EPA/600/A-92/184
PROCEEDINGS
Water Quality
for the New Decade
ANNUAL CONFERENCE
Philadelphia, Pennsylvania
June 23-27, 1991
Sponsored by
American Water Works Association
;a
6666 W. Quincy Ave., Denver, CO 80235

-------
MICROBIOLOGICAL CHANGES IN SOURCE VttTER TREATMENT:
REFLECTIONS IN DISTRIBUTION WATER QUALITY
Edwin E. Geldreich,
EPA Senior Research Microbiologist
INTRODUCTION
Microbial barriers in treatment processes are the major block to the
passage of waterborne pathogens from raw source waters. Many of the
processes utilized in water supply treatment have saae inpact on
microbial densities and survival. For instance, raw water storage for
24 to 48 hours can generally reduce the bacterial load by 501,
coagulation-sedimentation by 60% and filtration by 99.9 percent under
favorable conditions. - Ccnbining these processes in a series of
successive treatments can provide a emulative reduction in waterborne
organisms so that the burden on final disinfection to achieve a 6 log
reduction of bacteria, 4 log reduction in virus and 3 log reduction in
protozoan cysts is possible on a continuous basis. The key variable in
this case is fluctuating source water quality that can inpact treatment
barrier effectiveness.
Treatment barrier effectiveness can also change as a consequence of
operational changes at the plant.;—Many water utilities are seriously
reviewing the need to modify treatment operations in an atterpt to
reduce the formation of disinfectant by-products either by reducing
organic precursors or changing the type of disinfectant for less
reaction products. These moves oust be considered carefully because of
possible adverse repercussions on treatment barrier effectiveness and
ultimately on coliform compliance for the water supply, the worst case
scenario being a waterborne outbreak.
FACTORING SOURCE WATER QQALITY IN DISTRIBUTION COMPLIANCE
It is obvious that without adequate treatment, there will be non-
compliance problems with distribution water. Treatment needs to be
designed to adequately process any given water supply with a wide margin
of public health safety. Excluding those situations where filtration
has been found to be an essential treatment ccnponent for Giardia
control in many "protected" surface water supplies, there are other more
subtle water characteristics that may degrade treatment barriers and
release coliforms into the distribution system.
CHLORINE DEMAND IN SOURCE WATER
Sudden chlorine demand changes in raw source water that are not adjusted
for during treatment operations may result in ineffective disinfection
for some time interval, thereby providing opportunities for coliforms
to pass into the distribution system. Springtime snow melt, major
storms over the watershed, seasonal turnovers in inpourxinents, algal
blooms, drought conditions and applications of agricultural fertilizers
to fields in the watershed can introduce a variety of substances ttat
exert a chlorine dsiand and reduce chlorine availability to treat water.
Of course, immediate attention to increasing chlorine dosage is
necessary but may not be applied pronptly to adjust to the changing
water quality. Two case histories illustrate possible scenarios that
lead to coliform biofilm problems in distribution systems.
GELDREICH
107

-------
Case History Mo. 1; This city utility is in Hew York State. Most of
the raw source is obtained from two lakes (Hemlock and Canadice) with
a lesser amount derived from Lake Ontario. The water is chlorinated,
fluoridated and gravity fed into the city's storage and distribution
system. Ninety percent of the monthly turbidity averages for a recent
five year period were less than 2 HTU, with a maximum monthly average
of 3 NTU. However, in late winter {1986} there were periods of
exceptionally heavy rains resulting in elevated turbidity and increased
chlorine demand, that at tiroes exceeded 1 mg/L. Conpensating increases
in chlorine applied were not promptly made to these changing souree
water conditions. By February, coliform occurrences started to increase
substantially to the point that the coliform 1CL was exceeded for two
consecutive months. Coliforms had penetrated the unfiltered treatment
process and within the next few months these organisms proceeded to
became established in the nutrient rich pipe sediments. With onset of
warn water conditions of late spring and summer, colonisation was
evident by releases of coliforms from the biofilm throughout the
distribution network.
Case History 2: Source water for this utility is a small stream, the
White River, in Indiana. Upstream of the water intake there are few
municipal and industrial discharges to the river; however, the watershed
is an intensive agricultural area that is the cause of spring time
elevated turbidities, high bacterial counts and anraonia nitrogen
concentrations frcro fertilizers that reach 7 mg/L. This seasonal change
in raw source water created an elevated chlorine demand, as well as a
troublesome taste and odor problem. In the past, the conventional
treatment plant carried a combined chlorine residual throughout the
treatment train and into the distribution system during this seasonal
period of high aaocnia levels. In this situation breakpoint chlori-
nation had been successful. Chlorine dioxide (0.5 mg/L) was applied as
a partial substitute for the normal 2 mg/L combined chlorine in past
disinfection to minimize taste and odor formation. The first coliform
positive samples appeared in the distribution systan shortly after the
seasonal application of chlorine dioxide-combined chlorine disinfectants
aided and the utility was back to a detectable free chlorine residual.
This situation suggested that the application of chlorine dioxide as a
partial substitute for combined chlorine had not been fully effective,
resulting in coliforms passing through the treatment barrier in late
February or early March during the first heavy spring runoff.
Subsequently toe distribution system was colonized with an active
biofilm that went into accelerated growth during seasonal warm water
with bacterial releases into the main flow of distribution water.
RAW WATER pH SHIFTS
Seasonal changes in raw water pH may contribute to the instability of
sediments coating the pipe walls. This change in sediment stability can
lead to sporadic releases of coliform bacteria and other viable
organisms in the attached biofilm or entrapped in the accumulation of
particulate deposits.
Case History: Such a situation was created at one utility in Illinois,
serving 32,000 people. Attention to a problen began when coli forms were
noted to appear only during the cold water months of Decenber to June.
This winter occurrence of coliforms vas unusual since irost coliform
biofilm release problems have been observed to take place during warm
water periods. An on-site review of treatment practices and plant
records indicated that there had been a pronounced shift in the Lake
108

-------
Michigan source water pH during the winter. Inspection of data on raw
water characteristics revealed water pH of 7.7 in sumner shifted to pH
8.2 by December, followed by a rapid decline to pH 7.4 during January
to March each year. The reason for these seasonal changes in water pH
was thought to be related to near shore turnover of bottom water
containing partially decayed vegetation debris (humic matter). Water
treatment measures used to process the lake water had little iirpact on
stabilizing the water pH, so this characteristic was passed on into the
distribution system.
Inplementation of reccrmendations to adjust the water to pH 8.3 prior
to release from the plant and to add lime slowly in the process basin
to form a more stable, firm coating on the pipe walls, apparently
resolved the collform occurrence problem in the following year.
Successful follow-up treatment may also have been aided by the
suggestion to increase the disinfectant concentration during cold water
periods to conpensate for the increased chlorine denand and reduced
disinfectant effectiveness at near-freezing water tenperatures.
OOLD SOURCE WATER TEMPERATURES
Cold water temperatures have an influence on disinfectant effectiveness.
At tanperature of 5% and below, inactivation of organisms by disin-
fectants requires a longer contact time or an increase in concentration
to achieve the same kill rate as at 22°C. The problem is of particular
concern for surface water supplies in rv -hem latitudes. Since contact
time is generally fixed the only meth available to achieve adequate
contact time (C-T) values is to increas iie disinfectant concentration
applied in the contact basin.
Case History: A utility experience in Alaska provided an example of
this situation. In this instance, the utility is part of a military
base that supplies water to two separate military camunities. Surfaae
water is generally processed by conventional treatment most of the year.
However, in late autumn surface source water is passed through a rapid
sand filter without the benefit of coagulation, then disinfected prior
to entry into the two distribution systems.
With raw water temperatures in January and February stabilized to
approximately 1 to 5% at the intake, the applied chlorine dosage in the
contact basin was not increased. As a consequence, a species of
Klebsiella began to be detected in one of the distribution systetts
nearest to the utility and the water supply was in non-carpiiance for
2 to 3 months. The other distribution system, serving the second
military base was not affected. This response was probably a result of
the extended contact time for chlorine exposure before water arrived at
the first customer location, miles further away from the treatment
plant. Under winter-time conditions in northern latitudes, the applied
chlorine dosage to surface source waters should be increased to achieve
effective inactivation of bacteria and viruses.
TREAIMPTT MODIFICATIONS REFLECTED IN DISTRIBUTION WATER QUALITY
Modifications in water treatment unit.processes or in their sequential
placement to optimize reductions in disinfection by-product formation,
nust be cautiously evaluated and monitored for inpact on microbial
barriers and on distribution water quality. Four major treatment
concepts either in pilot plant or full scale, may cause changes in
microbial quality: (a) changing the point of free chlorine application;
109

