Page 14: Guidelines for Canadian Drinking Water Quality: Guideline Technical Document – Turbidity

• Previous Page
• Table of Contents
• Next Page

Appendix B: Log removal credits

Table B.1 shows the average potential removal credits estimated for Giardia, Cryptosporidium and viruses when treated water meets the turbidity values specified in this guideline technical document. These log removals are adapted from the removal credits established by the U.S. EPA as part of the LT2ESWTR (U.S. EPA, 2006b) and the Long Term 1 Enhanced Surface Water Treatment Rule (LT1ESWTR) Disinfection Profiling and Benchmarking Technical Guidance Manual (U.S. EPA, 2003a). Exact pathogen removal efficiencies will be dependent on the particulars of the water to be treated and the treatment process. Specific log reduction rates can be established on the basis of demonstrated performance or pilot studies. Facilities that believe they can achieve a higher log credit than is automatically given can be granted a log reduction credit based on a demonstration of performance by the appropriate regulatory agency. Under the multi-barrier approach to drinking water treatment, pathogen physical log removal credits should be used in conjunction with disinfection credits to meet or exceed overall treatment goals. Specific information pertaining to disinfection requirements can be found in the technical documents for enteric protozoa (Health Canada, 2012) and enteric viruses (Health Canada, 2011).

Cryptosporidium removal credits

When developing the LT2ESWTR, the U.S. EPA conducted an assessment of the available information on Cryptosporidium occurrence and treatment (U.S. EPA, 2005). As with the previous IESWTR and LT1ESWTR, the U.S. EPA indicated that the focus was preferentially placed on Cryptosporidium owing to the difficulty of removal through treatment, high infectivity and high chlorine resistance. For conventional filtration, the U.S. EPA concluded that plants in compliance with IESWTR turbidity requirements (combined filter effluent 95th-percentile value of = 0.3 NTU; maximum value of 1.0 NTU) will achieve a minimum 2-log removal of Cryptosporidium and that most filtration plants will achieve median reductions close to 3 log (U.S. EPA, 2006b). Making its recommendations, the U.S. EPA cited recent studies on the performance of various treatment technologies in the removal of Cryptosporidium (McTigue et al., 1998; Patania et al., 1999; Emelko et al., 2000; Huck et al., 2000; Dugan et al., 2001).

The U.S. EPA (2005, 2006b) also discussed studies of treatment plant performance in removing total particle counts and aerobic spores as indicators for estimating Cryptosporidium removal, citing the findings of Dugan et al. (2001), Nieminski and Bellamy (2000) and McTigue et al. (1998). Dugan et al. (2001) reported that aerobic spores and total particle counts were conservative indicators of Cryptosporidium removal across sedimentation and filtration, with full-scale plants reporting average reductions of close to 3 log for both parameters. Nieminski and Bellamy (2000) found that aerobic spores were good indicators of treatment effectiveness, but that spore removals did not entirely correlate with protozoa removals. In an evaluation of raw and finished water from 24 utilities, the authors noted average removals of 2.8 log for aerobic spores, 2.6 log for Giardia and 2.2 log for Cryptosporidium. McTigue et al. (1998) found a strong relationship between removals of Cryptosporidium and particles larger than 3 µm, with data showing a median particle removal of approximately 3 log.

The U.S. EPA also communicated the findings of their assessment of studies on Cryptosporidium removal when effluent turbidity was in the range of 0.1-0.2 NTU. The rationale was that treatment plants, in order to ensure compliance with a filter effluent turbidity value of 0.3 NTU, would target turbidity values in this range for their operations. Analyzing the summary data from four pilot-scale investigations, Patania et al. (1995) reported a median Giardia removal of 3.3 log when filter effluent turbidity was greater than 0.1 NTU. A similar relationship was observed for Cryptosporidium (Patania et al., 1995). Dugan et al. (2001), in a pilot-scale assessment of conventional filtration conditions, observed filter runs with effluent turbidity between 0.1 and 0.2 NTU and corresponding removals of greater than 3.2 log and 3.7 log for Cryptosporidium oocysts.

An evaluation of existing conventional filtration data was also conducted by the Dutch KIWA research group (KIWA, 2007). The exercise was performed as part of a project to develop estimates of the microorganism removal capacity for processes used in drinking water treatment. Under the group's assessment, individual studies were weighted according to the scale of the process involved (full scale, pilot scale or bench scale), the quality of the experimental conditions and the overall quality of the data. The weighted studies were then interpreted to generate values for the microorganism elimination credits (MECs) for individual filtration technologies. Average MECs for conventional filtration were estimated at 3.4 log (range of 2.1-5.1 log) for Giardia, 3.2 log (range of 1.4-5.5 log) for Cryptosporidium and 3.0 log (range of 1.2-5.3 log) for viruses.

