echnologies
and
Costs
for
Control
of
Microbial
echnologies
and
Costs
for
Control
of
Microbial
Contaminants
and
Disinfection
Byproducts
Contaminants
and
Disinfection
Byproducts
T
PREPARED
FOR:

U.
S.
ENVIRONMENTAL
PROTECTION
AGENCY
Office
of
Ground
Water
and
Drinking
Water
PREPARED
BY:

THE
CADMUS
GROUP,
INC.
1901
North
Fort
Myer
Drive
Suite
900
Arlington,
VA
22209
MALCOLM
PIRNIE,
INC.
1900
Polaris
Parkway
Suite
200
Columbus,
OH
43240
US
EPA
CONTRACT:
68­
C­
02­
026
Work
Assignment:
1­
21
June
2003
Technologies
and
Costs
for
Control
of
Microbial
Contaminants
and
Disinfection
Byproducts
June
2003
i
Table
of
Contents
Chapter
1:
Introduction
1.1
Purpose
of
Technology
and
Cost
Document
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
1­
1
1.2
Existing
Regulations
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
1­
2
1.2.1
Surface
Water
Treatment
Rule
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
1­
2
1.2.2
Information
Collection
Rule
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
1­
3
1.2.3
Interim
Enhanced
Surface
Water
Treatment
Rule
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
1­
3
1.2.4
Stage
1
Disinfectants
and
Disinfection
Byproducts
Rule
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
1­
4
1.2.5
Long
Term
1
Enhanced
Surface
Water
Treatment
Rule
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
1­
4
1.2.6
Filter
Backwash
Recycling
Rule
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
1­
5
1.3
Public
Health
Concerns
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
1­
5
1.3.1
Pathogenic
Microorganisms
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
1­
5
1.3.2
Disinfectants/
Disinfection
Byproducts
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
1­
6
1.4
Proposed
Regulations
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
1­
6
1.4.1
Long
Term
2
Enhanced
Surface
Water
Treatment
Rule
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
1­
6
1.4.2
Stage
2
Disinfectants/
Disinfection
Byproducts
Rule
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
1­
7
1.5
Technologies
Evaluated
for
the
Control
of
Pathogens
and
Disinfection
Byproducts
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
1­
7
1.6
Document
Organization
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
1­
9
Chapter
2:
Technologies
for
DBP
and
Microbial
Contaminant
Control
2.1
Introduction
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
2­
1
2.2
Alternative
Disinfection
Strategies
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
2­
2
2.2.1
Chloramination
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
2­
2
2.2.1.1
Efficacy
Against
Pathogens
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
2­
2
2.2.1.2
DBP
Formation
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
2­
3
2.2.1.3
Factors
Affecting
Performance
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
2­
4
2.2.2
Chlorine
Dioxide
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
2­
6
2.2.2.1
Efficacy
Against
Pathogens
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
2­
6
2.2.2.2
DBP
Formation
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
2­
7
2.2.2.3
Factors
Affecting
Performance
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
2­
8
2.2.3
Ultraviolet
Light
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
2­
8
2.2.3.1
Efficacy
Against
Pathogens
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
2­
10
2.2.3.2
DBP
Formation
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
2­
12
2.2.3.3
Factors
Affecting
Performance
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
2­
12
2.2.4
Ozone
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
2­
14
2.2.4.1
Efficacy
Against
Pathogens
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
2­
16
2.2.4.2
DBP
Formation
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
2­
18
2.2.4.3
Factors
Affecting
Performance
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
2­
20
Technologies
and
Costs
for
Control
of
Microbial
Contaminants
and
Disinfection
Byproducts
June
2003
ii
2.2.5
Microfiltration
and
Ultrafiltration
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
2­
21
2.2.5.1
Efficacy
Against
Pathogens
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
2­
22
2.2.5.2
DBP
Formation
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
2­
26
2.2.5.3
Factors
Affecting
Performance
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
2­
26
2.2.6
Bag
and
Cartridge
Filtration
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
2­
27
2.2.6.1
Efficacy
Against
Pathogens
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
2­
28
2.2.6.2
Factors
Affecting
Performance
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
2­
30
2.2.7
Bank
Filtration
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
2­
31
2.2.7.1
Efficacy
Against
Pathogens
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
2­
31
2.2.7.2
Factors
Affecting
Performance
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
2­
32
2.2.8
Second
Stage
Filtration
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
2­
32
2.2.8.1
Efficacy
Against
Pathogens
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
2­
33
2.2.8.2
Factors
Affecting
Performance
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
2­
34
Filter
Type
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
2­
34
Filter
Media
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
2­
34
Filter
Hydraulics
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
2­
35
2.2.9
Pre­
Sedimentation
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
2­
35
2.2.9.1
Efficacy
Against
Pathogens
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
2­
35
2.2.9.2
Factors
Affecting
Performance
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
2­
36
Short
Circuiting
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
2­
36
Coagulant
Dose
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
2­
36
2.2.10
Watershed
Control
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
2­
36
2.2.10.1
Efficacy
Against
Pathogens
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
2­
36
2.2.10.2
Factors
Affecting
Performance
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
2­
38
2.2.11
Combined
Filter
Performance
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
2­
38
2.2.11.1
Efficacy
Against
Pathogens
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
2­
38
2.2.11.2
Factors
Affecting
Performance
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
2­
39
Coagulant
Dose
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
2­
39
Filter
Ripening
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
2­
39
Filter
Breakthrough
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
2­
40
Filtration
Rate
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
2­
40
Backwashing
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
2­
40
2.3
DBP
Precursor
Removal
Strategies
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
2­
40
2.3.1
Granular
Activated
Carbon
Adsorption
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
2­
40
2.3.1.1
DBP
Precursor
Removal
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
2­
42
2.3.1.2
Factors
Affecting
Performance
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
2­
42
2.3.2
Nanofiltration
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
2­
44
2.3.2.1
Efficacy
Against
Pathogens
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
2­
45
2.3.2.2
DBP
Precursor
Removal
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
2­
47
2.3.2.3
Factors
Affecting
Performance
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
2­
49
Technologies
and
Costs
for
Control
of
Microbial
Contaminants
and
Disinfection
Byproducts
June
2003
iii
Chapter
3:
Technology
Design
and
Criteria
3.1
Introduction
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
3­
1
3.2
Base
Treatment
Plant
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
3­
1
3.3
Alternative
Disinfection
Strategies
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
3­
2
3.3.1
Chloramination
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
3­
2
3.3.2
Chlorine
Dioxide
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
3­
3
3.3.3
Ultraviolet
Light
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
3­
5
3.3.4
Ozone
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
3­
7
3.3.5
Microfiltration
and
Ultrafiltration
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
3­
8
3.3.6
Bag
and
Cartridge
Filtration
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
3­
10
3.3.7
Bank
Filtration
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
3­
11
3.3.8
Second
Stage
Filtration
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
3­
11
3.3.9
Pre­
Sedimentation
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
3­
12
3.3.10
Watershed
Control
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
3­
12
3.3.11
Combined
Filter
Performance
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
3­
14
Installing
Backwash
Water
Polymer/
Coagulant
Feed
Capability
.
.
.
.
.
.
.
.
.
.
.
.
.
.
3­
15
Installing
Additional
Coagulant
Feed
Points
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
3­
15
Adding
Filter
Media
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
3­
15
Adding
Filter
to
Waste
Capabilities
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
3­
16
Installing
or
Replacing
Filter
Rate­
of­
Flow
Controllers
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
3­
16
Increasing
Plant
Staffing
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
3­
16
Increasing
Staff
Qualifications
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
3­
16
Purchasing
or
Replacing
Bench­
Top
Turbidimeters
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
3­
16
Purchasing
or
Replacing
Jar
Testing
Apparatus
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
3­
16
Purchasing
or
Replacing
a
Particle
Counter
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
3­
16
Staff
Training
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
3­
17
3.4
DBP
Precursor
Removal
Technologies
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
3­
17
3.4.1
Granular
Activated
Carbon
Adsorption
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
3­
17
3.4.2
Nanofiltration
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
3­
19
Chapter
4:
Technology
Costs
4.1
Introduction
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
1
4.2
Approach
for
Cost
Estimates
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
2
4.2.1
Cost
Components
and
Capital
Cost
Multipliers
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
3
O&
M
Costs
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
4
4.2.2
Cost
Indices
and
Unit
Cost
Inputs
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
5
4.2.3
Cost
Build­
up
Approach
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
7
4.2.4
Lump
Sum
Estimates
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
7
4.2.5
Cost
Modeling
Approach
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
7
4.2.5.1
VSS
Model
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
8
Technologies
and
Costs
for
Control
of
Microbial
Contaminants
and
Disinfection
Byproducts
June
2003
iv
4.2.5.2
Water
Model
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
8
4.2.5.3
W/
W
Cost
Model
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
9
4.2.6
Indirect
Capital
Costs
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
9
Permitting
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
9
Piloting
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
10
Land
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
10
Housing
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
11
Operator
Training
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
12
Public
Education
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
12
4.3
Estimation
of
Annualized
Costs
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
12
4.4
Alternative
Disinfection
Strategies
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
13
4.4.1
Chloramination
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
13
4.4.1.1
Summary
of
Chloramine
Capital
Cost
Assumptions
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
14
Process
Costs
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
14
Capital
Cost
Multipliers
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
15
Indirect
Capital
Costs
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
15
4.4.1.2
Summary
of
Chloramine
O&
M
Cost
Assumptions
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
15
4.4.2
Chlorine
Dioxide
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
21
4.4.2.1
Summary
of
Chlorine
Dioxide
Capital
Cost
Assumptions
.
.
.
.
.
.
.
.
.
.
.
4­
21
Process
Costs
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
21
Feed
Equipment
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
21
Capital
Cost
Multipliers
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
22
Indirect
Capital
Costs
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
22
4.4.2.2
Summary
of
Chlorine
Dioxide
O&
M
Cost
Assumptions
.
.
.
.
.
.
.
.
.
.
.
.
4­
23
Feed
Equipment
(
systems
smaller
than
2.0
mgd)
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
23
Chemical
Usage
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
23
Materials,
Electricity,
and
Labor
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
23
4.4.3
Ultraviolet
Light
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
27
4.4.3.1
Summary
of
UV
Disinfection
Capital
Cost
Assumptions
.
.
.
.
.
.
.
.
.
.
.
.
4­
27
Process
Costs
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
27
Capital
Cost
Multipliers
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
28
Indirect
Capital
Costs
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
28
4.4.3.2
Summary
of
UV
Disinfection
O&
M
Cost
Assumptions
.
.
.
.
.
.
.
.
.
.
.
.
4­
28
4.4.4
Ozone
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
38
4.4.4.1
Summary
of
Ozonation
Capital
Cost
Assumptions
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
38
Process
Costs
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
38
pH
Adjustment
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
43
Capital
Cost
Multipliers
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
43
Indirect
Capital
Costs
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
43
4.4.4.2
Summary
of
Ozonation
O&
M
Cost
Assumptions
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
44
4.4.5
Microfiltration
and
Ultrafiltration
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
52
Technologies
and
Costs
for
Control
of
Microbial
Contaminants
and
Disinfection
Byproducts
June
2003
v
4.4.5.1
Summary
of
MF/
UF
Capital
Cost
Assumptions
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
52
Process
Costs
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
52
Capital
Cost
Multipliers
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
55
Indirect
Capital
Costs
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
55
4.4.5.2
Summary
of
MF/
UF
O&
M
Cost
Assumptions
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
58
Membrane
Replacement
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
58
Performance
Monitoring
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
59
Clean­
in­
Place
Chemicals
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
59
Materials
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
59
Power
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
60
Labor
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
60
POTW
Surcharge
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
61
4.4.6
Bag
and
Cartridge
Filtration
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
64
4.4.6.1
Summary
of
Bag
and
Cartridge
Filter
Capital
Cost
Assumptions
.
.
.
.
.
.
4­
64
Process
Costs
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
64
Capital
Cost
Multipliers
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
66
Indirect
Capital
Costs
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
66
4.4.6.2
Summary
of
Bag
and
Cartridge
Filter
O&
M
Cost
Assumptions
.
.
.
.
.
.
4­
66
Bag
and
Cartridge
Replacement
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
66
Power
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
67
Labor
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
67
4.4.7
Bank
Filtration
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
70
4.4.8
Second
Stage
Filtration
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
70
4.4.9
Pre­
Sedimentation
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
71
4.4.10
Watershed
Control
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
71
4.4.11
Combined
Filter
Performance
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
72
4.4.11.1
Installing
Backwash
Polymer
Feed
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
75
4.4.11.2
Installing
Additional
Coagulant
Feed
Points
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
75
4.4.11.3
Filter
Media
Addition
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
75
4.4.11.4
Filter
to
Waste
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
76
4.4.11.5
Filter
Rate­
of­
Flow
Controller
Replacement
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
77
4.4.11.6
Increase
Plant
Staffing
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
78
4.4.11.7
Update
Plant
Staff
Qualifications
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
78
4.4.11.8
Purchase
Turbidimeter
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
78
4.4.11.9
Purchase
Jar
Test
Apparatus
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
79
4.4.11.10
Purchase
Particle
Counters
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
79
4.4.11.11
Staff
Training
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
79
4.4.11.12
Average
Plant
Cost
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
81
4.5
DBP
Precursor
and
Microbial
Removal
Technologies
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
83
4.5.1
Granular
Activated
Carbon
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
83
4.5.1.1
Summary
of
GAC
Capital
Cost
Assumptions
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
84
Technologies
and
Costs
for
Control
of
Microbial
Contaminants
and
Disinfection
Byproducts
June
2003
vi
Process
Costs
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
84
Capital
Cost
Multipliers
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
86
Indirect
Capital
Costs
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
86
4.5.1.2
Summary
of
GAC
O&
M
Cost
Assumptions
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
88
GAC
Usage
Rate
and
Replacement
Costs
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
88
Labor
Costs
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
89
Natural
Gas
Costs
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
91
Performance
Monitoring
Costs
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
91
Maintenance
Materials
Costs
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
92
VSS
Model
Costs
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
93
4.5.2
Nanofiltration
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
100
4.5.2.1
Summary
of
NF
Capital
Cost
Assumptions
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
100
Process
Costs
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
100
Capital
Cost
Multipliers
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
102
Indirect
Capital
Costs
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
102
4.5.2.2
Summary
of
NF
O&
M
Cost
Assumptions
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
104
Clean­
in­
Place
Chemicals
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
104
Acid/
Anti­
Scalant
and
Caustic
Chemicals
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
104
NF
Membrane
Replacement
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
105
Cartridge
Filter
Replacement
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
105
Repair,
Maintenance
and
Replacement
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
105
Performance
Monitoring
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
106
Power
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
106
Labor
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
106
POTW
Surcharge
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
107
4.6
Annualized
Costs
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
110
Chapter
5:
References
Technologies
and
Costs
for
Control
of
Microbial
Contaminants
and
Disinfection
Byproducts
June
2003
vii
List
of
Exhibits
Exhibit
2.1:
Comparison
of
CT
Values
for
Free
Chlorine
and
Chloramine
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
2­
3
Exhibit
2.2:
Comparison
of
CT
Values
for
Free
Chlorine
and
Chlorine
Dioxide
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
2­
6
Exhibit
2.3:
Summary
of
Chlorine
Dioxide
CT
Values
for
Cryptosporidium
Inactivation
.
.
.
.
.
.
.
.
2­
7
Exhibit
2.4:
Comparison
of
UV
Lamps
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
2­
9
Exhibit
2.5:
UV
Dose
Requirements
for
Inactivation
of
Cryptosporidium,
Giardia,
and
Viruses
During
Validation
Testing
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
2­
11
Exhibit
2.6:
Comparison
of
Air
and
Liquid
Oxygen
Systems
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
2­
15
Exhibit
2.7:
Comparison
of
CT
Values
for
Free
Chlorine
and
Ozone
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
2­
16
Exhibit
2.8:
Reported
Ozonation
Requirements
for
2
log
Inactivation
of
Cryptosporidium
Oocysts
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
2­
18
Exhibit
2.9:
CT
Considerations
for
Cryptosporidium
Inactivation
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
2­
21
Exhibit
2.10:
Pressure­
Driven
Membrane
Separation
Spectrum
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
2­
22
Exhibit
2.11:
MF
and
UF
Studies
Documenting
Bacteria
Removal
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
2­
23
Exhibit
2.12:
MF
and
UF
Studies
Documenting
Giardia
Removal
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
2­
24
Exhibit
2.13:
MF
and
UF
Studies
Documenting
Cryptosporidium
Removal
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
2­
24
Exhibit
2.14:
MF
and
UF
Studies
Documenting
Virus
Removal
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
2­
25
Exhibit
2.15:
Summary
of
Bag
Filter
Performance
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
2­
29
Exhibit
2.16:
Bank
Filtration
Studies
Measuring
Coliform
and
Spore
Removal
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
2­
32
Exhibit
2.17:
NF
Studies
Documenting
Microbial
Removal
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
2­
46
Exhibit
2.18:
NOM
Removal
Through
NF
Processes
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
2­
48
Exhibit
2.19:
Bromide
Removal
Through
NF
Processes
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
2­
49
Exhibit
3.1:
Base
Plant
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
3­
2
Exhibit
3.2:
Chloramines
for
Secondary
Disinfection
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
3­
3
Exhibit
3.3:
Disinfection
with
Chlorine
Dioxide
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
3­
4
Exhibit
3.4:
UV
Disinfection
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
3­
6
Exhibit
3.5:
Water
Quality
Assumptions
for
UV
Disinfection
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
3­
6
Exhibit
3.6
Number
of
Assumed
UV
Reactors
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
3­
7
Exhibit
3.7:
Ozone
Disinfection
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
3­
8
Exhibit
3.8:
Microfiltration
and
Ultrafiltration
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
3­
9
Exhibit
3.9:
Bag
and
Cartridge
Filtration
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
3­
11
Exhibit
3.10:
GAC
Filtration
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
3­
17
Exhibit
3.11:
Nanofiltration
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
3­
19
Exhibit
4.1:
Technologies
Costed
and
Methodology
Used
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
2
Exhibit
4.2:
Summary
of
Capital
Cost
Multiplier
Components
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
4
Exhibit
4.3:
Costs
Indices
Used
in
the
Water
and
W/
W
Cost
Models
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
5
Exhibit
4.4:
Unit
and
General
Cost
Assumptions
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
6
Exhibit
4.5:
Chemical
Costs
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
6
Exhibit
4.6:
Summary
of
Piloting
Cost
Assumptions
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
10
Technologies
and
Costs
for
Control
of
Microbial
Contaminants
and
Disinfection
Byproducts
June
2003
viii
Exhibit
4.7:
Summary
of
Land
Cost
Assumptions
(
as
a
percentage
of
Capital
Cost)
.
.
.
.
.
.
.
.
.
.
.
4­
11
Exhibit
4.8:
Amortization
Factors
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
13
Exhibit
4.9:
Chloramines
as
Secondary
Disinfectant
Cost
Summary
­
Ammonia
Dose
=
0.15
mg/
L
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
17
Exhibit
4.10:
Chloramines
as
Secondary
Disinfectant
Cost
Summary
­
Ammonia
Dose
=
0.55
mg/
L
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
19
Exhibit
4.11:
W/
W
Cost
Model
Electricity
Usage
and
Required
Labor
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
24
Exhibit
4.12:
Chlorine
Dioxide
Cost
Summary
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
25
Exhibit
4.13:
UV
Disinfection
Cost
Summary
(
40
mJ/
cm2
Without
UPS)
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
30
Exhibit
4.14:
UV
Disinfection
Cost
Summary
(
200
mJ/
cm2
Without
UPS)
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
32
Exhibit
4.15:
UV
Disinfection
Cost
Summary
(
40
mJ/
cm2
with
UPS)
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
34
Exhibit
4.16:
UV
Disinfection
Cost
Summary
(
200
mJ/
cm2
with
UPS)
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
36
Exhibit
4.17:
Ozone
Piloting
Assumptions
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
44
Exhibit
4.18:
Ozonation
O&
M
Cost
Assumptions
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
45
Exhibit
4.19:
Ozonation
Cost
Summary
(
0.5
log
Cryptosporidium
Inactivation)
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
46
Exhibit
4.20:
Ozonation
Cost
Summary
(
1.0
log
Cryptosporidium
Inactivation)
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
48
Exhibit
4.21:
Ozonation
Cost
Summary
(
2.0
log
Cryptosporidium
Inactivation)
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
50
Exhibit
4.22:
Summary
of
MF/
UF
Vendor
Estimates
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
53
Exhibit
4.23:
Summary
of
MF/
UF
Interstage
Pumping
Assumptions
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
54
Exhibit
4.24:
MF/
UF
Land
Cost
Assumptions
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
56
Exhibit
4.25:
Summary
of
MF/
UF
Operator
Training
Assumptions
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
57
Exhibit
4.26:
Summary
of
Backwash
Disposal
Pipeline
Assumptions
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
57
Exhibit
4.27:
Summary
of
Membrane
Replacement
Costs
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
59
Exhibit
4.28:
Summary
of
MF/
UF
Labor
Assumptions
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
61
Exhibit
4.29:
Microfiltration/
Ultrafiltration
Cost
Summary
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
62
Exhibit
4.30:
Design
Criteria
for
Bag
and
Cartridge
Filters
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
65
Exhibit
4.31:
Summary
of
Bag
and
Cartridge
Filter
Pump
Cost
Data
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
65
Exhibit
4.32:
Bag
Filter
Cost
Summary
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
68
Exhibit
4.33
Cartridge
Filter
Cost
Summary
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
69
Exhibit
4.34:
Bank
Filtration
Cost
Estimates
for
Three
System
Sizes
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
70
Exhibit
4.35:
Second
Stage
Filtration
Cost
Estimates
for
Three
System
Sizes
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
70
Exhibit
4.36:
Pre­
Sedimentation
Cost
Estimates
for
Three
System
Sizes
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
71
Exhibit
4.37:
Watershed
Cost
Categories
for
Three
System
Sizes
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
72
Exhibit
4.38:
Summary
of
Filtration
Improvement
Design
Assumptions
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
74
Exhibit
4.39:
Valve
Actuator
Horsepower
Assumptions
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
78
Exhibit
4.40:
Capital
Unit
Costs
for
Combined
Filter
Performance
Components
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
80
Exhibit
4.41:
O&
M
Unit
Costs
for
Combined
Filter
Performance
Components
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
80
Exhibit
4.42:
Percentages
of
Plants
Using
Each
Filter
Improvement
Option
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
81
Exhibit
4.43:
Capital
Cost
Estimates
for
the
Combined
Filter
Performance
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
82
Exhibit
4.44:
O&
M
Costs
for
the
Combined
Filter
Performance
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
82
Exhibit
4.45:
GAC
Contactor
Assumptions
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
83
Technologies
and
Costs
for
Control
of
Microbial
Contaminants
and
Disinfection
Byproducts
June
2003
ix
Exhibit
4.46:
Summary
of
GAC
Costs
(
EBCT
=
10
minutes,
360
day
reactivation
frequency)
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
94
Exhibit
4.47:
Summary
of
GAC
Costs
(
EBCT
=
20
minutes,
90
day
reactivation
frequency)
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
96
Exhibit
4.48:
Summary
of
GAC
Costs
(
EBCT
=
20
minutes,
240
day
reactivation
frequency)
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
98
Exhibit
4.49:
Percent
Distribution
of
NF
Equipment
Cost
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
101
Exhibit
4.50:
Summary
of
NF
Housing
Cost
Assumptions
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
103
Exhibit
4.51:
NF
Land
Cost
Assumptions
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
103
Exhibit
4.52:
NF
Operator
Training
Cost
Assumptions
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
104
Exhibit
4.53:
Summary
of
NF
Technical
Labor
Assumptions
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
107
Exhibit
4.54:
Nanofiltration
Cost
Summary
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
108
Exhibit
4.55:
Annualized
Cost
Summary
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
111
Technologies
and
Costs
for
Control
of
Microbial
Contaminants
and
Disinfection
Byproducts
June
2003
x
List
of
Appendices
Appendix
A
Very
Small
Systems
Model
Capital
Cost
Breakdown
Summaries
Appendix
B
Water
Model
Capital
Cost
Breakdown
Summaries
Appendix
C
W/
W
Cost
Model
Capital
Cost
Breakdown
Summaries
Appendix
D
Technology
Cost
Curves
Technologies
and
Costs
for
Control
of
Microbial
Contaminants
and
Disinfection
Byproducts
June
2003
xi
List
of
Acronyms
AWWA
American
Water
Works
Association
AWWARF
American
Water
Works
Association
Research
Foundation
AWWSC
American
Water
Works
Service
Company
BAT
Best
Available
Technology
BCI
building
cost
index
BDOC
biodegradable
organic
carbon
BLS
Bureau
of
Labor
Statistics
BV
bed
volume
/
C
degrees
Celsius
CAP
total
capital
costs
CDC
Centers
for
Disease
Control
and
Prevention
CFE
combined
filter
effluent
CFR
Code
of
Federal
Regulations
CIP
clean­
in­
place
cm
centimeter
CT
measured
disinfectant
residual
×
contact
time
CWS
community
water
system
D/
DBP
disinfectant/
disinfection
byproduct
DBP
disinfection
byproduct
DBPR
Disinfectants
and
Disinfection
Byproducts
Rule
DES
designated
flow
DNA
deoxyribonucleic
acid
DOC
dissolved
organic
carbon
E&
I
electrical
and
instrumentation
EA
economic
analysis
Technologies
and
Costs
for
Control
of
Microbial
Contaminants
and
Disinfection
Byproducts
June
2003
xii
EA
environmental
assessment
EBCT
empty
bed
contact
time
EIS
environmental
impact
statement
ENR
Engineering
News
Record
EPA
United
States
Environmental
Protection
Agency
ES
effective
size
ESWTR
Enhanced
Surface
Water
Treatment
Rule
FACA
Federal
Advisory
Committee
Act
FBRR
Filter
Backwash
Recycling
Rule
fps
feet
per
second
ft
feet
ft2
(
sf
or
sq
ft)
square
feet
FTW
filter
to
waste
GAC
granular
activated
carbon
gfd
gallons
of
filtrate
per
day
per
square
foot
of
membrane
area
gpd
gallons
per
day
gpm
gallons
per
minute
GWUDI
ground
water
under
the
direct
influence
of
surface
water
HAA
haloacetic
acid
HAA5
sum
of
five
haloacetic
acids
HAA6
sum
of
six
haloacetic
acids
HIV
human
immunodeficiency
virus
Hp
horsepower
HPC
heterotrophic
plate
count
hr
hour
HVAC
heating,
ventilation,
and
air
conditioning
i
discount
rate
Technologies
and
Costs
for
Control
of
Microbial
Contaminants
and
Disinfection
Byproducts
June
2003
xiii
I&
C
instrumentation
and
controls
ICR
Information
Collection
Rule
IESWTR
Interim
Enhanced
Surface
Water
Treatment
Rule
in
inch
kgal
thousand
gallons
kgpd
thousand
gallons
per
day
kW
kilowatt
kWh
kilowatt
hour
lb
pound
LOX
liquid
oxygen
LP
low
pressure
LPHO
low
pressure
high
output
LPUV
low
pressure
ultraviolet
light
LT1ESWTR
Long
Term
1
Enhanced
Surface
Water
Treatment
Rule
LT2ESWTR
Long
Term
2
Enhanced
Surface
Water
Treatment
Rule
MCL
maximum
contaminant
level
MCLG
maximum
contaminant
level
goal
M­
DBP
microbial­
disinfection
Byproduct
MF
microfiltration
mg/
kg
milligrams
per
kilogram
µ
g/
L
micrograms
per
liter
mg/
L
milligrams
per
liter
mgal
million
gallons
MGD
or
mgd
million
gallons
per
day
mJ
milliJoules
mJ/
cm2
milliJoules
per
square
centimeter
µ
m
micrometer
Technologies
and
Costs
for
Control
of
Microbial
Contaminants
and
Disinfection
Byproducts
June
2003
xiv
mm
millimeter
MP
medium
pressure
MRDL
maximum
residual
disinfectant
level
MRDLG
maximum
residual
disinfectant
level
goal
MWCO
molecular
weight
cut­
off
MWDSC
Metropolitan
Water
District
of
Southern
California
N
number
of
years
NDWAC
National
Drinking
Water
Advisory
Council
NF
nanofiltration
NIPDWR
National
Interim
Primary
Drinking
Water
Regulation
nm
nanometers
NOM
natural
organic
matter
NPDWR
National
Primary
Drinking
Water
Regulation
NSF
National
Science
Foundation
NTNCWS
nontransient
noncommunity
water
system
NTU
nephelometric
turbidity
units
O&
M
operations
and
maintenance
OGWDW
Office
of
Ground
Water
and
Drinking
Water
OH&
P
overhead
and
profit
OSHA
Occupational
Safety
and
Health
Administration
P&
V
pipes
and
valves
PAC
powder
activated
carbon
PLC
programmable
logic
controller
POTW
publicly
owned
treatment
works
ppb
parts
per
billion
ppm
parts
per
million
PPI
Producer
Price
Index
(
for
Finished
Goods)
Technologies
and
Costs
for
Control
of
Microbial
Contaminants
and
Disinfection
Byproducts
June
2003
xv
PSA
pressure
swing
absorption
psi
pounds
per
square
inch
psig
pounds
per
square
inch
gauge
PUV
pulsed
ultraviolet
PVC
polyvinyl
chloride
PWS
public
water
supply
RIA
regulatory
impact
analysis
RNA
ribonucleic
acid
RO
reverse
osmosis
SAB
Science
Advisory
Board
SCADA
Supervisory
Control
and
Data
Acquisition
scf
standard
cubic
feet
SDS
simulated
distribution
system
SDWA
Safe
Drinking
Water
Act
sf
(
ft2
or
sq
ft)
square
feet
SOC
soluble
organic
carbon
SOC
synthetic
organic
compound
sq
ft
(
or
sf
or
ft2)
square
feet
SWAT
surface
water
analytical
tool
SWTR
Surface
Water
Treatment
Rule
TDH
total
dynamic
head
TDP
Technology
Design
Panel
TDS
total
dissolved
solids
THM
trihalomethane
THMFP
trihalomethane
formation
potential
TMP
transmembrane
pressure
TNCWS
transient
noncommunity
water
system
Technologies
and
Costs
for
Control
of
Microbial
Contaminants
and
Disinfection
Byproducts
June
2003
xvi
TOC
total
organic
carbon
TOX
total
organic
halide
TOXFP
total
organic
halide
formation
potential
TSS
total
suspended
solids
TTHM
total
trihalomethane
TWG
Technical
Work
Group
UC
uniformity
coefficient
UF
ultrafiltration
UPS
uninterrupted
power
supply
UV
ultraviolet
UVT
ultraviolet
transmittance
UV
254
ultraviolet
absorbance
at
254
nm
VSS
Very
Small
Systems
Best
Available
Technology
Cost
Document
wk
week
WTP
water
treatment
plant
W/
W
water
and
wastewater
yr
year
Technologies
and
Costs
for
Control
of
Microbial
Contaminants
and
Disinfection
Byproducts
June
2003
1­
1
1.
Introduction
1.1
Purpose
of
Technology
and
Cost
Document
This
document
provides
information
on
costs
and
treatment
effectiveness
of
technologies
and
treatment
strategies
available
to
public
water
systems
(
PWSs)
to
remove
or
inactivate
pathogenic
microorganisms,
specifically
Cryptosporidium,
and/
or
reduce
the
formation
of
disinfection
byproducts
(
DBPs).
This
information
is
developed
solely
for
use
in
conducting
Economic
Analyses
(
EAs)
for
the
proposed
Stage
2
Disinfectants
and
Disinfection
Byproducts
Rule
(
DBPR)
and
Long
Term
2
Enhanced
Surface
Water
Treatment
Rule
(
LT2ESWTR).
Please
note
that
the
information
provided
by
this
document
is
of
a
general
nature.
It
is
not
intended
to
guide
PWSs
in
selecting
or
designing
technologies
for
compliance
with
existing
or
proposed
rules.

The
proposed
LT2ESWTR
will
require
systems
to
provide
additional
Cryptosporidium
treatment
if
Cryptosporidium
concentrations
in
their
source
waters
exceed
specified
levels.
Cryptosporidium
is
resistant
to
chlorine
but
can
be
inactivated
with
certain
alternative
disinfectants
or
can
be
physically
removed
through
filtration
processes.

The
proposed
Stage
2
DBPR
will
require
PWSs
to
reduce
the
formation
of
trihalomethanes
(
THMs)
or
haloacetic
acids
(
HAAs)
if
they
exceed
specified
levels.
THMs
and
HAAs
form
primarily
through
reactions
between
chlorine
and
natural
organic
matter
(
NOM).
Their
formation
can
be
reduced
with
alternative
disinfectants
or
disinfection
practices
or
through
increases
in
NOM
removal
prior
to
chlorine
application.

Issues
associated
with
microbial
disinfection
and
the
formation
of
DBPs
are
interwoven;
PWSs
should
not
undercut
microbial
protection
in
their
efforts
to
reduce
DBP
levels.
Several
of
the
alternative
disinfectants
that
systems
could
choose
to
reduce
the
formation
of
THMs
and
HAAs
can
provide
increased
protection
against
chlorine­
resistant
pathogens
like
Cryptosporidium.
For
these
reasons,
PWSs
should
have
the
ability
to
make
decisions
regarding
compliance
strategies
for
the
Stage
2
DBPR
and
LT2ESWTR
at
the
same
time.
Consequently,
the
United
States
Environmental
Protection
Agency
(
EPA)
is
developing
these
regulations
as
a
paired
rulemaking
and
is
addressing
compliance
technologies
for
both
rules
in
a
single
document.