-------
(b) applying granular activated carbon (GAC) adsorption for organic
renoval; (c) use of biological activated carbon (BAC) for further
reduction of dissolved organics through microbial activity; and (d)
employment of alternative disinfectants (chloraroines, chlorine dioxide
and ozone) to reduce trihalcmethane (THM) formation.
DISINFECTION - POINT-OF-APPLICATION
In the trade-off to minimize disinfection by-product formation while
maintaining microbial integrity, (by moving the point of disinfection
application) some migration of organisms deeper into the treatment train
may occur or changes in the microbial flora of process waters will
evolve. In the worst case scenario, microbial colonization of the
process media materials may result in periodic releases of biofilm
aggregates containing coliform bacteria into the finished water.
Case History: Changing the site for chlorine application was
investigated at the Cincinnati, Ohio Water Works (Table 1) in a series
of two week study periods (1). During routine treatment plant
operations, chlorine was applied to the source water after 48 hours of
open reservoir storage (Table 1). Adequate retention time of raw source
waters is a beneficial first step in microbial population reductions
through self-purification processes and can be a buffer against
taiporary impairment of water quality fran an upstream accidental spill
of industrial chemicals. In the Cincinnati water treatment operation,
coagulant has added to the water as it exited the open-reservoir and
chlorine was routinely applied ahead of in-plant treatment processes.
In the modified treatment operation, chlorination was delayed (Table 1)
until after an additional four-hour clarification process consisting of
coagulation and settling.
The results of both the routine and modified treatment schgnes showed
that 48 hour source water storage with alum treatment reduced the total
coliform densities fcy approximately 97 percent and turbidities by
approximately 90 percent. The coagulation and settling process,
however, had little effect on further turbidity reductions and further
decrease in the coliform population was only approximately 50 percent
when chlorination was delayed until after this process (Table 1).
Moving the point of chlorination to after coagulation and settling
resulted in an intrusion of coliforms into early stages of water
treatment. This change placed increased importance on providing a high
quality process water at this point, so that final disinfection would
be effective in the inactivation of residual densities of various
organisms of public health concern. Neither a measurable change in the
bacterial quality of the finished water nor any apparent in-plant
problems developed.
GRANULAR ACTIVATED CARBON (GAC) ADSORPTION
Carbon filtration, either in the GAC or BAC (biological activated
carbon) treatment process, may provide opportunities for specialized
microbial populations to became predominant, seme of which may be less
effectively controlled by conventional disinfection practices. This
consideration, coupled with the fact that there are health risk
limitations on chlorine dioxide concentration and slower inactivation
rates for chloranines, make it apparent that disinfection concentration
and contact time values are of critical importance.
110

-------
t
Granular activated carbon (GAC) has been in use for many years to remove
a variety of synthetic organics and naturally occurring taste and oacr
compounds. Optimizing removal of organics in process water to lessen
disinfection product formation would suggest the GAC process be placed
early in the progressive treatment of polluted surface waters. In fact,
powdered carbon is often applied in source water impoundments to control
taste and odor problems but is not adequate for removal of many other
organic compounds (2) . Obviously, the GAC filtration process cannot be
applied directly to raw surface waters with significant turbidities
(above 1 NTU) because silt in these raw waters quickly coats the carbon
particles and rapidly reduces organic adsorption capacity. Thus,
settling of raw water and chereical treatment with clarification,
generally precedes the GAC process. These treatment processes also
remove much of the turbidity-associated, microbial flora which include
a wide range of environmental organisms, some of which are capable of
aggressive colonization of GAC particles.
In the adsorption of organic substances, including those that may be
trihalcmethane precursors, granular activated carbon particles became
focal points for bacterial nutrients and also provide suitable
attachment sites for habitation. Although the portion of organic
renoval in the GAC process possibly attributed to biodegradation is
snail (cccpared with physical adsorption to the activated carbon
surface) there is a substantial microbial population present at the
water-activated carbon surface interface. This process can, therefore,
be of concern in that the treatment barriers must remain effective
against bacterial population densities that can include regrowth of
indicator organisms and selective adaptation by some organisms known to
be disinfectant resistant, opportunistic pathogens or antagonists to
coliform detection.
Case History: Installation of a GAC filter adsorber on-line after
chemical clarification (Figure 1) did result in an approximate 85
percent reduction in turbidity to values ranging from 0.3 to 0.9 NTU
(Table 2) and also reduced the chlorine residual to virtually zero, at
Jefferson Parish, Louisiana (3). The occasional wide differences in
residual turbidities reflect the entrapment of coagulant particles on
the OC bed and their migration through the filter prior to backwashing.
Application of chloramines to the clarified water, before passage
through the GAC filter adsorber, did not result in a complete reduction
of total coliforms in the influent to below the one organism per 100 nL
detection level, except for the autunn 1979 period. Disinfectant
concentration and contact time become more critical when chloramines are
applied, since these agents are slower acting than free chlorine. Total
chlorine residual data included not only the active disinfectant
components but also same complexes that have no disinfection power.
Consequently, a few coliforms were often found in the GAC filter
adsorber effluent except during the winter period of 1979.
Moving the GAC adsorber treatment process to a point following sand
filtration (Table 3) resulted in an inprovement (0.1 to 0.3 NTU) of
effluent turbidity. The most beneficial effect was an inprovement in
the bacteriological quality of the influent. Heterotrophic bacterial
densities were below 75 organisms per mL and no total coliforms were
found in any of the effluent samples. This water quality inprovement
was a result of sand filtration effectiveness and increased contact time
with chloramines added after clarification. As a result of better
quality influent water to the adsorber, no coliforms were detected
in the OC effluent over a three year study period. However, little
111

-------
difference was observed in the cyclic rise and decline of the hetero-
trophic bacterial population in the adsorber effluents associated with
either treatment arranganents.
While granular activated carbon is often used in conventional treatment
beds designed originally for sand filtration, SC may also be used in
pressure contactors. In pressure contactors, GAC bed depth is usually
more than 36 inches (0.9 m) to provide the contact time necessary to
renove aertain classes of organics as the water is punped through each
unit. Factors to be considered were the extent of bacterial
colonization in fast flowing water through a deeper GAC bed and
containment in a closed cylinder.
Two aspects of microbial response in GAC contactors were explored:
salification potential for heterotrophic bacteria, including coliform
persistence, and species profile for various Pseudcmonas and
Flavobacterium strains that could be opportunistic pathogens.
Inspection of data in Figure 2 reveals that diloraminated influent
source water (used as feed water through the GAC contactors in series)
had standard plate counts (SPC) ranging from 4 to 170 organisms per mL.
Lower maxinun densities occurred during late autism and winter cold
water taperature conditions. After the influent water passed through
the first GAC contactor, densities of heterotrophic bacteria increased
by 3 to 4 logs and remained at these higher levels with subsequent
serial passage through following contactors.
Residual coliform populations surviving the inpact of source water
chloramination recovered sufficiently to pass from one contactor to the
next resulting in colonization and occasional release in the effluents
from each contactor. Apparently, assimilable nutrients in the GAC
contactors were the limiting factor that prevented a more aggressive
growth of coliforms and higher densities of heterotrophic bacteria in
a stepwise fashion from one contactor to another. Conversely, once the
limiting density of bacteria was reached in the first contactor, there
was no decline in bacterial numbers that might suggest that lower
nutrient levels were available after passage of process water to the
third contactor in series. Fran these data, it quickly becanes obvious
there are a large number of other organisms present in this process
water ttet we know little about in terms of public health significance.
Does the (AC contactor environment become a habitat for various
Pseudomonas species? Periodic speciation of isolated bacteria frcm GAC
contactor effluent water indicate that a variety of Pseudomonas species
(Table 4) colonized the GAC and persisted for months in the contactor
environment with recurrent or continuous releases to the process water.
Occasional quantitation of these organisms over a three year period
(1985 to 1987) suggested that all Pseudomonas species represented only
one to two percent of the entire population of heterotrophic bacteria
detected in effluent waters from GAC contactors. The extent of regrowth
and density artplif ication of Pseudomonas species were not measurable
because of nvmerous indeterminate high counts in the contactor
effluents. However, these events appeared to increase as the water
passed through contactors in series, suggesting that expanding
colonization occurred as a consequence of increased inputs of
accvsnulated organisms from the preceding contactors.
One of the areas of greatest confusion in studying the microbial ecology
of GAC adsorbers has been the selection of a cultural protocol (medium,
incubation time and tenperature) to optimize recovery of these
112

