United States
Environmental Protection
Agency
National Risk Management
Research Laboratory
Cincinnati, OH 45268
Research and Development
EPA/600/SR-97/025 July 1997
SEPA Project Summary
Removal of Cryptosporidium and
Giardia through Conventional
Water Treatment and Direct
Filtration
Eva C. Nieminski
Pilot- and full-scale evaluations of
Giardia and Cryptosporidium cyst re-
movals through direct filtration and con-
ventional water treatment were con-
ducted by the Utah Department of En-
vironmental Quality. Cysts were seeded
continuously in a step dose at a 0.5
gpm pilot plant, and in a spike at a 900
gpm full-scale plant; both plants were
operated under conventional treatment
and direct filtration regime. The results
of 20 pilot-scale cyst seeding trials and
8 full-scale trials indicated that source
water quality (turbidity and algal con-
tent), as well as treatment effective-
ness in removing turbidity, controlled
the removal of seeded Giardia and
Cryptosporidium. Changes in source
water quality influenced removal rates
more than the mode of treatment.
Higher removal rates were consistently
observed for Giardia cysts (3.3-log) than
for Cryptosporidium oocysts (3.0-log).
A high correlation was found between
cyst removal rates and removal of the
respective size particles; poorer corre-
lation existed between cysts and tur-
bidity removal, while no significant cor-
relation was established between the
removals of cysts and heterotrophic
bacteria.
To assure that the best available de-
tection method was used in enumera-
tion of the cysts in raw and treated
water, two versions of the immunofluo-
rescence staining method were evalu-
ated for their efficiencies in detecting
Giardia cysts and Cryptosporidium oo-
cysts seeded at known concentrations
in water: (1) the ASTM method for de-
tection of Giardia cysts and Crypto-
sporidium oocysts in low-turbidity wa-
ter and (2) a modified Sauch's proce-
dure employing sampling by 2.0 um mem-
brane filters, Percoll/Percoll step gradient
flotation, and immunofluorescence stain-
ing on 2.0 |jm porosity polycarbonate mem-
brane filters. The second method was se-
lected, since it was characterized by higher
recovery rates in all three types of waters
tested: raw surface water, partially treated
water from a flocculation basin, and fil-
tered water. Cyst and oocyst recovery effi-
ciencies decreased with increasing water
turbidity regardless of the method used.
Recoveries of seeded Giardia cysts ex-
ceeded those of Cryptosporidium oocysts
in all types of water sampled.
This Project Summary was developed
by EPA's National Risk Management
Research Laboratory, Cincinnati, OH,
to announce key findings of the re-
search project that is fully documented
in a separate report of the same title
(see Project Report ordering informa-
tion at the back).
Introduction
Problem Statement and Study
Objectives
The Surface Water Treatment Rule re-
quires all public water system treating sur-
face water to effectively remove enteric
viruses and Giardia cysts. The removal of
microbial contaminants by filtration is be-
ing re-evaluated by the U.S. Environmen-
tal Protection Agency in conjunction with
-------
the publication of the proposed Interim
Enhanced Surface Water Treatment Rule.
While the removal requirements for Giar-
dia may be increased depending on cyst
concentration in raw water, additional, more
stringent regulations may be developed to
control Ciyptosporidium in response to health
concerns surrounding this pathogen. As new
rules are being developed, allowable filtration
credits should be revisited and possibly re-
vised. Also, as the analytical methods for
detection of Giardia and Ciyptosporidium cysts
in water are being improved and new meth-
ods developed, surrogate water quality pa-
rameters should be established to allow for
an accurate, economical, and practical evalu-
ation of cyst removal effectiveness through
treatment.
The project objectives were designed to
address some of the questions associ-
ated with the development of the new
regulations. The specific tasks were to
examine the most critical relationships in
removal of Giardia and Cryptosporidium
by comparing the effectiveness of Giardia
and Cryptosporidium removal through con-
ventional treatment with that resulting from
direct filtration, the effectiveness of Giar-
dia removal with Cryptosporidium removal,
and the effectiveness of Giardia and
Cryptosporidium removals with the remov-
als of turbidity, cyst-size particles, and het-
erotrophic bacteria.
To enable accurate evaluation of Giar-
dia and Cryptosporidium removal efficiency
in water treatment processes, a reliable
method for measuring the concentration
of these pathogens in water must be used.
Therefore, another objective of this
project was to select an analytical method
capable of measuring the concentration of
Giardia cysts and Cryptosporidium oocysts
that would be accurate, reliable, flexible,
and verifiable, and could be applied to
measuring the cyst/oocyst removal perfor-
mance of water treatment processes.