The
EAs
for
the
LT2ESWTR
and
Stage
2
DBPR
evaluate
the
total
impact
of
a
regulation
in
terms
of
costs
associated
with
additional
treatment
requirements
and
benefits
associated
with
reduced
risk.
This
evaluation
requires
the
following
types
of
information:

°
National
occurrence
of
the
regulated
contaminant(
s)

°
Existing
level
of
treatment
for
the
contaminant
provided
by
PWSs
Technologies
and
Costs
for
Control
of
Microbial
Contaminants
and
Disinfection
Byproducts
June
2003
1­
2
°
Unit
costs
and
efficacy
of
treatment
strategies
available
for
compliance
with
the
proposed
regulation
°
Number
and
sizes
of
PWSs
that
will
select
a
particular
treatment
strategy
for
regulatory
compliance
°
Benefits
and
costs
resulting
from
changes
to
existing
treatment
This
document
supports
the
EA
by
describing
the
design
criteria
necessary
for
a
technology
to
achieve
a
desired
level
of
treatment
and
the
cost
associated
with
that
technology
as
a
function
of
the
design
criteria.
Information
on
unit
costs
and
treatment
performance
is
critical
to
projecting
technology
usage
stemming
from
a
regulation
and
to
evaluating
national
compliance
costs
and
benefits.
No
information
is
given
here
on
the
national
compliance
costs
(
that
information
is
provided
in
the
EA)
or
on
the
numbers
of
PWSs
that
will
adopt
various
treatment
strategies
to
comply
with
the
proposed
regulations.

Process
design
criteria
for
alternative
disinfection
strategies
and
DBP
precursor
removal
technologies
were
developed
in
large
part
using
water
quality
data
gathered
under
the
Information
Collection
Rule
(
ICR)
and
best
engineering
judgement.
Where
appropriate,
EPA
used
ICR
data
to
generate
statistics
regarding
water
quality
parameters
that
affect
technology
performance.
These
water
quality
statistics
were
used
to
estimate
costs
for
technology
options
presented
in
this
document.
Costs
were
developed
using
EPA
cost
models,
manufacturer
price
data,
and
recent
literature.
Unit
prices
and
cost
indices
for
model
input
were
based
upon
vendor
information,
prevailing
rates,
and
published
values
in
the
trade
literature
(
e.
g.,
Engineering
News
Record,
Bureau
of
Labor
Statistics).
These
costs
were
reviewed
by
the
Technical
Work
Group,
which
was
convened
by
EPA
to
assist
in
the
Stage
2
DBPR
and
LT2ESWTR
regulatory
development
process.
Subsequent
revisions
have
also
been
made
to
respond
to
comments
from
outside
reviewers,
particularly
the
National
Drinking
Water
Advisory
Council
(
NDWAC)
and
EPA's
Science
Advisory
Board
(
SAB).

1.2
Existing
Regulations
The
following
are
existing
regulations
that
address
risks
posed
by
microorganisms
and
DBPs
in
public
water
systems.

1.2.1
Surface
Water
Treatment
Rule
Under
the
Surface
Water
Treatment
Rule
(
SWTR),
finalized
in
1989,
EPA
set
Maximum
Contaminant
Level
Goals
(
MCLGs)
of
zero
for
Giardia
lamblia,
viruses,
and
Legionella;
and
promulgated
National
Primary
Drinking
Water
Regulations
(
NPDWRs)
for
all
PWSs
using
surface
water
or
ground
water
under
the
direct
influence
of
surface
water
(
GWUDI).
Unfiltered
systems
were
required
to
comply
with
the
SWTR
by
1991
and
filtered
systems
by
1993.
The
SWTR
includes
treatment
technique
requirements
for
filtered
and
unfiltered
systems
that
are
intended
to
protect
against
Technologies
and
Costs
for
Control
of
Microbial
Contaminants
and
Disinfection
Byproducts
June
2003
1­
3
the
adverse
health
effects
of
exposure
to
Giardia,
viruses,
and
Legionella,
as
well
as
other
pathogenic
microorganisms
(
63
FR
69478
December
1998b).
Briefly,
those
requirements
include
the
following:

°
Maintenance
of
a
disinfectant
residual
in
the
distribution
system
°
Removal/
inactivation
of
3
log
(
99.9
percent)
for
Giardia
and
4
log
(
99.99
percent)
for
viruses
°
Combined
filter
effluent
turbidity
performance
standards
°
Watershed
protection
and
raw
water
quality
requirements
for
unfiltered
systems
1.2.2
Information
Collection
Rule
The
ICR
is
a
monitoring
and
data
reporting
rule
that
was
promulgated
in
1996.
The
purpose
of
the
ICR
was
to
collect
occurrence
and
treatment
information
to
help
evaluate
the
need
for
possible
changes
to
the
SWTR
and
microbial
treatment
practices
and
to
help
evaluate
the
need
for
future
regulation
of
DBPs.
The
ICR
provided
EPA
with
information
on
the
occurrence
of
pathogenic
microorganisms,
including
Cryptosporidium,
Giardia,
and
viruses,
as
well
as
the
occurrence
of
DBPs
and
water
quality
parameters
that
impact
DBP
formation.
The
ICR
also
provided
engineering
data
on
how
PWSs
control
such
contaminants
(
65
FR
19046
April
2000).

1.2.3
Interim
Enhanced
Surface
Water
Treatment
Rule
The
Interim
Enhanced
Surface
Water
Treatment
Rule
(
IESWTR)
was
finalized
in
December
1998
and
applies
only
to
surface
water
and
GWUDI
PWSs
serving
10,000
or
more
people.
The
purposes
of
the
IESWTR
were
to
improve
control
of
microbial
pathogens,
specifically
Cryptosporidium
and
to
address
risk
trade­
offs
between
pathogens
and
disinfection
byproducts
(
65
FR
19046
April
2000).
Key
provisions
of
the
rule
include
the
following:

°
MCLG
of
zero
for
Cryptosporidium
°
2
log
(
99
percent)
Cryptosporidium
removal
requirements
for
systems
that
filter
°
Strengthened
combined
filter
effluent
turbidity
standards
°
Requirements
for
individual
filter
turbidity
monitoring
°
Disinfection
benchmark
provisions
to
ascertain
the
level
of
microbial
protection
provided
as
systems
take
steps
to
comply
with
new
DBP
standards
Technologies
and
Costs
for
Control
of
Microbial
Contaminants
and
Disinfection
Byproducts
June
2003
1­
4
°
Inclusion
of
Cryptosporidium
in
the
definition
of
GWUDI
and
in
the
watershed
control
requirements
for
unfiltered
systems
°
Requirements
for
covers
on
new
finished
water
reservoirs
°
Requirements
for
sanitary
surveys
for
all
surface
water
and
GWUDI
systems,
even
those
serving
fewer
than
10,000
people
1.2.4
Stage
1
Disinfectants
and
Disinfection
Byproducts
Rule
The
Stage
1
Disinfectants
and
Disinfection
Byproducts
Rule
was
promulgated
in
1998.
The
Stage
1
DBPR
applies
to
all
PWSs
that
are
community
water
systems
(
CWSs)
or
non­
transient
noncommunity
water
systems
(
NTNCWSs)
and
that
treat
their
water
with
a
chemical
disinfectant
for
either
primary
or
secondary
disinfection.
In
addition,
certain
requirements
for
chlorine
dioxide
apply
to
transient
non­
community
water
systems
(
TNCWSs).
Surface
water
and
GWUDI
systems
serving
at
least
10,000
people
were
required
to
comply
with
the
Stage
1
DBPR
by
January
2002.
All
ground
water
systems,
as
well
as
surface
water
and
GWUDI
systems
serving
fewer
than
10,000
people,
must
comply
with
the
Stage
1
DBPR
by
January
2004.

The
Stage
1
DBPR
established
the
following
provisions:

°
Maximum
residual
disinfectant
level
goals
(
MRDLGs)
for
chlorine,
chloramines,
and
chlorine
dioxide
°
MCLGs
for
three
trihalomethanes
(
bromodichloromethane,
dibromochloromethane,
and
bromoform),
two
haloacetic
acids
(
dichloroacetic
acid
and
trichloroacetic
acid),
bromate,
and
chlorite
°
Maximum
residual
disinfectant
levels
(
MRDL)
for
chlorine,
chloramines,
and
chlorine
dioxide
°
MCLs
for
total
trihalomethanes
(
TTHM),
five
haloacetic
acids
(
HAA5),
bromate,
and
chlorite
The
rule
also
includes
monitoring,
reporting,
and
public
notification
requirements
for
the
listed
compounds.
EPA
estimates
that
the
rule
will
provide
public
health
protection
for
an
additional
20
million
households
not
previously
covered
by
drinking
water
rules
for
DBPs
(
65
FR
19046
April
2000).
Technologies
and
Costs
for
Control
of
Microbial
Contaminants
and
Disinfection
Byproducts
June
2003
1­
5
1.2.5
Long
Term
1
Enhanced
Surface
Water
Treatment
Rule
The
Long
Term
1
Enhanced
Surface
Water
Treatment
Rule
(
LT1ESWTR)
(
67
FR
1812
January
2002),
finalized
in
January
2002,
extends
the
requirements
of
the
IESWTR
to
surface
water
and
GWUDI
systems
serving
fewer
than
10,000
people.

1.2.6
Filter
Backwash
Recycling
Rule
The
Filter
Backwash
Recycling
Rule
(
FBRR)
(
66
FR
31086
June
2001)
regulates
systems
in
which
filter
backwash
is
returned
to
the
treatment
process.
The
rule,
promulgated
in
June
2001,
applies
to
surface
water
and
GWUDI
systems
that
use
direct
or
conventional
filtration
and
recycle
spent
filter
backwash
water,
sludge
thickener
supernatant,
or
liquids
from
dewatering
processes.
The
rule
requires
that
these
recycled
liquids
be
returned
to
a
location
such
that
all
steps
of
a
system's
conventional
or
direct
filtration
process
are
employed.
The
rule
also
requires
systems
to
notify
the
state
that
they
practice
recycling.
Finally,
systems
must
collect
and
maintain
information
for
review
by
the
state.

1.3
Public
Health
Concerns
1.3.1
Pathogenic
Microorganisms
In
1990,
EPA's
SAB,
an
independent
panel
of
experts
established
by
Congress,
cited
drinking
water
contamination
as
one
of
the
most
important
environmental
risks
and
indicated
that
diseasecausing
microbial
contaminants
(
e.
g.,
bacteria,
protozoa,
and
viruses)
are
probably
the
greatest
remaining
health
risk
management
challenge
for
drinking
water
suppliers
(
EPA/
SAB
1990).
Information
on
the
number
of
waterborne
disease
outbreaks
from
the
U.
S.
Centers
for
Disease
Control
and
Prevention
(
CDC)
underscores
this
concern.
CDC
indicates
that,
between
1991
and
2000,
145
drinking
water­
related
disease
outbreaks
were
reported,
with
more
than
431,000
associated
cases
of
disease
(
This
includes
outbreaks
in
individual
water
systems,
which
are
not
PWSs.
About
400,000
cases
of
illness
were
from
one
outbreak.)
During
this
period,
a
number
of
agents
were
implicated
as
the
cause,
including
protozoa,
viruses,
and
bacteria.

Waterborne
diseases
are
usually
acute
(
i.
e.,
sudden
onset
and
typically
lasting
a
short
time
in
healthy
people),
and
most
waterborne
pathogens
cause
gastrointestinal
illness,
with
diarrhea,
abdominal
discomfort,
nausea,
vomiting,
and/
or
other
symptoms.
Some
waterborne
pathogens
cause,
or
are
associated
with,
more
serious
disorders
such
as
hepatitis,
gastric
cancer,
peptic
ulcers,
myocarditis,
swollen
lymph
glands,
meningitis,
encephalitis,
and
other
diseases.

Cryptosporidium,
a
protozoan
parasite,
is
of
particular
concern
as
a
waterborne
pathogen
because
it
is
highly
resistant
to
inactivation
by
chlorine
and
chloramines.
In
addition,
no
therapeutic
treatment
currently
exists
for
cryptosporidiosis,
the
infection
caused
by
Cryptosporidium.
Technologies
and
Costs
for
Control
of
Microbial
Contaminants
and
Disinfection
Byproducts
June
2003
1­
6
Cryptosporidiosis
usually
causes
7­
14
days
of
diarrhea,
sometimes
accompanied
by
a
low­
grade
fever,
nausea,
or
abdominal
cramps
in
healthy
individuals
(
Juranek
1995).
It
may,
however,
cause
the
death
of
individuals
with
compromised
immune
systems.
In
1993,
Cryptosporidium
caused
more
than
400,000
people
in
Milwaukee
to
experience
intestinal
illness.
More
than
4,000
were
hospitalized,
and
at
least
50
deaths
were
attributed
to
the
cryptosporidiosis
outbreak.
Nevada,
Oregon,
and
Georgia
have
also
experienced
cryptosporidiosis
outbreaks
over
the
past
several
years.

Despite
filtration
and
disinfection,
Cryptosporidium
oocysts
have
been
found
in
filtered
drinking
water
(
LeChevallier
et
al.
1991),
and
many
of
the
individuals
affected
by
waterborne
disease
outbreaks
caused
by
Cryptosporidium
were
served
by
filtered
surface
water
supplies
(
Solo­
Gabriele
and
Neumeister
1996).
Surface
water
systems
that
filter
and
disinfect
may
still
be
vulnerable
to
Cryptosporidium,
depending
on
the
source
water
quality
and
treatment
effectiveness.

1.3.2
Disinfectants/
Disinfection
Byproducts
While
the
use
of
chemical
disinfectants
is
highly
effective
in
reducing
the
risk
of
waterborne
disease,
disinfectants
are
known
to
react
with
NOM
to
form
byproducts
that
may
pose
a
public
health
risk.
In
addition,
the
disinfectants
themselves
may
pose
a
public
health
risk
at
high
concentrations.

The
assessment
of
public
health
risks
from
chlorination
of
drinking
water
currently
relies
on
inherently
difficult
and
incomplete
empirical
analysis.
Nevertheless,
while
recognizing
these
uncertainties
and
taking
into
account
the
large
number
of
people
exposed
to
DBPs
and
the
different
potential
health
risks
that
may
result
from
exposure
to
DBPs
(
e.
g.,
cancer
and
adverse
reproductive
and
developmental
effects),
EPA
believes
that
the
weight
of
evidence
represented
by
the
available
epidemiology
and
toxicology
studies
support
a
hazard
concern
and
a
protective
public
health
approach
to
regulation.

1.4
Proposed
Regulations
1.4.1
Long
Term
2
Enhanced
Surface
Water
Treatment
Rule
In
September
2000,
an
Agreement
in
Principle
was
reached
by
EPA
and
members
of
the
Stage
2
Microbial­
Disinfection
Byproduct
(
M­
DBP)
Federal
Advisory
Committee
Act
(
FACA)
Committee
regarding
the
requirements
of
the
proposed
LT2ESWTR
(
65
FR
83015
December
2000).
Under
the
agreement,
the
LT2ESWTR
will
require
all
surface
water
systems,
including
GWUDI,
that
serve
at
least
10,000
people
to
conduct
two
years
of
source
water
monitoring
for
Cryptosporidium.
Conventional
systems
whose
annual
average
Cryptosporidium
concentrations
are
at
least
0.075,
1.0,
or
3.0
oocysts
per
liter
would
be
required
to
achieve
an
additional
1,
2,
or
2.5
logs,
respectively,
of
Cryptosporidium
removal
or
inactivation
beyond
conventional
treatment.
Systems
could
meet
these
additional
treatment
requirements
through
the
use
of
various
options
including:
enhanced
filtration
Technologies
and
Costs
for
Control
of
Microbial
Contaminants
and
Disinfection
Byproducts
June
2003
1­
7
performance,
watershed
control,
alternative
disinfectants,
membranes,
various
types
of
filters,
and
demonstrations
of
performance.

1.4.2
Stage
2
Disinfectants/
Disinfection
Byproducts
Rule
The
Stage
2
DBPR,
which
will
be
proposed
along
with
the
LT2ESWTR,
will
apply
to
all
CWSs
and
NTNCWSs
that
add
a
disinfectant
other
than
ultraviolet
(
UV)
light
or
deliver
disinfected
water.
Under
the
Stage
2
M­
DBP
Agreement
in
Principle
(
65
FR
83015
December
2000),
the
Stage
2
DBPR
will
retain
the
MCLs
of
80
:
g/
L
for
TTHM
and
60
:
g/
L
for
HAA5
established
by
the
Stage
1
DBPR.
However,
the
Stage
2
DBPR
will
change
the
way
compliance
with
these
MCLs
is
determined.
Under
Stage
1,
compliance
with
the
TTHM
and
HAA5
MCLs
is
based
on
a
running
annual
average
of
all
monitoring
points
within
a
distribution
system.
Under
the
Stage
2
DBPR,
compliance
would
be
based
on
a
locational
running
annual
average,
which
means
that
the
running
annual
average
at
each
monitoring
point
within
a
distribution
system
would
have
to
be
less
than
the
MCL.
The
Stage
2
DBPR
would
also
require
systems
to
conduct
an
initial
distribution
system
evaluation
which
would
identify
the
areas
with
the
highest
concentrations
of
TTHM
and
HAA5;
compliance
monitoring
will
be
conducted
at
those
locations.

1.5
Technologies
Evaluated
for
the
Control
of
Pathogens
and
Disinfection
Byproducts
Systems
required
to
provide
additional
treatment
for
Cryptosporidium
under
the
LT2ESWTR
can
use
two
basic
mechanisms:
inactivation
and
removal.
While
chlorine
and
chloramines
are
not
effective
against
Cryptosporidium
at
doses
used
in
drinking
water
treatment,
chlorine
dioxide,
ozone,
and
UV
light
have
been
demonstrated
to
inactivate
this
pathogen.
Chlorine
dioxide
and
ozone
generally
require
higher
doses
to
inactivate
Cryptosporidium
than
those
necessary
for
Giardia
and
viruses;
the
use
of
these
disinfectants
is
limited
by
the
formation
of
regulated
byproducts
like
chlorite
and
bromate.
UV
has
been
shown
to
achieve
high
levels
of
Cryptosporidium
inactivation
at
relatively
low
doses
but
is
currently
not
widely
used
in
the
United
States
for
drinking
water
treatment.
Nevertheless,
EPA
believes
that
ozone,
chlorine
dioxide,
and
UV
are
available
to
PWSs
to
inactivate
Cryptosporidium.
Consequently,
EPA
has
evaluated
these
technologies
in
this
document.

PWSs
can
increase
the
physical
removal
of
Cryptosporidium
in
their
treatment
plants
by
using
additional
physical
barriers
like
microfiltration
(
MF)
or
bag
and
cartridge
filtration.
These
technologies
have
been
shown
to
achieve
high
log
reductions
of
Cryptosporidium
when
properly
designed
and
implemented.
This
document
addresses
Cryptosporidium
removal.

Utilities
can
also
take
steps
to
reduce
the
concentration
of
Cryptosporidium
entering
the
treatment
plant
through
strategies
such
as
watershed
control,
pre­
sedimentation
basins,
and
bank
Technologies
and
Costs
for
Control
of
Microbial
Contaminants
and
Disinfection
Byproducts
June
2003
1­
8
filtration.
Costs
for
these
technologies
were
obtained
from
the
M­
DBP
FACA
Committee
and
are
provided
in
Chapter
4.
However,
these
costs
were
too
uncertain
to
use
in
the
EA
for
the
LT2ESWTR.

Systems
required
to
reduce
the
formation
of
TTHM
and
HAA5
for
compliance
with
the
Stage
2
DBPR
can
use
two
approaches.
One
approach
is
to
reduce
the
use
of
free
chlorine
by
switching
to
disinfectants
that
do
not
form,
or
form
only
low
concentrations
of,
TTHM
and
HAA5.
Such
disinfectants
include:
chloramines,
ozone,
chlorine
dioxide,
and
UV.
Systems
may
also
reduce
free
chlorine
doses
by
using
physical
barriers
like
microfiltration;
microfiltration
removes
more
microorganisms
so
that
less
disinfection
is
needed.
This
document
evaluates
chloramines,
ozone,
chlorine
dioxide,
UV,
and
MF
as
alternative
disinfection
strategies
for
reducing
TTHM
and
HAA5
formation.
(
Note
that
several
of
these
disinfection
strategies
were
also
evaluated
for
Cryptosporidium
treatment
as
described
above.)

The
second
approach
for
systems
to
reduce
TTHM
and
HAA5
formation
is
to
increase
the
removal
of
DBP
precursors
(
i.
e.,
NOM)
prior
to
disinfection.
Systems
can
remove
precursors
by
increasing
coagulation
dosages
in
a
process
termed
enhanced
coagulation,
or
softening,
or
by
installing
granular
activated
carbon
(
GAC)
or
nanofiltration
(
NF).
For
the
purposes
of
this
document,
it
was
assumed
that
utilities
will
have
already
optimized
coagulation
or
softening
practices
to
meet
the
requirements
of
the
Stage
1
DBPR.
As
a
result,
this
document
evaluates
only
GAC
and
NF
as
precursor
removal
strategies.

In
summary,
this
document
provides
an
analysis
of
the
following
technologies:

Alternative
disinfection
strategies
°
Chloramination
°
Chlorine
dioxide
°
Ultraviolet
(
UV)
light
°
Ozone
°
Microfiltration
and
ultrafiltration
°
Bag
and
cartridge
filters
°
Bank
filtration
°
Second
stage
filtration
°
Pre­
sedimentation
basins
Technologies
and
Costs
for
Control
of
Microbial
Contaminants
and
Disinfection
Byproducts
June
2003
1­
9
°
Watershed
control
°
Combined
Filter
Performance
Alternative
DBP
precursor
removal
strategies
°
Granular
activated
carbon
adsorption
°
Nanofiltration
1.6
Document
Organization
This
remainder
of
this
document
contains
the
following
sections:

Chapter
2
­
Technologies
for
DBP
and
Microbial
Contaminant
Control:
Presents
comprehensive
discussions
of
all
disinfection,
Cryptosporidium
removal,
and
DBP
precursor
removal
strategies
considered
in
this
document.
Includes
technology
descriptions,
effectiveness
of
technologies
for
DBP
precursor
and/
or
microbial
control,
and
factors
affecting
the
performance
of
each
technology.

Chapter
3
­
Technology
Design
Criteria
:
Discusses
the
rationale
behind
development
of
the
design
criteria
for
which
costs
are
presented
in
Chapter
4.
Includes
design
approach,
assumptions
and
additional
factors
(
e.
g.,
residuals
handling)
which
may
impact
design.

Chapter
4
­
Technology
Costs:
Presents
capital,
operations
and
maintenance,
and
total
annualized
costs
for
each
disinfection
strategy
and
DBP
precursor
removal
technology
considered.
Also
includes
discussion
of
estimation
methods
(
e.
g.,
cost
models
and
vendor
information).

Chapter
5
­
References:
Provides
a
comprehensive
bibliography
of
all
literature
used
in
the
compilation
of
this
document.

Appendices:
Contain
capital
cost
breakdown
summaries
for
technologies
for
which
cost
models
were
used.
Technologies
and
Costs
for
Control
of
Microbial
Contaminants
and
Disinfection
Byproducts
June
2003
2­
1
2.
Technologies
for
DBP
and
Microbial
Contaminant
Control
2.1
Introduction
Public
water
systems
may
employ
various
treatment
strategies
to
reduce
chlorinated
DBPs
and
to
provide
better
physical
removal
or
inactivation
of
Cryptosporidium
for
compliance
with
the
proposed
Stage
2
DBPR
and
LT2ESWTR.
EPA
considers
the
following
treatment
strategies
as
being
available
for
compliance
with
these
two
proposed
regulations:

Alternative
disinfection
strategies
°
Chloramination
(
section
2.2.1)

°
Chlorine
dioxide
(
section
2.2.2)

°
Ultraviolet
light
(
section
2.2.3)

°
Ozone
(
section
2.2.4)

°
Microfiltration
and
ultrafiltration
(
section
2.2.5)

°
Bag
and
cartridge
filtration
(
section
2.2.6)

°
Bank
filtration
(
section
2.2.7)

°
Second
stage
filtration
(
section
2.2.8)

°
Pre­
sedimentation
(
section
2.2.9)

°
Watershed
control
(
section
2.2.10)

°
Combined
filter
performance
(
section
2.2.11)

DBP
precursor
removal
strategies
°
Granular
activated
carbon
adsorption
(
section
2.3.1)

°
Nanofiltration
(
section
2.3.2)
Technologies
and
Costs
for
Control
of
Microbial
Contaminants
and
Disinfection
Byproducts
June
2003
2­
2
2.2
Alternative
Disinfection
Strategies
2.2.1
Chloramination
Chloramines
are
formed
by
reactions
of
ammonia
with
aqueous
chlorine.
These
reactions
may
result
in
the
formation
of
monochloramine
(
NH2Cl),
dichloramine
(
NHCl2)
and
trichloramine
(
NCl3).
The
relative
concentrations
of
these
species
depend
upon
the
pH
of
the
water
and
the
relative
proportion
of
chlorine
and
ammonia.
At
chlorine­
to­
ammonia
mass
ratios
of
3:
1
to
5:
1
(
Cl2:
NH3­
N)
and
neutral
pHs,
conditions
common
to
drinking
water
treatment,
the
principal
chloramine
species
formed
is
monochloramine
(
USEPA
1999b).

One
of
the
least
expensive
methods
for
controlling
DBP
formation
is
the
use
of
monochloramine,
instead
of
free
chlorine,
to
maintain
a
distribution
system
residual.
After
the
appropriate
free
chlorine
contact
time,
ammonia
is
added
to
quench
the
residual
free
chlorine
and
to
retard
DBP
formation.
This
reduces
the
free
chlorine
contact
time
and,
thus,
DBP
formation,
without
compromising
microbial
protection.
The
initial
free
chlorine
contact
time
and
chloramine
together
provide
sufficient
disinfection.
A
survey
conducted
by
the
American
Water
Works
Association
Research
Foundation
(
AWWARF)
has
shown
that
most
of
the
utilities
that
changed
disinfection
practices
to
lower
distribution
system
THM
levels
have
done
so
by
switching
to
chloramine
as
the
secondary
disinfectant
(
McGuire
1989).

Systems
that
do
not
use
free
chlorine
for
primary
disinfection
(
e.
g.,
that
use
ozone
or
UV
light)
must
add
chlorine
prior
to
ammonia
addition.
For
most
systems,
the
free
chlorine
residual
needs
to
be
increased
prior
to
the
point
of
ammonia
addition
to
maintain
the
desired
chloramine
residual
in
the
distribution
system.
This
can
be
accomplished
by:
1)
simultaneous
addition
of
chlorine
and
ammonia
(
after
primary
disinfection
with
free
chlorine
or
ozone)
or
2)
the
addition
of
ammonia
after
chlorine
addition.

Further
information,
including
case
studies
of
systems
converting
from
free
chlorine
to
chloramine,
is
summarized
in
Optimizing
Chloramine
Treatment
(
Kirmeyer
et
al.
1993).
This
reference
supplies
additional
information
on
the
reason(
s)
for
switching
to
chloramine
and
contains
information
on
chloramination
changeover
and
start­
up
procedures,
nitrification,
and
impact
on
taste
and
odor.

2.2.1.1
Efficacy
Against
Pathogens
Chloramine
is
less
effective
than
free
chlorine
for
the
disinfection
of
most
pathogenic
microorganisms.
At
pH
7
and
below,
monochloramine
is
approximately
200
times
less
effective
than
free
chlorine
for
coliform
inactivation
under
the
same
contact
time,
temperature,
and
pH
conditions.
For
viruses
and
cysts,
the
combined
chlorine
forms
(
e.
g.,
monochloramine
and
dichloramine)
are
considerably
less
effective
than
free
chlorine
(
USEPA
1999b).
Historical
studies
have
found
time
Technologies
and
Costs
for
Control
of
Microbial
Contaminants
and
Disinfection
Byproducts
June
2003
2­
3
factors
(
monochloramine
contact
time:
free
chlorine
contact
time)
from
20:
1
to
80:
1
for
the
same
bacterial
inactivation
efficiency.
For
the
same
conditions
of
contact
time,
temperature,
and
pH,
combined
chlorine
(
monochloramine)
doses
are
approximately
25
times
higher
than
free
chlorine
for
the
same
bacterial
inactivation
efficiency
(
White
1999).
There
is
evidence
that
dichloramine
may
be
twice
as
effective
as
monochloramine;
however,
dichloramine
is
generally
avoided
because
it
contributes
to
taste
and
odor
problems.

The
Guidance
Manual
for
Compliance
with
the
Filtration
and
Disinfection
Requirements
for
Public
Water
Systems
Using
Surface
Water
Sources
(
SWTR
Guidance
Manual
 
USEPA
1990)
presents
CT
(
contact
time
multiplied
by
residual
disinfectant
concentration)
values
for
multiple
disinfectants,
pathogens,
pH
and
temperature
ranges.
Exhibit
2.1
compares
CT
requirements
for
chloramine
with
those
of
free
chlorine
over
a
range
of
temperature
and
pH
values.

Exhibit
2.1:
Comparison
of
CT
Values
for
Free
Chlorine
and
Chloramine
Log
Remova
l
Giardia
Viruses
<
1o
C
10o
C
20o
C
<
1o
C
10o
C
20o
C
Cl
NH2Cl
Cl
NH2Cl
Cl
NH2Cl
Cl
NH2Cl
Cl
NH2Cl
Cl
NH2Cl
0.5
40
635
21
310
10
185
­­
­­
­­
­­
­­
­­

1
79
1270
42
615
21
370
­­
­­
­­
­­
­­
­­

2
158
2535
83
1230
41
735
6
1243
3
643
1
321
3
237
3800
125
1850
62
1100
9
2063
4
1067
2
534
Note:
­­
Data
not
available.
Source:
USEPA
1990.

Exhibit
2.1
demonstrates
that
chloramine
is
relatively
ineffective
compared
to
free
chlorine
for
Giardia
and
virus
inactivation.
In
addition,
chloramine
is
ineffective
for
inactivation
of
Cryptosporidium
(
Peeters
et
al.
1989,
Korich
et
al.
1990).
Several
studies
have
evaluated
whether
disinfection
with
ozone
followed
by
chloramination
(
Liyanage
et
al.
1997a,
Driedger
et
al.
1999)
has
a
synergistic
effect
on
Cryptosporidium
inactivation
(
i.
e.,
the
inactivation
achieved
using
both
disinfectants
combined
is
greater
than
what
is
expected
for
each
of
the
disinfectants
separately).
The
results
of
these
studies
are
inconclusive
but
indicate
that
some
synergism
may
exist
for
ozone/
chloramine
applications.
Technologies
and
Costs
for
Control
of
Microbial
Contaminants
and
Disinfection
Byproducts
June
2003
2­
4
2.2.1.2
DBP
Formation
The
byproducts
formed
by
chloramination,
for
the
most
part,
are
identical
to
those
produced
during
chlorination
and
include
THMs,
HAAs,
haloacetonitriles,
and
cyanogen
chloride.
With
the
possible
exception
of
cyanogen
chloride,
chloramination
does
not
preferentially
form
any
of
the
halogenated
DBPs
compared
to
free
chlorine.
In
fact,
studies
have
demonstrated
that
chloramines
produce
much
lower
levels
of
DBPs
than
free
chlorine
(
Kirmeyer
et
al.
1993,
Symons
et
al.
1996).
This
is
the
primary
reason
water
systems
implement
chloramines
for
secondary
disinfection
rather
than
free
chlorine.

The
formation
of
DBPs
resulting
from
chloramination
is
influenced
by
the
following
treatment
variables
(
Kirmeyer
et
al.
1993,
Carlson
and
Hardy
1998):

°
Contact
time
and
chloramine
dosage
°
Point
of
ammonia
application
°
pH
and
temperature
°
Total
organic
carbon
°
Chlorine­
to­
ammonia
ratio
°
Mixing
and
reaction
time
for
chloramine
formation
The
point
of
ammonia
application
after
chlorine
addition
generally
impacts
the
length
of
time
free
chlorine
reacts
with
NOM.
For
most
plants
using
chlorine
as
a
primary
disinfectant,
the
point
of
ammonia
application
depends
on
disinfection
requirements
and
goals.
Once
ammonia
is
added,
the
rate
of
DBP
formation
is
significantly
reduced
(
Kirmeyer
et
al.
1993).

Within
the
range
of
chloramine
residuals
commonly
used
in
the
water
industry
(
1
to
5
milligrams
per
liter
(
mg/
L)),
chloramine
dose
does
not
appear
to
be
a
significant
factor
in
DBP
formation;
the
chlorine­
to­
ammonia
ratio
appears
to
be
more
significant.
TTHM
concentrations
remain
quite
low
at
chlorine­
to­
ammonia
weight
ratios
less
than
5:
1,
then
increase
dramatically
above
the
5:
1
ratio
(
Kirmeyer
et
al.
1993).
Most
utilities
use
chlorine­
to­
ammonia
ratios
of
3:
1
to
5:
1
because
dichloramine
and
trichloramine
form
at
higher
ratios.
These
species
are
unstable
and
cause
taste
and
odor
problems.
Technologies
and
Costs
for
Control
of
Microbial
Contaminants
and
Disinfection
Byproducts
June
2003
2­
5
2.2.1.3
Factors
Affecting
Performance
When
chlorine
and
ammonia
are
added
simultaneously,
good
mixing
can
reduce
the
time
free
chlorine
has
to
react
with
NOM.
With
complete
mixing
at
neutral
pHs
(
7
to
9)
and
temperatures
of
20
to
25
degrees
Celsius
(
/
C),
the
reaction
of
ammonia
and
chlorine
to
form
monochloramine
takes
less
than
3
seconds.
This
eliminates
the
free
chlorine
almost
immediately
and
reduces
the
potential
for
DBP
formation
(
Kirmeyer
et
al.
1993).
Efficient
mixing
and
dispersion
of
chemicals
(
chlorine
and
ammonia)
at
the
point
of
addition
determines
the
extent
of
free
chlorine
contact
and,
thus,
substantially
impacts
the
formation
of
DBPs.

As
noted
above,
pH
is
important
for
rapid
formation
of
chloramine.
Symons
et
al.
(
1996)
showed
that
DBP
formation
decreased
with
increasing
pH.
Exceptions
to
the
trend
are
noted
in
some
instances
at
pH
8,
where
Symons
et
al.
noted
that
the
complexity
of
chloramine
chemistry
may
cause
water­
specific
responses.