-------
organisms. In nany of the pilot- and full-scale studies reported here
and elsewhere in the literature, the standard plate count procedure
(plate count agar (PCA), 35^ incubation for 48 hours) has been used
until recent years. A conparative analysis of the same process waters
by two different culture media and extended incubation time illustrates
the probler. (Table 5). The traditional standard plate count procedure
does not adequately detect either the magnitude of bacterial growth in
adsorber beds or in other process waters. These organisms need a median
with a diversity of nutrients in lew concentration, such as found in the
R2A agar formulation (4). Increasing the length of incubation time at
a lower tenperature (28^C) further enhances the recovery of a wide
spectrum of organisms that may be present in GAG adsorber effluents,
other stages of water treatment, finished water, and water in
distribution.
ENHANCED BIOLOGICAL DEGRADATION
Many of the industrial chemical conpounds found in polluted source
waters and naturally occurring organics released to surface waters by
decaying vegetation and algal blooms are non-biodegradable and are
poorly adsorbed in (AC filtration. Since these organics may also
react with disinfectants to foes undesirable by-products, attention has
been directed toward the conversion of these complex, refractory
impounds into more readily biodegradable substances that can be
adsorbed by GAC (5-8) or consisted by microorganisms established in the
filter. This combined process is sometimes described as biological
activated carbon (BAG) treatment (9-10), which frequently involves the
use of ozone to enhance the BAC process.
Case History: A 10 gpn pilot plant (Figure 3) was constructed at the
Shreveport Treatment Facility to evaluate THM precursor removal through
a conventional treatment train using a "Waterboy" package plant (without
disinfection) plus GAC adsorption, or with ozonation prior to (AC
adsorption in a biological activated carbon (BAC) mode (27). The
purpose of investigating BAC in this pilot study was to enhance the
biodegradation of high levels of TW1 precursors in Cross Late water, the
principal water supply for Shreveport, LA. The microbiological concern
was the possible loss of effective barriers to coliform penetration
further into the system, so that final disinfection was the only barrier
to colifono migration and elevated heterotrophic bacterial densities
reaching the distribution system.
While preozooation of the process water was used primarily to convert
recalcitrant organics to shorter chain carbon ccnpounds, the process,
as applied, did teve sane inpact on bacterial population densities and
profile of organisms entering the,GAC filter bed (Table 6). During the
cold water periods of autism and winter, there was a 2 to 3 log
reduction in the source water bacterial densities applied to BAC as
ccopared to non-ozonated raw water. Smaller reductions were noted
during warn water months and on one occasion (July 16, 1981) the
ozonated influent contained 4 times the density of heterotrophic
bacteria found in non-ozonated influent water. Ozonation exposure may
have caused the breakup of bacterial aggregates or algal masses in the
source water with a release of individual bacteria.
The biological activated carbon mode intentionally encourages greater
microbial activity in the BAC bed for the purpose of assimilating much
of the recalcitrant organic conversions by ozonation. While ozonation
exposure initially suppressed heterotrophic bacterial densities in the
113

-------
effluent released from the detention basin, there was an expected 10-
fold increase in these organisms in the effluent from the second BAC
contactor because of the increase in biodegradable organics created by
ozonation. Coliform growth was not detected in the BAC contactors
probably because the general population of other heterotrophic organisms
rapidly became dominant in this environnent and suppressed coliform
development and detection. In both GAC and BAC treatment nodes, the
pilot study revealed no treatment barrier protection was provided- at
these latter stages of water processing. In such situations, final
disinfection nust be 100% effective at all tiroes to achieve the
necessary 6 log reduction of bacteria for a safe water supply.
Examination of process water examinations for coliform bacteria was also
done at Shreveport, Louisiana (12) and provided evidence that total
coliforms may persist in both BAC and GAC columns (Table 6). The raw
source water contained 10^ to 104 coliforms per 100 mL which were not
completely inactivated in the pretrestment (package-plant processed) of
influent waters going to the pilot plant, nor by preozonation in BAC
treatment. Coliforms were also occasionally isolated from BAC treated
in a similar pilot plant study conducted in Philadelphia, Pennsylvania
(12). In the Philadelphia study, the river source water contained
48,000 total coliform organisms per 100 nL and pretreatment was not very
effective in the inactivation of coliforms. It is important to note
that the ozone concentration used was selected to obtain maximum removal
of dissolved organic carbon and was not necessarily optiimm for
disinfection of the raw source water.
Identifying the coliform strains isolated from the GAC filter effluents
revealed that Klebsiella, Enterobacter and Citrobacter were the genera
involved. These organisos are the sane coliforms that have been
reported to predomirate in biofilm growth within some water distribution
systems (13). How nuch of a case can be made for limiting coliform
occurrences in distribution systans to treatment barrier penetration by
coliforms from a process water is unknown, but should not be overlooked.
The observation that profiles of bacterial groups and species in BAC
effluents shew a rsnarkable similarity to those present in (AC effluents
is not surprising. The reason for the similarity in bacterial profiles
is a reflection of the way BAC technology is studied in this country.
Rather than extending service life of the filter to encourage
development of specialized bacterial populations that are more efficient
in assimilation of dissolved organics, greater reliance was placed on
the carbon adsorption aspect with more frequent reactivation of the
carbon media. The Shreveport pilot study (14) denonstrated that after
52 weeks of BAC operation, only microbial metabolism was responsible for
renoval of organics and the rate of trihalomethane formation potential
was not sufficiently reduced to meet a 0.1 mg/L maximum contaninant
level (MX) for trihalomethanes.
Anplification and acclimatization of a diverse and specialized microbial
population through extended service life of a BAC filter presents
another concern. Under these conditions, a biofilm of specialized
organisms develops through successional changes in dominant species and
is similar to biologically active floe development in sewage treatment
processes. ttiile final disinfection can be effective in inactivating
many of these diverse organisms, others will be resistant to applied
disinfection and will pass into the distribution system. Neither the
health effect significance of this diversified population of organisms
entering the potable water supply nor the contribution these organisms
114

