EPA 570/9-87-001
APRIL 23-25, 1985

Joseph Cotruvo, EPA
Craig Vogt, EPA
panel chaimen
William Ford
Michael Kavanauqh
Peter Rodgers
panel scribes
Paul Berger, EPA
Arden Calvert, EPA
Steve Clark. EPA
Henry Erbis, EPA
Stig Regli, EPA
EDITORS: Stig Regli and Paul Beraer
Office of Drinking Wat^r

In April 1985, the American Water Works Association Research
Foundation (AWWA), under a grant from the U.S. Environmental
Protection Agency (EPA), Office of Drinking Water, convened a
workshop in Baltimore, Maryland, to discuss options for regulating
microbial contaminants, and treatment requirements such as filtration
and disinfection. The primary objectives of the workshop were to
provide EPA with advice and detailed comments and recommendations
on (1) specific monitoring requirements for coliforms/ turbidity,
heterotrophic bacteria, viruses, and Giardia, (2) definitions of
filtration and disinfection for regulatory purposes, and (3)
appropriate variance criteria for the treatment technique regulations
(filtration and disinfection of surface waters and disinfection
of ground waters).
Participants included environmental microbiologists and
engineers; federal, state, and local public officials; public
water supply personnel; and AWWA and EPA personnel. The
participants were assigned to one of three workgroups, and each
workgroup addressed a list of specific issues prepared by EPA.
The participants on each workgroup and their affiliations are
identified in Sections 1A and IB, respectively. The workgroup
response to each issue is presented in Section 2. The three
workgroups, two of which responded to the same list of issues,
and their chairmen are indicated below:
Microbiological Monitoring and Disinfection of Ground Waters —
William Ford
Mandatory Filtration and Disinfection of Surface Supplies (Group 1) --
Michael Kavanaugh
Mandatory Filtration and Disinfection of Surface Supplies (Group 2) —
Peter Rodgers
The recommendations contained herein do not necessarily
reflect the views of the EPA, but will be influential in the
development of revised microbiology and treatment regulations.


SECTION 1 A: Panel Participants
SECTION IB: Workshop Participants and Affiliations
SECTION 2: Issues and Panel Responses

Slmer Akin, Ph.D.
Cincinnati, Ohio
A. Amirtharaja, Ph.D.
Atlanta, Georgia
Stephen Bishop
Woburn, Massachusetts
Robert Chapman
Denver, Colorado
Charles Gerba, Ph.D.
Tucson, Arizona
Jerry Healey
Boston, Massachusetts
David Hendricks, Ph.D.
Fort Collins, Colorado
Ray Jarema
Hartford, Connecticut
:'Ucnael Xavar.augn, Ph.D.
Walnut Creek, California
John Kirner
Tacoma, wasmngton
Larry McReynolds
Los Angeles, California
James Pluntze
Olympia, Washington
Robert Reinert
Bridgeport, Connecticut
Edward C. Scheader
New York, New York
Kenneth Stone
Burlington, Vermont
Richard Tobin, Ph.D.
Ottawa, Ontario
Robert Willis
Portland, Oregon

Joseph Calabro, Ph.D.
Scranton, Pennsylvania
John Cleasby, Ph.D.
Ames, Iowa
Ellas Cooney
Wellesley, MAssachusetts
John Courchene
Seattle, Washington
Henry Graeser
Dallas, Texas
James Geoff
Washington, D.C.
Alan Hess
Paramus, New Jersey
Richard Karlin
Denver, Colorado
Gary Logsdon, Ph.D.
Cincinnati, Ohio
Fran Ludwig
New Haven, Connecticut
Fred Marrocco
Harnsburg, Pennsylvania
Richard Moser
Haddon Heights, New Jersey
John O'Connor, Ph.D.
Columbia, Missouri
Vincent OLivieri, Ph.D.
Baltimore, Maryland
William Opfer
Washington, D.C.
Peter Rogers
Sacramento, California
Peter Smith
Albany, New York
Joseph Taylor
Portland, Maine
Heather Wicke
Washington, D.C.
Larry Worley
Seattle, Washington

Blaise Brazos
Columbia, Missouri
Dean Cliver, Ph.D.
Madison/ Wisconsin
William Ford
Jefferson City, Missouri
Edwin Geldreich
Cincinnati, Ohio
Richard Kolish
Baltimore, Maryland
Donald Kuntz
Charleston, West Virginia
Ray Lee
Haddon Heights, New Jersey
Gordon McFeters, Ph.D.
Bozeman, Montana
Betty Olson, Ph.D.
Irvine, California
William Pamsh
Baltimore, Maryland
Wesley Pipes, Ph.D.
Philadelphia, Pennsylvania
John Regnier
Montgomery, Alabama
William Ring
Washington, D.C.
Billy Turner
Atlanta, Georgia

Paul Berger, Ph.D.
Washington, D.C.
Arden Calvert
Washington, D.C.
Steve Clark
Washington, D.C.
Joseph Cotruvo, Ph.D.
Washington, D.C.
Henry Erbes
Washington, D.C.
Victor J. Kimm
Washington, D.C.
John B. Manmon
Denver, Colorado
Edward Ohanian, Ph.D.
Washington, D.C.
Art Perler
Washington, D.C.
Stig Regli
Washington, D.C.
John Trax
Washington, D.C.
Craig Vogt
Washington, D.C.


Elmer Akin, Ph.D.
Drinking Water Researcn Div.
Health Effects Research Laboratory
Environmental Protection Agency
26 W. St. Clair Street
Cincinnati, OH 45268
A. Amirthara^ah, Ph.D.
School of Civil Engineering
Georgia Institute of Technology
Atlanta, GA 30332
Paul Berger, Ph.D.
Office of Drinking Water (WH-550)
Environmental Protection Agency
401 M Street, S.W.
Washington, DC 20460
Mr. Stephen Bishop
MeteaIf & Eddy
P.O. Box 4043
Woburn, MA 10888-4043
Mr. Blaise Brazos
Research Associate
College of Engineering
1047 Engineering Building
University of Missouri—Columbia
Columbia, MO 65211
Joseph Calabro, Ph.D.
Manager, Water Quality and
Purification Facility
135 Jefferson Avenue
Scranton, PA 18503
Arden Calvert
Program Analyst
Office of Drinking Water (WH-550)
Environmental Protection Agency
401 M Street, S.W.
Washington, DC 20460
Robert Chapman
Asst. Director '.varer Engineering
CH2M Hill
P.O. 3ox 22508
Denver, CO 80222
Steve Clark
Environnental Engineer
Office of Drinking Water iWH-550)
Environmental Protection Agency
401 M Street, S.W.
Washington, DC 20460
John Cleasby, Ph.D.
Iowa State University
492 Town Engineering Bldg.
Ames, IA 50010
Dean Cliver, Ph.D.
Food Research Institute
University of Wisconsin
1925 Willow Drive
Madison, WI 53706
Mr. Elias Cooney
Senior Vice President
Whitman & Howard
45 William Street
Wellesley, MA 02181
Joseph Cotruvo, Ph.D., Director
Criteria & Standards Div.
Office of Drinking Water (WH-550)
Environmental Protection Agency
401 M Street, S.W.
Washington, DC 20460
Mr. John Courchene
Water Quality Division
Seattle Water Department
1509 S. Spokane Street
Seattle, WA 98144
Mr. Henry Erbes
Environmental Engineer
Office of Drinking Water (WH-550)
Environmental Protection Agency
401 M Street, S.W.
Washington, DC 20460

Mr. William Ford, Former Director
Public Drinking Water Program
Missouri Dept. of Natural Res.
P.O. Box 176
Jefferson City, MO 65102
Edwin Geldreich, Senior Scientist
Drinking Water Research Division
Water Engineering Research Laboratory
Environmental Protection Agency
26 W. St. Clair Street
Cincinnati, OH 45268
Charles Gerba, Ph.D.
Dept. of Microbiology & Immunology
University of Arizona
Tucson, AZ 85721
Mr. Henry Graeser
Vice President
Black & Veatch
5728 LBJ Freeway
Dallas, TX 75240
James B. Groff
Deputy Executive Director
Government Affairs Office
American Water Works Association
1010 Vermont Ave., N.W., Suite 810
Washington, PC 20005
Jerry Healey
Chief, Water Supply Branch
Region I
Environmental Protection Agency
J. F. Kennedy Federal Building
Boston, MA 02203
David Hendricks, Ph.D.
Engineering Research Center
Colorado State University
Fort Collins, CO 80523
Mr. Alan Hess
Manager, Water Quality
Malcolm Pirnie, Inc.
P.O. Box 36
Paramus, NJ 07652
Mr. Ray Jarema, Chief
Water Supplies Section
Connecticut Dept. of Health
79 Elm Street
Hartford, CT 06115
Mr. Richard Karlin, Chief
Drinking Water Section
Colorado Department of Health
4210 East 11th Avenue
Denver, CO 80220
Michael Kavanaugh, Ph.D.
Vice President
James M. Montgomery Cons. Engrs.
2255 Ygnacio Valley Road, Ste. C
Walnut Creek, CA 94598
Victor J. Kimm, Former Director
Office of Drinking Water (WH-550)
Environmental Protection Agency
401 M Street, S.W.
Washington, DC 20460
Mr. John Kirner
Water Resource Coordinator
Tacoma Dept. of Public Utilities
P.O. Box 11007
Tacoma, WA 98411
Mr. Richard Kolish
City of Baltimore
Asburn Filtration Plant
3001 Druid Park Drive
Baltimore, MD 21215
Mr. Donald Kuntz, P.E., Chief
Water Supply Program
Div. of Sanitary Engineering
State Department of Health
Environmental Health Services
State Office Bldg. No. I
1800 Washington Street, East
Charleston, WV 25305
Mr. Ray Lee
Assistant Director
American Water Works Service Co.
500 Grove Street
Haddon Heights, NJ 08035