Carlson
and
Hardy
(
1998)
evaluated
the
effects
of
various
water
quality
variables,
such
as
pH,
temperature,
chlorine
dosage,
and
total
organic
carbon
on
THM
and
HAA
formation
for
waters
from
five
utilities.
Of
the
variables
studied,
the
free
chlorine
contact
time
was
found
to
be
the
most
important
in
forming
chlorinated
DBPs.
Chlorine
contact
time
must
be
balanced
to
provide
disinfection
and
to
control
byproduct
formation.
The
type
of
DBP
precursor
was
also
found
to
be
important.
Based
on
this
study,
the
authors
proposed
the
concept
of
two
sets
of
precursors:
those
that
form
DBPs
quickly
and
those
that
form
DBPs
slowly.
The
precursor
material
that
rapidly
reacts
with
chlorine
to
form
DBPs
(
i.
e.
the
quick
formers)
are
of
greater
importance
when
chloramine
is
used
to
maintain
a
residual.
These
quick
formers
are
less
affected
by
reaction
conditions
than
are
the
slow
formers.
Relatively
consistent
THM
and
HAA
concentrations
formed
quickly
after
the
addition
of
chlorine.
Temperature,
chlorine
dosage,
and
pH
had
a
greater
effect
on
precursor
materials
that
formed
DBPs
slowly.

White
(
1999)
summarizes
the
effect
of
contact
time
and
dose
on
the
disinfection
properties
of
chloramines.
Generally,
chloramines
require
much
longer
contact
times
than
other
chemical
disinfectants
(
e.
g.,
free
chlorine
and
ozone).
This
is
one
reason
they
are
more
suitable
for
secondary
disinfection
in
the
distribution
system,
where
residence
times
can
be
several
days.
Chloramines
are
a
less
powerful
oxidant
than
many
other
chemical
disinfectants
and
can
require
substantially
higher
doses
to
achieve
the
same
level
of
disinfection
(
White
1999).
Because
longer
contact
times
and
higher
doses
are
required
for
effective
chloramine
disinfection,
residual
stability
is
of
major
importance.
Monochloramine,
the
preferred
chloramine
form,
is
the
dominant
species
at
pH
levels
greater
than
5.5
and
is
essentially
the
only
species
present
at
pH
levels
around
7.5
(
Kirmeyer
et
al.
1993).
Systems
using
chloramines
for
secondary
disinfection
should
try
to
maintain
a
distribution
system
pH
of
approximately
7.5.

A
primary
concern
for
systems
using
chloramines
is
nitrification
in
the
distribution
system.
Nitrification
is
a
microbiological
process
by
which
free
ammonia
is
converted
to
nitrite
and
nitrate.
Ammonia
oxidizing
bacteria
and
nitrobacter,
which
are
naturally
present
in
distribution
system
biofilms
Technologies
and
Costs
for
Control
of
Microbial
Contaminants
and
Disinfection
Byproducts
June
2003
2­
6
and
may
infiltrate
leaking
or
corroding
pipes,
convert
free
ammonia
to
nitrite
and
(
in
the
presence
of
sufficient
dissolved
oxygen)
nitrate,
respectively.
Among
the
effects
of
nitrification
are
a
depletion
of
the
chloramine
residual
and
an
increase
in
heterotrophic
plate
counts
(
HPC)
(
Kirmeyer
et
al.
1995).
To
prevent
nitrification,
it
is
important
to
optimize
the
chlorine:
ammonia
ratio
and
minimize
free
ammonia
in
the
distribution
system.
Nitrification
is
most
likely
to
occur
in
distribution
system
dead
ends,
areas
of
low
demand,
and
storage
tanks.
As
a
result,
the
potential
for
nitrification
can
also
be
minimized
by
improving
distribution
system
piping
configurations
(
e.
g.,
looping
to
eliminate
dead
ends
and
increasing
flow
in
low
demand
areas)
and
by
increasing
storage
tank
turnover.

2.2.2
Chlorine
Dioxide
Chlorine
dioxide
has
been
used
for
drinking
water
treatment
in
the
United
States
for
more
than
50
years,
primarily
to
control
taste
and
odor
problems.
However,
chlorine
dioxide
has
received
attention
lately
because
of
its
potential
application
for
Cryptosporidium
inactivation
(
Finch
et
al.
1995,
Li
et
al.
1998)
and
for
reduced
formation
of
THMs
or
HAAs
during
disinfection
(
White
1999).
However,
chlorine
dioxide
degrades
to
form
chlorite
and
chlorate.
Chlorite
is
considered
to
have
public
health
implications
and
is
a
regulated
DBP.

Chlorine
dioxide
cannot
be
transported
because
of
its
instability
and
explosiveness.
Therefore,
it
is
generated
on­
site.
The
five
common
methods
for
producing
chlorine
dioxide
are
as
follows:
1)
sodium
chlorite
reaction
with
acid,
2)
chorine
solution
reaction
with
chlorite
solution,
3)
chlorine
gas
reaction
with
chlorite
solution,
4)
reduction
of
sodium
chlorate
using
hydrogen
peroxide
and
concentrated
sulfuric
acid,
and
5)
chlorine
gas
reaction
with
solid
chlorite
(
White
1999).
The
yield,
purity,
and
production
capacities
of
chlorine
dioxide
vary
for
the
five
types
of
methods.
The
most
common
chlorine
dioxide
generation
technique
is
chlorine
solution
reaction
with
chlorite
solution.
Chlorine
dioxide
dosages
that
can
be
used
in
drinking
water
treatment
are
constrained
by
regulatory
limits
on
the
production
of
chlorite
and
chlorine
dioxide
residual.

2.2.2.1
Efficacy
Against
Pathogens
The
SWTR
Guidance
Manual
presents
CT
values
for
inactivation
of
Giardia
and
viruses
for
both
free
chlorine
and
chlorine
dioxide.
The
values
indicate
that
chlorine
dioxide
is
approximately
four
times
more
effective
that
chlorine
for
the
inactivation
of
Giardia
at
most
conditions.
Chlorine,
however,
is
more
effective
for
the
inactivation
of
viruses.
Exhibit
2.2
summarizes
CT
values
contained
in
the
guidance
manual.
Technologies
and
Costs
for
Control
of
Microbial
Contaminants
and
Disinfection
Byproducts
June
2003
2­
7
Exhibit
2.2:
Comparison
of
CT
Values
for
Free
Chlorine
and
Chlorine
Dioxide
Log
Remova
l
Giardia
Viruses
<
1o
C
10o
C
20o
C
<
1o
C
10o
C
20o
C
Cl
ClO2
Cl
ClO2
Cl
ClO2
Cl
ClO2
Cl
ClO2
Cl
ClO2
0.5
40
10
21
4
10
2.5
­­
­­
­­
­­
­­
­­

1
79
21
42
7.7
21
5
­­
­­
­­
­­
­­
­­

2
158
42
83
15
41
10
6
8.4
3
4.2
1
2.1
3
237
63
125
23
62
15
9
25.6
4
12.8
2
6.4
Note:
­­
Data
not
available.
Source:
USEPA
1990.

Chlorine
dioxide
has
been
compared
to
other
oxidants
for
inactivating
Cryptosporidium
(
Korich
et
al.
1990);
chlorine
dioxide
and
ozone
are
found
to
be
more
effective
in
inactivating
Cryptosporidium
compared
to
chlorine
and
monochloramine.
However,
unlike
ozone,
the
degradation
byproducts
of
chlorine
dioxide
do
not
contribute
to
the
inactivation
of
Cryptosporidium
(
Liyanage
et
al.
1997b).
The
American
Water
Works
Service
Company
(
AWWSC)
evaluated
the
effectiveness
of
chlorine
dioxide
for
the
inactivation
of
Cryptosporidium
(
AWWSC
1998).
AWWSC
found
that
chlorine
dioxide
is
effective
for
warm,
high
pH
waters
(
pH
of
approximately
8
and
temperature
around
20
degrees
Celsius).
Finch
et
al.
(
1995)
summarized
the
chlorine
dioxide
research
regarding
the
inactivation
of
Cryptosporidium.
A
summary
of
CT
values
for
Cryptosporidium
is
presented
in
Exhibit
2.3.

Exhibit
2.3:
Summary
of
Chlorine
Dioxide
CT
Values
for
Cryptosporidium
Inactivation
Log
Inactivation
AWWSC
From
Summary
by
Finch
et
al.
(
1995)
10o
C
20o
C
1
99
48
60
2
257
115
80
3
­­
­­
140
Note:
­­
Data
not
available.
All
values
are
for
pH
8.
Temperature
for
Finch
et
al.
is
unknown.
Technologies
and
Costs
for
Control
of
Microbial
Contaminants
and
Disinfection
Byproducts
June
2003
2­
8
Chlorine
dioxide
has
also
been
proven
effective
for
the
inactivation
of
selected
bacteria
over
a
pH
range
of
3.0
to
8.0
(
Junli
et
al.
1997,
White
1999)
and
is
a
stronger
disinfectant
than
chlorine
for
bacteria,
requiring
lower
CT
values.
Some
of
the
bacteria
evaluated
in
Junli
et
al.
(
1997)
are
E.
coli
(
A
and
B),
Staphylococcus
aureus,
Sarcina,
Chloropseudomonas,
Bacillus
subtilis,
and
Shigella
dysenteriae.

2.2.2.2
DBP
Formation
Studies
have
demonstrated
that
chlorine
dioxide
does
not
produce
THMs
(
White
1999);
under
proper
generation
conditions
(
i.
e.,
no
excess
chlorine),
halogen­
substituted
DBPs
are
not
formed.
The
application
of
chlorine
dioxide
produces
only
a
small
amount
of
total
organic
halide
(
TOX)
(
Werdehoff
and
Singer,
1987).
The
use
of
chlorine
dioxide
aids
in
reducing
the
formation
of
TTHMs
and
HAAs
by
oxidizing
precursors.
By
moving
the
point
of
chlorination
downstream
in
the
plant
after
coagulation,
sedimentation,
and
filtration,
the
quantity
of
NOM
is
reduced.
This
results
in
a
lower
chlorine
dosage
during
post­
chlorination
of
the
water
which,
in
turn,
results
in
fewer
THMs.

In
normal
pH
ranges
(
6
to
9),
chlorine
dioxide
undergoes
a
variety
of
oxidation
reactions
with
NOM
to
form
oxidized
organic
species,
such
as
chlorinated,
brominated,
or
polysubstituted
organic
byproducts
and
chlorite
(
ClO2
­).
Chlorite
concentrations
can
account
for
up
to
70
percent
of
the
chlorine
dioxide
consumed
(
American
Water
Works
Association
(
AWWA)
1999;
Werdehoff
and
Singer
1987).
Chlorite,
and
chlorate
(
ClO3
­)
are
formed
when
chlorine
dioxide
is
added
to
water.
All
three
oxidized
chlorine
species
(
chlorine
dioxide,
chlorite,
and
chlorate)
are
considered
to
have
adverse
health
effects
and
are
of
concern
in
finished
water
(
AWWA
1999).

Chlorine
dioxide
may
also
facilitate
a
number
of
chlorine
substitution
reactions.
Studies
evaluating
drinking
water
and
NOM
have
shown
that
TOX
concentration
increases
upon
application
of
chlorine
dioxide
at
normal
treatment
dosages
(
AWWA
1999).

2.2.2.3
Factors
Affecting
Performance
Temperature
dramatically
affects
Cryptosporidium
inactivation
by
chlorine
dioxide
(
Li
et
al.
1998).
At
1
oC,
a
0.5
log
inactivation
is
observed
at
a
CT
of
150
milligrams
*
minutes
/
liter
(
mgmin
L),
compared
to
a
2.0
log
inactivation
for
the
same
CT
at
22"
C.
Chlorine
dioxide
can
effectively
inactivate
bacteria
over
a
pH
range
of
3.0
to
8.0.
Because
it
is
a
more
effective
disinfectant
for
bacteria
than
free
chlorine,
lower
CT
values
are
required.
Caution
must
be
taken,
however,
when
selecting
the
appropriate
dose,
as
excessive
dosages
can
lead
to
chlorite
formation
above
permissible
levels.
Purity
and
generator
yields
are
two
of
the
most
critical
factors
that
effect
chlorine
dioxide
use.
Chlorine
and
the
oxychlorine
species
(
i.
e.,
chlorite
and
chlorate)
are
typically
present
in
the
impurities
of
chlorine
dioxide
(
White
1999).
Therefore,
the
purity
of
the
chlorine
dioxide
generated
should
be
considered
to
avoid
a
violation
of
the
chlorite
maximum
contaminant
level
(
MCL).
Technologies
and
Costs
for
Control
of
Microbial
Contaminants
and
Disinfection
Byproducts
June
2003
2­
9
2.2.3
Ultraviolet
Light
The
use
of
UV
light
for
disinfection
of
drinking
water
has
recently
received
much
attention
because
of
new
developments
regarding
Cryptosporidium
inactivation
at
low
UV
light
doses
(
Bukhari
et
al.
1999)
and
because
it
creates
very
few
DBPs.
Disinfection
is
accomplished
by
irradiating
water
with
UV
light.
UV
light
is
electromagnetic
radiation
between
wavelengths
of
100
and
400
nanometers
(
nm).
The
specific
range
of
UV
wavelengths
and
the
level
of
irradiance
depend
on
the
type
of
UV
lamp
system
used.
The
effective
germicidal
wavelength
range
for
most
microorganisms
is
generally
considered
to
be
between
200
and
300
nm
(
Malley
1998).

UV
systems
consist
of
UV
reactors
with
an
associated
control
panel.
Commercial
UV
reactors
used
for
drinking
water
applications
are
closed
reactors
containing
UV
lamps,
quartz
sleeves,
UV
intensity
sensors,
quartz
sleeve
wipers,
and
temperature
sensors.
UV
lamps
are
housed
within
the
quartz
sleeves,
which
protect
and
insulate
the
lamps.
Some
reactors
include
automatic
cleaning
mechanisms
to
keep
the
quartz
sleeves
free
of
deposits
that
may
form
due
to
contact
with
the
water.
UV
intensity
sensors,
flow
meters,
and
in
some
cases,
UV
transmittance
monitors
are
used
to
monitor
dose
delivery
by
the
reactor.

UV
lamps
can
be
divided
into
two
categories:
continuous
wave
and
pulsed
wave.
Currently,
continuous
wave
UV
lamps
are
most
widely
used
for
drinking
water
treatment.
The
types
of
continuous
wave
lamps
are
low
pressure
mercury
vapor
(
LP),
low
pressure
high
output
(
LPHO),
and
medium
pressure
mercury
vapor
(
MP).
"
Pressure"
refers
to
the
pressure
of
mercury
vapor
within
the
lamp
casing.
A
comparison
of
the
LP,
LPHO,
and
MP
lamps
is
shown
in
Exhibit
2.4.
Technologies
and
Costs
for
Control
of
Microbial
Contaminants
and
Disinfection
Byproducts
June
2003
2­
10
Exhibit
2.4:
Comparison
of
UV
Lamps
Parameter
LP
LPHO
MP
Germicidal
UV
light
Monochromatic
at
254
nm
Monochromatic
at
254
nm
Polychromatic,
including
germicidal
range
(
200
­
300nm)
Mercury
Vapor
Pressure
(
torr)
Optimal
at
0.007
Optimal
at
0.007
100
­
10,000
Operating
Temperature
(
oC)
Optimal
at
40
130
­
200
600
­
900
Electrical
Input
(
W/
centimeter
(
cm))
0.5
1.5
­
10
50
­
150
Germicidal
UV
Output
(
W/
cm)
0.2
0.5
­
3.5
5
­
30
Electrical
to
Germicidal
UV
Conversion
Efficiency
(%)
35
­
38
30
­
40
10
­
20
Arc
Length
(
cm)
10
­
150
10
­
150
5
­
75
Relative
Number
of
Lamps
Required
for
a
Given
Dose
High
Intermediate
Low
Lifetime
(
hours(
hrs))
8,000
­
10,000
8,000
­
12,000
3,000
­
5,000
Source:
EPA
UV
Disinfection
Guidance
Manual
(
USEPA
2003).

The
light
emitted
by
LP
and
LPHO
lamps
is
essentially
monochromatic
at
253.7
nm,
which
is
in
the
range
of
the
most
germicidal
wavelengths
for
microorganisms.
MP
lamps
emit
at
a
higher
intensity
than
LP
lamps
but
at
a
wide
range
of
wavelengths.
Therefore,
LP
and
LPHO
lamps
convert
power
to
germicidal
light
more
efficiently
than
MP
lamps.
Theoretically,
LPHO
lamps
have
the
same
efficiency
as
LP
lamps
because
they
operate
at
similar
vapor
pressures.
However
in
practice,
LPHO
lamp
efficiency
can
be
significantly
lower
when
operating
at
different
power
settings.
The
main
differences
between
LP
and
MP
lamps,
as
shown
in
Exhibit
2.4,
are
the
vapor
pressure,
operating
temperatures,
electrical
input,
and
germicidal
UV
output.

Pulsed
ultraviolet
(
PUV)
systems
irradiate
a
high
intensity
UV
light
in
flashes
at
approximately
50
flashes
per
second.
PUV
systems
have
limited
operating
experience
on
the
full­
scale
and
are
not
costed
in
this
document.

The
UV
lamp
ballast
controls
the
amount
of
electricity
supplied
to
the
lamp
and
should
ensure
a
consistent
and
constant
power
delivery.
Power
supplies
and
ballasts
can
be
supplied
in
many
different
configurations
and
are
tailored
to
a
unique
lamp
type
and
application.
UV
systems
may
use
electronic
ballasts,
magnetic
ballasts,
or
transformers.

UV
intensity
sensors
are
photosensitive
detectors
that
measure
the
UV
intensity
at
a
point
within
the
UV
reactor.
This
intensity
information
is
used
to
indicate
dose
delivery.
Intensity
sensors
can
be
classified
as
wet
or
dry.
Dry
sensors
monitor
UV
light
through
a
monitoring
window
whereas
wet
UV
intensity
sensors
are
in
direct
contact
with
the
water
flowing
through
the
reactor.
Monitoring
Technologies
and
Costs
for
Control
of
Microbial
Contaminants
and
Disinfection
Byproducts
June
2003
2­
11
windows
and
the
wetted
ends
of
the
wet
sensors
can
become
fouled
over
time
and
require
cleaning,
similar
to
quartz
sleeves.

The
lamp
cleaning
mechanism
used
for
a
UV
system
depends
on
the
lamp
type,
system
size,
and
fouling
potential
of
the
water.
Both
manual
and
automatic
cleaning
regimes
have
been
developed.
Manual
cleaning
is
primarily
used
for
smaller
systems
with
relatively
few
sleeves
and
lower
fouling
potential.
Automatic
cleaning
approaches
may
be
classified
as
flush
and
rinse
systems,
mechanical
wipers,
or
physical­
chemical
wipers.
LPHO
systems
typically
use
flush
and
rinse
systems,
and
MP
systems
typically
use
wipers
because
the
higher
lamp
temperatures
accelerate
fouling
under
certain
water
qualities.
The
cleaning
frequency
of
the
lamps
is
a
function
of
the
lamp
temperature
and
the
concentration
of
dissolved
organic
and
inorganic
species
that
can
cause
lamp
fouling.

2.2.3.1
Efficacy
Against
Pathogens
When
UV
light
is
applied
to
a
microorganism,
deoxyribonucleic
acid
(
DNA)
and
ribonucleic
acid
(
RNA)
absorb
the
light
energy
and
their
structure
is
altered,
thereby
interfering
with
replication
of
the
microbe.
The
UV
dose
necessary
for
inactivation
of
microorganisms
varies
from
species
to
species
and
across
microorganism
classifications.
Inactivation
of
microorganisms
increases
with
increasing
UV
dose,
although
it
does
not
always
follow
the
typical
log­
linear
relationship.

Of
the
pathogens
of
interest
in
drinking
water,
viruses
are
most
resistant
to
UV
disinfection,
followed
by
bacteria
and
protozoa.
Exhibit
2.5
presents
UV
dose
requirements
for
inactivation
of
Cryptosporidium,
Giardia,
and
viruses
(
as
derived
in
the
USEPA
UV
Disinfection
Guidance
Manual,
Appendix
B).
The
UV
dose
requirements
presented
in
Exhibit
2.5
are
the
minimum
required;
operational
UV
doses
will
likely
be
two
to
four
times
higher
after
application
of
a
safety
factor.

Exhibit
2.5:
UV
Dose
Requirements
for
Inactivation
of
Cryptosporidium,
Giardia,
and
Viruses
During
Validation
Testing
Log
Inactivation
0.5
1.0
1.5
2.0
2.5
3.0
3.5
4.0
Cryptosporidium
1.6
2.5
3.9
5.8
8.5
11.7
­
­

Giardia
1.5
2.1
3.0
5.2
7.7
10.8
­
­

Virus
39.4
58.1
79.1
100.1
120.7
142.6
163.1
186.0
Note:
All
values
presented
in
mJ
/
cm2
Source:
USEPA
UV
Disinfection
Guidance
Manual,
Appendix
B.
Technologies
and
Costs
for
Control
of
Microbial
Contaminants
and
Disinfection
Byproducts
June
2003
2­
12
Based
on
the
analysis
presented
in
Appendix
B
of
the
EPA
UV
Disinfection
Guidance
Manual,
the
sensitivities
of
Giardia
and
Cryptosporidium
to
UV
disinfection
are
very
similar;
viruses,
however,
are
more
difficult
to
inactivate.
Battigelli
et
al.
(
1993)
performed
bench
scale
UV
collimated
beam
experiments
to
determine
the
relationship
between
UV
dose
and
inactivation
of
Hepatitis­
A
virus
(
strain
HM­
175),
coxsackievirus
type
B­
5,
rotavirus
strain
SA­
11,
and
bacteriophages
MS­
2
and
fX174.
MS­
2
bacteriophage
required
the
highest
dose
of
25
milliJoules
per
square
centimeter
(
mJ/
cm2)
for
less
than
1
log
inactivation.
With
the
other
viruses,
4
log
inactivation
is
achieved
at
doses
ranging
between
16
and
42
mJ/
cm2.
The
most
UV­
resistant
viruses
of
concern
in
drinking
water
are
adenovirus
Type
40
and
Type
41.
Meng
and
Gerba
(
1996)
report
a
dose
of
23.6
to
30
mJ/
cm2
for
a
1
log
inactivation
in
adenovirus
and
a
dose
of
111.8
to
124
mJ/
cm2
for
4
log
inactivation.

Because
microbes
that
have
been
exposed
to
UV
light
still
retain
metabolic
functions,
some
are
able
to
repair
the
damage
done
by
UV
light
and
regain
infectivity.
Repair
of
UV
light­
induced
DNA
damage
includes
photoreactivation
and
dark
repair
(
Knudson
1985).
Photoreactivation
(
or
photorepair)
is
an
enzymatic
DNA
repair
mechanism
wherein
the
DNA
damage
is
repaired
when
exposed
to
light
between
310
and
490
nm.
Dark
repair
is
an
enzymatic
repair
mechanism
that
does
not
require
light.
Not
all
microorganisms
contain
the
necessary
cellular
mechanisms
to
repair
UV­
damaged
DNA.
One
study
has
shown
that
Cryptosporidium
contains
the
capability
to
undergo
some
DNA
repair.
However,
even
though
the
DNA
was
repaired,
infectivity
was
not
restored
(
Oguma
et
al.
2001).
Another
study,
by
Shin
et
al.
(
2001),
did
not
observe
photorepair
with
Cryptosporidium
parvum.
Linden
et
al.
(
2002a)
did
not
observe
photoreactivation
or
dark
repair
of
Giardia
at
UV
doses
typical
for
UV
disinfection
applications
(
16
and
40
mJ/
cm2).
However,
unpublished
data
from
the
same
study
showed
Giardia
reactivation
in
light
and
dark
conditions
at
very
low
UV
doses
(
0.5
mJ/
cm2;
Linden
2002a).
Shaban
et
al.
(
1997)
found
that
viruses
also
lack
the
repair
enzymes
necessary
for
photoreactivation.
However,
photorepair
of
viral
DNA
can
occur
using
the
enzyme
systems
of
their
host
cells.
Knudson
(
1985)
found
that
bacteria
were
able
to
repair
in
light
and
dark
conditions,
suggesting
that
bacteria
may
have
the
enzymes
necessary
for
photorepair
and
dark
repair.
As
such,
photoreactivation
is
generally
limited
to
bacteria.

E.
coli
and
HPC
inactivation
by
UV
light
are
well
documented,
particularly
with
respect
to
wastewater
disinfection
(
Chang
et
al.
1985,
Wilson
et
al.
1992).
Photoreactivation
of
bacteria
has
been
documented
with
E.
coli,
S.
aureus,
and
coliphage,
while
dark
repair
has
been
documented
with
S.
aureus
and
coliphage
(
Shaban
et
al.
1997).
One
study
(
Knudsen,
1985)
examined
two
different
strains
of
E.
coli:
one
that
had
the
enzymes
necessary
for
repair
(
B/
R
strain)
and
one
that
lacked
the
necessary
repair
enzymes
(
recA­
uvr­
strain).
The
difference
in
UV
dose
needed
for
1­
log
inactivation
of
the
strain
capable
of
repair
was
over
two
orders
of
magnitude
higher
than
the
dose
needed
for
1­
log
inactivation
of
the
repair
deficient
strain,
indicating
that
dark
repair
impacts
the
UV
dose­
response
of
microorganisms.
Unlike
bacteria,
viruses
do
not
have
the
enzymes
necessary
for
dark
repair.
However,
viruses
can
repair
in
the
host
cell
using
the
host
cells'
enzymes
(
Rauth
1965).
Technologies
and
Costs
for
Control
of
Microbial
Contaminants
and
Disinfection
Byproducts
June
2003
2­
13
2.2.3.2
DBP
Formation
Several
studies
have
been
conducted
to
determine
if
DBPs
are
formed
as
a
result
of
UV
light
irradiation.
Zheng
et
al.
(
1999)
found
that
TTHM
and
HAA9
formation
did
not
increase
when
UV
light
was
applied
to
chlorinated
water
at
a
dose
of
100
mJ/
cm2.
Linden
et
al.
(
1998)
investigated
DBP
formation
in
wastewater
secondary
effluent
that
is
irradiated
with
LP
and
MP
UV
lamps
and
found
no
evidence
of
photochemical
reactions
or
DBP
formation.
Malley
et
al.
(
1996)
examined
the
effects
of
post­
UV
disinfection
(
chlorination
and
chloramination)
on
DBP
formation
and
found
no
significant
impact
by
UV
on
DBP
levels
formed
by
chemical
disinfection.
Malley
et
al.
(
1995)
also
observed
no
significant
change
in
THM,
HAA,
bromate,
or
other
halogenated
DBP
concentrations
following
disinfection
with
UV
light.
A
study
performed
with
filtered
drinking
water
indicated
no
significant
change
in
aldehydes,
carboxylic
acids,
or
TOX
(
Kashinkunti
et
al.,
2003).
However,
a
low
conversion
rate
(
about
one
percent)
of
nitrate
to
nitrite
by
UV
light
has
been
observed
(
von
Sonntag
and
Schuchman,
1992).
Conversion
of
nitrate
to
nitrite
is
lower
with
LP
lamps
than
with
MP
lamps
because
the
UV
absorbance
of
nitrate
is
higher
below
240
nm
than
it
is
at
254
nm.
Due
to
the
low
conversion
rate
of
nitrate
to
nitrite
by
UV
light,
it
is
of
minimal
concern
in
drinking
water
applications.
Several
studies
have
shown
low­
level
formation
of
non­
regulated
DBPs
(
e.
g.,
aldehydes)
as
a
result
of
applying
UV
light
to
wastewater
and
raw
drinking
water
sources.
The
difference
in
results
can
be
attributed
to
the
difference
in
water
quality,
most
notably
the
higher
concentration
of
organic
material
in
raw
waters
and
wastewaters.

2.2.3.3
Factors
Affecting
Performance
Particle
content
can
impact
UV
disinfection
performance.
Particles
may
absorb
and
scatter
light,
thereby
reducing
the
UV
intensity
delivered
to
the
microorganisms.
Particle­
associated
microbes
also
may
be
shielded
from
UV
light,
effectively
reducing
disinfection
performance.
Particles
in
source
waters
are
diverse
in
composition
and
size
and
include
large
molecules,
microbes,
clay
particles,
algae,
and
flocs.

Recent
research
by
Linden
et
al.
(
2002b)
indicates
that
the
UV
dose­
response
of
microorganisms
added
to
filtered
drinking
waters
was
not
altered
by
variation
in
turbidity
that
met
regulatory
requirements.
For
unfiltered
raw
waters,
Passantino
and
Malley
(
2001)
found
that
source
water
turbidity
up
to
10
nephelometric
turbidity
units
(
NTU)
did
not
impact
the
UV
dose­
response
of
separately
added
(
seeded)
organisms.
In
these
experiments,
however,
organisms
were
added
to
waters
containing
various
levels
of
treated
or
natural
turbidity.
Therefore,
it
was
not
possible
to
examine
microbes
associated
with
particles
in
their
natural
or
treated
states.
Consequently,
these
drinking
water
studies
can
only
suggest
the
impact
of
turbidity
on
dose­
response
as
it
relates
to
the
impact
of
UV
light
scattering
by
particles.
The
studies
cannot
predict
the
effect
on
UV
disinfection
of
microbes
attaching
to
particles.
Technologies
and
Costs
for
Control
of
Microbial
Contaminants
and
Disinfection
Byproducts
June
2003
2­
14
UV
absorbance,
often
exerted
by
dissolved
organic
matter
in
drinking
water
applications,
affects
the
design
of
the
UV
system.
Water
that
absorbs
a
significant
amount
of
UV
light
(
i.
e.,
high
UV
absorbance
and
low
transmittance)
will
need
a
higher
UV
irradiance
or
longer
exposure
to
achieve
the
same
level
of
inactivation
as
water
with
lower
UV
absorbance.
As
UV
absorbance
increases,
the
intensity
throughout
the
reactor
decreases
for
a
given
lamp
configuration.
This
results
in
a
reduction
in
delivered
dose
and
measured
UV
intensity
for
a
given
lamp
output.
In
a
situation
with
a
fixed
UV
output,
lower
UV
absorbance
values
result
in
more
UV
energy
being
available
in
the
water
column,
causing
a
higher
log­
inactivation
of
microorganisms
than
a
water
with
a
higher
UV
absorbance.
For
systems
with
high
levels
of
dissolved
organic
matter
(
high
UV
absorbance),
it
is
more
efficient
to
apply
UV
light
after
unit
processes
that
remove
organic
matter.

Several
chemicals
used
in
water
treatment
processes
can
increase
the
UV
absorbance
of
water
(
e.
g.,
Iron
(
Fe+
3)).
However,
some
oxidants
(
including
ozone)
can
reduce
the
UV
absorbance
(
APHA
et
al.
1998).
Water
treatment
processes
upstream
of
the
UV
reactors
can
be
operated
to
control
and
reduce
UV
absorbance,
thereby
optimizing
the
design
and
costs
of
the
UV
system.

Depending
on
the
water
quality
(
e.
g.,
dissolved
ions,
hardness,
alkalinity,
and
pH
levels)
and
lamp
temperature,
scale
can
form
on
the
UV
lamps.
MP
lamps
tend
to
scale
more
easily
than
LP
and
LPHO
lamps
because
the
operating
temperature
of
MP
lamps
is
considerably
higher.
Scale
can
reduce
the
UV
energy
being
transmitted
through
the
lamp
sleeve
into
the
water
and
potentially
compromise
disinfection.
Lamp
cleaning
is
an
important
consideration
for
the
design
of
UV
systems
to
control
lamp
scaling
and
to
ensure
consistent
disinfection
performance.
Water
pH
may
also
affect
lamp
scale
formation,
but
inactivation
of
microorganisms
with
UV
light
is
not
pH
dependent
(
Malley
1998).

UV
inactivation
of
microorganisms
is
not
directly
affected
by
water
temperature.
However,
the
performance
of
UV
lamps
is
dependent
on
the
lamp
temperature.
Most
UV
lamps
have
sleeves
(
usually
made
of
quartz)
that
insulate
the
lamps,
maintain
optimal
temperature,
and
provide
maximum
irradiance.
If
the
lamp
temperature
deviates
from
optimal,
the
lamp
irradiance
will
be
reduced.
This
is
especially
true
with
LP
UV
lamps
in
cold
waters
(
Mackey
et
al.
2000).
Therefore,
the
water
temperature
variation
should
be
considered
when
designing
a
low
pressure
system.
However,
MP
lamps
have
a
significantly
higher
operating
temperature
compared
to
the
water
temperature.
Thus,
as
long
as
an
insulating
quartz
sleeve
is
in
place,
the
water
temperature
has
little
effect
on
the
operating
temperature
of
the
MP
lamp
and
MP
lamp
performance.

Hydraulics
are
an
important
part
of
the
UV
equipment.
Ideally,
the
UV
reactor
should
exhibit
plug­
flow
characteristics.
In
plug
flow,
water
that
enters
the
reactor
is
completely
mixed
axially
and
moves
through
the
reactor
as
a
single
plug
with
no
dispersion
in
the
direction
of
flow.
However,
"
real
world"
hydraulics
in
a
full­
scale
reactor
are
never
plug
flow.
UV
reactors
are
typically
equipped
with
baffles
to
reduce
the
amount
of
short­
circuiting
through
the
reactor
and
to
encourage
plug
flow,
although
these
baffles
can
increase
head
loss
through
the
reactor.
Staggered
lamp
arrays
also
promote
mixing
within
the
reactor
and
minimize
short­
circuiting
of
flow.
Alternatively,
vortex
mixers
can
be
used
to
increase
lamp
spacing,
thereby
reducing
head
loss
through
the
reactor.
Technologies
and
Costs
for
Control
of
Microbial
Contaminants
and
Disinfection
Byproducts
June
2003
2­
15
Inlet
and
outlet
conditions
can
have
a
significant
impact
on
reactor
hydrodynamics.
Straight
inlet
conditions
with
gradual
changes
in
cross
sectional
area
can
be
used
to
develop
flow
for
optimal
dose
delivery.
Straight
inlets
with
no
sharp
bends
or
sudden
changes
in
cross
sectional
area
optimize
dose
deliveries.

It
may
be
necessary
to
characterize
the
reactor
flow
regime
in
order
to
determine
the
level
of
disinfection
provided.
Tracer
tests
are
typically
not
feasible
because
the
hydraulic
residence
time
in
the
reactor
is
too
short
(
i.
e.,
on
the
order
of
seconds
or
fractions
of
a
second).
However,
hydraulic
models
are
available
to
understand
the
behavior
of
the
UV
reactor.