Procedure
Evaluation of Analytical
Methods
Two methods for finding Giardia cysts
and Cryptosporidium oocysts in water were
compared. Both methods follow flotation
steps and immunofluorescence staining.
These methods were 1) The American
Society for Testing and Materials (ASTM)
method and 2) another immunofluores-
cence antibody (IFA) method, referred to
as the alternate method, and applied prin-
cipally by Ongerth and Stibbs. The objec-
tive was to evaluate the two IFA methods
using three factors as criteria for compari-
son. First, the applicability to cyst seeding
experiments in full- or pilot-scale water
treatment plant was evaluated. Second,
the applicability of the methods to cyst
detection in environmental water samples
of varying water quality was assessed.
The third criterion was the economics as-
sociated with the two methods. The supe-
rior method was then used in the cyst
seeding experiments in the pilot- and in
the full-scale treatment plant.
The ASTM method involves sampling
100 L or more of water through a 1.0 |im
porosity polypropylene yarn cartridge fil-
ter, extracting the particulates from the
cartridge filter, and concentrating the ex-
tracted particulates by centrifugation. The
concentrated particulates are then pro-
cessed to selectively concentrate cysts
and oocysts by flotation in 50 ml_ tubes
on a Percoll/sucrose gradient. The par-
ticulates recovered at the interface of the
Percoll/sucrose are stained with fluores-
cent-tagged antibodies on 25 mm diam-
eter, 0.2 urn pore size cellulose acetate
filters. After mounting on slides, the mem-
brane filters are scanned using a UV
epifluorescent microscope for objects of
the right size, shape, and fluorescence
characteristic as Giardia cysts and Crypto-
sporidium oocysts. On finding such ob-
jects the microscope optics are switched
to phase contrast to look for internal char-
acteristics of the organisms.
The alternate method involves filtration
of the water sample through either a 293
or 142 mm diameter, 2.0 |im pore-size
polycarbonate membrane filter; recovery
of particles from the filter by rinsing and
scraping them from the surface; and con-
centration of the particulates by centrifu-
gation. The cysts and oocysts are then
selectively concentrated from other par-
ticulates by flotation in 15 ml_ tubes on a
two-step Percoll/Percoll gradient, followed
by IFA staining on 13 mm diameter, 2.0
H.m pore-size polycarbonate membrane fil-
ters. After mounting on slides, the mem-
brane filters are scanned using an UV
epifluorescent microscope for objects of
the right size, shape, and fluorescence as
Giardia cysts and Cryptosporidium oocysts.
Confirmation of internal structures is not
performed in this method.
Cyst Seeding and Sampling
Procedures
Monitoring of raw and filtered water quality
was conducted throughout the seeding trials.
In addition to monitoring the major water qual-
ity parameters, particle counting in four size
ranges (2-4 urn, 4-7 nm, 7-14 ^m, and 14-25
H-m), was performed during the seeding trials.
Raw water sources were sampled and ana-
lyzed for background count of Giardia and
Cryptosporidium, naturally occurring in the
two watersheds.
Inactivated, formalin-fixed Giardia
lamblia cysts and Cryptosporidium parvum
oocysts were used in seeding experiments
in a pilot- and a full-scale treatment plant.
A total of 20 trials were conducted in the
pilot plant, and 8 trials in the full-scale
plant, alternating between conventional
treatment and direct filtration.
The first site for testing was a pilot plant
residing at the 180 MGD Jordan Valley
Water Treatment Plant in Bluffdale, UT.
The pilot plant simulated the actual, con-
ventional treatment plant, treating Provo
River water downstream from Deer Creek
Reservoir. Water flow rate was maintained
at 0.5 gpm. Alum was used as a coagu-
lant, at dosages established through jar
testing. After each seeding trial using the
conventional treatment train, the water was
re-routed through the direct filtration train
for seeding trials the following day. Alter-
nating between the conventional treatment
and direct filtration allowed for a compari-
son of treatment effectiveness of the wa-
ter of comparable quality.
The second site was a 900 gpm Hun-
tington Water Treatment Plant, situated
near Price, UT. The plant was operated at
600 gpm for cyst seeding experiments.
Polyaluminum chloride was used as a co-
agulant. The Huntington Plant was oper-
ated by conventional treatment during the
first four seeding trials. After converting
the plant to direct filtration mode, another
four seeding trials were performed.
Detection of Giardia and
Cryptosporidium Cysts
The alternate IFA method for sampling,
processing, and detection of Giardia and
Cryptosporidium cysts was chosen for the
seeding trials, based on results of the method
comparison. Samples were collected by mem-
brane filtration through 2.0 urn porosity, 293
mm diameter polycarbonate membrane fil-
ters, processed on Percoll/Percoll step gradi-
ent in 15 mL centrifuge tubes, stained on 2.0
(im porosity, 13 mm diameter polycarbonate
membrane filters, and enumerated under an
epifluorescent microscope.