Gary Logsdon, Ph.D., Chief
Microbiological Treatment Branch
Drinking Water Research Division
Environmental Protection Agency
26 W. St. Clair Street
Cincinnati, OH 45268
Ms. Fran Ludwig
Environmental Affairs
New Haven Reg. Water Authority
90 Sargent Drive
New Haven, CT 06510-5966
Mr. John 8. Mannion
Deputy Executive Director
AWWA Research Foundation
6666 W. Quincy Avenue
Denver, CO 80235
Mr. Fred Marrocco, Chief
Division of Water Supply
Dept. of Environ. Resources
P.O. Box 2357
Harrisburg, PA 17120
Gordon McFeters, Ph.D.
Dept. of Microbiology
Montana State University
Bozeman, MT 59717
Mr. Larry McReynolds
Water Quality Manager
Los Angeles Dept. of Water & Power
111 North Hope Street
Los Angeles, CA 90051
Mr. Richard Moser
Vice President
American Water Works Service Co.
500 Grove Street
Haddon Heights, NJ 08035
John O'Connor, Ph.D. "Dr. O'Connor"
Professor of Civil Engineering
College of Engineering
University of Missouri—Columbia
1047 Engineering Building
Columbia, MO 65211
Edward Ohanian, Ph.D., Chief
Health Effects Branch
Office of Drinking Water (WH-550)
Environmental Protection Agency
40? M Street, S.W.
Washington, DC 20460
Vuicent Olivieri, Sc.D.
Johns Hopkins School of Hygiene &
Public Health
615 N. Wolfe Street
Baltimore, MD 21205
Betty Olson, Ph.D.
Associate Professor
Program in Social Ecology
University of California
Irvine, CA 92717
Mr. William Opfer
Official Representative
U.S.D.A. Forest Service
Room 1108 RPE, P.O. Box 2417
Washington, DC 20013
Mr. William Parrish
Chief, Division of Water Supplies
Maryland Office of Envir. Programs
201 W. Preston Street
Baltimore, MD 21201
Art Perler, Chief
Science & Technology Branch
Office of Drinking Water (WH-550)
Environmental Protection Agency
401 M Street, S.W.
Washington, DC 20460
Wesley Pipes, Ph.D.
Dept. of Civil Engineering
Drexel University
Philadelphia, PA 19104
Mr. James Pluntze
Section Head
Water Supply & Waste Section
Washington Dept. of Soc. & Hlth Serv.
Mail Stop LD—11
Olympia, WA 98504

Stig Regli
Environmental Engineer
Office of Drinking Water (WH-550)
Environmental Protection Agency
401 M Street, S.W.
Washington, DC 20460
Mr. John Regnier
Program Manager
Alabama Rural Water Association
4556 S. Court Street
Montgomery, AL 36105
Mr. Robert Reinert
Vice President
Bridgeport Hydraulic Co.
835 Main Street
Bridgeport, CT 06609
Mr. William Ring
Government Affairs Office
American Water Works Association
1010 Vermont Ave., N.W., Ste 810
Washington, DC 20005
Mr. Peter Rogers, Chief
Sanitary Engineering Branch
California Department of Health
714 P Street, Room 600
Sacramento, CA 95814
Edward C. Scheader
Deputy Director
New York Department of
Environmental Protection
1250 Broadway
New York, NY 10001
Mr. Peter Smith, Director
Bureau of Public Water Supply Prot.
New York State Department of Health
Empire St. Plaza, Tower Bldg., 4th Fir
Albany, NY 12237
Mr. Kenneth M. Stone, Chief
Division of Environmental Health
Vermont Department of Health
60 Main Street
Burlington, VT 05401
Mr. Joseph Taylor
General Manager
Portland Water District
P.O. Box 3553
Portland, ME 04104
Richard Tobin, Ph.D.
Canada Health & Welfare
Environmental Health Center
Tunney's Pasture
Ottawa, ON X1A OL2
John Trax, Chief
Drinking Water Branch
State Programs Division
Office of Drinking Water (WH-550)
Environmental Protection Agency
401 M Street, S.W.
Washington, DC 20460
Mr. Billy Turner
Vice President—Operations
Jordan, Jones & Goulding, Inc.
2000 Clearview Ave., N.E.
Atlanta, GA 30340
Craig Vogt, Deputy Director
Criteria & Standards Div.
Office of Drinking Water (WH-550)
Environmental Protection Agency
401 M Street, S.W.
Washington, DC 20460
Heather Wicke
Senate Staff
SD 410
Dirksen Senate Office Bldg.
Washington, DC 20510
Mr. Robert Willis
Water Engineering Supervisor
Portland Water Bureau
1120 S.W. 5th Ave., 6th Floor
Portland, OR 97204-1926
Mr. Larry Worley
Chief, Tech. Support Section
Drinking Water Branch, Region X
Environmental Protection Agency
1200 Sixth Avenue
Seattle, WA 98101


Responses of Work Group to Issues on Microbiological
Monitoring and Disinfection of Ground Waters
1. Compliance
ODW is considering the use of a presence/absence concept
for regulating total coliforms. Several compliance
issues must be addressed.
a.	Question: Should a few positive samples be permitted?
If so,
(1)	should compliance be based on the number of
positive samples or percentage of positive
samples during the reporting period? How many
(or what percent) should be allowed? How
about consecutive positive samples?
(2)	Should the maximum allowable number (or percent)
of positive samples depend on the number of
samples collected, number of people served by
system, or some other basis?
b.	Question: What should the reporting period(s) be?
12 consecutive months, 1 month, both, variable based
on monitoring frequency, other?
Response: The consensus was that no more than 5%
of the coliform samples should be positive over a
consecutive 12-month period and over one month.
No system sampling less than 20 times/month should
have more than one positive coliform sample per
There was extensive debate over the use of a presence-
absence (P-A) concept (also called "frequency of
occurrence monitoring"). Most felt that the present
system has served us well over many years and should
not be changed. Others supported the P-A concept
because of its conceptual simplicity, reduced data
truncation problems, and reduced sensitivity to
changes in coliform densities during transit/storage
periods. Part of the concern over the P-A concept
appeared to be based on the feeling by some that
more sensitive coliform media might be specified
for use with this concept in the revised regulations.

Sampling Frequency
a.	Question: Should the present minimum sampling
frequency scheme be retained?
Response: See Response to question 2c below.
b.	Question: Should frequency be based on population
served (present regulations), length of pipe, volume
of water pumped, source water quality and type, or
some other basis?
Response: It was quickly agreed that
frequency should be based on population served.
No reason to change was offered.
c.	Question: What should the minimum sampling frequency
be for small communities (under 5000 people served)?
Response: No consensus was reached. Some members
felt that 5 samples/month would be the minimum
acceptable, especially if the P-A concept was
employed. (According to studies performed by Dr.
Pipes, if 5x12 = 60 samples are collected over one
year, and at least 95% of them are negative, then
there is a 95% probability that less than 10% of
the water has coliforms present). Part of the
resistance to the P-A concept (mentioned earlier)
appeared to be this accompanying increase in
monitoring frequency for the smallest systems.
Some were concerned about the increased State
resources that would be needed if 5 samples/month
monitoring is required. The question arose of
whether one negative coliform sample/month is
meaningful or provides a false sense of security.
One individual indicated that many small systems
would ignore a mandated increased monitoring frequency
due to lack of information, greater expense, and
the additional resources needed. Alternatives to
higher minimum monitoring frequencies were discussed.
These included mandating on-site sanitary surveys
for small systems (especially if MCL violations
had occurred) and requiring disinfection if fewer
than 5 samples/month are collected.
d.	Question: Should samples be collected over regular
intervals or can collection be skewed (e.g., all
samples for a month collected at one time at
different locations)?
Response: No consensus was reached. For small
systems, collecting all samples during a single
day/month is cheaper and easier, and different

sections of the distribution are examined. However,
this approach would leave systems at risk over too
long a time (one month). Some members felt this
issue should be left to State discretion.
e.	Question: Should the same minimum monitoring
frequency apply to both community and non-community
Response: The same minimum monitoring frequency
should apply to both community and non-communitv
supplies; for non-communities, the number of samples
collected should be based on the expected number of
persons served during a month.
f.	Question: Should State discretion be provided for
reduced monitoring for certain systems (e.g.,
protected groundwater source serving a small
Response: Minimal time was spent on this issue and
no consensus was reached. Some participants
supported State discretion.
3.	Sampling location
a. Question: Present regulation at 141.21(a) reads:
"The samples shall be taken at points which are
representative of the conditions within the
distribution system." Should this be changed?
Response: The present definition concerning sampling
location should not be changed. Guidance should be
provided in the regulations that representative
samples should include high risk areas.
4.	"Check" sampling policy
a. Question: How many repeat samples should be required
after a positive original sample and should any be
taken at locations different from the site at which
the positive original sample was collected?
Response: If a coliform sample is positive, follow-
up action must be taken. This should include at
least one repeat sample from the same location as
the original sample. To diagnose the extent of a
possible problem, it is worthwhile to collect
additional repeat samples both in the vicinity of
the original positive sample and throughout the
distribution system, in order to characterize the
contamination. One participant indicated that
repeat samples from locations other than the original
sampling point should be at State discretion.

In no circumstances can a positive sample
be ignored.
b.	Question: Should compliance be based on a positive
original sample or only on a positive reDeat sample?
Should repeat samples be included in calculating
the total number of samples collected each reporting
period to determine compliance?
Response: Repeat samples should be included in MCL
calculations. This provision encourages additional
monitoring. If the P-A concept is used, then at
least 5 initial samples must be collected. One
State participant stated that an original positive
sample should be excluded for compliance purposes
if repeat samples are negative.
c.	Question: What type of response should be required
of a supply to a positive original sample? To a
positive repeat sample?
Response: A positive original sample should require
an attempt to locate and remedy the problem. A
repeat sample(s) should be required to assist in
this process.
One individual indicated that a positive repeat
sample suggests a localized persistant problem.
Another member suggested that repeat samples should
be examined for fecal contamination. Several
panel members suggested that public notification
should be reserved only for those systems which
cannot remedy their problem quickly. One participant
suggested that the degree of action response should
be left to State discretion, but that any evidence
of fecal contamination should require a boil-water
d.	Question: Should the number of colonies or number
of tubes be counted on original or repeat samples?
Response: The group presumed an EPA P-A approach
would obviate need for enumeration, but that States
could retain the right to require enumeration
at their discretion.
Chlorine substitution policy
a. Question: The present regulation at 141.21(h)
allow substitution of chlorine residual monitoring
for not more than 75% of the required coliform
samples. Should this provision be retained, changed,
or deleted? If changed, how?
Response: Chlorine substitution policy should be
deleted, since no State is making use of it. No