2.2.4
Ozone
In
recent
years,
the
use
of
ozone
technology
in
water
treatment
has
dramatically
increased.
In
1991,
approximately
40
water
treatment
plants
in
the
United
States,
each
serving
more
than
10,000
people,
utilized
ozone
(
Langlais
et
al.
1991).
As
of
April
1998,
this
number
had
grown
to
264
operating
plants
(
Rice
et
al.
1999).
The
main
reasons
for
the
escalating
use
of
ozonation
are
the
strong
oxidizing
properties
of
ozone
and
the
absence
of
the
formation
of
chlorinated
DBPs
during
disinfection
(
however,
bromated
DPBs
are
formed).

In
water,
ozone
reacts
with
hydroxide
ions
(
OH­)
to
form
hydroxyl
free
radicals
(
HO1);
therefore,
pH
is
a
very
important
parameter
in
determining
the
extent
and
rate
of
contaminant
oxidation.
Oxidation
with
ozone
is
also
influenced
by
other
water
quality
characteristics,
such
as
temperature,
alkalinity,
and
the
concentration
of
reduced
chemical
species
(
i.
e.,
iron
and
manganese).
Other
important
considerations
include
ozone
dose
and
contact
time.

Ozone
is
commonly
added
to
raw
water
(
pre­
ozonation)
or
settled
water.
To
take
advantage
of
ozone's
ability
to
improve
flocculation
and
NOM
removal,
ozone
must
be
applied
to
raw
water.
Application
of
ozone
to
raw
or
settled
water
is
considered
to
be
equally
effective
for
primary
disinfection.
However,
larger
doses
may
be
necessary
for
raw
water
application
due
to
the
higher
NOM
and
particulate
matter
concentrations.

There
are
two
basic
types
of
ozone
generation
equipment:
liquid
oxygen­
based
systems
and
airbased
systems.
Liquid
oxygen
feed
systems
are
relatively
simple
(
e.
g.,
there
is
no
air
pretreatment
equipment),
less
capital
intensive,
and
yield
a
higher
ozone
concentration
than
air­
based
systems.
The
liquid
oxygen
feed
system
components
include
a
storage
tank,
an
evaporator
to
convert
the
liquid
to
a
gas,
filters
to
remove
impurities,
and
pressure
regulators
to
limit
the
gas
pressure
to
the
ozone
generators.

Air­
fed
systems
require
air
pretreatment
equipment
to
prevent
damage
to
the
ozone
generator.
Air
needs
to
be
dry,
free
of
contaminants,
and
with
a
dew
point
between
­
50"
and
­
60"
C.
Air
Technologies
and
Costs
for
Control
of
Microbial
Contaminants
and
Disinfection
Byproducts
June
2003
2­
16
pretreatment
equipment
consists
of
compressors,
after
coolers
(
optional),
refrigerant
dryers,
desiccant
dryers,
air
filters,
and
pressure
regulators.
Power
consumption
is
higher
for
air
feed
systems
(
8­
12
kWh/
lb
O3)
than
for
oxygen
feed
systems
(
4­
8
kWh/
lb
O3).
Exhibit
2.6
presents
a
comparison
of
the
advantages
and
disadvantages
of
the
two
types
of
ozonation
systems
(
USEPA
1999b).

Exhibit
2.6:
Comparison
of
Air
and
Liquid
Oxygen
Systems
System
Advantages
Disadvantages
Air
Commonly
used
equipment
Proven
technology
Suitable
for
small
and
large
systems
More
energy
consumed
per
ozone
volume
produced
Extensive
gas
handling
equipment
required
Maximum
ozone
concentration
of
1­
5
%
Higher
power
consumption
Fairly
complicated
technology
Liquid
Oxygen
Less
equipment
required
Simple
to
operate
and
maintain
Suitable
for
small
and
large
systems
Can
store
excess
oxygen
to
meet
peak
demands
Higher
ozone
concentration
(
14­
18%)
Approximately
doubles
ozone
production
for
same
generator
Lower
power
consumption
Variable
liquid
oxygen
costs
Storage
of
oxygen
onsite
(
i.
e.,
safety
concerns)
Loss
of
liquid
oxygen
in
storage
when
not
in
use
Oxygen­
resistant
materials
required
Ozone
is
usually
applied
in
one
of
three
flow
configurations:
1)
co­
current
(
ozone
and
water
flowing
in
the
same
direction),
2)
counter­
current
(
ozone
and
water
flowing
in
the
opposite
direction),
or
3)
alternating
co­
current/
counter­
current.
Ozone
application
systems
include
fine
bubble
diffusers,
injectors/
static
mixers,
and
turbine
mixers
(
Langlais
et
al.
1991).
The
fine
bubble
diffuser
system
is
the
most
common
and
offers
high
ozone
transfer
rates,
process
flexibility,
operational
simplicity,
and
no
moving
parts.
The
injector/
static
mixer
system
applies
ozone
in
an
in­
line
or
a
sidestream
configuration.
Ozone
is
injected
under
negative
pressure,
created
by
a
venturi
section,
and
then
mixed
to
enhance
dispersion
of
ozone
in
the
water
stream.
The
turbine
mixer
systems
feed
ozone
in
the
contactor
and
mix
ozone
with
the
water
in
the
contactor.
The
turbine
mixer
can
either
project
outside
of
the
ozone
contactor
or
be
submerged.

Hoigne
and
Bader
(
1976)
described
ozone
decomposition
in
water.
Once
ozone
enters
solution,
it
follows
one
of
two
reaction
pathways:
1)
direct
oxidation,
which
is
slow
and
selective
in
its
oxidation
of
organic
compounds,
and
2)
autodecomposition
to
the
hydroxyl
free
radical
(
HO°),
which
is
extremely
fast
and
nonselective.
The
hydroxyl
free
radical
is
scavenged
by
carbonate
and
bicarbonate
ions,
commonly
measured
as
alkalinity,
to
form
carbonate
and
bicarbonate
free
radicals.
These
radicals
do
not
affect
the
organic
reactions.
The
hydroxyl
radicals
produced
by
the
autodecomposition
react
with
organics
and
other
radicals
to
reform
hydroxyl
radical
in
an
autocatalytic
process.
Technologies
and
Costs
for
Control
of
Microbial
Contaminants
and
Disinfection
Byproducts
June
2003
2­
17
The
stability
of
dissolved
ozone
is
affected
by
pH,
ultraviolet
light,
ozone
concentration,
and
the
concentration
of
radical
scavengers
(
Langlais
et
al.
1991).
Conditions
of
low
pH
favor
the
direct
oxidation
pathway,
and
high
pH
conditions
favor
the
autodecomposition
pathway
described
earlier.
At
pH
levels
between
3
and
6,
the
ozone
is
present
primarily
in
its
molecular
form
(
O3),
and
direct
oxidation
dominates.
However,
as
the
pH
rises,
the
autodecomposition
of
ozone
to
produce
the
hydroxyl
free
radical
(
HO°)
becomes
increasingly
rapid.
At
pH
levels
greater
than
10,
the
conversion
of
molecular
O3
to
HO°
is
virtually
instantaneous.
In
general,
better
disinfection
would
be
expected
at
lower
pHs,
since
free
hydroxyl
radicals
are
short­
lived
compared
to
molecular
ozone.
Studies
have
shown
that
increasing
the
temperature
from
0"
to
30"
C
reduces
the
solubility
of
ozone
and
increases
its
decomposition
rate
(
Kinman
1975).

2.2.4.1
Efficacy
Against
Pathogens
Ozone
is
one
of
the
most
potent
biocides
used
in
water
treatment.
It
is
effective
against
a
wide
range
of
pathogenic
microorganisms
including
bacteria,
viruses,
and
protozoa.
Ozone
shows
greater
efficiency
inactivating
most
types
of
pathogenic
microorganisms
than
chlorine,
chloramine,
and
chlorine
dioxide
(
Clark
et
al.
1994).
This
is
demonstrated
by
the
CT
values
found
in
the
SWTR
Guidance
Manual
presented
in
Exhibit
2.7.
The
resistance
of
pathogenic
microorganisms
to
ozone
increases
in
the
following
order:
bacteria,
viruses,
protozoa
(
Camel
and
Bermond
1999).

Exhibit
2.7:
Comparison
of
CT
Values
for
Free
Chlorine
and
Ozone
Log
Remova
l
Giardia
Viruses
<
1o
C
10o
C
20o
C
<
1o
C
10o
C
20o
C
Cl
O3
Cl
O3
Cl
O3
Cl
O3
Cl
O3
Cl
O3
0.5
40
0.48
21
0.23
10
0.12
­­
­­
­­
­­
­­
­­

1
79
0.97
42
0.48
21
0.24
­­
­­
­­
­­
­­
­­

2
158
1.9
83
0.95
41
0.48
6
0.9
3
0.5
1
0.25
3
237
2.9
125
1.43
62
0.72
9
1.4
4
0.8
2
0.4
Note:
­­
Data
not
available
Source:
USEPA
(
1990)

Small
concentrations
of
ozone
are
usually
effective
against
bacteria.
E.
Coli
levels
were
reduced
by
4
log
(
99.99
percent
removal)
in
less
than
one
minute
at
an
initial
ozone
concentration
of
9
micrograms
per
liter
(
µ
g/
L)
(
Wuhrmann
and
Meyrath
1955).
Legionella
pneumophila
levels
were
reduced
by
2
log
(
99
percent
removal)
in
less
than
five
minutes
at
an
initial
ozone
concentration
of
0.21
milligrams
per
liter
(
mg/
L)
(
Domingue
et
al.
1988).
Technologies
and
Costs
for
Control
of
Microbial
Contaminants
and
Disinfection
Byproducts
June
2003
2­
18
Typically,
viruses
are
more
resistant
to
ozone
than
bacteria,
although
ozone
is
still
effective
against
viruses.
Ozone
dosages
of
0.2
to
1.5
mg/
L
consistently
achieved
2
log
inactivation
of
poliomyelitis
viruses
with
a
contact
time
of
40
seconds
(
Katzenelson
et
al.
1974).
Katzenelson
et
al.
(
1974)
also
observed
that
poliomyelitis
inactivation
increased
to
nearly
5
log
at
a
dose
of
1.5
mg/
L
and
a
contact
time
of
approximately
100
seconds.
Coxsackie
virus
inactivation
is
more
than
5
log
with
an
initial
ozone
dosage
of
1.45
mg/
L
(
Keller
et
al.
1974).
The
sensitivity
of
human
rotavirus
to
ozone
was
found
to
be
similar
to
that
of
coxsackie
virus
(
Vaughn
et
al.
1987).

Protozoan
cysts
are
more
resistant
to
ozone
than
bacteria
and
viruses.
Data
available
for
inactivation
of
Cryptosporidium
oocysts
suggest
that,
among
protozoans,
this
pathogen
is
the
most
resistant
to
ozone
(
Peeters
et
al.
1989;
Langlais
et
al.
1990).

Ozone
inactivation
kinetics
of
Cryptosporidium
are
evaluated
by
Gyurek
et
al.
(
1999).
The
observed
inactivation
behavior
of
Cryptosporidium
by
ozone
is
characterized
by
a
"
tailing­
off"
effect.
At
22"
C
and
a
5
minute
contact
time,
an
initial
ozone
residual
of
1.2
mg/
L
was
required
to
provide
2
log
inactivation.
For
contact
times
less
than
5
minutes,
a
relatively
small
increase
in
the
applied
contact
time
significantly
decreases
the
required
initial
ozone
residual;
however,
for
contact
times
greater
that
10
minutes
an
increase
in
the
applied
contact
time
provides
a
negligible
decrease
in
the
required
initial
ozone
residual.
Hence,
the
influence
of
contact
time
on
the
inactivation
kinetics
decreases
as
Cryptosporidium
is
progressively
exposed
to
ozone.

CT
values
for
ozone
inactivation
of
Cryptosporidium
are
still
under
development.
Initial
studies
have
demonstrated
that
CT
values
may
be
as
much
as
25
times
higher
than
those
required
for
Giardia
(
Rennecker
et
al.
1999).
These
preliminary
studies
also
demonstrate
that
CT
requirements
for
Cryptosporidium
inactivation
increase
by
an
average
factor
of
approximately
three
for
every
10"
C
decrease
in
temperature.
A
summary
of
reported
ozonation
requirements
for
2
log
inactivation
of
Cryptosporidium
oocysts
is
presented
in
Exhibit
2.8.
Technologies
and
Costs
for
Control
of
Microbial
Contaminants
and
Disinfection
Byproducts
June
2003
2­
19
Exhibit
2.8:
Reported
Ozonation
Requirements
for
2
log
Inactivation
of
Cryptosporidium
Oocysts
Experimental
Protocol
Initial
Ozone
Residual
(
mg/
L)
Temperature
(
0C)
Contact
Time
(
min)
CT
(
mg­
min/
L)
Reference
Batch
liquid,
batch
ozone
0.77
0.51
Ambient
6
8
4.6
4.0
Peeters
et
al.
1989
Batch
liquid,
continuous
gas
1.0
25
5­
10
5­
10
Korich
et
al.
1990
Batch
liquid,
batch
ozone
0.50
0.50
7
22
18
7.8
9.0
3.9
Finch
et
al.
1993
Flow
through
contactor,
continuous
gas
22­
25
7.4
5.5
Owens
et
al.
1994
Batch
liquid,
batch
ozone
0.7
22
3.2
3.2
Gyurek
et
al.
1999
Note:
Owens
et
al.
do
not
report
residual
dose.

2.2.4.2
DBP
Formation
Ozone
does
not
produce
chlorinated
DBPs;
however,
through
the
oxidation
of
natural
organic
precursor
materials,
ozone
can
alter
the
reactions
between
chlorine
and
NOM
and
affect
the
formation
of
chlorinated
DBPs
when
chlorine
is
added
downstream.
Ozonation
of
natural
waters
produces
aldehydes,
haloketones,
ketoacids,
carboxylic
acids,
and
other
types
of
biodegradable
organic
material
which
must
be
adequately
controlled
(
often
with
a
granular
media
biofilter).

Ozonation
often
increases
the
biodegradability
of
NOM
in
the
treated
water.
Increasing
biodegradability
could
be
beneficial
if
a
biological
filtration
process
follows
the
ozonation
step.
A
biological
filtration
step
can
remove
the
biodegradable
fraction
of
NOM,
increasing
organic
precursor
removal.
Biological
filters
remove
NOM
by
using
it
as
a
substrate.
Biological
filtration
can
be
employed
on
adsorptive
media,
such
as
GAC,
and/
or
non­
adsorptive
media,
such
as
sand
and
anthracite.
Conversely,
if
the
biodegradable
fraction
is
not
removed,
it
can
increase
the
regrowth
of
microorganisms
in
the
distribution
system.

Haag
and
Hoigne
(
1983)
have
shown
that
ozone
oxidizes
bromide
to
form
hypobromous
acid
and
hypobromite
(
HOBr
and
OBr­)
under
water
treatment
conditions.
Hypobromite
was
found
to
be
further
oxidized
to
bromate
or
to
a
species
that
regenerates
bromide,
whereas
HOBr
reacts
with
NOM
to
form
brominated
organic
byproducts
in
waters
containing
bromide.
Technologies
and
Costs
for
Control
of
Microbial
Contaminants
and
Disinfection
Byproducts
June
2003
2­
20
Changes
in
pH
can
have
a
dramatic
effect
on
the
concentrations
of
HOBr
and
OBr­
and,
therefore,
the
species
of
byproducts
formed.
An
increase
in
pH
increases
the
relative
concentration
of
Br­,
which,
in
turn,
leads
to
increased
bromate
formation.
Reduced
pH
levels
are
often
accompanied
by
a
reduction
in
bromate
concentrations;
the
lower
pH
enhances
formation
of
bromoform
and
other
organic
brominated
DBPs.

Krasner
et
al.
(
1989)
found
that
an
ozone
residual
is
necessary
to
produce
detectable
levels
of
bromate.
Siddiqui
and
Amy
(
1993)
found
that
the
bromoform
concentration
first
increased
then
diminished
at
higher
dosages.
Song
et
al.
(
1995)
demonstrated
that
lower
ozone
dosage
and
longer
contact
time
should
produce
less
bromate
than
higher
dosages
and
shorter
contact
times.

Halogenated
organic
compounds
are
formed
when
NOM
reacts
with
free
chlorine
or
free
bromine.
Free
bromine
can
be
formed
in
ozone
disinfection
whenever
bromide
is
present
in
the
raw
water
source.
The
level
of
brominated
byproducts
formed
during
oxidation
is
dependent
on
the
concentration
of
bromide
in
the
raw
water
source
and/
or
the
relative
amount
of
bromide
present
compared
to
organic
precursors.

Ozonation
followed
by
chlorination
has
been
observed
to
produce
higher
levels
of
haloketones
than
chlorination
alone
(
Jacangelo
et
al.
1989b).
Chloral
hydrate
occurs
primarily
as
a
result
of
chlorination,
although
ozonation
followed
by
chlorination
has
been
observed
to
increase
levels
beyond
those
observed
with
chlorination
only.
Ozonation
followed
by
chlorination
or
chloramination
can
increase
chloropicrin
levels
above
those
observed
with
chlorination
or
chloramination
alone.
Ozonation
followed
by
chloramination
has
been
observed
to
increase
cyanogen
chloride
levels
beyond
those
observed
with
chloramination
only.
Cyanogen
bromide,
the
brominated
analog
of
cyanogen
chloride,
has
been
detected
after
ozonation
of
water
containing
high
bromide
levels
(
McGuire
et
al.
1990).

Much
less
is
known
about
non­
halogenated
disinfection
byproducts
than
the
halogenated
organic
compounds.
Among
the
major
ozonation
byproducts,
aldehydes
and
carboxylic
acids
have
the
highest
concentrations
(
Glaze
et
al.
1993).
Ozonation
followed
by
chlorination
has
been
found
to
yield
the
highest
levels
of
acetaldehyde
and
formaldehyde.
In
addition,
ozonation
prior
to
chloramination
is
shown
to
produce
more
of
these
aldehydes
than
chloramination
alone.
Najm
and
Krasner
(
1995)
report
that
the
formation
of
ketoacids
is
proportional
to
the
amount
of
dissolved
organic
carbon
(
DOC)
in
the
water.
Ketoacid
concentrations
are
largely
unaffected
by
bromide
concentration.

Ammonia
addition
has
been
used
to
limit
the
formation
of
some
ozonation
byproducts.
In
one
study
(
Siddiqui
and
Amy
1993),
bromoform
concentrations
decrease
by
approximately
30
percent
when
ammonia
is
added
at
a
NH3­
to­
ozone
ratio
of
0.25
mg/
mg.
The
reason
for
this
reduction
is
because
HOBr
reacts
with
ammonia
to
form
bromamines,
presumably
making
HOBr
unavailable
for
reaction
with
NOM.

Conflicting
results
of
ammonia
addition
on
bromate
formation
have
been
observed
(
Glaze
et
al.
1993,
Krasner
et
al.
1993).
Siddiqui
et
al.
(
1995)
explained
the
percentage
of
bromate
reduction
upon
Technologies
and
Costs
for
Control
of
Microbial
Contaminants
and
Disinfection
Byproducts
June
2003
2­
21
adding
ammonia
is
more
dependent
upon
pH
and
bromide
concentration
than
on
ammonia
concentration
(
Siddiqui
et
al.
1995).
High
bromide
levels
trap
more
oxidizing
equivalents
to
give
higher
bromine
yields
and
scavenge
more
radicals,
thus
reducing
the
radical
processes
that
may
cause
bromate
formation.
Siddiqui
et
al.
(
1995)
demonstrated
that
(
at
similar
ammonia
concentrations)
bromate
formation
decreased
by
more
than
80
percent
upon
increasing
the
bromide
concentration
from
0.1
to
1.0
mg/
L.

2.2.4.3
Factors
Affecting
Performance
Ozone
decays
rapidly
at
high
pH
and
warm
temperatures.
Krasner
et
al.
(
1993)
noted
that
as
the
ozonation
pH
decreases,
the
required
dose
to
meet
CT
requirements
of
the
IESWTR
drops
and
less
bromate
is
formed.
For
one
of
the
waters
evaluated
during
bromide
spiking
experiments,
bromate
concentrations
ranged
from
24
to
68
µ
g/
L
at
pH
8.
For
the
same
water,
bromate
concentrations
ranged
from
less
than
5
to
7
µ
g/
L
when
the
pH
was
decreased
to
6.
Better
disinfection
is
expected
at
pH
levels
between
6
and
8
where
molecular
ozone
dominates.

Temperature
and
alkalinity
also
affect
formation
of
byproducts
during
ozonation.
Increased
temperature
will
increase
the
levels
of
bromate,
bromoform,
and
total
organic
bromide.
It
also
increases
the
decomposition
of
ozone.
Conversely,
increasing
alkalinity
has
been
shown
to
reduce
the
formation
of
bromoform
and
total
organic
bromide
and
increase
the
formation
of
bromate.
Bicarbonate
scavenges
OH
radicals,
suggesting
that
the
OH
radical
may
play
a
role
in
the
formation
of
brominated
species
by
affecting
the
level
of
HOBr,
which
is
presumed
to
be
an
active
species
for
total
organic
bromide
formation
(
Glaze
et
al.
1993).

Total
organic
carbon
(
TOC)
concentration
can
have
significant
impacts
on
Cryptosporidium
CT
requirements.
It
has
been
demonstrated
that
ozone­
to­
TOC
ratios
greater
than
1
are
required
for
Cryptosporidium
inactivation;
whereas
ozone­
to­
TOC
ratios
are
typically
less
than
0.5
for
Giardia
inactivation.
As
previously
discussed,
temperature
can
also
drastically
affect
the
solubility,
decomposition
rate
and
biocidal
effectiveness
of
ozone.
Exhibit
2.9
presents
CT
requirements
for
Cryptosporidium
inactivation
at
multiple
temperatures
and
for
inactivation
ranging
from
0.5
to
3
log.
Exhibit
2.9
also
compares
the
Cryptosporidium
CT
requirements
with
those
of
Giardia
and
presents
the
ratio
of
the
Cryptosporidium
requirement
to
the
Giardia
requirement.
Technologies
and
Costs
for
Control
of
Microbial
Contaminants
and
Disinfection
Byproducts
June
2003
2­
22
Exhibit
2.9:
CT
Considerations
for
Cryptosporidium
Inactivation
Log
Inactivation
Crypto
CT
at
Temperature
(
C)
1
Giardia
CT
at
Temperature
(
C)
2
Multiplier
at
Temperature
(
C)
3
1
°
13
°
22
°
1
°
13
°
22
°
1
°
13
°
22
°
0.5
6.00
2.00
0.60
0.48
0.19
0.10
12.5
10.6
5.8
1.0
12.00
4.00
1.50
0.97
0.38
0.21
12.4
10.4
7.2
1.5
24.00
8.00
3.00
1.50
0.58
0.31
16.0
13.9
9.6
2.0
40.00
11.00
4.40
1.90
0.76
0.42
21.1
14.5
10.6
2.5
45.00
15.00
6.00
2.40
0.95
0.52
18.8
15.7
11.5
3.0
62.00
22.00
8.00
2.90
1.14
0.62
21.4
19.3
12.8
1
Values
reported
to
be
acceptable
for
a
pH
range
of
6
to
9.
2
Giardia
CT
required
numbers
are
based
upon
the
CT
table
included
in
the
SWTR
Guidance
Manual.
3
Multiplier
=
Crypto
CT
at
a
given
temperature
/
Giardia
CT
at
the
same
temperature.

Source:
Summary
from
Finch
1999.

2.2.5
Microfiltration
and
Ultrafiltration
Membranes
act
as
selective
barriers,
allowing
some
constituents
to
pass
through
while
blocking
the
passage
of
others.
The
movement
of
these
constituents
across
a
membrane
requires
a
driving
force
(
i.
e.,
to
overcome
the
potential
difference
across
the
membrane).
Only
pressure­
driven
processes
are
discussed
in
this
document
due
to
their
feasibility
for
DBP
precursor
and
microbial
control
and
their
popularity
in
the
drinking
water
field.

There
are
four
categories
of
pressure­
driven
membrane
processes:
microfiltration,
ultrafiltration
(
UF),
nanofiltration,
and
reverse
osmosis
(
RO).
Low­
pressure
membrane
processes,
MF
and
UF,
are
typically
applied
for
the
removal
of
particulate
and
microbial
contaminants
and
can
be
operated
under
positive
or
negative
(
i.
e.,
vacuum)
pressure.
Positive
pressure
systems
typically
operate
between
3
and
40
pounds
per
square
inch
(
psi),
whereas
vacuum
systems
operate
between
­
3
and
­
12
psi.
RO
and
NF
are
typically
applied
for
the
removal
of
dissolved
contaminants,
including
both
inorganic
and
organic
compounds.
These
processes
operate
at
pressures
significantly
greater
than
the
applied
pressure
in
MF
and
UF
processes,
between
100
and
150
psi.
Desalination
applications
can
operate
at
pressures
as
high
as
1,200
to
1,500
psi.

The
ability
of
a
membrane
to
remove
a
particular
contaminant
is
influenced
by
its
molecular
weight
cut­
off
(
MWCO)
or
pore
size.
MWCO
is
a
manufacturer
specification
that
refers
to
the
molecular
mass
of
a
macrosolute
(
e.
g.,
glycol
or
protein)
for
which
a
membrane
has
a
retention
capacity
greater
than
90
percent.
The
pore
size
refers
to
the
diameter
of
the
micropores
on
the
membrane
surface.
The
true
pore
size
is
difficult
to
measure,
and,
as
a
result,
membrane
manufacturers
typically
use
some
measure
of
performance
to
categorize
the
pore
size
of
a
membrane.
The
nominal
pore
size
is
typically
based
upon
a
given
percent
removal
of
a
marker
(
e.
g.,
microsphere)
of
a
known
diameter.
The
absolute
pore
size
is
typically
characterized
as
the
minimum
diameter
above
which
100
percent
of
a
Technologies
and
Costs
for
Control
of
Microbial
Contaminants
and
Disinfection
Byproducts
June
2003
2­
23
0.001
µ
0.01
µ
0.
1
µ
1.0
µ
10
µ
100
µ
1000
µ
Dissolved
Organics
Bacteria
Sand
Cysts
Viruses
Salts
Colloids
Media
Filtration
Microfiltration
Ultrafiltration
Nanofiltration
Reverse
Osmosis
marker
of
a
specific
size
is
removed
by
the
membrane.
Exhibit
2.10
presents
the
MWCO/
pore
size
ranges
for
membrane
processes,
as
well
as
the
relative
size
of
common
drinking
water
contaminants.

MF
and
UF
are
primarily
used
for
particle
and
microbial
removal,
either
following
granular
media
filtration
or
as
a
replacement
for
media
filters.
Chemical
disinfection
may
be
required,
depending
upon
the
approach
of
the
State
regulatory
agency
and
the
class
of
membrane
used
(
i.
e.,
MF
or
UF).
MF
pore
sizes
are
generally
too
large
for
virus
removal
and
many
States
require
a
minimum
0.5
log
chemical
inactivation
as
part
of
a
multiple
barrier
approach
to
disinfection.

The
major
components
of
a
typical
MF
or
UF
membrane
system
include
cartridge
filters,
low
pressure
feed
pumps,
membrane
modules,
high­
pressure
backwash
pumps,
a
chemical
cleaning
system,
a
chlorination
feed
system,
and
a
concentrate
handling
and
disposal
system.

Exhibit
2.10:
Pressure­
Driven
Membrane
Separation
Spectrum
Note:
µ
=
Microns.

2.2.5.1
Efficacy
Against
Pathogens
MF
and
UF
have
shown
excellent
capabilities
in
turbidity,
particulate
matter,
and
microbial
removal.
MF
and
UF
processes
remove
contaminants
through
physical
straining
of
the
feed
water
as
it
passes
through
the
membrane.
In
this
respect,
microbial
contaminants
that
are
larger
than
a
given
membrane
pore
will
be
retained
and
prevented
from
entering
the
treated
water.
Since
the
size
and
Technologies
and
Costs
for
Control
of
Microbial
Contaminants
and
Disinfection
Byproducts
June
2003
2­
24
shape
of
microorganisms
varies
among
species
and
since
the
size
and
shape
of
membrane
pores
varies
among
membrane
types,
the
removal
of
a
particular
microorganism
by
MF
and
UF
may
vary.
Many
States
have
adopted
disinfection
log
removal
credits
for
MF
and
UF
processes.
States
grant
removal
credits
on
a
case­
by­
case
basis
for
up
to
3
log
Giardia
removal
and
4
log
virus
removal.
However,
virus
removal
credits
are
typically
0.5
log
or
less
due
to
the
smaller
size
of
viruses
relative
to
MF/
UF
pores.

MF
and
UF
offer
disinfection
capabilities
that
are
much
improved
over
conventional
media
filtration.
Exhibits
2.11
through
2.14
summarize
observed
removals
of
bacteria,
Giardia,
Cryptosporidium,
and
viruses,
respectively.

Exhibit
2.11:
MF
and
UF
Studies
Documenting
Bacteria
Removal
Reference
Process
Membrane
Pore
Size
Bacteria
Type
Log
Removal
Hofmann
et
al.
(
1998)
MF
150,000
to
200,000
Daltons
HPC,
coliforms,
thermotolerant
coliforms,
SSRC
2.5
to
3.5
Jacangelo
et
al.
(
1997)
MF
100,000
Daltons
P.
Aeruginosa
>
8.7*

Jacangelo
et
al.
(
1997)
MF
0.2
:
m
P.
Aeruginosa
>
8.2*

Jacangelo
et
al.
(
1997)
MF
0.2
:
m
Coliforms
>
1.8*

Jacangelo
et
al.
(
1997)
MF
0.2
:
m
E.
Coli
>
7.8*

Jacangelo
et
al.
(
1997)
MF
0.2
:
m
HPC
>
1.8*

Clair
et
al.
(
1997)
MF
0.2
:
m
HPC
2.4
Clair
et
al.
(
1997)
MF
0.2
:
m
Total
Coliforms
>
3
Glucina
et
al.
(
1997)
MF
0.2
:
m
HPC
and
total
Coliforms
>
3
Glucina
et
al.
(
1997)
UF
100,000
Daltons
Total
Coliforms
>
3
Jacangelo
et
al.
(
1997)
UF
100,000
Daltons
Coliforms
>
2.1*

Jacangelo
et
al.
(
1997)
UF
100,000
Daltons
E.
Coli
>
7.8*

Luitweiler
(
1991)
MF
­­
HPC
1.7
Jacangelo
et
al.
(
1991)
UF
­­
Total
Coliforms
>
3
Heneghan
and
Clark
(
1991)
UF
­­
HPC
>
3.4
Jacangelo
et
al.
(
1989a)
UF
­­
HPC
2.8
Note:
*
Indicates
removal
to
detection
limit.
­­
Data
not
available.
Technologies
and
Costs
for
Control
of
Microbial
Contaminants
and
Disinfection
Byproducts
June
2003
2­
25
Exhibit
2.12:
MF
and
UF
Studies
Documenting
Giardia
Removal
Reference
Process
Membrane
Pore
Size
Log
Removal
Scheider
et
al.
(
1999)
MF
0.2
:
m
>
4.8
Scheider
et
al.
(
1999)
MF
0.1
:
m
>
4.8*

Scheider
et
al.
(
1999)
MF
0.1
:
m
>
4.8*

Trussel
et
al.
(
1998)
MF
0.2
:
m
>
5.1*

Jacangelo
et
al.
(
1997)
MF
0.2
:
m
>
5.2*

Jacangelo
et
al.
(
1997)
MF
0.2
:
m
>
6.8*

Hagen
(
1998)
UF
100,000
Daltons
>
8*

Trussel
et
al.
(
1998)
UF
100,000
Daltons
>
5.1*

Jacangelo
et
al.
(
1997)
UF
100,000
Daltons
>
5.2*

Jacangelo
et
al.
(
1997)
UF
100,000
Daltons
>
6.8*

Jacangelo
et
al.
(
1991)
UF
 
>
4*

Jacangelo
et
al.
(
1989a)
UF
100,000
Daltons
>
5*

Note:
*
Indicates
removal
to
detection
limit.
­­
Data
not
available.

Exhibit
2.13:
MF
and
UF
Studies
Documenting
Cryptosporidium
Removal
Reference
Process
Membrane
Pore
Size
Log
Removal
Scheider
et
al.
(
1999)
MF
0.2
:
m
4.2
Scheider
et
al.
(
1999)
MF
0.1
:
m
>
4.2
Scheider
et
al.
(
1999)
MF
0.1
:
m
>
4.2
Trussel
et
al.
(
1998)
MF
0.2
:
m
>
4.7*

Jacangelo
et
al.
(
1997)
MF
0.2
:
m
>
4.9*

Jacangelo
et
al.
(
1997)
MF
0.2
:
m
>
6.4*

Trussel
et
al.
(
1998)
UF
100,000
Daltons
>
5.1*

Hagen
(
1998)
UF
100,000
Daltons
>
8*

Jacangelo
et
al.
(
1997)
UF
100,000
Daltons
>
4.9*

Jacangelo
et
al.
(
1997)
UF
100,000
Daltons
>
6.4*

Jacangelo
et
al.
(
1989a)
UF
100,000
Daltons
>
5*

Jacangelo
et
al.
(
1997)
UF
100,000
Daltons
>
6.4*

Jacangelo
et
al.
(
1997)
UF
100,000
Daltons
>
6.4*

Note:
*
Indicates
removal
to
detection
limit.
Technologies
and
Costs
for
Control
of
Microbial
Contaminants
and
Disinfection
Byproducts
June
2003
2­
26
Exhibit
2.14:
MF
and
UF
Studies
Documenting
Virus
Removal
Reference
Process
Membrane
Pore
Size
Log
Removal
Scheider
et
al.
(
1999)
MF
0.2
:
m
0.5
Scheider
et
al.
(
1999)
MF
0.1
:
m
1.1
Scheider
et
al.
(
1999)
MF
0.1
:
m
2.3
Trussel
et
al.
(
1998)
MF
0.2
:
m
0.4
to
3.1
Jacangelo
et
al.
(
1997)
MF
0.2
:
m
>
1
Jacangelo
et
al.
(
1997)
MF
0.2
:
m
>
1.5
Kruithof
et
al.
(
1997)
MF
­­
0.7
to
2.3
Trussel
et
al.
(
1998)
UF
100,000
Daltons
>
6.9*

Jacangelo
et
al.
(
1997)
UF
100,000
Daltons
>
6
Kruithof
et
al.
(
1997)
UF
­­
>
5.4
Jacangelo
et
al.
(
1989a)
UF
100,000
Daltons
>
8*

Jacangelo
et
al.
(
1989a)
UF
­­
>
6
Note:
*
Indicates
removal
to
detection
limit.