In calculations of cyst removal through
treatment, a direct ratio of the difference
between the cysts seeded and the cysts
detected was calculated for each run in
the pilot plant. Two cyst removal rates
were determined based on two different
-------
initial cyst concentrations: concentration
in the seeding solution prior to being
pumped into the raw influent, and concen-
tration in the seeded influent already mixed
with the raw influent. The cyst removal
rates, achieved in the full-scale plant, were
also calculated as relative differences be-
tween the influent and effluent concentra-
tions, but the influent concentrations were
adjusted for dilution of cysts in respective
basins.
Three conservative assumptions were
made for data interpretation in calcula-
tions of the removal rates, both in the
pilot- and full-scale plant. First, the remov-
als based on the cyst concentrations found
in the seeded influent were used in data
analysis. Otherwise, the higher cyst con-
centrations found in the seeding solutions
would result in higher removal rates re-
ported. Secondly, cyst removal rates were
calculated only for trials in which cysts
were detected in both influent and effluent
samples. Otherwise, calculations of remov-
als during trials when cysts were not de-
tected in the effluent would be based on
very low detection limits, and therefore
would result in higher removal rates. Fi-
nally, no adjustments were made for the
differences in cyst recovery efficiencies in
turbid raw influent samples versus clean
filtered effluent samples. Adjusting for low
recovery rates in influent samples would
also result in higher removal rates being
calculated.
Results and Discussion
Evaluation of Two IFA Methods
for Detection of Giardia and
Cryptosporidium
The effectiveness of cyst recovery from
spiked water samples was impacted pri-
marily by the number of analytical steps
involved in the cyst detection. Both IFA
methods were characterized by low re-
covery efficiency, when seeded raw water
samples were filtered, then concentrated,
transferred to gradients, stained, and enu-
merated. Higher recovery rates were ob-
served in detecting Giardia cysts when
the alternate method was employed (Fig-
ure 1). In spiked raw water samples, an
average 12% of the seeded Giardia cysts
were detected by the ASTM method, while
the alternate method was characterized
by an average 49% Giardia cysts recov-
ery efficiency. Recovery rates for Crypto-
sporidium oocysts in spiked raw water av-
eraged 8% detected by the ASTM method
and 9% detected by the alternate IFA
method. The recovery rates in filtered wa-
ter were 14% and 52% for Giardia and
12% and 12% for Cryptosporidium, using
the ASTM, and the alternate method, re-
spectively. In flocculated water samples,
an average 22% of Giardia cysts were
detected using the ASTM method and 40%
using the alternate method. The recover-
ies of Cryptosporidium in flocculated wa-
ter were 7% with the ASTM method and
1% with the alternate method.
Losses were demonstrated to occur due
to incomplete yarn cartridge filtration. In
sampling raw water, 5% Giardia cysts and
6% Cryptosporidium oocysts were cap-
tured from the yarn cartridge filtrate by
passing it through a 293 mm diameter 2.0
|im pore size Nuclepore membrane filter.
An average 7% of Giardia and 8% of
Cryptosporidium were recovered from
membrane filters after cartridge filtration.
The sampling step resulted in a high loss of
seeded Giardia cysts and Cryptosporidium
oocysts. When the sampling step was elimi-
nated and cysts were seeded directly onto
flotation gradients, the resulting recovery rates
increased dramatically (Figure 2). An average
53% of Giardia cysts and 27% of Cryptospori-
dium oocysts was detected from seeded
Percoll/sucrose gradients used in the ASTM
method. The alternate method employing
Percoll/Percoll flotation, yielded recoveries of
82% for seeded Giardia cysts and 69% for
Cryptosporidium oocysts.
The highest cyst recovery rates were
reported when both sampling and flotation
steps were avoided and spiked with Giar-
dia cysts and Cryptosporidium oocysts
samples were stained directly onto the
membrane filters used for IFA assay. The
ASTM method resulted in recoveries of
72% and 56% for Giardia cysts and Crypto-
Giardia Cryptosporidium
ASTM method
Giardia Cryptosporidium
Alternate method
Flocculated
Water source
I I Raw, 5NTU
Filtered
Figure 1. Water quality vs. recovery of cysts seeded into water.
-------
sporidium oocysts, respectively. The al-
ternate method was characterized by an
average recovery of 86% for Giardia cysts
and 78% for Cryptosporidium oocysts.
Testing of cyst recovery efficiencies from
raw, flocculated, and filtered water indi-
cated that any increase in water turbidity,
whether due to presence of algae or to
that of chemical floe, resulted in a signifi-
cant decrease in parasite recoveries. The
results also indicate both methods were
more effective in detecting seeded Giar-
dia cysts than Cryptosporidium oocysts.