support was voiced for the concept. One panelist
stated that coliforms are often found in waters
containing 0.2 mg/1 free chlorine residual.
6. Heterotrophic bacteria
ODW is considering integrating heterotrophic bacteria
limits into the total coliform regulations on the basis
that high levels interfere with coliform analyses. If
heterotrophic bacteria counts (HPC) are above a specified
number (e.g., 500 colonies/ml), then a coliform-negative
original sample would be declared invalid and another
original sample sought. If HPC is incorporated into
the coliform regulation:
a. Question: What should the minimum monitoring
frequency be? Should this depend on size of system,
quality or type of source water, percentage of
coliform samples collected, or other factors?
Response: The following consensus was reached:
Because the HPC test is not currently
feasible for routine use because of the 6-8 hour
holding time limit at ambient temperatures, it is
recommended as a follow-up test when problems with
the coliform test occur (i.e., confluent growth or
"too numerous to count" situations with MF or turbid
gas-negative tubes (MPN). Any coliform sample
which manifests one of these problems would be
considered invalid and a repeat sample required.
The repeat sample would be transported within 6-8
hours (up to 24 hours if refrigerated) and analyzed
for both coliforms and HPC. If the HPC is 500
colonies/ml or less, then the repeat sample is the
valid sample. If the HPC is greater than 500
colonies/ml, then the coliform sample is invalid
(the data from samples previously taken are also
questionable/ which in the absence of subsequent
valid coliform results would signify violation of
monitoring requirements. The State will have to
follow-up to ensure that the system takes action
to re-validate their coliform analyses. In written
comments after the meeting, one participant recommended
that if a plate with dense or confluent growth
exhibits at least 5 distinguishable sheen colonies,
that this sample be considered valid and no
HPC test required. Another participant subsequently
suggested that an action plan for high densities of
heterotrophic bacteria should include flushing the
distribution lines and increasing disinfection;
alternatives might include anaerobic incubation for
the coliform test or use of an optional coliform
medium, both of which might reduce background
levels. This individual indicated that a transit
time of up to 24 hours was reasonable only if
samples are held below 10°C.

b.	Question: Where should samples be taken—water
plant, distribution system, Dlumbing systsns of
health care facilities, etc?
Response: Samples should be collected from the
same location as was the problem sample.
There was some debate on the effectiveness of HPC
as an indicator of water quality. If HPC is used in
such a capacity, then samples could be taken at
dead ends or at points associated with mean
water residency times.
c.	Question: Which analytical procedures or other
criteria will obviate requirement? (e.g., anaerobic
incubation, use of multiple fermentation tube test).
Response: If confluent growth or "too numerous to
count" situations develop for the MF test, or turbid
gas-negative tubes for the MPN test, then either
anaerobic incubation or increased water treatment
may be attempted to remedy problem.
d.	General:
There was some discussion on the narrow context of
HPC monitoring framed by the Agenda (i.e., HPC
limits based only on interference with total coliform
testing). Some felt that HPC should be a water
quality index in its own right, especially as
related to treatment effectiveness. Moreover, many
heterotrophic bacteria are opportunistic pathogens
which may be responsible for hospitalacquired
infections via the waterborne route. One individual
stated that AWWA has an HPC goal of less than 10
colonies/ml at the utility and less than 1000/ml in
the distribution system. Another individual stated
that a substantial change in HPC densities was more
significant than absolute HPC levels. One State
official stated that HPC monitoring should be
guidance rather than a regulation due to additional
resources needed, lack of laboratory facilities,
and long transportation times. With regard to the
holding time problem, HPC tests are being developed
which will greatly reduce this as an issue. Several
individuals were concerned about HPC limits based
on data where older, less sensitive media and
methods (i.e., pour vs. spread plate) were used.
1. Compliance
Currently there are two MCLs—a monthly average of 1
NTU and a two-consecutive day average of 5 NTU. Both

apply to systems using surface water sources in ^hole
or in part.
a.	Question: Should either reporting period oe
discontinued or changed?
Response: It was auickly agreed that there is no reason
to change the current reporting periods.
b.	Question: Should the two-consecutive day average
be retained as an MCL? If so, should the 5 NTU
value be retained?
Response: It was quickly agreed that there is no
reason to change the current MCL.
c.	Question: If the turbidity MCL is set within the
0.1-0.5 NTU range, should a few samples exceeding
this value be permitted during the reporting period?
If so, how many or what percent of the total samples?
Response: The group consensus was that none are
permitted now and there is no reason to change
this. One member pointed out, however, that if this
MCL represents a monthly average, that by definition
a few values above the MCL value are permitted.
d.	Question: Can turbidity be effectively measured as
low as 0.1 NTU on a routine basis, by treatment
plant personnel? What is the lower limit of reliable
quantification using equipment and techniques
appropriate for treatment plant laboratories.
Response: One member stated that measurement of
0.1 NTU is routinely achievable, while others
asserted 0.2 NTU was the lowest level practical to
e.	Question: Should turbidity standards be applied to
groundwater systems?
Response: Consensus was quickly reached that
turbidity monitoring should not apply to groundwater
systems. One individual stated that groundwater
turbidity is usually inorganic and thus less important
to human health.
2. Sampling frequency
a. Question: Current regulations require daily
monitoring. Should this be changed? To what?
Should system size be a factor? Should continuous
monitoring be required for larger systems? If so,
how large?

Response: Regulations should permit eliner daily
monitoring or continuous monitoring. Some felt it was
necessary for a State to specify a uniform time of the
day for sample collection to standardize measurements.
b. Question: Should waivers be allowed for curbiaitv
monitoring? If so, what criteria should be used?
Response: The idea of a waiver was quickly rejected.
Sampling locations
a. Question: Current regulations require samples to
be taken "... at a representative entry point(s) to
the water distribution system..." Should this be
changed to include distribution system monitoring?
Response: Turbidity monitoring regulations should
include entry point and not distribution system
There was some debate on this issue. Individual
panel members observed that distribution system
monitoring relates more closely to water served to
the customer, that health aspects of turbidity
increases in the system were ill-defined, additional
monitoring is more expensive, and sample holding
times must be limited to six hours at ambient
temperatures. One individual suggested the use of
HPC monitoring instead. The group felt that the
association of turbidity and health should be a
research item.
"Check" sample policy
a.	Question: If a daily sample exceeds the standard,
a repeat sample must be taken. Presently, the
repeat sample is the sample used for calculating
the monthly average. Should this practice be
allowed in the revised regulations?
Response: The group quickly concluded a change in
the current regulations is not warranted.
b.	Question: Present regulations require that if
the turbidity MCLs are exceeded, the supply must
report to the State within 48 hours and notify the
public. Should the revised regulation specify
additional requirements if a repeat sample exceeds
the MCL?
Response: The revised regulations should specify that
if a repeat sample is positive, remedial action
must be taken.

5. General
The merits of turbidity monitoring were debated.
One individual stated that turbidity removals die not
correlate at all with removals of bacteria (see minority
report at the end of this document). Other individuals
disagreed. One individual indicated that a turbidity
level at or below 0.1 NTU reflects the reduction of
microbial densities to levels which guarantiee the
effectiveness of disinfection. Mo one disagreed
with the use of turbidity monitoring as an indicator
of Giardia removal, although one individual wondered
whether this alone would justify a national turbidity
standard. The comparative effectiveness of turbidity
vs. HPC monitoring was briefly discussed/ but this
issue was not resolved. One individual noted that
turbidity monitoring provided more rapid results and
was cheaper than HPC monitoring, and the ease of
measurement promoted more comprehensive ttionitoring.
Turbidity MCL values were also discussed. One member
noted that the absolute level of turbidity may be less
important than the percentage reduction of turbidity.
Another panelist consider smaller, gradual and consistent
changes of greater significance than occasional large
sharp increases.
In spite of recent advances in monitoring methods for
viruses and Giardia, certain technical difficulties
remain. These include the sophistication and training
required to do the analysis (but not necessarily
sampling), the delay of days or weeks in receiving
results from the analysis, the accuracy and reproducibility
of the test results (i.e., false negatives, and false
positives for Giardia) and the need to either perform
multiple tests, or select an indicator.
a.	Question: In view of these difficulties, is it
technically and economically feasible to initiate
wide scale monitoring of viruses and Giardia in
drinking water on a routine basis?
Response: It is not yet technically and economically
feasible to initiate routine virus and Giardia
b.	Question: Is there sufficient basis for using
either the classical enteroviruses or coliphage as
an indicator for the presence of other waterborne
pathogenic viruses?

Response: There is not yet a sufficient basis for
using either the classical enteroviruses or coliphage
as an indicator for pathogenic viruses.
1. Question: What is the definition of "disinfection" for
purposes of specifying appropriate techniques in the
a. Question: Which disinfection technologies are
appropriate? What design characteristics and
operating parameters, such as disinfectant
concentration and contact times or CxT values should
be included in the definition?
Response: Disinfection should be operationally
defined as any process which reduces the level of
viruses and nonsporeforming bacteria by 4 logs
(99.99%). The following parameters may be used to
reflect this goal.
Dis infectant

free chlorine
<_ 7.5

> 7.5
£ 7-5

b.	Question: What parameters and monitoring requirements
should be specified in the regulations to assure
consistent and effective performance of the
disinfection technique?
Response: Free disinfectant residual levels should be
measured daily at the plant effluent. Disinfectant
levels should reflect CxT values specified above.
C&T values vary with pH, temperature, and turbidity
c.	Question: In what circumstances might application
of a certain type of disinfection be inappropriate?
Response: A disinfectant may be inappropriate if
such high levels are needed that human toxicity
becomes a factor. This includes chlorine, chlorine
dioxide, and iodine.
Chloramines may pose a hazard for individuals on
dialysis and for fish (individuals with aquariums,
especially fish suppliers, would need to protect

fish). Ultraviolet light may be ineffective if the
irradiated water is subsequently exposed to ambient:
light for any length of time, due to photoreact1vation.
d. Question: Should there be a requirement for
maintenance of a residual in tne system? For the
various disinfectants, what residual should be
Response: Systems using chloramines snould
monitor for residual in the distribution system.
State discretion should be allowed for other
There was some debate on how effective free chlorine
is in the distribution system. One individual
stated that a free chlorine residual was not
particularly germicidal, while others felt it
helped control slime and problems associated with
cross-connections, and also suppressed growth of
heterotrophic bacteria which interfere with coliform
2. A variance may be granted if the public water system
demonstrates to the satisfaction of the State that
treatment is not necessary to protect the health of
persons because of the nature of the raw water source.
The variance is to be conditioned on meeting specific
requirements, such as monitoring, as appropriate.
a. Question: What criteria would have to be met to
demonstrate to the State that disinfection was not
needed? What monitoring or other treatment should
be specified as a condition for receiving a variance?
Response: The following criteria would have to be
met for a variance from disinfection (some individuals
objected to the term "variance" as being pejorative
and preferred another term).
"there has not been any waterborne disease outbreak
reported to the State with the water system in its
present configuration (i.e., if the reason for a
previous outbreak had been identified and the system
substantively upgraded to remedy the situation,
then this variance criterion would be satisfied).
"an on-site sanitary survey is performed as a
condition of variance and periodically thereafter
as specified by the primacy agent. The sanitary
survey results must support finding that the
specified treatment is not necessary.