­­
Data
not
available.

As
shown
in
Exhibits
2.11
through
2.14,
both
MF
and
UF
systems
are
capable
of
significant
log
removal
of
bacteria,
Giardia
cysts,
and
Cryptosporidium
oocysts.
The
data
presented
indicate
that
MF/
UF
are
capable
of
bacteria
removals
of
nearly
9
log
and
Giardia
and
Cryptosporidium
removals
in
excess
of
8
log.
In
fact,
in
nearly
all
cases,
the
log
removal
demonstrated
is
simply
a
function
of
the
influent
microbe
concentration,
since
bacteria
and
cysts
are
typically
removed
to
detection
limits.
As
shown
in
Exhibit
2.14,
however,
MF
and
UF
are
differentiated
by
virus
removal.
The
maximum
virus
removal
reported
for
MF
membranes
is
approximately
3
log,
but
the
average
reported
removal
is
nearer
to
1
log.
UF
membranes
typically
remove
viruses
to
detection
limits.

Note
that
the
studies
summarized
in
Exhibits
2.11
through
2.14
are
conducted
with
intact
membranes
(
i.
e.,
the
membranes
are
not
compromised).
Had
a
fiber
from
one
of
these
membranes
been
broken,
either
deliberately
or
accidentally,
the
results
could
be
significantly
different,
since
the
potential
would
exist
for
microorganisms
to
pass
into
the
treated
water.
For
this
reason,
it
is
important
to
include
membrane
integrity
testing
when
assessing
the
ability
of
a
membrane
to
act
as
a
barrier
against
microorganisms.
Many
types
of
membrane
integrity
tests
exist.
These
tests
fall
into
two
categories:
1)
direct
methods
and
2)
indirect
methods.
Indirect
methods
include
monitoring
the
treated
water
for
parameters
such
as
particle
counts
or
turbidity.
Direct
methods
include
tests,
such
as
air
pressure
decay
and
diffusive
airflow,
that
directly
assess
the
integrity
of
the
membrane
itself.
Integrity
testing
represents
an
important
aspect
of
a
membrane
system
from
a
regulatory
perspective,
since
it
provides
a
measurement
of
the
integrity
of
the
filter.
Commercial
manufacturers
have
recognized
this,
and
most
systems
are
now
provided
with
automatic
integrity
testing
that
can
be
conducted
frequently
(
e.
g.,
hourly).
Technologies
and
Costs
for
Control
of
Microbial
Contaminants
and
Disinfection
Byproducts
June
2003
2­
27
2.2.5.2
DBP
Formation
Disinfection
by
MF/
UF
is
achieved
through
physical
removal.
Because
of
this,
no
DBPs
are
formed
during
disinfection
by
MF/
UF.
Chlorine
or
chloramines
must
be
added
subsequent
to
MF/
UF
to
maintain
a
disinfectant
residual.
Chlorination
and
chloramination
can
produce
DBPs
as
discussed
in
section
2.1.2.

MF
and
UF
alone
are
generally
not
effective
for
DBP
precursor
removal.
The
pore
sizes
are
typically
large
enough
to
allow
most
NOM
to
pass
through
these
membranes,
thus
removing
little
NOM.
Some
tight
UF
membranes
with
MWCOs
on
the
order
of
10,000
Daltons
may
be
capable
of
removing
some
NOM,
but
significant
NOM
removal
cannot
be
achieved
by
MF
or
UF
alone.
MF/
UF
systems
may
be
combined
with
other
processes
to
aid
in
removing
DBP
precursors.
By
associating
the
NOM
with
a
filterable
particulate
matter
(
e.
g.,
powder
activated
carbon
(
PAC)
or
coagulant
floc),
the
membranes
can,
in
effect,
reject
some
NOM.
Adsorption
of
organics
onto
PAC
depends
on
the
type
and
dose
of
PAC,
the
contact
time
available,
and
the
type
of
NOM.
Similarly,
the
efficiency
of
incorporating
NOM
into
coagulant
flocs
depends
on
the
type
and
dose
of
coagulant,
the
operating
conditions,
and
the
type
of
NOM.

2.2.5.3
Factors
Affecting
Performance
Membrane
pore
size
greatly
affects
microorganism
removal.
To
illustrate
this,
Exhibit
2.10
shows
the
size
of
several
microbes
of
concern
against
different
membrane
filtration
options.
As
shown
in
Exhibit
2.10,
cysts
(
including
Giardia
and
Cryptosporidium)
are
larger
than
essentially
all
MF
and
UF
pore
sizes.
Consequently,
these
processes
are
capable
of
large
log
removal
of
cysts.
On
the
other
hand,
as
shown
in
Exhibit
2.10,
viruses
are
larger
than
most
UF
pore
sizes,
but
smaller
than
most
MF
pore
sizes.
For
this
reason,
UF
is
capable
of
large
virus
removal,
but
MF
typically
is
not.

Membrane
pores
are
typically
a
distribution
of
sizes
(
Mallevialle
et
al.
1996),
only
as
accurate
as
the
manufacturing
process
allows.
At
the
present
time,
no
precise
techniques
for
membrane
pore
size
determination
are
available.
For
these
reasons,
a
membrane
of
a
given
MWCO
may
have
pores
that
are
larger
and
smaller
than
the
given
MWCO.
Imperfections
in
the
membrane
module
or
membrane
system
may
result
in
passage
of
microorganisms
into
the
treated
water.

Imperfections
can
arise
through
manufacturing
imprecision,
allowing
microbes
to
penetrate
orings
end
seals,
or
spacers.
Conversely,
microbial
contaminant
removal
may
be
increased
by
the
cake
layer,
which
forms
on
the
membrane
surface
during
a
filtration
cycle.
This
cake
layer
consists
of
contaminants
rejected
by
the
membrane,
including
particles,
organic
matter,
and
microorganisms.
As
this
layer
builds,
it
can
aid
filtration
of
suspended
particulates,
such
as
microorganisms,
as
water
passes
across
the
membrane.
In
this
way,
microorganisms
that
might
normally
pass
through
a
membrane
pore
can
be
filtered
from
the
feed
water
stream.
Technologies
and
Costs
for
Control
of
Microbial
Contaminants
and
Disinfection
Byproducts
June
2003
2­
28
One
of
the
critical
design
parameters
for
a
membrane
process
is
flux,
which
is
typically
expressed
in
gallons
of
filtrate
per
day
per
square
foot
of
membrane
area
(
gfd).
The
design
flux
determines
the
membrane
area
required
for
a
specific
plant
capacity.
Thus,
flux
has
a
significant
impact
on
capital
cost
and
results
in
a
competitive
motivation
for
design
engineers
to
use
a
higher
membrane
flux,
thereby
reducing
the
area
requirements.
Although
increasing
the
membrane
flux
can
reduce
the
capital
cost,
it
will
increase
operational
costs
due
to
higher
operating
pressure,
more
frequent
chemical
cleaning,
and
a
potential
increase
in
membrane
replacement
costs.

Another
important
design
parameter
is
recovery,
the
ratio
of
feed
water
to
product
water.
Recovery
for
MF
and
UF
systems
is
typically
85
to
97
percent,
and
a
function
of
the
backwash
method
and
frequency.
Recovery
can
play
a
significant
role
in
the
design
of
membrane
facilities,
particularly
in
water­
scarce
regions.

Feed
water
quality
can
also
have
a
significant
impact
on
membrane
system
design,
operation,
and
performance.
Suspended
solids
and
other
contaminants
(
e.
g.,
iron,
calcium,
barium,
or
silica)
can
result
in
more
rapid
fouling
of
the
membrane,
decreases
in
flux,
and
increases
in
transmembrane
pressure
(
TMP).
TMP
is
the
pressure
applied
to
drive
water
through
the
membrane.
As
a
result,
most
membrane
systems
include
some
level
of
pretreatment
to
reduce
the
concentration
of
these
foulants,
with
the
level
of
pretreatment
dependent
upon
raw
water
quality.

2.2.6
Bag
and
Cartridge
Filtration
Like
MF
and
UF,
bag
and
cartridge
filters
act
as
selective
barriers
and
are
used
to
remove
particles,
including
pathogens,
in
water
treatment.
As
water
passes
through
the
bag
or
porous
cartridge,
particulate
matter
and
organisms
whose
size
exceeds
the
largest
pore
size
are
retained
on
the
filter.
The
nature
of
the
filter
material
and
the
direction
of
flow
are
two
features
that
differentiate
bag
from
cartridge
filtration
(
AWWA
1999).

Bag
filters
can
be
either
woven
or
felt
and
made
of
materials
such
as
polypropylene,
polyester,
nylon,
or
teflon.
Typically,
only
felt
filters
will
display
nominal
pore
size
ratings
as
low
as
0.5
to
1
:
m,
which
are
values
likely
to
be
associated
with
high
removal
of
pathogens.
Bag
filters
can
also
comprise
a
sealing
system
on
their
open
end
in
order
to
ensure
flow
integrity
between
the
water
inlet
and
the
bag
filter.

The
bag
is
housed
in
a
pressure
vessel
and
supported
by
a
mesh
basket.
The
pressure
vessel
is
made
of
carbon
steel
or
stainless
steel.
The
water
flow
is
from
inside
the
bag
filter
to
outside.
As
filtered
material
(
i.
e.,
suspended
solids)
accumulates
on
the
filter
surface,
head
loss
increases,
and
a
pressure
differential
develops
between
both
sides
of
the
filter.
Technologies
and
Costs
for
Control
of
Microbial
Contaminants
and
Disinfection
Byproducts
June
2003
2­
29
A
number
of
bag
filter
configurations
are
commercially
available.
Pressure
vessels
exist
in
single,
duplex
or
four­
plex,
series
or
parallel
modules,
or
as
multi­
filter
vessels.
Manufacturers
claim
that
a
single
vessel
can
filter
flow
rates
from
10
to
approximately
2,000
gallons
per
minute
(
gpm),
depending
on
its
configuration.
The
standard
pressure­
rating
for
vessels
has
been
observed
to
be
150
psi.

Cartridge
filters
are
typically
composed
either
of
fiberglass
or
ceramic
membranes
supported
by
a
rigid
core
or
are
made
from
strings
of
polypropylene,
acrylics,
nylon,
or
cotton
wrapped
around
a
filter
element.
Nominal
pore
size
ratings
generally
range
from
0.3
to
200
microns.
With
regard
to
membranes,
the
number
of
pleats
in
a
cartridge
filter
is
typically
larger
relative
to
a
bag
filter,
thus
providing
greater
surface
area.
The
cartridge
is
housed
in
a
pressure
vessel
made
of
carbon
steel
or
stainless
steel,
similar
to
the
bag
filter,
but
the
direction
of
the
flow
is
from
the
outside
to
the
inside
of
the
cartridge.
Accumulation
of
particulate
matter
on
the
surface
and
in
the
depth
of
the
cartridge
element
leads
to
increased
pressure
loss
across
the
cartridge.
Operation
of
the
cartridge
filter
beyond
the
recommended
maximum
pressure
drop
would
result
in
the
structural
failure
of
the
cartridge
and
potential
damage
to
the
cartridge
filter
vessel.

Commercially
available
cartridge
filter
single
vessels
allow
for
housing
of
1
to
approximately
200
cartridges.
It
is
possible
to
connect
these
vessels
in
series
(
for
multiple­
stage
filtration)
or
parallel
(
for
treatment
capacity
increase
and/
or
continuous
operation).

2.2.6.1
Efficacy
Against
Pathogens
Because
their
mode
of
operation
is
based
on
a
size­
exclusion
mechanism,
bag
and
cartridge
filters
with
the
proper
pore
size
rating
can
remove
Cryptosporidium,
Giardia,
and
other
pathogens,
depending
on
their
size.
Available
studies
assessing
the
efficacy
of
bag
and
cartridge
filters
against
pathogens
have
frequently
utilized
polystyrene
beads
as
surrogates
for
the
Cryptosporidium
oocysts
and
Giardia
cysts
(
Li
et
al.
1997,
Goodrich
et
al.
1995,
Long
1983).
Cysts
and
oocysts
are
suspected
to
fold
and
deform,
eventually
passing
through
filtration
pores
that
are
smaller
than
their
nominal
diameter.
In
an
effort
to
account
for
this
flexibility,
investigators
have
used
polystyrene
beads
smaller
than
the
pathogens
they
represent.

In
a
study
by
Li
et
al.
(
1997),
log
removals
of
Cryptosporidium
oocysts
and
4­
6
µ
m
polystyrene
microspheres
by
bag
filters
were
determined
and
compared.
The
investigators
concluded
a
linear
correlation:
1
log
removal
of
4­
6
µ
m
polystyrene
microspheres
is
equivalent
to
1.040
log
removal
of
Cryptosporidium.
This
is
attributed
to
similar
size
distributions.

The
EPA
Risk
Reduction
Engineering
Laboratory
assessed
the
ability
of
bag
filtration
to
remove
Cryptosporidium
and
surrogates
under
various
flow
(
12.5
and
25
gpm)
and
pressure
drop
(
0,
7,
15,
and
25
psi)
conditions
(
Li
et
al.
1997).
The
study
evaluated
three
polypropylene
bag
filters.
The
surrogates
tested
were
turbidity,
1­
25
:
m
particle
counts,
4­
6
:
m
particle
counts,
and
4­
6
:
m
Technologies
and
Costs
for
Control
of
Microbial
Contaminants
and
Disinfection
Byproducts
June
2003
2­
30
polystyrene
microspheres.
The
study
found
the
polystyrene
microspheres
to
be
"
accurate
and
precise"
indicators
of
filter
performance
with
respect
to
Cryptosporidium.
The
results
of
this
study
are
summarized
in
Exhibit
2.15.

Exhibit
2.15:
Summary
of
Bag
Filter
Performance
Filter
Type
Nominal
Pore
Size
Contaminant
Log
Removal
(
Average)

Multi­
layer
polypropylene
1­
:
m
4.5­
:
m
microspheres
1.14
­
1.88
(
1.39)

Cryptosporidium
1.35
­
1.48
(
1.41)

Single­
layer
polypropylene
1­
:
m
4.5­
:
m
microspheres
0.14
­
0.72
(
0.46)

Cryptosporidium
0.26
­
0.64
(
0.42)

Multi­
layer
polypropylene
99%
removal
of
2.5
:
m
particles,
95%
removal
of
1.5
:
m
particles
4.5­
:
m
microspheres
0.93
­
3.42
(
2.08)

Cryptosporidium
3.00
­
3.63
(
3.29)

Source:
Li
et
al.
(
1997).

The
results
presented
in
Exhibit
2.15
may
indicate
a
benefit
in
removal
efficiency
associated
with
multi­
layering
of
the
filter
fabric.
Based
on
this
study,
a
multi­
layer
fabric
bag
filter
can
achieve
1
to
2
log
Cryptosporidium
removal
under
proper
operation
conditions.
One
interesting
result
of
these
tests
is
that
experimental
controls
performed
with
Cryptosporidium
showed
that
0.1
to
0.2
log
removal
can
be
attributed
to
the
pressure
vessels
alone
without
bag
filters.
This
is
assumed
to
reflect
the
ability
of
Cryptosporidium
oocysts
to
adhere
to
the
surface
walls
of
the
vessel.

Another
study
by
the
Risk
Reduction
Engineering
Laboratory
(
Goodrich
et
al.
1995)
evaluated
cartridge
filters
for
the
removal
of
4­
6
:
m
polystyrene
spheres.
The
results
of
this
study
indicate
that
a
single
cartridge
filter,
with
a
2
:
m
rating,
achieved
an
average
microsphere
removal
of
3.6
log.

A
study
conducted
by
Long
(
1983)
evaluated
the
log
removal
of
17
different
cartridge
filters
for
Giardia
surrogates.
These
cartridge
filters
were
tested
using
the
same
pressure
vessel
at
a
pressure
of
45
psi
and
a
flow
rate
of
0.5
gpm.
The
microspheres
used
as
surrogates
for
Giardia
cysts
had
an
average
diameter
of
5.7
:
m,
with
a
standard
deviation
of
1.5
:
m.
The
filters
were
made
of
a
variety
of
materials
(
cotton,
cellulose,
glass
fiber,
polypropylene,
polyester)
and
configurations
(
majority
pleated
or
spirally
wound).
The
pore
ratings
ranged
from
0.2
to
10.0
:
m.

According
to
a
scanning
electron
microscopy
analysis
that
allowed
visual
counting
of
the
microspheres
passing
through
the
filter,
ten
cartridge
filters
out
of
seventeen
had
a
microsphere
removal
of
99.99
percent
(
4
log
reduction).
The
lower
performances
seemed
to
be
associated
with
the
absence
of
end
seals
on
the
cartridges
and
the
use
of
cotton
or
polyester
as
the
main
filtering
material
(
Long
Technologies
and
Costs
for
Control
of
Microbial
Contaminants
and
Disinfection
Byproducts
June
2003
2­
31
1983).
Note
that
the
tests
were
conducted
at
a
flow
rate
of
0.5
gpm,
which
is
significantly
lower
than
the
expected
operation
flow
rate
(
typically
20
gpm
per
unit).
The
impact
of
this
reduced
flow
rate
on
removal
performance
is
unclear.

2.2.6.2
Factors
Affecting
Performance
Feed
water
quality
is
the
primary
factor
affecting
the
performance
of
both
bag
and
cartridge
filters.
Although
these
filters
can
operate
at
turbidity
levels
from
0.1
to
10
NTU,
it
is
recommended
that
turbidity
be
minimized
to
extend
the
filter
lifetime.
If
turbidity
of
the
feed
water
is
above
1
NTU,
bag
filters
may
operate
properly
for
only
a
few
hours
(
USEPA
1998).
Thus,
use
as
a
secondary
barrier
following
conventional
treatment
is
a
preferred
mode
of
operation.
Granular
media
filters
can
reduce
feed
water
turbidity
to
less
than
0.1
NTU
and
provide
a
feed
water
stream
of
appropriate
quality
for
bag
and
cartridge
filters.

Feed
water
should
also
contain
very
low
levels
of
sand,
silt,
or
algae
to
prevent
clogging
of
the
filters.
If
raw
water
quality
is
such
that
the
concentrations
of
these
parameters
are
high,
pretreatment,
such
as
sand,
multimedia
filters,
or
preliminary
bag
or
cartridge
filters
with
larger
pore
size
(
e.
g.,
10
:
m),
is
encouraged.

The
appropriate
choice
of
the
pore
size
rating
is
an
important
issue.
Giardia
cysts
and
Cryptosporidium
oocysts
are
suspected
to
deform
and
fold,
enabling
them
to
pass
through
pores
that
are
nominally
smaller
than
the
pathogen.
The
selected
pore
size
should
be
sufficient
to
achieve
significant
removal
of
microorganisms
while
maximizing
the
expected
filter
lifetime,
based
upon
raw
water
quality
and
filter
loading.
Likewise,
the
quality
of
the
system's
seals
will
greatly
impact
the
level
of
performance.
The
most
critical
seals
appear
to
be
between
the
filter
and
the
pressure
vessel
and
within
the
structure
of
the
filter
itself.
A
faulty
seal
is
a
way
for
pathogens
to
partially
or
completely
bypass
filtration.

Pilot
testing
(
i.
e.,
challenge
studies)
is
frequently
recommended
to
assess
the
performance
of
bag
and
cartridge
filters.
However,
the
costs
associated
with
pilot
testing,
particularly
for
small
systems,
can
represent
a
significant
portion
of
the
installation
costs.
As
a
result,
pilot
testing
may
not
be
affordable
for
small
systems
and
may
limit
the
use
of
these
technologies
where
pilot
testing
is
necessary.
Some
States
(
e.
g.,
Oregon)
accept
manufacturer
data
regarding
removal
efficiency
and
permit
systems
to
operate
in
a
demonstration
mode,
with
additional
monitoring
requirements.

The
skill
level
required
to
operate
bag
or
cartridge
filters
is
typically
described
as
basic
(
AWWA
1999,
Campbell
et
al.
1995a).
Turbidity,
head
loss,
and
total
number
of
gallons
filtered
should
be
monitored
daily
to
evaluate
the
need
to
replace
the
bag
or
cartridge
(
AWWA
1999).
For
example,
cartridges
are
generally
replaced
when
the
pressure
differential
reaches
35
psi,
after
one
to
six
months
of
operation
(
Malcolm
Pirnie
1993).
The
maximum
allowable
pressure
differential
is
typically
recommended
by
the
manufacturer.
Technologies
and
Costs
for
Control
of
Microbial
Contaminants
and
Disinfection
Byproducts
June
2003
2­
32
Cartridges
and
bags
are
easily
damaged
at
the
time
of
installation.
Bags
should
be
replaced
with
caution
to
prevent
tearing
of
the
material.
Likewise,
the
operator
should
carefully
install
new
cartridges,
as
the
filter
seal
can
be
damaged
and
induce
leakage.

Because
of
their
rigid
structure
and
multi­
layer
design,
cartridge
filters
are
generally
more
sturdy
and
offer
more
operational
flexibility
than
bag
filters.
However,
this
higher
performance
is
typically
associated
with
higher
cost.
As
mentioned
previously,
cysts
and
oocysts
can
adhere
to
and
accumulate
on
the
surface
walls
of
the
system.
As
a
consequence,
the
inward
flow
of
water
in
the
cartridge
filter
requires
that
the
housing
be
cleaned
entirely
when
replacing
the
cartridge,
which
is
not
the
case
with
bag
filters.

2.2.7
Bank
Filtration
Bank
filtration
is
a
water
treatment
process
that
uses
a
river
bed
or
the
bank
of
a
river
or
lake
as
a
natural
filter.
Water
from
a
river
or
stream
flows
through
the
bank
and
draws
from
one
or
more
wells.
Microorganisms
and
other
particles
are
removed
by
contact
with
the
aquifer
materials
as
the
water
travels
through
the
subsurface,
either
horizontally
or
vertically.
High
removal
occurs
when
ground
water
velocity
is
slow
and
the
aquifer
consists
of
granular
materials
with
open
pore
space,
allowing
water
flow
around
the
grains.
In
these
granular
porous
aquifers,
the
flow
path
is
very
tortuous,
thereby
providing
ample
opportunity
for
the
microorganism
to
contact
and
attach
to
a
grain
surface.
Although
detachment
from
the
grains
can
occur,
it
typically
occurs
at
a
very
slow
rate.
When
ground
water
velocity
is
exceptionally
slow,
or
when
little
or
no
detachment
occurs,
most
microorganisms
become
inactivated
before
they
can
enter
a
well.
Thus,
bank
filtration
provides
physical
removal
and,
in
some
cases,
inactivation
to
protect
wells
from
pathogen
contamination.

2.2.7.1
Efficacy
Against
Pathogens
Due
to
the
low
recovery
rate
of
Cryptosporidium
oocysts
in
influent
and
effluent
samples,
full
scale
treatment
data
are
of
limited
utility
for
assessing
removal
of
Cryptosporidium
via
bank
filtration.
However,
measurement
of
other
parameters
indicate
the
potential
for
pathogen
removal.
Exhibit
2.16
summarizes
bank
filtration
studies
that
measured
coliform
and
spore
removal.
Cryptosporidium
removal
is
site­
specific
and
highly
dependent
on
the
aquifer
characteristics;
therefore,
these
data
are
only
an
indication
of
contaminant
removal
that
can
be
achieved
by
bank
filtration.
Technologies
and
Costs
for
Control
of
Microbial
Contaminants
and
Disinfection
Byproducts
June
2003
2­
33
Exhibit
2.16:
Bank
Filtration
Studies
Measuring
Coliform
and
Spore
Removal
Log
Removal
Reference
Travel
Distance
(
m)
Travel
Time
(
days)
Total
Coliform
Thermotoleran
t
Coliform
Spores1
Havelaar
et
al.
(
1995)
30
15
>
5.0
>
4.1
>
3.1
Havelaar
et
al.
(
1995)
25
63
>
5.0
>
4.1
>
3.6
Medema
et
al.
(
2000)
13
7
N/
A
4.1
3.3
25
18
N/
A
4.5
3.9
150
43
N/
A
6.2
5.0
Wang
et
al.
(
2000)
0.6
N/
A
N/
A
N/
A
2.0
1.6
2.0
3.0
2.0
16
3.0
1
Spore
data
are
sulphite­
reducing
clostridium
for
all
references
except
Wang
et
al.
(
2000),
where
spore
data
are
aerobic
endospores.

2.2.7.2
Factors
Affecting
Performance
The
main
factor
affecting
the
performance
of
bank
filtration
is
the
type
of
aquifer
material
through
which
the
water
is
filtered.
Granular
media
is
the
most
effective,
while
fractured
rock
or
gravel
with
large
pore
sizes
may
be
the
least
effective
and
allow
Cryptosporidium
to
pass
through
without
contacting
a
grain
surface.
The
flow
rate
is
also
an
important
factor
in
determining
performance.
Too
high
a
flow
rate
can
cause
oocysts
to
detach
from
the
aquifer
material.
Low
flow
rates,
however,
may
make
it
difficult
to
meet
volume
demands.

2.2.8
Second
Stage
Filtration
Second
stage,
or
secondary,
filtration
requires
the
use
of
rapid
sand,
dual
media,
GAC,
or
other
fine
grain
media
in
a
separate
stage
following
rapid
sand
or
dual
media
filtration.
A
cap,
such
as
GAC,
on
a
single
stage
of
filtration
is
not
considered
second
stage
filtration.

Filtration
processes
are
standard
in
the
water
treatment
process,
and
much
design
and
operational
information
is
available.
However,
the
use
of
a
second
filtration
stage
is
not
as
common,
and
little
information
is
available.
Technologies
and
Costs
for
Control
of
Microbial
Contaminants
and
Disinfection
Byproducts
June
2003
2­
34
2.2.8.1
Efficacy
Against
Pathogens
There
is
relatively
little
published
data
on
the
removal
of
Cryptosporidium
by
second
stage
filtration.
Results
based
on
a
number
of
single
stage
filtration
studies
demonstrate
that
rapid
sand
filtration,
when
preceded
by
coagulation,
can
achieve
significant
removal
of
Cryptosporidium.
While
these
studies
evaluated
only
a
single
stage
of
filtration,
the
same
mechanisms
of
removal
would
occur
with
a
second
filtration
stage.
Studies
have
also
shown
that
Cryptosporidium
breakthrough
occurs
after
the
first
stage
of
filtration;
therefore,
a
second
stage
of
filtration
is
likely
to
provide
a
barrier
to
these
oocysts.

Many
studies
(
Dugan
et
al.
2001
and
Emelko
et
al.
1999)
have
demonstrated
that
aerobic
spores
are
a
conservative
indicator
of
Cryptosporidium
removal
by
granular
media
filtration
when
preceded
by
coagulation.
Consequently,
EPA
believes
that
data
on
spore
removal
by
a
second
stage
filtration
process
are
indicative
of
the
capacity
of
this
process
to
remove
Cryptosporidium.

Between
1999
and
2000,
the
Cincinnati
Water
Works
collected
spore
and
turbidity
removal
data
from
their
GAC
system.
The
specifics
of
their
system
are
provided
below.

°
11­
foot
deep
GAC
filter
following
dual
media
filter
°
Loading
Rate
=
3.4
­
3.9
gpm/
ft2
(
average);
7.1
gpm/
ft2
(
design)

°
12*
40
mesh
°
d10
=
0.5
­
0.75
millimeters
(
mm);
d10
is
the
diameter
through
which
10
percent
of
the
media
will
pass
°
Uniformity
Coefficient
(
UC)
<
2;
UC
is
the
uniformity
coefficient
of
the
media
°
Media
age
­­
new
to
7
years
old;
carbon
reactivation
two
times
per
year
°
Empty
Bed
Contact
Time
(
EBCT)
=
22
minutes
at
120
million
gallons
per
day
(
mgd)
(
average
flow);
12
minutes
at
220
mgd
(
design
flow)

A
median
incremental
spore
removal
of
0.92
log
was
observed
in
their
GAC
filter.
Additionally,
the
secondary
GAC
filters
were
observed
to
dampen
or
eliminate
turbidity
spikes
from
preceding
dual
media
filters
that
occurred
during
ripening,
breakthrough,
etc.
These
data
indicate
that
0.5
log
or
greater
removal
of
Cryptosporidium
can
be
achieved
by
a
secondary
filtration
process
like
GAC
contractors.
Technologies
and
Costs
for
Control
of
Microbial
Contaminants
and
Disinfection
Byproducts
June
2003
2­
35
Based
on
information
presented
by
Hall
et
al.
(
1994),
up
to
a
50
percent
improvement
in
turbidity
removal
was
observed
when
using
a
second
stage
filter.
However,
no
improvement
in
Cryptosporidium
removal
was
observed
due
to
the
second
stage
filter.
This
information
was
collected
after
spiking
500
oocysts/
L
into
the
raw
water
of
a
conventional
filter
followed
by
a
secondary
filter
consisting
of
GAC.

2.2.8.2
Factors
Affecting
Performance
Filter
Type
There
are
several
types
of
filters.
Fine
sand
filters,
dual
media
filters,
and
multimedia
filters
are
the
main
types
of
filters
used
in
conventional
filtration
plants.
In
order
to
encourage
penetration
of
solids
into
the
depth
of
the
bed,
the
dual
media
filter,
consisting
of
a
layer
of
coarser
anthracite
coal
on
top
of
a
layer
of
finer
silica
sand,
was
developed.
Studies
conducted
by
many
researchers
(
Conley
and
Pitman
1960a,
Conley
1961,
Tuepker
and
Buescher
1968)
showed
the
benefits
of
dual
media
filters
in
reducing
the
rate
of
head
loss
development,
which
lengthened
the
filter
run.
Although
dual
media
is
presumed
to
improve
the
quality
of
the
filtrate,
this
benefit
has
not
been
well
demonstrated
(
Water
Quality
and
Treatment
1999).
Research
conducted
by
Robeck,
Dostal,
and
Woodward
(
1964)
demonstrated
that
the
head
losses
in
dual
media
filters
were
lower
than
the
head
losses
in
traditional
fine
sand
filters.
When
a
typical
dual
media
filter
and
a
fine
sand
filter
are
operated
at
the
same
filtration
rate
on
the
same
influent
water,
the
head
loss
development
rate
for
the
typical
dual
media
filter
should
be
about
half
the
rate
of
the
fine
sand
filter
(
Water
Quality
and
Treatment
1999).
Multimedia
filters
add
a
layer
of
garnet
to
the
media
which
allows
for
a
finer
layer
of
media
at
the
bottom
of
the
filter.

Filter
Media
As
with
all
filters
(
first
or
second
stage),
various
properties
of
a
filter
medium,
such
as
size,
shape,
density,
and
hardness,
affect
filtration
performance.
Filter
media
are
defined
by
their
uniformity
coefficient
(
UC)
and
effective
size
(
ES).
The
porosity
of
the
filter
bed
formed
by
the
grains
is
also
important
(
Water
Quality
and
Treatment
1999).
Filter
media
should
be
coarse
enough
to
retain
large
quantities
of
floc,
yet
fine
enough
to
prevent
passage
of
suspended
solids.
The
filter
bed
should
also
be
deep
enough
to
allow
long
filter
runs
and
graded
to
permit
backwash
cleaning.
In
order
to
obtain
high
rates
of
filtration,
coarse
sands
and
dual
media
beds
of
anthracite
overlying
sand
have
been
used
in
the
recent
past
(
Water
Supply
and
Pollution
Control
1993).

The
bed
porosity
and
the
ratio
of
the
bed
depth
to
media
grain
diameter
affect
the
filter
efficiency.
The
larger
the
depth
of
the
filter
bed
(
L),
the
more
opportunities
exist
for
particle
capture;
the
larger
the
average
diameter
of
the
media
(
d),
the
more
of
the
media
is
available
to
capture
particles
over
the
depth
of
the
filter
bed.
The
two
most
commonly
used
methods
in
selecting
the
optimal
filter
bed
depth
and
media
size
are
pilot
plant
studies
and
existing
data
from
filtration
facilities
treating
similar
waters.
Technologies
and
Costs
for
Control
of
Microbial
Contaminants
and
Disinfection
Byproducts
June
2003
2­
36
Filter
Hydraulics
Hydraulic
surges
occur
when
the
flow
through
a
filter
changes
rapidly
(
e.
g.,
during
either
filter
backwashing
or
servicing).
Hydraulic
shifts
can
lead
to
significant
particle
detachment.
To
ensure
that
the
second
stage
filtration
unit
is
unaffected
by
any
hydraulic
surges
caused
by
the
backwashing
of
the
first
stage
filtration
unit,
the
first
stage
filters
should
be
hydraulically
isolated
during
backwashing
and
servicing.

2.2.9
Pre­
Sedimentation
Pre­
sedimentation
is
a
preliminary
treatment
process
used
to
remove
particulate
material
from
the
source
water
before
the
water
enters
the
main
treatment
plant.
Because
pre­
sedimentation
reduces
particle
concentrations,
it
is
also
expected
to
reduce
Cryptosporidium
levels.
In
addition,
by
reducing
variability
in
water
quality
of
the
source
water,
pre­
sedimentation
may
improve
the
performance
of
subsequent
processes
in
the
treatment
plant.
To
remove
pathogens
through
floculation
and
sedimentation,
it
is
necessary
to
add
coagulant.

Sedimentation
processes
are
standard
in
the
water
treatment
process,
and
much
design
and
operational
information
is
available.
However,
the
use
of
a
pre­
sedimentation
basin
is
not
as
common,
and
little
information
is
available.

2.2.9.1
Efficacy
Against
Pathogens
There
is
relatively
little
published
data
on
the
removal
of
Cryptosporidium
by
presedimentation
Consequently,
EPA
analyzed
studies
that
investigated
Cryptosporidium
removal
by
conventional
sedimentation
basins.
The
removal
efficiency
in
conventional
sedimentation
basins
may
be
greater
than
in
pre­
sedimentation
due
to
differences
in
surface
loading
rates,
coagulant
doses,
and
other
factors.
To
supplement
these
studies,
EPA
reviewed
data
provided
by
utilities
on
removal
of
other
types
of
particles,
primarily
aerobic
spores,
in
the
pre­
sedimentation
processes
of
full­
scale
plants.
Studies
have
shown
that,
in
the
presence
of
a
coagulant,
spore
removal
is
a
conservative
indicator
of
Cryptosporidium
removal
(
Dugan
et
al.
2001).