The results of testing the processing
and detection methods indicated that re-
coveries of cysts were substantially and
consistently higher with the alternate
method, employing Percoll/Percoll step
gradient combined with IFA on 13 mm
polycarbonate filters, than with the ASTM
method, in all three types of water. Con-
sequently, the alternate method was se-
lected as the method of choice for seed-
ing experiments. Added support for this
choice was provided by a comparison of
the qualitative advantages and disadvan-
tages of the two methods. In summary,
the main advantage of the ASTM method
was its ability to confirm presumptive cysts
and oocysts. The most serious disadvan-
tages of this method were its relatively
high cost and the amount of time required
to complete it. The alternate method, on
the other hand, was found to be less ex-
pensive and required less time to com-
plete than the ASTM method. Attractive
features of the membrane filter sampling
method include relatively small sample
volumes, flexibility, and compatibility with
frequent seeded controls. The major limi-
tation of the alternate method was its lack
of a confirmation step.
Experience with detecting Giardia cysts
and Cryptosporidium oocysts in the wa-
ters tested during this study and the re-
sults generated during this part of the
study, indicate that the alternate method
be recommended to evaluate water treat-
ment processes that use high concentra-
tions of seeded parasites in which algae,
occurring in concentrations much lower
than the seeded parasites, are not of con-
cern. The alternate method can be con-
sidered for analysis of environmental
samples, particularly for low-turbidity wa-
ters. When high water turbidity requires
higher-volume samples to be collected and
examined, and when cross-reacting algae
should be differentiated from the organ-
isms of interest, the ASTM method should
be used with environmental samples.
The results of this stage of the study
have indicated Giardia cysts and espe-
cially, Cryptosporidium oocysts are lost
during the gradient flotation steps of both
methods. Therefore, it is recommended
that the flotation step should be avoided,
whenever possible when processing
treated (filtered) water samples.
A hybrid method, combining the most
efficient steps from the two methods,
should be investigated. Such a hybrid
method should include sampling by mem-
brane filtration only for low-turbidity wa-
ters. High-turbidity waters should be
sampled by the ASTM cartridge sampling
method. Since the Percoll/Percoll step gra-
dient in 15 ml_ tubes is more economical
and had higher cyst recovery than the
Percoll/sucrose gradient, it should be used.
Staining on cellulose acetate membranes,
as opposed to polycarbonate membranes,
allows the demonstration of the internal
morphological characteristics of the organ-
isms. Consequently, staining on cellulose
acetate membranes should be incorpo-
rated into a hybrid method. Elvanol mount-
ing medium should not be incorporated
into a hybrid method. As a water-based
medium, it is not compatible with the de-
hydrated cellulose acetate membrane and
does not allow the membrane to be cleared
so that the cyst's internal structure can be
visualized by contrast microscopy.
Removal of Giardia and
Cryptosporidium through
Conventional Water Treatment
and Direct Filtration
A general observation about removal of
seeded Giardia and Cryptosporidium was
made that was valid in both pilot- and full-
scale plant throughout the entire seeding
studies. Consistent removal rates of Giar-
dia and Cryptosporidium were achieved,
when the treatment plant was producing
water of consistently low turbidity (0.1-0.2
CO
to
T3
S.D. =32
l
-S_D.=-8-
Giardia Cryptosporidium
ASTM method
Giardia Cryptosporidium
Alternate method
Water source
Flocculated I I Raw, 5 NTU
Filtered
Figure 2. Water quality vs. recovery of cysts seeded into flotation gradients.
-------
NTU). As soon as the plant's performance
changed, and resulting filtered water tur-
bidity fluctuated, a high variability in cyst
concentrations was detected in collected
samples.
Removal of Cryptosporidium was com-
pared with removal of Giardia by both
methods of treatment, conventional and
direct filtration, at both the full-scale plant,
and the pilot plant. Trials at the full-scale
plant were impacted by the change in
seasonal water temperature and algal con-
tent. Because of the need for a construc-
tion to bypass the sedimentation basin,
the first four trials by conventional treat-
ment were conducted from June through
September, while the four trials using di-
rect filtration were conducted in Novem-
ber and December. Greater flexibility of
the pilot plant allowed trials by both treat-
ment methods to be conducted within one
day of each other and enhanced not only
the comparison between removal of
Cryptosporidium with Giardia removal, but
also the comparison between conventional
treatment and direct filtration, as well as
the comparison of cyst removal with re-
moval of other water quality indicators.
