Search

This article consolidates and updates available information on Cryptosporidium removal by filtration, with a focus on optimal and suboptimal filtration performance. Emphasis was placed on integrating various strategies for optimizing filtration processes with the most current information regarding pathogen removal by filtration. As this review underscores, in response to regulations such as the Interim Enhanced Surface Water Treatment Rule (IESWTR) and the Long Term 2 ESWTR, there is a need to develop methods to compare the relative effectiveness of treatment alternatives and to understand the most significant operational effects on filter performance. In many cases, filters operated at or near optimal conditions can achieve in the vicinity of 3-log (or better) removal of Cryptosporidium. Operational conditions such as coagulant upsets and end-of-run (early breakthrough) filtration may significantly decrease oocyst removal by drinking water filters, even when filter effluent turbidities are low (

The goal of this project was to evaluate the removal of emerging pathogens and pathogen indicators by pilot-scale coagulation, flocculation, sedimentation, and granular media filtration. The emerging pathogens included Cryptosporidium parvum oocysts, Encephalitozoon intestinalis spores, Escherichia coli 0157:H7, and Aeromonas hydrophila. Bacteriophage MS2 and turbidity were used as pathogen indicators. This work revealed that some emerging pathogens were removed much more effectively than others. A. hydrophila was removed more effectively than C. parvum, and E. intestinalis spores and E. coli 0157:H7 were the least effectively removed. For the water tested in this study, the results suggest that a change in filter effluent turbidity requirements from 0.5 to 0.3 ntu would not achieve a significant improvement in the reliability of pathogen removal. However, by setting filter effluent turbidity goals below 0.2 ntu, significant improvements in microbiological quality could be obtained. In general, the pilot-plant data suggest that good removal of both turbidity and natural organic matter (NOM) decreases the risk of achieving poor emerging pathogen removal. In other words, optimizing the coagulation process for reduction of turbidity and NOM would improve removal of emerging pathogens.

Because much rier Cryptosporidium against attention of the this limitations has pathogen. oocysts been of directed and The many the ripening chemical fear toward of or waterborne filtration disinfectants maturation as disease the period for primary inactivating outbreaks, of the barfilCryptosporidium oocysts and the fear of waterborne disease outbreaks, much attention has been directed toward filtration as the primary barrier against this pathogen. The ripening or maturation period of the filter run may be the Achilles heel of this barrier. Several research groups have already examined Cryptosporidium removal during the filter-ripening period and found removals that were typically only slightly less than those of the same filters following the ripening period (Huck et al, 2000; Health Stream , 1999; Emelko et al, 1998; Patania et al, 1995). Many studies only seeded oocysts for a brief period (if at all) of the preceding filter run and allowed few oocysts to be present in the backwash remnant water that is responsible for a significant portion of the turbidity peak that follows backwashing. Amburgey and co-workers (2001a, 2001b) did seed the entire preceding filter runs of both conventional and biological filters and performed extensive sampling during the ripening period to more accurately assess the risk of Cryptosporidium passage. These researchers found that biological filters usually had effluent oocyst concentrations approximately equal to those found later in the filter run, but the conventional filter had oocyst concentrations from 3 to 18 times higher than the biological filters at the peak of the ripening period following an efficient backwash procedure. Hall et al (2000) found similar results following continuously seeded conventional filter runs. Overall, there appears to be only a brief period of increased Cryptosporidium passage during the ripening period of optimized pilot-scale filter runs, but this may not be the case for full-scale filters operating at suboptimal coagulation conditions and/or employing less efficient backwash techniques. Several enhancements of the normal filter-operating process have been explored in recent years in response to the risks encountered during the filter

Utility managers looking for strategies to assess pathogen removal by filtration with tools that are easier to use and more cost-effective than monitoring Cryptosporidium parvum directly will find this article of interest. The article discusses how oocyst-sized microspheres are easier, safer, and less expensive to work with than Cryptosporidium parvum and are also a reliable tool for assessing C. parvum removal by filtration It also shows that the ability of filters to remove C. parvum is not the same during all phases of the filter cycle. Even when a filter is in compliance with drinking water regulations, its C parvum removal capacity can be substantially worse near the end of a filter cycle compared with the earlier portions of the cycle This work also emphasizes that process evaluations must consider all operating periods, not just optimized operation. Further, it suggests that there can be a real difference in water quality even when differences in turbidity are very small.