-------
r
sake to the development of biofilm in the distribution pipe network and
associated reservoirs has been clearly demonstrated.
APPLICATION Of PISIKFSCTAhT ALTERNATIVES
Another abroach to minimize trihalomethane production in water
treatment is the use of a disinfectant alternative to free chlorine.
Preformed chloramines (chloramination), chlorine dioxide and ozone have
been proposed as practical disinfectant alternatives. In addition,
potassium permanganate has been suggested as a pre-oxidant for some raw
source raters. Because of the desire to maintain a disinfectant
residual in distribution water, chloramines and chlorine dioxide have
received considerable attention. Sane surface water systems are
seriously considering ozonation because of its more favorable
disinfection C-T values, particularly with regard to its effectiveness
to oxidize recalcitrant organics and inactivate Giardia cysts. While
ozone is a powerful inactivating agent for waterbome pathogens, it does
not have a lasting residual to provide protection in distribution water
and is known to create acre assimilable organics that stimulate the
growth of heterotrophic bacteria. Each alternative disinfectant
candidate has specific advantages over free chlorine application but
also some significant disadvantages that oust be understood in the
trade-off.
CHLORINE DIOXIDE
Case History Mo. 1: Hie Western Pennsylvania Water Cenpany, Hays Mine
plant presented an opportunity to study the alternative use of chlorine
dioxide as the primary disinfectant (15). For this investigation, the
routine practice (Table 7) was source water chlorination, potassium
permanganate treatment, coagulation, settling, activated carbon
filtration/adsorption and free chlorine application in the clear-well.
Later, the treatment train was modified (Table 7) to inject chlorine
dioxide and potassium permanganate into the source water entering the
coagulation basin, with free chlorine used as a secondary disinfectant
in the clearwell prior to distribution. Chlorine dioxide dosage to the
source water wis 1.5 mg/L and contained less than 0.1 mg/t chlorine.
Bacteriological data indicated that in the source water, 1.5 mg/L of
chlorine dioxide was not as effective a disinfectant as 2.6 mg/L
chlorine. During source water chlorination, mean total coliform and
standard plate count densities in the activated carbon/filter adsorber
influent were one per 100 mL and 50 per raL, respectively. When chlorine
dioxide was the applied disinfectant prior to coagulation and settling,
a disinfectant residual could not be maintained. As a result, mean
bacterial densities reaching the activated carbon filter/adsorber were
43 total coliforms per 100 mL and 7,100 standard plate count organisms
per mL- In-plant survivors of the total coliform population passed
through the two and one-half year old granular activated carbon
filter/adsorber essentially unchanged in density. In both treatment
trains, the secondary application of chlorine in the clearwell was,
however, an effective barrier to total coliform penetration into the
distribution system.
Front these data, 1.5 mg/L of chlorine dioxide was not equal to the
disinfection effectiveness of free chlorine during source water
disinfection. Increasing the dose of chlorine dioxide was not
economically feasible aid might exceed the limit of 0.5 mg/L residual
chlorine dioxide, chlorite, and chlorate recattnended by the U.S.
115

-------
Environmental Protection Agency (14). New information on disinfection
by-product formation may further reduce this total oxidant limit to 0.3
mg/L in the future.
A further modification of treatment that utilized source water
disinfection with a low concentration of both disinfectants was
effective in reducing the bacterial densities in the GAC filter/adsorber
influent at the Hays Mine plant, although seme regrowth of total
coliforms and the heterotrophic bacterial population did occur in the
filter/adsorber and appeared in the effluent. With the application of
chlorine at the clearwell, however, the finished water net the
bacteriological standard for total coliforms and a low mean standard
plate count of eight organisms per mL was present.
Case History No. 2: In a similar experience, the Evansville (Indiana)
water utility has been successful in substituting chlorine dioxide for
chlorine as a predisinfectant to their raw source water. With an
average chlorine dioxide dosage of 1.2 mg/L applied in pretreatment,
total oxidants of 2.1 mg/L applied in the clearwell provided an
acceptable bacteriological quality in the distribution systen (16).
POTASSIUM PESMANGANATE
Potassium permanganate is most often used in the water supply utility
for taste and odor control or for renoval of iron and manganese (17,
18). Since it is an oxidant other applications suggested have been the
disinfection of process basins (concrete, cement mortar lining, asbestos
cement surfaces) and water lines after repairs (19, 20). Because
potassium permanganate has a limited disinfection efficacy, (21)
application in the disinfection of water lines is not as effective as
use of chlorinated water (22). Nevertheless, there may be seme
measurable benefit achieved in using potassium permanganate as a pre-
oxidant in early stages of the treatment train. In this situation the
pre-oxidant may reduce growth of algae and slime bacteria in the
treatment basins plus provide sane abatenent in the bacterial
population.
Case History No. 1: St. Joseph, Missouri Water Ccrspany uses clari-
fication, sedimentation, filtration and chlorine disinfection in the
processing of Missouri River water (23). In August 1982, 1.1 mg/L
potassium permanganate was applied for four weeks (Table 8) at the
discharge of the clarifiers (Fig. 4), prior to entering settling basin
No. 1. On the fifth week no potassium permanganate was applied so as
to provide percent reduction data for settling alone. Data in this
preliminary study suggest that settling produced 59% of the bacterial
reduction. Application of potass ion permanganate apparently accounted
for an additional 40.9%. While potassium permanganate would not be
satisfactory for application in final disinfection, use as a pre-
oxidant early in the treatment train provides some early in-plant
bacterial reductions in addition to controlling interference front algal
blooms and bacterial slimes. The inpact that this pre-disinfectant has
on shaping the resultant microbial flora entering the distribution
system is unknown.
Case History No. 2; The Davenport, Iowa Water Conpany processes raw
water from the Mississippi River (Figure 5) using clarification,
sedimentation, filtration through GAC and disinfection. In 1983, 0.61
mg/L potassium permanganate was added to control odor in the flocculator
basins and keep the sedimentation basin sludge from turning septic (23).
116

-------
Data collected over ten months (Table 9) indicate that the combination
of pre-oxidant application and settling for 35.9 hours could provide a
significant reduction in both total conforms and the standard plate
count. How much of this reduction was due to settling vs pre-oxidant
contact time cf 35.9 hours was not determined. The treatment approach
did eliminate odor in the flocculator buildings which appeared after the
discontinuance of prechlorination. The changes this pre-oxidant might
have had on selective survival of bacteria and conversion of various
chemical complexes to assimilable organic compounds released to the
distribution system is not known.
OZONATION
Ozone has frequently been used in water supply treatment to remove
taste, odor and color because many of the compounds responsible for
these characteristics are unsaturated organics (24). Other uses include
the removal of iron and manganese, or as a coagulant aid to reduce
coagulant requirements and increase filtration rates (25, 26).
Since ozone does not react with organic residuals found in source waters
for water supply to produce trihalomethanes, there is a growing interest
in the use of ozone as a disinfectant. For instance, ozone is also far
more effective in the inactivation of Giardia cysts than is chlorine.
Unfortunately ozone residuals are quickly dissipated with a lifetime of
less than an hour in most drinking water systsns (26). As a result,
secondary application of chlorine is necessary to provide disinfectant
residual protection in the distribution system.
Treatment train application of ozone generally includes ®C filter-
sorbers. Ozone exposure maximizes the breakdown of ccnplex organics to
shorter chain ccnpounds which are then either absorbed in the GfcC filter
bed or degraded by the bacterial flora in a biologically activated
carbon filter. The net result will be less THM precursors to react with
chlorine in final disinfection. However, there is a trade-off to
consider: increased bacterial densities released from the GfiC contactor
and increased levels of assimilable organic carbon in the finished water
support seasonal regrowth of the heterotrophic bacteria in the pipe
network.
IMPACT OH DISTRIBUTION SYSTEM VOTER QUALITY
Major changes in source water quality, treatment modifications and
operational practices are reflected in distribution water quality. The
beneficial aspects of some treatment modifications may, in the long
term, lead to reduced assimilable organic carbon and biofilm development
in the pipe environment. However, adverse effects as a result of
reduction in treatment barrier redundancies may eventually lead to
biofilm colonization of pipe sections, taste and odor complaints and
increased coliform occurrences. Therefore, it is essential to carefully
monitor the microbial quality of water in distribution, particularly at
the end of the system and in areas of slow flow where disinfectant
residuals are marginal or non-existent. Several examples will help
illustrate this concern.
Case History; The Cincinnati Water Works stopped chlorination of the
Ohio River source water and began chlorination at the influent to the
treatment plant (Table 1) on July 14, 1975, as an initial step in
changing the in-plant water treatment process to control trihalo-
methane concentrations. Chlorination at the clearwell was used to
117