"total coliform monitoring is performed at least 5
times/month for systems monitoring less than that
now. If a sample produces confluent growth on
MF or neavy growth in gas-negative MPN tubes, then
riPC monitoring is Derformed (as per response to
question 6 under Total Coliforms).
°there is a history of total coliform monitoring
with no MCL violations within the last 12 months or
last 60 samples whichever time is greater.
There was no discussion on monitoring virus levels.
3.	A variance may also be granted if an alternative
treatment technique is shown to be at least as efficient
in lowering the level of the contaminant(s) as the
required disinfection techniques.
a. Question: What might be considered equivalent
treatment and upon what basis? What data would
necessarily be available to demonstrate that the
equivalent technique is at least as efficient in
contaminant reduction? What monitoring or other
requirements, such as measurement of performance,
should be specified?
Response: Any alternate treatment which demonstrates
a 4-log reduction in bacteria and viruses should
4.	a. Question: Should the disinfection treatment technique
requirement apply to non-community systems? Would
different variance criteria and monitoring requirements
be appropriate? If so, please specify.
Response: The disinfection requirement should
apply to non-community systems. Variance criteria
and monitoring requirements should be the same as
those for community systems.
5. General
There was some debate as to whether groundwater systems
should be required to disinfect. Some felt the majority of
non-disinfecting groundwater systems pose no threat to public
health, and that requiring them to justify a request for a
variance imposes an unnecessary burden. State regulators
would incur a corresponding burden of processing and deciding
upon formal variances under this scheme. As the problem
systems are seen to be the exception rather than the rule,
these participants felt the proper regulatory scheme should
reserve mandatory provisions for only systems with demonstrated
problems, with the burden of proof on the regulators. EPA

representatives explained the SDWA provided for no such
"selective" application of a treatment regulation, except
through the variance procedure outlined.
One individual suggested that groundwater disinfection
should be a "Category II" regulation (as explained in the
Phase II ANPRM), and thus apply only to those locales
experiencing viruses in their source water.
t ives
A vote was requested to gauge the participants opinion of
mandatory disinfection requirement applied to ground-water
source systems. A majority of those voting (EPA representa
and others abstained) did not favor the requirement.
1.	Define relationship of P-A concept to existing coliform
2.	Develop and validate more sensitive total coliform media.
3.	Define relationship between turbidity levels and
heterotrophic bacteria densities in potable water.
4.	Define sensitivity of turbidimeters in measuring low
turbidity levels.
5.	Investigate utility of turbidity monitoring in the
distribution system.
One panel member submitted a minority report questioning
the wisdom of turbidity monitoring. His reasons are:
1. There is no correlation between turbidity and bacteria
in drinking water at 1 NTU and below. Therefore, turbidity
is not a reliable parameter by which to measure filtration
plant efficiency (Chemical and Microbiological Evaluations
of Drinking Water Systems in Missouri, AWWA WQTC, Denver 1984).

2.	Turbidity does not interfere with disinfection due to
reduction in the chlorine residual. The chlorine demand
in drinking water is in the dissolved fraction of the total
organic carbon and on the surface of pipes.
3.	Turbidity is a qualitative measurement of light
scattering and has never been chemically quantitated.

Responses of Workgroup # 1 to Issues on Mandatory Filtration
And Disinfection of Surface Supplies
Viruses and Giardia Monitoring
1. In spite of recent advances in monitoring methods for viruses
and Giardia, certain technical difficulties remain- These
include: the sophistication and training required to do the
analysis (but not necessarily sampling), the delay of days
or weeks in receiving results from the analyses, the accuracy
and reproducibility of the test results (i.e., false negatives,
and false positives for Giardia), and the need to either per-
form multiple tests, or select an indicator.
a. Question: In view of these difficulties, is it techni-
cally and economically feasible to initiate monitoring
of viruses and Giardia in drinking water on a routine
Response: It is not technically feasible or economically
practical to initiate monitoring of viruses and Giardia
in regard to a maximum contaminant level (MCL) regulatory
approach. The cost of monitoring is not a key factor
in this conclusion.
While it is highly desirable to be able to determine
the efficiency of processes, current technology is not
at that stage. As a research goal, accurate and repro-
ducible methods should be sought.
The status of virus monitoring is such that in many
tissue culture methods, the analytic costs are $125 to
$500 per sample. With extensive monitoring, the cost
can be reduced, as it was in the United Kingdom, to $70
to $80 per sample.
The cost of Giardia monitoring ranges from $40 to $300
per sample using concentration and microscopic examina-
tion. The Riggs/EPA method shows promise as being a
better method. The current method has a drawback in
that it is not feasible to distinguish viable cysts from
nonviable cysts.
Both pathogens occur in nonrandom distributions. Patch-
iness of occurrence poses severe sample frequency
problems. There are data interpretation difficulties
(i.e., absence is not an indication of a "safe" supply).
The existing data base on frequency of occurrence and
distribution are very limited. Low turbidity values
(less than 1 NTU) do not necessarily mean absence of
Giardia cysts.

Because of these factors, the Work Group also made these
I.	A water supply susceptible to contamination from
v iruses/Giardia requires treatment.
II.	Encouragement should be given to research in devel-
opment of methods, especially Giardia, and to
monitoring to expand the existing data base.
Question: Is there sufficient basis for using either
the classical enteroviruses or coliphage as an indicator
for the presence of other waterborne pathogenic viruses?
Response: There is not sufficient basis for the use
of enteroviruses or coliphages as an indicator organism,
but they may be useful adjuncts for process control.
Disinfection Treatment Techniques
1. Question: What is the definition of "disinfection" for pur-
poses of specifying appropriate techniques in the regulation?
a.	Which disinfection technologies are appropriate? What
design characteristics and operating parameters,
such as disinfectant concentration and contact times
or CxT values, should be included in the definition?
b.	What parameters and monitoring requirements should
be specified in the regulations to assure consistent
and effective performance of the disinfection technique?
c.	In what circumstances might application of a certain
type of disinfection be inappropriate?
d.	Should there be a requirement for maintenance of a
residual in the system? For the various disinfectants,
what residual should be required?
Response: EPA should define concentration and contact time
criteria for free chlorine disinfection of unfiltered, low
turbidity source water to give 99.9% destruction of
Giardia 1., as a function of pH and temperature.
For filtered sources, less stringent CT minimum values will
be acceptable. EPA should specify CT values for removal
of viruses to levels providing a microbiologically safe
water using free chlorine (e.g., CT greater than or equal
to 15, @ pH less than 8, at some specified temperature).
Alternative or equivalent disinfectants that meet the three-
log reduction criteria for Giardia should be approved by
states, without a variance procedure (i.e., definition dis-
infection includes free chlorine or "equivalent oxicants").

EPA should provide guidance on what procedure(s) should
be used for demonstrating equivalency. Other sources of
information should also be employed (e.g., American Water
Works Association Research Foundation).
Prior to development of these responses, the group had
significant discussion regarding disinfection, including
a definition. Disinfection was defined as a process of
destruction or inactivation of human pathogens to provide
a microbiologically safe water for human consumption, using
acceptable levels of chemical oxidants, or equivalent agents.
Further discussion brought forth the following points:
° Numerous oxidants have been shown to be capable of
destruction of human pathogens (e.g., free chlorine,
chloramines, chlorine dioxide, ozone). Other processes
such as ultraviolet light also show this capability
to destroy human pathogens.
° The effectiveness of disinfection is dependent on many
factors, including: target organism, temperature,
pH, oxidant demand, oxidant mixing, contact time,
and pretreatment.
° Giardia 1. is the most resistant known organism to
free chlorine disinfection.
° Disinfection is the essential final barrier.
° Concern for by-product formation may limit the doses of
certain disinfectants.
The group developed a regulatory definition of disinfection
which would assist EPA in writing a mandatory disinfection
Disinfection is the process of reduction of pathogens
to a level such that they are not a significant source
of disease. Free chlorine is the disinfectant most
commonly used and will be the reference disinfec-
tant for comparative studies. Other disinfectants
will be permitted, provided that they are as effective
as free chlorine. The primacy agencies will determine
equivalency of disinfectants.
Because of their resistance characteristics, Giardia
species and viruses are the target organism for disin-
fection (and related water treatment techniques).
In general, all other pathogens will be inactivated
and/or removed by processes effective against these

