figure S, which contained some of the highest initial phage concentrations among the samples collected in this research, exposure to a UV dose of 20 mJ/cm^sup 2^ resulted in viable phage concentrations that were at or below the limit of detection.

Measurements of nucleic acid content were used as an index of phage diversity in disinfected samples. Table 4 provides a summary of nucleic acid composition measurements for surviving phages from selected samples from this portion of the research. In general terms, UV irradiation yielded much less diverse phage populations than did chlorination for the conditions of disinfection used in these experiments.

In general terms, the results of these experiments indicate that the conditions of disinfection (based on chlorination with either combined chlorine or free chlorine, or UV irradiation) that are needed to accomplish compliance with discharge regulations used in conventional disinfection operations yield incomplete inactivation of phages. Phage inactivation responses by UV irradiation were on the order of 2 log^sub 10^ units; phage inactivation by chlorine was less than this value. By extension, this suggests that these conditions of disinfection used for compliance with conventional disinfection may yield poor inactivation of enteric viruses.

Conclusions

The first of the two central questions that formed the basis of this research was “should municipal wastewater effluents be disinfected before discharge?” It is clear that no single response can appropriately answer this important question for all circumstances. The information presented above suggests that “conventional disinfection” of municipal wastewater effluents, as commonly practiced in the United States, is probably not as effective in preventing communicable disease transmission as is generally assumed. It appears that control of bacterial populations is generally effective in receiving waters only within a relatively short distance from the point of discharge. Moreover, viral inactivation accomplished by most systems (particularly those that use chlorination) is probably minimal. Therefore, in situations where direct human contact is likely or when ingestion of indigenous microorganisms in a near-outfall area is likely, it appears that disinfection of municipal wastewater effluents may yield some direct benefits. Anecdotally, it is interesting to note that human contact does occur in many such situations, and it is not obvious that the incidence of disease associated with these situations is abnormally high. Therefore, it, may be that the risks of disease transmission associated with these effluents may be less than expected. In situations where direct human contact is unlikely, it is not obvious that disinfection should be used as a default treatment process, at least not using the approaches that are common today.

With this in mind, it is also important to consider the second central question of this research, which is “under circumstances where disinfection is necessary, how should it be accomplished?” In applying any disinfectant, it is critical to strike a balance between minimizing risks associated with microbial pathogens and those associated with disinfection byproducts and related (chemical) lexicological issues. The data presented in this research indicate that UV irradiation and chlorination/dechlorination, when applied with the goal of complying with conventional effluent discharge regulations, are similar in terms of their ability to inactivate waterbome bacteria, although total bacterial populations generally recover to a greater extent in chlorinated effluents than in UV- irradiated effluents. Perhaps more importantly, the conditions that are used to accomplish bacterial (indicator) inactivation based on chlorination/ dechlorination appear to be relatively ineffective for control of waterborne viruses compared with UV irradiation. Therefore, in circumstances where wastewater disinfection is to be applied, it appears that UV irradiation is the method of choice, based on antimicrobial efficacy. However, disinfection practices that are consistent with the objectives of conventional disinfection, as defined herein, do not appear to be effective for inactivation of all pathogens. Decisions regarding the design, implementation, and operation of a disinfection system must be made on a site-specific basis taking into account these and other relevant factors.

Credits

This project was completed with the cooperation of management and operators from several municipal wastewater treatment facilities. Their participation in this project was both necessary and appreciated. This work was supported by a grant from the Water Environment Research Foundation (WERF) (Alexandria, Virginia). A complete report of the findings and recommendations of this investigation has been published and is available from WERF (project number 99-HHE-l).

Submitted for publication July 18, 2005; revised manuscript submitted December 23, 2005; accepted for publication January 24, 2006.

The deadline to submit Discussions of this paper is April 15, 2007.

References

Adams, M. H. (1959) Bacteriophages. Interscience: New York.

Amann, R. I.; Ludwig, W.; Schleifer, K. H. (1995) Phylogenetic Identification and In Situ Detection of Individual Microbial Cells Without Cultivation. Microbiol. Rev., 59 (1), 143-169.

American Public Health Association; American Water Works Association; Water Environment Federation (1998) Standard Methods for the Examination of Water and Wastewater, 20th ed.; American Public Health Association: Washington, D.C.

Belkin, S.; Dukan, S.; Levi, Y.; Touati D. (1999) Death by Disinfection: Molecular Approaches to Understanding Bacterial Sensitivity and Resistance to Free Chlorine, hi Microbial Ecology and Infectious Disease, Rosenberg, E. (Ed.); American Society for Microbiology: Washington, D.C., p 133-142.

Blatchley III, E. R. (1997) Numerical Modeling of UV Intensity: Application to Collimated-Beam Reactors and Continuous-Flow Systems. Water Res., 31, 2205-2218.

Brock, T. (1987) The Study of Microorganisms In Situ: Progress and Problems. In Ecology of Microbial Communities, Fletcher, M. (Ed.); Cambridge University Press: New York, p 1-17.

Donnermair, M. M.; Blatchley III, E. R. (2003) Disinfection Efficacy of Organic Chloramines. Water Res., 37, 1557-1570.