Removal of Seeded Giardia and
Cryptosporidium During Pilot-
Scale Seeding Trials
Table 1 summarizes cyst removal rates
calculated based on cyst concentration in
seeded influent (after mixing the cysts with
the incoming raw water). These removal
rates, based on seeded influent concen-
Table 1. Removal of Giardia and Cryptosporidium Through Conventional Treatment
and Direct Filtration at Jordan Valley
Giardia removal
Cryptosporidium removal
Trial No.
Date
Percent
removal
Log
removal
Percent
removal
Log
removal
Conventional treatment
1-C
2-C
3-C
4-C
5-C
6-C
7-C
8-C
9-C
10-C
4/27/93
5/11/93
5/25/93
6/8/93
6/22/93
7/6/93
7/20/93
8/4/93
8/17/93
8/31/93
ND
99.16
ND
99.98
ND
99.95
99.95
ND
99.91
99.98
ND
2.20
ND
3.90
ND
3.69
3.69
ND
3.03
3.90
99.65
98.66
99.87
99.95
ND
99.88
99.45
ND
99.69
99.96
2.81
1.94
2.94
3.98
ND
2.94
2.64
ND
2.84
3.78
Average log removal
Standard deviation
Direct Filtration
3.40
0.67
2.98
0.64
1-D
2-D
3-D
4-D
5-D
6-D
7-D
8-D
9-D
10-D
4/29/93
5/13/93
5/28/93
6/15/93
6/23/93
7/8/93
7/22/93
8/5/93
8/19/93
9/2/93
Average log removal
Standard deviation
ND
ND
99.78
ND
ND
ND
99.90
ND
ND
99.99
ND
ND
2.90
ND
ND
ND
3.00
ND
ND
4.00
3.30
0.77
99.95
ND
92.06
99.96
ND
ND
99.80
ND
99.92
99.84
3.60
ND
1.31
3.78
ND
ND
2.90
ND
3.31
2.93
2.97
0.89
ND indicates that cysts were not detected in filter effluent.
tration, were consistently lower than the
removal rates based on cyst concentra-
tions in seeding solution (indicating 99.99%
or 4 log removal for both Giardia and
Cryptosporidium regardless of the treat-
ment mode). Calculations of cyst remov-
als, observed during seeding experiments,
can also be highly impacted by measure-
ments of cyst concentration in filter efflu-
ent samples. When cysts were not de-
tected in filter effluent samples, their con-
centration could be estimated based on
analytical detection limits, determined for
each sample batch. Such estimates lead
to underestimation of cyst concentration
in finished water samples and in turn,
result in overestimation of calculated cyst
removal rates.
Removal of Seeded Giardia and
Cryptosporidium During Full-Scale
Seeding Trials
Table 2 presents a summary of results
and removal rates calculated only from the
trials where cysts were detected both in
influent and effluent in the full-scale plant.
Similar to the pilot-scale experiments, the
removal of Giardia and Cryptosporidium can
be overestimated when calculations are
based on estimated effluent concentrations.
Removal rates based on cyst concentra-
tions detected and enumerated both in in-
fluent and effluent can be considered con-
servative.
Several factors impacted the results of
the full-scale seeding trials, which made
the comparison between conventional
treatment and direct filtration more depen-
dent on uncontrolled variables. Changes
in raw water quality, observed from the
time the plant was in operation by the
conventional mode, compared to raw wa-
ter quality during operation by the direct
filtration mode, influenced removal rates
more than the mode of treatment. The
water was treated in the conventional plant
during summer, when treatability was more
difficult, while direct filtration was used in
late fall, when the water was easier to
treat. The presence of prolific algal blooms
in samples collected during the first four
trials, and the lack of algal content in
samples from the last four trials, was an-
other variable making the comparison of
removal data problematic.
The results of the pilot-plant experiments
indicate that Giardia cysts were removed
more effectively than were Cryptospori-
dium oocysts. This observation was valid
regardless of the treatment mode. The
difference between log removals of Giar-
dia and Cryptosporidium ranged from 0.1
-------
Table 2. Removal of Giardia and Cryptosporidium Through Conventional Treatment and
Direct Filtration at Huntington
Giardia removal
Cryptosporidium removal
Trial No.
Date
Percent
removal
Log
removal
Percent
removal
Log
removal
Conventional treatment
1-C
2-C
3-C
4-C
6/11/92
7/7/92
8/5/92
10/6/92
99.95
ND
ND
99.66
Average log removal
Standard deviation
Direct Filtration
1-D
2-D
3-D
4-D
11/10/92
11/20/92
12/8/92
12/22/92
99.97
ND
99.97
ND
Average log removal
Standard deviation
3.7
ND
ND
2.82
3.26
0.67
3.87
ND
3.87
ND
3.87
0.00
99.60
99.05
97.87
ND
99.75
99.82
99.37
ND
2.78
2.07
1.89
ND
2.25
0.47
2.88
2.92
2.57
ND
2.79
0.19
ND indicates that cysts were not detected in filter effluent.
to 1.1 log and averaged 0.3 log, as calcu-
lated across all seeding runs. The differ-
ence between cysts and oocysts remov-
als were even more pronounced in the
full-scale plant than those observed in the
pilot plant.