Design Goals
Tables prepared by EPA will be used for guidance in
disinfecting with free chlorine in water treatment
plants. Other tables will be required to be developed
on an ongoing basis for other disinfectants. It is
realized that byproducts from disinfection processes
must be considered, and the MCL for THMs must be
me t.
a.	For treatment without filtration:
The objective of disinfection is a three-log reduc-
tion of Giardia species.
b.	For treatment including filtration processes:
The objective of disinfection is to inactivate
viruses (as Giardia are removed by filtration).
Question: A variance may be granted if the public water
system demonstrates to the satisfaction of the state that
the treatment is not necessary to protect the health of
persons because of the nature of the raw water source.
The variance is to be based on meeting specific require-
ments, such as monitoring, as appropriate.
a.	What criteria would have to be met to demonstrate to
the state that disinfection was not needed?
b.	What monitoring or other requirements should be speci-
fied as a condition for receiving a variance?
Response: No variances for disinfection of surface sources
should be allowed. There was almost no discussion on
this consensus. Variances for disinfection of ground water
were not addressed.
Question: A variance may also be granted if an alternative
treatment technique is shown to be at least as efficient
in lowering the level of the contaminant(s) as the required
disinfection techniques?
a. What might be considered equivalent treatments and
upon what basis? What data would necessarily be
available to demonstrate that the equivalent technique
is at least as efficient in contaminant reduction?
b. What monitoring or other requirements, such as measure-
ment of performance, should be specified?
Response: As mentioned before, the group felt that alterna-
tive disinfection methods should be approved or authorized

by the primacy agency, without requiring a variance oroce-
dure. Monitoring requirements were not discussed.
4. Quest ion; Should the disinfection treatment technique
requirement apply to noncommunity systems? /Jould different
variance criteria and monitoring requirements be appropriate?
If so, please specify.
Response; Requirements should be the same for both community
and noncommunity systems. Almost no discussion was required
to reach this consensus opinion.
Filtration Treatment Technique
1. Question: What is the definition of filtration for purposes
of spec ify i ng appropriate techniques in the regulation?
a.	Should the definitions of appropriate filtration tech-
nologies for the Federal Regulations be based upon
design and operating parameters or upon some type of
performance criteria using an indicator? Would this
choice change with different technologies? Would it
be appropriate to have both bases for certain technol-
b.	Which filtration technologies are appropriate? For
each technology, which design and operating parameters
or performance parameters should be included in the
definition? What values should these parameters be?
c.	What parameters and monitoring requirements should
be specified in the regulations to assure consistent
and effective performance of the different filtration
d.	In which circumstances might a certain filtration tech-
nique be inappropriate?
Response: "Filtration" is defined as an engineered process
for removal of particulate matter (natural or anthropogenic
origin) by passage through porous media. For regulatory
purposes, the group decided to suggest a definition of fil-
tration for slow sand filtration, diatomaceous earth
filtration, and rapid granular filtration. Each of
these were defined as follows:
Slow Sand Filtration
Slow sand filtration is a filtration technique which is
applicable for nonclay bearing, low-turbity, low-algae
waters and is generally defined as follows:
° No coagulating chemicals are employed as filter aids.

Application rates are less than 0.4 cubic meters per
hour per square meter of surface area (10 gal/sq ft/hr)
unless pilot plants demonstrate effective treatment
can be achieved at higher rates.
The sand or filtration media
0.4 mm or less, unless pilot
has an effective size of
plants demonstrate other-
° Filter to waste facilities should be provided.
° Whenever the depth of the filter bed becomes less than
60 cm due to removal of the surface layer ("schmutz-
decke"), the sand bed is cleaned by a process known
as "throwing over" or equivalent and additional
media is added to the bed.
The filtered water shall have a turbidity of less than
1.0 turbidity units (NTU).
Diatomaceous Earth Filtration
Diatomaceous earth filtration is a filtration technique
which is appropriate for relatively low turbidity waters
and is generally defined as follows:
° A precoat cake consisting of 3 to 5 mm thick layer
of powder-sized diatomaceous earth filter media is
deposited on a support membrane (septum).
while the water is filtered by passing through the
cake on the septum, additional filter media, known
as body feed, is continuously added to the feed water,
in order to maintain the permeability of the filter
° The filter media may or may not be coated with alum.
Filtered water shall meet MCLs for turbidity (currently 1
Rapid Granular Filtration
Granular filtration is a technique which is generally
defined as having the following characteristics:
° Primary coagulant required for adequate removals,
coagulant aid may be used.
° Direct filtration may be used consisting of rapid
mix and flocculation prior to filtration.
° In-line filtration may be used consisting of rapid mix
(RM) only.

0 Conventional pretreatment may be used consisting of
RM, flocculation, solids separation by sedimentation
or equivalent.
° All processes may use granular media, using various
designs as appropriate.
° The average turbidity of the filtered water, which
will be introduced into the distribution system, meets
the MCL (currently 1 NTU).
0 If the raw water turbidity is less than 1 NTU, the
filter system should demonstrate "substantial"
(greater than 50%) reduction of turbidity.
° The filters are periodically backwashed. When the
filter is returned to service, the first filtered waters
treated are wasted.
As is apparent from the above definitions, the group agreed
on performance type standards, except for slow sand filtra-
tion. Slow sand filters do not have any real operating
flexibility and they have a long history of use so that
design parameters can be established. Of course, with
pilot studies which demonstrate performance it may be
possible to exceed these design parameters.
Question: A variance may be granted if the public water
system demonstrates to the satisfaction of the State that
the treatment is not necessary to protect the health of
persons because of the nature of the water source. The
variance is to be based on meeting specific requirements
such as monitoring, as appropriate.
a.	What criteria would have to be met to demonstrate to
the states that filtration was not needed?
b.	What monitoring or other requirements would be specified
as a condition for receiving a variance?
Response: The group suggested a set of variance criteria
as follows:
(1) Watershed control
° Watershed management program as approved by the
primacy agency may include:
Restricted use
-	Land use control
-	Mammal control
-	Extended raw water storage
(This item is difficult to quantify.)

(2) Design of equivalent process
° Disinfection must meet requirements (as minimum)
for Giardia 1. 1nactivation, three-log kill (99.9%
° Complete system redundancy of all components,
except contact basin
° System should provide operational flexibility
such as off-line storage, alternate emergency
(3)	Finished water monitoring
° Continuous free CI2 residual monitoring or equiva-
lent oxidant monitoring, with redundancy
0 Monitoring of heterotrophic plate count with
target level of less than or equal to 10/ml
0 Frequent coliform analysis required with frequency
to be determined by the primacy agency.
(4)	Distribution system monitoring
0 Frequent monitoring in distribution system of HPC,
with a target of less than or equal to 500/ml at a
frequency to be determined by the primacy agency.
(5)	Raw water monitoring
0 Routine monitoring for bacterial and Giardia
contamination of the raw water source will be
conducted. The purpose of this monitoring is
to (1) establish the quality of the water, (2)
detect any trends in water quality that could
require modifications to treatment or disinfection
practices, and (3) to establish that the level
of microbiological contamination is within the
capacity of the disinfection methods proposed.
(No absolute limits on the degree of raw water
contamination should be specified by the EPA.)
(6)	State requirements
0 System must meet all additional primacy requirements.
There was extensive discussion as to whether those systems
which control their watersheds and have effective disinfec-
tion against Giardia should have to go through a variance
procedure, since a variance implies something less than
what is required is being provided. On the other hand,
unresolved issues temper enthusiasm for variance criteria

for a number of participants. Is a single barrier concept
acceptable even though risk can't be quantified? Is the
database adequate to define minimum disinfection design
criteria for 99.9% inactivation of Giardia 1.? With
these questions, the setting of criteria becomes a public
policy decision rather than a scientific/technical decision
in view of lack of quantifiable risks. Some people felt
that no variances should be permitted because of potential
health hazards and uncertainty in the database. Other
members object to the procedure because of the stigma
(perception) of wrongdoing on the part of the utility.
These members preferred a treatment regulation which
would permit primacy agencies to authorize other treatment
techniques, such as disinfection of Giardia in conjunction
with watershed management, without requiring a variance
procedure. It was agreed that whatever the formal and
legal procedures, a detailed critical examination would
have to be made and the technical, management, and opera-
tional aspects of the evaluation would be essentially the
same for all systems. The multiple barrier approach is
desirable when treating surface waters known to be
contaminated with human pathogens. It was also stated
that if you look long enough, you will find Giardia in
any water supply.
Three regulatory approaches were discussed for surface
(a)	Mandatory filtration, no variances
(b)	Mandatory filtration, with variance procedure
(c)	Mandatory treatment for Giardia removal, states to
determine equivalent treatment (e.g., disinfection
rather than filtration plus disinfection with no
variance requirement).
A straw poll showed 6 out of 16 for (a), 7 out of 16 for
(b), and 3 out of 16 for (c).
The majority opinion of the group was that filtration should
be mandatory, but provisions for variances should be estab-
lished. The benefits of filtration were broadly accepted
(a)	Improved aesthetic quality of the water
(b)	Reduced particle content
(c)	Reduced chlorine demand
(d)	Reduced organic content, most obviously THM precursors.

It was also broadly accepted that filtration is a desirable
practice from numerous viewpoints (barrier concept, debris
control/ etc.)/ cost considerations aside. However, if
the existing MCL for turbidity is maintained and can be
met, and if the effectiveness of chlorine for Giardia control
can be rigorously established, there do not appear to be
clearly definable conditions under which a variance should
not be granted. The granting of variances will thus require
the exercise of considerable professional judgment, and
this process must obviously allow for full public disclosure
and opportunity for input to the decision process.
Therefore, the group recommends that a rigorous set of
criteria be developed which should be considered in the
decision process whereby a variance might be granted. This
approach will best assure that both responsible regulatory
professionals (with the primacy agency in each state) and
the public have sufficient information on which to make
prudent decisions. Those entities seeking a variance should
be required to prepare a rigorous assessment of the impacts
of not filtering the water to the satisfaction of the
primacy agency. The assessment should include, as a
minimum/ the following elements:
(a)	Water resource evaluation
Water quality summary
Sanitary survey
Land use and recreational use practices
Storage considerations
Multiple source flexibility
(b)	Proposed disinfection practices
Chlorine residual and contact time criteria
(or alternate disinfectant)
Provisions for continuous monitoring of residual
Redundancy provisions to assure reliability
Demonstrated capability to operate and maintain
the disinfection equipment
(c)	Direct comparisons of the impact of the filtration/no
filtration decision on measureable parameters such as:
Particle size distribution
Trihalomethanes (THMs)
Total Organic Carbon (TOC)
Total Halogenated Organics (TOX)
Taste and odor
Chlorine demand/required residual
Heterotrophic plate count and coliforms in the
system (regrowth potential)
Iron/ manganese
Asbestos particles