-------
inactivate any residual coliform population that might have penetrated
other processes in the treatment chain. With careful control of
chlorine dose, point of application, and water pH, a significant
decrease in trihalanethane concentration was realized. The inpact that
this treatment modification might have on the bacteriological quality
of drinking water at the distribution system dead-ends and other slow-
flow sections in the distribution network was determined from an
intensive 2-year study.
With the cooperation of the Cincinnati Water Works Water Distribution
Maintenance Section, samples from 32 dead-end water mains were examined
on a rotating basis of eight sites per week. These sites are among a
number of troublesome dead-end water mains that are flushed out each
week to clear accural ated sediments and bring fresher water with free
chlorine residuals into these distribution lines. Sarples from these
flushes were iced immediately and processed within 5 hours of
collection. Analyses of 613 water samples over the 2 year period
included a 10 tube, three dilution total coliform most probable number
(MPN) test and a standard plate count incubated at 35% (95°F) for 48
hours. Physical/chenical paraneters measured were free chlorine
residual, turbidity, water tauperature, and pH.
Changes in distribution system water quality were not observed
immediately on the day of the treatment change. Approximately 15 days
passed before sane decrease in free chlorine residual concentrations,
turbidity, and pH occurred. Before the change in the point of
disinfection application, increased chlorine residuals were inconsistent
in limiting sane coliform occurrences, probably because of sediment
accumulations that resulted in an average turbidity of 20.7 NTC in these
dead-end sections. The most extreme example occurred during one week
in Decenber 1974, when the total coliform density averaged 138 organisms
per 100 mL in the eight sanples collected frcm selected dead-end
flushings. Once the turbidity decreased to an average of 10.1 NTU,- this
interference with disinfection was not apparent. Why the turbidity in
the dead-ends was reduced following the treatment change is not known;
the protocol and frequency of main flushing renained unchanged. Perhaps
this reduction in turbidity was a result of more water flow with
increased tap-ins from residential developments or it may have been a
result of more stable scale formation on the pipe walls (pH shifted from
8.0 to 7.8) following treatment modifications.
After the point of chlorination was moved, a free chlorine residual
concentration of at least 0.2 mg/L was effective in controlling colifonn
occurrences in the dead-end sections of the distribution network. When
free chlorine residual concentrations declined to 0.1 mg/L or less
during warm water periods, hcwever, viable coliforms in these protected
pipe habitats were detected in densities as great as 30 organisms per
100 mL. Water tsiperatures during these periods of -low free chlorine
residual concentrations fluctuated frcm 20 to 25°t (68 to 77°F). Sudden
increases in standard plate count densities often occurred a few days
to a week in advance of the appearance of colifozros in these waters.
Thus, increased standard plate counts could serve as an early signal of
a loss of disinfection effectiveness or other undesirable quality
changes occurring in water distribution systems.
The effects that GAC or BAC treatment have on distribution water quality
are largely undocumented. Several colifonn species. (Klebsiella,
Enterobacter, and Citrobacter) have been found to colonize GAC filters,
regrcw during warm water periods, and discharge into the process
118

-------
effluent. Carbon particles have also been detected in finished water
from several water plants using powdered carbon or GAC treatment. Over
17 percent of finished water sanples examined from nine water treatment
facilities contained carbon particles colonized with coliform bacteria
(28). These findings confirm that carbon fines provide a mechanisrr. by
which microorganisms penetrate treatment barriers and reach the
distribution system. Other mechanisms that could be involved in
protected transport of bacteria include aggregates or clunps of
organisms from colonization sites in GAC or sand filtration and by the
protected nature of particulates in water.
Case History; Another important finding was that full-scale GAC
treatment (Manchester, New Hampshire) resulted in a statistically
significant increase in heterotrophic bacterial densities in
distribution water as compared to a similar water treatment operation
(Concord, NH) that does not employ ac (29). Furthermore, water
temperature, pH, and turbidity had a positive influence on heterotrophic
bacterial densities (30). These physical-cheaical conditions of water
are key factors that also impact disinfection effectiveness. Stability
of disinfectant residuals during water distribution is iiqportant for a
number of purposes; particularly to prevent colonization of surviving
organisms and to disinfect contaminants that intrude into the pipe
network. Microbial colonization may lead to corrosive effects in the
distribution system and aesthetic changes in taste, odor, and
appearance. Regrowth of potential health-related opportunistic
organisms and their inpact on coliform detection should not be dismissed
as a trivial problem. Further, the maintenance of a disinfectant
residual to the consumer's tap keeps the system clean and protects
against some cross-connection contamination. The sudden disappearance
of disinfectant residuals is a sensitive indication of distribution
system problems. Although the maintenance of a disinfectant residual
in the distribution system will not combat massive levels of external
gross contamination that are detectable through odors, color and milky
turbidity changes, the residual may quickly inactivate pathogens in
situations that are involved with contaminants seeping into large
volumes of high quality potable water (31).
Distribution systsn problems associated with the use of combined
chlorine residual or no residual have been documented in several
instances (32-34). In these cases, the use of combined chlorine is
characterized by an initial satisfactory phase in which chloramine
residuals are easily maintained throughout the system and bacterial
counts are very low. Over a period of years, however, problems may
develop including increased densities of heterotrophic bacteria, loss
of chloramine residuals in the pipe network extrenities, increased taste
and odor complaints necessitating more frequent flushing of the system.
Where treatment modifications are highly effective in reducing the
concentrations of dissolved organic ccopounds in water, there will be
less trihalomethane production and also less bacterial regrewth because
of the reduction of assimilable organic carbon in the distribution
system. The reduced potential for bacterial regrewth may, however be
slow to appear in many systaos because of the untold years of organic
accunulations in pipe sediments and tubercle material.
SlM9,K£
Microbial quality in the distribution system is a reflection of raw
source water characteristics, treatment process configurations and their
i
i
119

-------
modifications and the physical conditions within the distribution system
itself. Based on case history experiences there may at tines be a
microbial breakthrough that is caused by fluctuations in raw surface
water turbidity, chlorine dsaand and water pH. These situations call
for appropriate changes in operational practices to compensate for water
quality degradations.
In the effort to reduce THM production, operational practices should
not abandon the concept of multiple barriers nor the necessity to
produce a high quality process water that can be effectively
disinfected. A growing data base from many systems suggests there may
be some microbial migration deeper into the treatment train while
achieving better organic contaminant reductions. This situation makes
disinfectant concentration and contact time values of critical
importance.
Changes in water supply treatment practices to reduce the formation of
disinfectant by-products oust be carefully monitored for microbial
breakthrough. Increases in microbial populations in finished water may
"lead to biofilm development in distribution pipe networks and the
potential for more frequent colifonn occurrences.
None of these issues are beyond control using reasonable treatment
precautions by water plant operations. Due to the canplex interaction
of many significant variables within the treatment train, it can be
reasoned that there will always be at least sane biological activity in
the final effluent from any treatment system, assuming traditional
disinfection procedures and doses. Vfoat is required is a revised
monitoring program for water treatment processes that will provide more
useful microbiological information by which to fine-tune treatment
effectiveness and provide better quality waters entering the
distribution system.
REFERENCES
1.	Ohio River Valley Sanitation Comaission. Water Treatment Process
Modifications for Trihalomethane Control and Organic Substances in
the Ohio River. U.S. Bivirotmental Protection Agency, EPA 600/2-
80-028, Cincinnati, Ohio (March 1980).
2.	Epton, M. A. and J. F. Becnel. Evaluation of Powdered Activated
Carbon for Removal of Trace Organics at New Orleans, Louisiana.
U.S. Envirocmental Protection Agency, EPA-600/52-81-027, Cincinnati,
Ohio (1981).
3.	Lykins, Jr., B. et al. Field Scale Granular Activated Carbon
Studies for Rsnoval of Organic Contaminants Other Than Trihalo-
methanes, fran Drinking Water. U.S. Environmental Protection
Agency, EPA-600/2-84-165, Cincinnati, Ohio (1984).
4.	Reasoner, D. J. and E. E. Geldreich. A New Medium for the
Enumeration and Subculturing of Bacteria from Potable Water. Appl.
Environ. Microbiol. 49:1-7 (1985).
5.	Benedek, A. et al. The Effect of 03 on Activated Carbon Adsorption:
A Mechanistic Analysis of Water Treatment Data. Ozonews 6(1): 1-3
(1979).
120