Among the seeding trials, where the
cysts were detected both in influent and
effluent, the following average removals
were calculated for the pilot plant runs:
average removal of Giardia
through conventional treatment:
3.40 log; S.D. = 0.67
average removal of Cryptosporidium
in conventional treatment:
2.98 log; S.D. = 0.64
average removal of Giardia through
direct filtration:
3.30 log; S.D. = 0.77
average removal of Cryptosporidium
through direct filtration:
2.97 log; S.D. = 0.89
Similarly, the following average remov-
als were reported from the full-scale seed-
ing trials:
average removal of Giardia through
conventional treatment:
3.26 log; S.D. = 0.67
average removal of Cryptosporidium
in conventional treatment:
2.25 log; S.D. = 0.47
average removal of Giardia through
direct filtration; 3.87 log
average removal of Cryptosporidium
through direct filtration:
2.79 log; S.D. =0.19
Taking into consideration that the re-
ported removal rates of Giardia and
Cryptosporidium were calculated very con-
servatively, the values presented above
can be interpreted as expected removals,
resulting from a consistent performance
and steady operation of the treatment
plants.
Surrogate Parameters for
Evaluation of Giardia and
Cryptosporidium Removal
Both surface water sources treated in the
treatment plants were characterized by a high
quality water, with low levels of inorganic,
organic, and microbial contamination. Both
sources, however, were tested positive for
both Giardia and Cryptosporidium cysts, con-
firming previous hypothesis and observations
about these pathogens being ubiquitous in
surface waters.
Both effectiveness and consistency of re-
moval of seeded Giardia and Cryptospori-
dium cysts depended on the effectiveness
and consistency of the removal of turbidity.
When raw water turbidity was high and it
could not be removed by direct filtration, re-
sulting removals of seeded cysts were low
and inconsistent. On the other hand, if treat-
ment by direct filtration consistently produced
low turbidity effluent, resulting cyst removals
were comparable to those achieved from con-
ventional treatment.
The results of seeded cyst removals,
generated throughout the study regard-
less of treatment mode, were compared
with the respective results from particle
counting, turbidity measurements, and het-
erotrophic bacteria counts. Correlation be-
tween Giardia and Cryptosporidium cyst
removal and removal of these potential
surrogates are presented in Figures 3
through 5.
The analysis of correlation between cyst
removal and particle removal was per-
formed separately for Giardia cyst and
cyst-size particles, and for Cryptospori-
dium oocysts and oocyst-size particles
(Figure 3). High correlation was reported
between both sets of data. A correlation
coefficient of 0.82 was calculated (p<0.1)
for the relationship between Giardia cyst
removal and removal of particles ranging
between 7 |j,m and 11 um Similarly, a
correlation coefficient for the relationship
between Cryptosporidium oocyst removal
and removal of 4 |im to 7 |im particles
was 0.79.
The results indicated that particle count-
ing could serve as a reliable indicator of
cysts and oocysts removal. Particle
counters, even though capital intensive,
are cheap to operate and are more sensi-
tive than the assays used in Giardia and
Cryptosporidium analyses.
Much lower correlation was established be-
tween removals of Giardia and Cryptospori-
dium and removal of turbidity (correlation co-
efficients of 0.64 and 0.55, respectively). As
presented in Figure 4, log removal of tur-
bidity can be used as an indicator of cyst
and oocyst removals, but with lower accu-
racy than particle counting. The most pro-
nounced differences between removal of
Giardia and Cryptosporidium and the ex-
pected removals of turbidity were observed,
when very high cysts removals (4-log) were
reported.
-------
Heterotrophic plate count was not shown
to be a surrogate in evaluation of cyst
removals (Figure 5). No correlation was
found between log removal of seeded cyst
and log removal of HPC. Despite the fact
that the filters in both pilot- and full-scale
plants were backwashed with chlorinated
water, a growth of heterotrophic bacteria
was reported in the filters during the seed-
ing experiment and plant run.
Plant performance evaluation using par-
ticle counting and turbidity measurement
can be an effective tool in evaluating ex-
pected removals of Giardia and Crypto-
sporidium. The search for a biological sur-
rogate for Giardia and Cryptosporidium
should continue and result in identification
of a parameter that defines both occur-
rence and removal of Giardia and Crypto-
sporidium. The results of the study coin-
cide with previously reported relationships
between cysts and particulates. Effective
removal of Giardia and Cryptosporidium
from the water would, however, require
treatment plants to consistently produce
very low turbidity (0.1-0.2 NTU)much
lower than the levels currently required.