(d) Cost evaluations
Incremental capital and O&M cost of filtration
(including cost saving emerging technology)
Total dollars
Cost per customer (rate impacts)
Ability to finance
A minority felt that a provision for variance would be
wrong and gave the following arguments to support this:
(a) Virtually all mountain watersheds have background
levels of Giardia cysts. Also in these watersheds,
turbidity levels may be less than 1 NTU most of the
time. In addition to Giardia cysts these low tur-
bidity waters have background levels of a host of
other microscopic biological substances which may
render the water not palatable to many persons,
though they are not a health hazard.
(b) The designation "controlled watershed" is a conceptual
fallacy. It carries the connotation that the water
is protected from contamination. Any watershed is
accessible to contaminants whether naturally occur-
ring or from human sources. While restrictions to
access may reduce the probability of contamination
by pathogens there is, nevertheless, a large measure
of uncertainty. When a contamination event occurs,
the single barrier approach may be inadequate.
(c)	The double barrier approach, i.e., filtration and
disinfection, provides the means for a water utility
to assure production of potable water regardless of
continually occurring activities and random events
which may occur on the watershed, whether natural or
human in origin. The double barrier philosophy is
advocated by a substantial number of professionals
in the waterworks industry. This philosophy is an
interpretation of national norms with respect to the
question of risk aversion in drinking water.
(d)	Filtration is defined by the work group as "the
engineered process for removal of particulate matter
(natural or anthropogenic origin) by passage through
porous media." Filtration, properly designed and
operated, will assure that low levels of background
particulates will be removed to produce a "palatable"
finished water, and a "safe" water when combined
with chlorination. Further, high levels of particu-
lates which occur probabalistically, above background
levels, will be removed. The disinfection step then
has a lesser burden and can function to provide an
additional measure of assurance and protection, as

(e) The turbidity standard and the coliform standard are
necessary, but are not sufficient to assure safe,
palatable water. An analysis of virtually any
natural water may reveal a variety of naturally
occurring microscopic substances, even though
turbidity levels are less than 1 NTU, and indeed,
coliform levels may be 0-40 organisms/100 ml. These
include eggs of insects, fecal matter from animals,
nematodes, algae, etc. Although not visible to the
eye, the presence of these organic substances must
affect palatability if consumers were aware.
(f) Disinfection may be an adequate treatment process
for countries where costs, benefits, and risks must
be weighed carefully (e.g., less developed countries
where investment capital is not available).
(g) While local utilities should have the right to
interpret local norms for risk aversion and priorities
on use of capital, this changes when national drinking
water standards are adopted. It is not fair to
states and localities having similar water sources
who have undertaken programs for mandatory filtration.
Nor is it fair to the traveling public who should be
able to take for granted that all public water
supplies in this country are both palatable and
safe. They should be warned by notices if a given
city is a purveyor of an unfiltered water supply.
When we have a national policy, as exemplified by
P.L. 93-523, the affairs of water utilities are no
longer strictly local.
It was also expressed by the minority that if it is poli-
tical reality that the variance provision is necessary,
then the regulations should defer to the states on defin-
ing specific provisions for the variance. The guidelines
developed in this Workshop should not be incorporated
as regulations, but should be only advisory guidelines
available to the states. The reason for this is that
incorporation of a variance specification legitimizes the
variance concept. Also, it undercuts the programs of
those states already enforcing the mandatory filtration
of surface waters with no provision for variance. It
does not seem logical that a few utilities having loud
voices should be able to undermine the programs of those
states that have elected to give economic priority for
compliance with national norms for drinking water safety.
If variances are granted, it should be with respect to
time only. Perhaps 5, 10, even 20 years could be permitted
to give utilities an opportunity to plan their capital
expenditures in a reasonable fashion.

3. Question: A variance may also be granted if an alternative
treatment technique is shown to be at least as effective
in lowering the level of the contaminants as the required
treatment technique.
a.	What treatments might be considered equivalent to
filtration and upon what basis? What data would
necessarily be available to demonstrate that the
equivalent technologies are at least as effective in
contaminant reduction?
b.	What monitoring or other requirements, such as measure-
ment of performance, should be specified?
Response: In light of the above discussion, alternative
treatment technologies such as disinfection for Giardia
are suggested to be included in the mandatory treatment
techniques regulation. Alternate filtration technologies
or emerging technologies were not discussed.
4. Question: Should	the filtration treatment technique
requirement apply	to noncommunity systems? Would different
variance criteria	and monitoring requirements be appropriate?
If so, specify.
Response; Requirements should be the same for both community
and noncommunity systems. Again, no discussion was needed
to reach this consensus.

Responses of Workgroup $ 2 to Issues on Mandatory Filtration
And Disinfection of Surface Supplies
In writing the regulations, EPA's responsibilities should
include the mandate for requiring treatment, definition of
treatment, performance standards, and variance criteria. EPA
should not incorporate engineering design and operational
characteristics into the regulations. Treatment operation
and design criteria should be developed by State agencies
based on guidelines provided by EPA.
Giardia and virus monitoring is useful for some agencies
and large utilities but not for general use because of high
costs and technical difficulties.
Proper pretreatment is essential for effective Giardia
cyst removal by conventional and direct filtration.
EPA should develop guidelines from which state regulatory
agencies can evaluate disinfection processes with chlorine,
chloramines, chlorine dioxide, and ozone. CT curves should
be provided for each disinfectant at different pH's and
Disinfection can be defined as an organism destruction
process which aims to eliminate pathogenic organisms in
drinking water.
Regulations should not specify monitoring or residual
disinfectant requirements (i.e., requiring a concentration
level to be maintained) •
Mandatory disinfection should be required for all
community and non-community surface water systems, and no
variances should be allowed.
The majority of the work group felt all ground waters
must maintain a disinfectant residual to the extent feasible,
allowing variances where appropriate. A minority of the work
group felt it would be appropriate for some ground water
systems to not maintain a residual and that the regulation
should allow states to grant such exclusion without needing
to go through a complicated, time consuming "variance"
procedure. Criteria suggested by minority group members for
allowing such exclusion included: no record of disease
outbreak, monitoring history for coliforms (no less than 5
coliform samples per month), heterotrophic plate count, and
an adequate sanitary survey.
Recognized treatment technologies that should be speci-
fied in the required treatment regulation for surface waters
include conventional treatment, slow sand filtration, direct
filtration (with pretreatment), and diatomaceous earth.

Alternative technologies must be determined through a "variance"
procedure which should be much more simple than that which
currently exists.
A majority of the group felt that simplified variance
procedure is especially important for many small and non-
community systems which will be trying to meet the treatment
requirement with alternative treatments (e.g., cartridge
filtration). A minority of the group felt that no variances
should be granted under a mandatory filtration requirement.
Conventional treatment should be defined as a treatment
process including coagulation, flocculation, sedimentation,
and filtration.
Slow sand filtration should be defined as a treatment
process involving the passage of raw water without coagulants
through a bed of fine sand at low velocity (generally less
than 0.4 m/h approach velocity) resulting in substantial
particulate removal by physical and biological mechanisms and
changes in chemical parameters by biological actions. It can
be used for raw waters of low turbidity, color, and algae
content without additional pretreatment; but is sometimes
preceded by other treatment steps to extend the range of
acceptable raw waters.
Direct filtration should be defined as a clarification
process which includes coagulation, filtration, and may or
may not include flocculation. It is appropriate only for low
turbidity, low color, and low algae waters.
Diatomaceous earth filtration can be defined as a fil-
tration technique appropriate for waters with low turbidity,
color, and algae which meets the following requirements:
a) a precoat cake of diatomaceous earth filter media is
deposited on a support membrane (septum), b) while the water
is filtered by passing through the cake on the septum, addi-
tional filter media, known as body feed is continuously added
to the feed water, in order to maintain the permeability of
the filter cake, c) the filter media may or may not be coated
with alum or coagulated substrate.
Recommended variance criteria for mandatory filtration
of surface waters include the following:
Watershed Control
an annual on-site survey is performed to identify and
satisfactorily control potential sources of contami-
nation to prevent significant adverse impacts upon
water quality.
an active program is maintained to monitor mammals
ich can be potentially infected with Giardia within
e watershed.		

Waterborne Disease Outbreak History
- the system has no history of waterborne disease out-
breaks with the current treatment system.
an active waterborne disease monitoring program is
instituted and maintained.
Source Water Monitoring
raw water does not exceed xx fecal coliforms per 100
ml. (monitoring frequency to be determined by state),
Limits of yy total coliforms and fecal Streptococcus
may also be appropriate.
Finished Water Monitoring
-	the water is routinely monitored for heterotrophic
plate count (HPC) before it enters the distribution
turbidity at the point of entry into the distribution
system is monitored and recorded continuously.
exiting turbidity MCLs are met consistently.
system has redundant disinfection units.
daily coliform monitoring is required at entry to the
distribution system.
Distribution System Monitoring
-	the coliform MCLs are met consistently.
the HPC is routinely monitored in the distribution
a disinfectant residual is maintained throughout the
State Requirements
all additional requirements established by the states
are met consistently.
Under existing regulations, a utility can obtain a
"variance" from a treatment requirement by: a) demonstrating
that such treatment is not necessary to protect the health of
the community, or b) demonstrating and using an alternative
treatment technique that can provide an equivalent degree of
safety. The existing concept of "variance" has negative
connotations (associated with variances for not meeting MCLs)
and requires a lengthy processing procedure. For "variances"

to treatment requirements a new terminology is needed to
avoid this stigma. A new procedure needs to be established
that will give states greater flexibility and facility in
processing "variances."
"Variance," as defined in (a), should be substituted
with "Demonstration of Equivalency" (DOE) or other term. The
process for implementing DOEs is recommended as follows: EPA
specifies the criteria for DOE in the regulations. The states
adopt these criteria and decide on DOE on a case by case
basis. In accordance with the normal state process the public
is given notice by the utility of its intention and rationale
for receiving a DOE. The public is given an opportunity to
comment on this (depending on participation of commentors, a
public hearing may be held). Following the public comment
period, one-time notification of the decision is made to the
public. DOE approval is included in the permit issued by the
state. The procedure for obtaining a DOE is not repeated.
Rather, the DOE is renewed periodically (without public
participation) as part of the permit renewal.
"Variance," as defined in (b) above should be substituted
with "Equivalent Technology Demonstration" (ET) or other term.
The process for implementing ETs, which should be specified
in the regulations, is recommended as follows. EPA includes
the language in the regulations (no "criteria" needed) that
will allow states to determine equivalent treatment tech-
nologies and to allow their usage in public systems. The
state determines the data needed to support the demonstration.
The state reviews the data submitted by the utility and
decides if equivalency of treatment exists. Public partici-
pation in the process is optional as determined by the state.
If equivalency of treatment is determined by the state it
issues approval through a permit (approval may be provisional
based on meeting certain criteria on an ongoing basis).