In assessing the available data to establish log removal credits for various technologies, both the U.S. EPA and KIWA assessments and findings from pilot-scale studies by Patania et al. (1995) and Huck et al. (2002) were reviewed. The log removal credits established are listed in Table B.1. In a 2002 study, Huck et al. found that the average Cryptosporidium removals were significantly reduced under suboptimal coagulation. The authors further observed that when coagulant was absent for a short duration, Cryptosporidium removals were impaired by several log units, but that a 2-log oocyst removal was still maintained. During pilot-scale seeding experiments, Patania et al. (1995) noted that during breakthrough, Giardia removal decreased by 0.5 log, but a 3-log cyst removal was still maintained. No impact was observed on the capacity for removal of Cryptosporidium oocysts. Nieminski and Ongerth (1995) observed average log reductions for Cryptosporidium and Giardia of 3.0 log and 3.4 log, respectively, for pilot scale and 2.3 log and 3.3 log, respectively, for full scale, when the treatment plant was producing water with filter turbidity ranging from 0.1 to 0.2 NTU. In a summary of the data from an investigation of the removal of waterborne pathogens by pilot-scale conventional filtration, Xagoraraki et al. (2004) reported that filter effluent turbidity samples (occurring during ripening and filter breakthrough) ranging from 0.2 to 0.3 NTU corresponded with a median Cryptosporidium removal of 1.8 log (range of 1.2-2.6 log). Decreasing turbidity to below 0.2 NTU resulted in improved log reductions. However, an increased turbidity of up to 0.5 NTU did not produce a noticeable change in the median, maximum or minimum removal values.

Pilot-scale studies in which filter log removal capabilities were determined using low pathogen concentrations were also assessed in establishing log removal credits for various technologies, outlined in Table B.1. Concerns raised in the literature (Assavasilavasukul et al., 2008) that the estimation of filter log removal capabilities can be limited by the influent pathogen concentration in full-scale studies and that concentrations used in pilot-scale seeding studies may not reflect typical full-scale conditions were also considered. McTigue et al. (1998) conducted a series of pilot-scale spiking experiments with varying influent Cryptosporidium and Giardia concentrations. This study demonstrated that removals averaged 4 log for the filters and were not shown to be significantly affected by cyst and oocyst concentrations, which ranged from 101-103 per litre using both grab samples and continuous sampling. Pilot-scale studies measuring removals of Cryptosporidium and Giardia at low influent concentrations were also conducted by Assavasilavasukul et al. (2008). Mean log removals calculated from grab samples at low concentrations (100-103 pathogens per litre) ranged from 1.2 to 2.0 log for Cryptosporidium and from 1.5 to 2.6 log for Giardia. Mean log removals calculated from continuous sampling (sampling 288 L) at low concentrations (100-103 pathogens per litre) ranged from 1.4 to 2.3 log for Cryptosporidium and from 1.8 to 3.2 log for Giardia. Log removals for treatment trains that achieved pathogen concentrations below the detection limits were observed (undetectable cysts/oocysts per 10 L for grab sampling and undetectable cysts/oocysts per 288 L for continuous sampling), suggesting the capability for greater log removals, but only data with detectable cysts/oocysts were included in the analysis.

Based on its review, Health Canada agrees with and has subsequently adopted assumptions for Cryptosporidium log removal credits for conventional filtration similar to those of the U.S. EPA. It was concluded from the review that there still exists some uncertainty in the literature regarding the possible magnitude of additional log removal credits for full-scale facilities achieving turbidities of less than 0.1 NTU (Ongerth and Pecoraro, 1995; Assavasilavasukul et al., 2008). As a result, additional credits are not specified at this time.

For slow sand and diatomaceous earth filtration, the U.S. EPA in its LT2ESWTR assessment concluded that both technologies, when well designed and properly operated and in compliance with turbidity performance standards established under the 1989 Surface Water Treatment Rule (SWTR) (U.S. EPA, 1989) (= 1 NTU in at least 95% of measurements; and a maximum of 5 NTU), will be able to achieve Cryptosporidium removals similar to those attained with conventional filtration plants. As with conventional filtration, the U.S. EPA asserted that these technologies are capable of median Cryptosporidium removals close to 3 log.

KIWA's (2007) assessment of the microorganism removal capacity of slow sand filters estimated considerably higher log removal credits for Cryptosporidium (average 4.8 log, range of 2.7 to greater than 6.5 log).

In reviewing the small number of studies available for slow sand filters (Schuler and Ghosh, 1991; Hall et al., 1994; Timms et al., 1995; Hijnen et al., 2007) and diatomaceous earth filters (Schuler and Ghosh, 1990; Ongerth and Hutton, 1997) as well as the U.S. EPA (2006b) and KIWA (2007) assessments, it was determined that the U.S. EPA approach is appropriate for estimating the log removal credits for Cryptosporidium achievable through these filtration technologies. This approach subsequently adopted similar assumptions for log removal credits achievable with these technologies assigned in Table B.1.