Gong, W. L.; Blatchley IU, E. R. (2002) Capillary Flow UV Reactor: Validation and Analysis by Chemical Actinometry and Point Source Summation. Proceedings of the 3rd IWA World Congress, Melbourne, Australia, April 7-12; International Water Association: London.

Gong, W.-L.; Sears, K. J.; Alleman, J. E.; Blatchley III, E. R. (2004) Toxicity of Model Aliphatic Amines and Their Chlorinated Forms. Environ. Toxicol. Chem., 23 (2), 239-244.

Guenero, M. A.; Jones, R. D. (1996) Pholoinhibition of Marine Nitrifying Bacteria. I. Wavelength-Dependent Response. Marine Ecol. Prog. Series, 141 (1-3), 183-192.

Lisle, J.; Broadaway, S.; Prescott, A.; PyIe, B.; Pricker, C.; McFeters, G. (1998) Effects of Starvation on Physiological Activity and Chlorine Disinfection Resistance in Escherichia coli O157:H7. Appl. Environ. MicmbioL 64 (12), 4658-4662.

Lisle, J.; Pyle, B.; McFeters, G. (1999) The Use of Multiple Indices of Physiological Activity to Assess Viability in Chlorine Disinfected Escherichia coli O157:H7. Letters Appl. Microbiol, 29 (1), 42-47.

Lue-Hing, C.; Lynam, B. T.; Zenz, D. R. (1976) Wastewater Disinfection: The case Against Chlorination, Report No. 76-17; Metropolitan Sanitary District of Greater Chicago: Illinois.

Matin, A.; Harakeh, S. (1990) Effect of Starvation on Bacterial Resistance to Disinfectants. In Drinking Water Microbiology, McFeters, G. (Ed.); Springer-Verlag: New York, p 88-103.

Oison, B. H.; Stewart, M. (1987) Factors that Change Bacterial Resistance to Disinfection. In Water Chlorination: Chemistry, Environmental Impact and Health Effects, Jolley, R. L. (Ed.); Lewis Publishers: Chelsea, Michigan, p 885-904.

Parkhurst, J. D. (1977) Pomona Virus Study: Final Report. Prepared for California State Water Resources Control Board and U.S. Environmental Protection Agency, Sanitation Districts of Los Angeles County, California.

Reasoner, D.; Geldreich, E. (1985) A New Medium for t\he Enumeration and Subculture of Bacteria from Potable Water. Appl. Environ. Microbiol., 49(1), 1-7.

Russell, A.; Furr, J.; Maillard, J. (1997) Microbial Susceptibility and Resistance to Biocides. ASM News, 63, 481-487.

Sedita, S. J.; Lue-hing, C.; Haas, C. N. (1987) Effects of Ceasing Chlorination on Selected Indicator Populations Downstream of Metropolitan Chicago’s Major Wastewater treatment Facilities, Report No. 87-17; Metropolitan Sanitary District of Greater Chicago: Illinois.

Sedita, S. J.; Zenz, D. R.; Lue-Hing, C.; O’Brien, P. (1987) Fecal Colifom Levels in the Man-Made Waterways of the Metropolitan Sanitary District of Greater Chicago Before and After Cessation of Chlorination at the West-Southwest, Calumet, and North Side Sewage Works, Report No. 87-22; Metropolitan Sanitary District of Greater Chicago: Illinois.

Shang, C.; Blatchley III, E. R. (1999) Differentiation and Quantification of Free Chlorine and Inorganic Chloramines in Aqueous Solution by MIMS. Environ. Sd. Technol., 33 (13), 2218-2223.

Shuler, M. L.; Kargi, F. (1992) Bioprocess Engineering: Basic Concepts. Prentice-Hall: New York.

Tchobanoglous, G.; Burton, F. L. (1991) Wastewater Engineering: Treatment, Disposal, and Reuse. McGraw-Hill: New York.

U.S. Environmental Protection Agency (2001) Method 1062: Male- Specific (F+) and Somatic Coliphage in Water by Single Agar Layer (SAL) Procedure, EPA-821/R-01-029; U.S. Environmental Protection Agency: Washington, D.C.

U.S. Environmental Protection Agency (2005) Analytical and Monitoring Requirements, 40 CFK 141. 74. Part 141-National Primary Drinking Water Regulations. Title 40-Protection of the Environment; U.S. Environmental Protection Agency: Washington, D.C.

Ernest R. Blatchley III1*, Woei-Long Gong2, James E. Alleman3, Joan B. Rose4, Debra E. Huffman5, Masahiro Otaki6, John T. Lisle7

1* Professor, School of Civil Engineering, Purdue University, 550 Stadium Mall Drive, West Lafayette, IN 47907-2051; e-mail: blatch@ ecn.purdue.edu.

2 Graduate Student, School of Civil Engineering, Purdue University, West Lafayette, Indiana.

3 Department Chair and Professor of Civil, Construction, and Environmental Engineering, Iowa State University, Ames.

4 Homer Nowlin Chair, Department of Fisheries and Wildlife, Michigan State University, East Lansing.

5 Research Associate, College of Marine Science, University of South Florida, Tampa.

6 Ochanomizu University, Tokyo, Japan.

7 U.S. Geological Survey, St. Petersburg, Florida.

Copyright Water Environment Federation Jan 2007

(c) 2007 Water Environment Research. Provided by ProQuest Information and Learning. All rights Reserved.