Viruses and Giardia Monitoring
Is it technically or economically feasible to initiate
monitoring of viruses and Giardia in drinking water on a
routine basis?
Monitoring is currently being conducted by some State and
large utility laboratories. Estimated cost per sample is
about $200. Colorado has been able to "have analyses conducted
at $50/sample through a university laboratory but high frequency
monitoring is not available.
The procedure requires about one hour to filter the water
and to extract the filtrate with distilled water, overnight
sedimentation with refrigeration (or centrifugation), and 1 to
2 "hours for preparation (depending on the quality of the water)
and microscopic analysis of the slide.
In-house monitoring costs can be reduced greatly if the
staff is trained to microscopically identify the cysts. If
many samples are analyzed consecutively/ eye fatigue can be a
problem. In such cases two analysts may be required to avoid
this limiting factor.
The major shortcoming of the analysis is not being able
to readily determine the viability of the cysts that are
identified to be present. Also, one cannot tell whether the
cyst originated from an animal or human host based on its
morphology. Another concern is not knowing the distribution
of the cysts in the water being analyzed. In some cases
cysts appear in bunches and in others not at all.
According to EPA s-tudies, samples spiked with cysts show
60-70% recovery on membrane filters and that 1-5 cysts per
liter are detectable.
Because of the difficulty and uncertainties in monitoring,
Colorado is taking a treatment approach. Utilities which do
not practice filtration and which detect cysts in their raw
water are being asked to super-chlorinate with a free chlorine
residual of 3 ppm for 1 hour. If this treatment is not prac-
ticed, a boiled water decree is issued.
The limitations on the sensitivity of the analysis makes
it difficult to monitor the effect of treatment change (e.g.,
the addition of filtration) on removing cysts. Studies using
bench scale filtration, using raw water spiked with cysts,

have shown 3-4 orders of magnitude of cyst removal and qood
correlation between turbidity removal, cyst removal, and
virus removal. However, we know from laboratory and pilot
scale work that treatment works and that its effectiveness
for removing cysts can be monitored using turbidity as an
Monitoring conducted in California in one system using
direct filtration (no pre-treatment) has shown consistently
higher cyst counts in the filtered water versus the raw
water. It appears that the cysts are coming through in bunches
rather than showing a homogenous distribution. These apparent
anomalies have raised questions on the recovery and reliability
of the analytical methodology and on the distribution of cysts
in water.
Use of polyelectrolytes can be critical on the effective-
ness of direct filtration for removing cysts.
Some raw waters with turbidities as low as 0.1 NTU have
been shown to contain cysts. It is difficult to convince
people with such low turbidity waters that they need to
filter. Can percent turbidity removal be a good indicator of
cyst removal in such low turbid waters?
In pilot scale work that we have conducted (Iowa State
University), good correlations exist between particle and
turbidity removal and cyst-sized particle removal using
direct filtration (with coagulant). These correlations have
been shown to be good in waters from 0.35 NTU (winter) to
0.1 NTU, but not for raw water turbidities less than 0.1 NTU.
The consensus of the group is that Giardia monitoring is
useful for some agencies and large utilities but not for
general use because of costs and technical difficulties.
Regarding monitoring for viruses, participants of the
group did not feel they had enough experience to adequately
discuss this issue. The consensus was that such monitoring
was not appropriate for general use because of costs and
technical difficulties with the analyses.
Disinfection Treatment Techniques
What is the definition of "disinfection" for purposes of
specifying approprate techniques in the regulations?
If we are to define disinfection in terms of concentration
dosages and contact times, the concept of "contact time"
needs elaboration. Reactors should be defined as plug flow

or mixed. Contact time in the pipeline should be "true,"
actual time it takes water with mean disinfectant residual
(concentration considered) to reach the first customer.
Disinfection criteria should not be written directly into
the regulation. States should be given flexibility to determine
these requirements. (This comment was supported by a majority
of participants in the work group.)
Disinfection should be defined in terms of organism
removal. The group consensus felt that disinfection should
be defined as "an organism destruction process which aims to
eliminate pathogenic organisms in drinking water" and that
all surface waters (community and non-community) should at
least be required to disinfect.
Guidelines should be provided by EPA to States regarding
C x T values (disinfectant concentration times contact time).
These values should be given for a range of pHs, temperatures,
and disinfectants (chlorine, chloramines, ozone, and chlorine
dioxide) for the different organisms of concern (bacteria,
viruses, and cysts) to provide guidelines for State disinfec-
tion requirements.
In EPA's proposed variance criteria (from worksheets that
were distributed), waters with pH >8.0 are not considered.
In its development of guidelines, disinfection criteria
should be provided for all possible pH conditions.
Without the availability of data it is inappropriate for
the group to recommend C x T values to EPA.
It may be more appropriate to define disinfection in terms
of disinfectant species (e.g., 0C1~ versus HOC1 for chlorine,
and NHCI2 versus NH2CI for chloramines). This would eliminate
the need for defining C & T conditions in terms of pH.
Should disinfection criteria be the same for unfiltered
versus filtered waters? Criteria should depend on the'organisms
of concern (e.g., bacteria, virus, Giardia) and not on the
prior existing treatment. It may be appropriate for some
systems to only use disinfection to kill potential Giardia cysts
present in raw water. In such cases disinfection criteria
should be much more stringent than for systems with no concern
for Giardia.
In its guidelines EPA should provide criteria for using
high pHs (e.g., lime softening) as a disinfection alternative
and for using combinations of disinfectants.
Iodinators are often used in non-community systems.
Iodination should thus be considered a viable alternative
disinfectant for such systems.

What monitoring requirements should be specified in the
disinfection regulatory requirement?
Monitoring requirements should not be specified in the
Federal regulations. They should be given within the State
Operation and Maintenance Guidelines.
Continuous monitoring is ideal but should not be mandated.
Monitoring requirements should depend on the reliability
of the disinfection system. If a back-up system is online
for automatic startup in case of system failure then the
required monitoring frequency can go down.
Monitoring frequency should depend on water quality and
objectives of disinfection. If the system depends solely on
disinfection for Giardia protection (i.e., has no filtration),
a very high monitoring frequency should be required.
Colorado allows decreased turbidity monitoring for
non-community systems which filter.
In what circumstances might application of a certain type
of disinfection be inappropriate?
High chlorine dosages and log contact times to address
the problem for potential Giardia presence (systems not
having filtration) may be inappropriate because of conflict
with the THM regulation.
Chloramines are probably inadequate for destroying the
Giardia cyst.
Disinfection with chloramines is a concern for dialysis
Should there be a requirement for maintenance of a
residual to the system?
Residual requirements are not enough. A cross-connection
control program should also be required.

In New York State mandatory disinfection is required for
all waters. Exceptions called "waivers" (not variances or
exemptions) are allowed to systems with active cross-connection
control programs and which conduct more monitoring than is
ordinarily required.
The benefits of maintaining a residual is not obvious.
Systems with residuals do not ensure the absence of coll forms.
In the event of gross contamination in the distribution line
a low residual concentration, as normally practiced, will not
adequately address such contamination.
We (New Haven Regional Water Authority) have a system in
which we maintain a free chlorine residual of 6 mg/1 in the
distribution system but still detect coliforms.
Nutrients from finished water entering the distribution
system are instrumental in causing growth problems. Nutrient
presence can be indicated by TOC. The requirement to maintain
a disinfectant residual could be dependent on some TOC concen-
tration, possibly 1 mg/L.
Basing a residual requirement on a TOC concentration in
finished water cannot be justified at this time because data
is not available to support this. Prevention of regrowth
should only be considered one benefit of residual disinfection.
Other benefits include limited protection from cross-connection
contamination and providing a marker for potential contamination
(i.e., absence of residual at particular location indicates
EPA should not require a specific residual concentration
to be maintained but leave this to State discretion.
Buildup of organisms in pipeline of new buildings can be
very significant. Standard plate counts as high as 6 x 10®/100
ml have been recorded in the tap water from such buildings.
Residual disinfection would mitigate such problems.
The majority of the group wanted a residual requirement
to apply to all distribution systems. All people felt
that variances should not apply to surface water systems.
A minority felt this requirement could take exception to some
groundwater systems, depending upon local conditions. Among
this minority, some felt that the concept of variance would
not be feasible to implement for small community and non-
community systems. Because of the limited resources available
to deal with such systems, States should be allowed discretion
to grant exclusions or waivers without having to go through a
complicated "variance" process. Suggested variance criteria,
among the minority that thought this concept to be appropriate,
included no reported outbreaks of waterborne disease, monitoring
history for coliform with no MCL violations, no history of high
heterotrophic plate counts, sanitary survey, and monitoring
for total coliform of no less than 5 samples per month.

Filtration Treatment Technique
What is the definition of filtration for purposes of
specifying appropriate techniques in the regulation?
Defining filtration by performance criteria using an
indicator has many limitations. Work done by the University
of Missouri indicates that turbidity removal is not a good
indicator for bacteria removal (see attachment for minority
statement). Based on microscopic bacterial counting, bacteria
in water treatment plant water tend to be much more free
floating than being attached to particulate matter. Little
correlation has been established between the density of
microorganisms and turbidity. HPC and direct microscopic
counting should be considered better indicators than turbidity
for monitoring pathogen removal.
No good correlation has yet been shown to exist between
HPC or bacteria counts and pathogen presence. Bacterial
counting or HPC may be a good indicator for monitoring treat-
ment plant performance but there is no data to demonstrate
this. No correlations have yet been demonstrated between
these removals and pathogen removal.
Measurement of the particulate fraction of TOC is a better
indicator than turbidity for organic removal (chlorine demand)
from water.
Good correlations have been demonstrated between turbidity
removal and Giardia cyst removal by filtration. Since this
has been demonstrated and no correlations have yet been shown
between HPC and bacteria count removals versus pathogen
removal, turbidity removal should be considered, until proven
otherwise, as a better indicator of treatment performance.
Design and operational criteria should not be incorporated
into the regulation. EPA should only specify which treatment
technologies are permitted under the requirement. States,
based on guidance from EPA, should be given the responsibility
for defining design and operating parameters.
Which filtration technologies are appropriate? How should
these be defined?
Comments t
The consensus of the group was that conventional treatment
of coagulation, flocculation, sedimentation, and filtration