The
literature
studies
reviewed
by
EPA
show
Cryptosporidium
log
removals
of
0.6
to
3.8
(
Dugan
et
al.
2001,
Payment
and
Franco
1993)
and
mean
Bacillus
subtilis
and
total
aerobic
spores
log
removals
of
0.6
to
1.1
(
data
collected
independently
by
the
Cincinnati,
OH,
and
St.
Louis,
MO,
water
utilities)
by
sedimentation
processes.
The
removal
of
aerobic
spores
through
sedimentation
basins
in
full­
scale
plants
demonstrate
that
pre­
sedimentation
is
likely
to
achieve
mean
reductions
of
greater
than
0.5
log
Cryptosporidium
removal
under
routine
operating
conditions
and
over
an
extended
time
period.
Technologies
and
Costs
for
Control
of
Microbial
Contaminants
and
Disinfection
Byproducts
June
2003
2­
37
2.2.9.2
Factors
Affecting
Performance
Short
Circuiting
Short
circuiting
in
the
sedimentation
tank
occurs
when
a
portion
of
the
influent
flow
reaches
the
outlet
of
the
sedimentation
basin
much
faster
than
the
designed
detention
time
of
the
basin.
Short
circuiting
increases
the
operational
surface
loading
rate
since
the
true
settling
area
available
for
a
portion
of
the
flow
is
reduced.
If
short
circuiting
causes
the
basin
to
operate
at
an
effective
loading
rate
greater
than
1.6
gpm/
ft2,
the
basin
would
not
receive
Cryptosporidium
removal
credit.
High
wind
velocities
and
density
and
temperature
differentials
between
the
influent
water
and
the
water
in
the
sedimentation
basin
cause
short
circuiting.
Additionally,
the
design
or
configuration
of
both
the
inlet
and
outlet
are
important
factors
that
can
affect
short­
circuiting
and
turbulence.
Systems
can
minimize
short
circuiting
by
adding
baffles
or
making
other
modifications
to
the
flow
pattern.

Coagulant
Dose
The
principle
goal
of
coagulation
is
to
destabilize
the
particles
so
that
they
can
be
more
easily
aggregated
into
flocs.
The
commonly
used
coagulants
are
alum,
ferric
chloride,
polyaluminum
chloride
(
PACl),
activated
charcoal,
and
activated
silica.
The
coagulant
dose
required
to
treat
an
influent
stream
depends
on
the
chemical
composition
of
the
influent,
the
characteristics
of
the
colloids
and
suspended
matter
in
the
influent,
the
water
temperature,
and
mixing
conditions.
The
use
of
a
coagulant
improves
the
pathogen
removal
capabilities
of
the
pre­
sedimentation
process,
although
some
pathogen
removal
is
expected
without
coagulant
addition.
Optimizing
a
coagulation
scheme
for
a
two­
stage
sedimentation
process
is
site­
specific
and
not
simple.
It
is
therefore
not
possible
to
prescribe
the
type
of
coagulant
and
appropriate
dose
for
an
aggregate
of
source
waters.
To
account
for
an
additional
sedimentation
process,
the
standard
jar
test
can
be
modified
to
a
two­
stage
process
reflecting
the
two
stages
of
sedimentation.

2.2.10
Watershed
Control
A
well­
designed
watershed
control
program
can
reduce
overall
microbial
risk.
The
risk
reduction
would
be
associated
with
the
implementation
of
practices
that
reduce
Cryptosporidium,
as
well
as
other
pathogens.
Knowledge
of
the
watershed
and
factors
affecting
microbial
risk,
including
sources
of
pathogens,
fate
and
transport
of
pathogens,
and
hydrology
can
also
help
a
system
reduce
microbial
risk.

2.2.10.1
Efficacy
Against
Pathogens
No
data
are
available
on
the
ability
of
watershed
control
programs
to
reduce
Cryptosporidium
loading
to
surface
water.
This
is
partly
because,
until
recently,
most
watershed
programs
have
focused
Technologies
and
Costs
for
Control
of
Microbial
Contaminants
and
Disinfection
Byproducts
June
2003
2­
38
on
improving
water
quality
for
recreational
and
ecological
uses
rather
than
for
drinking
water
protection.
Thus,
studies
of
the
success
of
such
programs
frequently
monitor
parameters
such
as
phosphorus
and
sediment
levels.
Watershed
programs
that
do
have
drinking
water
protection
as
a
goal
frequently
track
fecal
coliform
bacteria
levels
but
do
not
regularly
monitor
Cryptosporidium.
Fecal
coliform
concentrations
do
not
always
correlate
with
Cryptosporidium,
but
better
indicator
data
are
not
usually
available.
E.
coli
may
be
a
better
indicator
of
fecal
contamination
than
fecal
coliform
bacteria,
but
monitoring
for
E.
coli
is
not
common
practice.

Most
water
systems
that
do
monitor
Cryptosporidium
have
been
doing
so
for
only
a
few
years
and
would
not
have
enough
data
to
show
a
change
in
water
quality
resulting
from
watershed
management.
In
addition,
because
Cryptosporidium
occurs
in
such
low
concentrations
and
is
often
undetected,
reductions
in
microbiological
contamination
are
difficult
to
demonstrate.

Regardless
of
the
constituents
monitored,
it
is
difficult
to
show
that
a
watershed
control
program
in
its
entirety
has
improved
water
quality.
Often,
reductions
in
contamination
from
one
source
can
be
overshadowed
by
increases
from
other
sources,
especially
in
urban
areas.
However,
various
components
of
a
watershed
control
program
have
been
shown
to
have
a
positive
effect
on
microbiological
water
quality
at
a
local
level,
at
least
for
fecal
coliform.
Combined,
these
components
should
theoretically
contribute
to
an
overall
decrease
in
microbiological
contamination.

For
instance,
Thurston
et
al.
(
2001)
showed
that
a
constructed
wetland
could
reduce
fecal
coliform
levels
in
wastewater
treatment
plant
effluent
by
98
percent
(
where
effluent
had
previously
received
secondary
treatment).
Cryptosporidium
reductions
of
64
percent
were
also
achieved
through
this
study.
A
similar
pilot­
scale
study
with
untreated
wastewater
indicated
an
overall
removal
of
microorganisms
of
90
percent
by
constructed
wetlands
(
Quinonez­
Diaz
et
al.
2001).
Preliminary
results
of
a
watershed
restoration
program
in
Vermont
showed
that
streambank
stabilization,
fencing
of
riparian
zones
to
prevent
grazing,
and
protected
stream
crossings
reduced
bacterial
levels
(
Meals
2001).
A
fencing
program
in
Virginia
suggested
some
reduction
in
fecal
coliform
levels,
and
the
proportion
of
fecal
streptococci
strains
traced
to
livestock
was
reduced
(
Hagedorn
et
al.
1999).

Another
way
to
reduce
microbiological
contamination
of
an
urban
watershed
is
to
upgrade
wastewater
collection
systems.
The
Fairfax
County,
Virginia,
Wastewater
Collection
Division
decreased
inflow
and
infiltration
into
its
sewers
and
increased
the
sewers'
capacity
through
a
rehabilitation
and
maintenance
program.
Between
1995
and
2001,
the
utility
reduced
the
number
of
sanitary
sewer
overflows
throughout
the
county
by
67
percent
and
reduced
the
peak
flow
to
one
of
its
wastewater
treatment
plants
by
35
mgd
(
USEPA
2001).
Similar
programs
throughout
the
United
States
are
contributing
to
reduced
effluent
volumes
from
sanitary
sewer
overflows
and
combined
sewer
overflows.
Technologies
and
Costs
for
Control
of
Microbial
Contaminants
and
Disinfection
Byproducts
June
2003
2­
39
2.2.10.2
Factors
Affecting
Performance
A
combination
of
interventions
such
as
those
described
above
is
expected
to
result
in
an
overall
decrease
of
Cryptosporidium
in
source
water.
However,
many
factors
can
negatively
affect
the
success
of
a
watershed
control
program.
The
interventions
a
system
implements
depend
on
the
types
of
contamination
sources
in
the
watershed.
Control
of
point
source
discharge
(
e.
g.,
waste
water
treatment
plants
and
industrial
discharges)
can
be
straightforward.
Agricultural
and
urban
nonpoint
sources
are
the
most
difficult
to
control.
Reduction
of
Cryptosporidium
from
these
sources
generally
depends
on
the
voluntary
cooperation
of
urban
residents
and
farmers.

Urban
watersheds
are
subject
to
increasing
development,
which
increases
surface
imperviousness
and
the
amount
of
runoff
entering
surface
waters,
along
with
the
pollutant
load.
Acquisition
of
undeveloped
land,
particularly
that
closest
to
the
source
water
and
its
tributaries,
is
one
of
the
best
ways
to
prevent
degradation
of
the
water
quality,
but
it
may
not
be
feasible
in
some
watersheds.
Other
restrictions
on
development,
such
as
zoning
requirements,
can
also
control
urban
runoff
to
some
extent,
but,
again,
these
may
not
be
feasible
or
may
not
have
the
support
of
the
public
or
other
government
agencies.

Another
problem
facing
PWSs
is
that
the
watershed
may
extend
beyond
the
municipal
boundaries
into
other
jurisdictions.
A
higher
authority
(
e.
g.,
State
or
county
government)
may
be
needed
to
regulate
activities
outside
a
PWS's
jurisdiction
that
could
affect
water
quality.

2.2.11
Combined
Filter
Performance
Combined
filter
performance
reduces
Cryptosporidium
levels
by
enhancing
filter
performance
to
produce
very
low
turbidity
water.
It
is
defined
specifically
as
producing
0.15
NTU
turbidity
water
in
the
combined
filter
effluent
(
CFE)
95
percent
of
the
time.
Methods
that
systems
may
use
to
improve
filter
performance
and
lower
turbidity
include
adding
polymer,
optimizing
the
filtration
process
by
adding
media
or
installing
filter­
to­
waste
capabilities,
and
improving
staff
ability
to
optimize
the
process
by
additional
training,
hiring
new
operators,
and
buying
new
laboratory
equipment.

Systems
likely
to
use
this
technology
are
those
which
operate
conventional
filtration
or
softening
plants
and
which
are
already
operating
well
below
the
current
turbidity
limits
of
0.3
NTU.
These
systems
more
than
likely
target
their
effluent
under
0.15
NTU
already
but
are
not
currently
hitting
that
target
more
than
95
percent
of
the
time.
These
plants
are
assumed
to
be
able
to
reach
the
target
95
percent
of
the
time
with
relatively
minor
adjustments
to
their
process.
Technologies
and
Costs
for
Control
of
Microbial
Contaminants
and
Disinfection
Byproducts
June
2003
2­
40
2.2.11.1
Efficacy
Against
Pathogens
There
have
been
a
number
of
studies
examining
the
removal
of
pathogens
by
conventional
filtration.
Several
of
these
studies
have
examined
the
relationship
between
finished
water
turbidity
and
protozoa
removal.
Studies
by
Dugan
et
al.
(
2001)
and
Patania
et
al.
(
1995)
showed
that
turbidity
is
an
adequate
indicator
of
pathogen
removal.
Although
the
correlation
between
turbidity
removal
and
pathogen
removal
is
not
one
to
one,
removal
of
turbidity
is
a
conservative
indicator
of
pathogen
removal.

Under
the
IESWTR
and
LT1ESWTR,
conventional
and
direct
filtration
plants
may
claim
2.0
log
Cryptosporidium
removal
credit
if
their
CFE
turbidity
never
exceeds
1
NTU
and
is
less
than
or
equal
to
0.3
NTU
in
95
percent
of
samples
taken.
Under
the
LT2ESWTR,
systems
using
conventional
filtration
treatment
or
direct
filtration
treatment
may
claim
an
additional
0.5
log
Cryptosporidium
removal
credit
for
any
month
that
a
plant
demonstrates
CFE
turbidity
levels
less
than
or
equal
to
0.15
NTU
in
at
least
95
percent
of
the
measurements
taken
each
month,
based
on
sample
measurements
collected
under
§
§
141.73,141.173(
a)
and
141.551.

EPA
expects
plants
that
rely
on
complying
with
a
0.15
NTU
standard
to
consistently
operate
below
0.1
NTU.
Results
from
studies
conducted
by
Patania
et
al.
(
1995),
Emelko
et
al.
(
1999),
and
Dugan
et
al.
(
2001)
show
that
plants
consistently
operating
below
0.1
NTU
can
achieve
an
additional
0.5
log
or
greater
removal
of
Cryptosporidium
than
when
operating
between
0.1
and
0.2
NTU.

2.2.11.2
Factors
Affecting
Performance
Many
factors
can
affect
removal
of
pathogens
through
sedimentation
and
filtration
and
hinder
a
plant's
ability
to
achieve
0.15
NTU
in
its
CFE.
In
order
to
achieve
0.15
NTU
95
percent
of
the
time,
plants
will
need
to
have
tight
control
of
their
process.
The
areas
which
require
specific
attention
include:
control
of
coagulant
dosing
and
mixing,
control
of
dosing
of
other
chemical
additions,
filter
hydraulics
and
media,
and
backwashing
procedures.

Coagulant
Dose
Insufficient
coagulant
can
lead
to
colloidal
particles
remaining
in
suspension,
while
too
much
coagulant
can
lead
to
inefficient
settling.
Therefore,
coagulant
must
be
optimized
for
the
entire
plant.
It
must
also
be
adjusted
as
influent
water
quality
varies
or
if
there
are
other
major
changes
in
plant
operation.

Filter
Ripening
During
the
period
immediately
after
a
backwash,
the
lack
of
particles
on
the
filter
media
can
make
capture
of
the
particles
by
the
media
more
difficult
and
lead
to
breakthrough
of
particles
and
Technologies
and
Costs
for
Control
of
Microbial
Contaminants
and
Disinfection
Byproducts
June
2003
2­
41
turbidity.
Hall
and
Croll
(
1996)
studied
removal
in
a
pilot
plant
and
saw
peaks
in
both
turbidity
and
oocysts
in
the
filtered
water
for
an
hour
after
backwashing.
West
et
al.
(
1994)
found
that
removal
increased
from
2
log
to
of
3
log
once
the
filters
had
ripened,
and
the
turbidity
had
dropped
from
an
initial
value
of
0.2
NTU
to
a
value
less
than
0.1
NTU.

Filter
Breakthrough
During
filter
runs,
particles
can
collect
in
the
filter
and,
if
not
backwashed,
will
reach
the
point
where
an
increased
amount
of
particles
pass
through
(
referred
to
as
breakthrough).
Emelko
et
al.
(
2000)
studied
the
performance
of
filters
throughout
a
typical
run
cycle.
They
found
that
removal
was
5.5
log
when
the
filters
were
operating
at
0.04
NTU.
When
the
turbidity
began
to
climb,
removal
dropped
to
2.1
log
even
while
turbidities
were
still
less
than
0.1
NTU.
By
the
time
turbidity
had
reached
0.3
NTU,
the
removal
had
dropped
to
1.4
log.

Filtration
Rate
If
the
filtration
rate
is
too
high,
filtration
effluent
water
quality
can
suffer.
McTigue
et
al.
(
1998)
found
that
removal
dropped
by
2
log
when
the
filtration
rate
was
doubled.
West
et
al.
(
1994),
however
found
no
difference
in
removal
between
filtration
rates
of
6
and
14
gpm/
ft2.

Backwashing
The
flow
rate
used
for
backwashing
is
important
in
maintaining
effluent
quality.
Too
low
a
rate
can
leave
the
media
dirty
and
lead
to
mudballs
and
eventual
particle
breakthrough.
Too
high
a
rate
can
cause
loss
of
filter
media
and
also
lengthen
filter
ripening
times.
A
surface
wash
can
also
help
detach
particles
from
the
media
and
improve
backwash
performance.

2.3
DBP
Precursor
Removal
Strategies
2.3.1
Granular
Activated
Carbon
Adsorption
Removal
of
undesired
compounds
from
water
supplies
can
be
achieved
through
adsorption
onto
solids.
GAC
is
used
in
water
treatment
to
adsorb
a
variety
of
organic
and
inorganic
compounds.
Important
properties
of
GAC
that
determine
its
effectiveness
include
particle
size,
specific
surface
area,
pore
size
distribution,
and
chemical
nature
of
the
surface.
GAC
adsorption,
as
practiced
in
water
treatment,
is
an
non­
steady
state
process,
with
the
effluent
concentration
of
the
contaminant
increasing
with
time.
Once
the
effluent
concentration
meets
the
maximum
allowable
concentration
for
a
contaminant,
the
GAC
column
must
be
taken
off­
line,
and
the
GAC
must
be
replaced
with
reactivated
or
fresh
GAC.
The
operation
time
to
this
maximum
effluent
concentration
is
termed
the
reactivation
interval.
Technologies
and
Costs
for
Control
of
Microbial
Contaminants
and
Disinfection
Byproducts
June
2003
2­
42
The
EBCT
is
defined
as
the
volume
of
media
divided
by
the
flow
rate.
GAC
contactors
should
be
used
when
longer
EBCTs
are
required,
while
sand
filters
with
a
GAC
cap,
where
the
top
portion
of
the
sand
is
replaced
by
GAC,
can
be
used
when
shorter
EBCTs
are
feasible.
These
GAC­
capped
filters
are
often
called
filter­
adsorbers.
Filter­
adsorbers
can
also
be
filtration
units
which
contain
GAC
alone.
Because
of
their
shorter
EBCTs,
filter­
adsorbers
meet
desired
water
quality
goals
for
a
much
shorter
period
of
time
than
GAC
contactors.
For
the
purposed
of
treating
seasonal
changes
in
water
quality
or
contaminant
shock
loads,
filter­
adsorbers
may
have
an
economic
advantage
over
post­
filter
GAC
contactors.
One
disadvantage
of
filter­
adsorbers
is
that
GAC
losses
are
high
during
backwashing,
and
reactivation
and
equipment
separating
GAC
from
sand
may
be
required
before
reactivation.

GAC
contactors
operate
in
either
downflow
or
upflow
configurations.
Downflow
fixed­
bed
contactors
offer
the
simplest
and
most
common
contactor
configuration
for
drinking
water
treatment.
Upflow
beds
are
typically
used
in
situations
where
very
long
contact
times
(
greater
than
120
minutes)
are
required
and/
or
where
the
level
of
suspended
solids
is
high.
Flow
through
GAC
contactors
can
be
either
gravity
or
pressure
driven.

The
hydraulic
constraints
of
a
given
system
govern
the
selection
between
pressure
or
gravity
contactors.
Pressure
contactors
may
be
more
applicable
for
ground
water
systems,
since
these
systems
already
are
pumping
their
water.
Gravity
contactors
are
generally
found
in
surface
water
systems,
if
sufficient
head
is
available.
Downflow
contactors
are
typically
placed
downstream
of
the
plant
filters
to
minimize
the
solids
loading
to
the
contactor.

The
GAC
in
a
contactor
is
usually
replaced
when
the
effluent
concentrations
exceed
the
treatment
objective.
At
this
point,
however,
only
a
portion
of
the
GAC
is
fully
utilized,
and
replacement
of
the
media
will
result
in
unnecessarily
high
carbon
usage
rates.
Operating
multiple
GAC
contactors
in
either
series
or
parallel
configurations
are
the
two
common
methods
to
reduce
GAC
usage
rates.

For
contactors
configured
in
series,
the
GAC
in
the
first
contactor
is
reactivated
when
the
effluent
from
it
no
longer
meets
the
treatment
objective.
The
first
contactor
is
taken
offline
while
the
second
contactor
continues
operation.
After
the
GAC
in
the
first
contactor
is
replaced,
it
is
brought
back
online
downstream
from
the
operating
contactor.
That
is,
the
position
of
the
two
contactors
is
reversed,
with
what
was
originally
the
second
contactor
becoming
the
first
contactor
and
vice
versa.
For
efficient
operation,
the
mass
transfer
zone
should
be
contained
within
the
bed
length
of
one
contactor.
This
can
be
achieved
using
reasonable
bed
lengths
for
adsorption
of
micropollutants,
but
the
mass
transfer
zone
for
TOC
removal
and,
therefore,
DBP
precursor
removal
is
usually
too
long.
The
use
of
two
contactors
in
series
does
not
result
in
significantly
longer
run
times
over
single
contactor
operation
(
USEPA
1999a).

For
contactors
configured
in
parallel,
multiple
GAC
beds
are
operated
with
a
staggered
reactivation
pattern.
The
effluent
from
individual
contactors
may
contain
contaminants
at
concentrations
higher
than
the
treatment
objective,
since
they
may
be
blended
with
effluent
from
other
contactors
with
Technologies
and
Costs
for
Control
of
Microbial
Contaminants
and
Disinfection
Byproducts
June
2003
2­
43
little
or
no
breakthrough.
The
combined
effluent
concentration,
from
all
the
GAC
beds,
can
thus
be
maintained
below
the
specified
treatment
objective,
further
reducing
carbon
usage
rates.
For
DBP
precursor
removal,
contactor
effluents
should
be
blended
prior
to
disinfection.

The
choice
between
a
single
contactor
and
contactors
in
series
or
parallel
is
site
specific
and
depends
on
the
type
and
concentration
of
the
contaminant
to
be
removed
and
its
rate
of
adsorption.
This
choice
also
depends
on
the
type,
concentration,
and
adsorption
rate
of
competing
contaminants.

2.3.1.1
DBP
Precursor
Removal
In
many
circumstances,
GAC
is
an
effective
process
for
the
removal
of
NOM
from
drinking
water
sources.
With
an
EBCT
of
15
minutes
and
a
reactivation
interval
of
180
days,
GAC
can
remove
35
to
70
percent
of
the
influent
TOC
on
a
running
average
basis.
Running
average
TOC
removals
of
55
to
85
percent
can
be
achieved
with
an
EBCT
of
30
minutes
and
a
reactivation
interval
of
180
days.

2.3.1.2
Factors
Affecting
Performance
The
removal
of
NOM
by
GAC
adsorption
depends
on
a
large
number
of
factors
including
the
following:

°
Molecular
size,
polarity,
and
concentration
of
NOM
entering
the
GAC
process
°
Water
quality
characteristics
such
as
pH
and
ionic
strength
°
GAC
characteristics
such
as
pore
size
distribution
and
surface
chemistry
°
Operational
characteristics
such
as
EBCT
and
GAC
usage
rate
°
Treatment
processes
used
prior
to
the
GAC
process
°
Configuration
of
GAC
contactors
This
section
briefly
describes
the
impacts
of
these
factors
as
seen
in
several
GAC
studies.

Constituents
of
NOM
are
adsorbed
within
the
GAC
bed
in
a
manner
proportional
to
their
adsorption
potential.
Weakly
adsorbing
components
of
NOM
may
irreversibly
preload
the
GAC
at
the
downstream
end
of
the
bed
and
may,
therefore,
reduce
the
capacity
of
the
bed
for
stronger
adsorbing
components
at
the
end
of
the
bed.
Technologies
and
Costs
for
Control
of
Microbial
Contaminants
and
Disinfection
Byproducts
June
2003
2­
44
The
impacts
of
pH
on
adsorption
of
NOM
and
humic
extracts
have
been
well
documented
in
equilibrium
studies
using
powdered
activated
carbon
(
Weber
et
al.
1983,
Randtke
and
Jepsen
1982,
McCreary
and
Snoeyink
1980,
Summers
1986).
All
of
these
studies
showed
increased
removal
of
TOC
with
decreased
pH
levels.
Unfortunately,
some
of
the
work
has
been
done
with
different
initial
TOC
concentrations,
and
the
increased
performance
attributed
to
low
pH
may
be
because
of
the
lower
initial
TOC.
A
relationship
between
the
relative
adsorption
capacity
for
TOC
at
the
same
initial
TOC
and
pH
has
been
established
for
13
different
source
waters
and
a
bituminous
coal­
based
GAC
(
Hooper
et
al.
1996b).
Within
the
pH
range
of
5
to
10,
a
decrease
in
the
pH
of
one
unit
yielded
a
six
percent
increase
in
adsorption
capacity.
However,
the
number
of
continuous
flow
evaluations
of
pH
impacts
is
limited.

The
relationship
between
GAC
pore
size
distribution
and
NOM
molecular
size
distribution
has
been
shown
to
be
important
(
Summers
and
Roberts
1988,
Lee
et
al.
1983,
Semmens
and
Staples
1986,
El­
Rehaili
and
Weber
1987,
Chadik
and
Amy
1987).
In
general,
investigators
have
found
the
GAC
process
to
favor
removal
of
NOM
molecules
of
low
to
moderate
size
even
though
the
adsorption
process
was
expected
to
favor
removal
of
large
molecules.
This
phenomenon
occurs
because
small
GAC
pores
physically
exclude
large
NOM
molecules
from
adsorbing.
Thus,
GAC
with
a
greater
quantity
of
large
pores
can
be
expected
to
remove
more
NOM
than
GAC
with
a
smaller
quantity
of
large
pores.

The
impacts
of
EBCT
on
GAC
usage
rate
for
NOM
removal
have
been
studied
in
numerous
continuous
flow
evaluations.
The
trend
observed
in
all
studies
is
that
increasing
EBCT
can
reduce
the
carbon
usage
rate.
One
study
(
Miller
and
Hartman
1982)
saw
significant
reduction
in
usage
rates
as
the
EBCT
is
increased
from
2.8
to
15.2
minutes.
Summers
et
al.
(
1997)
evaluated
EBCTs
of
10
and
20
minutes
for
a
number
of
water
sources
and
concluded
that
EBCT
had
a
definite
effect
in
prolonging
the
bed
life
of
a
GAC
contactor.
However,
the
carbon
usage
rate
is
relatively
unaffected
by
EBCTs
at
the
ranges
evaluated.
They
also
noted
that
the
balance
between
EBCT
and
the
frequency
of
GAC
replacement
or
reactivation
is
primarily
a
choice
between
larger
capital
investment
(
i.
e.,
longer
EBCTs)
and
greater
operational
complexities
(
i.
e.,
more
frequent
reactivation).
Another
study
indicated
that
GAC
usage
rate
decreased
with
an
increase
in
EBCT
from
7.5
to
30
minutes.
However,
a
further
increase
in
EBCT
from
30
to
60
minutes
did
not
influence
the
GAC
usage
rate
(
McGuire
et
al.
1989).

GAC
systems
may
require
some
kind
of
pretreatment
to
prevent
clogging
of
the
GAC
bed,
to
minimize
the
organic
loading
on
the
GAC,
and
to
improve
cost
effectiveness.
Clogging
of
the
GAC
bed
could
be
caused
by
suspended
solids
in
the
raw
water
or
by
precipitation
of
calcium
carbonate,
iron,
and
manganese
on
the
GAC.
Suspended
solids
typically
cause
problems
in
surface
water
systems,
while
carbonate
scaling,
iron,
and
manganese
precipitation
may
occur
in
both
surface
and
ground
waters.
When
the
GAC
bed
life
is
long,
clogging
may
also
be
caused
by
biological
growths.
Pretreatment
methods
include
coagulation,
filtration,
or
softening
ahead
of
the
GAC
system.
Conventional
coagulation,
clarification,
and
filtration
processes
may
be
optimized
for
the
removal
of
organic
material
to
reduce
natural
organic
loading
to
the
GAC
bed.
Technologies
and
Costs
for
Control
of
Microbial
Contaminants
and
Disinfection
Byproducts
June
2003
2­
45
The
impacts
of
coagulation
on
NOM
adsorption
have
also
been
well
documented
in
batch
experiments
studying
adsorption
equilibria
(
Weber
et
al.
1983,
Randtke
and
Jepsen
1981,
Lee
et
al.
1981,
El­
Rehaili
and
Weber
1987,
Harrington
and
DiGiano
1989).
Coagulation
processes,
as
a
pretreatment
to
GAC,
can
both
reduce
influent
TOC
concentration
and
decrease
the
influent
pH
to
the
adsorber,
thus
leading
to
improved
GAC
performance.

Several
investigators
have
reported
better
GAC
performance
for
TOC
control
after
coagulation
or
after
increasing
the
coagulant
dose
(
i.
e.,
enhanced
coagulation).
Hooper
et
al.
(
1996a,
1996b,
1996c)
have
shown
that
the
increase
in
GAC
run
time
after
enhanced
coagulation
can
be
attributed
to
the
lower
pH
and
lower
initial
TOC
concentration
associated
with
the
enhanced
coagulated
water.
This
improvement
is
most
often
attributed
to
a
decrease
in
solubility
of
NOM
at
lower
pH
(
Symons
et
al.
1998).

In
most
GAC
applications
of
any
significant
size,
multiple
contactors
are
operated
in
a
parallel
configuration.
Parallel
GAC
contactors
are
operated
in
a
staggered
mode
wherein
each
contactor
has
been
in
operation
for
different
lengths
of
time.
In
this
mode
of
operation,
one
contactor
at
a
time
is
taken
off­
line
when
the
blended
effluent
exceeds
the
target
effluent
concentration,
and
a
column
with
fresh
or
reactivated
GAC
is
then
placed
on­
line.
The
effluent
from
the
contactor
in
operation
the
longest
can
be
higher
than
the
target
breakthrough
concentration,
as
it
is
blended
with
water
from
the
contactors
that
have
effluent
concentrations
much
lower
than
the
target
concentrations.
Consequently,
the
effluent
of
parallel
contactors
are
blended
prior
to
disinfection.
Thus,
parallel
operation
in
a
multiple
contactor
configuration
will
result
in
longer
GAC
bed­
life
and
the
time
between
reactivation
will
be
longer.
Under
ideal
conditions,
staged
blending
with
multiple
parallel
contactors
leads
to
near
steadystate
effluent
concentration
(
Roberts
and
Summers
1982).

Experimental
and
modeling
methods
for
predicting
the
blended
effluent
concentration
from
GAC
contactors
were
developed
by
Summers
et
al.
(
1997).
The
authors
observed
during
this
study
that
the
time
to
GAC
performance
goals
can
be
significantly
extended
by
blending
the
effluent
from
multiple
contactors.
For
the
three
waters
examined,
blending
increased
the
run
time
by
an
average
of
150
percent
for
both
TOC
and
TTHM.

The
research
described
above
demonstrates
how
the
performance
of
GAC
systems
can
be
influenced
by
many
process
variables.
In
general,
the
process
can
be
modified
to
provide
the
same
level
of
NOM
removal
at
lower
GAC
usage
rates
by
the
following:

°
Maintaining
low
pH
conditions
through
the
process
°
Increasing
NOM
removal
in
processes
that
precede
GAC
adsorption
°
Using
an
EBCT
greater
than
or
equal
to
10
minutes
Technologies
and
Costs
for
Control
of
Microbial
Contaminants
and
Disinfection
Byproducts
June
2003
2­
46
Ozonation
prior
to
GAC
does
not
guarantee
improved
NOM
removals
because
it
can
either
decrease
or
increase
the
ability
to
adsorb
and
increase
the
biodegradability
of
NOM.
The
overall
impact
of
preozonation
on
NOM
removal
in
GAC
contactors
depends
on
the
efficiency
of
biotreatment
to
remove
the
weakly
adsorbing
hydrophilic
fraction.

2.3.2
Nanofiltration
Nanofiltration
is
a
high­
pressure
membrane
process
that
has
been
traditionally
used
as
a
softening
process
to
remove
hardness
ions.
Generally,
NF
membranes
reject
divalent
ions
(
e.
g.,
Mg2+,
Ca2+),
but
pass
monovalent
ions
(
e.
g.,
Na+,
Cl­).
Recently,
NF
has
been
used
more
extensively
for
removal
of
DBP
precursors
and
color,
particularly
in
brackish
waters,
as
well
as
other
surface
waters.
Although
NF
processes
remove
nearly
all
turbidity
in
feed
water,
they
cannot
be
used
for
turbidity
removal
in
the
same
manner
as
MF
and
UF
due
to
smaller
pore
sizes
(
Mallevialle
et
al.
1996).
Smaller
pore
size
makes
NF
membranes
more
prone
to
fouling.
The
application
of
NF
for
surface
waters
is
generally
not
accomplished
without
extensive
pretreatment
for
particle
removal
and
possibly
pretreatment
for
dissolved
constituents.

The
percentage
of
treated
water
that
can
be
produced
from
the
feed
water
is
known
as
the
recovery.
Recovery
is
an
important
factor
for
cost
of
membrane
processes
and
is
one
measure
of
the
efficiency
of
a
system.
Recovery
for
NF
systems
is
typically
75
to
90
percent
and
is
impacted
by
feed
water
characteristics,
membrane
properties,
and
operating
conditions,
such
as
TMP.
Since
treatment
and
disposal
of
the
reject
stream
(
i.
e.,
waste
stream)
can
be
a
significant
portion
of
the
overall
cost
of
a
system,
recovery
can
greatly
affect
cost
efficiency.