Conclusions and
Recommendations
Analytical Procedures for
Detection of Cysts in Water
Based on the results generated during
the evaluation of the effectiveness of the
IFA methods in enumeration of Giardia
cysts and Cryptosporidium oocysts, it was
concluded that the alternate IFA method
was more suitable for meeting the project
objectives. This method employed sam-
.ro
S
to
CD
4.5-
4-
3.5-
3-
2.5-
2-
1.5-
1-
0.5-
0
y = 0.9246X + 0.7221
R2 = 0.8247
0123
Log removal of 7-14jjm particles
I
:§
I
o
4.5
4-
3.5-
3-
2.5-
2-
1.5-
1-
0.5-
y = 0.8856X + 0.4647
R2= 0.7935
2 4
Log removal of 4-7 jjm particles
Figure 3. Relationship between removal of cysts and particles.
4.5'
5
g
5
CD
4-
2.5-
1.5-
°> X
o 1-
0.5-
0
y = 1.0093x + 1.6697
R2= 0.6429
0123
Log removal of turbidity
4.5
1 41
'§ 3.5-
I 3-
2-
1.5-
1-
y = 0.9631 x +1.1000
R2= 0.5623
1 2
Log removal of turbidity
Figure 4. Relationship between removal of cysts and turbidity.
-------
pling through a 2.0 |im polycarbonate
membrane filter, centrifugation in 15 mL
tubes, flotation on a two-step Percoll/
Percoll gradient, IFA staining on 13 mm
diameter, 2.0 |im pore-size polycarbonate
membrane filters, and enumeration under
an UV epifluorescent microscope. The fol-
lowing conclusions were formed:
The membrane filtration for low tur-
bidity samples outperformed the
polypropylene yarn sampling method
in terms of recovery efficiency of
seeded organisms.
The Percoll/Percoll step gradient had
better recoveries of seeded organ-
isms then the Percoll/sucrose gradi-
ent. Since 15 mL tubes were used in
place of 50 mL tubes, the procedure
of the step gradient flotation was more
economical.
The alternate method has proven
more effective in recovering seeded
cysts, and therefore, was considered
more suitable in parasite seeding ex-
periments, where evaluation of water
treatment process efficiencies was
conducted using high concentrations
of seeded cysts.
The ASTM method, employing staining
on cellulose acetate membranes, had the
advantage since the gradients could be
cleared and the internal structure of the
organisms could be visualized under the
phase- or differential-interference contrast
microscopy. The ASTM method, due to its
ability to confirm presumptive cysts and
distinguish between algal cells and the
cysts by contrast microscopy, was found
very applicable in testing of the environ-
mental water samples.
Based on the above results, it is recom-
mended that the alternate method should
be used in evaluating water treatment pro-
cess efficiencies using high concentrations
of seeded parasites. The ASTM method
is recommended in analyses of environ-
mental samples where the confirmation
step is essential.
A hybrid method, combining the most effi-
cient steps from the two methods, should be
investigated. The membrane filtration for low
turbidity samples, which far outperformed the
polypropylene yarn sampling method in terms
of recovery efficiency of seeded organisms,
shows promise in sample collection. Similarly,
Percoll/Percoll step gradient, used in the alter-
nate method, had better recoveries of seeded
organisms then the Percoll/sucrose gradient.
Since 15 mL tubes were used in place of 50
mL tubes used in the ASTM method, the
procedure of the step gradient flotation was
more economical. On the other hand, staining
on cellulose acetate membranes used in the
ASTM method, had the advantage since the
gradients could be cleared and the internal
structure of the organisms could be visualized
under the phase- or differential- interference
contrast microscopy.
The results of the testing of IFA method
effectiveness have indicated that Giardia cysts
and especially Cryptosporidium oocysts are
lost during the gradient flotation steps of both
methods. Recovery efficiencies increased dra-
matically in samples, did not contain much
debris, and could be processed without the
flotation steps and stained directly on mem-
branes. Therefore, it is recommended that the
flotation step should be avoided when pro-
cessing treated (filtered) water samples when-
ever possible.
Removal of Giardia and
Cryptosporidium through
Conventional Treatment and
Direct Filtration
The following conclusions were formed
from the pilot- and full-scale study on Gia-
rdia and Cryptosporidium cysts removal
through conventional treatment and direct
filtration:
In a properly operated treatment plant
effectively removing turbidity to 0.1-
0.2 NTU, either conventional treat-
ment or direct filtration can result in a
3-log removal of Giardia.