(since rapid sand filtration by itself is not effective),
direct filtration, slow sand filtration, and diatomaceous
earth filtration should be considered appropriate to fill the
treatment requirement. Recommended definitions of these
technologies are given in the summary.
Definition in the regulations of the treatment technolo-
gies should not include operating or performance criteria.
These criteria should be defined by the States in their
issuance of permits to the public utilities. EPA should
provide guidance to the States for issuing such criteria.
Allowance of a monthly average turbidity in the effluent
(e.g., 0.2 NTU for conventional treatment) gives too much
flexibility to the plant operator. This allows daily higher
turbidities and invites trouble. A daily upper limit is
Some plants use only polymers for coagulation, no aluminum
or ferric salts. This should be allowed by EPA in their guide-
lines for pre-treatment.
Recommended guidelines for conventional treatment (coagu-
lation, flocculation, sedimentation, filtration) and direct
filtration are as follows. Finished water should meet a
turbidity goal (e.g., 0.2 NTU). Effective pre-treatment must
be achieved since it is essential for filtration. Goals for
percent reductions in turbidity should be recommended by EPA
for different raw water turbidity levels (sliding scale,
e.g., for NTU >1 greater than 80-90%, for NTU <1 greater than
60-70% might be required). States should be given flexibility
in regulations to use % turbidity reduction and an MCL or
lower turbidity goal (e.g., 0.1 NTU) in issuing permits.
Requiring % turbidity reduction in addition to meeting an MCL
is essential for ensuring effective filtration in low turbidity
Some utilities are now reducing their coagulant dosage to
the point of just meeting the turbidity MCL. To discourage
this practice a % turbidity reduction requirement should be
imposed by the States. This additional requirement is espe-
cially critical for systems with low turbidity raw waters.
States should have permit regulations to ensure that the
treatment plant is run properly. Operation and maintenance
plans should be included to ensure the plant is operating
according to how it was designed to be run.
All States may not have the resources (and expertise) to
issue permits specifying performance, operation and maintenance
criteria for the different technologies under consideration.
It is unreasonable to expect all utilities to meet
performance criteria 100% of the time on a continuing basis.

Some leeway is needed such as only requiring compliance 95%
of the time.
The majority felt the effluent turbidity level should be
consistent with the MCL, which should be the same for all
technologies. Exceedance of the turbidity MCL could be
allowed if the system demonstrates, through the variance
procedure, equivalent pathogen removal to that achieved in
conventional treatment. Discontent may arise between adjacent
towns if they have different standards of performance
because of having different treatments.
Slow sand filtration is a biological process which kills
pathogens and, therefore, should be considered differently
from the other technologies. Effective removal of microorgan-
isms can be achieved while the effluent still has a relatively
high turbidity. Data have shown 2 to 4 orders of magnitude
reductions of microorganisms in waters being reduced in
turbidity from 8 to 3 NTU.
If turbidity effluent requirements are low for slow sand
filtration, e.g., at the 0.2 NTU proposed level for conven-
tional treatment, many systems will not be able to apply this
technology even though effective pathogen removal could be
expected. Turbidity requirements for indicating slow sand
filtration performance should thus be different than for other
Because of their inferior performance during the ripening
period, slow sand filters should be kept off-line during this
time. States need to include specific operating criteria in
their permits and/or regulations to cover such concerns.
The majority of the group felt that for diatomaceous earth
filtration, the effluent turbidity standard should be the same
as for conventional treatment. No design or operation criteria
should be included in the regulations.
Cartridge filters should be considered for inclusion among
those technologies allowed in the treatment regulation. This
technology offers an economical way for noncommunity and very
small systems to meet the regulation.
The great variety of cartridge filters with differing
specifications on the market (some perform well and others
poorly) makes it inappropriate to define them as one technology.
Judgment on their suitability needs to be used on a case by
case basis.
The existing variance mechanisms under which a cartridge
filter might be allowed is too difficult for small systems
and non-community systems.

A testing program needs to be defined under which such
technologies can be evaluated.
The State of Pennsylvania recognizes pomt-of-use filters
that have been tested and certified by the National Sanitation
Foundation as an acceptable filtration technology for control-
ling Giardia.
Cartridge filters are effective for removing protozoans.
Their effectiveness for virus removal has not yet been
Technical options, other than going through the variance
procedure, are essential to facilitate non-community and very
small systems compliance with the treatment regulation.
Quest ion:
What criteria would have
States that filtration is not
(i.e., that the system can ob
forego such treatment)?
to be met to demonstrate to
needed for surface waters
ain a variance allowing it to
Group consensus variance criteria are given in the summary.
The following comments pertain to the development of these
An annual sanitary survey should be required to identify
potential sources of contamination. It is inappropriate to
require that a system's watershed contain no sewage discharges,
septic tank fields, or sanitary landfills. Many watersheds
are very large and the distance between such potential sources
of contamination is often great enough to allow sufficient die-
off of pathogens before they would reach the raw water influent
at the treatment plant. It is sufficient to require, through
meeting criteria within the sanitary survey, that potential
sources of contamination are controlled to the extent that they
do not impact adversely upon the quality of the raw water to
the plant.
Systems with watersheds containing significant human
habitat recreational activities should not be allowed to
obtain a variance.
Mammals which can potentially be infected with Giardia
should not be required to be eliminated from the watershed.
Such mammals, however, should be monitored to assure they are
not carriers of Giardia.
Systems with raw water exceeding a fecal coliform con-
centration (concentration and monitoring frequency to be
determined by EPA) should not be considered eligible for

receiving a variance. Cutoff points for fecal streptococci
and total coliform concentrations should also be considered
for inclusion in the variance criteria.
Giardia monitoring of raw water by utilities would be
greatly limited because of costs. Since the significance of
no Giardia being detected among limited samples is questionable,
a Giardia monitoring requirement for variance eligibility is
i nappropriate.
Heterotrophic plate count should be monitored routinely in
the finished water before it enters the distribution system.
Permissable counts and the monitoring frequency should be
determined by the State. The methodology for measuring HPC
will greatly affect the sensitivity of the analyses and should
be considered by the State when setting the criteria.
Turbidity should be monitored continuously at the entry
into the distribution system. One half of the participants
felt the turbidity should never exceed 1.0 NTU. The other
participants felt that the higher turbidity levels should be
allowed, depending upon other data and system considerations.
Disinfection should be practiced with greater reliability.
Backup equipment of all components should be in place to ensure
continuous disinfection in case the primary disinfection system
A minimum of daily coliform monitoring at the entry to
the distribution system should be required. Coliform measure-
ments at points in the distribution system should consistently
meet the MCLs. Heterotrophic plate count should routinely be
monitored and a disinfectant residual should be maintained
throughout the distribution system.
What variance criteria are appropriate for allowing a
utility to demonstrate to the State an equivalent technology?
The existing process for obtaining a variance for equiva-
lent technology needs to be greatly simplified. States
should be.allowed to determine which systems qualify for a
variance. A recommended process for how this could be done
is given in the summary. The issue of a simplified variance
proceeding is especially critical for non-community systems,
many of which would like to use cartridge filtration and
iodination for disinfection.

ATTACHMENT; Statement from John T. O'Connor, Chairman, and
C.W. LaPierre > Professor of Civil Engineering,
University of Missourl-Colurapia
"The Microscope versus the Turbidimeter for the Evaluation
of LMicrobiological Quality of Water — A Dissenting View"
Turbidity is a primary microbiological drinking water
standard for surface waters. It is ranked up there with
cadmium and 2,4 D as primary, health-related concerns in
drinking water. Presumably, if drinking water contains
turbidity in excess of 1 NTU, there is reason to question its
microbiological quality. Oddly, turbidity has never been
quantified or characterized. In some waters, turbidity is
caused by silt and clay while in others, it is due exclusively
to iron oxides. In softened waters, the turbidity is most
often due to calcium carbonate. In coagulated surface waters,
turbidity may be due to the post-precipitation of aluminum
oxide. In wastewater, it may be due to microorganisms and
organic matter. Simply put, 1 NTU of turbidity is not equal
to 1 NTU of turbidity. Turbidity is different in all waters.
Some very low turbidity waters have recently created a regu-
latory crisis since they bear the disease, Giardiasis, despite
being cold and exceptionally clear.
There is good reason, from an aesthetic viewpoint, to
have low turbidity drinking water. However, why is turbidity
presently employed as an indicator of microbiological quality?
The reasons have never been definitively stated. However,
the arguments for using turbidity, instead of microorganisms,
to evaluate the microbiological quality of drinking waters
are as follows:
1.	Turbidity is related to the microbiological content
of water. Removal of turbidity results in propor-
tionate removal of the microorganisms present in raw
water sources since most organisms are attached to
the solids which comprise raw water turbidity.
2.	The removal of turbidity reduces the chlorine demand
of treated drinking water.
3.	Turbidity interferes with the coliform determination.
Both the published literature and recent research
performed at the University of Missouri-Columbia, Department
of Civil Engineering, contradicts the first two of these
First, microscopic examination of raw waters shows that
most microorganisms themselves contribute little to the
measured turbidity value. There would need to be in excess of

1010 cells per liter before bacteria would contribute 1 NTU
to water. Secondly, conventional water coagulation, sedi-
mentation and filtration removes "turbidity" very effectively,
but may be far less effective in the removal of bacterial
cells. In other words, it is possible to produce filtered
water which is visually very clear, but which contains a
significant portion of the microorganisms which were initially
present in the raw water. This failure to remove microorganisms
effectively is of particular concern in the winter when water
temperatures are low, since both coagulation and filtration
processes are less effective at low temperature.
Still more recent research indicates that turbidity
contributes only a small share of the chlorine demand. Most
chlorine depletion occurs within distribution piping. It is
true, however, that turbidity interferes with the enumeration
of coliform organisms. However, this last argument would
hardly appear to be adequate justification for using turbidity
as a primary microbiological standard.
The support for the use of turbidity is based on tradition
and ease of measurement. Qualitatively, turbidity can be
measured by eye.
An alternate method of evaluating the effectiveness of
water treatment processes in removing microorganisms is the
direct microscopic count. Currently viewed as expensive and
impractical, the direct count requires the use of a microscope
for counting bacterial cells. In soliciting research proposals,
the AWWA Research Foundation has precluded the use of the
direct microscopic count from use in studies of microbial
regrowth in water distribution systems since the method is
not in common use in the waterworks industry.
Even if the dialog over the use of turbidity as a primary
microbiological drinking water standard is not joined, the
issue of the effectiveness of removal of microorganisms during
water treatment may soon be resolved. Although not in the
waterworks literature, the methods for enumerating total
bacterial populations have been developed and described. It
would be truly ironic if data on removal of microorganisms
from drinking water emerge from high school science fairs
before before local waterworks utilities are aware of the
effectiveness of their treatment processes.