In an effort to better control formation of disinfection by-products (DBPs), the US Environmental Protection Agency's recently promulgated Disinfectants/DBP Rule requires that US surface water treatment plants whose influent and effluent waters meet certain criteria practice enhanced coagulation. Although this treatment technique has been shown to effectively reduce natural organic matter, and consequently DBPs, questions have been raised concerning its effect on other aspects of water treatment such as particle and pathogen removal. The current study was designed to investigate the influence of decreased coagulation pH levels (an integral component of enhanced coagulation) on removal of Cryptosporidium oocysts as well as on reduction of total organic carbon (TOC), turbidity, and particle counts. A series of pilot-plant trials was conducted in which commonly used coagulants (ferric chloride, alum, and polyaluminum chloride) were used at various pH levels to treat river water spiked with large numbers of Cryptosporidium oocysts. The results showed that TOC removal is significantly enhanced by coagulation at lower pH levels and that all three coagulants are effective in removing Cryptosporidium oocysts during conventional treatment (mean removal = 4.3 log units). However, turbidity and particle counts appear to be unreliable indicators of oocyst removal. Finally, the investigation suggested that lowering coagulation pH does not interfere with removal of Cryptosporidium. However, questions remain concerning the use of alum at pH 5.

Recent findings that Cryptosporidium inactivation occurs at technically and economically feasible ultraviolet (UV) doses have generated a wellspring of interest in UV disinfection for drinking water. Many utilities are considering UV disinfection as a means to meet future regulatory requirements at their existing facilities. Most utilities evaluating UV disinfection will be retrofitting existing treatment plants. In this article, capital, operations and maintenance, and total annualized costs were estimated for retrofitting existing facilities with UV disinfection after the filters and before the clearwell. Cost estimates were developed for a UV dose of 40 mJ/cm 2 , which would ensure at least 2-log inactivation of Cryptosporidium, based on current research, and for a range of flows and filtered water qualities. The cost estimates presented indicate that UV disinfection is a relatively inexpensive method to achieve a high level of Cryptosporidium inactivation. As expected, annualized costs increased as system size decreased, although the costs were lower than those of other technologies providing similar levels of Cryptosporidium removal/inactivation.

This research examined removals of Giardia and Cryptosporidium by clarification either dissolved-air flotation [OAF] or lamella [plate] sedimentation) and by dual-media filtration following clarification. Studies were conducted for challenge conditions of cysts and oocysts (i.e., high influent concentrations) and for process design detention times and hydraulic loadings. To evaluate the effect of water temperature on performance, sets of experiments were performed for winter and spring seasons. DAF clarification performed better than lamella sedimentation and consistently resulted in lower turbidity levels and particle counts Overall log removals of Giardia and Cryptosporidium by clarification and filtration were about the same for DAF versus lamella sedimentation. Although DAF is not recognized under the Surface Water Treatment Rule or the Interim Enhanced Surface Water Treatment Rule, the authors argue that based on these findings, DAF plants should receive Giardia and Cryptosporidium removal credits at least equal to those received by sedimentation plants.

Faced with a state mandate to cover a 125-mil-gal- (0.4 × 10 6 -m 3 -) capacity finished water reservoir and with public opposition to using a floating cover, the Pittsburgh Water and Sewer Authority explored the possibility of postreservoir treatment options. The regulatory agency indicated that any acceptable posttreatment must ensure 4-log reduction of viruses, 3-log reduction of Giardia, and 2-log reduction of Cryptosporidium. The authority studied ozonation and membrane filtration for this application. Ozone inactivation trials resulted in a 1.0- to 1.2-log inactivation of Cryptosporidium oocysts at bench scale and 0.1- to 0.5-log inactivation at pilot scale. Membrane filter challenge studies demonstrated a 5-log removal of Giardia cysts and 6-log removal of Cryptosporidium oocysts. Based on the results of pilot testing and additional considerations, the authority is constructing a 20-mgd (0.07 x 10 6 -m 3 /d) microfiltration plant to treat water leaving the finished water reservoir.

The impact of storage of potable water in open reservoirs was assessed by examining inlet and effluent water samples from six open finished water reservoirs used by four New Jersey utilities. Water quality parameters investigated included Giardia cysts, Cryptosporidium oocysts, total and fecal coliforms, bacteriophage, heterotrophic plate count bacteria, turbidity, particle counts, chlorine residuals and other parameters. Fifteen percent of inlet samples and 25 percent of effluent samples contained the organisms. When data for cysts and oocysts were combined, the difference in concentrations between the inlet and effluent was statistically significant, even when results were adjusted for analytical recovery efficiency. Although the concentration of protozoa increased following open reservoir storage, the analytical method used does not permit assessment of the organisms' public health significance. Nearly all of the cysts and oocysts were empty or contained undiscernible internal structures, suggesting the health risk is low. The implication of these findings for watershed control programs is discussed.