-------
6.	Benedek, A. et al. The Effect of Ozone on the Biological Degrada-
tion and Activated Carbon Adsorption of Natural and Synthetic
Organics in Water, Part II: Adsorption. Ozone, Sci, and Engr.
1:347-356 (1979).
7.	Rice, R. G., G. W. Miller, C. M. Robson and W. Kuhn. A Review of
the Status of Pre-Ozonation of Granular Activated Carbon for Removal
of Dissolved Organics and Amnonia From Water and Wastewater. In:
Carbon Adsorption Handbook, P.N. Cheranisinoff and F. Ellerbusch
(Eds.). Ann Arbor Science, Ann Arbor, Mich, pp 485-537 (1978)
8.	Stephenson, P. et al. The Effect of Ozone on the Biological
Degradation and Activated Carbon Adsorption of Natural and Synthetic
Organics in Water, Part 1: Ozonation and Biodegradation Ozone. Sci.
and Engr. 1:263-279 (1979).
9.	Miller, G. W., R. G. Rice and C. M. Robson. Large Scale Applica-
tions of Granular Activated Carbon with Ozone Pretreatinent. In:
Activated Carbon Adsorption of Organics from the Aqueous Phase, Vol.
2, M. M. McGuire and I. H. Suffet (Eds.), Ann Arbor Science, Ann
Arbor, Michigan, pp 323-347 (1980).
10.	Rice, R. G. et al. Biological Processes in the Treatment of
Municipal Water Supplies. U. S. Environmental Protection Agency,
EPA-600/2-82-020, Cincinnati, Ohio (1982).
11.	Neukrug, H. M. et al. Removing Organics From Philadelphia Drinking
Water by Combined Ozonation and Adsorption, U.S. Environmental
Protection Agency Cooperative Agreenent No. CR-806256, National
Technical Information Service, PB83-223370, Springfield, Virginia
(1983).
12.	Glaze, W. H. et al. Evaluation of Biological Activated Carbon for
Ranoval of Trihalomethane Precursors. U.S. Bivironnental Protec-
tion Agency, Contract No. CR-806157, National Technical Information
Servioe, PB82-230301, Springfield, Virginia. (1982).
13.	Geldreich, E. E. Microbiological Quality Control in Distribution
Systens. F. W. Pontius and J. M. Synons (eds). Water Quality and
Treatment. McGraw Hill Publishers, New York, NY. (1990).
14.	Carswell, J. K., R. G. Eilers, D. J. Reasoner. Pilot Scale
Extramural Research on Biological Activated Carbon: A Summary
Report to the Office of Drinking Water. Drinking Water Research
Division, U.S. Environmental Protection Agency, Cincinnati, Ohio
(Sept. 1984).
15.	Symons, J. M., et al. Treatment Techniques for Controlling
Trihalomethanes in Drinking Water. U.S. Environmental Protection
Agency, EPA-600/2-81-156, Cincinnati, Ohio (1981).
16.	Lykins, Jr., B. W. and N. G. Griese. Using Chlorine Dioxide for
Trihalomethane Control. Jour. Amer. Water Works Assoc. 78:88-93
(1986).
17.	Cherry, A.K. Use of Potassium Permanganate in Water Treatment.
Jour.- Amer. Water Works Assoc. 54:417-424 (1962).
121

-------
18.	Shull, K.E. Operating Experiences at Philadelphia Suburban Treat-
ment Plants, Jour. Amer. Water Works Assoc. 54:1232-1240 (1962).
19.	Sonneborn, M. and Bohn, B. Formation and Occurrence of Haloforms
in Drinking Water in the Federal Republic of Germany. In: Water
Chlorination - Environmental Iirpact and Health, pp. 537-542. Eds:
R. Jolley, H. Gorchev, and D. Hamilton, Jr. Ann Arbor Science, Am
Arbor, MI., March 1978.
20.	Hamilton, J.J. Potassium Permanganate As A Main Disinfectant,
Jour. Amer, Water Works Assoc. 66:734-735 (1974).
21.	Cleasby, J. L., Bauroan, E.R., Blade, C.D. Effectiveness of
Potassium Permanganate for Disinfection. Jour. Amer. Water Works
Assoc. 56:466-474 (1964).
22.	Buelov, R.W., Taylor, R.H., Geldreich, E.E., et al. Disinfection
of New W&ter Mains. Jour. Amer. Water Works Assoc. 68:283-288
(1976).
23.	Blanck, C.A. Total Coliform Reduction During Treatment of
Mississippi River Water Using Potassitm Permanganate, pp 309-318.
In: Proceeding - Water Quality Technology Conference, Norfolk, VA,
Dec. 4-7, 1983, American Water Works Association, Denver, CO.
24.	Anselme, C., Suffet, I.H. and Mallevialle, J. Effects of Ozonation
on Tastes & Odors. Jour. Amer. Water Works Assoc. 80:45-51 (1988).
25.	Prendivilla, P.W. Ozonation at the 900 cfs Los Angeles Water
Purification Plant. Ozone Sci. and Engineer. 8:77-93 (1986).
26.	Glaze, W.H. Drinking Water Treatment with Ozone. Environ. Sci.
and Technol. 21:224-230(1987).
27.	Glaze, W.H., W&llace, J.L., Dickson, K.L., et al. Pilot Scale
Evaluation of Biological Activated Carbon for the Removal of THM
Precursors. Project Siaanary EPA-600/S2-82-046, Municipal
Environmental Research Laboratory, U.S. E.P.A. (1982).
28.	Can*>er, K., et al. Bacteria Associated with Granular Activated
Carbon Particles in Drinking Water. Appl. Environ. Microbiol.
52:434-438 (1986).
29.	Haas, C. N., M. A. Meyer and M. S. Paller. Microbial Dynamics in
ffiC Filtration of Potable Water. Proc. Amer. Soc. Civil Eng.,
Jour. Environ. Eng. Div. 109:956-961 (1983).
30.	Haas, C. N., M. A. Meyer and M. S. Paller. Microbial Alterations
in Water Distribution Systems in their Relationships to Physical-
Chemical Characteristics. Jour. Amer. Water Works Assoc. 75:475-
481 (1983).
31.	Snead, M. C., V. 0. Olivieri, K. Kawata, and C. W. Kruse. The
Effectiveness of Chlorine Residuals in Inactivation of Bacteria and
Viruses Introduced by Post-Treatment Contamination. Water
Research, 14:403-408 (1980).
122

-------
32.	Vendryes, J. H. Experiences with the Use of Free Residual
Chlorination in the Water Supply of the City of Kingston, Jamaica.
In: Proceedings AIDIS Congress of Washington, D.C. 1962.
33.	Buelow, R. W. and G. Walton. Bacteriological Quality vs Residual
Chlorine. Jour. Amer. Water Works Assoc. 63:28-35 (1971).
34.	3rodeur, T. P., J. E. Singley and J. C. Thurrott. Effects of a
Change to Free Chlorine Residual at Oaytona 3each. In: Proceeding
- Fourth Water Quality Technology Conference, San Diego,
California, Dec. 6-7, 1976, paper 34-5. Amer. Waterworks Assoc.,
Denver, Colorado.
123

-------
Table I. CHLORINE APPLICATION POINT STUDY
CINCINNATI, OHIO WATER WORKS
BEFORE TREATMENT MODIFICATION
SAMPLE POINT (MEAN VALUES)
AFTER TREATMENT MODIFICATION
SAMPLE POINT (MEAN VALUES)
PARAMETER
SOURCE
STOREO
SOURCE
GOAQULATED
* 8ETTLE0
FILTEREO
FINISHED
SOURCE
STORED
SOURCE
COAOULATEO
t SETTLED
FILTERED
FINISHED
FLOW TIME MRS.
0
48
62
52.5
55.5
0
48
52
52.5
55.5
TURBIDITY, NTU
32
1.0
1.2
0.1
0.1
14
0.80
1.1
0.07
0.06
TOTAL COLIFORM
PER 100 mL
9600
200
« 1
< 1
< 1
84000
2400
1400
< 1
< 1
8PC PER mL
NR
NR
600
< 1
5
NR
NR
5500
15
< 1
pH
7.3
7.1
8.5
8.3
8.7
7.6
7.2
8.1
8.1
8.2
FREE Clt
RESIDUAL. mg/L
NR
NR
1.8
1.8
1.5
NR
NR
0
1.8
1.4
TOTAL Cl|
RESIDUAL, mg/L
NR
NR
2.0
1.8
1.6
NR
NR
0
2.0
1.5
CHLORINE (3.6 mg/L) (BEFORE MODIFICATION)
STORED
SOURCE
FILTER
FINISHED
COAGULATION
& SETTLING
(AFTER MODIFICATION! (CHLORINE 3.6 mg/L)
OHIO RIVER