2.3.2.1
Efficacy
Against
Pathogens
As
would
be
expected
based
on
MF
and
UF
microbial
removal
efficiencies,
NF
processes
are
capable
of
excellent
disinfection
by
removing
nearly
all
microbial
contaminants
in
feed
water,
including
Giardia,
Cryptosporidium,
and
viruses.
Historically,
NF
processes
have
not
been
used
as
a
primary
means
of
disinfection,
since,
in
large
part,
they
have
been
used
to
treat
ground
water
or
have
been
coupled
with
pretreatment
processes
such
as
MF
or
UF.
When
only
disinfection
is
required,
MF
and
UF
processes
are
typically
used
instead
of
NF,
since
they
are
less
costly
and
can
achieve
the
required
level
of
pathogenic
rejection
(
Mallevialle
et
al.
1996).
Because
of
this,
relatively
few
studies
documenting
microbial
removal
with
NF
membranes
have
been
conducted
in
comparison
to
MF
and
UF
processes.
Because
NF
and
RO
processes
represent
systems
that
are
very
similar
in
terms
of
disinfection
capabilities,
available
studies
documenting
microbial
removal
with
RO
as
well
as
NF
membranes
are
presented
in
Exhibit
2.17.
Technologies
and
Costs
for
Control
of
Microbial
Contaminants
and
Disinfection
Byproducts
June
2003
2­
47
Exhibit
2.17:
NF
Studies
Documenting
Microbial
Removal
Reference
Process
Membrane
Giardia
Log
Removal
Crypto
Log
Removal
MS2
Virus
Log
Removal
Gagliardo
et
al.
(
1999)
RO
HR
­­
­­
3.0
Gagliardo
et
al.
(
1999)
RO
DOW
­­
­­
5.4
Gagliardo
et
al.
(
1999)
RO
ESPA
­­
­­
4.7
Gagliardo
et
al.
(
1998)
RO
ULP
­­
­­
3.4
Seyde
et
al.
(
1999)
NF
(
Pilot)
Acumem­
4040
>
51
>
61
4.2
to
5.0
Colvin
et
al.
(
1999)
RO
(
bench)
FilmTec
BW30
­­
­­
>
42
Colvin
et
al.
(
1999)
RO
(
bench)
FilmTec
BW30
­­
­­
>
71
Trussel
et
al.
(
1998)
RO
(
MF
pretreat)
FilmTec
BW30
­­
­­
4.1
to
5.9
Trussel
et
al.
(
1998)
RO
(
MF
pretreat)
Hydranautics
4040
UHA
 
­
­
3.7
to
5.7
Trussel
et
al.
(
1998)
RO
(
MF
pretreat)
Fluid
Systems
TFLC/
M48
20HR
­­
­­
2.1
to
3.3
Trussel
et
al.
(
1998)
RO
(
MF
pretreat)
Fluid
Systems
TFCL/
ULP
 
­
­
2.9
to
4.3
Gagliardo
et
al.
(
1997)
RO
(
pilot)
TFC
>
5.7
>
5.7
3.0
to
4.0
Gagliardo
et
al.
(
1997)
RO
(
pilot)
CA
>
5.7
>
5.7
3.3
to
5.1
Note:
1
Indicates
removal
to
detection
limit.

2
0.02
:
m
Fluospheres
 
Data
not
available
As
shown
in
Exhibit
2.17,
NF
and
RO
processes
are
capable
of
significant
log
removals
of
cysts
and
viruses,
which
is
to
be
expected
since
these
microbes
are
much
larger
than
the
pore
size
of
the
membranes.
However,
the
data
in
Exhibit
2.17
show
that
NF
and
RO
systems
are
not
an
absolute
barrier;
they
can
allow
microorganisms
to
pass
through
the
membrane
into
the
treated
water.
For
this
reason,
it
is
important
to
consider
membrane
integrity
testing
when
assessing
the
ability
of
a
membrane
to
act
as
a
barrier
to
microorganisms.
Although
no
standard
NF
integrity
testing
method
exists,
some
tests
that
have
been
proposed
include
vacuum
testing
and
monitoring
effluent
water
quality
parameters
such
as
chloride,
UV­
254
absorbance,
microorganisms,
and
particle
counts
(
Spangenberg
et
al.
1999).
Vacuum
testing
entails
taking
the
membrane
off­
line.
This
has
the
disadvantage
of
being
unable
to
provide
on­
line
integrity
monitoring.
Should
a
system
become
compromised,
it
would
not
be
realized
until
the
module
is
taken
off­
line
and
tested.
Effluent
water
quality
monitoring
does
provide
real­
time
results.
However,
the
parameters
being
monitored
must
be
sensitive
enough
to
provide
an
alert
if
the
Technologies
and
Costs
for
Control
of
Microbial
Contaminants
and
Disinfection
Byproducts
June
2003
2­
48
system
is
compromised.
Sensitivity
of
various
parameters
will
depend
on
the
influent
level
of
that
particular
parameter
along
with
the
amount
of
removal
accomplished
by
the
membrane.
The
parameter
acting
as
a
surrogate
for
membrane
integrity
must
be
removed
to
a
significant
degree
such
that
a
noticeable
increase
in
effluent
concentration
would
be
seen
if
the
membrane
system
were
compromised.

NF
processes
are
also
capable
of
reducing
biodegradable
organic
carbon
(
BDOC)
(
Escobar
and
Randall
1999).
Since
BDOC
serves
as
substrate
for
microorganisms
in
the
distribution
system,
reducing
BDOC
can
reduce
the
potential
for
regrowth
in
a
distribution
system,
disinfectant
doses,
and
DBPs.
A
recent
full­
scale
study
was
performed
to
document
the
microbiological
and
disinfection
benefits
derived
from
implementing
NF
where
conventional
treatment
had
previously
been
practiced
(
Laurent
et
al.
1999).
The
results
of
this
study
showed
significant
decreases
in
chlorine
residual
fluctuations,
microbiological
counts,
DOC,
and
BDOC
in
treated
water
and
in
the
distribution
system.
In
effect,
this
created
greater
water
quality
stability
in
all
areas
of
the
distribution
system,
particularly
in
areas
with
high
residence
times.
In
addition,
the
finished
water
chlorine
dose
required
was
lowered
from
about
1
mg/
L
to
0.2
mg/
L
by
the
use
of
NF.

2.3.2.2
DBP
Precursor
Removal
Membrane
processes
can
remove
DBP
precursors
through
filtration
and
adsorption
of
organics.
Membranes
remove
NOM
through
filtration
(
i.
e.,
sieving)
when
NOM
molecules
are
larger
than
a
given
membrane
pore
size,
causing
them
to
be
rejected.
Size,
however,
is
only
one
factor
that
influences
NOM
rejection.
Shape
of
the
NOM
molecules
and
membrane
pores,
along
with
chemical
characteristics
of
the
NOM
molecules
and
membrane
also
play
important
roles
in
the
permeation
of
NOM
across
a
membrane
(
Mallevialle
et
al.
1996).
Membranes
may
also
remove
NOM
through
adsorption
of
organics
directly
on
the
membrane
surface.
The
level
of
adsorption
to
the
membrane
surface
depends
on
the
chemical
characteristics,
particularly
charge
and
hydrophobicity,
of
both
the
membrane
material
and
the
NOM.
Unfortunately,
organic
adsorption
is
generally
undesirable
since
it
has
proven
to
be
a
primary
cause
of
irreversible
fouling
of
membranes,
thereby
shortening
membrane
life.

Without
pretreatment,
NF
processes
remove
NOM
to
varying
degrees.
NOM
removals
for
NF
and
RO
processes
are
typically
on
the
order
of
50
to
99
percent.
NOM
removal
depends
on
many
factors,
including
membrane
MWCO
and
hydrophobicity,
characteristics
of
the
NOM,
and
membrane
system
operating
parameters
such
as
recovery
and
operating
pressure.
Results
from
several
studies
on
NOM
removal
by
NF
processes
are
provided
in
Exhibit
2.18.
Technologies
and
Costs
for
Control
of
Microbial
Contaminants
and
Disinfection
Byproducts
June
2003
2­
49
Exhibit
2.18:
NOM
Removal
Through
NF
Processes
Reference
Design
Criteria
Conclusions
of
Study
Taylor
et
al.
(
1987
and
1989)
Operating
pressure:
98­
141
psi
Flux:
8.9­
16.4
gpd/
sf
Recovery:
50­
79%
°
MWCO
of
100
to
500
are
needed
for
DOC
removal
up
to
90%.
°
MWCOs
of
1000
to
3000
may
achieve
50%
DOC
removal.
°
Trihalomethane
formation
potential
(
THMFP)
and
total
organic
halide
formation
potential
(
TOXFP)
reductions
up
to
95%
could
be
achieved
with
300
MWCO.
°
Operating
pressure
had
a
negligible
impact
on
NOM
removal.
°
TDS1
and
hardness
rejection
are
increased
by
increased
operating
pressure.

Conlon
and
McClellan
(
1989)
Operating
pressure:
90­
100
psi
Recovery:
75%
°
NOM
removal
greater
than
90%
for
200
MWCO.

Allgeier
and
Summers
(
1995)
Operating
pressure:
95
psi
Flux:
15­
24
gpd/
sf
Recovery:
30­
87%
°
66­
94%
TOC
removal
for
200
MWCO.
°
TOC
removal
decreased
by
up
to
15%
as
recovery
approached
90%.

Lozier
et
al.
(
1997)
Operating
pressure:
70
psi
Flux:
10
gpd/
sf
Recovery:
85%
°
69­
98%
TOC
removal
using
MF
pre­
treated
water.

Chellam
et
al.
(
1997)
Operating
pressure:
70
psi
Flux:
10
gpd/
sf
Recovery:
85%
°
90­
95%
TOC
removal
with
200
MWCO
on
MF
and
UF
pretreated
water.
°
95­
99%
SDS
THM
precursor
removal.
°
96­
99%
SDS
HAA6
precursor
removal.

Mulford
et
al.
(
1999)
Operating
pressure:
100
psi
Flux:
15
gpd/
sf
Recovery:
82%
°
96%
DOC
removal
with
200
MWCO.

Fu
et
al.
(
1995)
Operating
pressure:
80
psi
Flux:
15­
20
gpd/
sf
Recovery:
75­
90%
°
85­
97%
TOC
removal
with
100
to
500
MWCO.

Yoon
et
al.
(
1999)
Not
reported
°
60­
90%
TOC
removal
with
200
to
8,000
MWCO.
°
Slightly
higher
NOM
removal
is
achieved
at
pilotscale
than
at
bench­
scale.

Legube
et
al.
(
1995)
Not
reported
°
79­
91%
DOC
removal.
°
91­
95%
TOXFP
reduction.
°
93­
94%
THMFP
reduction.

1TDS
=
total
dissolved
solids
Technologies
and
Costs
for
Control
of
Microbial
Contaminants
and
Disinfection
Byproducts
June
2003
2­
50
In
addition
to
NOM
removal,
NF
processes
are
capable
of
some
DBP
removal,
although
little
work
has
been
performed
in
the
area.
Bromide
removal
is
also
important
for
the
reduction
of
brominated
DBPs.
NF
membranes
are
capable
of
significant
bromide
removal.
Several
studies
documenting
the
use
of
NF
processes
for
bromide
removal
are
summarized
in
Exhibit
2.19.

Exhibit
2.19:
Bromide
Removal
Through
NF
Processes
Reference
Conclusions
of
Study
Amy
and
Siddiqui
(
1999)
38­
41%
bromide
removal
with
150
to
300
MWCO.

Mulford
et
al.
(
1999)
50­
63%
bromide
removal
with
200
MWCO.

Allgeier
and
Summers
(
1995)
40­
61%
bromide
removal
with
200
MWCO.

Fu
et
al.
(
1995)
24­
38%
bromide
removal
with
100
to
500
MWCO.

Prados­
Ramirez
et
al.
(
1993)
63%
bromide
removal.

Conlon
and
McClellan;
Taylor
et
al.
(
1989)
60­
70%
chloride
removal,
with
bromide
removal
expected
to
be
nearly
identical.

As
shown
by
the
data
in
Exhibit
2.19,
NF
is
capable
of
high
percentage
bromide
removal.
Overall,
however,
bromide
removal
using
NF
would
probably
not
be
cost
effective
if
used
only
for
that
purpose.
If
the
process
were
incorporated
into
a
treatment
train
and
used
for
other
contaminant
removal,
membrane
removal
of
bromide
may
become
cost
effective
(
Amy
and
Siddiqui
1999).
It
is
important
to
note
that,
if
bromide
is
not
removed
sufficiently
but
TOC
levels
are
reduced,
the
bromideto
TOC
ratio
will
increase
considerably
and
will
cause
a
net
shift
in
speciation
of
DBPs
to
the
more
brominated
compounds.
In
the
worst
case,
such
a
scenario
could
cause
a
net
increase
in
the
absolute
level
of
brominated
DBPs
(
i.
e.,
bromoform)
after
chlorination
(
Amy
and
Siddiqui
1999).

2.3.2.3
Factors
Affecting
Performance
NF
is
gaining
popularity
as
a
DBP
precursor
removal
process,
since
production
costs
are
comparable
with
competing
processes
(
Mallevialle
et
al.
1996).
Due
to
the
small
pore
size
associated
with
NF,
other
feed
water
constituents
will
also
be
removed.
For
example,
divalent
salts,
some
metals,
and
some
soluble
organic
carbon
(
SOCs)
may
be
rejected
by
these
membranes
and,
therefore,
be
concentrated
in
the
waste
stream.
This
may
increase
the
cost
associated
with
disposing
of
the
waste
stream
compared
to
disposal
costs
associated
with
MF,
UF,
and
conventional
treatment
processes.
If
regulatory
limits
prohibit
sending
the
waste
stream
to
a
receiving
body,
costs
for
waste
handling
and
disposal
can
be
a
substantial
portion
of
the
overall
treatment
cost.
Technologies
and
Costs
for
Control
of
Microbial
Contaminants
and
Disinfection
Byproducts
June
2003
2­
51
MWCO
is
a
key
characteristic
affecting
membrane
performance.
Membranes
with
MWCOs
in
the
100
to
500
range
appear
to
be
very
effective
as
a
means
of
DBP
precursor
removal.
TOC,
THMFP,
and
TOXFP
reductions
of
70
to
95
percent
are
commonly
achieved
in
systems
using
such
membranes.
These
processes
can
effectively
remove
bromide
as
well,
with
reductions
up
to
95
percent.
Larger
MWCO
membranes
(
i.
e.,
MWCO
near
and
above
10,000),
however,
will
not
be
as
effective
for
NOM
reduction.

Commercial
NF
(
as
well
as
MF
and
UF)
membranes
are
available
in
many
types
of
material
(
e.
g.,
cellulose
acetate
and
polysulphone)
and
in
various
configurations
(
e.
g.,
spiral
wound
and
hollow
fiber).
The
chemistry
of
the
membrane
material,
particularly
surface
charge
and
hydrophobicity,
can
play
an
important
role
in
rejection
properties,
since
membranes
can
remove
contaminants
through
adsorption
on
the
membrane
surface
as
well
as
through
sieving
across
the
membrane
pores.
These
factors
must
be
taken
into
consideration
to
accommodate
source
water
characteristics
and
removal
requirements.

Source
water
quality
can
also
dictate
pretreatment
requirements.
The
small
pore
size
of
NF
and
RO
membranes
makes
them
more
prone
to
fouling
than
UF
or
MF
membranes,
necessitating
higher
quality
feed
water.
The
application
of
NF
and
RO
for
surface
water
treatment
is
generally
not
accomplished
without
extensive
pretreatment
for
particle
removal
and
possibly
pretreatment
for
dissolved
constituents.
For
example,
the
rejection
of
scale­
forming
ions,
such
as
calcium
and
silica,
can
lead
to
precipitation
on
the
membrane
surface
since
these
ions
are
concentrated
on
the
feed
side
of
NF
and
RO
membranes.
Organic
constituents
and
metal
compounds,
such
as
iron
and
manganese,
can
promote
fouling
through
precipitation
and
adsorption
as
well.
Precipitation
and
adsorption
can
result
in
irreversible
fouling
and
must
be
avoided
through
appropriate
pretreatment,
including
anti­
scaling
chemical
and/
or
acid
pretreatment
and
possibly
pretreatment
for
organics
removal.

In
terms
of
contaminant
removal,
membrane
performance
can
also
be
influenced
by
the
operating
pressure
and
percent
recovery,
depending
on
the
mechanism
of
rejection.
(
This
is
true
for
NF
and
RO
systems,
but
generally
not
true
for
MF
and
UF
systems.)
Contaminant
rejection
by
NF
and
RO
systems
generally
increases
with
decreasing
operating
pressure
and
with
decreasing
recovery.
Thus,
rejection
can
be
enhanced
by
changing
operating
parameters,
but
not
without
corresponding
increases
in
operating
costs.
To
increase
recovery,
membranes
are
often
staged
(
i.
e.,
the
concentrate
of
one
stage
of
membranes
is
treated
by
another
stage
of
membranes).
Two
to
three
stages
are
common
for
NF
and
RO
systems.
(
Staging,
however,
is
generally
not
used
for
MF
and
UF.)
Staging
is
also
used
to
keep
the
fluid
velocity
across
the
membranes
at
a
specified
rate.
The
maximum
attainable
percent
recovery
is
usually
governed
by
the
degree
to
which
the
water
can
be
concentrated
without
the
occurrence
of
precipitation
for
NF
and
RO.
Technologies
and
Costs
for
Control
of
Microbial
Contaminants
and
Disinfection
Byproducts
June
2003
3­
1
3.
Technology
Design
and
Criteria
3.1
Introduction
This
Chapter
provides
assumptions
related
to
the
overall
design
for
each
technology
addressed
in
this
document.
Types
of
information
provided
in
this
Chapter
include:

°
Assumed
water
quality
conditions
(
e.
g.,
median
filter
water
quality
assumptions
for
UV
design)

°
Chemical
doses
(
e.
g.,
ozone
dose
for
Cryptosporidium
inactivation)

°
Equipment
type
(
e.
g.,
types
of
UV
lamps
for
various
system
sizes)

Chapter
4
builds
on
this
Chapter
by
providing
more
detailed
design
assumptions
for
technology
components
and
presents
the
costs
for
each
technology.

Section
3.2
describes
the
assumed
based
treatment
plant
used
for
all
technology
modifications.
Sections
3.3
and
3.4
describe
the
design
approach
for
alternative
disinfectant
and
DBP
precursor
removal
technologies,
respectively.

3.2
Base
Treatment
Plant
The
base
treatment
plant
is
assumed
to
represent
the
existing
treatment
configuration.
All
modifications
with
alternative
disinfection
strategies
and
removal
of
DBP
precursors
are
assumed
to
be
retrofitted
from
this
base
treatment
plant.
The
base
plant
is
represented
by
a
conventional
treatment
plant,
employing
the
basic
processes
of
coagulant
addition
and
mixing,
flocculation,
clarification,
granular
media
filtration,
and
chlorination
for
both
primary
disinfection
and
maintenance
of
a
distribution
system
residual.
A
schematic
of
the
base
plant
is
shown
in
Exhibit
3.1.
Technologies
and
Costs
for
Control
of
Microbial
Contaminants
and
Disinfection
Byproducts
June
2003
3­
2
Cl
Cl
2
Coagulant
Coagulant
Rapid
Mix
Flocculation/
Sedimentation
Filtration
Storage
Caustic
Caustic
Cl
Cl
2
Exhibit
3.1:
Base
Plant
3.3
Alternative
Disinfection
Strategies
Pertinent
to
compliance
with
the
Stage
2
DBPR
and
the
LT2ESWTR,
alternative
disinfection
strategies
may
be
selected
to
provide
additional
treatment
for
Cryptosporidium
and/
or
to
limit
the
formation
of
DBPs.
This
section
describes
the
overall
design
approach
used
for
costing
a
number
of
alternative
disinfection
strategies
capable
of
achieving
these
goals.

3.3.1
Chloramination
Chloramines
can
be
used
for
secondary
disinfection
to
limit
DBP
formation
in
the
distribution
system.
Chloramines
are
less
effective
for
microbial
inactivation
than
chlorine
and
are
typically
ineffective
as
a
primary
disinfectant;
however,
they
may
be
used
in
combination
with
other
technologies
discussed
in
this
section
(
e.
g.,
ozone
for
primary
disinfection)
to
reduce
DBP
formation
in
the
distribution
system.
Typically,
ammonia
is
added
after
filtration
(
or
possibly
after
storage)
to
quench
the
chlorine
residual
and
form
chloramines.
A
schematic
of
a
chloramine
system
is
shown
in
Exhibit
3.2
Technologies
and
Costs
for
Control
of
Microbial
Contaminants
and
Disinfection
Byproducts
June
2003
3­
3
Cl
Cl
22
Rapid
Mix
Flocculation/
Sedimentation
Filtration
Storage
Caustic
Caustic
NH
NH
3
Description
of
Process:
Pre­
chlorination
for
primary
disinfection;
add
ammonia
after
filtration
at
a
residual
chlorine
to
ammonia
ratio
of
4:
1.
Coagulant
Coagulant
Exhibit
3.2:
Chloramines
for
Secondary
Disinfection
A
range
of
finished
water
chlorine
residuals
were
derived
using
the
ICR
database.
The
10th
and
90th
percentile
finished
water
chlorine
residuals
from
the
ICR
database
are
approximately
0.6
and
2.2
mg/
L,
respectively.
From
these
residuals,
the
ammonia
dosages
of
0.15
and
0.55
mg/
L
were
derived
assuming
a
4:
1
chlorine
to
ammonia
ratio
(
typical
chlorine
to
ammonia
ratios
are
between
3:
1
and
5:
1
to
ensure
monochloramine
formation).
Upgrade
costs
were
generated
only
for
ammonia
storage
and
feed
systems
(
the
base
plant
is
assumed
to
provide
the
necessary
chlorine).
It
is
assumed
that
all
chloramination
can
be
accomplished
at
the
plant
and
that
no
distribution
system
booster
stations
are
required.

Aqueous
ammonia
is
assumed
for
small
systems
(<
1
mgd),
and
anhydrous
ammonia
is
assumed
for
large
systems
(>
1
mgd).
Anhydrous
ammonia
is
generally
more
cost
effective
for
larger
utilities;
however,
safety
and
handling
issues
with
anhydrous
ammonia
also
need
to
be
considered.
The
aqueous
ammonia
system
consists
of
a
chemical
storage
container,
metering
pumps,
an
on­
line
process
analyzer,
piping,
and
valves.
The
anhydrous
ammonia
system
consists
of
bulk
storage
pressure
vessels,
a
vacuum
feed
system,
an
on­
line
process
analyzer,
piping,
and
valves:
The
larger
systems
may
also
include
a
vaporizer
and
an
emergency
scrubber
system.

3.3.2
Chlorine
Dioxide
Chlorine
dioxide
is
an
effective
oxidant/
disinfectant
that
is
frequently
used
to
control
THM
formation.
It
has
also
been
shown
to
inactivate
Cryptosporidium,
as
described
in
Chapter
2.
Thus,
chlorine
dioxide
can
replace
chlorine
(
or
other
oxidants)
as
the
primary
disinfectant
and
potentially
achieve
a
greater
level
of
pathogen
inactivation
while
decreasing
THM
and
HAA
formation.
However,
controlling
the
formation
of
chlorite
ions
can
be
a
considerable
challenge
in
chlorine
dioxide
treatment
implementation.
Technologies
and
Costs
for
Control
of
Microbial
Contaminants
and
Disinfection
Byproducts
June
2003
3­
4
ClO
ClO
2
Rapid
Mix
Flocculation/
Sedimentation
Filtration
Storage
Caustic
Caustic
Description
of
Process:
Replace
chlorination
with
chlorine
dioxide
addition.
Point
of
addition
may
be
1)
prior
to
rapid
mix,
or
2)
prior
to
flocculation,
or
3)
prior
to
filtration,
or
4)
post
filtration
Coagulant
Coagulant
Because
of
the
significant
operator
attention
required
to
monitor
and
control
chlorite
formation
as
well
as
to
address
safety
concerns,
it
is
assumed
that
systems
serving
fewer
than
500
people
will
not
have
the
expertise
necessary
to
use
this
technology.
Therefore,
costs
are
only
developed
for
systems
with
a
design
flow
of
0.091
mgd
or
greater.

Many
plants
add
chlorine
dioxide
as
a
pre­
oxidant,
but
it
can
also
be
added
after
filtration.
For
the
analysis
presented
here,
it
is
assumed
that
chlorine
dioxide
can
be
added
at
any
point
in
the
process
train.
(
A
schematic
of
the
chlorine
dioxide
system
is
shown
in
Exhibit
3.3.)
Chlorine
dioxide
costs
do
not
include
construction
of
a
basin
for
additional
chlorine
dioxide
contact
time.
It
is
assumed
that
plants
can
achieve
adequate
contact
time
with
their
existing
configuration.

Exhibit
3.3:
Disinfection
with
Chlorine
Dioxide
All
chlorine
dioxide
cost
analyses
presented
in
this
document
are
based
on
an
applied
dose
of
1.25
mg/
L.
This
is
close
to
the
maximum
dosage
of
chlorine
dioxide
that
can
be
added
while
remaining
in
compliance
with
a
1.0
mg/
L
MCL
for
chlorite,
conservatively
assuming
a
70
percent
conversion
of
chlorine
dioxide
to
chlorite
and
a
safety
factor
to
account
for
impurities,
such
as
unreacted
chlorine,
in
the
chlorine
dioxide
feed.
This
analysis
evaluated
chlorine
dioxide
costs
at
the
maximum
dosage
because
chlorine
dioxide
is
being
considered
here
for
inactivation
of
Giardia
and
Cryptosporidium.
Protozoa
inactivation
by
chlorine
dioxide
typically
requires
high
CT
values
as
described
in
Chapter
2.
Additionally,
evaluating
the
maximum
chlorine
dioxide
dose
provides
a
degree
of
conservatism
to
these
cost
estimates.
The
level
of
Cryptosporidium
inactivation
that
would
be
achieved
by
this
dose
depends
on
water
quality
and
contact
time
and
is
not
assessed
in
this
cost
analysis.
Higher
doses
that
would
necessitate
the
removal
of
chlorite
are
not
evaluated
at
this
time
due
to
uncertainty
about
the
applicability
and
efficacy
of
chlorite
removal
processes.
Technologies
and
Costs
for
Control
of
Microbial
Contaminants
and
Disinfection
Byproducts
June
2003
3­
5
For
all
systems,
the
use
of
an
automatic
generator
is
assumed.
Key
design
assumptions
for
large
systems
are
presented
below.

°
Chlorine
dioxide
generation
is
accomplished
through
addition
of
sodium
chlorite
to
a
chlorine
solution
created
by
dissolution
of
chlorine
gas
in
water.

°
A
sodium
chlorite
metering
and
mixing
system
is
provided.

°
A
chlorine
dioxide
generator
(
detention
time
=
0.2
minutes)
is
provided.

°
A
polyethylene
day
tank
and
mixer
are
provided
to
store
chlorine
dioxide
prior
to
its
addition
to
the
process.

°
A
dual
head
metering
pump
is
provided
to
add
chlorine
dioxide
to
the
process.

°
A
1:
1
mass
ratio
of
chlorine
gas
to
sodium
chlorite
is
assumed
to
ensure
that
the
sodium
chlorite
is
completely
utilized.
(
The
additional
chlorine
serves
to
lower
the
pH
for
reaction
through
creation
of
hypochlorous
acid.)

It
is
assumed
that
small
systems
(<
2
mgd)
will
rent
the
ClO2
generation
equipment
and
only
incur
capital
costs
for
instrumentation
and
piping
and
valves.

3.3.3
Ultraviolet
Light
UV
light
is
an
effective
disinfectant
for
bacteria,
viruses,
Giardia,
and
Cryptosporidium
and
does
not
form
THMs
or
HAAs
(
see
Chapter
2).
For
cost
estimates
in
this
document,
a
conceptual
design
for
retrofitting
the
base
plant
with
a
UV
disinfection
system
was
developed
based
on
plant
flow
(
i.
e.,
system
size
category)
and
water
quality.
Because
particulate
matter
may
affect
the
performance
of
UV
systems,
the
cost
estimates
assume
that
the
UV
system
is
installed
downstream
from
the
filter.
Exhibit
3.4
presents
a
schematic
of
a
conventional
water
treatment
plant
(
WTP)
with
UV
disinfection.
As
shown
in
the
schematic,
interstage
pumping
is
assumed
because
many
utilities
will
not
have
sufficient
hydraulic
head
to
support
the
addition
of
UV
disinfection
facilities
without
significantly
affecting
plant
operation.
Technologies
and
Costs
for
Control
of
Microbial
Contaminants
and
Disinfection
Byproducts
June
2003
3­
6
Rapid
Mix
Flocculation/
Sedimentation
Filtration
Caustic
Caustic
Storage
UV
Light
Description
of
Process:
Replace
chlorination
with
UV
light
for
disinfection.
Coagulant
Coagulant
Interstage
Pumping
Exhibit
3.4:
UV
Disinfection
The
filtered
water
quality
conditions
assumed
for
all
UV
costs
are
based
on
median
values
reported
in
the
ICR,
as
indicated
in
Exhibit
3.5.

Exhibit
3.5:
Water
Quality
Assumptions
for
UV
Disinfection
Parameter
Value
UV
254
absorbance1
(
cm­
1)
0.051
UVT
(%)
1
89
Turbidity
(
0.1
Alkalinity
(
mg/
L
as
CaCO3)
2
60
Hardness
(
mg/
L
as
CaCO3)
2
100
1
Median
of
maximum
filtered
water
UVT
(
minimum
UV
absorbance)
from
the
ICR
data
2
Median
of
all
ICR
filtered
water
data
Source:
ICR
Data
Cost
estimates
for
UV
are
provided
for
two
UV
doses:
40
and
200
mJ/
cm2.
As
discussed
in
Chapter
2,
a
UV
dose
of
40
mJ/
cm2
has
been
shown
to
be
sufficient
for
3
log
inactivation
of
Cryptosporidium
and
Giardia
and
1
to
2
log
inactivation
of
viruses.
Studies
have
shown
that
a
UV
dose
of
200
mJ/
cm2
is
adequate
for
4
log
inactivation
of
viruses.

Low
pressure
UV
lamp
based
systems
have
been
used
for
small
treatment
plants
but
are
not
typically
installed
at
larger
facilities
due
to
the
high
number
of
lamps
that
would
be
required.
Medium
pressure
lamp
systems
are
not
typically
used
for
smaller
utilities
due
to
higher
capital
costs
in
comparison
to
LP
systems
at
low
flow
rates.
Therefore,
UV
reactors
utilizing
LP
lamps
are
assumed
for
the
small
system
(<
1
mgd)
designs.
Depending
upon
the
manufacturer,
LPHO
and/
or
MP
reactors
are
provided
in
the
large
system
(>
1
mgd)
cost
estimates.
Technologies
and
Costs
for
Control
of
Microbial
Contaminants
and
Disinfection
Byproducts
June
2003
3­
7
All
UV
systems
are
designed
with
an
equipment
redundancy
of
one
extra
UV
reactor
(
n+
1)
or
15
percent
capacity
above
design
flow,
whichever
is
greater.
The
number
of
reactors
costed
for
each
system
size
is
shown
in
Exhibit
3.6
below.
The
number
of
reactors
for
each
design
flow
is
based
on
currently
available
UV
reactor
sizes
and
flows.

Exhibit
3.6
Number
of
Assumed
UV
Reactors
Design
Flow
(
mgd)
Duty
UV
Reactors
Standby
UV
Reactors
Total
Number
of
UV
Reactors
0.022
­
3.5
1
1
2
17
3
1
4
76
5
1
6
210
11
2
13
430
22
4
26
UV
disinfection
systems
are
sensitive
to
power
interruptions
and
fluctuations.
When
a
UV
reactor
goes
down,
it
can
take
from
four
to
ten
minutes
for
the
UV
lamps
to
regain
full
power.
A
utility
with
poor
power
quality
might
have
problems
with
their
UV
systems
going
down
too
frequently.
One
way
to
prevent
this
problem
is
to
install
a
uninterruptible
power
supply
(
UPS),
which
is
essentially
a
battery
that
smooths
out
the
power
interruptions
and
fluctuations.
Because
some
systems
may
need
UPS
systems,
cost
estimates
in
Chapter
4
are
completed
at
UV
doses
of
40
and
200
mJ/
cm2,
with
and
without
UPS
systems.

3.3.4
Ozone
Ozone
can
be
used
to
replace
chlorine
for
primary
disinfection
and
can
provide
a
higher
level
of
inactivation
of
certain
pathogens,
such
as
Cryptosporidium,
while
reducing
formation
of
THMs
and
HAAs.
Ozone
is
one
of
the
most
powerful
oxidants
available
for
water
treatment
(
second
only
to
the
hydroxyl
free
radical).
Disinfection
with
ozone
is
influenced
by
water
quality
characteristics
such
as
pH,
temperature,
alkalinity,
TOC,
and
certain
inorganic
compounds
like
iron
and
manganese.
The
use
of
ozone
can
be
limited
by
raw
water
bromide
levels
and
consequent
bromate
formation.
These
factors,
in
conjunction
with
the
CT
necessary
for
the
desired
level
of
pathogen
inactivation,
impact
the
design
and
operation
of
the
ozone
system.

A
schematic
of
the
ozone
configuration
is
shown
in
Exhibit
3.7.
The
costing
process
allows
for
ozone
application
to
either
raw
or
settled
water
(
settled
water
application
is
depicted
in
Exhibit
3.7).
To
control
bromate
formation
during
ozonation,
it
may
be
necessary
to
lower
the
pH
in
certain
waters.
Separate
costs
are
estimated
for
pH
adjustment
so
that
this
cost
may
be
added
to
the
costs
of
ozonation,
where
appropriate.
The
pH
adjustment
costs
include
addition
of
a
chemical
feed
system.
To
reduce
the
pH,
sulfuric
acid
is
used
and
caustic
(
after
ozonation)
is
used
to
raise
pH.
Technologies
and
Costs
for
Control
of
Microbial
Contaminants
and
Disinfection
Byproducts
June
2003
3­
8
Rapid
Mix
Flocculation/
Sedimentation
Filtration
Caustic
Caustic
Storage
Ozone
Generator
&
Contact
Basin
Description
of
Process:
Replace
chlorination
with
ozonation.
Coagulant
Coagulant
Exhibit
3.7:
Ozone
Disinfection
Costs
for
ozone
treatment
systems
are
directly
related
to
the
dose
applied.
For
the
purposes
of
the
LT2ESWTR
and
the
Stage
2
DBPR,
three
ozone
doses
are
costed
based
on
the
three
levels
of
Cryptosporidium
inactivation:
0.5,
1.0,
and
2.0
log.
The
Surface
Water
Analytical
Tool
(
SWAT)
model
is
used
to
calculate
the
ozone
dose
required
for
each
inactivation
level,
based
on
CT
tables
in
Chapter
2
(
Exhibit
2.13)
and
assuming
an
ozone
CT
of
12
minutes.
For
each
plant
in
the
ICR
survey,
and
for
each
month
with
data,
the
SWAT
model
was
used
for
raw
water
characteristics
and
existing
plant
configurations
to
determine
the
dose
required
to
achieve
the
desired
Cryptosporidium
inactivation.
Mean
and
maximum
doses
were
then
determined
for
each
ICR
plant.

