United States
Environmental Protection
Agency
Municipal Environmental Research
Laboratory
Cincinnati OH 45268
Research and Development
EPA-600/S2-83-062 Sept. 1983
Project Summary
Microbiological Alterations in
Distributed Water Treated with
Granular Activated Carbon
Charles N. Haas, Michael A. Meyer,
Marc S. Pa Her, Michael A. Zapkin,
and Donald B. Aulenbach
Granular activated carbon (GAC)
treatment of drinking water was invest-
igated to determine its effects on the
microbiological characteristics of pot-
able water distribution systems. Though
GAC is increasingly used to remove
organic chemicals and reduce trihalo-
methane precursors from drinking
water, concern exists that the practice
may lead to increased microbial con-
centrations in the finished water.
Microbial parameters examined were
total coliforms (TC), endotoxins, stand-
ard plate count (SPC), membrane
standard plate count (m-SPC), and
acid-fast organisms (AF).
An 18-month field investigation was
held at two full-scale water treatment
plants, and a pilot-scale study was
conducted on GAC columns receiving
prechlorinated, coagulated, and filtered
water.
The 18-month field investigation
showed that such treatment does
increase SPC and m-SPC organisms
but has no effect on any of the other
microbial parameters investigated.
Furthermore, conventional postchlo-
rination following GAC treatment
appeared to mitigate any impact on the
distributed water quality. Within the
water distribution system itself, pH,
temperature, and turbidity appeared to
be positively correlated with SPC and
m-SPC densities, with only sporadic in-
verse correlations to chlorine concen-
tration.
The pilot plant GAC studies using
coagulated, postfiltration water were
conducted to determine the effect of
empty bed contact time (EBCT), velocity,
and prechlorination on the microbial
dynamics in GAC columns. Increasing
EBCT by providing greater bed depth
tended to stimulate microbial growth,
but a decrease in velocity tended to
depress microbial growth. Increasing
levels of prechlorination tended to yield
greater microbial concentrations on the
carbon itself and to effect a slight
increase in effluent microbial densities.
This Project Summary was developed
by EPA's Municipal Environmental
Research Laboratory. Cincinnati. OH,
to announce key findings of the research
project that is fulry documented in a
separate report of the same title (see
Project Report ordering information at
back).
Introduction
Recent concerns over the byproducts of
potable water disinfection and the
presence of synthetic organic chemicals
in water intakes have led to proposals for
introducing GAC treatment into many
water treatment plants. But some experi-
mental evidence suggests that introducing
surfaces such as GAC into dilute envi-
ronments may lead to increased microbial
concentrations. This study was conducted
to determine the existence and magnitude
of such microbial increases, and to
investigate how manipulation of GAC
design and operational variables might
affect such increases.
As one component of this study, an 18-
month, biweekly monitoring program
-------
was conducted at two water treatment
plants—Manchester, New Hampshire
(which employs GAC treatment), and
Concord, New Hampshire (a similar plant
that does not employ GAC). Samples of
raw, partially treated, and finished waters
were collected from each system. A
second, parallel effort was conducted
using pilot-scale (2 54-cm inside dia-
meter) GAC columns receiving prechlori-
nated, coagulated, and filtered water at the
Troy, New York, treatment plant. Four
parallel columns were operated: one with
no GAC served as a control, and the other
three columns were each operated with a
different flow rate, bed depth, and
influent concentration of free chlorine.
Total coliforms (TC) and standard plate
count organisms (SPC) were determined
using Standard Methods. The membrane
standard plate count (m-SPC) was deter-
mined using the procedure of Taylor and
Geldreich. Endotoxms were determined
using the spectrophotometric Limulus
amoebocyte lysate (LAL) procedure. Acid-
fast organisms (AF) were determined
using the membrane filter procedure of
Engelbrecht and Haas.
Results and Conclusions
1 Full-scale GAC treatment as prac-
ticed at Manchester, New Hamp-
shire, resulted in a statistically
significant increase in SPC and m-
SPC organisms averaging 0.65 and
0.74 logs, respectively. For AF, TC,
and total endotoxins, no statistically
significant effects were noted
(Table 1). But conventional post-
chlorination as practiced at Man-
chester appears to mitigate the
impact of these effects on the
distribution system.
2. At Manchester, growth of m-SPC
organisms in the GAC column
appears to be positively correlated
with both influent free chlorine and
combined chlorine levels (Table 2).
Similarly, the pilot plant experiments
showed a positive relationship
between influent free chlorine and
carbon-associated microbial levels.
3. In pilot plant experiments (Table 3),
increases in EBCT as a result of
increased carbon depth tended to
increase effluent microbial densities,
whereas a decrease in influent flow
rate, maintaining a constant carbon
depth, tended to decrease effluent
microbial densities High effluent
microbial concentrations appeared
to be associated with low tempera-
ture, high dissolved oxygen, and
high effluent and influent turbidities
Table 1. Pairwise t- Tests Between Influent and Effluent Microbial Parameters in the
Manchester, New Hampshire. GAC Contactors
Variable
Log (SPC)
Log(m-SPC)
Log(AF)
Total
Endotoxins
Mean
(Effluent-Influent)
+0.649
+0743
-0. 106
-0.098
Number of
Observations
27
17
14
20
P*
0.0012
0.0002
07316
0.9995
*TheP value is the probability (student t-statistic) that the given deviations between pairedsamples
could have been equalled or exceeded when no underlying differences between samples existed.