A two-year evaluation of Giardia cyst and Cryptosporidium oocyst removal was conducted at a full-scale 900-gpm water treatment plant and at a 0.5-gpm pilot plant, both operated under conventional treatment and direct filtration regimes. Raw water quality (turbidity and algal content), as well as treatment effectiveness in removing turbidity, controlled the removal of seeded cysts more than the mode of treatment. A high correlation was found between cyst removal rates and removal of the respective size particles; poorer correlation was found between the cysts and turbidity removal, whereas no correlation was established between the removals of cysts and heterotrophic bacteria.

Bench-scale studies were used to investigate the effectiveness of dissolved-air flotation (DAF) for the removal of Cryptosporidium parvum oocysts from a drinking water supply. Oocysts were spiked into a natural water at a concentration of 3-4 x 10 5 oocysts/L. Results indicate that DAF achieved a >2-log removal of oocysts under a variety of conditions. In addition, flotation was a superior clarification process to sedimentation for the conditions tested. Coagulation and DAF conditions that minimize residual turbidity and maximize the removals of organic matter are conditions that produce high log removals of Cryptosporidium.

The American Water System has conducted extensive monitoring of its operations since 1988. Analysis of 347 surface water samples collected between 1988 and 1993 showed that the prevalence rate of Giardia and Cryptosporidium was 53.9 percent and 60.2 percent, respectively. But because the parasite assay does not indicate viability or virulence, these results do not necessarily indicate that these water systems were at risk from waterborne pathogens. To supplement coagulation and filtration, the average system will have to apply sufficient disinfection to reduce viable Giardia levels by 3.1 log 10 . An analysis of existing disinfection practices shows that most systems are already applying disinfectant at a level sufficient to reduce Giardia levels. However, the proposed Disinfectants/Disinfection By-products (D/DBP) Rule may hamper the ability of water utilities to apply sufficient disinfection under current operating conditions. Careful integration of the D/DBP and the Enhanced Surface Water Treatment rule is encouraged.

The authors conducted a blind survey of 16 commercial laboratories in order to evaluate their ability to test for Giardia and Cryptosporidium using the American Society for Testing and Materials (ASTM) method. After filters were spiked with approximately 740 Giardia cysts and Cryptosporidium oocysts or 500 cells of the alga Oocystis, they were sent to the laboratories for analysis

Between September 1986 and May 1988 a study of an operating slow sand water filtration plant was undertaken to determine its efficiency in removing Giardia and Cryptosporidium from a source water. The limited quantified data demonstrated an average 93 percent removal of Giardia cysts by the filtration plant, with this protozoan being detected in only one of 35 unfiltered water samples. An average 48 percent of Cryptosporidium oocysts was removed. Cryptosporidium was detected in 46 percent of the filtered samples. These findings suggest that Cryptosporidium may not be adequately removed from a contaminated source water under very cold operating conditions or if the filtration plant does not comply with accepted design standards. Other variables studied were turbidity and total and fecal coliform bacteria

This study examines concentrations of Giardia and Cryptosporidium, turbidity, and particles in raw and filtered waters as indicators of treatment plant effiiciency. Results indicate that the probability of detecting cysts and oocysts in finished water supplies depends on the number of organisms in the raw water. Particle counts are useful indicators of plant performance and can indicate when filters are not functioning properly.

Diatomaceous earth filtration was evaluated for removal of Giardia muris , Cryptosporidium , turbidity, and coliform bacteria from a surface water. The effect of cationic polymers and alum on this removal was also investigated. Essentially all of the Giardia cysts and >99.9 percent of Cryptosporidium oocysts were removed from waters containing approximately 2,000 cysts/L. The removal of Cryptosporidium was further improved by using alum. Greater removals of coliforms and turbidity, especially small particles (3 /xm or less), were also achieved when chemical coagulants were added. Use of these chemicals reduced the length of filter runs, even for low-turbidity waters, because of higher head losses. Filtration has long been used to remove particulates, including human pathogens, from water supplies. But filtration of surface waters became mandatory only recently with the passage of the