For
costing
purposes,
two
doses
were
established
for
each
of
the
three
Cryptosporidium
inactivation
levels
(
0.5,
1.0,
and
2.0
logs).
The
median
of
all
plant­
mean
ozone
doses
(
1.78,
2.75,
and
3.91
mg/
L,
respectively)
were
used
to
calculate
operation
and
maintenance
costs.
This
is
the
dose
which
will
be
the
most
common
for
all
plants
achieving
the
given
inactivation
and
the
dose
most
representative
of
daily
plant
flows.
To
determine
capital
costs,
the
median
of
the
plant­
maximum
doses
(
3.19,
5.0,
and
7.0
mg/
L,
respectively)
are
used,
as
systems
will
be
designed
to
meet
the
maximum
dose
that
could
be
required
under
typical
conditions.

The
primary
components
of
the
ozone
process
include
in­
plant
pumping,
ozone
generation
system,
ozone
contactor,
off­
gas
destruction
facilities,
effluent
ozone
quench,
stainless
steel
piping
(
including
valves
and
ductwork),
electrical
and
instrumentation
(
E&
I),
and
chemical
storage
facilities.
Components
not
related
directly
to
the
process
(
e.
g.,
for
which
indirect
costs
are
calculated)
include
piloting,
permitting,
land,
operator
training,
and
housing.
Technologies
and
Costs
for
Control
of
Microbial
Contaminants
and
Disinfection
Byproducts
June
2003
3­
9
Rapid
Mix
Flocculation/
Sedimentation
Filtration
Caustic
Caustic
Storage
Micro/
Ultrafiltration
Description
of
Process:
Addition
of
microfiltration
or
ultrafiltration
following
granular
media
filtration.
It
may
be
necessary
to
move
the
point
of
chlorination
to
after
MF/
UF,
as
some
membranes
can
be
damaged
by
chlorine.
Coagulant
Coagulant
Interstage
Pumping
Cl
Cl
2
3.3.5
Microfiltration
and
Ultrafiltration
Microfiltration
or
ultrafiltration
can
be
added
to
the
base
plant
process
train
to
enhance
particle
and
microbial
removal,
including
removal
of
Cryptosporidium.
MF/
UF
may
also
allow
treatment
plants
to
reduce
DBP
formation
by
decreasing
the
disinfectant
dose
required
to
meet
plant
CT
requirements.
MF/
UF
can
be
added
to
the
treatment
process
following
conventional
media
filtration,
or,
in
some
cases,
may
be
added
as
a
replacement
for
media
filtration.
In
certain
applications
(
e.
g.,
low
total
suspended
solids
(
TSS)
surface
waters
or
groundwaters),
MF/
UF
can
replace
the
entire
conventional
treatment
process.
However,
the
design
assumptions
and
costs
presented
in
this
document
assume
addition
of
MF/
UF
to
an
existing
conventional
treatment
plant
for
enhanced
removal
of
Cryptosporidium
and/
or
DBP
control.
Consequently,
the
costs
presented
in
Chapter
4
do
not
include
all
of
the
components
that
would
be
required
to
replace
a
conventional
treatment
train.
A
schematic
of
the
MF/
UF
treatment
process
is
shown
in
Exhibit
3.8
As
discussed
in
section
2.2.5,
flux
is
a
critical
design
parameter
for
membrane
applications
and
is
often
used
in
membrane
procurements
as
a
specification.
However,
the
configuration
of
one
membrane
is
often
very
dissimilar
to
that
of
another.
Membrane
fiber
diameter,
pore
size,
flow
configuration
(
i.
e.,
cross­
flow
vs.
dead­
end,
pressure
vessels
vs.
submersible
membranes),
and
other
membrane­
specific
factors
can
impact
flux
and
other
design
and
operating
parameters.
As
a
result,
membrane
feed
water
quality
is
used
as
the
basis
of
design
for
the
membrane
portion
of
the
costs
presented.

Exhibit
3.8:
Microfiltration
and
Ultrafiltration
Cost
estimates
are
based
upon
a
design
feed
water
temperature
of
10
°
C.
As
previously
discussed,
temperature
can
have
a
significant
impact
on
membrane
system
design.
As
the
water
temperature
decreases,
water
viscosity
increases.
This,
coupled
with
temperature
effects
on
the
membranes
themselves,
can
result
in
the
need
for
increased
membrane
area
and/
or
increased
operating
Technologies
and
Costs
for
Control
of
Microbial
Contaminants
and
Disinfection
Byproducts
June
2003
3­
10
pressures
to
maintain
the
desired
level
of
production.
It
is
important
to
note
that
this
effect
can
vary
from
membrane
to
membrane,
and
many
manufacturers
have
developed
membrane­
specific
correction
factors.

Membrane
system
costs
were
approximated
using
estimates
provided
by
four
manufacturers
(
all
pressure
vessel
systems).
The
only
criteria
given
to
the
manufacturers
was
the
feed
water
temperature
of
10
°
C.
Since
the
design
assumes
a
post­
filtration
retrofit,
the
effect
of
solids
loading
on
the
membrane
is
considered
minimal
and
was
not
specified
for
manufacturer
estimates.
Each
manufacturer
then
used
its
own
flux
specifications
and
temperature
correction
factor
to
provide
cost
estimates
for
design
flows
ranging
from
0.01
to
430
mgd.
Estimates
for
design
flows
of
0.007
and
520
mgd
were
extrapolated
from
these
estimates.

The
membrane
costs
from
the
manufacturers
include
skid­
mounted
membrane
modules
with
associated
piping,
feed
pumps,
backwash
and
recirculation
pumps
(
where
appropriate),
chemical
cleaning
feed
tanks
and
pumps,
and
instrumentation
and
control
for
proper
operation.
Additional
instrumentation
and
control
and
pipes
and
valves
were
included
in
process
costs
for
interconnection
with
existing
plant
control
systems
and
processes.
Interstage
pumping
was
also
added
based
on
the
assumption
that
the
existing
plant
may
not
have
sufficient
hydraulic
head
to
accommodate
the
membrane
process.
O&
M
costs
include
replacement
membranes
(
membrane
life
is
5
years),
process
power,
chemicals
for
cleaning,
and
labor.

For
the
purposes
of
design,
it
was
assumed
backwash
and
reject
water
could
be
discharged
to
a
sanitary
sewer
for
treatment
at
a
publicly
owned
treatment
works
(
POTW).
This
assumes
the
sanitary
sewer
has
sufficient
capacity
to
accommodate
the
increase
in
flow,
and
the
POTW
is
able
to
handle
the
increase
in
daily
flow.
However,
in
many
cases,
the
reject
and
backwash
water
can
be
recycled
to
the
head
of
the
treatment
plant.
In
some
instances,
recycle
may
be
a
lower
cost
option
than
discharge
to
a
POTW.
In
other
cases,
recycle
may
require
additional
pumping
and
site
piping,
modification
or
addition
of
chemical
feed
systems,
installation
of
equalization
basins,
or
expansion
of
other
process
components.
Therefore,
the
costs
associated
with
POTW
discharge
represent
a
conservative
estimate
in
some
cases
(
i.
e.,
where
recycle
requires
few
process
improvements)
and
may
underestimate
costs
in
others
(
i.
e.,
where
extensive
improvements
are
necessary).
However,
for
the
purposes
of
approximating
treatment
costs,
POTW
discharge
represents
an
approximate
average
cost
per
utility.

3.3.6
Bag
and
Cartridge
Filtration
Bag
and
cartridge
filters
may
be
an
attractive,
low
cost
option
for
small
systems
to
improve
microbial
removal.
Filter
bags
and
cartridges
are
available
in
a
number
of
different
materials
and
a
wide
range
of
pore
sizes.
The
removal
efficiency
can
be
affected
by
the
filter
material,
pore
size
distribution,
and
filter
durability.
Filter
durability
affects
how
a
filter
stands
up
to
routine
cleaning
and
affects
replacement
frequency.
Technologies
and
Costs
for
Control
of
Microbial
Contaminants
and
Disinfection
Byproducts
June
2003
3­
11
Description
of
Process:
Addition
of
bag
filters
OR
cartridge
filters
following
granular
media
filtration.
Rapid
Mix
Flocculation/
Sedimentation
Filtration
Caustic
Storage
Bag
or
Cartridge
Filters
Coagulant
Interstage
Pumping
It
is
assumed,
for
the
purposes
of
this
document,
that
bag
or
cartridge
filters
are
installed
downstream
of
existing
granular
media
filters.
Exhibit
3.9
presents
a
schematic
of
bag
and
cartridge
filtration.

Exhibit
3.9:
Bag
and
Cartridge
Filtration
Costs
for
different
bag
and
cartridge
filter
construction
materials
were
used
to
develop
a
range
of
costs.
The
frequency
of
replacement
depends
upon
the
durability
of
construction
and
water
quality
and
can
vary
from
a
few
weeks
to
as
long
as
a
year.
This
can
have
a
significant
impact
on
O&
M
costs.
Filter
housings
are
available
in
carbon
steel
for
approximately
half
the
cost
of
a
stainless
steel
unit.
However,
for
drinking
water
application,
stainless
steel
is
more
likely
to
be
the
material
of
choice.
As
a
result,
only
stainless
steel
housing
was
considered
in
development
of
costs.

3.3.7
Bank
Filtration
Bank
filtration
may
be
advantageous
for
systems
that
currently
have
surface
intake
from
a
stream
which
is
underlain
by
a
granular
media.
Such
a
system
would
essentially
drill
a
well
below
the
water
table
created
by
the
surface
water
source.
The
well
would
replace
the
existing
surface
water
intake.
Particles
and
other
contaminants
would
be
trapped
in
the
pores
of
the
river
bed
material
or
adsorb
onto
the
river
bed
material.
The
river
bed
material
thus
acts
as
a
pre­
filtration
step
for
the
treatment
process.

3.3.8
Second
Stage
Filtration
Second
stage
filtration
may
be
a
desirable
option
for
systems
with
frequent
fluctuations
in
hydraulics
and
turbidity.
Second
stage
filtration,
like
single
stage
filtration,
operates
by
depth
removal.
Technologies
and
Costs
for
Control
of
Microbial
Contaminants
and
Disinfection
Byproducts
June
2003
3­
12
Depth
filtration
is
when
the
solids
are
removed
within
the
granular
media.
The
surface
area
of
the
media
provides
attachment
sites
for
the
particles
suspended
in
the
influent
water.

To
meet
EPA's
proposed
0.5
log
credit
for
Cryptosporidium
removal,
second
stage
filtration
must
have
the
following
characteristics:

°
First
stage
of
filtration
must
be
preceded
by
a
coagulation
step.

°
Both
filtration
stages
must
treat
100
percent
of
plant
flow.

3.3.9
Pre­
Sedimentation
Pre­
sedimentation
basins
will
be
useful
for
systems
with
high
influent
turbidities
and
high
particle
counts.
EPA
is
proposing
to
give
pre­
sedimentation
basins
with
coagulant
addition
0.5
log
credit
if
the
following
criteria
are
met:

°
All
flow
must
pass
through
basin.

°
Continuous
flow
through
basin
and
coagulant
addition
near
the
influent
of
the
presedimentation
basin
while
plant
is
in
operation.

°
Maximum
day
settling
surface
loading
rate
of
1.6
gpm/
ft2.

°
Annual
mean
influent
turbidity
>
10
NTU
or
maximum
daily
influent
turbidity
>
100
NTU.

Systems
with
existing
pre­
sedimentation
basins
may
monitor
after
the
pre­
sedimentation
basin
and
prior
to
the
main
treatment
plant
for
the
purpose
of
determining
LT2ESWTR
bin
assignment.
Costs
in
Chapter
4
were
determined
assuming
that
the
basin
met
all
the
above
specifications.

3.3.10
Watershed
Control
Each
PWS's
watershed
control
program
plan
is
expected
to
be
site­
specific
and
will
depend
on
the
hydrology
and
land
use
in
the
watershed,
the
location
and
type
of
Cryptosporidium
sources
in
the
watershed,
the
population
served,
size
of
the
watershed,
funding,
and
other
issues.
Watershed
programs
may
include
the
following:

°
Monitoring
for
Cryptosporidium
or
indicator
organisms
throughout
the
watershed
°
Fencing
or
otherwise
restricting
access
to
the
source
water
Technologies
and
Costs
for
Control
of
Microbial
Contaminants
and
Disinfection
Byproducts
June
2003
3­
13
°
Land
acquisition
°
Managing
land
owned
by
the
PWS
°
Working
with
sewer
or
stormwater
utilities
to
develop
plans
to
upgrade
treatment
or
otherwise
reduce
pollutant
loads
°
Working
with
municipal
governments
to
regulate
land
use
and
development,

°
Conducting
outreach
to
other
stakeholders
To
receive
credit
for
removal
of
Cryptosporidium,
a
watershed
control
program
must
have
the
following
elements:

°
It
must
be
reviewed
and
approved
by
the
primacy
agency.

°
It
must
include
an
analysis
of
the
system's
source
water
vulnerability
to
the
different
sources
of
Cryptosporidium
identified
in
the
watershed.
The
vulnerability
assessment
must
include
a
characterization
of
the
watershed
hydrology
and
identification
of
an
"
area
of
influence
on
source
water
quality"
(
i.
e.,
the
area
to
be
considered
in
future
watershed
surveys).
The
assessment
must
also
address
sources
of
Cryptosporidium,
seasonal
variability,
and
the
relative
impact
of
the
sources
of
Cryptosporidium
on
the
system's
source
water
quality.
It
is
likely
that
water
systems
will
obtain
much
of
the
information
to
be
provided
in
the
vulnerability
assessment
from
the
source
water
assessment
performed
as
part
of
the
State
source
water
assessment
program.

°
It
must
present
an
analysis
of
sustainable
interventions
and
an
evaluation
of
their
relative
effectiveness
in
reducing
Cryptosporidium
in
source
water.
Interventions
may
include
anything
from
outreach
to
point
source
elimination.

°
It
must
address
goals
and
define
and
prioritize
specific
actions
to
reduce
source
water
Cryptosporidium
levels.
The
plan
must
1)
explain
how
actions
are
expected
to
contribute
to
specified
goals,
2)
identify
partners
and
their
roles,
resource
requirements
and
commitments,
and
3)
include
a
schedule
for
plan
implementation.

°
It
must
include
submission
of
an
annual
report
performance
of
a
watershed
survey,
and
submission
of
a
request
for
review
and
reapproval.

A
watershed
control
program
could
include
interventions
such
as
1)
the
elimination,
reduction,
or
treatment
of
discharges
of
contaminated
wastewater
or
storm
water,
2)
treatment
of
Cryptosporidium
contamination
at
the
site
of
generation
or
storage,
and
3)
prevention
of
Cryptosporidium
migration
from
the
source
(
e.
g.,
farms
or
wastewater
treatment
plants).
The
Technologies
and
Costs
for
Control
of
Microbial
Contaminants
and
Disinfection
Byproducts
June
2003
3­
14
feasibility
and
sustainability
of
various
interventions
may
depend
on
the
cooperation
of
other
stakeholders
in
the
watershed.
Stakeholders
that
have
some
level
of
control
over
activities
that
could
contribute
to
Cryptosporidium
contamination
include
municipal
government,
private
operators
of
wastewater
treatment
plants,
livestock
farmers,
and
other
government
and
commercial
organizations.

The
LT2ESWTR
does
not
specifically
mandate
any
interventions
that
must
be
included
in
a
watershed
control
program
plan.
The
only
required
elements
are
those
submitted
with
an
application
for
approval
of
the
watershed
control
program
plan.
These
are
the
delineation
of
an
"
area
of
influence
on
water
quality"
and
a
vulnerability
assessment.
Watershed
delineation
and
susceptibility
analyses
are
already
required
under
the
Source
Water
Assessment
Program;
data
gathered
under
this
program
can,
in
many
cases,
be
used
in
preparing
information
required
for
the
application.

3.3.11
Combined
Filter
Performance
Combined
filter
performance
is
not
a
single
technology
but
many
different
activities
that
can
improve
existing
filtration
processes
to
enhance
performance.
Plants,
which
can
operate
their
filters
in
such
a
way
to
produce
0.15
NTU
or
lower
turbidity
water
95
percent
of
the
time,
will
receive
a
0.5
log
Cryptosporidium
reduction
credit
under
the
LT2ESWTR.

The
Regulatory
Impact
Analyses
(
RIAs)
for
the
IESWTR
and
LT1SWTR
identified
35
actions
that
facilities
could
take
to
lower
the
finished
water
turbidity
from
the
SWTR
standard
of
0.5
NTU
to
the
IESWTR
standard
of
0.3
NTU.
These
tasks
were
examined
and
professional
judgement
was
applied
to
determine
which
of
these
actions
would
be
helpful
in
further
lowering
turbidity
from
0.3
to
0.15
NTU.

In
determining
processes
that
could
further
reduce
filtered
water
turbidity,
systems
that
would
select
this
Cryptosporidium
removal
option
were
assumed
to
be
conventional
filtration
or
softening
plants
which
were
already
operating
well
within
the
0.3
NTU
standard
currently.
These
plants
would
likely
have
to
make
only
minor
modifications
to
the
treatment
process
to
meet
the
0.15
NTU
standard.
These
plants
were
also
assumed
to
be
operating
under
0.15
NTU
less
than
95
percent
of
the
time
or
to
be
capable
of
achieving
0.15
NTU.

Based
on
these
assumptions,
the
filter
improvements
listed
in
the
IESWTR
were
reviewed
for
applicability
to
this
treatment
option.
The
following
were
considered
as
possible
actions
that
systems
may
take
to
implement
this
option:

°
Installing
backwash
polymer
feed
capability
°
Installing
coagulant
feed
points
°
Adding
filter
media
Technologies
and
Costs
for
Control
of
Microbial
Contaminants
and
Disinfection
Byproducts
June
2003
3­
15
°
Adding
filter
to
waste
capabilities
°
Replacing
the
filter
rate­
of­
flow
controller
°
Increasing
plant
staffing
°
Increasing
staff
qualifications
°
Purchasing
or
replacing
bench­
top
turbidimeters
°
Purchasing
or
replacing
jar
test
apparatus
°
Purchasing
or
replacing
a
particle
counter
or
streaming
potential
meter
°
Staff
training
It
is
not
assumed
that
each
system
using
this
technology
will
use
all
eleven
tasks.
Instead,
it
is
assumed
that
each
system
would
have
to
use
at
least
one
of
these
tasks
and,
most
likely,
two
or
more
to
meet
the
turbidity
targets
(
successfully).
To
develop
costs
for
this
technology,
the
percentage
of
the
plants
choosing
each
action
was
determined.
The
percentage
of
systems
choosing
a
particular
task
was
then
multiplied
by
the
unit
cost
for
that
task
to
arrive
at
an
average
unit
cost
for
all
plants.
Further
details
of
the
percentages
and
costs
are
given
in
Chapter
4
of
this
document.

The
assumptions
for
each
filter
improvement
action
is
discussed
below.

Installing
Backwash
Water
Polymer/
Coagulant
Feed
Capability
Adding
coagulant
to
backwash
water
aids
in
filter
ripening
and
helps
to
reduce
post
backwash
turbidity
spikes.
Systems
choosing
backwash
polymer
to
lower
turbidity
were
assumed
to
not
have
this
capability
currently.
Costs
were
for
a
dry
polymer
feed
system
that
can
be
loaded
with
a
seven­
day
polymer
supply.

Installing
Additional
Coagulant
Feed
Points
Installing
additional
coagulant
feed
points
can
help
to
improve
coagulation
of
particles
and
their
removal
by
settling.
Capital
costs
were
based
on
feeding
an
additional
5
parts
per
million
(
ppm)
dose
of
primary
coagulant.
The
primary
coagulant
is
assumed
to
be
ferric
chloride,
ferric
sulfate,
or
alum.
Thirty
days
of
bulk
storage
are
assumed
for
ferric
chloride
or
ferric
sulfate
(
equivalent
to
approximately
fifteen
days
of
storage
for
alum).

Adding
Filter
Media
Technologies
and
Costs
for
Control
of
Microbial
Contaminants
and
Disinfection
Byproducts
June
2003
3­
16
Often
during
routine
operation
of
filters,
media
is
lost
either
through
attrition
and
passage
out
the
underdrains
or
through
the
backwash.
If
too
much
media
is
lost,
filter
performance
will
suffer.
Therefore,
adding
additional
media
can
often
improve
turbidity
in
the
effluent.

Adding
Filter
to
Waste
Capabilities
Filter
turbidity
often
spikes
immediately
after
backwashing.
Installing
filter
to
waste
capabilities
allows
water
to
be
wasted
after
a
backwash
instead
of
sending
the
high
turbidity
water
to
the
CFE.
Costs
included
piping,
valves,
and
fittings.

Installing
or
Replacing
Filter
Rate­
of­
Flow
Controllers
Flow
surges
can
cause
spikes
in
filter
turbidity.
Installing
a
rate­
of­
flow
controller
or
replacing
a
faulty
one
can
improve
performance.
Costs
were
for
replacing
a
unit
and
were
based
on
assumed
24­
hour
operation.

Increasing
Plant
Staffing
Systems
which
only
have
part
time
staff
or
are
understaffed
may
have
trouble
controlling
filter
conditions
closely
enough
to
meet
the
0.15
NTU
turbidity
target.
Hiring
additional
staff
or
extending
current
staff's
hours
may
help
systems
to
more
finely
control
filter
operations.

Increasing
Staff
Qualifications
Better
trained
staff
may
be
able
to
recognize
conditions
which
lead
to
filter
turbidity
breakthrough
and
to
prevent
it.
Costs
for
this
option
were
based
on
the
cost
of
sending
an
operator
to
a
training
class.
Costs
include
class
registration
fees
to
attend
an
operator
certification
class.

Purchasing
or
Replacing
Bench­
Top
Turbidimeters
Typically,
every
plant
has
at
least
one
bench­
top
or
on­
line
turbidimeter.
However,
some
of
these
units
may
be
obsolete
to
meet
the
monitoring
requirements
of
the
LT2ESWTR
for
combined
and
individual
filter
effluents.
Bench­
top
turbidimeters
do
not
appear
to
be
suited
to
fulfill
these
monitoring
tasks.
Therefore
the
purchase
of
up­
to­
date
on­
line
turbidimeters
with
electronic
data
acquisition
interface
was
costed.

Purchasing
or
Replacing
Jar
Testing
Apparatus
A
jar
testing
apparatus
is
necessary
for
optimizing
coagulant
and
polymer
dosing.
Old
units
will
need
to
be
replaced,
and
new
units
purchased
if
a
facility
does
not
have
one.
Systems
serving
greater
than
100,000
people
were
assumed
to
buy
two
units,
and
those
serving
more
than
1,000,000
people
were
assumed
to
purchase
three
units.
Technologies
and
Costs
for
Control
of
Microbial
Contaminants
and
Disinfection
Byproducts
June
2003
3­
17
Rapid
Mix
Flocculation/
Sedimentation
Filtration
Caustic
Caustic
Storage
Coagulant
Coagulant
GAC
Filter
Description
of
Process:
Install
GAC
filter
following
granular
media
filter.
Interstage
Pumping
Purchasing
or
Replacing
a
Particle
Counter
Instruments
such
as
particle
counters,
zetameters,
and
streaming
current
monitors
can
be
used
to
optimize
filter
processes.
The
cost
for
this
option
assumes
the
purchase
of
one
of
these
instruments.
The
cost
of
a
particle
monitor
was
used
as
a
surrogate
for
any
one
of
these
three
instruments.

Staff
Training
Better
trained
staff
will
be
better
able
to
spot
and
fix
problems
in
filter
performance.
The
costs
for
this
option
were
based
on
hiring
a
consultant
to
provide
on­
the­
job
training
for
10
to
140
hours.

3.4
DBP
Precursor
Removal
Technologies
A
strategy
for
reducing
DBP
formation
is
removal
of
DBP
precursors
(
e.
g.,
natural
organic
matter).
The
technologies
discussed
in
this
section
may
not
be
applicable
for
all
systems.
Each
technology
section
presents
the
approach
and
assumptions
used
to
develop
the
costs
presented
in
Chapter
4.

3.4.1
Granular
Activated
Carbon
Adsorption
GAC
filters
reduce
DBP
formation
by
removing
organic
carbon.
For
the
purposes
of
this
document,
installation
was
assumed
after
the
existing
filters.
A
schematic
of
the
GAC
process
is
shown
in
Exhibit
3.10.

Exhibit
3.10:
GAC
Filtration
The
application
of
GAC
adsorption
involves
the
following
process
design
considerations:

°
Empty
bed
contact
time,
volume
of
empty
contactor
divided
by
flow
rate
Technologies
and
Costs
for
Control
of
Microbial
Contaminants
and
Disinfection
Byproducts
June
2003
3­
18
°
Reactivation
interval
or
frequency,
which
affects
the
GAC
usage
rate
(
pounds
of
GAC
used
per
gallon
of
water
treated)
°
Pre­
treatment
°
Contactor
configuration
(
e.
g.,
downflow
versus
upflow,
pressure
versus
gravity,
singlestage
versus
multi­
stage
or
parallel,
filter
adsorber
versus
post­
filter
GAC
contactor)

°
Method
of
GAC
reactivation
(
e.
g.,
on­
site
versus
off­
site)

°
Interstage
pumping
°
Performance
monitoring
(
for
TOC)

EBCTs
of
ten
and
twenty
minutes
were
chosen
for
the
cost
evaluation
based
upon
an
analysis
of
EBCTs
and
NOM
removal.
This
analysis
indicated
that
EBCTs
lower
than
10
minutes
do
not
remove
sufficient
NOM
to
warrant
installation
as
a
control
for
DBP
precursors.
Similarly,
EBCTs
in
excess
of
20
minutes
do
not
provide
significant
improvements
in
NOM
removal.
Accordingly,
10
minutes
and
20
minutes
were
selected
to
represent
the
upper
and
lower
bounds
of
appropriate
EBCTs
for
NOM
removal.

Reactivation/
replacement
frequencies
vary
based
on
water
quality
and
the
number
of
contactors
in
parallel.
For
the
purposes
of
this
document
frequencies
of
90,
240,
and
360
days
were
evaluated.
Ninety
days
was
selected
as
a
minimum
value
based
upon
best
professional
judgement
that
reactivating
at
intervals
lower
than
90
days
is
impractical
from
an
operational
standpoint.
Three
hundred
and
sixty
days
was
selected
as
the
maximum
reactivation
frequency
since
the
cost
of
GAC
technology
increases
insignificantly
for
reactivation
frequencies
of
greater
than
1
year.
High
operating
costs
were
captured
by
considering
90­
day
regeneration
frequency
for
the
GAC
facility
with
EBCT
of
20
minutes.
Low
operating
costs
were
captured
by
considering
360­
day
regeneration
frequency
for
the
GAC
facility
with
EBCT
of
10
minutes.
An
intermediate
operating
cost
was
also
captured
by
considering
240­
day
regeneration
frequency
for
the
GAC
facility
with
EBCT
of
20
minutes.

Based
upon
best
professional
judgement,
it
was
decided
that
small
systems
are
unlikely
to
regenerate
on­
site,
since
it
requires
more
substantial
capital
investment
and
operator
attention.
As
a
result,
small
systems
(
less
than
1
mgd)
were
assumed
to
operate
on
a
replacement
basis
(
i.
e.,
when
the
carbon
is
spent,
it
is
discarded
and
replaced
with
new
carbon).
Large
systems
(
greater
than
1
mgd)
were
assumed
to
regenerate
on­
site
using
multiple
hearth
furnaces.

Very
small
system
GAC
installations
(<
0.1
mgd)
include:
pressure
GAC
contactors,
virgin
GAC,
pressure
booster
pumps,
pipes
and
valves,
and
instrumentation
and
controls.
O&
M
is
a
function
of
regeneration
frequency.
Technologies
and
Costs
for
Control
of
Microbial
Contaminants
and
Disinfection
Byproducts
June
2003
3­
19
Description
of
Process:
Addition
of
nanofiltration
following
granular
media
filtration,
OR
replacement
of
granular
media
filters
with
nanofiltration
Rapid
Mix
Flocculation/
Sedimentation
Filtration
Caustic
Caustic
Storage
Nanofiltration
Coagulant
Coagulant
Interstage
Pumping
Small
system
GAC
installations
(>
0.1
mgd
and
<
1
mgd)
include:
pressure
vessels
designed
for
working
pressure
of
50
psi;
factory
assembled
units
mounted
on
steel
skid
12
feet
high
and
varying
diameter
depending
on
the
EBCT;
access
for
filling
and
removing
carbon;
pressure
booster
pump,
valves,
piping
and
pressure
gauges,
initial
charge
of
activated
carbon,
supply
and
backwash
pump,
and
electrical
control
panels.

Large
system
GAC
installations
(>
10
mgd)
include:
concrete
gravity
contactors
8.3
feet
high;
loading
rate
5
gpm/
ft2;
troughs
and
pipes
for
carbon
removal
as
a
slurry;
other
pipe
gallery;
pressure
booster
pump;
flow
measurement
and
instrumentation;
master
operations
control
panel;
building;
initial
virgin
carbon;
single
multiple­
hearth
furnace
for
carbon
regeneration­
loading
rate
of
50
pounds
per
square
foot
per
day;
and
two
TOC
analyzers.

3.4.2
Nanofiltration
Nanofilters
remove
NOM,
thereby
reducing
DBP
formation.
NF
is
an
advanced
treatment
process
which
typically
requires
higher
levels
of
pre­
and
post­
treatment
than
traditional
water
treatment
processes.
For
this
cost
analysis,
nanofilters
were
assumed
to
be
located
downstream
of
existing
filters.
A
schematic
of
the
NF
technology
is
shown
in
Exhibit
3.12.

Exhibit
3.11:
Nanofiltration
Typically,
NF
requires
both
physical
and
chemical
pre­
treatment.
Pre­
treatment
is
usually
required
for
NF
treatment
of
all
surface
waters
and
some
ground
waters.
Physical
pre­
treatment
often
includes
a
component
to
remove
particles,
typically
multi­
media
filtration,
microfiltration,
or
cartridge
filtration.
Chemical
pre­
treatment
often
includes
acid
or
anti­
scalant
addition
to
reduce
the
fouling
potential
of
the
feed
water.
Particle
removal
and
softening
with
chemical
addition
are
also
used
as
pretreatments
Attention
should
be
paid
to
the
compatibility
of
coagulant
and
the
membrane
for
such
situations.
Technologies
and
Costs
for
Control
of
Microbial
Contaminants
and
Disinfection
Byproducts
June
2003
3­
20
Post­
treatment
may
also
be
required,
depending
on
the
characteristics
of
the
product
water.
NF
product
waters
usually
have
low
pH
and
total
dissolved
solids
levels.
This
creates
the
potential
for
an
unstable
and
corrosive
finished
water.
Chemical
post­
treatment
may
be
required
to
create
a
more
stable
and
non­
corrosive
water.
Commonly
used
post­
treatments
include
addition
of
caustic
(
to
raise
the
pH),
soda
ash
(
to
raise
pH
and
alkalinity),
and
poly/
ortho
phosphates
for
stabilizing
the
water.
Blending
a
portion
of
raw
water
with
finished
water
can
also
be
used
to
stabilize
the
finished
water.

The
design
criteria
in
this
document
assume
that
the
NF
system
is
an
"
add­
on"
process
to
an
existing
treatment
plant
which
is
generating
a
water
that
can
be
fed
directly
to
the
NF
process
without
further
pre­
treatment.
It
is
assumed
that
100
percent
of
the
design
flow
is
passing
through
the
NF
membranes.
Recoveries
of
85
percent
and
operating
pressures
of
90­
110
psi
were
assumed.
Costs
were
developed
assuming
a
design
feed
water
temperature
of
10
degrees
Celsius.
Like
MF,
the
cost
of
a
NF
system
can
vary
significantly
with
temperature
because
the
membrane
productivity,
or
flux
(
gallons/
ft2­
day),
is
strongly
dependent
on
feed
water
temperature.
Empirical
relations
are
available
to
estimate
the
flux
at
a
design
temperature
using
the
flux
at
a
reference
temperature
(
i.
e.,
10
degrees
Celsius).
These
relations
are
available
both
in
published
literature
and
with
membrane
manufacturers.

NF
system
cost
quotations
were
obtained
from
manufacturers
for
all
NF
equipment
items,
including
membrane
elements,
online
instruments,
booster
pumps,
clean­
in­
place
systems
and
acid/
antiscalant
addition
systems.
Unlike
other
treatment
processes,
membrane
systems
are
typically
supplied
by
the
equipment
vendor
as
package,
skid­
mounted
units;
therefore,
smaller
multipliers
are
assumed.
Capital
cost
multipliers
of
1.67
and
2.0
were
used
respectively
for
small
and
large
systems
to
estimate
total
capital
cost.
It
was
assumed
that
a
unit
NF
skid
can
produce
up
to
2
mgd
of
product
water.
NF
systems
smaller
than
2
mgd
were
assumed
to
have
fewer
membrane
modules
and
membranes.

The
O&
M
costs
include
chemical
usage,
membrane
replacement
(
assumed
membrane
life
of
five
years),
process/
building
power,
additional
labor
hours,
and
process
monitoring.
Efforts
were
made
to
capture
the
drop
in
prices
of
the
membranes,
modules,
and
associated
equipment
over
the
past
few
years
due
to
increasing
use
of
the
NF
systems.
Where
necessary,
the
costs
for
retrofitting
and
operating
an
NF
plant
were
verified
with
data
from
various
surveys,
including
Florida's
softening
plants
(
Bergman
1996)
and
the
Bureau
of
Reclamations
(
BOR
1997)
surveys.
The
cost
curves
presented
in
Chapter
4
were
verified
with
real­
plant
data
for
different
flow
levels.

NF
design
criteria
developed
here
include
handling
of
the
brine
stream
generated
by
the
NF
process.
This
handling
assumes
direct
discharge
of
the
brine
to
a
receiving
body,
ocean
outfall,
sanitary
sewer,
storm
drain,
or
a
salinity
interceptor.
The
costs
presented
in
Chapter
4
pertain
only
to
plants
located
in
areas
where
brine
can
readily
be
discharged
to
either
a
receiving
water
body,
a
sewer/
storm
drain,
or
a
salinity
interceptor.
Plants
located
in
areas
where
this
is
not
an
option
will
have
significantly
higher
waste
stream
treatment
and
handling
costs.