Table 2. Statistically Significant Regressions for Microbial Densities in the Manchester, New
Hampshire. GAC Contactors
Dependent
Variable
Log (Effluent SPC)
Log /Effluent m-SPC \
Relationship
= -0.233 + 0.093 TEMP*
- n -)-3o -7 1 10 ri c 4- a •)•> ri T*
N
30
IK
Coefficient of
Determination (r2)
0515
n 7-?o
I Influent m-SPC f
"TEMP = °C; CLF = free chlorine; CLT = total chlorine.
Table 3. Effect of Operational Variables on Microbial Changes in GA C Pilot Columns at Troy,
New York
Effluent — Influent
Carbon
Depth
(cm)
RUN#1 —
O(control)
30.5
61
122
RUN #2 —
O(control)
61
61
61
RUN #3 —
Ofcontrol)
61
61
61
Infli
EBCT Tola
(min) (mg
lont imean t
1 C/2 Log
/L) (SPC)
12/19/80—4/1/81:
—
2.5
5
10
4/18/81—6/30/81:
—
6.66
5
0.0778
-0.1799
0.4061
0.3739
0. 1374
-0.1799
0.3843
2.5 0.4754
8/14/81 — 11/18/81
— * -0.0786
5 +1.0 3.3321
5 * 2.3841
5 +0.7 2.9381
jr rairea uirreren
Log
(m-SPC)
0.4871
1.4116*.
5. 10241
5.4171
-0.2313
1.41161
2.2001
0.9778
1.2150
5.399
ID11
ID11
cesj
Endotoxins
(vg/L)
-0.3897
-0.2657
0.5982
0.5320
-0.0589
0.1975
0.2144
0.1667
-0.0708
0.16051
0.0608
0.0693
"The feed to the columns contained 0.3 to 0.5 mg/L free chlorine in Runs til and #2; in Run #3,
increments of 0 7 and 1.0 mg/L were dosed to specific columns.
1 These means are statistically different from zero at the 5% confidence level using the paired
comparisons student t-test.
11 Indeterminate Insufficient observations were recorded to determine sample statistics.
Total organic carbon (TOC) does not
appear to be important in governing
microbial populations in the water
treatment and distribution systems
examined.
Increased values of pH, temperature,
and turbidity were positively asso-
ciated with SPC and m-SPC popula-
tion densities at both Manchester
and Concord.
Though both prechlormation and
post-chlorination have been found
to reduce microbial concentrations,
data collected in this study do not
consistently support the hypothesis
that chlorine residuals are inversely
related to microbial concentrations
in distribution systems.
For virtually all subsets of the field
data examined, a high, statistically
significant correlation existed be-
tween SPC and m-SPC values. In
many cases, however, there is a
numerical difference between the
values of these parameters recorded
at a given station ortime This result
suggests that though the m-SPC
procedure is a useful substitute for
the SPC procedure, the numerical
relationship between these two
-------
parameters is not a universal one.
Nevertheless, the m-SPC specific
standard curve is established in
much the same way that TOC is
used as an indicator for biochemical
oxygen demand in wastewater
treatment.
8. Though there was a statistically
significant, nonzero correlation
between endotoxins and various
indicator organisms, the absolute
magnitude of this correlation is very
low. Thus from a practical point of
view, the utility of the spectropho-
tometric LAL endotoxin assay as a
surrogate for indicator organisms is
minimal.
The full report was submitted in
fulfillment of Cooperative Agreement R-
807172 by Rensselaer Polytechnic Insti-
tute under sponsorship of the U.S.
Environmental Protection Agency.
Charles N. Haas is with the Illinois Institute of Technology, Chicago. IL 60616;
Michael A. Meyer. MarcS. Paller. Michael A. Zapkin, and Donald B. Aulenbach
are with Rensselaer Polytechnic Institute. Troy. NY 12181.
Raymond H. Taylor was the EPA Project Officer (see below).
The complete report, entitled "Microbiological Alterations in Distributed Water
Treated with Granular Activated Carbon." (Order No. PB 83-247 163; Cost:
$17.50, subject to change) will be available only from:
National Technical Information Service
5285 Port Royal Road
Springfield. VA 22161
Telephone: 703-487-4650
For further information. Eugene W. Rice can be contacted at:
Municipal Environmental Research Laboratory
U.S. Environmental Protection Agency
Cincinnati, OH 45268
GOVERNMENT PRINTING OFFICE 1983-659-017/7174
-------
United States
Environmental Protection
Agency
Center for Environmental Research
Information
Cincinnati OH 45268
Postage and
Fees Paid
Environmental
Protection
Agency
EPA 335
Official Business
Penalty for Private Use $300
RETURN POSTAGE GUARANTEED
AGENCY
CHICAGO IL 60604
------- |