For cartridge and bag filtration, it was determined that the U.S. EPA's assessment (U.S. EPA, 2006b) is appropriate--that the performance of these alternative filtration technologies varies among individual manufacturers and that it is currently not possible to propose general removal credits for these technologies.

Giardia and virus removal credits

The log removal credits for Giardia and viruses in Table B.1 are adapted from the removal credits established by the U.S. EPA in the LT1ESWTR Disinfection Profiling and Benchmarking Technical Guidance Manual (U.S. EPA, 2003a). In developing the LT2ESWTR, the U.S. EPA extended the requirements in this manual to this rule. In reviewing the available data, it was noted that the Giardia and virus credits in the manual were derived from a small number of studies that were available at the time of the U.S. EPA assessment (AWWA, 1991). It was also noted that the credits were intended for a higher 95th-percentile turbidity performance standard of 0.5 NTU, recommended under the U.S. EPA's SWTR. Lastly, in the documentation, the U.S. EPA used a conservative approach to assigning filtration credits and requiring the remainder of the total inactivation credits to be achieved through disinfection, as part of the multiple barrier concept (AWWA, 1991).

Recent studies of pathogen removal with conventional, direct, slow sand and diatomaceous earth filtration technologies have demonstrated log reduction capabilities for Giardia similar to, and in many cases greater than, those demonstrated with Cryptosporidium (Schuler and Ghosh, 1990,1991; Nieminski and Ongerth, 1995; Patania et al., 1995; McTigue et al., 1998; Nieminski and Bellamy, 2000; DeLoyde et al., 2006; Assavasilavasukul et al., 2008). On the basis of the currently available data, it has been determined that the data support filtration log removal credits for Giardia being the same as the log removal credits established for Cryptosporidium.

For viruses, information on the removal capabilities of the different filtration technologies has been limited. Recent data available on the removal of viruses and their surrogates (MS2 phage) by pilot-scale studies using conventional filtration (Rao et al., 1988; Harrington et al., 2001; Xagoraraki et al., 2004) are supportive of the virus log removal credits established in the LT1ESWTR Disinfection Profiling and Benchmarking Technical Guidance Manual (U.S. EPA, 2003a). Estimated virus removal credits established by KIWA (2007) were 3.0 log (range of 1.2-5.3 log) for conventional filtration and 2.2 log (range of 0.6-4.0 log) for slow sand filtration. Based on this information, it has been determined that there is not sufficient evidence to suggest the need to revise the virus filtration credits based on those in the LT1ESWTR Disinfection Profiling and Benchmarking Technical Guidance Manual (U.S. EPA, 2003a).

For bag, cartridge and membrane filtration, it was determined that the U.S. EPA's (2006b) assessment is appropriate--that the performance of these alternative filtration technologies varies among individual manufacturers and that it is currently not possible to propose general removal credits for these technologies.

Challenge testing

As indicated in Table B.1, cartridge and bag filtration can obtain log removal credits for Cryptosporidium and Giardia with the appropriate challenge testing. Under the U.S. EPA's LT2ESWTR, bag and cartridge filtration processes may receive a 1-log removal credit for Cryptosporidium for bag filtration that shows a minimum of 2-log removal in challenge testing and a 2-log removal credit for cartridge filtration showing a minimum of 3-log removal in challenge testing (U.S EPA, 2006b). Challenge testing is product specific; therefore, each manufacturer or third party can challenge test the filter unit to obtain a removal rating. For facilities wishing to find out more information, resources describing requirements for bag and cartridge filter challenge testing procedures include the U.S. EPA's LT2ESWTR Toolbox Guidance Manual (U.S. EPA, 2010) and the joint NSF/U.S. EPA Protocol for Equipment Verification Testing for Physical Removal of Microbiological and Particulate Contaminants (NSF, 2005).

Microfiltration, ultrafiltration, nanofiltration and reverse osmosis can receive log removal credit for Cryptosporidium, Giardia and nanofiltration and reverse osmosis can receive log removal credit for viruses if the process establishes a removal efficiency through challenge testing that can be verified by direct integrity testing and undergoes periodic direct integrity testing and continuous indirect integrity monitoring during use. The maximum removal credit that a membrane filtration process is eligible to receive is equal to the lower value of either the removal efficiency demonstrated during challenge testing or the maximum log removal value that can be verified through the direct integrity test (i.e., integrity test sensitivity) used to monitor the membrane filtration process (U.S. EPA, 2005, 2006b). For facilities wishing to find out more information, resources describing requirements for membrane filter challenge testing procedures include the U.S. EPA's Membrane Filtration Guidance Manual (U.S. EPA, 2005) and the joint NSF/U.S. EPA Protocol for Equipment Verification Testing for Physical Removal of Microbiological and Particulate Contaminants (NSF, 2005).

• Previous Page
• Table of Contents
• Next Page