-------
Table 2. Microbial Quality of GAC Filter Adsorber Influent and Effluent
Receiving Clarified and Chlorinated Process Water*
Water
Water Treatment Process
Year/
/Season**
Terp.
°C
J*L
Turbidity
(7TU
SPC
per mL
Tota 1
Coli form
per 100 inL
Total Chlorine
Residual
	mg/L
Clarified & Chlorinated
Processed Water
(Influent to GAC filter
adsorber)
1978
Winter
Spring
Sumner
Autumn
1979
Winter
Spring
Sumner
Autunm
4.8
16.1
30.0
18.3
7.1
20.9
28.7
24.0
8.7
8.5
8.4
7.7
7.6
7.3
7.5
7.5
2.5
2.0
2.8
3.9
2.9
2.7
2.8
2.7
1,300
530
440
120
210
140
60.0
9.6
15.3
8.3
22.1
< 0.1
1.64
1.62
1.57
1.57
1.90
1.45
1.29
2.13
Filter Adsorber Effluent
1978
Winter
Spring
Sunnier
Autism
1979
Winter
Spring
Sunnier
Autunvi
8.7
8.4
8.0
7.5
7.3
7.1
7.2
7.1
0.5
0.3
0.3
0.8
0.9
0.4
0.3
0.4
24
62
2,800
750
425
2,850
98
110
0.4
2.7
5.8
1.5
< 0.1
6.0
12.7
11.8
0.83
0.63
0.00
0.00
0.00
0.00
0.00
0.00
*Data from full scale operation, Jefferson Parish, (Jl.
**Seasonal geometric means based on 46 sanples per season.

-------
Table 3. Microbial Quality of GAC Filter Adsorber Influent and Effluent


Water



Total
Total Chlorine

Year/
Tcnp.

Turbidity
SPC
Coliform
Residual
Water Treatment Process
/Season**
°C
PH
rrru
per mL
per 100 rt.
iwj/L
Sand Filter Process Water
1977






(Influent to Adsorber)
Winter
19.9
10.0
0.4
10
< 0.1
1.60

Spring
23.5
9.9
0.2
1.7
< 0.1
1.63

Simmer
30.S
9.9
0.2
10
< 0.1
1.57

Autum
13.0
8.7
0.9
34
< 0.1
1.79

1978







Winter
4.8
8.6
0.4
10
< 0.1
1.67

Spring
16.1
8.3
0.3
10
< 0.1
1.62

Sumwr
30.0
8.4
0.4
17
< 0.1
1.53

Autum
18.3
7.7
0.9
33
< 0.1
1.83

1979







Winter
7.1
7.5
1.1
73
< 0.1
1.90

Spring
20.9
7.3
0.5
27
< 0.1
1.40

Siamer
28.7
7.5
0.3
26
< 0.1
1.27

Autum
24.0
7.4
0.5
23
< 0.1
2.17
Adsorber Effluent








1977







Winter
—
9.9
0.3
16
< 0.1
0.00

Spring
—
9.6
0.2
96
< 0.1
0.00

Sumner
—
9.6
0.2
685
< 0.1
0.00

Autum
—
8.5
0.8
138
< 0.1
0.01

1978







Winter
—
8.5
0.4
31
< 0.1
0.20

Spring
—
8.1
0.3
40
< 0.1
0.00

Sumner
—
7.9
0.2
647
< 0.1
0.00

Autum
—
7.4
0.7
305
< 0.1
0.00

1979







Winter
—
7.3
0.8
415
< 0.1
0.00

Spring
—
7.1
0.4
2,500
< 0.1
0.00

Sumner
—
7.2
0.3
130
< 0.1
0.00

Autum
—
7.1
0.4
63
< 0.1
0.00
** Seasonal geometric means based on 46 sanples per season.

-------
Table 4.
Characterizing the Pseudoroonas Population in
GAC Contactor Effluents*
Maximum Density/irL	Yearly Occurrences
Process
Season
SPC
Pseudcntonas
Pseudaraonas Species
Plant Influent
1986



Chlorinated
Winter
52
22
Ps. alcaligenes

Spring
100
—
Ps. pseudoflava

Sumner
44
>100
Ps. pickittii

Autum
53
44
Ps. pseudoalcaligenes

1987




Winter
38
33
Ps. maltophila

Spring
170
57
Ps. paucimobilis

Sumner
110
23


Autixnn
23
—

Contactor #1
Effluent
Contactor #2
Effluent
1986



Winter
2,300
86
Ps.
Spring
80,000
>100
Ps.
Sumner
53,000
>100
PS.
Autumn
55,000
>100

1987



Winter
130,000
>100
Ps.
Spring
85,000
—
Ps.
Sumner
19,000
70
Ps.
Autum
—
—

1986



Winter
62,000
>100
Ps.
Spring
27,000
>100
Ps.
Sumner
77,000
>100
Ps.
Autum
110,000
>100

1987



Winter
43,000
>100
Ps.
Spring
22,000
73
Ps.
Sunner
—
90

Autum
—
—

Ps. maltoshilia
Ps. paucinobilis
Contactor 13
1986


Winter
64,000
>100
Spring
38,000
>100
Sumner
11,000
>100
Autum
95,000
>100
1987


winter
85,000
>100
Spring
55,000
>100
Sumner
61,000
71
Autum
4,400
—
Ps. ma1tophila
Ps. pickittii
Ps. pseudoalcaligenes
Ps. maltophila
~Data fran full scale operation, Jefferson Parish, LA.
127

-------
Table 5. Bacterial Populations In Water Treatment Processes Using Standard Plate Count
Medium or R-2A Medium with Extended Incubation Times*
(OrganIsma/mL)
Lime-softened water		Sand filter effluent		GftC adsorber effluent
Sampling
day
SPC,
2 days
SPC,
6 days
R2A,
6 days
SPC,
2 days
SPC,
6 days
R2A,
6 days
SPC,
2 days
SPC,
6 days
R2A,
6 days
Initial
120
350
510
890
1,200
1,500
< 1
140
220
7
31
202
510
820
22,000
35,000
1
24,000
95,000
14
7
7
130
< 1
1,200
9,400
< 1
600
4,400
21
7
18
150
2,200
2,500
33,000
< 1
5,200
16,000
28
3
39
530
700
7,800
67,000
1
11,000
55,000
35
< 1
490
330
100
6,000
25,000
< 1
12,000
74,000
42
70
120
1,700
1,200
71,000
22,700
N.D.
56,000
52,000
49
9
1,200
23
5,000
41,000
3,000
80
4,200
100
56
< 1
10
< 1
< 1
700
12,000
N.D.
1,900
50,000
63
29
190
170
170
2,000
3,000
N.D.
5,000
48,000
•Data revised from Symons (15), All cultures incubated at 35^.
SPC = Standard Plate Count
N.D. » Not Done

-------
Table 6. Heterotrophic Plate Counts and Total Coliform Densities In
Pilot Water Treatment Facility, Shreveport, LA
Non-Ozonated Water Ozortated Water
Date Tenp Influent	> GftC	> Effluent Influent	> GAC —> Kffluent
(1980}
(C")
HPC
T.C.
HPC
T.C.
HPC
T.C.
HPC
T.C,
July 8
32
3,300
< 1
4,100
4
290
< 1
13,000
1
Aug 12
31
2,500
1
550
8
30
< 1
1,100
< 1
Sept 15
29
400
9
1,600
7
78
1
3,400
2
Oct 6
21
130
4
750
2
9
< 1
550
1
Nov 12
14
1,700
9
300
1
240
5
1,500
< 1
Dec 8
14
1,900
N.D.
7,600
N.D.
2
< 1
1,300
< 1
(1981)