Cryptosporidium oocysts are more dif-
ficult to remove than Giardia cysts,
3.5-
ro
? 3H
2.5-
2-
1.5-
1-
0.5-
0
y = 0.7675X + 2.5245
R2= 0.0841
0 0.5
Log removal of HPC
E
.g
'c
o
CL
U)
IL
O
'o
"ra
o
E
0)
D)
O
3.5-
3-
2.5-
2-
1.5-
1-
0.5-
n -
^_
^^ ^^fc^^
^* *^* ^
^ *"
*
y = 0.7586X + 2.2524
R2= 0.0753
0.5 1
Log removal of HPC
Figure 5. Relationship between removal of cysts and heterotrophic bacteria.
-------
both in a conventional plant and
through direct filtration (up to 1.0-log
difference).
Removals of cyst-size particles and
removal of turbidity can be used as
indicators of cyst removal effective-
ness.
A general observation about removal
of seeded Giardia and Cryptosporidium
cysts was made that was valid in both
pilot- and full-scale plant throughout the
entire seeding studies. Both effectiveness
and consistency of removal of seeded
Giardia and Cryptosporidium cysts de-
pended primarily on the effectiveness and
consistency of the removal of turbidity.
When treatment by direct filtration con-
sistently produced low turbidity effluent
(0.1-0.2 NTU), the resulting Giardia and
Cryptosporidium cyst removals were con-
sistent and comparable to these achieved
from conventional treatment. As soon as
the plant's performance changed, and re-
sulting filtered water turbidity fluctuated,
a high variability in cyst concentrations
was detected in collected samples. When
raw water turbidity was high, and it could
not be removed by direct filtration, result-
ing removals of seeded cysts were low
and inconsistent.
A high correlation coefficient was calcu-
lated for the relationship between Giardia cyst
removal and removal of particles ranging be-
tween 7 urn and 14 urn, and similarly, for the
relationship between Cryptosporidium oocyst
removal and removal of particles of 4 \m to 7
|im in size. Much lower correlation was estab-
lished between removals of Giardia and
Cryptosporidium and removal of turbidity. Het-
erotrophic plate count was not shown to be a
surrogate in evaluation of cyst removals, with
no correlation found between log removal of
seeded cyst and log removal of HPC.
A combination of particle counting and
turbidity measurement was shown to be
an effective tool in water treatment plant
performance evaluation in terms of pre-
dicting removals of Giardia and Crypto-
sporidium. Effective removal of Giardia and
Cryptosporidium from the water would,
however, require treatment plants to con-
sistently produce very low turbidity (0.1-
0.2 NTU)much lower than the levels
currently required. The results of the study
indicate that the removal of particulates,
measured through particle counting and
turbidity monitoring, should be a critical
factor used in the evaluation of plant per-
formance in Giardia and Cryptosporidium
removal. Continuous and consistent re-
moval of particulates should be monitored
by continuous particle counting and tur-
bidity monitoring.
The results of the project imply that the
credits given for Giardia cyst removal in
direct filtration plants, may be similar to
credits obtained in conventional treatment
plants, and also may be higher than the
credits applicable under the current regu-
lations. Since Cryptosporidium is more dif-
ficult to remove than Giardia, and it is also
more resistant to disinfection than Giar-
dia, new requirements need to be devel-
oped to control this pathogen. Finally, due
to the need of further defining the credits
given to treatment plants for physical re-
moval of Giardia and Cryptosporidium, a
study on evaluation of removal of these
pathogens through pre-sedimentation
should be conducted.
The full report was submitted in fulfill-
ment of CR818895-010 by the Utah De-
partment of Environmental Quality under
the sponsorship of the U.S. Environmen-
tal Protection Agency.
-------
Eva C. Nieminski is with the Utah Department of Environmental Quality, Salt Lake
City, UT84114.
Kim Fox is the EPA Project Officer (see below).
The complete report, entitled "Removal ofCryptosporidium anc/Giardia through
Conventional Water Treatment and Direct Filtration," (Order No. PB97-162507;
Cost: $35.00, subject to change) will be available only from:
National Technical Information Service
5285 Port Royal Road
Springfield, VA 22161
Telephone: 703-487-4650
The EPA Project Officer can be contacted at:
National Risk Management Research Laboratory
U. S. Environmental Protection Agency
Cincinnati, OH 45268
United States
Environmental Protection Agency
Center for Environmental Research Information
Cincinnati, OH 45268
Official Business
Penalty for Private Use $300
BULK RATE
POSTAGE & FEES PAID
EPA
PERMIT NO. G-35
EPA/600/SR-97/025
------- |