Jan 13
10
82
2
67
2
3
< 1
3,600
< 1
Feb 10
9
110
4
70
5
5
< 1
1,700
15
Mar 24
19
660
1
150
1
55
5
2,900
1
Apr 15
23
170
< 1
240
1
230
7
2,200
N.D.
May 14
24
300
8
130
7
200
1
4,700
28
June 24
34
550
33
140
455
870
135
2,900
80
July 16
31
700
25
160
215
2,900
81
2,400
91
Raw Lake Water Quality: 102 - 10* Total Coliform per 100 mL
HPC ° Heterotrophic Plate Count per mL using soil extract agar
T.C. = Total Coliform Density per 100 mL
N.D. ¦ No Data

-------
>. CHLORINE APPLICATION POINT STUDY
WESTERN PENNSYLVANIA WATER COMPANY
BEFORE TREATMENT MODIFICATION AFTER TREATMENT MODIFICATION
SAMPLE POINT (MEAN VALUES)	SAMPLE POINT (MEAN VALUES)
PARAMETER
SOURCE
PLANT COAOU-
INPLUENT LATION
QAO
SETTLED PILTERED
FINISHGO
SOURCE
PLANT CQAQU* OAC
INFLUENT LATION SETTLED FILTERED FINISHED
PLOW TIME, MRS.
0
0.6
3.76
12.6
13.6
14.8
0
0.6 3.8 12.6
13.6
14.7
TURBIDITY. NTU
61
38
6.7
8.5
0.0
0.2
6.8
6.2 8.3 2.3
0.3
0.2-
TOTAL COLIFORM
PER WO mL
21000
4
1
1
8
« 1
14000
4200 100 43
44
« 1
•PC PER mL
NR
490
200
60
160
3
NR 29000 4790 7100
860
1
P«
7.2
7.1
7.3
7.1
7.2
7.1
7.1
7.1 7.6 7.4
8.9
8.8
FREE CI,
RESIDUAL. mg/L
NR
0.4
« 0.1
« 0.1
« 0.2
0.8
NR
« 0.1 « 0.1 < 0.1
0.1
« 0.4
TOTAL CI, (CIO,)
RESIDUAL mfl/L
NR
0.8
0.4
0.3
0.2
0.8
NR
« 0.1 < 0.1 « 0.1
« 0.1
« 0.1
CHLORINE (2.6 mg/L; 1.1 mg/L) IBEFORE MODIFICATION!
FINISHED
(AFTER MODIFICATION!
GAG
FILTER
PLANT
INFL.
COAGU-
LATION
SETTLING
CHLORINE DIOXIDE (1.5 mg/L)	CHLORINE (1.4 mg/L)
MONONGAHELA RIVER
Table 8,

-------
MONONGAHELA RIVER
Table 8. •
Potassium Permanganate As A Pre-Oxidant
St. Joseph, MO Water Treatment Application*

Source**
Settling
1 Reduction
Source**
Settling
% Reduction


Basin #1


Basin #1

Retention time
(hrs)
8.5
-
—
10.8
•
Turbidity. (NTU)
755
30
96.0
2,520
62
97.5
Total Coliform
52,000
86
99.9
87,000
35,700
59.0
(per 100 mL)






Data from Blanck (23)	KMnO^	KMnO^
1.1 mg/L Allied	Not Applied
~Average of 4 runs (Aug.
~~Missouri River
5-10, 1982) Applied; one run (Aug. 15) as control.

-------
Table 9.
Potass inn Permanganate As A Pre-Oxidant
Davenport, Iowa Water Treatment Application*
Source**
Flocculator #2
Effluent
% Reduction
Retention time (hrs)
Turbidity (NTU)	11.6
Total Coliform	8,660
(per 100 nSL}
Standard Plate Count 7,960
(per 1 mL)
35.9
1.9
17
1,030
93.9
99.8
99.9
Data from Blanck (23)
KMnO|
0.61 mg/L Applied
•Average values for 10 months (19833
••Mississippi River
132

-------
MISSISSIPPI RIVER
LIME, FERROUS SULFATE
CATIONIC
POLYMER
KMnO
FLASH HIXINC CHAMBER
CONTACTOR
PILOT
COLUMNS
SAND FILTER
v_y VsJ	
ADSORBER
PILOT
COLUMN
CIII.OR INR
AMMONIA
ADSORB!R
RESERVOIR
ADSORBER
FILTER
fS FILTER ADSORBER
PILOT COLUMN
DISTRIBUTION
SYSTEM
CLEARWELI
FICURE 1. FULL-SCALE FILTER AND PILOT COLUMN FLOW CHART FOR JEFFF.RSON PARISH. LA.

-------
HETEROTROPHIC
BPCTERIO
5.700-13,000
PER mL
COUrpRHS
2,200-4,tOO
PER 100 ml
TURBIDITY
37-7S NTU
MISSISSIPPI RIVfR
HETEROTROPHIC
B0CTER1B
1.200-27.000 PER 100 mi
coliroRns
< I t PER ml
rtRWius suLimc
CPU IONIC
POLYMER
CONIOCTOk
PILOT
COLUMNS
SAND
riLTER
HETEROTROPHIC
BBCTERIR
110-270 PER ml
coliroRns
1-4.a PER 100 mL
TURBIDITY
2.0*2.9 NTU
CHLORINE,
nrtujNin
HETEROTROPHIC BPCTERIO
4-170 PER mL
COL!FORMS
< 1-1.0 PER 100 mL
TURBIOITY
O.S-O.O NTU
HETER0TR1PHIC
BQCTERIP
710-09,000 PER ml
COL IFORMS
< 1-8 PER 100 mL
HETEROTROPHIC
BRCTERIO
1,000-110,000 PER mL
COL IFORMS
< 1-4 PER 100 mL
HETEROTROPHIC BOCTERIfl
170-190,000 PER mL
CPU FORMS
« 1-7 PER 100 mL
FIGURE 2. MICROBIAL QUALITY OF EFFLUENTS FROM CONTACTORS IN SERIES

-------
MICROFLOC
•WATER BOY"
HETEROTROPHIC
bacte¥«a
1,700 PER mL
COLIFORMS
8.0 PER 100 mL
FILTER
BACKWASH
BASIN
o3
CONTACTOR
O
A
GAC
3 DETENTION
BASIN
(35-40 mln)
A
HETF,ROTHO»»HIC
BACTERiA
269 PER mL
COI.irORMS
10.1 PER 100 mL
BAC
COLIFORMS

2.0 PER 100 mL

co i. iron ms
10 PER 100 mL
SAC
BAC
NOTE; BACTERIAL DENSITIES
ARE AVERAGE VALUES
OVER STUDY PERIOD
X
HETEROTROPHIC
bacteria
1,800 PER mL
COLIFORMS
64.0 PER 100 mL
HETEROTROPHIC
BACTERIA
3.400 PER mL
COLIFORMS
18.3 PER 100 mL
Figure 3. MICROBIAL QUALITY OF EFFLUENTS-GAC VS 03 ~ GAC
SHREVEPORT LA. PILOT TREATMENT FACILITY

-------
Figure 4. ST:
CLARIFIER
RAW
WATER
WEIL
JOSEPH WATER COMPANY
PURIFICATION PLANT
FLOW DIAGRAM
r—CARBON
LIME
ALUM
FLASH MIX
CHLORINE
L- POTASSIUM PERMANGANATE
^DISTRIBUTION
FILTERS
PUMPS
BASIN
BASIN
BASIN
CLEARWELL

-------
Figure 5.
DAVENPORT, IOWA FLOW DIAGRAM - EAST RIVER STATION
POLYMER
LIME
ALUM
FILTER
AIO
POTASSIUM PERMANGANATE
CHLORINE
AMMONIA
FLUORIDE
GAC
FILTERS
CLEARWELL
RAW
WATER
WELL
BASIN
BASIN
FLOCCULATOR
FLOCCULATOR
1 1
DISTRIBUTION
<4
•>i

-------
TECHNICAL REPORT DATA
tPUsse reed !ns:rj:::ons on :ht r.—.rrj
-------